M. Gunthorpe, S. Cai, F. Chehab;
UCSF Medical Center, San Francisco, CA, United States.

Few clinical laboratories test for b -thalassemia, an anemia which results from decreased or absent production of the b -globin polypeptide chain of hemoglobin. The molecular diagnosis of b -thalassemia is complex due to the vast number and types of characterized mutations. Protocols that rely on PCR amplification of the b -globin gene followed by hybridization to allele-specific probes, such as in the reverse-dot blot hybridization technique, require production of numerous primers and are limited to screening for the most common 30-40 mutations. In order to screen in clinical samples for all the mutations known to cause b -thalassemia, we devised a DNA sequencing strategy based on sequencing the entire b -globin gene (about 2 kb) with dye-terminator chemistry. This procedure involves the amplification of the gene first by generating two PCR products, and then determining their DNA sequence with a total of 8 primers. Characterized mutations and polymorphisms were rapidly compiled by an assembly software and tabulated for inclusion in a clinical report. The assay was validated by blindly sequencing patients with previously characterized mutations. In addition, this technique will uncover b -globin gene DNA polymorphisms, which are valuable for ruling out gene deletions when homozygosity for a mutation is found in a patient. Mutation analysis by DNA sequencing greatly simplifies the diagnosis of b -thalassemia disorders caused by subtle changes in the b -globin gene structure. However, this approach is refractory to the detection of the few rare deletions which encompass a large region of the b -globin gene cluster. This DNA sequencing strategy is currently being implemented in a clinical setting and is used as an adjunct to the determination of b -thalassemia mutations that are not identified by standard allele-specific probe assays.

D. Bond, A. Swei, K. Lee, R. Nutter, C. Swartz, A. Yang, K. Louie, Y. Lei, J. Lee, K. Perry, J. Teare, E. Sun, J. Vuong, C. Heiner, J. Cassel;
Applied Biosystems, Foster City, CA, United States.

Fluorescent dideoxy sequencing has advanced greatly since the early days of automated DNA sequencing about 15 years ago. Cycle sequencing using thermally stable DNA polymerase enzymes provided a robust method of sequencing plasmids over a much wider and lower range of concentrations than was previously possible. Fluorescent dideoxy terminators reduced the number of steps, equipment and supplies required to perform sequencing reactions, and BigDye terminators provided several fold additional sensitivity. These advances greatly facilitated the rapid sequencing of the human genome. Researchers today are challenged to finish the difficult-to-sequence areas of the human genome and other organisms, and to obtain highly accurate sequences of individuals in studying polymorphisms. We have recently released two new sequencing kits that will help to address these challenges. Here we present BigDye^® Terminators v3.1 and 1.1 Cycle Sequencing kits. We will illustrate improvements with both our work as well as the results of our conformation test sites.

N. Panayotova, E. C. Almira, W. G. Farmerie;
University of Florida Biotechnology Program, Gainesville, FL, United States.

High-throughput capillary array DNA sequencing instruments have been in service since their introduction in the late 1990s, yet few improvements have been made to standard operating conditions and protocols. Many small facilities, especially those with a single sequencer, continue using the same operating conditions they used initially. In slab-gel sequencing, most common consumables, such as gel matrices, running buffers, and glass plates, may come from various sources. However, capillary sequencer consumables are mainly proprietary, making it difficult for core facilities to find potential alternatives. Still, in an effort to improve services while keeping the costs to a minimum, many budget-conscious core facilities hope to discover compatible but cost-effective consumables. We recently tried and tested three alternative running buffers (A, B and D) in our MegaBACE 1000 using standard plasmids and production DNA templates generated via the Templiphi rolling circle amplification process. Compared to the current running buffer (C), buffer D caused a moderate but significant increase in the number of total phred Q20 bases per plate. Data analysis revealed an increase in sequence readlength rather than passing rate. Results with buffers A and B were either similar or inferior to those of C. A much more consistently improved sequence data is now being generated after switching to buffer D.

D. A. Bintzler, S. I. Sherwani, Y. Song, M. Jordan, J. Ackerman;
DNA Core Facility, University of Cincinnati, Cincinnati, OH, United States.

In order to use a fine quality of DNA for various molecular biology techniques like sequencing, cDNA synthesis and cloning, RNA transcription, transfection, nucleic acid labeling (random prime labeling), etc., quantitation of DNA is extremely important to have the defined template concentration. Failure to produce results from the above techniques can sometimes be attributed to an incorrect estimate of the DNA template used.
A resource facility that provides automated DNA sequencing as a service can apply several techniques to measure the quality and the quantity of DNA templates submitted for sequencing. Scanning a template using a spectrophotometer is one technique used in various resource facilities to determine contaminants such as phenol and salt and can be applied quickly. However, this technique is limited to locating, but not quantifying, the amount of contaminant in the template. We have also found that it is quite difficult to quantify smaller PCR fragments due to the sensitivity of the spectrophotometer. Several commercial suppliers also provide quick strips that can quantify PCR products with a certain degree of accuracy. However, the process can be labor-intensive and expensive. Therefore, we decided to further develop two known techniques that can be used in a resource facility that can be less labor-intensive and can improve sensitivity. Further, we have found pre-testing a sample to be a small source of extra income, while it saves clients the cost of sequencing contaminated templates by reducing sequence failures.
Further research is warranted so that resource facilities could introduce such highly sensitive yet simple services at a minimum cost to the customer.

A. Greco, I. Y. Hlede, R. T. Pon;
University of Calgary, Calgary, AB, Canada.

University core service laboratories constantly attempt to provide the best possible results for the wide variety of DNA templates they receive. DNA templates that give poor or limited sequencing results are always a major problem for our clients and us. Our facility performed a series of reactions with the recently released new formulation of BigDye Terminator sequencing kit, version 3.1 from Applied Biosystems which is expected to produce a marked improvement in sequencing difficult templates. We have examined a variety of GC rich, AT rich, and other templates, which sequenced poorly or not at all with the BigDye v.3.0 kit. Results will be presented comparing how the v.3.0 and v.3.1 sequencing kits perform on the ABI 377 DNA sequencer.

B. T. Coullahan¹, K. D. Eluwa², O. Zimerman²;
¹Genomic Analysis and Technology Core, University of Arizona, Tucson, AZ, United States, ²Edge BioSystems, Inc., Gaithersburg, MD, United States.

The Genomic Analysis and Technology Core at the University of Arizona has evaluated a new magnetic dye terminator removal technology on a Biomek FX workstation in conjunction with an ABI 3730XL DNA Analyzer. In this method, sequencing reaction products are adsorbed to magnetic particles, washed with alcohol, and eluted in water. The system is adaptable for single tube as well as 96 and 384 well plate configurations. In a 96 well plate configuration, processing times were typically less than 20 minutes. Sequencing reaction volumes of 5 to 20 μL were evaluated, with DNA template ranging from 25 to 500 ng, and BigDye™ premix volume ranging from 0.125 μL to 8 μL. Sequence quality is excellent throughout much of the range of template, volume and premix described, with long read lengths. This technology offers a convenient and inexpensive alternative to costly and/or unautomatable purification methods and provides substantially better reproducibility than ethanol precipitation.

T. Mamone;
Amersham Biosciences, Piscataway, NJ, United States.

The TempliPhi™ DNA sequencing template amplification kit relies on Phi29 DNA polymerase to drive rolling circle amplification (RCA) of circular input DNA. Several micrograms of template DNA are generated from picogram amounts of input. The method requires only two steps to generate sequence quality templates, potentially bypassing both the requirement for liquid culture growth and the conventional steps for preparing plasmid and phage DNA for use as templates in sequencing applications.
This method is of enormous use to sequencing facilities in several aspects. Time is saved by omitting traditional multistep plasmid preps. Significant cost savings are achieved by reduced plasticware. Space devoted to bacterial growth and prep can be put to better use. Most significantly, success rates can be near 100% for plasmids, regardless of the source, due to the normalization of template amount. Supporting data will be shown, in addition to difficult template sequencing improvements.
An extension of this method in development, called GenomiPhi™, is appropriate for amplifying large constructs, bacterial genomes, and human chromosomal DNA.

M. J. Reagin, D. M. Meyers;
Amersham Biosciences, Piscataway, NJ, United States.

The majority of failures in core lab sequencing facilities are due to either poor quality of customer plasmid DNA prep or the miscalculation of DNA concentration. For core labs that offer resequencing of failed reactions, not only is the cost passed onto the customer, but also valuable time is spent by lab personnel to reproduce data that might once again fail. TempliPhi DNA amplification is a process that employs rolling circle amplification (RCA) and Phi29 DNA polymerase to produce microgram amounts of circular template from picograms of DNA in as little as 6 hours. We will show that TempliPhi can produce enough template from an empty tube that once held plasmid DNA, to generate a DNA sequence of at least 500 bp. We will also show sequence produced on a capillary electrophoresis instrument that failed due to an insufficient amount of starting plasmid template. By adding small aliquots of the failed sequencing reactions to TempliPhi, improved sequence quality and longer readlengths upon resequencing are obtained. Carryover primer from the failed sequencing reaction does not interfere with RCA and the rapid amplification by the Phi29 DNA polymerase provides enough template for a successful sequence by the next day with minimal manipulation.

E. J. Thompson, A. Anderson, R. Pershad;
The University of Texas MD Anderson Cancer Center, Houston, TX, United States.

At The University of Texas M. D. Anderson Cancer center, mouse models are frequently used to help identify genes related to cancer. The ability to genotype mice in high throughput format has been mostly limited to manual PCR. This process is time consuming and laborious. Separate pcr assays must be assembled and multiple thermocyclers are required to accommodate the unique annealing temperatures of different primer sets. In addition a Southern analysis must be performed to identify copy number.
The DAC at M. D. Anderson has developed a method to rapidly genotype mice using Taqman hydrolysis probes in 5' exonuclease based real-time polymerase chain reaction assays. This method allows mice to be rapidly screened in 96 well format. The key advantages of this technology are that multiple assays can be designed to work under the same conditions and that assays can be multiplexed.
Here we show a method to differentiate wild type, heterozygote and null mice using Taqman Assays.

M. A. Robertson, R. Scholl, M. Klein, R. M. Cawthon;
University of Utah, Salt Lake City, UT, United States.

Single nucleotide polymorphisms (SNPs) are abundant in the human genome and appear at a frequency of about 1 in every 500 to 1000 bases. SNPs can be used as genetic markers for genome scans in pedigree-based linkage analysis of families, in population studies where the haplotype distributions and linkage disequilibria can be used to map genes by association methods, and in case control studies to identify functional SNPs associated with a particular phenotype. Many more SNPs than microsatellites are required for these kinds of large studies so the methodology used has to be as automated and as cost effective as possible. This poster will compare two different methods of SNP genotyping using the 7900HT instrument as a detection platform. One method uses allele specific-PCR and melting curve analysis with Sybr Green I dye to distinguish alleles. Another method makes use of the Taqman assay with fluorescent primers and probes from Applied Biosystems. Both methods will be compared with a "mini-sequencing method" using the SnapShot™ kit from Applied Biosystems. This poster will address the ease of use, the robustness of the technology, the quality of the results and the ability for automation for each of the methods.

A. Khvorova;
Dharmacon Research, Lafayette, CO, United States.

RNA-mediated interference (RNAi) is an increasingly important method for analyzing gene function. Chemically synthesized double-stranded, short interfering RNA (siRNA), can effectively bypass the mammalian interferon response and sequence-specifically silence gene expression. Some of the factors that greatly affect the success of RNAi experiments are the quality of synthetic RNA, rapid access to candidate siRNA duplexes and the parameters implemented in the design of the duplexes.
2’-ACE RNA technology allows for rapid, dependable and scalable synthesis of any sequence regardless of nucleotide content. In addition, the technology provides a stable, well-behaved intermediate that is readily deprotected in aqueous buffer. The speed and versatility of 2’-ACE chemistry coupled with laboratory automation facilitates high-throughput siRNA applications for genome wide projects.
By analysis of a large number of functional and nonfunctional siRNAs and coupling the information with a systematic high throughput screening approach we developed a method allowing identification of highly active siRNA’s. Empirical testing of the method demonstrated that a >50% reduction in gene expression occurred with 99.97% of the selected sequences. Furthermore, in 97% of our experiments, rationally designed gene silencing reagents delivered >80% reduction in gene expression.
The ability to knock-out gene function via siRNA-mediated methods represents an exciting and valuable tool for gene functional analysis that will accelerate critical investigations across a broad range of biomedical and biological research.

M. Swope Willis, T. Fox;
Vertex Pharmaceuticals, Cambridge, MA, United States.

Chemogenomics, a driving force of drug-discovery research at Vertex Pharmaceuticals, is a parallel process allowing the evaluation of a wide variety of targets within a single gene family. The protein requirements to drive this process are substantial in their breadth and number. We have optimized our protein production process by modifying or automating several steps from expression to purification. However, when the expression of soluble protein is not possible, refolding of the inclusion bodies may be required to obtain functionally active protein. Refolding proteins is often a difficult and highly unpredictable process. Published conditions resulting in the successful refolding of one protein often tend to be protein-specific and are not always applicable to other proteins. Using semi-automated dilution equipment, we have developed a system for evaluating the effects of varying pH, osmolytes, reducing and thiol shuffling reagents, and chaperones, on protein refolding. Applying this method to multiple proteins representing five gene families, we have successfully refolded members of each family. This rapid dilution method has enabled the systematic screening of an expanded number of refolding conditions. We have identified trends in the data correlating the key additives with various proteins being refolded.

M. Kuschel¹, P. Barthmaier², E. Chernokalskaya³, E. Benson³;
¹Agilent Technologies, Waldbronn, Germany, ²Agilent Technologies, Palo Alto, CA, United States, ³Millipore Corporation, Danvers, MA, United States.

The Agilent 2100 bioanalyzer was the first commercially available microfluidic device for the molecular biology lab. Developed by Agilent Technologies in collaboration with Caliper Technologies the bioanalyzer can be used for the automated analysis of proteins, nucleic acids and cells. As with all microfluidic devices smaller sample sizes and faster run times are achieved on the bioanalyzer, compared to SDS-PAGE. With its speed and low sample consumption the Agilent 2100 bioanalyzer is the ideal instrument to use in conjunction with the Millipore Zip Tip_Ni. To aid in the purification of His tagged proteins Millipore has developed Zip Tip_Ni. The affinity interaction between Ni⁺⁺ ions and histidine is well—known. Nickel ion chelated chromatography resins have proven to be effective in capturing recombinant proteins that have 4 — 8X histidines fused to the N or C terminus of a protein. The His-tag /Ni⁺⁺ is becoming increasingly important in proteomics for expressing/isolating low abundant proteins and generating sufficient quantities of proteins for interaction studies and structure analysis. We report on the use of a 10 μl pipette tip containing ca. a 0.5 μl of Ni⁺⁺ chelated resin immobilized at the distal end (ZipTip_Ni). This device was effective in isolating 50 — 100 ng of native or denatured His-tagged protein from ca. 10 to 50 μl of cleared cell lysate in about 5 minutes. The pipette tip format offers a convenient sized tool for the rapid isolation of proteins that can subsequently be characterized on the Agilent 2100 bioanalyzer all under 1 hour.

A. T. Yeung¹, C. A. Oleykowski¹, A. Godwin¹, S. Henikoff², E. Sokurenko³, E. Nicolas¹;
¹Fox Chase Cancer Center, Philadelphia, PA, United States, ²Fred Hutchison Cancer Research Center, Seattle, WA, United States, ³University of Washington, Seattle, WA, United States.

CEL I, a nuclease from celery, has high specificity for DNA mismatch base-substitutions, insertions and deletions. It has been used to develop a robust mutation detection assay that is easy to automate. We have demonstrated its utility in the screening of unknown mutations and polymorphisms of BRCA1, BRCA2, ARSC, MED1, and TDG genes, of hundreds of individuals. Multiple unknown mutations in one PCR fragment can be simultaneously identified. This fragment analysis assay has been adopted for high throughput screening of EMS induced mutations in several plant genomes (Tilling procedure, over 600 Arabidopsis mutants have been obtained) to facilitate plant genetics studies. Similar approach is ongoing for zebra fish and mice in user labs. In contrast to knockouts that are less informative, Tilling produces a panel of mutants for investigating the multiple domains and phenotypes of a protein. CEL I mismatch endonuclease is also used for the detection of mutations in large genomic regions of microorganisms (Giraff procedure). Current effort is aimed at applying this assay for the genomic scanning of mutations in mice, the detection of mutations in pathogens such as Bacillus anthracis, and early cancer detection. The enzyme can be obtained through Transgenomic Inc. by sending an email to AT_Yeung@fccc.edu.
http://web-apps.fccc.edu/fccc/yeung/index.html

A. Swei, C. Kosman, P. Baybayan, S. Bay, S. Chang, S. Dodgen, L. Esfahani, S. Fang, T. Hatch, J. Labrenz, J. Pone, R. Sefanov, P. Suri, D. Xu, K. Yung;
Applied Biosystems, Foster City, CA, United States.

The Applied Biosystems 3730 DNA Analyzer and the accompanying control and analysis software packages are specifically designed to increase throughput and efficiency while reducing the data entry and processing complexity of diverse DNA projects. To streamline sample input and extraction, a new Data Collection Software architecture is implemented which automatically tracks and stores plate records, run folders, and analysis parameters within a searchable database. Via an auto-analysis function the sample files are processed using one of several application-specific Applied Biosystems software packages: Sequencing Analysis Software v5.0 contains an improved base calling algorithm that provides per-base Quality Value assignment and sample trimming via these Quality Value scores, along with automatic length of read estimation and several output file formats offering compatibility with various downstream post-analysis programs; SeqScape™ v2.0 Software provides integrated base calling, trimming, sequence assembly and alignment for fast and accurate sequence comparisons and variant identification; GeneMapper™ v3.0 Software offers automated peak detection and allele calling with process quality values to simplify the confirmation and review process for microsatellite analysis. With the introduction of the Applied Biosystems 3730 DNA Analyzer, the data input and analysis pathways have been simplified in order to increase the sample processing efficiency of diverse laboratories.

R. J. Chalkley¹, L. Huang¹, K. C. Hansen¹, N. P. Allen², M. Rexach², A. L. Burlingame¹;
¹UCSF, San Francisco, CA, United States, ²Stanford University, Stanford, CA, United States.

This work presents a detailed analysis of a complex mixture of proteins using a selection of multi-dimensional separation strategies. Our group has previously studied the protein-protein interactions of yeast nuclear pore complex proteins using 1D SDS-PAGE followed by mass spectrometric analysis of discrete bands observed. This work has now been extended to study changes in interactions during the cell cycle.
Isolated protein complexes have been characterized using the same 1D SDS-PAGE followed by mass spectrometry approach employed before, but also have been analyzed using multi-dimensional liquid chromatography followed by mass spectrometry. Following strong cation exchange as a first dimension of separation samples were analyzed both by LC-MALDI-MS (TOF-TOF) and by LC-ESI-MS (QSTAR). The new cleavable ICAT technology (Applied Biosystems) was also employed to quantify changes in protein ratios.
The results show a substantial overlap between the proteins identified with the different approaches, but each approach provides additional data. LC-ESI-MS provided more comprehensive information on the sample content, but the off-line analysis of sample by LC-MALDI-MS allowed a more focused analysis of those peptides which are of particular interest. The 1D gel provides a clearer representation of dramatic changes in protein levels between samples. However, ICAT technology is better for studying subtle changes, and does not suffer from the potential complication of a protein that is changing in concentration co-migrating with another protein.
In conclusion, for comprehensive analysis of a complex mixture a combination of approaches is most informative.
This work was supported by NIH grants NCCR 01614 and RR14606.

K. G. Huwiler, D. F. Mosher, M. M. Vestling;
University of Wisconsin, Madison, WI, United States.

While establishing the location of the disulfide bonds in thrombospondin type 1 repeats (1), using partial proteolysis, high performance liquid chromatography, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS), and reduction of disulfide bonds, we found that the choice of matrix can significantly enhance the appearance of cystine- and cysteine-containing peptides. The ready in-source-decay (ISD) of disulfides using 2,5-dihydroxybenzoic acid (dhb) and a-cyano-4-hydroxycinnamic acid (CHCA) as matrices is well known (2-6). The appearance of ions corresponding to the thiols making up the disulfide in the spectrum of a disulfide is often annoying and complicates spectral interpretation. In our hands, the cleanest spectra of disulfides were obtained using mixtures of CHCA or dhb and 2-(4-hydroxyphenylazo)benzoic acid (HABA). Experiments using different mixtures and methods of mixing points up the need to consider the role of radicals and electrons in the MALDI process. In the presence of HABA, a dramatic reduction in the intensity [M].+ with respect to [MH]+ is seen for CHCA. With our conditions, disulfide containing peptides and recombinant proteins can now be easily characterized using MALDI-TOF-MS.
(1) K. G. Huwiler, M. M. Vestling, D. S. Annis, D. F. Mosher, Biochemistry, in press.
(2) S. D. Patterson, V. Katta, Anal. Chem. 66: 3727-3732 (1994).
(3) D. L. Crimmins, M. Saylor, J. Rush, R. S. Thoma, Anal. Biochem. 226: 355-361 (1995).
(4) M. D. Jones, S. D. Patterson, H. S. Lu, Anal. Chem. 70: 136-143 (1998).
(5) V. Katta, D. T. Chow, M. F. Rohde, Anal. Chem. 70: 4410-4416 (1998).
(6) V. Schnaible, S. Wefing, A. Resemann, D. Suckau, A. Bucker, S. Wolf-Kummeth, D. Hoffmann, Anal. Chem. 74: 4980-4988 (2002).

O. Chertov, T. Boronina, V. M. Hoang, T. P. Conrads, T. D. Veenstra, R. J. Fisher;
SAIC Frederick, NCI Frederick, Frederick, MD, United States.

Most reliable methods of protein identification by mass spectrometry require the generation of peptides using highly specific proteases. The confidence of protein identification improves with the increase of the number of analyzed protease-specific peptides. To generate the maximum number of peptides, the disulfide bonds should be disrupted to ensure a complete enzymatic digestion. The most popular method of disulfide bond cleavage is the reduction of proteins using thiols with subsequent alkylation of the free sulfhydryl groups. To maintain protein solubility high concentrations of chaotropic agents are used, however, they must be removed prior to analysis. These steps can lead to losses of analyte. Another method for disulfide bond cleavage is based on oxidation of proteins by performic acid. Cysteine and cystine are oxidized to cysteic acid, methionine to sulfone, tryptophan is transformed into multiple forms including N-formylkynureine. The phenolic group of tyrosine and hydroxyl groups of serine and threonine can also be oxidized and other residues could be formylated.
We investigated if this method could be modified to reduce undesirable modifications, yet still be sufficiently effective for disulfide bond cleavage yielding well-defined oxidized derivatives of Cys, Met and Trp. We found that the treatment of different proteins with 0.3 % hydrogen peroxide in concentrated formic acid for 15 min at 4° C cleaves intramolecular disulfide bonds, while converting cysteine and cystine into cysteic acid, Met into Met sulfone, and Trp is oxidized to Trp-2O.
We have utilized this method as a quick and efficient preparatory step before digestion of proteins for mass spectrometric identification. Contract NO1-CO-12400.

T. Lee¹, K. A. Resing¹, N. G. Ahn^1,2;
¹University of Colorado at Boulder, Boulder, CO, United States, ²Howard Hughes Medical Institute, Boulder, CO, United States.

Signaling pathways involving protein kinases are critical for many aspects of cell function. One regulatory mechanism for these pathways involves enzyme docking interactions in protein complexes. Recent studies identified two potential docking domains in ELK1 that target unknown docking sites on extracellular signal regulated kinase (ERK2). Mutations of the ELK1 docking domain increases the Km of ERK2, suggesting that docking interactions regulate catalysis. In order to identify the docking site on ERK2 and gain insight into how catalysis is activated, we carried out hydrogen exchange mass spectrometry (HX-MS), where the protein is incubated in D₂O for various times to measure the deuterium exchange at the back-bone amides. HX reflects conformational flexibility in proteins, as well as changes in solvent protection upon ligand binding. Experiments were carried out with activated ERK2 (ppERK) and two peptides with sequences derived from the two candidate docking domains of ELK1 (D-domain and FXFP-motif). Measurements were made with a QSTAR® (API), using LC/MS to resolve peptides; the high resolution and signal sensitivity of this system improved data quality. We also implemented a Visual Basic program which determines the weighted mass average for each peptide at each time point; this information was then fit to a three exponential model using DataFit (Oakdale Engineering). HX-MS data show these peptides each bind in distinct regions of ppERK, and both peptides altered the exchange rates near the active site and on the opposite side of ppERK. We propose that conformational mobility changes in the hinge and near the activation lip induced by docking interactions are important for ERK kinetic mechanism.

D. Chelius, C. Shieh;
ThermoFinnigan corp., San Jose, CA, United States.

The detection limits for peptide detection by mass spectrometery has improved significantly. This is important for proteomics research. Higher sensitivity allows the detection of low abundance proteins. Nano-electropray coupled with mass spectrometry is the method of choice for maximum sensitivity and separation of peptides. Peptides are separated on a reversed-phase column (75 μm ID), which is placed directly in front of the mass spectrometer, limiting peak dispersion after the column. In this paper, we describe a method that allows fast loading and highly efficient separation of peptide fragments. Peptide fragments are loaded on a peptide trap at high flow rate. After loading, the trap is switched to a low flow, reversed phase gradient with a C18 packed tip and ion trap mass spectrometer. Peptides from a bovine serum albumin digest can be detected at 500 amol. This method allows fast loading and high sensitivity detection of peptide fragments.

F. A. Abdi, S. Bhardwaj, S. Hattan, M. Lin;
Applied Biosystems Inc., Framingham, MA, United States.

Global analysis of cellular proteomics recently gained ample interest of many researchers. However, conventional methodologies for protein separation and identification are still not comprehensive enough as low abundant proteins and membrane proteins are less represented and more difficult to detect. In addition, the speed, accuracy and sensitivity at which such analyses are carried out play important role in such studies. To address this, multidimensional liquid chromatography interfaced with MS has been introduced as new technology to generate more comprehensive data. Applied Biosystems’ 4700 Proteomic Analyzer with TOF/TOF® optics was developed to provide high speed and sensitivity associated with MALDI with the ability to analyze selected ions by high-energy fragmentation. We demonstrate a mechanism enabling LC MALDI MSMS by spotting eluent from LC separation onto MALDI plates followed by analysis on ABI’s 4700 Proteomic Analyzer.A protein mixture from a living organism, obtained from the Swiss Proteomics Society, was used as a collaborative study to independently evaluate members’ abilities in the field of protein identification from complex mixtures by MS. Sample was separated using one-hour gradient on LC Packings Ultimate micro-capillary HPLC system interfaced with LC packings Probot and deposited on a single 576 spot MALDI plate. MS and MSMS analyses of each spot were carried out using ABI’s 4700 Proteomic Analyzer. Protein identifications were accomplished using ABIs’ GPS Explorer software in concert with Mascot database search engine. Sample origin was found to be Snake Venom as more than 25 proteins were identified belonging to this origin. Additional protein families were identified by database search and De Novo sequencing. This study illustrates the advantage of ABI’s 4700 Proteomic Analyzer and its versatility in LC/MALDI MS/MS studies.

C. Van Pelt;
Advion BioSciences, Inc., Ithaca, NY, United States.

Detection of non-covalent interactions by electrospray ionization mass spectrometry (ESI/MS) provides important information on molecular mass and binding stoichiometry of protein-ligand complexes. Conventional methods employ pulled-capillary nanoelectrospray; a tedious and time-consuming technique that requires a highly skilled user. The NanoMateä 100 and ESI Chipä, an automated nanoelectrospray system, have been developed to improve the efficiency and quality of such ESI/MS methods. Advantages of the system include low sample consumption, conservation of sample not consumed in the analysis, one-time spray optimization, enhanced spray stability, and no carryover.
The system aspirates samples from a 96-well microtiter plate using disposable, conductive pipette tips, and then delivers samples to the inlet of the ESI Chip, a microarray of 100 ESI nozzles microfabricated in silicon. A pressure seal is formed around a through-chip channel at one nozzle and nanoelectrospray is initiated by applying a spray voltage and psi pressure to the sample in the pipette tip.
A non-covalent interaction experiment for the binding of cytidine 2'-monophosphate to ribonuclease A was performed to demonstrate the capability of using the NanoMate system to detect non-covalent interactions. Ribonuclease A was titrated with cytidine 2'-monophosphate in 10 mM ammonium acetate, pH6.8. Scatchard plot was used to determine the dissociation constants. The result is in good agreement with the previous report. The same approach is used to study the dissociation constants for an endocelluase system. An inactive mutant of endoglucanse catalytic domain from the T. fusca bacterium was titrated with ligands cellotetrose and cellopentose. This is the first report using an ESI/MS approach to study the interaction of cellulase and its oligosaccharides. The results demonstrate that the system can be used to rapidly screen potential drug candidates in discovery programs.

U. Schweiger-Hufnagel¹, M. Lubeck¹, D. Suckau¹, P. Kovalski², C. Baessmann¹;
¹Bruker Daltonik GmbH, Bremen, Germany, ²Bruker Daltonics Inc., Billerica, MA, United States.

In proteomic research de-novo sequencing of peptides becomes a central issue when database search fails. This might happen due to post-translational modifications, sequence errors in the database or protein sequences missing in the database. For those cases a new de-novo sequencing was integrated in proteome analysis software.
Proteins from 2-D gels from various organisms were enzymatically digested, and MS- and MS(n)-spectra were acquired on an electrospray ion trap mass spectrometer and on a MALDI TOF/TOF mass spectrometer. The software used the automatically created peaklists to generate peptide sequence proposals, considering all provided hints including possible modification. The resulting peptides sequences were scored against the experimental spectrum or used for homology searches.
The presented results include the the recognition and localization of modifications, e.g. the double oxidation of tryptophane. Homology searches based on the obtained de-novo sequences were applied to detect errors in the sequence database and to identify proteins whose sequence differed from the one contained in the database.

C. Sauber;
Agilent Technologies, Waldbronn, Germany.

Oligonucleotides (ONs) are widely used as primers for polymer chain reactions (PCR) and in antisense therapy. Antisense ONs can provide useful levels of gene-specific inhibition and have, therefore, become important functional genomics tools. Significant ON-mediated inhibition has been demonstrated. ONs are analyzed with various analytical techniques including gel electrophoresis, capillary electrophoresis, MALDI/TOF and HPLC.
All these techniques could be used either for qualitative or quantitative analysis and do not allow a simultaneous confirmation and quantification of ONs. Only the on-line coupling of a separation step with mass spectrometric detection allows a rapid, sensitive and fully automated identification and quantification of ONs. LC separation is preferably used over capillary electrophoresis for this application, as LC allows using more MS compatible buffer and larger injection volumes could be used. However the sensitive analysis of oligonuclotides with LC/MS needs a carefully developed method. The hydrophilic nature of these analytes allows a gradient elution from reversed-phase columns at low organic content only, while mass spectrometric detection with electrospray requires a high pH and organic content for good ionization efficiency.
We describe a methodology based on ion-pair reversed-phase capillary LC/Ion Trap mass spectrometry for the low femtomole quantitation of ONs in biological matrices.
The method for the sensitive quantitation of oligonucleotides in biological matrices shows a limit of quantitation of 30 fmol injected. Retention on capillary LC 300 um reversed-phase columns could be achieved by using a gradient of 5 mM triethylamine acetate, pH 7 and acetonitrile, starting at 1 % organic. For best MS sensitivity 5 mM triethylamine/10 mM imidazole in 90% acetonitrile /10 % water was added post-column to further suppress adduct ion formation and aid the desolvation process with a higher content of organic modifier.

X. Guo, C. Liu, L. M. Fell, T. Covey;
Applied Biosystems / MDS Sciex, Concord, ON, Canada.

The combination methods of mass spectrometry (MS) and nanoflow chromatography have become the method of choice for the characterization of low level of proteins in biological samples, an example of such tryptic digest crude cellular mixtures. In proteomics, not only is the task of protein identification important but increasingly research has focused on the study of the biological function and specificity of proteins, such as post-translational modifications. The complete LC gradient has proven to be an essential for the better coverage of protein due to the wide distribution of degree of hydrophobicity of peptides in protein digests. Therefore, the robustness and reliability of sprayer technology has come to be an issue for convenience and throughput when performing these analyses. In this poster, we present the results of a nanoLC interface which proves sensitive and robust on a QqTOF mass spectrometer. A number of nano-LC/MS experiments of standard proteins digests are used to evaluate the interface in both positive and negative scan modes.

S. Bhardwaj, F. Abdi, S. Hattan, M. Lin;
Applied Biosystems, Inc., Framingham, MA, United States.

Proteomics has developed as the principal technology in the drug discovery area. Mass Spectrometry has become the technology of choice for analysis of biological samples from a variety of sources. One major challenge in this field is analysis of low abundance proteins from complex mixtures. As demonstrated in this study, the recently developed ABI 4700 Proteomics Analyzer with TOF/TOF® optics can overcome this challenge. It provided high sensitivity associated with MALDI and has the capacity to follow the acquisition of MALDI-TOF mass spectrum with automated tandem time of flight fragmentation and analysis of selected ions. Even though this technology seems to be an ideal alternative to traditional LC-MS/MS, LC separation is still needed for MALDI based mass spectrometry. In this study, by comparing protein identifications from sample spots containing un-separated mixtures, we demonstrated that LC separation is essential to overcome ion suppression effect on mass spectrometers. It allows in depth study of Proteomics by identifying low abundant proteins. It also largely increases the confidence of protein ID from relatively simple mixtures, such as those from 1D gels.
Protein digest mixtures that contain different number and amount of proteins were analyzed using ABI 4700 Proteomics analyzer. Comparing the MS/MS analysis from single sample spots that contain undigested protein mixtures, LC-MALDI MSMS technique not only identified more proteins, it also positively identified low abundant proteins presented at different dynamic range in each mixture. Using this approach, more peptides were also identified for each protein and therefore improved the confidence of protein identifications by providing higher sequence coverage. Protein digests from an E. Coli 1-D gel sample were also analyzed by LC-MALDI MSMS technique.

D. Suckau¹, S. Hahner¹, P. Hufnagel¹, A. Resemann¹, J. P. Kowalski², A. Holle¹;
¹Bruker Daltonik GmbH, Bremen, Germany, ²Bruker Daltonics Inc., Billerica, MA, United States.

A novel TOF/TOF design has been developed in which a potential lift followed by fragment ion post acceleration (i.e., the source of a 2^nd TOF) is the central functional element. The technique is dubbed LIFT-TOF/TOF. Besides the capability of MS on intact proteins or protein mass fingerprints, the instrument offers three MS/MS modes of operation: 1. LID (laser-induced decomposition) using the metastable ions generated by the HCCA matrix, the standard mode for protein identification; 2. CID (collision-induced decomposition) high energy fragmentations due to a gas filled into the collision cell; and 3) reflector-ISD (in-source decay) prompt fragmentation, providing monoisotopically resolved precision fragment masses of intact proteins up to 40 kDa and higher. All MS/MS spectra are obtained as single scan spectra and acquisitions for proteomics data take a few seconds. Data-dependent MS/MS acquisition and processing is fully automated.
On protein digests MS sensitivity was in the 200 attomole range and for MS/MS in the 1-10 fmole range. The high sensitivity and ready availability of MS/MS data on tryptic peaks allowed the detection of sequence errors, modifications and an overall significant improvement of the extractable information content. De novo sequencing combined with homology searching (MS BLAST) was applied to these data. Typically, LID was used for de novo sequencing and CID only, if the Ile/Leu ambiguity needs to be resolved.
ISD was used for internal sequencing of intact proteins. ISD provides monoisotopically resolved fragment ions of N-terminal residues ~10-40 based on assigned c-type fragment ions and, therefore, allowed to obtain near N-terminal sequence information irrespective of a blocked N-terminal.

I. P. Smirnov, X. Zhu, I. A. Papayanopoulos, D. J. Pappin;
Applied Biosystems, Framingham, MA, United States.

MALDI-TOF Mass Spectrometry has become one of the major analytical tools in the analysis of proteins and peptides. It offers a very high throughput rate with data acquisition speed up to few seconds per spectrum. The successful identification of the peptides depends on the signal intensity and the absence of interfering signals. MALDI is susceptible to alkali salt contamination, which reduces sensitivity and causes the matrix cluster formation (widely reported for CHCA matrix) observed as signals dominating in the range below 1100 Da in the spectrum. One way to remove these background signals especially for small peptide concentrations lower then 10 fmol/ul is to wash matrix/sample spots after crystallization on the MALDI plate with deionized water prior to analysis. This method takes advantage of the low water solubility of CHCA compare to it’s alkali salts. We report here that the application of some ammonium salt solutions, such as citrates and phosphates instead of deionized water greatly improves efficiency of such approach. Besides total matrix cluster suppression due to more complete salt removal, such approach provides substantial (~3-5 fold) improvement in the sensitivity of MALDI detection compared to unwashed sample spots. This makes the analysis of protein digest in concentration below 1 fmol/ul easy to achieve. The effect of washings of the sample spot with various ammonium salts has been studied as a function of the their type, pH and concentration. This sample preparation method provided much better peptide coverage in database search and better MS-MS data on the 4700 proteomics analyzer for subnanomolar concentration protein digests, thus is becoming a valuable approach in every day sample preparation in our laboratory.

X. Zhu, I. Smirnov, I. A. Papayannopoulos;
Applied Biosystems, Framingham, MA, United States.

Introduction
Alpha-cyano-4-hydroxycinnamic acid (CHCA) has been widely used as matrix for the analysis of proteins and peptides by Matrix Assisted Laser Desorption Ionization (MALDI) mass spectrometry. CHCA also ionizes and matrix cluster ions and their salt adducts often interfere with the detection and measurement of peptide ions. We found that these CHCA adduct ions are suppressed in MALDI mass spectra when certain ammonium salts are added to the CHCA matrix.
Method
Proteins were digested with trypsin. Samples were diluted to a series of concentrations from 1 fmol/μL to 1 pmol/μL. Each of a number of ammonium salt solutions was mixed, at various concentrations, with CHCA to make MALDI matrix solutions. Protein samples were mixed with the matrix solutions and spotted on MALDI targets ("dry-droplet" method). Samples were analyzed with the Applied Biosystems Voyager DE™-STR mass spectrometer or the 4700 Proteomics Analyzer. Mass spectral data were interpreted using the GPS Explorer™ software.
Results
CHCA sodium and potassium adduct ions in MALDI-TOF mass spectra were significantly suppressed if protein digests were mixed with CHCA containing an ammonium salt such as ammonium phosphate or ammonium citrate. The signals of peptides present in low fmol to high amol levels were enhanced with this method. Both MALDI MS and MS/MS data were readily matched to proteins using GPS Explorer™ and the MASCOT™ search engine.
Conclusions
The addition of certain ammonium salts to CHCA results in suppression of matrix salt adducts in MALDI-TOF mass spectra and the enhancement of peptide signals as well, which benefits especially low concentration samples.

S. Ackloo¹, E. Denoyer², J. Dicesare², A. Loboda¹;
¹PerkinElmerSCIEX, Concord, ON, Canada, ²PerkinElmerSCIEX, Shelton, CT, United States.

