Phosphoproteome Analysis of Capacitated Human Sperm

Before fertilization can occur, mammalian sperm must undergo capacitation, a process that requires a cyclic AMP-dependent increase in tyrosine phosphorylation. To identify proteins phosphorylated during capacitation, two-dimensional gel analysis coupled to anti-phosphotyrosine immunoblots and tandem mass spectrometry (MS/MS) was performed. Among the protein targets, valosin-containing protein (VCP), a homolog of the SNARE-interacting protein NSF, and two members of the A kinase-anchoring protein (AKAP) family were found to be tyrosine phosphorylated during capacitation. In addition, immobilized metal affinity chromatography was used to investigate phosphorylation sites in whole protein digests from capacitated human sperm. To increase this chromatographic selectivity for phosphopeptides, acidic residues in peptide digests were converted to their respective methyl esters before affinity chromatography. More than 60 phosphorylated sequences were then mapped by MS/MS, including precise sites of tyrosine and serine phosphorylation of the sperm tail proteins AKAP-3 and AKAP-4. Moreover, differential isotopic labeling was developed to quantify phosphorylation changes occurring during capacitation. The phosphopeptide enrichment and quantification methodology coupled to MS/MS, described here for the first time, can be employed to map and compare phosphorylation sites involved in multiple cellular processes. Although we were unable to determine the exact site of phosphorylation of VCP, we did confirm, using a cross-immunoprecipitation approach, that this protein is tyrosine phosphorylated during capacitation. Immunolocalization of VCP showed fluorescent staining in the neck of noncapacitated sperm. However, after capacitation, staining in the neck decreased, and most of the sperm showed fluorescent staining in the anterior head.

After ejaculation, sperm are able to move actively but lack fertilizing competence. They acquire the ability to fertilize in the female genital tract in a time-dependent process called capacitation (1). Capacitation is accompanied by a cAMPprotein kinase-dependent increase in tyrosine phosphorylation of a subset of proteins (2,3). Because a protein kinase cascade is involved in the regulation of the sperm fertilizing ability, it is important to characterize the proteins that undergo phosphorylation and examine how these changes relate to capacitation.
Post-translational protein phosphorylation by protein kinases plays a role in many cellular processes including transduction of extracellular signals, intracellular transport, and cell cycle progression. The use of two-dimensional gel electrophoresis followed by tandem mass spectrometry (MS/MS) 1 provides a comprehensive approach to the analysis of proteins involved in cell signaling (4). Specifically, changes in tyrosine phosphorylation can be monitored using two-dimensional gel electrophoresis followed by Western blot analysis with antiphosphotyrosine (␣-PY) antibodies (5). Proteins that undergo changes in tyrosine phosphorylation during cellular processes can be then isolated from a complementary gel and sequenced by MS/MS. In the present study, we have used this approach to identify several sperm proteins that undergo tyrosine phosphorylation during capacitation.
Identification of the site at which a particular protein is phosphorylated is important in identifying the physiologically relevant protein kinase involved in a particular pathway. In addition, to understand the function of a phosphoprotein, phosphorylation sites are excellent candidates for site-directed mutagenesis. Determination of individual phosphorylation sites in vivo often requires the purification to homogeneity and/or mutation analysis of the phosphoprotein. Recently, several methods for the selective detection and enrichment of phosphopeptides have been developed (6 -8); however, most of them have been applied only on a protein-by-protein basis. In the present work, we have used Fe 3ϩ -immobilized metal affinity chromatography (IMAC) prior to MS/MS to enrich digests for peptides containing phosphoamino acids. To increase the selectivity of the IMAC column for phosphopeptides, we have used a recently developed modification of this technique in which acidic resi-dues are converted to methyl esters to block their binding to iron before IMAC is employed (9). Using this methodology, 5 sites of Tyr, 56 of Ser, and 2 of Thr phosphorylation have been characterized.
Although the combination of IMAC and MS/MS is ideally suited for the characterization of phosphorylation sites on proteins in complex mixtures, it is also important to determine which sites are phosphorylated in response to a particular stimulus. In the present paper, we have used differential isotopic labeling to quantify phosphorylation of defined sequences and to determine which sites suffer increased or decreased phosphorylation during human sperm capacitation. The evaluation of differential phosphorylation, added to the knowledge of the exact phosphorylated sequence, goes beyond the sperm capacitation field and could be used to understand signaling mechanisms in a multiplicity of cell systems.
Identification of the tyrosine phosphorylation sites of AKAP-3 and AKAP-4 confirmed the previous findings using a two-dimensional gel approach, suggesting that these proteins are tyrosine phosphorylated in a capacitated sperm population. Nevertheless, because of the relative abundance of tyrosinephosphorylated phosphopeptides compared with phosphoserine phosphopeptides, most of the tyrosine-phosphorylated sites remained unknown. Among them, sites in the NSF homolog VCP remained unidentified. Because VCP is associated with a role in membrane fusion in other cell types and because several aspects of sperm physiology required membrane fusion events, VCP was chosen for further investigation. Anti-VCP antibodies confirmed that this protein is tyrosine phosphorylated during capacitation. In addition, immunolocalization of VCP was determined and showed different localization patterns before and after human sperm capacitation.

EXPERIMENTAL PROCEDURES
Preparation of Spermatozoa-The basic medium used for all experiments was modified human tubal fluid (Irvine Scientific, Santa Ana, CA). Semen specimens were obtained by masturbation (approved by the UVA Human Investigation Committee). Individual semen samples were allowed to liquefy at room temperature (0.5-3 h), and mature sperm were purified by Percoll (Amersham Biosciences) density gradient centrifugation as described previously (5). Sperm presenting Ͼ 90% motility were treated immediately to obtain a noncapacitated population or capacitated at 37°C in 5% CO 2 incubator. After overnight capacitation, sperm were concentrated by centrifugation at 20,000 ϫ g for 2 min at room temperature and then washed once in 0.5 ml of phosphate-buffered saline (PBS) at room temperature. The sperm pellet was then resuspended in the appropriate buffer depending on the experiment. For two-dimensional gel analysis, sperm were resuspended in Celis buffer (10). For immunoprecipitation and/or for direct MS/MS, cells were resuspended in 0.1% SDS, 25 mM Tris-HCl, pH 7.5, and boiled for 5 min, centrifuged at 20,000 ϫ g for 5 min, and the supernatant recovered.
