| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
INSERM U.567 CNRS-UMR 8104, Département d'Hématologie, Maternité de Port-Royal4
Département de Maladies Infectieuses, Institut Cochin,5 Université René Descartes, 75014 Paris, France
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
ankyrin, apoptosis, endoplasmic reticulum, gamete biology, spermatogenesis, testis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We have recently described the expression of a gene called Phtf1 (putative homeodomain transcription factor 1) in male germ cells. Three phtf genes have been characterized to date, one in insects (d-phtf in Drosophila) and two (Phtf1 and Phtf2) in mammals. Both d-phtf and Phtf1 are highly expressed in male germ cells [1] and Phtf2 is mainly expressed in skeletal muscular tissues [1, 2]. PHTF proteins do not display any sequence similarities with known or predicted proteins, but their conservation among species suggests an essential function. The 84 kDa PHTF1 protein was found to be an integral membrane protein associated with a domain of the endoplasmic reticulum (ER) juxtaposed to the Golgi apparatus. Immunolocalization of PHTF1 reveals it is present during meiosis and spermiogenesis, and makes it possible to follow the modifications that occur in the Golgi apparatus and ER system. By the end of spermiogenesis, PHTF1 is released from the mature spermatic cell within the residual bodies [3].
To get some insight into the molecular function of Phtf1 we searched for potential partners in testis, and isolated the testis-specific ortholog of the Caenorhabditis elegans feminization factor 1 (fem-1), Fem1b. The fem genes of C. elegans were initially identified as being necessary for male fate determination. Null mutations of the fem-1, fem-2, or fem-3 genes prevented all aspects of male development, and transform both XO and XX animals into females [4–6]. Most of the actors involved in C. elegans sex determination have been identified, but their biochemical roles are not entirely clear. It has been suggested that fem-2 is a member of the PP2C family of the serine/threonine phosphatases, and that its binding to fem-3 may be necessary for male development [7, 8]. Unlike fem-2 and fem-3, fem-1 is highly conserved among species, with one known member in Drosophila (Flybase number CG11896) and three known mammalian orthologs, Fem1a, Fem1b, and Fem1c [9–11]. The nematode fem-1 protein is widely expressed in both hermaphrodites and males. In the mouse, duplication events have created divergent expression patterns, with Fem1a more abundant in the heart and skeletal muscle, and Fem1b more abundant in the testis [9, 11]. The Fem1c gene was recently described, and is also abundant in testis [9, 11]. Homologies between the fem-1 family members are mainly in their N-terminal domain, which was recognized as containing six repeats matching the ankyrin (ANK) consensus [9, 12, 13]. The ANK repeat is a motif of 33 amino acids, and four or more ANK repeats form an ANK domain that mediates specific protein-protein interactions [14, 15]. Human FEM1B is 99% identical to mouse protein, and is reportedly able to associate with the intracellular tail of the death membrane receptors Tnfrsf6 (i.e., tumor necrosis factor receptor superfamily, member 6; also known as Fas) and Tnfrsf1a (also known as TNFR1) [16]. The region responsible for this interaction has not yet been identified. The authors have demonstrated that FEM1B can induce apoptosis when overexpressed in some cell lines and, that the domain shown to be both necessary and sufficient included the second ANK repeat in the middle of the protein (amino acids 82 to 342). Moreover, cleavage of human FEM1B by a caspase 3-like activity liberates the N-terminal region, that seems to be necessary for apoptosis to be induced [16]. However, no correlation between FEM1B expression and apoptosis in germinal cells has been investigated.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A region of PHTF1, amino acids 1–77, was appended to Gal4BD coding sequence (pGBT10-Phtf11–77) of the expression vector pGBT10. The corresponding coding sequence was cloned by polymerase chain reaction (PCR) amplification from mouse Phtf1 cDNA (GenBank accession number AJ133721). The amplification primers used for this purpose were ATGGCCTCAAATGAGAGAGATGCTATATCATGG as sense, and CCCCTTCCGAGTGAGCGAGGTCCATGG as antisense. Screening was performed by cotransformation of AH109 cells with pGBT10-Phtf11–77 and DNA from testis cDNA Library cloned into the pACT2 vector (Matchmaker GAL4; Clontech, Saint Quentin Yvelines, France). Briefly, the double transformants were plated on yeast dropout media containing 5 mM 3-aminotriazole and lacking tryptophan, leucine, and histidine. They were grown for 3 days, and then His+ colonies were patched on selective media, replica-plated on Whatman-40 filters, and tested for ß-galactosidase activity as described [17]. Some residual growth on selective media were completely prevented by the presence of 15 mM 3-aminotriazole, an inhibitor of histidine biosynthesis required to suppress the leakiness of histidine deficiency.
