Two-Dimensional Microcolumn Separation Platform for Proteomics

from the CIEF column to the CEC column will stay tightly adsorbed to the inlet top of the CEC column until it will be eluted and separated into its pr...
5 downloads 8 Views 306KB Size
Download PDF
Two-Dimensional Microcolumn Separation Platform for Proteomics Consisting of On-Line Coupled Capillary Isoelectric Focusing and Capillary Electrochromatography. 1. Evaluation of the Capillary-Based Two-Dimensional Platform with Proteins, Peptides, and Human Serum Minquan Zhang and Ziad El Rassi* Department of Chemistry, Oklahoma State University, Stillwater, Oklahoma 74078-3071 Received April 22, 2006

In this report, an on-line coupling of capillary isoelectric focusing (CIEF) to capillary electrochromatography (CEC) is developed via a nanoinjector valve for performing two-dimensional (2D) proteomics separation. CIEF constitutes the first separation dimension, while CEC operates as the second separation dimension. Besides the orthogonal migration mechanisms of the two capillary-based separation modes, which lead to a 2D system whose overall peak capacity is the product of the peak capacity of the individual modes, the solvent of the CIEF mode is a weak eluent for the reversed-phase CEC (RP-CEC) mode, thus, allowing the transferring of focused fractions from CIEF to CEC without inducing band broadening, and instead zone sharpening would result. In fact, the transferred focused protein fraction from the CIEF column to the CEC column will stay tightly adsorbed to the inlet top of the CEC column until it will be eluted and separated into its protein components with a hydro-organic mobile phase. The theoretical peak capacity of the CIEF-CEC 2D platform is estimated at nCIEF () 560) × nCEC () 97) ) 54 320. This peak capacity is more than needed for proteomics profiling. Also, only a fraction of this peak capacity is needed when looking at heart cuts for performing subproteomics. The 2D platform described here offers the convenience to generate the needed peak capacity to solve a given proteomic separation problem. This is facilitated by the RP-CEC dimension, which ensures rapid isocratic separation of proteins and peptides and rapid solvent change and column equilibration and avoids lengthy gradient elution. The RP-CEC column is based on neutral C17 monolith, which offers high separation efficiency and relatively high column permeability. To the best of our knowledge, the proposed 2D platform combining CIEF and CEC is reported for the first time for proteins and proteomics. Keywords: Capillary Isoelectric Focusing • Capillary Electrochromatography • Capillary-Based Two-Dimensional Separations • Proteomics

1. Introduction The inherent complex nature of proteomics samples has been a serious challenge to all separation techniques. Although high-resolution separation techniques such as gel electrophoresis, high-performance liquid chromatography (HPLC), capillary electrophoresis (CE), and capillary electrochromatography (CEC) are very suitable for separating moderately complex mixtures, they are unable to resolve proteomics mixtures when used in a single separation mode, that is, one dimension. However, these separation techniques offer various separation modes that are based on orthogonal mechanisms, which can be exploited in two-dimensional (2D) formats to maximize resolution and consequently solve the complex separation problem. O’Farrell first demonstrated the 2D separation strategy for proteins in 1975 in his landmark report on * To whom correspondence should be addressed. Tel, 405-744-5931; fax, 405-744-1235; e-mail, [email protected]. 10.1021/pr060185u CCC: $33.50

 2006 American Chemical Society

2D map analysis of proteins from lysed Escherichia coli cells.1 In this work, 1100 different proteins were resolved and detected on a single 2D map, and it was suggested that the maximum resolution capability of the 2D system might have been as high as 5000 different proteins. O’Farrell combined isoelectric focusing (IEF) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) at a right angle to yield a 2D map in which the protein spots were separated according to their charges (or isoelectric point pI) in one dimension and size in the second dimension. This 2D gel electrophoresis separation strategy was done in much the same way as in other planar chromatography or electrophoresis techniques (i.e., at a right angle) such as 2D paper chromatography,2 thin-layer chromatography,3,4 and paper electrophoresis,5,6 which were reported many years earlier. The appellations multidimensional and 2D are currently used interchangeably, although most of the complex separations are carried out in two dimensions. Although our aim here is not to provide an overview of 2D Journal of Proteome Research 2006, 5, 2001-2008

