Regulation by pH of the Alternative Splicing of the Stem Cell Factor Pre-mRNA in the Testis*

Proliferation and differentiation of progenitor stem cells are mainly controlled by diffusible and adhesion molecules. Stem cell factor (SCF), an essential regulator of spermatogenesis produced by Sertoli cells, utilize both modes of cell to cell communication. Indeed, SCF exists in soluble (SCFs) and membrane-bound (SCFm) forms, which are required for a complete spermatogenesis, and are generated by alternative splicing of optional exon 6, encoding sites of proteolysis. We show that in the mouse testis, the alternative splicing of SCF is developmentally regulated. SCFs predominates in fetal and neonatal gonads and is then replaced by SCFm in the prepubertal and adult gonads. By sequencing SCF exon 6, we show that the flanking intronic sequences perfectly follow the gt-at rule, suggesting that the basal splicing machinery might not be responsible by itself for exon 6 skipping. Moreover, freshly isolated Sertoli cells mainly express SCFm, but a switch to SCFs occurs after 48 h of culture. We found that this change can be prevented by acidification of the culture medium at pH 6.3 or by addition of lactate. The sustained synthesis of SCFm at low pH was no longer observed in the presence of cycloheximide, suggesting that SCF exon 6 skipping requires de novo protein synthesis. Accordingly, UV cross-linking experiments show that nuclear Sertoli cell protein(s) bind in a sequence-specific manner to exon 6. Together, our data allow the proposal of an integrated mechanism in which the synthesis of lactate by Sertoli cells is used in the same time as an energetic substrate for germ cells and as a promoter of their survival/proliferation through the production of SCFm.

Stem cells (from hematopoietic, nervous, and gonadal systems) are the subject of increasing interest because of their biological and medical importance (1). Specifically, in the testis, the conversion of stem cell spermatogonia into differentiated haploid spermatozoa is a complex process highly dependent upon the somatic Sertoli cells. One of the major questions is the identification of the Sertoli cell-derived factors driving quiescent stem cells into proliferation. Although these factors are largely unknown, some growth factors and cytokines have been suggested to be at play (2)(3)(4)(5). Among them, the stem cell factor (SCF) 1 appears as a key factor, as demonstrated by the pleiotropic effects of the SCF gene mutation, leading to the depletion of three embryonic migratory lineages: hematopoietic stem cells, neural crest-derived melanocytes, and primordial germ cells (6).
SCF has been detected both in membrane-bound (SCFm) or soluble (SCFs) forms (7,8). The soluble form is generated from an integral membrane protein precursor, by proteolytic cleavage at a site located in the proximal extracellular domain (9). Because the main proteolytic site involved in this process is encoded by a short 84-base pair-long alternative exon (exon 6), the final localization of SCF is ultimately dictated by differential splicing (10). Although SCFm and SCFs transcripts are equally abundant in spleen and heart (9 -11), SCFs is predominant in all other organs tested, including brain, bone marrow, kidney, lung, liver, and thymus of adult mouse (9 -11). Interestingly, in the testis, where SCF and its cognate receptor c-kit are produced, respectively, by Sertoli cells (12)(13) and spermatogonia (14 -17), the major form of SCF found in the adult mouse is SCFm (10,11). Accordingly, SCFm has been shown to be of major functional importance for germ cells, as demonstrated by the infertility of mice homozygous for a SCF gene mutation named Steel-Dickie (Sl d ) (6,18). This mutation consists in an intragenic deletion of the transmembrane-and cytoplasmic domain-coding regions, resulting in the constitutive production of SCFs (10,19).
In the present study, we demonstrate that SCF pre-mRNA splicing is developmentally regulated in the male gonad and that an acidic microenvironment, probably because of the high amounts of lactate in testicular Sertoli cells, might be responsible for the switch of SCF splicing in favor of SCFm, which can promote germ cell survival in the contact of Sertoli cells and proliferation of spermatogonia type A. ase/dispase was obtained from Boehringer Mannheim. Sigma was the source for transferrin, gentamicin, nystatin, insulin, ␣-tocopherol, HEPES, random hexanucleotides, and deoxynucleotides triphosphate (dNTP). Taq polymerase was purchased from Appligene-Oncor (Illkirch, France). Amersham Pharmacia Biotech was the source for [␣-32 P]UTP. In vitro transcription kit was purchased from Promega (Madison, WI), and ABI Prism® dye terminator kit was from Perkin-Elmer.
