In testis, several RNA binding proteins have been shown to play a role in the translational regulation of specific transcripts. The human protein TRBP (TAR RNA binding protein) is the homologue of the mouse Prbp (Prm-1 RNA binding protein) involved in the protamine mRNA translational delay. TRBP is known to activate the HIV-1 long terminal repeat but this protein has never been investigated during spermatogenesis. The aim of this work was to analyse the TRBP expression in human testis. By Northern blot analysis, we demonstrated a major 1.5 kb transcript present at a high level in human testis and, to a lesser extent, in some other tissues. In-situ hybridization revealed that this transcript was present only in elongating spermatids. Antibodies raised against a 27 amino acid TRBP-specific peptide revealed a single protein of 43 kDa expressed in the cytoplasm of elongated spermatids. At the ultrastructural level, quantitative analysis of both TRBP mRNA and protein, using electron microscopy in-situ hybridization and immunocytochemistry, showed that TRBP is expressed mainly in spermatids at steps 3–4 of spermiogenesis. These results are in agreement with the probable role of TRBP in the control of human protamine mRNA translation.
Previous studies have established that, in man, protamines appear in spermatid nuclei at the end of the elongation phase (step 4–5 of spermiogenesis) (Roux et al., 1988; Lescoat et al., 1993), whereas transcripts are already present in round spermatids (Siffroi et al., 1999). These results indicate that, as in rodents, human protamine mRNA are regulated at a post-transcriptional level and are stored for several days in an inactive state before being translated. Such a phenomenon has been described for the majority of spermatid-specific transcripts (Penttilä et al., 1995; Kleene, 1996; Hecht, 1998; Cataldo et al., 1999; Steger, 1999) and is thought to involve specific RNA binding proteins (reviewed by Venables and Eperon, 1999). However, a large number of spermatid mRNA are also stored as free ribonucleoprotein particles by a mechanism acting independently of their nucleotide sequence (Schmidt et al., 1999).
Using transgenic mice carrying a fusion of the mouse protamine 1 (Prm-1) and human growth hormone (hGH) genes, it was shown (Braun et al., 1989) that 156 nucleotides of the Prm-1 3′ untranslated region (3′UTR) are sufficient to confer a translational regulation of the hybrid mRNA. Additional studies revealed that the most-3′ 62 nucleotides but not the most terminal 23 nucleotides, which contain the polyadenylation signal, are able to mediate a translational control (Braun, 1990). These results suggest that regulatory factors bind to the 3′UTR of protamine mRNA and prevent translation from occurring at the 5′ end of the transcripts. This concept is supported by the fact that transgenic male mice lacking the normal Prm-1 3′UTR show a premature accumulation of Prm-1 in their spermatid nucleus, leading to an arrest in the differentiation of these cells (Lee et al., 1995).
A number of proteins that interact with the 3′UTR of testis-specific mRNA have been described either by RNA–protein assays (Kwon and Hecht, 1991; Fajardo et al., 1994) or by the screening of male germ cell cDNA expression libraries (Schumacher et al., 1995a,b; Lee et al., 1996). The developmental expression of one of them, the testis brain RNA binding protein (TB-RBP), has already been investigated in mouse testis (Gu et al., 1998).
Among these RNA binding proteins, the mouse Prm-1 RNA binding protein (Prbp) is expressed in the cytoplasm of late-stage meiotic cells and round spermatids (Lee et al., 1996). It contains two copies of a double-stranded RNA binding domain which interact with the Prm-1 RNA 3′UTR portion implicated in the translational regulation. Prbp exhibits strong homologies with the human TAR (Trans-Activation-Responsive) RNA binding protein (TRBP), a cellular factor which activates the HIV-1 long terminal repeat (Gatignol et al., 1991). The gene encoding TRBP (TARBP2) has been located on human chromosome 12 (Kozak et al., 1995). However, the expression of the protein during spermatogenesis has never been investigated. Assuming that this protein might be involved in the regulation of protamine mRNA translation, the aim of the present work was to study the expression of the gene TARBP2 in human testis.
