Anal. Chem. 2003, 75, 5306-5316
Low-Attomole Electrospray Ionization MS and MS/MS Analysis of Protein Tryptic Digests Using 20-µm-i.d. Polystyrene-Divinylbenzene Monolithic Capillary Columns Alexander R. Ivanov, Li Zang, and Barry L. Karger*
Barnett Institute and Department of Chemistry, Northeastern University, Boston, Massachusetts 02115
This work explores the use of 20-µm-i.d. polymeric polystyrene-divinylbenzene monolithic nanocapillary columns for the LC-ESI-MS analysis of tryptic digest peptide mixtures. In contrast to the packing of microparticles, capillary columns were prepared, without the need of high pressure, in fused-silica capillaries, by thermally induced in situ copolymerization of styrene and divinylbenzene. The polymerization conditions and mobile-phase composition were optimized for chromatographic performance leading to efficiencies over 100 000 plates/m for peptide separations. High mass sensitivity (∼10 amol of peptides) in the MS and MS/MS modes using an ion trap MS was found, a factor of up to 20-fold improvement over 75-µmi.d. nanocolumns. A wide linear dynamic range (∼4 orders of magnitude) was achieved, and good run-to-run and column-to-column reproducibility of isocratic and gradient elution separations were found. As samples, both model proteins and tissue extracts were employed. Gradient nano-LC-MS analysis of a proteolytic digest of a tissue extract, equivalent to a sample size of ∼1000 cells injected, is presented. The high mass sensitivity identification and quantitation of large numbers of peptides from protein digests is a major goal in proteomics. Nanoflow liquid chromatography, using commercially available 75- and 100-µm-i.d. reversed-phase columns, offers the advantages of high resolution, high mass sensitivity, and low sample and mobile-phase consumption. However, analysis of a limited amount of sample (e.g., laser capture microdissected cells, immunoprecipitated proteins, 2-D gel spots, etc.) can still be challenging with the above columns. For a fixed limited amount of sample injected, columns with smaller inner diameter can decrease chromatographic band dilution1-3 and thus increase the signal for concentration-sensitive ESI-MS.4 However, narrow-bore columns (particularly less than 50-µm i.d.) are difficult to pack with conventional microparticles because of the very high pressure required to overcome the low column permeability.4-6 * To whom reprint requests should be sent. E-mail:
[email protected]. (1) Introduction to Microscale High-Performance Liquid Chromatography; Ishii, D., Ed.; VCH Publishers Inc.: New York, 1988. (2) Haskins, W. E.; Wang, Z.; Watson, C. J.; Rostand, R. R.; Witowski, S. R.; Powell, D. H.; Kennedy, R. T. Anal. Chem. 2001, 73 (21), 5005-14. (3) Unger, K. K. In Packings and Stationary Phases in Chromatographic Techniques; Unger, K. K., Ed.; Marcel Dekker: New York, 1990. (4) Shen, Y.; Zhao, R.; Berger, S. J.; Anderson, G. A.; Rodriguez, N.; Smith, R. D. Anal. Chem. 2000, 274 (16), 4235-49.
5306 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
Monolithic nanocapillary columns of 20-µm i.d. or less can be a good alternative to microparticle-packed columns because of their relative ease of manufacture, without the need of high pressure, and their high-performance characteristics.7-13 Relative to packed columns, porous monolithic nanocapillary columns also result in potentially less clogging of the ESI tip and do not require frits. Several approaches for synthesis of polymeric monoliths in capillaries of 100-300-µm i.d. have been published in which the reaction mixture is placed in the capillary, followed by UV or thermally induced in situ polymerization.7-9,11-14 High efficiencies of the narrow-bore capillary columns are found, due to decreased flow dispersion and a homogeneous packing bed structure. The bulk liquid flow in ultranarrow-i.d. capillary columns is reduced by 1 order of magnitude (15-50 nL/min), relative to 75-100-µmi.d. columns, which results in analytes being dissolved in much lower eluent volume with lower amount of ionic and neutral species of the mobile phase in the chromatographic band. The effect is higher mass sensitivity and higher electrospray ionization ability.1,2,4,15-20 Thus, highly sensitive detection is achievable when using ultra-narrow-bore LC-ESI-MS due to the fact that ESI is a primarily concentration-sensitive technique over a wide range of flow rate.20 (5) MacNair, J. E.; Patel, K. D.; Jorgenson, J. W. Anal. Chem. 1999, 71 (3), 700-8. (6) Wu, N.; Lippert, J. A.; Lee, M. L. J. Chromatogr., A 2001, 911 (1), 1-12. (7) Premstaller, A.; Oberacher, H.; Walcher, W.; Timperio, A. M.; Zolla, L.; Chervet, J. P.; Cavusoglu, N.; van Dorsselaer, A.; Huber, C. G. Anal. Chem. 