New Insights into the Rat Spermatogonial Proteome

Despite the essential role played by spermatogonia in testicular function, little is known about these cells. To improve our understanding of their biology, our group recently identified a set of 53 spermatogonial proteins using two-dimensional (2-D) gel electrophoresis and mass spectrometry. To continue this work, we investigated a subset of the spermatogonial proteome using narrow range immobilized pH gradients to favor the detection of less abundant proteins. A 2-D reference map of spermatogonia in the pH range 4–9 was created, and protein entities fractionated in a pH 5–6 2-D gel were further processed for protein identification. A new set of 156 polypeptides was identified by peptide mass fingerprinting and tandem mass spectrometry. These polypeptides corresponded to 102 different proteins, which reflect the complexity of post-translational modifications. Seventy-nine of these proteins were identified for the first time in spermatogonia. All identified proteins were classified into functional groups. This work represents a first step toward the establishment of a systematic spermatogonia protein database.

In mammals, spermatogenesis takes place in the seminiferous tubules of the testis and is the process in which the germ line stem cells named spermatogonia become mature functional spermatozoa. This process includes stem cell renewal, various steps of germ cell proliferation and differentiation, and germinal apoptosis events. Spermatogonia initially undergo successive mitotic divisions, leading them to differentiate into spermatocytes. The spermatocytes then divide by meiosis, giving rise to haploid spermatids. Finally the spermatids are matured into spermatozoa in a process called spermiogenesis. These events are regulated by different molecular and cellular mechanisms, which involve hormones (1), paracrine communications between the different testicular cell components located within the seminiferous tubules, includ-ing the Sertoli cells (2,3), and the expression of specific genes by the germ cells themselves (4,5).
Although a number of laboratories have recently focused on spermatogonia (6 -9), the biology of these cells remains largely unknown. Complex techniques are required to isolate and enrich them (10,11), and it is very difficult to culture them (6,12). However, several groups have identified a number of genes and proteins characterizing spermatogonia in different mammals (13,14). In the latter study (14), we combined two-dimensional electrophoresis (2-DE) 1 and mass spectrometry to investigate the protein content of spermatogonia. This led to the identification of 53 spermatogonial polypeptides, two of which were further studied in detail within the testis because of their possible implication in proliferation and differentiation events (translationally controlled tumor protein (15) and stathmin (16)).
The aim of the present study was to provide additional information concerning the proteome of spermatogonia by using narrow range 2-DE (zooming gels) that improve the detection of low abundance proteins (17)(18)(19). We have established a spermatogonial proteome map from pH 4 -9 with different narrow range immobilized pH gradient (IPG) gels (i.e. 4 -5, 4.5-5.5, 5-6, 5.5-6.7, and 6 -9) and decided as a first step to focus on the 5-6 pI range map. We report here the identification of 156 additional polypeptides by mass spectrometry and searches of protein sequence databases. The identified proteins have been classified according to their main known/postulated functions and these data, together with those obtained in our previous study (14), will be used to create a Web-based proteomic database of the rat testis (TESTIS-2DPAGE).

EXPERIMENTAL PROCEDURES
Isolation of Spermatogonia-Seminiferous epithelial cells from 9-day-old male Sprague-Dawley rats (Elevage Janvier, Le Genest Saint Isle, France) were separated as described previously (20) with the minor modifications suggested by Dym et al. (21). The decapsu-lated testes were resuspended in phosphate-buffered saline (PBS) containing collagenase (1 mg/ml) and DNase (1 g/ml) and were incubated at 32°C for 15 min in an orbital incubator with shaking (100 rpm). The resulting small fragments were washed twice in PBS and incubated in PBS containing collagenase (1 mg/ml), hyaluronidase (1.5 mg/ml), trypsin (0.5 mg/ml), and DNase (1 g/ml) for 25 min at 32°C with shaking (100 rpm). The dispersed cells were washed three times with PBS and filtered through nylon meshes with 80-and 40-m pores. Cells from the dissociated tubules were separated by sedimentation velocity at unit gravity at 4°C using a 2-4% (w/v) bovine serum albumin (BSA) gradient in Ham's F12/Dulbecco's modified Eagle's medium. In brief, the cells were bottom-loaded into an SP-120 chamber (STAPUT) in 30 ml of Ham's F12/Dulbecco's modified Eagle's medium containing 0.5% BSA, and a gradient was simultaneously generated using 275 ml of each of two media, one supplemented with 2% BSA and the other with 4% BSA. The cells were allowed to sediment for a standard period of 2.5 h, and 35 fractions of 15 ml were then collected from the bottom of the gradient. Fractions 16 -21, corresponding to 2.8 -3% BSA, were pooled, washed several times with PBS, and centrifuged at 100 ϫ g for 10 min. The spermatogonia in the enriched preparation were counted, and an aliquot of each cell suspension was stained with 0.2% trypan blue in PBS for viability assessment. Purity of the cell suspension was assessed by transmission electronic microscopy (14), which showed that it contained a mean of 95% spermatogonia (80% type B and 15% type A spermatogonia) and an average of 5% Sertoli cell fragments, peritubular cells, and Leydig cells (Fig. 1). The cell pellets were dried and stored at Ϫ80°C until protein extraction.