Image Analysis of Two-dimensional Gels-Gel electrophoresis was performed concurrently to ensure equivalent electrophoretic conditions. Gels were stained with silver (5) or electroblotted to polyvinylidene difluoride membrane and probed with ␣-PY (clone 4G10, Upstate Biotechnology). The silver-stained gel and x-ray films (short and long exposure of ECL) were scanned at 300 dpi using a desktop Hewlett Packard scanner. Digitized images were overlaid in Adobe Photoshop 6.0 using different percents of transparency. Using known "landmarks" such as fibrous sheath proteins and tubulins, the silver image was aligned with the ECL images. After marking all reactive spots on the ECL image with arrows, the ECL image was hidden and the arrows identified corresponding silver-stained spots. These spots were then cored from the silver-stained gel and submitted for mass spectrometry analysis.
Immunofluorescence Microscopy-Noncapacitated and capacitated sperm were air dried onto slides, washed three times with PBS, permeabilized with methanol, washed with PBS, and then blocked with 10% normal goat serum in PBS. Incubations were then carried out with ␣-PY or ␣-VCP/p97 antibodies (1:250) diluted in PBS with 1% normal goat serum, washed, and incubated with FITC-conjugated F(abЈ) 2 fragments of donkey ␣-mouse IgG (1:200) or donkey ␣-rabbit IgG (1:200) (Jackson ImmunoResearch) in normal goat serum in PBS. Slides were washed with PBS and mounted with Slow-Fade Light (Molecular Probes, Eugene, OR). Sperm were observed by differential interference contrast and epifluorescence microscopy using a Zeiss Axiophot microscope (Carl Zeiss, Inc. Thornwood, NY). Results depicted in Fig. 1 represent the mean Ϯ S.E. of three independent experiments with triplicate determination. Statistical differences between the groups were determined using Student's t test, comparing the capacitated with the noncapacitated population after arcsin transformation (12).
Preparation of Samples for MS/MS and for Immunoprecipitation-Sperm were directly treated or capacitated for 18 h as described above; 1 ϫ 10 8 sperm were centrifuged, washed in PBS, and then boiled in lysis buffer containing 150 mM NaCl, 0.1% SDS, and 25 mM Tris-HCl, pH 7.5, for 5 min. This treatment was efficient in solubilizing tyrosine-phosphorylated proteins. Proteins solubilized by this method were subsequently treated with trypsin or V8 protease (Glu-C) and tryptic peptides analyzed by MS/MS. This solubilization procedure was also used for immunoprecipitation with ␣-PY (4G10). Briefly, 10 g of ␣-PY antibody was added to the suspension and incubated at room temperature for 1 h. Then 100 l of a 30% suspension of protein A-Sepharose (PAS) was added and incubated further for 1 h. The PAS was then washed five times by repetitive centrifugation with lysis buffer. Finally the washed PAS beads were resuspended in 1 mM phenyl phosphate and incubated further for 1 h. The suspension was then centrifuged and the remaining supernatant divided; 90% of the sample was then used for MS/MS, and the remaining supernatant was analyzed with ␣-PY and ␣-VCP/p97. ␣-AKAP-3 immunoprecipitation was performed using 5 l of ␣-AKAP-3 serum (13) in a similar manner. In the experiments depicted in Fig. 2, ␣-PY and ␣-AKAP-3 immunoprecipitates were boiled in sample buffer (14) and analyzed by Western blot using ␣-PY RC20 (Transduction Laboratories), ␣-VCP, and ␣-AKAP-3 antibodies.
Preparation of Peptide Methyl Esters-Digests were evaporated to dryness. Methanolic 2 N HCl or DCl was prepared by adding 160 l of acetyl chloride (Aldrich) to 1 ml of anhydrous d 0 -methyl alcohol (Aldrich) or d 3 -methyl d-alcohol (Aldrich) dropwise with stirring. After 5 min, 200 l of the reagent was added to lyophilized peptide mixtures. This solution was incubated at room temperature for 2 h and lyophilized. If necessary, the procedure was repeated to obtain full conversion of carboxyl groups to methyl esters. Lyophilized peptide methyl esters were reconstituted in 0.1% acetic acid (standard protein digests) or 1:1:1 acetonitrile, methanol, and 0.1% acetic acid (sperm protein digests).
IMAC-IMAC was performed as described previously (16) with modifications (9). IMAC columns were constructed by packing 8-cm POROS 20 MC (PerSeptive Biosystems, Framingham, MA) into fused silica microcapillaries (360 ϫ either 200-or 100-m inner diameter; Polymicro Technologies, Phoenix, AZ). Metal ions were removed from the column by washing with 50 mM EDTA. Excess EDTA was removed by rinsing with water. The column was then activated with 100 l of 100 mM FeCl 3 . Excess iron atoms were removed by rinsing with 20 l of 0.1% acetic acid. Peptide mixtures, pH Ϸ 3.5, were loaded onto the IMAC column at a rate of 1 l/min. Nonbinding peptides were removed by rinsing with 25 l of acetonitrile/water/acetic acid (25/74/1 v/v/v) containing 100 mM NaCl (Aldrich). The IMAC column was re-equilibrated with 10 -20 l of 0.1% acetic acid and was then connected to a second fused silica column (360 ϫ 75-m inner diameter or 360 ϫ 100-m inner diameter, fused silica). Phosphopeptides were eluted with 5 l (360 ϫ 75-m inner diameter IMAC column) or 8 l (360 ϫ 100-m inner diameter IMAC column) of 50 mM Na 2 HPO 4 , pH 9.0. The reversed phase column was then disconnected and rinsed with 0.1% acetic acid to remove salts before subsequent MS analysis.