Plasmids
The mouse Phtf2 amino acids 1–77 and the Fem1a entire coding sequence were cloned by PCR amplification from a mouse skeletal muscle cDNA library (Marathon-Ready; Clontech). The specific primers Phtf21–77 used were ATGGCGTCGAGAGTCACAGAT and CCCTTTCCTTGTCAGAGAAGT as sense and antisense, respectively. The fragment was inserted into the pGBT-10 plasmid vector (pGBT10-Phtf21–77). The amplification primers used for Fem1a were GCATGGATCTGCACACAGCGGTGTACAATGCGG as forward and GGCTCAATGCAACTGGATGAGGCCTCCAGC as reverse, and the Fem1a fragment was cloned into the pACT2 vectors (pACT2-Fem1a).
Expression vectors encoding Fem1b (GenBank accession number AK032338) deletions were made by cloning appropriate restriction fragments obtained by partial or total digestions. The pACT2 empty vector was prepared using BglII, The vector containing only the HA tag with the use of BglII/EcoRI; the plasmids designated 1–109, 1–216, and 1–435 using NotI; the 436–627 using SacI/SmaI; and the 565–627 with EcoRI/ SacI. The numbers denote the encoding amino acids of FEM1B remaining in the construction. The correctness of the open reading frame was verified by sequencing.
The entire HA-Fem1b sequence was recovered from pACT2 vector using BglII and inserted into a pCS3-MT mammalian expression vector deleted from its myc tags (pCS-HA-Fem1b). The HA control vector was then obtained by deleting the fem1b sequence from the HA-Fem1b plasmid. The plasmids designated pCS-HA.1–471, and pCS-HA.1–383 were obtained with the use of BamHI/BglII, and the pCS-HA.472–627 using BglII. The pcDNA-Phtf1-myc vector was previously described [3].
Cell Culture and Transfections
Cells lines were maintained in Dulbecco modified Eagle medium supplemented with 10% fetal calf serum. Cells were transfected on cells at 30%–40% confluence in agreement with the ratio of 1 µg of plasmid per 10-mm well by using 2 µl of Fugene reagent as recommended by the supplier (Roche, Les Ulis, France).
Coupled Transcription-Translation and In VitroBinding Assays
Coupled transcription-translation was performed by using pCS-HA-fem1b or luciferase vectors in the presence of 20 µCi of [35S]-Methionine and by using pcDNA-Phtf1-myc or pcDNA-myc empty vector in the presence of canine microsome membranes as indicated by the manufacturer (Promega, Charbonnières, France). After translation, 10 µl were kept as input controls. For binding assays 20 µl of PHTF1 and FEM1B (or luciferase) reactions were mixed and rocked for 60 min. at room temperature. The volume was adjusted to 1 ml with TBS (NaCl 150 mM, Tris 10 mM, pH 7.4) and microsomes were recovered by centrifugation for 30 min at 105 000 x g in a TFT-55.5 rotor. Pellets were twice washed and supernatants were acetone precipitated before being recovered in Laemmli buffer. Reactions were subjected to 8% SDS-PAGE for radioactive counting or for immunoblot with the anti-HA and the anti-myc antibodies.
