No Circadian Rhythms in Testis: Period1 Expression Is Clock ...

0888-8809/03/$15.00/0 Printed in U.S.A.

Molecular Endocrinology 17(1):141–151 Copyright © 2003 by The Endocrine Society doi: 10.1210/me.2002-0184

No Circadian Rhythms in Testis: Period1 Expression Is Clock Independent and Developmentally Regulated in the Mouse DAVID MORSE*†, NICOLAS CERMAKIAN*‡, STEFANO BRANCORSINI, MARTTI PARVINEN, PAOLO SASSONE-CORSI

AND

Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire, Centre National de la Recherche Scientifique-Institut National de la Sante´ et de la Recherche Me´dicale-Universite´ Louis Pasteur (D.M., N.C., S.B., P.S.-C.), B.P. 10142, 67404 Illkirch-Strasbourg, France; and Department of Anatomy (M.P.), University of Turku, 20520 Turku, Finland Spermatogenesis is a process whereby haploid spermatozoa differentiate through meiosis from precursor stem cells. We examined the expression of circadian clock genes in the testis, to assess clock control over the timing of different developmental events. Clock genes are known to oscillate with circadian rhythmicity in the central clock structure, the suprachiasmatic nucleus of the hypothalamus, but also in peripheral tissues. Here we show that Per1 gene expression in the testis is constant over a 24-h period and that the Per1 transcript is expressed at a level higher than the peak values of the Per1 oscillations observed for other tissues. Bmal1, another clock gene whose expression oscillates in other tissues, also shows constant expression levels in the testis. In addition, the

levels and phosphorylation state of the PER1 protein are not oscillating at all times of day. Strikingly, Per1 is restricted primarily to step 7 to 10 spermatids and thus appears to be developmentally regulated. The expression of the Clock transcript is also developmentally regulated, but it is found principally in spermatogonia and spermatocytes up until the time of the first meiotic division. Per1 expression is not altered in testes from Clock mutant mice, suggesting that CLOCK does not activate Per1 in male germ cells, in contrast to what it does in other mouse tissues. Taken together, our observations suggest that the testis, in contrast to all other peripheral tissues, lacks a functioning circadian clock. (Molecular Endocrinology 17: 141–151, 2003)

S

velopmental steps found adjacent to one another (3). In cross-section, the spermatogonia, spermatocytes, and spermatids are each uniformly distributed in distinct concentric rings or layers. Spermatogonia are found at the periphery, in contact with the basal lamina of the tubule, whereas the spermatids are closer to the lumen. The development of a particular germ cell type is synchronized with its neighbors, so that in a given cross-section, all the cells in a given concentric ring are in the same developmental phase (3, 6). Remarkably, although one might think a cross-section could contain any mixture of cell types, only certain combinations of germ cell types are ever found. These associations of different cell types are called stages (numbered in roman numerals from I to XII in the mouse, or from I to XIV in the rat) (6). The constant association of particular cell types in stages suggests that the developmental events occurring in spermatogonia, spermatocytes, and spermatids are all coordinated with one another during spermatogenesis. Although the mechanism by which this could occur is not well understood, one possibility is that internal time-keepers, or biological clocks, might influence developmental decisions taken by each of the different cell types. The most widely studied timekeeper is the circadian (daily) clock found in most cells, whether they be free living or part of a multicel-

PERMATOGENESIS IS A complex sequence of events that results in the formation of haploid spermatozoa from precursor stem cells. The process starts with the proliferation and differentiation of diploid spermatogonial stem cells to give rise to diploid spermatocytes, which undergo meiosis to produce round haploid spermatids. Diploid primary spermatocytes spend more than 90% of the time in prophase I (leptotene, zygotene, pachytene, diplotene, and diakinesis) with most of it in pachytene (1). The haploid round spermatids resulting from the second meiotic division are sculpted into their final elongate mature shape in the process of spermiogenesis, which involves extensive biochemical and morphological remodeling (2). The entire process requires about 35 d in mice (3). This differentiation program involves a number of highly specific transcriptional regulatory events (2, 4, 5). The development of germ cells within the seminiferous tubule is tightly regulated. In longitudinal sections, spermatogenesis occurs in synchronized waves within the seminiferous epithelium with successive deAbbreviations: BMAL1, Brain and muscle ARNT-like protein; CRE, cAMP response element; CREM, CRE modulator; LD, light-dark; ORF, open reading frame; RPA, ribonuclease protection assay; SCN, suprachiasmatic nucleus. 141

