Phosphopeptide Isomer Separation Using Capillary Zone

Merck Frosst Canada & Co., 16711 Trans Canada Highway, Kirkland, Quebec, H9H 3L1, Canada. Methods for the rapid separation of phosphopeptide isomers ...
3 downloads 0 Views 97KB Size
Anal. Chem. 1999, 71, 3469-3476

Phosphopeptide Isomer Separation Using Capillary Zone Electrophoresis for the Study of Protein Kinases and Phosphatases Tanya N. Gamble, Chidambaram Ramachandran, and Kevin P. Bateman*

Merck Frosst Canada & Co., 16711 Trans Canada Highway, Kirkland, Quebec, H9H 3L1, Canada

Methods for the rapid separation of phosphopeptide isomers (peptides with the same sequence but with phosphates on different residues) were developed using capillary zone electrophoresis with ultraviolet (CZE-UV) detection. Uncoated, cationic and neutral capillaries were used with both acidic and basic peptides. These methods enabled the assay of several protein kinases (mitogen activated protein kinase, protein kinase A, GST-tyrosine kinase) and phosphatases (acid, alkaline, and protein tyrosine phosphatase) and the determination of the sites of phosphorylation and dephosphorylation. Incubations of nonphosphorylated or phosphorylated peptide with kinases or phosphatases took place directly in the instrument’s autosampler and were monitored over several hours using CZE-UV. The state of cellular protein phosphorylation is a dynamic process that depends on the activities of both protein kinases and protein phosphatases.1-3 One aspect of understanding signal transduction and its modulation by phosphorylation (kinases) and dephosphorylation (phosphatases) is the mechanism by which these enzymes recognize their substrates and the degree to which they act on those substrates.4,5 Knowledge of the sites of phosphorylation including the number of sites and what specific amino acids (Thr/Ser or Tyr) are modified is necessary to fully understand how phosphorylation affects cellular signal transduction. Further, the ability to identify and quantify individual sites of phosphorylation may lead to an understanding of the role of the primary sequence of the peptide in relation to the specificity of protein kinases and phosphatases. Several analytical techniques have been developed to assess the level of protein phosphorylation and to identify specific sites of phosphorylation in the primary structure of phosphoproteins. These include metabolically [32P]phosphate labeling of phosphoproteins followed by enzymatic or chemical cleavage and two* Corresponding author: (phone) (514) 428 8689; (fax) (514) 428 8615; (email) Kevin•[email protected]. (1) Barford, D.; Das, A. K.; Egloff, M.-P. Annu. Rev. Biophys. Biomol. Struct. 1998, 27, 133-64. (2) Posada, J.; Cooper, J. A. Mol. Biol. Cell 1992, 3, 583-92. (3) Hunter, T. Cell 1995, 80, 225-36. (4) Till, J. H.; Annan, R. S.; Carr, S. A.; Miller, W. T. J. Biol. Chem. 1994, 269, 7423-28. (5) Songyang, Z.; Cantley, L. C. Methods Enzymol. 1995, 254, 523-35. 10.1021/ac990276t CCC: $18.00 Published on Web 07/16/1999