The widening scope of proteomics and the increasing demands on high-throughput analysis has driven the need to design an instrument with improved performance and ruggedness. A new MALDI-TOF instrument — equipped with both collisional cooling and orthogonal injection — has been developed. This design effectively decouples ion production from time-of-flight measurement making mass analysis more robust. Collisional cooling conditions, laser energy and pulse rate may be varied to optimize ionization conditions, suppress fragmentation, and increase ion yield for different matrix/analyte combinations. At the same time, high mass accuracy is achieved and maintained without the need for internal calibration; a critical success factor for protein analysis.
Decoupling ionization conditions from mass analysis permits more flexibility in the design of a sample target. Traditionally, samples are prepared on reusable metal targets where sample carryover is a concern. Single-use sample plates, in microtiter format, were developed to assure reproducible analysis without contamination. The new single-use plates perform similarly to traditional stainless steel targets.
The performance of this new instrument will be demonstrated using proteins and protein digests prepared using traditional methods. Performance criteria including mass accuracy, mass resolution, and sensitivity will be examined. Mass accuracy in the low ppm range and sensitivity in the low fmol range provide the opportunity to successfully analyze low quantities of sample. Mass spectra of a mixture of proteins, ranging in mass from 1.5 kDa to 30 kDa and acquired at one set of instrument parameters, will be presented. These data show that a wide mass range can be acquired without sacrificing mass resolution; a unique capability of the instrument and which is directly applicable to biomarker identification.

R. Zeng¹, C. Shieh², H. Tran², B. Wu²;
¹Shanghai Institute for Biological Science, Shanghai, China, ²ThermoFinnigan Corp., San Jose, CA, United States.

Mass spectrometry has become an important tool for the protein analysis and peptide sequence determination. Typically, a protein mixture is digested by trypsin, analyzed by mass spectrometry, and then searched against a relevant protein database where identification can be made by comparing experimental MS/MS peptide spectra with those predicted by theory. In this paper, we report the analysis of a yeast protein digest by an AP MALDI ion source coupled to an Ion Trap Mass Spectrometer. Proteins from the yeast cell lysate were extracted by protein G beads followed by one-dimensional SDS-PAGE gel electrophoresis. Ten gel bands were cut and digest by trypsin. The resulting peptide fragments were analyzed using both MALDI-TOF and AP MALDI-Ion Trap. The MS results from MALDI-TOF were searched by Mascot software. Two proteins were identified. The MS/MS spectra from the AP MALDI-Ion Trap system were searched by SEQUEST™ software. Ten proteins were identified. The ability to generate MS/MS spectra by coupling MALDI ionization with ion trap mass spectrometry provides higher confidence and greater accuracy for protein identification compared to MALDI-TOF mass fingerprinting.

P. Perkins, D. Yi, J. Meza, C. A. Miller;
Agilent Technologies, Palo Alto, CA, United States.

With the introduction of atmospheric pressure matrix assisted laser desorption/ionization ion trap mass spectrometry (AP-MALDI trap), it is possible to easily obtain MS/MS data on a protein digest using a MALDI source. Several different protein digests were analyzed using both AP-MALDI trap and ESI-LC/MS/MS. As expected, the MS/MS spectra of doubly charged ions from ESI showed more complete fragmentation and predominantly b and y ions. AP-MALDI MS/MS of singly charged ions also generated either b or y ions, but the MS/MS spectra showed other uninformative mass losses such as water. The protein sequence coverage was incomplete for both techniques at low levels, with a large overlap in detection between AP-MALDI and ESI.

J. W. Finch, J. C. Gebler, W. Chen, S. J. Berger, S. A. Cohen;
Waters Corporation, Milford, MA, United States.

The technique of static nanospray utilizes loading of 1-3 μL of sample into needle emitters which have been fabricated from borosilicate glass capillaries pulled to a fine tip (1-2 micron i.d.) and coated with a conductive material. One key advantage of this technique is the ability to introduce samples into a mass spectrometer by electrospray ionization (ESI) at very low flow rates, typically between 20-40 nanoliters/minute, where many MS/MS spectra can be acquired from a single sample over a period of 1-2 hours. While this technique is highly sensitive, samples must be relatively free of high concentrations of buffers, salts, detergents, and other contaminants which suppress ionization of peptides and proteins, and can cause the emitter to plug. Sample pre-treatment methods currently employed to remove these contaminants include use of ZipTips™ and microcolumn purification techniques. However, these methods are cumbersome, making it difficult to handle and transfer microliter volumes without sample loss. A further disadvantage is that sample recovery from packed beds is often poor.
In this study we present results demonstrating an alternative sample preparation method for static nanospray, which utilizes a surface with unique properties that allow both pre-concentration and removal of contaminants. The surface facilitates manipulation of low sample volumes without loss of sample, and the process can be easily automated. Results are presented for intact proteins and protein digests from 1D and 2D gel spots analyzed on a quadrupole-time-of-flight hybrid (Q-Tof) instrument. Excellent sensitivity can be achieved when a data-directed analysis (DDA) protocol is utilized to automatically generate MS/MS data for each individual sample infused by static nanospray.

D. J. Rousell, S. M. Dutta, K. K. Murray;
Louisiana State University, Baton Rouge, LA, United States.

Desorption/ionisation on silicon, DIOS, has the advantage in that its matrix free approach makes sample preparation simpler. In addition, the lack of matrix reduces the number of low mass peaks considerably simplifying this section of the mass spectrum. The UV-DIOS technique requires a porous silicon surface. It has been shown that for IR-DIOS a porous surface is not necessary.
The instrument under development combines a commercial mass spectrometer and IR laser to produce a system capable of employing DIOS and MALDI techniques with IR and UV lasers. The mass spectrometer used is a Bruker "Omniflex" reflectron instrument that has been modified to be capable of triggering an external laser and sample targets were milled to accept silicon wafers. An Er:YAG (2.94 micrometer) laser is positioned so that the IR laser beam enters the vacuum chamber co-axially with the UV beam. With this configuration the system can be easily switched between IR and UV analysis.
Spectra obtained with IR-DIOS contain fewer low mass peaks when compared to spectra of identical compounds under IR-MALDI conditions. The ease of sample preparation is evident with biological analytes. Tissue can be applied directly to the silicon surface for immediate mass spectrometry analysis. Bacteria have also been nebulized from an aqueous solution directly onto silicon. Recorded mass spectra contain numerous peaks below m/z 20,000 as previously reported. In addition our mass spectra also contain peaks of significantly higher mass.
Tobacco smoke particulate has also been analyzed. Using a single stage Andersen N6 impactor sidestream smoke was directly collected onto silicon. Mass spectrometric analysis was achieved without the need for sample extraction or matrix addition.

M. Kostrzewa¹, T. Elssner¹, K. Fahr¹, D. Peters¹, T. Wenzel¹, J. Wurmbach², W. Pusch²;
¹Bruker Daltonik GmbH, Leipzig, Germany, ²Bruker Daltonik GmbH, Bremen, Germany.

After completion of the Human Genome Project, intensive exploration of genome function diversity is one of the main objectives. The variety of genomes and their function can be observed at several levels like genetic polymorphisms, mRNA expression, protein abundance and protein modification. Mass spectrometry has evolved to a key technology for proteomic approaches in the last few years. Here, we present GENOLINK, a SNP genotyping module expanding the capabilities of a MALDI-TOF MS system to create a platform for functional genomics studies. GENOLINK is the combination of a novel method for SNP genotyping, genoSNIP, a robotic system for automated sample preparation, and a bioinformatics package. The genoSNIP method is based on extension primers containing a non-nucleoside UV-cleavable building block and a biotin label at the 5’ part. After amplification of template DNA by PCR, these primers hybridize adjacent to the polymorphic site and will be converted to allele specific products by single nucleotide extension. Subsequently, primer extension products are desalted using genostrep, a new DNA purification system based on streptavidin coated microtiter plates. This sample preparation is a simple pipetting protocol performed by the puredisk, a novel 96fold pipetting robot. The purified molecules are cleaved by UV- irradiation releasing very small molecules containing the genotype information. These small products are measured in the mass spectrometer with very high precision and sensitivity. The mass spectra obtained are translated into genotypes by the genotools software. Thereby, the GENOLINK system comprises a complete solution for automated, highly accurate SNP genotyping by MALDI-TOF MS.

G. Zurek¹, A. Germanus¹, C. Baessmann¹, A. Ingendoh¹, H. U. Muccitelli²;
¹Bruker Daltonik GmbH, Bremen, Germany, ²Bruker Daltonics Inc, Billerica, MA, United States.

The major task of metabolite identification is rapid and reliable elucidation of structural changes. Liquid chromatography hyphenated to atmospheric pressure ionization mass spectrometers (LC-MS) has become an integral part of metabolite identification. Ion trap mass spectrometers have proven excellent tools, especially due to the multiple stages of MS capability. Because data acquisition is now performed in an automated fashion, the new bottleneck in the identification process is data interpretation.
We present here that LC-MS data sets of control and serum/urine samples are automatically interpreted by Metabolite Tools, which also provides significant input downstream for the processes of data acquisition and interpretation. Full integration of Metabolite Tools with the Esquire Control software enables an effortless process.
Examples are shown where the Metabolite Predict module generates candidates of biotransformation based on metabolism rules of the parent drug structure. Molecular weight information of predicted metabolites are used to generate a mass list for further AutoMSn experiments or can be transferred to Metabolite Detect for use in an LC-MS data extraction.
The Metabolite Detect module performs a sophisticated comparison of two LC-MS chromatograms using the eXpose algorithm, and rapidly delivers even low-level metabolites that are barely detectable using radioactive tracers. Retention time and the m/z value are then used for an MS/MS experiment that yields higher quality spectra than data dependant experiments. Identification of metabolites is then completed using the MSn data.
Metabolite Tools is not only essential in drug metabolite detection, but also useful in detecting subtle changes in LC-MS analysis of peptides, nucleotides and other biomolecules.

V. Pham, W. J. Henzel;
Genentech, South San Francisco, CA, United States.

We have developed a rapid method for digesting and identifying blocked proteins electroblotted onto polyvinylidene difluoride (PVDF) membranes. The unbound protein-binding sites on the membrane are blocked using zwittergent 3-16. The proteins are then digested with trypsin, Lys-N, Lys-C, Asp-N, Glu-C, or clostripain in appropriate buffers at 37 C for 5-30 min. Peptides generated by these digestions remain on the membrane, which are then identified by N-terminal sequencing. The mixture of sequences obtained is used for database searching which leads to unambiguous protein identification. This approach can also be used for blocked proteins that have been subjected to the Edman chemistry. The membrane is removed from the sequencer, blocked and digested with Asp-N, Glu-C or clostripain.

V. P. Hanko, J. S. Rohrer;
Dionex Corporation, Sunnyvale, CA, United States.

Integrated pulsed amperometric detection with a highly selective anion-exchange column (AAA-Direct™) is a powerful technique for amino acid analysis. This technique enables the direct detection of virtually all amino acids without pre- or post-column derivatization and without compromising sensitivity. The AminoPac™ PA10 anion exchange column is designed to separate amino acids with baseline resolution. Using high eluent strength and elevated column temperature, this column provides a rapid (<12 min) separation of tryptophan (Trp) from other amino acids and sample components. We determined Trp in alkaline hydrolyzed protein and peptide samples and cell culture media/fermentation broths. High levels of carbohydrates, typical of some samples, do not interfere with this method.
In this poster we present methods for preparing protein samples for hydrolysis and subsequent Trp analysis by AAA-Direct. Techniques commonly used to increase Trp recovery during alkaline hydrolysis were also evaluated for their suitability with AAA-Direct. Direct detection allowed us to easily optimize hydrolysis conditions because we did not have to be concerned with differences in derivatization efficiencies between samples complicating our results.

J. A. Moss, A. R. Coyle, K. D. Janda;
The Scripps Research Institute, La Jolla, CA, United States.

Immobilized metal affinity chromatography (IMAC) has rapidly become one of the most widespread affinity purification techniques employed in recombinant protein expression. The efficacy of this technique relies on the stable coordination sphere provided by an immobilized metal (Ni²⁺, Zn²⁺) and a hexa-histidine tag appended to an expressed protein. While this method usually affords material in high yield and purity, our purifications of an antibody single chain variable fragment (scFv) expressed in E. coli were contaminated by a hydrophobic 80 kDa protein and lipopolysaccharide (LPS) endotoxin. Implementation of a diphenyl reversed-phase HPLC purification step following IMAC purification effectively removed both of these contaminants without serious adverse effects on the antibody. In addition, this orthogonal purification step furnished material that could be lyophilized and obtained as a salt-free solid amenable to long-term storage without perceivable loss in activity.

E. Chertova, J. Bess, B. Crise, R. Sowder, II, T. Schaden-Ireland, L. Henderson, J. Lifson, L. Arthur;
SAIC-Frederick, NCI-Frederick, Frederick, MD, United States.

Objective: We characterized factors influencing the gp120^SU content of retroviral virions.
Methods: To determine the Gag and Env glycoproteins content of purified virions and to calculate the ratio of Gag /Env proteins in viral samples, we used microscale HPLC methods, coupled with SDS-PAGE, immunoblot analysis, mass spectrometry, and protein sequence analysis of eluted proteins.
Results: All HIV-1 and most SIV isolates examined had Gag:Env ratios of approximately 60:1, corresponding to only 7-14 envelope-containing trimers per virion, given 1200-2500 Gag molecules per virion. SIVmne E11S showed a Gag:Env ratio of approximately 6:1, corresponding to 70-140 envelope trimers per virion. TM truncation (gp32, rather than gp41) was associated with higher envelope glycoprotein content. All HIV-1 and SIV isolates examined, except one SIV clone, had surface SU:TM molar ratios of 1:1 indicating that gp120 does not generally shed from the virus once TM is inserted into the membrane; shedding would yield a molar excess of TM over SU. Virus SU content and SU:TM ratios were not significantly changed during sucrose gradient purification.. Heating purified HIV and SIV to temperatures of 55°C or greater for one hour resulted in loss of most of the gp120 from the virus but retention of TM.
Conclusions: Envelope glycoprotein incorporation and not shedding of surface glycoprotein gp120 is the primary determinant of gp120 content of purified HIV-1 SIV.
Contract No. N01-CO-12400.

J. Chervet, Sr.;
LC Packings, Amsterdam, Netherlands.

Capillary and Nano LC columns with inner diameters ranging from 50 to 300 µm are typically packed with 3 to 5 µm particle size stationary phases. Although these column are robust and yield reasonable efficient separations there are a few limitations: the slow mass transfer in the stationary phase limits the separation efficiency and the packed bed of this granular phases results in high back pressures. In contrast to conventional stationary phases, monolithic structures are highly porous, characterized by macro and meso pores which results in better mass transfer properties and in reduced back pressure.
In this poster we report on the chromatographic performance of 200 µm i.d. polystyrene/divinylbenzene monolithic capillary columns. The improved mass transfer in the stationary phase will be illustrated by H/u curve. Efficiencies up to a 1/4 million plates per meter are obtained routinely. To achieve these high efficiencies special attention must be given to the chromatographic system. The response time and sampling rate of the data system must be adapted accordingly.
Other advantages using monolithic bed structures are the high mechanical stability of the column bed (no voiding) resulting in a superior lifetime.
Ultra-fast separation of proteins and peptides including protein digests will be shown to illustrate the performance of these monolithic capillary columns. For proteins and peptides peak width at half heights of 1 to 3 seconds are achieved routinely. Peak capacities with up to 100 peaks in less than 15 minutes in gradient mode are now possible.

H. Schwartz¹, R. Swart², I. Dragan², R. Van Ling², J. Chervet², R. Van Soest³;
¹Dionex Corporation, Sunnyvale, CA, United States, ²LC Packings, a Dionex Company, Amsterdam, Netherlands, ³LC Packings USA / Dionex, San Francisco, CA, United States.

During the last decade, capillary and nano LC columns (i.d.s ~ 50-300 um) were typically packed with 3 to 5 um particles. Although these columns are robust and yield fairly efficient separations, they do have limitations: 1) slow mass transfer in the stationary phase limits the separation efficiency and 2) the packed bed of these phases results in high back pressure. In contrast, monolithic structures are highly porous, and have macro- and mesopores resulting in better mass transfer properties and reduced back pressure. Another advantage of a monolithic bed structure is the high mechanical stability of the column bed (no void formation), resulting in superior column lifetime.
Here we report on the performance of 200 um i.d. polystyrene/divinylbenzene monolithic capillary columns. The improved mass transfer in the stationary phase is illustrated by an H/u plot. Efficiencies up to a 0.25 million plates/m can be obtained routinely. To achieve these high efficiencies, special attention must be given to the hardware of the chromatographic system: connecting tubing, flow cell, injector, and other fluidic parts must have minimal dead volume. In addition, the response time and sampling rate of the data system must be adapted. A dedicated Capillary/Nano LC system, UltiMate (LC Packings, a Dionex company) meets these strict requirements. The sample capacity for proteins and peptides, and the influence of mobile phase additives also will be discussed. Finally, ultra-fast separations of proteins and peptides will be shown. For proteins and peptides gradients, peak widths of 1-3 s can be obtained with peak capacities of ~ 100 peaks/15 min.

R. Moritz¹, N. O'Reilly²;
¹Agilent Technologies, Waldbronn, Germany, ²Cancer Research UK, London, United Kingdom.

Synthetic peptides become more and more important as drug candidates in the treatment of a variety of diseases. In order to keep pace with the growing number of newly synthesized peptides, peptide purification should not represent the bottleneck in the drug discovery process.
Mass-based fraction collection is an efficient technique for the purification of compounds with well-known masses. In contrast to less specific detectors, in each run only the compound of interest is being purified. Hence, it is not necessary to pick out target compounds out of a series of redundant fractions that have been collected. Here we demonstrate mass-based purification of a series of synthetic peptides with the Agilent 1100 Series Purification System. The modular set-up of the system, a reliable fully automated delay volume calibration and a comprehensive software package assemble a versatile platform for purification tasks in the flow range from 0 to 100 mL/min. We could successfully purify crude peptides by reverse phase HPLC from less than 1 kDa to more than 10 kDa, covering a pI range from 4 to 13.

T. Neumann, R. Moritz, M. Kuschel;
Agilent Technologies, Waldbronn, Germany.

Liquid chromatography in analytical and preparative scale plays a key role in the purification of proteins and peptides. Due to its high resolving power reverse phase high performance chromatography (RP-HPLC) is often employed as a last polishing step in a purification workflow. Besides appropriate chromatographic conditions also reliable fraction collection is indispensable. Herein we present an LC system with a fully automated delay calibration and low delay volume for reliable fraction collection of valuable samples. In a combined approach we demonstrate how microfluidics can be integrated efficiently into a protein purification workflow as an orthogonal method for purity checks. In addition the microfluidic system allows to monitor and optimise the protein purification process in a fast and easy way.

F. A. Chen¹, M. Santos¹, D. Binger²;
¹Life Sciences Division, Beckman Coulter Inc., Fullerton, CA, United States, ²Diagnostics Division, Beckman Coulter Inc., Chaska, MN, United States.

Monoclonal antibodies obtained from mouse ascites fluid and cell culture demonstrate differences in isoelectric range despite their identical genetic imprint from the hybridoma. This may influence susceptibility to chemical modification and aspects of functional behavior in immunoassays. Presumably the post-translational modification of the monoclonal antibodies by sugar moieties results in such differences. In this report, we demonstrate a practical approach for the analysis of N-linked oligosaccharide in IgG monoclonal antibodies obtained from mouse ascites fluid and cell culture and provide comparative isoelectric focussing profiles for two production methods. Purified IgG was denatured in the presence of a reducing agent and the N-linked oligosaccharides were released in the presence of peptide N-glycosidase (PNGase F; EC3.5.1.52). The released reducing sugars were derivatized with 8-aminopyrene-1,3,6-trisulfonate (APTS, 1) under mild reductive amination conditions to avoid desialylation and the detachment of fucose residues. The reaction adducts mixture was analyzed directly without any prior clean-up procedures by capillary electrophoresis (CE) with a laser-induced fluorescence (LIF) detection.
Significant differences in the oligosaccharide profiles were observed between the antibody obtained from the ascites fluid and the cell culture. These results are compared with the microheterogeneity of protein species obtained by capillary isoelectric focussing of the same monoclonal antibodies. Differential behavior of each monoclonal antibody and relationship to altered isoelectric point will be discussed. The present method of oligosaccharide profiling of IgG monoclonal antibody appears to be suitable for routine analysis.
Reference:
1) F-T. A. Chen and R. A. Evangelista. Profiling N-linked oligosaccharides of Glycoproteins. Electrophoresis, 15, 1892-1898 (1998).

A. K. Shukla, M. M. Shukla;
Glygen Corp., Columbia, MD, United States.

The desalting and cleanup of small water-soluble molecules such as carbohydrates is a challenging task. Due to their smaller size and similar molecular properties, these molecules cannot be cleaned easily by dialysis, because the small pore size (100, 500 or 1000Dalton) of the membranes allows only a very slow dialysis and may take 24 hrs or longer. Furthermore, the molecular weight cut off’s (MWCO) of these membranes of smaller pore sizes are not very sharp, therefore, it is difficult to get the sample completely desalted for further analysis (such as MALDI, MS, capillary electrophoresis, capillary HPLC). By using the activated charcoal, porous carbon or ion-exchange resin in a small filter-less column with volatile elution buffer, the sample can be cleaned up efficiently for further analysis. Using active charcoal and porous carbon, the different mono, di and oligo-saccharides can be easily cleaned up from phosphate buffer and the samples are analyzed by HPLC. The binding capacity of the active charcoal is higher than that of the porous carbon, however, the reproducibility of porous carbon is better, because porous carbon particles are commercially available with controlled particle size. However, commercially available active charcoal particles are not of controlled particle size and are a mix of various shapes and sizes. By using the same technique, other small water-soluble molecules, such as nucleotides, water-soluble dyes, peptides, glycopeptides can also be cleaned up for further analysis.

A. M. DeSantis;
Virginia Tech Center for Genomics, Blacksburg, VA, United States.

Several methods are currently available for enriching phosphopeptides for analysis by liquid chromatography-tandem mass spectrometry. Most protocols involve enrichment of the phosphopeptides by immobilized metal-ion affinity chromatography (IMAC) prior to separation by reverse-phase chromatography and subsequent analysis by mass spectrometry. A methylation step for minimizing contamination of phosphopeptide solutions by acidic, non-phosphorylated peptides has also been reported. As we desired to have a routine protocol in place to rapidly enrich and analyze for phosphopeptides, we embarked on a study to compare several strategies to see which protocol, if any, offers both convenience and breadth of recovery. Two phosphopeptide purification techniques, an Fe-IMAC column procedure and the commercially available gallium-based phosphopeptide isolation discs are assessed for their abilities to obtain enriched phosphopeptide solutions from non-casein standards as well as tryptic digests. Both methylated and nonmethyated samples are used in the investigation.

T. H. Steinberg, K. J. Martin, B. J. Agnew, K. R. Gee, W. Leung, T. Goodman, B. Schulenberg, J. M. Beechem, R. P. Haugland, W. F. Patton;
Molecular Probes, Inc., Eugene, OR, United States.

Fluorescence detection methods are likely to offer the best solution to detection and quantitation of total protein profiles and post-translational modifications in proteomics. Reversible protein phosphorylation plays a critical regulatory role in biological systems, as demonstrated in the context of carcinogenesis and other disease states. Our newly developed Pro-Q Diamond phosphoprotein dye technology is suitable for the fluorescent detection of phosphoserine-, phosphothreonine- and phosphotyrosine-containing proteins displayed on SDS-polyacrylamide gels, 2-D gels and protein microarrays. This technology is based upon small-molecule-fluorophore, phospho-amino acid recognition, and does not require antibodies. Additionally, the technology is appropriate for the determination of protein kinase and phosphatase substrate preference. DNA, RNA and sulfated glycoproteins fail to be detected with Pro-Q Diamond dye. The staining is rapid, simple to perform, readily reversible and fully compatible with modern microchemical analysis procedures, such as MALDI-TOF mass spectrometry. Pro-Q Diamond dye technology currently can detect as little as 1-2 ng of b-casein, a pentaphosphorylated protein, and 8 ng of pepsin, a monophosphorylated protein, in SDS minigels with a linear response over a 500-1000 fold concentration range. Less than 1 pg of phosphorylated protein and/or peptide can be detected in microarray format using a microarray optimized procedure for Pro-Q Diamond stain. Through combination of Pro-Q Diamond phosphoprotein stain with Pro-Q Emerald glycoprotein stain and SYPRO Ruby protein gel stain, our Multiplexed Proteomics platform permits quantitative, multi-color fluorescence detection of phosphorylated proteins, glycoproteins, and total-protein in the same 2-D gel or microarray.

K. A. Lee, J. B. Hurley;
University of Washington, Seattle, WA, United States.

On stimulation, rhodopsin, the light-sensing protein in the rod cells of the retina, is phosphorylated at several sites on its C terminus as the first step in deactivation. We have developed a mass spectrometry-based method to quantify the kinetics of phosphorylation in vivo. After exposing a mouse to a light stimulus, phosphorylation and dephosphorylation reactions are terminated by rapidly homogenizing the eye in 8M urea. The C-terminal peptide containing all known phosphorylation sites is cleaved from rhodopsin, partially purified by ultracentrifugation, and analyzed by LCMS. The relative sensitivity of the mass spectrometer to peptides with zero to five phosphates was determined using purified phosphopeptide standards, and an equation was generated to calculate these sensitivity coefficients for other phosphopeptides. Three primary phosphorylation sites, Ser334, Ser338, and Ser343 were distinguished by liquid chromatography coupled with tandem mass spectrometry (LCMS/MS). Although peptides monophosphorylated at Ser338 and Ser343 coeluted, the relative amounts of each species were determined by monitoring the ratio of specific daughter ions characteristic of each peptide. Doubly-phosphorylated and triply-phosphorylated rhodopsin peptides with different sites of phosphorylation also were distinguished by LCMS/MS. These methods were used to study the kinetics of rhodopsin phosphorylation and dephosphorylation in response to a continuous light stimulus. Sixty to ninety minutes of light adaptation were required to achieve steady-state phosphorylation. Upon light exposure, Ser343 is phosphorylated most rapidly, followed by Ser338 and Ser334. However, at the steady state, rhodopsin peptides are monophosphorylated on Ser334 seven times more than Ser343 and three times more than Ser338. After reaching the steady state of phosphorylation, mice were placed in darkness, and the kinetics of dephosphorylation were determined for each phosphorylation site.

G. M. Hathaway, J. Zhou, F. Rusnak;
Caltech, Pasadena, CA, United States.

The increasing number of known sequences makes it possible to use a new approach to determine sites of post-translational modifications (PTM) by chemically targeting them for specific proteolytic cleavage. Analysis of the digest allows one to obtain correct structures by inference. We call this approach chemically-targeted- identification or CTID.
Peptides, phosphorylated or glycosylated on serine and threonine residues, underwent ß-elimination in the presence of the nucleophile 2-aminoethanethiol to produce analogs of lysine. Both serine and threonine derivatives were specifically cleaved with lysine endopeptidase.
The products of digestion were analyzed by mass spectrometry and/or Edman sequencing. Prior knowledge of the sequence allowed site identification either from peptide mass, or from a few residues sequenced from the newly liberated amino terminus.
Oxidation of the modified peptide allowed its detection in digest mixtures by precursor ion or neutral loss linked scans. Starting with 2 picomoles of a 3 kDa glycopeptide, the CTID method successfully assigned the site of glycosylation to a single threonine residue.

M. Zou¹, C. Shi², X. Hou², R. H. Cohen²;
¹The University of Tennessee, Knoxville, TN, United States, ²Boston University, Boston, MA, United States.

We have previously reported that peroxynitrite (ONOO^-), a potent oxidant formed from nitric oxide (NO) and superoxide anions (O₂^.-) at a diffusion-controlled rate, activates 5’-AMP-activated kinase (AMPK), resulting in phosphorylation of endothelial nitric oxide synthase serine¹¹⁷⁹-P (eNOS-Ser¹¹⁷⁹-P) and acetyl CoA carboxylase (Ser⁷⁹, ACC), two known downstream targets of AMPK. Here we investigated how ONOO^- activates AMPK in cultured bovine aortic endothelial cells (BAEC). In parallel with increased AMPK (Thr¹⁷²) phosphorylation as well as that of eNOS and ACC, ONOO^- significantly increased c-Src (Tyr⁴¹⁶) phosphorylation. Inhibition of c-Src activity by PP2 or overexpression of a c-Src dominant-negative mutant blocked ONOO^--enhanced phosphorylation of AMPK, eNOS-Ser¹¹⁷⁹-P, and ACC-P. In addition, ONOO^- significantly enhanced levels of phosphorylated PDK1 and p70S6 kinase, indicating that ONOO^- activated PI-3 kinase. Furthermore, inhibition of PI-3 kinase by either wortmannin or LY294002, or overexpression of an inactive p-PDK1 mutant attenuated ONOO^--induced AMPK phosphorylation as well as that of eNOS-Ser¹¹⁷⁹-P and ACC-P. Inhibition of PI-3 kinase also prevented the co-immunoprecipitation of AMPK and eNOS induced by ONOO^-, suggesting that the oxidant, through its actions on upstream kinases, favors the physical association of AMPK and eNOS. Taken together, our results indicate that ONOO^- activates AMPK in a PI-3 kinase-dependent manner and suggest that ONOO^--induced activation of AMPK might thereby regulate metabolic enzymes, such as ACC, HMG-CoA reductase, and glycogen synthase.

W. L. McEldoon, M. J. Horn;
BioMolcular Technologies, Inc., Sunnyvale, CA, United States.

We previously have described easy methodologies employing our Pi^3™ solid-phase reagents to selectively separate and isolate methionine-, tryptophan-, or phosphotyrosine-containing peptides from mixtures. We have applied these methodologies to digests of mixtures of proteins, peptides derived from in-gel digestion of proteins, as well as digestions of lysates. We report here the development of an improved, simplified and quicker method for phosphotyrosine, using a novel water-insoluble, solid-phase reagent that forms a covalent bond with phosphotyrosine-containing peptides. The reagent provides for the selective isolation of phosphotyrosine peptides from peptide mixtures and from mixtures containing phosphoserine- and phosphothreonine-containing peptides. These separated phosphotyrosine- containing peptides subsequently may be regenerated and analyzed by mass spectrometry or liquid chromatography. When coupled with mass spectrometric analysis and more focused database searching, these methodologies provide a significant increase in simplicity and accuracy of protein identification. Additionally, sites of tyrosine phosphorylation may be readily identified in known and unknown proteins. This new methodology has been shown to allow for the identification of phosphotyrosine peptides at the 2% occurrence level. Compared with IMAC, the phosphopeptide yields of this described method are substantially higher, and the method provides the specificity of phosphotyrosine only.

D. C. Watson, N. M. Young, J. R. Brisson, J. Kelly, H. C. Jarrell, L. Tessier, P. H. Lanthier, N. Cadotte, F. St Michael, E. Aberg, C. M. Szymanski;
National Research Council, Instiute of Biological Sciences, Ottawa, ON, Canada.

Campylobacter jejuni is a Gram-negative bacterium that is a leading cause of gastroenteritis, which in a small number of cases can be followed by a neuroparalytic disorder, Guillain Barre Syndrome. Mass spectrometry investigations of partially purified C. jejuni protein PEB3 showed it to be partially modified with an Asn-linked glycan of mass 1406Da and comprised of one hexose, five HexNAc, and a species of mass 228Da, consistent with a trideoxydiacetamido hexose. The attachment site was an Asn residue within a sequence DFNVSK, thus matching the eukaryotic N-linked glycoprotein sequon Asn-Xaa-Ser. By means of soybean lectin affinity chromatography, a mixture of glycoproteins was obtained from a glycine extract, and 2D-gel proteomics analysis led to the identification of at least 22 glycoproteins, predominantly annotated as periplasmic proteins, all of which contained putative Asn sequons with either Ser or Thr. Glycopeptides were prepared from the glycoprotein mixture by pronase digestion and gel filtration. The structure of the glycan was determined by using nano probe magic angle spinning-NMR techniques to be
GalNAc-α1,4-[Glcβ1-3-]GalNAc-α1,4-GalNAc-α1,4-GalNAc-α1,4-GalNac-α1,3-Bac-β1,N-Asn-Xaa, where Bac is bacillosamine, 2,4-diacetamido-2,4,6-trideoxyglucopyranose. Protein glycosylation was abolished when the pglB gene was mutated, providing evidence that the enzyme encoded by this gene is responsible for formation of the glycopeptide N-linkage. Comparison of the pgl locus with that of Neisseria meningitidis suggested that most of the homologous genes are probably involved in the synthesis of bacillosamine.

A. Cordoba, B. Shyong, D. Breen, S. Shire, R. Harris;
Genentech, South San Francisco, CA, United States.

Monoclonal antibodies (MAbs) typically undergo fragmentation over time in the liquid state. The purpose of this study was to investigate whether this fragmentation is due to enzymes or kinetic forces that break the weak link that connects Fab to Fc. Fragmentation occurs in the hinge region near the papain cleavage site, resulting in Fab and Fab-Fc forms. Four different Genentech MAbs were exchanged into the same buffer and stored at —20°C, 5°C, 30°C, and 40°C for 1 month. Protease inhibitors were added to one of the MAbs. Cleavage rates were measured after incubation using size exclusion chromatography (SEC). Peaks that were separated using SEC were analyzed by electrospray mass spectrometry to identify the clipping sites. The effect of host cell proteins on fragmentation was also studied. Different pools of an antibody (affinity pool, cation exchange pool, and final bulk pool) were incubated at different temperatures. After incubation their cleavage rates were measured by SEC. Cleavage rates were observed to be equivalent between antibodies. MAb clipping does not always produce complementary Fab and Fc fragments. In all the MAbs cleavage takes place within the same heavy chain hinge sequence: Ser-Cys-Asp-Lys-Thr-His-Thr. The main clip sites are between Asp-Lys and His-Thr (the papain clip site). No clipping was observed in the light chain. The use of protease inhibitors did not affect rates of cleavage or modify sites of fragmentation. Degradation was not affected by host cell protein content. In summary, fragmentation does not appear to be caused by proteases.

Z. Wu^1,2, K. Nakanishi³, J. W. Crabb^1,2;
¹Cleveland Clinic Foundation, Cleveland, OH, United States, ²Cleveland State University, Cleveland, OH, United States, ³Columbia University, New york, NY, United States.

CRALBP (cellular retinaldehyde-binding protein) serves as an 11-cis-retinol acceptor and as a modulator of 11-cis-retinol dehydrogenase in the mammalian rod visual cycle. Mutations in the CRALBP gene cause progressive retinal degenerations that lead to blindness. To better understand the molecular interactions between CRALBP and its ligand, photoaffinity labeling with retinoid analogue 3-diazo-4-keto-11-cis-retinal has been pursued and labeled sites identified by mass spectrometry. Purified human recombinant CRALBP was labeled in the dark with 3-diazo-4-keto-11-cis-retinal. Covalent incorporation of this ligand was evaluated following photolysis with UV-light (254 nm) at —196⁰C for 5s to 20 min. Protein bound retinal was reduced to retinol with NaB³H₄, radiolabeling the ligand incorporation sites. Labeled rCRALBP was denatured in 8 M urea, alkylated with iodoacetamide, digested with trypsin and peptides fractionated by RP-HPLC. Radioactive HPLC fractions were identified by scintillation counting and peptide sites of incorporation identified by MALDI-TOF MS and LC MS/MS. Short (5-40s) irradiation times yielded relatively constant incorporation levels (~1%) therefore 5s photolysis times were chosen to minimize nonspecific protein modifications. MALDI TOF MS and LC MS/MS analyses of tryptic peptides from photolabeled rCRALBP have accounted for 100% of the rCRALBP sequence. Eight photoaffinity modified residues were identified, all with variable mass additions. This variability may be due to free radical migration throughout the conjugated double bonds of the retinoid analogue. Four of the photoaffinity modified sites are in authentic retinoid binding pocket residues based upon previous studies. The other four modified residues may represent newly identified CRALBP ligand binding pocket components. The functional significance of these residues is under investigation.

M. A. Ritchie, N. Johnson, J. B. Hoyes, A. Millar, R. Carruthers, C. Jones, J. Langridge;
Micromass MS Technologies, Manchester, United Kingdom.

Asparagine (N-) linked glycosylation is one of the most common, complex, and highly variable post-translational modification. As glycoforms are the product of a series of biochemical modifications, perturbations within a cell can have profound effects on their structure. As glycosylation also plays an important role in cell signaling and recognition it’s detection and characterisation are of great importance.
Here we present a method for the detection and characterisation of glycopeptides. A proteolytic digest is analysed by reverse phase HPLC-ESI using a Q-Tof mass spectrometer. The instrument is switched at one-second intervals between low and high collision energy on the collision cell. The quadrupole operates in the non-selective rf only mode. The first data set at low energy (7eV) shows only the normal pseudo molecular ions. The second at higher energy shows their fragments. Wherever a specified product occurs in the high-energy data all its possible precursors are revealed by the corresponding 7eV data. Upon detection of the carbohydrate oxonium ions at m/z 204 (HexNAc), 366 (HexHexNAc) and 274/292 (NeuAc) the instrument is switched into MS/MS mode and ions from the low energy spectra are selected by the quadrupole for fragmentation. The high mass of glycopeptides gives rise glycopeptide ions typically of 4 or more positive charges, a feature that is exploited for their preferential selection for MS/MS. The glycosidic bonds tend to be more labile than the peptide bonds hence MS/MS spectra produce predominately glycopeptide Y-type fragment ions. Interpretation of MS/MS data is facilitated through the use of software tools.

Y. Zhang¹, R. R. Cocklin¹, K. D. O'Neil², N. X. Chen², S. M. Moe^2,3, K. R. Bidase⁴, M. Wang¹;
¹Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, IN, United States, ²Department of Medicine, Indiana University School of Medicine, Indianapolis, IN, United States, ³3Richard L. Roudebush VAMC, Indiana University School of Medicine, Indianapolis, IN, United States, ⁴Department of Pharmacology, University of Nebraska Medical Center, Omaha, NE, United States.

Beta-2-Microglobulin (β2M) amyloidosis is a destructive osteoarticular disease that affects patients undergoing long-term hemodialysis. Some of the β2M in the amyloid deposits, and circulating β2M in dialysis patients, are modified with advanced glycation end products (β2M-AGEs). Advanced glycation end products, which are made up of various glucose or carbohydrate adducts, are thought to be responsible for several diabetic and age related complications. However, to date, specific sites of AGE modification on the β2M protein remain largely unknown. We report here the use of MALDI-MS to determine the type and location of several AGEs on human β2M. A PERL script software, that was developed in-house, makes the relatively large amount of data generated by the MALDI much more manageable. The outstanding sensitivity of MALDI-MS coupled with the afore mentioned software allows such an approach to be a very useful tool in detecting AGEs and other post-translational modifications. We believe that this method could be an important option when analyzing naturally occurring AGEs in other human diseases.

C. L. Hunter¹, S. T. Weintraub², T. A. Settineri¹;
¹Applied Biosystems, Foster City, CA, United States, ²University of Texas Health Science Center, Dept. Biochemistry, San Antonio, TX, United States.