142 Mol Endocrinol, January 2003, 17(1):141–151

lular organism (7–11). The circadian clock has a genetic basis, and several genes have been implicated in its function (7–11). On a molecular level, the time of day for a cell is dictated by the cellular levels of one or more oscillating clock gene products. In essence, the clock gene products oscillate because of a delayed negative feedback loop involving both transcriptional and posttranscriptional steps (12, 13). The negative feedback loop is believed to involve transcription of the three Period (Per) and two Cryptochrome (Cry) genes, whose protein products accumulate approximately 6 h after their transcripts (14, 15). The PER and CRY proteins become hyperphosphorylated and appear to form heteromultimeric complexes, which are translocated to the nucleus (12, 16, 17). Nuclear CRYs interfere with binding of a heterodimeric transactivator to E box elements in their promoters, thus repressing transcription of both Per and Cry genes (16, 17). The positively acting heterodimer is constituted by two basic helix-loop-helix/period兾aryl hydrocarbon receptor nuclear translocator兾single-minded (PAS)-containing transcription factors called CLOCK and brain and muscle aryl hydrocarbon receptor nuclear translocator (ARNT)like protein 1 (BMAL1) (18–20). There is also evidence suggesting that PER2 may positively regulate the transcription of Bmal1, thus adding a positive feedback element to the clockwork (13). In mammals, Bmal1 transcript levels also cycle, but peak about 12 h out of phase with the Per transcript levels (21) [the Clock transcript levels do not cycle (13, 22)]. The approximately 24-h period of the oscillations in transcript levels appears to be due to a number of different posttranscriptional processes such as RNA and protein stability (23, 24), protein phosphorylation (12, 23, 25), and regulated nuclear entry (12, 24, 26). Although the circadian variables of different tissues in mammals appear synchronized in a hierarchical manner by a master clock within the suprachiasmatic nucleus (SCN) in the lower hypothalamus (27), nonneuronal tissues have been shown themselves to contain functioning clocks (28–30). For example, the same oscillations of clock gene transcripts observed in the SCN can also be observed in tissues such as liver, kidney, or muscle, albeit with a delay compared with the phase of the oscillation in the SCN (21, 31, 32). Moreover, rat tissues maintain oscillations of a Per1-driven reporter for several cycles after being put in culture (30, 33). Lastly, cultured cells, such as fibroblasts, can experience several cycles of clock gene transcript oscillations when treated with a serum shock (15, 28). Of most relevance to the current study, one group has reported Per gene products oscillating in the testis (32), suggesting that this organ contains a functioning clock similar to that in other tissues. We report here that, in contrast to what was previously reported (32) and unlike the situation in other tissues, we do not observe any change in the levels of the Per1 and Bmal1 transcripts in the testis. We also observe that PER1 protein levels do not oscillate, and

Morse et al. • Per1 Expression in Mouse Testis

PER1 does not exhibit a change in molecular weight suggestive of changes in phosphorylation state. In contrast to other tissues, Per1 and Clock transcripts are found in different cell types (spermatids for Per1; spermatogonia and pachytene spermatocytes for Clock) at all times, suggesting that expression of these genes may be developmentally controlled. Lastly, the levels of the Per1 transcript are the same in wild-type and in Clock c/c mutant mouse testis, but are reduced in cAMP response element modulator (CREM) ⫺/⫺ mouse testis. We conclude from these observations that Per1 gene expression in testis is not driven by CLOCK/BMAL1 dimers and suggest that the testis might represent the first described circadian clockfree tissue in mammals.

RESULTS Per1 Transcript Levels Are Elevated and Constant in the Testis We first used ribonuclease protection assays (RPAs) to quantitate the levels of Per1 RNA in extracts of mouse testes maintained under 12-h light, 12-h dark (LD) cycles. No significant difference was found in the amount of Per1 transcript at six different times under the LD cycle (Fig. 1A; shaded bars in panel C). Furthermore, semiquantitative in situ hybridization shows a similar Per1 pattern and expression level at all the different times (Fig. 1E). These constant Per1 transcript levels are in striking contrast with the daily variation observed in kidney or skeletal muscle (Fig. 1D). Furthermore, transcript levels in the testis are even higher than the peak Per1 levels in these other tissues, which occur around ZT12 (Fig. 1D). Based on the lack of variation of Per1 expression, we predicted constant levels of the Bmal1 transcript, which normally oscillates out of phase with the Per1 transcript with an amplitude of up to 10-fold in peripheral tissues (13, 21, 32). Indeed, no robust rhythmicity is seen in Bmal1 transcript levels measured in the same RNA preparations (Fig. 1B; solid bars in panel C). We conclude from these data that regulation of clock gene transcripts in the testis may be distinct from that in other tissues. PER1 Protein Levels Are Constant in the Testis The results of the Per1 RPA suggested that levels of the PER1 protein in testis might also be constant. To test this, Western blot analysis of testis extracts was performed with an affinity-purified PER1-specific antibody. Interestingly, two or three protein bands were regularly detectable with the anti-PER1 antibody in a testis extract, whereas in other tissues this was never observed (Fig. 1F). Possibly, as often seen in male germ cells, posttranscriptional modifications of the Per1 transcript could generate multiple isoforms. It would be interesting to explore this issue and to establish whether these putative isoforms are all nuclear


Mol Endocrinol, January 2003, 17(1):141–151

143

Fig. 1. Per1 Transcript and PER1 Protein Levels Do Not Oscillate in the Testis Mice under a 12:12 LD schedule were killed at 4-h intervals, and testes were frozen for RNA or protein extractions and in situ hybridization. For RPA, the total RNA for each sample was preequilibrated by ethidium staining and the protections performed with a ␤-actin probe included in each assay. RPA was performed to measure levels of Per1 (A) or Bmal1 (B) RNAs in testis RNA at six different times. C, The mean and SD deviation were calculated from data acquired with a PhosphorImager and represent either Per1/␤-actin (shaded bars; four experiments) or Bmal1/␤-actin (solid bars; two experiments). D, RPA was performed to compare Per1 RNA levels in the testis with levels in kidney and muscle at peak (ZT 12) and trough (ZT 0) times in the Per1 oscillation. A differently sized fragment of the ␤-actin probe is protected in skeletal muscle. All RPAs include 4 ␮g Escherichia coli tRNA (t) as control. E, Cross-sectioned testes were used for semiquantitative in situ hybridization with a Per1 sense or antisense probe. Representative autoradiograms are shown. F, Western blot was done on extracts from Per1 ⫺/⫺ and Per1 ⫹/⫹ mouse testes with an anti-PER1 antibody. An immunoreactive band with an apparent molecular mass of 175 kDa is present in wild-type and absent in Per1 ⫺/⫺ mice. G, Western blot was done on testis extracts prepared from mice killed at 4-h intervals under a LD cycle with an anti-PER1 antibody.