Cell Fractionation
Cells transfected for 24 to 36 h (5 x 106) were washed twice in TBS (10 mM Tris pH 7.4, 150 mM NaCl), scraped with a spatula, recovered in TBS-sucrose (NaCl 150 mM, Tris 10 mM pH 7.4, 5 mM EDTA, 3 mM EGTA, and 0.5 M sucrose) with antiproteases (0.7 µg/ml pepstatin, 0.5 µg/ml leupeptin, 40 µg/ml bestatin, and 100 µg/ml PMSF) and broken by 12 passages through a 25-gauge needle. Nuclei and unbroken cells (<1%) were separated by centrifugation at 3000 x g for 20 min. Low-speed membranes, P28, were harvested after centrifugation at 28 000 x g for 60 min in a TFT-55.5 rotor. High-speed pellets, P150, were collected by centrifugation for 2 h at 150 000 x g. The supernatant was precipitated with two acetone volumes, and pellets were finally recovered in 100 µl of Laemmli buffer.
For the membrane experiment, the high-speed pellet was recovered from transfected cells and washed with 0.1 M Na2CO3 or TBS solution.
Animal Procedures and Preparation of TesticularPurified Cells
All animal procedures were conducted in accordance with European guidelines for the care and use of laboratory animals. Male adult Wistar rats and CD-1 mice were obtained from Harlan France (Gannat, France). Animals were killed by i.p. injection of sodium pentobarbital.
Spermatogonia were obtained by sucrose gravity sedimentation, and pachytene spermatocytes and round spermatids were prepared by centrifugal elutriation as previously described [3].
Immunoprecipitation
Exogenous proteins Transiently transfected cells were washed twice with PBS 1x buffer, scraped, and lysed in 400 µl of NP-40 lysis buffer (1% NP-40, 0.5% sodium deoxycholate, 150 mM NaCl, 5 mM EDTA, and 3 mM EGTA) containing the antiproteases indicated in the cell fractionation section. After 30 min on ice, lysates were centrifuged at 13 000 x g for 15 min in a microfuge, and supernatants were cleared with 20 µl of protein G-Sepharose beads (Amersham, Saclay, France). Cleared lysates were incubated with 4 µg of antibodies anti-myc (9E10; Roche, Les Ulis, France), anti-HA (Y-11; Santa Cruz, Le Perray en Yvelines, France), or mp71 (anti-PHTF1, [3]) plus 20 µl of beads overnight at 4°C with rocking. The immunocomplexes were recovered by a 30-sec spin, and washed five times with NP-40 buffer containing 350 mM NaCl.
Endogenous proteins
Testes were decapsulated and tubules were minced with two scalpels. Cells were separated from intact tubules by gravity sedimentation, washed three times with PBS, and concentrated at 1500 x g. NP40-modified lysis buffer (0.1% Triton X-100 instead of sodium deoxycholate) was added at about 400 µl per 50-µl of packed cells volume. Anti-FEM1B antibodies ab801 (Abcam, Cambridge, U.K.) or anti-F1A
(Calbiochem, France Biochem, Meudon) were used at 4 µg per 400-µl of cleared lysate.
The immunoprecipitates were boiled in Laemmli buffer and loaded onto 8% or 10% SDS-polyacrylamide gels. The proteins were transferred to Hybond-C or Hybond-P membranes (Amersham).
Western Blot
Membranes were incubated in 5% nonfat milk in TBS-T (10 mM Tris pH 7.4, 150 mM NaCl, 20% Tween-20) for 1 h and the antiserum was added as follows: 1:2500 dilution of anti-myc 9E10 (Roche) or anti-HA Y-11 (Santa Cruz), 1:500 dilution of anti-F1A anti-FEM1B; Calbiochem) and ab801 (anti-FEM1B, Abcam), 1:2000 of anti-calnexin (a gift from Dr. P. Fanen, Créteil, France) or 1C9 anti-calmegin (a gift from Dr. D. Bunick, University of Illinois, Urbana, IL [18]), 1:1500 of anti-porin31HL (Calbiochem), 1:2000 of RY11 anti-µ1 antibody (a gift from Dr. L. Traub, University of Pittsburgh School of Medicine, Pittsburgh, PA [19]), 1:100 of anti-Actin H-300 (Santa Cruz), and 1:1500 of mp71 (anti-PHTF1).
