B94, a primary response gene inducible by tumor necrosis factor-alpha, is expressed in developing hematopoietic tissues and the sperm acrosome.

B94 was originally described as a novel tumor necrosis factor-alpha-inducible primary response gene in endothelial cells which was also induced in an in vitro model of angiogenesis. To further characterize its expression, we cloned the mouse homologue and mapped its developmental and tissue specific expression. The predicted amino acid sequence of mouse B94 was found to be 83% similar to its human homologue. The gene was localized to mouse chromosome 12 just centromeric to the immunoglobulin heavy chain locus, in a region that is often rearranged in T-cell neoplasms. To explore the possibility that B94 is expressed during vasculogenesis and other developmental processes, the expression of its transcript was determined during mouse development by in situ hybridization. In 10-day embryos B94 was expressed prominently in the myocardium and in the aortic arch. By the 15th day of gestation, expression was restricted largely to the liver, the bone forming regions of the jaw, the aortic endothelium, and the nasopharynx: a pattern that was maintained until just prior to birth. Postnatally, expression shifted to the red pulp of the spleen and the thymic medulla. B94 expression was extinguished in most adult tissues but was detectable in lymphopoietic tissues including the spleen, tonsil, and lymphatic aggregates in the gut. Consistent with this was the finding that mononuclear progenitor cells in bone marrow and mature peripheral blood monocytes expressed B94. A truncated testis-specific transcript previously identified by Northern blot analysis was determined to result from the use of an alternate polyadenylation signal which was surprisingly located within the open reading frame. This shorter transcript was expressed at high levels exclusively in late stage spermatids. Immunostaining with an affinity-purified polyclonal antiserum revealed B94 to be localized to the acrosomal compartment of mature sperm. These studies demonstrate that B94 expression is tightly regulated during development and suggests distinct roles for B94 in myelopoiesis and spermatogenesis.

pression, we cloned the mouse homologue and mapped its developmental and tissue specific expression. The predicted amino acid sequence of mouse B94 was found to be 83% similar to its human homologue. The gene was localized to mouse chromosome 12 just centromeric to the immunoglobulin heavy chain locus, in a region that is often rearranged in T-cell neoplasms. To explore the possibility that B94 is expressed during vasculogenesis and other developmental processes, the expression of ita transcript was determined during mouse development by in s i t u hybridization. In 10-day embryos B94 was expressed prominently in the myocardium and in the aortic arch. By the 16th day of gestation, expression was restricted largely to the liver, the bone forming regions of the jaw, the aortic endothelium, and the nasopharynx: a pattern that was maintained until just prior to birth. Postnatally, expression shifted to the red pulp of the spleen and the thymic medulla. B94 expression was extinguished in most adult tissues but was detectable in lymphopoietic tissues including the spleen, tonsil, and lymphatic aggregates in the gut. Consistent with this was the finding that mononuclear progenitor cells in bone marrow and mature peripheral blood monocytes expressed B94. A truncated testis-specific transcript previously identified by Northern blot analysis was determined to result from the use of an alternate polyadenylation signal which was surprisingly located within the open reading frame. This shorter transcript was expressed at high levels exclusively in late stage spermatids. Immunostaining with an affinity-purified polyclonal antiserum revealed B94 to be localized to the acrosomal compartment of mature sperm. These studies demonstrate that B94 expression is tightly regulated during development The influence of cytokines upon their target cells is mediated by biochemical and genetic events which lead to characteristic alterations in cellular behavior. Genes transcriptionally activated by cytokines and growth factors without the requirement of intervening protein synthesis are known as primary response genes and encode key regulatory proteins including transcriptional factors such as c-fos and cjun, paracrine factors such as interleukin-8, and cell surface molecules that mediate adhesive interactions such as E-selectin (1). Overall, these alterations in biochemical pathways and cell behavior results in the phenotypic change associated with a particular growth factor or cytokine. Tumor necrosis factor-a (TNF)' is a multifunctional cytokine implicated in diverse processes, including acute inflammation (21, angiogenesis (31, blood cell differentiation (4), bone resorption (5), cell proliferation (61, and cell killing (7). It is likely that the primary response genes that TNF induces mediate a subset of these responses. For example, the TNF" inducible protein A20 has been shown to protect cells from TNF-mediated programmed cell death (8).