and all regulate transcription as the full-length PER1 protein. To establish whether the multiple PER1 immunoreactive proteins were all bona fide Per1 products, we compared testis extracts from Per1 ⫹/⫹ and ⫺/⫺ mice. All PER1-immunoreactive protein bands are

lacking in the extract from Per1 ⫺/⫺ mice, including the protein with an apparent molecular mass of 175 kDa corresponding to the full-length PER1 (Fig. 1F). Next, extracts taken at six different times under the LD regimen were analyzed, and in agreement with the lack of RNA oscillation, PER1 did not change in


amount over the daily cycle (Fig. 1G). We also note that the apparent molecular mass of the immunoreactive protein was the same at the different time points, suggesting that the phosphorylation state of the PER1 did not change with time. The constant cellular levels and phosphorylation state of PER1 suggests that a lack of Per1 mRNA rhythmicity may be due to an inability of PER1 to inhibit its own transcription. Per1 Expression Is Stage Specific The in situ hybridization analysis (Fig. 1E) showed that, whereas the overall level of Per1 expression was similar in cross-sections of the testis taken at different times, not all tubules in the cross-sections hybridized equally well to the probe. In fact, each testis section contains both strongly and weakly hybridizing tubules. This difference suggests that Per1 expression might be developmentally regulated. To address this issue, tubules from rat testes were dissected into different stages, RNA was prepared from pools of the different stages, and the levels of the Per1 transcript were measured in each pool using RPA. A peak of Per1 transcript abundance was observed in approximately the middle of the spermatogenic wave (stages VII and VIII), which represents the stages harboring elongated spermatids and spermatozoa just before their release in the lumen (Fig. 2, A and B). This is in clear contrast to the ␤-actin levels, which show a gradual increase over the course of the developmental cycle. We then analyzed Per1 expression in the mouse, using three pools representing three different groups of mouse tubule stages. The level of Per1 transcript is higher in stages VII-VIII than in earlier or later stages (Fig. 2C). Because the resolution of tubule dissection in mice is lower than in the rat, we examined emulsions of individual tubules after in situ hybridization with the Per1 probe to complement the dissection. We observe the highest level of hybridization in stages VII–X, with the amount of the signal dropping off dramatically both at earlier and later stages (Fig. 2D). Furthermore, it is clear that the signal is highest in spermatids, the cell type nearest to the lumen, and lowest in the spermatogonia and pachytene spermatocytes, which are closer to the tubule periphery. Taken together, we conclude from these results that Per1 expression is a developmentally regulated process in the testis. Per1 Expression Is Not Controlled by CLOCK in Male Germ Cells The stage specificity of Per1 expression suggested the possibility that transcription of this clock gene was not driven by CLOCK/BMAL1 dimers, as in most other tissues, but by another positive regulator, perhaps one unique to spermiogenesis. To address this issue, we analyzed the distribution of the Clock transcript in testis by in situ hybridization. Low magnification of testis cross-sections shows that the distribution of


Clock expression is quite different from that of Per1 (Fig. 3A). First, the Clock transcript is found in all tubules at roughly similar levels. Second, the location of the Clock transcript is much closer to the tubule periphery than it is for Per1. On emulsion and at higher magnification, the Clock transcript can be seen in spermatogonia and in primary spermatocytes before, but not after, meiosis (compare Fig. 3, panels C and D). Thus Clock expression occurs at an earlier developmental time in the spermatogenic process than does Per1 expression, and thus in different cell types. This is summarized schematically (Fig. 3E), with Per1 expression in red and Clock expression in green. This suggested that Per1 expression was not mediated by a CLOCK-containing heterodimer. However, the possibility remained that CLOCK protein translated from the Clock transcript in pachytene spermatocytes persisted until the spermatid stage when Per1 is induced. To rule out this possibility, we examined the level of the Per1 transcript in Clock c/c mutant mice. We reasoned that if CLOCK was required for expression of Per1, then levels of this RNA should be reduced in the mutant mice. Figure 3F shows that the levels of Per1 transcript in the mutant are high and comparable to the levels found in wild-type mice. We conclude from these results that CLOCK is unlikely to be involved in the high level, stage-specific Per1 transcription. Per1 Expression Is Reduced in CREM-Deficient Mouse Testis If developmentally regulated Per1 expression in mouse seminiferous tubules is not regulated by the CLOCK-BMAL1 dimer, what induces its transcription? The induction of the Per1 transcript expression coincides with translation of the transcript encoding the transcriptional activator CREM␶, which shows peak protein levels in spermatids of stages VII and VIII (34). As the Per1 promoter contains a cAMP response element (CRE) (35), the site recognized by CREM, we thus investigated the possibility that Per1 expression is induced by CREM. RPA shows that Per1 transcript levels are reduced in testes of mice homozygous for a disrupted allele of the CREM gene, whereas the transcript levels for the synaptonemal complex protein SCP3, expressed in primary spermatocytes (36), remain high (Fig. 4A). Per1 levels normalized to either rRNA or ␤-actin are significantly different in CREM ⫺/⫺ and wild-type littermates (Fig. 4B). This result is confirmed by semiquantitative in situ hybridization experiments comparing the Per1 signal in CREM ⫺/⫺ mice testes with that in testes from wild-type animals (Fig. 4C). We find that, in agreement with our RPA results, tubules expressing Per1 to high levels are clearly absent from the CREM ⫺/⫺ testis. Interestingly, we also observe that the basal level of Per1 expression appears slightly higher in the CREM ⫺/⫺



