Requirement of Tissue-Selective TBP-Associated Factor TAFII105 in Ovarian Development

See allHide authors and affiliations

Science  14 Sep 2001:
Vol. 293, Issue 5537, pp. 2084-2087
DOI: 10.1126/science.1061935


Transcription factor TFIID, composed of TBP and TAFIIsubunits, is a central component of the RNA polymerase II machinery. Here, we report that the tissue-selective TAFII105 subunit of TFIID is essential for proper development and function of the mouse ovary. Female mice lacking TAFII105 are viable but infertile because of a defect in folliculogenesis correlating with restricted expression of TAFII105 in the granulosa cells of the ovarian follicle. Gene expression profiling has uncovered a defective inhibin-activin signaling pathway in TAFII105-deficient ovaries. Together, these studies suggest that TAFII105 mediates the transcription of a subset of genes required for proper folliculogenesis in the ovary and establishes TAFII105 as a cell type–specific component of the mammalian transcriptional machinery.

The control of transcription in a spatial and temporal fashion is essential for the proper development and differentiation of specialized cell types that define multicellular organisms. The intricate regulation of transcription relies on the coordinate assembly of large multiprotein complexes at promoter and enhancer regions of specific genes. The transcription factor TFIID is responsible for core promoter recognition and for directing RNA polymerase II to select genes in response to upstream activators (1, 2). Although TFIID was initially thought to be ubiquitous in expression and universal in function, the discovery of putative tissue-restricted components of TFIID prompted us to reevaluate the gene-specific function of this transcription complex. The first cell type–specific subunit of TFIID, called TAFII105, was identified as a polypeptide that coprecipitated with TATA-binding protein (TBP) and the other TAFIIs from a highly differentiated human B cell line but not other cell lines (3). The primary amino acid sequence of TAFII105 revealed that it is related to the more broadly expressed human TAFII130 (4, 5) and its Drosophila homolog dTAFII110 (6). The recent identification of yeast TAFII48 has revealed a weak similarity to the COOH-terminal third of TAFII105; however, the NH2-terminal coactivator domain is absent in TAFII48, suggesting that the metazoan subunits function in regulating programs of gene expression specific to multicellular organisms (7). Indeed, the circumstances of TAFII105's discovery suggest that it might be involved in regulating B cell–specific gene expression. Recent experiments demonstrating that human TAFII105 can interact with nuclear factor kappa B and OCA-B (also called OBF-1) support this hypothesis (8,9). To examine further whether TAFII105 indeed functions in a tissue- and gene-specific manner in mammals, we set out to characterize the biological role of TAFII105 in the mouse.

To determine the tissue-selective nature of TAFII105 expression in mice, we measured the relative levels of TAFII105 and TAFII130 mRNA in selected tissues. Ribonuclease (RNase) protection assays (Fig. 1A) revealed high levels of TAFII130 in all tissues examined, except for the liver. In contrast, TAFII105 transcripts were expressed most highly in the testes and ovary, whereas lower levels were detected in most other tissues. On the basis of restricted expression of TAFII105 mRNA largely in the gonads, we turned to identifying tissue-selective functions of TAFII105.

Figure 1

TAFII105 is differentially expressed in mouse tissues. (A) TAFII105 is highly expressed in the testes and ovary. RNase protections were carried out as described (25). Protected products specific for TAFII105, TAFII130, and an 18SrRNA control are indicated on the right. Mouse tissues from which total RNAs were derived are shown. A yeast total RNA sample was included in each experiment as a negative control. (B) Mouse TAFII105 specifically coprecipitates with TFIID in extracts derived from mouse ovaries. Immunoprecipitation (IP) of ovarian extracts with a nonspecific antibody (lane 1) or anti-TBP (lane 2) followed by Western analysis with anti-TAFII250, anti-TAFII105, and anti-TBP is shown. Molecular mass markers are indicated on the left; mouse TAFII250-, TAFII105-, and TBP-specific bands are indicated on the right.

First, we confirmed that TAFII105 is a component of mouse TFIID in ovaries (Fig. 1B). Total cell extracts prepared from mouse ovaries were precipitated with monoclonal antibodies to TBP (anti-TBP); the presence of TAFII250, TAFII105, and TBP was confirmed by Western blot analysis (Fig. 1B, lane 2) (10). No TFIID subunits were detected in the control precipitation with a nonspecific antibody (lane 1). These data establish that TAFII105 is a bona fide TFIID component in murine ovaries.