B94 is a cytokine-driven primary response gene that was originally cloned from TNF stimulated endothelium (9). The B94 gene encodes a novel 73-kDa intracellular protein that exists as a single copy gene on human chromosome 14 near the immunoglobulin heavy chain locus. It is also known that B94 is activated by factors other than TNF; endothelial cells stimulated with interleukin-lp or lipopolysaccharide-induced B94 and Northern blot analysis of murine embryonic tissues revealed it to be expressed in embryonic liver and kidney, organs in which TNF is not present during development (16). Interestingly, a second smaller transcript was highly expressed in the testis.
In the present studies, the expression of B94 was examined in greater detail. The predicted mouse B94 amino acid sequence was compared to its human homologue, and its expression during murine embryonic development was mapped by in situ hybridization. In addition, the expression of B94 in hematopoiesis and male germ cell maturation was examined in detail. The results suggested that B94 may play multiple roles in development, including vasculogenesis, blood cell differentiation, and spermatogenesis. dayW post-coitum; ZGH, immunoglobulin heavy chain locus; Aut-1, a-l-The abbreviations used are: TNF, tumor necrosis factor-a; dpc, antitrypsin, kb, kilobases; PBS, phosphate-buffered saline; PDGF, platelet-derived growth factor; FGF, fibroblast growth factor.

MATERIALS AND METHODS
Cloning and Sequence Analysis-A random-primed cDNA library was constructed in the pCDNAl vector (Invitrogen) from mouse embryonic kidney poly(A)+ RNA. 1 x lo5 colonies were screened with a 32P random-labeled PstI restriction fragment of the human B94 cDNA open reading frame encompassing nucleotides 103C1660 (9) using standard techniques (10). The testis form of mouse B94 was cloned by reverse transcriptase-polymerase chain reaction using adult mouse testis RNA as template. The 3' primer, 5"GTGTGACTCGAGTCGACCAGCTG- contained an oligo(dT) stretch for hybridization to poly(A) sequences and upstream XhoI, SalI, and HincII restriction endonuclease sites for subcloning. The 5' primer, 5"TGGCl"CGA-CACCCTGCTCC-3' was specific for mouse B94 sequences just upstream of an EcoRI site in the open reading frame. Full-length cDNA clones for the testis form of B94 were isolated from an adult testis cDNA library in the pCDNAl vector. Plasmids containing mouse B94 cDNA sequences were purified by CsCl ultracentrifugation and sequenced on both strands with Sequenase (United States Biochemical Corp.) and synthetic deoxyoligonucleotides as primers. Sequence was assembled and analyzed using MacVector release 4.0 (IBI) and the Genetics Computer Group Sequence Analysis SoRware package (Version 7.0).
Mouse Chromosomal Localization-C3WHeJ-gld/gld and Mus spretus (Spain) mice and (C3HEIeJ-gldlgld X Mus spretus)F, X C3WHeJgldlgld interspecific backcross mice were bred and maintained as described previously (11). Genomic DNA isolated from mouse organs was digested with a panel of restriction endonucleases and 10 pg of each digest was resolved in 0.9% agarose gels, transferred to nylon membranes, hybridized to 32P-radiolabeled DNA probes under high stringency conditions, and washed at high stringency as described previously (11). A 2.4-kb XhoI fragment of mouse B94, representing most of the open reading frame and all the 3'-untranslated region, was used as a probe. Gene linkage was determined by segregation analysis with other published markers (11).
Tissue Preparation and in Situ Hybridization-CD1 male mice were mated with virgin females, and embryonic day 1 was established by the presence of a vaginal plug. Embryos and postnatal organs snap-frozen in OCT were cut at 8 pm and collected on acid-washed silane-treated slides. Sections were stored at -80 "C prior to use. Bone marrow cells isolated from an adult murine tibia were washed two times in RPMI media with 1% bovine serum albumin, and 100,000 cells were attached to acid washed slides by cytocentrifugation. The cells were briefly airdried and then fixed in fresh 2% paraformaldehyde, 0.1% glutaraldehyde (EM Sciences) in phosphate-buffered saline, pH 7.4 (PBS). Fresh human peripheral blood leukocytes were isolated by centrifugation through Ficoll-Paque (Pharmacia LKB Biotechnology Inc.) for 30 min at 400 x g and were separated by elutriation into fractions enriched in lymphocytes, monocytes, and polymorphonuclear cells. Elutriation was done in elutriation buffer (PBS with 0.5% dextrose, 0.35% bovine serum albumin, and 1 m~ EDTA) with a JE-GB rotor (Beckman) at 2030 rpm and a flow rate varying upwards from 4 to 10 d m i n . Fractions were monitored for specific cell types by morphology and through nonspecific esterase staining for monocytes. The cells were then cytocentrifuged and fixed in the same manner as bone marrow cells.
