Two distinct forms of the 64,000 Mr protein of the cleavage stimulation factor are expressed in mouse male germ cells
- A. Michelle Wallace*,
- Brinda Dass*,
- Stuart E. Ravnik*,
- Vijay Tonk†,
- Nancy A. Jenkins‡,
- Debra J. Gilbert‡,
- Neal G. Copeland‡, and
- Clinton C. MacDonald*,§,¶
- Departments of *Cell Biology and Biochemistry and †Pediatrics and Pathology and §Southwest Cancer Center at University Medical Center, Texas Tech University Health Sciences Center, Lubbock, TX 79430; and ‡Mammalian Genetics Laboratory, Advanced BioScience Laboratories/Basic Research Program, National Cancer Institute/Frederick Cancer Research and Development Center, Frederick, MD 21702
-
Communicated by Thomas E. Shenk, Princeton University, Princeton, NJ (received for review February 3, 1999)
Abstract
Polyadenylation in male germ cells differs from that in somatic cells. Many germ cell mRNAs do not contain the canonical AAUAAA in their 3′ ends but are efficiently polyadenylated. To determine whether the 64,000 M r protein of the cleavage stimulation factor (CstF-64) is altered in male germ cells, we examined its expression in mouse testis. In addition to the 64,000 M r form, we found a related ≈70,000 M r protein that is abundant in testis, at low levels in brain, and undetectable in all other tissues examined. Expression of the ≈70,000 M r CstF-64 was limited to meiotic spermatocytes and postmeiotic spermatids in testis. In contrast, the 64,000 M r form was absent from spermatocytes, suggesting that the testis-specific CstF-64 might control expression of meiosis-specific genes. To determine why the 64,000 M r CstF-64 is not expressed in spermatocytes, we mapped its chromosomal location to the X chromosome in both mouse and human. CstF-64 may, therefore, be absent in spermatocytes because the X chromosome is inactivated during male meiosis. By extension, the testis-specific CstF-64 may be expressed from an autosomal homolog of the X chromosomal gene.
Polyadenylation, the process of 3′ end formation in eukaryotic mRNAs, is required for synthesis, transport, translation, and stability of most mRNAs (1–4). The sequence AAUAAA specifies accurate and efficient addition of poly(A) to the mRNA 3′ end (5, 6). Substitutions at any position in this sequence diminish polyadenylation in vivo and in vitro (7–9), and 90–95% of sequenced mRNAs have AAUAAA in their 3′ ends (1, 2). However, many mRNAs expressed in male germ cells do not have AAUAAA (refs. 10–13 and Table 1), but are nevertheless efficiently polyadenylated. Sequences such as UAUAAA, AUUAAA, UACAAA, and GAUAAA might substitute for the normal polyadenylation signal (10, 11), but no experimental evidence support these assertions.
Non-AAUAAA polyadenylation signals in germ cell transcripts
One hypothesis to account for the use of non-AAUAAA signals is that a protein involved in polyadenylation is altered in germ cells. A candidate is the 64,000 M r subunit of the cleavage stimulation factor (CstF-64) (14, 15), one of three polypeptides of CstF (16). CstF-64 has been shown to be essential for growth and viability of avian cells (17) and acts by binding to a U- or G/U-rich region downstream of the cleavage site (18) whereas a 160,000 M r protein of the cleavage and polyadenylation specificity factor binds to the AAUAAA signal (19, 20). CstF-64 governs polyadenylation site choice in adenovirus (21) and is involved in the immunoglobin switch from a membrane to secretory form during B-cell maturation (22–24). CstF also is involved in the cooperation of polyadenylation with splicing (25), and recent results show that it interacts with the C-terminal domain of RNA polymerase II and thus couples polyadenylation with transcription (26).