Sequence Analysis of SCF Splice Sites-SCF intron 5 and 6 were amplified using a mouse DNA library as a template and Expand ® long template PCR system (Boehringer Mannheim). Intron 5 was amplified using the following primers (see Fig. 1A): S5 (5Ј-TGGTGGCATCTGA-CACTAGTGA-3Ј) and S6b (5Ј-GCTACTGCTGTCATTCCTAAGG-3Ј); intron 6 was amplified using the primers: S6a (5Ј-CCAGAGTCAGTGT-CACAAAACC-3Ј) and S7 (5Ј-CTTCCAGTATAAGGCTCCAAAAGC-3Ј). PCRs were done in the presence of 300 nM primers in the following conditions: 94°C for 5 min; 10 cycles of 94°C for 10 s, 60°C for 30 s, 68°C for 8 min; 20 cycles of 94°C for 10 s, 60°C for 30 s, 68°C for 8 min (with an increment of 20 s at each cycle); and 68°C for 14 min; 90 ng of purified PCR products were used as template for DNA sequencing using the ABI Prism dye terminator kit in a DNA Thermal Cycler (Perkin-Elmer) for 30 cycles: 96°C for 10 s, 50°C for 5 s, 60°C for 4 min. After cycling, PCR products were purified and dried, and just before loading, the sample was resuspended in 4.0 l of deionized formamide, 50 mM EDTA, pH 8.0 (5:1). Fluorescence-based DNA sequence analyses were obtained on an ABI 373 DNA sequencer fitted with a 6% polyacrylamide gel using the manufacturer's version 2.1.0 software.
RT-PCR Analysis-Total RNAs were extracted from mouse testis or cultured Sertoli cells with TRIzol ® reagent, a monophasic solution of phenol and guanidine isothiocyanate.
Single strand cDNA was synthesized by reverse transcription starting from 3 g of total RNA and using 5 M random hexanucleotides, 0.2 mM dNTP, and 1 units/l Moloney murine leukemia virus. PCR reactions were done using Taq polymerase (0.01 units/l), 0.2 mM dNTP, and 0.01 g/l S5 and S7 primers. The mixture was first heated at 94°C for 5 min and then 20 cycles of 94°C for 40 s, 57°C for 1 min (with 0.02 s and 0.3°C decrease each cycle), 72°C for 40 s followed with 12 cycles of 94°C for 40 s, 51°C for 15 s, 72°C for 40 s, then 72°C for 5 min. SCF amplification products containing exon 6 (SCFs) or lacking exon 6 (SCFm) were, respectively, 251 bp and 167 bp long (Fig. 1A). They were analyzed on 2% agarose gel and visualized by ethidium-UV staining. For ␤-actin RT-PCR we used the primers: 5Ј-GACAGGATG-CAGAAGGAGAT-3Ј and 5Ј-TTGCTGATCCACATCTGCTG-3Ј, and ␤-actin cDNAs were amplified according the following conditions 94°C for 2 min, 28 cycles of 94°C for 30 s, 59°C for 30 s, 72°C for 30 s and 72°C for 5 min.
Sertoli Cell Nuclear Extract Preparation-Buffers A, C, and D used for preparation of nuclear extracts were described by Dignam et al. (21). All steps were done on ice, and proteolysis was minimized by addition of 0.5 mM phenylmethysulfonyl fluoride and 0.3 g/ml antipain and leupeptin to all buffers. Aprotinin (0.5 g/ml) was also added to buffers A and C. Nuclear extracts were prepared as described by Dehbi et al. (22). The extract was removed and stored at Ϫ80°C. Protein concentration was determined by the method of Bradford at 545 nm, using bovine serum albumin as a standard (23).