2001

Published on Web 07/04/2006

research articles separations, it is essential to emphasize their recent evolution, especially in the field of proteins, peptides, and proteomics. Although so far the 2D gel electrophoresis maps represent the most popular technique in proteome analysis (for recent reviews, see refs 7, 8), we are currently witnessing an increasing interest in nongel-based 2D maps for proteomics due to the facts that 2D gel electrophoresis has some drawbacks including the limited dynamic range and molecular mass range and the lack of on-line integration with mass spectrometry. In the category of nongel-based 2D maps, 2D HPLC9-12 and microand nano-2D LC9-11 have been described where size-exclusion chromatography (SEC) or anion/cation exchange chromatography served as the first dimension and reversed-phase chromatography (RPC) as the second dimension. Other nongel 2D platforms include capillary isoelectric focusing (CIEF) coupled to transient capillary isotachophoresis/capillary zone electrophoresis (CZE),13 integration of CIEF with capillary RPC,14,15 coupled µLC (RPC) and CZE,16,17 and on-line coupling of SEC or MEKC with imaged CIEF;18,19 for a recent review on capillary separations in proteomics, see ref 20. Furthermore, attempts of 2D separation on microchips for peptides and proteins have been recently reported.20-23 While the separation abilities of microfluidics 2D separation systems (i.e., microchips) are still awaiting further development and are currently limited to a narrow range of peptides and proteins, the separation abilities of 2D HPLC and capillary-based 2D platforms are far superior. However, the 2D LC systems consisted either of (i) biphasic columns (i.e., packed with strong cation exchange and reversedphase packing material) that were not true two-dimensional systems24 or (ii) true 2D systems based on multiple valves, columns, detectors, and costly sophisticated pumps,10,11 and required (both i and ii) lengthy gradient elution which increased analysis time. In some similar fashions, the CIEF-based 2D systems required multiple trap columns followed by transferring the trapped fractions to the second dimension and analysis by lengthy gradient elution using nano-RPC.14,15,25 Furthermore, when CIEF constituted the second dimension after either a SEC or MEKC first dimension, the 2D system necessitated a dialysis interface to eliminate salt and other unwanted first-dimension effluent components and subsequent adding of carrier ampholyte, which would lead to band broadening. As noted above, although significant progress has been made in the field of nongel-based 2D separations, the exploitation of their full potentials is yet to come. This necessitates further developments, whereby two important undertakings, among other things, must be achieved: (i) avoid the intermediate trapping of fractions while transferring from first to second dimension and (ii) eliminate lengthy gradient elution. It is a well-known fact that CIEF is an important separation process that simultaneously yields high resolution and relatively high concentration factor for low-abundance proteins in the first dimension (for recent reviews, see refs 26, 27). For high throughput, the second dimension, which is the rate-limiting dimension in any 2D platform, must be judiciously selected as to allow fast separation, fast solvent exchange, and isocratic elution. Meeting these requirements constituted the central elements of the proposed 2D platform in this investigation. We propose reversed-phase capillary electrochromatography (RPCEC) with neutral monolithic capillary columns as the second dimension. Monolithic capillary columns are characterized by their relatively high permeability and fast equilibration time during solvent change. In addition, isocratic elution allows the separation of proteins and peptides in relatively short time,28-30 2002

Journal of Proteome Research • Vol. 5, No. 8, 2006

Zhang and El Rassi

thus, eliminating the use of lengthy gradient elution. Very recently, a neutral monolithic C17 column was introduced by our laboratory for rapid protein separation providing relatively strong electroosmotic flow (EOF) and high separation efficiency in the absence of electrostatic interactions due to the fact that the neutral monolith is void of fixed charges on its surface.29 In summary, in this initial investigation, we are reporting a “proof of concept” that demonstrates the viability of combining CIEF and CEC in a 2D platform. Further investigation is under way to expand the capability of the 2D platform and produce a prototype that will include an array of 12 CEC columns for higher throughput format. This future aim should be readily achieved, since the concept of monolith allows the preparation of several identical columns very conveniently due to the relatively high reproducibility of monolithic columns when the conditions are kept under good control.29-32 Other advantages of CEC include (i) its hybrid nature that combine differential electromigration to chromatographic partitioning thus offering a unique selectivity and high resolution and (ii) its readily compatibility for coupling with mass spectrometry.33,34