The direct combination of triple quadrupole and ion trapping capabilities in the
Q TRAP™ hybrid linear ion trap LC/MS/MS System presents new opportunities for the automated investigation of posttranslational modifications (PTMs). For the first time, the unique specificity of precursor ion (PI) and neutral loss (NL) scans for PTM identification can be automatically linked to high sensitivity ion trap MS/MS scans on an LC time scale. In addition, the ability to switch rapidly between positive and negative ion detection while performing these scans provides unique capabilities and maximum flexibility for automated PTM discovery. To identify the PTMs, Pro ID software was used to determine the protein of origin for each modified peptide and to identify the site(s) of modification on the peptides.
In this study, phosphopeptide standards containing S, T or Y phosphorylation were spiked into a characterized protein digest (500 fmole) and analyzed by nanoflow LC/MS/MS on the Q TRAP System. Several different concentration levels (20-100 fmole) were used to simulate a range of phosphorylation levels known to exist in nature. Automated PTM discovery was performed using information-dependent acquisition (IDA), where a negative ion PI scan for m/z 79 was used as a survey scan, followed by a positive ion enhanced resolution (ER) scan and a positive ion enhanced product ion (EPI) scan. At 50 fmole injected, all phosphopeptides were detected by PI scanning, and the resulting MS/MS spectra yielded positive sequence identification of 75% of these phosphopeptides. Further results will be presented on the identification and characterization of PTM sites on other modified peptides.

S. Wu, A. Huhmer, I. Jardine;
ThermoFinnigan, San Jose, CA, United States.

The analysis using peptide mass maps has become an important technique for proteomic studies. By digesting a protein from a gel spot or a separated subcellular fraction with a specific enzyme and comparing the measured fragment ions of the peptides (e.g. in tandem MS or ion trap) to those predicated by the sequence, one could elucidate the corresponding protein identification through a database search. For a targeted protein in a complex mixture, any changes in the sequence can be pinpointed to specific locations and the structures of the modifications (e.g. phosphorylation, glycosylation, methylation, and disulfide linkage) can be determined by further fragmentation (e.g. MS to 3 in an ion trap).
In this report, we will present how to use ion-trap mass spectrometer (MSⁿ) coupled withprotein identification (TurboSEQUEST^®), denovo sequencing (DenovoX™), and structure modification (Salsa) software to study the modifications of phosphorylation, glycosylation, and methylation in a complex mixture of biological fluid.

S. P. Yadav;
The Lerner Research Institute, Cleveland, OH, United States.

Unlike oligonucleotide synthesis, peptide synthesis has inherent complexities due to incredible chemical diversity in amino acid side chains. In fact each peptide made is unique in its own way. Despite the availability of several types of in situ coupling reagents and activators in solid-phase synthesis, amino acid deletion remains a major pitfall during the synthesis. To ensure that the coupling reactions reach completion, several fold excess Na-Fmoc protected amino acids are used in solid-phase synthesis. Other approaches, such as increasing the coupling cycle time for val, ile, ser, thr, gly and his including double coupling at long stretches of ser and thr have been proposed to avoid deletions. At present there are no well-defined guidelines or database available for predicting potential pitfalls in a given sequence to be synthesized. We have analyzed several synthetic peptides of various lengths by MALDI-TOF to seek information on peptide deletion products, methionine oxidation, and incomplete side chain deblocking. These preliminary data suggest that arg and asn were deleted more frequently than other amino acids in sequences that were synthesized using diisopropylcarbodiimide (DIC) as an activator.

B. Lee, S. Gupta, I. Morozova;
University of Illinois at Chicago, Chicago, IL, United States.

We have developed a novel three-dimensional sodium dodecyl sulfate polyacrylamide cube gel electrophoresis (3-D SDS-PACGE) technique to increase the resolution of protein separation. The three-dimensional technology involves the separation of proteins by isoelectric point (PI) in the first dimension, by molecular weight with a high percentage (12%) SDS-PAGE in the second dimension, and by molecular weight with a low percentage (7.5%) SDS-PACGE in the third dimension. The high percentage SDS-PAGE enhances the separation of the low molecular weight proteins. The low percentage SDS-PAGE cube gel enhances the separation of the high molecular weight proteins. The utilization of two SDS-PAGE steps in 3-D SDS-PACGE compared to one single step in 2-D SDS-PAGE has increased the separation between proteins. Three mixtures of multi-colored protein molecular weight standards containing broad, high, and low molecular weight ranges were used to provide a proof-of-concept study. We clearly observed the improvement in the protein separation. Also, the bovine serum albumin was used to demonstrate the improvement of the protein separation. Different steps including making and running of the cube gel, covalently coupling a colored dye to the proteins for visualization, and documenting the cube gel using digital camera, have been optimized to make the 3-D SDS-PACGE technology a reality.

M. R. Pisano¹, M. Saxton¹, K. J. Biederman¹, D. L. Allen¹, R. Nunez²;
¹Proteomic Research Services, Inc, Ann Arbor, MI, United States, ²Invitrogen Corp., Carlsbad, CA, United States.

Two dimensional protein gel separation is an established technique that is still considered to be the best option for protein profiling when one considers resolution. The analysis of a complex mixture of proteins using traditional large format 2D gel electrophoresis is tedious, time consuming, and can be expensive. The ZOOM® IPGRunner™ System (Invitrogen) enables researchers to profile proteins faster and easier. This system utilizes a mini-gel format for 2D separations that makes the first and second dimension gel separation fast, simple and easy. Rat liver tissue extracts were profiled using both traditional large format gels and the mini gel system. Image analysis and mass spectrometric analysis were performed in order to compare the results from these 2 formats. The methods were compared in the context of time, cost, and results of the image and MS analyses. The data acquired on the ZOOM gels at the lower protein loads was found to be superior to that from the large format gels. It is evident from the analysis that the ZOOM® Proteomic System is of great value and complements large format 2D gel profiling.

L. E. Cammish;
NextGen Sciences Ltd, Huntingdon, United Kingdom.

Current technologies used for protein profiling include the use of 2-dimensional electrophoresis (2DE) with subsequent identification via mass spectrometry. The physical process of separating proteins via 2DE has, until now, remained long, multi-step, labour intensive and often results in production of irreproducible data. In this presentation we will describe the design and use of a totally unique, highly flexible, reproducible and fully automated system to resolve these problems and simplify this separation process. It enables solubilised protein samples to be hydrated, focused, equilibrated on conventional IPG strips and finally resolved in the second dimension by SDS-PAGE without user intervention. It provides consistency of processing with thorough and robust methods development capability, eliminating the issues that can result from day-to-day and user-to-user variability. The innovative automation and microfluidic technologies incorporated in the system, which remove the complexities and variation associated with the traditional manual process resulting in the ability to perform protein separation in a highly reproducible fashion, will be described.

N. Laird;
Amersham Biosciences Corporation, San Francisco, CA, United States.

The introduction of immoblized pH gradient gels has significantly improved the quality, simplicity and reproducibility of 2-D electrophoresis. Extended strip lengths and new pH gradient gels have further advanced the technique to provide even higher resolution separation of proteins. IPG separations at basic pH have not always produced optimum results. In this study, proteins extracted from wheat seeds were used to represent a "difficult" sample in order to explore ways to improve IEF separation of basic proteins. Conventional rehydration solutions containing dithiotreitol resulted in significant horizontal streaking of the proteins making identification difficult. Adding a thiol blocking reagent, 2,2´-dithiodiethanol, resulted in superior separations of basic wheat storage proteins.

J. Li, Y. Zhang, T. Osterman, F. Witzmann, F. Witzmann;
IUPUI, Indianapolis, IN, United States.

In our routine proteomic efforts, collaborators and clients often desire access to the 2D images and protein identification information generated by our laboratories. The 2D analysis platform, PDQuest™, is powerful but its output and annotation tools have significant limitations. Even though a free viewer version can be downloaded, for those without a PDQuest™ background or experience, it is difficult to learn. In this poster presentation, we demonstrate a novel method to output PDQuest™ results in html format. With this method, 1) the result can be exported into two files with total size less than 100kb, 2) neither installation nor knowledge of PDQuest™ or the viewer version is needed, 3) the 2D image and highlighted protein spots can be viewed with Internet Explorer or any internet browser, and 4) information about the spot, such as SSP, name, PI, MW, GI access, and additional information from NCBI can be obtained with a mouse click.

Z. Lin, D. Crockett, M. Lim, K. Elenitoba-Johnson;
ARUP Laboratories, Salt Lake City, UT, United States.

Background: The identification of the interacting proteins within complexes is key to understanding the regulation of cell signaling pathways, and discovery of novel disease markers. Immunoprecipitation of the protein complex followed by LC-MS/MS has been used to identify targets that bind to a protein of interest. Here, we report a high-throughput NSI-LC-MS/MS method for analysis of protein/peptide complexes.
Approach: To overcome the large dwell volume from connecting lines and the injector when using an autosampler in microcapillary LC-MS/MS, we employed a valve-controlled variable flow method with a peptide trap. This method enables fast trapping of peptides from samples injected by an auto-sampler within minutes followed by a split-controlled nanoflow of acetonitrile gradient through a microcapillary C18 column to separate and elute peptides into the ion-trap MS/MS. Over 40 proteins/peptides samples at fmol levels can be analyzed continuously using the same in-house packed microcapillary C18 column.
The p38 MAP kinase is a signaling protein that is involved in regulation of apoptosis, cell cycle and tumorigenesis. Immuno-complex of monoclonal anti-p38 antibodies from a 2 mg total lysates of lymphoma cells was resolved in 1-D-PAGE gel followed by silver staining of the gel. Protein bands were excised and digested with trypsin. Peptides were extracted and analyzed by the automated LC-MS/MS method.
Results and Conclusions: From 37 excised protein bands in the 1-D-page gel, we identified more than 50 proteins, including cytoskeletal proteins, ribosomal proteins, transcription factors and KIAA potential signaling proteins. These proteins are the potential targets that may interact directly or indirectly with p38 MAP kinase. Our studies demonstrate the utility of automated nanoflow LC-MS/MS for sensitive and high-throughput analysis of protein/peptide complexes.

P. T. Jedrzejewski, G. Ganshaw, A. Gaertner;
Genencor International, Inc., Palo Alto, CA, United States.

The study of the proteome is a critical step towards the functional annotation of the genome. In contrast to genomic analyses, i.e. gene arrays, robust, sensitive, and comprehensive methods for proteome separation and identification are not yet available. Current proteomic methodologies, such as 2D-PAGE, ICAT suffer from limitations (e.g., difficult to automate, sample throughput, sensitivity, dynamic range, not comprehensive). In order to overcome these limitations, we have developed alternate methods, specifically, multidimensional chromatographic, for proteome analyses. These methods allow for high degree of automation in addition to other advantages (e.g., higher sample capacity, dynamic range).
In this presentation, we will show the application of multidimensional methods for the analysis of complex biological samples. Several multidimensional approaches were evaluated. As a model biological system, we have selected the exoproteome of a cellulase producing strain of the filamentous fungus Trichoderma reesei. We have evaluated the effectiveness of multidimensional chromatography (ion-exchange chromatography followed by reversed-phase chromatography) on the protein level. In order to improve throughput, we also investigated single dimension chromatography followed by solid phase extraction in a microtiter format. Following separation, fractions were processed by a liquid handler and analyzed on nanoLC-MS/MS system. Protein identification was performed using the database search software.
Using these approaches we successfully identified over 35 proteins from this fungal exoproteome. In less than a third of the time of a typical 2D-PAGE analysis (ca 26hrs), approximately 4-fold greater number of proteins were identified. The advantages and limitations of these techniques for the characterization of proteomes as compared to other methods, such as 2D-PAGE will be presented.

K. Meyer-Arendt^1,2, A. Mendoza², L. Aveline^1,2, N. Ahn^1,2, K. Resing²;
¹Howard Hughes Medical Institute, Boulder, CO, United States, ²University of Colorado, Boulder, CO, United States.

Expression profiling of the proteome by proteolysis of cell lysates and mass spectrometric (MS) identification of the peptides (shotgun proteomics) requires multidimensional chromatographic separation of the peptides. As multidimensional protein identification technology (MudPIT) MS analysis moves into a high-throughput domain, automated protein identification becomes increasingly important. Identification based on database search methods as done by Sequest and Mascot has become the de facto standard. Repeated analyses of peptides from a mammalian proteome show that approximately 80% of MS/MS spectra are identified incorrectly. We have implemented software which bases our peptide identification on a combination of database search programs and our own algorithms. A preprocessing step removes +2/+3 decoy dtas, recognizes implausible results by database search programs, and corrects elution time values across samples. Our software then groups and scores dtas based on spectral similarity. Finally, we combine our analyses with validated peptide identification obtained from the database search programs in order to provide evidence for protein identification. The program also supports visualization of comparative protein profiles from different experimental samples. Evaluation of our software results in fewer, but far more believable, identified proteins.

C. Shieh, R. Kiyonami, R. Shen, B. Wu;
ThermoFinnigan corp., San Jose, CA, United States.

Protein identification using peptide mapping has become an important technique for proteomic studies. Proteins were digested with a site-specific enzyme and analyzed by mass spectrometry. By comparison of the measured fragment ions of the peptides (e.g. in tandem MS or ion trap) to those predicated by the sequence, one could elucidate the corresponding protein identification through a database search. Previous analysis of proteins relies on 1D- or 2D-gel separation, followed by MS analysis. In this paper, we reported the analysis of protein mixture using an automated 2D-LC/MS/MS system equipped with a new orthogonal source in the LCQ ion trap mass spectrometer to increase the sensitivity of analysis.
Protein mixture from the human plasma is digested and loaded onto a strong cation exchange column and then gradually released to a 0.18 mm diameter C18 column by stepwise elution with salt steps of increasing molarity. After reversed phase HPLC separation, the peptides are analyzed by a LCQ^TM Deca XP mass spectrometer with a microflow electro-spray interface. Using Sequest™ software, more than 600 proteins were identified in one experiment. Compared to the current two dimensional electrophoresis method, this on-line 2D LC-MS/MS system has the advantages of higher identification capacity, higher sensitivity, higher throughput and a higher degree of automation.

R. Van Soest¹, R. Swart², M. Smoluch², G. Mitulovic², J. Chervet², H. Schwartz³, D. Patterson⁴;
¹LC Packings USA, San Francisco, CA, United States, ²LC Packings, a Dionex Company, Amsterdam, Netherlands, ³Dionex Corporation, Sunnyvale, CA, United States, ⁴Applied Biosystems, Framingham, MA, United States.

Here we present the use of a high precision X-Y-Z robotic system (Probot™) allowing direct interfacing of Capillary/Nano LC separation with MALDI/MS. By moving only the collection table of the robotic system and not the collection needle, precision of ±2 µm is achieved routinely. The use of a static needle allows for the collection of extremely small, nanoliter volumes (impossible with a moving needle due to capillary forces). For optimal crystallization the needle set-up allows for co-axial addition of matrix solution. This can be done during or prior to fraction collection and generates a perfect sweet spot. For high-throughput using MALDI/TOF/TOF MS, up to 6 high density targets can be placed on the table deck. With Nano LC (~200 nl/min), collection times as small as 2 sec are possible resulting in collection (spotting) volumes down to 7 nl. Under these conditions no chromatographic dispersion is observed resulting in the highest MALDI/MS sensitivity. Another advantage of the interfacing of Capillary/Nano LC with MALDI/MS is the storage capability of targets (up to several months, thus permitting optimization of work flows and sample re-analysis).
PROBOT is a versatile instrument, allowing the use of MALDI targets from different vendors including MALDI/TOF/TOF high-throughput analyzers. In this poster data will be shown using the PROBOT Micro Fraction Collector as the perfect interface between Capillary/Nano LC and MALDI/MS for high-throughput Proteomics.

S. J. Berger¹, K. Millea², A. Chakraborty¹, I. S. Krull², S. A. Cohen¹;
¹Waters Corporation, Milford, MA, United States, ²Department of Chemistry, Northeastern University, Boston, MA, United States.

The complexity of proteomic samples is highly conducive to the use of chromatographic prefractionation strategies at the intact protein level. A primary benefit derived from prefractionation is that simplified mixtures of analytes can be presented to a detection system (e.g. gel or LC/MS) or analysis technique (e.g. ICAT). In many cases, prefractionation also addresses dynamic range issues with proteomic samples, by resolving lower abundance from higher abundance components. When using mass spectrometry directly for intact protein analysis, another important benefit arises from grouping proteins of similar characteristics. The capabilities of ion exchange chromatography (IEX) to enrich proteins of similar charge density (basic/neutral/acidic proteins), and denaturing size exclusion chromatography (SEC) to enrich proteins of similar MW, creates protein subpopulations less likely to be obscured by competitive ionization effects encountered during ESI and MALDI MS experiments. Reversed phase separation conditions can also be optimized for the individual resolved fractions, leading to more effective LC/ESI-MS and LC/MALDI-MS analyses.
While affinity selection typically exhibits the greatest selectivity of prefractionation techniques, it is often used only to answer very focused biological questions. More typically, a scientist requires a generalized methodology to accomplish a global proteomic analysis. In this work, we will demonstrate a general approach to developing strategies for proteomic sample prefractionation, using two basic modes of biochromatography. In particular, the implementation of the AutoBlend™ (Warren et.al. (1989) Amer. Biotechnol Lab 6:120-6) technique for systematic ion exchange method development will be presented, and the capabilities of denaturing versus nondenaturing SEC will be described with regard to multiple global proteomic samples.

M. L. Shen¹, D. A. Sarracino², P. Kowalski¹, R. van Soest³;
¹Bruker Daltonics, Inc., Billerica, MA, United States, ²Matritech, Newton, MA, United States, ³LC Packings USA/Dionex, San Francisco, CA, United States.

Breast cancer is one of the leading causes in cancer deaths in women. Alternative methods that could improve the early detection and diagnosis are being explored. With recent developments in nanoscale separation technology and mass spectrometry, this task can be achieved with relative ease. We report a technology which could be employed to discover novel disease markers from human tissues, which could be extended to screen for breast cancer development at an earlier stage.
Specifically, human serum samples were collected from patients in various stages of breast cancer development. Proteins in the serum were extensively prefractionated with cation exchange chromatography to enrich and concentrate potential biomarkers which are present at very low abundance. Fractions were collected from the cation exchange chromatography, and reduced, alkylated and digested. The digestion products were then subjected to two-dimensional chromatography with an LC Packings/Dionex UltiMate™ system followed by MS/MS analyses on mass spectrometers. The data collected from the two-dimensional analyses were automatically processed, combined and searched against human protein database. The results obtained from cancerous serum samples were directly compared with those from normal samples to identify potential biomarkers.
We demonstrate here that protein markers could be identified in cancerous serum while they were not observed in normal serum. The experimental approach that we used in this pilot study has proved to be indispensable in the detection of disease markers which are present in extremely low concentration. The strategy of combining extensive prefractionation with multidimensional LC-MS/MS can dramatically enhance the detectability of biomarkers which were, otherwise, masked by hugely abundant serum proteins.

B. Glatz¹, P. Hoerth¹, M. Vollmer¹, P. Goodley²;
¹R+D, Waldbronn, Germany, ²R+D, Palo Alto, CA, United States.

The major limitations in the analysis of low abundance proteins from complex mixtures of complex biological fluids can be attributed to an over abundance of co-eluting peptides in a limited time domain. Over crowding and "stacking" of peptides inhibits the effective MS/MS identification even with data dependent MS/MS detection. This is especially true for low-level peptides, which are over whelmed by the larger abundance peptides. In order to reduce the sheer number of co-eluting peptides from the biological extracts, both gel-bands and fluids, multi-dimensional separations have been preformed using shallow gradients, longer columns. One current approach is to use salt steps to elute the peptides from strong cat ion exchanger columns (SCX) however the salt step process limits the separation and the selectivity. This work will demonstrate the effective separation, collection, and reinjection to achieve additional peptides and protein identifications. Optimized linear SCX salt gradients were used with micro-fraction collection and re-injection onto high performance reverse phase analytical columns, which significantly increased the observed number of low level peptides. These additional peptides allow for increase in the number of proteins from trypsin digest of Plastids. An about 30% increase in the number of protein identifications of the e-coli was observed even though the micro fraction and reinjection system was not completely optimized.

M. H. Simonian, E. Betgovargez;
Beckman Coulter, Inc., Fullerton, CA, United States.

The discovery stage of proteome profiling typically involves the comparison of different states of a cell or tissue such as diseased vs. normal, mature vs. immature or drug-treated vs. non-treated. This profiling traditionally has been done by two-dimensional gel electrophoresis. While this technique has successfully resolved proteomes, it is labor intensive, time consuming and, at best, semi-quantitative. This paper presents a two-dimensional chromatographic alternative for proteome profiling and fractionation. A complete system, the ProteomeLab* PF 2D, has been developed with integrated hardware, software and chemistry. The first dimension separation is done by chromatofocusing over a pH range from 8.5 to 4.0 where proteins are separated by their isoelectric points. In this dimension, fractions are collected based on pH intervals as detected by a pH monitor. The proteins are detected by UV absorbance at 280 nm. The first dimension fractions are then resolved in the second dimension by high resolution reversed phase chromatography with a gradient of trifluoroacetic acid (TFA) in acetonitrile and TFA in water. The proteins are detected by a second UV detector at 214 nm. The two dimensions are linked by a combination fraction collector/auto-injector. This allows the second dimension to commence automatically when the first dimension is complete. The proteome profiles of human plasma from a fasting and non-fasting state were compared with this new system to determine if there were any qualitative and/or quantitative differences. Data from this comparative study with this new proteome profiling and fractionation system will be presented.
*All trademarks are property of their respective owners.

L. A. Wolf^1,2, K. Jonscher¹, A. Mendoza¹, K. Meyer-Arendt^1,2, K. Pierce¹, N. Ahn^1,2, K. Resing¹;
¹University of Colorado, Boulder, CO, United States, ²Howard Hughes Medical Institute, Boulder, CO, United States.

The aim of this project is to optimize the analysis of a human proteome using multi-dimensional chromatography and mass spectrometry. Proteins are extracted from human K562 erythroleukemia cells, desalted, alkylated on Cys, and proteolyzed with trypsin. The protocol for alkylation was optimized using bovine serum albumin; the final protocol yielded <10% unalkylated peptides as observed by database searching with Cys-alkylation as a variable modification. Peptides from the K562 extract were separated by strong cation exchange chromatography (SCX), comparing several columns and run conditions. Fractions from the SCX column were collected offline and peptides were separated further by a reverse phase column, then analyzed by an LCQ classic ion trap mass spectrometer. Data dependent MS/MS were then searched with TurboSequest using an NCBI non-redundant human protein database. Turbosequest output summaries viewed in Excel, were employed to evaluate the efficiency of the cell lysate preparation, the resolution and reproducibility of our chromatography and the level of protein detection. After running a series of reverse phase LC-MS/MS using this procedure we can identified 100-500 peptides per SCX fraction with cross-correlation „ 3.3 and RSp = 1.0. In addition, we have used Mascot to carry out database searches with the same peptides; the combined information allowed identification of over 6000 peptides, which correlates to ~2000 proteins from 8mg of K562 lysate. We continue to improve our resolution and sensitivity with the goal of identifying proteins in the range of 100 copies per cell.

H. Liu, R. S. Plumb, J. W. Finch, S. A. Cohen;
Waters Corporation, Milford, MA, United States.

LC/MS/MS of global protein digests is rapidly gaining favor in proteomics for the identification of large numbers of proteins in complex mixtures. We have we evaluated the effects of the chromatographic process in a LC/MS/MS system for the analysis of protein digests using a yeast ribosomal digest and an enolase digest as model systems. We examined key chromatographic parameters including gradient duration, sample loading amounts, column dimensions, and the effect of using trapping columns. The LC-MS results were evaluated using factors such as peak capacity, the number of proteins identified and the average sequence coverage for the identified proteins. In general the studies showed (1) peak capacity of the system was significantly increased by lengthening the gradient duration but there were limits to the benefits of longer analysis times, (2) trapping columns yielded somewhat broader peaks than direct loading, although there was a significant savings in sample analysis time and (3) larger sample amounts led to greater sequence coverage and more identified proteins, although again, this effect was not unlimited. The effect of column size was more complicated to interpret, with the quality of the analysis depending on the interaction between the column size and the sample amount loaded on a particular column. We will discuss how these results allow chromatographic system optimization depending on the analysis needs.

D. B. Wall¹, C. Stumpf¹, J. W. Finch¹, R. Plumb¹, S. A. Cohen¹, J. N. Haselden²;
¹Waters Corporation, Milford, MA, United States, ²GlaxoSmithKline, Ware, United Kingdom.

Plasma contains thousands of endogenous peptides, some of which may be useful biomarkers for detection of a disease state or toxicological drug response. Biomarkers are extremely important in pharmaceutical discovery and development, as they aid HTS and clinical efficacy evaluation. However, biomarker detection generally requires that the sample components are separated, detected and then plotted as a 3-D image and/or analyzed by statistical methods. During the last 5 years a new approach using proton NMR and principal components analysis has been pioneered; it provides a rapid method for the detection and identification of small molecule biomarker, this work has recently been extended to encompass LC/MS data. Traditionally 2-D gels are used to separate and detect proteins for biomarker identification; unfortunately for peptides the 2-D gel method is not effective and so alternative separation and detection methods need to be employed. In this work LC-MS methodology is applied to the endogenous peptide sample of rat plasma to separate and detect potential peptide biomarkers resulting from the administration of a candidate pharmaceutical. Ten untreated and ten treated rat plasma samples were analyzed by LC-MS and the data was then subjected to principal components analysis to detect significant differences between the two animal groups. The peptide peaks responsible for the group differentiation were then identified LC-MS/MS analysis.

S. Hahner¹, A. Resemann¹, S. Liedtke², D. Suckau¹, J. Stoerker³, M. Lubeck¹;
¹Bruker Daltonik GmbH, Bremen, Germany, ²LC Packings - A Dionex Company, Amsterdam, Netherlands, ³Bruker Daltonics, Billerica, MA, United States.

A novel platform has been developed for coupling of liquid chromatography (LC) with MALDI-TOF/TOF mass spectrometry. Peptides from complex protein mixtures were separated and directly eluted onto a 384er microtitre format MALDI target off-line for subsequent MS/MS analysis. The use of hydrophobic coated MALDI targets on which an array of hydrophilic spots of accurate size (600 µm) is prepared allow the precise fixation of the LC separated peptides onto discrete spots. No additional liquid flow or advanced techniques for sample deposition were required for deposition of the LC trace directly onto the hydrophilic spots even for LC flows in the nano range (200 nl/min). A microfraction collector designed for on-line collection of discrete LC fractions enables the preparation of eluent directly on the MALDI target by moving the collection table holding the target. The analysis of the fractionated peptides was performed fully automatically using the novel MALDI-TOF/TOF mass spectrometer capable for MS and MS/MS analysis.

M. Vollmer, E. Naegele, P. Hoerth;
Agilent Technologies, Waldbronn, Germany.

Depending on the carbohydrate which is supplied as sole source of carbon and electron donor, different metabolic enzymes are induced for degradation in E. coli. Using two-dimensional nano LC/MS/MS, we analysed the proteome of E. coli cells grown in a minimal medium containing different carbohydrates.
The total protein content of E. coli cells was subjected to tryptic digestion and further separated by 2D nano LC followed by a nano electrospray MS and MS/MS. Complex peptide maps originating from different culture conditions were constructed by comparative database analysis. Significant differences in expression patterns are presented and further discussed.

M. Lubeck, U. Schweiger-Hufnagel, C. Baessmann;
Bruker Daltonik GmbH, Bremen, Germany.

Two dimensional polyacrylamid gel electrophoresis (2D-PAGE) followed by mass spectrometry is the most widely used method for protein separation, quantification and identification. However this approach is time-consuming, and membrane proteins as well as proteins with extremes in pI and molecular weight are rarely seen in 2D-Gel studies. As an alternative two-dimensional liquid chromatography (2D-LC) prior to mass spectrometry is a promising high resolution separation technique. As a model system of limited complexity (<200 proteins) ribosomal proteins were investigated in respect of handling and analyzing protein mixtures efficiently. For the 1^st dimension of the 2D chromatography the tryptic peptides were bound to a strong cation exchange column. By injecting salt solutions of increasing concentration, peptide fractions were eluted (microflow). For the 2^nd dimension the peptides were directly used without collecting fractions. They were further separated by reversed phase chromatography (nanoflow). An esquire₃₀₀₀^plus Ion Trap Mass Spectrometer (Bruker Daltonik GmbH) was used to acquire MS and MS(n) data of the complete peptide mixture automatically. Strategies will be shown how to reduce protein losses during identification and how reliable two only slightly different proteome samples can be compared using 2D LC.

M. Vollmer, P. Hoerth, E. Naegele;
Agilent Technologies, Waldbronn, Germany.

Exposure of E. coli to heat shock results in a change of protein expression pattern. Heat shock proteins (HSPs) act as molecular chaperones and proteases to prevent cell damage under harmful conditions.
Using a fully automated 2D nano LC/MS System, we constructed comprehensive peptide maps from E.coli grown under heat shock vs. non-heat shock conditions. Tryptically digested bacterial samples were separated in a two dimensional manner by combining strong cation exchange and reversed phased chromatography. Peptides were eluted online into an ion trap MS equipped with a nanoelectrospray ionsource and further analyzed by fully automated intelligent data dependent MS/MS fragmentation. Bacterial proteins originating from the different samples were determined by a MASCOT search using the Swiss Prot Database.
In our study, we present data where we were able to detect several hundred different proteins significantly expressed under heat shock and non-heat shock conditions, respectively. Differences in expression patterns are shown and further discussed.

I. Jordon-Thaden¹, R. Hilger¹, T. Elthon², R. Cerny¹;
¹Dept of Chemistry, Lincoln, NE, United States, ²School of Biological Sciences, University of Nebraska, Lincoln, NE, United States.

A library of monoclonal antibodies has been developed at random to Zea mays mitochondrial proteins. We are identifying the proteins in the library using tandem mass spectrometry. The antibody identity can be associated with a function to explore biological processes occuring in plant mitochondria. Once these proteins are identified, the antibodies can be used as tools to probe mitochondrial functions. We have developed a reproducible 2D gel map of the proteins expressed in maize mitochondria. Protein identification is based on nano HPLC MS/MS methods using a Q-TOF Ultima hybrid tandem mass spectrometer. The maize proteome is not well defined within existing public domain databases. Information can be gained from other plant species, such as Arabidopsis and Oryza, that may have proteins with homology to Zea mays. This does not, however, allow one to identify the actual gene within maize. If the complete gene that codes the protein in question is not in current databases, we attempt to construct a putative gene from ESTs. We have utilized both the identified homology as well as database searching of ESTs to identify the specific gene and therefore the full length protein sequence for many of the antibody targets. We have selected 33 proteins associated with the antibody library; 85% of these have been positively identified in replicate experiments. Our work complements that previously performed on Arabidopsis mitochondrial proteins (1,2), and is yielding new insights into plant mitochondrial biochemistry. 1. Kruft, V. et al. 2001 Plant Physiology, Vol. 127, pp. 1694-1710. 2. Millar, A.H. et al. 2001, Plant Physiology, Vol. 127, pp. 1711-1727.

R. Van Soest¹, M. Smoluch², G. Mitulovic², R. Swart², J. Chervet², H. Schwartz³;
¹LC Packings, a Dionex Company, San Francisco, CA, United States, ²LC Packings, a Dionex Company, Amsterdam, Netherlands, ³Dionex Corporation, Sunnyvale, CA, United States.

2-D Capillary LC has become an attractive alternative to 2-D PAGE for the analysis of protein samples. For complex samples such as cell lysates, both techniques often are combined. 2-D gel spots are excised, digested, and the extracted peptides analyzed by 2-D Capillary/Nano LC/MS/MS. The 2-D Capillary LC typically involves strong cation-exchange chromatography (SCX) with salt plugs, followed by reversed phase (RP) chromatography. A disadvantage of this approach is the limited selectivity achievable, resulting in the spreading of a particular peptide over different salt fractions (carry-over).
To overcome this limitation, a continuous gradient can be performed with the SCX column (as is done with the RP column). This requires a Capillary LC system with dual gradient pumping capabilities. In this presentation we will show the use of a dual-gradient Capillary/Nano LC system (UltiMate, LC Packings, a Dionex Company) for comprehensive 2-D LC/MS/MS of complex protein samples. Comparison with the former 2-D approach using salt plugs will be given. Applied to complex proteomics samples such as bacterial lysates, this technique yields much better sequence coverage, identification of more proteins, and a drastically reduced carry-over of high-abundant peptides. Experimental conditions for optimal use of the dual-gradient approach will be presented.

K. J. Biederman, M. Summers, R. D. Edmondson, M. R. Pisano, B. L. Allen;
PRS, Inc, Ann Arbor, MI, United States.

The comprehensive identification of proteins from serum is challenging due to the wide dynamic range of proteins present in the sample. We have previously shown that this dynamic range can be reduced, and the number of proteins identified by 2D/LC/MS/MS increased, by performing one of three albumin depletion strategies: cibachron blue (CB), Con A, or immunoaffinity fractionation (IA). The most significant gains in the number of proteins identified were achieved following CB or IA fractionation. Additionally, several proteins unique to each method were identified versus a non-depleted control.
In this study the CB and IA samples were analyzed by SDS-PAGE/LC/MS/MS and compared to whole serum. Data from both studies will be compiled and the depletion strategies will be evaluated in the context of the total number of proteins identified as well as proteins unique to each sample.

E. Naegele, M. Vollmer, P. Hoerth;
Agilent Technologies, Waldbronn, Germany.

Today, two-dimensional LC/MS methods play a key role in proteomics research for the separation of complex peptide mixtures from digested proteins. This is mainly due to the fact that 2D LC has been found to overcome some of the limitations inherent to 2D gel ectrophoresis.
In our two-dimensional chromatographic approach, we combined strong cation exchange (SCX) chromatography as a first dimension with reversed phase (RP) separation in the second dimension. The eluted peptides were sprayed directly into an ion trap MS equipped with a nanoelectrospray ionsource and were subsequently analysed by fully automated intelligent data dependent MS/MS fragmentation. Our data were evaluated by performing a MASCOT search of the Swiss Prot Database.
We first successfully demonstrated the excellent performance of our developed methodology by identifying the complete set of proteins from an artificial medium-complex protein mixture digested with trypsin. In order to show that it can also cope with a "real-life" sample, we characterized the proteome of a cell lysate from E. coli.

K. R. Jonscher, L. Aveline, A. Mendoza, K. Meyer-Arendt, K. Resing, N. Ahn;
University of Colorado, Boulder, CO, United States.

Multi-dimensional chromatography is rapidly becoming a major analytical tool in the mass spectral analysis of cellular protein content. Separation of peptides by charge, typically using strong cation exchange, is followed by separation of peptides by hydrophobicity using reverse-phase HPLC. The primary benefit of performing these orthogonal separations is simplification of the peptide mixtures so that the combination of HPLC elution times and mass analysis times can be most efficiently combined to analyze co-eluting molecules. Resolution of the reverse-phase column is therefore one of the keys to detecting peptides generated from low-abundance proteins that co-elute with peptides from higher copy number proteins.
In this work we compared the resolution obtained using Phenomenex Columbus resin with 5 micron particles and Phenomenex Jupiter resin with 10 micron particles. A tryptic digest of myoglobin was used to optimize conditions, then a tryptic digest of a whole cell lysate of K562 erythro-leukemia cells was separated by strong cation exchange. Aliquots of one fraction were loaded onto each column and analyzed using automated LC-MS/MS on an LCQ Classic ion trap. Following processing with TurboSequest, high-scoring peptides were compared between the different conditions. As expected, by increasing chromatographic resolution, hence decreasing the number of peptides presented to the mass spectrometer at a given time, more high Xcorr peptides could be observed.

J. P. Murphy, III, G. A. Valaskovic;
New Objective, Woburn, MA, United States.

A widely used method for protein identification incorporates two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) followed by nanoelectrospray ionization mass spectrometry (nESI-MS). Nanospray in combination with nanobore chromatography (75 µm i.d. columns) provides typical limits of detection in the subfemtomole to attomole range whem combined with ion trap mass spectrometry.
Typically, an in-gel digested protein sample is evaporated to dryness and resuspended in several microliters of solvent. Sample inejction is typically carried out by one of two methods; direct on-column injection using a pressure bomb, or on-line sample pre-concentration in a sample trap cartridge. The time intensive bomb injection method is highly sensitive, minimizing sample handling and maximizing sample utilization. Low back pressure trapping cartridges permit higher loading rates and compatibility with autosampler methods. Trap cartirdges also yield guard-column functionality, as in-line enrichment and desalting of samples increases nanobore column lifetime.
A study was performed on there nanobore injection methods to determine the advantages and pitfalls for each method. Tryptic digests were loaded onto PicoFrit columns (New Objective, Inc). A PicoFrit column combines a typical nanobore column with a fritted electrospray emitter to eliminate post column losses and band broadening. This isolated any change in sensitivity to the injection method. Analysis were performed by an LC/MS system composed of an 1100 Cap LC pump (Agilent) interfaced with a LCQ Deca (Thermo Finnigan) with a PicoView ion source (New Objective). Figures of merit include sequence coverage, analysis time, limit of detection and column lifetime.

J. I. Langridge¹, S. Da Cruz², C. Hughes¹, A. Millar¹, J. Martinou²;
¹Micromass MS Technologies, Manchester, United Kingdom, ²Département de Biologie Cellulaire, Genève University, Switzerland.

Mitochondria have received considerable attention over the past few years because of their physiological role in cell homeostasis and in apoptosis. Moreover, they have been implicated in the pathogeny of many diseases including Alzheimer’s disease, Parkinson’s disease, amyothrophic lateral sclerosis, stroke, myocardial infarction, optical atrophy as well as in ageing.
The inner mitochondrial membrane of mitochondria is of particular interest since it contains the respiratory chain and various ion channels, including calcium, potassium and proton channels. Many studies have demonstrated a key role of these channels in the control of mitochondrial function and respiration. Moreover, several data indicate that potassium-ATP-dependent channels play a role in cell injury following ischemia, in particular in the heart. The identification of a majority of inner mitochondrial proteins and the understanding of their function would therefore be invaluable since it would allow a better understanding of the mechanism underlying various pathologies.
The success to obtaining the most complete and accurate proteomic profile possible will depend on the quality of the starting material (purification of inner mitochondrial membranes), on the MS techniques used for peptide analysis and of the performance of the bioinformatic tools used to analyse the data produced.
In this study inner mitochondrial membranes (IMM) from adult mouse liver were prepared allowing us to obtain highly enriched membranes. IMM was subsequently digested, in-solution, with trypsin and analysed by state of the art LC-MS/MS on a Q-Tof mass spectrometer. Results will be presented and discussed from the analysis of the IMM preparation.