2001, 73 (11), 2390-6. (8) Huang, X.; Zhang, S.; Schultz, G. A.; Henion, J. Anal. Chem. 2002, 74 (10), 2336-44. (9) Xie, S.; Allington, R. W.; Svec, F.; Fre´chet, J. M. J. Chromatogr., A 1999, 865 (1-2), 169-74. (10) Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Soga, N.; Nagayama, H.; Hosoya, K.; Tanaka, N. Anal. Chem. 2000, 72 (6), 1275-80. (11) Moore, R. E.; Licklider, L.; Schumann, D.; Lee, T. D. Anal. Chem. 1998, 70 (23), 4879-84. (12) Petro, M.; Svec, F.; Fre´chet, J. M. J. J. Chromatogr., A 1996, 752, 59-66. (13) Meyers, J. J.; Liapis, A. I. J. Chromatogr., A 1999, 852, 3-23. (14) Rohr, T.; Hilder, E. F.; Donovan, J. J.; Svec, F.; Fre´chet, J. M. J. Macromolecules 2003, 36, 1677-84. (15) Wilm, M.; Mann, M. Int. J. Mass Spectrom. Ion Processes 1994, 136, 16780. (16) Valaskovic, G. A.; Kelleher, N. L.; McLafferty, F. W. Science 1996, 273, 1199-202. (17) Gale, D. C.; Smith, R. D. Rapid Commun. Mass Spectrom. 1993, 7, 101721. (18) Fenn, J. B. J. Am. Soc. Mass Spectrom. 1993, 4, 524-35. (19) Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972A-86A. (20) Emmett, M. R.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 1994, 5, 60513. 10.1021/ac030163g CCC: $25.00
© 2003 American Chemical Society Published on Web 09/09/2003
In this paper, we report the use of 20-µm-i.d. monolithic nanocapillary columns for the high mass sensitivity analysis of small amounts of protein digests of model mixtures and tissue extracts by nano-LC-ESI-MS. We demonstrate the MS/MS analysis of peptides at the low-attomole level, high efficiency (over 100 000 plates/m), ease of manufacture, robustness (at least 1000 injections/column), and a wide linear dynamic range. These column characteristics suggest that ultra-narrow-bore monolithic columns are highly promising for proteomics and other applications. EXPERIMENTAL SECTION Materials. Fused-silica capillary tubing (20-µm i.d., 360-µm o.d.), with a polyimide outer coating, was purchased from Polymicro Technologies (Phoenix, AZ). Divinylbenzene, styrene, 1-octanol, tetrahydrofuran (THF), 2,2′-azobisisobutyronitrile (AIBN), sodium hydroxide, hydrochloric acid, vinyltrimethoxysilane, formic acid, and 2-propanol were obtained from Aldrich (Milwaukee, WI). Sequencing-grade TPCK-treated trypsin and standard peptides and proteins were obtained from Sigma (St. Louis, MO), and HPLC-grade methanol and acetonitrile were from Fisher (Fair Lawn, NJ). Breast ductal normal and carcinoma tissue sections were kindly provided by Dr. D. E. Palmer-Toy and Dr. Dennis Sgroi (Massachusetts General Hospital, Boston, MA). Deionized water (18.2 MΩ) was prepared using a Milli-Q system from Millipore (Bedford, MA). Caution: Styrene is known as a sensitizing and carcinogen agent. Proper precautions should be taken when handling this chemical. Column Preparation. Pretreatment and vinylization of the fused-silica capillaries were based on our previously published procedure.21 In brief, capillaries 50 cm long were flushed with 0.5 M sodium hydroxide for 30 min and then washed with deionized water for 5 min, followed by methanol for 5 min. Next, the inner surface of the fused-silica capillaries was treated with vinyltrimethoxysilane to attach covalently the anchoring vinyl sites for the grafting of the polymeric stationary phase. Capillaries were then rinsed for 40-50 min with a solution consisting of 0.36 M hydrochloric acid and 14.3% (v/v) methanol in vinyltrimethoxysilane. The rinsing solution was mixed and degassed by ultrasonication for 5 min before use. Subsequently, the capillaries were continually purged with helium for 1 h at 100 °C. Next, a polymerization mixture containing 16.5% (v/v) styrene, 16.5% (v/v) divinylbenzene, 7.9% (v/v) THF, 59.1% (v/v) 1-octanol, and 1% (w/v) AIBN was prepared, followed by filtering using a singleuse syringe with an 0.02-µm Millipore PVDF filter (Bedford, MA). The solution was degassed (air removal) by ultrasonication for 5 min, and then the vinylsilanized capillaries were filled with the polymerization mixture. Separate tests revealed column performance for helium degassing similar to that for ultrasonication for the small volume of the polymerization mixture of 2.4 mL. Both ends of the capillary were immersed in a 4-mL vial half-filled with the polymerization mixture and sealed by a screw cap with two silicon rubber septa (Alltech Associates, Inc., Deerfield, IL). The vial was pressurized (30-40 psi) with helium, and the vial with the capillary was heated in a GC oven at 75 °C for 3 h and then 85 °C for another 16 h. Next, 3-5-cm-long column sections were (21) Goetzinger, W.; Kotler, L.; Carrilho, E.; Ruiz-Martinez, M. C.; Salas-Solano, O.; Karger, B. L. Electrophoresis 1998, 19 (2), 242-8.