General LC/MS Parameters-All HPLC experiments employed a gradient of 0 -60% B in 40 min (unless noted otherwise) composed of solvent A (0.1 M acetic acid) and solvent B (70% acetonitrile with 0.1 M acetic acid). All microcapillary column connections were made with 1 cm of 0.152 ϫ 0.03 cm inner diameter Teflon tubing (Zeus, Orangeburg, SC). In experiments performed with IMAC enrichment, peptides were loaded onto an IMAC column and selectively eluted to a C18 microcapillary column (see above). In experiments performed without IMAC enrichment, peptides were loaded directly onto a C18 microcapillary column.
LC/MS Parameters on the LCQ Quadrupole Ion Trap Mass Spectrometer-Peptide mixtures were analyzed as described (16,17). A C18 microcapillary column containing peptides of interest was connected to an analytical column with an integrated ESI emitter tip (1-5-m diameter). Peptides were gradient eluted into an LCQ quadrupole ion trap mass spectrometer (spray voltage ϭ 1.6 kV). All MS/MS scans (both targeted and data-dependent) were performed with an isolation window of 3 Da (precursor m/z Ϯ 1.5 Da). For data-dependent analyses, the dynamic exclusion option was selected with a repeat count of 1, a repeat duration of 0.5 min, and exclusion duration of 1 min.

LC/MS Parameters on the Fourier Transform-Ion Cyclotron Resonance (FT-ICR) Mass
Spectrometer-Peptide mixtures were also analyzed on a home-built FT-ICR mass spectrometer (17). C18 microcapillary columns containing the peptide of interest were connected to analytical columns with integrated ESI emitter tips. Peptides were eluted into the mass spectrometer with the above gradient. Full scan mass spectra (m/z 300 -5,000) were acquired at a rate of ϳ1 scan/s. Mass resolving power ranged from 5,000 to 10,000.
Phosphoproteome Analysis of Capacitated Sperm Total Protein Digests-Aliquots containing 700 g and 2 mg of capacitated sperm total protein were digested with trypsin and Glu-C (1:20 enzyme:substrate ratio), respectively, for 18 h at 37°C. Both digests were performed in 0.1 M ammonium acetate, pH 8.5. Peptides resulting from each digestion were incubated twice with methanolic HCl as described. The resulting peptide methyl esters were loaded onto activated IMAC columns (360 ϫ 200-m inner diameter, fused silica). Peptides were gradient eluted into the LCQ ion trap mass spectrometer with an HPLC gradient (0 -60% B in 40 min; 60 -100% B in 5 min).
Analysis of Immunoprecipitated Sperm Protein Digests-Aliquots containing 50 g of immunoprecipitated sperm proteins were digested separately with trypsin and Glu-C (1:20 enzyme:substrate ratio) for 18 h at 37°C in 100 mM ammonium bicarbonate, pH 8.5. From 5 to 45 g of the resulting digest was desalted and analyzed by IMAC/RP-HPLC/MS as described above with or without prior methyl ester modification.
Post-IMAC Dephosphorylation of Peptides-Alkaline Phosphatase (AP) columns were constructed by packing 360 ϫ 200-m inner diameter fused silica microcapillaries with 12 cm of immobilized calf intestine AP (matrix F7m; 50-m polyvinyl spheres) from Mobitec (Marco Island, FL). Columns were first rinsed with 20 l of reaction buffer (supplied by the manufacturer; 50 mM Tris HCl, 0.1 mM ZnCl 2 , 1 mM MgCl 2 , pH 8.0). The AP column was then connected to an IMAC column charged with the phosphopeptides of interest. A C18 microcapillary column (360 ϫ 100-m inner diameter packed with 6 cm of 5-20-m irregular C18 particles) was connected to the AP column. Phosphopeptides were eluted from the IMAC column by rinsing with 10 l of modified reaction buffer (same as above except that the pH was raised to 9.0 with NH 3 and EDTA). The IMAC column was disconnected, and the AP/C18 column array was rinsed with 15 l of 0.1% acetic acid. Finally, the C18 column was disconnected and rinsed with 0.1% acetic acid before MS analysis of dephosphorylated peptides.
Standard Protein Digests for Phosphopeptide Quantitation-Standard protein solutions for phosphopeptide quantitation were made from stock solutions of ␤-casein (50 pmol/l). The indicated amount of protein was added to 100 mM NH 4 HCO 3 , pH 8.5 (500 l, final volume). To each solution was added 1 g of trypsin (in a volume of 2 l). Solutions were incubated for 18 h at 37°C, and the resulting peptides were converted to the corresponding methyl esters as described above. Derivatized peptides were dissolved in 500 l of 0.1% acetic acid. IMAC/RP-HPLC/ MS/MS analyses of derivatized digests were performed on a quadrupole ion trap and a FT-ICR mass spectrometer as described above.
Global Quantitation of Phosphopeptides from Capacitated and Noncapacitated Sperm Total Protein Digests-Approximately 800 g each of capacitated and noncapacitated sperm total protein was digested with trypsin (1:20 enzyme:substrate ratio) for 18 h at 37°C. The protein digest of capacitated sperm was treated with d 3 -methanolic DCl, whereas the protein digest of noncapacitated sperm was treated with d 0 -methanolic HCl. Peptide methyl esters were dissolved in 1:1:1 acetonitrile/methanol/0.1% acetic acid, and 100 g of each digest was loaded onto a 360 ϫ 200-m inner diameter IMAC column. Phosphopeptides were eluted to separate 360 ϫ 100-m inner diameter C18 columns and gradient eluted into the mass spectrometer. In another experiment, equal portions of each digest were combined, and 40 g of total peptide was loaded onto the IMAC column and gradient eluted into a FT-ICR instrument.