Indirect Immunofluorescence
Bouin-fixed testis slides were obtained and treated as previously described [3] using 1:100 for mp71 or anti-F1A (Calbiochem); 1:25 for FITC-mp71 antibodies. Secondary antibody was the TRITC-conjugated anti-rabbit (1:750) from Jackson Laboratories (Cambridgeshire, U.K.). Controls were always performed by obviating one out of the two primary antibodies or by using preimmune sera.
Tissues were examined with a Leica TCS SP2 confocal microscopy device with a 63x objective (plan apo, NA 1.4) and with a Zeiss Axioplan 2 (plan apo, NA 1.3). Data were processed with Metamorph and Photoshop (Adobe Systems, Mountain View, CA).
Reverse Transcription-PCR Analysis
RNA was isolated by using Trizol reagent (Invitrogen, Paris, France). Reverse transcriptions were performed using random hexamers and Super Superscript II reverse transcriptase as indicated by the manufacturer (Invitrogen). PCRs were performed by using specific fem1b primers (CCGGCACCGTCCGCTTCGACGG and GCTGCGATCATTAGGCAGG as the sense and the antisense primers, respectively). Primers for Phtf1 and actin control are described elsewhere [3]. Each PCR was performed as follows: an initial 96°C denaturation for 5 min; 94°C for 15 sec, 60°C for 15 sec, and 72°C for 30 sec for 28 cycles; and a final extension at 72°C for 7 min. Amplified products (258 base pairs [bp] for Fem1b, 289 bp for Phtf1, and 1020 bp for actin) were analyzed on agarose gels.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
PHTF1 displays the properties of an integral membrane protein anchored to a cell membrane by six to eight trans-membrane domains (Fig. 1a). To search for proteins interacting with PHTF1, we used a yeast two-hybrid system to screen a murine testis cDNA expression library with the PHTF1 N-terminal part fused to the Gal4-DNA binding domain (pGBT10-Phtf11–77). Out of a total of 1.6 x 106 transformants, 15 were positive for nutritional selection and ß-galactosidase activity and, 6 were independent clones of Fem1b, the isoform of the C. elegans feminization factor 1 expressed in the mouse testis.
|
Murine FEM1B contains an N-terminal domain comprising six ANK repeats that span residues 45 to 249 and, in the C-terminal domain, two additional ANK repeats are predicted between amino acids 483 and 570 (Fig. 1b). Ankyrin domains are constituted by at least four ANK repeats in tandem array, and are believed to mediate protein-protein interactions. Because all the Fem1b clones started around the translation initiation site, leaving all the ANK repeats intact, this motif could be necessary for interaction. We then made a series of deletions in the sequence of FEM1B and used them to map the region that could associate with the PHTF1 amino terminal region by two-hybrid assay (Fig. 1c). We found that the C-terminal part of FEM1B or N-terminal fragments bearing a truncated ANK domain were not sufficient to permit association (Fig. 1c). Only the fragment spanning amino acids 1 to 435 that includes the six ANK repeats permitted an efficient association with PHTF1 (Fig. 1c).
Because the two paralogs Phtf2 and Fem1a are both expressed mainly in skeletal muscle [2, 9], we tested their ability to interact in the two-hybrid assay. Phtf21–77 failed to interact with FEM1A (Fig. 1, day). In this assay, the muscle-expressed gene Phtf21–77 interacted moderately with FEM1B, and the testis-expressed PHTF11–77 associated with both FEM1A and FEM1B.
PHTF1-FEM1B Interaction in TransfectedMammalian Cells
To analyze interactions in mammalian cells, transient transfections with the HA-tagged FEM1B and the fusion Phtf1-myc expression vectors were carried out using HEK293T cells. Extracts from cells transfected with Phtf1-myc and HA-Fem1b were immunoprecipitated using anti-myc, and Western blot analysis revealed that FEM1B was associated with PHTF1 in immunoprecipitated complexes. No FEM1B was coimmunoprecipitated by anti-myc antibodies when FEM1B and PHTF1 were expressed separately (Fig. 2a, top panel).
|
Conversely, PHTF1-FEM1B association was detected by the 9E10 anti-myc antibody when the HA-FEM1B protein was immunoprecipitated from cotransfected cells using an anti-HA antibody (Fig. 2a, bottom panel). The PHTF1-FEM1B association was dose dependent, and transfection with rising quantities of HA-FEM1B increased the size of the FEM1B reactive band when immunoprecipitated with anti-PHTF1 (mp71) or anti-myc antibodies (Fig. 2b). Using increased PHTF1 concentration of PHTF1 confirmed this finding (Fig. 2c).