Given the highly restricted pattern of TAFII105 expression in mice, it seemed plausible that disruption of the endogenous TAFII105 gene by homologous recombination would not compromise viability. A TAFII105 genomic targeting vector (Fig. 2A) was constructed carrying the neomycin (NEO) resistance gene inserted in reverse orientation to the 3′ end of the TAFII105 gene (11). This construct was transfected into mouse embryonic stem (ES) cells, generating several clones containing one copy of the wild-type (WT) gene and one copy of the mutant TAFII105 gene (Fig. 2B). Heterozygous ES cells were injected into mouse blastocysts, and chimeric mice were established that transmitted a mutant copy of the TAFII105 gene through the germ line. A polymerase chain reaction (PCR)–based genotyping assay (Fig. 2C) was performed on the progeny of heterozygous matings to identify homozygous knockout (KO) animals. Western blot analysis of B cell extracts derived from mice of each genotype (Fig. 2D) confirmed that TAFII105 protein expression was abrogated (10). The TAFII105-specific band was detected only in cells derived from WT and HET mice and was absent in the KO mice. These data demonstrate that we successfully disrupted both copies of the endogenous TAFII105 gene and produced mice lacking TAFII105 protein.

Figure 2

Generation of a TAFII105 KO mouse. (A) The strategy used to disrupt the endogenous TAFII105 gene. Double arrows of 8.0 and 6.8 kb represent the WT and mutant alleles of TAFII105, respectively. Restriction sites of the genomic locus are indicated (B, Bam HI; Xb, Xba I; and H, Hind III), and a small black box indicates the probe used for Southern blot analysis. (B) Southern blot analysis of several ES clones containing both WT and mutant copies of the TAFII105 locus. (C) PCR-based genotyping of DNA derived from the progeny of TAFII105 heterozygous crosses. (D) Mouse TAFII105 protein is absent in cells derived from homozygous TAFII105 knockout mice. Western blot analysis of LPS-stimulated splenic B cell extracts with anti-TAFII105 is shown. Genotypes of protein sources are shown. The mouse TAFII105-specific protein band is indicated, and a nonspecific cross-reacting protein band is marked with an asterisk.

Thus far, heterozygous matings have yielded 242 progeny, with TAFII105 genotypes of 63 WT, 119 HET, and 60 KO. This distribution approximates the expected Mendelian ratio of 1:2:1 from heterozygous crosses and indicates that the lack of TAFII105 has no deleterious effect on viability. Because TAFII105 was originally identified in human B cells, several parameters of immune function were examined and shown to be unaffected by the deletion of TAFII105 (12). Thus, TAFII105 may play a redundant role in the immune system with another factor, such as TAFII130, that can compensate for the absence of TAFII105 in our KO mice.

Female homozygous KO mice exhibited a striking abnormality in fertility. In all the natural matings of hybrid TAFII105 KO female mice, no progeny were obtained. In contrast, male homozygous TAFII105 KOs were fertile. The infertility of hybrid homozygous KO females was consistent through six generations, suggesting that the defect is due to disruption of the TAFII105 gene and not some spurious mutation. To elucidate the basis of this infertility, we assessed whether null females cycled, exhibited behavioral estrous, and mated. All eight of the –/– females that were test-mated, bred, as evidenced by the presence of a copulatory plug. This observation suggested that the defect in TAFII105-null females may be ovarian and/or uterine. To distinguish between these possibilities, we induced pseudopregnancy in null females and performed embryo transfers (13). Four out of the five null females that carried transferred eggs became pregnant, gave birth to viable litters, lactated, and displayed normal maternal behavior. This finding suggested that the defect in fertility was not uterine but ovarian in origin.

To define the ovarian defect, we assessed whether null females ovulated naturally or were responsive to hormonal treatment to induce ovulation. Four nonhormone-treated –/– females that bred with stud males ovulated (∼seven eggs per female); however, none of the 26 eggs collected developed in culture. To assess whether null females could be induced to ovulate, we hormonally primed age-matched 8-week-old +/– and –/– females. Whereas eight treated +/–females ovulated 120 eggs, only a single egg was recovered from the null females. The failure to retrieve more than a single egg from treated null females could be a consequence of a defect in oocyte development and/or maturation. To distinguish between these possibilities, we isolated and weighed the ovaries from hormonally primed females and assessed ovarian histology. The ovaries from null females were half the weight (4.3 mg ± 1.3 mg; n = 8 mice) of the ovaries from age-matched fertile +/– females [8.9 mg ± 2.2 mg; n = 8 mice (14)]. Histologically, ovaries from treated null females had very few maturing follicles and only an occasional antral follicle (Fig. 3A). In contrast, ovaries from +/+ and +/– females displayed follicles at all developmental stages. Whereas 2-month-old null females failed to undergo induced ovulation, 5- to 6-month-old females did respond, albeit poorly, to hormone treatment. From 13 treated older null females mated to stud males, 36 eggs were collected, none of which cleaved in culture. This suggests a defect in oocyte maturation and/or fertilization. These data are consistent with the complete lack of fertility seen in TAFII105 KO females at all reproductive ages.