removed from both ends of the capillary, and the column was purged with nitrogen, washed with methanol, and purged again with nitrogen to eliminate any remaining polymerization solution. The capillary was cut into 10-cm-long pieces, and the column was then ready for use. The column was stored in a dry state before and after use. Tryptic Protein Digestion. Proteins (cytochrome c (horse), trypsinogen (bovine), myoglobin (horse), serum albumin (bovine), ovalbumin (chicken), β-casein (bovine), R1-acid glycoprotein (human), β-lactoglobulin (bovine), R-lactoglobulin (human), and catalase (bovine)) were dissolved in 1 mL of 6 M guanidine chloride, 50 mM Tris-HCl (pH 8.0), and 5 mM DTT and incubated for 1 h at 60 °C for denaturation and reduction. Then, an aliquot of protein solution was diluted 10-fold by 50 mM ammonium bicarbonate, pH 8.0, at a concentration of ∼1 mg/mL. Trypsin was added at a substrate-to-enzyme ratio of 100:1, and the solution was incubated overnight at 37 °C. The digest was vacuum-dried and reconstituted in water without additional cleanup steps before analysis. Preparation of Breast Tissue Samples: A breast ductal carcinoma tissue section (∼106 cells) was solubilized in 100 µL of PicoPure protein extraction buffer (Arcturus, Mountain View, CA, Catalog KIT 0104) with complete protease inhibitor (Roche Diagnostics GmbH, Mannheim, Germany) at 65 °C for 3 h. The solution was centrifuged at 15 000 rpm for 5 min. The suspension was then reduced and alkylated according to standard procedures (Promega, Madison, WI; Sigma). The sample was precipitated in 80% acetone at -20 °C overnight. After precipitation, the sample was centrifuged for 5 min at 15 000 rpm, and the supernatant was carefully decanted from the vial. The precipitate was reconstituted in 100 µL of 100 mM ammonium bicarbonate, and an aliquot (0.5 µL) of 20 µg/µL trypsin was added to the solution, followed by digestion at 37 °C for 10 h. A second aliquot (0.5 µL) of trypsin was added to the sample for an additional overnight digestion at 37 °C. Mass Spectrometry. An LCQ Deca XP+ quadrupole ion trap mass spectrometer (ThermoFinnigan, San Jose, CA) was used in all experiments. On-line ESI-MS was performed in the positiveion mode with the ESI voltage set at 0.5-1.4 kV and the heated inlet capillary maintained at 160 °C. All nano-LC-MS experiments were carried out with a maximum in-source sample injection time of 50 ms, and three microscans were summed for each scan. In all data-dependent MS/MS experiments, a full MS scan between 400 and 2000 m/z was followed by three full MS/MS scans for the three most intense ions. The relative collision energy was set to 35%, with an activation time of 30 ms. Dynamic exclusion was used with a repeat count of 2 and a repeat duration of 1 min, with a 3-min exclusion duration window. The activation time was set at 30 ms, with the default charge state of +2 and an isolation width of 3 m/z. The results of data-dependent MS/MS scanning were submitted for a search against a database created for the selected model proteins and the nonredundant protein database nr.fasta or, for the tissue extract sample, against the human NCBInr database using the TurboSequest algorithm (Bioworks 3.1, ThermoFinnigan). Nano-LC-ESI-MS. An UltiMate Micro Pump (Dionex, Sunnyvale, CA) equipped with an Accurate stream splitter was used to Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5307
