The advantages of microcapillary HPLC over conventional, larger scale HPLC have long been recognized. They include increased chromatographic resolution, greater sensitivity, and greatly reduced solvent consumption (for reviews see references 1 and 2). In addition to being used to separate and collect picomole to femtomole amounts of proteins and peptides by reversed-phase chromatography, microcapillary HPLC has been used in affinity and ion-exchange chromatography and as an on -line, upstream component of electrospray mass spectrometry. This workshop was held to introduce some of the most recent and advanced applications of microcapillary HPLC.
Kristine Swiderek, "Fast screening of protein samples by automated microspray LC-MS/MS analysis: A practical application for core facilities?"
Two main issues were addressed in the first presentation: how to set up a microcapillary HPLC system and the application of such a system to automated microcapillary LC-microspray MS/MS analysis.
A capillary HPLC system consists of several elements: a column, a detector, and a solvent delivery system. All components are commercially available or can be manufactured in your own laboratory. The preparation of columns with 100-500 mm internal diameters (ID) has been described in detail (1-3) using fused silica, glass-lined stainless steel or plastic tubing such as PEEK. Packing fused silica columns requires tubing with various IDs, fast-setting epoxy glue, chromatography matrices (bulk material or reclaimed resin from old, large-scale columns), and an HPLC system for the actual packing process (1, 2). Commercial UV detectors with small-volume flow cells are available for capillary HPLC, but sometimes the volumes of the detector's inlet and outlet lines must be lowered to accommodate low flow rates and to prevent loss of resolution. Alternatively, the smallest volume flow cells can be made in the laboratory by burning the polyimide coating off a piece of fused silica tubing (2). At present, solvent delivery systems capable of flow rates below 10 ml/min are not commercially available. However, existing HPLC systems can be easily modified with solvent splitters to provide low flow rates (1-3). Extra care must be taken to minimize dead volume between the point of gradient formation and the column as much as possible to avoid overly long delays in delivery of the gradient to the column. An alternative to a split-flow system is the gradient loop system, which can be built in the laboratory (4).
Changing from a conventional 2.1 mm to a 0.5 mm ID reversed-phase column will increase sensitivity 30-fold and will improve resolution. Typical flow rates for 0.5 mm ID columns are about 20 ml/min, and as little as 10 pmol of sample can provide enough material to collect fractions for further structural analysis. If column ID is reduced further to 0.35 mm, flow rates can be lowered to 2 ml/min for MALDI-MS sample collection or for on -line electrospray mass spectrometry. For flow rates less than 0.5 ml/min, 100 mm ID columns are used. These microflow columns in combination with electrospray mass spectrometry allow the structural analysis of peptides and proteins (microspray -microcapillary LC-MS) at the fmol level.
Microspray LC-MS will give a 2- to 3-fold increase in sensitivity over conventional electrospray, will lower sample consumption, and will reduce spectral background due to solvents. The combination of automated, data-controlled tandem mass spectrometry with microspray LC-MS (LC-MS/MS) demonstrates the full potential of this technique and several applications were shown to illustrate this (5). Using a 100 mm ID column, 1 pmol of cytochrome c digested with Lys-C was analyzed by automated LC-MS/MS, and the obtained collision-induced dissociation (CID) data were correlated with a protein database using SEQUEST (10). Eight peptides, accounting for 84% of the entire protein sequence, could be structurally identified by their CID spectra. Injecting 50 fmol of the same digestion mixture gave CID spectra still covering 55% of the complete sequence.
In a second example, an unknown 90 kDa protein in an SDS -PAGE gel was digested in situ with trypsin, and 3% of the digestion mixture was analyzed by LC-MS/MS. In a single analysis, 15 peptides were structurally matched to human endoplasmin, corresponding to 26% of the entire protein sequence. 97% of the digestion mixture was separated on a 0.5 mm ID reversed-phase column, and some of the fractions were subjected to automated Edman degradation. Two fractions gave unambiguous sequence matches with endoplasmin, and one fraction gave two sequences that could be interpreted clearly once the protein had been identified. Based on the Edman degradation results, the estimated amount of peptides analyzed by LC-MS/MS was 200 fmol or less.
