Journal of Cell Science partnership with Dryad

Journal of Cell Science makes data accessibility easy with Dryad


We analysed the phenotypic outcome of a Stra8-null mutation on male meiosis. Because the mutant spermatocytes (1) underwent premeiotic DNA replication, (2) displayed cytological features attesting initiation of recombination and of axial-element assembly, and (3) expressed Spo11 and numerous other meiotic genes, it was concluded that STRA8 is dispensable for meiotic initiation. The few mutant spermatocytes that progressed beyond leptonema showed a prolonged bouquet-stage configuration, asynapsis and heterosynapsis, suggesting function(s) of STRA8 in chromosome pairing. Most importantly, a large number of mutant leptotene spermatocytes underwent premature chromosome condensation, within 24 hours following the meiotic S phase. This phenomenon yielded aberrant metaphase-like cells with 40 univalent chromosomes, similar to normal mitotic metaphases. From these latter observations and from the wild-type pattern of Stra8 expression, we propose that, in preleptotene spermatocytes, STRA8 is involved in the process that leads to stable commitment to the meiotic cell cycle.


In meiosis, a single round of DNA replication – the meiotic S phase – precedes two rounds of chromosome segregation, resulting in the production of haploid gametes from diploid cells. The first meiotic division, termed meiosis I, is a crucial stage in which alignment and synapsis of homologues (i.e. the maternal and paternal homologous chromosomes) occurs, allowing for exchange of genetic material, which contributes to genetic diversity (Cohen et al., 2006).

The timing of the commitment to enter meiosis shows a striking difference in male and female mammalian germ cells. In female mice, for instance, the entire gonocyte pool starts meiosis in utero on embryonic days (E)12.5 to 14.5 and meiosis definitively stops at birth (Speed, 1982). By contrast, male meiosis is only initiated at about postnatal day (P)10 and continues throughout life (Nebel et al., 1961). Many other aspects of mammalian meiosis show sexual dimorphism (Handel and Eppig, 1998), and there are several mouse mutants in which meiosis is differentially impacted in oocytes and spermatocytes (Morelli and Cohen, 2005; Cohen et al., 2006).

In the female mouse, the retinoic-acid signalling pathway triggers the entry into meiosis by promoting the expression of STRA8 (stimulated by retinoic acid 8) (Bowles et al., 2006; Koubova et al., 2006; Baltus et al., 2006). STRA8, a 45 kD vertebrate-specific product of unknown molecular function and without significant homology to any other protein, is present exclusively in gonads (Oulad-Abdelghani et al., 1996). In female mice, Stra8 is expressed in embryonic gonocytes during a restricted window of time (Menke et al., 2003) and is indispensable for their entry into the meiotic S phase (Baltus et al., 2006). In males, Stra8 expression is first detected in primitive spermatogonia at P5, then, from P10 onwards, becomes prominent in premeiotic (i.e. preleptotene) spermatocytes (Vernet et al., 2006a). Here, we thoroughly investigated the phenotype of STRA8-deficient spermatocytes. We found that STRA8 is not required for meiotic initiation, but participates as a fundamentally positive regulator in the commitment of spermatocytes to meiosis and that it also regulates progression through the early stages of meiotic prophase.


Generation of mice with disrupted Stra8 alleles

To assess the role of STRA8 in vivo, we generated a Stra8 mutant allele in the mouse through homologous recombination in embryonic stem (ES) cells (Fig. 1A,B). Heterozygous mating (n 31; mean is seven pups per litter) yielded 26% (n 56) wild-type (WT, Stra8+/+), 51% (n 111) heterozygous (Stra8+/L–) and 23% (n 50) homozygous (Stra8L–/L–) mice. To check that gene disruption was efficient, Stra8 mRNA expression was analysed by reverse transcription of mRNA coupled to PCR amplification (RT-PCR) of the cDNA obtained. In RNA extracts from P15 WT testes, two fragments of 479 bp and 208 bp were detected, corresponding to the two major Stra8 isoforms reported in databases ( (Fig. 1C, lanes 4-6). In RNA extracts from Stra8L–/L– testes, a single, shorter, fragment of 120 bp was amplified (Fig. 1C, lanes 7-10). It corresponded to a truncated mRNA derived from the splicing between exons 1 and 5, as expected from the structure of the mutant (L–) locus (Fig. 1A). Importantly, the translation-initiation codons (located in exons 2 and 4, respectively) were missing in the mutant transcript (see Fig. 1C), therefore making synthesis of any truncated STRA8 protein unlikely.

Fig. 1.

Targeted disruption of the Stra8 gene. (A) Structure of the targeting vector and partial restriction map of the WT Stra8 locus before [WT (+) allele] and after (L3 allele) homologous recombination, as well as after Cre-mediated excision (L– allele). Black boxes (1-5) indicate exons. The locations of restriction sites (E, EcoRI; N, NcoI; 47, Eco47III, X, XbaI) are indicated. Arrowhead flags represent loxP sites. neo, neomycin-resistance cassette; pBS-SK, pBluescript II SK+. Arrows indicate the location of primers #1, #2 and #3, which were used for genotyping. (B) Long-range PCR analysis of tail genomic DNA from mice with the indicated Stra8 genotype using primers #1 and #2 (upper panel), and #1 and #3 (lower panel). The identities of the different alleles are indicated on the right. Note that a large fragment corresponding to the (+) allele is amplified from DNA of WT mice using primers #1 and #3 (lower panel). The asterisk points to an unspecific product. (C) Schematic representation of the Stra8 transcripts that are synthesised from the WT or the mutant loci (left). Open and black segments in boxes 1-5 indicate non-coding and translated regions of the exons, respectively. The location of the ATG codon is specified for each Stra8 isoform. The size of the fragments amplified from the corresponding cDNA using primers #4 and #5 is also indicated. (Right) Results of RT-PCR assays obtained from P15 WT (+/+) and homozygous mutant (L–/L–) testis RNA using primers #4 and #5 (right panel). The size of the fragments is indicated on the right. Negative and positive controls were obtained using water (H2O) and a STRA8-expressing plasmid (pStra8), respectively. As expected, a shorter fragment corresponding to a truncated mRNA is amplified from the Stra8-null-testis RNA samples. The asterisk points to primer dimers.