Because of its essential role in somatic cell polyadenylation, we examined CstF-64 expression in male germ cells to test the hypothesis that it might contribute to polyadenylation of non-AAUAAA-containing mRNAs. We found that two different proteins immunologically related to CstF-64 are expressed in testis, and to a lesser extent in brain, whereas a single form is expressed in all other tissues. We found that, of the two forms, the somatic form is expressed in germ cells before and after, but not during meiosis. In contrast, the testis-specific form is expressed exclusively during meiosis and afterward. Mapping of the CstF-64 gene to the X chromosome in both mouse and human suggests that the somatic form of CstF-64 is absent in meiotic cells because of X chromosome inactivation. It further suggests that a second, autosomal gene encodes the testis-specific CstF-64. Exclusive, stage-specific expression of an essential mRNA processing protein during meiosis could be important in the global regulation of gene expression in those cells.
MATERIALS AND METHODS
Protein Analysis.
HeLa nuclear extract was made as described (29). Testicular tubules were prepared by decapsulating freshly harvested mouse testes in PBS and washing to remove interstitial cells. Nuclei were prepared by centrifugation over a sucrose pad as described (30), were boiled and sonicated in loading buffer (31), were separated by 10% SDS/PAGE, and were transferred to PVDF-Plus (Micron Separations, Westborough, MA) (32) for immunoblotting. Membranes were stained with Ponceau S and were incubated for 1–16 hours in block solution (TBS with 10% nonfat dry milk). Monoclonal antibodies 3A7 and 6A9, specific for the human CstF-64, are described elsewhere (16) and were used at a dilution of 1:50. Membranes were incubated with primary antibody in block followed by incubation with horseradish peroxidase-coupled sheep anti-mouse IgG (1:10,000). After washing in TBS, immunoreactive bands were visualized by chemiluminescence by using the Pierce SuperSignal kit.
Immunohistochemistry.
Immunohistochemistry of paraformaldehyde-fixed paraffin-embedded sections of adult mouse testis was done according to Ravnik and Wolgemuth (12) by using the Vectastain ABC kit (Vector Laboratories). Sections were blocked with 1 × PBS/0.1% Triton X-100/2.5% normal horse serum for two hours and were incubated with primary antibody at 1:100 in 1 × PBS/0.1% Triton X-100/1.5% horse serum for 16–18 hours at 4°C. After washing, sections were incubated with secondary biotinylated antibody (anti-mouse IgG, 1:200) and were incubated with avidin–biotin complex according to the manufacturer’s directions. Antibody staining was developed in 0.1 M Tris⋅HCl (pH 7.2) with 0.2 mg/ml of 3, 3′ diaminobenzidine. Slides were counterstained by using Harris’ Hematoxylin and were viewed with an Olympus (New Hyde Park, NY) BX-60 photomicroscope.
Fluorescence in Situ Hybridization (FISH).
A P1 artificial chromosome containing the human genomic CstF-64 [dJ347M6 (33), GenBank accession no. Z95327], provided by Gareth Howell (The Sanger Centre, Cambridgeshire, U.K.), was labeled with digoxigenin-11-dUTP (Boehringer Mannheim) by using the Oncor large fragment labeling kit (Oncor). Probe was hybridized with human Cot-1 DNA (GIBCO/BRL), was resuspended in hybridization solution [50% formamide/10% dextran sulfate/2× standard saline citrate (SSC)], and was denatured at 80°C before use. Metaphase spreads from human peripheral white blood cells (34) were denatured in 2× SSC and 70% formamide at 72°C, were dehydrated in an ethanol series, and were air dried. Chromosomes were hybridized at 37°C for 12–16 hours. Washes were for 5 min at 65°C in 2× SSC (pH 7.0) followed by 1× phosphate-buffered detergent (Oncor). Slides then were incubated with FITC-antidigoxigenin antibody for 45 min at 37°C, were washed in 1× phosphate-buffered detergent, and were mounted in Antifade (Oncor) with propidium iodide counterstain. Chromosomes were visualized by using a Zeiss Axioskop microscope with epifluorescence using FITC and propidium iodide single-band pass filters. Images were captured by using a charge-coupled device camera and a macktype 5.4 imaging system (Perceptive Scientific Instruments, League City, TX) with the 100×, 1.3 numerical aperture objective under oil.
Interspecific Mouse Backcross Mapping.
