Created: 28th February 1999, last updated: 7th April 1999, © 1999 ABRF
Donald G. Sheer and Aldo M. Pitt
Millipore Corporation, Danvers, MA
Reprinted from the electronic version of JBT available at http://www.abrf.org/JBT/JBT.html accession #0009.
Address correspondence and reprint requests to Aldo M. Pitt, Millipore Corporation, 17 Cherry Hill Drive, Danvers, MA 01923 (email: aldo_pitt@millipore.com).
High-performance liquid chromatography (HPLC) combined with mass spectrometry has become the method of choice for identifying and characterizing cell-expressed biomolecules. This technology has evolved so rapidly that efficient sample preparation in a high-throughput mode has become a rate-limiting step. Applications using C18 resin, 200 Å pore size, and 15-µm bead size silica and sulfonated divinylbenzene particles were tested. The convenient, solvent-resistant, 96-well MultiScreen filter plates with the Multiscreen Column Loader provided efficient removal of salts and detergents and excellent sample recovery for small volumes. The 96-well simultaneous, uniform loading of dry powders, beads, or resins in 45-, 80-, or 100-µL volumes easily accommodates various media capacities and elution volumes. Recovered eluates demonstrated high well-to-well reproducibility during analyte adsorption, washing, and elution. Sample recovery was analyzed by HPLC for a variety of proteins, peptides, and proteolytic digests. The utility of MultiScreen 96-well mini-columns in performing sample clean-up was also demonstrated for HPLC and mass spectroscopy. (J Biomol Tech 1999;10:21-25)
Key words: sample preparation, 96-well microplate format, structural characterization, mass spectroscopy, high-throughput screening.
High-throughput screening technology has rapidly evolved at most levels of assay development, although sample preparation remains a challenge. The use of a 96-well microplate format in combination with column loading for preparation of equivalent amounts of dry resin into 96 wells provides versatility in performing selective sample preparation. Similar to adsorptive column chromatography, optimization of sample recovery and reproducibility depends on the sample load, particle size, and the selected high-throughput screening mini-column. To demonstrate reproducibility and validation in the 96-well format as a sample preparation device, various sizes of functional resins were loaded in solvent-resistant MultiScreen plates using the 45-µL column loader. The 96 mini-columns containing dry resin were wetted and packed by centrifugation, bound with sample, washed, and eluted. Using cytochrome c, bovine serum albumin (BSA), and trypsinized cytochrome c, MultiScreen performance results are presented and summarized to demonstrate high-throughput sample preparation for downstream processing.
All assays were performed in MultiScreen-Resist LCR, 0.4-µm, hydrophilic, polytetrafluoroethylene (PTFE) resistant plates (Millipore Cat. #MAR4NO410) using the centrifuge ring (Cat. #MAFCO98F4). The 96 mini-columns were simultaneously loaded with the 45-µL MultiScreen Column Loader (Cat. #MACL09645) (Fig. 1).
FIGURE 1. The MultiScreen Column Loaders enable sample preparation for various amounts of sample using selected media and resins.
All centrifugations were performed on Jouan centrifuges at 2500 to 3000 g for 5 minutes using the CR312 or CR412 models with E4 or 4/ST rotors, respectively. The following adsorptive media, pore sizes, and bead sizes were used for MultiScreen loading. For reverse phase (C18), we used Amicon-Matrex C18 (300 Å, 15 µm and 300 Å, 40 µm); Waters-Microbondapak (300 Å, 10 µm and 300 Å, 37 to 55 µm); Polymer Labs PLRP-S (300 Å, 15 to 20 µm [divinylbenzene]).
For strong cation exchange (SCX), we used Purolite (300 Å, 10 to 22 µm [divinylbenzene]).
Dry C18 silica or polymeric (ie, divinylbenzene) resins were poured into the wells of the column loader, distributed with the beveled acrylic scraper (see Fig. 1), and loaded into MultiScreen as described in the operating instructions. The resultant MultiScreen mini-columns were packed by adding 300 µL of isopropanol to each well, followed by centrifugation at 3000 g for 5 minutes. Residual isopropanol was removed as columns were equilibrated in the respective binding solutions after receiving 200 µL of 0.1% trifluoroacetic acid (TFA, the reverse-phase resin) or 20 mM hydrochloric acid (HCl, the SCX resin) and centrifuged.