© 1999 American Chemical Society

dimensional phosphopeptide mapping.6 This method is used to elucidate phosphopeptide content but provides no direct sequence information. More recently, several techniques based on mass spectrometry have been reported for the specific identification of phosphopeptides.7-9 Both on-line and off-line analysis with or without chromatographic separation have been described. The detection of fragments characteristic for phosphopeptides and phosphoamino acids permit the sequencing of peptides containing one or more phosphate groups. The monitoring of kinase and phosphatase reactions at a single site in a peptide substrate was reported using time-of-flight mass spectrometry; however, quantitation was not possible.10 An on-line enzyme reactor has been used to identify peptides phosphorylated on tyrosine residues. Tyrosine-phosphorylated peptides were identified by subtractive analysis of peptide patterns generated with or without phosphatase treatment.11 Phosphopeptides generated by enzymatic or chemical cleavage can contain multiple sites of phosphorylation, and understanding the order (if any) in which these sites are phosphorylated or dephosphorylated will lead to an understanding of how protein kinases and phosphastases recognize specific substrates.5 A study of the sequential dephosphorylation of a multiply phosphorylated peptide by Ramachandran and co-workers used high-performance liquid chromatography (HPLC) and solid-phase sequencing to determine the order of dephosphorylation.12 Quantitation of the individual peptide isomers (three doubly phosphorylated and three single phosphorylated) was not possible because of the inability of HPLC to resolve phosphopeptide isomers. Phosphatase-mediated dephosphorylation of a doubly phosphorylated peptide was monitored using HPLC by Denu et al.13 Quantitation of the peptide isomers by HPLC required a 1-h gradient elution that did not fully resolve the singly phosphorylated isomers. (6) van der Geer, P.; Hunter, T. Electrophoresis 1994, 15, 544-54. (7) Hunter, A. P.; Games, D. E. Rapid Commun. Mass Spectrom. 1994, 8, 55970. (8) Annan, R. S.; Carr, S. A. Anal. Chem. 1996, 68, 3413-21. (9) Carr, S. A.; Huddleston, M. J.; Annan, R. S. Anal. Biochem. 1996, 239, 18092. (10) Craig, A. G.; Hoeger, C. A.; Miller, C. L.; Goedken, T.; Rivier, J. E.; Fischer, W. H. Biol. Mass Spectrom. 1994, 23, 519-28. (11) Amankwa, L. N.; Harder, K.; Jirik, F.; Aebersold, R. Protein Sci. 1995, 4, 113-25. (12) Ramachandran, C.; Aebersold, R.; Tonks, N. K.; Pot, D. A. Biochemistry 1992, 31, 4232-38. (13) Denu, J. M.; Zhou, G.; Wu, L.; Yuvaniyama, J.; Saper, M. A.; Dixon, J. E. J. Biol. Chem. 1995, 270, 3796-803.

Analytical Chemistry, Vol. 71, No. 16, August 15, 1999 3469

Table 1. Synthetic Peptides Used in This Study set

peak no.

name

amino acid sequence

1

1 2 3 4

SPS pSPS SPpS pSPpS

Gly Ser Pro Ser Pro Pro Pro Glu Glu Glu Ser Arg Gly pSer Pro Ser Pro Pro Pro Glu Glu Glu Ser Arg Gly Ser Pro pSer Pro Pro Pro Glu Glu Glu Ser Arg Gly pSer Pro pSer Pro Pro Pro Glu Glu Glu Ser Arg

2

5 6 7 8

TY pTY TpY pTpY

Glu Lys Ile Gly Glu Gly Thr Tyr Gly Val Val Tyr Lys Gly Arg His Lys Glu Lys Ile Gly Glu Gly pThr Tyr Gly Val Val Tyr Lys Gly Arg His Lys Glu Lys Ile Gly Glu Gly Thr pTyr Gly Val Val Tyr Lys Gly Arg His Lys Glu Lys Ile Gly Glu Gly pThr pTyr Gly Val Val Tyr Lys Gly Arg His Lys