Protein Extraction-Spermatogonia pellets were dissolved in a 20 mM HEPES buffer, pH 7.5, containing 1 mM EDTA, 0.5 mM DTT, 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, and 10 M trans-epoxy-succinyl-L-leucylamido-(4-guanidino) butane. The suspension was homogenized by sonication on dry ice. The sample was centrifuged at 15,000 ϫ g for 30 min at 4°C, and the supernatant was ultracentrifuged at 105,000 ϫ g for 1 h at 4°C. The resulting supernatant was desalted and concentrated using a Centricon YM-10 centrifugal filter device (cut-off, 10 kDa; Millipore, Saint Quentin en Yvelines, France). The protein concentration of the concentrate was determined by the bicinchoninic acid assay (Pierce) according to the manufacturer's instructions.
Sample Preparation and 2-DE-Prior to electrophoresis, spermatogonial proteins were incubated in a rehydration buffer containing 6 M urea, 2 M thiourea, 1% DTT, 4% CHAPS, 0.5% IPG Buffer (Amersham Biosciences), and bromphenol blue for 1 h at room temperature. For the 6 -9 IPG strips, a modified rehydration buffer (7 M urea, 2 M thiourea, 2.5% DTT, 4% CHAPS, 10% isopropanol, 5% glycerol, 2% IPG Buffer, and bromphenol blue) was used (22). The first dimension was performed using an IPGphor isoelectric focusing apparatus for the 3-10NL, 4.5-5.5, 5-6, and 5.5-6.7 18-cm IPG strips or a Multiphor II electrophoresis unit for the 4 -5 and 6 -9 18-cm IPG strips according to the manufacturer's instructions (Amersham Biosciences). The conditions used for rehydration and isoelectric focusing are summarized in Table I. For the 6 -9 IPG strips, a "paper bridge," which had been soaked with the modified rehydration buffer containing 3.5% DTT, was applied at the cathode (22). After isoelectric focusing, the IPG strips were equilibrated for 15 min at room temperature in an equilibration solution (0.05 M Tris-HCl (pH 8.8), 6 M urea, 30% glycerol, 2% SDS, and bromphenol blue) containing 65 mM DTT. The strips were equilibrated a second time in the same solution containing 250 mM iodoacetamide and a few grains of bromphenol blue. The second dimension was then performed at 15°C with a flat bed Multiphor II electrophoresis unit (Amersham Biosciences) using a precast ExcelGel XL SDS-polyacrylamide 12-14% gradient gel (Amersham Biosciences) with precast anode and cathode buffer strips (Amersham Biosciences). The electrophoresis conditions were as described previously (23): 150 V, 20 mA for 1 h following by 800 V, 40 mA for 3 h.
Gel Staining-For image analysis, 2-D gels were silver-stained as described previously (24). After staining, gels were scanned at 300 dpi resolution, and the digitized images were analyzed using 2D Elite Image Master software (Amersham Biosciences).
For protein identification, 2-D gels were silver-stained according to a specific protocol compatible with in-gel digestion of proteins and mass spectrometry (25). In brief, gels were fixed in 50% methanol, 5% acetic acid for 20 min, washed for 10 min in 50% methanol, and then thoroughly washed in MilliQ water for 2 h. Gels were sensitized for 1 min in 0.02% Na 2 S 2 O 3 and washed twice for 1 min in MilliQ water. Gels were then silver-stained in ultragrade AgNO 3 for 20 min at 4°C, developed in 0.04% formaldehyde, 2% Na 2 CO 3 , and incubated in 5% acetic acid to stop staining. Gels were stored at 4°C in 1% acetic acid until in-gel digestion.