UV Cross-linking Analysis-A SCF cRNA containing the last 57 nucleotides of exon 5, whole exon 6, and the first 46 nucleotides of exon 7 (named SCF 567) was synthesized by in vitro transcription from a pGEM-T plasmid containing the corresponding cDNA fragment after linearization with NcoI and using the SP6 RNA polymerase. A SCF cRNA containing the last 57 nucleotides of exon 5 and 110 nucleotides of exon 7 but lacking exon 6 (named SCF 57) was transcribed using the T7 RNA polymerase after linearization of the pGEM-T vector with NotI. A 241-nucleotide-long cRNA Trk B, corresponding to a mRNA species produced by Sertoli cells but unrelated to SCF, was used as a control and generated by transcription from a pGEM-T vector containing the trkB cDNA using the SP6 RNA polymerase after linearization with NcoI. Radiolabeled RNAs were obtained by adding 50 Ci of [␣-32 P]UTP in the in vitro transcription mixtures. The specific activity of 32 P-SCF 567 is 7.9 ϫ 10 8 cpm/g. Proteins interacting with radiolabeled precursor RNAs were detected by UV cross-linking methods (24). Radiolabeled RNAs were incubated with 20 g of nuclear protein extracts for 10 min at 30°C in binding buffer (1 mg/ml poly(G), 0.5 mg/ml tRNA, 0.1 mM dithiothreitol, 50 mM KCl, 10 mM Tris, pH 8.0, 3 mM MgCl 2 ). After binding, samples were cross-linked at 2 J/cm 2 using an Appligene UV cross-linker. Then, RNA was digested for 15 min at 37°C in the presence of RNase A (1 mg/ml). After the addition of SDS sample buffer, the samples were heated to 100°C for 5 min and then run on SDS, 10% PAGE. Gels were dried and exposed 1-2 days with Fuji RX film.

Alternative Splicing Pattern of SCF during Mouse Testicular
Development-SCFs and SCFm mRNAs were discriminated by exon connection RT-PCR, using primers defined in exons 5 and 7 and named S5 and S7, respectively, as shown in Fig. 1A. The SCFs cDNA fragment appears longer, with an additional 84bp-long exon 6. Using this method, we have first determined the evolution of the SCFm/SCFs ratio during mouse testicular development (Fig. 1B). Mouse testes were removed at different key stages of testicular development i.e. fetal (13, 15, 18 days post-coitum (dpc); day 0 post-coitum was the day where the vaginal plug was detected), neo-natal (2 and 8 days post-partum (dpp), pre-pubertal (10, 15, and 21 dpp), pubertal (28, and 42 dpp), and adult (60 dpp). SCFm (lacking exon 6) appears predominant in 13 dpc fetal testis, whereas SCFm and SCFs (containing exon 6) mRNAs appear equally abundant in 15 dpc testis, and SCFs becomes predominant at the end of gestation (18 dpc) and in neonatal testis (2 dpp) (Fig. 1B). At the prepubertal and adult periods, male gonad contains predominantly SCFm. At 42 and 60 dpp, the total level of expression of SCF was lower because of the dilution of Sertoli cells (which are in a limited number) by the increasing amounts of adult germ cells (Fig. 1B).
Sequence Analysis of SCF Gene Exon 6 and Flanking Intronic Sequences-Because most cases of alternative splicing are the consequence of suboptimal splice sites, we determined the genomic sequences surrounding exon 6. Surprisingly, this analysis showed that 5Ј and 3Ј splice sites of introns 5 and 6 perfectly follow the gt-ag rule, but in addition, exactly fit mammalian 5Ј and 3Ј splice site consensus motifs, which are, re-FIG. 1. Alternative splicing of SCF pre-mRNA. A, exon 6, encoding proteolysis sites (bold vertical bars) is either maintained or eliminated from the cytoplasmic SCF mRNAs, generating, respectively, SCFs or SCFm. The localization of the RT-PCR primers used in this study, S5, S6a, S6b, and S7, is shown. S5-S7 RT-PCR products corresponding to SCFs and SCFs were 251 and 167 bp long, respectively. The region of exon 7 encoding the transmembrane domain is represented with horizontal lines. B, SCF alternative splicing pattern during development. Expression of SCF mRNAs was analyzed by RT-PCR in testis from fetus (13, 15, 18 dpc) and from 2-to 60-day-old mice. ␤-Actin mRNA was measured as an internal control from the same input of reverse transcription mixtures. spectively, GTRAGT and YAG preceded by a polypyrimidine stretch (Fig. 2). The presence of these optimal splice sites suggests that the basal splicing machinery can not be responsible by itself for exon 6 skipping. Hence, we have looked for the presence of other cis elements that may be the targets for more specific RNA binding factors. The particularly high conservation degree of the exon 6 nucleotide sequence, when compared with upstream and downstream exons (Table I), suggest that, in addition to its coding capacity, the nucleotide sequence of exon 6 may contain a cis element involved in its own splicing. Scanning of the intronic sequences against data bases revealed that only the intronic region flanking exon 6 in 3Ј (accession number AF083887) is related to a previously identified cDNA sequence. This cDNA (accession number A50814) corresponds to a human SCF mRNA species likely to have retained intron 6 during splicing.