2. Materials and Methods 2.1. Instruments. The CE/CEC system was assembled inhouse from commercially available components. It consisted of two 30 kV dc power supplies of positive and negative polarity, models MJ30P400 and MJ30N400, respectively, from Glassman High Voltage (Whitehouse Station, NJ) and a UV-vis variable wavelength detector, model 200, from Linear Instruments (Reno, NV) equipped with a cell for on-column detection. The electrochromatograms were recorded with a Shimadzu data processor model C-R5A (Kyoto, Japan). A microsyringe pump 74900 series was from Cole-Parmer Instrument Co. (Vernon Hills, IL). A 6-port nanoinjector valve model M485 offering 100% ceramic internal pathway (i.e., metal-free flow path) was from Upchurch Scientific (Oak Harbor, WA). Fused-silica capillaries of 50 µm i.d. and 100 µm i.d., and 360 µm o.d. were from Polymicro Technology (Phoenix, AZ). 2.2. Chemicals and Reagents. Reagent grade glacial acetic acid, hydrochloric acid, sodium hydroxide, ethylene glycol (EG), cyclohexanol, and acetone were from Fisher Scientific (Pittsburgh, PA), and ammonium hydroxide was from Pharmco (Brookfield, CT). Sodium phosphate monobasic was purchased from Mallinckrodt (Paris, KY). HPLC grade acetonitrile was obtained from Baxter (McGraw Park, IL). Hydroxypropyl cellulose (HPC), hydroxypropyl methylcellulose (HPMC), N,N,N′,N′tetramethylethylenediamine (TEMED), 3-(trimethoxysilyl) propyl methacrylate (bind silane), pentaerythritol diacrylate monostearate (PEDAS), and azobisisobutyronitrile (AIBN) were purchased from Aldrich Chemical Co. (Milwaukee, WI). γ-Glycidoxypropyltrimethoxysilane was from Huls America, Inc. (Bristol, PA). Pharmalyte with a pH range of 3-10 was from Amersham Pharmacia (Uppsala, Sweden). All protein standards including horse heart myoglobin, baboon hemoglobin, human erythrocyte carbonic anhydrase I, bovine erythrocyte carbonic anhydrase II, bovine milk β-lactoglobulin A and B, bovine serum albumin, bovine milk R-lactalbumin, soybean trypsin inhibitor, bovine milk β-casein, Aspergillus nigers amyloglucosidase, human transferrin, bovine pancreas insulin, bovine liver β-galactosidase, A. nigers glucose oxidase, human serum, and trichloroacetic acid (TCA) were all purchased from Sigma Chemical Co. (St. Louis, MO). Synthetic peptide pI markers listed in Table 1 and covering the pI range from 3.38 to 10.17

research articles

2D Microcolumn Separation Platform for Proteomics Table 1. Number, Amino Acid Sequence, and pI Values of Peptide pI Makers no.

peptides

pI

28 29 30 31 33 34 35 36 37 38 39 40 41 42 43

H-Trp-Tyr-Lys-Arg-OH H-Trp-Tyr-Lys-Lys-OH H-Trp-Tyr-Tyr -Lys-Lys-OH H-Trp-Tyr-Tyr-Tyr-Lys-Lys-OH H-Trp-Glu-Tyr-Tyr-Lys-Lys-OH H-Trp-Glu-His-His-His-Arg-OH H-Trp-Glu-His-Arg-OH H-Trp-Glu-His-His-OH H-Trp-Glu-Arg-OH H-Trp-Glu-His-OH H-Trp-Asp-Asp-His-His-OH H-Trp-Glu-Glu-His-OH H-Trp-Asp-Asp-Arg-OH H-Trp-Glu-Glu-OH H-Trp-Asp-Asp-Asp-OH