A series of HA-Fem1b deletions were then cotransfected with a Phtf1-myc-expressing plasmid to identify the FEM1B domains responsible for the interaction with PHTF1. As in the two-hybrid assay, the C-terminal part of FEM1B failed to interact (Fig. 2, day). The 1–383 region of FEM1B, including all the N-terminal ANK repeats, was sufficient for binding to occur (Fig. 2, day). Moreover, the segment containing amino acids 1–472 exhibited greater PHTF1 immunoprecipitate than the 1–383 segment, suggesting that the 383–472 region may be involved in the proper conformation of the ANK domain. PHTF1 could not be recovered after larger FEM1B deletions (data not shown).
Interaction Between Endogenous PHTF1 and FEM1B in Germ Cells
Northern blot studies have previously shown that human and mouse FEM1B are abundantly expressed in the testis [9, 10]. To find out which germinal cells were expressing Fem1b, RNA from mouse testis and isolated germ cell populations was subjected to reverse transcription-PCR analysis. The specific Fem1b band appeared in spermatogonia, displayed greater intensity in spermatocytes, and was still found in spermatids (Fig. 3a). Fem1b RNA appeared earlier than Phtf1 RNA, but both were present in meiotic and postmeiotic germ cells (Fig. 3a). We measured their protein levels using anti-PHTF1 (mp71) and the two commercial anti-FEM1B antibodies (F1A and sc-801). We used calmegin, a testis meiotic and postmeiotic ER chaperone [20], and actin as loading controls. The FEM1B antibodies revealed a 68–70 kDa immunoreactive band matching the expected size (only anti-F1A is shown in Fig. 3b). FEM1B could not be detected in the spermatogonia, but was markedly induced in spermatocytes and spermatids. This distribution pattern reflected the presence of the PHTF1 and calmegin proteins (Fig. 3b).
|
Because FEM1B and PHTF1 were coexpressed in germ cells, in vivo binding assays were performed using protein extracts from testis. Mouse and rats testicular extracts were immunoprecipitated using anti-FEM1B, and Western blot analysis revealed that PHTF1 was associated with the FEM1B complexes (Fig. 3c), indicating that these two proteins interacted in the germ cells of the testis. We were not capable to detect FEM1B after immunoprecipitation of PHTF1. This effect was possibly due to the low sensitivity of anti-FEM1B antibodies, to the stability of the proteins, or to the occupancy of the anti-PHTF1 binding sites.
PHTF1 Enhances the Binding of FEM1B to Cell Membranes
We have established that the integral membrane protein PHTF1 binds FEM1B in vitro and in vivo. We further investigated how PHTF1 interacts with the predicted soluble protein FEM1B.
FEM1B/PHTF1 association was initially tested using an in vitro assay. We incubated microsomal membranes, some containing PHTF1, others not; with 35S-labeled FEM1B or firefly luciferase as a control soluble protein. After 1 h of incubation we centrifuged the samples to obtain membrane and soluble fractions, we separated proteins in denaturing polyacrylamide gel, and we measured the presence of each protein in the different fractions. Neither FEM1B nor luciferase sedimented when they were incubated with a translation reaction without microsomes (data not shown), indicating that both proteins were correctly folded in vitro, that they did not aggregate, and that FEM1B was indeed a soluble protein. When analyzing test tubes containing microsomes we found that a fraction of FEM1B could bind these membranes (Fig. 4a). This finding suggested that FEM1B finds docking sites in microsomal membranes. Furthermore, when FEM1B reactions were incubated with microsomal vesicles containing PHTF1 we observed that more FEM1B could be detected in pellets. Densitometric analysis indicated that FEM1B was 1.8 times (two experiments) more abundant in the pellet containing PHTF1, indicating that PHTF1 would readily increase the possibility of binding FEM1B to membranes (Fig. 4a).