The separation of peptides using capillary zone electrophoresis (CZE) is a well-established technique which offers extremely high resolution and peak efficiency. Resolution of phosphopeptide isomers would permit the quantitation of each peptide as it is produced by the phosphorylation or dephosphorylation of an appropriate substrate. The effect of peptide sequence on the ability of kinases and phosphatases to recognize substrates could also be studied. Micellar electrokinetic chromatography has been used to separate a mixture of eight phosphopeptide isomers,14 and singly phosphorylated peptide isomers have been separated using high ionic strength phosphate buffer.15 The study of kinase and phosphatase reactions has been monitored using CZE, but only at a single site of phosphorylation, and therefore did not require isomer separation.10,15 The goal of this research was to investigate the use of capillary electrophoresis for phosphopeptide isomer separations. Several approaches have been developed using a variety of columns and electrolyte systems. Key requirements where set in order to develop a method that was easy to use, was relatively fast, and provided baseline resolution of all components in order to quantitate individual species. Application of the methods developed to the study of protein kinases and phosphatases is presented. EXPERIMENTAL SECTION Reagents and Materials. Fused-silica capillary was purchased from Polymicro Technologies (Phoenix, AZ). Sodium phosphate monobasic and sodium phosphate dibasic salts were purchased from American Chemicals Ltd. (Montreal, PQ, Canada). Fischer Scientific (Nepean, ON, Canada) provided the sodium hydroxide, and orthophosphoric acid was obtained from BDH Chemicals Ltd. (Toronto, ON, Canada). Sodium dodecyl sulfate (SDS, 99%) was purchased from Lancaster Synthesis (Morecambe, England). Acetic acid (99%), formic acid (99%), ammonium persulfate (98%), N,N,N′,N′-tetramethylethylenediamine (TEMED, 99.5%), poly(vinyl alcohol) (PVA, 99%), acrylamide (99%), and 3-(trimethylsilyl)propyl methacrylate (97%) were purchased from Aldrich (Milwaukee, WI). Silar Laboratories (Wilmington, NC) provided 7-oct-1enyltrimethoxysilane. [(Acryloylamino)propyl]trimethylammonium chloride (called BCQ by the manufacturer) was supplied by Chemische Fabrik Stockhaussen (Krefeld, Germany). Two sets of phosphopeptide isomers were used for developing the isomer separations (Table 1). Set 1 peptides had the sequence GSPSPPPEEESR and were nonphosphorylated or phosphorylated (14) Tadey, T.; Purdy, W. C. Electrophoresis 1995, 16, 574-79. (15) Dawson, J. F.; Boland, M. P.; Holmes, C. F. B. Anal. Biochem. 1994, 220, 340-45.

3470

Analytical Chemistry, Vol. 71, No. 16, August 15, 1999

at serine 2 and/or 4. Set 2 consisted of the nonphosphorylated peptide EKIGEGTYGVVYKGRHK, the doubly phosphorylated peptide (residues seven (threonine) and eight (tyrosine)), and the two singly phosphorylated peptides with either the 7-phosphothreonine or the 8-phosphotyrosine. Set 2 peptide pY and the four set 1 peptides were synthesized by California Peptide Research (Napa, CA). The other three set 2 peptides were obtained from Research Genetics (Huntsvillle, AL). Several protein kinases and protein phosphatases were used to study the phosphorylation events. Flag protein tyrosine phosphatase 1B (Flag-PTP-1B)16 and bead-immobilized GST-tyrosine kinase (GST-TK) fusion protein17 were prepared as described. Mitogen activated protein (MAP) kinase and protein kinase A (PKA) were obtained from Calbiochem-Novabiochem Corp. (La Jolla, CA). The alkaline phosphatase (calf intestine) was supplied by Boehringer Mannhein GmbH, and the acid phosphatase (potato) was supplied by Sigma Chemical Co. (St. Louis, MO). Capillary Coatings. Capillary was typically coated in 1.5-5-m lengths and cut as needed for use in CZE experiments. All rinses were carried out using a GC column cleaning kit at 20-40 psi. BCQ Coating. The method used for preparing the BCQ coating was described previously.18,19 The capillary was rinsed sequentially with 1 M NaOH, deionized water, and methanol, each for 1 h at 20 psi. A solution of 7-oct-1-enyltrimethoxysilane (20 µL) and glacial acetic acid (20 µL) in methanol (4 mL) was passed through the column overnight (8-12 h) at 20 psi. The capillary was subsequently rinsed with methanol and deionized water (1 h each, 20 psi). TEMED (8 µL) and aqueous ammonium persulfate (15% (w/v), 56 µL) were added to a solution of BCQ in deionized water (2% (v/v), 4 mL), and this solution was immediately rinsed through the column for 8 h (or overnight) at 20 psi. The capillary was flushed with deionized water for 1 h and then stored. Prior to use, the column was flushed with CZE buffer for 5-10 min. PVA Coating. Poly(vinyl alcohol)-coated capillaries used a modified version of a method described previously.20 A solution of PVA (10% (w/v) in water) was warmed to just below boiling and filtered using a Millex-HA 0.45-µm syringe filter. An aliquot (100 µL) was diluted to 5% (w/v) in distilled water. Typically 1-2 (16) Huyer, G.; Liu, S.; Kelly, J.; Moffat, J.; Payette, P.; Kennedy, B.; Tsaprailis, G.; Gresser, M. J.; Ramachandran, C. J. Biol. Chem. 1997, 272, 843-. (17) Huyer, G.; Li, Z. M.; Adam, M.; Huckle, W. R.; Ramachandran, C. Biochemistry 1995, 34, 1040-. (18) Bateman, K. P.; White, R. L.; Thibault, P. J. Mass Spectrom. 1998, 33, 110923. (19) Bateman, K. P.; White, R. L.; Thibault, P. Rapid Commun. Mass Spectrom. 1997, 11, 307-15. (20) Clarke, N. J.; Tomlinson, A. J.; Schomburg, G.; Naylor, S. Anal. Chem. 1997, 69, 2786-92.

