BIOLOGY OF REPRODUCTION 70, 1080±1087 (2004) Published online before print 17 December 2003. DOI 10.1095/biolreprod.103.022947

Developmental Distribution of the Polyadenylation CstF-64 and the Variant ␶CstF-64 in Mouse and Rat Testis1

A. Michelle Wallace,3 Toni L. Denison, Ebtesam N. Attaya, and Clinton C. MacDonald2 Department of Cell Biology & Biochemistry, Texas Tech University Health Sciences Center, Lubbock, Texas 79430

ABSTRACT was expressed, although to a lesser extent, in mouse brain.

The somatic CstF-64 protein cannot be expressed during Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 Messenger RNA polyadenylation is one of the processes that meiosis, because the CstF-64 , CSTF2, is on the X control gene expression in all eukaryotic cells and tissues. In , which is inactivated during male meiosis [12, mice, two forms of the regulatory polyadenylation protein CstF- 64 are found. The gene Cstf2 on the X chromosome encodes this 13]. This led us to the hypothesis that ␶CstF-64 was en- form, and it is expressed in all somatic tissues. The second form, coded by a second gene on an autosome. The mouse gene ␶CstF-64 (encoded by the autosomal gene Cstf2t), is expressed for ␶CstF-64, Cstf2t, was cloned and mapped to chromo- in a more limited set of tissues and cell types, largely in meiotic some 19, con®rming its existence [11], and human CSTF2T and postmeiotic male germ cells and, to a smaller extent, in was mapped to [14]. brain. We report here that whereas CstF-64 and ␶CstF-64 ex- The CstF-64 is one of many X-linked that are pression in rat tissues resembles their expression in mouse tis- inactivated during male meiosis [12]. Somatic X-chromo- sues, significant differences also are found. First, unlike in mice, some inactivation in female eutherian mammals has been in which CstF-64 was expressed in postmeiotic round and elon- known for much longer [15, 16], and it likely evolved to gating spermatids, rat CstF-64 was absent in those cell types. compensate for gene dosage in the homogametic sex [17, Second, unlike in mice, ␶CstF-64 was expressed at significant 18]. Transcriptional silencing of sex-determining chromo- levels in rat liver. These differences in expression suggest inter- somes is a phenomenon unique to mammals [19]. Unlike esting differences in X-chromosomal gene expression between somatic X-chromosome inactivation that occurs in females these two rodent species. for reasons of dosage compensation [15], X-chromosome gametogenesis, gene regulation, meiosis, spermatogenesis, testis inactivation in male eutherian mammals occurs only in germ cells. According to the leading hypothesis, because the X and Y have only limited regions of INTRODUCTION homology, synapsis is prevented by sequestration of the X and Y chromosomes in the microscopically visible ``sex Messenger RNA processing is one of the most important vesicle'' or ``XY body'' during the pachytene stage of sper- mechanisms by which diversity of gene expression is con- matogenesis [12]. Interestingly, although the Xist transcript trolled in eukaryotes [1]. In our investigation of mRNA can be detected in the XY body at this time [20, 21], tar- processing in male germ cells, we found that one of the geted deletion of the Xist gene, which defeats somatic X most highly conserved RNA-processing elements, the ca- inactivation in females, has little or no effect in males [22, nonical polyadenylation signal AAUAAA, was absent from 23], suggesting strongly that the mechanism for male sex the 3Ј-ends of a large number of mRNAs expressed in male germ cells [2, 3]. Furthermore, a number of signi®cant chromosome inactivation differs from that in females. mRNAs that were expressed in both germ cells and somatic The CstF-64 is one of three CstF subunits required for cells used a different polyadenylation site in germ cells than polyadenylation. The CstF, along with poly(A) polymerase, in somatic cells [4±10]. This suggested key differences in the cleavage-speci®city factor (CPSF), and two cleavage the mechanism of polyadenylation in germ cells compared factors (CFI and CFII), can recapitulate polyadenylation in with that in somatic cells. In examining candidate vitro [24±26]. During polyadenylation, CstF interacts with that might be involved in altering germ cell polyadenyla- the pre-mRNA through the RNA-binding domain of CstF- tion, we discovered that one essential polyadenylation pro- 64 [27] at U- or GU-rich sequences [28] within 10 to 30 tein, the 64 000-M subunit of the cleavage stimulation fac- nucleotides downstream of the cleavage site [29]. Interac- r tion of CstF with CPSF promotes the cooperative binding tor (CstF-64), had an approximately 70 000-Mr variant form that was expressed in meiotic and postmeiotic male germ of the 160 000-Mr subunit of CPSF to the polyadenylation cells in mice [2]. The variant protein, ␶CstF-64 [11], also signal, usually AAUAAA in somatic cells [30±32], fol- lowed by subsequent cleavage of the pre-mRNA and 1Supported by the NIH (1 R01 HD37109-01A1), the South Plains Foun- poly(A) addition [33±35]. dation, and Houston Educational Institute. We were interested to learn whether the developmental 2Correspondence: Clinton C. MacDonald, Department of Cell Biology & patterns of CstF-64 and ␶CstF-64 that we saw in mice was Biochemistry, Texas Tech University Health Sciences Center, 3601 4th a common pattern or one that differed among species, be- Street, Lubbock, TX 79430. FAX: 806 743 2990; ginning with other rodents. We found that in rats, distri- e-mail: [email protected] 3 bution of ␶CstF-64 in meiotic and postmeiotic germ cells Current address: Lexicon Genetics, Inc., The Woodlands, TX 77381. seemed to be nearly identical to that in mice, supporting its proposed role in those cells [2, 11]. Therefore, we were Received: 4 September 2003. surprised to ®nd that unlike in mice, the somatic CstF-64 First decision: 25 September 2003. Accepted: 1 December 2003. was not expressed after meiosis in rats [36]. We were fur- ᮊ 2004 by the Society for the Study of Reproduction, Inc. ther surprised to observe that unlike in mice, rat liver ex- ISSN: 0006-3363. http://www.biolreprod.org pressed high levels of ␶CstF-64. Finally, we observed that 1080 CstF-64 ISOFORMS IN RAT AND MOUSE DEVELOPMENT 1081