Data Base Searching of Phosphopeptide MS/MS Spectra and Sequence Assignments-MS/MS spectra were matched to sequences in various protein data bases using SEQUEST (18). Spectra from sperm protein analyses were searched against the nonredundant protein data base (NRPD) from NCBI and with a subdata base of proteins from NRPD which contained the phrase testis-specific. In these searches, differential modification of 80 Da to Ser, Thr, and Tyr residues was selected. For searches of MS/MS spectra recorded during analysis of peptide methyl esters, a differential modification of 14 Da to Glu and Asp acid and a static modification of 14 Da to the C terminus were also selected. Rapid identification of phosphopeptide candidates from the 1,000 MS/MS spectra acquired during a typical HPLC gradient was accomplished with an in-house computer program (neutral loss tool) (16). This program screens MS/MS spectra for losses characteristic of phosphorylated peptides (98, 49, and 32.6 Da from single, double, and triple charged precursor phosphopeptides, respectively). Neutral loss of phosphoric acid from the peptide precursor mass is a common feature of ion trap MS/MS spectra (19). For all phosphotyrosine-containing peptides and one peptide containing two phosphoserine residues, synthetic peptide MS/MS spectra matched experimentally obtained MS/MS spectra, confirming our assignments. For all sequences reported, spectra were manually validated and contained sufficient information to assign not only the sequence but also the site of phosphorylation (unless otherwise noted).

Characterization of Proteins That Undergo Tyrosine
Phosphorylation during Capacitation-To identify the proteins that serve as substrates for tyrosine phosphorylation during capacitation, human sperm proteins were extracted before and after overnight capacitation and separated by two-dimensional gel electrophoresis. As described previously (20), after transfer of the two-dimensional gels to Immobilon P, a capacitationassociated increase in protein-tyrosine phosphorylation was observed by Western blot using ␣-PY (Fig. 1A). Immunofluorescence experiments also showed an increase in ␣-PY fluorescent staining, indicating an increase in tyrosine phosphorylation (Fig. 1B). Before capacitation, 18% of the sperm displayed a low intensity ␣-PY signal in the tail only. After overnight capacitation, there was a significant increase in the number of sperm that show fluorescent staining of the head. Although we did not use an imaging program to quantify fluorescence intensity, it was possible to observe qualitatively an increase in the fluorescence of the tail in capacitated sperm compared with a noncapacitated sperm population as observed in Fig. 1B.
To identify tyrosine-phosphorylated proteins, extracts from capacitated human sperm were separated by two-dimensional gel electrophoresis. In each experiment, two gels were run in parallel. One gel was stained with Coomassie Blue and subsequently with silver, and the other was transferred and probed with ␣-PY (Fig. 1C). Both the silver-stained gel and the Western blot were scanned and compared. Protein spots showing ␣-PY staining were excised, digested, and sequenced. The results of this analysis are summarized in Table I and their exact localization shown in Fig. 1C.
Phosphoproteome Analysis of Capacitated Human Sperm Cells-Identification of phosphorylation substrates using twodimensional gel electrophoresis and immunoblotting is a powerful approach. However, this methodology has several limitations. First, it is difficult to identify proteins that undergo phosphorylation on serine or threonine residues because antibodies against those phosphoamino acids are not sensitive enough to detect most proteins phosphorylated on these residues. Second, MS/MS of proteins obtained from polyacrylamide gels has detection limits several orders of magnitude higher than MS performed on proteins not embedded in gels. Third, the use of this approach strongly suggests that a given protein is phosphorylated on tyrosine residues; nevertheless, a full demonstration requires the use of an independent method (e.g. cross-immunoprecipitation, direct sequencing, mutagenesis analysis). Fourth, although in some cases it is possible to obtain the exact site of phosphorylation of a candidate protein, in general the phosphorylation site remains elusive because of the aforementioned lack of sensitivity. In addition, to determine the site of phosphorylation is, in most cases, a very important goal in a phosphorylation study and the strongest demonstration that a protein is phosphorylated on a particular amino acid.
As shown above, several proteins undergo tyrosine phosphorylation during capacitation. However, no sites of phosphorylation were defined. To understand further the role of phosphorylation in human sperm capacitation, we have analyzed the phosphoproteome of capacitated human sperm. Toward this goal, we have improved the enrichment of phosphopeptides by IMAC converting acidic residues to methyl esters and adapting this technology to the analysis of phosphorylation sites directly from capacitated human sperm total protein extracts.
Phosphorylated peptides were identified by screening MS/MS spectra for an abundant neutral loss of phosphoric acid from the peptide precursor mass. This process is commonly observed in ion trap MS/MS spectra of phosphorylated peptides (19). When peptides (1 mg) were analyzed by this method, more than 200 distinct phosphorylated species were detected. Manual and SEQUEST (18) interpretation of MS/MS spectra led to the identification of 18 sites of Ser phosphorylation and a single site of tyrosine phosphorylation on a total of 7 different proteins (Table II, trypsin). Fragment ions in the spectra allowed unambiguous assignment of the phosphorylation sites. MS/MS spectra recorded on synthetic peptides confirmed the sequence assignment of the phosphotyrosine-containing peptide (data not shown). Although 200 phosphopeptides were detected, a majority of the spectra (Ͼ75%) showed multiple (up to four) losses of phosphoric acid from the precursor mass and were difficult to interpret. Because trypsin digestion can generate phosphopeptides that are too large or too small to be compatible with RP (C18) chromatography, another protease with a different specificity was used. V8 protease (Glu-C) cleaves substrate proteins on the C-terminal side of Asp and Glu. Using this protease combined with IMAC and MS/MS, 40 phosphopeptides were detected. Data base searching and de novo sequencing efforts elucidated new sites of phosphorylation including 11 on Ser, 1 on Thr, and 2 on Tyr (Table II, Glu-C). Fragment ions in spectra allowed phosphorylated residues to be assigned unambiguously. Alternative enzymes (i.e. chymotrypsin) may be necessary to define the entire capacitated human sperm phosphoproteome. No nonphosphorylated sperm peptides were detected in either analysis. This illustrates that conversion of the sample to peptide methyl esters prior to IMAC increases significantly the selectivity of this technique toward phosphorylated peptides.
IMAC Analysis of Immunoprecipitated Protein Digests-

FIG. 1. Analysis of the capacitationassociated increase in protein-tyrosine phosphorylation by two-dimensional gels and immunofluorescence.