Samples were loaded in 200-µL aliquots of appropriate binding solution and centrifuged. Loosely bound sample, salts, and detergents were removed by performing a 200-µL wash step with the respective binding solutions during the 5-minute centrifugation.
Resin-bound analyte was recovered with 2 X 75 µL centrifugations containing desorption buffer of 90% acetonitrile in 0.1% TFA (ie, reverse phase) or 1.5 N ammonium hydroxide and 50% methanol in DI water (ie, SCX). Analysis for sample recovery was performed as described in the figure legends using visible (VMax plate reader, Molecular Devices, CA, USA) or ultraviolet 214 absorbance (HPLC, Shimadzu, Columbia, MD, USA).
To evaluate reproducibility and consistency of the MultiScreen mini-columns, three different types of reverse-phase media were packed into separate plates to receive various amounts of cytochrome c. Results were obtained for 25, 50, and 100 µg of sample applied to each plate (Fig. 2).
FIGURE 2. Comparison of C18-like resin performance in MultiScreen plates as a function of cytochrome c loading and bed volume. Samples were loaded in 200 µL of 0.1% TFA and recovered in 2 x 75 µL elutions by centrifugation using 90% acetonitrile and 0.1% TFA. Quantitation was determined against a standard curve by 405-nm absorbance in a 96-well plate reader or by reverse-phase high-performance liquid chromatography. Coefficients of variation were lower than 7% for all samples (n = 36 for all loads).
Day-to-day reproduciblility of cytochrome c recovery in the C18 loaded MultiScreen plates was compared by using different plates on separate days (Fig. 3). As expected, C18 silica exhibited higher capacity and recovery for cytochrome c at all loads (Table 1).
FIGURE 3. Day to day variability for C18 (300 Å pore size, 15-µm bead size) silica-loaded columns with increasing loads of cytochrome c. MultiScreen plates were prepared and assayed on consecutive days. Coefficients of variation were less than 7% for both days at all loads (n = 36).
TABLE 1.
Sample Recovery
|
|
||||||||||
Sample |
Applied (µg) |
Recovered (µg) |
Recoverya (%) |
CV (%) |
Capacity (µg/µL resin) |
|||||
|
|
||||||||||
| Cytochrome c | 40 | 33 | 84 | 5.17 | ||||||
| 20 | 18.5 | 91.3 | 7.32 | 1.45 | ||||||
| 10 | 8 | 80 | 10.34 | |||||||
| 5 | 2.5 | 50 | 11 | |||||||
| 1 | 0.45 | 45 | 11 | |||||||
| BSA | 40 | 12.5 | 30.8 | 20 | ||||||
| 20 | 5.5 | 28.2 | 15.4 | 0.8 | ||||||
| 10 | 4.3 | 44.1 | 9.1 | |||||||
| 5 | 2.38 | 47.6 | 5.1 | |||||||
| 2.5 | 1.85 | 662 | 7.4 | |||||||
|
|
||||||||||
aSample recovery is a function of cytochrome c and bovine serum albumin (BSA) loading by centrifugation of 45-µL packed mini-columns in MultiScreen plates. Samples were bound and washed in 0.1% trifluoroacetic acid (TFA) and water and eluted twice with 50 µL of 90% acetonitrile and 0.1% TFA. The cytochrome c concentration was determined directly by absorbance at 410 nm in a spectrophotometer and BSA by ultraviolet light at 280 nm or by BCA total protein assay (Pierce Chemical Co., Rockford, IL, USA), visible at 562 nm (n = 16).
The wide range of available derivatized resins allowed us to investigate the relation of cytochrome c loading and particle size. The small, spherical, high-capacity, 15-µm C18 resin is compared with the lower-capacity, 40-µm C18 resin in Figure 4. Results demonstrate that, in sample loads greater than 25 µg, cytochrome c binding significantly decreased in the 40-µm C18 columns.