Figure 1. CZE-UV analysis of set 2 peptides (100 µg/mL each) using (a) 0.1, (b) 0.5, (c) 1.0, and (d) 2.0 M formic acid. Separations were carried out at +15 kV on a 57 cm × 50 µm uncoated capillary. Peak labels as per Table 1.

m of capillary was washed with water (20 min) followed by the warm PVA solution (20 min). The excess PVA solution was then purged from the capillary with nitrogen gas at 40 psi. The capillary was then placed inside a GC oven maintained at 145 °C and baked for 3 h with a steady flow of nitrogen gas (30 psi). Capillary Zone Electrophoresis. CZE experiments were carried out using a Beckmann P/ACE System 5000 capillary electrophresis system (Beckmann Instruments, Fullerton, CA). For all experiments, a 57 cm × 50 µm i.d. column was used. Separations performed using an uncoated capillary used varying concentrations of formic acid as the electrolyte. Initial conditioning of the capillary used 0.1 M NaOH (5 min), water (5 min), and electrolyte (10 min). The capillary was rinsed with 0.1 M NaOH (1 min), water (1 min), and electrolyte (2 min) between analyses. Depending on the set of phosphopeptides being separated on the PVA-coated capillary, either formic acid or 25 mM phosphate buffers were used as the medium during separation. The phosphate buffers were prepared with sodium phosphate monobasic and titrated with either phosphoric acid or sodium phosphate dibasic to reach the desired pH. Separation on the BCQ-coated capillary employed varying concentrations of formic acid as the electrolyte. For both the PVA-coated capillary and the BCQ-coated capillary, the column was rinsed for 2 min with the running electrolyte prior to injection. Detection for separations using the uncoated, BCQ-coated and PVA-coated capillaries was by UV absorbance at 200 nm.