A, human sperm were treated immediately (NON) or capacitated overnight (CAP), and sperm proteins were analyzed using two-dimensional gel electrophoresis, transferred to polyvinylidene difluoride membranes, and Western blots were performed with ␣-PY (4G10) and developed as described. The arrow in the right panel indicates the VCP spot. B, different patterns of Tyr phosphorylation observed in human sperm before (NON) and after capacitation (CAP). Human sperm were fixed in paraformaldehyde, dried out, and immunofluorescence was performed as described under "Experimental Procedures." Values represent the mean Ϯ S.E. of three independent experiments. C, silver-stained (left panel) and parallel Western blots (middle panel, 30-s exposure; right panel, 5-min exposure) using ␣-PY. Both images were overlaid using Adobe Photoshop, and silver-stained spots exhibiting Tyr phosphorylation also were cored and sequenced by MS/MS. IMAC/MS analyses of total protein extracts facilitated the discovery of three sites of tyrosine phosphorylation. Western blotting (Fig. 1), however, suggests that many more proteins are tyrosine phosphorylated. In an effort to identify sites of tyrosine phosphorylation selectively, total capacitated human sperm protein extracts were immunoprecipitated with ␣-PY antibodies (clone 4G10). Elution was performed using 1 mM phenyl phosphate to ensure selective elution from the ␣-PY antibody after precipitation with PAS. Immunoprecipitated proteins were digested separately with trypsin or Glu-C and TABLE I Proteins that undergo Tyr phosphorylation during capacitation Proteins were cut out from silver-stained two-dimensional gels and microsequenced using MS/MS. To assign a particular protein to the respective cut band, at least five peptides from a single spot matched the data-base sequence for the assigned protein.
Protein assigned from blast search of sequenced peptides NCBInr. 12.5.2001 Accession no. analyzed by IMAC and MS/MS. De novo and SEQUEST interpretation of the data identified many new sites of phosphorylation, including 24 on Ser, 1 on Thr, 1 on Tyr, and 1 ambiguous site (phosphoserine or phosphothreonine) ( Table II, IP).
Post-IMAC AP Treatment of Phosphopeptides-Many MS/MS spectra recorded during IMAC analysis of sperm peptides had peaks at m/z values corresponding to multiple losses of phosphoric acid from the precursor mass. Such spectra contain few ions indicative of amino acid sequence because amide bond cleavage cannot effectively compete with gas phase dephosphorylation during MS/MS. To assess whether phosphate removal prior to MS analysis would facilitate peptide identification, we converted sperm tryptic peptides to methyl esters. After IMAC enrichment, phosphopeptides were eluted on-line to an AP column before capture on C18 particles and MS/MS. For comparison, a similar analysis was performed without AP treatment. These parallel analyses each detected several peptides (Tables III and IV). The MS/MS spectrum of SVESVK, recorded during analysis of AP-treated peptides, contains peaks at m/z values corresponding to y4, b4, and b5 ions, whereas the phosphorylated analog, pSVEpSVK, does not (Fig.  2, A and B, respectively). This illustrates that phosphate removal prior to MS analysis aids in peptide sequence determination. Although sequences of dephosphorylated peptides themselves cannot reveal where or to what extent a peptide is phosphorylated, complementary information provided by phosphopeptide MS/MS spectra obtained through parallel analysis without AP treatment may provide enough information to assign the phosphorylation sites.
The peptides PLASSPPR and VSGSSQSPPNLK were also detected after AP treatment. These peptides are derived from PLApSpSPPR and VSGpSSQpSPPNLK observed in analyses of trypsin-digested sperm proteins (see Table II). Not all phosphorylated peptides detected without AP treatment were detected after AP treatment. This is probably the result of incomplete elution of phosphopeptides from the IMAC column in the AP experiment because phosphate was not included in the IMAC elution buffer to avoid inhibition of the AP enzymatic activity. In contrast, the use of the AP column allows analysis of peptides not amenable to positive ion MS in their phosphorylated form. For example, the peptide SPSAP-PAKPPSTQR, detected during MS analysis of dephosphorylated sperm peptides (Table IV), was not found in samples without AP treatment. Retention of this peptide by IMAC indicates that this stretch of amino acids was phosphorylated within the parent protein (AKAP-4). The phosphorylated form of this peptide may not ionize well in the positive ion mode, preventing the identification of this peptide during analysis performed without the use of AP. These results suggest that dephosphorylation of phosphopeptides after IMAC enrichment prior to MS analysis is a useful tool to derive sequences of multiple phosphorylated peptides.
Quantitative Phosphorylation Analysis: Protein Standards-We have identified 56 sites of phosphorylation in several capacitated human sperm proteins. Although this information describes phosphorylation in sperm on a global scale, it would be useful to discern sites of phosphorylation induced during capacitation. This goal could be achieved by comparison of the ratio of any particular phosphopeptide present in digests of proteins from capacitated and noncapacitated sperm cells. To adapt IMAC/MS methodology to display phosphopeptides differentially from different sperm capacitated states, we utilized an isotopic labeling strategy (Fig. 3A). Tryptic peptides from two samples of cells are converted to peptide methyl esters with deuterated (d 3 ) and nondeuterated (d 0 ) methanol, respectively. Both samples are then mixed in equal proportions and the mixture purified by IMAC to ensure that only phosphopeptides were retained. Signals for phosphopeptides present in both samples appear as doublets separated by n(3 Da)/z (where n is the number of carboxylic acid groups in the peptide and z is the charge on the peptide). The ratio of the two signals in the doublet changes as a function of expression level of the particular phosphoprotein in each sample. Peptides of interest are then targeted for sequence analysis subsequently performed on the ion trap instrument.
To validate this approach, we used ␤-casein because tryptic   digestion of this protein produces a phosphorylated peptide, FQpSEEQQQTEDELQDK, which is detectable by positive ion electrospray. Briefly, 50 and 250 pmol of tryptic peptides from the phosphorylated protein ␤-casein was converted to d 0 -peptide methyl esters, and 500 pmol of tryptic peptides from the same protein was converted to d 3 -methyl esters. Equal portions were combined to produce known concentrations of the d 0 -and d 3 -tryptic peptide methyl esters from ␤-casein and analyzed by MS after IMAC enrichment. Because the phosphorylated ␤casein peptide contained seven free carboxyl groups, the mass difference between nondeuterated (d 0 ) and deuterated (d 21 ) analogs was 21 Da. Mass spectrometry of these samples revealed d 0 :d 21 ratios within 11% of the actual ratio in two separate experiments, indicating that this method is useful for the quantitation of phosphopeptides (Fig. 3B). Subsequent MS/MS confirmed the sequence of the peptides to be the deuterated and nondeuterated FQpSEEQQQTEDELQDK. The fact that deuterated and nondeuterated peptides nearly coelute (deuterated forms having a slightly lower retention time) coupled with the accurate mass measurements (typically Յ20 ppm in our experiments) of FT-ICR allows for quick correlation of related species. The number of carboxyl groups can be calculated by Equation 1, where ⌬mass is the mass difference between deuterated and nondeuterated isotopic distributions of the same charge state.