Materials and methods
Preparation of labelled TRBP probes
A cDNA clone (clone IMAGE 327516) containing a 1020 bp TRBP insert, corresponding to nucleotides 423 to 1443 of the TRBP sequence and cloned into a modified pT7T3D vector (Pharmacia Biotech, Sweden), was obtained from the HGMP resource center (Hinxton, UK) and checked by DNA sequencing.
For Northern blot analysis, the 1020 bp insert was isolated by enzymatic digestion using NotI and EcoRI restriction enzymes and purified by agarose gel electrophoresis. Following electro-elution, the insert was labelled with [α32P]dCTP (3000 Ci/mmol, Amersham, UK), at a specific activity of 1.2×109 c.p.m./μg, by using a Random Priming DNA Labelling Kit (Gibco BRL Life Technologies, Gaithersburg, USA).
For light microscopy in-situ hybridization, sense and antisense 33P-labelled single strand RNA probes were transcribed from TRBP linearized plasmids using T7 (antisense) or T3 (sense) RNA polymerases and [33P]UTP (>1000 Ci/mmol, Amersham, Uppsala, Sweden) respectively according to the supplier's protocol (Promega, Madison, USA). After transcription, DNA templates were removed by RNase-free DNase digestion [RQ1 (Rnase free Qualified) Dnase1, Promega, USA]. The probes were then purified on Sephadex G50 spin column.
For electron microscopy in-situ hybridization, probes labelled with digoxigenin-11 dUTP were obtained by polymerase chain reaction (PCR) reactions using plasmid cDNA as template, as described previously (Siffroi et al., 1999). For this purpose, two oligonucleotide primers were designed from the TRBP cDNA sequence: A: 5′-GTCTGTGTCAGAGGTAGAGATGG-3′ (nucleotides 163 to 185); B: 5′-CTGCAGTACCTCAAGATCATGGC-3′ (nucleotides 295 to 317).
PCR amplification of a 154 bp fragment was performed on a GeneAmp 2400 thermocycler (Perkin Elmer, Foster City, CA, USA) using 200 ng of DNA template and a mixture of dNTP containing 1 mmol/l dATP, 1 mmol/l dGTP, 1 mmol/l dCTP, 0.9 mmol/l dTTP and 0.1 mmol/l digoxigenin-11 dUTP. Estimation of the yield of labelled probes, and efficiency of labelled nucleotide incorporation, were checked by electrophoresis on 1% agarose gel containing ethidium bromide, in comparison with non-labelled PCR products.
Preparation of polyclonal antibodies to TRBP
A 27 amino acid TRBP (100–125-Y) peptide was synthesized by a conventional solid phase method using a 432A peptide synthesizer model (Applied Biosystems, Foster City, CA, USA). After completion of synthesis, the identity and purity of the peptide were checked first by amino acid analysis on a Beckman analyser and then by high performance liquid chromatography (HPLC) and microsequence of the HPLC peak using an automated 477A protein sequencer (Applied Biosystems, USA).
The peptide was then conjugated to keyhole limpet haemocyanin (KLH) using benzidine as a coupling agent. The peptide-carrier conjugate was purified by gel filtration on a PD 10 column (Pharmacia Biotech) and analysed by amino acid analysis to calculate the coupling molecular ratio (hapten/carrier), corresponding to 1470.
Two rabbits were immunized by i.d. injection of 150 μg equivalent peptide in complete Freund's adjuvant on day 0. Injections with 25 μg equivalent peptide in incomplete Freund's adjuvant were performed on days 21, 42 and 49. Bleedings were performed 7 days after each injection. Sera were tested at various dilutions in enzyme-linked immunosorbent assays for their capacity to react with TRBP peptide.
Human testicular biopsies were obtained from men, aged 60–65 years, undergoing orchidectomy for prostate cancer. None of them had received hormonal therapy or chemotherapy before surgical treatment. Testis fragments were either immediately processed for microscopic studies or frozen in liquid nitrogen.