Modulation of the SCF Alternative Splicing Pattern in Cultured Sertoli Cells-We wanted then to test the ability of Sertoli cells to synthesize SCF when isolated from the whole testis context. As shown in Fig. 3, freshly isolated Sertoli cells (from 16 -18-day-old animals) predominantly expressed SCFm mRNA, as observed in vivo. By contrast, during the culture, Sertoli cell SCF alternative splicing pattern changed in favor of the expression of SCFs mRNA (Fig. 3). For example, at 0.5 and 2.5 h of culture, SCFm mRNA is still predominantly expressed (ratio SCFm/SCFs Ͼ 1), whereas at 48 and 72 h of culture, SCFs becomes predominant (ratio SCFm/SCFs Ͻ 1). The point of equivalence of SCFm and SCFs mRNAs is around 24 h (ratio SCFm/SCFs Х 1).
Role of Low pH and Lactate in the Regulation of SCF Alternative Splicing-The shift of SCF alternative splicing pattern observed in cultured Sertoli cells has led us to look for the mechanisms regulating the presence of exon 6 in mature SCF mRNAs. While testing various culture conditions for their ability to maintain the predominance of SCFm synthesis, we found that the SCFm mRNA synthesis is strikingly dependent on pH. Lowering the culture medium pH by HCl addition or sodium bicarbonate withdrawal prevents the shift to the SCFs mRNA form. Sertoli cells cultured for 48 h at acidic pH (6.3) predominantly expressed SCFm, whereas those cultured at pH 7.2 and pH 7.6 presented an alternative processing in favor of SCFs (Fig. 4). Furthermore, kinetic studies show that at pH 6.3, SCFm predominates at each time tested (2 to 72 h), whereas it is progressively replaced by SCFs after culture shift at pH 7.6 ( Fig. 5). Indeed, at pH 6.3, at each time tested, the SCFm/SCFs ratio is higher than 1, whereas at pH 7.6, the ratio decreased progressively until 0.6 at 72 h (Fig. 5) The action of low pH on SCFm expression was detected at 12 h and maintained during the other times tested.
As Sertoli cells are known to produce lactate, an acidic metabolite in the context of Sertoli cell-germ cell metabolic cooperation, this metabolite was tested. As shown in Fig. 6, lactate in cultured Sertoli cells maintained the SCFm expression. Indeed, after 48 h of culture in the presence of lactate, the SCFm/SCFs ratio is higher than 1, whereas in the absence of lactate, the ratio is lower than 1 (Fig. 6). Altogether, these results suggested that the trans-acting factors regulating exon 6 skipping are somehow regulated by acidity. In this respect, one must notice that pH remained neutral at the end of the experiment shown in Fig. 3, suggesting that an intracellular factor responsible for the production of SCFm is lost from the Sertoli cells maintained in culture unless an experimental acid-FIG. 2. DNA sequence of 5 and 3 splice sites of SCF intron 5 and 6. Introns 5 and 6, enclosing the optional exon 6, were cloned and sequenced as described under "Experimental Procedures." The intron boundaries sequences close to exon 6 are represented in bold letters, and polypyrimidine stretches are underlined. GenBank accession numbers are AF083885 and AF083886, for beginning and end of intron 5, respectively, and AF083887 and AF083888 for beginning and end of intron 6, respectively. ification is imposed. Fig. 7, Sertoli cells cultured at pH 6.3 for 48 h expressed predominantly SCFm (SCFm/SCFs ratio Ͼ1), whereas a progressive increase of the SCFs is observed when cycloheximide is added into the culture medium (SCFm/ SCFs ratio ϭ 1). This result indicates that protein synthesis is required for sustained production of SCFm at low pH.