STRA8 protein is no longer expressed in Stra8L–/L– mutant testes

That our Stra8 mutation was functionally a null mutation was assessed by immunocytochemistry. In this assay, all preleptotene spermatocyte nuclei in WT adults and a large number of WT primitive spermatogonia nuclei at P5 were strongly labelled, as expected (Fig. 2A,C) (Vernet et al., 2006a). By contrast, mutant testes (hereafter designated as Stra8–/–), which were histologically indistinguishable from WT testes at P5 (not shown) and which all contained preleptotene spermatocytes at adulthood (see below), never displayed STRA8 immunoreactivity (Fig. 2B,D). Likewise, spreads generated from testes showed that preleptotene nuclei were positive for STRA8 in WT, but negative in Stra8–/– mice (compare Fig. 2E-G with 2H-J). To further verify that we generated a null allele, we performed western blot analysis on total proteins from WT and Stra8–/– testis extracts using an anti-STRA8 polyclonal antibody. A single protein band with a molecular mass of ∼50 kDa was detected in testes of WT mice, whereas no protein was detected in extracts from Stra8–/– mice (Fig. 2K, arrow in upper panel). Expression of actin (∼43 kDa) was subsequently determined to monitor protein loading among the samples (Fig. 2K, arrow in lower panel). Because, in addition, ovaries from adult mice homozygous for the Stra8 mutation (n 3) were devoid of oocytes and follicles [compare Fig. 2L,M with 2N,O and Baltus et al. (Baltus et al., 2006)], our disruption mutation is a genuine null mutation.

Meiotic prophase I is arrested in Stra8-null mutants

Histological analyses of testes from young (8-week old) Stra8–/– adults (n 5) revealed that spermatogenesis was severely disrupted (compare Fig. 3A with 3B). About 20% of the seminiferous-tubule sections contained only spermatogonia and supporting cells (i.e. Sertoli cells), whereas the majority harboured spermatocytes with preleptotene and leptotene chromatin morphologies (Fig. 3B,E and data not shown). Only a few spermatocytes exhibited zygotene and pachytene nuclear morphologies (Fig. 3C; compare Fig. 3G,I with 3F,H). Mid-pachytene spermatocytes (i.e. containing an XY body, Fig. 3H) and diplotene spermatocytes, as well as all post-meiotic germ cells (i.e. spermatids), were absent (Fig. 3B and data not shown).

To further characterise the spermatocyte stages present in Stra8–/– testes, we analysed the expression of meiotic prophase-I markers. On histological sections of WT mice, threads of the synaptonemal complex (SC) protein SYCP3 were first visible in nuclei at early pachynema (Fig. 3J), whereas the transcriptional regulator TRIM24 was detected in Sertoli cells, spermatogonia and prophase-I spermatocytes from mid-pachynema and beyond (Ignat et al., 2008) (Fig. 3L). In seminiferous tubules of Stra8–/– mutants, expression of TRIM24 was not altered in Sertoli cells or spermatogonia. Few spermatocyte nuclei displayed SYCP3 threads (Fig. 3K), but none ever stained positively for TRIM24 (Fig. 3M). Altogether, these data indicate that, in Stra8–/– mutants, meiosis reaches the early-pachytene stage (Fig. 3K,M) but does not progress to mid-pachynema.

In accordance with an arrested maturation of spermatocytes, the seminiferous tubules exhibited large amounts of TUNEL-positive and pycnotic nuclei scattered among those with a pachytene-like morphology (arrowheads in Fig. 4P,Q; Fig. 5K). Therefore, Stra8–/– early-pachytene-like spermatocytes undergo apoptosis.

STRA8-deficient preleptotene spermatocytes can achieve an S phase and enter meiosis

In WT adult testes, Stra8 transcripts and proteins were detected in differentiating A spermatogonia that were present at stages VII to XI (i.e. differentiating A1, A2 and possibly A3 spermatogonia) (Chiarini-Garcia and Russell, 2001). The immunostaining and in situ hybridisation signals in spermatogonia were obviously specific to particular stages of the seminiferous-epithelium cycle, indicating that most, if not all, undifferentiated spermatogonia did not express Stra8. Stra8 transcripts and proteins were absent from B spermatogonia, which are the immediate precursors of preleptotene spermatocytes (Fig. 4A-G) (Vernet et al., 2006a), but both were highly expressed in all preleptotene spermatocytes that are present at stages VII and VIII of the seminiferous epithelium cycle (see Russell et al., 1990). Moreover, STRA8 was present in preleptotene spermatocytes from the earliest time of their appearance, at P7 (data not shown). Low amounts of STRA8 protein, but no Stra8 transcripts, were also detected in leptotene spermatocytes, whereas meiotic spermatocytes beyond this stage and post-meiotic germ cells did not express Stra8. Importantly, this expression pattern strongly suggests that STRA8 could play a crucial role during preleptonema.

Fig. 2.

The Stra8 mutation is a null mutation. (A-K) Absence of the STRA8 protein in germ cells from male mutants. (A,B) Show immunodetection of STRA8 (red signal) on histological sections of testes at 8 weeks of age, and the DAPI nuclear counterstain (blue). (C,D) Show immunodetection of STRA8 (red signal) on histological sections of testes at P5 and the DAPI nuclear counterstain (blue). (E-J) Show immunodetection of STRA8 (red signal) on nuclear spreads from testes at 8 weeks of age and the DAPI counterstain (blue). Note that: (1) in G,J, the superimposition of the immunohistochemical signals (red false colour) and of the DAPI nuclear stain (blue false colour) were converted to a bright-field image using Photoshop, and (2) in E-J, the three panels in each column represent the same nuclei. (K) Western blot analysis of testis proteins (100 g) from adult mice with the indicated genotypes, using an anti-STRA8 polyclonal antibody (upper panel). Excision of exons 2-4 resulted in a null allele, as confirmed by the absence of any STRA8 in Stra8–/– testes (arrow). The asterisk points to an unspecific signal that was detected both in WT and Stra8–/– extracts. The lower panel, showing detection of actin (ACTA), reveals that similar amounts of protein are loaded on each lane (arrow). (L-O) Absence of oocytes and follicles in ovaries from mutant females at 8 weeks of age. Histological sections were stained with haematoxylin and eosin. The genotypes are as indicated. AF/PF/PRF, antral/primary/primordial follicles; E, ovarian surface epithelium; LY, Leydig cells; PR, preleptotene spermatocytes; ST, ovarian stroma; T, seminiferous tubules. Scale bar: 80 m (A-D); 15 m (E-J); 160 m (L,N); 40 m (M,O).