145

Fig. 2. Per1 Transcript Levels Are Dependent on Developmental Stage A, RPA with Per1 and ␤-actin riboprobes using total RNA from dissected rat tubules of different stages. B, Densitometric scans of the RPA in (A) showing Per1 (circles) and ␤-actin (squares) levels. C, RPA with Per1 and ␤-actin riboprobes using total RNA from dissected mouse tubules of different stages. D, In situ hybridization with a Per1 riboprobe on mouse testis cryosections. Toluidine blue (upper panels) and the corresponding dark-field images of the emulsions (bottom panels) for individual tubules at the stages shown.

mice testes, which suggests the lack of an inhibitory factor. As a caveat to these observations, it must be noted that spermatids beyond step 4 are absent in CREM knock-out mice, and thus the low levels of Per1 transcript in these mutant mice might be due to the fact that this gene is mainly expressed after this developmental step. To provide additional support for the idea that the high levels of Per1 expression might be due to the action of CREM, we asked whether nuclear extracts from testis contained a protein that could bind the CRE in the Per1 promoter. Gel mobility shift assays with an oligonucleotide probe encompassing the Per1

CRE (Fig. 4D) show that wild-type nuclear extracts retard the mobility of the oligonucleotide probe, as does a recombinant CREM␶ control. As expected, nuclear extracts of CREM ⫺/⫺ testis lack this binding protein. There is no binding to an oligonucleotide mutated in the CRE by either recombinant CREM␶ or the wild-type nuclear extract, confirming that binding to this oligonucleotide occurs via the CRE. Furthermore, binding to the CRE by both recombinant CREM␶ protein or testis nuclear extracts was blocked by preincubation with an antibody directed against CREM␶. Thus, the CREM protein naturally expressed in germ cells binds the Per1 CRE.



Fig. 3. Clock and Per1 Transcripts Are Found in Different Cell Types A–D, Micrographs of toluidine blue-stained mouse testis cryosections (left panels) and the corresponding dark-field images of the emulsions (right panels) are shown for in situ hybridization with either a Clock antisense probe (A, C, and D) or a Clock sense probe (B). Higher magnification views of single tubules are from the stages indicated (C and D). E, A schematic view of spermatogenesis is used to summarize Per1 (red) and Clock (green) gene expression derived from the in situ hybridization results. Lighter colors reflect lower levels of expression. F, RPA with Per1 and ␤-actin riboprobes on RNA extracted from wild-type and Clock c/c mouse testes.

DISCUSSION Our interest in clock gene expression in the testis derived from the possibility that the circadian clock might play a role in timing some of the developmental events that take place during spermatogenesis. Mammals have a master clock in the SCN in the hypothalamus, which dictates the rhythmic phenotype of the animals (8, 27). This master oscillator induces rhythmicity in clock gene transcript levels in other tissues in the intact animal (21, 30, 32, 37). Furthermore, SCN neurons in culture not only have a robust oscillation in their own clock gene components but can maintain robust oscillations of clock gene transcripts in cocultured cells that would rapidly disappear if the cells were cultured alone (38). These results leave the im-

pression that circadian clock gene oscillations will be observed in all mammalian cells. However, our results indicate that clock components do not oscillate in the testis. This conclusion is based on a number of lines of evidence. First, the levels of Per1 RNA, as measured either by RPA in tissue extracts or by in situ hybridization on tissue sections, are constant over the daily cycle. Next, levels of the Bmal1 transcript, expected to cycle antiphase to Per1, are also constant over the daily cycle, as are the levels of the PER1 protein itself. With respect to the mechanism underlying the high and constant Per1 expression in spermatids, we have determined that the CLOCK/BMAL1 heterodimer (39), the usual transactivator of Per1 expression, is probably not involved in its regulation. This conclusion is based on the observations that Clock and Per1 tran-



147

transcription factor, perhaps one restricted to the first stages of spermiogenesis, might be involved in Per1 expression instead of CLOCK. One such transcription factor specific to spermiogenesis is CREM, whose protein levels peak in spermatids of stages VII and VIII within the spermatogenic cycle (34), paralleling the cell types with the highest Per1 expression (Fig. 2). Moreover, we show that the CRE site in the Per1 promoter can be bound by the CREM␶ present in nuclear extracts from wild-type testis (Fig. 4D). In agreement with these observations, we find that the testis of mice devoid of CREM show reduced Per1 transcript levels and lack a CRE binding protein by gel shift assays (Fig. 4). These data are thus consistent with a model in which the transcriptional activator CREM␶ drives Per1 expression in spermatids even in the absence of the CLOCK/BMAL1 heterodimer (Fig. 5C). It is interesting, in this regard, that the developing male germ cells in

Fig. 4. Per1 Transcript Levels Are Low in CREM Knock-Out Mouse Testis A, RPA with Per1, SCP3, and ␤-actin riboprobes on RNA extracted from wild-type and CREM⫺/⫺ mouse testes. RNA was equilibrated by ethidium staining (bottom panel). B, The mean and SD of the relative levels are plotted for Per1 normalized for ␤-actin levels and RNA levels (n ⫽ 6). Student’s t test: *, P ⫽ 0.0035; **, P ⫽ 0,0009, for the difference between wild-type and CREM⫺/⫺ values. C, Cross-sectioned testes from wild-type and CREM⫺/⫺ mice were used for in situ hybridization with a Per1 antisense probe. Representative autoradiograms are shown in the panels at left. Toluidine blue-stained mouse testis cryosections (center panels) and the corresponding dark-field images of the emulsions (right panels) are also shown. D, Gel mobility shift assays using either Per1 CRE or mutated Per1 CRE oligonucleotides together with nuclear extracts from wild-type or CREM⫺/⫺ mice testis or recombinant CREM␶ (left panel). For specific inhibition of CREM␶ binding, anti-CREM␶ was added 30 min before the labeled oligonucleotide.

scripts peak in different cell types within the tubules, and that Per1 expression is not decreased in Clock c/c mice. This latter observation suggests that another