No. of carboxyl groups ϭ (⌬ mass ϫ charge state)/3 (Eq. 1) The information gained by FT-ICR analysis of phosphopeptides (accurate mass and number of carboxyl groups) can be coupled with the complementary information provided by datadependent or targeted ion trap MS/MS analyses (fragment ions, sequence tags, and minimum number of phosphorylated residues from neutral losses of phosphoric acid) to identify peptide sequences from protein data bases rapidly.
To identify capacitation-associated sites of phosphorylation, 2 ϫ 10 8 human sperm were separated in equal aliquots, one aliquot was immediately used, the remaining sperm aliquot was capacitated overnight, and then protein extracts were obtained. Because sperm are unable to synthesized proteins de novo and both samples contained the same amount of sperm, quantitation of phosphopeptides in each sample reflected changes in phosphorylation which occurred during capacitation. Both noncapacitated and capacitated human sperm extracts were then digested with trypsin, and the resulting peptides were converted to the corresponding peptide methyl esters using d 0 and d 3 methanol, respectively. Peptide pools (derived from capacitated and noncapacitated digests) were then mixed in equal proportions, and an aliquot was analyzed by IMAC/RP-HPLC/ESI/MS on a home-built FT-ICR instrument. The data were examined manually to identify singlet peaks, i.e. phosphopeptide species unique to the capacitated or noncapacitated sample. Parallel IMAC/RP-HPLC/ESI/MS(/ MS) experiments on a quadrupole ion trap mass spectrometer were used to define the sequences of these peptides.
Using this method, the peptide pSVEpSVK from a novel Ca 2ϩ -binding protein, CABYR, was present in the capacitated and noncapacitated sperm total protein digests at about the same level (Fig. 3C). In contrast, the peptide INApSTDpSLAK, derived from AKAP-4, was found at a level 23 times greater in capacitated sperm digests (Fig. 4C), indicating a capacitationdependent phosphorylation of this peptide. The charge state of these peptides can be determined by the difference in mass between the C12 and C13 isotope peaks observed by MS analysis. Because the actual mass difference between these isotope peaks is 1 Da, the observed mass difference of 0.5 Da between peaks in the isotopic envelopes of pSVEpSVK (Fig. 3D) and INApSTDpSLAK (data not shown) indicates that both peptides were double charged. Using Equation 1, it was deduced that both phosphopeptides (pSVEpSVK and INApSTDpSLAK) possessed two free carboxyl groups because double charged isotopic envelopes corresponding to deuterated and nondeuterated analogs were separated by 3 Da (Fig. 4D and data not shown). Total ion chromatograms were obtained by plotting the sum of the intensities of ions in mass spectra versus time (data not shown). Single ion chromatograms were acquired by plotting the sum of the intensities of ions within a small mass window (i.e. 418.67 Ϯ 0.1 Da, versus time) derived from MS analysis of sperm tryptic peptide methyl esters using an FT-ICR mass spectrometer (Fig. 3E).
Approximately 500 peptide species were observed in FT-ICR MS analysis of IMAC-enriched modified peptides. Most of these species were observed as doublets, indicating similar levels of phosphorylation between capacitated and noncapacitated sperm populations; 20 unique species were differentially phosphorylated in these populations. Most of these spectra showed neutral losses characteristic of phosphopeptides, however, pep-

FIG. 3. Quantitation of phosphorylated peptides in protein digests.
A, schematic representation of the procedure for phosphopeptide quantitation. B, ratios of deuterated and nondeuterated peptide FQp-SEEQQQTEDELQDK present in trypsin digests of ␤-casein. QIT, quadrupole ion trap. C, capacitated and noncapacitated sperm total protein extracts measured using the approach described above. D, mass spectrum recorded during coelution of deuterated and nondeuterated pS-VEpSVK peptides. E, single ion chromatograms (SIC; obtained by plotting the sum of the intensities of ions within a small mass window, i.e. 418.67 Ϯ 0.1 Da, versus time) derived from MS analysis of differentially methyl ester-modified tryptic sperm peptides using an FT-ICR mass spectrometer. Double charged ions corresponding to deuterated (d 6 ) and nondeuterated (d 0 ) forms of peptides P1 (pSVEpSVK) and P2 (INAp-STDpSLAK) were observed. Note that both peptides contain two carboxyl groups so that derivatized analogs differ in mass by 6 Da. Peptide sequences were derived from parallel MS/MS experiments performed on an ion trap mass spectrometer. tide sequences could not be derived. Further experiments will define additional phosphopeptides unique to the capacitated total protein digests.
Valosin-containing Protein (VCP/p97) Is Tyr Phosphorylated and Changes the Immunofluorescence Pattern during Capacitation-Because capacitation prepares sperm to undergo the acrosome reaction, a form of regulated exocytosis, phosphorylation of proteins involved in fusion events may regulate this process and is of particular interest for further investigation. VCP, also known as p97 (VCP/p97), has been implicated in several fusion events in yeast and mammalian cells (21,22). This protein appears to undergo tyrosine phosphorylation during capacitation (Fig. 1C and arrow in Fig. 1A). However, we were unable to characterize the VCP site of tyrosine phosphorylation. Several possibilities could explain the lack of phosphotyrosine phosphopeptides from VCP. First, the VCP tyrosine phosphorylation site could be masked by the complexity of the MS/MS spectra. Second, although the global analysis of phosphorylation sites is a very powerful approach, it has limitations. One relevant limitation is that abundant phosphopeptides will be sequenced at higher rates, masking the ability of the method to detect nonabundant phosphopeptides. This limitation also explains why we have detected more phosphorylated Ser residues than Tyr residues. Third, it is also possible that the tyrosine-phosphorylated sequence of VCP is present in a tryptic peptide that is not detectable with this methodology.