Enzyme Incubations. Peptides from both sets were incubated with various protein kinases and protein phosphatases. The phosphorylation and dephosphorylation were monitored using the CZE methods discussed below for the PVA-coated capillary. The separation of set 1 peptides involved a 25 mM phosphate running buffer (pH 6.3) and set 2 peptides using a running electrolyte of 2 M formic acid. Total reaction volume for all incubations was 100 µL, and sample injection was performed using pressure for 10 s. Phosphatase Incubations. The fully phosphorylated set 1 peptide (100 µg/mL) was incubated with two separate phosphatases (alkaline phosphatase and acid phosphatase). The alkaline phosphatase (∼0.1 µg/mL) was incubated with the substrate at room temperature in 50 mM Tris HCl (pH 8.0) and 5 mM magnesium chloride. The acid phosphatase (∼0.8 µg/mL) was incubated with substrate at room temperature in 50 mM Bis Tris (pH 6.3) and 2 mM EDTA. The incubations were repeated with fully phosphorylated set 2 peptide (100 µg/mL) using identical conditions. The fully phosphorylated peptide from set 2 (100 µg/mL) was also incubated at room temperature with flag-PTP-1B. The reaction mixture consisted of 50 mM Bis Tris (pH 6.3), 2 mM EDTA, 5mM DTT, 20% glycerol, and 0.1 µg/mL flag-PTP-1B. Kinase Incubations. The nonphosphorylated peptides from set 1 (50 µg/mL) and set 2 (50 µg/mL) were individually incubated with two separate kinases, MAP kinase and PKA at room temperature. The reaction mixture for the incubation with MAP Analytical Chemistry, Vol. 71, No. 16, August 15, 1999

3471

Figure 3. Peptide mobility as a function of pH for the separation of set 1 peptides on a PVA-coated capillary (57 cm × 50 µm) with 25 mM phosphate buffer.

Figure 2. Optimized separation on a BCQ-coated capillary (57 cm × 50 µm, -15 kV) with 2 M formic acid for (a) set 1 peptides (100 µg/mL each) and (b) set 2 peptides (100 µg/mL each). Peak labels as per Table 1.

kinase consisted of 40 µg/mL MAP kinase, 20 mM HEPES buffer (pH 7.55), 5 mM MgCl2, 5 mM β-mercaptoethanol, and 1 mM ATP. The reaction mixture for the incubation with PKA differed only in the amount of kinase used (1000 units/mL). The nonphosphorylated set 2 peptide was also incubated with the immobilized GST-TK at 27 °C. Approximately 50 µL of the bead suspension was washed three times with wash buffer (50 mM imizadole pH 7.2, 10 mM DTT) and twice with reaction buffer (50 mM imidazole pH 7.2, 10 mM DTT, 30 mM MgCl2, 1 mM MnCl2, 1 mM NaVO3, 0.05% Triton X-100). The reaction buffer was then added to the beads along with 5 mM ATP and 1 mg/ mL peptide. The mixture was continually shaken. For reaction sampling, the mixture was centrifuged for 10 s at 14 000 rpm and 10 µL of the supernatant was transferred to a separate vial and quenched with 2 M formic acid in preparation for CZE analysis. RESULTS AND DISCUSSION The peptides used to develop standard conditions for the separation of phosphopeptide isomers are shown in Table 1. The set 1 peptides are acidic in nature with a pI of 4.25 for the nonphosphorylated peptide. The peptide sequence is from a cAMP-specific phosphodiesterase that is known to be phosphorylated in vivo.21,22 The nonphosphorylated set 2 peptide has a pI of 9.41 and is from the enzyme cell division cycle 2 kinase, which (21) Lenhard, J. M.; Kassel, D. B.; Rocque, W. J.; Hamacher, L.; Holmes, W. D.; Patel, I.; Hoffman, C.; Luther, M. Biochem. J. 1996, 316, 751-58. (22) Sette, C.; Conti, M. J. Biol. Chem. 1996, 271, 16526-34.