10.17 9.99 9.68 9.50 8.40 7.27 7.00 6.66 5.91 5.52 5.31 4.28 4.05 3.78 3.38

were donated by Dr. Kiyohito Shimura, Teikyo University, Sagamiko, Kanagawa, Japan.35 2.3. CIEF. 2.3.1. Inner Wall Coating of CIEF Capillaries. The capillary inner wall was coated with hydroxypropylcellulose (HPC) of average Mw 100 000 to suppress the EOF and solute-wall interaction. First, the capillary was washed with 1.0 M sodium hydroxide for 30 min, flushed with 0.10 M hydrochloric acid for 30 min, rinsed with deionized water for 30 min, and dried with helium. Then the capillary inner wall was reacted, at room temperature, with a solution of 50% (v/v) γ-glycidoxypropyltrimethoxysilane in acetone for 10 h. Thereafter, the capillary was flushed with DMF and reacted with 1% w/v HPC in DMF in the presence of boron trifluoride etherate for 30 min. After rinsing with DMF and water and drying with helium, the capillary was filled and flushed with 5% w/v HPC in water for 10 min. Then, the solution was blown out using 60 psi helium, and the capillary was heated in a GC oven from 30 to 140 °C at a rate of 5 °C/min and then kept for 28 min at 140 °C while nitrogen was blown through the capillary. The procedure for the second HPC coating resembles somewhat that previously described by Shen and Smith.36 2.3.2. CIEF Procedures. The CIEF separations were performed using the coated capillary described above with an i.d. of 50 µm and a total length of 44 cm (an effective length of 38 cm). The entire capillary was filled with samples mixed with the Pharmalyte pH 3-10. A voltage of 20 kV was used for the focusing step, while hydrodynamic mobilization was obtained by elevating the anode reservoir to a desired height and was monitored by UV absorbance at 280 nm. In one case, 20 mM phosphoric acid containing 0.1% HPMC, pH 2.03, and 20 mM sodium hydroxide, pH 11.87, were used for anolyte and catholyte, respectively, while in another case, 1% v/v acetic acid, pH 2.5, and 1% v/v ammonium hydroxide pH 10.7 were used as the anolyte and catholyte, respectively. Between runs, the capillary was rinsed with deionized water for ∼2 min. 2.4. CEC. 2.4.1. Monolithic Column Preparations for CEC. Capillaries with an i.d. of 100 µm and a desired length were used to prepare the CEC columns. The inner wall of the fused silica capillary was treated with 1.0 M sodium hydroxide for 30 min, flushed with 0.10 M hydrochloric acid for 30 min, rinsed with deionized water for 30 min, and dried with helium. Then, a 50% (v/v) bind silane solution in acetone was introduced and kept for 1 h to vinylize the inner wall of the capillary.37-39 Finally, after rinsing with acetone, the capillary was swept and dried by purging with nitrogen for 30 min before polymerization.

Neutral stearyl-acrylate monolithic columns (neutral C17Monolith) for CEC were made by in situ polymerization according to a previously described procedure from our laboratory.29 Polymerization solution weighing 1.48 g was prepared from monomer PEDAS and porogenic solvents in a ratio of 30: 70 (w/w) monomer/solvents. The monomer was dissolved in a ternary porogenic solvent consisting of 79.2 wt % cyclohexanol, 17.2 wt % EG, and 3.6 wt % water. The mixture was vortexed to dissolve the PEDAS, and then AIBN (1 wt % with respective to monomer) was added to the solution as an initiator. The mixture was sonicated to obtain a clear solution and then purged with helium for 5 min. A 45 cm pretreated capillary having vinylized walls (see above) was filled with the mixture solution up to 37 cm by immersing the inlet of the capillary in the solution vial and applying vacuum to the outlet. The capillary ends were sealed in an oxygen/propane flame. Then, the capillary was put in a GC oven and heated from 30 to 75 °C at a rate of 0.5 °C/min and kept for 16.5 h at 75 °C. The resulting monolithic column was washed with an 80:20 (v/v) ACN/H2O mixture using an HPLC pump. An optical window for UV detection was made at 1-2 mm after the end of the monolithic material by burning off the external polyamide coating of the capillary with thermal wire stripper. Finally, the capillary column was cut to a desired length. 2.4.2. CEC Separations. The CEC monolithic column was flushed with the running mobile phase for ∼30 min and was preconditioned by gradually applying voltage up to 10 kV over the column until a stable current was observed before the experiment started. CEC was performed with the neutral stearyl-acrylate monolithic column, 35 cm (to detector), 41 cm (total length) × 100 µm i.d., using a hydro-organic mobile phase: 40% (v/v) ACN in 20 mM sodium dihydrogen phosphate, pH 7.0. 2.5. Construction and Operation of CIEF-CEC 2D Separation Platform. As shown in Figure 1, a 6-port nanoinjector valve with port-to-port volume of 25 nL was used for on-line combination of CIEF with CEC in a 2D platform. In experiments with standard proteins, a 34-cm long CIEF capillary column (50 µm i.d., denoted segment b in Figure 1) was housed in the nanoinjector valve by connecting the two ends of the capillary to ports 1 and 4 of the valve. The length of segment b was changed to 41 cm in the case of human serum sample. In all cases, one 4-cm long capillary (50 µm i.d., denoted segment a in Figure 1) was connected to the sample port 6 of the injector and immersed in the inlet reservoir containing an anolyte of 20 mM phosphoric acid (or 1% acetic acid), while another 6 cm long capillary (50 µm i.d.) was connected to port 5 of the injector and immersed in the outlet reservoir containing a catholyte of 20 mM sodium hydroxide (or 1% ammonium hydroxide). Thus, the total CIEF column length was either 44 cm (with standard proteins) or 51 cm (with human serum), while the effective CIEF column length was either 38 cm (with standard proteins) or 45 cm (with human serum) consisting of a 4 cm segment (segment a) and either a 34 cm segment or 41 cm segment (segment b), respectively. A platinum wire was inserted into the inlet reservoir (i.e., anolyte reservoir) and connected to a high-voltage power supply, while another platinum wire was inserted into the outlet reservoir (i.e., catholyte reservoir) and connected to the common ground. After the CIEF capillary was filled with the sample containing Pharmalyte (pH 3-10) with the nanoinjector in the load position (see Figure 1a), and the inlet and outlet capillary ends were immersed in the anolyte and catholyte reservoirs, respecJournal of Proteome Research • Vol. 5, No. 8, 2006 2003