deliver mobile phase through the column. The monolithic nanocapillary column was attached to a manual injector, VICI Model C4-0004-01 with a 10- or 200-nL internal loop (Valco Instruments Co., Houston, TX). ESI voltage (+0.5 to +1.4 kV) was applied to the inlet of the column at the injector. A simple design of a sheathless electrospray emitter was used to produce a stable spray. The outlet of the monolithic column was connected as close as possible to a spray capillary (20-µm i.d., 365-µm o.d., 5-µm tip orifice i.d., 1.5-2 cm, New Objective, MA) butt to butt, and the connection was held together with Teflon tubing (250-µm i.d., 1.6-mm o.d., 1.5 cm, Dionex). The estimated dead volume was calculated to be less than 5 nL. The ends of all capillaries were ground by a fine diamond sand lapper (VWR Scientific Products, West Chester, PA) before use. The position of the ESI emitter was carefully optimized within a distance of 0.3-1 mm to the mass spectrometer inlet orifice with use of a micromanipulator for maximum signal intensity. RESULTS AND DISCUSSION Preparation and Characterization of Monolithic Columns. As we have noted, for a given amount injected, reducing the column dimensions and the electrospray flow rate can result in an increase in analyte mass sensitivity at high separation efficiency.1-3 A decrease in the column dimensions from 75- to 20-µm i.d increases column resistance and thus raises the pressure requirements to pack such columns with microparticles. To overcome this problem, a polystyrene-divinylbenzene (PS-DVB) monolithic stationary phase was examined for use with the ultralow-i.d. capillary columns because high pressure is not required in their construction, and high chromatographic performance is potentially achievable.22-25 The procedure for producing the monolithic column consisted of (1) covalent modification of the fused-silica capillary inner wall by vinyltrimethoxysilane, (2) filling the narrow-i.d. capillary with a degassed polymerization mixture containing monomer, cross-linker, and inert porogens, (3) thermally induced in situ polymerization in the presence of initiator in a capillary pressurized from both ends to form a macroporous, rigid, and uniform structure covalently attached to the capillary inner surface, and (4) washing out the porogens and remaining polymerization mixture. Polymerization and crosslinking led to a phase separation to create permanent channels in the rigid polymeric material. The composition of the polymerization mixture was related to earlier studies7,12,22-24 but was carefully optimized to achieve high chromatographic performance. Covalent immobilization helped reduce shrinkage of the monolith during the polymerization7 and eliminated the need for retaining frits. Either thermal or UV-initiated polymerization could be used for synthesis of the monoliths in capillaries.24,25 Typically, heatinduced polymerization is a longer process,7,12,22-25 but it was chosen here since 20-µm-i.d. fused-silica tubing with a UVtransparent outer coating is not at present commercially available. We plan to explore UV-induced polymerization of nanocapillary monoliths soon. (22) Enlund, A. M.; Ericson, C.; Hjerte´n, S.; Westerlund, D. Electrophoresis 2000, 122 (3), 511-7. (23) Seidl, J.; Heitz, W. Adv. Polym. Sci. 1967, 5, 113-213. (24) Zhang, S.; Huang, X.; Zhang, J.; Horva´th, C. J. Chromatogr., A 2000, 887 (1-2), 465-77. (25) Peters, E. C.; Petro, M.; Svec, F.; Fre´chet, J. M. Anal. Chem. 1998, 70 (11), 2296-302.