Interspecific backcross progeny were generated by mating (C57BL/6J × Mus spretus) F1 females and C57BL/6J males as described (35). A total of 205 N2 mice were used to map the Cstf64 locus. Southern blot analysis was performed as described (36) with Hybond-N+ nylon membranes (Amersham Pharmacia). The probe, a 342-bp fragment of mouse CstF-64 cDNA from the 3′ untranslated region [nucleotides 1809–2151 (B.D. and C.C.M., unpublished work)], was labeled with [α-32P]dCTP by using a random primed labeling kit (Stratagene); washing was done to a stringency of 0.5× SSCP (120 mM NaCl/5 mM sodium citrate/20 mM sodium phosphate, pH 6.8) and 0.1% SDS at 65°C. A fragment of 7.9 kilobases was detected in TaqI-digested C57BL/6J DNA, and a fragment of 6.0 kilobases was detected in M. spretus DNA. Presence or absence of the 6.0-kilobase TaqI M. spretus-specific fragment was followed in backcross mice. Probes and restriction fragment length polymorphisms for the loci linked to Cstf64 have been reported (37). Recombination distances were calculated by using map manager 2.6.5 (Roswell Park Cancer Institute, Buffalo, NY; http://mcbio.med.buffalo.edu/mapmgr.html). Gene order was determined by minimizing the number of recombination events required to explain the allele distribution patterns.
RESULTS
Two Forms of CstF-64 Are Expressed in Mouse Testis.
Monoclonal antibodies to human CstF-64 were used to examine CstF-64 in mouse tissues (Fig. 1). Protein immunoblots of nuclear extracts from various mouse tissues were incubated with the monoclonal antibodies 3A7 and 6A9 (16). The 3A7 antibody detected a protein of ≈64,000 M r in HeLa cell (Fig. 1 A, lane 1) and all mouse tissues examined (lanes 2–8), corresponding to CstF-64 first described in HeLa cells (14, 16).
Protein immunoblot detection of CstF-64 in mouse tissues. (A and B) HeLa nuclear extract (lane 1), nuclei from indicated mouse tissues (lanes 2–6 and 8), or whole testicular tubules (lane 7) were examined by protein immunoblotting. Testis tubules were used rather than nuclei because of the difficulty in obtaining germ cell nuclei (66). Immunoblots were incubated with 3A7 (A) or 6A9 (B) monoclonal antibodies. (C) A single immunoblotted membrane was split along one lane (5a/5b), was incubated with 3A7 (Left) or 6A9 (Right), and was reassembled. Lanes contain HeLa nuclear extract (lanes 1 and 6), whole testicular tubules (lanes 2, 5a/5b, and 7), liver (lanes 3 and 8), or thymus (lanes 4 and 9) nuclei. Arrows indicate approximate sizes of CstF-64 proteins.
With the 6A9 antibody, a 64,000 M r reactive protein was detected in HeLa nuclear extract (Fig. 1 B, lane 1). In contrast, a protein of ≈70,000 M r was detected in mouse testicular extracts (Fig. 1 B, lane 7) and in small amounts in brain (lane 2). No protein was detected in other tissues (Fig. 1 B, lanes 3–6 and 8). To ensure that the ≈70,000 M r protein detected with 6A9 did not comigrate with that detected by 3A7, we repeated immunoblots with 3A7 and 6A9 but bisected a lane containing testis extracts (Fig. 1 C, lane 5a/5b). After realignment of the filters, it was evident that the protein detected with 6A9 had a substantially different mobility than that detected with 3A7.
The ≈70,000 Mr CstF-64 Is the Only Form Expressed in Spermatocytes.
Expression of 3A7-reactive (64,000 M r) and 6A9-reactive (≈70,000 M r) CstF-64 were examined immunohistochemically in mouse testis sections. The 3A7 antibody detected immunoreactivity in nuclei of testicular somatic cells, primarily Leydig, Sertoli, and macrophage cells (Fig. 2 A) and detected protein in the nuclei of kidney, liver, thymus, and epididymal cells (data not shown). This is consistent with 3A7 immunoblot data showing the 64,000 M r form in every tissue examined.