E. J. Unsworth, Jr., D. M. Maynard, S. P. Markey;
NIH/NIMH, Bethesda, MD, United States.

Complex biological matrices, such as those deriving from whole cell lysates, represent a formidable proteomic challenge. Those proteins present in high abundance are readily identified while the more important, low abundance proteins go largely undetected using conventional methods. Efforts to ameliorate the dilemma have included subcellular fractionation, two-dimensional electrophoresis, two-dimensional HPLC, nanoLC and numerous combinations of those mentioned.
Recently, Zuo and Speicher adapted the work of Righetti et al. to develop a scaled-down version of Righetti’s multicompartment, solution IEF procedure. The advantage of the Zuo technique is that it utilizes low volume chambers, 100 to 500 microliters, separated by Immobiline II-polyacrylamide pI discs. The low chamber volumes required are more manageable for subsequent procedures. However, all electrophoretic procedures utilize detergents for maintenance of protein solubility when at their isoelectric point. Detergents, whether ionic or nonionic, tend to foul HPLC columns and interfere with MS detection and subsequent database searching.
This work details a microsolution IEF procedure lacking detergent for the prefractionation of a yeast lysate. No protein precipitation was observed following protein focusing in the absence of detergent. A solution-based double proteolytic digestion was performed on the fractions as described by Yates et al. followed by two-dimensional HPLC/MS/MS. The spectra generated were searched for known proteins using the Mascot database search engine. The results demonstrate the advantage of reducing the complexity of the sample matrix prior to LC/LC/MS/MS analysis.

A. F. Hühmer¹, R. G. Biringer¹, H. Amato¹, M. Harrington²;
¹Thermo Finnigan, San Jose, CA, United States, ²Huntington Research Institutes, Pasadena, CA, United States.

High-resolution and high-throughput techniques such as 2-D LC coupled to mass spectrometry (MS) are currently explored for large scale proteomic expression studies of heterogeneous protein mixtures. Monitoring protein expression levels in biological fluids, such as cerebrospinal fluid (CSF) is an important step towards understanding disease states associated with these biological fluids. We apply two quantitation strategies to measure protein levels in CSF. In the first method CSF proteins are covalently tagged with an ICAT reagent and proteolyzed followed by identification and quantitation by 2-D LC/MS/MS. In the second method CSF proteins are proteolyzed, identified and quantitated by 2-D LC/MS/MS using peak areas from selected ion chromatograms. Both methods are contrasted with respect to their ability to quantitate proteins accurately and reproducibly. Systematic identification and quantitation of proteins using an ICAT strategy allows the quantitation of a large number of peptides from CSF, but protein information obtained in those experiment is restricted to cysteine containing proteins. For some proteins from CSF the reduced sequence coverage inherent to the ICAT labeling strategy affects the confidence level of positive protein identification considerably. Quantitation of proteins from CSF using selected ion chromatograms is not biased towards cysteine containing proteins and allows quantitation of fewer proteins from CSF more reproducibly and with higher confidence.

R. G. Biringer, H. Amato, A. F. Hühmer;
Thermo Finnigan, San Jose, CA, United States.

In an effort to define protocols for quantitating proteins in complex biological mixtures, both cleavable and non-cleavable ICAT reagents were employed to evaluate a mixture of human transferrin, human albumin, and human α-macroglobulin. Both reagents were investigated with respect to their ability to quantitate and identify peptides reproducibly from this simple protein mixture. Chromatographic behavior of the product peptides and the particular modified peptides identified were reagent dependent. The non-cleavable ICAT reagent produced modified peptides for which the chromatographic behavior is apparently dominated by the large attached tag. The MS² spectra for these peptides were complex and generally required higher fragmentation energy to obtain searchable MS² spectra. On the other hand, the cleavable ICAT reagent produced modified peptides with chromatographic behavior that is more dependent on the sequence than the tag and for which the MS² spectra were generally high quality. For these reasons, significantly more peptides were identified with the cleavable reagent than the non-cleavable reagent. Comparison of the datasets from both ICAT reagents indicate that only one peptide for human transferrin was common to both datasets under the conditions employed. We conclude that analysis conditions for cleavable ICAT reagents is significantly different from non-cleavable ICAT reagent,which allow for consistent and reproducible protein identification and quantitation.

I. Mohtashemi, M. Chen, D. Arnott;
Genentech Inc, South San Francisco, CA, United States.

Digestion of unfractionated protein mixtures followed by LC-MS/MS and database searching is an increasingly common approach to identifying proteins in complex mixtures. Incorporation of stable isotopes into the proteins or peptides has extended this approach to allow quantitative comparisons between samples, e.g. response to external stimuli or between normal and disease states. In the isotope-coded affinity tag experiment, for example, cysteine residues are labeled with a reagent incorporating zero or eight deuterium atoms. Samples to be compared are labeled with the light and heavy forms of the label, mixed, and analyzed by mass spectrometry; corresponding peptides are recognized by their 8 Da mass shifts, and peak heights or areas reflect relative abundances. More general strategies involve metabolic incorporation of labeled amino acids, digestion in 18O-enriched water, or acylation of primary amines. In the latter case tryptic peptides terminating in lysine are labeled twice, whereas those terminating in arginine are labeled once, complicating interpretation of results.
We describe a strategy to selectively label the n-termini of tryptic peptides for the quantitative comparison of proteins in biological mixtures. In this procedure, lysine residues are guanidinated, leaving the n-terminal amine free for subsequent reductive methylation. Use of deuterated formaldehyde in this reaction produces a four Da mass shift relative to the undeuterated version. A six Da shift is accomplished using D213C-formaldehyde. These reactions are rapid, efficient, and highly specific. Labeled peptides produce sequence-specific fragments in MS/MS, and are amenable to database searching. Furthermore, all n-terminal product ions contain the isotopic label, facilitating spectral interpretation. This method has been used to quantify selected protein in complex mixtures such as plasma membrane preparations of mammalian cells

C. E. Haydon^1,2, P. A. Eyers^3,4, J. L. Maller^3,4, K. A. Resing¹, N. G. Ahn^1,2;
¹University of Colorado, Boulder, CO, United States, ²Howard Hughes Medical Institute, Boulder, CO, United States, ³University of Colorado Health Sciences Centre, Denver, CO, United States, ⁴Howard Hughes Medical Institute, Denver, CO, United States.

Protein phosphorylation has evolved as a highly dynamic mechanism for regulating protein function. Mass spectrometric analysis of proteolytically derived phosphopeptides has developed into a widespread technique for the identification of phosphorylated amino acids. However, the identification of phosphopeptides in a complex peptide mixture is often suppressed due to the electronegativity of the phosphoryl groups.
Here we introduce a simple modification to an existing phosphopeptide enrichment method that allows enhanced identification of phosphopeptides. We investigated the pH-dependent binding of phosphate groups in synthetic peptides to an immobilized metal-affinity chromatography (IMAC) column. In order to determine the efficiency of phosphopeptide binding by our improved IMAC methodology, we chose to evaluate Xenopus Aurora A, a Ser/Thr protein kinase that is active and hyperphosphorylated when overexpressed and purified from E.coli. Aurora A has been implicated in early mitotic events such as the centrosome cycle and multiple phosphorylated forms are generated during Xenopus oocyte meiotic maturation in response to progesterone. We therefore decided to use this method to identify novel sites of Aurora A phosphorylation.
We demonstrate that IMAC performed at near physiological pH levels significantly enhances binding of phosphopeptides and improves phosphopeptide enrichment from a simple peptide mixture. Subsequent identification of the phosphorylation sites in these peptides by LC-ESMS is therefore less problematic.
Furthermore, we have extended our methodology development for phosphopeptide enrichment from simple mixtures in order to analyze the mammalian phosphoproteome.

A. B. Chakraborty, H. Liu, S. A. Cohen, S. J. Berger;
Waters Corporation, Milford, MA, United States.

The growing interest in global proteomics has stimulated the effort to develop better analytical approaches capable of separating, identifying, and characterizing proteins in complex mixtures. Effective fractionation and separation methods have become a critical element in this effort. While the most traditional approach for the analysis of complex protein mixtures uses 2-D gel methodologies, the inherent shortcomings of this approach have triggered the development of non-gel based alternatives. The recent interest in "top-down" methodologies have focused on obtaining comprehensive intact protein MS from complex mixtures, rather than identifying proteins by "bottom-up" approaches using one or more signature peptides. We have shown the synergistic power of combining both "top-down" and "bottom-up" approaches, using multidimensional chromatography coupled with both online MS analysis and fraction collection, and applied this strategy to yeast ribosomal protein samples. In work presented here, we have applied multidimensional separations (SCX/RP) coupled online with ESI-TOF-MS to identify and characterize intact proteins from global soluble protein fractions obtained from E. coli. Additionally, we have implemented rapid automated processing (trypsin digestion and MALDI plate spotting) of collected fractions by incorporating a mass spectrometry compatible protein solubilization factor (RapiGest™SF) into a robotic automation procedure. This permitted us to confirm protein identity and putative modifications without the need for post-digestion sample cleanup.

T. N. Krogh¹, T. G. Klenø², P. F. Nielsen¹;
¹Protein Enginering, Novo Nordisk A/S, Bagsværd, Denmark, ²Applied Trinomics, Novo Nordisk A/S, Måløv, Denmark.

This study is designed to evaluate the potentials of differential proteome display using either cleavable ICAT (isotope-coded affinity tag) reagent labelling combined with MALDI-TOF/TOF MS (matrix assisted laser desorption mass spectrometry) or 2-D DIGE (Difference Gel Electrophoresis) with flourescence labelling and image analysis.
The experimental set-up uses rat liver protein extract from Sprague Dawley rats. The protein extract from rats treated with hydrazine (90 mg/kg b.wt., 72 h) is labelled with heavy ICAT reagent and mixed with liver extracts from control rats (vehicle, 72 h) labeled with light ICAT reagent. The pooled protein extract were separated by 2-DE (12% homogenous gels). Prior to spot picking, proteins are visualized by Sypro Ruby stain. The gel spots are in-gel digested with trypsin using the Proteineer system for multiply parallel digestions and proteins identified by combining MALDI-MS and MS/MS spectra.
In a parallel set-up samples are processed using 2-D DIGE fluorescence based imaging and quantitation. In order to compensate for increased MW (and possible changed 2-D position) in the ICAT treated samples, the protein extract used for DIGE experiments is alkylated using un-cleavable ICAT reagent instead of a conventional alkylating reagent. Thus making it easier to "align" 2-D gels from the two different sets of experiments and ensuring comparison of the same spots.
Finally 10-20 spots from the 2-D gels are selected for a direct comparison.

B. J. Boucher, M. Minkoff, S. B. Daniels, S. W. Yuen, P. Yuan, A. J. Tomlinson;
Applied Biosystems, Framingham, MA, United States.

A comprehensive understanding of biological systems at the molecular level is required by many disciplines. Deciphering aberrant protein expression or interactions of diseased states is central to the development of next generation therapeutics. Likewise, understanding the biology of bacteria may lead to novel antibiotics, and the biochemistry of plants will aid the development of higher yielding, disease resistant crops. Protein identification and quantification is a necessary starting point for these investigations. Simplification of protein digest mixtures is essential if more than structural and house-keeping genes are to be identified. Use of isotope coded affinity tag reagents conveniently and specifically isolates cysteine containing peptides from complex mixtures, thereby simplifying peptide mixtures and extending the dynamic range of mass spectrometric based protein identification methods. In the current presentation we will describe application of cleavable ICAT™ reagents that incorporate an acid cleavable linker for protein identification and quantification in various biological systems.

J. L. Bundy¹, T. Freeman¹, N. VerBerkmoes², J. L. Stephenson, Jr¹;
¹Research Triangle Institute, Research Triangle Park, NC, United States, ²Oak Ridge National Laboratory, Oak Ridge, TN, United States.

Rapid large-scale protein identification via the direct analysis of complex peptide mixtures by LC-MS based methods for is of considerable current interest to the proteomics community. One of the most popular algorithms for the analysis of large tandem mass spectrometry datasets is SEQUEST, developed by Yates and co-workers. A typical 1-D or 2-D LC-MS run can generate thousands of tandem mass spectra, presenting a considerable data reduction challenge.
One software tool that has been developed to organize and filter large SEQUEST datasets is DTASelect, developed by David Tabb of the Yates group. The program in its current form is a very powerful tool; however, in its current form it is run from the command line or a manually generated batch file, which may make it inaccessible to some potential users. We have written a graphical user interface for this program in Visual Basic that allows parameters from DTAselect to be easily set by the user and passed to the program via an automatically generated batch file. This interface greatly simplifies the use of the DTASelect program for the novice user, and allows multiple filter setttings to be easily implemented on a given dataset.

K. Ashman¹, P. Metalnikov¹, C. Lock², A. Lau¹, C. Lau¹;
¹Samuel Lunenfeld Institute, Toronto, ON, Canada, ²MDS Sciex, Concord, ON, Canada.

Introduction:
The advent of the field of proteomics has resulted in unprecedented interest in analysing protein complexes by mass spectrometry. To meet this demand there is an urgent need to develop software capable of automating the analytical process.
Methods and Instrumentation:
A new suite of control scripts has been designed and implemented on a Sciex Qstar hybrid mass spectrometer, equipped with a research grade orthogonal MALDI ion source. The scripts initially instruct the mass spectrometer to perform a TOF scan to identify peptides and then perform MS/MS on the highest peptide peak. The spectrum is sent to a search engine such as MASCOT to identify the protein. The sequence of the identified molecule is converted to tryptic peptides and the original MS spectrum checked for matching masses. These peaks are now subjected to MS/MS and database searching to confirm the peptide originates from the expected protein. If two additional peptides from the identified protein are found the program generates an exclusion list and moves on to peptide masses that have not been excluded. This process is repeated until the mass spectrum has been exhaustively analysed. Finally a report is generated showing a summary of the protein(s) identified and a log of the analytical path. The software has been used to successfully identify up to 6 proteins in a single MALDI spot in a fully automated fashion, with no operator intervention required.

A. M. Mendoza¹, T. Cheung¹, K. Meyer-Arendt^2,1, L. Wolf^2,1, N. Ahn^2,1, K. Resing¹;
¹University of Colorado, Boulder, CO, United States, ²Howard Hughes Medical Institute, Boulder, CO, United States.

Data sets generated from a shotgun proteomics project are enormously complex. Shotgun proteomics samples are proteolytic digests of full cell or fractionated lysates containing several thousand proteins. They require separation of peptides, typically by strong cation exchange followed by reverse phase LC coupled to a mass spectrometer for analysis (multi-dimensional protein identification technology: MudPIT). Samples may resolve into more than a hundred fractions, each of which may require 3-10 LC-MS/MS analyses and database correlation. Output from the database search algorithms must be captured and correlated in a centralized data repository. Commercially available MS analysis tools are not capable of integrating the vast amounts of data generated by shotgun proteomics into a dynamic entity for high throughput data processing. This creates the need for data management systems and the development of database technologies. By assessing requirements, processes, and data flow it is possible to construct a conceptual model that provides a blueprint for process optimization and database system design. We consolidated information from multiple LC-MS/MS analyses, as well as Sequest and Mascot database searches, into a central data repository by developing parsers (Perl, Java) that interact with the output files and extract result information into Oracle relational tables. This process of transforming and organizing captured data into a relational database not only optimized the laboratory’s current processing capabilities, but also allowed rapid prototyping and implementation of new technologies. By representing data in relational database tables, comparison to other types of data, such as DNA microarray and dynamic querying provides high-level data mining capabilities across multiple aspects of the project.

J. M. Asara, J. Adamec, X. Zhang, T. R. Londo, M. Hincapie, S. Naylor;
Beyond Genomics, Inc., Waltham, MA, United States.

Quantitation of differentially expressed proteins continues to be a challenge in differential protein expression profiling between healthy and diseased states. It is a critical step involved in the discovery of diagnostic markers, biological pathways, and drugs. Proteomic approaches for differential profiling have been shown using isotopic labeling techniques. However, the need for improved strategies continues to be an important goal in quantitative proteomics. We have applied a proprietary technology, Global Internal Standard Technology (GIST) to our biomarker discovery program. GIST is a peptide-level, isotopic labeling technique that has been shown to be effective by acetylating peptides at primary amine groups and lysine residues. Serum and plasma are ideal sources for biomarker discovery due to accessibility, and we have demonstrated that GIST can be used to detect changes in serum and plasma protein levels by LC/MS on a Micromass QTOF mass spectrometer. Peptides digested with trypsin from experimental and control samples were isotopically labeled with acetate(²H)- or propionate(¹³C)- based reagents and equal aliquots of the respective experimental and control solutions were combined. Ratios of the resulting light and heavy peptide pairs were determined by GISTool, an in-house software package. Peptides demonstrating a significant change in signal were then targeted for MS/MS or identified from a data dependent LC/MS/MS experiment. Labeling efficiency is shown to be greater than 95 %. ¹³C-based reagents have an advantage since no chromatographic shift is observed, as is the case for deuterium-labeled peptides. This allows a single data dependent LC/MS/MS experiment to be used for both quantitation and sequence determination since the real-time ratio of the isotopic pairs is accurate.

W. Chen¹, P. J. Lee¹, J. W. Finch¹, E. S. Bouvier¹, J. C. Gebler¹, J. Brown², E. Claude², D. Gostick², J. Langridge²;
¹Waters Corporation, Milford, MA, United States, ²Waters Corporation, Manchester, United Kingdom.

The MALDI mass analysis has a reputation of relative tolerance to contaminants in biological samples. It has been proved that desalting of applied sample can greatly improve analysis results, especially when sample concentration is low. Common techniques to remove contaminants in very small sample volume, such as ZipTips and microcolumns, are often time consuming and give low sample recovery. We have developed a novel MALDI target plate that possesses both concentrating and desalting capabilities as a new sample preparation technology for MALDI MS analysis. When a large volume of diluted sample is applied to the target plate, it focuses, during drying down, to a confined region having peptide binding property. The region is then subsequently washed to remove the contaminates (for example, salts, buffers, detergents, etc.) that are contained in sample and may interfere with MALDI analysis.
In-gel digests from faint silver-stained spots were evaluated using the MALDI target, and the results are compared with other common purification methods. Results have shown that excellent sensitivity can be obtained using the target plate (subfemotomole), and desalting efficiency is at least as efficient as other commonly used methods. Moreover, examples of micro-scale chemical reactions performed directly on the target demonstrate additional benefits of the technology.

F. P. Pratt¹, A. K. Sater², F. P. Pratt², J. Z. DeGnore¹, K. Dong¹, M. Chen¹, A. Krowczynska¹, F. Deng², H. Gohil²;
¹Applied Biosystems, Framingham, MA, United States, ²Dept. of Biology and Biochemistry University of Houston, Houston, TX, United States.

Peptide mass fingerprinting (PMF) is a powerful mass spectral approach to identify proteins from 2D gels. Coupling PMF with MS/MS on a 4700 Proteomic Analyzer with TOF-TOF™ Optics enables higher confidence in the PMF results and affords identification of low level protein bands that yield too few peptides for PMF identification. Xenopus laevis is a key model system for studying the mechanisms underlying the choice between neural and epidermal fate, which embryonic ectoderm cells make during gastrulation. We have initiated a proteomic analysis comparing cells committed to forming neural ectoderm (st. 11 neural plate), cells committed to forming epidermis (st 11 animal cap), and cells isolated before this choice is made (st. 9). These cells are identified by their position in the developing embryo and are collected by micro dissection. The cells are lysed and the proteins were extracted. Two-dimensional gel electrophoresis was chosen to separate proteins because of its high resolving capability. Each Coomassie Blue stained spot of interest was cut out of the gel then digested with trypsin. The tryptic digests were spotted to a MALDI plate and Time-of-Flight (TOF) spectra were generated. Various databases were searched including dbEST database for protein identification.
Protein profiles are compared to study the protein expression patterns in frog early developmental stages. The goal of the study is to follow initial protein identification with comparisons of expression levels of identified proteins across the different tissue samples, using 2-D image comparison software and a variety of other Proteomic techniques.

F. Maroto¹, M. Scigelova², C. Rudd¹, A. Huhmer¹, R. Biringer¹, J. Vazquez³;
¹ThermoFinnigan, San Jose, CA, United States, ²ThermoFinnigan, UK, United Kingdom, ³Centro de Biologia Molecular Severo Ochoa, Madrid, Spain.

Abstract: A rigorous probabilistic and statistic model for evaluation of protein identifications obtained from database search of MS/MS spectra is presented. The probability of misidentification is derived and compared with other proposed methods of validation. The model is applied to data acquired using two-dimensional chromatography techniques for large-scale protein identification.
Methods: A complex mixtures of yeast and human proteins were analyzed using a commercial two-dimensional chromatography set up. The confidences for the protein and peptide identifications were evaluated using a computer program integrated into the data base search engine.
Results: We show how the methods for validating protein identifications that have been proposed previously fail when databases are small or when the volume of data analyzed is very large. We show how the method proposed here remains valid, and we discus the limits of protein identification using database search.
Conclusions: The confidence for protein and peptide identifications using database search of MS/MS spectra can be calculated. Under some circumstances the confidence will not be high enough. In that cases other complementary methods such as de novo sequencing verification become necessary.

A. Wallace, M. A. Ritchie, C. Jones, S. Leicester, J. Langridge;
Micromass MS Technologies, Manchester, United Kingdom.

Identification of proteins from complex mixtures can be performed by tryptic digestion on the whole protein mixture and analysis by electrospray LC-MS/MS. This approach is limited due to peptide co-elution and the preferential selection of abundant peptides for MS/MS, resulting in poor dynamic range and redundant data from identified proteins. If protein identification is obtained from a databank search of an MS/MS spectrum, it is potentially valuable to exclude the rest of the theoretical tryptic peptides to "mine" deeper into the protein complex being studied.
We have introduced a new protein databank search engine capable of matching a tryptic peptide from the Swissprot/TrEMBL databank to an MS/MS spectrum in one second. Using this search engine we are able to generate dynamic tryptic peptide exclude lists, based upon the theoretical tryptic peptides from the identified protein, which can be passed to the acquisition software of our Q-Tof mass spectrometer in real time. Thus, we are able to automatically steer the Q-Tof, during acquisition, to prevent selection of peaks that belong to a protein already identified for MS/MS. A new isotope detection routine has been implemented such that the high mass accuracy obtained from the Q-Tof is used to allow a tight mass tolerance on the peptide exclude list. This results in extremely specific exclusion without total exclusion of the m/z range.
To illustrate this methodology we show examples, both on standard samples and complex protein mixtures where Q-Tof data acquisition has been directed based upon the results from a databank search. This data will be compared and contrasted to data acquired in the normal automated LC-MS/MS mode.

R. Mineki, H. Taka, T. Fujimura, M. Kikkawa, N. Shindo, K. Murayama;
Juntendo Univ. Sch. Med. Cent. Lab. Med. Sci., Tokyo, Japan.

Cysteinyl residues in proteins were alkylated with acrylamide during sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) to yield a thioether derivative, cys-S-β-propionamide (PAM cys). The process was termed in situ alkylation with acrylamide. Using this method, the recovery of PAM-cys peptides from bovine serum albumin (BSA) was 88.6 % at 10 picomol in 1D SDS-PAGE and 97.1% at 50 picomol in 2-dimensional (2D) SDS-PAGE. The coverage of tryptic peptide of BSA in 1D/2D SDS-PAGEs was 83.7% and 81.1%, respectively. The advantages of in situ alkylation with acrylamide were the following. (i) Cysteinyl peptides were effectively derived in a single PAM cys and then proteins were precisely identified using databases. (ii) Marked reduction of salts compared with post alkylation, e.g., using carboxymethylamide (CAM), resulting in higher signal intensity and wider coverage of cysteinyl peptides from PAM cys, compared with those of CAM derivatives, in mass spectrometry peptide mapping. (iii) Shorter duration by excluding the processes of post alkylation and desalting before peptide mapping.
We used the method for the automated in gel digestion without reduction and alkylation. And then, the desalting process was also unnecessary before mass spectrometry analysis for protein identification.

J. J. Cummings, L. Kochanski, E. Rohde, J. Mehl, G. Salituro;
Merck & Co. Inc., Rahway, NJ, United States.

Abstract
Hannig & Heidrich first described the use of continuous free flow electrophoresis (FFE) for purifying proteins almost three decades ago.¹ We describe the use of the Tecan FFE for first dimension isoelectric focusing of plasma as a means of enrichment for low abundance proteins. Whole plasma and plasma depleted of albumin and immunoglobulin G were subjected to continuous horizontal isoelectric focusing. SDS-PAGE, 2D-PAGE and RP-HPLC further resolved the resulting fractions. Selected proteins were identified by μLC-ESI-IT-MS/MS and AP-MALDI-IT-MS/MS with SEQUEST database correlation.
FFE has the ability to confine abundant plasma proteins such as albumin in as few as four of seventy fractions collected. For this reason it is a useful first step in plasma profiling. Reducing the complexity of the plasma (through depletion of albumin and IgG) prior to fractionation by FFE greatly improves the ability to identify proteins not previously observed with conventional depletion methods. We identified proteins from Coomassie blue stained gels that are present in plasma in the ng/ml range. These proteins include ceruloplasmin and serum amyloid protein. Applying each FFE fraction to an RP-HPLC can effect further resolution of the plasma proteome. This method allows for the rapid screening of a greater number of plasma samples than if gel electrophoresis were employed for the second dimension separation.
References:
¹Hannig, K and Heidrich, H. G. The use of continuous preparative free flow electrophoresis for dissociating cell fractions and isolation of membranous components. Methods in Enzymology. 3, 746. 1974.

U. Schneider, A. Posch, G. Weber, P. Weber, M. Nissum, C. Obermaier, R. Wildgruber, S. Kuhfuss, C. Eckerskorn;
Tecan Munich GmbH, Kirchheim, Germany.

Hydrophilic interaction chromatography (HILIC) is a long standing variant of normal phase chromatography, which binds proteins to a strongly hydrophilic support based on interaction of hydrophilic parts of the proteins with the support.
Binding is strongly promoted in highly concentrated organic solvent, whereas elution occurs, when the support is flushed with aqueous solutions.Here we show the utility of HILI-SP in the context of Free Flow Electrophoresis post-sample processing. We demonstrate the concept of "compatible recovery" which allows recovery of bound proteins from the HILISP support by elution with the downstream analytics buffer, e.g. if 2D-PAGE is required as downstream analytics, elution takes place with 2D-PAGE sample buffer.

D. L. Wong, P. S. Niedle, N. L. Chow, B. G. Fryksdale, B. S. Miller, A. L. Gaertner;
Genencor International, Inc., Palo Alto, CA, United States.

Abstract: Two-dimensional (2D) gel electrophoresis and matrix-assisted laser desorption ionization-time of flight-mass spectrometry (MALDI-TOF/MS) are widely used tools in proteomic analyses. With the availability of genomic databases, industrialization of research tools and the development of data analysis algorithms, the quantity of data generated has increased exponentially in the past few years. Unfortunately, data quality has not always kept pace. This is in part due to the fact that many current methodologies for protein identification have not been optimized for high-throughput proteome analysis. We have evaluated and made improvements in a number of sample preparation methods, including in-gel protease digestion, peptide extraction and sample spotting onto MALDI targets to address this problem.
A set of proteins was selected to represent realistic challenges faced in today’s proteomics research. The proteins included: bovine serum albumin (BSA), the highly hydrophobic bacteriorhodopsin and the highly glycosylated Tricoderma reesei cellobiohydrolase I (CBH-I). Our optimization scheme tested the effect of 1) various in-gel digestion methods. 2) solvents for peptide extraction. 3) various MALDI matrices and 4) different sample spotting techniques. Mass spectra of the samples were evaluated for total number of peptides detected, average peptide intensity corresponding to its hydrophobicity value, percent protein sequence coverage, and final protein identification score obtained from NCBI protein database searching. We optimized the conditions for both manual preparation and via the use of an automated digestion robot. We report here the optimized methods for high-throughput proteome analysis.

P. R. Levison, M. Streater;
Whatman International Ltd., Maidstone, United Kingdom.

Stable storage of functionally active proteins is challenging often requiring storage at <-20ºC or the use of lyophilised preparations. In the present study we report the development and utility of Protein Saver™ a novel storage medium to which can be applied protein samples, allowed to dry and then be stably stored at room temperature for several months. In order to stably store a proteome we applied a sample of human plasma and stored it at room temperature for 2 weeks. The sample was eluted from the medium and analysed by 2D-gel electrophoresis. The protein profile was similar to that obtained for plasma stored at -80ºC. Trypsin is an enzyme routinely used for protein digestion following gel electrophoresis. Trypsin has poor stability yet we have stably stored bovine and porcine pancreatic forms for more than 3 months at room temperature on Protein Saver. Trypsin previously stored dry on Protein Saver at room temperature and a freshly made solution were used to digest several proteins punched from a 2D-gel of human heart left ventricle. Analysis by MALDI-TOF MS demonstrated similar peptide fragmentation for both enzyme preparations. In a related study tryptic digests of bovine serum albumin could also be stably stored on Protein Saver at room temperature for at least 7 days and gave similar mass spectra to digests stored at 4ºC. These data demonstrate that Protein Saver has utility as a storage medium being both fast and simple to use with evaporation of water being the physical event occurring in each example. Furthermore the stored proteins can be readily eluted from Protein Saver without the use of specialised elution systems.

A. J. Alpert¹, A. K. Shukla²;
¹PolyLC Inc., Columbia, MD, United States, ²Glygen Corp., Columbia, MD, United States.

In serum and other complex biological samples, a few proteins often constitute a high percentage of the total protein. These high-abundance proteins can mask the numerous proteins of lower abundance in proteomics analyses, whether intact or in tryptic digests. Affinity materials can remove specific proteins such as albumin or immunoglobulins, but it would be tedious and expensive to develop such materials for the various high-abundance proteins in every biological sample. If the proteins and peptides of interest are smaller than ~ 30 KDa, then a simple alternative is to add organic solvent to the sample sufficient to precipitate proteins larger than this size. This would include the six most abundant proteins in serum, which represent > 90% of the total protein. The supernatant could then be subjected to proteomics analysis. We demonstrate this approach using serum as well as standards of the most abundant serum proteins. Proteins and peptides remaining in solution are analyzed by Size Exclusion Chromatography to demonstrate the decrease in their average mol. wt. upon addition of increasing amounts of solvent. They can be concentrated for further fractionation by Hydrophilic Interaction Chromatography (HILIC) or by other methods such as ion-exchange chromatography.

A. Posch, G. Weber, P. Weber, U. Schneider, C. Eckerskorn;
Tecan Munich GmbH, Kirchheim, Germany.

Facing the complexity of higher eurkaryotic cells, the complete characterization of the corresponding proteomes is almost impossible to achieve without applying pre-fractionation strategies like subcellular fractionation. Although density-gradient centrifugation has been known for a long time as a successful method to isolate subcellular fractions like organelles in high amounts, their purity is usually not sufficient for in-depth proteome analyses due to interferences by cross contaminating organelles, which are very often partially co-purified. In the present study, we evaluated the use of Free-Flow Electrophoresis (FFE) as a tool for purifying mitochondria from Saccharomyces cerevisiae for subsequent proteome analysis. Whereas mitochondrial preparations isolated by means of differential centrifugation included a considerable degree of non-mitochondrial proteins (16%), such contaminations could be effectually removed by the inclusion of a FFE purification step (2%) as demonstrated by 2-D electrophoresis and MALDI-MS analysis. This higher degree of purity led to the identification of many more mitochondrial proteins from FFE-purified mitochondrial protein extracts (n=129), as compared to mitochondrial protein extracts isolated by differential centrifugation (n=80). Moreover, a marked decrease of degraded proteins was found in the ZE-FFE-purified mitochondrial protein extracts. Therefore these results represent a feasible approach for an in-depth proteome analysis of mitochondria and possibly other organelles.

Y. Yu, M. Gilar, P. J. Lee, J. C. Gebler;
Waters Corporation, Milford, MA, United States.

Integral membrane proteins play vital roles in many cellular and physiological processes. Due to their hydrophobic nature, characterization of the integral membrane proteins is very difficult. Surfactants are typically used to enhance their solubility prior to enzymatic digestion. However, addition of surfactant usually causes reduction in proteolytic enzyme activity and compromises the LC/MS analysis. We applied an acid degradable, trypsin compatible denaturant, RapiGest™ SF, to improve the solubility of a model membrane protein, bacteriorhodopsin. Bacteriorhodopsin was suspended in RapiGest™ SF solution (0.05%, w/v) prior to tryptic digestion. The digested protein sample was separated and mass analyzed via capillary HPLC/MS. Tryptic peptides were identified by their masses with high accuracy via TOF MS analysis. Our preliminary results show that complete mapping of bacteriorhodopsin was achieved with less than 1 µg of sample. This study shows that RapiGest™ SF dramatically improves the characterization of membrane proteins by LC/MS.

D. Sarracino¹, P. Murphy¹, V. Saenz-Vash¹, K. Grant¹, J. Stoerker²;
¹Matritech, Inc., Newton, MA, United States, ²Bruker Daltonics, Billerica, MA, United States.

In discovery proteomics, many classical analytical methods suffer severe deficiencies. For example, HPLC methods provide excellent partitioning but generally fail to provide throughput required for unbiased population or association studies. Surface capture methods provide minimal sampling capacity and may only yield information on abundant proteins in complex mixtures. Affinity methods generally require assumptions about the nature of the target protein.
We demonstrate a method that allows multiple samples to be simultaneously prepared in an automated format. This method surveys both proteins and peptides from a half-milliliter of serum. MALDI friendly buffers allow reproducible and detailed partitioning for rapid analysis. Essential elements of the method include passivation of plasticware, non-foaming buffers to eliminate crossover contamination, and high capacity resins to promote reproducible distributions of proteins and peptides.
In this example, sera obtained from 6 healthy and 6 prostate cancer patients were partitioned. Albumin is stripped from each serum sample using Blue Sepharose 6 Fast Flow in a 96 well plate. Each stripped sample was divided in half and the pH adjusted for anion and cation exchange resins. MacroPrep Q and S were in alternate rows in a second 96 well plate. Stepwise elutions were made of proteins and peptides in 13 increasingly concentrated, MALDI friendly buffers. Portions of each of the 26 elutions were spotted directly onto HTP AnchorChip targets and allowed to dry before the automated addition of matrix for either proteins (HABA) or peptides (HCCA). MALDI-TOF and TOF/TOF analysis were performed on the partitioned samples. Peaklists from the samples were compiled for statistical analyses. Candidate biomarkers from the analyses will be discussed.

M. Minkoff, R. Pintal, K. Davis, S. Zobbi;
Applied Biosystems, Framigham, MA, United States.

Abstract
Isotope coded affinity tags are currently used by many laboratories for protein identification and quantitation. Complex samples such as serum or plasma present the challenge of identifying relatively low abundant proteins in the presence of high abundance proteins such as Albumin, IgG and Transferrin. Upfront sample preparation can enhance any proteomics experiment by reduced sample and reagent use. Simple depletion of these proteins from the sample aids in the total number of proteins identified by removing a layer of complexity and by enriching the sample for lower abundance proteins.
Ligand absorption chromatography, such as Heparin and Concanavalin A, can help enrich for protein classes of interest. Additional 1D gels can be used as a technique that can both remove interfering proteins and isolate classes of proteins.
LC/MS/MS data shows that by removing non-relevant proteins or enriching for proteins of interest, we can reduce the complexity of mass spectra, improving identification and interpretation. Our results also indicate that we were able to delve further into the Proteome using these simple sample preparation techniques.
Both depletion and absorption affinity chromatography increases the relevant quantity of low abundant proteins available for identification. Increasing protein coverage leads to enhanced identification and quantitation.

P. Hufnagel¹, C. Koester¹, M. Schuerenberg¹, H. Langen², A. Wattenberg³, M. Blueggel³, C. Stacey⁴, D. Suckau¹;
¹Bruker Daltonik GmbH, Bremen, Germany, ²Roche, Basel, Switzerland, ³Protagen GmbH, Dortmund, Germany, ⁴Bruker Daltonics Inc., Billerica, MA, United States.

Gel-imaging, spot selection and protein digestion, MS analysis by MALDI-TOF and subsequent data-dependent acquisition of MALDI-TOF/TOF and LC-ESI-MS/MS were integrated into a single platform. It consists of a spot picking and an eight-needle digestion robot, which provide the samples on a 384 sample plate and four 96-well MTPs for MALDI and ESI, respectively. All information about the samples, spots, gels and their respective images and mass spectra including search results are stored in a proteomics results database that also controls the workflow downstream the initial acquisition of protein mass fingerprints. Failed PMF identifications cause the system to automatically acquire TOF/TOF fragment ion spectra from selected peaks and, if these fail again, to acquire full LC-ESI-MS/MS datasets from the corresponding well in the MTP.
The digest robot sample preparation is adapted to AnchorChips, i.e., sample plates with hydrophobic coating and on which hydrophilic spots are arrayed that predefine the size and location of each sample. In situ purification procedures are run on the AnchorChip that make expensive or labour-intense microcolumn purifications obsolete. In addition, they provide a sensitivity for protein identification of phosphorylase b on gel for PMF ca. 50 fmol and for MS/MS ca. 100 fmol even though only 30 % of the sample was used for those analyses. 70 % is saved for downstream electrospray work. A system throughput of 2*384 2D gel spots/day was achieved with the system and the ready availability of MS/MS data allowed to detect sequence errors in the database and protein modifications.

E. Chernokalskaya¹, C. E. Murphy², A. J. Tomlinson², W. Kopaciewicz¹;
¹Millipore Corp., Danvers, MA, United States, ²Applied Biosystems, Framingham, MA, United States.

Even with the most sophisticated technology, the depth of proteome coverage is usually low with abundant proteins dominating the analysis. There are likely to be many more proteins in the gel than are visualized by commercially available stains. These proteins are often accessible to mass spectral analysis.
In this paper, we describe a system for analyzing low abundant proteins from cell lysates. Gel pieces were excised from unstained regions of a 2-D gel and processed using a system composed of an integrated 96 well sample preparation and presentation system that spots digested proteins directly onto a MALDI target. The sample preparation system (ZipPlate™) is a 96 well solid phase extraction device containing 300 nl of C18 membrane. Gel pieces are added to the wells where they are destained and digested with trypsin. The resulting peptides are desalted and concentrated on the C18 media and the eluant from the ZipPlateC18 device is directly captured on a Voyager DE™-STR MALDI-TOF-MS compatible target. This new target has been manufactured such that 4 targets fit neatly under the ZipPlate device, which is dimensioned to SBS standards.
A human carcinoma cell lysate was fractionated on a 2-D gel, visualized, and gel plugs were cut from low intensity and unstained gel regions. These samples were processed in the ZipPlate device, directly spotted on the new Voyager system target and analyzed by MALDI-TOF-MS and tandem mass spectrometry using an OMALDI™ interface on a hybrid quadrupole-time of flight MS system. Both sample preparation and analysis were completed in about seven hours. Detection and identification of unvisualized proteins was demonstrated.

T. Slyker¹, M. Nguyen¹, B. Gillece-Castro², L. Sapp², M. Brubacher¹, W. Strong¹;
¹Bio-Rad Laboratories, Hercules, CA, United States, ²Waters Corporation, Beverly, MA, United States.

2D gel electrophoresis (2DE) is a widely used, proven method for Proteome analysis. The quality and value of the information obtained from a 2DE experiment is highly dependent upon the initial sample preparation. In order to identify the most complete array of cellular proteins it is often necessary to reduce the complexity of the protein sample. A strategy for reduction in sample complexity is especially important when analysis of low abundant and membrane proteins is the goal. Ideally the method(s) employed should be simple, reproducible, general to a wide variety of cell types, and result in a low conductivity protein sample that is free of substances that interfere with 2DE. With these aims in mind, we will present several solutions for convenient and efficient extraction of cellular proteins into discrete, more easily manageable fractions that are enriched in certain classes of proteins such as cytosolic, nuclear, membrane and signaling. The majority of these procedures are intended to provide tools that simplify the preparation of membrane proteins that are generally considered difficult to isolate. These proteins are of considerable interest due to their roles in signal transduction and cell-to-cell interactions among other functions. 2DE and MALDI peptide mass fingerprinting data will be presented to illustrate the effective application of these techniques to improved sample pre-fractionation and protein identification.

