Distinct Roles for TBP and TBP-Like Factor in Early Embryonic Gene Transcription in Xenopus

See allHide authors and affiliations

Science  22 Dec 2000:
Vol. 290, Issue 5500, pp. 2312-2315
DOI: 10.1126/science.290.5500.2312


The TATA-binding protein (TBP) is believed to function as a key component of the general transcription machinery. We tested the role of TBP during the onset of embryonic transcription by antisense oligonucleotide–mediated turnover of maternal TBP messenger RNA. Embryos without detectable TBP initiated gastrulation but died before completing gastrulation. The expression of many genes transcribed by RNA polymerase II and III was reduced; however, some genes were transcribed with an efficiency identical to that of TBP-containing embryos. Using a similar antisense strategy, we found that the TBP-like factor TLF/TRF2 is essential for development past the mid-blastula stage. Because TBP and a TLF factor play complementary roles in embryonic development, our results indicate that although similar mechanistic roles exist in common, TBP and TLF function differentially to control transcription of specific genes.

The TATA-binding protein (TBP) is often considered an essential component of the general transcription machinery, being involved in transcription by all three eukaryotic RNA polymerases. TBP is essential in yeast, binds to a variety of TATA boxes, and is recruited to TATA-less promoters via protein-protein interactions [reviewed in (1)]. Recently, however, TBP-related factors have been identified in metazoans. Drosophila has two such proteins (TRF1 and TRF2). TRF1 plays a role in the transcription of tudor and transfer RNA (tRNA) genes (2,3). TRF1 has thus far not been identified in human, mouse,Xenopus, or Caenorhabditis elegans. TRF2, also known as TBP-like factor (TLF), is found in many metazoans, including Xenopus (4). The protein interacts with TFIIA and TFIIB and does not bind to canonical TATA boxes but may positively contribute to promoters similar to the SV40 promoter (5–7).

Here we examine whether either TBP or TLF/TRF2 is required for the onset of embryonic transcription or embryogenesis of Xenopus laevis. In Xenopus embryos, transcription begins after the egg has divided into 4000 cells, at the mid-blastula stage (8, 9). Before the mid-blastula transition (MBT), embryos contain very little TBP (10, 11). However, TBP is translated from maternally stored RNA just before the MBT to better match the levels of other basal transcription factors (11). TBP is rate-limiting for transcription before the MBT in both extracts and embryos (11–13), and precocious translation of synthetic TBP RNA in the embryo facilitates transcription before the MBT (11). The rapid rise in TBP protein levels correlates well with the onset of embryonic transcription.

TBP protein accumulation was prevented by an antisense oligonucleotide–based approach. We injected embryos with chemically modified oligonucleotides that are more stable in vivo than unmodified oligonucleotides, in order to reduce the amount of oligonucleotide required for effective mRNA turnover. These oligonucleotides contained a backbone modified with a cationic diethyl-ethylenediamine (DEED) group (14), except for six or seven nucleotides in the middle that have normal phosphodiester bonds in their backbone. The 5′ and 3′ modified bases protect against exonuclease activity, whereas six phosphodiester bonds in the middle fulfill ribonuclease H (RNase H) substrate requirements to cleave the targeted mRNA (15). One of these oligonucleotides, TBP-AS3 (Fig. 1A), efficiently degraded maternal TBP mRNA in the embryo (Figs. 1B and 3B), effectively preventing the accumulation of TBP protein at the mid-blastula transition (Fig. 1C). Although these embryos lacked detectable levels of TBP, they were able to initiate gastrulation (Fig. 2C). However, development arrested before the completion of gastrulation (Fig. 2D). We have also tested a more fully DEED-modified TBP-AS3 oligonucleotide. By leaving only three instead of six phosphodiester bonds in the middle of the oligonucleotide, hybrids formed with mRNA are not degraded by RNase H (15). Embryos injected with this oligonucleotide developed normally (16). In addition, we constructed a mutant TBP RNA that encodes the wild-type TBP protein but is resistant to TBP-AS3–mediated degradation (Fig. 1B). This mutant RNA was translated in the embryo (Fig. 1C) and rescued the effects of the TBP-AS3 depletion of endogenous TBP mRNA. Rescued embryos not only completed gastrulation but developed into larvae (Fig. 2, E and F). Because transcription is required for the onset of gastrulation (8, 17), it is remarkable that embryos without detectable levels of TBP initiated gastrulation. It is possible that levels of TBP that are undetectable by Western blotting contribute to the initiation of gastrulation in TBP-AS3 embryos. However, the significant reduction of transcription from a number of genes well before the developmental arrest suggests that TBP was functionally deficient in these embryos.

