A Novel Germ Line-specific Gene of the Phosducin-like Protein (PhLP) Family

We identified a new member of the phosducin-like (PhLP) protein family that is predominantly, if not exclusively, expressed in male and female germ cells. In situ analysis on testis sections and analysis of purified spermatogenic cell fractions evidenced a stage-specific expression with high levels of RNA and protein in pachytene spermatocytes and round spermatids. Three mRNA species were detected, which correspond to different polyadenylation sites and vary in abundance during germ cell maturation. Only low levels of RNA were detected in whole ovary extracts, but expression of the protein became detectable within hours after hormonal induction of superovulation. The gene (Mgcphlp) is located on mouse chromosome 5 in the immediate vicinity of the Clock locus. The predicted amino acid sequence shows extensive similarities not only with the known mammalian PhLP proteins but also with the yeast phosducin-like protein Plp2, required for the production and growth of haploid cells. Expression of the murine protein was found to complement the defect of a yeastplp2Δ mutant. We propose that MgcPhLP/Plp2 proteins exert a function in germ cell maturation that is conserved from yeast to mammals.


From the University of Nice, INSERM U470, 06108 Nice, France
We identified a new member of the phosducin-like (PhLP) protein family that is predominantly, if not exclusively, expressed in male and female germ cells. In situ analysis on testis sections and analysis of purified spermatogenic cell fractions evidenced a stage-specific expression with high levels of RNA and protein in pachytene spermatocytes and round spermatids. Three mRNA species were detected, which correspond to different polyadenylation sites and vary in abundance during germ cell maturation. Only low levels of RNA were detected in whole ovary extracts, but expression of the protein became detectable within hours after hormonal induction of superovulation. The gene (Mgcphlp) is located on mouse chromosome 5 in the immediate vicinity of the Clock locus. The predicted amino acid sequence shows extensive similarities not only with the known mammalian PhLP proteins but also with the yeast phosducin-like protein Plp2, required for the production and growth of haploid cells. Expression of the murine protein was found to complement the defect of a yeast plp2⌬ mutant. We propose that MgcPhLP/Plp2 proteins exert a function in germ cell maturation that is conserved from yeast to mammals.
Phosducin, a protein highly expressed in the retina and pineal gland, has been considered as playing a role in retinal phototransduction by interacting with the ␤␥ subunits of G proteins and thereby modulating their signaling functions (1). A partially similar, widely expressed phosducin-like protein (PhLP) 1 has been identified, which also inhibits G␤␥ function (2,3). Two related genes identified in the yeast Saccharomyces cerevisiae were designated PLP1 and PLP2 (4). The Plp1 protein was shown to bind efficiently the G␤␥ subunits. Binding of Plp2 was also evidenced but with a lesser affinity. On the other hand, genetic analysis evidenced the role of Plp2 in the generation of viable haploid cells.
The mammalian phosducin and phosducin-like proteins are expressed ubiquitously (5). Whether tissue-specific homologues exist remains an open question. In the course of screening a mouse testis cDNA library, we have identified a novel germ cell-specific phosducin-like protein, designated MgcPhLP (for "mouse germ cell-specific phosducin-like protein"), which exhibits significant similarities to both the mouse phosducin and phosducin-like proteins. Expression is strictly restricted to the male and female germ cells in a regulated manner depending on the stage of germ cell maturation. The murine gene complemented the defect of a yeast plp2⌬ mutant, suggesting an evolutionarily conserved function in meiotic and/or post-meiotic cells.

EXPERIMENTAL PROCEDURES
Mice-Mice used in all experiments were C57BL/6 ϫ DBA/2 F1. Eggs were collected from either naturally ovulated females or after hormonal induction of superovulation according to standard procedures (6).
Cell Cultures and Preparation of Germ Cells-Cultures of the 15P-1-established Sertoli cell line and primary Sertoli cell cultures, preparation of total germ cells, and fractionation by elutriation centrifugation were performed as described previously (7).
RNA Analysis-Total RNAs were extracted and analyzed by Northern blot hybridization as described previously (7). Dot blot hybridization was performed on Mouse RNA Masterblot TM (Clontech).
DNA Sequencing-Purified PCR products were first subcloned in pGEM-T Easy vector System I (Promega, catalog no. A1360). The purified plasmids were then sequenced with a DNA sequencing system kit (Big Dye Terminator, PerkinElmer Life Sciences, catalog no. 4303149) according to manufacturer's instructions. Sequences were analyzed with the ABI PRISM 310 genetic analyzer (PerkinElmer Life Sciences).
