INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Calcium plays a central role in numerous biological processes, including cell proliferation, protein secretion, and muscle contraction. Many of these cellular effects are mediated by the ubiquitous intracellular calcium receptor calmodulin, which when bound to calcium can activate a variety of enzymes, including protein kinases, phosphatases, and phosphodiesterases (6). Because calmodulin is present in all tissues, cell-type-specific functions are determined by the complement of its downstream targets.

Calmodulin is especially abundant in the testis, which led to the identification of the testis-specific binding protein, calspermin. Calspermin was initially purified from rat and pig testes as a potent inhibitor of the calmodulin-dependent cyclic nucleotide phosphodiesterase (11). Purified calspermin binds calmodulin in the presence of Ca²⁺ (10, 11) and contains a calmodulin-binding domain close to the N terminus (12). Calspermin and Ca²⁺-calmodulin-dependent protein kinase IV (CaMKIV) are both products of the Camk4 gene (7, 9) and are derived by alternative transcriptional initiation (7). The calspermin promoter and its testis-specific first exon are located within the 10th CaMKIV intron, and the two proteins share the final two exons of the Camk4 gene (15).

In vitro transcription assays with the rat calspermin promoter demonstrated that a fragment spanning from 80 to +361 (with +1 as the transcriptional initiation site) yielded maximal activity (16). Within this region, two cyclic AMP response element (CRE) motifs bind the testis-specific transcriptional activator CREM (16), which regulates transcription of several male germ cell-specific genes (5). Mice deficient in CREM fail to express calspermin even if heterozygous, suggesting that calspermin is exquisitely sensitive to levels of CREM (8). In transfection assays, CREM must be phosphorylated for full activity, and in vitro both protein kinase A (PKA) and CaMKIV can phosphorylate CREM. Cotransfection of CREM with either PKA or CaMKIV can stimulate transcriptional activity from the calspermin promoter in NIH 3T3 cells (16). That the calspermin promoter functions in vivo has been demonstrated by transgenic mice bearing the -galactosidase gene driven by the calspermin promoter, which exhibit male germ cell-specific X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside) staining (15). Furthermore, -galactosidase activity can be activated in mouse embryonic fibroblasts from these mice by transfection of CREM with PKA or CaMKIV (15).

These results led to the prediction that CaMKIV might regulate expression of calspermin and other CRE-dependent testis-specific genes by phosphorylation and activation of CREM (14). However, this model has since been refuted by two lines of evidence. (i) CREM is not phosphorylated in vivo in germ cells. Rather, CREM appears to interact with the transcriptional regulator ACT to stimulate transcription in postmeiotic spermatids (4). (ii) Expression of CRE-containing genes, including calspermin, is unaltered in CaMKIV-deficient mice (18). Instead, the absence of CaMKIV results in the failure of spermatogenesis and reflects a requirement for CaMKIV in the replacement of transition protein by protamine during spermiogenesis (18).

CaMKIV and calspermin are both expressed in spermatids (17) and, although CaMKIV is not required for the proper expression and localization of calspermin, the relationship between these two proteins is not understood. Indeed, the function of calspermin in the testis remains unknown. One plausible role that it may serve is to regulate the abundance of calmodulin levels within the testis. Calmodulin is essential during cell replication, but it may be inhibitory during differentiation in the last stages of sperm production (3). CREM^/ mice lack calspermin and exhibit defective spermatogenesis, but they also fail to express several other CRE-containing genes, including those encoding the protamines and transition proteins, which are also required for sperm production (1, 8).

To clarify the role of calspermin in the testis, we have specifically deleted calspermin gene expression in mice. In doing so, it was critical to leave the kinase intact since CaMKIV has been shown to have important effects in several tissues, including the testis. We report here that calspermin is not required for fertility and that neither the expression of CaMKIV nor its targeting to the nuclear matrix are dependent on calspermin.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of calspermin-null (CaSKO) mice. A  clone spanning 15 kb of the 3' end of the Camk4 gene was obtained from a mouse 129/Sv genomic DNA library and subcloned into the pBluescript vector. This fragment was used to generate a targeting construct for calspermin. The 400 bp spanning the minimal calspermin promoter and the testis-specific exon were replaced with the thymidine kinase and neomycin selectable markers flanked with loxP sites. The targeting construct was electroporated into 129/Sv-derived embryonic stem (ES) cells. Genomic DNA from G418-resistant clones was digested with KpnI and screened by Southern blot analysis using a 5' probe. This probe detects a 12-kb band from the wild-type allele and a 7-kb band from the mutant allele. Correctly targeted clones were transfected with Cre recombinase, resulting in the efficient removal of the selectable markers. The resulting ES cells were used to generate chimeric mice, which were then bred to wild-type C57BL/6J mice to obtain germ line transmission. The resulting mice were genotyped by PCR, with primers A and C yielding a 400-bp wild-type band and primers B and C producing a 750-bp mutant band. All mice were housed at the Duke University Vivarium under a 12-h light, 12-h dark cycle. Food and water were provided ad libitum, and all care was given in compliance with National Institutes of Health (NIH) guidelines on the use of laboratory and experimental animals.