5308
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
Standard procedures7,11,12,26 for column polymerization in lowi.d. capillaries were not found to be successful. Polymerization of monoliths narrower than 50-µm i.d. without column pressurization led to the appearance of gaps and a dramatic shrinkage of the polymer structure due to the fast kinetics of the exothermic polymerization reaction in the narrow-i.d. capillaries together with the relatively high surface-to-volume ratio. Pressurization during polymerization27 reduced monolith structure irregularities by minimizing both shrinkage27,28 and bubble formation with porous monoliths in the 20-µm-i.d. column. Typically, 40-50 psi applied from both ends of the capillary filled with polymerization mixture helped to achieve a uniform continuous monolithic structure. Constant pressurization during polymerization supplied additional portions of the mixture of monomers, which compensated for shrinkage. Pressurization also helped to improve column-tocolumn reproducibility. In preliminary experiments, we have found pressurization also enables synthesis of 15-µm-i.d. and narrower polymeric monolithic nano-LC columns. Careful optimization of the polymerization conditions was necessary for the preparation of high-performance monoliths. Careful control of the polymerization kinetics was also important for the specific morphology of the monolith. Temperature, reaction time, and polymerization composition all affected performance. A number of parameters were significant in producing a specific pore size including monomer type: solubility and reactivity, the amount of cross-linker, the amount and type of porogens, the solubility of polymer in porogens, and polymerization temperature and pressure.12,23,27 The middle 25-40-cm sections of a longer (50-60 cm) synthesized monolithic nanocapillary column were found to be similar in chromatographic performance when the polymerization process was carefully controlled. Scanning electron micrographs were acquired from ∼ 1-cmlong cuts of different column sections to examine the column morphology and surface structure. Figure 1 presents scanning electron microscopy (SEM) images of the monolith, demonstrating a well-ordered polymeric quasi-bead structure with attachment to the capillary wall. The monolith contained relatively large flowthrough channels (∼1.5-3 µm) that allowed rapid convective mass transport between the mobile phase and surface of the polymer. The relatively low backpressure for monoliths for a given column i.d. and pseudo-bead size (∼0.3-0.5 µm) at a typical flow rate of 20-50 nL/min was a result of this increased porosity. SEM images of different sections of the column appeared to be similar in structure. A variety of chromatographic studies were conducted to characterize the 20-µm-i.d. monolithic columns under isocratic conditions. The PS-DVB monolithic columns with pseudobeads smaller than 1 µm demonstrated high efficiency, over 100 000 plates/m (calculated from extracted ion chromatograms) for peptides over a range of capacity factors (3 e k′ e 10) using a 10-cm-long column, with typical peak widths at half-height of 5-15 s (Figure 2), as well as reasonable back pressure of 200-300 bar. To demonstrate the resolving power of the 20-µm-i.d. monolith, Figure 3 shows gradient nano-LC-ESI-MS analysis of a tryptic digest of a model 10-protein mixture containing hundreds of peptides at the (26) Svec, F.; Fre´chet, J. M. J. Macromolecules 1995, 28, 7580-2. (27) Bente, P. F., III; Myerson, J. U.S. Patent 4,810,456, 1989. (28) Huber, C.; Oberacher, H.; Premstaller, A. U.S. Patent Application 0088753 A1, 2002.
Figure 1. Scanning electron micrographs of the polystyrene-divinylbenzene monolithic packing in a 1-cm section of fused-silica capillaries of 20-µm i.d. The images correspond to 5000× (A) and 12500× (B) magnification. The monolith was synthesized as described in the Experimental Section.
Figure 2. Extracted ion chromatograms for selected tryptic peptides for isocratic nano-LC-ESI-MS on the monolithic column 20 µm i.d. × 10 cm (5 fmol injected on the column). Mobile phase: 15% (v/v) acetonitrile, 0.1% (v/v) formic acid in water. Flow rate: ∼25 nL/min. Column back pressure: ∼220 bar. Average efficiency exceeds 100 000 plates/m. Peak widths at half-height (w1/2): 12 s for DAQLFIQKKAVK, 6 s for FSTVAGESGSADTVR, 10 s for ADNRDPASDQMK, and 9 s for VLNEEQRK.
level of 10-40 fmol. Separation with subsequent on-line ESI-MS was performed with a 30-min gradient. A 3-D overlay (Figure 3A) and a planar ion density map (Figure 3B) demonstrate the complexity of the sample and the high resolving power of the method. Figure 3C presents examples of extracted ion chromatograms for several peptides with typical narrow elution windows of 15-40 s (at the retention time 8 e tR e 30 min). MS/MS datadependent scanning followed by TurboSEQUEST searching against a database created for the selected model proteins allowed
identification of all proteins in the mixture with a high score. The total number of fully tryptic peptides identified in a single run with 25-min gradient was 96 with a SEQUEST delta correlation (∆Cn) greater than 0.08 and a correlation (Xcorr) greater than 2.0 and, for singly charged ions, Xcorr greater than 1.5 for doubly charged ions and Xcorr greater than 3.3 for triply charged ions.29 (29) Peng, J.; Elias, J. E.; Thoreen, C. C.; Licklider, L. J.; Gygi, S. P. J. Proteome Res. 2003, 2, 43-50.
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5309
Figure 3. Gradient nano-LC-ESI-MS of a tryptic digest of a 10-protein mixture on the monolithic column (10-40 fmol injected on the column). Proteins: cytochrome c (horse), trypsinogen (bovine), myoglobin (horse), serum albumin (bovine), ovalbumin (chicken), β-casein (bovine), R1acid glycoprotein (human), β-lactoglobulin (bovine), R-lactoglobulin (human), and catalase (bovine). Mobile phase: (solvent A) 2% (v/v) acetonitrile, 0.1% (v/v) formic acid in water; (solvent B) 10% (v/v) water, 5% (v/v) 2-propanol, 0.1% (v/v) formic acid in acetonitrile. Gradient: 5% B, 0 min; 40% B, 25 min; 90% B, 26 min; gradient steepness parameter b ) 0.28. Injection was made and data acquisition commenced 10 min after the start of the gradient. (A) 3-D overlay of LC-MS chromatogram; (B) planar ion density map; (C) extracted ion chromatograms for selected peptides of the tryptic digest of the 10-protein mixture.