In Situ Hybridization-In situ hybridization was performed as described previously (8) using single-stranded digoxigenin-labeled RNA probes corresponding to sense or antisense Mgcphlp messengers.
Purification of the MgcPhLP Protein-Mgcphlp cDNA was inserted into the pGEX-5X vector (Roche Molecular Biochemicals) by taking advantage of a BamHI site in the amino-terminal region (amino acid 5), generating a GST (glutathione S-transferase) fusion gene in which expression was induced in Escherichia coli cultures during overnight growth at 25°C in the presence of 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside. Bacteria from 500 ml of culture were washed and extracted in 50 ml of 100 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 50 g/ml of both leupeptin and aprotinin, by sonication (six times for 10 s) at 4°C. Extract was clarified by high-speed centrifugation (30 min, 30,000 rpm, 4°C in a Beckman Ti-50 rotor), and submitted batchwise to adsorption on glutathione-Sepharose beads (0.50 ml packed volume, prewashed with extraction buffer) with gentle agitation for 2 h at 4°C. Beads were washed five times at 4°C with 1 ml of 20 mM Hepes, pH 7.5, 1.0 M NaCl, 1 mM EDTA, 2 mM dithiothreitol, 0.1% Nonidet P-40 and four times with proteolysis buffer (20 mM Hepes, pH 7.5, 100 mM NaCl, 2 mM CaCl 2 , 2 mM dithiothreitol, 10% glycerol, and 20 g/ml bovine serum albumin). Biotinylated factor Xa (Roche Molecular Biochemicals) was added (10 g in 1 ml final volume), and reaction was allowed to proceed for 2 h at 4°C under gentle agitation. Phenylmethylsulfonyl fluoride was then added (to a final concentration of 1 mM), and, after the beads were spun down, the supernatant was cleared of protease by incubation for 1 h at 4°C with streptavidin beads (50 l packed volume, Roche Molecular Biochemicals). The latter were removed by centrifugation, and the final supernatant was dialyzed against phosphate-buffered saline (400 ml, two changes, 4 h, 4°C). The final preparation was analyzed by SDS-PAGE (see Fig. 6) and stored frozen at Ϫ70°C. In parallel, control buffer was prepared in the same way using bacteria that express GST.
Immunological Techniques-GST-MgcPhLP fusion protein, prepared bound to glutathione-Sepharose beads as described above, was eluted by phosphate-buffered saline containing 20 mM reduced glutathione and used to raise antiserum in rabbits. The resulting antiserum recognized both GST and MgcPhLP in Western blot analysis (data not shown). Specific antibodies, prepared by depleting the rabbit serum with Sepharose beads carrying the GST protein, were used as primary antibodies in immunofluorescence and Western immunotransfer analyses.
Western transfers were obtained from mouse organs minced and boiled in Laemmli's buffer by a standard procedure (9); and mouse oocytes were collected and treated in the same way. Secondary antibody was a mouse monoclonal anti-rabbit IgG Fc fragment antibody coupled to horseradish peroxidase (Sigma). Staining was performed by the ECL technique (Amersham Pharmacia Biotech).
Yeast Strain, Media, and Plasmid-The mutant S. cerevisiae strain (plp2::URA3,plp2⌬) used in this study was obtained from Dr. H. G. Dohlman (Yale University School of Medicine, New Haven, CT). For expression in yeast, the full-length Mgcphlp cDNA was inserted in the expression vector pRS314 (10).

RESULTS
A New Phosducin-like mRNA-During the course of an unrelated expression-based screen of a mouse testis library, we identified a cDNA uniquely expressed in germ cells. Fig. 1 shows 818 base pairs of cDNA sequence with an open reading frame of 240 amino acids, encoding a putative protein of 27.824 kDa, designated MgcPhLP, which exhibits 30 -36% amino acid identity to the mouse and bovine phosducins and to the rat phosducin-like protein in a 143-amino acid core region (Fig. 2). Important amino acid residues conserved between the phosducin and phosducin-like proteins of various species (human (11), rat (12), mouse (13)) are also conserved in MgcPhLP protein. As documented in more detail under "Evolutionary Conservation," significant similarities were also detected with two phosducinrelated proteins of the yeast S. cerevisiae.