In situ hybridization was camed out as described previously with modifications (12). Sections were fxed in fresh 4% paraformaldehyde in PBS, rinsed in PBS, digested in 10 mdml prpteinase K for 5 min, and refured in 4% paraformaldehyde in PBS. Sections were placed in 0.25% acetic anhydride containing 0.1 M triethanolamine for 10 min, rinsed in PBS then 0.15 M NaCl for 5 min each, and dehydrated through an ethanol series. An XbaI-EcoRI fragment of mouse B94 encompassing nucleotides 1318-1607 was subcloned into the plasmid pGEM7zf (Promega) and subsequently linearized with either EcoRI or XbaI. A human B94 template encompassing nucleotides 1347-1656 was generated by polymerase chain reaction using the oligonucleotides 5'-gtaatacgactcactatag AATGAA'MTCTGGAGAGAGG-3' (T7 promoter in lower case) and 5'ggatttaggtgacactata GCTCTGAGAACTCAGGCAGC-3' (SP6 promoter in lower case). Single-stranded RNA transcripts uniformly labeled with [35S]UTP (Du Pont NEN) were transcribed with either SP6 or T7 RNA polymerase according to the manufacturer's instructions (Promega). Probes purified by ethanol precipitation were used for hybridization the same day.
The slides were dipped in NTB-2 emulsion (IBI) diluted 1:l with 2% glycerol, and exposed for 2-4 weeks at 4 "C. The emulsion was developed for 2.5 min in Dl9 developer at 18 "C, then fixed, and counterstained with hematoxylin and eosin. Leukocytes were counterstained with giemsa (Sigma) diluted 1:20 in water for 6 h and washed in 2% acetic acid twice for 1.5 min and 100% ethanol twice for 45 s. The slides were photographed on a Wild M420 darkfield microscope.
Cell Culture and RNA Analysis-Swiss-3T3 cells were cultured and treated as described previously (13). Neonatal Schwann cells were prepared according to Manthorpe and Varon (14) and passaged in DMEM containing 10% fetal bovine serum. The myelomonocytic cell lines HL-60 was obtained from the American Type Culture Collection (Rockville, MD). HL-60 cells were routinely cultured in RPMI-1640 supplemented with 10% horse serum. Before treatment cells were washed twice in PBS. Phorbol 12-myristate 13-acetate (16 n~) and dimethyl sulfoxide (1.25%) were added to the cells, and RNAwas extracted at the times indicated. Northern blot analysis was done as described previously (9).
Immunocytochemistry-Testes from a young adult macaque were processed in the same manner as the mouse embryo sections. Adult human sperm from ejaculate was washed by pelleting the cells three times through PBS, freezing in OCT in hexane, and sectioning to 8-pm thickness. The sections were fixed in 1% paraformaldehyde in PBS for 2 min, washed with PBS, and blocked for 20 min in normal goat serum (Vector Labs) diluted 1:20 with PBS. Sections were then incubated for 1 h with affinity-purified rabbit polyclonal antisera raised against the carboxyl-terminal 137 amino acids of the human B94 protein (9). Serial sections were incubated with blocked antisera to control for antibody specificity. Sections were then washed three times with PBS and incubated for 30 min with fluorescein-conjugated goat anti-rabbit antibody (Sigma) diluted 1:20 in PBS. Sections were rinsed five times with PBS and then examined and photographed on a Leitz Orthoplan fluorescence microscope.