G. A. Musso, X. Dai, C. Du, J. G. Coles;
Hospital for Sick Children, Toronto, ON, Canada.

Protein phosphorylation is one of the most vital aspects of signal transmission in all living cells. It plays a crucial role in cell growth, function and apoptosis. To date, the protein phosphorylation pathways present during myocardial ischemia have not been well delineated. The purpose of this study was to study protein phosphorylation during myocardial ischemia through Surface Enhanced Laser Desorption-Ionization Time of Flight (SELDI-TOF) spectrometry. The study utilized Immobilized Metal Affinity Chromatography (IMAC) protein chip arrays as a platform for SELDI-TOF analysis. The IMAC chips were pre-incubated with Gallium, allowing them to specifically bind phosphorylated proteins. SELDI-TOF analysis allows high-throughput simultaneous quantification of hundreds of phosphorylated proteins, and thus represents a useful tool for studying widespread protein signatures specific to ischemia. Amounts of the phosphorylated form of proteins in both control and 14 hour ischemically-stressed fetal rat cardiomyocytes were compared. Thus it was possible through analysis of the spectrometry readings to compare relative expression of phosphorylated proteins between the two groups. In the mass range of 8000 to 30000 Daltons, three proteins were found to consistently have decreased expression with ischemia (p {Mann-Whitney} < 0.0003). Identification of these selectively de-phosphorylated proteins using affinity chromatography and mass spectrometry is currently underway. These results represent the first successful trial of IMAC-Gallium coated chips in the field of cardiovascular proteomics. This method represents a novel and generically informative way to study apoptosis and other pathways highly dependent on protein phosphorylation.

H. U. Shekhar, C. Itagaki, T. Isobe, T. Ichimura;
Tokyo Metropolitan University, Tokyo, Japan.

One of the strategies of functional proteomics is the analysis of functional cellular machinery, e.g. cell signaling complexes and organelles, to better understand the basis of cell functions. Among the methods to isolate the cellular complexes, an affinity tag based pull-down experiment has been widely used for its versatility; Where , the gene with an affinity purification tag is transfected and expressed in the cultured cells, and the protein complex formed is pull-downed using an appropriate ligand immobilized on beads. We report here a novel multiple affinity tag, called myc-tev-flag (MEF), as a tool for rapid purification of functional protein complexes.
MEF consists of 135 nucleotides and allows three consecutive purification steps, giving high purity of the target protein and its associated complex. Due to its small size, it can easily be inserted into different position of the cDNA without hampering the biological activity . We tested this MEF tag to the analysis of RasV12, RasN17(in HEK 293T and PC12), TrkA (in nnr5), and C-Raf-1 (in 293T) cDNA complex. It was observed that tagging of MEF did not interfere with the biological activity of these proteins. C-Raf-1 kinase complex was purified from 293T as well as PC12 cell line after inserting MEF at the N-terminal of C-Raf-1. The associated proteins identified by nano LC-MS/MS system included the known associated proteins HSP90, tubulin ,14-3-3 and CDC homolog, as well as twenty other novel components.

M. L. Johnston, J. Banas, N. Laliberte, W. R. Stochaj, J. Corbo, J. Nagel, M. Meys;
Beyond Genomics, Waltham, MA, United States.

A major obstacle to proteomic analysis of complex samples such as plasma and serum is the presence of highly abundant proteins which are present at levels ten orders of magnitude higher than those proteins at the low concentration end of the spectrum. The removal of abundant proteins can facilitate the detection of lower concentration proteins by altering the dynamic range of the sample. We describe a technique which uses multi-column immuno-affinity and on-line reversed phase chromatography to deplete four abundant proteins from human plasma and serum and to desalt the resulting solution. The presence of abundant proteins and subsequent removal are illustrated by the disappearance of densely stained areas that are observed on 2-D electrophoresis gels corresponding to areas where proteins such as serum albumin and transferrin migrate. In addition to serum albumin and transferrin, IgGs and alpha-1-antitrypsin are quantitatively removed from 50 uL aliquots of serum or plasma. The specificity of the technique is demonstrated and evidence is shown that non-specific proteins are not removed with the targeted proteins. The benefit of this technique is an increase in the dynamic range of protein detection by mass spectroscopy.

T. Yokono¹, R. Mineki², H. Taka², H. Kotaniguchi¹, K. Murayama²;
¹M&S Instruments Trading Inc. Tech. Div., Tokyo, Japan, ²Juntendo Univ. Sch. Med. Cent. Lab. Med. Sci., Tokyo, Japan.

We have improved the automation of in gel digestion system, DigestPro 96 (Intavis AG, Germany) using is situ alkylation with acrylamide. The in situ alkylation was done during 1D SDS-PAGE　(Mineki, R. et al., Proteomics, in press). We used alkylated and CBB stained BSA　as a sample for the method. The improved method included the steps of destaining, dehydration, trypsin digestion, and extraction. The improved method did not include reduction and alkylation since these steps were performed prior to and during the electrophoresis run. Extracted peptide mixtures were directly loaded onto the micro LC column (Magic, C18) combined with the Mass spectrometer (AB-QSTAR pulsar i hybrid)(LC-MS) and resulted spectra were processed with Mascot software to estimate the sequence coverage of the BSA. Original method supplied by Intavis AG was designed for Laemmli method’s 1D SDS gel and contained the steps of reduction and post-alkylation with CAM. The original method also needed desalting step that was essential for the mass spectrometry, especially MALDI-TOF-MS.
We compared both methods using BSA (3 picomol loaded to the gel, one third of digested peptide mixture was injected to LC-MS). The original method gave both CAM and propionicamide (PAM) derivatives. The latter was adducts made by unpolymerized acrylamide during electrophoresis. The coverage of CAM derivatives of BSA by the original method was 10% with desalging and 19% without desalting. The coverage of PAM derivative by the improved method was 32%. The improved automated in gel digestion method for in situ PAM alkylated protein had the advantage on recovery over the original method with CAM post-alkylation.

C. Eckerskorn, G. Weber, P. Weber, A. Posch, U. Scheider, K. Rein;
Tecan Munich, Kirchheim, Germany.

Current traditional proteomics techniques can be insufficient in regards to: sensitivity (e.g. difficulties in detecting unknown, low abundant proteins), reproducibility (e.g. lack of robust, consistent methods), throughput (lack of automation systems and methods).
To address these needs, Tecan has developed an automated modular platform. which consists of, a Free Flow Electrophoresis System (a novel semi-preparative, charged based, liquid separation technology for fractionation of cells, cell organelles and complex protein mixtures), a 2D-PAGE System (a fully automated system from IEF, Gel casting, SDS-PAGE, and Staining), and a Protein Processing System (consisting of a spot picking device, in gel digestion, interface to MS). The instrumentations can be integrated in various combinations (e.g. FFE combined directly with 2D-PAGE or LC-MS) and allows researchers to construct a customized platform for different proteomics applications.
In the presentation the benefit of an integrated automated platform where shown by means of different examples from human serum, yeast and liver. A special accentuation will be the combination of Free Flow Electrophoresis with classical 2D-PAGE. Since FFE can be operated in different modes like IEF, ZE, and ITP, this technology is capable to reduce the complexity of the samples by fractionation, which would allow to increase the number of low abundance proteins that can be visualized and further analysed by appropriate down-stream technologies. ZE was applied to cellular subfractionation, e.g. the isolation of mitochondria or lysozomes, native and denaturing IEF or ITP were used for fraction of complete protein mixtures. This approach, showed a persuasive new protein separation pathway that provides highly dissected 2-D protein patterns, as well as very fast fractionations in combination with nearly complete sample recoveries

M. Swartz¹, D. Hazen-Martin¹, A. Bland¹, J. Arthur^2,1;
¹Medical University of South Carolina, Charleston, SC, United States, ²Ralph H. Johnson VA Medical Center, Charleston, SC, United States.

Regions of the renal tubule serve different functions and express proteins appropriate to their function. Knowledge of specific proteins expressed in each region of the nephron will provide a basis for understanding function, the way that changes in renal protein abundance occur in specific diseases and assist in drug targeting. Primary cultures of renal proximal tubule cells were isolated from normal sections of human kidneys that were removed for medical reasons. Cells were grown until confluent, harvested and separated into Tris-soluble supernatant and pellet. The pellet was resuspended in a buffer containing 5 mM Tris, 8 M urea, 4% CHAPS, 0.2% ampholytes (3-10), 2 mM tributyl phosphine. Proteins were separated by 2-dimensional gel electrophoresis and stained with Sypro Ruby. Ninety-six spots were chosen for identification. These spots were processed using robotic workstations for picking, digestion, ZipTip cleanup and MALDI target spotting. Protein digests were analyzed in a MALDI-TOF mass spectrometer in reflectron mode and the output was analyzed using Protein Lynx Global Server software. Gel images were annotated with appropriate identifications. Thirty-five of the spots (36.5%) were identified including: 14-3-3 Protein zeta/delta; Actin bundling protein; Actin, Alcohol dehydrogenase; Annexins I, III, IV; Dipeptidyl peptidase IV; FUSE binding protein GRP 78; Heat shock 70 kDa protein; Heterogeneous nuclear ribonucleoproteins H1, C; IMP dehydrogenase II; Lamin C; Manganese SOD; Phosphoglycerate kinase; Phosphoglycerate mutase; Pyrophosphatase; Pyruvate kinase; Stratifin; TCP-1-epsilon; TGF-Beta receptor interacting protein 1; Transformation upregulated nuclear protein; Tropomyosin; Tubulin; Villin 2 and Vimentin. These data demonstrate that a primary human cell line and robotic workstations can be used for rapid identification of physiologically relevant proteins.

M. Meys, S. Naylor, T. N. Plasterer, M. Oresic, M. Hincapie, C. B. Clish, E. Davidov;
Beyond Genomics, Waltham, MA, United States.

Systems Biology is an integrated approach to studying biological systems-intracellular networks, cells, organs, and biological entities by measuring and integrating genetic, genomic, metabolic, proteomic and other data. This approach analyzes cellular and pathway events that are in flux and interdependent. The application of Systems Biology to drug discovery includes utilizing clinical samples from diseased and healthy patients to discover BioSystem Markers™ and BioSelective Targets™ that are indications of disease and targets for therapeutic intervention. Essential to extracting knowledge from the component inputs is an integrated informatics environment which assembles sets of gene, protein and metabolite data into biochemical pathway and whole system knowledge. We present selected data from both animal and human studies which illustrate the application of Systems Biology to several important disease states and examples of pattern recognition and data display tools to interpret such data.

T. Tremblay, J. K. Sandhu, Y. Li, K. Mcrae, D. Stanimirovic, M. Sikorska;
National Research Council of Canada, Ottawa, ON, Canada.

Mitochondrial Encephalopathy, Lactic Acidosis, Stroke-like episodes (MELAS) syndrome is caused by a point mutation in the mitochondrial tRNA^Leu gene at A3243G. This mutation is thought to disrupt protein synthesis from genes encoded by the mtDNA and results in the defective respiratory chain. The mechanisms underlying the complex relationship between mtDNA genotype and clinical phenotype remain to be established. To better understand the pathological consequences of mitochondrial dysfunction, we have generated mtDNA-null ρ° cells from U87MG glioblastoma cells. Subsequently, we used these cells to develop cybrids by fusing them with fibroblasts from a patient carrying the MELAS mutation. Two cybrid clones, R9 and R921, were isolated and found to carry a different proportion of mutated mtDNA, 55-60% and 70-80%, respectively. One control cybrid, R15, was also established and found to contain 100% of normal mtDNA. This study was undertaken to investigate any potential effects of these mutations on the cellular proteome in order to identify a link between genotype and phenotype. Proteins from parental, ρ° and cybrids were extracted and separated by two-dimensional electrophoresis (2D). Differentially expressed proteins were identified by mass spectrometry. Comparison of the parental and ρ° cells showed that key mitochondrial enzymes were missing from the 2D gel maps obtained from ρ° cells, confirming the disruption of mitochondrial function. However, anti-oxidative enzymes as well as enzymes involved in the glycolytic metabolism were up-regulated in these cells indicating that the cells were trying to compensate for their loss in mitochondrial anti-oxidative and respiratory function. By applying the same proteomics approach on the cybrids, we hope to better understand functional changes in brain cells expressing the MELAS mutation.

E. D. Decker, Y. Zhang, R. R. Cocklin, F. A. Witzmann, M. Wang;
Indiana University, Indianapolis, IN, United States.

Cells treated with UV irradiation undergo cell cycle arrest at S-phase and G1/S boundary, allowing DNA repair to occur. It is known that UV-induced inhibition of replication is efficiently reversed after 8 hours of treatment. Previous studies have used different cell lines to link proteins responsible for this observation. In this study, we used 2-DE based proteomic approach to identify specific proteins that may be responsible for UV-induced replication arrest. Briefly, HeLa cells were exposed to 10 J/m² of UV light and cells were harvested at different time points after UV treatment. The proteins were analyzed using 2-dimensional gel electrophoresis and MALDI-TOF Mass Spectrometry. Unique changes in the amount of protein expression for more than 30 proteins were observed over 24-hour time course. Several proteins were significantly up- or down-regulated in the first 8 hours following UV treatment. The level of these proteins precisely correlated to DNA replication activity, suggesting involvement of these proteins in modulating DNA replication and repair activities. This proteomic approach provides a good tool for such studies.

A. Zhukov¹, M. Schürenberg², Ö. Jansson¹, D. Areskoug¹, J. Buijs¹;
¹Biacore AB, Uppsala, Sweden, ²Bruker Daltonics, Bremen, Germany.

In the light of the recent developments in functional proteomics the cell environment is emerging as a complex network of interacting biomolecules. We are presenting a tool which is a result of integrating two well-established techniques, mass spectrometry (MS) and surface plasmon resonance (SPR). This combination can be used in functional proteomics for ligand fishing or studying novel biomolecular interactions and macromolecular complexes.
SPR technology provides a sensitive real time detection of the binding and dissociation of biomolecules on a sensor chip. When a complex biological mixture is injected, specific interaction partners are captured by the molecule immobilized on the chip. The bound material can then be eluted and its identity established by MS. Thus SPR-MS allows characterization of biomolecular interactions, affinity purification, and identification of interaction partners in one single and fully automated experiment with a minimum of sample consumption.
To demonstrate the general applicability of the SPR-MS tandem, examples will be given in which proteins of interest were fished from complex biological mixtures such as cell lysates. In a typical experiment 100 to 500 femtomoles was captured on the sensor chip and then eluted in a volume of 2 μl. After elution the samples were automatically prepared for MS analysis in various ways including direct application on MALDI target, on-target purification, and enzymatic digestion prior to MS sample preparation. MS was used to characterize the captured proteins by their intact molecular weight and to identify them by peptide mass fingerprints.

M. G. Kolonin, C. I. Vidal, R. Pasqualini, W. Arap;
M.D. Anderson Cancer Center, Houston, TX, United States.

Organ development and maintenance depends on signaling pathways that originate in tissue-specific endothelial beds and are then transduced into underlying cells of the tissue. These processes are operated by protein interactions occurring in the vasculature, such as those between growth factors or extracellular matrix molecules and their receptors expressed by the endothelium. We had previously developed an in vivo phage display selection method, in which peptides homing to specific vascular beds are identified after administration of a phage display random peptide library into mice. This strategy had revealed a receptor-mediated address system that allows targeting of the vascular endothelium. Several vascular receptors for individual organ-homing peptides have been successfully identified, however new efficient approaches to systematically characterize the ligand-receptor map are clearly needed. Recently, we reported the first in vivo screen of a phage library in a human subject (Nature Medicine, 2002, 8, 121-7). To validate the receptors recognized by the ligands isolated in the human in vivo biopannings, we have developed a phage-based technology to register interactions between peptides and proteins. Our approach will be used for the high-throughput analysis of the proteome, and will eventually help constructing a detailed map of ligand-receptor interactions in the human vasculature.

M. Iqbal¹, N. R. Pumford¹, K. Lassiter¹, Z. X. Tang¹, T. Wing², M. Cooper², W. G. Bottje¹;
¹University of Arkansas, Poultry Science Department, Center of Excellence for Poultry Science, Fayetteville, AR, United States, ²Cobb-Vantress, Inc., Siloam Spring, AR, United States.

The objectives of this study were to determine the effects of low or high feed efficiency (LFE, HFE) on; a) the activities of various electron transport chain (ETC) complexes, b) ubiquitin conjugated proteins, and c) various ETC protein subunits encoded by mitochondrial (mt-) and nuclear (n-) DNA in broilers. Tissue homogenate or mitochondria were isolated from breast muscle or liver of broilers with HFE (0.80 ± 0.01, n = 7-8) and LFE (0.62 ± 0.02, n = 5-8). The activities of the ETC complexes were measured spectrophotometrically, while the levels of ubiquitin and immunoreactive proteins were analyzed using western blots. Activities of all complexes (I, II, III & IV) were higher in HFE compared to LFE broilers for both breast muscle and liver mitochondria. The levels of ubiquitin conjugated proteins were higher in LFE compared to HFE broilers breast muscle. While the levels of the ETC immunoreactive proteins were higher in LFE muscle mitochondria for core I, cyt c1, cyt b (Complex III proteins) and COX II (Complex IV), but the levels of subunits ND3, ND4, ND5, ND6, ND7 (Complex I), 70S (Complex II), core II, ISP (Complex III), α—ATPase (Complex V) were not different between the groups. We hypothesize that higher steady state expression of certain ETC proteins (core 1, cyt c1[n-encoded], cyt b and COX II [mt-encoded) might be a compensatory response to lower ETC complex activity possibly to overcome the increased protein oxidation in LFE birds.

D. P. Pashby, J. Prime, A. Alban, B. Hughes;
Amersham Biosciences, Amersham, United Kingdom.

Two-dimensional electrophoresis (2-DE) is the leading tool in proteomics research, capable of visualising many components of complex proteomes in a single gel. Accurate quantitative proteome analysis of two or more different samples using conventional "one sample per gel" 2-DE is limited. Many software packages have attempted to address the problems of gel-to-gel variability and the limitations in the dynamic range of some protein stains (e.g. silver), however these do not resolve the underlying issue of experimental variability.
In two-dimensional difference gel electrophoresis, the Ettan™ DIGE system, protein samples are labelled with mass and charge-matched, spectrally resolvable fluorescent dyes. Once labelled the samples may be combined and run on a single 2-D gel enabling the inclusion of an internal standard on every gel. These dyes, combined with the appropriate experimental design and the novel DeCyder™ analysis software, allow accurate determination of changes in protein levels to be made and reduce experimental variation to exceptionally low levels.
Conventional 2-DE analysis is compared with the Ettan DIGE system using E. coli lysates ‘spiked’ with four different proteins, at eight different amounts to give 8 sample types — equating to ‘time points’, and run in triplicate for statistical analysis. Each protein spike was added to the lysate in such a way that the trend of the level of the spike over the eight points was different for each spike (i.e. one spike showed a linear increase, another a linear decrease etc.).
The Ettan DIGE system approach, with internal standard, showed the spike results were much closer to the predicted values than those from the conventional "one sample per gel" analysis.

G. J. Rucklidge¹, M. D. Reid¹, J. H. Beattie¹, M. Apostolova^1,2;
¹Rowett Research Institute, Aberdeen, United Kingdom, ²Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, ON, Canada.

Mechanisms to prevent and reduce coronary heart disease (CHD) have been the focus of intense research for many years. Homocysteine is a sulfur-containing amino acid playing a pivotal role in methionine metabolism and enzyme defects or dietary deficiency of B-vitamins involved in this metabolism result in elevated homocysteine levels that have been associated with increased risk of CHD and/or atherosclerosis. Whether this association is causal is uncertain.
We have used a model system, allied to proteomics, involving endothelial cells (EA.hy 926) and/or monocytes and metal depleted media to assess targets of hCys action.
We demonstrated that following 24h treatment with 100 uM hCys, confluent EA.hy 926 cells exhibited alterations in four major protein groups associated with mitochondrial function, apoptosis, cellular matrix and cytoskeleton. Using an anti-phospho-tyrosine antibody, phosphorylated proteins that participate in the early activation of signal transduction pathways following hCys treatment were detected within 20 min. These included Hiap-1, BiP, TCF3, filensin, tubulin vimentin, testican 3,Cdc42, Erk, PKC z/d showing that the effects on cellular matrix proteins can be related to cytoskeletal reorganization. To understand how such re-organisation impacts on diapedesis of monocytes through the endothelial layer, EA.hy 926 monolayers were co-cultured for up to 30 min with U937 cells (pro-monocytic cell line) in the presence of 100 µM hCys. The results showed there is an increase (P<0.001) in the number of spreading, transmigrating or infiltrated monocytes
At such concentrations of hCys there are direct effects on signal cascades, mitochondrial respiration and matrix proteins of endothelial cells that then allows invasion of monocytes a precursor event to atherosclerotic plaque formation.

L. E. Cammish;
NextGen Sciences Ltd, Huntingdon, United Kingdom.

Following completion of the first draft of the Human Genome, the focus in research and development is shifting to protein expression and the function of proteins in healthy and diseased states. Resultantly, there is a growing need for new tools to produce many hundreds of proteins for disease identification and drug discovery research, for uses including high throughput screening, antibody production and characterisation, protein microarray applications and crystallography. NextGen Sciences describe their unique combination of tools that enable the whole process of protein expression and subsequent purification to be automated, resulting in the systematic, parallel production of many hundreds of purified proteins with minimum input and hands-on time. A range of unique expression vector systems will be discussed which incorporate a series of fusion partners, enabling optimal protein expression of any protein from any organism, together with affinity tags for subsequent purification. Using the fully automated system that will be described in this presentation, it is possible to start with cDNA clones and automate the whole process of sub-cloning, expression and purification of hundreds of proteins in parallel for use in a whole range of different proteomics research applications

R. A. Grant, A. M. Fieno, M. P. Lacey, T. W. Keough, Y. Sun, C. E. Catrenich;
The Procter & Gamble Company, Cincinnati, OH, United States.

The development of bacterial resistance to common antimicrobials (e. g., biocides and antibiotics) is of increasing public concern with respect to human health and our ability to treat infectious organisms. Of particular interest is the relationship of biocide and antibiotic resistance and the resulting clinical implications. To better understand the mechanism by which bacteria may develop resistance to the biocide triclosan (TCS), we characterized laboratory-derived variants of E. coli that were generated following exposure to increasing concentrations (0.15-2000 μg/ml) of this antimicrobial. Despite resistance to TCS, the susceptibility of the variants to a number of commonly used antibiotics was essentially unchanged. However, changes in protein profiles, metabolism, and morphology were observed. Using a combination of 2D gel electrophoresis and mass spectrometry, we determined that the FabI protein isolated from all of the resistant strains was mutated (Gly93Val). The most resistant variant (P10A-B) overexpressed this protein by about two-fold compared to the parent strain and also showed changes in the expression of numerous other proteins. Proteins that were up-regulated by TCS include those involved in protein and RNA synthesis, stress, and protein secretion, and those that were down-regulated include membrane porins and periplasmic binding proteins. E. coli variants had slower growth rates and filamentous morphologies. Interestingly, all strains maintained their TCS-resistant phenotype and the Gly93Val mutant form of the FabI protein following subsequent passage in TCS-free media. Our results suggest that laboratory-induced TCS resistance is a complex phenomenon and can arise via more than one route.

P. Lanthier¹, C. Szymanski¹, E. Taboada¹, B. Allan², K. Amoako², S. Gomis², L. Tessier¹, R. Verhulp¹, W. Wakarchuk¹, J. Nash¹, J. Kelly¹;
¹National Research Council Canada, Ottawa, ON, Canada, ²Veterinary Infectious Disease Organization, Saskatoon, SK, Canada.

Campylobacter jejuni is the leading bacterial cause of human enterocolitis in the developed world with 2-3 million cases reported in North America annually. We are examining the ability of C. jejuni to colonize chickens, their commensal host. We observed that the NRC version of the genome-sequenced strain, NCTC 11168, is a poor colonizer of chicks when compared with the same strain from VIDO. In addition, the VIDO strain is capable of spreading to uninfected chicks while the NRC version lacks this ability. This suggested that NRC’s strain has become lab adapted and less virulent even though both lineages arose from the same human isolate. A microarray-based genomic comparison shows that the two strains are genetically similar, consistent with a common origin, and suggests that phenotypic differences are likely due to differences in gene/protein expression levels. Electron microscopy and motility studies demonstrate that the VIDO strain is significantly more flagellated and motile, characteristics that are key to colonization. Genomic and proteomic data have also shown that the expression levels of flagellin-related genes/proteins are significantly higher in the VIDO strain. Consistent with these findings FliA, a sigma factor responsible for the up-regulation of flagellar genes, is more transcriptionally active in the VIDO strain. Furthermore, when both strains are induced to adhere equally to CaCo-2 cells by centrifugation, the NRC strain is much less invasive compared to the VIDO strain suggesting that other virulence factors and/or the type III secretion system has been affected. Results from this ongoing study will be presented here.

W. H. Vensel¹, C. K. Tanaka¹, W. J. Hurkman¹, J. H. Crawford²;
¹US Department of Agriculture, Albany, CA, United States, ²Nonlinear Dynamics Ltd., New Castle-upon-Tyne, United Kingdom.

High temperature during wheat (Triticum aestivum L. cv. Butte 86) grain-fill accelerates grain development, storage protein accumulation, and starch deposition. A proteomics approach is being used to identify endosperm proteins and determine the effects of temperature on metabolic events central to grain development. Plants were grown in a climate-controlled greenhouse that had an average daily maximum daytime temperature of 25°C and nighttime temperature of 17°C. At 10 and 20 days post anthesis, plants were transferred to a second greenhouse with identical growing conditions except that the average daily maximum daytime temperature was 37°C and nighttime temperature was 28°C. Developing grain from plants grown under the three regimens was harvested and endosperm collected. Salt soluble proteins were partitioned from storage proteins and separated by 2-D gel electrophoresis. Computer-based image analysis was used to determine qualitative and quantitative changes in gel-separated proteins. Peptides from in-gel tryptic digests are initially screened by MALDI MS and MS/MS (QSTAR). Samples not identified are subjected to ESI MS/MS (LCQ) or another round of MALDI MS/MS after concentration using ZipTips. This strategy successfully identified about 70% of the proteins processed. The identified proteins are involved in transport and metabolism of amino acids and carbohydrates, defense against predators, posttranslational modification, and transcription. Protein accumulation profiles were varied and reflected protein function. For example, many proteins involved in transport and metabolism of amino acids and carbohydrates were in highest amount in early to mid development and proteins involved in defense, late in development. Protein accumulation profiles reflected the shortened duration of grain development caused by the high temperature regimes.

X. Gu¹, K. Shadrach¹, M. Sun², K. A. West¹, L. Shan³, S. L. Hazen³, R. G. Salomon², J. G. Hollyfield¹, J. W. Crabb¹;
¹Cole Eye Institute, Cleveland Clinic Foundation, Cleveland, OH, United States, ²Chemistry Department, Case Western Reserve University, Cleveland, OH, United States, ³Lerner Research Institute, Cleveland Clinic Foundation, Cleveland, OH, United States.

Bruch’s membrane, located beneath the retina, thickens and losses permeability in age-related macular degeneration (AMD). We are probing the possible role of lipid peroxidation and protein modification in this pathology. Bruch’s membrane was isolated free of the adjacent tissues from 5 normal and 4 AMD donor eyes and in the presence of antioxidants. Lipids were extracted with chloroform/methanol and analyzed by LC MS. Proteins were subjected to 1D SDS-PAGE, and either blotted to PVDF membrane or gel bands were excised and proteins identified by LC MS/MS. Western analyses were used to screen for oxidative protein modifications. Lipid oxidation products from free radical-induced oxidation of docosahexaenoyl phosphatidylcholine (DHA-PC), arachidonoyl (AA)-PC, and linoleyl (LA)-PC were detected in relatively amounts from AMD than age matched healthy donor eyes. Many of the proteins we recently identified in drusen (2002 PNAS 23, 14682-14687) were also detected in Bruch’s membrane, including tissue inhibitor metalloproteinase-3, clusterin, vitronectin, and serum albumin. By Western analysis, carboxyethyl pyrrole protein adducts, generated from the oxidation of DHA containing lipids, were more abundant in Bruch’s membrane proteins. These preliminary observations support a role for oxidative damage in the pathology of AMD. Efforts are focused upon developing methods for characterizing low abundancy lipid oxidation products and oxidative protein modifications in ocular tissues.

S. K. Bhattacharya¹, Z. Wu¹, K. A. West¹, Z. Jin¹, M. Nawrot², J. C. Saari², J. W. Crabb¹;
¹Cleveland Clinic Foundation, Cleveland, OH, United States, ²University of Washington, Seattle, WA, United States.

The rod visual cycle is the process by which all-trans-retinal released from rhodopsin during bleaching is enzymatically isomerized to 11-cis-retinal in the retinal pigment epithelium (RPE), then shuttled back to the rod photoreceptor cells for rhodopsin regeneration. The mechanism for regeneration of bleached visual pigments in cone photoreceptors appears to be different. Protein-protein interactions were sought in bovine retinal pigment epithelial (RPE) microsomes by reciprocal immunoprecipitations using antibodies to several known visual cycle proteins. Proteins were identified by Western analyses, MALDI-TOF MS and LC MS/MS (Qtof2). Kinetic parameters for 11-cis-retinol dehydrogenase (RDH5) activity with and without CRALBP were measured. CRALBP, RDH5, RPE65, RGR opsin, IRBP, CRBP and RBP co-precipitate from RPE microsomes with anti-CRALBP antibodies. Similar results were obtained with anti-RDH5, anti-CRBP and anti-RPE65 antibodies. CRALBP was found to enhance the affinity of purified RDH5 for 9-cis- and 11-cis-retinoids, supporting a functional interaction. The results support the existence of an RPE retinoid metabolizing protein complex. Further proteomic analyses are directed toward identifying other possible components of visual cycle protein complexes.
Supported in part by NIH grants EY06603, EY014239, EY01730, EY02317, a Research Center Grant from The Foundation Fighting Blindness, and funds from the Cleveland Clinic Foundation and Research to Prevent Blindness, Inc.

S. Gutierrez, E. Chernokalskaya, A. M. Pitt;
Millipore corporation, Danvers, MA, United States.

The standard method for protein identification entails techniques including two-dimensional electrophoresis (2-DE), tryptic in gel digestion, and mass spectrometry. These techniques combined with ultrafiltration were utilized to both compare and characterize the different protein expression profiles between malignant epithelial cell lines. Cell lysates of a highly invasive (MDA-MB-231, ATCC # HTB-26) and a non-invasive (MCF-7, ATCC# HTB-22) human breast cancer cell lines were prepared. Proteins and other molecules over 10-20,000 daltons were concentrated using centrifugal ultrafiltration prior to loading and running the 2-DE. Resolution of the low abundance proteins was found to be enhanced by first concentrating the cell lysates prior to 2-DE. Stained gel spots were imaged and analyzed. A number of proteins appeared to be differentially expressed or have changes indicative of post-translational modifications. Spots of interest were excised and the proteins in-gel digested with sequencing grade trypsin. In-gel digestion, reversed-phase (C18) purification and sample preparation were all performed in a novel flow through SPE 96-well plate. Proteins were identified by MALDI-TOF mass spectrometry and database searches. The application of these proteomics techniques combined with the additional cell lysate and extracted peptide concentration allowed improved protein detection and identification. These methods of cell expression analysis were found to be a reliable and convenient means to identify and characterize differences between the proteins and peptides present in these cell lines.

R. J. McNALL, M. J. Adang;
University of Georgia, Athens, GA, United States.

The crystal proteins of Bacillus thuringiensis (Bt) are widely used in transgenic crops and commercially available pesticides. When these toxins are ingested by a susceptible insect, death occurs after binding events in the midgut of the insect. Manduca sexta is the model insect for Bt mode-of-action studies and has been used to identify two candidate receptors. However, despite its widespread use, the protein components of M. sexta midgut are mostly unknown. Vesicles prepared from the brush border of M. sexta midgut were analyzed using one- and two-dimensional gel electrophoresis (2DE). Gels were silver stained as well as blotted to filters for Western blot analysis. Proteins separated by 2DE revealed multiple species at a given molecular size, and 2DE blots probed with biotinylated Bt Cry1Ac toxin were more complex than blots from one-dimensional separations. Vesicle proteins that were treated with PIPLC were similar between 1DE and 2DE blots analyzed with anti-CRD antibody. Mass peptide fingerprints (MPF) were generated for several spots from 2DE gels and databases were searched. Search results generally did not produce hits to M. sexta proteins; however hits with high Z-scores were produced. To verify MPF results, Western blots were done. From these blots, actin, aminopeptidase N, and membrane alkaline phosphatase were confirmed as accurate protein identifications. This study demonstrated that using 2DE, an assortment of proteins is revealed that are not seen in one-dimensional analyses. The 2DE technique also demonstrated that Cry1Ac binds to actin and alkaline phosphatase in addition to the previously reported aminopeptidase N. This study also underscores the need for additional genomic information of agriculturally important pests.

K. C. Hansen, G. Schmitt-Ulms, R. J. Chalkley, M. A. Baldwin, A. L. Burlingame;
UCSF, San Francisco, CA, United States.

For low level protein samples of high complexity it can be difficult, using current ICAT protocols, to obtain protein assignments of satisfactory confidence. Multi-dimensional liquid chromatography coupled to downstream mass spectrometry is a powerful tool for high sensitivity protein identification but suffers from an intrinsic lack of differential profiling capabilities.
In an attempt to compare the protein content of low quantity samples of high complexity a protocol was established that combines the differential profiling strength of cleavable isotope-coded affinity tag (cleavable ICAT) reagents with the high sequence coverage provided by multi-dimensional liquid chromatography. Prominent concerns during protocol optimization were to minimize sample losses and establish a robust procedure that solely employs volatile buffer systems and therefore is highly compatible with mass spectrometry. The strategy was applied to three different types of samples: (i) immunoaffinity purified crosslinked complexes containing the prion protein, (ii) the membrane fraction of lung alveolar type II cells and (iii) bait purified nucleopore complexes. ICAT labeled and non-labeled tryptic peptides were analyzed by nanoflow LC-ESI MS/MS and LC-MALDI MS/MS. In all three studies a large number of proteins were identified using stringent match criteria, including several that have been identified by previous experiments.
The technical methodology developed in this work allows for analysis of large protein complexes and improves the sequence coverage obtained on low level samples, thereby resulting in more protein identifications of higher confidence. Mass spectrometry was supported by NIH NCRR grant #RR01614

R. Shen, C. Shieh;
ThermoFinnigan corp., San Jose, CA, United States.

Reversible phosphorylation is an important post-translational modification associated with many proteins that have a regulatory function in cell cycle control, receptor-mediated signal transduction, cell differentiation, and nutrient metabolism. Phosphoproteins are generally found in low quantities within cells. This low abundance, coupled with the higher acidity of phosphopeptides, may compromise the efficiency of their ionization in a mass spectrometer operated under positive ESI mode. Phosphoprotein analysis in a complex mixture is, therefore, technically challenging. An efficient method of enriching phosphopeptides can greatly facilitate LC/MSⁿ characterization of phosphoproteins.
Immobilized metal ion affinity chromatography (IMAC) is a technique that has been used to purify phosphoproteins and phosphopeptides^(1-3). In this procedure, the phosphate moieties of phosphoproteins/phosphopeptides coordinate with Fe⁺³ or Ga⁺³ ions chelated on a matrix of chromatographic beads containing iminodiacetic or nitrilotriacetic acid ligands.
In this paper, we present a fast, simple protocol for enriching phosphopeptides in a complex mixture using magnetic bead-based IMAC prior to LC/MSⁿ analysis in an ion trap mass spectrometer. The newly developed software also will automatically perform MS ³ analysis on all phorsphopeptide fragments. The results from MS/MS/Ms analysis will analyze and searched by Sequest software. Phorsphoproteins can bed quickly identified by this method.

J. S. Vogel, D. Hillegonds, M. Palmblad, B. Buchholz, P. G. Grant;
Lawrence Livermore National Laboratory, Livermore, CA, United States.

Fluorophosphate (FP) moities bind covalently to serine, providing a marker for a common hydrolase function in protein isolates. Previous studies used dissected and homogenized tissues exposed to excesses of FP to discover the complement of hydrolases . Accelerator mass spectrometry (AMS) affords such high sensitivity for 14C-modified proteins that in vivo targets of FP ingested at assured sub-toxic doses can be discovered and quantified as functions of tissue, dose, or chemical challenges. Mice were exposed through food to [14C]-diisopropylfluorophosphate (DFP) at 0.1 to 500 microgram per kilogram, with euthanasia and tissue harvesting 48 hours post dose, after all unbound DFP had cleared. No signs of physical or behavioral toxicity were noticed. Separate IEF gel strip and PAGE molecular weight analyses provided virtual 2D separation of targeted proteins for AMS quantitation to 0.1 attomole sensitivity. Dominant binding proteins were identifiable from the 14C levels at particular pI/MW loci in the two gel strips, but can also be eluted for MS analysis. Butyrlcholinesterase, plasminogen, and chymotrypsin were among the 10 or so prominent targets in plasma, while brain cytosol shows a wider range of nearly equivalent targets. The use of virtual 2D separation reduces AMS usage to a manageable 1 day of measurements (200 samples) from the 10 weeks needed for a full 100 by 100 grid of a real 2D gel (10,000 samples). Protein targets of drugs, nutrients, or modifications can be discovered at low 14C exposures in vivo with AMS. This work was performed in part under the auspices of the U.S. Department of Energy by University of California Lawrence Livermore National Laboratory under Contract No. W-7405-Eng-48.

A. S. Woods;
NIDA IRP, NIH, Baltimore, MD, United States.

Cation-¼ bonds and amines- or ammonium-aromatic interactions are important factors in ligand-receptor interaction. Many drugs of abuse, neurotransmitters and ganglionic blockers are tertiary and quaternary ammonium compounds. Cation-¼ bonds and amines- or ammonium-aromatic interactions are important factors in ligand-receptor interaction. We have previously shown that quaternary amines or two adjacent Arg could form non- covalent complexes with the Carboxyl group of Glu or Asp. Using MALDI MS we have found that peptides containing aromatic amino acids such as Phe, Trp and Tyr form non-covalent complexes with quaternary ammonium compounds or peptides containing two or more adjacent Arg. If the aromatic residue is a phosphorylated Tyr, two rather than one molecule will add. The relative intensity (RI) of the first addition is about 15-20 % of the second one. However if the peptide contains an Arg residue within 1-2 residues of the phosphorylated Tyr, the addition of one or two molecules gives peaks of equivalent RI. The data suggests that the phosphate group forms a more stable complex with quaternary amines or guanido groups than the ¼-cloud of the aromatic residues. However if the side chain of the Arg on the same peptide as phosphorylated Tyr is within the interaction sphere of the phosphate group, it probably partially interacts with it, thus making its acidity equivalent to that of the aromatic ring.