in both rats and mice, spleen and thymus express low but detectable levels of ␶CstF-64, which is consistent with a plausible hematopoietic or immunological role for that pro- tein. These differences between mice and rats suggest that in at least some tissues in some species, functions of the two forms of CstF-64 can substitute for one another, where- as in other tissues, the two forms might have complemen- tary functions.

MATERIALS AND METHODS Animals In all studies, animal protocols were approved by the Institutional An- imal Care and Use Committee, and NIH guidelines for animal care were Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 followed. For developmental studies, male CD-1 mice (at 7, 13, and 17 days postpartum or retired breeders [6±12 mo of age]) and male Wistar rats (at 10, 20, and 40 days postpartum) were anesthetized with CO2 and killed by cervical dislocation. For protein analyses, tissues were removed and washed in ice-cold PBS before nuclei preparation (discussed below). Testes were quickly removed, ®xed in 4% paraformaldehyde in PBS, and sectioned into 5-␮m sections for histological evaluation. For studies of ␶ adult animals, male CD-1 mice (age, Ͼ35 days) and male Wistar rats (age, FIG. 1. Protein blots comparing CstF-64 (A) and CstF-64 (B and C) ex- pression in nuclei from rat (A and B) and mouse (C) tissues. Nuclei (lanes 62±67 days) were anesthetized with CO2, killed by cervical dislocation, and perfused by ventricular puncture with PBS followed by 4% parafor- 2–5 and 7) or testicular tubules (lanes 1 and 6) were prepared as de- maldehyde in PBS. Testes and livers were removed and sectioned for scribed previously [2]. Loading was normalized to nuclear DNA content histological evaluation. as described in Materials and Methods. Samples were mouse testis (lane 1), brain (lane 2), liver (lane 3), lung (lane 4), spleen (lane 5), testis (lane 6), and thymus (lane 7). Antibodies The speci®cities of the monoclonal antibodies 3A7 and 6A9, which were raised against the human CstF-64 [37], were described previously caused by inactivation of Cstf2 during spermatogenesis and [2]. In mice, the 3A7 antibody is speci®c for the somatic CstF-64, whereas activation of the autosomal Cstf2t gene for ␶CstF-64 [2, 6A9 is speci®c for ␶CstF-64. A control antibody, ␣100k-1 is a monoclonal 11]. As further support for this hypothesis, we wished to antibody to the adenovirus 100k DNA-binding protein and is not expressed examine CstF-64 expression in other species, including oth- in uninfected mouse tissues. er rodents, starting with rats. Detection of the two forms of CstF-64 was facilitated, Immunohistochemistry because in mice, the 3A7 monoclonal antibody detected Immunohistochemistry of paraformaldehyde-®xed, paraf®n-embedded exclusively somatic CstF-64 whereas the 6A9 monoclonal sections of developmental rat, adult rat, and mouse testis as well as adult antibody detected exclusively the variant ␶CstF-64 [2]. To rat and mouse liver was done as described previously [2, 9] using the determine the distribution of the two CstF-64 isoforms in Vectastain ABC kit (Vector Laboratories, Burlingame, CA). Immunore- rat tissues, we used the 3A7 and 6A9 antibodies to perform active proteins were developed in 0.1 M Tris-HCl (pH 7.2) with 0.2 mg/ ml of 3,3Ј-diaminobenzidine. Slides were counterstained using Harris he- protein immunoblot analysis of nuclei from several rat tis- matoxylin, mounted with Permount (Fisher Scienti®c, Houston, TX), and sues and seminiferous tubules (Fig. 1, A and B) and com- viewed with an Olympus BX-60 photomicroscope (Olympus, Melville, pared those with tissues from mice (Fig. 1C). Protein sam- NY). ples from mouse testis (lane 1) and rat or mouse brain, liver, lung, spleen, testis, and thymus nuclei (lanes 2±7, respec- Protein Analysis tively) were normalized to DNA content, separated by 10% Nuclei from rat and mouse tissues and testicular tubules were prepared SDS-PAGE, transferred to PVDF membranes, and exam- as described previously [2]. Samples were sonicated brie¯y in loading ined by immunoblotting (see Materials and Methods). buffer [38], boiled brie¯y, separated by 10% SDS-PAGE, and transferred Membranes were incubated with either the 3A7 (Fig. 1A) to polyvinylidene ¯uoride (PVDF) membranes for protein immunoblot or 6A9 (Fig. 1, B and C) monoclonal antibodies. analysis [2, 39]. In an effort to adjust for differing cellular protein contents As previously shown [2], 3A7 reacted with the approx- in somatic tissues, sample loading was normalized to the DNA content imately 64 000-M CstF-64 protein in mouse seminiferous for each tissue using the Bio-Rad Fluorescent DNA Quantitation Kit (Bio- r Rad, Hercules, CA) to control for cell number. Because stages within tubules (Fig. 1A, lane 1). Similarly, CstF-64 was detected seminiferous tubules have differing DNA contents, tubule samples were in nuclei prepared from rat brain, liver, lung, spleen, testis, compared to other tissues comparing glyceraldehyde phosphate dehydro- and thymus (Fig. 1A, lanes 2±7, respectively). Slight dif- genase, and the loading was adjusted accordingly (data not shown). Protein ferences in levels of CstF-64 were detected among rat tis- immunoblotting was performed as described previously [2]. sues (Fig. 1A, compare lanes 2±7), although the signi®- cance of the differences is not clear. Also, slight differences RESULTS were observed in the mobility of CstF-64 among the rat tissues (Fig. 1A, compare lanes 5±7). Because these differ- Distribution of CstF-64 and ␶CstF-64 Proteins ences do not re¯ect cross-reactivity of 3A7 with the more in Rat Tissues slowly migrating ␶CstF-64 protein (data not shown; see be- To summarize our previous ®ndings in mice, CstF-64 low), differences in phosphorylation might be the cause was found in the nuclei of somatic cells in every tissue [37, 40]. Finally, differences were observed in relative mo- examined, whereas ␶CstF-64 was restricted to the nuclei of bility between rat and mouse CstF-64 that might be caused testicular germ cells and brain [2]. Because Cstf2, the gene by differential phosphorylation of the CstF-64 proteins or for CstF-64, was found on the X chromosome, we hypoth- that might represent a species difference between the pro- esized that the pattern of expression in mouse testis was teins (Fig. 1A, compare lanes 1 and 2). 1082 WALLACE ET AL.