Northern blot analysis
Two human multiple tissue Northern blots (Clontech, Palo Alto, USA), containing ~2 μg of polyA+ RNA from various human tissues, were pre-hybridized for 6 h at 65°C under agitation in ExpressHyb solution (Clontech). The blots were then hybridized separately overnight under agitation at 65°C in the same buffer with 3×107 c.p.m. of the 32P-labelled probe added for each blot. The membranes were then washed in 0.5×SSC (1×SSC: 150 mmol/l NaCl; 15 mmol/l sodium citrate, pH 7); 0.1% sodium dodecyl sulphate (SDS) for 1 h at 65°C. The blots were exposed for 15 days on a fluoroscreen and the hybridization signal was detected on a phosphofluoro-imager Storm (Molecular Dynamics, Sunnyvale, USA).
Western blot analysis
Germ cell suspension was obtained by mechanical dissociation of testis biopsies and was solubilized in 1×Laemmli reducing sample buffer. Approximately 25 μg of the fractions were run on a denaturing SDS-10% polyacrylamide gel and blotted onto nitrocellulose (Towbin et al., 1979). After saturation with 5% non-fat milk powder in phosphate-buffered saline (PBS) containing 0.1% Tween-20 for 2 h at room temperature, membranes were incubated with the antibody to TRBP or pre-immune serum diluted 1/100 followed by an incubation with a peroxydase conjugated goat anti-rabbit Ig diluted 1/20 000 (Jackson ImmunoResearch Laboratories, Baltimore, USA). Enzymatic activity was detected with ECL substrate (Amersham).
As a control of specificity, antibody to TRBP was pre-incubated for 18 h at 4°C with the TRBP peptide diluted 1/10–4 mol/l. The solution was then applied to membranes which were processed as described above.
Light microscopy in-situ hybridization and immunohistochemistry
For light microscopy in-situ hybridization, cryostat sections (5-8 μm) obtained from testis biopsies were collected on silicane-coated slides and fixed in 4% paraformaldehyde in PBS for 20 min at 4°C. After two washes in PBS, slides were air-dried and stored at –80°C. Before hybridization, slides were thawed slowly for 1 h at room temperature, treated with proteinase K (0.1 mg/ml) for 10 min at 37°C and then post-fixed in 4% paraformaldehyde for 5 min. After washing twice in 4×SSC, slides were acetylated in 4×SSC-triethanolamide-acetic anhydride for 10 min at room temperature, rinsed three times in 2×SSC, dehydrated in ethanol gradients and air-dried. Sections were then pre-hybridized in 50% formamide; 4×SSC; 1% Sarcosyl; 10 mmol/l dithiothreitol; 1×Denhardt's [0.2% bovine serum albumin (BSA); 0.02% Ficoll; 0.02% polyvinylpyrrolidone] in PBS for at least 4 h at 63°C, rinsed twice in 2×SSC, dehydrated in ethanol and air-dried. Hybridization was realized by covering sections with 50 μl drops of a solution made of the pre-hybridization mix containing 10% Dextran sulphate and 106 c.p.m. of the labelled antisense single strand RNA probe. Incubation was performed in a moist chamber at 63°C overnight. After hybridization, slides were rinsed twice in 2×SSC for 10 min at room temperature and incubated 30 min at 37°C in 0.1 mol/l Tris pH 8, 1 mmol/l EDTA, 0.5 mol/l NaCl containing 20 μg/ml RNase A and 20 U/ml RNase T1. Slides were then washed twice in 2×SSC for 30 min at 63°C, twice in 1×SSC for 30 min at 63°C, once in 0.5×SSC for 30 min at 63°C and finally once in 0.1×SSC for 15 min at room temperature. Sections were dehydrated in ethanol and air-dried before being coated with Ilford RNP40 nuclear emulsion and exposed for 8 days. Before examination, slides were counterstained with haematoxylin-eosin. Controls were performed by the hybridization of a labelled sense single strand RNA probe under the same conditions.