RESULTS
Isolation a n d Analysis of Mouse B94 cDNAs-Mouse B94 has been shown previously by Northern analysis to be expressed in a variety of tissues including the embryonic kidney (9). Screening of day 17 post-coitum (dpc) embryonic kidney cDNA library with a human cDNA probe resulted in the isolation of eight overlapping clones, the longest two of which were characterized further. The composite cDNA sequence2 contained one long open reading frame encoding a 650-amino acid polypeptide and a long nonconserved 3"untranslated sequence terminating in a consensus polyadenylation signal and poly(A) tail (data not shown). Comparison of the predicted amino acid sequence with its human homologue showed that the sequences were 73% identical and 83% similar (Fig. 1). Interestingly, cysteine residues which are usually invariant were relatively unconserved.
Chromosomal Localization-Genomic DNA samples prepared from mice which were the progeny of an interspecific backcross was utilized to determine the chromosomal localization of the mouse B94 gene. Initially, DNA from each of the parental mice was digested with various restriction enzymes and hybridized with labeled cDNA to identify a polymorphic restriction site within the B94 locus. A polymorphism detected with the enzyme TaqI (Fig. 2 A ) was subsequently used to monitor the segregation of B94 with reference to known positional markers in a haplotype analysis of the backcross mice.  Developmental Expression Patterns-Expression patterns of B94 during murine development were determined by in situ hybridization of 35S-labeled antisense RNA to sagittal sections of embryos aged 10-19 dpc. Serial sections were routinely probed with 35S-labeled sense RNA as a control for nonspecific hybridization.
B94 was expressed in a spatially and temporally restricted pattern that was for the most part consistent throughout the gestational period. At 10 dpc, B94 was present in the forming myocardium and in the aortic endothelium (Fig. 3A 1. Persistent myocardial expression of B94 was evident throughout embryonic development albeit at reduced levels. The aortic endothelium demonstrated marked expression through 19 dpc.
B94 was also detectable at 10 dpc in somites, which form muscle and bone. At later stages various ossifying tissues expressed B94. Bone producing osteoblasts in forming vertebrae and the perichondrium expressed low levels of B94 at day 15 but much higher levels by the 19th day. Chondrocytes appeared not to synthesize B94. Odontoblasts and the forming pulp cavity of the mandible as well as the superficial layer of the olfactory epithelium and associated bone forming regions also expressed considerable B94. Other tissues positive for B94 at 10 dpc include the allantois and the pia mater.
The liver offered the most striking expression pattern of B94 during prenatal development. B94 became detectable in the liver 13 dpc, peaked by day 15, and gradually decreased through day 19. B94 was undetectable in the adult liver (data not shown). Expression was apparently uniform throughout the organ at all stages. In addition to the rapid proliferation of hepatocytes, the liver is the major site of hematopoiesis during  15 (C, D), 17 (E, F ) , and 19 (G, H ) days hybridized with either antisense ( A ,   C, E, G ) or sense (B, 0, E H )  this developmental period: blood cell production moves from the yolk sac to the liver around gestational day 13 and gradually to the spleen at day 17 (15).
Other tissues positive for B94 expression during embryonic development included the epithelium of the trachea and the oropharynx 17-19 dpc, the hypophysis 15-19 dpc, the submandibular gland 17-19 dpc, and discrete cells in the thymus 19 dpc. Finally, the developing kidney expressed B94 mainly in developing glomeruli, although diffuse signal was present throughout the organ. By adulthood renal expression of B94 was extinguished.
Since the role of TNF in embryonic development is primarily limited to the thymus (161, other factors were likely to be responsible for promoting B94 expression in embryogenesis. To determine if growth factors were capable of inducing B94, quiescent fibroblasts were stimulated with bovine calf serum or purified growth factors and induction of B94 transcript was monitored by Northern analysis. B94 was inducible by serum and to a lesser extent by platelet-derived growth factor and fibroblast growth factor (Fig. 4). Maximal induction occurred 4 h after treatment, was not inhibitible by cycloheximide, and gradually decreased through 48 h.