Immunohistochemical localization of CstF-64 in mouse testis sections. Sections of mouse testis were incubated with 3A7 (A), 6A9 (B), a mixture of 3A7 and 6A9 (C), or an irrelevant control antibody, 2A6 (D). Roman numerals refer to the stage of the seminiferous epithelium. lc, Leydig cell; sc, Sertoli cell; sg, spermatogonia; spc, spermatocyte; rs, round spermatid; es, elongating spermatid. (×100.) (E) Summary of CstF-64 protein expression in mouse germ cells (see also Fig. 5). 3A7 (somatic form) reactivity is in red; 6A9 (testis-specific) is in blue. The diagram is adapted from Russell et al. (38).
Within the seminiferous epithelium, different germ cell types were characterized by their association with each other and with types found in different cross sections and were categorized into 12 stages of the seminiferous epithelium (38). 3A7 detected strong immunoreactivity in all spermatogonia and early spermatocytes. The intensity of 3A7 staining diminished in pachytene spermatocytes in stage II–IV tubules and was not detected again until after meiosis in round spermatids of stage IV tubules (Fig. 2 A and summarized in Fig. 2 E). Spermatid staining with 3A7 remained until spermatid elongation (stage X) at approximately the same time as when protein translation in spermatids ends (see Fig. 5).
In marked contrast, 6A9 recognized an epitope beginning in stage V pachytene spermatocytes and continuing in all spermatocytes to early spermatids (Fig. 2 B, summarized in Fig. 2 E). Expression continued in spermatocytes through meiosis and terminated in elongating spermatids at stage X. No staining with 6A9 was detected in early germ cells or any somatic cells (Fig. 2 B), including those from the liver, thymus, and kidney (data not shown). When testis sections were incubated with a mixture of both 3A7 and 6A9, staining was observed in all testicular cell types except late-elongating spermatids and spermatozoa (Fig. 2 C), suggesting that no cell type in the testis lacks one or the other form of CstF-64. The 2A6/E1B antibody is of the same isotype (IgG1) as 3A7 and 6A9 and was used as a control (16). This antibody displayed no appreciable nuclear or cytoplasmic staining of any testicular cell type (Fig. 2 D).
Mapping of CSTF64 to the Human X Chromosome.
The P1 artificial chromosome dJ347M6 contains the CstF-64 genomic clone and maps to the long arm of the human X chromosome (Xq22). This genomic clone is entirely colinear with the human cDNA, as confirmed by sequence comparison and Southern hybridization (data not shown).
Further confirmation of X chromosome assignment of CSTF64 was obtained by FISH analysis of human metaphase spreads by using dJ347M6 (see Methods and Materials). Over 150 spreads were examined in samples from females, and 75 were used from males. FISH showed two high-intensity signals from females (Fig. 3 A) and one from males (Fig. 3 B) on medium-sized (Group C) chromosomes that are either metacentric or submetacentric, consistent with X-chromosomal localization of CSTF64. Several lower-intensity signals also can be seen at other locations in interphase cells (Fig. 3 B).
FISH analysis of human metaphase chromosomes. Metaphase spreads from peripheral blood were prepared as described in the text. Hybridization was with dJ347M6 containing the human CstF-64 gene. (A) Representative metaphase spread from a female. (B) Metaphase spread from a male. Yellow arrows indicate X chromosomal signals from probe hybridization. The blue arrow indicates signal in an intact interphase cell.
Cstf64 Is X Chromosomal in Mouse, and Cstf64-Related Genes Are on Other Chromosomes.
The mouse chromosomal location of Cstf64 was determined by interspecific backcross analysis using progeny derived from matings of [(C57BL/6J × M. spretus) F1 × C57BL/6J] mice. This panel has been typed for >2,800 loci that are distributed among all autosomes as well as the X chromosome (35). C57BL/6J and M. spretus DNAs were digested with several enzymes and were analyzed for informative restriction fragment length polymorphisms by using a mCstF-64 3′ untranslated region probe. The 6.0-kilobase TaqI M. spretus restriction fragment length polymorphism (see Materials and Methods) was used to follow segregation of the Cstf64 locus. Results indicated that Cstf64 is located in the central region of the X chromosome linked to Pgk1 and Btk. One-hundred and thirty-five mice were analyzed for every marker (Fig. 4), and up to one-hundred and eighty-five mice were typed for some pairs of markers. Each locus was analyzed for recombination frequencies by using the additional data. Ratios of the number of mice with recombinant chromosomes to the number analyzed for each pair of loci and the most likely gene order are: centromere—Pgk1—6/135—Cstf64—1/185—Btk. The recombination frequencies (genetic distances in centimorgans ± the standard error) are Pgk1—4.4 ± 1.8—Cstf64—0.5 ± 0.5—Btk. Hybridization with a full-length cDNA probe detected multiple restriction fragments, at least one of which is autosomal (data not shown). These results suggest that there are several Cstf64-related genes in the mouse genome.