FIG. 2. CstF-64 and ␶CstF-64 expression during mouse spermatogonial develop- ment. Male mice were killed at 7 days (A and B), 13 days (C and D), 17 days (E and F), or greater than 6 mo (‘‘adult’’; G and H) after birth, and testes were removed and prepared for immunohistochemistry (see Materials and Methods). Paraffin sec- tions of paraformaldehyde-fixed tissue were incubated with either the 3A7 (CstF- 64; A, C, E, and G) or 6A9 (␶CstF-64; B, D, F, and H) monoclonal antibodies, and antibody interactions were detected with 3,3Ј-diaminobenzidine (DAB; leaves a brown staining pattern). After antibody de- tection, sections were counterstained with Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 Harris hematoxylin (blue). Sections incu- bated with an irrelevant control monoclo- nal antibody showed little or no antibody staining (not shown). For a discussion of predominant cell types at each stage, see Results. Magnification ϫ40.

In agreement with our earlier ®ndings [2], the 6A9 an- results, we examined CstF-64 and ␶CstF-64 in the testes of tibody detected a protein of approximately 70 000 Mr in mice at different ages after birth (Fig. 2), during which time mouse seminiferous tubules (Fig. 1, B, lane 1, and C, lane germ cells are developing and speci®c known cell types are 6). Also, in agreement with those ®ndings, a similarly sized present [41]. At 7 days after birth, the 3A7 antibody (which protein was detected in rat brain (Fig. 1B, lane 2) and testis detects the somatic CstF-64 in mice) detected immunore- (Fig. 1B, lane 6). These ®ndings are consistent with the active protein staining in nuclei of nearly every cell type, pattern seen for the ␶CstF-64 protein in mouse (Fig. 1C, including primitive type A spermatogonia and Sertoli cells lanes 2 and 6). However, unlike what we reported earlier (Fig. 2A). At Day 13, 3A7 staining continued in type A in the mouse [2], ␶CstF-64 was detected in rat spleen and and B spermatogonia and in preleptotene, leptotene, and thymus (Fig. 1B, lanes 5 and 7, respectively). zygotene spermatocytes (Fig. 2C). However, staining was More unexpectedly, the 6A9 antibody detected high lev- diminished in early to midpachytene spermatocytes. This is els of ␶CstF-64 in rat liver (Fig. 1B, lane 3), although very consistent with our earlier ®nding that CstF-64 protein was little was detected in mouse liver (Fig. 1C, lane 3). Similar not expressed in stage V pachytene spermatocytes [2]. Sim- high levels of ␶CstF-64 were detected in liver extracts pre- ilar patterns were seen at Day 17 (Fig. 2E), at which time pared from several different rats from several different spermatogonia and early spermatocytes are stained but preparations, whereas little or none has ever been detected pachytene spermatocytes become more abundant and are in mouse liver extracts from three different strains of mouse not stained [41]. Finally, in agreement with our earlier (Fig. 1C, lane 3, and data not shown). study [2], sections from adult mouse testis show 3A7 stain- ing in resident somatic cells (Leydig, macrophage, and Ser- CstF-64 Was Detected from Earliest Times (Day 7) toli), spermatogonia, and early spermatocytes as well as in Immature Mouse Testis, But ␶CstF-64 Was Not round and early elongating spermatids (Fig. 2G). Detected until Day 13, When Pachytene Spermatocytes At Day 7 (Fig. 2B), few, if any, cells were stained with Appeared in Mice the 6A9 antibody, suggesting that ␶CstF-64 was not ex- Earlier, we examined CstF-64 and ␶CstF-64 expression pressed in spermatogonia, which are developing at this time in testes of adult mice [2]. To con®rm and extend these [41]. However, by Day 13 (Fig. 2D), although not earlier, CstF-64 ISOFORMS IN RAT AND MOUSE DEVELOPMENT 1083