research articles

Zhang and El Rassi

min. The supernatant was collected, and 1 mL of ice-cold acetone was added to it. The mixture was incubated on ice for 15 min and centrifuged as above. The supernatant was collected and evaporated at -85 °C for 30 min to get rid of acetone. The resulting sample was kept at -20 °C for analysis.

3. Results and Discussion

Figure 1. Schematic illustration of the 2D platform, which couples CIEF as the first dimension to CEC as the second dimension using a nanoinjector valve with 25 nL port-to-port volumes. The effective CIEF capillary length (e.g., 38 cm) is the sum of segment a of 4 cm and segment b of 34 cm.

tively, focusing was performed at a constant voltage of 20 kV over the entire CIEF capillary. Once the current was reduced to ∼10% of the original value, the focusing was considered to be complete, usually within 20 min. The second dimension was accomplished following the sequence: (1) After CIEF focusing, a fraction of a predetermined size of the focused CIEF column was transferred into the CEC column (see Figure 1b), which was already equilibrated with a plain buffer by using a microsyringe pump filled with this plain buffer (12 mM sodium dihydrogen phosphate pH 7.0) used in the CEC mobile phase, while the nanoinjector is in the injection position. (2) Thereafter, the injector was returned to the load position, and the microsyringe pump was activated to displace out the ampholyte from the transferred fraction and leave the focused proteins tightly bound at the CEC column inlet. (3) The CEC column was eluted separately (see Figure 1c) by a hydro-organic mobile phase consisting of 40% (v/v) ACN, 20 mM sodium dihydrogen phosphate, pH 7.0, and a running voltage of 10 kV. (4) While in load position, the CEC run was stopped by turning off the high voltage power supply, and the microsyringe pump containing plain buffer was used to return the CEC column to the initial conditions and be ready for transferring a second fraction from the CIEF following the sequential events (1), (2), (3), and (4). This cycle of events was repeated until all the CIEF fractions were processed in the CEC dimension. 2.6. Albumin Removal from Human Serum. Albumin was removed from human serum by a modified method.40 Briefly, 20 µL of human serum was precipitated by a rapid addition of 80 µL of ice-cold 10% (w/v) TCA solution in acetone. After immediately being gently vortexed, the mixture was incubated at -20 °C for 90 min and centrifuged at 15 000g at 4 °C, for 20 2004