For light microscopy immunohistochemistry, cryostat sections were fixed in 4% paraformaldehyde in PBS for 20 min at room temperature, rinsed in PBS and then stored at –80°C. After thawing, slides were incubated overnight at 4°C with the antibody to TRBP diluted 1/100 in PBS pH 7.2. Slides were then rinsed three times in PBS for 10 min. Immunoreaction was revealed by using a commercially available kit (LSAB2; Dako, USA) according to the manufacturer's instructions. As a control of specificity, antibody was diluted 1/100 in PBS pH 7.2, 0.1% Tween, 5% milk and pre-absorbed by the immunizing peptide, diluted 1/10–4 mol/l, for 18 h at 4°C. Slides were then treated as described above.
Electron microscopy in-situ hybridization and immunocytochemistry
For electron microscopy in-situ hybridization, testis fragments were fixed by immersion in 4% paraformaldehyde; 0.1% glutaraldehyde in 0.1 mol/l Sorensen buffer (0.2 mol/l Na2HPO4, 0.2 mol/l NaH2 PO4, pH 7.4) for 1 h at 4°C. Samples were washed several times in the same buffer, dehydrated in increasing gradients of N,N-dimethylformamide and embedded in Lowicryl K4M medium under UV irradiation for 2 days at 4°C and then for 3 days at room temperature (Polysciences, Paris, France). Ultrathin sections were mounted on gold grids and incubated for 4 h at 37°C by floating on 2 μl drops of the hybridization solution (50% formamide, 0.6 mol/l NaCl, 10 mmol/l Tris–HCl pH 7.5, 1 mmol/l EDTA, 10% Dextran sulphate) containing 10 ng/μl of the heat-denatured digoxigenin-labelled probe. After hybridization, sections were rinsed three times in deionized water for 5 min and three times more in PBS pH 7.2. They were then incubated for 5 min at room temperature in Tris–HCl pH 8 and in Tris–HCl pH 8; 0.1% BSA for 20 min. Detection of hybrids was made by incubating the grids for 1 h at room temperature with a sheep antibody to digoxigenin (Boehringer Mannheim, Germany) coupled to 10 nm gold particles, diluted 1/30 in Tris–HCl pH 8, 0.1% BSA. After several washes in Tris–HCl pH 8; 0.1% BSA and in deionized water, sections were air-dried before staining with uranyl acetate. They were then examined using a Jeol 120CX electron microscope at 60 kV. Controls for specificity were performed by pre-incubating the sections with RNase A (1 mg/ml), by hybridizing them in buffer without labelled probes or by not including the antibody to digoxigenin during immunocytochemical development.
For electron microscopy immunocytochemistry, testicular fragments were fixed in 2% paraformaldehyde; 0.1% glutaraldehyde in PBS for 1 h at 4°C and embedded in Lowicryl K4M medium as described for the in-situ hybridization experiments. Ultrathin sections were mounted on gold grids, incubated with goat serum diluted 1/30 in PBS pH 7.4 for 2 h at 4°C and then overnight at 4°C in a moist chamber with 10 μl drops of PBS pH 7.4 containing the antibody to TRBP diluted 1/100 and goat serum diluted 1/30. Sections were then rinsed successively five times for 1 min at room temperature in PBS pH 7.4, once in 0.02 mol/l Tris pH 8.2 for 5 min and once in 0.02 mol/l Tris pH 8.2, 1% BSA for 20 min. Immunoreaction was revealed by incubating sections with goat anti-rabbit Ig (British BioCell International, Cardiff, UK) diluted 1/30 in Tris–HCl, pH 8, 0.1% BSA and coupled to 10 nm gold particles. After washing in deionized water, sections were air dried before staining with uranyl acetate. They were then examined using a Jeol 120CX electron microscope at 60 kV. Controls for specificity were performed by omitting the antibody to IRBP in the incubation of sections.