3472 Analytical Chemistry, Vol. 71, No. 16, August 15, 1999

Figure 4. Separation of set 1 peptides (100 µg/mL each) on a 57 cm × 50 µm PVA-coated capillary using 25 mM phosphate pH 6.36 and an applied voltage of -22.5 kV. Peak labels as per Table 1.

is also known to be phosphorylated in vivo.23 The peptides contain serine, threonine, and/or tyrosine residues that are phosphorylated, representing the most common sites of protein phosphorylation. Several conditions for separating each set of peptides were investigated in order to have conditions that may apply to other sets of phosphopeptides not explicity studied here. Uncoated Capillary. Initially an uncoated capillary with various concentrations of formic acid as the separation electrolyte was used for the analysis of the set 1 and set 2 peptides. The low pH (1.5-2.5) resulted in reduced electroosmotic flow and long analysis times for the acidic set 1 peptides. As the concentration of the electrolyte was increased, the resolution of the two singly phosphorylated peptides was increased for both sets of peptides. At the highest concentration of formic acid (2.0 M), the peptides could be baseline resolved. However, an impurity could not be resolved and the analysis time was over 40 min for the set 1 (23) Coleman, T. R.; Dunphy, W. G. Curr. Opin. Cell Biol. 1994, 6, 877-82.

Figure 5. CZE-UV analysis of set 2 peptides (100 µg/mL each) using (a) 0.1, (b) 0.5, (c) 1.0, and (d) 2.0 M formic acid. Separations were carried out at +15 kV on a 57 cm × 50 µm PVA-coated capillary. Peak labels as per Table 1.

peptides (data not shown). The effect of increasing electrolye concentration on the resolution of the set 2 peptides is shown in Figure 1. The increase in acid strengths results in decreased EOF and improved peak resolution, especially for peak 8 in Figure 1.24 BCQ-Coated Capillary. An alternative approach was investigated for the set 1 peptides in order to decrease the analysis time. Previously we showed that BCQ-coated capillaries with formic acid electrolyte provided excellent separation conditions for peptides and glycopeptides.24 Optimized conditions for the separation of set 1 peptides is shown in Figure 2a. The analysis time is less than half that required for the uncoated capillary; however, baseline resolution of the isomers could not be achieved. Separation of the set 2 peptides on the BCQ-coated capillary is shown in Figure 2b and is similar to the separation on the uncoated capillary. It should be noted that the migration order of the peptides is reversed because of the cationic coating. PVA-Coated Capillary. A third approach for separating the peptides was developed using PVA-coated capillaries. This coating eliminates the electroosmotic flow so that separations are driven purely by the mobility and differences in mobility of the peptides. For the set 1 peptides, separations were carried out using 25 mM phosphate buffer at various pH’s. These peptides are acidic, and at the pH’s used for separation are anionic; therefore, separations were effected in reverse-polarity mode. A plot of mobility versus (24) Bateman, K. P.; White, R. L.; Yaguchi, M.; Thibault, P. J. Chromatogr. A 1998, 794, 327-44.

pH (Figure 3) for the four peptides in the mixture revealed that the optimal pH for separation was between 5.5 and 6.5. The optimized separation of the set 1 peptides is shown in Figure 4. The two singly phosphorylated peptides are baseline resolved using these conditions. Separation conditions for the set 2 peptides were also developed using the PVA-coated capillary. As with the uncoated and the BCQ-coated capillary, formic acid at various ionic strengths was used. The separation of the peptides at four different concentrations of formic acid is shown in Figure 5. The phosphopeptide isomers are resolved using 1.0 M formic acid (Figure 5c) and increasing the acid concentration to 2.0 M further enhances the resolution (Figure 5d). This effect is similar to what was seen for the uncoated capillary. Since the PVA-coated capillary was able to separate both sets of peptides it was used for all subsequent enzyme incubation experiments. The apparent electrophoretic mobilities for the peptides at the optimized conditions for each capillary are listed in Table 2 (set 1) and Table 3 (set 2). Mobility differences for the singly phosphorylated peptide isomers is consistent among the three columns used. The mobility difference for the set 1 isomers on the BCQ column is less than for all other conditions as expected since they were not resolved using these conditions. It has been reported that the separation of peptides with the same residues but different sequences can be attributed to “nearest-neighbor” Analytical Chemistry, Vol. 71, No. 16, August 15, 1999

3473

Table 2. Apparent Mobilities for Set 1 Peptides at Optimized Conditions apparent mobility (×10-6 in m2/V‚s) peptidea

uncoated

PVA-coated

BCQ-coated

1 2 3 4

1.132 0.757 0.782 0.466

0.435 0.723 0.700 0.921

1.932 2.321 2.301 2.639

|∆µ|, singly phosphorylated peptide

0.024

0.023

0.020

a

Peptides numbered according to Table 1.