Fig. 5. Schematic of Per1 Transcription in Somatic and Germ Cells A, In somatic cells, Per1 transcription is clock controlled through CLOCK/BMAL1 binding to the E box elements, with binding at the CRE reserved for modulating transcription as a result of signal transduction pathways. B, In germ cells before meiosis, either the presence of an inhibitory CREM isoform or the absence of a functional CLOCK/BMAL1 heterodimer may impair rhythmic Per1 transcription. C, In spermatids, the developmental switch to an activatory CREM isoform could allow Per1 transcription even in the absence of CLOCK or BMAL1.


the fly do not express per (40), given that CRE-binding proteins apparently play no role in Drosophila spermiogenesis (41). Taken together, our data are consistent with a model in which CREM␶ regulates Per1 transcription. However, as CREM knock-out mice lack the cell types normally expressing Per1 to high levels, the involvement of other transcription factors cannot be ruled out. Additional studies will validate whether CREM is directly activating Per1 transcription in male germ cells. How are the low and constant levels of Per1 expression explained in spermatocytes, where we have shown Clock gene expression (Fig. 5B)? One possibility is that Clock may simply be untranslated (42, 43). A second possibility is that the levels of Bmal1 expression in the testis may be too low to allow any significant activity of CLOCK on the Per1 promoter. Low Bmal1 transcript levels might be an expected consequence (13) of the low levels of Per2 expression in the testis (Ref. 32 and our unpublished results). Third, another factor important for Per1 expression may be missing in the testis. Lastly, these cells may express an factor inhibiting CLOCK/BMAL1 activity, such as the CREM␣ or the CREM␤ isoform (44). Our results thus differ from those of a previous report in which a circadian rhythm in both Per1 and Per3 transcripts was described in testis (32). However, unlike this previous report, we have used total RNA in RPAs instead of Northern blots, and ␤-actin instead of glyceraldehyde-3-phosphate dehydrogenase to standardize RNA in different samples. Our results dovetail nicely with other observations made in our laboratory using zebrafish. In this vertebrate, the Clock, Bmal1, and Bmal2 transcripts are rhythmic in all tissues except the testis in which Clock transcript levels are constitutively low (45) and Bmal transcripts are constitutively high (46). This suggests that zebrafish testis also lacks a functioning circadian clock, raising the possibility that a clockless testis may be a common theme in vertebrates. Strikingly, reproductive organs of insects also present analogous cases. In Drosophila, per RNA cycles in a various tissues, but not in ovaries (47). In the codling moth, although Per is rhythmically expressed in some parts of the reproductive system, no Per mRNA was detected in cells of the walls of the testes or in sperm bundles (48). What special feature of the testis requires the absence of the circadian clock so pervasive in other tissues? The testis has a number of characteristics that make it quite different from other tissues. The testis contains spermatogenic cells that perform a constant and complex cell differentiation program in which reductive cell divisions occur. It may be that the complex pattern of gene expression engendered by the circadian clock leads to unfavorable interactions with the developmental process of paramount importance to the testis. Alternatively, the normal oscillation of clock gene expression may be distorted by other transcriptional regulators or coactivators present uniquely in seminiferous tubules (49). Clock gene ex-


pression may thus be accidental and without any specific consequences. Indeed, it has been suggested that the high levels of the RNA polymerase II complex in testis (50) might result in a transcription from poor promoters or even DNA sequences that do not normally act as promoters (51). It must be borne in mind that none of the clock mutants examined to date have been observed to exhibit a reduced fertility in the laboratory. If expression of any clock gene in testis fulfills a role apart from the maintenance of rhythmicity, this role is clearly either not vital or functionally redundant. It is of course important to note that the circadian timing system does play an important role in determining the reproductive capacity of seasonally breeding mammals. For example, the syrian hamster experiences testicular atrophy during long-night photoperiods (L:D 12:12) but is normal during short-night photoperiods (L:D 14:10) (52). The involvement of the circadian clock can be seen from the photoperiodic response of the ␶ mutant hamster, which has a short free-running period of 20 h. Interestingly, these mutant hamsters experience testicular atrophy with a 10-h short-night dark period (53). It is thought that the circadian clock measures melatonin levels, as pinealectomized mutant hamsters receiving melatonin infusions for 10 h every 20 h experienced gonadal atrophy (53). It is unlikely however, that the testis is reacting directly to these melatonin signals. It is known that the levels of LH and FSH, which have profound effects on testis development, are affected by photoperiod (54), suggesting that a circadian clock in the testis is not required to mediate the effects of the photoperiod. In any event, our results indicate that the testis is devoid of the basic molecular circadian clockwork observed both in the SCN and in numerous nonneuronal tissues. Although we cannot rule out the possibility that a circadian clock involving different molecular components may be operating in the testis or that a very small subgroup of cells within this tissue may yet be rhythmic, our results, taken at face value, indicate that the testis is unlikely to employ a circadian clock in timing developmental processes. Looking for the mechanism and the consequences of this unexpected situation may allow the discovery of novel unexpected roles of so-called clock genes as well as to provide crucial data on the timing of male germ cell differentiation.