A computer search indicated that in the mouse genome, the Mgcphlp gene is located within a 204-kb segment of chromo-some 5 (GenBank TM accession number AF146793), which also includes the Clock locus and has been designated Pdcl2 (Gen-Bank™ accession number AAD30564). The existence of two potential proteins, named protein B and PDCL2 (corresponding, respectively, to amino acids 3-204 and 3-195 of MgcPhLP) had been predicted from the genomic DNA sequence. The complete nucleotide sequence derived from the initial cDNA clone and the 5Ј-and 3Ј-RACE products now allows us to determine the correct position of the initiating methionine and thus the complete amino acid sequence.
An analysis of mouse EST libraries showing identities with Mgcphlp sequences (Unigene cluster Mm.143764 plus additional clones) strongly suggested that the gene was at least preferentially expressed in the testis. All identities were found in testis-derived library sources and none among clones derived from other tissues, with the best match found in a purified spermatocyte library (BG100925).
Expression of Mgcphlp in the Testis-Northern blot analysis on a limited series of tissues detected expression only in the testis (Fig. 3A). A larger screen performed by dot blot analysis on a membrane spotted with poly(A)ϩ RNA from 22 mouse organs (Fig. 3B) again showed expression only in the testis. Mgcphlp RNA was not detected in any of the somatic organs tested. The same conclusion was reached using the most sensitive RT-PCR assay, which could not detect Mgcphlp RNA sequences in two somatic tissues (Fig. 3C). The RNA was present in purified germ cells (see below and expression of the gene cannot be excluded either in yet another tissue, or in a minor fraction of cells, possibly under specific physiological conditions, it would appear to occur mostly in testicular germ cells. In fact, as shown in a subsequent section, Mgcphlp is also expressed in female germ cells during a limited period of meiotic maturation. In whole testis extracts, three mRNA species were detected. Analysis of cDNA ends by 5Ј-and 3Ј-RACE revealed that these RNAs share the same 5Ј extremity but differ by their 3Ј ends. Three major polyadenylation sites were identified, corresponding to three canonical polyadenylation signals in the genomic sequence (Fig. 4). As shown in Fig. 3D, distribution of the three isoforms significantly differed between the pachytene spermatocyte, the round spermatid and the elongated spermatid fractions prepared by elutriation centrifugation. During post-natal development, RNA was first detected at 20 days post-partum, and the maximum level of expression was reached at day 30. The midsize mRNA was the only one identified at day 20, whereas the other two isoforms appeared only at the later time points.
Stage-specific expression in the testis was confirmed by in situ hybridization. The probe consisted of the entire coding sequence of Mgcphlp and thus detected all three mRNA iso-forms. As shown in Fig. 5, expression was detected in pachytene to round spermatids at all spermatogenesis stages.
Expression of MgcPhLP was examined at the protein level by using a rabbit antiserum raised against a GST fusion protein (see "Experimental Procedures"). Immunofluorescence staining of purified fractions of male germ cells was clearly positive from the meiotic to late haploid stages of spermatogenesis, and the protein was also present in the mature spermatozoa of epididymal sperm. It was not detectable by Western blotting at 10 days post-partum but was clearly present in the testis at day 18 (Fig. 5, B-F).
Expression in Ovary-Despite the facts that no EST sequences corresponding to Mgcphlp were present in ovarian libraries from various species and that RNA was not detected by dot blot hybridization (Fig. 3B), low levels of expression were evidenced by RT-PCR analysis in total extracts from adult ovary (Fig. 6A). Western blot analysis (Fig. 6B) failed to detect expression of the protein in extracts from either ovary or unfertilized eggs. It was, however, detected in fertilized eggs (Fig.  6C). Taking into account the smaller proportion of germ cells in the female gonad, these results did not exclude the possibility of stage-specific expression during meiotic maturation. This was shown to be the case by Western blot analysis following injection of human chorionic gonadotropin as part of a superovulation regime (6) (Fig. 6D). The protein was detected as early as 3 h after hormone injection, a time corresponding to the nuclear breakdown step of preovulatory meiotic maturation.