Quantitative study of gold particle distribution and statistical analysis
For both in-situ hybridization and immunocytochemistry, spermatids, as classified by Holstein and Roosen-Runge (1981), were divided into three groups: Rs: round spermatids (steps 1 and 2); Els1: spermatids at the beginning of the elongation phase (steps 3 and 4); and Els2: spermatids at the end of the elongation phase (steps 5 and 6). Corresponding steps during mouse spermiogenesis are steps 1–8 for Rs and steps 9–13 for Els1 and Els2 from the 16 characteristic steps in this species.
For each group, 30 cells were counted by a single operator. From negative prints of electron micrographs, taken at a standard magnification of ×6700, nuclear and cytoplasmic surface areas of cells were calculated using an image analysis instrument (Starwise Morphostar Imstar, Paris, France). Into these arbitrary frames, the labelling densities of nuclei and cytoplasms consisted of the number of gold particles per 100 μm2. A similar evaluation of gold particle density in spermatogonia was used as a control.
One-way analysis of variance (ANOVA) was used to determine the statistical significance of differences in labelling between the cellular types observed. The Scheffé multiple range test was used to analyse specific comparisons (PC Statview Program).
Hybridization of Northern blots containing polyA+ RNA from various human tissues revealed a major transcript at 1.5 kb and present at high levels in testis (Figure 1). However, a lower detectable signal of the same size was also observed in other tissues such as heart, placenta, liver and pancreas. Tissues such as brain, lung, thymus, small intestine, colon and blood leukocytes exhibited a level of TRBP mRNA expression that was at the lower limit of detection. Two other barely detectable transcripts were visible at 3.2 and 5.4 kb in most samples.
Incubation of Western blots containing human testis proteins with the antibody raised against the TRBP peptide revealed a major band at 43 kDa and weaker background bands of higher molecular weight (Figure 2, lane 2). This major band was not recognized by the pre-immune serum (Figure 2, lane 3). Signal corresponding to the major band was pre-absorbed specifically by the immunizing peptide (Figure 2, lane 4). In contrast, the weaker background bands were not, or marginally, reduced by pre-absorption.
In testis cryostat sections, in-situ hybridization of a 33P-labelled antisense single strand RNA probe showed a silver grain distribution restricted to elongating spermatids (Figure 3A) although some round spermatids were also labelled. Silver grains were not observed over spermatogonia, spermatocytes, Sertoli or interstitial cells. In sections hybridized with a labelled sense single strand RNA probe, the background appeared to be very low (Figure 3B). Immunolocalization of the protein TRBP was in agreement with in-situ hybridization data. A specific and strong labelling was observed in the cytoplasm of elongating spermatids (Figure 3C). This labelling was not present in sections incubated after pre-absorbtion of the antibody to TRBP by the immunizing peptide (Figure 3D).
At the ultrastructural level, detection of TRBP transcripts and protein, as revealed by in-situ hybridization and immunocytochemistry labellings, showed that the distribution of gold particles in spermatids appeared at random, without any association with particular cellular organelles (Figure 4A,B). Controls for specificity were negative for both in-situ hybridization and immunocytochemistry experiments (data not shown).
Quantitative evaluation of the gold particle distribution for TRBP mRNA in the different groups of spermatids is shown in Figure 5A. ANOVA revealed that this distribution, per nucleus area and per cytoplasmic area, was not random between the different defined groups of spermatid (P < 0.0001).
The labelling density in round spermatid (Rs) nuclei was statistically different from that observed in spermatogonia, taken as control (P < 0.0001). The labelling density became significantly higher in the nuclei of spermatids at the beginning of the elongation phase (Els1) (P < 0.0001) before sharply decreasing in spermatids at the end of the elongation phase (Els2) (P < 0.0001), while persisting at a different level from controls (P < 0.01).
Cytoplasmic labelling appeared in the Rs group at a level different from controls (P < 0.0001) and increased dramatically in spermatid cytoplasm in the Els1 group (P < 0.0001). It then decreased significantly in Els2 spermatid cytoplasm (P < 0.0001) while persisting at a different level from the control group (P < 0.01).
