A novel 14-base-pair regulatory element is essential for in vivo expression of murine beta4-galactosyltransferase-I in late pachytene spermatocytes and round spermatids

Abstract

During murine spermatogenesis, beginning in late pachytene spermatocytes, the beta4-galactosyltransferase-I (beta4GalT-I) gene is transcribed from a male germ cell-specific start site. We had shown previously that a 796-bp genomic fragment that flanks the germ cell start site and contains two putative CRE (cyclic AMP-responsive element)-like motifs directs correct male germ cell expression of the beta-galactosidase reporter gene in late pachytene spermatocytes and round spermatids of transgenic mice (N. L. Shaper, A. Harduin-Lepers, and J. H. Shaper, J. Biol. Chem. 269:25165-25171, 1994). We now report that in vivo expression of beta4GalT-I in developing male germ cells requires an essential and previously undescribed 14-bp regulatory element (5'-GCCGGTTTCCTAGA-3') that is distinct from the two CRE-like sequences. This cis element is located 16 bp upstream of the germ cell-specific start site and binds a male germ cell protein that we have termed TASS-1 (transcriptional activator in late pachytene spermatocytes and round spermatids 1). The presence of the Ets signature binding motif 5'-GGAA-3' on the bottom strand of the TASS-1 sequence (underlined sequence) suggests that TASS-1 is a novel member of the Ets family of transcription factors. Additional transgenic analyses established that an 87-bp genomic fragment containing the TASS-1 regulatory element was sufficient for correct germ cell-specific expression of the beta-galactosidase reporter gene. Furthermore, when the TASS-1 motif was mutated by transversion, within the context of the original 796-bp fragment, transgene expression was reduced 12- to 35-fold in vivo.

FIG. 1

Schematic representation of the 5′ end of the murine β4GalT-I gene. Shown is the upstream genomic DNA (solid box), the corresponding 5′-untranslated region (727 nt) of the male germ cell-specific transcript (hatched box), the coding sequence of exon 1 (open box), and a portion of the first intron (thin line). The bent arrows denote the three β4GalT-I transcriptional start sites. The male germ cell-specific start site (Gc) is used exclusively in late pachytene spermatocytes and round spermatids. The 4.1-kb start site (4.1) is used exclusively in spermatogonia and predominantly in all somatic cells. The 3.9-kb start site (3.9) is used predominately in the mammary gland during lactation. The locations of the first two in-frame ATGs are shown. Numbering is relative to the first in-frame ATG, which is designated +1. Exon 1 is not drawn to scale. Translation of the 4.1- and 3.9-kb β4GalT-I mRNAs results in two catalytically identical, trans-Golgi resident protein isoforms with NH₂-terminal cytoplasmic domains of 24 and 11 amino acids, respectively.

FIG. 2

Schematic representation of the LacZ-based constructs used to generate the first three transgenic lines. The original β4GalT[−1270/−474]LacZ construct contains a 796-bp fragment that includes 543 bp of the genomic sequence upstream of the male germ cell start site (Gc) and 253 bp of the flanking downstream sequence (38). In the β4GalT[−1085/−474]LacZ construct, the sequence spanning −1270 to −1086 was deleted, while in the β4GalT[−1270/−619]LacZ construct, the sequence spanning −618 to −474 was deleted. The β4GalT[−1085/−628]LacZ construct combines both 5′ and 3′ deletions. The solid, hatched, and open boxes represent the upstream genomic DNA, the 5′-untranslated region of the male germ cell transcript, and the LacZ coding sequence, respectively.

FIG. 3

Expression of the β-galactosidase reporter gene in the testes of transgenic mice. Testes from a nontransgenic control and the indicated transgenic lines were paraformaldehyde fixed, incubated in X-Gal for 48 h, and visualized directly (parts 1 to 4). To establish which male germ cell populations expressed the β-galactosidase reporter gene, the fixed and stained testes were paraffin embedded, sectioned, counterstained with nuclear fast red, and inspected under a microscope. (A) Parts: 1, a testis from a nontransgenic mouse; 2, a testis from a β4GalT[−1085/−474]LacZ-346 mouse; 3, a testis from a β4GalT[−793/−707]LacZ-320 mouse; 4, a testis from a β4GalT[mut−756/−743]LacZ-374 mouse. Note that the blue staining, indicative of the expression of the β-galactosidase reporter gene (LacZ) is concentrated in the seminiferous tubules. (B) Section of a seminiferous tubule obtained from the testis of a β4GalT[−1085/−474]LacZ-346 mouse. The purple arrows indicate the pachytene spermatocytes, the yellow arrows indicate the round spermatids, and the black arrows indicate the elongated spermatids (final magnification, ×625). (C) Section of a seminiferous tubule obtained from the testis of a β4GalT[−793/−707]LacZ-320 mouse (final magnification, ×625). Note that this 87-bp genomic fragment which contains the TASS-1 regulatory element is sufficient to drive male germ cell-specific expression of the reporter gene in a pattern that is comparable to that of the endogenous β4GalT-I gene.