MATERIALS AND METHODS Animals Mice were routinely kept under a LD cycle (12 h-light,12-h dark) and were fed ad libitum. Under LD cycles, the start of the light period is termed ZT 0 (for Zeitgeber time). Clock c/c mutant mice (55) were obtained from J. Takahashi, and the Per1⫺/⫺ and CREM⫺/⫺ mice developed in house have been described previously (31, 56). For those experiments requiring constant conditions, mice were transferred to con-


stant darkness at the start of the normal light period, a time corresponding to CT 0 (for circadian time). For tubule dissections, isolated tubules were cut into sections in between the darkly colored regions containing the mature spermatids as described (57, 58). RPAs and in Situ Hybridization Total RNA was extracted from all tissues by homogenizing the frozen tissue in RNAsolv (Omega BioTek, Doraville, GA) and the concentrations standardized by eye on ethidium bromide-stained gels. The miniaturized RPA was as described previously (31), using riboprobes prepared from either nucleotides 1–336 of the Per1 open reading frame (ORF) (GenBank sequence AF022922), nucleotides 1256–1881 of the Bmal1 ORF (AB014494), nucleotides 432–853 of the SCP3 ORF (Y08485), and nucleotides 193–331 of ␤-actin (X03672) using an in vitro transcription kit (Promega Corp., Madison, WI). The Per1 and ␤-actin probes were incubated in the same reaction. Autoradiographs were quantitated either by densitometry or by using a PhosphorImager (Amersham Biosciences, Piscataway, NJ). For in situ hybridization, sense and antisense 35S-labeled riboprobes were made from the same sequence as for RPA, whereas the Clock probe spans the entire cDNA (GenBank AF000998). Frozen sections of the testis were prepared and processed as described (59). Western Blotting Anti-PER1 antibodies were raised in rabbits against the peptide sequence GSSSQDSAMEEEEQGGGS conjugated to keyhole limpet hemocyanin (Pierce Chemical Co., Rockford, IL), and purified by affinity on Sulfolink Coupling Gel (Pierce Chemical Co.) coupled with the same antibody. Testes were homogenized directly in Laemmli sample buffer, and the protein extracts were electrophoresed and transferred to nitrocellulose as described. Nuclear Extracts and DNA-Binding Assays Nuclear extracts from the testis of both wild-type and CREM⫺/⫺ mice were prepared as described (60). For gel retardation assays, oligonucleotides for the Per1 CRE (5⬘tccgcttTGACGTCAcctccct-3⬘) and its mutated form (5⬘tccgcttTCACAAAAcctccct-3⬘) were labeled and used as described (61). Recombinant CREM␶ was produced as described (62). For blocking CREM␶ binding to the Per1 CRE, a rabbit polyclonal antibody raised against recombinant CREM␶ was incubated with the nuclear extracts for 30 min at room temperature before addition of the labeled oligonucleotides.

Acknowledgments We thank Lucia Monaco for discussions and expert advice; Estelle Heitz and Ste´ phanie Roux for technical assistance; J. Takahashi and M. Vitaterna for the gift of the Clock cDNA and the Clock mutant mouse line; H. Tei for the gift of the Per1 cDNA; and all the members of the Sassone-Corsi laboratory for assistance, reagents, and helpful discussions.

Received May 21, 2002. Accepted October 18, 2002. Address all correspondence and requests for reprints to: Paolo Sassone-Corsi, Institut de Ge´ ne´ tique et de Biologie Mole´ culaire et Cellulaire, 1, rue Laurent Fries, 67404 IllkirchStrasbourg, France. E-mail: [email protected]. This work was supported by grants from Centre National de la Recherche Scientifique, Institut National de la Sante´ et


149

de la Recherche Me´ dicale (INSERM), Centre Hospitalier Universitaire Regional, Human Frontier Science Program, Organon Akzo/Nobel, Fondation pour la Recherche Me´ dicale, and Association pour la Recherche sur le Cancer. N.C. was supported by a Human Frontier Science Program long-term fellowship and a Canadian Institutes of Health Research postdoctoral fellowship; S.B. was supported by a Fondation de la Recherche Me´ dicale postdoctoral fellowship. D.M. was the recipient of an INSERM Poste Orange fellowship. * D.M. and N.C. contributed equally to this work. † D.M. is on leave from the De´ partement des Sciences Biologiques, Universite´ de Montre´ al, Montre´ al, Que´ bec H3C 3J7, Canada. ‡ Present address: Douglas Hospital Research Center, McGill University, 6875 LaSalle Boulevard, Montre´ al, Que´ bec H4H 1R3, Canada.

REFERENCES 1. Cobb J, Handel MA 1998 Dynamics of meiotic prophase I during spermatogenesis: from pairing to division. Semin Cell Dev Biol 9:445–450 2. Sassone-Corsi P 1997 Transcriptional checkpoints determining the fate of male germ cells. Cell 88:163–166 3. Parvinen M 1982 Regulation of the seminiferous epithelium. Endocr Rev 3:404–417 4. Eddy EM, O’Brien DA 1998 Gene expression during mammalian meiosis. Curr Top Dev Biol 37:141–200 5. Hecht NB 1998 Molecular mechanisms of male germ cell differentiation. Bioessays 20:555–561 6. Clermont Y 1972 Kinetics of spermatogenesis in mammals: seminiferous epithelium cycle and spermatogonial renewal. Physiol Rev 52:198–236 7. Allada R, Emery P, Takahashi JS, Rosbash M 2001 Stopping time: the genetics of fly and mouse circadian clocks. Annu Rev Neurosci 24:1091–1119 8. Cermakian N, Sassone-Corsi P 2000 Multilevel regulation of the circadian clock. Nat Rev Mol Cell Biol 1:59–67 9. Dunlap JC 1999 Molecular bases for circadian clocks. Cell 96:271–290 10. Reppert SM, Weaver DR 2001 Molecular analysis of mammalian circadian rhythms. Annu Rev Physiol 63: 647–676 11. Young MW, Kay SA 2001 Time zones: a comparative genetics of circadian clocks. Nat Rev Genet 2:702–715 12. Lee C, Etchegaray JP, Cagampang FR, Loudon AS, Reppert SM 2001 Posttranslational mechanisms regulate the mammalian circadian clock. Cell 107:855–867 13. Shearman LP, Sriram S, Weaver DR, Maywood ES, Chaves I, Zheng B, Kume K, Lee CC, van der Horst GT, Hastings MH, Reppert SM 2000 Interacting molecular loops in the mammalian circadian clock. Science 288: 1013–1019 14. Field MD, Maywood ES, O’Brien JA, Weaver DR, Reppert SM, Hastings MH 2000 Analysis of clock proteins in mouse SCN demonstrates phylogenetic divergence of the circadian clockwork and resetting mechanisms. Neuron 25:437–447 15. Yagita K, Tamanini F, van Der Horst GT, Okamura H 2001 Molecular mechanisms of the biological clock in cultured fibroblasts. Science 292:278–281 16. Griffin Jr EA, Staknis D, Weitz CJ 1999 Light-independent role of CRY1 and CRY2 in the mammalian circadian clock. Science 286:768–771 17. Kume K, Zylka MJ, Sriram S, Shearman LP, Weaver DR, Jin X, Maywood ES, Hastings MH, Reppert SM 1999 mCRY1 and mCRY2 are essential components of the negative limb of the circadian clock feedback loop. Cell 98:193–205 18. Gekakis N, Staknis D, Nguyen HB, Davis FC, Wilsbacher LD, King DP, Takahashi JS, Weitz CJ 1998 Role of the