B94 Expression in Postnatal Murine Tissues-Postnatal patterns of B94 expression were also explored by in situ hybridization. Although many tissues expressed little or no B94 postnatally, including the heart, kidney, liver, and lung (data not shown), tissues involved in hematopoiesis and lymphoid development displayed B94-specific hybridization (Fig. 5). In postnatal day 8 spleen, B94 was highly expressed throughout the red pulp and in the white pulp marginal zone but not in the white pulp. By adulthood, expression was limited to cells in the red pulp surrounding the white pulp and to groups of cells just inside the splenic capsule. B94 was also present in the thymic medulla a t postnatal day 15 and adulthood and in the luminal region of lymphoid aggregates in the adult small intestine, an area rich in phagocytic antigen presenting cells and T-cells. Additionally, high levels of B94 were detected in human adolescent tonsil epithelium. A small population of cells in the tonsillar germinal centers also produced B94.
B94 Expression in Hematopoietic Cells-In order to determine if B94 was expressed by blood cells or their precursors, in situ hybridization was carried out on mouse bone marrow hematopoietic precursor cells and on mature human peripheral blood leukocytes. B94 expression was evident in large mono- nuclear cells in the bone marrow, likely to belong to the myelomonocytic lineage, but not in smaller mononuclear cells of the lymphoid lineage nor in more differentiated granulocytic or erythroid cells (Fig. 6, A and B). Consistent with this result mature peripheral blood monocytes but neither lymphocytes demonstrated that B94 was present at low levels in the basal state (data not shown). In the pluripotential cell line HL60, increased expression of B94 was observed when treated with phorboi 12-myristate 13-acetate, which promotes differentiation down the myelomonocytic pathway (Fig. 6G ) To determine how this transcript differed from the originally described and cloned 4.2-kb species, short 32P-labeled fragments of mouse cDNA were used to probe mouse testis Northern blots at high stringency. Probes including sequence from the 5"untranslated region and the coding region hybridized with both RNA species, whereas probes from the 3"untranslated region did not hybridize with the smaller testis-specific transcript (data not shown). Reverse transcriptase polymerase chain reaction of testis RNA with an oligo(dT) adapter and a B94 coding region-specific primer demonstrated that the 2.5-kb transcript was polyadenylated just five nucleotides downstream of the open reading frame termination codon (Fig. 7). Surprisingly, the consensus polyadenylation signal se-

6, C-F).
quence was found to be present 13 nucleotides upstream of the stop codon, within the open reading frame. Comparison of human and mouse nucleotide sequences in this region showed that the polyadenylation signal and the sequence surrounding the polyadenylation site were evolutionarily conserved. Fulllength cDNA clones of the 2.5-kb species isolated from oligo(dT)-primed testis cDNA library were found by sequence analysis to be identical other than in the placement of the poly(A) tail (data not shown). The 3"untranslated region has been implicated in affecting transcript stability as well as translational efficiency (17). Since 3"untranslated region sequences were not present in the 2.5-kb form of B94, it was possible that one of these two processes was affected. To explore transcript stability it was necessary to find a cell line that expressed both forms of B94. Of many cell lines examined, primary embryonic mouse Schwann cells were the only one found to express both species (Fig. 7B 1. Initial actinomycin D chase experiments did not reveal a significant difference in stability of the two transcripts (data not shown).
B94 Expression in Reproductive Tissues-In situ hybridization was used to localize B94 expression within the testis. B94 transcript was detectable in a subset of seminiferous tubules over late stage rounded spermatids (Fig. 8A 1. This localization was supported by hybridization of B94 cDNA to a Northern blot of RNA derived from morphologically staged rat seminiferous tubules where B94 was found to be highly expressed in stages 7 and 8, to a lesser extent in stages 9-12, and was undetectable in stages 1-6 and stage 13 (data not shown). Only the 2.5-kb species of B94 transcript was evident in the isolated seminiferous tubule Northern blot, confirming that the smaller transcript was the one present in rounded spermatids.
To determine whether B94 was also present in female reproductive organs, in situ hybridization was carried out on sections of an adult mouse ovary. Although no B94 was detectable in the ovary, the fallopian epithelium expressed high levels of transcript (Fig. 8, B and C).