As an alternative, to confirm that VCP is tyrosine phosphorylated during capacitation we used an independent approach. An ␣-VCP (S), donated by Dr. Samelson (23) was used to analyze the presence of this protein after ␣-PY immunoprecipitation of noncapacitated as well as of capacitated sperm protein extracts. As described previously (20,24), ␣-PY Western blots from total extracts show that there is an increase in tyrosine phosphorylation after human sperm capacitation (Fig. 4A, left  panel). 20 million sperm of each population were extracted in a SDS 0.1% buffer as described under "Experimental Procedures" and immunoprecipitated using ␣-PY (clone 4G10). The immunoprecipitates were then probed with either ␣-PY (RC20) or ␣-VCP (S) by Western blot (Fig. 4B, left two panels). Because RC20 is already labeled with peroxidase, it does not need a secondary antibody avoiding detection of IgG in the immunoprecipitates. This experiment shows that VCP was more abun-dant in the ␣-PY immunoprecipitates from the capacitated population, confirming the tyrosine phosphorylation of this protein during the capacitation process observed using twodimensional gel analysis (Fig. 1A, see arrow in right panel). A similar experiment was performed to confirm tyrosine phosphorylation of AKAP-3. In this case, 2 ϫ 10 7 of either noncapacitated or capacitated human sperm were extracted and immunoprecipitated with ␣-AKAP-3 (13). The immunoprecipitates were then probed with ␣-AKAP-3 and with ␣-PY (RC20) (Fig. 4B, right panels). As predicted, the AKAP-3 signal is similar in both immunoprecipitates from noncapacitated or capacitated sperm. However, the ␣-PY signal is higher in the capacitated population confirming that this protein is tyrosine phosphorylated during capacitation.
Because VCP/p97 has been shown to change subcellular localization after phosphorylation (25), the localization of VCP/ p97 in human sperm was analyzed before and after capacitation. To evaluate the localization of VCP/p97, we have used two antibodies from independent sources, ␣-VCP (S), donated by Dr. Samelson (23), and ␣-VCP (T), donated by Dr. Tonks (26). Both antibodies recognized a single protein by Western blot (Fig. 4A). Using these antibodies, it was possible to observe that before capacitation VCP/p97 localized to the neck region of human sperm. However, after overnight capacitation, the appearance of fluorescent staining in the anterior head of human sperm was observed (Fig. 5). It is noteworthy that the signal in the neck decreased at the same time that the signal in the anterior head increased. In contrast, another protein, SP-10, which is present in the sperm acrosomal matrix, localized in the anterior head in capacitated as well as in noncapacitated human sperm (Fig. 6). Similarly, AKAP-3 localized in the flagellum in both capacitated and noncapacitated human sperm (Fig. 6). Because capacitation prepares the sperm to undergo a ligand-dependent exocytosis and VCP/p97 has a role in membrane fusion events in other biological systems, our findings suggest that the regulation of VCP/p97 might be a link between capacitation and the acrosome reaction. DISCUSSION The physiological changes that render mammalian sperm able to fertilize are collectively known as capacitation. Capacitation has been correlated with the increase in tyrosine phos- phorylation of several proteins (1). With the exception of two members of the AKAP family (13,27) and CABYR (28), proteins that undergo tyrosine phosphorylation during capacitation have not yet been characterized. Capacitation prepares the sperm to undergo the acrosome reaction and also is associated with changes in sperm motility (e.g. hyperactivation) in a number of species (1). Therefore, one may postulate that components of the sperm exocytotic and motility machinery are modified during capacitation (e.g. phosphorylation of specific proteins, changes in protein localization, and/or modification of protein-protein interactions). In the present work, we have analyzed the phosphoproteome of capacitated human sperm using 1) two-dimensional gels followed by ␣-PY Western blot and 2) direct MS/MS sequencing of phosphopeptides in total protein extracts.
To understand the link between capacitation and the acrosome reaction, an increased knowledge of the mechanisms that regulate this exocytotic event in sperm is necessary. Sperm homologs of SNARE (29) as well as SNARE-associated proteins have been detected in sea urchin (30) and mammalian sperm (31,32). These observations support the idea that the sperm acrosome reaction might be regulated in ways similar to the exocytotic processes in somatic cells. Among the proteins that undergo tyrosine phosphorylation during capacitation, we identified VCP/p97 in this study. VCP/p97, a member of the AAA family (ATPases Associated with various cellular Activities) (33), along with NSF and the Golgi t-SNARE syntaxin 5, mediate the fusion of Golgi membranes (21,34). Although VCP/p97 and NSF are highly homologous, they appear to act in distinct fusion events, presumably because of additional specific cofactors (35). VCP/p97 has a role as a chaperone and aids in the assembly, disassembly, and functional operation of protein complexes. VCP/p97 undergoes tyrosine phosphorylation during T cell activation, and although this phosphorylation did not alter its ATPase activity (23), tyrosine phosphorylation regulates the subcellular localization of this protein (25). Moreover, a membrane fusion process such as the transitional endoplasmic reticulum assembly in vitro requires tyrosine phosphorylation of VCP/p97 (36). We have demonstrated that VCP/ p97 undergoes tyrosine phosphorylation during capacitation. In addition, we have shown that prior to capacitation VCP/p97 localizes in the neck of human sperm, whereas after overnight incubation, immunofluorescence experiments showed a decreased staining in the neck and the appearance of this protein in the anterior head of capacitated sperm. At least three hypotheses can be made to explain these different immunofluorescent patterns. First, modifications of the sperm during capacitation allowed ␣-VCP antibodies to enter the anterior head of the sperm. Although possible, experiments showing that SP-10 can be recognized before and after capacitation added to the observation of a decrease in fluorescent staining in the neck of capacitated sperm argue against this explanation. Second, the epitopes recognized by two ␣-VCP antibodies are unmasked in the anterior head, but they are masked in the neck of capacitated sperm. Third, there is a translocation of VCP dur- FIG. 5. Immunolocalization of VCP/p97 in human sperm before and after capacitation. Left panels show the immunofluorescence (IF) of air-dried human sperm at ϫ40 magnification using two different rabbit anti-VCP/p97 antibodies, ␣-VCP (S) and ␣-VCP (T), before (Non) and after (Cap) capacitation. The antibodies were visualized using FITC ␣-rabbit secondary antibody as described under "Experimental Procedures." Right panels are the corresponding bright fields. Insets in the upper right bright fields show the ␣-VCP immunofluorescence in either noncapacitated or capacitated sperm at higher magnification (ϫ100). Controls were performed using normal rabbit serum.