Figure 1

The TBP-AS3 antisense oligonucleotide abolishes TBP protein accumulation during early embryogenesis. (A) Region of the TBP RNA targeted with the AS3 oligonucleotide. Point mutations were made (residues shown in bold) to render a synthetic RNA resistant to AS3-dependent turnover. The six nucleotides in the middle of the AS3 oligonucleotide feature normal phosphodiester bonds (shaded box). (B) Northern blot showing that the maternal TBP message (first lane) was effectively degraded in pre-gastrula embryos by TBP-AS3 oligonucleotide (second lane). A synthetic mutant TBP mRNA (0.5 ng per embryo) was resistant to AS3-mediated turnover (third lane). Staining of 28S RNA is a loading control (lower panel). (C) Western blot showing that the AS3 oligonucleotide prevents accumulation of TBP protein at the MBT (first and second lanes). A mutant TBP mRNA resistant to AS3-mediated turnover efficiently restored TBP protein levels to the embryo (third lane). Details on the methods are provided in (26).

Figure 2

TBP accumulation is required for embryonic development but is not essential for the onset of gastrulation. Embryos injected at the one-cell stage with water (A and B), 2 ng of partially modified TBP-AS3 oligonucleotide (C and D), or 2 ng of TBP-AS3 plus 0.25 ng of mutant TBP RNA (E and F) are shown at stages 12 (advanced gastrula, left panels) and 25 (larva, right panels). More than 95% of TBP-AS3–injected embryos showed a phenotype similar to the phenotype shown in (C) and (D).

To determine to what extent transcription was affected in TBP-AS3–injected embryos, α32P-UTP was injected to visualize the most abundant transcripts at the MBT. TBP-AS3 reduced most transcripts (Fig. 3A). Specifically, tRNA synthesis, which accounts for more than 85% of all transcription in the embryo after the MBT, decreased about 10-fold in TBP-deficient embryos. As for the other identifiable RNA polymerase III–dependent transcripts, 7S but not 5Stranscription was down-regulated in AS3-injected embryos. U1 and U2, two abundant RNA polymerase II–dependent transcripts involved in splicing, were also affected in these embryos (Fig. 3A). Northern blot analysis was performed to investigate the requirement for TBP protein accumulation for the embryonic onset of transcription of protein-encoding genes (Fig. 3B). The genes selected for analysis are transcribed primarily after the MBT. Xbra is a frog brachyurytranscription factor homolog involved in mesoderm formation (18), whereas GS17 is a gene of unknown function that is activated at the MBT (19). Neither gene was significantly affected in TBP-AS3–injected embryos (Fig. 3B). In contrast, transcription of the genes encoding translation elongation factor EF1α (20), Xenopus keratins XK81A1 and XK70A (21), and MyoD (22) decreased significantly in embryos lacking detectable amounts of TBP (Fig. 3B). These effects were rescued by injected TBP message (Fig. 3C).

Figure 3

TBP is required for the onset of transcription of a subset of embryonically transcribed genes. (A) Embryos were injected at the one-cell stage with [α32P]UTP with or without TBP-AS3 oligonucleotide, and RNA was isolated at the stages of development indicated. Developmental stages are as follows: Stages 6 to 9, blastula; stage 8.5, mid-blastula; stages 10.5 to 13, gastrula. Transcripts were identified by size as previously described (8, 9). (B) Northern blots were labeled, stripped, and reprobed multiple times with DNA fragments of the genes shown on the right. Three embryo equivalents of RNA from uninjected (first and third lanes) or TBP-AS3 oligonucleotide–injected (second and fourth lanes) embryos were loaded per lane. (C) TBP mRNA rescues the effects of the AS3 oligonucleotide on gene expression.