FIG. 3. Immunohistochemical localiza- tion of CstF-64 and ␶CstF-64 expression in adult rat testis sections. Adult male rats (age, 62–67 days) were killed and per- fused with paraformaldehyde, and testes were removed for immunohistological sec- tioning and examination. Testes were incu- bated with either the 3A7 (CstF-64; A–C) or 6A9 (␶CstF-64; D–F) monoclonal anti- bodies, and antibody interactions were de- tected with 3,3Ј-diaminobenzidine (DAB). Sections in A, B, D, E, and G were coun- terstained as for Figure 2; Sections in C, F, and H did not receive counterstain. Con- trol sections (G and H) were incubated with ␣100k-1, an irrelevant antibody that Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 does not react with mouse tissues. Roman numerals refer to the stage of the seminif- erous epithelium. ES, Elongating sperma- tid; LC, Leydig cell; SC, Sertoli cell; SPC, spermatocyte. Magnification ϫ40 (A, C, D, and F and ϫ100 (B, E, G, and H).

a number of centrally located cells showed nuclear staining staining diminished in pachytene spermatocytes in stage II± with 6A9, suggesting ␶CstF-64 expression in pachytene III tubules and was undetectable in any other meiotic or spermatocytes. This result was further seen in developing postmeiotic cell types. This was in sharp contrast to 3A7 pachytene spermatocytes at Day 17 (Fig. 2F). Results with staining in mice, which was not detected in pachytene sper- the 6A9 antibody at Days 9 and 11 postpartum were con- matocytes but which resumed in round spermatids and early sistent with these observations (data not shown). In adult elongating spermatids (Fig. 2G) [2]. (These data are also mice, ␶CstF-64 was detected in stage V pachytene sper- summarized in Fig. 6.) matocytes continuously through stage XI elongating sper- matids (Fig. 2H and data not shown), exactly as described earlier [2]. (These data are summarized in Fig. 6.) ␶CstF-64 Was Detected in Pachytene Spermatocytes and in Round and Early Elongating Spermatids in Rats Unlike in Mice, CstF-64 Was Not Expressed in Meiotic Using the 6A9 antibody, ␶CstF-64 protein was not de- or Postmeiotic Cells in Rat Testis tected in resident somatic cells in adult rat testis, nor was To examine possible differences in rodent species, we it detected in spermatogonia (Fig. 3, D±F). The ␶CstF-64 examined CstF-64 and ␶CstF-64 in rat testicular sections was ®rst detected in stage III±IV pachytene spermatocytes using the 3A7 and 6A9 antibodies. The 3A7 antibody de- and increased in intensity through stage XIV. Similarly, tected CstF-64 in nuclei of testicular somatic cells, primar- ␶CstF-64 was detected strongly in round spermatids and ily Leydig, Sertoli, and macrophage cells (Fig. 3, A±C), diminished in early elongating spermatids (stage X). In which is consistent with our previous ®ndings in mice [2]. these stages of spermatogenesis, the somatic CstF-64 is ab- In the seminiferous epithelium, different rat germ cell types sent; this is very similar to the pattern seen previously in were characterized by their association with each other and mouse testis (Fig. 2) [2]. However, unlike in mice, no cell with types found in different cross-sections and then cate- type in rat germ cells expresses CstF-64 and ␶CstF-64 si- gorized into 14 stages [42]. The 3A7 detected strong im- multaneously (data summarized in Fig. 6). No signal was munoreactive protein in spermatogonia and in leptotene and detected in histochemical sections using a control antibody zygotene spermatocytes (Fig. 3, A±C). The intensity of 3A7 (Fig. 3, G and H). 1084 WALLACE ET AL.

FIG. 4. CstF-64 and ␶CstF-64 expression during rat spermatogonial development. Male rats were killed at 10 days (A and B), 20 days (C and D), 40 days (E and F), and 65 days (‘‘adult’’; G and H) after birth, and testes were removed and prepared for immunohistochemistry (see Materials and Methods). Paraffin sections of paraformal- dehyde-fixed tissue were incubated with either the 3A7 (CstF-64; A, C, E, and G)or 6A9 (␶CstF-64; B, D, F, and H) monoclo- nal antibodies as for Figure 2. Sections in- cubated with an irrelevant control mono- clonal antibody showed little or no anti- body staining (see Fig. 3). For a discussion of predominant cell types at each stage, Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 see Results. Magnification ϫ40.