In conclusion, microcapillary HPLC is essential for chromatography of pmol amounts of samples. Microcapillary systems can be built from commercially available or homemade components or with a combination of both. Implementing microspray-microcapillary LC-MS/MS is more demanding. However, if the instrumentation is available, this technique represents one of the most powerful approaches for the structural analysis of proteins and peptides in the fmol range currently available.
Axel Ducret, "Capillary LC-ESI-MS/MS for the identification of proteins with known sequence and de novo protein sequence analysis"
The second presentation discussed the automation of capillary HPLC and how it can be used in combination with mass spectrometry to identify proteins and to determine de novo protein sequences.
The on-line tandem mass spectrometric analysis of peptides purified by reversed-phase chromatography provides information about sample purity and concentration without requiring additional manipulations. Furthermore, tandem mass spectra can be correlated with sequence databases and often supply enough sequence information for protein identification.
A modular instrument configuration was described allowing sample analysis either in [i] collection mode, where 90% of the purified peptides remain available for further investigations, or in [ii] automated mode, where all the sample is analyzed by ESI tandem MS. The instrumentation consists of: a refrigerated autosampler, capable of unattended injection of up to 96 samples; a microbore HPLC system featuring in configuration [i] a 1 mm ID column operating at a flow rate of 50 ml/min or in configuration [ii] a 0.5 mm ID column at 15 ml/min; an on-column UV-cell; and a tandem mass spectrometer equipped with a regular electrospray ionization interface.
In configuration [i], a Tee flow splitter is placed between the UV-cell outlet and the mass spectrometer, and the split ratio is set by adjusting the lengths of the lines (7). The mass spectrometer automatically analyzes by tandem MS every parent ion that reaches a software-set intensity threshold. Based on the mass of the parent ion, the mass-dependent fragmentation energy is automatically calculated and set (7). At the end of each LC run, an automated version of SEQUEST (6, 8) correlates the obtained CID spectra with a sequence database and summarizes the results in a one-page, user-friendly report.
Using the configuration [ii], the tryptic digests of 70 spots originating from a 2D-SDS-PAGE separation of a total cell extract from yeast were successfully investigated in a completely automated manner. Two series of analysis were performed with 40 and 30 samples each and a LC run time of 36 minutes per analysis. SEQUEST identified 62 proteins from 58 digests, indicating that some spots contained more than one protein. One digest was found to contain a protein whose sequence was not deposited in a sequence database, and 11 digests did not give enough material for analysis by this system. Based on spot intensities and experiments performed with standard proteins, the detection limit of this instrument configuration is estimated to be 50-100 fmol, with approximately 500 fmol needed to obtain interpretable tandem mass spectra. This corresponds roughly to 1-2 pmol protein loaded onto SDS-PAGE. Configuration [i] is used to analyze proteins when a de novo sequencing by Edman degradation is necessary. Using this configuration, eluting peptides are routinely analyzed by mass spectrometry to assess their purity before sequencing them. In addition, tandem mass spectrometry usually provides an excellent means to proofread and complete chemical sequencing results. Combining Edman sequencing and tandem mass spectrometry allows correct sequence assignment with a mixture of two peptides, if the CID spectrum of at least one peptide is available. Although not extremely sensitive, this configuration allows us to routinely analyze unknown proteins loaded onto SDS-PAGE at the 10 pmol level with a level of confidence not obtained by either of these two techniques alone.
Dan Kassel, "Multi-dimensional HPLC-MS methods for optimizing receptor-ligand and enzyme-substrate binding"
The third presentation of the workshop covered the application of multidimensional capillary HPLC.
This approach in combination with electrospray ionization mass spectrometry was used to characterize complex combinatorial libraries for receptor binding and substrate optimization. It illustrated that capillary HPLC can be used not only for the separation and collection of fractions but also for studies of protein-peptide interaction through affinity-binding assays.
A microcapillary affinity column was prepared as follows: A biotinylated site was engineered at the carboxyl-terminus of the SH2 domain of pp60c-src. The biotinylated protein was then incubated with avidin-coated polystyrene beads slurried in Tris buffer, and a capillary affinity column was constructed by packing the slurry into a PEEKTm column (750 mm x 10 cm) at high pressure.