Journal of Proteome Research • Vol. 5, No. 8, 2006

CIEF as the first separation dimension provides both concentration and separation of multicomponent proteins based on differences in their pI values.27 However, for complex protein mixtures, the CIEF bands may contain several unresolved proteins of the same pI values, which would require another dimension of different migration mechanism for their further separation. CEC as the second separation dimension provides further separation based on differences in size, hydrophobicity, and electrophoretic mobility of the proteins. Thus, the separation mechanisms of CIEF and CEC are completely orthogonal, and the overall peak capacity of the 2D separation platform under investigation should be the product of the peak capacities of the individual 1D separations.41 In the following sections, we will demonstrate the separation capability of each dimension run separately first and then in the coupled 2D format with standard proteins and peptides and thereafter with human serum sample. 3.1. Evaluation of the Components of the 2D Platform. 3.1.1. Evaluation of Coated Capillaries with Standard Peptides and Proteins. In this investigation, a hydrophilic coating with a top layer of hydroxypropyl cellulose (HPC) on the inner walls of fused silica capillaries was performed. It is a multilayered coating consisting of a primary layer (i.e., bottom layer) of γ-glycidoxypropyltrimethoxysilane, which was further reacted with HPC via the epoxide ring opening with BF3 (as the catalyst), thus, establishing ether bonds between the primary layer and the top HPC layer. This covalently attached HPC layer was further thickened by reacting it with another HPC layer(s) by thermal treatment, see Materials and Methods. In addition to the covalent attachment of the bilayered HPC coating, one can also envision the existence of some hydrogen bonding attachment between both HPC layers as well as the silica surface. HPC-coated capillaries by direct attachment of HPC to the inner walls of fused silica capillaries by thermal treatment have been found useful in CIEF.36,42 Also, HPC as adsorptive coating sieving matrix was found efficient in DNA separation,43 and a slightly different multilayered coating of HPC reported earlier by our laboratory44 was described for CZE of proteins with negligible EOF. Before performing any CIEF run, the HPC-coated capillary columns were first evaluated in the CZE mode to determine the magnitude of the EOF exhibited by these capillaries. The EOF was measured under conditions of relatively low ionic strength running electrolyte, since in CIEF the ionic strength is also low27 consisting of 20 mM phosphoric acid as the anolyte, 20 mM sodium hydroxide as the catholyte, and 1% (v/v) Pharmalyte as the focusing medium. In this regard, the capillaries were run in CZE with 5 mM sodium dihydrogen phosphate, pH 7.0, and acetone was injected as the EOF marker. Under these conditions, the velocity of the EOF marker was ∼0.1 mm/s at 300 V/cm. Also, the HPC-coated capillaries were tested in CZE with basic standard proteins (e.g., lysozyme, cytochrome c, ribonuclease A, and R-chymotrypsinogen A) using 0.10 M sodium dihydrogen phosphate, pH 6.0, as the running electrolyte. The observed separation efficiency at applied field strength of 300 V/cm exceeded 1.1 million

2D Microcolumn Separation Platform for Proteomics

Figure 2. CIEF separation of a set of 15 peptide pI markers using HPC-coated fused-silica capillary, 38 cm (to detector), 44 cm (total length) × 50 µm i.d.; anolyte, 20 mM phosphoric acid, pH 2.03, containing 0.1% HPMC; catholyte, 20 mM sodium hydroxide, pH 11.87; injection, by filling the whole capillary column with peptides 5 µM each dissolved in 1% (v/v) Pharmalyte containing 0.4% HPMC, and 0.3% TEMED; focusing for 15 min at 20 kV; gravity mobilization by elevating the anode reservoir by 12 cm. Wavelength, 280 nm. The numbers above the peaks are the peptide numbers listed in Table 1.

theoretical plates/m. The low magnitude of EOF and the high separation efficiency obtained are indications of the effectiveness with which the HPC coating reduces the EOF and the protein-wall interactions, respectively. To evaluate the effectiveness of the HPC-coated capillaries in CIEF, a mixture of 15 synthetic peptides of known pI values (i.e., pI markers) were separated by CIEF. A typical electropherogram is shown in Figure 2. The average separation efficiency for the 15 peptides was nearly 3 million theoretical plates/m. This high separation efficiency is attributed to the focusing effect, which arises from the fact that axial diffusion of an ampholyte (e.g., peptide) out of the zone results in acquisition of charge, and electromigration returns the ampholyte to the zone. The purpose of adding HPMC into the anolyte solution is to control the mobilization speed of the anodic side of the pH gradient,35 while the use of the TEMED is for extending the separation range from pI 10 to pI 12 (the pH range of Pharmalyte is 3-10).13 Thus, the two most basic peptides, no. 28 (pI 10.17) and no. 29 (pI 9.99) were wellseparated. Under these CIEF conditions, a peak capacity of ∼560 (Rs ) 1) is realized within the time frame of 40 min of separation time extending between 5 min (appearance of first peak, pI near 12) and 45 min (end of mobilization). The peak capacity, n, is calculated by the following equation: n)1+

x16N × ln t

tω R

where N is the plate number, tω is the mobilization time of the last eluting peak, and tR is the mobilization time of the first eluting peak. Assuming equal peak width at base for the focused zones, the calculated peak capacity n ) 560 yields a width of ∼0.6 mm for each focused zone. In this calculation, the capillary is considered as being a string of pearls where each pearl is touching its neighbor with no unoccupied space between the pearls. The retraction of the protein band from an initial length of 440 mm (total column length) to 0.6 mm

research articles

Figure 3. CIEF of 15 standard proteins, using the same conditions as with peptides in Figure 2. Injection by filling the whole capillary column with proteins at 0.020 mg/mL each dissolved in 1% (v/v) Pharmalyte containing 0.4% HPMC, and 0.3% TEMED; gravity mobilization by elevating the anode reservoir by 12 cm. Peak numbering as in CEC. Peak identifications form left to right: 5, myoglobin; 10, hemoglobin; 2, carbonic anhydrase I; 4, carbonic anhydrase II; 12, transferrin; 3, β-lactoglobulin B; 1, insulin; 8, β-lactoglobulin A; 11, albumin; 7, R-lactalbumin; 9, trypsin inhibitor; 14, β-galactosidase; 15, glucose oxidase; 6, β-casein; 13, amyloglucosidase.