REC8 and SMC1B are meiotic-specific cohesins that first appear in preleptotene and leptotene spermatocytes, respectively (Fig. 4R,S,V,W) (Eijpe et al., 2003). Both Rec8 and Smc1b transcripts were readily detectable in preleptotene spermatocytes of Stra8–/– mutants (Fig. 4T,U,X,Y), thus confirming our histological observation that preleptotene spermatocytes can differentiate from B spermatogonia in the absence of STRA8.

A unique property of preleptotene spermatocytes is to perform the single round of DNA replication; this replication is indispensable for the two rounds of chromosome segregation during meiosis. The capacity of Stra8–/– preleptotene spermatocytes to replicate DNA was assessed through 5-bromo 2′-desoxyuridine (BrdU)-incorporation and immunodetection assays. In Stra8–/– mutants, as in WT mice, all preleptotene spermatocytes, which localised along the circumference of seminiferous tubules, had incorporated BrdU at 2 hours after its injection (Fig. 4H-K), and BrdU-labelled germ cells at 8 days post-injection displayed a zygotene or early-pachytene chromatin morphology (Fig. 4L-Q), but never a preleptotene (nor a leptotene) morphology.

RT-PCR analysis of genes involved in DNA double-strand break (DSB) formation (Spo11, Mei1), DSB repair and recombination (Atm, Rad51, Msh4, Mlh1, Mlh3), sister-chromatid cohesion (Rec8, Smc1b), and SC structure (Sycp1, Sycp3, Syce2, Fkbp6) during meiotic prophase 1 (Cohen et al., 2006) revealed that they were all expressed in testes from P15 Stra8–/– mutants (not shown). Altogether, these data clearly indicate that STRA8-deficient preleptotene spermatocytes replicate DNA and enter meiosis.

The normal assembly of the SC is impaired in the absence of STRA8

To identify defects underlying the developmental arrest and apoptosis of Stra8–/– spermatocytes at zygonema or early pachynema, we examined the dynamics of chromosome synapsis and recombination using antibodies directed against SC proteins, components of the meiotic nodules and chromatin markers. Synapsis of homologues, which begins at zygonema and is completed at pachynema in WT autosomes (Cohen et al., 2006), was monitored by co-immunostaining nuclear spreads with antibodies against SYCP1 and SYCP3. During WT spermatogenesis, SYCP3 proteins begin to assemble along each sister-chromatid pair at leptonema to form the axial element, which represents the precursor of the SC and subsequently, during zygonema, becomes a major component of the SC lateral elements (Cohen et al., 2006). SYCP1, which appears at zygonema, is a major component of the SC central element, and is used as a marker of fully synapsed chromosome segments (for a review, see Cohen et al., 2006). Numerous Stra8–/– spermatocyte nuclei with short strands of fine-caliber SYCP3, characteristic of leptonema, were present (Fig. 5A). The Stra8–/– zygotene spermatocytes either showed a wild-type SC configuration (compare Fig. 5D with 5F) or a SC in a `bouquet configuration' (Fig. 5A-C). It is noteworthy that the latter, which reflects a brief and transient nuclear-organisation motif (for a review, see Scherthan, 2007), was not identified in WT spermatocytes with the protocol that we used to generate nuclear spreads. In WT pachytene spermatocytes, SYCP1 and SYCP3 decorated the axes of all 19 completely synapsed autosomes (Fig. 5E). The few, most-advanced, Stra8–/– nuclei that were presumed to be at a pachytene-like stage showed at least one fully synapsed chromosome pair (Fig. 5G). However, synapsis never approached completion in these spermatocytes, because they all contained a mixture of single axial elements (corresponding to univalents; asterisk in Fig. 5G,K), and completely synapsed (arrowhead, Fig. 5G) and partly synapsed (thin white arrows, Fig. 5G,P) chromosomes. They also displayed intricate networks of SCs, indicative of heterosynapsis (large blue arrows, Fig. 5P,Q). Altogether, these observations indicate that SYCP3 is loaded as normal on chromosome cores at leptonema in Stra8–/– spermatocytes. SYCP3 also participates in the formation of axial elements at zygonema but, subsequently, the assembly of the SC is both deficient (leading to asynapsis) and aberrant (yielding heterosynapsis) in Stra8–/– spermatocytes.

Fig. 3.

Spermatogenesis in Stra8–/– mutants can progress to an early pachytene-like stage. Histological sections through the testes of WT and Stra8–/– mutants at the age of 8 weeks are shown. (A,B) Overview of seminiferous-tubule sections stained with haematoxylin and eosin. (C) Percentages of the most advanced spermatocytes detected on haematoxylin- and eosin-stained histological sections of 8-week-old WT and Stra8–/– mutants. PR (preleptotene), L (leptotene), Z (zygotene), P (pachytene), D (diplotene) and M (metaphase I) indicate the most advanced spermatocytes found in a given seminiferous-tubule section, whereas 0 indicates tubule sections containing only Sertoli cells, spermatogonia and eventually abnormal metaphase-like cells (M*, see the main text for further details). Stra8–/– mutants display a relative excess in premeiotic (i.e. preleptotene) and early meiotic (i.e. leptotene) spermatocytes, together with a marked deficit in the more advanced zygotene or pachytene spermatocytes and no diplotene spermatocytes. (D-I) Semi-thin histological sections stained with toluidine blue demonstrate the presence, in Stra8–/– mutants, of spermatocyte nuclei with zygotene (Z) and early-pachytene (eP) nuclear morphologies. The latter represent the most developmentally advanced germ-cell types present in Stra8–/– testes, because nuclei with an XY body (the condensed chromatin structure containing the sex chromosomes) that are characteristic of mid-pachynema (see XY in H) were never seen in Stra8–/– testes. S, Sertoli cells. (J-M) Immunofluorescence staining for SYCP3 and TRIM24. (J,K) The SYCP3 immunolabelling pattern (red), which is characteristic of early-pachytene spermatocytes (eP), is shown. The DAPI nuclear counterstain is blue. R, round spermatids. (L,M) TRIM24 (red), which is not detected in WT maturing spermatocytes prior to mid-pachynema, is undetectable in mutant meiocytes. The DAPI nuclear counterstain is blue. SG, spermatogonia. Roman numerals in D,F,H,L designate stages of the seminiferous-epithelium cycle (Russell et al., 1990). Scale bar: 80 m (A,B); 15 m (D-K); 40 m (L,M).