Association with 14-3-3 Protein(s)-The 14-3-3 family includes a series of closely related proteins that bind phosphorylated components of signal transduction pathways and modulate their interactions (reviewed in Ref. 14). Several of them, prominently the 14-3-3 protein, are expressed in a stage-de-pendent manner in the spermatogenic differentiation pathway (15). As shown in Fig. 7, a complex of the MgcPhLP protein with 14-3-3 protein(s) was evidenced by immunoprecipitation of testicular protein extracts with polyvalent anti-14-3-3 antibodies followed by Western blot analysis of the precipitated complexes with anti-MgcPhLP antiserum. Further experiments are in progress to identify more precisely the protein(s) present in these complexes.
Evolutionary Conservation-The Mgcphlp coding sequence appears to have been relatively well conserved throughout evolution. Coding regions of the murine gene are 87-92% identical at the nucleotide level to human EST sequences contained in Unigene cluster Hs.223712, derived exclusively from germ cells or testes. The cluster sequences are located in human chromosome 4q11 region, again in close proximity to the human homologue of the mouse Clock gene. In the S. cerevisiae genomic sequence, two genes, PLP1 and PLP2, encode proteins with a clear similarity to the mammalian phosducins and phosducin-like sequences (4). Alignments of phosducin-like sequences (Fig. 2) show that the murine MgcPhLP protein is more closely related to Plp2. One may note that, in the aminoterminal part of the sequences, an 11-amino acid region implicated in G␤␥ binding (16) is completely conserved between the phosducin proteins and Phlp1 but present neither in Plp2 nor in MgcPhLP. That would distinguish two groups of phosducinlike proteins: on one hand, the mammalian phosducins with the region that interacts with G␤␥ proteins, and on the other, the MgcPhLP and PLP2 proteins, with a distinct pattern of conserved amino acids. Other discrete patches of pairwise similarities could also be observed in the central and carboxylterminal parts of the sequences.

Complementation of a Yeast plp2 Mutant by the Murine Mgcphlp Gene-A yeast strain bearing mutations in the TRP1
and URA3 nutritional markers and in which one PLP2 allele has been replaced by URA3 (plp2::URA3,plp2⌬) was established previously (4). The deletion of PLP2 prevented the recovery of mutated haploid clones upon induction of sporulation in diploid heterozygotes. The mutated strain was transformed with an expression vector for Mgcphlp and TRP1 (see "Experimental Procedures"), and the resulting colonies were induced to sporulate. Isolated spores were tested in duplicate in media selective for either URA3 (plp2⌬) or TRP1 (Mgcphlp). Haploid colonies were then revealed by replica plating with yeast strains of either the a or ␣ mating types carrying a mutation in the nutritional HIS gene and selection in histidine-free medium. Haploid derivatives identified as carrying both the URA3 (plp2D) allele and TRP1 (Mgcphlp) were grown. Analysis by PCR amplification confirmed the expected absence of the wild type PLP2 and the presence of the Mgcphlp allele (Fig. 8A). These complemented haploid clones were successfully grown for successive generations. As expected from the published data, transfer of the empty vector did not result in the production of viable haploid cells.
A Meiotic Function Conserved from Yeast to Mammals?-It was initially reported (4) that haploid derivatives could not be grown from the mixture of meiotic products generated by the diploid heterozygous genotype (plp2::URA3,plp2⌬). One could not on this basis distinguish between a function of the gene during meiosis (or at an early stage of spore formation) and a general requirement for cellular growth. In view of the results shown in Fig. 8B, we may now conclude that the gene is not required for growth. Genotyping of haploid clones that had been maintained in culture for 30 -40 generations after their isolation from the complemented parent strain showed that the mouse gene was eventually lost. Growth properties of the clones in the absence of a functional PLP2 gene remained, however, unaffected. It is therefore most likely that the protein, not necessary for growth, was required for the establishment of the haploid state, a conclusion that is consistent with the expression of the mouse gene being restricted to the meiotic and early post-meiotic stages. DISCUSSION Phosducins are regulators of G protein activity in the retina, and the phosducin-like proteins are considered to be potential ubiquitous regulators of G␤␥ signaling. We describe in this report a novel phosducin-like mRNA specific of the meiotic and post-meiotic germ cells, which shows significant amino acid sequence similarities to the phosducin and phosducin-like proteins of various species. Searching existing EST clones from both mouse and human suggested germ cell specificity, because related sequences were found in EST libraries from testes but not from other tissues. In the mouse, expression of Mgcphlp was at least predominantly observed in male and female germ cells, in both sexes