Northern blots. Total RNA was extracted from the mouse testes with Ultraspec RNA (Biotecx Laboratories). Equal amounts (10 µg) of RNA were loaded on a formaldehyde gel, subjected to electrophoresis, and transferred to a Zeta-Probe membrane (Bio-Rad) in 10× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate). The membrane was hybridized overnight in Church buffer (500 mM Na₂HPO₄, pH 7.2; 1 mM EDTA; 7% sodium dodecyl sulfate [SDS]) at 65°C and washed with 40 mM sodium phosphate-5% SDS-1 mM EDTA at 65°C. A probe for calspermin was labeled with Klenow and [³²P]dCTP (Amersham Pharmacia Biotech, Piscataway, N.J.).

Western blotting. Tissues were homogenized in a buffer with 25 mM HEPES, 1 mM EGTA, 1 mM EDTA, 0.5 mM dithiothreitol (DTT), 10% glycerol, and 10 µg of aprotinin, 1 µg of leupeptin, 20 µg of trypsin inhibitor, and 0.1 µg of Pefablock per ml. Then, 75 µg of protein was subjected to electrophoresis on a 12% gel, transferred to a polyvinylidene difluoride membrane, and blocked with 5% milk in Tris-buffered saline with 0.05% Tween 20. Membranes were incubated with anticalspermin at 1:5,000, detected with horseradish peroxidase (HRP)-conjugated antibody, and developed with the ECL System (Amersham Pharmacia Biotech).

Sperm counts. The epididymis was dissected out at the vas deferens, and sperm was expressed from the cauda epididymis into phosphate-buffered saline and counted with a hemacytometer.

Histology and immunohistochemistry. Testes were fixed in Bouin's fixative and paraffin embedded. Next, 7-µm sections were cut and stained with periodic acid-Schiff-hematoxylin (Poly Scientific, Bay Shore, N.Y.). For immunohistochemistry sections were incubated in 3% H₂O₂ to block endogenous peroxidase activity, subjected to antigen retrieval by microwaving in 10 mM sodium citrate (pH 6.0) for 10 min, and incubated with anticalspermin at 1:200 overnight at 4°C. Following three washes in phosphate-buffered saline, sections were incubated with biotinylated secondary antibody and streptavidin-HRP (Vector Laboratories, Burlingame, Calif.), detected with diaminobenzidine (Sigma, St. Louis, Mo.) or NovaRed substrate (Vector Laboratories), and counterstained with hematoxylin.

Nuclear matrix preparation. Isolation of nuclear matrix was performed as described earlier (13) with modifications. Testes were homogenized in cytoskeletal buffer (CSK): 10 mM PIPES, pH 6.8; 100 mM NaCl; 300 mM sucrose; 3 mM MgCl₂; 1 mM EGTA; 1 mM DTT; 0.5% Triton X-100; 1 µg of leupeptin, 1 µg of aprotinin, and 1 µg of pepstatin per ml; 1 mM phenylmethylsulfonyl fluoride; 1 mM Na₃VO₄; and 25 mM NaF. After incubation at 4°C for 3 min, samples were centrifuged at 5,000 × g for 3 min. Chromatin was released by digestion with 1,000 U of RNase-free DNase I (Boehringer Mannheim) per ml in CSK buffer for 15 min at 37°C. Ammonium sulfate in CSK buffer was added to a final concentration of 0.25 M. Samples were incubated at 4°C for 5 min and centrifuged. The pellet was extracted with 2 M NaCl in CSK buffer for 5 min at 4°C and then centrifuged. The resulting nuclear matrix fraction was solubilized in urea buffer (8 M urea; 0.1 M NaH₂PO₄; 0.01 M Tris, pH 8).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Our approach to the selective deletion of calspermin was to use Cre/lox technology to excise the calspermin promoter and testis-specific exon while leaving CaMKIV exons intact. A 400-bp fragment spanning the minimal calspermin promoter and testis-specific intron was replaced with a cassette containing neomycin and thymidine kinase selectable markers (Fig. 1A). Because we wanted CaMKIV transcription to proceed normally, we flanked the selectable markers with loxP sites so that they could later be removed. The targeting construct was transfected into ES cells. A gene encoding diphtheria toxin was used to negatively select nonhomologous recombination events. Homologous recombination in ES cells was confirmed by Southern blotting, with the wild-type allele yielding a 12-kb band and the mutant allele yielding a 7-kb band (Fig. 1B). Correctly targeted ES cell clones were transfected with Cre recombinase, resulting in the efficient removal of the selectable markers and leaving only a single 50-bp lox site. These clones were used to generate chimeric mice, which were bred for germ line transmission. CaSKO mice were successfully generated without disruption of CaMKIV. Northern blot analysis of testis RNA demonstrated that calspermin mRNA is absent in homozygous null mice and decreased in heterozygous littermates, while CaMKIV RNA levels are unaffected (Fig. 1C). Western blot analysis of thymus lysates confirmed that CaMKIV protein expression was similarly unchanged in CaSKO mice (Fig. 1D).