We also monitored the first meiotic prophase using antibodies directed against the phosphorylated form of the histone variant H2AX and the RAD51 recombinase. Phosphorylated H2AX (called H2AX) is a marker of the DSB during leptonema and zygonema (Mahadevaiah et al., 2001), and it also occurs at pachynema as part of a signalling pathway involved in transcriptional silencing of unpaired chromatin (Turner et al., 2005). In Stra8–/–, as in WT, spermatocytes, H2AX was distributed throughout the nuclei at zygonema (Fig. 5H,J). However, the normal restriction of H2AX to the chromatin of the XY body, typical of WT pachytene spermatocytes (Fig. 5I), was never seen in the most advanced Stra8–/– spermatocyte nuclei (Fig. 5K). Instead, H2AX was retained on almost every chromosome, with the notable exception of synapsed segments (arrows in Fig. 5K). RAD51, which catalyses strand invasion during homologous recombination, was detected in numerous foci at zygonema in WT spermatocytes (Fig. 5L), whereas its expression markedly decreased at pachynema (Shinohara et al., 1997; Barlow et al., 1997) (Fig. 5M). This decrease was not observed in the most advanced Stra8–/– spermatocytes (Fig. 5N,O). The presence of H2AX and of RAD51 foci on chromosomes from Stra8–/– mutants demonstrates that the formation of genetically programmed DSBs, the initiating event of meiotic recombination, takes place in the absence of STRA8. The retention of these two meiotic markers in Stra8–/– spermatocytes correlates with asynapsis [see above, and Turner et al. (Turner et al., 2005) and references therein].

We finally probed meiotic chromosomes with an antibody directed against the mismatch-repair protein MLH3, a marker of chiasmata (Marcon and Moens, 2003). WT mid-pachytene chromosomes displayed, as expected, one to two MLH3 foci per synapsed homologue, whereas no MLH3 signal was detected on chromosomes of Stra8–/– zygotene or early-pachytene-like spermatocytes (not shown). Thus, although Mlh3 transcripts were expressed in Stra8–/– spermatocytes (see above), the protein was not detected, probably because meiotic progression of mutant spermatocytes arrests prior to the stage at which late recombination nodules are assembled. In any event, absence of MLH3 foci indicates that meiotic recombination is not achieved in Stra8–/– males.

Stra8-null mutants display aberrant mitotic figures of spermatocyte origin

In testes from adult Stra8–/– mutants, 22±6% (mean ± s.d.; n 3) of the seminiferous-tubule sections were filled with aberrant cells containing patches of condensed chromatin that were reminiscent of apoptotic bodies or of chromosomes (M* in Fig. 3B,E and Fig. 6A,B). These aberrant cells appeared to be embedded in the seminiferous epithelium and were not sloughed off in the lumen of the seminiferous tubules. Their dense patches of chromatin were not labelled in TUNEL assays (Fig. 6K). Instead, they strongly reacted with an antibody against the phosphorylated form of histone H3 (M* in Fig. 6E,F; compare with normal meiotic metaphases, M in Fig. 6C,D), indicating that they represented highly condensed chromosomes (Cobb et al., 1999b). These chromosomes displayed a metaphase-like topography and were associated with an abnormal, monopolar and CDK5-negative spindle (compare M* and M in Fig. 6G-J) (Session et al., 2001). Importantly, BrdU was detected in the metaphase-like cells at 24 hours post-injection (M*, Fig. 6P,Q), which is the time required for WT as well as for Stra8–/– spermatocytes to enter meiosis (i.e. to move from preleptonema into leptonema, Fig. 6L-O). The metaphase-like cells and the WT metaphase spermatocytes both showed SYCP3 immunostaining on centromeres (M*, Fig. 7C,D,K, compare with M, Fig. 7A,B,I) [also see Kouznetsova et al. (Kouznetsova et al., 2005) and references therein], but they differed with respect to H2AX expression and chromosome number. In fact, all metaphase-like cells were immunostained for H2AX (M*, Fig. 7G,H,L), in contrast to WT cells in meiotic metaphase (M, Fig. 7E,F,J). In addition, the metaphase-like cells present in testes of Stra8–/– mutants all contained 40 unpaired univalents (n 200 nuclei from four mutant mice), similarly to WT mitotic metaphases (Fig. 7N), but unlike the configuration of 19 bivalents (plus X and Y chromosomes) normally found at metaphase I (Fig. 7M). Finally, 53±20% (mean ± s.d., n 3 mice) of the seminiferous-tubule sections containing metaphase-like cells also displayed leptotene spermatocyte nuclei, identified according to chromatin morphology (L, Fig. 6A) and diffuse nuclear distribution of SYCP3 (L, Fig. 7A-D); otherwise, the metaphase-like cells were the only meiocytes present in the tubule sections. Thus, the detection of SYCP3 and the kinetics of BrdU incorporation demonstrate together that the metaphase-like cells present in testes of Stra8–/– mutants are derived from leptotene spermatocytes, whereas detection of H2AX and the kinetics of BrdU incorporation demonstrate together that Stra8–/– leptotene spermatocytes containing unrepaired DSBs can undergo premature chromosome condensation.

Fig. 4.

Stra8–/– preleptotene spermatocytes can achieve a premeiotic S phase and express the meiotic cohesins Rec8 and Smc1b. Histological sections through the testes of WT and Stra8–/– mutants at the age of 8 weeks are shown. (A-G) Photomicrographs illustrating that Stra8 transcripts (purple and red signals) are expressed (1) in all preleptotene spermatocytes and (2) in differentiating A spermatogonia that were identified as A1-A3 on the basis of their presence at precise stages of the seminiferous-epithelium cycle (Chiarini-Garcia and Russell, 2001). (H,J) Immunohistochemical detection of BrdU (red signal) in WT and Stra8–/– preleptotene spermatocytes at 2 hours after its administration. (I,K) Show the same seminiferous tubules as in H,J, but stained with haematoxylin and eosin to allow identification of the cell types on the basis of their nuclear morphology. (L-Q) Immunohistochemical detection of BrdU in seminiferous tubules 8 days after its administration: BrdU that was initially incorporated in preleptotene spermatocytes has been transferred to spermatocytes with pachytene nuclear morphologies both in WT and Stra8–/– males. Note that M,P correspond to superimpositions of fluorescent signals (brown false colour), and of an haematoxylin and eosin stain of the tubule sections displayed in L,O. Also note that N,Q are high-magnification views of the areas delineated by boxes in L,M,O,P. (R-Y) Detection of Rec8 and Smc1b transcripts in the seminiferous epithelium. A1/A2/B, A1/A2/B spermatogonia; P, pachytene spermatocytes; PR, preleptotene spermatocytes; R, round spermatids; S, Sertoli cells; St10/St15/St16, step 10/15/16 spermatids. Roman numerals designate stages of the seminiferous-epithelium cycle. Note that, in C,E,G,S,U,W,Y, the in situ hybridisation signals were converted to a red false colour, and the DAPI nuclear stain was converted to a bright-field image and then to a blue false colour using Photoshop. Scale bar: 80 m (A,H-P); 30 m (B-G,N,Q); 40 m (R-Y).