19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

CLOCK protein in the mammalian circadian mechanism. Science 280:1564–1569 Hogenesch JB, Gu YZ, Jain S, Bradfield CA 1998 The basic-helix-loop-helix-PAS orphan MOP3 forms transcriptionally active complexes with circadian and hypoxia factors. Proc Natl Acad Sci USA 95:5474–5479 King DP, Zhao Y, Sangoram AM, Wilsbacher LD, Tanaka M, Antoch MP, Steeves TD, Vitaterna MH, Kornhauser JM, Lowrey PL, Turek FW, Takahashi JS 1997 Positional cloning of the mouse circadian clock gene. Cell 89: 641–653 Oishi K, Sakamoto K, Okada T, Nagase T, Ishida N 1998 Antiphase circadian expression between BMAL1 and period homologue mRNA in the suprachiasmatic nucleus and peripheral tissues of rats. Biochem Biophys Res Commun 253:199–203 Sun ZS, Albrecht U, Zhuchenko O, Bailey J, Eichele G, Lee CC 1997 RIGUI, a putative mammalian ortholog of the Drosophila period gene. Cell 90:1003–1011 Keesler GA, Camacho F, Guo Y, Virshup D, Mondadori C, Yao Z 2000 Phosphorylation and destabilization of human period I clock protein by human casein kinase I epsilon. Neuroreport 11:951–955 Yagita K, Tamanini F, Yasuda M, Hoeijmakers JH, van Der Horst GT, Okamura H 2002 Nucleocytoplasmic shuttling and mCRY-dependent inhibition of ubiquitylation of the mPER2 clock protein. EMBO J 21:1301–1314 Lowrey PL, Shimomura K, Antoch MP, Yamazaki S, Zemenides PD, Ralph MR, Menaker M, Takahashi JS 2000 Positional syntenic cloning and functional characterization of the mammalian circadian mutation ␶. Science 288:483–492 Yagita K, Yamaguchi S, Tamanini F, van Der Horst GT, Hoeijmakers JH, Yasui A, Loros JJ, Dunlap JC, Okamura H 2000 Dimerization and nuclear entry of mPER proteins in mammalian cells. Genes Dev 14:1353–1363 Pando MP, Morse D, Cermakian N, Sassone-Corsi P 2002 Phenotypic rescue of a peripheral clock genetic defect via SCN hierarchical dominance. Cell 110: 107–117 Balsalobre A, Damiola F, Schibler U 1998 A serum shock induces circadian gene expression in mammalian tissue culture cells. Cell 93:929–937 McNamara P, Seo S, Rudic RD, Sehgal A, Chakravarti D, FitzGerald GA 2001 Regulation of clock and mop4 by nuclear hormone receptors in the vasculature. A humoral mechanism to reset a peripheral clock. Cell 105:877–889 Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, Ueda M, Block GD, Sakaki Y, Menaker M, Tei H 2000 Resetting central and peripheral circadian oscillators in transgenic rats. Science 288:682–685 Cermakian N, Monaco L, Pando MP, Dierich A, SassoneCorsi P 2001 Altered behavioral rhythms and clock gene expression in mice with a targeted mutation in the Period1 gene. EMBO J 20:3967–3974 Zylka MJ, Shearman LP, Weaver DR, Reppert SM 1998 Three period homologs in mammals: differential light responses in the suprachiasmatic circadian clock and oscillating transcripts outside of brain. Neuron 20: 1103–1110 Abe M, Herzog ED, Yamazaki S, Straume M, Tei H, Sakaki Y, Menaker M, Block GD 2002 Circadian rhythms in isolated brain regions. J Neurosci 22:350–356 Delmas V, van der Hoorn F, Mellstrom B, Jegou B, Sassone-Corsi P 1993 Induction of CREM activator proteins in spermatids: down-stream targets and implications for haploid germ cell differentiation. Mol Endocrinol 7:1502–1514 Travnickova-Bendova Z, Cermakian N, Reppert SM, Sassone-Corsi P 2002 Bimodal regulation of mPeriod promoters by CREB-dependent signaling and CLOCK: BMAL1 activity. Proc Natl Acad Sci USA 99:7728–7733