M. Pelzing, C. Neusüß, P. Palloch;
Bruker Daltonik GmbH, Leipzig, Germany.

Beside LC-MS and MALDI-TOF, CE-MS has become a main analytical method for proteomic studies. CE is known for its high separation efficiency and low sample amount, however being classified to be less sensitive than above mentioned techniques. The direct "sheathless" CE-MS coupling is known to overcome the sensitivity problems by increasing the detection limits by an order of magnitude¹, however loosing robustness due to difficulties in generating a stable spray and in realizing a robust conductivity connection of the tip. Here, we show that the coupling of CE via an orthogonal triplet ESI-sprayer to an ion trap mass spectrometer is a robust and sensitive method for proteomics. By optimization of sheath liquid parameters, separation conditions and ion trap parameters peptide standards with an injected amount of less than one fmol could be detected.² On-line sample stacking can increase the concentration sensitivity significantly. LODs in the low nmol/l-range have been achieved. This sensitivity is close to published values for sheathless CE-MS and nano-LC-MS, respectively. Moreover, with capillaries of about 50 cm in length separations in less than 10 min are possible resulting in a throughput of up to four analyses per hour. This is a factor of 4—12 times faster than nano-LC separation, being the state-of-the-art techniques for proteomic studies.
A comparison of CE-MS/MS and nano-LC-MS/MS regarding sensitivity, sequence coverage and separation time for analysis of tryptic digests will be presented.
References:
¹ J. F. Kelly et al., Anal. Chem., 69 (1997), 51-60
² C. Neusüß, M. Pelzing and M. Macht, Electrophoresis, 23 (2002), 3149-3159

E. C. Almira, N. Panayotova, W. G. Farmerie;
University of Florda Biotechnology Program, Gainesville, FL, United States.

Deciding which electrokinetic injection parameters to be applied on any given sample plate is not always easy especially in a core facility setup where users’ sample DNA concentration and quality vary widely among wells, plates, projects and laboratories. To get the maximum sequence output from any given run, one has to use the optimum injection conditions to minimize or at least balance the number of wells giving low sequence signals (i.e. underloading) and those showing low capillary current (i.e. overloading). Maintaining a balance between overloaded and underloaded samples has been a problem and oftentimes leads to several reruns of the same sample plate. We approached this problem initially by normalizing template DNA concentration among wells to a narrower range. We later replaced our time-consuming normalization protocol with TempliPhi amplification method, which reduces both the steps of template preparation and variation in DNA concentration. Despite the improvement in passing rate, we still encounter problematic plates on a regular basis. Based on a recent report (1), we tested the effect of resuspending the sequencing reaction products in low concentrations of agarose prior to electrokinetic injection. After several unsuccessful attempts, we modified the resuspension solution by adding MegaBACE loading solution to it. Compared to plain loading solution, the new mixture resulted in 20-50% improvement in sequence yield. Agarose appeared to minimize overloading and its negative effects on sequence quality. Preliminary analysis shows the significant rise in total phred20 bases is attributed to increase in either passing rate or read length, or both, depending on the overall quality of DNA in the sample plate. 1. Vatcher, G. et al, BioTechniques 33:3, pp532-539, 2002.

A. Wheaton, L. Joe, S. Bay, J. Briggs, T. Hatch, Y. Lou, C. Brown, O. Ostadan, L. Krotzer;
Applied Biosystems, Foster City, CA, United States.

Core laboratories using traditional methods for high throughput microsatellite analysis can face multiple obstacles, such as assay design, sample loading and detection, and data interpretation. With an increasing emphasis on complex diseases, laboratories need cost effective, precise tools to rapidly scan the genome for variations. Applied Biosystems has taken a system-wide approach to remove these barriers by: developing a pre-configured assay (the ABI PRISM® Linkage Mapping Set v2.5); and introducing a production level capillary electrophoresis instrument, along with highly automated analysis software.
The Applied Biosystems 3730 DNA Analyzer, a 48 capillary system, offers the ideal platform for high throughput microsatallite analysis. With the capacity for 40 runs per day, the 3730, in concert with 5-dye chemistry, can generate over 38,000 genotypes per day. The system is completed with GeneMapper™ version 3.0, a fully integrated analysis solution that simplifies the allele calling process. We describe the benefits of this system for high throughput automated microsatellite analysis.

B. F. Johnson, S. R. Berosik, S. J. Bay;
Applied Biosystems, Foster City, CA, United States.

Run protocols using a new separation medium, POP-7™ polymer, on the Applied Biosystems® 3730 DNA Analyzer have been developed. Sequencing data from the Applied Biosystems 3730 and ABI PRISM® 3700 DNA Analyzer was analyzed with ABI PRISM® DNA Sequencing Analysis Software v 5.0. Basecalling error for a defined test sequence was determined for the basecalls generated by the ABI (old basecaller) and KB™ basecallers (new basecaller) in DNA Sequencing Analysis software version 5.0. Quality values (QV) for the test sequence analyzed by Sequencing Analysis Software v 5.0 were obtained by secondary analysis with Phred software. Another set of QV produced by the KB™ basecaller was obtained for comparison. The performance of POP-7™ polymer on the Applied Biosystems 3730 DNA Analyzer versus POP-5™ polymer on the ABI PRISM 3700 DNA Analyzer was compared by both the actual error rate and by sequence quality values. A rapid protocol for the 36cm array on the Applied Biosystems 3730 yields a Phred Q20 cutoff of over 550 nucleotides in a run time of 35 minutes. A standard protocol for the 36 cm array delivers over 650 Phred Q20 bases in a run time of 60 minutes. A long read protocol for the 50 cm array on the Applied Biosystems 3730 yields a Phred Q20 cutoff of better than 800 nucleotides in a run time of 120 minutes. The read length obtained with the Applied Biosystems 3730 is consistently 200 nucleotides greater than the ABI PRISM 3700 DNA Analyzer. Read lengths based upon quality values obtained with the KB™ basecaller are equal to or greater than those obtained with Phred software.

S. Bay¹, B. Johnson¹, E. Nordman¹, J. Goudberg¹, J. Burrows¹, J. Briggs¹, K. Yung¹, O. Ostadan¹, J. Yeager¹, M. Kojima²;
¹Applied Biosystems, Foster City, CA, United States, ²Hitachi High Technologies, Naka, Japan.

The Applied Biosystems 3730 DNA Analyzer is a new production-scale capillary electrophoresis instrument designed to increase throughput of large-scale sequencing and genotyping projects. The 3730 DNA Analyzer is a 48-capillary array instrument that combines highly sensitive detection optics, and automated fluid and sample handling with Applied Biosystems’ genetic analysis reagents to yield high-quality data.
Key features for 24-hour unattended operation include: an integrated autosampler; a plate stacker that accommodates up to sixteen 96- and/or 384-well plates; highly sensitive detection optics; automated polymer replacement; internal and external barcode readers; and extended temperature control of the capillaries. A new, high resolution and high speed separation matrix, POP-7™ Performance Optimized Polymer, combined with BigDye® Terminator sequencing reagents typically yields trimmed Phred Q20 read lengths greater than 800 bases in 2 hours using a 50-cm capillary array and 550 bases in 35 minutes with a 36-cm capillary array. Using the same separation matrix on a 36-cm capillary array, microsatellites up to 350 base pairs are sized with less than 0.15 bp precision in 35 minutes.

K. M. Gunning, E. E. McIntyre, K. M. Chansky, A. M. Wheaton, Q. C. Doan, C. L. Brown, C. M. Wike;
Applied Biosystems, Foster City, CA, United States.

The ABI Prism® Genetic Analyzer has become the instrument of choice for the core laboratory. With the recent release of new Data Collection software, v 1.1, the new features are providing researchers with additional benefits to streamline their workflow and generate quality data. Flexibility in the handling of applications and data distribution, coupled with the addition of new Big Dye Terminator™ Chemistry has allowed the service laboratory industry to provide a timely turnaround and reliability that is expected by the customer.
We will discuss the benefits of the new features, the quality of data, and the implementation of easy to use downstream application software such as Seqscape™ Software v 2.0 and Genemapper™ Software v 3.0 that further enhance the productivity of the service laboratory.

J. Frenck, K. Sperle, G. N. Picerno, D. L. Stabley, A. Davis-Williams, G. Hobson, K. Sol-Church;
Alfred I duPont Hospital for Children, Wilmington, DE, United States.

Capillary Electrophoresis has recently become a multifaceted tool, specifically in the area of quantitative PCR. We have shown that Capillary Electrophoresis can be used as an accurate and reproducible method to assess gene duplication in patients affected with Pelizaeus-Merzbacher disease (PMD). PMD is an X-linked recessive dysmyelinating disorder of the central nervous system. More than half of PMD cases are due to duplication of a genomic region that includes the proteolipid protein gene (PLP1) located at Xq22, while point mutations account for 5 %-25% of cases and complete deletion for <1%. The remaining cases are undiagnosed at the molecular level. By using an internal standard (dystrophin) and comparing the relative peak areas of the PLP and dystrophin gene PCR products, a ratio is determined from which duplication and deletion can be determined. Compared to other quantitative analysis methods used in our lab (Eagle Eye and STORM analysis) we have shown that the capillary electrophoresis method is the most precise and reproducible. We have also found this technology to be quite powerful in the high-resolution mapping of the extent of duplication and deletion of the PLP region in patients with PMD.

D. A. Iacobas, D. C. Spray;
Albert Einstein College of Medicine, New York, NY, United States.

Users of spotted gene arrays face many challenges to minimize non-biological variability and increase the sensitivity and the specificity of the method. In this report we analyze the principal sources of errors and limitations of the cDNA microarray technique. Three experimental strategies (Label Flipping, Single Label Reference and Multiple Yellow methods) and their related experimental designs and normalization procedures are presented. All three strategies considerably improve data accuracy and reliability as compared to the original method. Each strategy is illustrated by an experiment performed in our laboratory using the mouse 27.4k cDNA arrays in which various normalization procedures were tested. The experiments compared gene expression in mouse neuroblastoma N2A cells transfected with gap junction (connexin) cDNA, in brains of wild type and connexin knockout mice and in control and mechanically stressed mouse osteoblasts. The Multiple Yellow method has been found to be the most cost-effective dual label technique yet developed, and the normalization it provides is most effective in reducing the interchip variability.

T. Cheung¹, P. J. Collins², T. B. Doan², K. W. Shannon², X. Liu¹;
¹University of Colorado at Boulder, Boulder, CO, United States, ²Agilent Technologies, Palo Alto, CA, United States.

Transforming growth factor-beta is a multifunctional growth factor whose best-known function is to inhibit cell growth and suppress tumor formation. Loss of TGF-β growth inhibition is one of the most common cellular events in the pathogenesis of human breast, pancreatic and colon cancers. TGF-β signals through a heteromeric signaling complex consisting of Smad2, 3 and 4. Disruption of the Smad signaling complex often leads to tumor formation. We have used both 60-mer oligonucleotide and cDNA microarrays to investigate the consequences of Smad4 inactivation to the TGF-β controlled transcription program in tumor cells. These high density DNA microarrays, generated using Agilent’s SurePrint inkjet technology, were used to profile global transcriptional regulation in breast, colon and pancreatic Smad4-null tumor cell lines in response to TGF-β. Data from both microarray types showed a high degree of correlation in demonstrating that TGF-β induces transcriptional activation and repression of genes involved in signal transduction, cell adhesion and transcriptional regulation across the range of cell lines tested. Comparison with expression profiles from Smad4-positive cell lines indicates that inactivation of Smad4 does alter transcriptional programs of TGF-β signaling but is not absolutely required for TGF-β signaling from the cell surface to the nucleus. The composition of the Smad signaling complex may control the specificity of TGF-β signaling to control expression.

P. J. Collins¹, T. Cheung², T. B. Doan¹, K. W. Shannon¹, X. Liu²;
¹Agilent Technologies, Palo Alto, CA, United States, ²Department of Chemistry and Biochemistry, UCB215, University of Colorado-Boulder, Boulder, CO, United States.

Transforming growth factor-β is a multifunctional growth factor whose best-known function is to inhibit cell growth and suppress tumor formation. Loss of TGF-β growth inhibition is one of the most common cellular events in the pathogenesis of human breast, pancreatic and colon cancers. TGF-β signals through a heteromeric signaling complex consisting of Smad2, 3 and 4. Disruption of the Smad signaling complex often leads to tumor formation. We have used both 60-mer oligonucleotide and cDNA microarrays to investigate the role of Smad4 in the TGF-β controlled transcription program in tumor cells. These high density DNA microarrays, generated using Agilent’s SurePrint inkjet technology, were used to profile global transcriptional regulation in breast, colon and pancreatic Smad4-null tumor cell lines in response to TGF-β. Data from both microarray types showed a high degree of correlation in demonstrating that TGF-β induces transcriptional activation and repression of genes involved in signal transduction, cell adhesion and transcriptional regulation across the range of cell lines tested. Data from a number of studies is presented comparing expression profiles from Smad4-null tumor cell lines to those from either Smad4-transfected cell lines or normal cell lines. These data indicate that the composition of the Smad signaling complex controls the specificity of TGF-β signaling.

G. S. Huang¹, y. Liu²;
¹China Medical College, Taichung, Taiwan Republic of China, ²National Chung Hsing University, Taichung, Taiwan Republic of China.

We incorporated gene expression information from cDNA microarray into flux analysis to simulate yeast diauxic growth. Expression ratios of both growth phases were applied to assign the split ratio at glyoxylate shunt during simulation, in which the equation was mathematically unsolvable due to the singularity and artificial split ratios, which were traditionally introduced without biological evidence. In addition, the directionality of microarray dataset was used as a further constraint during simulation. Metabolic fluxes obtained by this modified approach are in general consistent with microarray analysis. However, discrepancies occurred when the quantity of fluxes was compared, probably due to the substantial reduction of substrates at phase II in which the increase in the enzymatic levels was not proportional to the increase of substrate flow, as would be predicted from microarray dataset. The modified flux analysis might have brought a new approach to investigate other cellular pathways.

J. Tso, R. Parker, T. Sana, J. Fredrick, A. Thompson, P. K. Wolber, J. McMillan, A. Schleifer, M. Caren;
Agilent Technologies, Palo Alto, CA, United States.

Whole genome screening using current high-density oligonucleotide microarrays has yielded valuable information to help researchers identify key biomarkers or pathways of interest. Lower density microarrays capable of screening a few hundred to a few thousand genes of interest can be used to perform detailed screening of specific disease states or evaluate the toxicity of certain drugs against target organs. This progress has prompted the development of a novel hybridization platform that would accommodate lower sample volumes and permit parallel screening of multiple microarrays on a single slide.
Agilent Technologies new hybridization technology, coupled with Agilent’s flexible SurePrint microarray manufacturing technology, allows simultaneous processing of multiple microarrays on a single 1"x3" microscope slide. Initial studies with this format enable the ability to easily perform up to 8 multiple microarray hybridizations within a single slide experiment using less than 50 microlitre volume of sample material per microarray. Hybridizations are conducted using Agilent’s current hybridization conditions and standard hybridization oven. The feasibility of using this format was demonstrated by evaluation of intra- and inter-array hybridization consistency using cyanine labeled K562 and HeLa cRNA targets, and control oligonucleotide targets. Log Ratio data from subsets of specific pathways in each of the 1.9K feature count microarrays were compared to larger, 8.5K feature count microarrays containing the same probe content. Our results demonstrate that inter-array (same slide) consistency is highly correlated (r=0.85-0.95). Furthermore, the correlation between identical probes in both the large and small feature count configuration was highly correlated (r=0.90). Comparison of signal intensity from a target dilution series showed linearity (r>0.95) over a 10-fold concentration range.

P. J. Collins, T. B. Doan, D. Amorese, K. W. Shannon;
Agilent Technologies, Palo Alto, CA, United States.

In recent years, DNA microarrays have become key tools in providing an understanding of cellular gene expression patterns. Gene transcripts have typically been represented on microarrays by whole cDNA molecules or by multiple short oligonucleotides, but in general selection of these sequences is not based on their performance in expression profiling experiments. We report here on the development of the Human 1 Oligonucleotide microarray using a novel experimental method for selection of 60-mer oligonucleotide probes and demonstrate the performance of these probes in profiling global gene expression. One probe is present on this microarray for each of more than 17,000 full-length genes in Incyte Genomics’ LifeSeq Foundation database. These probes were designed to consensus regions across all gene transcripts, so are splice variant non-specific. Each is annotated with information on protein function, as well as data describing its performance in validation experiments. Probes are printed in a 22,575 feature microarray using Agilent’s SurePrint technology. Data are presented which demonstrate the performance of these microarrays with regard to their consistency, accuracy and sensitivity. Furthermore, data from a range of expression profiling experiments is compared with that derived from Human 1 cDNA microarrays. The results presented demonstrate the power of this experimental method for selection of microarray probes to provide extremely accurate and reliable gene expression data.

G. Probst¹, J. Posch¹, O. Hagenbuchle², J. Wyniger², G. L. Porter³;
¹Tecan Austria, Grodig, Austria, ²Swiss Institute for Experimental Cancer Research (ISREC), Lausanne, Switzerland, ³Tecan U.S., Research Triangle Park, NC, United States.

High signal to response ratios, low background and slide-to-slide reproducibility of microarray experiments are all desirable to increase the range and reliability of conclusions drawn from expression profiling experiments. Automation of the hybridization process has the potential to reduce experimental variation as well as increase the number of experimental repetitions possible. In this study, we compared the reproducibility of the probe signal, background non-specific signal and the signal-to-noise response performed by normal manual hybridization protocol and the hybridization done by an automated hybridization processor (Tecan HS 4800 Hybridization Station). The effect on these responses of three agitation treatments was also investigated. The results show that there is a 6-fold reduction in the signal variation and a 1.3-fold improvement of signal-to-noise of the hybridization process when using the Tecan HS 4800 Hybridization Station in contrast to the manual method, and in addition that the periodic agitation mode of the hybridization station improves the performance significantly.

J. J. Broz, C. P. Weibel;
Point Technologies, Inc., Boulder, CO, United States.

Microarray expression analysis requires the application of reproducible protocols to obtain consistent spot morphology and fluorescence intensity; however, due to the different combinations of spotting environments, microarray pins, spotting buffers, arrayed products, slide chemistries, and deposition robots the process has inherent variability. During microarray printing, spotting pins use capillary action to uptake μLs of buffered solution and surface tension interactions to "print" onto a coated slide. Pin diameter, wicked volume, liquid surface tension, pin and coating surface energies, dwell time, and pin z-force affect delivery volume. Spot morphology variations prior to hybridization (i.e., coalescence, size and shape variations, non-uniformity) can dramatically affect fluorescence intensity and compromise expression analysis accuracy. Little attention has been paid to customizing pin characteristics for optimal fluid delivery to produce a "Known Good Spot™". In this work, proprietary micromachining techniques that can alter spotting pin surface properties and shape characteristics are combined with metrology data from a novel image analysis method. Multiple "off-the-shelf" slotted microarray pins (Point Technologies, Telechem, and Majer) were used to print several microarrays of buffered solution onto poly-L-sine coated slides in an environmentally controlled chamber (22°C and 40-45% RH). For each pin type, baseline spot quality CVs were determined from the unhybridized microarrays. Subsequently, the PT pins were customized and the spot shape CVs were incrementally improved. Image derived quality metrics of the unhybridized spots provided key information for modifying the pin geometry to reduce CV values. Precision customization can be used to reduce pin-to-pin variations and optimize pin characteristics for individual microarraying application such that "Known Good Spots™" are consistently produced.

D. Luk, A. Tuyet-Doan, V. Le, K. Gulyas, R. Guettler;
GeneMachines, San Carlos, CA, United States.

Oligonucleotides are fast becoming the probe molecule of choice in microarray research. GeneMachines is uniquely positioned to support both the synthesis and fabrication of oligonucleotide arrays. Traditionally, self-spotted microarrays have utilized cDNA for expression features. The availability of low-cost oligos and high-quality genomic sequence is driving a transition to oligonucleotide arrays. The GeneMachines PolyPlex oligo synthesizer and OmniGrid microarrayers combine to deliver straightforward, high-quality, and cost-effective oligonucleotide array production. These two products put the power and flexibility of oligonucleotide arrays easily within reach of production facilities.
As microarray researchers increase the desired number of genes to monitor in a single expression experiment, the cost associated with producing the required oligos has increased dramatically. Although synthesis costs have fallen substantially for commercial oligos, the cost to purchase array-ready commercial oligos remains high. The PolyPlex synthesizer and OmniGrid microarrayer together deliver high-quality oligo arrays at a fraction of commercial costs by utilizing crude 5'-amino-modified oligos on OmniGrid microarray slides. As a result, tedious and expensive oligo purification steps are eliminated. Synthesis in 96-well format provides error-free handling in the transition between synthesizer and microarrayer. During the processing of the printed slide, the truncated oligos that did not receive the amino linker are washed off during the pre-hybridization steps. This complete synthesis and microarray package offers a flexible in-house alternative to costly commercial oligo arrays.

D. Ilsley¹, S. Coates²;
¹Agilent Technologies, Palo Alto, CA, United States, ²Agilent Technologies, Little Falls, DE, United States.

Fluorescent labeling of nucleic acids has become the preferred method of DNA detection for a wide variety of applications, including gene expression profiling using DNA microarrays. One method of generating fluorescently labeled targets for gene expression profiling is direct incorporation of dye-dNTPs during the reverse transcription reaction. Unfortunately, dye-dNTPs are poor substrates for reverse transcriptases. We have designed tiling microarrays to gain a better understanding of dye-dNTP incorporation by several reverse transcriptases. These microarrays contain 60 mer probes spanning the 5’ to 3’ sequence of genes. The reverse transcription reaction products were generated by varying time, substrates, enzyme concentration, and reaction components and were analyzed using these microarrays. Microarray analysis offers several advantages over traditional gel electrophoresis methods including high sensitivity, ability to examine multiple substrates simultaneously, and individual product analysis. These results were consistent with previous studies performed using a simple model system to kinetically characterize dye-dNTP incorporation.

L. E. Cammish;
NextGen Sciences Ltd, Huntingdon, United Kingdom.

Expression analysis of key indicator proteins can provide a useful prognostic tool in breast cancer studies. A problem associated with use of traditional analytical techniques, however, is the amount of scarce breast cancer sample that is required for effective processing and the current lack of a ‘tool kit’ for high throughput, parallel analysis. The use of protein microarrays to miniaturise and enable high throughput protein expression profiling is a valuable tool for studies of this description. NextGen Sciences are developing a number of protein microarrays for use in this application. Aspects of NextGen Sciences work that will be presented will include experimental data from studies of complex breast cancer samples illustrating optimised methods of attachment of antibodies and functional protein to biochip substrates, methods of accessing antibody cross reactivity via use of antigen biochips, and methods by which assays

K. Martin, T. H. Steinberg, J. M. Beechem, W. F. Patton;
Molecular Probes, Inc., Eugene, OR, United States.

Ultrasensitive detection of minute amounts of phosphorylated proteins/peptides is a key requirement for unraveling many of the most important signal transduction pathways in mammalian systems. Historically, phosphorylation detection methods have relied on either radioisotopes [primarily 32P] or antibodies (to phosphorylated amino-acids). Both of these methods suffer from relatively well-known defects. In this study, a small-molecule-fluorophore detection technology is described (Pro-Q Diamond dye), which is capable of ultrasensitive detection of phosphorylated amino-acids in peptides and proteins on both microarrays (this study) and on 1-D/2-D gels (see Steinberg, T., et. al., abstract, this meeting). Pro-Q Diamond dye, can function as a universal sensor of phosphorylation. DNA, RNA and sulfated glycoproteins fail to be detected with Pro-Q Diamond dye. Pro-Q Diamond dye staining is rapid, simple to perform, readily reversible and fully compatible with modern microchemical analysis procedures, such as MALDI-TOF mass spectrometry. Pro-Q Diamond dye phosphorylation detection technology is demonstrated on phosphoproteins, phosphopeptides and in kinase reactions in microarray format (~ 185 - 200 µm spots). The detection sensitivity of phosphoproteins on a variety of different commercially available protein array substrates was 1 - 2 pg for monophosphorylated proteins and at least 650 fg - 1 pg for multiphosphorylated proteins (2 - 5 phosphates per protein). Pro-Q Diamond dye characterization of enzymatic phosphorylation of immobilized peptide targets, differentiated specific and non-specific peptide labeling at picogram to subpicogram levels of peptide.

S. Meyer-Plath;
Advalytix AG, Brunthal, Germany.

Protein and DNA hybridization has become a standard tool in proteomics / genomics research. In order to guarantee fast and reproducible hybridization results the diffusion limit must be overcome. SAW (surface acoustic wave) micromixer chips efficiently agitate smallest sample volumes (down to 10µl and below) without introducing any dead volume. The advantages are
• reduced reaction time (typically a factor of 5)
• increased signal-to-noise ratio (typically a factor of 3)
• improved homogenuity across the microarray
• better slide to slide reproducibility
The SAW micromixer chips are the heart of the ArrayBooster™ which is compatible with all microarrays based on the microscope slide format. The instrument is small, lightweight and easy to use. It offers maximum flexibility as up to four slides with different protocols can be hybridized in parallel. The ArrayBooster™ has also been shown to substantially improve the results of protein microarrays.
We also demonstrate that high speed agitation can be used to measure binding kinetics without a flow cell. The streaming velocity induced by the SAW chips is sufficient to prevent sample depletion in the vicinity of the spots and equilibrate the sample solution. As a result the binding constant and the reaction kinetics rates can be determined for many different molecular reactions in parallel.

H. McManus, P. Patel, J. Koehler, G. Dantsker;
Nanostream Inc., Pasadena, CA, United States.

Microfluidic systems have the potential to revolutionize biochemical analysis and high throughput screening. In order to respond to the dynamic nature of this expanding industry, we have developed a "modular" approach to microfluidic development. A wide range of microfluidic systems can be economically constructed using our microfluidic componentry in a manner of hours rather than weeks, greatly reducing the product development cycle.
Drug companies are coming under increasing market pressure to broaden their pipelines. Many companies are searching for novel technologies that will automate routine, but time consuming, steps in the drug discovery/development cycle. Increasingly demands are been made to increase throughput in sample analysis.
This talk will focus on the development of a microfluidic-system that allows 24 simultaneous separations. The separations chip has 24 columns with onboard control of sample injection, flow and solvent mixing to produce chromatograms of excellent quality but much faster (10x) than a standard HPLC. Preliminary results from the analysis of a pharmaceutical library will be presented.

C. Buhlmann, T. Preckel, M. Kuschel, M. Valer;
LSCA Agilent Technologies, Waldbronn, Germany.

The goals of Lab-on-a-chip technology are to achieve a drastic decrease of analysis time by automation of lab processes, and a reduction of the required sample and reagent volume. We here present a new implementation of cell assays, like antibody staining for cell surface protein expression, on disposable microfluidic chips. The applications based on the controlled movement of cells by pressure-driven flow in microfluidic channels and two color fluorescence detection of single cells. This new technology makes simple flow cytometric studies of fluorescently labeled cells possible in a microfluidic chip-based system. Importantly, we developed staining procedures working ‘on-chip’, eliminating time-consuming washing steps. Cells and staining-reagents are loaded directly onto the microfluidic chip and analysis is started after a short incubation time. These procedures require only a fraction of staining reagents generally needed for flow cytometry and 30,000 cells per sample, thereby demonstrating the advantages of microfluidic technology. The specific advantage of an on-chip staining reaction is the amount of time, cells and reagents saved. This is of great importance when working with limited numbers of cells, e.g. primary cells or when needing to perform routine tests of cell cultures as a quality control step. Applications of the technology are antibody staining of proteins and determination of cell transfection efficiency by GFP expression. For all applications we used reagents commercially available for conventional flow cytometry. Results obtained with microfluidic chips using standard cell lines and primary cells showed good correlation with data obtained using a standard flow cytometer.

S. J. Lightfoot¹, O. Mueller²;
¹Agilent Technologies, Palo Alto, CA, United States, ²Agilent Technologies, Waldbronn, Germany.

Advances in microarray technology have led to the increased demand for techniques appropriate for the analysis of RNA. The two main parameters to be assessed are RNA quality and the quantitation of RNA. Absorbance and fluorescence based measurements are established techniques for RNA quantitation. A new technique for the measurement of RNA is the Agilent 2100 bioanalyzer in combination with the RNA 6000 Nano LabChip® kit. The Agilent 2100 bioanalyzer, the first commercially available microfluidics device, offers the advantage of gauging RNA integrity, in addition to a quantitative measurement. In the current work, we have investigated different techniques for RNA quantitation. The Agilent 2100 Bioanalyzer is compared with UV spectrometry, and the Ribogreen® quantitation system accessing the linear dynamic range, reproducibility, and the effects of common sample contaminants for each quantitation method.

M. Kuschel¹, T. Neumann¹, P. Barthmaier²;
¹Agilent Technologies, Waldbronn, Germany, ²Agilent Technologies, Palo Alto, CA, United States.

Antibodies are commonly produced for diagnostics, as research tools or as therapeutics and require precise testing and quality control (QC). The key quality criteria include protein identification, quantitation, purity and stability. The standard methods for antibody QC include besides others, gel electrophoresis (GE) and capillary gel electrophoresis (CGE). GE has not changed significantly over the past 30 years and does not allow for a high degree of automation. CGE is a well-known alternative to GE and offers important advantages for protein analysis: automation and faster sample throughput. Lab-on-a-chip (LOAC) technology has extended these advantages even further by transferring CGE principles into a glass chip format. It integrates multiple manual experimental procedures such as sample handling, separation, staining, detection and analysis in a single automated process. The first commercial LOAC instrument, the Agilent 2100 bioanalyzer, was used to verify the suitability of this technology for antibody QC. It allows sizing and analysis of proteins from 5 to 200 kDa, depending on the application, and to determine relative concentration based on internal standards, or absolute quantitation based on user-defined standards. Antibodies were analyzed for size, concentration, purity, and stability. The sizing and quantitation accuracy as well as reproducibility was determined and compared to standard methods. The LOAC protein analysis is comparable in sensitivity, sizing accuracy and reproducibility to gels stained with Coomassie. Resolution and linear dynamic range are improved. Absolute quantitation accuracy and reproducibility is superior to gels and is comparable to Lowry or Bradford. The LOAC analysis allows performing antibody analysis under both reducing and non-reducing conditions in one run. Additional advantages include fast analysis times, ease of use, automated data analysis, and digital data.

C. A. Miller, B. D. Miller, J. Roark;
Agilent Technologies, Palo Alto, CA, United States.

Mass spectrometry has become a core technology for proteomics research and the increasing interest in large-scale proteome characterization, e.g. shotgun proteomics, can readily lead to a bottleneck in data interpretation and review. An advanced informatics tool is introduced which provides an environment for the rapid, comprehensive review of data from large-scale proteomics experiments. Within this environment, protein database search, result review and validation of database search results can be performed. In addition, large datasets can be compared across multiple experiments and the results can be readily summarized at the protein level. For those peptides not identified through database search, an automated de novo sequencing tool is available and the results from this tool are integrated into the informatics environment to enable cross correlation with database search results. The software also offers an open platform for analysis of data from multiple MS techniques and vendors. This work will demonstrate the features and performance of the new informatics software for analyzing data from typical proteomics experiments.

K. A. West^1,2, J. Sun^1,2, L. Yan³, K. Shadrach^1,2, A. Hasan^1,2, M. Miyagi⁴, J. S. Crabb^1,2, J. G. Hollyfield^1,2, A. D. Marmorstein^1,2, J. W. Crabb^1,2;
¹Cleveland Clinic Foundation, Cleveland, OH, United States, ²Lerner Research Institute, Cleveland, OH, United States, ³University Medicine Dentistry of New Jersey, Newark, NJ, United States, ⁴University of North Dakota, Grand Forks, ND, United States.

The retinal pigment epithelium (RPE) is a single cell layer that separates the photoreceptor cells of the retina from their principal blood supply in the choroid. In all vertebrates, the RPE is responsible for vectorial transport of nutrients to rod and cone photoreceptors, removal of waste products to the blood, absorption of scattered light and regeneration of bleached visual pigment. To facilitate studies of RPE physiology, we have initiated the development of a human RPE protein database. RPE cells were isolated from normal adult human donor eyes, subcellular fractions prepared and proteins fractionated by 1D and 2D electrophoresis. Following in-gel proteolysis, tryptic digests were analyzed by LC MS/MS peptide sequencing and/or MALDI TOF peptide mass mapping and the data used to query protein sequence databases. Preliminary analyses have identified 278 proteins and provide a starting point for building a database of the human RPE proteome. Bioinformatic challenges encountered so far in this project will be emphasized.
Supported in part by NIH grants EY06603, EY02362, EY13160, EY014239, EY014240, a Research Center grant from The Foundation Fighting Blindness, and funds from the Cleveland Clinic Foundation.

R. A. Welch, M. Yeager, B. Staats, H. Sicotte, A. Crenshaw, S. Yadavalli, A. Hutchinson, C. Glaser, T. Bandey, S. Chowdhury, B. Packer, E. Miller, M. Kiley, S. Chanock;
NCI/ATC, Gaithersburg, MD, United States.

The National Cancer Institute’s Core Genotyping Facility (CGF) is a high-throughput genotyping laboratory dedicated to provide intramural genotyping for investigators studying the effects of variation in candidate genes in the etiology of cancer. SNP genotyping methodologies are highly dependant on the sequence surrounding the SNP of interest. Public databases describing this sequence context do not always depict allele frequencies or any surrounding polymorphisms that may occur in populations, and using this sequence information can pose potential problems in assay performance including genotyping errors. The CGF has adopted a strict pipeline for assay validation and subsequent genotyping, including 1.) Annotation -- complete annotation of SNP and 600 base pairs of surrounding genomic sequence ("contig"), 2.) Sequencing -- each contig is sequenced (forward and reverse reads) in the SNP500Cancer panel, which consists of 102 individuals from 4 ethnic groups, and determine allele frequencies of requested SNP and other surrounding SNPs, 3.) Genotyping assay development and optimization -- assays are optimized on the SNP500Cancer panel, 4.) Genotyping assay validation -- For each optimized assay, the results are compared to the sequencing results per SNP500Cancer individual. There must be 100% concordance between specific assay genotypes and sequencing genotypes for an assay to be considered validated. This pipeline has proven to improve the throughput and accuracy of genotyping data, and has resulted in the discovery and annotation of genetic variants that could have posed serious genotyping assay problems if left unidentified. All validated assays and SNP sequence and frequency information are publicly available at http://snp500cancer.nci.nih.gov.

C. Kosman, J. Labrenz, C. Gehman, J. Sorenson, C. Heiner, L. Johnston-Dow, G. Fry, S. Schneider, A. Pradhan, A. Pegadaraju, Q. Doan;
Applied Biosystems, Foster City, CA, United States.

New basecalling algorithms have been developed at Applied Biosystems to address current requirements of sequencing customers performing de novo and re-sequencing applications with the goal to improve the accuracy of analysis while retaining the flexibility and robustness of the original ABI basecaller. The KB basecaller has been developed independently of existing basecallers and is offered as an integrated component of the Sequencing Analysis Software v5.0 and SeqScape™ Software v2.0 applications. In this initial release, calibrations are provided to support for the Applied Biosystems 3730 and 3730xl DNA Analyzers. Anticipated software updates will extend this support to the ABI PRISM® 3100-Avant Genetic Analyzer and 3100 Genetic Analyzer. New or improved features beyond current industry-standard solutions include: calibrated quality value predictions Q on each base call and optional mixed-base detection, increased length-of-read for de novo sequencing, reduced clear range error, detection of failed samples and improved robustness to common chemistry or instrument related data anomalies that tend to induce basecalling error. Statistical analyses shown here on diverse data sets demonstrate these features and improvements along with a comparison of the KB basecaller and existing tools. An overview of the user-definable settings provided in the Sequencing Analysis Software v5.0 and SeqScape™ Software v2.0 user interfaces for arranging the integrated basecalling and sequence quality estimation is also provided.

S. Ghosh, G. Delenstarr, S. Connell, P. Wolber, N. Sampas;
Agilent Technologies, Palo Alto, CA, United States.

Microarrays frequently employ samples labeled with fluorescent dyes in two or more colors to compare biological sequences with a control condition. The dyes are typically incorporated into RNA targets in separate labeling reactions potentially causing differential enzymatic incorporation. Inherent chemical difference in dyes with respect to their quantum efficiencies, scanner response, and propensity for non-specific binding to array surface potentially introduce systematic inaccuracies in gene-expression measurements. Consequently, data normalization is a major bioinformatics’ challenge; its success is strongly coupled to the, essential steps preceding it:
1. Background subtraction;
2. Selection of probes (array features) for calculation of offset parameters to remove bias in estimated background baseline.
3. Selection of probes for calculation of normalization parameters.
The Rank Consistency Filter has been developed as a robust probe selection tool. Elimination of pixel and feature level outliers is a pre-condition for filtering. An initial intensity-to-rank transformation is performed, per channel, on all non-control inlier probes. The algorithm clusters those probes populating the central tendency zone — typically defined by no differential expression - of the background-subtracted intensity distribution. Probes populating the low-end of the intensity spectrum are re-clustered for background adjustment. The drivers of the algorithm are coupling of the feature intensity across the dye channels and a Euclidean distance based similarity metric.
The parameters computed using these optimized probes enhance the accuracy and reproducibility of log-ratios for two-channel data. The average standard deviation of log-ratio, across 4 self-self arrays, computed on an intra-array, inter-feature basis are 0.052 and 0.064 for only rank-consistent and all non-control probes respectively. The same metric on an inter-array, intra-feature basis yields 0.023 and 0.027 respectively.

J. Corson¹, J. McMillan²;
¹Agilent Technologies, Santa Clara, CA, United States, ²Agilent Technologies, Palo Alto, CA, United States.

The microarray processing work-flow contains multiple, discrete steps each contributing differing amounts of error to the end-point measurement. A microarray scanner holds a key place in this work-flow, but is often ignored as a source of experimental error. Scientists are frequently bewildered by the selection of scanners available, and lack a comparison method or data needed to make an assessment about impact of the scanner's error on the total experimental error. Since downstream decisions will be made based upon the outcome of these microarray experiments, recognizing the magnitude and minimizing the source of error due to the scanning apparatus is critical.
Typical sources of error from a microarray scanner include noise in the background light, non-uniformity of the scan field, and spectral cross-talk between dye channels. These characteristics are sometimes difficult for the end-user to measure individually, so analytic approaches have been developed to quantify the impact of these sources of error using standard microarray metrics such as detection limit, feature CV%, signal to noise, and log ratio consistency.
Simple methods available to any researcher that will permit the measurement of these parameters for any scanner or set of scanners will be described. Further, a means for relating these performance parameters to biological expression data will be provided.

W. Kopaciewicz, M. Kavonian;
Millipore, Danvers, MA, United States.

ABRF 2003 marks the 5^th year anniversary of ZipTip® pipette tips for the micro-volume sample preparation of peptides and proteins prior to MALDI TOF MS. Over the past five years, researchers have experimented with the devices developing creative ways to use them. In this poster, we will summarize the most notable of these "tricks of the trade" that protein chemists have employed to solve their sample preparation challenges. Such tips include: how does one load a large sample volume? How can the sample be eluted in the highest concentration ? How can the tip be eluted into an electrospray needle. When to use C4 vs. C18 ? When to use SCX vs. C18. In addition, we will discuss the most commonly encountered usage problems and how they have typically been solved. This information should serve as a valuable contribution regarding how micro-scale sample preparation has been adapted to solve a variety sample handling issues in protein chemistry.