Fig. 5.

Abnormal synapsis in Stra8–/– spermatocytes. Immunostaining of surface-spread nuclei from 4-week-old males using antibodies to SYCP1, SYCP3, H2AX and RAD51, as indicated, is shown. L/Z/P, leptotene/zygotene/pachytene spermatocytes; Z/P, designates the most advanced spermatocytes that were present in the mutant testes – these cells shared characters common to zygonema and pachynema. (A-G,P,Q) Formation of the SC is assessed by immunofluorescence on nuclear spreads. The axial elements and the lateral elements of the SC are stained with anti-SYCP3 antibodies (green). The central elements of the SC are stained with anti-SYCP1 antibodies (red). The SC appears yellow from overlapping of SYCP1 and SYCP3 signals. (A) Early leptonema is indistinguishable between WT (not shown) and Stra8–/– spermatocytes. (A-C) Bouquet configurations are unique to Stra8–/– spermatocytes. (D,F) Short fragments of the SC (yellow) have begun to form both in WT and Stra8–/– spermatocytes at early zygonema. (E) In WT pachytene spermatocytes, 19 fully synapsed bivalents are visible along with the XY pair. (G,P,Q) In the Stra8–/– Z/P spermatocytes, the chromosomes are either not synapsed (asterisk) or partially synapsed (white arrows), although fully synapsed chromosome pairs are occasionally seen (arrowheads). In addition, lateral elements are frequently heterosynapsed (blue arrows). (H,J) WT and Stra8–/– spermatocytes express H2AX (red signal) throughout the nucleus during zygonema. (I) In WT spermatocytes during pachynema, H2AX expression becomes restricted to the (almost) unsynapsed XY body. (K) In Stra8–/– spermatocytes at the Z/P stage, H2AX remains widely distributed throughout the nucleus, except in the synapsed regions (arrowed thick green strands), reflecting the near-absolute lack of homologous pairing in these spermatocytes. (L,M) In WT spermatocytes, the number of RAD51 foci (green signal) sharply decreases during the progression from zygonema to pachynema. (N,O) In the most advanced Stra8–/– spermatocytes, the density of RAD51 foci is abnormally elevated but, similar to the WT situation, displays an inverse correlation with the extent of synapsis. Scale bar: 10 m (A-O); 4 m (P,Q).

Stra8–/– mutants at P10 displayed only rare seminiferous-cord sections (1.1±0.4%; mean ± s.d., n 3) containing centrally located metaphase-like cells, whereas, at P15, these cells populated 33±11% (mean ± s.d.; n 3) of the cord sections. Therefore, the timing of appearance of the metaphase-like cells matches the normal onset of the meiotic prophase during the pre-pubertal wave of spermatogenesis (Nebel et al., 1961).


STRA8 is not required for initiating meiosis in the male germ line

Initiation of meiosis can be (immuno-)cytologically recognised at leptonema because chromosome condensation begins, cohesin proteins are loaded onto the chromosomes and installation of the axial element (the precursor of the SC), as well as DSB formation, are initiated.

Stra8–/– testes at P15 and in young adults contain a large number of spermatocytes that have features characteristic of leptonema, namely nuclear morphologies, extensive H2AX phosphorylation, and the presence of short strands of immunolabelled SYCP3, attesting initiation of (1) meiotic chromosome condensation, (2) meiotic recombination and (3) axial element formation, respectively. The mutant testes also show an abundance of cells with a preleptotene nuclear morphology that, similarly to their WT counterparts, undergo premeiotic DNA replication, as assessed by BrdU-labelling assays, and express the meiotic-specific Rec8 and Smc1b cohesin transcripts. Stra8–/– preleptotene and leptotene spermatocytes also express Spo11, the product of which initiates meiosis by inducing the formation of DSBs (for a review, see Cohen et al., 2006). Moreover, because numerous RAD51 foci form and stay along the chromosomal cores, the removal of SPO11 and resection of 5′ DSB ends must not be affected in Stra8–/– males.

Fig. 6.

Stra8–/– aberrant meiocytes, displaying metaphase-like aspects and abnormal spindles, are derived from leptotene spermatocytes. Histological sections at 8 weeks of age are shown. (A,B) Staining with haematoxylin and eosin (A) and toluidine blue (B): the aberrant cells display evenly matched patches of condensed chromatin. (C-J) Immunofluorescence staining (red signal) and DAPI nuclear counterstain (blue) of histological sections for detection of: (C-F) the phosphorylated form of histone H3 (pH3), a marker of highly condensed (i.e. diplotene and metaphase) chromosomes; (G,H) -tubulin, a marker of mitotic spindles; and (I,J) CDK5, a marker of the meiotic spindle. Note that the abnormal metaphase-like chromosomes (in H) are spherically arranged in rosette configurations instead of being neatly aligned at the cell equator (in G). (K) Superimposition of two adjacent sections processed for detection of apoptotic cells (TUNEL assay, green signal) and for detection of pH3 (red signal). The absence of colocalisation of the green and red signals indicates that the Stra8–/– aberrant meiocytes do not undergo apoptosis. (L-Q) Immunohistochemical detection of BrdU (red signals) in seminiferous tubules 24 hours after its administration. (L-O) BrdU that was incorporated into preleptotene spermatocytes (see Fig. 3) has been transferred to leptotene spermatocytes in both WT and Stra8–/– testes. (P,Q) In Stra8–/– testes, BrdU has also been transferred to the aberrant metaphase-like meiocytes during this 24-hour period. M,O,Q correspond to the same tubule sections as in L,N,P, respectively, now stained with haematoxylin and eosin. The insets in M,O,Q correspond to high magnifications of the meiocytes arrowed in L,N,P, respectively. L/D, leptotene/diplotene spermatocytes; M, cells in meiotic metaphase (WT); M*, metaphase-like (Stra8–/–) cells; S, Sertoli cells. Roman numerals designate stages of the seminiferous-epithelium cycle. Scale bar: 40 m (A,C-F,I,J,L-Q); 20 m (B); 10 m (G,H); 160 m (K).