36. Yuan L, Liu JG, Zhao J, Brundell E, Daneholt B, Hoog C 2000 The murine SCP3 gene is required for synaptonemal complex assembly, chromosome synapsis, and male fertility. Mol Cell 5:73–83 37. Sakamoto K, Nagase T, Fukui H, Horikawa K, Okada T, Tanaka H, Sato K, Miyake Y, Ohara O, Kako K, Ishida N 1998 Multitissue circadian expression of rat period homolog (rPer2) mRNA is governed by the mammalian circadian clock, the suprachiasmatic nucleus in the brain. J Biol Chem 273:27039–27042 38. Allen G, Rappe J, Earnest DJ, Cassone VM 2001 Oscillating on borrowed time: diffusible signals from immortalized suprachiasmatic nucleus cells regulate circadian rhythmicity in cultured fibroblasts. J Neurosci 21: 7937–7943 39. Darlington TK, Wager-Smith K, Ceriani MF, Staknis D, Gekakis N, Steeves TDL, Weitz CJ, Takahashi JS, Kay SA 1998 Closing the circadian loop: CLOCK-induced transcription of its own inhibitors per and tim. Science 280:1599–1603 40. Beaver LM, Gvakharia BO, Vollintine TS, Hege DM, Stanewsky R, Giebultowicz JM 2002 Loss of circadian clock function decreases reproductive fitness in males of Drosophila melanogaster. Proc Natl Acad Sci USA 99: 2134–2139 41. Fuller MT 1998 Genetic control of cell proliferation and differentiation in Drosophila spermatogenesis. Semin Cell Dev Biol 9:433–444 42. Kleene KC 1996 Patterns of translational regulation in the mammalian testis. Mol Reprod Dev 43:268–281 43. Kleene KC 2001 A possible meiotic function of the peculiar patterns of gene expression in mammalian spermatogenic cells. Mech Dev 106:3–23 44. Foulkes NS, Mellstrom B, Benusiglio E, Sassone-Corsi P 1992 Developmental switch of CREM function during spermatogenesis: from antagonist to activator. Nature 355:80–84 45. Whitmore D, Foulkes NS, Strahle U, Sassone-Corsi P 1998 Zebrafish Clock rhythmic expression reveals independent peripheral circadian oscillators. Nat Neurosci 1:701–707 46. Cermakian N, Whitmore D, Foulkes NS, Sassone-Corsi P 2000 Asynchronous oscillations of two zebrafish CLOCK partners reveal differential clock control and function. Proc Natl Acad Sci USA 97:4339–4344 47. Hardin PE 1994 Analysis of period mRNA cycling in Drosophila head and body tissues indicates that body oscillators behave differently from head oscillators. Mol Cell Biol 14:7211–7218 48. Gvakharia BO, Kilgore JA, Bebas P, Giebultowicz JM 2000 Temporal and spatial expression of the period gene in the reproductive system of the codling moth. J Biol Rhythms 15:4–12 49. Sassone-Corsi P 2002 Unique chromatin remodeling and transcriptional regulation in spermatogenesis. Science 296:2176–2178 50. Schmidt EE, Schibler U 1995 High accumulation of components of the RNA polymerase II transcription machinery in rodent spermatids. Development 121: 2373–2383 51. Schmidt EE 1996 Transcriptional promiscuity in testes. Curr Biol 6:768–769 52. Eskes GA, Zucker I 1978 Photoperiodic regulation of the hamster testis: dependence on circardian rhythms. Proc Natl Acad Sci USA 75:1034–1038 53. Stirland JA, Hastings MH, Loudon AS, Maywood ES 1996 The ␶ mutation in the Syrian hamster alters the photoperiodic responsiveness of the gonadal axis to melatonin signal frequency. Endocrinology 137:2183–2186 54. Lincoln GA, Peet MJ, Cunningham RA 1977 Seasonal and circadian changes in the episodic release of folliclestimulating hormone, luteinizing hormone and testosterone in rams exposed to artificial photoperiods. J Endocrinol 72:337–349


55. Vitaterna MH, King DP, Chang AM, Kornhauser JM, Lowrey PL, McDonald JD, Dove WF, Pinto LH, Turek FW, Takahashi JS 1994 Mutagenesis and mapping of a mouse gene, Clock, essential for circadian behavior. Science 264:719–725 56. Nantel F, Monaco L, Foulkes NS, Masquilier D, LeMeur M, Henriksen K, Dierich A, Parvinen M, Sassone-Corsi P 1996 Spermiogenesis deficiency and germ-cell apoptosis in CREM-mutant mice. Nature 380:159–162 57. Parvinen M, Vanha-Perttula T 1972 Identification and enzyme quantitation of the stages of the seminiferous epithelial wave in the rat. Anat Rec 174:435–449 58. Martianov I, Brancorsini S, Gansmuller A, Parvinen M, Davidson I, Sassone-Corsi P 2002 Distinct functions of TBP and TLF/TRF2 during spermatogenesis: requirement of TLF for heterochromatic chromocenter formation in haploid round spermatids. Development 129: 945–955


151

59. Fimia GM, Morlon A, Macho B, De Cesare D, SassoneCorsi P 2001 Transcriptional cascades during spermatogenesis: pivotal role of CREM and ACT. Mol Cell Endocrinol 179:17–23 60. Dignam JD, Lebovitz RM, Roeder RG 1983 Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res 11:1475–1489 61. Foulkes NS, Borrelli E, Sassone-Corsi P 1991 CREM gene: use of alternative DNA-binding domains generates multiple antagonists of cAMP-induced transcription. Cell 64:739–749 62. Delmas V, Laoide BM, Masquilier D, de Groot RP, Foulkes NS, Sassone-Corsi P 1992 Alternative usage of initiation codons in mRNA encoding the cAMP-responsive-element modulator generates regulators with opposite functions. Proc Natl Acad Sci USA 89:4226–4230