R. Martin¹, C. Sutton¹, K. Jackson¹, E. de Oliveira², A. Carne³;
¹Shimadzu Biotech, Manchester, United Kingdom, ²Proteomics Facility, Barcelona, Spain, ³Proteomika SL, Univ. of Barcelona, Barcelona, Spain.

In-gel trypsin digest of protein bands or spots, followed by extraction of the resulting peptides represents a fundamental strategy in proteomics. Peptide mass fingerprinting by MALDI mass spectrometry is used for protein identification. Of the signals generated in the spectrum, only a few are used to identify the protein (typically 30% sequence coverage). When there are a number of members of a protein family, this coverage is not sufficient for accurate identification. The remaining signals are often ignored, but contain considerable information about the absolute identity of the protein and possible structural modifications. MS/MS data is important for accurate protein identification in such cases. A novel MALDI instrument with an ion trap for MS^n analysis and time-of-flight separation of ions for accuracy, was used to focus on signals that did not correlate with the identified protein sequence.
The strategy was applied to a specific model, in which albumin fragments, that are present in biological fluids of patients with various diseases, were separated by 2D gel electrophoresis and characterised for significant biological/structural function. Albumin fragments were typical examples of samples for which the measured mass (from SDS PAGE) bore no significance to the protein sequence database entries, and MS/MS data was required for confirmation. However, common tryptic peptide masses between protein spots, were used for identification-by-association, thereby indicating the value for both MS and MS/MS data in thorough proteomics investigation.

H. Kang, J. Kim, Y. Kim, K. Kwak, J. Kim, J. Lee;
Chonnam National University, Gwangju, Republic of Korea.

Posttranscriptional regulations of gene expression in eukaryotes are of crucial importance for growth and development. The levels of regulation include pre-mRNA splicing, polyadenylation, mRNA transport, translation and stability. Regulation is mainly achieved either directly by RNA-binding proteins (RBPs) or indirectly, whereby RNA-binding proteins modulate the function of other regulatory factors. In recent years, a number of studies has demonstrated that the expression of RBP increased in stress conditions, such as cold, drought, wounding, high salinity, and UV-irradiation, which imply that RNA-binding proteins play roles in plant responses to environmental stresses. To understand the mechanism of RBP-mediated posttranscriptional regulation of genes in plants under various environmental stimuli, a number of target RBPs was identified among complete genome of Arabidopsis and the function of some of the selected genes was investigated. A quantitative real-time RT-PCR was employed for an efficient screening of target genes induced upon various stress treatment. Among 14 potential genes tested, six RBP-encoding genes were highly induced by environmental stimuli, especially cold treatment. To search for the potential target RNA sequences for the RBP, the recombinant RBP was expressed in E. coli, and the recombinant protein was used to screen the target RNA sequences in vitro. To test the role of RBP in response to environmental stresses, the recently developed RNAi knockdown method and overexpression approach were employed. Results indicated that the RBPs were induced by cold treatment, and the specific sequence of the RNA was not required for an efficient binding of the RBPs. The present study would provide decisive clues to understand the function of RBP in stress responses and the mechanism of posttranscriptional regulation of gene expression in plants.

C. W. Sutton¹, R. Martin¹, A. Gooley², F. Hopwood²;
¹Shimadzu Biotech, Manchester, United Kingdom, ²Proteome Systems Limited, Woburn, MA, United States.

Valuable samples from patients, such as tissues obtained by invasive surgery, or biological fluids obtained as part of time course of treatment or disease progression, will require a new approach for proteomics because the opportunities to obtain additional biopsies will be extremely limited. Two-dimensional gel electrophoresis still represents the most convenient method for displaying a large number of proteins. However, gels are not a stable medium for long-term storage of proteins. Continual freeze-thawing of stored gels results in tears in the polymer making them difficult to handle, especially after multiple protein spots have been removed. Further, to analyse samples from the gel it is necessary to use the whole protein spot for protein identification by mass spectrometry.
A new strategy has been developed for the long term archiving of protein arrays separated by 1D- and 2D-gels. Samples from patients can be stored for a prolonged period by electroblotting the proteins from gels onto polyvinyldifluoride (PVDF) membranes. The membranes are supported on MALDI mass spectrometry targets and subjected to a chemical printing method, which has been designed to allow multiple queries of the electroblotted proteins, with a variety reagents (proteases, glycosidases or antibodies). Using piezo-based dispensing only a small portion of any protein spot is modified, sufficient for subsequent analysis by MALDI mass spectrometry. After modification, the XY co-ordinates of the treated part of the protein spots are seamlessly transferred to the MALDI mass spectrometer so that when the membrane (on the MALDI target) is placed in the instrument the positions for analysis are automatically avaliable. The process will be illustrated using trypsin-digested proteins from a number of sources.

D. A. Bintzler¹, P. S. Adams², Y. Bao³, R. Keefe⁴, L. O. Petukhova⁵, C. Rosato⁶, R. Scholl⁷, K. Sol-Church⁸;
¹DNA Core Facility, University of Cincinnati, Cincinnati, OH, United States, ²Trudeau Institute, Saranac Lake, NY, United States, ³University of Virginia School of Medicine, Charlottesville, VA, United States, ⁴Wadsworth Center/NYDOH, Albany, NY, United States, ⁵Rockefeller Center, New York, NY, United States, ⁶Oregon State University, Corvallis, OR, United States, ⁷University of Utah, Salt Lake City, UT, United States, ⁸Alfred I. duPont Hospital for Children, Wilmington, DE, United States.

The members of the Fragment Analysis Research Group (FARG) conducted a study to survey resource facilities conducting DNA fragment analysis in current methods and instrumentation. The goal of the study was to look at the resources used in the service facility conducting fragment analysis and to determine how the trends have changed (or advanced). The questions in the study addressed a broad range of topics relating to personnel, technologies and services provided. Study participants were notified of the study availability by mass mailing and by email. The study was conducted anonymously as an on-line questionnaire with questions related to laboratory personnel, equipment, cost and types of services provided. Results from the study were analyzed by members of FARG and compared to the results in a similar study conducted and presented by FARG during the ABRF conference in 1999.

L. M. Vaughan;
Corning Life Sciences, Acton, MA, United States.

The industrialization of genomics and proteomics demands the miniaturization of filtration devices to a more automation-compatible format. In response to this need, Corning has developed the FiltrEX 96 and 384 filter plate platforms. These products are based on unique patented technology that allows individual filter disks to be integrally sealed in a microwell format. The objective of the current study was to demonstrate the utility of these filter plates in high throughput genomic and proteomic sample preparation methods using standard vacuum and centrifugation techniques. High throughput methods were developed for plasmid, PCR fragment and sequencing reaction clean up as well as protein size separation and de-salting/concentration. The results clearly demonstrate that these products can be used to perform rapid, low cost, high quality purification of both DNA and protein and that they are compatible with standard laboratory equipment and automation.

I. Horsey, J. Yan, A. Devenish, T. Stone;
Amersham Biosciences, Amersham, Bucks, United Kingdom.

With the advances in genomics made in the last 50 years we have begun to unravel the complex link between genes and function in many organisms. However, it has become apparent that not all the information is encoded in the genome and in order to enrich our understanding, we must also study the relationship between proteins and function.
Two-dimensional gel electrophoresis remains the state of the art in complex protein separation, and can be combined with peptide mass fingerprinting for accurate protein identification. Whilst this powerful technique has the ability to visualise many components of complex proteomes in a single gel, accurate quantitative comparative analysis of two (or more) different samples using this technique is limited. The limited dynamic range of some protein stains (e.g. silver) and the gel-to-gel variability of the spot patterns produced are both major contributors to this lack of accurate comparisons. Many software packages have been produced to try and address this issue, but the underlying experimental variation has proved to be a difficult obstacle to overcome.
The Ettan^TM DIGE system allows multiplexed labelling of samples with CyDye™ DIGE Fluor dyes followed by concurrent separation of proteins on a single 2-D gel. For comparative proteomics, this approach has obvious advantages. Use of the multiplexed fluors enables the inclusion of an internal standard on every gel in each experiment. When the use of these fluors is combined with the appropriate experimental design and analysed using the novel DeCyder™ Differential Analysis Software, accurate determination of changes in protein levels can be made.
This poster will outline the Ettan DIGE system and will demonstrate its application to analyse differential protein expression in Escherichia coli.

C. R. Cisar¹, J. M. Balog¹, N. B. Anthony², M. Iqbal², W. G. Bottje², A. M. Donoghue¹;
¹Poultry Production & Product Safety Research Unit, ARS, USDA, Fayetteville, AR, United States, ²University of Arkansas, Fayetteville, AR, United States.

Ascites syndrome, also known as pulmonary hypertension syndrome, is a significant problem for producers in the poultry industry. Meat-type birds such as broilers can develop ascites if exposed to cold, fed a high energy diet, or raised at high altitude. Ascites syndrome in chickens is characterized by chronic elevated pulmonary blood pressure resulting in enlargement of the heart, accumulation of fluid in the abdomen, and eventually death of the animal. Ascites syndrome was induced in commercial broilers and in chickens from ascites resistant and susceptible broiler lines developed previously by placing birds at hatch in a hypobaric chamber (simulated altitude of 2,900 m). Over a six week period, the heart and lung tissues of birds that exhibited symptoms of cyanosis and whose abdomens were enlarged were collected. Tissues were also collected from birds that were symptomless after six weeks in the hypobaric chamber. RV:TV ratios (right ventricle weight to total ventricle weight) were calculated and used to confirm whether birds were ascitic or non-ascitic. Using commercial antibodies to mammalian bone morphogenetic receptor type II (BMPR-II) protein and the 70 kDa and 30 kDa proteins from mitochondrial complex II, we examined the expression of these proteins in heart and lung tissues using SDS PAGE and Western blots. Putative chicken BMPR-II protein is upregulated in the lung and heart tissues of ascitic birds. Our experiments indicate that the mitochondrial proteins from complex II are also differentially expressed. Future experiments include 2D gel electrophoresis and identification of differentially expressed proteins by mass spectrometry.

A. T. Yeung¹, P. S. Adams², G. A. Buck³, B. P. Holloway⁴, K. Mills⁵, S. Scaringe⁶, G. L. Shipley⁷;
¹Fox Chase Cancer Center, Philadelphia, PA, United States, ²Trudeau Institute, Saranac Lake, NY, United States, ³Virginia Commonwealth University, Richmond, VA, United States, ⁴CDC, Atlanta, GA, United States, ⁵Millenium Pharmaceuticals, Inc., Cambridge, MA, United States, ⁶Dharmacon Research, Inc., Lafayette, CO, United States, ⁷University of Texas-Health Science Center, Houston, TX, United States.

Real-Time or quantitative (q) PCR technology is of increasing importance in genomic research. The high cost of FRET DNA probes for experiments has long impeded the full utilization of qPCR. The commercial cost of dual-labeled probes for qPCR reactions is high because of the post-synthesis HPLC and/or gel purification steps required by limitations in the traditional synthesis chemistry. The recent availability of CPG quencher reagents to core DNA synthesis facilities has opened up the possibility that probes may be used without extensive post-synthesis purification, if carefully prepared. This would substantially reduce the cost, making the synthesis of qPCR probes feasible and more affordable for any DNA synthesis laboratory. The NARG tested the proposal that all DNA synthesis labs are able to make quality dual-labeled probes suitable for qPCR reactions without gel and/or HPLC purification, by inviting members of the DNA synthesis community to synthesize 5'-FAM, 3'-BHQ1 or -TAMRA quenched human β-actin probes and submit them for quality analysis. The NARG members performed quality analysis on the probes using CE, DHPLC, ESI-MS and PAGE. Effectiveness in Real-Time PCR experiments was determined over a five log range of standard template concentration to assess the effect on assay efficiency and sensitivity compared to highly purified probes.

H. A. Remmer¹, N. Ambulos², L. F. Bonewald³, J. J. Dougherty, Jr.⁴, E. Eisenstein⁵, E. Fowler⁶, J. Johnson⁷, A. Khatri⁸, M. O. Lively⁹, N. Ritter¹⁰, S. T. Weintraub¹¹;
¹The University of Michigan, Ann Arbor, MI, United States, ²The University of Maryland, Baltimore, MD, United States, ³University of Missouri, Kansas City, MO, United States, ⁴Eli Lilly and Company, Indianapolis, IN, United States, ⁵Center for Advanced Research in Biotechnology, Rockville, MD, United States, ⁶Millennium Pharmaceuticals Inc., Cambridge, MA, United States, ⁷Texas A&M University, College Station, TX, United States, ⁸Massachusetts General Hospital, Charlestown, MA, United States, ⁹Wake Forest University, Winston-Salem, NC, United States, ¹⁰NMR Biotech, Germantown, MD, United States, ¹¹The University of Texas Health Science Center, San Antonio, TX, United States.

The goal of the Peptide Standards Project is the production, characterization, validation and establishment of three synthetic peptides as registered and certified peptide reference standards. This project is conducted by the Peptide Project Standards Committee (PSPC)of the ABRF in collaboration with the National Institute of Standards and Technology (NIST).
The project entails the following: (1) Large-scale synthesis of the standard peptides in high (>95%) purity; (2) Tests for stability and shelf life of the peptides; (3) Packaging of the peptides in small quantities for distribution and analysis by ABRF member laboratories; (4) Use of the ABRF member analyses to yield the required Certificate of Analysis.
The Peptide Standards Project was originally initiated by joint interactions between the Quality Compliance Committee (QCC) and the Peptide Synthesis Research Group (PSRG). It is the first ABRF project funded by NIST. Detailed information about large-scale synthesis and purification of the peptides as well as the current status and scope of this project will be presented.

M. Alterman¹, D. Chin², R. Harris³, P. Hunziker⁴, A. Le⁵, S. Linskens⁶, L. Packman⁷, J. Schaller⁸;
¹Dept. of Chemistry, University of Kansas, Lawrence, KS, United States, ²Chiron Corp., Emeryville, CA, United States, ³Genentech, South San Francisco, CA, United States, ⁴Institute of Biochemistry, University of Zurich, Zurich, Switzerland, ⁵Transgenomic, Inc., San Jose, CA, United States, ⁶Lanais-Pro, Buenos Aires, Argentina, ⁷Dept. of Biochemistry, University of Cambridge, Cambridge, United Kingdom, ⁸Dept. of Chemistry and Biochemistry, Unversity of Bern, Bern, Switzerland.

The AAA2003 study is a continuation of a previous study focusing on quantitation of proteins by amino acid analysis and dye-binding assays. Each participating laboratory received five separate solutions of pure proteins for the determination of their protein concentration by AAA. The concentration of each sample was also determined by the BCA or Bradford dye-binding assays. An additional standard sample was supplied for the calibration of the dye-binding assay. The amino acid analysis results were used by Research Group members to calculate the protein concentration in each sample and to compare these data with the relative precision and accuracy of the colorimetric assays. The goal of this study is to show whether colorimetric assays are a reliable method of quantitating unknowns when using a standard protein as reference.

J. W. Hawes¹, G. S. Grills², E. Jackson-Machelski³, H. Escobar⁴, T. Hunter⁵, K. L. Knudson⁶, R. Pershad⁷, D. Spicer⁸, T. W. Thannhauser⁹;
¹Indiana University School of Medicine, Indianapolis, IN, United States, ²Harvard University, Boston, MA, United States, ³Washington University School of Medicine, St Louis, MO, United States, ⁴Brigham Young University, Provo, UT, United States, ⁵Vermont Cancer Center, Burlington, VT, United States, ⁶University of Iowa, Iowa City, IA, United States, ⁷U.T. M.D. Anderson Cancer Center, Houston, TX, United States, ⁸Prolinx Inc.,, Bothell, WA, United States, ⁹Cornell University, Ithaca, NY, United States.

The DSRG is performing a study that tests the ability of laboratories to analyze DNA templates containing difficult repeat sequences. Such templates are common and often present a difficult problem to DNA sequencing core laboratories. Three templates ranging in difficulty have been presented to participating laboratories with appropriate primers. Test samples were chosen from a well-characterized collection of clones provided by Harvard Partners Genome Center (65 Landsdowne Street, Cambridge, MA. The goal of the study is to assess the ability of core laboratories to successfully analyze difficult repeat sequences and to gather data relating to which reaction conditions, chemistries and instrument formats are best for such templates. A survey has been developed and distributed to collect data regarding the reaction conditions and chemistry choices that each participant employs in this study. All of the data will be distributed to each member of the DSRG. Each member of the DSRG will have input into the analysis of this data. This study will compliment well the previous studies conducted by the DSRG including the analysis of GC-rich sequences (sasha and lunatic) and the NES (never ending story) which focuses on the analysis of pGEM plasmid as a standard template.

R. Pershad¹, J. Hawes², T. Hunter³, E. Jackson-Machelski⁴, K. Knudtson⁵, H. Escobar⁶, D. Spicer⁷, G. Grills⁸, T. Thannhauser⁹;
¹University of Texas M.D. Anderson Cancer Center, Houston, TX, United States, ²Indiana University, Indianapolis, IN, United States, ³University of Vermont, Burlington, VT, United States, ⁴Washington University School of Medicine, St. Louis, MO, United States, ⁵University of Iowa, Iowa City, IA, United States, ⁶Brigham Young University, Provo, UT, United States, ⁷Prolinx Inc., Bothell, WA, United States, ⁸Harvard University, Cambridge, MA, United States, ⁹Cornell University, Ithaca, NY, United States.

The goal of this study by the ABRF DNA Sequencing Research Group is to identify the changes in the composition and configuration of DNA sequencing facilities. DNA Sequencing Core Facilites serve as centralized resources, providing expertise in the area of DNA analysis in both academic and commercial institutions. As new instrumentation and methods are developed the make up of Sequencing Core facilities is directly impacted and subject to change. Information was collected from participating laboratories on the size, staffing, funding, instrumentation, chemistries and facility services provided. This study will help identify new trends in the DNA Sequencing Facilites and will present a current profile of the typical DNA Core Sequencing Facility.

A. I. Brooks¹, K. Knudston², C. Griffin³, A. Iacobas⁴, K. Johnson⁵, G. Khitrov⁶, S. Levy⁷, A. Massimi⁴, N. Nowak⁸, A. Viale⁹, G. Grills¹⁰;
¹University of Rochester Medical Center, Rochester, NY, United States, ²University of Iowa, Iowa City, IA, United States, ³University of San Francisco, San Francisco, CA, United States, ⁴Albert Einstein College of Medicine, New York, NY, United States, ⁵Jackson Laboratories, Bar Harbour, ME, United States, ⁶Rockefeller University, New York, NY, United States, ⁷Vanderbilt University, Nashville, TN, United States, ⁸Roswell Park Cancer Center, Buffalo, NY, United States, ⁹Memorial Sloan Kettering Cancer Center, New York, NY, United States, ¹⁰Harvard University, Boston, MA, United States.

Custom spotted microarrays are an important tool for many laboratories. Oligonucleotide libraries have been developed by industry to aid in the production of custom spotted arrays. This approach promises increased sensitivity and specificity over cDNA array platforms while keeping the overall cost of array production low. The MARG has engaged in a study to identify the most important variables when utilizing high density oligonucleotide libraries. Seven commercial vendors were given genbank accession numbers in order to design and provide 130 long oligonucleotides to test their performance in a microarray platform. This study reports on the following factors that contribute to the generation of these libraries: (1) bioinformatic design; (2) oligonucleotide length; (3) oligonucleotide modifications; (4) oligonucleotide synthesis and purification; and (5) genetic coverage. For these experiments, microarrays were printed at a single laboratory site and processing was performed at multiple laboratory facilities in order to investigate the reproducibility of results. The data presented will provide valuable information on important factors to consider when constructing or purchasing an oligonucleotide library for use in microarray production and analysis.

K. Knudston¹, A. I. Brooks², C. Griffin³, A. Iacobas⁴, K. Johnson⁵, G. Khitrov⁶, S. Levy⁷, A. Massimi⁴, N. Nowak⁸, A. Viale⁹, G. Grills¹⁰;
¹University of Iowa, Iowa City, IA, United States, ²University of Rochester Medical Center, Rochester, NY, United States, ³University of California, San Francisco, San Franciso, CA, United States, ⁴Albert Einstein College of Medicine, Bronx, NY, United States, ⁵Jackson Laboratories, Bar Harbour, ME, United States, ⁶Rockefeller University, New York, NY, United States, ⁷Vanderbilt University, Nashville, TN, United States, ⁸Roswell Park Cancer Institute, Buffalo, NY, United States, ⁹Memorial Sloan Kettering Cancer Center, New York, NY, United States, ¹⁰Harvard University, Boston, MA, United States.

New microarray-based methods and instrumentation are constantly being introduced and the number of investigators using these technologies is rapidly expanding. This study is the third annual general survey conducted by the ABRF Microarray Research Group (MARG). Goals include: (1) to build a current profile of microarray facilities; (2) to compare the current survey results with previous profiles to illustrate how microarray technology is evolving; and (3) to provide insight as to where this rapidly developing technology is going. This survey focused on spotted DNA microarray and Affymetrix GeneChip technologies. Information was also requested regarding the use of other microarray platforms. Data was requested from laboratories by posting instructions for participation on microarray related electronic discussion groups. The survey was aimed at gathering information from academic, pharmaceutical, and commercial laboratories that offer microarray technologies as a shared resource. Individual laboratories that have these technologies could also participate. A web based survey was used to collect information such as instrumentation, protocols, staffing, funding, costs, and throughput. The resulting analysis gives a picture of the state of the art of microarray analysis.

T. Neubert¹, D. Arnott², M. Gawinowicz³, R. Grant⁴, L. Packman⁵, K. Speicher⁶, K. Stone⁷, C. Turck⁸;
¹New York University, New York, NY, United States, ²Genentech, Inc., South San Francisco, CA, United States, ³Columbia University, New York, NY, United States, ⁴The Procter & Gamble Co., Cincinnati, OH, United States, ⁵Cambridge University, Cambridge, United Kingdom, ⁶The Wistar Institute, Philadelphia, PA, United States, ⁷Yale University, New Haven, CT, United States, ⁸Max Planck Institute of Psychiatry, Munich, Germany.

A fundamental aspect of proteomics is the analysis of posttranslational modifications, including phosphorylation. Modern mass spectrometric techniques now make possible phosphorylation site determination at high sensitivity. It is not known, however, how many laboratories routinely take on such projects, which methods are most often applied, and how successful the laboratories and particular methods are. The Proteomics Research Group has undertaken a study to help answer these questions.
A sample containing one or more proteins and two synthetic phosphopeptides was distributed in the form of a tryptic digest, with participating laboratories challenged to identify the component protein(s), identify the phosphorylated peptides, and determine the sites of phosphorylation. The phosphopeptides were included at low stoichiometries relative to the unphosphorylated protein since this is a common challenge. It was anticipated that strategies employed for the detection of phosphopeptides would include subtractive analysis, enrichment by affinity chromatography, diagnostic mass differences, and characteristic fragmentation patterns in PSD or CID. Phosphorylation site determination could involve MS/MS or Edman degradation. This study has the following goals: 1) Provide a mechanism for participating laboratories to evaluate their capabilities in locating sites of phosphorylation in non-radioactive samples; 2) Provide an introduction to phosphorylation site analysis for laboratories that do not currently do such analyses; 3) Compare strategies for phosphorylation site mapping; and 4) Help establish realistic expectations for these analyses.

D. Bintzler¹, Y. Bao², R. Keefe³, L. Petukova⁴, C. Rosato⁵, R. Scholl⁶, K. Sol-Church⁷;
¹University ofCincinnati, Cincinnati, OH, United States, ²University ofVirginia School of Medicine, Charlottesville, VA, United States, ³Wadsworth Center/NYSDOH, Albany, NY, United States, ⁴The Rockefeller University, New York, NY, United States, ⁵Oregon State University, Corvallis, OR, United States, ⁶University ofUtah, Salt Lake City, UT, United States, ⁷Alfred I. DuPont Hospital forChildren, Wilmington, DE, United States.

The Fragment Analysis Research Group (FARG) will report the results of their on-line survey of laboratories conducting DNA fragment analysis. The survey questions developed by FARG members were meant to comprehensively address current methods and instrumentation used in genotyping labs, and included questions related to the services provided by core facilities. The following survey areas will be discussed: the number of staff involved in fragment analysis; types of technologies and specific equipment used; types of genotyping software used; nature of the DNA fragments analyzed (snps, microsatellites, etc.) and sample throughput; other services provided by core facilities that carry out DNA fragment analysis; fees charged for services, the availability of bioinformatics support; what quality control procedures are implemented; and the nature of financial support received. The summarized survey data will hopefully provide staff at genotyping facilities with information useful for making constructive evaluations of their own services and methodologies in light of input provided by survey participants. A round table discussion will follow the presentation of the survey results. Any meeting attendee may ask questions of a panel of FARG members regarding trouble shooting problems encountered in a genotyping lab or core facility. Using the collective experience of both FARG panelists and audience members, it is hoped that some solutions can be offered.

J. M. Neveu¹, J. Fernandez², S. Buckel³, M. Crawford⁴, R. Cook⁵, B. Madden⁶, N. D. Denslow⁷, L. Steinke⁸;
¹Harvard University, Cambridge, MA, United States, ²Rockefeller University, New York, NY, United States, ³Amgen, Boulder, CO, United States, ⁴Yale, New Haven, CT, United States, ⁵Baylor College of Medicine, Houston, TX, United States, ⁶Mayo Clinic, Rochester, NY, United States, ⁷University of Florida, Gainsville, FL, United States, ⁸University of Nebraska, Omaha, NE, United States.

The ABRF-2003ESRG sample is the 15^th in a series of studies designed to asses the overall performance of participating laboratories in protein sequencing using automated Edman degradation. The sample for the 2003 study consisted of a low level electroblotted 43kDa protein, with expected initial yields of about 2 pmol. The objective was to evaluate each laboratory’s initial yield, repetitive yield, length of correct sequence, identification of the protein, overall performance of instruments, and run conditions or other factors that may affect the final results. Participating laboratories may also compare their results to other responders, as well as a control data set generated by ESRG members.

J. L. Bleibaum¹, C. Nicolet², G. Sarath³, J. Simpson⁴, S. Yadav⁵;
¹Roche Bioscience, Palo Alto, CA, United States, ²University of Wisconsin, Madison, WI, United States, ³University of Nebraska, Lincoln, NE, United States, ⁴SAIC-Frederick, Frederick, MD, United States, ⁵Cleveland Clinic Foundation, Cleveland, OH, United States.

The ABRF was conceived as an organization to serve the needs of scientists working in research core facilities. Since its inception, the ABRF has grown steadily and specific activities have been developed to benefit members. Among these activities are: the annual meeting; the organization’s journal, Journal of Biomolecular Techniques (JBT); the research groups; the electronic bulletin board/discussion group; and the travel award. The ABRF Survey Committee wanted to know how well the ABRF was serving members and potential members via these activities. We also wanted to know why members chose to be members and how the ABRF could improve. We solicited opinions from ABRF members and non-members regarding ABRF activities and membership issues. Respondents were encouraged to comment on the importance and/or relevance of ABRF activities as well as on membership in general. Preliminary results indicate that all of the ABRF activities resonate strongly with respondents, especially the ABRF bulletin board/discussion group. Comments regarding membership have been very specific in support of the ABRF continuing in its mission of serving scientists working in core facilities. Exchange of ideas with other core facility scientists doing similar work was also frequently mentioned as an ABRF asset. Comments from non-members have specified cost constraints, residence outside of the U.S., and eligibility issues as barriers to membership. It is important that these concerns be addressed as the ABRF continues to grow. Nevertheless, it is clear that the ABRF is a necessary organization with an active, interested, and vocal membership.

L. Faas¹, M. Delaney², B. Ellgas³, S. Mowbray⁴;
¹Seattle Public Utilities, Seattle, WA, United States, ²Massachusetts Water Resources Authority, Winthrop, MA, United States, ³East Bay Municipal Utility District, Oakland, CA, United States, ⁴County Sanittion Districts of Orange County, Fountain Valley, CA, United States.

Laboratory managers face increasing pressure to provide cost-effective service, meet higher expectations from customers and deal with the complexity of changing analytical and regulatory requirements, and benchmarking can be a useful tool to help laboratory directors address these challenges. This presentation describes a rigorous benchmarking process undertaken by nine municipal government laboratories. Goals of the study included:
•
Improve functional laboratory operation
•
Minimize costs and maximize efficiencies
•
Develop a rigorous model for ‘apples to apples’ cost comparisons between laboratories
•
Use the cost model to compare changes from year to year
•
Identify best practices of the individual labs
•
Identify laboratory costs associated with the various stages of the wastewater treatment process
The major product of the benchmarking process was a detailed cost per test spreadsheet that helped lab managers determine their own costs and compare costs with each other. This presentation will explain how to develop a rigorous cost model to calculate the cost of laboratory tests and services and how to use the cost model to make ‘apples to apples’ comparison of costs between laboratories. Other study findings that will be discussed include:
•
Best laboratory practices in the areas of workload management, customer service, staffing strategies and employee development and morale
•
Factors associated with increasing efficiency and minimizing cost per test
•
Strategies to improve laboratory efficiency, flexibility and cost-effectiveness
•
Tools to help laboratories improve customer service and employee morale

D. A. Bintzler¹, P. S. Adams², Y. Bao³, R. Keefe⁴, L. O. Petukhova⁵, C. Rosato⁶, R. Scholl⁷, K. Sol-Church⁸;
¹DNA Core Facility, University of Cincinnati, Cincinnati, OH, United States, ²Trudeau Institute, Saranac Lake, NY, United States, ³University of Virginia School of Medicine, Charlottesville, VA, United States, ⁴Wadsworth Center/NYDOH, Albany, NY, United States, ⁵Rockefeller Center, New York, NY, United States, ⁶Oregon State University, Corvallis, OR, United States, ⁷University of Utah, Salt Lake City, UT, United States, ⁸Alfred I. duPont Hospital for Children, Wilmington, DE, United States.

The members of the Fragment Analysis Research Group (FARG) conducted a study to survey resource facilities conducting DNA fragment analysis in current methods and instrumentation. The goal of the study was to look at the resources used in the service facility conducting fragment analysis and to determine how the trends have changed (or advanced). The questions in the study addressed a broad range of topics relating to personnel, technologies and services provided. Study participants were notified of the study availability by mass mailing and by email. The study was conducted anonymously as an on-line questionnaire with questions related to laboratory personnel, equipment, cost and types of services provided. Results from the study were analyzed by members of FARG and compared to the results in a similar study conducted and presented by FARG during the ABRF conference in 1999.

J. Van Ee¹, P. S. Adams², R. H. Lyons³, K. I. Mitchelhill⁴, M. Pamukcu⁵, L. Steinke⁶, T. W. Thannhauser¹, C. D. Walls⁷;
¹Cornell University, Ithaca, NY, United States, ²Trudeau Institute, Saranac Lake, NY, United States, ³University of Michigan, Ann Arbor, MI, United States, ⁴Peter MacCallum Cancer Institute, East Melbourne, Australia, ⁵MPA, Sunnyvale, CA, United States, ⁶University of Nebraska, Omaha, NE, United States, ⁷University of Utah, Salt Lake City, UT, United States.

The ABRF has embraced the concept of electronic communication since it's inception. The initial form of communication was the still popular and very valuable "Roundtable" discussion group which was announced in the ABRF news in September 1994 . An FTP site for exchange of information was introduced shortly thereafter. The first ABRF web site was unveiled in August 1995 and the roundtable archives were initiated in 1996 . The Web Site Committee was formed in June of 2000 as an ad-hoc group charged with reviewing the ABRF’s current web presence and exploring avenues for improvement. Currently the ABRF membership is enjoying an attractive and functional new site that provides information about all functions of the ABRF and its Research Groups activities, archives of the roundtable discussions, presentations from past meetings, the white pages of membership listing, and the yellow pages of facilities and the services they offer. Improvements include: a new list server system which has solved chronic problems with spam, virus/worm propagation and subscription issues; automatic archiving; an improved search engine; member tools for managing individual and facility information and Research Group pages. The ABRF Web committee recently compiled results showing web site activity over the past year that should aid in further refinements.

J. K. Westwick;
Odyssey Pharmaceuticals, Inc., San Ramon, CA, United States.

The transmission of cellular signals occurs through regulated protein-protein interactions. Until recently, high throughput and quantitative methods for discovering and monitoring interactions among mammalian signaling proteins did not exist. The protein-fragment complementation (PCA) approach utilizes a dissected reporter enzyme cloned in-frame with any known cDNA pair (or a specific gene and a cDNA expression library). Expression and folding of the reporter enzyme occurs only when the linked proteins bind.
PCA has several distinct advantages over traditional protein interaction analysis techniques. Using PCA, proteins are expressed in their true sub-cellular context with native post-translational modifications — in fact, PCA has been used to visualize and quantify sub-cellular localization and trafficking of protein complexes. PCA has enabled the elucidation of biological function for proteins on a genome-wide scale, and provides a unique approach for high-throughput screening of novel targets directly in living cells.

S. Miick;
Nanogen, Inc., San Diego, CA, United States.

Nanogen has developed the ElectroCapture™ technology, an electronic, fluorescent kinase assay, for use in primary and secondary screening. The ElectroCapture™ assay is configured for 384-well microtiter plates and uses no antibodies and no radioactivity. The ElectroCapture™ technology involves kinase phosphorylation to invert the charge of a fluorophore-labeled peptide substrate, followed by electrophoresis to capture the phosphorylated product. We have also developed a proprietary approach for substrate identification to quickly identify specific fluorophore labeled substrates for kinases. The substrate identification technology and the ElectroCapture™ assay enable faster assay development time for an independent platform which can be particularly useful in lead optimization.

L. M. Brill, S. B. Ficarro, A. R. Salomon, M. Mukherj, Q. T. Phung, C. Ericson, D. M. Horn, A. Brock, E. C. Peters;
Genomics Institue of the Novartis Research Foundation, San Diego, CA, United States.

Reversible phosphorylation is a key mechanism for regulating protein activity. Because our goal is to understand the effects of this modification at the molecular level, residues that are phosphorylated must be defined. Tandem mass spectrometry (MS²) based approaches are well suited to this task and have been used for several years to study phosphorylation on a protein-by-protein basis. More recently, we have developed methods using immobilized metal affinity chromatography coupled to reversed phase-HPLC/micro-electrospray/MS² to determine phosphorylation sites on multiple proteins in a single analysis. Because phosphotyrosine residues are important in cell signaling, immunoprecipitation of phosphotyrosine-containing proteins was used to enrich for these low-abundance residues. We found several dozen phosphorylation sites on signaling proteins from Leukemia cells and activated T-cells; some sites were known and some were novel. We conclude that these procedures resulted in determination of a large number of phosphorylation sites in individual analyses, and could be applied to profiling of protein phosphorylation within cells in states of health and disease.

D. W. Speicher¹, K. Champion², S. Gygi³, K. Lilley⁴, M. Tan⁵, J. Yates⁶;
¹The Wistar Institute, Philadelphia, PA, United States, ²Genentech, Inc., South San Francisco, CA, United States, ³Harvard Medical School, Boston, MA, United States, ⁴University of Cambridge, Cambridge, United Kingdom, ⁵Zyomyx, Hayward, CA, United States, ⁶Scripps Research Institute, La Jolla, CA, United States.

The most commonly pursued type of proteome studies involves quantitative comparison of protein levels in two or more experimental samples of a cell, tissue, or organism. However, highly complex proteomes such as those from cells or tissues from higher eukaryotes contain far more components than any single current technology platform can reproducibly separate and quantitatively compare. This roundtable discussion will compare alternative methods for conducting quantitative protein profile comparisons of two or more samples. The strengths and weaknesses of 2-D gels, 2-D difference gel electrophoresis (2-D DIGE), antibody arrays, and non-gel based LC-MS/MS methods will be briefly presented, followed by an open discussion. A subsequent roundtable will focus on a more in-depth discussion of alternative MS-based quantitative protein profiling methods.

P. S. Adams¹, G. L. Shipley²;
¹Trudeau Institute, Saranac Lake, NY, United States, ²University of Texas Health Sciences Center-Houston, Houston, TX, United States.

Whether you utilize Real-time PCR technology as an official Core function or in a research laboratory, this roundtable discussion atmosphere should provide an opportunity to share your experiences and find out how other scientists utilize this versatile technology. Potential topics include: sample preparation; the reverse transcription (RT) reaction; different quantitative PCR (qPCR) platforms available; various real-time fluorescent detection mechanisms; assay design; software and hardware issues; robotics and lab automation; data analysis; long term data storage; sample tracking and record keeping and core facility management.
ABRF meeting attendants are encouraged to bring their questions and expertise to this discussion group so that we all can benefit from each other’s knowledge, or obtain help trouble shooting problems presently experienced in our own laboratories. You may present/support your question(s) with overheads or other media.

S. P. Gygi¹, T. Addona², S. A. Gerber¹, M. Mann³;
¹Harvard Medical School, Boston, MA, United States, ²Millenium Pharmceuticals Inc, Cambridge, MA, United States, ³University of Southern Denmark, Odense, Denmark.

This roundtable will compare the strengths and weaknesses of several mass spectrometry-based methods for the quantification of proteins and posttranslationally modified proteins. Quantitative technologies to be discussed include i) stable isotope incorporation (both pre and post protein isolation), ii) creating a protein abundance index based on the numbers of peptides sequenced for each protein, and iii) new technologies for the absolute quantification of proteins and phosphoproteins (AQUA). An earlier roundtable will focus on more in-depth discussion of non MS-based quantitative protein profiling methods.

X. Lin;
ProteomTech, Inc., Emeryville, CA, United States.

With the completion of the sequencing of human genome, and the near-complete collection of human full-length cDNAs, high-throughput expression and purification of recombinant proteins become the next bottle-neck of life science. The most efficient high-throughput expression system is the E. coli expression. However, the most serious problem in E. coli expression is the formation of inclusion bodies, which cannot be avoided in most of the expressed proteins. Protein refolding is one of the most complicated problems in biology. The refolding success is heavily dependent on the techniques or "artistic" abilities of individual scientist. ProteomTech is working on solving the refolding problem experimentally, trying to develop general method for refolding various proteins. In addition, ProteomTech is developing automation procedures for refolding screening. With this approach, protein refolding may eventually be transformed from "art" to science.

T. C. Hunter¹, H. Escobar², G. Grills³, J. Hawes⁴, E. Jackson-Malcheski⁵, K. Knudtson⁶, R. Pershad⁷, T. Thannhauser⁸;
¹University of Vermont, Burlington, VT, United States, ²Brigham Young University, Provo, UT, United States, ³Harvard University, Cambridge, MA, United States, ⁴Indiana University School of Medicine, Indianapolis, IN, United States, ⁵Washington University School of Medicine, St. Louis, MO, United States, ⁶University of Iowa, Iowa City, IA, United States, ⁷University of Texas MD Anderson Cancer Center, Houston, TX, United States, ⁸Cornell University, Ithaca, VT, United States.