In-gel Digestion-Silver-stained protein spots were excised from 2-D gels and digested as described previously (26) with minor modifications. In brief, gel pieces were washed twice in MilliQ water, dehydrated in acetonitrile, and dried in a vacuum centrifuge. Gel pieces were then rehydrated at 4°C for 45 min in a digestion buffer containing 50 mM NH 4 HCO 3 and 6.7 ng/l trypsin (modified, sequencing grade, Promega, Charbonniè res, France). The supernatant was then replaced by 35 l of 50 mM NH 4 HCO 3 , and the samples were incubated overnight at 37°C.
MALDI-TOF MS-The digested peptides contained in the supernatant were purified using a home-made micropurification column containing 0.1 l of a slurry consisting of 20R2 reversed-phased material (PerSeptive Biosystems, Framingham, MA) packed in a Ge-Loader tip (Eppendorf, Hamburg, Germany) and equilibrated with 1% trifluoroacetic acid as described previously (27). Ten microliters of the mixture were loaded onto the column, and after washing with 1% trifluoroacetic acid, adsorbed peptides were eluted directly on the MALDI target with 0.8 l of 70% acetonitrile, 0.1% trifluoroacetic acid containing saturated ␣-cyano-4-hydroxycinnamic acid. MALDI-TOF mass spectra were acquired using a Reflex II Instrument (Bruker, Bremen, Germany), processed using the MoverZ software (Proteo-Metrics, New York, NY), and internally calibrated using trypsin autodigestion peptides.
ESI-MS/MS-Digested proteins were purified using a home-made micropurification column containing 0.3 l of a slurry consisting of 20R2 reversed-phased material (PerSeptive Biosystems) packed in a GeLoader tip (Eppendorf) and equilibrated with 5% formic acid. About 25 l of the peptide samples were loaded onto the column, washed with 5% formic acid, and directly eluted by brief centrifugation in a medium nanoelectrospray needle (nanoelectrospray capillaries, Protana, Odense, Denmark) with 0.8 -1.2 l of 50% methanol, 0.1% formic acid. Tandem mass spectra were acquired with a Q-TOF Ultima mass spectrometer (Micromass, Manchester, UK) and processed using MassLynx software.
Protein Identification-Protein identification by mass spectrometry was performed as described previously (28). Monoisotopic masses of tryptic peptides observed in the MALDI-MS spectra were used to query NCBInr and SWISS-PROT sequence databases using Mascott (www.matrixscience.com) (29) and Profound (prowl.rockefeller.edu/ cgi-bin/ProFound) (30) search programs. Search conditions used were as follows: initial rather open mass window of 70 ppm for an internal calibration and 200 ppm for an external calibration, one missed cleavage allowed, modification of cysteines by iodoacetamide, and methionine oxidation and N-terminal pyroglutamylation as variable modifications. To avoid incorrect identifications, each matching was carefully checked manually as described by other groups (28,31) by considering the mass accuracy of matched peptides (i.e. higher than 50 ppm for an internal calibration and higher than 150 ppm for an external calibration (see Table II)) and the missed cleavage motifs that are frequently observed as described previously (32). Furthermore the peptides with N-terminal glutamine or with oxidized methionine were only taken into account if, respectively, the pyroglutamylated peptide or the non-oxidized peptide was also matched. Cross-species identifications were allowed for mammalian proteins only when all the conditions described above were gathered and when rat homologous proteins were not present in databases at the time the searches were performed. When peptide mass fingerprinting gave no or ambiguous results, sequencing by tandem mass spectrometry was performed. MS/MS spectra were used to query NCBInr and SWISS-PROT sequence databases using the MS/MS ion search tool of Mascott program (www.matrixscience.com) (29). Moreover peptide sequence tags were manually designed from MS/MS spectra (33) and analyzed with the Pepsea sequence tag program (pepsea.mdsproteomics.com). For validating an identification by peptide sequence tags, we checked that 1) the identified peptide contained the sequence tag, 2) it was terminated by a tryptic cleavage site, and 3) it matched with the whole MS/MS spectra (28). To ascertain unambiguous identification, a minimum of three different sequenced peptides was required for one protein. If less than three peptides could be sequenced, masses corresponding to other tryptic peptides were required in the MALDI-MS spectra.