Exon 6 RNA Recognition by Proteins from Sertoli Cell Extracts-
To test whether exon 6 may be bound by a specific RNA binding factor, [␣-32 P]UTP RNAs containing exon 6 (SCF567) were cross-linked by UV to nuclear protein extracts from Sertoli cells freshly isolated. A single cross-linked protein-RNA complex was detected, with an apparent molecular mass of 69 kDa (Fig. 8A, lane 1). Competition experiments showed that the 69-kDa complex formation can be prevented by previous addition of an excess of cold RNA containing exon 6 in a dose-dependent manner (Fig. 8A, lanes 2-4) but not with SCF lacking exon 6 (SCF57) or the unrelated trkB mRNA (Fig. 8B). The 69-kDa complex is no longer obtained when using RNA probes unrelated to exon 6 as SCF57 or TrkB RNA (data not shown). Altogether, these results suggested the exon 6 sequence specificity of the 69-kDa RNAbinding protein(s). DISCUSSION The key role of SCF action in spermatogenesis is dramatically dependent on its membrane-bound localization, as demonstrated by several observations. (i) Adult mutant mice, which express only SCF soluble form (Steel-Dickie mutation (19)) or an abnormal SCFm devoid of cytoplasmic domain (Steel 17 H mutation with a skipping of exon 8, which encodes the cytoplasmic tail of SCF (25)) are infertile; ii) Disruption of spermatogenesis using a Sertoli cell toxicant such as 2,5-hexanedi-  3. Progressive change of the SCF pre-mRNA splicing in cultured Sertoli cells. Sertoli cells were freshly isolated (0) or cultured for different times (0.5 to 72 h) in defined medium (DME/F12, pH 7.38, with HEPES 15 mM and bicarbonate 1.2 g/liter) and supplemented as described under "Experimental Procedures." Exon connection analysis was performed using the S5 and S7 primers defined in exons 5 and 7 located apart from the optional exon 6. The histogram in the lower panel represents the SCFm/SCFs ratio.

FIG. 4. Effect of low pH on SCF alternative splicing pattern.
Sertoli cells were cultured in DME/F12 medium for 48 h, and the pH was adjusted from 6.3 to 7.6. The histogram in the lower panel represents the SCFm/SCFs ratio.
FIG. 5. Time course of SCF alternative splicing following pH changes. Sertoli cells were cultured for different times (2 to 72 h) in DME/F12 medium, and the pH was adjusted to 6.3 or 7.6. The histogram in the lower panel represents the SCFm/SCFs ratio. one was accompanied with a switch in alternative splicing from SCFm to SCFs mRNA (26); iii) In vitro experiments have shown that SCFm promotes anchoring of germ cells to Sertoli cells (11) as well as proliferation of germ cells (19). SCF thus provides a striking example for the biological importance of certain post-transcriptional regulations. The aim of the present study was to examine the splicing responsible for the synthesis of SCFm and SCFs during gonadal development and to identify the regulatory mechanisms involved in alternative splicing of SCF pre-mRNA in Sertoli cells.
Two other laboratories (11,27) have reported, by using RT-PCR approach, the testicular expression of SCFs and SCFm during the post-natal period (but not to the fetal period). These data were conflicting. Rossi et al. (27) showed that testis from 8 to 13 dpp expressed predominantly SCFm transcripts, whereas 18-to 60-day-old testis expressed the SCFs form. By contrast, Marziali et al. (11) showed that at 6 dpp, SCFs and SCFm were equally abundant, whereas SCFm form was predominantly expressed from 12 dpp to the adult period. Because of these discrepancies, we first investigated the developmental pattern in SCF expression during the testicular fetal, postnatal, prepubertal pubertal, and adult periods. Our results, at least in adult animal, appears to correlate with those of Marziali et al. (11). In the fetal mouse testes, we have shown that SCFm mRNAs are predominant at 13 dpc when primordial germ cells proliferate. Indeed, primordial germ cells proliferate from 7.5 to 13 dpc (28). Then SCFs mRNAs are predominant at 18 dpc to 2 dpp, when germ cells are quiescent (28). In this context, the two SCF forms are likely to play important and distinctive roles during the fetal gonad development. SCFs may promote migration and survival of primordial germ cells, whereas SCFm allows survival and proliferation of primordial germ cells (29,30). Finally, from 8 dpp to adulthood, the SCFm form predominates. The switch from SCFs to SCFm coincides with the first wave of multiplication of spermatogonia, occurring at 6/8 dpp (31).