Quantitative evaluation of gold particle distribution for the protein TRBP is shown in figure 5B. Nuclear labelling densities were not significantly different from that observed in controls except for the Els1 group (P = 0.01). In cytoplasm, labelling density in the Rs group was significantly different from controls (P < 0.0001) and increased dramatically in the Els1 group (P < 0.0001) before decreasing slightly in the Els2 group (P = 0.03). In this latter group, cytoplasmic labelling density persisted at a high level in comparison to the control group (P < 0.0001).
Taking into account a strong homology between mouse Prbp and human TRBP proteins (93% identical at the amino acid level), we analysed the tissue distribution of TRBP mRNA and protein in human testis, as well as in other tissues, and studied their localization in male germ cells. To our knowledge, this work represents the first contribution to the study of the expression pattern of a RNA binding protein potentially involved in protamine mRNA translational regulation in human testis.
Northern blot analysis showed that TRBP transcripts are predominantly found in testis and are the same size as the mouse Prbp mRNA (1.6 kb). They are also present, at low but detectable levels, in other tissues such as heart, liver, pancreas, kidney and placenta, whereas the other organs barely express the TRBP2 gene. Additional minor transcripts, found at 3.2 and 5.4 kb, represent probably either a non-specific hybridization signal or non-spliced mRNA. They could also represent differences between somatic and germinal mRNA due to the use of alternative transcription start sites or polyadenylation signals. This pattern of expression is similar to the distribution of mouse Prbp mRNA (Lee et al., 1996). As for Prbp, TRBP trancripts are translated into a 40 kDa protein. However, Prbp protein exists in multiple forms in testis, as revealed by a cluster of four major bands around 43 kDa after Western blotting (Lee et al., 1996), whereas the TRBP protein is present as a single band which is also detected around 43 kDa. Therefore, molecular characterization of TRBP mRNA and protein is very similar to that observed for Prbp in mouse and suggests that both proteins share identical functions in the two species.
Indeed, the location of TRBP mRNA and protein mainly observed in the elongating spermatids, as shown by light microscopy in-situ hybridization and immunohistochemistry, suggests a role for this protein in the terminal differentiation of male germ cells, as has been shown for Prbp in mouse. Such a function can be considered on the basis of our results obtained by electron microscopy in-situ hybridization and immunocytochemistry. It has been previously demonstrated that mRNA coding for human Prm-1 are transcribed first in round spermatids, where transcripts are found in both the nucleus and cytoplasm (Siffroi et al., 1999). Then the level of cytoplasmic mRNA decreases progressively from steps 3 to 5 of spermiogenesis, the time at which several immunocytochemical studies have established that protamines deposit in spermatid nucleus (Roux et al., 1988; Le Lannic et al., 1993; Lescoat et al., 1993). These observations are similar to those made in rodents (Morales et al., 1991). An RT-PCR analysis of human Prm-1 and Prm-2 mRNA showed that protamine transcripts in spermatids at step 3 of spermiogenesis were detected only after proteinase K digestion of samples prior to reverse transcription. These results suggested that regulatory proteins may bind to both transcripts in step 3 spermatids. Since the highest levels of TRBP mRNA and protein are observed in intermediate spermatids at steps 3–4 of spermiogenesis, it can be suggested that TRBP is synthesized concurrently to protamine mRNA translation and participates in the regulation of this process, although there is as yet no evidence that TRBP actually binds to protamine mRNA.