FIGURE 4. Comparison of cytochrome c loading as a function of particle size using C18 silica packed MultiScreen plates. Mini-columns were prepared in separate plates. Relative amounts of cytochrome c were determined by absorbance at 405 nm using a plate reader (n = 32 for each concentration of cytochrome c).
The efficiency of divinylbenzene resins (C18 and SCX) for binding and elution with complex peptide mixtures is shown in Figure 5 for tryptic peptides from cytochrome c. With loads greater than 25 µg, sulfonated divinylbenzene exhibited a significantly increased capacity compared with PLRP-S. Selectivity for peptide binding and reproducibility from well to well was investigated using 50 pmol (~3.5 µg) of loaded trypsinized cytochrome c. Random wells were assayed as described in Figure 6.
FIGURE 5. Comparison of peptide binding and elution with C18-like and SCX resins in MultiScreen plates. Tryptic digest of cytochrome c was bound in 0.1% TFA (C18) or 20 mM HCl (SCX) and washed once with these binding solutions. Elution was performed twice with 50 µL of 90% acetonitrile and 0.1% TFA (C18) or 1.5 N ammonium hydroxide and 50% acetonitrile and water (SCX). Relative amounts of recovered peptides were determined by absorbance at 405 nm and subsequently quantitated by reverse-phase high-performance liquid chromatography (see Fig. 6). Coefficients of variation were less than 10% (n = 16).
FIGURE 6. Chromatogram comparison plots of cytochrome c tryptic peptides after binding and desalting in C18 MultiScreen plates. Eluates were recovered, speed vacuum dried, and resuspended in 0.1% TFA for reverse-phase high-performance liquid chromatography (HPLC) analysis. Peptides were separated on a 150 X 2 mm C18 Amicon-Matrex column (300 Å pore size, 15-µm column) using a 5% to 45% acetonitrile gradient over 40 minutes. Approximately 100 pmol (~7 µg) were loaded onto the column using a Shimadzu HPLC system with autosampler. Results were comparable for C18 silica- and resin-based PLRP-S (data not shown for C18 silica). Samples are as follows. (A) Starting solution and peptide digest loaded onto MultiScreen. (B, C, D, and E) Randomly picked eluates after MultiScreen desalting. The initial salt peak is reduced from MultiScreen-processed samples, but resolution is significantly improved for a variety of peptides (eg, 22 to 28 minutes in gradient).
HPLC analysis by reverse-phase methods demonstrates highly reproducible chromatogram comparisons of eluted cytochrome c digests. As expected, early HPLC-eluted hydrophilic peptides did not bind well in the C18-prepared columns.
The reported separations were performed by centrifugation to deliver the required uniformity for column packing, washing, and elution. For the small-size resins studied, vacuum filtration is not recommended. Particle size and sample loading were optimized for sample recovery. Of the reverse-phase matrices investigated, similar-size resins exhibited comparable performance. Reproducible sample recovery of complex peptide mixtures was demonstrated after sample desalting with C18 and SCX, and SCX was more selective for hydrophilic peptides.
Increased capacity is also available with 80- and 100-µL column loaders under optimized conditions. The combination of salt or detergent removal and stepwise batch separation (ie, by pH or hydrophobicity) can be performed with the 96-well format for downstream analysis (ie, mass spectroscopy, HPLC, and capillary electrophoresis). Results from this study demonstrate that consistent and reproducible chromatography in high-throughput screening can be achieved with the 96-well MultiScreen format.
We are grateful to Maria Lurantos and Jeffrey Busnach for their excellent technical contributions.
Dr. Don Sheer, Consulting Scientist for the Analytical Division of Millipore Corporation, was aboard Swissair Flight SR111 when it crashed on September 2, 1998. His wife, Diane, was also on the flight. Don was an active member of ABRF and The Protein Society for many years and made many valuable contributions in analytical sample preparation. Don will be missed by his colleagues and friends for his technical expertise, dedication, and numerous collaborative interactions with scientists in many fields.