However, our data indicate that as soon as Sertoli cells were isolated from their in vivo context, they exhibit a dramatic decline in their ability to synthesize SCFm, suggesting that a factor(s) from their local in vivo environment is required for active exon 6 skipping. We propose that low pH might initiate this process. Indeed, acidity (pH 6.3) was able to maintain SCFm synthesis by cultured Sertoli cells. Our findings were consistent with examples in the literature showing that some cases of pre-mRNA splicing are regulated by pH (32,33). For instance, in the case of the ATP synthase that is involved in energy production in heart and skeletal muscle tissues, an exon skipping in ATP synthase ␥ subunit is specifically observed in heart and skeletal muscle but not in the liver (33). Whether an acidic (micro)environment may occur in Sertoli cells in physiological conditions is still to be established. However, it is of interest to note that, in the testis, lactate is produced in very large amounts by Sertoli cells, to fulfill the energetic requirements of germ cells that are unable to metabolize glucose (34). Lactate is an energetic substrate preferentially used by postmeiotic germ cells and particularly spermatids (35). Recently, we have shown that these germ cells control and direct glucose metabolism in Sertoli cells toward lactate formation through some signaling molecules that increase lactate dehydrogenase A (an enzyme that favors the convertion of pyruvate into lactate (36)). Based on the present observation that in acidic conditions, Sertoli cells preferentially express the SCFm, we hypothesize that at certain specific stages during the seminiferous epithelium cycle, which particularly associates Sertoli cells and spermatids, an acidic microenvironment might be generated by an elevated concentration of lactate. Therefore, lactate synthesis, in addition to supplying spermatids with their energetic substrate, can also, by lowering the pH, allow the synthesis of SCFm that stimulates through c-kit the proliferation of spermatogonia and therefore triggers a new spermatogenic cycle. The hypothesis that terminally differentiated germ cells (spermatids) might send signals to stem germ cells to initiate a new spermatogenic cycle, has been proposed some 40 years ago by Roosen-Runge (37). Although some candidates (e.g. residual bodies (37)) have been suggested as such signals between differentiated and stem germ cells, the precise determining factor remains unknown. This report proposes that an increased acidity through an increase in lactate production at the level of Sertoli cells in contact to spermatids may represent the triggering signal for SCFm expression.
To investigate the mechanisms involved in SCF exon 6 skipping in the male gonad, we used as an experimental model, purified mouse Sertoli cells cultured in defined medium and have taken advantage from the observation related to a switch occurring from SCFm to SCFs after 12 h of culture of Sertoli cells. With regard to the mechanisms involved in alternative splicing of pre-mRNA, examples available in the literature indicate that exon skipping is most often the consequence of suboptimal cis elements at the exon-intron boundaries. These weak sites fail to bind components of the basal splicing machinery, including U1 small nuclear RNP (38,39) and U2AF 65 (39), thus leading to constitutive exon exclusion. Exon inclusion then requires additional trans-acting factors, which specifically recognize the weak exons at the level of so-called splicing enhancers (40). Interestingly, an opposite situation is likely to occur in the case of SCF. Because our determination of intron sequences surrounding the alternative exon 6 has revealed the presence of 5Ј and 3Ј intron sequences, exactly fitting the splice site consensus for the major splicing pathway of nuclear pre-mRNA introns (41), exon 6 of SCF would be constitutively retained in mature SCF mRNAs by the basal spliceosome. In such a situation, one may hypothesize that an exclusion factor is involved to specifically eliminate exon 6 during the synthesis of SCFs mRNA. This possibility is strongly supported by the fact that de novo protein synthesis is required for predominant synthesis of SCFm, i.e. for exon 6 elimination. The particularly high degree of sequence conservation of exon 6 during evolution raises the hypothesis that exon 6 RNA sequence may be the target for a sequence-specific RNA binding factor(s). Consistently, cross-linking experiments show that a nuclear protein(s) specifically interacts with SCF alternative exon. Although we have not yet obtained direct evidence that the exon 6-binding 69-kDa complex is the one actually involved in the acidity-mediated splicing, this possibility is highly suggested by the strict sequence conservation of exon 6 between mammals and even birds (42), in parallel with the evolutionary conservation of the roles of SCFs and SCFm. Identifying the nuclear protein(s) interacting with SCF exon 6 would be of great interest for elucidating a central turning point of the testis development.