Analysis of the Cytosolic Spermatogonial Proteome by 2-DE-
To investigate the cytosolic spermatogonial proteome, we performed 2-D gels using broad range (3-10NL) and overlapping narrow range (i.e. 4 -5, 4.5-5.5, 5-6, 5.5-6.7, and 6 -9) IPGs for the first dimension (Fig. 2). The spots detected by image analysis in each silver-stained narrow range gel and in the corresponding regions of the broad range gel are presented in Fig. 3. Around 700 protein spots were detected in the pH 4 -9 region of the 3-10NL 2-D gel, whereas about 1750 distinct spots could be detected on the combined narrow range IPG gels when the overlap was excluded so that every protein was only referenced once (Fig. 3). Analysis of the 4 -5 2-D gel revealed the presence of around 400 spots, whereas only 170 could be detected in the same region of the 3-10NL 2-D gel. For the 4.5-5.5 2-D gel, 640 different spots were resolved and detected compared with 310 in the corresponding region of the 3-10NL 2-D gel. Similarly around 850 spots were revealed in the 5-6 2-D gel, and only 310 were revealed in the corresponding region of the 3-10NL 2-D gel. Four hundred distinct spots were counted in the 5.5-6.7 2-D gel compared with only 290 in the corresponding region of the 3-10NL 2-D gel. Concerning the basic region, which is the most difficult to separate, we detected 500 protein spots in the 6 -9 2-D gel and only 220 in the corresponding region of the 3-10NL 2-D gel. In summary, narrow range IPG gels allowed us to visualize between 1.3-and 2.7-fold more spots than classical broad range IPG gels. PH 5-6 Spermatogonial Proteome-As the 5-6 2-D gel contained the highest number of spots, we decided to further analyze this gel. We analyzed 230 of the 850 spots detected with peptide mass fingerprinting and/or sequencing by tandem mass spectrometry. In total, 156 polypeptides were successfully identified, which corresponds to 18% of the proteins on the whole 5-6 2-D gel. The polypeptides identified are listed in Table II and are annotated in the 2-D gel map presented in Fig. 4. The 156 identified proteins corresponded to 102 different polypeptides. Indeed 20 of the identified proteins were represented by at least two distinct spots on the gel (for example, adenosine kinase spots EC10 and EC11), and some of them were represented by as many as eight different spots (enolase 1␣ spots EC2, EC3, EC34, EC103, EC 194, EC200, EC203, and EC240; actin ␤ spots EC101, EC202, EC213, EC214, EC215, EC230, EC231, and EC236). Five of the spots analyzed corresponded to more than one single polypeptide according to tandem mass spectrometry and peptide mass fingerprints results (e.g. spots EC14, EC121, EC159, EC197, and EC198). It is also noteworthy that, for several polypeptides, the analyzed spot contained only a fragment of the identified protein. In other words, the relative mass of the protein as calculated from the 2-D gel was below the expected mass of this protein as calculated from its amino acid composition. For 30 of these proteins, it was possible to

Proteome Analysis of Rat Spermatogonia
identify fragments because the peptides detected by MALDI-TOF MS covered only a partial sequence in the N or C terminus of the protein. For one protein (spot EC106), MALDI-TOF MS analysis gave a significant result, but the most important peak of the spectrum did not match with any peptides of the corresponding protein. Tandem mass spectrometry revealed that this peptide corresponded to the N ␣ -terminal acetylated peptide with a mass increase of 42 Da of the first amino acid of the sequence (data not shown). Further analysis of the identified proteins revealed that, except for the ubiquitously expressed actin and tubulin, 44 proteins had already been detected in the testis, and among these, 23 have been shown to be present in spermatogonia, including the 19 proteins previously identified in the broad range 3-10NL 2-D gel such as stathmin and the phosphatidylethanolamine-binding protein (14) (Table III). Functional Analysis of the Identified Proteins-We then classified the identified polypeptides with regard to their main known/postulated function. Polypeptides were divided into the following categories: primary metabolism, protein synthesis and processing, cell structure, cell defense and detoxification, regulation of gene expression, signal transduction, calcium-binding proteins, vesicular transport, DNA replication, other, and unknown proteins (Table II). Information concerning the putative functions of these proteins was found in the SWISS-PROT and GenBank TM databases or in the literature. The distribution of the identified proteins in these functional groups is shown in Fig. 5. The most abundant class