As with Prbp in mouse, TRBP could bind to the 3′UTR of protamine transcripts and to regulate their translation. It has been shown that mouse Prbp protein binds in vitro to a 24 nucleotide sequence of the Prm-1 3′UTR, inside the 62 nucleotides known to mediate translational delay, but also to poly (I) and poly (C) suggesting that it could act as a non-specific translational repressor (Lee et al., 1996). In vivo, the Prbp protein appears first in pachytene spermatocytes but principally in the cytoplasm of round spermatids, in which Prm-1 mRNA are repressed. The protein is no longer detected in elongated spermatids, the step at which these mRNA are translated into protamines (Lee et al., 1996). However, a study of male mice carrying a disruption of the Prbp gene has shown a failure in the normal replacement of transition proteins by protamines rather than a premature translation of protamine mRNA (Zhong et al., 1999). This suggests that Prbp is required for the translational activation of the repressed transcripts coding for protamines and could function as a chaperone responsible for the assembly of translationally regulated protamine ribonucleoprotein particles. This is in agreement with the fact that separate elements in the 3′UTR of mouse Prm-1 mRNA mediate either translational repression or activation (Fajardo et al., 1997).
Our results, concerning the expression pattern of both TRBP mRNA and protein by electron microscopy, suggest a similar role for TRBP in the activation of human protamine mRNA translation, as shown by the persistence of an immuno-gold labelling in the spermatids at the end of the elongation phase. The lack of in-situ hybridization and immunocytochemical labelling of late-pachytene spermatocytes differs from data obtained for Prbp in mouse (Lee et al., 1996). However, in mouse spermatocytes, the protein was detected only at a very low level. Therefore, the expression pattern of TRBP in human testis is likely to be spermatid specific, which is different from that observed for another RNA binding protein, TB-RBP, the mRNA of which is highly expressed in mouse pachytenes (Gu et al., 1998).
Another difference between TRBP and Prbp lies in the intracellular localization. While Prbp is localized to the cytoplasm, a slight but significant labelling is found for TRBP in spermatid nucleus at the beginning of the elongation phase. TRBP has also been shown to be primarily nuclear in HeLa cells (Kozak et al., 1995) and this could account for some differences between the actual roles of both proteins.
Another feature of both Prbp and TRBP proteins is their tendency to form oligomers on target RNA (Gatignol et al., 1993; Lee et al., 1996). Since these proteins are thought to protect mRNA from premature translation, one could suggest a simple translational repression model in which they bind to specific sequences in the 3′UTR of protamine mRNA and then oligomerize to form filamentous inactive ribonucleoprotein particles (Lee et al., 1996). This would explain why TRBP mRNA transcription and translation is maintained at a high level up to steps 3–4, as seen by electron microscopy in-situ hybridization and immunocytochemistry.
A possible role for TRBP protein during spermiogenesis can be inferred from results obtained in cultured HeLa cells. It has been shown (Park et al., 1994) that TRBP inhibits a ribosome-associated protein kinase (PKR), which is responsible for the phosphorylation of the eukaryotic translation initiation factor 2 (eIF-2) leading to its inactivation and to translational blocking. TRBP has also been shown to dimerize with itself and with PKR in a yeast two-hybrid system (Cosentino et al., 1995). Therefore, the inhibition of PKR by TRBP would activate the translation of specifically repressed mRNA. In testis, PKR, which is an interferon (IFN)-induced kinase, has been studied in isolated seminiferous cells in the presence or in the absence of IFN alpha (Dejucq et al., 1997). Results showed that PKR is expressed constitutively only in Sertoli and peritubular cells but not in pachytene spermatocytes nor in early spermatids. Despite a lack of PKR expression in haploid cells, in this model, further studies need to be performed to verify the relationships between TRBP and the eIF-2/PKR-dependent mechanism of translational regulation in testis.
In conclusion, this work suggests that, like Prbp in mouse, human TRBP protein is implicated in the translational activation of repressed protamine transcripts. How TRBP acts in vivo with other proteins responsible for translational inhibition and/or activation of protamine mRNA remains to be established.
The authors wish to thank Dr D.Escalier. They also thank Ms A.Gonzales, C.Le Bourhis, S.Reposo, A.Triclin for their helpful technical assistance and Mr Ph.Nguyen for artwork. This work has been supported by grants from AP-HP (CRC 96053 and PHRC AOM 96142).