Special attention was paid to system dead volumes. Precolumn dwell volume and postcolumn dead volume were downscaled to 5310 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
∼200 and ∼5 nL, respectively. Injection volumes were maintained below ∼30% of the average eluting peak volume, as typical in
Figure 4. Linear dynamic range measurements for the monolithic column with a steep gradient (b ) 0.8, tg ) 10 min). A model peptide mixture containing equal amounts from 1 amol to 10 pmol each of Met-enkephalin (1, [), Leu-enkephalin (2, +), R-endorphin (3, ×), and neurotensin (4, b) was injected on the column. Table 1. Run-to-Run Reproducibility for the Retention Time of the Selected Catalase Tryptic Peptidesa tryptic peptides of catalase digest LGPNYLQIPVNCPYR m/z 903.5 IQALLDKYNEEKPK m/z 563.2 DALLFPSFIHSQK m/z 750.2 GAGAFGYFEVTHDITR m/z 582.3 GPLLVQDVVFTDEMAHFDR m/z 731.1 VWPHGDYPLIPVGKLVLNR m/z 1087.9 aA
run 1
run 2
retention time, min run 3
average
SD
RSD, %
28.48
28.15
28.36
28.33
0.17
0.59
29.68
29.26
29.46
29.47
0.21
0.71
29.68
29.33
29.51
29.51
0.18
0.59
26.87
26.51
26.68
26.69
0.18
0.67
38.16
38.08
38.08
38.11
0.05
0.12
38.85
38.82
38.82
38.83
0.02
0.04
total of 100 amol of digest injected on the monolithic column. For conditions, see Figure 3.
HPLC, to avoid band broadening. In the gradient elution mode, this injection limitation can be overcome by preconcentration either at the head of the column or with a vented trapping column placed at the head of the separation column.30 The monolithic column provided a wide linear dynamic range of loading capacity, as described as the maximum injection amount without the appearance of increased peak broadening. For example, the monolithic column resulted in a linear dynamic mass range of 3.5 orders of magnitude (100 fg-500 pg) for small peptides (Mw 555.6-1746.0, retention time 6 e tr e 18 min) on the column under a sharp gradient (b ) 0.8, tg ) 10 min; see Glossary for definitions), Figure 4. Peptides with higher retention (30) Licklider, L. J.; Thoreen, C. C.; Peng, J.; Gygi, S. P. Anal. Chem. 2002, 74 (13), 3076-83.
(10 e tR e 32 min) showed a linear dynamic range up to 4.5-5 orders of magnitude under shallower (b ) 0.42, tg ) 20 min) gradient conditions. This linear dynamic range of ∼4 orders of magnitude was found to be similar to that for a 75 µm i.d. × 15 cm packed capillary column under similar gradient elution conditions (0.4 e b e 0.8), but the mass range shifted by ∼20fold toward the lower injected amounts for the 20-µm-i.d. column. Retention reproducibility of the monolithic columns was tested for the separation of tryptic peptides of a bovine catalase digest (see Tables 1 and 2). Limited run-to-run and column-to-column repeatability studies were performed by measuring the relative standard deviation of peptide retention times from three independent sets of measurements. Run-to-run reproducibility was found to be better than 1%, and column-to-column (three columns) Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5311
Table 2. Column-to-Column Reproducibility for the Retention Time of the Selected Catalase Tryptic Peptidesa retention time, min tryptic peptides of catalase digest LGPNYLQIPVNCPYR m/z 903.5 IQALLDKYNEEKPK m/z 563.2 DALLFPSFIHSQK m/z 750.2 GAGAFGYFEVTHDITR m/z 582.3 GPLLVQDVVFTDEMAHFDR m/z 731.1 VWPHGDYPLIPVGKLVLNR m/z 1087.9 a
run 1
column 1 run 2
run 3
run 1
column 2 run 2
run 3
RSD, %
28.48
28.15
28.36
27.53
29.53
28.47
2.28
29.68
29.26
29.46
28.57
30.21
28.94
1.95
29.68
29.33
29.51
29.60
30.14
28.94
1.34
26.87
26.51
26.68
26.55
28.46
27.26
2.74
38.16
38.08
38.08
35.52
35.60
3.76
38.85
38.82
38.82
37.74
38.38
1.64
37.39
A total of 100 amol of digest injected on two different monolithic columns. For conditions, see Figure 3.