|
To distinguish between cell membranes targeted by PHTF1, FEM1B, or both, we transfected HEK293T cells with Phtf1-myc and HA-Fem1b expression vectors and fractionated cells in low- and high-speed membrane pellets as well as a soluble fraction (Fig. 4b). PHTF1 sedimented with the calnexin-positive ER fraction, as expected [3] (Fig. 4b). Fem1b was detected in the three fractions obtained (i.e., the low-speed P28 pellet comprising the ER and mitochondria, identified with anti-calnexin and anti-porin antibodies, respectively; the high-speed P150 pellet, characterized by the µ1 subunit of the AP1 clathrin aptamer; and the soluble fraction; Fig. 4b). A low amount of FEM1B was also observed in nuclear fractions but immunofluorescence analysis showed that it was mostly associated to the ER (data not shown). No size changes were observed in the HA-FEM1B protein associated with different fractions, indicating that the association with different membranes is not accompanied by major post-translational modifications.
To determine whether the membrane-associated FEM1B was covalently linked or electrostatically associated with these membranes, we treated postnuclear membrane pellets with sodium carbonate and separated the reaction into membranes and supernatant (Fig. 4c). This treatment dissociated FEM1B from the membranes into the soluble fraction, indicating that the association with these membranes was ionic. As expected for an integral membrane protein [3], PHTF1 remained in the pellet.
Because of the broad distribution of FEM1B in the cell cytoplasm and of the presence of PHTF1 in the ER, we intended to know whether PHTF1 can actively recruit FEM1B onto the ER surface. This was investigated by transfecting rising quantities of PHTF1 and studying the distribution of FEM1B in the different fractions. The FEM1B/low-speed membrane association was clearly dependent on the dose of PHTF1, as rising quantities of PHTF1 increased the size of the FEM1B-reactive band at the low speed (P28) membrane fraction (Fig. 4d). This result indicates that FEM1B is recruited, stabilized, or both in the ER surface through its association with PHTF1.
Colocalization of PHTF1 and FEM1B in RatSeminiferous Tubules
Our analysis of PHTF1 and FEM1B expression in the different populations of germinal cells showed that they would be present in the same cells and, immunoprecipitation assays demonstrated that they effectively are present and interact in vivo. We then wanted to know in which spermatic cells the interaction would take place by immunostaining rat testicular sections and analyzing them by confocal microscopy. Representative images obtained in this analysis are exposed in Figure 5. Specific FEM1B staining was first observed in leptotene and early pachytene spermatocytes as a faint labeling covering all the cytoplasm with a granular aspect, but more condensed in the Golgi pole (Fig. 5n). This labeling became intense in pachytene spermatocytes at stage IX-X and persisted throughout spermiogenesis (see Fig. 5, b, f, and j; and the schematic representation in Fig. 5m). At spermiation, most of the protein was associated with the residual bodies, but some staining persisted in the queues of mature spermatids (data not shown).