Our data indicate therefore that STRA8 is dispensable for initiating meiosis in the male mouse. By contrast, germ cells of female mice lacking Stra8 fail to undergo premeiotic DNA replication, to initiate meiotic recombination and to condense chromosomes (Baltus et al., 2006). Thus, although STRA8 is required for fertility in both males and females, it displays, similarly to a variety of other meiotic proteins, sexually dimorphic functions (for a review, see Morelli and Cohen, 2005).

STRA8 regulates the switch between mitotic and meiotic cell cycles

In Stra8–/– P15 and young-adult mutants, about one of four seminiferous-tubule segments are filled with mitosis-like figures containing highly condensed chromosomes, as assessed from their phosphorylated histone-H3 content. These metaphase-like cells express the meiotic-specific protein SYCP3 but, unlike normal meiotic metaphases, contain 40 hypercompacted univalents, display an aberrant H2AX content and are associated with structurally abnormal monopolar mitotic spindles that, unlike normal meiotic spindles, lack CDK5. The metaphase-like cells are generated within 24 hours following the mitotic S phase, which, both in WT and Stra8–/– testes, corresponds to the time of progression from preleptotene to leptotene. Moreover, within a given seminiferous-tubule segment, these cells are often associated with leptotene spermatocytes, but not with other meiocytes and, during pre-pubertal testes development, the appearance of these two cell-types occurs according to the same schedule. Altogether, these data indicate that the metaphase-like cells are derived from leptotene spermatocytes – the chromosomes of which have undergone rapid condensation and aberrant hypercompaction – prior to the time of synapsis and DSB resolution.

Extensive generation of univalent chromosomal configurations during meiosis I also occurs in REC8-, EXO1-, MLH1- and MLH3-deficient spermatocytes (Lipkin et al., 2002; Eaker et al., 2002; Wei at al., 2003; Xu et al., 2005). However, Rec8, Mlh1 and Mlh3 are expressed in Stra8–/– mutants, thereby suggesting that Stra8 is not epistatic to any of the three former genes. By contrast, the premature condensation of univalent chromosomes is a pathognomonic feature of the Stra8–/– phenotype, because this phenotype has never been reported in any mammalian meiotic mutant, including those that, similarly to Stra8–/– mutants, display univalent chromosomes at prophase I and/or extensive asynapsis or heterosynapsis (for a review, see Cohen et al., 2006). In this context, it is interesting to note that, although precocious condensation of meiotic chromosomes had, up to now, never been reported in vivo, WT leptotene spermatocytes have nevertheless the capacity to fully condense univalents upon treatment with the phosphatase inhibitor okadaic acid (Cobb et al., 1999a). This indicates that, already at the onset of prophase I, spermatocytes contain a biochemical machinery that is capable of triggering rapid chromosome condensation, but which is kept silent under physiological conditions.

Fig. 7.

Characterisation of the metaphase-like cells in Stra8–/– testes. (A-H) Immunodetection of SYCP3 (green signal), H2AX (red signal) and DAPI nuclear counterstain (blue) on histological sections. (I-L) Double immunolabelling for detection of SYCP3 (green signal), H2AX (red signal) and DAPI nuclear counterstain (blue) on nuclear spreads. (M-P) Giemsa-stained metaphase spreads. Note that panels M,N correspond to WT meiotic and mitotic (i.e. spermatogonial) metaphases, respectively. L/Z, leptotene/zygotene spermatocytes; M, cells in meiotic metaphase (WT); M*, metaphase-like (Stra8–/–) cells; XII, stage XII of the seminiferous-epithelium cycle. Scale bar: 30 m (A-H); 10 m (I-P).

That STRA8 might be a component of a yet unidentified cell-cycle surveillance system (i.e. a checkpoint) that senses meiotic errors and eliminates cells containing unresolved defects upon initiating a kind of `mitotic catastrophe' (for a review, see Bucher and Britten, 2008) is unlikely because such a function would be redundant with that of the evolutionarily conserved, okadaic-acid-sensitive pachytene checkpoint control (Roeder and Bailis, 2000). In fact, the pachytene checkpoint appears sufficient to inhibit cell-cycle progression in response to any defect in recombination or chromosome synapsis. Thus, the most probable explanation of the premature chromosome condensation is that Stra8–/– preleptotene spermatocytes are disabled in their capacity to depart from the mitotic cycle. According to this scenario, Stra8–/– spermatocytes entering leptonema rapidly condense chromosomes, which acquire a metaphase-like morphology within 24 hours instead of the 13 days normally required to reach the metaphase-I stage (Oakberg, 1956). Because these chromosomes fail to align properly on a mitotic spindle, the cells probably die in situ through transcriptional arrest at a persistent metaphase-like stage (for reviews, see Roninson et al., 2001; Blagosklonny, 2007). Therefore, we propose that STRA8 is required to regulate the switch from the mitotic pattern of cell division, which is that of their immediate precursors, the B spermatogonia, to a meiotic pattern. As pointed out in the introductory section, the timing of this switch is strikingly different in males and females. The notion that Stra8–/– spermatocytes retain a mitotic pattern of cell division is nevertheless consistent with the proposed role of STRA8 in committing female mouse germ cells to meiosis (Baltus et al., 2006; Pawlowski et al., 2007; Wolgemuth, 2006). This commitment occurs in females gonocytes during embryogenesis (Baltus et al., 2006), whereas it is established postnatally in the male – i.e. in a totally different cellular context, most probably in preleptotene spermatocytes but, in any event, not in B spermatogonia, which never express Stra8.

Interestingly, a similar function in controlling the switch between mitosis and meiosis has been ascribed to the maize ameiotic1 (Am1) gene product (for reviews, see Hamant et al., 2006; Pawlowsky et al., 2007). Furthermore, Am1 mutant meiocytes display sexually dimorphic abnormal phenotypes, which are strikingly similar to those of Stra8–/– meiocytes (Baltus et al., 2006) (this study). Thus, STRA8 and AM1 might be functionally homologous, even if they have no sequence homologues outside of vertebrates and plants, respectively.