The observation that not all embryonic transcription is TBP-dependent raised the possibility of an important developmental role of TLF (TRF2). The Xenopus homolog of TLF was isolated by reverse transcription–polymerase chain reaction (RT-PCR). We determined the expression of TLF during embryogenesis by quantitative RT-PCR. TLF mRNA was detected in oocytes and embryos, being relatively abundant during cleavage and blastula stages (Fig. 4A). TLF is widely expressed in the embryo, as was shown in a large-scale in situ hybridization screen using randomly picked cDNAs, one of which was TLF (clone 3.39) (23). We targeted the endogenous TLF mRNA for degradation with a DEED-modified oligonucleotide, TLF-AS93, similar to the way in which we targeted TBP mRNA with TBP-AS3. TLF-AS93 mediated efficient degradation of TLF mRNA in vivo (Fig. 4D). TLF-AS93–injected embryos developed normally until the onset of embryonic transcription between stage 8 and 9, the stage at which they arrested (Fig. 4, B and C). TLF-AS57, a different oligonucleotide capable of degrading TLF mRNA in vivo, caused a similar developmental arrest, whereas control oligonucleotides did not (16). Recently, a similar early embryonic arrest was observed in tlf-1(RNAi) embryos of C. elegans(24, 25). We examined the expression of embryonically transcribed genes in TLF-AS93–injected embryos, and we could not detect embryonic transcription of Xbra orEF1α, and GS17 RNA was also severely reduced [Fig. 4, D and E (16)]. Maternal TBP mRNA, which in normal embryos is translated at the MBT (11), was translated in TLF-AS93 embryos (Fig. 4D), albeit with slightly lower efficiency. We ruled out a possible contribution of lower TBP levels to the TLF-AS93 phenotype by overexpressing TBP with a synthetic TBP mRNA. Both the TLF-AS93–mediated developmental arrest and the transcriptional impairment of GS17 were independent of TBP levels (Fig. 4D). The expression of MyoD and GS17 was examined in more detail. In the absence of detectable amounts of TBP, MyoD was not transcribed de novo, although low levels of maternally deposited MyoD RNA were observed (Fig. 4E). In TLF-AS93–injected embryos, on the other hand, MyoD was expressed after their developmental arrest, most notably at the time normal embryos start gastrulation [stage 10 to 11 (Fig. 4E)]. GS17 expression by comparison was significantly affected in TLF-AS93 embryos (Fig. 4, D and E) but much less so in TBP-AS3 embryos (Figs. 3B and 4E).

Figure 4

Expression and function of TLF inXenopus embryogenesis. (A) TLF mRNA abundance in oocytes (O), and cleavage (stage 3), early and late blastula (stages 7 and 9), gastrula (stage 11), neurula (stage 16), and tailbud stage (stage 28) embryos, as detected by quantitative RT-PCR (26). Staining of 28S and 18S RNA was performed as a control for RNA amount (lower panel). (B andC) Phenotypes of normal and TLF-AS93–injected embryos at stage 12 (gastrula). The arrest was observed in close to 100% of all TLF-AS93–injected embryos. (D) Top panel: TLF mRNA is efficiently degraded in a TLF-AS93–dependent fashion in vivo, as assessed by quantitative RT-PCR. Second panel: TBP expression in control embryos (first lane), TLF-AS93 oligonucleotide–injected embryos (second lane), and embryos injected with TLF-AS93 and TBP RNA (third lane). Any TBP detected in the first and second lanes is the result of embryonic translation of maternally deposited TBP mRNA (11). Third panel: Northern blot analysis ofGS17 expression in normal and TLF-AS93 gastrula stage embryos. Methylene blue staining of 18S RNA was performed as a loading control (bottom panel). (E) RT-PCR of MyoD and GS17 RNA of normal, TBP-AS3–injected, and TLF-AS93–injected embryos at different stages (stages are listed in Fig. 3 legend). Details on the methods are provided in (26).

These data establish that TBP is essential for transcription of some but not all class II and class III genes during early embryogenesis. TBP is dispensable for transcription of a subset of genes and for the onset of gastrulation, whereas it is required for sustaining embryonic development. TLF, on the other hand, is essential for development beyond the mid-blastula stage. Our analysis of TBP and TLF function suggests that a remarkable functional dichotomy exists regarding general transcription factor requirements. Important developmental genes such as Xbra are transcribed in the absence of TBP, whereas a gene such as MyoD is strictly TBP-dependent. TBP and TLF fulfill distinct requirements for transcription and embryogenesis.

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


View Abstract

Navigate This Article