Figure 5. Gradient nano-LC-ESI-MS of a tryptic digest of bovine catalase on the monolithic column (10 amol of digest injected on the column). (A) Extracted ion chromatograms for selected peptides; (B) spectra at the peak maximum of the same peptides. See Figure 3 for conditions. 5312
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
Figure 6. SEQUEST match for the ESI-MS/MS spectrum of the selected catalase tryptic peptide LGPNYLQIPVNCPYR (10 amol of digest injected on the column). See Figure 5 for conditions.
Figure 7. Gradient nano-LC-ESI-MS of a tryptic digest of bovine catalase on the monolithic column (1 amol of digest injected on the column). Extracted ion chromatograms for selected peptides. See Figure 3 for conditions.
reproducibility was better than 4%. The columns could be flushed with solvents in either direction up to at least 300 bar pressure without any damage to the monolithic structure, and the columns could withstand prolonged exposure to mobile phases from pH 2.0 to pH 11.5 for at least several months (at least 1000 injections/ column). As noted by others, clogged columns can be reused after cutting a short piece (∼1-2 mm) from the clogged end or after grinding the end of the column with a diamond sand lapper. Monolithic nanocapillary columns also resulted in less clogging of the ESI tip in comparison to packed columns because of the absence either of loose particles in the column bed or frits and also because of higher chemical stability. Mass Sensitivity of Nano-LC-ESI-MS Using 20-µm-i.d. Monolithic Columns in Tryptic Peptide Analysis
A main characteristic of nanocolumns is high mass sensitivity, as a result of the decreased dilution of the chromatographic band.1-3 Theoretically, downscaling from conventional nano-LC columns of 75-20-µm i.d. should result in a gain in sensitivity of (d1/d2)2 ≈ 14 (for the same injected sample amount, linear column velocity, and column length). We observed a gain higher than 20-fold, which can be a result of the difference in efficiency and column length between the two columns. Monolithic 20-µm-i.d. nanocapillary columns operating at a flow rate 20-50 nL/min may also take advantage of nanospray ionization to increase MS sensitivity.20 Stable ESI conditions at ultralow flow rate can potentially result in the more efficient transfer of ions into the entrance of the mass spectrometer.2,4,15-19 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5313
Figure 8. SEQUEST match for ESI-MS/MS spectrum of the selected catalase tryptic peptide LGPNYLQIPVNCPYR (1 amol of digest injected on the column). See Figure 7 for conditions.
To illustrate the effectiveness of the nano-LC monolithic columns for the analysis of low-abundant analytes, a tryptic digest of bovine catalase was injected at the 10-amol level on the monolithic nanocapillary column. Figures 5A presents extracted ion chromatograms (EIC) for this level injected on the column, while Figure 6 illustrates MS/MS spectra with TurboSEQUEST database searching. Among 14 peaks corresponding to catalase tryptic peptides observed in the MS mode at 10 amol of the digest injected on the column, 3 provided good MS/MS fragmentation and high SEQUEST scores. Then, a tryptic digest of bovine catalase was injected at the 1-amol level on the column. Figure 7 presents EIC for this sample level. Correspondingly, among six peaks observed in the MS mode (S/N g 5) at 1 amol of catalase tryptic digest injected on the column, one provided good MS/ MS fragmentation and high SEQUEST score (see Figure 8). For the specific peptide LGPNYLQIPVNCPYR, reasonable fragmentation patterns, high SEQUEST scores, and high reliability of peptide identification29 were obtained at both levels (10 and 1 amol). Typically, a mass sensitivity of 5-10 amol was observed for tryptic peptides in the MS mode at S/N g 5; however, for some peptides, it was almost 1 order of magnitude higher, as expected from the dependence of the ESI-MS signal intensity on the peptide physical properties.31,32 A commercial 75 µm i.d. × 15 cm packed with 3-µm C18 beads showed mass sensitivity of ∼200-300 amol for the same bovine catalase tryptic peptides. The number of peptides identified in the MS and MS/MS modes at this level using the 75-µm-i.d. column was comparable to that at 5-10-amol level of injection on a 20-µm-i.d. monolithic column. To demonstrate the potential of the monoliths in the highsensitivity nano-LC-ESI-MS/MS system, a tryptic digest of protein from an extract of a breast ductal carcinoma tissue section was analyzed. We injected 200 nL (103 cell equivalent) of the digested protein extract onto the 20-µm-i.d. monolithic column, followed by a 1-h gradient. In data-dependent MS/MS scanning, a full MS (31) Cech, N. B.; Krone, J. R.; Enke, C. G. Anal. Chem. 2001, 73 (2), 208-13. (32) Cech, N. B.; Enke, C. G. Anal. Chem. 2000, 72 (13), 2717-23.