of proteins was that related to primary metabolism (39.7% of the identified polypeptides), which comprised enzymes involved for example in various metabolic pathways such as glycolysis (enolase 1␣, phosphoglycerate kinase testis-specific, etc.) and fatty acid biosynthesis (fatty-acid synthase, heart fatty acid-binding protein, etc.). A large number of identified proteins were involved in protein synthesis and processing (18%) and cell structure (17.3%). The protein synthesis and processing group included polypeptides involved in protein synthesis (eukaryotic translation initiation factor 3 and eukaryotic translation elongation factor 1␥), proteins that control protein folding (chaperonin containing the TCP-1␤ subunit, chaperonin subunits 5 and 7, etc.), and proteins involved in degradation processes (bendless protein, proteasome ␤-chain precursor, ubiquitin-activating enzyme E1X, etc.). The cell structure group contained polypeptides related to the cytoskeleton such as architectural proteins (actin ␤, tubulin ␣, vimentin, etc.) and proteins that regulate the stability of the polymers made by these molecules (cofilin, profilin, stathmin, etc.). The functional groups that contained the smallest numbers of polypeptides were: cell defense and detoxification (4.5%), regulation of gene expression (3.85%), signal transduction (3.2%), calcium-binding proteins (1.9%), vesicular transport (1.9%), and DNA replication (1.9%). The functional group named "cell defense and detoxification" included proteins involved in the protection of the cells against different toxins and injury (antioxidant protein 2, dismutase, glutathione Stransferase, etc.). The regulation of gene expression group contained proteins involved in the control of gene transcription (Baf53, COP9 subunit 5, etc.) or in recombination processes such as the structure-specific recognition protein (EC159). The signal transduction group was composed of FIG. 4. 2-DE protein map of the cytoplasmic spermatogonial extracts. 160 g of proteins were loaded on an 18-cm IPG strip with a linear pH 5-6 gradient for isoelectric focusing, and a 12-14% gradient SDS-polyacrylamide gel was used for SDS-PAGE. Proteins were stained with silver nitrate. Isoelectric point and molecular mass calibrations were performed using a calibration abacus (kindly provided by Amersham Biosciences) and molecular weight markers, respectively. Identified spots were assigned a number.
polypeptides involved in different cell signaling pathways (phosphatidylethanolamine-binding protein, protein phosphatase 2, RACK1, etc.). The calcium-binding protein group included polypeptides involved in the different calcium-mediated regulations (annexin, etc.). The main proteins in the vesicular transport group are involved in vesicle formation (dynamin and rab-GDI).
The two proteins that have DNA regulation properties were placed in the DNA replication group. Finally the "other" group (3.85%) contained proteins that could not be classified in the above-mentioned functional classes, such as the fertility protein SP22, which is involved in fertilization, and the vacuolar ATP synthase, which regulates the acidification of intracellular com- partments. The unknown proteins (4.5%) were related to mouse or human clones or to expressed sequence tags that have not yet been characterized in the rat.

DISCUSSION
The aim of this study was to extend the previously established 2-DE reference map of spermatogonial cytoplasmic proteins (14). The removal of cells from their normal physiological environment is intrinsically deleterious to all cell types, especially germ cells, which appear to be the less autonomous of all testicular cells in mammals (2). Additionally it is now generally accepted that there is no primary culture system of purified spermatogonia or of other categories of germ cells (12,34). Considering this and to fulfill with our objectives, it was necessary for us to use large amounts of a freshly isolated fraction of cells displaying the highest purity as possible. Due to the extreme complexity of the seminiferous tubule organization, "the most complex epithelium" according to Fawcett (35), the use of laser microdissection techniques, a method of choice for transcriptomic approaches, could not be used here because 1) the unequivocal recognition of spermatogonia on cryopreserved sections is impossible, 2) remnants of surrounding Sertoli cell cytoplasms would be present in the microdissected material because germ cells and Sertoli cells are extremely intricate from an anatomical point of view, and 3) a proteome analysis would require at least 10,000 -20,000 laser-microdissected spermatogonia, which would not be realistic to undertake.