View larger version (25K): [in this window] [in a new window]   FIG. 1.   Targeted deletion of calspermin. (A) Map of the Camk4 129/Sv genomic clone and construction of the targeting construct. The minimal calspermin promoter, which contains two CRE motifs, and the testis-specific exon (Ts) were replaced by neomycin (Neo) and thymidine kinase (TK) selectable markers flanked by loxP sites. Correctly targeted ES cells were transfected with Cre recombinase to delete the selectable markers, leaving a single loxP site. The probe used for Southern blot analysis is indicated by a black bar. (B) Southern blot analysis of KpnI(K)-digested genomic DNA from targeted and wild-type ES cells hybridized with the 5' probe. The wild-type allele is 7 kb, while the mutant allele is 13 kb. (C) Northern blot analysis of total RNA probed with a portion of the calspermin cDNA, which hybridizes to both the 2.1-kb Camk4 mRNA and the 1.1-kb calspermin mRNA. Equal amounts (10 µg) of RNA were loaded per lane. (D) Western blot analysis of thymus extracts from wild-type (+/+), heterozygous (+/), and homozygous null (/) animals detected with anti-CaMKIV antibody.

Mice deficient in calspermin were bred to wild-type mice to assess their ability to produce offspring. To our surprise, CaSKO mice exhibited no impairment of fertility. There was no difference in the testis weights (Fig. 2A) or sperm counts (Fig. 2B) between CaSKO and wild-type mice. Furthermore, in breeding studies, CaSKO males produced litters comparable in frequency and size to their wild-type littermates (data not shown). In all other respects as well CaSKO mice were normal, and female mice were also fertile. Histological analysis of tubules from CaSKO mice confirmed that spermatogenesis is able to proceed normally in the absence of calspermin (Fig. 3A and B). Specifically, there was no apparent defect in spermiogenesis despite the expression of calspermin in round and elongating spermatids.

View larger version (11K): [in this window] [in a new window]   FIG. 2.   CaSKO mice are fertile. (A) Testis weights of wild-type (+/+) and CaSKO (/) mice. (B) Sperm counts from wild-type (+/+) and CaSKO (/) mice. n = 6 for each genotype. Values are ± the standard deviation.

View larger version (138K): [in this window] [in a new window]   FIG. 3.   Histology of CaSKO seminiferous tubules. (A and B) Histological analysis of wild-type (A) and CaSKO (B) testes. Magnification, ×170. (C and D) CaMKIV localization is not dependent on calspermin. Testis sections from wild-type (C) and CaSKO (D) mice were analyzed by immunohistochemistry as described in Materials and Methods. CaMKIV expression patterns are unchanged in CaSKO testes Magnification, ×85.

These mice allowed us to investigate further the relationship between CaMKIV and calspermin in spermatids. We have already demonstrated that calspermin expression is unperturbed in mice lacking CaMKIV (18). Likewise, we found that CaMKIV expression and localization are not dependent on calspermin. Western blot analysis of testes lysates confirmed that calspermin protein is not produced in CaSKO testes and revealed that CaMKIV levels in the testis are not changed in the absence of calspermin (Fig. 4A). Interestingly, we find that the levels of calspermin are markedly lower than that of CaMKIV, suggesting that in the mouse, unlike the rat, calspermin is not very abundant in the testis. To examine the cellular localization of CaMKIV in CaSKO testes, immunohistochemistry experiments were performed on testis sections from wild-type and CaSKO mice. In wild-type mice CaMKIV is expressed in spermatogonia and spermatids in a stage-dependent manner. This pattern is identical to that seen in CaSKO testes (Fig. 3 C and D).

View larger version (29K): [in this window] [in a new window]   FIG. 4.   CaMKIV expression and targeting to the nuclear matrix are unaltered in CaSKO mice. (A) Western blot analysis of testes extracts blotted with anticalspermin, which detects both CaMKIV and calspermin. (B) Soluble and nuclear matrix preparations from wild-type (+/+), heterozygous (+/), and CaSKO (/) testes were immunoblotted for CaMKIV. (C) Testes extracts from these mice were also immunoblotted with an antibody against calmodulin.