5314
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
scan between 400 and 2000 m/z was followed by three full MS/ MS scans for the three most intense ions from the MS scan. Figure 9 presents the separation of the relatively complex sample in which many peaks are observed with narrow elution peak widths and high S/N. Even though a low amount of sample material was injected on the column, MS/MS data-dependent acquisition, followed by database searching with TurboSEQUEST (Bioworks 3.1, ThermoFinnigan) allowed protein identification. Fifty fully tryptic peptides were matched with SEQUEST ∆Cn greater than 0.08 and Xcorr greater than 2.0, 1.5, and 3.0 for charged states of +1, +2, and +3, respectively, with a search against the whole protein human NCBInr database. Thirty-six proteins were identified by this approach from the human NCBInr database, including vimetin, R-tubulin 6, glyceraldehyde-3phosphate dehydrogenase, and member M of the H2A histone family. These results open the possibility of utilizing such narrowbore monolithic columns for analysis of tissue samples with a limited number of cells, e.g., laser capture microdissected cell samples.33 CONCLUSIONS We report an approach for high-sensitivity analysis of small quantities of protein digests using ultralow flow nano-LC-ESI ion trap mass spectrometry using 20-µm-i.d. columns filled with a porous monolithic stationary phase. The effectiveness of the approach was demonstrated in the gradient elution nano-LC-ESIMS with data-dependent MS/MS fragmentation of low-attomole amounts of tryptic digest of a protein mixture. An examination of a model system of breast cancer tissue, with the equivalent of 1000 cells injected, presented the potential of the approach for analysis of a limited amount of sample. Of course, at these levels, one must exercise care that losses are minimized in sample (33) Wu, S. L.; Hancock, W. S.; Goodrich, G. G.; Kunitake, S. T. Proteomics 2003, 3 (6), 1037-46. (34) Stadalius, M. A.; Gold, H. S.; Snyder, L. R. J. Chromatogr. 1984, 296, 31. (35) Snyder, L. R.; Stadalius, M. A.; Quarry, M. A. Anal. Chem. 1988, 55, 1412A30A.
Figure 9. Gradient nano-LC-ESI-MS of a protein extract from a breast cancer tissue section digested with trypsin. 200 nL injected on the column; sample amount corresponding to an extract from ∼1000 cells. Mobile phase: (solvent A) 2% acetonitrile, 0.1% formic acid in water; (solvent B) 10% water, 5% 2-propanol, 0.1% formic acid in acetonitrile. Gradient: 0 min 5% B, 60 min 40% B; 65 min - 90% B; gradient steepness parameter b ) 0.12. (A) 3-D overlay LC-MS chromatogram; (B) planar ion density map.
processing. Nano- and capillary-LC columns (100-300-µm i.d.) would still be effective to use in protein tryptic digest analysis, if the sample amount were sufficient. Further development of the nanocapillary columns using trapping columns for sample preconcentration, UV-induced polymerization of porous monoliths, use of nanocolumns less than 20-µm i.d., and coupling of the ultranano-LC columns to more sensitive MS instrumentation (e.g., Fourier transform, linear ion trap, etc.) is planned.
ACKNOWLEDGMENT The authors thank NIH for supporting of this work under GM 15847. We are grateful to Dr. Bill Fowle for his assistance with scanning electron microscopy and to Drs. Roger Kautz, Tomas Rejtar, and Prof. William S. Hancock for fruitful discussions in the preparation of the manuscript. Special thanks to Drs. Darryl E. Palmer-Toy and Dennis Sgroi (Massachusetts General Hospital, Boston, MA), who kindly provided breast ductal carcinoma tissue Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5315
sections. The authors acknowledge Robert Kane (ThermoFinnigan) for software support. This is contribution no. 825 from the Barnett Institute.
F
flow rate
S
∼0.48Mw0.44 ≈ 10, ref 34
k′
capacity factor, k′ ) (tr - t0)/t0
GLOSSARY
tr
retention time
t0
column delay time, t0 ) V0/F
b
gradient steepness parameter, b ) ∆φV0S/tgF, refs 34 and 35
tg
gradient time
∆φ
difference between the initial and final organic composition in eluent
Received for review April 22, 2003. Accepted July 18, 2003.
V0
column dead volume
AC030163G
5316
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