|
Because both anti-FEM1B and anti-PHTF1 antibodies are raised in rabbit we cross-linked anti-PHTF1 antibodies to fluorescein isothiocyanate (FITC) and performed simultaneous staining of both proteins. Using the direct FITC-mp71 antibody resulted in a barely perceptible PHTF1 signal in regions previously characterized as being positive for PHTF1 in pachytene spermatocytes [3]. Because no balance between TRITC and FITC signals could be obtained in these cells, we show only conventional fluorescence microscope images in Figure 5n. Both proteins accumulate in the Golgi pole in these cells (arrows in Fig. 5n). In spermatids, where FEM1B and PHTF1 signals were intense, the colocalization could be observed clearly. In round spermatids of stages VI to VIII, pixel counts indicated that about 47% of the FEM1B signal and 39% of the PHTF1 signal (n = 82) matched (Fig. 5, g and h). The correspondence of this staining was higher in ER-positive regions surrounding the nuclear envelope than in regions adjacent to the plasma membrane (Fig. 5h). In the cytoplasmic lobes of stage XI– XIII spermatids, overlapping occurred in more than 85% of the PHTF1 signal versus only 20% of the FEM1B signal (n = 63). At these stages, PHTF1 labeling became stronger because of the condensation of the ER, and FEM1B do not show any particular site of accumulation. Figure 5m summarizes FEM1B and PHTF1 distribution in rat seminiferous epithelium. These findings suggest that PHTF1-FEM1B association takes place during meiosis and principally during the first steps of spermiogenesis. This could be indicative of an active participation in the morphogenesis of the sperm cell.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The fem-1 gene is broadly expressed in C. elegans, but the effect of null mutations seems to be limited to male germ cell development [4–6]. Studies of expression and function of fem-1 are not yet available in organisms such as insects, in which phtf would be first represented. In Drosophila, d-phtf expression is confined to the germinal cells [1]. In mammals, both Phtf1 and Fem1b are expressed in germinal cells [4–6]. Both genes harbor high degrees of conservation through evolution, and striking links with the spermatogenic process. Fem-1 appears earlier than Phtf during evolution, and it is possible to hypothesize that the complexity of spermatogenesis recalls the evolutionary process in the sense that more evolved organisms have recruited more molecules to "achieve" a more complex cell. A rapid comparison between the sperm cell FEM1B of C. elegans, insects, and mammals reveals striking changes as the presence of an acrosome, a flagellum, and a possible different mechanism of elimination of the cytoplasm remnant. New mechanisms of differentiation would intermingle with ancestral pathways but would be more favorably incorporated by adding functionalities than by modifying old ones. We can thus think Fem-1 as being part of a basic, general or primitive mechanism and PHTF1 of a latest, specialized or derived, that offers new capacities to the cell. The functions of one and the other are yet unknown, but is now possible that with the knowledge of one, we could rapidly infer the other.
Fem-1 was initially identified in the signaling pathway for sex determination, but its biochemical role is still unclear [4–6]. The human FEM1B has been shown capable of associating the intracellular tail of the death membrane receptors Tnfrsf6 and Tnfrsf1a, and its function was investigated by overexpression in some cultured mammalian cells [16]. It was reportedly able to induce apoptosis [16] and then, similar results were described for the C. elegans fem-1 and fem-2 [21, 22]. The Tnfrsf6 system has been proposed as being a strong candidate for elimination of defective cells during the spermatogenic process, and a good correlation was found between cells suffering apoptosis and expression of Tnfrsf6 [23, 24]. These studies showed that the apoptotic cells are very rare in the seminiferous tubules, and the cells affected are the spermatogonia and the spermatocytes [23, 24]. Here, we show FEM1B expression in every germinal cell from the beginning of meiosis to spermiation and, thus, both distribution and abundance of FEM1B in the seminiferous tubules contrast sharply with its function as an apoptosis inducer. An indirect role or a collaboration in this process is still possible, as several proteins involved in programmed cell death have been revealed in later steps of spermatogenesis and would be necessary for removing the residual body in mammals [25] or the waste bag in Drosophila [26]. However, a function as an apoptosis inducer is unlikely for FEM1B. Thus, a determination of the function of fem-1 in the cell and in particular, its role in germ cell fate determination, are still open questions.
PHTF1 is an integral membrane protein associated with the ER, and FEM1B is primarily a soluble protein. Our results indicate that FEM1B binds to different membrane fractions and that PHTF1 can recruit FEM1B onto the ER membrane surface. That FEM1B can be associated with several membranes constitutes a first hint into a function of FEM1B in intracellular signaling. About the function of PHTF1, one possibility is that PHTF1 acts as an anchoring (and even a sequestering) protein in the ER for FEM1B. Another idea is that PHTF1 is involved in the communication between both faces of the ER membrane via its multipass structure, with FEM1B modulating this function.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Natacha Raich, INSERM U 567-UMR 8104, 123 boulevard de Port-Royal, 75014 Paris, France. FAX: 33 01 4325 1167; raich{at}infobiogen.fr
3 Current address: UCSF Comprehensive Cancer Center, San Francisco, CA
Received: 5 September 2004.
First decision: 1 October 2004.
Accepted: 10 November 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