Table 3. Apparent Mobilities for Set 2 Peptides at Optimized Conditions apparent mobility (×10-6 in m2/V‚s) peptidea

uncoated

PVA-coated

BCQ-coated

5 6 7 8

1.841 1.613 1.589 1.358

1.539 1.315 1.291 1.067

0.987 1.211 1.236 1.464

|∆µ|, singly phosphorylated peptide

0.024

0.024

0.025

a

Peptides numbered according to Table 1.

effects.25 In this case, the position of the phosphate may affect the pKa of charged residues, resulting in a different net charge. Alternatively, the phosphate may induce a change in the conformation of the peptide due to intramolecular charge interactions.26 Whatever the mechanism, these effects lead to a change in overall mobility and permit the isomers to be resolved. Enzyme Incubations. Separation of peptide isomers permits the quantitation of phosphorylation and dephosphorylation at specific residues in the peptide substrates. This enables the study of, for example, the specificity and/or the rate of the phosphorylation/dephosphorylation of peptide substrates. The optimized conditions for the PVA-coated capillary were used to monitor both kinase and phosphatase reactions of the set 1 and set 2 peptides. Dephosphorylation of the fully phosphorylated substrates was studied using three different phosphatases, acid phosphatase, alkaline phosphatase, and PTP-1B (PTP-1B was used with set 2 only). For the phosphorylation reactions, the nonphosphorylated substrates were incubated with three different kinases, MAPK, PKA, and GST-TK (GST-TK was used with set 2 only). Phosphatase Incubations. The fully phosphorylated peptide from set 1 was incubated with both alkaline phosphatase and acid phosphatase in separate experiments. The reaction mixture was placed in the CZE instrument’s autosampler and injections occurred approximately every 40 min over a 6-h time period. The results for both reactions are plotted as normalized peak height versus time in Figure 6. In both experiments, the two singly phosphorylated and the nonphosphorylated peptides are generated. The fact that the singly phosphorylated peptides are (25) Landers, J. P. Trends Biochem. Sci. 1993, 18, 409-14. (26) Thorsteinsdo´ttir, M.; Beijersten, I.; Westerlund, D. Electrophoresis 1995, 16, 564-73.

3474 Analytical Chemistry, Vol. 71, No. 16, August 15, 1999

Figure 6. Time course study for the dephosphorylation of set 1 peptide 4 (GpSPpSPPPEEESR) using (a) alkaline phosphatase and (b) acid phosphatase.

observed indicates that, after removal of one phosphate group, the peptide is released from the enzyme’s active site. For both phosphatases, cleavage of the phosphate from the second phosphoserine appears to be favored, since the singly phosphorylated peptide GpSPSPPPEEESR is more abundant. For the alkaline phosphatase (Figure 6a), sampling of the reaction was terminated before the reaction went to completion. The activity of the acid phosphatase (Figure 6b) was lost before this reaction was completed. Repetition of this experiment with a more active enzyme preparation resulted in complete dephosphorylation of the peptides in less than 40 min (data not shown). Experiments using the singly phosphorylated peptides would need to be conducted to better evaluate any differences in affinity for these peptides. The set 2 peptide, EKIGEGpTpYGVVYKGRHK, contains both phosphothreonine and phosphotyrosine residues. This peptide was incubated with both alkaline and acid phosphatase, as well as the tyrosine specific phosphatase PTP-1B. Experiments were carried out directly in the autosampler and time courses are shown in Figure 7. The reaction using PTP-1B went to near completion with only the phosphotyrosine residue being dephosphorylated (Figure 7a), verifying that this enzyme is specific for these residues. The reaction time course for the acid phosphatase incubation (Figure 7b) indicates that the tyrosine residue is preferentially dephosphorylated, resulting in the appearance of the monophosphorylated peptide containing phosphothreonine in much greater abundance relative to the phosphotyrosine peptide. This peptide