at the meiotic and post-meiotic stages. It must be taken into account, however, that a detailed in situ analysis could not be performed on every possible tissue, and therefore we cannot exclude the possibility that the gene might be expressed in a minor cell population and/or only during a limited physiological or developmental period (as is in fact the case in the ovary). In the male gonad, three Mgcphlp RNA isoforms were identified at different stages of differentiation. These three specific mRNAs correspond to different polyadenylation sites in the locus. It is interesting to note the presence in the vicinity of the polyadenylation sites in genomic DNA of AU-rich sequences similar to the sequences described as cytoplasmic polyadenylation elements mediating polyadenylation and translation of messages during the oocyte release from the meiotic block at ovulation and prior to the activation of zygotic genome at the two cell stage (17,18). The presence of stage-specific isoforms of the Mgcphlp message may thus reflect a translational control during meiotic and postmeiotic maturation. The conclusions of RNA analysis were confirmed by direct determination of the protein by polyclonal antibodies specific for the mouse protein. This was especially informative in the ovary, in which expression is normally limited to the small number of oocytes undergoing meiosis but  1 and 3) derived from spores generated by a plp2::URA3,plp2⌬ yeast clone after transfer of the Mgcphlp expression vector. These clones were checked to be haploid on the basis of their ability to conjugate with reference a and ␣ yeast strains. PCR analysis was performed with primers Plp2-s and Plp2-r, generating a 525-bp product diagnostic of the wild type allele (WT) and a 1.3-kbp fragment from the deleted gene with the URA3 insert. B, loss of the mouse gene upon long term growth of the complemented haploid clones. PCR analysis was performed with primers Mgc271 and Mgc483 (lanes 1, Mgcphlp), Plp2-s and Plp2-r (lanes 2, PLP2), and Plp2-s and Ura3-r (lanes 3, URA3, plp2⌬) could readily be evidenced after hormonal stimulation leading to superovulation.
As is the case for the other phosducin-related proteins, the function of the protein at the molecular level remains largely to be established. A possibly significant feature in this respect is its association with at least one of the proteins of the 14-3-3 family. Binding may be mediated by the RSSVP motif (amino acids 119 -123, Fig. 1), which resembles the sites of interaction identified in other 14-3-3-binding proteins (14). 14-3-3 binds phosphorylated serine residues in a number of proteins active in signal transduction. In retinal photoreceptors, 14-3-3 is considered as regulating the binding of phosducin to G␤␥ by sequestering the phosphorylated phosducin molecules and blocking their binding. 14-3-3 was also recently shown to interact in the brain with a phosducin-like protein (19,20). The specificity of 14-3-3 binding and its relationship with the phosphorylation of serine residues in MgcPhLP are currently being studied.
Two phosducin-related genes were recently described in yeast, PLP1 and PLP2 (4). MgcPhLP displays a greater amino acid similarity with PLP2 than with PLP1. The inability of a plp2⌬ mutant to generate viable haploid products was successfully complemented by transfer of the mouse gene. Regarding the function of the yeast gene, published data (4) have left two possibilities open; the Plp2 protein could either be necessary for growth in general or specifically required for the generation of haploid products, either during or after meiosis. The observation that haploid clones bearing the plp2⌬ mutation could not be grown from a sporulating culture is compatible with both interpretations, thus making it impossible to evaluate the phenotype of the diploid homozygous mutant. Our data, in fact, rather favor the hypothesis of a meiotic function. We observed, upon long term growth of several of the complemented haploid strains, a loss of the murine gene that did not impair their growth ability. A meiotic function of the yeast gene is consistent with the restricted meiotic and post-meiotic expression of Mgcphlp in the mouse.
Mgcphlp is included in the 204-kb DNA fragment of the Clock locus (GenBank TM accession number AF146795). Taking into account that phosducin, which is expressed abundantly in the retina, has been considered to be involved in signal phototransduction cascades, one might speculate that Mgcphlp expression could be part of the same type of signaling cascade initiated by dark/light stimuli. In the mouse, the ovulation cycle is known to be dependent on light periodicity, and circadian periods have been reported for the ovarian melatonin and rhythm of cAMP accumulation. Increase in Mgcphlp expression within hours after induction of superovulation by human chorionic gonadotropin injection clearly points to hormonal regulation. It is clear, however, that beyond such speculations, the function of the mouse protein in germ cell differentiation will require the use of site-directed and/or temporally controlled mutagenesis technologies.