CTCF, a conserved, ubiquitous, and highly versatile 11-zinc-finger factor involved in various aspects of gene regulation, forms methylation-sensitive insulators that regulate X chromosome inactivation and expression of imprinted genes. We document here the existence of a paralogous gene with the same exons encoding the 11-zincfinger domain as mammalian CTCF genes and thus the same DNAbinding potential, but with distinct amino and carboxy termini. We named this gene BORIS for Brother of the Regulator of Imprinted Sites. BORIS is present only in the testis, and expressed in a mutually exclusive manner with CTCF during male germ cell development. We show here that erasure of methylation marks during male germ-line development is associated with dramatic up-regulation of BORIS and down-regulation of CTCF expression. Because BORIS bears the same DNA-binding domain that CTCF employs for recognition of methylation marks in soma, BORIS is a candidate protein for the elusive epigenetic reprogramming factor acting in the male germ line.

H1t is a novel linker histone variant synthesized in mid- to late pachytene spermatocytes. Its regulatory region is of interest because developmentally specific expression has been impressed on an otherwise ubiquitously expressed promoter. Using competitive band-shift assays and specific antisera, we have now shown that the H1t-60 CCTAGG palindrome motif region binds members of the RFX family of transcriptional regulators. The testis-specific binding complex contains RFX2, probably as a homodimer. Other DNA-protein complexes obtained from testis as well as somatic organs contain RFX1, primarily as a heterodimer. Western blots confirmed that RFX2 expression is greatly enhanced in adult testis and that RFX2 is equally prominent in highly enriched populations of late pachytene spermatocytes and round spermatids. Immunohistochemistry carried out on mouse testis showed that RFX2 is strongly expressed in pachytene spermatocytes, remains high in early round spermatids, and declines only in advance of nuclear condensation. Maximum expression correlates well with the appearance of H1t. In contrast, RFX1 immunoreactivity in germ cells was only detected in late round spermatids. RFX-specific band complexes were also identified for both the mouse lamin C2 and Sgy promoters, using either testis nuclear extracts or in vitro-synthesized RFX2. These results call attention to RFX2 as a transcription factor with obvious potential for the regulation of gene expression during meiosis and the early development of spermatids.

H1t is an H1 histone variant unique to late spermatocytes and early spermatids. Using gene targeting and embryonic stem cell technologies, we have produced mice with a disrupted H1t gene. Homozygous H1t-null mice have normal fertility and show no obvious phenotypic consequence due to the lack of this histone. Biochemical and immunohistochemical approaches were used to show that normal changes in chromosomal proteins occurred during spermatid development, including the appearance and disappearance of transition proteins 1 and 2. Both protamines 1 and 2 are present in normal amounts in sonication-resistant spermatid nuclei from H1t-null mice. Analysis of H1 histones by quantitative gel electrophoresis in enriched populations of pachytene spermatocytes and round spermatids showed that the lack of H1t is only partially compensated for by somatic H1s, so that the chromatin of these cells is H1 deficient. Because H1t is thought to create a less tightly compacted chromatin environment, it may be that H1-deficient chromatin is functionally similar to chromatin with H1t present, at least with respect to permitting spermatogenesis to proceed.

H1t is a testis-specific histone 1 variant restricted to the male germ line and expressed only in pachytene spermatocytes. Understanding the regulation of the H1t gene is an interesting challenge as its promoter shares all of the recognized control elements of standard somatic H1 genes, yet H1t is not expressed in somatic or in early spermatogenic cells. To investigate the mechanism of this apparent repression, we exchanged three promoter subregions between H1t and a major somatic H1 gene (H1d) by introduction of suitable restriction sites just 5 of the TATA box and 3 of the conserved H1 AC box. Hybrid promoters were joined to a lacZ reporter gene and assayed by transient transfection in NIH3T3 fibroblasts. In this system the wild type H1d promoter was 20-fold stronger than the H1t promoter. Much of this difference in activity was traced to inhibitory sequences immediately downstream of the TATA box in H1t, although sequences upstream of the H1t AC box and within the H1t 5-untranslated region played some role as well. A series of deletions and short oligonucleotide mutations scanned across the region between the TATA box and cap site identified two tracts of C (GC box 2) as the inhibitory sequences. While both Sp1 and Sp3 bind to this region weakly in vitro, they are unlikely to be responsible for the inhibitory effect of GC box 2, and additional binding proteins (CTB-4 and CTB-5) were identified by electrophoretic mobility shift assays as better candidates for mediating the repressive effect. When repression of the H1t promoter was relieved by mutation of GC box 2, additional mutations introduced into GC box 1 upstream of the CAAT box led to a large decrease in activity, indicating that these two G/C-rich elements have opposite effects on promoter activity.Spermatogenesis is the only example of cellular development in mammals that involves expression of tissue-specific histone variants (1-3). H1t 1 is a testis-specific linker histone variant, appearing late in the prophase of meiosis I in pachytene spermatocytes, retained in early haploid cells, and lost from the nucleus prior to release of mature sperm (3-5). While the amino acid sequence of H1t has a number of novel features (6, 7), it is clearly related to the standard somatic H1 histone family (8). Isolation of the H1t gene from several mammals (9 -12) revealed that the promoter region also shows a surprising similarity to those of standard somatic H1 variants (13,14). Homologies include a TATA box, a CAAT box, a GC box, and an H1-specific AC motif within 100 nucleotides of the cap site as well as an inverted AC motif located farther upstream (15). Despite these shared regulatory elements, H1t expression differs almost completely from that of the common somatic H1 variants. Common variants are produced during S phase of the cell cycle to accommodate the duplication of the chromosomes (14), whereas H1t is expressed only in pachytene spermatocytes, well after completion of replicative DNA synthesis (4,5,16,17). As the presence of H1t mRNA correlates with H1t ...

Histone H1t is synthesized only in male germ cells during the late pachytene stage of meiosis and is retained in spermatids until the nucleus elongates. Transgenic experiments suggest that spermatocyte-directing sequences lie within 140 base pairs of the cap site. To study the mechanism of this specificity we compared the DNase I footprints made on the immediate promoter regions of H1t and H1d (a typical somatic H1) by testis and liver extracts and observed both common and differentially protected regions. The common footprints of H1t included an Sp1 consensus (GC box 1) and a CCAAT motif. Electrophoretic mobility shift assays (EMSA) identified ubiquitous binding factors for GC box 1 and a binding factor for the CCAAT element that we identified immunologically as H1TF2. H1t-specific footprints occurred over the palindrome CCTAGG and a GC-rich sequence downstream of the TATA box (GC box 2). EMSA analysis of the palindrome identified testis-specific as well as ubiquitous binding factors. UV irradiation of a palindrome-binding reaction generated a cross-linked doublet of about 50 kDa from both testis and liver. Protein factors that bound to the GC box 2 sequence were similar from testis and liver, and GC box 1 and an Sp1 consensus competed for them. In vitro transcription directed by H1t occurred at comparable levels in testis and liver extracts. The importance of both GC box 1 and CCAAT elements was demonstrated by deletion analysis and by oligonucleotide competition. No dependence on the H1t palindrome was observed for in vitro transcription.