FIG. 4

Identification of a single 23-bp protected region within the 457-bp promoter region by DNase I footprinting analysis. Protected regions on the coding strand (A) and noncoding strand (B) are indicated by the black and gray bars, respectively. A DNA fragment corresponding to the region between −870 to −474 was end labeled with ³²P and digested with DNase I in the presence of BSA (lanes 1) or nuclear proteins from mouse L cells (lanes 2) or testes (lanes 3). (C) Sequence of the region spanning −768 to −637. The black rectangle indicates the region protected by the testis nuclear extract, while the gray rectangle indicates the region protected by both testis and L-cell nuclear extracts. The locations of the two CRE-like motifs are shown. The male germ cell transcription start site (Gc) is located at position −727.

FIG. 5

The two CRE-like sequences within the 457-bp promoter fail to bind a nuclear protein(s) when analyzed by EMSA. Double-stranded oligomers containing the CRE consensus binding site (left panel) or the CRE-like motif spanning −766 to −759 (middle panel) or −645 to −638 (right panel) were end labeled with ³²P and incubated with 10 μg of testis (lanes 3, 6, and 9) or mouse L-cell (lanes 2, 5, and 8) nuclear extract. FP designates the migration position of the free probe.

FIG. 6

Nuclear factor(s) binding to the DNA sequence spanning −756 to −734. (A) The −756/−734 oligomer was end labeled with ³²P using Klenow DNA polymerase and incubated with nuclear extracts from testes (lane 2), pooled male germ cells (lane 3), or mouse L cells (lane 4). Upon a lighter exposure, the band present in lane 4 can be resolved into two bands. (B) Specificity of the complex obtained with the testis nuclear extract (lane 2) was demonstrated by addition of a 100-fold molar excess of the unlabeled −756/−734 oligomer (lane 3) prior to addition of the labeled probe. Incubation with a 100-fold molar excess of a nonspecific oligomer containing the CRE-like motif spanning −766 to −759 (lane 4) or the CRE consensus motif (lane 5) did not prevent complex formation. FP indicates the migration position of the free probe.

FIG. 7

Effect of mutations within the −756 to −743 region on complex formation. (A) Sequences of the wild-type and mutated (mut 1 to mut 8) double-stranded oligomers used for EMSAs. The mutated bases are in boldface. (B) Eight mutated double-stranded oligomers (lanes 2 to 9) labeled with ³²P at their respective 5′ ends and incubated with 10 μg of testis nuclear extract. Complex formation observed with each mutated oligomer was compared with that observed with the −756/−734 oligomer (lane 1). FP designates the migration position of the free probe.

FIG. 8

Schematic representation of the β4GalT[−793/−707]LacZ and β4GalT[mut−756/−743]LacZ constructs. The original β4GalT[−1270/−474]LacZ construct is shown for comparison. In the β4GalT[−793/−707]LacZ construct, an 87-bp fragment containing 67 bp of the genomic sequence upstream from the β4GalT-I male germ cell start site and 20 bp of the flanking downstream sequence was fused to the bacterial LacZ coding sequence. In the β4GalT[mut−756/−743]LacZ construct, the nucleotides within the 14-bp TASS-1 motif (spanning −756 to −743) were mutated by transversion. The positions and sequences of the original and mutated sites are shown. The solid, hatched, and open boxes represent the upstream genomic DNA, the 5′-untranslated region of the male germ cell transcript, and the LacZ coding sequence, respectively.

FIG. 9

Model depicting the binding of TASS-1 and a putative cofactor (Co-F) to the 14-bp regulatory element. (A) High-affinity binding of TASS-1 to the 14-bp motif (open rectangle) is shown by the solid vertical lines. This binding leads to the recruitment of a putative Co-F which interacts with TASS-1 via protein-protein interactions. (B) TASS-1 binds weakly, as shown by the dashed vertical lines, to oligomer mut 1, in which the 5′ end of the binding motif has been mutated (solid box). Co-F binding to DNA-bound TASS-1 is unaffected, but the complex can easily dissociate; thus, the steady-state level of the complex is greatly reduced. (C) TASS-1 binds weakly to oligomers mut 4 and mut 5, in which the 3′ end of the binding motif has been mutated (solid box); however, Co-F cannot bind to the DNA–TASS-1 complex. (D) TASS-1 is unable to bind an oligomer in which the core sequence is mutated (mut 2 or mut 3) or the entire 14-bp motif is mutated (mut 8).