Figure 8. Time course study of the phosphorylation of set 1 peptide 1 (GSPSPPPEEESR) using MAP kinase.

Figure 7. Time course study for the dephosphorylation of set 2 peptide 8 (EKIGEGpTpYGVVYKGRHK) using (a) protein tyrosine phosphatase 1B and (b) acid phosphatase.

is subsequently dephosphorylated to produce the nonphosphorylated peptide. Kinase Incubations. The nonphosphorylated set 1 peptide, GSPSPPPEEESR, was incubated with MAP kinase and protein kinase A in separate reactions. The time course for the MAP kinase is shown in Figure 8; the protein kinase A reaction gave identical results (data not shown). At the concentration of enzyme used, the conversion of nonphosphorylated to phosphorylated peptide was quite slow, and neither reaction went to completion during the sampling period. There was no evidence to suggest that either of the singly phosphorylated peptides (peptides 2 and 3 in Table 1) was being produced or that the third serine residue was being phosphorylated. These results indicate that the substrate enters the active site of the enzyme, is phosphorylated on both serines, and then is released. The singly phosphorylated peptides may serve as substrates for the enzyme and this would need to be evaluated by incubating these peptides with the enzymes. The incubation of the nonphosphorylated set 2 peptide with MAP kinase and protein kinase A did not produce any peaks corresponding to phosphorylated peptides in the CZE-UV. Samples were analyzed over a 16-h period. When this peptide was incubated with GST-TK, a tyrosine-specific kinase, there was almost complete turnover of the peptide to a phosphorylated form (Figure 9). The peak generated comigrated with the singly phosphorylated peptide, EKIGEGTpYGVVYKGRHK, indicating that phosphorylation was specific for the first tyrosine residue. This residue is known to be phosphorylated in vivo,23 suggesting that the residues

Figure 9. Time course study of the phosphorylation of set 2 peptide 5 (EKIGEGTYGVVYKGRHK) using tyrosine kinase.

adjacent to the tyrosines play a role in how this kinase recognizes the substrate. CONCLUSIONS Several conditions for the separation of phosphopeptide isomers have been developed using two sets of peptide standards. Both coated (neutral and cationic) and uncoated capillaries have been used successfully with relatively simple separation conditions. The best conditions were found to be formic acid or phosphate buffer with a PVA-coated capillary. These techniques may prove valuable for studying both protein kinases and phosphatases by permitting the analysis of individual sites of phosphorylation. The resolution of phosphopeptide isomers permitted the application of the CZE method to the study of protein phosphatases and kinases. Incubation of phosphorylated or nonphosphorylated peptides with protein phosphatases or kinases resulted in the appearance of nonphosphorylated or phosphorylated pepAnalytical Chemistry, Vol. 71, No. 16, August 15, 1999

3475

tides. Specific residue dephosphorylation or phosphorylation could be determined because of the resolution provided by the CZE analysis. The development of standard conditions for the separation of phosphopeptide isomers is expected to assist in the study of protein kinases and phosphatases. The use of UV detection permits quantification of substrate and products. Application of

3476 Analytical Chemistry, Vol. 71, No. 16, August 15, 1999

this method to the characterization of enzymes in terms of their substrate specificity and kinetic parameters is ongoing.

Received for review March 15, 1999. Accepted May 21, 1999. AC990276T