␶CstF-64 Was First Detected at Day 20 after Birth 4F). This is consistent with the meiotic and postmeiotic in Rat Testis distribution of ␶CstF-64 seen in adult rat testis (Figs. 3, D± E, and 4H). To determine more accurately when CstF-64 and ␶CstF- 64 expression occurred, we examined their expression in ␶CstF-64 Was Detected in Hepatocytes But Not in Other testis sections from rats at different days after birth. As expected, even at the earliest time point (Day 10), the 3A7 Cell Types in Rat Liver antibody detected CstF-64 protein in somatic cell types and Because we detected ␶CstF-64 protein in rat liver ex- early germ cells, likely type A spermatogonia (Fig. 4A). At tracts but not in mouse liver, we wished to determine in Day 20, spermatogonia were stained as at Day 10, but stain- which hepatic cell types the CstF-64 and ␶CstF-64 proteins ing diminished with the appearance of later spermatocyte were expressed. Immunohistochemical analysis was per- cell types (Fig. 4C). The CstF-64 was not detected in round formed on paraformaldehyde-®xed mouse and rat liver sec- spermatids that appeared at Day 40 (Fig. 4E). tions using the 3A7 and 6A9 antibodies. Using the 3A7 In contrast, 6A9 did not detect protein in somatic cells antibody, we con®rmed that the somatic CstF-64 was ex- at Day 10 (Fig. 4B), suggesting that ␶CstF-64 was absent pressed in the nuclei of almost every hepatic cell type in at those times. However, light staining was seen in cells at both rat (Fig. 5A) and mouse (Fig. 5F) liver, including he- the periphery of seminiferous tubules (Fig. 4B, arrows); patocytes and endothelial cells. In both rat and mouse, some similar staining was not seen in control experiments (data sinusoidal endothelial cells are less highly stained using this not shown, but see Fig. 3, G and H). This ®nding suggests antibody (Fig. 5A, indicated as ``S''). Also, dark spots of a window of ␶CstF-64 expression early during rat sper- staining seen in mouse sinusoids probably were erythro- matogenesis in type A spermatogonia, which is surprising, cytic staining caused by endogenous peroxidase in eryth- because no staining of the corresponding cell types was rocytes, because they are seen in 3A7-stained (Fig. 5F) and seen in adult rat testis (Figs. 3, D±F, and 4H) or in mice 6A9-stained (Fig. 5G) sections as well as in controls (not (Fig. 2, B and H). shown). Rat sections (Fig. 5, A±E) lacked this pattern, be- Strong staining by 6A9 was seen in spermatocytes by cause the animals were more extensively perfused before Day 20 (Fig. 4D) and in round spermatids at Day 40 (Fig. ®xation. CstF-64 ISOFORMS IN RAT AND MOUSE DEVELOPMENT 1085

FIG. 5. Expression of CstF-64 and ␶CstF- 64 in rat and mouse liver. Paraformalde- hyde-fixed histological sections from rat (A–E) or mouse (F and G) liver were incu- bated with either the 3A7 (CstF-64; A, B, and F), 6A9 (␶CstF-64; C, D, and G), or a control antibody (␣100k-1; E) as for Figure 2. Some sections (A and C) were incubat- ed with hematoxylin for contrast. BD, Bile duct; CV, central vein; H, hepatocyte; HA, hepatic artery; K, Kuppfer cell; PV, portal vein; S, sinusoidal endothelial cell. Magni- fication ϫ100. Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021

Using the 6A9 antibody, little or no ␶CstF-64 was de- which is encoded by Cstf2t on mouse chromosome 19 [11] tected in any cell type in mouse liver (Fig. 5G), although and by CSTF2T on human chromosome 10 [14]. peroxidase staining of erythrocytes was apparent (blue col- To complement the description of CstF-64 and ␶CstF-64 or in the majority of mouse liver nuclei was caused by the expression in mouse somatic and testicular cells, we ex- hematoxylin-and-eosin counterstain). In contrast, the 6A9 amined expression of both proteins in rat tissues. Whereas antibody detected ␶CstF-64 in the nuclei of the majority of the overall patterns of expression of CstF-64 and ␶CstF-64 hepatocytes in rat liver (Fig. 5, C and D). The ␶CstF-64 were very similar in both rats and mice, we found a number was not detected in sinusoidal endothelial cells or in en- of striking differences between the patterns of expression dothelial cells lining the portal vein in rats. Irrelevant con- of CstF-64 and ␶CstF-64 in mice and in rats (summarized trol antibodies showed little or no staining in either mouse in Fig. 6). Like in mice, CstF-64 protein in rats was ex- (not shown) or rat (Fig. 5E) liver, although erythrocytes pressed in most, if not all, somatic cell types (Fig. 1A), and accumulated stain because of peroxidase activity in mouse ␶CstF-64 was expressed in rat testis and brain (Fig. 1B). (Fig. 5, F and G). However, unlike in mice, ␶CstF-64 was expressed at high levels in rat liver (Figs. 1B, lane 3, and 5, C and D), and DISCUSSION the somatic CstF-64 was not expressed in postmeiotic male We undertook this study to increase our understanding germ cells. Therefore, the major differences between rats of the expression of the two isoforms of the CstF-64 poly- and mice are the strong expression of ␶CstF-64 in rat liver adenylation protein in rodent testes. The gene for the so- and the lack of expression of CstF-64 in postmeiotic rat matic CstF-64 protein, Cstf2, is on the X chromosome [2] germ cells. Because of this lack of postmeiotic expression and, therefore, is inactivated during the pachytene stage of of CstF-64, at no time during rat spermatogenesis are CstF- spermatogenesis. This suggested the necessity for a second, 64 and ␶CstF-64 both expressed. autosomal isoform of CstF-64 that would function during It is not yet clear to us why CstF-64 should not be ex- spermatogenesis. Our previous studies identi®ed a variant pressed in postmeiotic rat germ cells when it is expressed isoform of CstF-64 that we called ␶CstF-64, the gene for in postmeiotic mouse germ cells, but we suspect the reason