Receptor-binding assays were performed on-column in a flow-through format. An eight-component mixture of phosphorylated peptides, representing the autophosphorylation sites for epidermal growth factor receptor (EGFR), were assayed for binding to the SH2 domain. The phosphorylated peptide mixture was applied to the affinity column and allowed to interact with the affinity matrix for 5 minutes at 4°C without flow. Of the eight potential phosphopeptide ligands, four were selectively bound to the affinity column. The bound ligands were displaced from the column by competition with a high concentration (10 mM) of the high-affinity phosphopeptide Ac-pYEEIE, where pY is phosphotyrosine, and eluted directly onto a reversed-phase capillary column. The peptides were then identified by gradient elution from the reversed-phase column and mass analysis.
The results of this microaffinity/RP-HPLC/MS assay were corroborated by both surface plasmon resonance using a BIAcore instrument and microcalorimetry. Optimal substrates for Src tyrosine kinase (SrcTK) have been defined previously (9) from combinatorial peptide libraries using metal-chelate chromato -graphy and Edman degradation. Hydrophobic residues (e.g., Ile) amino-terminal to the tyrosine residue and charged residues (e.g., Glu) carboxyl-terminal to the tyrosine residue were identified as important characteristics of good substrates.
A two-dimensional HPLC/mass spectrometry method was developed to further define the characteristics of good SrcTK substrates. Using 15-component peptide libraries, residues amino -terminal to tyrosine were evaluated by capillary reversed-phase HPLC/MS, and the "optimal" sequence was identified as AcLQIY -amide. Based on these results, residues carboxyl-terminal to tyrosine were probed with a 225-component library having the sequence AcLQIYXXI-amide, where X represents non-charged amino acid residues. Because peptides in this library have no net charge, a capillary anion-exchange column was used to selectively isolate "true" substrates (i.e., peptides that became phosphorylated through the action of SrcTK). These substrates were then eluted from the anion-exchange column directly onto a reversed-phase capillary and then mass analyzed, as described above for receptor -ligand binding assays. The results were corroborated by preparing a set of discrete peptides based on the LC-MS data and testing them individually in SrcTK substrate assays.
References
1. Davis, M.D., Stahl, D.C., Swiderek, K.M. and Lee, T.D. (1994) Methods, A Companion to Methods in Enzymology, 6: Micromethods for Protein Structure Analysis, 304-314, ed. by John E. Shively, Academic Press, Inc., San Diego.
2. Swiderek, K.M., Lee, T.D. and Shively, J.E. (1996) Trace Structural Analysis of Proteins. Methods of Enzymology, ed. by Barry L. Karger and William S. Hancock, Spectrum Publisher Services, 271, 68-86.
3. Moritz, R.L. and Simpson, R.J. (1992) J. Chromatogr. 599, 119-130.
4. Davis, M.T., Stahl, D.C. and Lee, T.D. (1995) J. Amer. Soc. Mass Spectrom. 6, 571-577.
5. Stahl, D.C., Swiderek, K.M., Davis, M.T. and Lee, T.D. (1996) J. Am. Soc. Mass Spectrom., in press.
6. Eng, J., McCormack, A.L. and Yates, J.R. (1994) J. Amer. Soc. Mass Spectrom. 5, 976-989.
7. Ducret, A., Foyn Bruun, C., Bures, E.J., Marhaug, G., Husby, G. and Aebersold, R. (1996) Electrophoresis 17, in press.
8. Yates, J.R., Eng, J., McCormack, A.L. and Schieltz, D. (1995) Anal. Chem. 67, 1426-1436.
9. Cantley, L.C. and Songyang, Z. (1994) J. Cell Sci. Suppl. 18, 121-126.
Kristine M. Swiderek may be contacted at the Beckman Research Institute, City of Hope Hospital, 1450 E. Duarte Rd., Duarte, CA 91010; Axel Ducret at the University of Washington, Department of Molecular Biotechnology, P.O. Box 357730, Seattle, WA 98195, E-mail: axeld@u.washington.edu, URL: http://weber.u.washington.edu/~mbt/groups/aebersold/aeber -sold.html; and Daniel B. Kassel at CombiChem, Inc., 9050 Camino Santa Fe, San Diego, CA 92121.
Return to the The ABRF Home Page