The DNA Sequencing Research Group will host a roundtable discussion. DSRG will invite ABRF members that currently serve as directors/managers of a facility offering DNA sequencing at their respective institutes to serve on this panel. The panel members will present methods and ideas on how to overcome difficult to sequence templates or "oddities" observed in DNA sequencing. Discussion will also include service related issues, cost saving measures, along with technical difficulties focusing primarily on capillary based systems.
This roundtable gathering will serve as a forum to promote and facilitate information exchange regarding DNA sequencing between the panel and audience. Audience participation is critical to a roundtable discussion and will be encouraged through solicitation of problematic sequences from ABRF members prior to the ABRF 2003 meeting.

C. Cantor;
Sequenom, Inc., San Diego, CA, United States.

SEQUENOM has developed a fully automated genotyping platform that allows up to 1 million SNPs to be measured in a single day. It is also possible to study allele frequencies of SNPs on population pools to provide a quick scans for SNPs of particular interest that may warrant individual genotyping.
SEQUENOM has worked with Incyte Pharmaceuticals and GlaxoSmithKline to develop a set of validated SNP assay portfolios that altogether number about 200,000 different polymorphic SNPs. These assay portfolios contain both gene-based and evenly spaced SNPs across the entire human genome. Allele frequency differences in phenotypically-stratified populations can reveal genes with strong associations to phenotypes. We have used an age-stratified healthy population as a gene discovery tool. We test the hypothesis that age is the major risk factor in complex disease; alleles that confer susceptibility to disease should thus decline in frequency as a function of age in a healthy population. This is true for some known disease markers and, from a scan of almost all of the human genome, we have discovered many new genes that appear to be linked to complex disease. Confirmation of disease gene hits is done by using additional, highly phenotyped populations. These populations are also used as discovery sets for additional whole genome scans. Thus far scans have been completed in Melanoma, Breast Cancer and Diabetes type II.
Recently we have developed two new applications that use the exact same automation and detection platform that Sequenom has developed and optimized. I will demonstrate how SNP discovery (re-sequencing) and gene expression analysis can also be done highly effectively by automated mass spectrometry.

D. J. Finley;
Dept. of Cell Biology, Boston, MA, United States.

Ubiquitin is conjugated via its C-terminus to lysine residues in a vast number of target proteins. The resulting isopeptide-linked conjugate is subject to rapid enzymatic breakdown. Conjugates are metabolized by the 2.5-Megadalton proteasome, which degrades the target protein, or by deubiquitinating enzymes, which reverse the isopeptide linkage. This pathway is currently estimated to comprise, in mammalian systems, at least 750 enzymes or components thereof. Substrates of the pathway are expected to be far more numerous, and overall we have only an initial understanding of the biological functions of ubiquitination.
Ubiquitin-dependent proteolysis underlies central regulatory mechanisms in the cell cycle, checkpoint control, inflammation, apoptosis, and other processes. Defects in ubiquitination underlie many diseases, and proteasome inhibitors appear to have potential in cancer therapy. Recently, nonproteolytic functions of ubiquitination have been described; ubiquitination mediates a variety of protein sorting decisions, and can also regulate protein activity states.
How can ubiquitination promote such different fates for its target proteins? One mechanism is through variations in the number and arrangement of ubiquitin groups decorating the target. For example, multiubiquitin chains are required on some substrates, and various chains can be distinguished by ubiquitin-receptor proteins on the basis of the alternative lysines in ubiquitin used for chain extension.
In addition to ubiquitin, ubiquitin-like proteins have been described in the last few years, further broadening the known scope of protein-protein modification pathways. For both ubiquitin and ubiquitin-like pathways, the identification of conjugative target proteins remains a formidable technical challenge. Physiologically important substrates are often modified at very low levels, due to rapid turnover of the isopeptide bond or by proteasomes. Proteomics- and genomics-based approaches will contribute significantly to our understanding of these vast pathways in the next few years.

C. A. Mizzen;
University of Illinois at Urbana Champaign, Urbana, IL, United States.

Histones H2A, H2B, H3 and H4 are small basic proteins that assemble to form octamers about which DNA is wrapped in nucleosomes, the fundamental subunit of eukaryotic chromatin. A fifth type of histone, H1, binds the outer surface of nucleosomes and contributes to the higher order folding of nucleosomal arrays. Nucleosomal structure and the higher order folding of chromatin compact eukaryotic genomes so that they fit within nuclei and constitute a primary mechanism for regulating DNA-templated activities. In the case of gene transcription, histone:DNA interactions can have either positive or negative effects that are modulated by post-translational modifications (PTMs) of histones. Despite intensive study, understanding the functional consequences of PTMs including acetylation, methylation, phosphorylation, poly(ADP-ribosylation) and ubiquitination has been complicated by the low abundance and lability of certain modified forms, the presence of molecules bearing similar levels of a given PTM at different sites, and the presence of molecules bearing different PTMs concurrently. Analyses of chromatin and histones in vitro have demonstrated that PTMs can influence DNA accessibility to regulatory factors and serve to recruit factors containing appropriate interaction motifs. Immunological approaches have demonstrated that the effects of acetylation and methylation on transcription in vivo can be site-specific and may be influenced by additional PTMs at nearby sites on the same molecule or at sites on different molecules but nearby in the context of chromatin. Together, these data suggest that combinations of PTMs are used to effect regulation of transcription and other nuclear processes. These data also suggest that new analytical techniques are required to reveal the true nature of this "histone code".

P. Roepstorff^1,2;
¹Department of Biochemistry and Molecular Biology, Odense M, Denmark, ²University of Southern Denmark, Odense, Denmark.

The first annotation of the human genome placed the number of protein coding genes at approximately 30 000 — 40 000, i.e., only approximately twice as many as that of the fly or worm and five times that of yeast. This strongly implies that protein complexity and regulation in higher eukaryotes is governed by co- and post-translational modifications and splicing events. It is becoming increasing obvious in biological studies that numerous proteins and protein complexes are regulated by post-translational modifications. Several protein modifications have been documented and methods to systematically identify post-translational modifications during protein identification studies are being developed. We have initiated an approach for specific detection of various types of protein modifications in proteomics, which we term Modification Specific Proteomics. The detection of PTM’s is very challenging because these are often transient and highly heterogeneous in terms of modified site (e.g. phosphorylation) as well as the modifying group (e.g. glycosylation). Our concept is based on specific detection of PTM’s in the 2D-gel, by specific "pull-out" of modified proteins/peptides, or by selective detection of the specific type of PTM in the mass spectrometer. It will be illustrated by our recent attempts to perform phospho- and glyco-specific proteomics. These two types of modifications are found in more that 50 % of all proteins in higher organisms. Due to their implications on protein structure and function, studies of biological phenomena must include detailed elucidation of these two types of modifications on a level compatible with proteomics studies.

A. T. Yeung;
Fox Chase Cancer Center, Philadelphia, PA, United States.

SNP, single nucleotide polymorphism, promises to be a key to the mapping of genes in complex genomes, as well as a major source of phenotypic differences among individuals of each species. In this scientific session, Dr. Arthur Holden, CEO of the SNP Consortium, Ltd., will provide an up to date perspective of the SNP initiatives. Next, as one example of successful application of SNP and related technologies, Dr. John Rioux, Director, Inflammatory Disease Research Group, Whitehead Institute, MIT, will present his first hand experience with SNP applications in disease gene mapping. I urge you to attend these highly informative presentations. If time permits, I will briefly introduce my CEL I mismatch endonuclease system that has become a powerful tool for the screening of unknown SNPs.
AT_Yeung@fccc.edu
http://web-apps.fccc.edu/fccc/yeung/index.html

D. Botstein;
Stanford University School of Medicine, Stanford, CA, United States.

Technology has been developed in our group to obtain and to analyze patterns of expression of many thousands of genes at once. Using this DNA microarray system, we have been studying patterns of gene expression in a variety of tumors, cell lines and normal tissues. Diverse kinds of tumors have been studied. In the cases of lymphomas, breast tumors, lung tumors, liver tumors, brain tumors and soft tissue tumors we have been able, by the application of clustering algorithms, to subclassify tumors on the basis of their gene expression patterns. These subclassifications often appear to be biologically meaningful; often sub-classes are statistically associated with different clinical outcomes. By clustering genes according to their pattern of expression, we have been able to associate particular genes with tumor features, including proliferation, presence of inflammatory, stromal or endothelial cell types, and we also have been able to associate with particular tumor subtypes strong expression of heretofore uncharacterized genes, including many that appear to encode either secreted or membrane associated proteins. By studying synchronized cells (budding yeast or human) growing in culture, we have identified many hundreds of human genes that are expressed periodically, at characteristically different points in the cell division cycle. When such an analysis is done in human cells, most of these genes are the same genes that comprise the "proliferation cluster", i.e. the genes whose expression is specifically associated with the proliferativeness of the tumors and tumor cell lines we have studied. Finally, we have been applying a variant of our DNA microarray technology (which we call "array comparative hybridization", or aCGH) to follow the DNA copy number of genes, both in tumors and in yeast cells undergoing adaptive evolution during hundreds of generations of growth in continuous culture. It has been well documented that many types of cancer involve characteristic chromosomal rearrangements, including especially specific translocations and gene amplifications. We have found similar rearrangements in yeast diploids in our adaptive evolution experiments, suggesting a similarity in mechanism between adaptive evolution in yeast and tumor progression in humans.

H. Steen¹, M. Gronborg², H. Molina², D. Kalume², O. N. Jensen³, M. Mann³, A. Pandey²;
¹Harvard Medical School, Boston, MA, United States, ²Johns Hopkins University, Baltimore, MD, United States, ³University of Southern Denmark, Odense M, Denmark.

Our goal is to develop proteomic methods that will enable a global analysis of signal transduction pathways. Post-translational modifications such as phosphorylation are a major mechanism for regulating protein function within cells. We have used antibodies against phosphotyrosine and phosphoserine/threonine residues to enrich phosphorylated proteins followed by analysis by nanoelectrospray tandem mass spectrometry or liquid chromatography coupled to tandem mass spectrometry. We have identified several known as well as novel proteins involved in tyrosine kinase signaling pathways and in serine/threonine kinase signaling pathways. A bioinformatic analysis of several of the novel phosphorylation sites that we have identified reveals that they exhibit significant deviation from the established ‘consensus’ phosphorylation motifs. Our studies indicate that such global studies are required for pathway elucidation and for improving the existing computational methods for prediction of phosphorylation sites.

P. Jorgensen;
Samuel Lunenfeld Research Institute, Toronto, ON, Canada.

Size homeostasis requires that budding yeast grow to a "critical" size before commitment to division at Start in late G1 phase. How critical cell size is determined is one of the least understood aspects of the cell cycle and has largely eluded genetic analysis. To systematically uncover pathways that coordinate growth and division, we determined size distributions for the complete set of ~6,000 Saccharomyces cerevisiae gene deletion strains. Alterations in cell size can be caused by a number of indirect effects. To identify lge mutants that are deleted for activators of the Start transition, we used Synthetic Genetic Analysis (SGA) to extract 9 lge mutants that are synthetically lethal/sick with deletion mutations in known Start activators. Plots of doubling time versus cell size were used to classify 25 of the whi mutants as potential repressors of the Start transition. Within this set of whi mutants are genes encoding an activator of the anaphase promoting complex (CDH1), a split zinc-finger transcription factor (SFP1), and a homolog of the oncogene AKT/PKB (SCH9). Epistasis analysis between our novel START repressors and known START activators revealed that previously unrecognized pathways feed into the Start transition. For instance, several lines of evidence suggest that rates of ribosome biogenesis modulate the critical cell size threshold required to pass Start. Using DNA microarrays, we have discovered that SFP1 encodes a transcriptional activator of a large suite of co-regulated genes, the majority of which appear to function in ribosome biogenesis. Given the conserved nature of many of the identified pathways, cell growth and division are likely to be coupled by conserved mechanisms.

U. Heberlein;
University of California San Francisco, San Francisco, CA, United States.

Alcohol abuse and alcoholism are among the most devastating social and medical problems in our society, yet little is known about the mechanisms by which ethanol regulates behavior. Our laboratory uses the fruitfly Drosophila, with its accessibility to behavioral, genetic, and molecular analysis, to help establish the missing links between genes and drug-induced behaviors.
Flies display many of the behaviors observed in mammals after both, acute and chronic exposure to ethanol and various psychostimulants. For example, ethanol activates locomotion at low doses, but depresses it at high doses. We have devised an automated locomotor tracking system that allows us to screen for mutants that show altered behaviors in the presence of ethanol. Different mutants affect distinct phases of the locomotor response to ethanol, highlighting the complexity of even a relatively simple behavior.
To define the neuroanatomical sites that regulate ethanol-induced behaviors we have used targeted expression of a transgene encoding a protein kinase A inhibitor to specific brain regions. These studies have revealed that a group of neurosecretory cells plays an important role in regulating the sedating effects of ethanol.

N. Perrimon;
Harvard Medical School, Boston, MA, United States.

After the completion of whole genome sequence of humans and major model organisms, one of the key challenges remains a functional characterization of gene products on a genome-wide scale. Computational analysis of genome sequences has predicted a large number of open reading frames without any functional classification. One of the current experimental challenges is to develop methodologies that allow a systematic and rapid analysis of every open reading frame in specific phenotypic assays.
In order to perform genome-wide reverse functional analysis, we have generated a library of 21,000 double-stranded RNA directed against every predicted Drosophila ORFs. RNA interference has been shown to be a powerful method to disrupt gene function by targeting specific messenger RNAs for destruction. In Drosophila cells, addition of double-stranded RNA gives penetrant and highly reproducible cellular phenotypes, making it an attractive and simple experimental system to establish loss-of-function phenotypes.
We have developed assays that monitor signaling pathways by reporter gene activity, cell death and survival phenotypes and other cell-biological readouts that can be monitored by luminescent or fluorescent read outs using plate reader or visual phenotypes that are monitored by automatic microscopy. In particular, we will present a genome-wide screen for cell death and cell growth defects showing that genome-wide RNAi screens yield reproducible results both on quantitative and qualitative levels that can be used for classifications of loss-of-function phenotypes.

P. R. Griffin, Y. Hamuro, S. Coales, J. Cawley, M. Southern;
ExSAR Corporation, Monmouth Junction, NJ, United States.

We have developed a novel approach to small molecule lead optimization, pre-clinical candidate selection and accelerating the discovery, development and production of biologics. Our approach involves the use of a non-invasive labeling strategy that is compatible with any type of biomolecule. This labeling approach involves hydrogen/deuterium exchange (H/D-Ex) of both amide hydrogens and carbon-centered hydrogens coupled with fully automated sample processing and LC-MS. We core discovery engine is combined with rapid physical measurements that allow us to correlate a compound structure with its interactions on-target and off-target.
Amide H/D-Ex coupled with mass spectrometry has become a powerful technique to study protein structure, protein-ligand interaction, and protein-protein interaction. Briefly, a protein is incubated in a deuterated environment followed by quenching at appropriate time intervals. The deuterated protein is degraded progressively by proteases to allow high sequence coverage and single amide resolution. The resultant peptide fragments are analyzed by LC-MS and the % deuteration is determined for each fragment. The experiment is repeated over various incubation times and the H/D-Ex rate of each amide or a region of amides is obtained. The rate of exchange for an amide(s) hydrogen affords the free energy of each amide(s) in the state H/D exchange was measured. This free energy map can give the structural and dynamic information of the protein.
Recently we have incorporated various improvements to this technology that allow high throughput and comprehensive analysis of protein ligand interactions and dynamics.

J. R. Yates¹, L. Florens¹, D. Carucci², A. Holden³, R. Sinden⁴;
¹The Scripps Research Institute, LaJolla, CA, United States, ²§Naval Medical Research Center, Malaria Program (IDD), Rockville, MD, United States, ³The Division of Parasitology, National Institute for Medical Research, London, United Kingdom, ⁴Imperial College of Science, Technology & Medicine, London, United Kingdom.

The completion of the Plasmodium falciparum genome provides a basis on which to conduct comparative proteomics studies of this human pathogen. Here, we applied a high throughput proteomics approach to identify new potential drug and vaccine targets and to better understand the biology of this complex protozoan parasite. We characterized four stages of the parasite life cycle (sporozoites, merozoites, trophozoites, and gametocytes) by Multidimensional Protein Identification Technology. Functional profiling of over 2,400 proteins agreed with the physiology of each stage. Unexpectedly, the antigenically variant var and rifin proteins, defined as molecules on the surface of the infected erythrocyte, were also largely expressed in sporozoites. The detection of chromosomal clusters encoding co-expressed proteins suggested a potential mechanism for controlling gene expression.

M. Patricelli, J. Wu, E. Okerberg, K. Blackford, J. J. Burnbaum;
ActivX Biosciences, La Jolla, CA, United States.

We present a complete platform for the characterization of proteomic samples based on fluorescent activity based probes (fABPs) and illustrate its utility for the identification of novel drug targets and for high-content drug screening. Fluorescent activity-based probes provide a means to simplify proteomic analyses by selectively reacting with members of specific protein families within a background proteome and afford the means for both sensitive detection and affinity purification. As a result, this approach circumvents many problems inherent to proteome-wide analysis with mass spectrometry, especially those problems resulting from the large dynamic range of protein expression and the low abundance of many pharmacologically important proteins. Moreover, this approach has proven useful for samples that are notoriously difficult to analyze, including membrane fractions and undiluted plasma. Because activity based probes react only with a single catalytic residues on each target protein, proteolytic digestion of labeled proteins does not increase sample complexity and permits quantitation of the resultant active site peptides with capillary electrophoresis (CE) and laser-induced fluorescence. Methods for reproducible, highly efficient peptide separations (<0.5% RSD in migration time, >200,000 theoretical plates) of fABP labeled whole cell lysate are presented. Additionally, a parallel method for molecular characterization utilizing nano-LCMS analysis of affinity purified fABP labeled material is detailed that permits identification of species from single femtomole quantities at abundance levels below 100 molecules/cell. Fraction collection of LC eluent on-line with each LCMS run provides a means to correlate CE signal with mass spectral identification and facilitates quantitative, high-content screening in a 96-channel CE format. Applications of this platform to the identification of novel drug targets and to high-content drug screening are presented.

R. M. Caprioli;
Vanderbilt University, Nashville, TN, United States.

Imaging Mass Spectrometry is a relatively new technology that takes advantage of the methodology and instrumentation of matrix-assisted laser desorption ionization (MALDI) mass spectrometry. It can locate specific molecules such as peptides and proteins up to about 80,000 Daltons directly from fresh frozen tissue sections or blots of a tissue. The data from such an analysis is a pictograph, in real x, y dimensions, of the location of a signal at any given molecular weight. Using a raster of mass spectra over a given area of a section, images of samples are produced in specific mass-to-charge (m/z) values, or ranges of values. Each spot on the sample irradiated by the laser is approximately 35 microns in diameter and the resulting spectrum contains hundreds of protein signals. Individual m/z values can be assembled from these mass spectra to produce selected m/z images. Tissue slices from both human and mouse brain, as well as from mouse prostate and colon, have been profiled and imaged to locate and map tissue specific peptides and proteins. For example, tumors present in these tissues, such as human glioblastomas, have been analyzed in this way. Data analysis is then able to reveal new proteins highly expressed in the tumor but not the corresponding normal tissue, and vice versa. Imaging Mass Spectrometry is an effective discovery tool for the analysis of proteins and peptides in tissues. This technology has also been applied to drug targeting and metabolic studies and the measurement of concomitant protein changes in specific tissues after systemic drug administration. The specific advantages and capabilities of the technique as well as its limitations will be discussed.

J. D. Aitchison;
Insitute for Systems Biology, Seattle, WA, United States.

Systems biology approaches are being employed to comprehensively understand all the cellular events involved in the formation, maturation and turnover of functional yeast peroxisomes.
Why study peroxisomes?
• They are medically relevant: Peroxisomes are essential organelles, a fact that is underscored by the lethality of the peroxisome biogenesis disorders, a group of genetic disorders attributable to peroxisome dysfunction through compromised peroxisome assembly. Furthermore, the roles of peroxisomes in diverse metabolic processes (e.g., the β-oxidation of fatty acids; the synthesis of cholesterol, bile acids and plasmalogens; and the decomposition of hydrogen peroxide and superoxides) link them to a number of other human health concerns including aging, cancer, heart disease, obesity, and diabetes.
• Their biology is poorly understood and their study can provide a model for more complex differentiation processes: Peroxisomes are dynamic organelles. They undergo a maturation process that requires the timely activity of a number of proteins and the import of specific enzymes. However the developmental program governing this maturation is poorly understood.
The approach: To identify proteins involved in peroxisome biogenesis and function we are applying novel large-scale proteomics techniques to provide an exhaustive list of peroxisomal proteins and microarray analysis to define the differential expression of genes underlying peroxisome assembly in yeast. Results from these approaches have yielded large lists of candidate proteins, which are being subjected to "high throughput" phenotypic screens, computational clustering and visualization tools to illuminate high priority proteins for "low throughput" cell biological characterization. This approach has permitted the identification of several novel proteins required for normal peroxisomal function and biogenesis.

Y. Shiio^1,2, A. Orian¹, R. Aebersold², R. N. Eisenman¹;
¹Fred Hutchinson Cancer Research Center, Seattle, WA, United States, ²Institute for Systems Biology, Seattle, WA, United States.

The Max transcription factor network comprises a group of highly conserved and widely expressed transcriptional regulatory proteins, which function in cell proliferation and differentiation. This network includes members of the Myc and Mad families, as well as Mnt, and Mga. All of these proteins possess basic-helix-loop-helix-zipper domains (bHLHZ), which mediate dimerization with the small bHLHZ protein Max, thereby forming heterodimers capable of recognizing specific DNA sequences. Recent evidence suggests that the transcriptional activities of these heterodimers derive from their ability to recruit chromatin modifying complexes to DNA. Importantly these heterodimers function as activators or repressors and appear to act to antagonize each others activity. These proteins influence cell behavior and their deregulation is associate with tumorigenesis.
In order to obtain a global view of how the network functions we have used two different approaches: chromatin profiling and quantitative proteomic analysis. For chromatin profiling we have employed a recently developed method to "mark" genomic regions bound by a specific transcription factor with a tethered Dam methyltransferase. Using the Drosophila dMyc, dMax and dMnt proteins fused to Dam we find a surprisingly large number of both overlapping and non-overlapping binding sites estimated at 15% of Drosophila coding regions. In addition we have employed quantitative proteomic analysis with ICATTM reagent labeling to examine changes in protein expression in whole cells and subcellular compartments following. Our findings define major functional pathways that are influenced by the expression of the Max network proteins. Together our findings indicate that Max network transcription factors exert a profound effect on global gene and protein expression.

T. Galitski;
Institute for Systems Biology, Seattle, WA, United States.

In response to low available nitrogen or carbon levels on a solid growth medium, S. cerevisiae cells differentiate from a yeast-form morphology to an adhesive filamentous form that invades the substrate. Adhesion and invasion during filamentous growth reflect changes in several cellular properties including budding pattern, cell elongation, and cell wall composition. Through various means, almost 200 yeast genes have been implicated in filamentous growth.
In order to map the genetic network involved in this process, we applied quantitative assays of filamentous growth phenotypes in a high-throughput screen for genetic interactions. In the filamentation-competent Sigma1278b background, we constructed sets of single-mutant and double-mutant strains for genes involved in filamentous growth. The mutations included gene deletions, plasmids bearing dominant alleles, and multicopy plasmids over-expressing wild-type genes. Agar adhesion and invasion phenotypes under different growth conditions were measured by imaging arrays of strains in plate-washing assays. Quantitative data were extracted using an image analysis program.
We identified and classified genetic interactions in the network by scoring the phenotypes of wild type, single mutants, and double mutant combinations. We developed computational methods and software to infer and visualize a network of genetic interactions and to relate the observed genetic interactions to current understanding of the filamentation network. The results of our analysis reveal information flow in the filamentation network and identify functional interactions that control adhesive and invasive behavior.

E. S. Kawasaki;
NATIONAL CANCER INSTITUTE, Gaithersburg, MD, United States.

With the completion of the Homo sapiens genome sequence it is now possible to analyze the expression of the entire human gene complement (>30,000). To accomplish this remarkable feat, DNA microarrays have come to the forefront as the main technological workhorse for gene expression studies. To date, the detection platform for most microarrays have relied on short (25 base) oligonucleotides synthesized in-situ, or longer, highly variable length DNAs obtained from cDNA libraries. A third choice, long oligo (50-80 bases) arrays, is now available and may eventually eliminate the cDNA variety. We are in the process of validating this platform by comparing its performance with the commonly used short oligonucleotide and cDNA arrays. The test platforms contain a ~21,000 70mer array from Operon; a ~17,000 25mer array from Affymetrix; and a ~10,000 cDNA array from Incyte. Prostate and breast cancer cell lines have been chosen for gene expression comparisons, and large amounts of total RNA have been isolated from each to ensure that RNAs from the same batch are used for probe production in all experiments. Genes in common to all platforms have been catalogued and primers for a select set of genes are being obtained to perform Quantitative RT-PCR for validation and determining absolute levels of expression. Very preliminary results indicate that ~97% of the common genes show the correct direction of upregulation or downregulation when compared at the 2X increase/decrease criterion. However, there appears to be large differences in expression ratios among the different platforms, with the oligo systems exhibiting a wider dynamic range. Further bioinformatic/computational methods are being applied to the data, along with QRT-PCR to verify these results.

M. X. Tan, H. T. Beernink, P. Wagner;
Zyomyx, Inc, Hayward, CA, United States.

Proteomics is increasingly dependent on analytical tools that focus on quantification of protein expression, biomolecular-protein interactions, and functional activity. To address these needs, Zyomyx has developed a novel protein biochip platform that facilitates rapid, precise, highly multiplexed analysis with minimal sample requirements. The biochip architecture employs a 3-dimensional array of raised features designed to provide consistent feature size and placement, while eliminating spot-to-spot cross contamination. Together with a fully automated assay station and fluorescence imaging scanner, the Zyomyx system provides a complete platform for robust, high-level quantitative protein analysis.
Based on this platform, our first product is a Human Cytokine Biochip capable of fully multiplexed measurement of 30 soluble proteins using minimal sample volume. The biochip uses sandwich-immunoassay methodology for sensitive and highly specific, quantitative measurement, and is compatible with a number of biological sample types including complex body fluids (serum, plasma, synovial fluid), tissue culture supernatants, and cell lysates. Applications of the human cytokine biochip will be discussed.

M. Fero;
Stanford University Medical School, Stanford, CA, United States.

The Stanford Functional Genomics Facility Produces gene expression microarrays for the Stanford Academic community. The demands of large scale production have shaped the nature of our facility. How we handle the large production volume (10-20,000 arrays per year), process tracking, the large amount of data produced, and how we have been able to use the considerable amount of data generated with our arrays to improve our process and detect process errors will be discussed. The error model for individual features on our arrays, our doping control regimen, and the development of our whole genome oligo array will be described as well.

J. C. Sanchez, A. Scherl, J. M. Deshusses, J. Burgess, N. Walter, G. L. Corthals, D. F. Hochstrasser;
Geneva University Hospital, Geneva, Switzerland.

The physical manifestation of organism including their morphology, physiology and behavior (i.e. their phenome) is characterized by complex interactions between genes, gene products, metabolites and environmental stimuli forming genetic-regulatory motifs and metabolic pathways. Further insights in the distinct levels of cell functions, structures and organizations are examined by genomics, transcriptomics, proteomics and metabolomics approaches. The qualitative and quantitative investigation of proteomes under different conditions to further unravel biological processes is often performed by 2-DE protein profiling and mass spectrometry identification. However, many of the low-copy number proteins as well as the membrane proteins are difficult to display on a 2-DE. It is thus clear that the separation of these particular proteins in amounts sufficient for post-separation analysis is an important issue in proteome studies and presents a challenge for 2-DE techniques.
One means of enriching rare polypeptides prior to 2-D PAGE separation is to use specific molecular properties of proteins of interest including their charge, size, hydrophobicity, solubility and location. Sub-cellular fractionation should provide amounts sufficient to help to gain insight into the multiple functions fulfilled by organelles. We report here, the purification and proteomics analysis of nucleoli to determine their protein content and potentially unravel complex regulatory events in cells.
Most of the cell membranes are easy to fractionate from other cellular compartment. However, membrane proteins are under-represented in 2-DE because of their intrinsic predilection to self-aggregate in the first dimension (IEF). We report here that addition of trifluoroethanol (TFE) significantly improves the 2-DE separation of proteins from intact cells as well as Escherichia coli and Staphylococcus aureus membranes.

R. A. VanBogelen;
Pfizer, Ann Arbor, MI, United States.

Proteomics has contributed to the parts list through many "Proteome Mapping" efforts. But proteomics largest contribution to systems biology will be in providing molecular information about the behavior of systems - the molecular physiology. The term proteomics signature is defined as the subset of proteins whose alterations in expression is characteristic of a response to a defined condition or genetic change. Signatures often relate to specific pathways or function. In drug discovery we are learning that signatures have enormous power to diagnose mechanism of action of compounds. In most cases we do not understand the rational for the signatures - why sets of proteins have altered expression when a cell process becomes dysfunction. However we have learned to trust the diagnosis through validation of our findings by other approaches.

A. I. Nesvizhskii, A. Keller, J. K. Eng, R. Aebersold;
Institute for Systems Biology, Seattle, WA, United States.

Over the past few years, tandem mass spectrometry (MS/MS) has been used increasingly for high throughput analysis of complex protein samples. A major challenge lies in the processing of the large amount of data generated in LC-MS/MS based protein profiling experiments. Acquired MS/MS spectra are first searched against sequence databases using automated database searching algorithms such as SEQUEST or Mascot. The database search results are then used to derive a list of identified peptides and their corresponding proteins. This task necessarily entails distinguishing correct from false identifications among a large number of assignments. We will describe our recent progress toward developing robust and accurate statistical models to access the validity of peptide and protein identifications made by MS/MS and database search. Our method applies machine learning techniques to distinguish correctly from incorrectly assigned peptides in the dataset and computes for each peptide assignment a probability of being correct (A. Keller, A.I. Nesvizhskii, E. Kolker and R. Aebersold, Anal. Chem., 74, 5383 (2002)). The list of peptide assignments and their corresponding probabilities is then used to estimate the likelihood that their corresponding proteins are present in the sample. Our method obviates the need to manually validate database search results in the case of all but most uncertain protein identifications. By using our method one can achieve much higher sensitivity for any given error rate compared to the results of using conventional filtering criteria. In addition, it can serve as a common standard by which the results of different experimental groups, using different mass spectrometers, and even different database searching tools, can be compared.

P. Braun, J. LaBaer;
Harvard Medical School, Boston, MA, United States.

The utility of protein microarray technology and other HT biochemical methods is currently limited by the lack of HT protein expression and purification methods. Structural proteomics as well requires methods for the HT screening and production of protein targets. The Harvard Institute of Proteomics is developing several key steps in recombinant protein production by building the FLEXGene cDNA repository and developing protein expression and purification methods.
The FLEXGene repository provides cDNAs for several model organisms including human in recombinational cloning systems and these will enable the rapid and HT assembly of expression clones for any protein expression system. In addition, the Institute has developed HT protein expression conditions for live bacteria and bacterial lysates. In developing bacterial protein expression several protein affinity tags have been explored. The data of 32 test set proteins indicate that the GST and MBP tag are suited to HT protein purification under non-denaturing conditions. Evaluation of the methods for denaturing and non-denaturing purification conditions on several hundred proteins indicates success rates of 67% under denaturing conditions and ~50% under non-denaturing conditions using the GST-tag. Biochemical and biophysical parameters were identified that influence protein purification success. These parameters may be used for target selection in structural proteomics.

C. von Mering;
European Molecular Biology Laboratory, Heidelberg, Germany.

Large-scale experimental and computational approaches are currently generating valuable protein-protein interaction data in many model organisms. One of the best-covered organisms is yeast, and a comparative assessment of yeast interaction datasets can help identify the potential of the various approaches, as well as their biases, strengths and weaknesses. Various ways to benchmark and compare interaction data are discussed, as well as approaches to integrate them into networks of interacting proteins. Particular emphasis is placed on computational methods for predicting functional interaction networks.

L. Salwinski;
UCLA - Howard Hughes Medical Institute, Los Angeles, CA, United States.

The recent development of high-throughput methods for detecting protein interactions resulted in rapid accumulation of genome-scale datasets. In order to integrate this vast amount of data with the results of conventional small-scale experiments the quality of high-throughput results has to be assessed. The expression profile reliability (EPR) index compares the characteristic features of the genome-scale datasets with the properties of the high-reliability interactions listed in the DIP database. The method estimates the biologically relevant fraction of protein interactions detected in a high throughput screen by comparing the RNA expression profiles for the proteins whose interactions are found in the screen with expression profiles for known interacting and non-interacting pairs of proteins. The method was applied to the recent high-throughput interaction datasets to asses their quality and thus to identify strengths and weaknesses of different high-throughput experimental approaches.

S. Bayliss;
Wellcome Trust Clinical Research Facility, Edinburgh, United Kingdom.

The University of Edinburgh Wellcome Trust Clinical Research Facility (WTCRF) Genetics Core has recently completed the set-up of a dedicated SNP typing laboratory using the ABI PRISM 7900HT Sequence Detection System as the detection platform and the Beckman Coulter Biomek FX to provide liquid handling during assay set-up. The tutorial will address the issues that led to this decision, including client genotyping strategy requirements, setting and assessing detection platform criteria as part of a European Union Tender, cost analysis and choosing automation. The tutorial will cover the use of the 7900HT and Biomek FX within the Genetics Core, especially our early experiences with the instruments, sample ID tracking efforts and the integration of Applied Biosystems Assays-by-Design and Assays-on-Demand products in to the Genetics Core genotyping pipeline. Pilot study results will be presented to show the accuracy and reproducibility of the 7900HT as well as further optimisation of reagent usage on the platform.

R. Spangler, N. Goddard;
The Rockefeller University, New York, NY, United States.

The tutorial will address most of the issues associated with establishing SYBR-PCR to measure small differences for many mRNAs in many samples of animal tissue. The tutorial will cover experimental design, tissue handling, high-throughput purification of RNA, optimizing cDNA synthesis, design of primers, Optic-PCR and low-level data analysis. Findings from studies employing this methodology will be presented to validate the comparative measures approach.

B. J. Madden;
Mayo Clinic, Rochester, MN, United States.

Although Edman sequencing is playing a lesser role for routine protein identification in many laboratories, it still provides an easy and reliable way to generate amino acid sequence when needed. Those performing Edman sequencing will often see a wide variety of samples submitted for analysis, with the resulting data ranging from simple straightforward assignments to chromatograms that are impossible to interpret. While homogenous samples pose minimal problems, heterogeneous samples create the challenge of being able to extract sequence assignments and ultimately, protein identifications from the complex chromatograms. The results of the ESRG 2002 sample illustrated drastic differences in each participating laboratory’s ability to identify a protein with a heterogeneous N-terminus, despite the sample being run on very similar instruments. This tutorial will touch on several aspects of generating, analyzing and interpreting Edman sequencing data and highlight strategies that can be employed when dealing with samples yielding multiple amino acids per cycle. Included will be the search algorithms and databases that might be used for different situations. Through the use of example sequencing runs of increasing complexity, the aim of the tutorial is to show that it may be possible to identify protein or proteins with data that, at first, may appear useless. A survey of reporting methods will also be presented based on the input from the ESRG 2003 sample study participants.

D. Bintzler¹, R. Scholl²;
¹1University of Cincinnati, Cincinnati, OH, United States, ²University of Utah, Salt Lake City, UT, United States.

In genetic studies, the identification of single nucleotide polymorphisms (SNPs) in coding, non-coding, or regulatory DNA sequences can reveal important information about the phenotype of an individual organism. A PCR-based amplification technique called single base extension (SBE) or single nucleotide extension (SNE) can be a useful tool in screening and validating known SNPs using slab gel or capillary DNA analyzers. The incorporation of a fluorescent-labeled dideoxy nucleotide by DNA polymerase to a primer that anneals immediately adjacent (and 5 prime) to a known SNP terminates the reaction. The resulting DNA fragment is then separated by electrophoresis, and the SNP alleles identified based upon size and color. Because this methodology employs the use of DNA analyzers, it can be easily integrated into existing DNA resource facilities or labs where these instruments are already available. Material will be presented highlighting some basic protocols used for sample preparation, data collection and analysis based on the use of the "SNaPshot" single nucleotide extension kit from Applied Biosystems and can be applied to other systems such as MegaBACE and the Transgenomics Wave system. Discussion will include primer design, multiplexing reactions, clean-up methods, workflow timelines, and pitfalls encountered.

A. I. Brooks¹, S. Levy², A. Viale³, G. Grills⁴;
¹University of Rochester Medical Center, Rochester, NY, United States, ²Vanderbilt University, Nashville, TN, United States, ³Memorial Sloan Kettering Cancer Center, New York, NY, United States, ⁴Harvard University, Boston, MA, United States.

Genomic technologies have become one of the most important core infrastructures that an academic institution can offer its faculty. The magnitude and complexity of emerging genomics technologies has given rise to the development of core facilities unlike any others. In addition, the integration of high-throughput approaches to investigating gene expression, protein expression and metabolites has forced the evolution of facilities that are expert in experimental design, sample preparation and most importantly, data analysis. This tutorial will provide an overview of the essential components of constructing a high-throughput genomics facility using gene and protein expression platforms as specific examples. More specifically this presentation will include presentations on microarray experimental design, management and propagation of genetic content, microarray production, sample processing, adaptation of commercial microarray technologies (including protein arrays), QA and QC measures for both array production and sample processing, validation techniques and an overview of data analysis approaches utilized by these core facilities. The proper construction and management of a genomics shared resource is becoming increasingly more important as data generated by these entities are rapidly becoming a part of national data repository efforts. This tutorial will provide useful information for institutions thinking of building a shared resource as well as those with established programs.

P. S. Cooper;
National Center for Biotechnology Information, Bethesda, MD, United States.

The NCBI has a growing collection of complete and draft-level complex genomes. Most notably this includes the human and mouse genomes as well as the sequence of the fruit fly, C. elegans, Arabidopsis thaliana, and Anopheles gambiae. Three important and integrated avenues of access to genome data are the Map Viewer, LocusLink and UniGene. The Map Viewer serves as a common graphical platform for displaying various kinds of maps and provides access to annotations, sequences and variations (SNPs). It allows both text queries and sequence similarity searching through the genomic BLAST pages. For several of the complete genomes as well as other important experimental organisms, the NBCI has created LocusLink, a central point of access to primary and reference sequences including assembled contigs, curated transcripts and proteins, gene models, expressed sequences, mapping information, functional data and nomenclature. The richest sources of sequence data for many organisms are expressed sequences in the form of ESTs. The NCBI organizes and reduces this highly redundant but important set of data in the UniGene collections. The MapViewer, LocusLink and UniGene are fully integrated with each other and are integrated with the Entrez / PubMed and BLAST systems. In this tutorial I will show how to access genome information using these tools by sequence similarity, marker interval, gene name, and sequence identifier and demonstrate how they can be used for gene discovery. During the tutorial I will also provide time to address questions from the audience and will be able to provide, live, online demonstrations of the NCBI tools and databases in answering these. Specific questions can be sent to me ahead of time through email (cooper@ncbi.nlm.nih.gov).