Cstf64 maps in the central region of the mouse X chromosome. (Right) Columns represent the chromosome identified in backcross progeny that inherited from the (C57BL/6J × M. spretus) F1 parent. Shaded boxes represent a C57BL/6J, and white boxes represent a M. spretus allele. The number of offspring inheriting each type of chromosome is listed at the bottom. (Left) Partial X chromosome linkage map showing Cstf64 in relation to linked genes. Recombination distances between loci are shown to the left, and positions of loci in human chromosomes are shown to the right. Human loci cited can be obtained from Genome Data Base, a computerized database maintained by The William H Welch Medical Library of The Johns Hopkins University (Baltimore).
DISCUSSION
We have shown that two distinct forms of the essential polyadenylation protein CstF-64 are expressed in mouse testis. The 64,000 M r somatic form, detected with the 3A7 monoclonal antibody, was found in all tissues examined and is undoubtably orthologous to CstF-64 described in HeLa cells (16). An ≈70,000 M r form, detected with the 6A9 antibody, was seen only in extracts from testis and brain. We also have shown that the CstF-64 gene is encoded by a locus on the X chromosome in both mouse and human. In mouse, Cstf64-related genes are present on other chromosomes.
We initially were surprised that we saw two forms of CstF-64 in mouse tissues using antibodies that recognize only a single form in HeLa cells. However, subsequent experiments have shown that the human cDNA for the somatic CstF-64 encodes determinants for both the 3A7 and 6A9 epitopes (B.D., K. W. McMahon, and C.C.M., unpublished work). In these same experiments, the mouse somatic cDNA encodes only the 3A7 epitope. Presumably, a similar analysis of the cDNA for the mouse testis-specific CstF-64 would have only the 6A9 epitope. This fortuitous segregation of the two epitopes has made it possible for us to examine the two forms individually in mouse, which would not be possible in human.
Immunohistochemical results (summarized in Figs. 2 E and 5) established expression of the somatic form of CstF-64 (3A7) in spermatogonia and early spermatocytes. In contrast, the testis-specific CstF-64 (6A9) was the only form expressed during most of meiosis. Subsequently, in round spermatids and early elongating spermatids, both the somatic and testis-specific forms were expressed. This suggests the possibility of three phases for polyadenylation in germ cells: phase I (premeiosis), in which normal polyadenylation is mediated by the somatic form of CstF-64. Phase II (meiosis) would be mediated by the testis-specific form. Phase III (postmeiosis) would be mediated by the simultaneous presence of both the somatic and testis-specific forms of CstF-64. This also suggests the possibility that the testis-specific form of CstF-64 is responsible for use of non-AAUAAA polyadenylation signals in germ cell mRNAs during and after meiosis (see Table 1).
Representation of essential events in mouse spermatogenesis with emphasis on sex chromosome transcription and CstF-64 expression. Stage and times (in days) of spermatogenesis are shown at top. Approximate durations of molecular events (Middle) and CstF-64 protein expression (Bottom) are expressed as bars.
The functional significance of these unconventional polyadenylation signals in male germ cells is not clear. Many mRNAs are expressed in multiple tissues, but the testicular message uses a different polyadenylation site (see Table 1). For example, the mRNA for the c-abl oncogene uses a different site during germ cell maturation (10), as does pim-1 (39), CREMτ (40, 41), and cyclin A2 (12). In each case, the polyadenylation mechanism overlooks a canonical AAUAAA polyadenylation signal for a germ cell-specific site, yet these mRNAs have normal poly(A) tails.