An affinity-purified polyclonal antiserum directed against human B94 was used to localize B94 protein in sections of monkey testis. The acrosomal compartment in the head of mature spermatids demonstrated specific reactivity to this antiserum (Fig. 9A). Acrosomal staining was not apparent in sections incubated with blocked antiserum (data not shown). To determine whether sperm competent for fertilization also contained B94 protein, frozen sections of sperm from human ejaculate were reacted with the B94 polyclonal antisera. Staining for B94 was again seen specifically in the acrosomal compartment (Fig 9, B and C ) . Thus the transcript for B94 was expressed at the sperm developmental stage immediately preceding that in which the protein was found.

DISCUSSION
In this paper we have described in detail the expression patterns of B94, a cytokine-and growth factor-inducible primary response gene. In order to carry out these studies, it was first necessary to clone and characterize the mouse B94 homologue. Although the derived amino acid sequence showed a fairly high degree of identity to its human counterpart, some potentially significant structural features of the mouse sequence were not conserved. First, the human B94 protein is predicted to contain 7 cysteine residues, whereas the mouse contains eight. Furthermore, only 5 cysteines are conserved between the two molecules. The nonconserved cysteine residues cluster at the NH2-and COOH-terminal regions of the proteins. This lack of conservation suggests that either disulfide bridging is not a determinant in B94 secondary structure or that B94 is a rapidly evolving protein. Second, a stretch of charged residues present in the NH2 terminus of the mouse predicted protein is reduced in charge density when compared with the human sequence. That this mouse sequence encodes the homologue of human B94 is supported by low stringency Southern blot analysis demonstrating the existence of a single gene in the mouse and human genome (9) and by chromosomal localization of mouse B94 to a region of chromosome 12 syntenic with human chromosome 14 where the human B94 gene is resident.
Expression of TNF during normal development is largely restricted to the thymus (18,16). When pregnant mice and their progeny are repeatedly injected with anti-TNF antisera, the thymus, spleen, and lymph nodes atrophy, suggesting a role for TNF in immune system development (19). Since TNF is expressed by T-cells in the thymus and B94 expression was apparently widespread during development, TNF may not be the only stimulus necessary for B94 expression. In fact, TNF may not control B94 expression during murine development.
Both TNF and its receptors are now recognized as members of expanding gene families. The TNF cytokine family includes lymphotoxin-a and lymphotoxin-P which are expressed almost exclusively on the surface of lymphocytes (20), and the ligands for the lymphocyte receptors CD27 and CD40 (21,221. Proteins with similarity to the p55 and p75 TNF receptors include the B cell antigen CD40 (23), the lymphocyte antigen APOl/Fas (24), and the widely expressed nerve growth factor receptor (25). Thus, B94 may be regulated by one or more of these related molecules.
That B94 expression can be modulated by factors other than TNF was demonstrated through its induction by mitogenic stimuli, including serum, PDGF, and FGF. Developmentally, PDGF is produced by a majority of cells of epithelial origin, and PDGF receptor is found on most mesenchymal cells (26). Similarly, acidic and basic FGF immunoreactivity is detectable in tissues of mesodermal and neuroectodermal origin, and basic FGF is a mesodermal inducer (27). Mesenchymal expression of B94 (vertebral osteoblasts, nasal conchae, mandible, and the forming vasculature) may thus be driven in part by these growth factors.
B94 expression patterns apparently followed the course of hematopoiesis through the developmental stages studied. Hematopoiesis shifts location a number of times during embryogenesis, originating in the yolk sac 8-10 dpc, shifting to the liver 12 dpc where hematopoiesis continues at reduced levels for a few weeks after birth, then to the spleen 16 dpc, which stays active throughout adulthood, and tinally to the bone marrow 19 dpc (15). Expression of B94 temporally and spatially mirrored the movement of hematopoiesis during development most prominently in the liver, but also in the allantois and the spleen and to a lesser extent in the bone marrow.
We have shown that B94 is expressed in myelomonocytic cells in the bone marrow, in peripheral blood monocytes, and in a cultured myelomonocytic precursor cell line. Furthermore, this cell line when induced to differentiate toward monocytes but not toward granulocytes exhibit increased expression of B94. This evidence suggests that B94 expression in bone marrow-derived and peripheral blood cells is largely restricted to cells of the myelomonocytic lineage. This finding also suggests that macrophages or other antigen presenting cells are likely the source of B94 expression in the thymic medulla, the adult spleen, the tonsilar germinal centers, and the gut-associated lymphatic aggregate. The hybridization patterns in these tissues correlates with that of antigen presenting cells. For example, in the gut-associated lymphatic tissues, macrophages and specialized epithelial M cells are resident in the luminal region where they sample antigens from the gut and present them to underlying lymphoid cells and in the spleen macrophages populate the red pulp and the white pulp marginal zone (28). Although B94 expression patterns in lymphoid tissues appear to recapitulate the distribution of antigen presenting cells, it was impossible at the resolution of the current analysis to confirm the identity of the B94 expressing cell populations.