FIG. 6. Comparison of CstF-64 and ␶- CstF-64 protein expression during mouse and rat spermatogenesis. Top) Timeline of events during mouse (above line, 34 days total) and rat (below line, 48 days total) spermatogenesis. Middle) Expression pat- terns of CstF-64 (dark gray) and ␶CstF-64 (light gray) in mouse germ cells. Bottom) Expression patterns of CstF-64 (dark gray) and ␶CstF-64 (light gray) in rat germ cells. 1086 WALLACE ET AL. is linked to the unusual dynamics of X-chromosome ex- species. These differences and others may be the keys to pression in male germ cells. The CstF-64 is among the X- uncovering answers to questions about the role of the CstF- linked genes that are inactivated during pachytene of mei- 64 gene family in tissue-speci®c control of gene expression. osis in spermatocytes [2]. Most X-linked genes that are inactivated during male meiosis remain inactive during sub- ACKNOWLEDGMENTS sequent spermiogenesis [19, 43±45]. It is more rare that genes become active postmeiotically [46, 47], although spe- The authors wish to acknowledge Dr. Sha®k Khan for samples of rat ci®c examples do exist, such as Akap4, Ube-1, and Smage testis, James Hutson and Stuart Ravnik for help in data analysis, Brandt [48±51]. In mice, CstF-64 falls into the latter category and Schneider for use of his ¯uorometer, and Wyatt McMahon for critical reading of the manuscript. is expressed postmeiotically as well as premeiotically [2]. In fact, in mice, CstF-64 mRNA is greatly overexpressed after meiosis [36]. This leads to questions about the post- REFERENCES meiotic function of CstF-64 in mouse germ cells and 1. Tupler R, Perini G, Green MR. Expressing the . Nature whether it is dispensable in rats. Furthermore, the differ- 2001; 409:832±833. Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 ences in postmeiotic CstF-64 expression in mice and rats 2. Wallace AM, Dass B, Ravnik SE, Tonk V, Jenkins NA, Gilbert DJ, suggest differences in X-chromosomal dynamics during Copeland NG, MacDonald CC. Two distinct forms of the 64 000 Mr protein of the cleavage stimulation factor are expressed in mouse male and after the period when it is inactive. In our attempts to germ cells. Proc Natl Acad Sci U S A 1999; 96:6763±6768. examine these differences between mice and rats, we have 3. MacDonald CC, Redondo J-L. Reexamining the polyadenylation sig- undertaken xenotransplantation experiments to transfer rat nal: were we wrong about AAUAAA? Mol Cell Endocrinol 2002; and mouse germ cells and to examine premeiotic and post- 190:1±8. meiotic expression of CstF-64 (unpublished results). Our 4. Meijer D, Hermans A, von Lindern M, van Agthoven T, de Klein A, preliminary results are consistent with a model in which Mackenbach P, Grootegoed A, Talarico D, Valle GD, Grosveld G. Molecular characterization of the testis speci®c c-abl mRNA in the postmeiotic activation of CstF-64 is a function intrinsic mouse. EMBO J 1987; 6:4041±4048. to the species from which the germ cells are derived and 5. Oppi C, Shore SK, Reddy EP. Nucleotide sequence of testis-derived not dependent on external cues. Other experiments (unpub- c-abl cDNAs: implications for testis-speci®c transcription and abl on- lished results) reveal a surprising amount of evolutionary cogene activation. Proc Natl Acad Sci U S A 1987; 84:8200±8204. divergence of CstF-64 and ␶CstF-64 among closely related 6. éyen O, Myklebust F, Scott JD, Cadd GG, McKnight GS, Hansson (Muridae) and more distantly related (Sciuromorphic and V, Jahnsen T. Subunits of cyclic adenosine 3Ј,5Ј-monophosphate-de- pendent protein kinase show differential and distinct expression pat- Caviomorphic) rodent species. This suggests that CstF-64 terns during germ cell differentiation: alternative polyadenylation in function might have species-speci®c functions that we have germ cells gives rise to unique smaller-sized mRNA species. Biol not yet uncovered. Reprod 1990; 43:46±54. Even more perplexing are the differences in ␶CstF-64 7. Wingett D, Reeves R, Magnuson NS. Characterization of the testes- expression in mouse and rat liver (Figs. 1, B and C and 5). speci®c pim-1 transcript in rat. Nucleic Acids Res 1992; 20:3183± Although many metabolic differences exist between mouse 3189. 