FISH analysis in human and interspecific backcross analysis in mouse maps the CstF-64 gene to the X chromosome in both species (Figs. 3 and 4). We compared our map of the X chromosome with a mouse linkage map of uncloned mouse mutations (Mouse Genome Database, The Jackson Laboratory). Cstf64 is in a region that lacks mutations with a phenotype that might be expected for an alteration in this locus. However, this region shares homology with the long arm of the human X chromosome (Fig. 4), consistent with the finding that the human homolog of Cstf64 maps to Xq.
CstF-64 is a phosphoprotein (refs. 22 and 42; C.C.M., unpublished work). However, the difference in mobility of the ≈70,000 M r and the 64,000 M r forms likely are not a result of phosphorylation; our experiments with one- and two-dimensional electrophoresis show that the ≈70,000 M r form does not comigrate with any phosphorylated form of somatic CstF-64 (data not shown). Therefore, three mechanisms might account for different protein forms of CstF-64 in germ cells: (i) the testis-specific protein could be the product of a different gene; (ii) it could be a post-translational modification of the primary protein that does not involve phosphorylation; or (iii) it could be an alternatively spliced product of the mRNA. Although we have not completely eliminated the second or third possibilities, we believe our data strongly support the first. Specifically, because the gene for the somatic CstF-64 is on the X chromosome, transcription of that gene is likely inactivated during male meiosis (28). For inactivation of essential X-linked genes, often an intronless retroposon on an autosome is activated during meiosis to replace that function, as is true for phosphoglycerate kinase-1 (27, 43), pyruvate dehydrogenase E1α subunit (44), and glucose-6-phosphate dehydrogenase (45). Based on those studies, we propose that the ≈70,000 M r CstF-64 is the product of an autosomal gene. In mouse, we have evidence for several autosomal homologs of Cstf64. We are exploring the possibility that one of these encodes the testis-specific CstF-64.
Curiously, the one other tissue in which we see low levels of ≈70,000 M r CstF-64 expression is brain. Many testis-expressed genes also are found in brain (46–56). Similarly, there are many examples of differential mRNA processing in brain, the best known of which is the switch between calcitonin and calcitonin gene-related peptide (57, 58). Here, inclusion or exclusion of exon 4 depends on the interaction of CstF-64 with downstream sequences and other elements.
It is also unusual that expression of the somatic form of CstF-64 resumes after meiosis, suggesting that the somatic CstF-64 gene escapes inactivation after meiosis. Although most inactivated genes remain so after meiosis, a rarer class is reactivated in spermatids (59). In fact, our preliminary data suggest that the mRNA for the somatic CstF-64 is greatly overexpressed in postmeiotic germ cells (B.D., A.M.W., and C.C.M., unpublished work), reminiscent of the mRNA for the TATA-binding protein (60, 61). Several recent studies have highlighted the interdependence of polyadenylation and transcription (26, 62–65). It is therefore interesting to speculate that future data might uncover unique functional interactions of these essential processes in male germ cells.
Acknowledgments
We thank Andreé Reuss for excellent technical assistance, Gareth Howell for the gift of reagents, and Daniel Hardy and Simon Williams for critical reading of the manuscript. This work was supported by grants from the American Heart Association, Texas Affiliate and the Wendy Will Case Cancer Fund, Inc. (to C.C.M.), by the National Cancer Institute, Department of Health and Human Services, under contract with Advanced BioScience Laboratories (to N.A.J. and N.G.C.), and by a Helen Hodges Educational Charitable Trust Scholarship (to B.D.).
Note
Beyer et al. (92) noted a 70,000 Mr protein in purified CstF from calf thymus that is recognized by antibodies to CstF-64, which might be the same protein we describe here.
Footnotes
-
↵ ¶ To whom reprint requests should be addressed at: Department of Cell Biology and Biochemistry, Texas Tech University Health Sciences Center, 3601 Fourth Street, Lubbock, TX 79430. e-mail: cbbccm2{at}ttuhsc.edu.
- ABBREVIATIONS:
- CstF,
- cleavage stimulation factor;
- FISH,
- fluorescence in situ hybridization
- Copyright © 1999, The National Academy of Sciences