STRA8-dependent promotion of homologue association probably occurs before the initiation of SC assembly

A small percentage of Stra8–/– leptotene spermatocytes escape the phenomenon of premature chromosome condensation and progress into later stages of meiosis, thereby reaching an abnormal state displaying both zygonema and pachynema characteristics. The latter cells exhibit a marked deficiency in chromosomal synapsis, attested by persistent axial elements, extensive H2AX immunostaining and highly abundant RAD51 foci (Kuroda et al., 2000; Turner et al., 2005; Kneitz et al., 2000). It must be emphasised that STRA8 is unlikely to play a direct role in SC assembly because it is not detected in germ cells beyond leptonema. By contrast, Stra8–/– leptotene spermatocytes often display bouquet-stage abnormalities, a condition that has been correlated with the occurrence of asynapsis and heterosynapsis at later meiotic stages [see Pandita et al. and Liebe et al. (Pandita et al., 1999; Liebe et al., 2006) and references therein].

Bouquet formation appears to be a consistent motif of meiotic prophase of the vast majority, if not all, of eukaryotic species (for a review, see Scherthan, 2007). `Bouquet' designates a three-dimensional nuclear-organisation motif that is set up at the leptotene-zygotene transition and results from the clustering of telomeres to a restricted site on the nuclear envelope. It is thought to promote homologous presynaptic alignment (i.e. pairing) (for a review, see Zickler, 2006), and thereby to facilitate the sorting process of homologues prior to their synapsis by aligning the ends of the chromosomes and restricting the homology search to a small volume of the nucleus. In the male mouse, the bouquet arrangement of chromosomes resolves soon after the initiation of synapsis (i.e. early zygonema) and renders only a very low percentage of bouquet nuclei readily detectable under optimal experimental conditions (for a review, see Scherthan, 2007).

The elevated occurrence of spermatocyte bouquet arrangements encountered in Stra8–/– testes suggests that the bouquet stage is maintained for a considerably longer period in the absence of STRA8 and indicates defects in chromosome pairing, which in turn can lead to improper synapsis because homologous chromosomes are not aligned correctly (Scherthan et al., 2000; Harper et al., 2004). As a result, defects in homologue pairing can be seen as asynapsis, incomplete synapsis or synapsis between non-homologous chromosomes on nuclear spreads of the few Stra8–/– spermatocytes exhibiting early-pachytene-like nuclear morphologies.

The complex nature of the Stra8–/– phenotype strongly suggests that STRA8 is required for essential early meiotic events, notably homologous-chromosome pairing, in addition to being involved in the switch from mitosis to meiosis. Identification of proteins that interact with STRA8 during preleptonema should provide invaluable information regarding these functions.

Materials and Methods


All mice, which had a mixed C57BL/6 (50%) 129/Sv (50%) genetic background, were housed in an animal facility licensed by the French Ministry of Agriculture (agreement N°B67-218-5) and all animal experiments were supervised by NBG (agreement N°67-205), in compliance with the European legislation on care and use of laboratory animals.

Generation and genotyping of the mice

To construct the targeting vector for homologous recombination in ES cells (see Fig. 1A), a 10-kb-long fragment isolated from a 129/Sv mouse genomic DNA library and containing exons 1-5 was inserted into pBluescript II SK+ (pBS-SK, Stratagene). A loxP-flanked neomycin-resistance cassette (neo) was cloned into the Eco47III site located upstream of exon 5. Oligonucleotides 5′-CATGGTCGACATAACTTCGTATAATGTATGCTATACGAAGTTATA-3′ and 5′-CATGTATAACTTCGTATAGCATACATTATACGAAGTTATGTCGAC-3′, containing SalI and loxP sites, were then inserted into the NcoI site located upstream of exon 2, which contains the ATG codon. This targeting vector was linearised and electroporated into P1 ES cells. One clone (#PT1) that targeted as expected was identified out of 372 G418-resistant clones screened by long-range PCR (L3 allele, Stra8+/L3) and then verified by Southern blot analysis (data not shown). Transient transfection of #PT1 ES cells with a Cre-expressing plasmid allowed excision of the loxP-flanked region, yielding cells bearing the excised, null, L– allele. This clone was injected into C57BL/6 blastocysts and the chimeras transmitted the modified allele (L–) to their germ line, yielding heterozygous mice (Stra8+/L–). Tail DNA was genotyped by PCR using Expand Long Template PCR System (Roche) with primers #1 (5′-AGCTTGCAGTCTACTGAAGG-3′) and #2 (5′-CTGGGTTGGTTGCCTTCTCC-3′) (see Fig. 1A) to amplify the WT allele (3798-bp long), and primers #1 and #3 (5′-CACATGTGTACACACTTGCCAAG-3′) (see Fig. 1A) to amplify the L– allele (3854-bp long). Conditions were ten cycles with denaturation at 92°C for 15 seconds, annealing at 55°C for 30 seconds and elongation at 68°C for 4 minutes, followed by 20 cycles with denaturation at 92°C for 15 seconds, annealing at 55°C for 30 seconds and elongation at 68°C for 4 minutes, extended by 20 seconds for each successive cycle.

RNA and protein analyses

Total RNA from P15 testes was extracted using Trizol Reagent (Invitrogen). Analysis of mRNA was carried out by reverse transcription coupled to PCR amplification (RT-PCR). Reverse transcription of 1 g total RNA was performed in 20 l using QuantiTect Reverse Transcription Kit (Qiagen). Then, 2-l aliquots of diluted cDNA (1:50) were amplified using Taq DNA polymerase (Roche) with primers #4 (5′-CACAAGTGTCGAAGGTGCAT-3′) and #5 (5-CTTCAGCATCTGGTGGAACA-3′) located in exons 1 and 5, respectively (see Fig. 1C). The other primers that were used to amplify meiotic-specific genes were as indicated in supplementary material Table S1. Conditions were 30 cycles with denaturation at 92°C for 15 seconds, annealing at 60°C for 15 seconds and elongation at 72°C for 30 seconds. The amplified fragments were resolved by electrophoresis on a 2% (w/v) agarose gel stained with ethidium bromide.

Total protein extracts (100 g) were resolved on 10% SDS-PAGE gels and blotted onto nitrocellulose membranes (Schleicher & Schuell). STRA8 and actin were detected using ab49602 (Abcam) and A-2066 (Sigma) rabbit polyclonal antisera, respectively, at a dilution of 1:500. Immunoreactions were visualised using protein A coupled to horseradish peroxidase (dilution 1:5000), followed by chemiluminescence according to the manufacturer's protocol (Amersham).