Figure 3

Expression of TAFII105 mRNA is restricted to the granulosa cells of the ovarian follicle. (A) TAFII105-deficient ovaries are smaller than WT ovaries and lack mature follicles. Hematoxylin-and eosin-stained ovary sections are shown (26). Genotypes of ovaries are indicated, and ovarian follicles are labeled as preovulatory follicle (Pf), antral follicle (af), growing follicle (gf), and corpus luteum (cl). (B) In situ hybridization with TAFII105- and TAFII130-specific antisense RNA probes (26). Genotypes of ovary sections and corresponding probes are shown.

To further confirm the relevance of the apparent lack of maturing follicles in hormone-primed females to the observed infertility, we localized the site of TAFII105 expression in the ovary. RNA in situ hybridization analysis revealed that TAFII105 is expressed exclusively in the granulosa cells surrounding the developing oocyte (Fig. 3B). In contrast, TAFII130 is expressed throughout the ovary, including the oocyte, granulosa cells, and corpus luteal cells. Whereas TAFII130 expression is readily detected in the TAFII105 KO ovary, there is a lack of TAFII105 expression in the mutant ovaries. There is a concomitant lack of maturing follicles in the KO ovary. Thus, the in situ data, together with the histologic observations, indicate that TAFII105 is likely critical for the timely development and/or proliferation of the granulosa cells of the ovarian follicle required for proper ovarian function.

To understand the molecular mechanism underlying the folliculogenesis defect caused by TAFII105 deletion, we set out to identify gene expression pathways that might be disrupted in our TAFII105-deficient ovaries. Oligonucleotide-based microarrays containing over 11,000 murine genes were probed with RNA derived from HET and KO ovaries dissected from 8-week-old age-matched and hormonally synchronized females. Only 132 genes (1%) of the genes assayed were down-regulated twofold or more (14). A number of genes known to function in folliculogenesis and known to be important for female fertility (15) were down-regulated (3- to 14-fold) in the TAFII105-deficient ovaries (Fig. 4). Most notably, the expression of multiple components of the inhibin-activin-follistatin pathway was severely compromised in the TAFII105 KO ovaries. Inhibins and activins are transforming growth factor–β (TGF-β) family members that modulate the release of follicle-stimulating hormone from the pituitary and regulate estrogen synthesis within the ovary (16). In addition, the expression of aromatase p450, a critical enzyme responsible for the conversion of androgens to estrogens in granulosa cells, decreased dramatically. Finally, cyclin D2, which is essential for granulosa cell proliferation and female fertility (17) was also down-regulated in the mutant ovaries. Consistent with the role of this specific set of genes, down-regulating their expression correlates with the improper regulation of folliculogenesis seen in the TAFII105-deficient ovaries and may explain the infertility observed in these mice.

Figure 4

Ovarian-specific genes down-regulated in TAFII105-deficient ovaries. Total RNA was isolated from dissected ovaries, and 7 μg was used to synthesize cRNA probes for hybridization to murine 11K microarrays (Affymetrix). Gene expression changes of over 11,000 genes were compared between the +/– and –/– ovaries, and genes that decreased by twofold or greater in the –/– ovaries were identified by genechip software (Affymetrix). Fold reductions in the TAFII105-deficient ovaries of a subset of these genes involved in female reproduction are listed. RNase protection assays (25) and corresponding genotypes of the RNA source are shown.

These studies establish that there are cell type–specific components of TFIID in mammals and that specialized transcription initiation complexes function in regulating tissue-specific programs of gene expression. Such cell type–specific general factors function analogously to sigma factors of the prokaryotic transcription system, which regulate promoter selectivity of RNA polymerase (18). These findings complement the recent findings of Fuller and colleagues, who have shown that a testes-specific homolog ofDrosophila TAFII80 called Cannonball is required for germ cell development in male flies (19). In addition, the TBP-homolog TRF2 has been shown to be required for embryogenesis inCaenorhabditis elegans and Xenopus, but it is required only for spermiogenesis in the mouse (20–24). Taken together, these studies emphasize the critical roles of the general transcriptional machinery in regulating cell type– and gene-specific transcription in metazoans.

The restricted expression of TAFII105 in the granulosa cells of fertile ovaries strongly correlates with the apparent defective function of this cell type in TAFII105 KO mice. In addition, we have identified a number of putative target genes, specifically down-regulated in the TAFII105-deficient ovaries, that are known to play a critical role in folliculogenesis (15). On the basis of TAFII105 expression in the granulosa cells of the ovary and TAFII105 being associated with TBP in ovary extracts, we envision that TAFII105 mediates the transcription of a subset of granulosa cell-specific genes, some of which have been identified in our microarray experiments (14). Future biochemical studies will examine whether TAFII105 directly mediates the expression of these target genes. It is tempting to speculate that TAFII105 is also involved in regulating ovarian gene expression in humans. Therefore, characterizing how TAFII105 functions in the mouse may help reveal the molecular basis of certain types of female infertility in humans.

  • * To whom correspondence should be addressed. E-mail: jmlim{at}


View Abstract

Stay Connected to Science

Navigate This Article