represents a concentration factor of ∼734. The resolution of such a system, ∆pI, is about 0.02. This is a very promising figure of merit for achieving a high resolving power isoelectric focusing in a capillary format. A plot of pI of the peptides versus the mobilization time is linear and is given by the equation pI ) -0.267t + 12.76 (R2 ) 0.994). This is a further indication that the HPC-coated capillary with negligible EOF provides a welldeveloped, continuous pH gradient. The use of pI markers such as the synthetic peptides used in this investigation is the only convenient approach to assess the pH profile in a capillary.35,45 In agreement with the findings by Shimura et al.35,45 and others (for a recent review see ref 27), unlike most proteins, which are unstable and may undergo changes in their respective pI, small peptides constitute stable pI markers for the accurate evaluation of a given CIEF system. The instability of proteins is mostly attributed to the hydrolysis of side chain amides of asparagine and glutamine residues27,46 and peptide bonds.35 The above CIEF conditions were applied to the focusing of 15 standard proteins, and the results are shown by a typical electropherogram in Figure 3. Although these proteins have been used as pI markers, one can observe that some of them yielded small additional peaks accompanying the main peaks. Recently, Shimura and co-workers reported this behavior35,45 and attributed it to the proteins’ instability and their tendency to become heterogeneous with respect to pI. To circumvent this ambiguity and assign the correct pI value for each protein, the strategy of peptide pI markers introduced by Shimura et al.35,45 was also applied in this study. The calibration of the CIEF system under investigation with peptide pI markers (see preceding paragraph) permitted the accurate determination of the pI of the 15 standard proteins, which were focused using the same CIEF conditions as with peptides. The differences between the pI values of the standard proteins reported in the literature and those values estimated using the calibration curve obtained for the peptide pI markers (see above) were significant Journal of Proteome Research • Vol. 5, No. 8, 2006 2005

research articles

Figure 4. Capillary electrochromatogram of 15 standard proteins on the neutral stearyl-acrylate monolithic column, (i.e., neutral C17 Monolith). Capillary, fused-silica capillary, 37 cm (to detector), 43 cm (total length) × 100 µm i.d.; hydro-organic mobile phase: 40% (v/v) ACN, 20 mM sodium dihydrogen phosphate, pH 7.0; electrokinetic injection for 6 s at 5 kV; running voltage, 10 kV. Peak identifications: 1, insulin; 2, carbonic anhydrase I; 3, β-lactoglobulin B; 4, carbonic anhydrase II; 5, myoglobin; 6, β-casein; 7, R-lactalbumin; 8, β-lactoglobulin A; 9, trypsin inhibitor; 10, hemoglobin; 11, albumin; 12, transferrin; 13, amyloglucosidase; 14, β-galactosidase; 15, glucose oxidase.

for acidic proteins in the pI range 5.0-3.5. For these acidic proteins (e.g., R-lactalbumin, trypsin inhibitor, β-galactosidase, glucose oxidase, β-casein, and amylglucosidase), the differences varied in the range 0.15-0.20 pI unit. For the rest of the proteins used in this study (pI > 5.0), the differences were less significant in the range of 0.01-0.05 pI unit. Returning to Figure 3, sharp peaks were obtained for the focused proteins with very high plate counts averaging ∼2 800 000 plates/m. Since the separation efficiency obtained with proteins is about the same order of magnitude as that of the peptides (2.8 vs 3.0 million plates), then the peak capacity of the two systems is about the same. This again demonstrates the ready development of continuous pH gradient signaling that the EOF is negligible. In fact, the EOF when measured under CIEF conditions is very reduced standing at 0.063 mm/s when HPMC is added to the anolyte, while the EOF in the absence of HPMC is slightly higher approaching 0.088 mm/s. 3.1.2. CEC Separation of Proteins. Recently, a neutral, nonpolar monolith (neutral C17-Monolith) having a relatively strong EOF, yet free of electrostatic interactions with charged solutes, was developed in our laboratory29 for the reversedphase CEC (PR-CEC) of neutral and charged species including peptides and proteins. Figure 4 shows the simultaneous separation of 15 standard proteins on the neutral monolithic column. Although the neutral C17-Monolith is void of fixed charges, and has no zeta potential with respect to water, a relatively strong EOF could be established across the monolith.29. The EOF can be explained by the adsorption of phosphate ions to the monolith thus becoming the zeta potential provider.32 This is an important development in the design of monolithic columns void of fixed charges but still yield a relatively strong EOF. The average separation efficiency is relatively high at ∼600 000 plates/m, and brought about the 2006