Of the several techniques designed to isolate highly purified spermatogonia (for a review, see Ref. 12), the procedure we used here is the only one allowing the collection of a large number of cells with a very high rate of purity, which is needed for proteomic analysis. Since the essential aim of our work was to focus on low expressed proteins, it remains that the 5% of contaminants present in our cell suspension (see "Ex-perimental Procedures") are unlikely to significantly interfere with our analysis.
In a previous series of experiments (14) our laboratory identified a set of 53 spermatogonial proteins from a 3-10NL 2-D gel. However, this type of gel can only detect the most abundant housekeeping proteins. To extend the analysis of the spermatogonial proteome and to detect the less abundant proteins produced by these cells, we used overlapping narrow range IPG gels to fractionate the sample. These ''zoom" gels markedly enhance the resolution of protein separation and improve the detection of polypeptides spots on 2-DE (17,18). The narrow range gels enabled us to detect around 1750 different spots compared with the 700 spots detected in the equivalent region of the broad range gel. This corresponds to an increase of more than 2-fold in the number of visualized spots. This improvement in protein resolution reduces the number of co-migrating proteins, which is more suitable for non-ambiguous identification by mass spectrometry, and makes it possible to characterize proteins that are present in low copy number (19). As a first step toward the establishment of a full spermatogonia protein database, this approach allowed us to identify 151 different spots in a 5-6 2-D gel, which corresponds to 65.6% of the spots analyzed. The failure to identify a larger number of known proteins (i.e. present in databases) could result from 1) there being too little protein in a given spot to allow mass spectrometry analysis, 2) the efficiency of peptide recovery being too poor for a significant identification by peptide mass fingerprinting or for analysis by tandem mass spectrometry, 3) the presence of very few tryptic cleavage sites in the protein of interest, or 4) the presence of too many post-transcriptional modifications in the expected tryptic peptides as described previously by others (31). Finally a large number of spots contained too much contaminating human keratin. Keratin contaminants can be introduced during the handling of 2-D gels. However, a recent work discussed the presence of epidermis-related keratin in the rat male germ cells throughout spermatogenesis, including spermatogonia, and particularly a human keratin 5 homolog, sak57 (36). As these proteins are not yet present in the different databases, we cannot exclude the possibility that, in addition to possible contamination by human keratins, some spots were in fact spermatogonial keratins.
The 151 identified spots in our 5-6 2-D gel corresponded to 156 polypeptides. Indeed five spots contained two different proteins (spots EC14, EC121, EC159, EC197, and EC198). For three of these spots (EC14, EC197, and EC198) tandem mass spectrometry was not required for the identification, and MALDI-TOF MS analyses were sufficient. This shows that peptide mass fingerprinting is a powerful technique and that, in good conditions, it can identify more than one polypeptide present in a single spot. Although the use of narrow range IPG gels reduced the number of co-migrating proteins, the resolution of this technique did not seem to be sufficient to separate and to detect all of the polypeptides present in our sample. This means that we may have failed to detect low copy number proteins that co-migrate with more abundant ones. The use of very narrow range IPG gels (i.e. 0.1 unit IPG gels) could partly resolve this problem.
Analysis of the 156 identified spots revealed that they corresponded to 102 different proteins. Proteins that give rise to multiple spots may have arisen due to multiple co-and posttranslational modifications as most eukaryotic proteins are modified, and these modifications are often essential for their function (31). Although we did not intend to characterize the post-translational modifications of the proteins appearing in multiple spots, we can deduce some of them. For instance, spots EC7 and EC8, which differ in their pI (respectively, 5.79 and 5.78) but not in their M r (around 50,000), were identified as being the same protein: aldehyde dehydrogenase 2, mitochondrial. These proteins should be differentially modified, and we hypothesize that they are differentially phosphorylated as this modification leads to a different pI with the same M r . For one of the proteins (N G ,N G -dimethylarginine dimethylaminohydrolase (EC106)) we identified a very common modification, N ␣ -terminal acetylation (37), using tandem mass spectrometry. For the other protein with the same identification (EC45), the peak on the spectrum that corresponded to the N ␣ -terminal acetylated peptide was also present in the MALDI-MS spectrum, suggesting that this isoform of the N G ,N G -dimethylarginine dimethylaminohydrolase is also acetylated. This hypothesis was confirmed by the NiceProt view of the protein in the SWISS-PROT database (accession number 008557). The two isoforms of the protein differed in their M r but also in their pI. This suggested that other modification(s) may have occurred in this polypeptide. Finally one of the other modifications visible in our 2-D gel was proteolysis with the experimental M r of the spot in the gel being below the theoretical M r calculated from the primary amino acid sequence. Indeed for 30 of the identified proteins it was possible to show that the matched peptides covered only a part of the sequence of the corresponding protein, generally the N or C terminus. This is specifically due to proteolytic processing because these spots were reproducibly detected in all the gels and because nonspecific processing always results in a smear on the gel. However, it is unclear whether these in vivo modifications correspond to a specific cleavage reaction that is necessary for protein function or whether they reflect an active degradation process in spermatogonia.