All histological observations, as well as analyses involving immunohistochemistry, in situ hybridisation and TUNEL reactions, were repeated on at least three mice per age group. All slides were examined at room temperature using a DMLA microscope (Leica) with 10×, 20×, 40× and 100× objectives with apertures of 0.3, 0.5, 0.7 and 1.3, respectively. Images were taken with a digital camera (CoolSnap; Photometrix) using CoolSnap v.1.2 software and then processed with Photoshop CS2 v.9.0.2 (Adobe). Testes destined for haematoxylin and eosin staining were fixed in Bouin's fluid for 48 hours and embedded in paraffin; sections were cut at a thickness of 5 m. Testes destined for epon embedding were perfusion-fixed with 2.5% (w/v) glutaraldehyde in phosphate-buffered saline (PBS) and processed according to standard procedures; semi-thin (i.e. 1-m thick) sections were stained with toluidine blue.

Preparation of germ-cell nuclear spreads

Nuclear spreads destined for immunostaining were obtained from testes at 4 weeks of age (i.e. when meiosis is complete in WT mice) to enrich for meiotic stages and were carried out as described in Peters et al. (Peters et al., 1997). Metaphase spreads of spermatocytes and spermatogonia destined for karyotype analyses were obtained from adult (8-week old) testes, were prepared as described in Peters et al. (Peters et al., 2001) and were stained with Giemsa.

BrdU incorporation and TUNEL assays

Detection of BrdU incorporation and of apoptotic cells was as described (Ghyselinck et al., 2006). In brief, BrdU (Sigma) dissolved in PBS was injected intraperitoneally at 50 mg per kg of body weight. Males were hemi-castrated 2 hours after the injection and killed either 24 hours or 8 days later. Testes were fixed in Bouin's fluid for 24 hours and then embedded in paraffin. BrdU incorporation was detected by using an anti-BrdU mouse monoclonal antibody (diluted 1:100; Roche) and indirect immunofluorescence labelling. For detection of apoptotic cells (TUNEL assays), testis samples were fixed in 4% (w/v) paraformaldehyde (PFA) in PBS for 16 hours at 4°C, then embedded in paraffin. TUNEL-positive cells were detected using the In Situ Cell Death Detection Kit, Fluorescein (Roche), and the sections were counterstained with 0.01% (w/v) 4′,6-diamidino-2-phenylindoledihydro-chloride (DAPI) in Vectashield (Vector Laboratories).


For immunodetection of STRA8, 5-m sections of freshly frozen testes were post-fixed for 5 minutes in ice-cold 4% (w/v) PFA in PBS. For immunodetection of CDK5, we used a similar protocol, except that the sections were treated for 5 minutes with acetone at –20°C prior to post-fixation for 3 minutes in ice-cold 4% (w/v) PFA in PBS. The sections were washed in PBS then incubated for 1 hour at room temperature with either the anti-STRA8 rabbit antibody [diluted 1:100 (Oulad-Abdelghani et al., 1996)], the anti-STRA8 ab49602 rabbit antibody (1:500, from Abcam) or with the anti-CDK5 rabbit antibody (diluted 1:100, sc-173, Santa Cruz Biotechnologies). For immunodetection of -tubulin, SYCP3, phosphorylated histone H3 and H2AX, testes were fixed by intracardiac perfusion of ice-cold 4% (w/v) PFA in PBS, then kept in the same fixative overnight at 4°C, washed in PBS, dehydrated and either frozen in nitrogen vapours, after cryoprotection through immersion in graded sucrose solutions, for detection of -tubulin or otherwise embedded in paraffin. Prior to immunostaining, the histological sections from paraffin-embedded tissues were collected on glass slides, immersed into 0.01 M sodium-citrate buffer and exposed to microwave treatment (power output 800 W; 2×5 minutes). The sections were incubated for 1 hour at room temperature with the mouse monoclonal anti- -tubulin antibody (diluted 1:400, T4026, Sigma Aldrich), the anti-SYCP3 antibody (diluted 1:100, ab15091, Abcam), anti-phosphorylated-histone-H3 (Thr3) rabbit antibody (diluted 1:100, 9714S, Upstate) and anti-H2AX mouse monoclonal antibody (diluted 1:500; clone JBW301, Upstate). Detection of bound primary antibodies was achieved by incubating the section for 45 minutes at room temperature using appropriate Cy3- or Alexa-Fluor-488-conjugated secondary antibodies. The sections were counterstained with DAPI.

For immunostaining of nuclear spreads using antibodies for SYCP1, SYCP3, RAD51, MLH3, phosphorylated histone H3 and H2AX, we followed the same protocol as above, but omitted microwave pretreatment of the samples. The anti-SYCP1 goat antibody (sc-20837, Santa Cruz Biotechnologies), anti-RAD51 rabbit antibody (PTC130T, Calbiochem) and anti-MLH1 mouse antibody (clone G168-728, Pharmingen) were diluted 1:50, 1:100 and 1:50, respectively.

In situ hybridisation assays

Testes were fixed by intracardiac perfusion of ice-cold 4% (w/v) PFA in PBS, and then kept in the same fixative overnight at 4°C, washed in PBS, cryoprotected and finally frozen in liquid-nitrogen vapours. Tissue sections were cut at a thickness of 10 m. In situ hybridisation using digoxigenin-labelled probes was performed as described (Vernet et al., 2006b). The templates were a 1.3-kb-long fragment of Stra8 cDNA encompassing the complete open reading frame (Oulad-Abdelghani et al., 1996), a 1.1-kb-long Rec8 cDNA fragment (encompassing position 828 to 1594 in NM_020002) and a 0.7-kb-long Smc1b cDNA fragment (encompassing position 3004 to 3703 in NM_080470) cloned into pDrive (Qiagen).


We acknowledge valuable discussions with Françoise Dantzer, Régine Losson and Yves Rumpler, as well as expert technical assistance from Nadia Messaddeq and Muriel Klopfenstein. We thank our twoanonymous referees for their helpful comments and corrections of the manuscript. This work was supported by funds from the Fondation pour la Recherche Médicale (FRM), the Centre National de la Recherche Scientifique (CNRS), the Institut National de la Santé et de la Recherche Médicale (INSERM) and the Hôpital Universitaire de Strasbourg. H.J. and N.V. were supported by French Ministry of Research and Technology (MRT) and Association pour la Recherche sur le Cancer (ARC) fellowships, respectively. C.-A.C. was a recipient of a post-doctoral `Charcot' grant allocated by the French Ministry of Foreign Affairs.


  • Accepted July 7, 2008.


View Abstract