The C terminus of CaMKIV is highly acidic, a characteristic of chromatin-associated proteins (2). Since calspermin is identical to this portion of CaMKIV, one might expect calspermin to colocalize with CaMKIV. We have previously demonstrated that in the testis CaMKIV associates with both chromatin and the nuclear matrix (17). It is not known whether calspermin has similar associations and, if so, whether this has any effect on the interactions of CaMKIV. We performed chromatin fractionation on testes lysates from CaSKO mice and found that CaMKIV targeting to the nuclear matrix is unaffected by calspermin deficiency (Fig. 4B). In addition, calspermin does not associate with either the chromatin or the nuclear matrix in testes of wild-type or CaMKIV-deficient mice (data not shown). We also found that protamine 2 and transition protein 2 expression patterns are normal in mice not expressing calspermin, a finding consistent with the lack of interaction between CaMKIV and calspermin in spermatids (data not shown).

If the physiological role of calspermin is to bind and sequester calmodulin in the testis, one possibility is that calmodulin levels have been altered in compensation for the absence of calspermin. We immunoblotted testes lysates from wild-type, heterozygous, and homozygous null mice and found that calmodulin levels are similar among these genotypes (Fig. 4C).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have successfully used Cre/lox gene knockout technology to specifically prevent calspermin gene expression without disrupting the transcription of CaMKIV in mice. In doing so we have conclusively demonstrated that the 400-bp fragment including two CRE motifs and the testis-specific exon are necessary and sufficient to drive calspermin expression in mouse testis. With the recent suggestions that the human genome may contain fewer genes than expected, it seems likely that there will be many instances of gene loci which encode multiple products, as is the case with the Camk4 gene. The ability to selectively delete individual gene products may prove to be a powerful tool in understanding their functions.

Early studies had found calspermin to be a highly abundant calmodulin-binding protein in the rat testis. We have recently demonstrated that in the mouse CaMKIV and calspermin are both expressed in postmeiotic spermatids (17) and that CaMKIV plays a critical role in their differentiation (18). We had predicted, based on these data, that the loss of calspermin would also have a negative impact on spermiogenesis. However, we found that mice carrying a targeted deletion of calspermin do not display any impairment of spermatogenesis or fertility. Furthermore, the generation of both CaSKO and CaMKIV-deficient mice has allowed us to fully demonstrate that the regulation and functions of CaMKIV and calspermin are completely independent of each other, despite their shared gene.

There are several potential reasons why mice lacking calspermin remain fertile. One obvious explanation is that the levels of calspermin in mouse testis are very low, i.e., only a fraction of the abundance of CaMKIV. In the rat, calspermin was found to be highly abundant, leading us to believe that it must play a significant role in the regulation of calmodulin levels. Given the low levels of expression in the mouse, it seems reasonable that the absence of calspermin would not have a great impact on calmodulin expression and/or function. In support of this contention, we have also shown that calmodulin levels are not appreciably changed in the absence of calspermin. Alternatively, it is possible that the calspermin transcript is a spurious product in murine male germ cells, with resulting low levels of translation into protein. Another possibility is that there are compensatory changes in levels of other calmodulin-binding proteins in the testis.

The most striking result from the CaSKO mice is that despite their shared gene locus and coexpression in spermatids, calspermin has no role in the regulation and function of CaMKIV. We have previously demonstrated that the expression of CaMKIV and calspermin in mouse seminiferous tubules is coordinated and stage dependent (17). Both proteins are first expressed in stage IV tubules. However, we have now definitively shown that CaMKIV and calspermin do not have an impact on each other's expression. Therefore, although a common factor(s) may be involved in initiating transcription, it appears to act independently on each promoter. Sertoli cells, which contact all cell types within a given staged tubule, may secrete some factor which simultaneously activates CaMKIV and calspermin gene expression. Alternatively, perhaps some genomic rearrangement within stage IV tubules exposes both genes for transcription. Whether these or other mechanisms are at work in driving expression of these two related but independently produced proteins remains to be determined.

We also report here that calspermin does not localize to the nuclear matrix with CaMKIV. This was somewhat surprising since the C terminus of CaMKIV, and therefore calspermin, had been predicted to mediate interaction with the chromatin and/or nuclear matrix. However, calspermin is not found at the nuclear matrix and, as such, should not be in a position to participate in the interactions proposed between CaMKIV, sperm basic nuclear proteins, and the nuclear matrix. Indeed, we show that the expression of CaMKIV and its involvement in regulating basic nuclear protein exchange are unimpaired in CaSKO testes. These results further suggest that, instead, the N terminus of CaMKIV may be required for targeting of the kinase to the nuclear matrix, a possibility currently under investigation.