Thus, other cell types in these organs may express B94.
Many genes have different size transcripts in the testis (29,301, suggesting that the transcriptional machinery may be specialized for the testicular environment. The near complete absence of 3'-untranslated sequence in spermatid B94 may iniluence protein expression levels in a number of ways, including transcript stability and translational competence. Given the high level of B94 protein in mature sperm acrosomes, it is likely that the truncated form of B94 in some way increases the ability of the cell to synthesize B94 protein. By in situ hybridization we have determined that B94 transcript in seminiferous tubules is present solely in late stage rounded spermatids. Since this stage follows the second meiotic division, it is likely that B94 is transcribed when the gamete is in the haploid state. Round spermatids are known to be transcriptionally competent as shown in heterozygous transgenic mice in which mutant and wild type alleles are expressed in distinct spermatids (31). Our finding that the 2.5-kb B94 transcript species is co-expressed with the 4-kb species in primary cultures of mouse embryonic Schwann cells provides a model for studying the posttranscrip tional regulation of this gene. The acrosome is derived from the Golgi compartment and is thought to be the sperm equivalent of the lysosome or an exocytic granule (28). Proteolytic enzymes involved in dismantling the egg zona pellucida and penetrating the egg membrane are stored in the acrosome as well as other proteins thought to be involved in fertilization (32, 33). The acrosome contents are released into the zona pellucida by the zona-induced acrosome reaction which is mediated by interaction between the sperm plasma membrane and ZP3, a protein component of the zona pellucida (34). This is followed by fusion of the sperm plasma membrane with the outer acrosomal membrane and punctate formation of pores (35). The pores then enlarge first neutralizing the acidic pH of the acrosome and allowing an influx of extracellular calcium, then later permitting macromolecular diffusion between the acrosomal compartment and the extracellular space. Eventually the pores join and the membrane dissipates as free vesicles, leaving the inner acrosomal membrane as the sperm plasma membrane. It is thus apparent that the suborganellar localization of acrosomal components can assist in predicting function, such as association with the inner acrosomal membrane suggests a role in s p e d e g g membrane fbsion. It will therefore be important to determine by immunoelectron microscopy the suborganellar localization of B94 in the acrosome. Preliminary results from transiently transfected 293T cells indicate that the B94 protein is associated with the membrane f r a~t i o n .~ Localization of the B94 gene (Tnfb94) to mouse chromosome 12 just proximal to the immunoglobulin heavy chain locus is in agreement with and further refines the previous localization of the human gene to human 14q32 (9). The distal telomeric region of mouse chromosome 12 is syntenic to the most telomeric band of human chromosome 14, and the placement of the B94 gene in this region on both chromosomes fiuther confirms their synteny. Additionally the mouse localization shows the B94 gene to be 3.5 centimorgans proximal to the ZGH locus, placing the gene in genomic context near to the akt-1 serine-threonine kinase oncogene (361, the putative ets-like gene Elk-2 (371, and the putative heat shock protein-90-like gene HSPCAIA (38).
The syntenic region of human chromosome 14 is often rearranged in T-cell neoplasms, and the gene or genes affected have not yet been defined, suggesting a potential role for B94 in malignancy.
It is now evident that cytokines and the primary response genes they induce are utilized not only in immune regulation but also in specification of embryonic and postnatal patterning. Here we have provided the first description of the expression patterns of a novel TNF-inducible primary response gene in mouse development. Important roles for B94 in development and in adulthood are suggested by the high level of expression in diverse biological processes, including bone formation, blood cell differentiation, and spermatogenesis. With the advent of embryonal stem cell techniques, which allow for the production of homozygous mice deficient in the gene of interest, it should be possible to directly test the role of B94 in development.