8. Foulkes NS, Schlotter F, PeÂvet P, Sassone-Corsi P. Pituitary hormone and rat liver metabolism, it is hard to see any connection FSH directs the CREM functional switch during spermatogenesis. Na- between those metabolic processes and nuclear mRNA pro- ture 1993; 362:264±267. cessing. However, we have seen ␶CstF-64 mRNA expressed 9. Ravnik SE, Wolgemuth DJ. The developmentally restricted pattern of at high levels in many mouse tissues, including liver, in expression in the male germ line of a murine cyclin A, cyclin A2, which protein expression was low or absent (unpublished suggests roles in both mitotic and meiotic cell cycles. Dev Biol 1996; results). This suggests that ␶CstF-64 expression is regulat- 173:69±78. 10. Edwalds-Gilbert G, Veraldi KL, Milcarek C. Alternative poly(A) site ed, at least partially, at the translational level. Perhaps in selection in complex transcription units: means to an end? Nucleic rat liver the translational repression is released, whereas in Acids Res 1997; 25:2547±2561. mouse liver it is not. 11. Dass B, McMahon KW, Jenkins NA, Gilbert DJ, Copeland NG, Mac- We detected intermediate amounts of ␶CstF-64 in rat Donald CC. The gene for a variant form of the polyadenylation pro- spleen and thymus (Fig. 1B, lanes 5 and 7). Although not tein CstF-64 is on chromosome 19 and is expressed in pachytene apparent in Figure 1C, on longer exposures we also have spermatocytes in mice. J Biol Chem 2001; 276:8044±8050. 12. Handel MA, Hunt PA, Kot MC, Park C, Shannon M. Role of sex detected small amounts of ␶CstF-64 in mouse spleen and chromosomes in the control of male germ-cell differentiation. Ann thymus (not shown). This suggests a potential role for N Y Acad Sci 1991; 637:64±73. ␶CstF-64 in cells of the immune system. For instance, 13. Handel MA, Park C, Kot M. Genetic control of sex-chromosome in- changes in the site of polyadenylation in immunoglobulin activation during male meiosis. Cytogenet Cell Genet 1994; 66:83± (Ig) heavy-chain mRNAs during B-cell development 88. 14. Dass B, McDaniel L, Schultz RA, Attaya E, MacDonald CC. The change IgG from a membrane-bound to a secreted form. gene CSTF2T encoding the human variant CstF-64 polyadenylation Increases in the levels of CstF-64 are suf®cient to affect protein ␶CstF-64 is intronless and may be associated with male ste- this change [52, 53] but might not be necessary [40, 54]. rility. Genomics 2002; 80:509±514. Perhaps ␶CstF-64 is involved in this change as well. 15. Lyon MF. Gene action in the X-chromosome of the mouse (Mus mus- These ®ndings raise many questions regarding polyade- culus L.). Nature 1961; 190:372±373. nylation function, X-chromosomal gene expression, and de- 16. Lyon MF. LINE-1 elements and X chromosome inactivation: a func- tion for ``junk'' DNA? Proc Natl Acad Sci U S A 2000; 97:6248± velopmental expression of germ cell-speci®c proteins. To 6249. answer some of these questions, we have undertaken to 17. Gartler SM, Riggs AD. Mammalian X-chromosome inactivation. clone and sequence CstF-64 and ␶CstF-64 from a large Annu Rev Genet 1983; 17:155±190. number of mammalian and nonmammalian species [14; un- 18. Heard E, Clerc P, Avner P. X-chromosome inactivation in mammals. published results]. We are ®nding that whereas CstF-64 is Annu Rev Biochem 1997; 31:571±610. conserved among all species, ␶CstF-64 is far less con- 19. Avner P, Heard E. X-chromosome inactivation: counting, choice and initiation. Nat Rev Genet 2001; 2:59±67. strained. The experiments described in the present study 20. Ayoub N, Richler C, Wahrman J. Xist RNA is associated with the suggest that developmental expression of these two proteins transcriptionally inactive XY body in mammalian male meiosis. Chro- also is highly variable, even between closely related rodent mosoma 1997; 106:1±10. CstF-64 ISOFORMS IN RAT AND MOUSE DEVELOPMENT 1087