Journal of Proteome Research • Vol. 5, No. 8, 2006

Zhang and El Rassi

Figure 5. Analysis of standard proteins by CIEF-CEC 2D separation platform with pI corresponding to (A) -7.3; (B) 7.2-5.6, peak order 2, 4, 5, 10, and 12 are corresponding to those in Figure 4 (CEC); (C) 5.4-4.4, peak order 1, 3, 7, 8, 9, 11, and 14 are corresponding to those in Figure 4; and (D) 4.2-3.6, peak order 6, 13, and 15 are corresponding to those in Figure 4. CIEF conditions are the same as in Figure 3. CEC conditions are the same as in Figure 4. Others conditions are as in Figure 2. After CIEF focusing, (1) CEC column was equilibrated with plain buffer (12 mM sodium dihydrogen phosphate pH 7.0); (2) sample was introduced into CEC column from CIEF column by pushing plain buffer with syringe pump; (3) the CEC column was rinsed with the buffer; and (4) the CEC was run with hydro-organic mobile phase: 40% (v/v) ACN, 20 mM sodium dihydrogen phosphate, pH 7.0; running voltage, 10 kV. (see Section 2.5. for experimental details).

baseline separation of β-lactoglobulin A (18 363 Da) and β-lactoglobulin B (18 276 Da), which differ only by three amino acid substitutions. For the time frame of Figure 4, i.e., between 8 and 18 min, the protein peak capacity is ∼97 (Rs ) 1), thus, clearly demonstrating the high resolving power of these neutral monoliths. This peak capacity is more than sufficient when the PR-CEC process functions as the second separation dimension following CIEF as the first dimension, which is characterized by an 8 times higher peak capacity (n ∼ 560). The overall peak capacity of the 2D platform will be 560 × 97 ) 54 320, which will be sufficient to tackle the complex nature of proteomics samples. 3.2. Evaluation of the 2D Platform with Proteins. 3.2.1. Standard Proteins. The 2D platform shown in Figure 1 was evaluated with standard proteins in the consecutive operation of CIEF and CEC separation dimensions. In this operation, after CIEF was completed, the focused zones were transferred to the CEC column previously equilibrated with plain buffer to afford the strongest interaction of the proteins with the C17 monolithic capillary column. The introduction of the focused zones into the CEC column was carried out by pushing with a plain buffer from a syringe pump connected to the nanoinjector interface while in the injection position (see Figure 1b), to allow the transfer of a given focused segment from the CIEF column making up the injection loop. Thereafter, and while in load position (Figure 1c), the CEC column was washed with the syringe pump with plain buffer to remove the

2D Microcolumn Separation Platform for Proteomics

Figure 6. CIEF electropherogram of albumin-depleted human serum using HPC-coated fused-silica capillary, 45 cm (to detector), 51 cm (total length) × 50 µm i.d. Injection by filling the whole capillary column with sample. Sample was dissolved in 2% (v/v) Pharmalyte in water; anolyte, 1% v/v acetic acid in water, pH 2.5; catholyte, 1% v/v ammonium hydroxide in water, pH 10.7; focusing for 15 min at 20 kV; gravity mobilization by elevating the anode reservoir by 5 cm. Wavelength, 280 nm.

ampholyte and other components of the focusing medium from the focused zones loaded onto the inlet top of the CEC column. After transferring the first ∼18 cm of the CIEF column (fraction A), which did not contain focused protein zones, to the CEC column at a flow rate of 0.14 mL/h for ∼9.3 s, three focused zones (denoted B, C, and D) were introduced into the CEC column at the same flow rate (i.e., 0.14 mL/h) for 3.3 s each. Each one of the three fractions corresponds to a 0.13 µL volume or about 17% of the effective length of the CIEF column (i.e., 6.65 cm) moved to the CEC column. Fraction D has only 2.65 cm of the pH gradient and 4 cm of plain buffer, since the last 4 cm (segment a, see Figure 1) are out of transfer to the CEC column. In the last 4 cm (segment a), there are no proteins, since it corresponds to pI values