Most of the identified proteins were detected for the first time in spermatogonia. A large proportion of the polypeptides already described in these cells (82.6%) was seen in our previous proteome work using a broad range 3-10NL 2-D gel (14). Among the identified proteins, 21 (20.6%) had already been described in the literature as being expressed within the testis and/or in testicular cells other than spermatogonia (e.g. Leydig cells, peritubular cells, Sertoli cells, spermatocytes, spermatids, and cytoplasmic lobes).
The proteins identified in the pH 5-6 range (i.e. 18% of the spots detected) were classified according to their putative functions in an attempt to improve our understanding of the biology of spermatogonia. As expected at this level of our analysis, most of the proteins identified were involved in primary metabolism. The two other groups that were highly represented, but at levels of only about half each of the previous group, were, not unexpectedly, involved in protein synthesis and processing and in cell structure. The underrepresented groups (cell defense and detoxification, regulation of gene expression, signal transduction, calcium-binding protein, vesicular transport, and DNA replication) were mostly composed of proteins involved in regulation processes, which are generally produced at a low copy number. Cell defense and detoxification proteins represent the necessity for the germ stem cells to defend themselves as the blood-testis barrier does not protect them. In fact, these stem cells are highly resistant to a number of toxicants, which is not surprising as their destruction by toxicants would lead to sterility, which would in turn prevent the species from perpetuating (38,39). The DNA replication proteins included polypeptides that can regulate the mitosis of spermatogonia. Obviously the unknown proteins are the most interesting group because new genes discovered by this proteomic approach could be spermatogonia-or germ cell-specific and may play a major role in spermatogenesis. The rat spermatogonial proteome is currently being systematically analyzed, and all remaining spots that can be recovered from the 5-6 2-D gel are being identified. The ultimate aim is to detect and to identify very scarce proteins and includes 1) the identification of spermatogonial proteins separated using other narrow range IPG gels (i.e. 4 -5, 4.5-5.5, 5.5-6.7, and 6 -9), 2) the use of very narrow range IPG gels (0.1 pH unit), 3) the subcellular fractionation of our samples, and 4) the use of free flow electrophoresis (40).
In this study, several new spermatogonial proteins were mapped. This confirms that the proteomic approach can comprehensively and globally analyze testicular proteins and in particular germ cell proteins, the biology of which remains poorly documented. We are currently concentrating our efforts on the identified proteins that may be involved in the control of spermatogonial activity (EC218, see Table II). Moreover the full characterization of unknown proteins and the cloning of their genes are underway in our laboratory (EC127, EC140, and EC167; see Table II). As discussed before, although unlikely to represent a significant interference with our proteomic analysis, at this stage we cannot totally exclude the possibility that the presence of a few testicular somatic contaminants in our spermatogonia preparation may be the origin of the identification of some of the proteins listed in Table II. We are therefore presently carefully checking, using specific antibodies, the origin of candidate proteins in situ.
Our results, together with those of our previous study (14), are being used as a basis for the construction of a Web-based proteomic database of the mammalian testis (TESTIS-2DPAGE). This database is a necessary step in our project aimed at producing a comprehensive protein map of major mammalian testicular cells. TESTIS-2DPAGE will also be an essential base line from which proteomics will be used to progressively improve our understanding of spermatogenesis, testicular physiology, and physiopathology.