21. McCarrey JR, Dilworth DD. Expression of Xist in mouse germ cells 38. Laemmli UK. Cleavage of structural proteins during the assembly of correlates with X-chromosome inactivation. Nat Genet 1992; 2:200± the head of bacteriophage T4. Nature 1970; 227:680±685. 203. 39. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins 22. Sheardown SA, Duthie SM, Johnston CM, Newall AE, Formstone EJ, from polyacrylamide gels to nitrocellulose sheets: procedure and some Arkell RM, Nesterova TB, Alghisi GC, Rastan S, Brockdorff N. Sta- applications. Proc Natl Acad Sci U S A 1979; 76:4350±4354. bilization of Xist RNA mediates initiation of X chromosome inacti- 40. Edwalds-Gilbert G, Milcarek C. Regulation of poly(A) site use during vation. Cell 1997; 91:99±107. mouse B-cell development involves a change in the binding of a gen- 23. Marahrens Y, Panning B, Dausman J, Strauss W, Jaenisch R. Xist- eral polyadenylation factor in a B-cell stage-speci®c manner. Mol Cell de®cient mice are defective in dosage compensation but not sper- Biol 1995; 15:6420±6429. matogenesis. Genes Dev 1997; 11:156±166. 41. Bellve AR. Puri®cation, culture, and fractionation of spermatogenic 24. Zhao J, Hyman L, Moore C. Formation of mRNA 3Ј-ends in eukary- cells. Methods Enzymol 1993; 225:84±113. otes: mechanism, regulation, and interrelationships with other steps in 42. Russell LD, Ettlin RA, Sinha Hikim AP, Clegg ED. Histological and mRNA synthesis. Microbiol Mol Biol Rev 1999; 63:405±445. Histopathological Evaluation of the Testis. Clearwater, FL: Cache Riv- 25. Colgan DF, Manley JL. Mechanism and regulation of mRNA poly- er Press; 1990. adenylation. Genes Dev 1997; 11:2755±2766. 43. McCarrey JR, Dilworth DD, Sharp RM. Semiquantitative analysis of 26. Wahle E. 3Ј-End cleavage and polyadenylation of mRNA precursors. X-linked gene expression during spermatogenesis in the mouse: eth- Downloaded from https://academic.oup.com/biolreprod/article/70/4/1080/2712970 by guest on 30 September 2021 Biochim Biophys Acta 1995; 1261:183±194. idium-bromide staining of RT-PCR products. Genetics Analysis Tech- nology and Applications 1992; 9:117±123. 27. MacDonald CC, Wilusz J, Shenk T. The 64-kilodalton subunit of the 44. Marin I, Siegal ML, Baker BS. The evolution of dosage-compensation CstF polyadenylation factor binds to pre-mRNAs downstream of the mechanisms. Bioessays 2000; 22:1106±1114. cleavage site and in¯uences cleavage site location. Mol Cell Biol 45. Pannuti A, Lucchesi JC. Recycling to remodel: evolution of dosage- 1994; 14:6647±6654. compensation complexes. Curr Opin Genet Dev 2000; 10:644±650. 28. Weiss EA, Gilmartin GM, Nevins JR. Poly(A) site ef®ciency re¯ects 46. Jegalian K, Page DC. A proposed path by which genes common to the stability of complex formation involving the downstream element. mammalian X and Y chromosomes evolve to become X inactivated. EMBO J 1991; 10:215±219. Nature 1998; 394:776±780. 29. Chen F, MacDonald CC, Wilusz J. Cleavage site determinants in the 47. Wang PJ, McCarrey JR, Yang F, Page DC. An abundance of X-linked mammalian polyadenylation signal. Nucleic Acids Res 1995; 23: genes expressed in spermatogonia. Nat Genet 2001; 27:422±426. 2614±2620. 48. Turner RM, Johnson LR, Haig-Ladewig L, Gerton GL, Moss SB. An 30. Murthy KGK, Manley JL. Characterization of the multisubunit cleav- X-linked gene encodes a major human sperm ®brous sheath protein, age-polyadenylation speci®city factor from calf thymus. J Biol Chem hAKAP82. Genomic organization, protein kinase A-RII binding, and 1992; 267:14804±14811. distribution of the precursor in the sperm tail. J Biol Chem 1998; 273: 31. Jenny A, Hauri H-P, Keller W. Characterization of cleavage and poly- 32135±32141. adenylation speci®city factor and cloning of its 100-kilodalton sub- 49. Hendriksen PJ, Hoogerbrugge JW, Themmen AP, Koken MH, Hoe- unit. Mol Cell Biol 1994; 14:8183±8190. ijmakers JH, Oostra BA, van der Lende T, Grootegoed JA. Postmeiotic 32. Murthy KG, Manley JL. The 160-kD subunit of human cleavage- transcription of X and Y chromosomal genes during spermatogenesis polyadenylation speci®city factor coordinates pre-mRNA 3Ј-end for- in the mouse. Dev Biol 1995; 170:730±733. mation. Genes Dev 1995; 9:2672±2683. 50. De Plaen E, De Backer O, Arnaud D, Bonjean B, Chomez P, Marte- 33. RuÈegsegger U, Blank D, Keller W. Human pre-mRNA cleavage factor lange V, Avner P, Baldacci P, Babinet C, Hwang SY, Knowles B, Boon T. A new family of mouse genes homologous to the human MAGE Im is related to spliceosomal SR proteins and can be reconstituted in vitro from recombinant subunits. Mol Cell 1998; 1:243±253. genes. Genomics 1999; 55:176±184. 51. Chang BH, Li WH. Estimating the intensity of male-driven evolution 34. De Vries H, Ruegsegger U, Hubner W, Friedlein A, Langen H, Keller in rodents by using X-linked and Y-linked Ube 1 genes and pseudo- W. Human pre-mRNA cleavage factor II(m) contains homologs of genes. J Mol Evol 1995; 40:70±77. yeast proteins and bridges two other cleavage factors. EMBO J 2000; 52. Takagaki Y, Seipelt RL, Peterson ML, Manley JL. The polyadenyla- 19:5895±5904. tion factor CstF-64 regulates alternative processing of IgM heavy 35. Wahle E. Poly(A) tail length control is caused by termination of pro- chain pre-mRNA during B cell differentiation. Cell 1996; 87:941±952. cessive synthesis. J Biol Chem 1995; 270:2800±2808. 53. Takagaki Y, Manley JL. Levels of polyadenylation factor CstF-64 con- 36. Dass B, Attaya EN, Wallace AM, MacDonald CC. Overexpression of trol IgM heavy chain mRNA accumulation and other events associated the CstF-64 and CPSF-160 polyadenylation protein messenger RNAs with B cell differentiation. Mol Cell 1998; 2:761±771. in mouse male germ cells. Biol Reprod 2001; 64:1722±1729. 54. Martincic K, Campbell R, Edwalds-Gilbert G, Souan L, Lotze MT, 37. Takagaki Y, Manley JL, MacDonald CC, Wilusz J, Shenk T. A mul- Milcarek C. Increase in the 64-kDa subunit of the polyadenylation/ tisubunit factor CstF is required for polyadenylation of mammalian cleavage stimulatory factor during the G0 to S phase transition. Proc pre-mRNAs. Genes Dev 1990; 4:2112±2120. Natl Acad Sci U S A 1998; 95:11095±11100.