FIG. 6. Immunolocalization of AKAP-3 and SP-10 in human sperm before and after capacitation. Left panels show the immunofluorescence (IF) of air-dried human sperm using a rat ␣-AKAP-3 (ϫ40) antibody or a mouse monoclonal ␣-SP-10 (ϫ100) antibody, ␣-VCP (S) and ␣-VCP (T), before (Non) and after (Cap) capacitation. The antibodies were visualized using FITC ␣-rat or FITC ␣-mouse secondary antibody as described under "Experimental Procedures." Right panels are the corresponding bright fields. Controls were performed using normal rat serum or normal mouse serum and showed no immunofluorescence staining (data not shown).
ing capacitation from the neck to the anterior head of human sperm. Ideally this hypothesis should be tested by a direct measurement of VCP in the neck and the anterior head after separation of these subcellular compartments; however, the impossibility of separating the neck from the anterior head prevented us from performing this experiment. Considering the role of VCP/p97 and other members of the AAA family of phosphatases in fusion processes, these results suggest that VCP/p97 and tyrosine phosphorylation of this protein could have a role as a link between capacitation and the acrosome reaction. Alternatively, this protein can act as a chaperone, bringing relevant membrane fusion proteins to the site of the acrosome reaction.
Capacitation is also linked to events that occur in the sperm flagellum. AKAPs represent a growing family of scaffolding proteins that function to tether the regulatory subunits of protein kinase A and other enzymes to organelles or cytoskeletal elements. These proteins permit the precise control of signal transduction in discrete regions of the cell (37). In the present work, we have confirmed tyrosine phosphorylation of AKAP-3 and AKAP-4 during human sperm capacitation and mapped eight phosphorylation sites of these proteins. Among these sites, the AKAP-4 phosphopeptide INApSTDpSLAK was found to be 23 times more abundant in capacitated sperm, suggesting that this phosphorylation site might be involved in the regulation of AKAP-4 function during capacitation.
Multiple methodologies have been used to study phosphorylation. Recently, mass spectrometry has become the preferred method to identify phosphopeptides because of its speed and high sensitivity. Nevertheless, because phosphorylation sites are usually substoichiometric, it has been necessary to devise new methods for selective detection and enrichment of phosphopeptides. Recently, although several methods have been developed, most have been applied only on a protein-by-protein basis (8) with a few notable exceptions (38,39). None of these methods has been used successfully to identify phosphotyrosine residues from complex mixtures probably because of multiple steps of derivatization that reduced the final recovery. On the other hand, IMAC has been used previously to enrich digests for phosphorylated peptides (6,16); however, the selectivity of this technique is poor because peptides containing acidic residues (i.e. Glu and Asp) bind to the immobilized iron atoms (40). To solve this lack of specificity, acidic residues were converted to methyl esters, eliminating the binding of nonphosphorylated species to the IMAC column. This procedure does not generate diasteromers and is compatible with phosphorylated Ser, Thr, and Tyr residues. This methodology was used successfully to map 60 sites of phosphorylation.
The sites of phosphorylation identified here defined in vivo sites of phosphorylation resulting from normal phosphorylation events in capacitated human sperm and were not obtained by kinase overexpression or by kinase activators or phosphatase inhibitors. Although most of the sites described in the present work have not been observed previously, phosphorylation of the C-terminal protein kinase A catalytic subunit peptide IRVpSINE has been demonstrated to be necessary for the catalytic activity of recombinant protein kinase A (41). Another observation derived from the capacitated human sperm phosphoproteome is that four of the five tyrosine-phosphorylated sequences contained phosphoserine in the proximity of the phosphotyrosine. This result raises the possibility that a dual specificity kinase is involved in the capacitation process and/or that phosphoserine is part of the substrate recognition motif for a sperm tyrosine kinase. Because the identity of tyrosine kinases present in sperm is at present not known, peptide sequences found to be tyrosine-phosphorylated can be used as substrates to purify tyrosine kinases from sperm.
After enrichment of phosphorylated peptides, the next step is to derive their sequences and define the sites of phosphorylation. In many cases, identification of the exact phosphorylation sequence was precluded because of the complexity of the phosphopeptide MS/MS spectra. To simplify interpretation of phosphopeptide spectra, we performed on-line AP dephosphorylation after IMAC phosphopeptide enrichment. AP treatment has been shown to enhance detection of multiply phosphorylated peptides (42). Analysis of phosphopeptides followed by AP treatment and reanalysis has been demonstrated as a technique for phosphopeptide identification based on 80-Da differences (43). This method allowed identification of one additional peptide that was not detected without AP treatment.
In summary, 16 proteins that are recognized by ␣-PY antibodies in two-dimensional Western blots were determined. From these proteins, VCP/p97 was also detected in ␣-PY immunoprecipitates and showed different immunolocalization patterns before and after capacitation. In addition, we have successfully employed IMAC enrichment of phosphopeptides coupled to MS/MS analysis to isolate and sequence phosphopeptides from total protein digests of human capacitated sperm. This is the first study in which this methodology has been used to map phosphopeptides in a human cell type. Moreover, we have coupled this methodology with differential isotopic labeling and IMAC enrichment to derive quantitative information on phosphorylation events that occur during capacitation. This technique goes beyond the field of reproductive biology and could potentially be used to map and compare sites of phosphorylation in a variety of biological systems.