Ubiquitin-Activating/Conjugating Activity of TAFII250, a Mediator of Activation of Gene Expression in Drosophila

See allHide authors and affiliations

Science  29 Sep 2000:
Vol. 289, Issue 5488, pp. 2357-2360
DOI: 10.1126/science.289.5488.2357


Ubiquitination of histones has been linked to the complex processes that regulate the activation of eukaryotic transcription. However, the cellular factors that interpose this histone modification during the processes of transcriptional activation are not well characterized. A biochemical approach identified the Drosophila coactivator TAFII250, the central subunit within the general transcription factor TFIID, as a histone-specific ubiquitin-activating/conjugating enzyme (ubac). TAFII250 mediates monoubiquitination of histone H1 in vitro. Point mutations within the putative ubac domain of TAFII250 abolished H1-specific ubiquitination in vitro. In the Drosophilaembryo, inactivation of the TAFII250 ubac activity reduces the cellular level of monoubiquitinated histone H1 and the expression of genes targeted by the maternal activator Dorsal. Thus, coactivator-mediated ubiquitination of proteins within the transactivation pathway may contribute to the processes directing activation of eukaryotic transcription.

Transcriptional activation is initiated by the binding of activator proteins to the enhancer region of their target genes and culminates in the recruitment of the general RNA polymerase II transcription machinery to core promoters of eukaryotic genes (1). However, within chromatin, the association of chromosomal DNA with histones (H1, H2A, H2B, H3, and H4) to form nucleosomes can inhibit the interaction of transcription factors and the general transcription machinery with target genes and, hence, transcriptional activation (2). To overcome this “nucleosome barrier,” transcription factors can recruit coactivators, which by prosttranslational modification of histones establish transcriptionally competent chromatin structures (3). Coactivator-mediated acetylation and methylation of histones has been intimately connected with activation of transcription (4, 5).

One other noteworthy posttranslational modification is the conjugation of histones with ubiquitin (6). Polyubiquitination represents a mark on proteins that identifies them for degradation and requires the involvement of three enzymes: (i) ubiquitin-activating enzymes (E1), which mediate the adenosine triphosphate (ATP)–dependent conjugation of E1 with ubiquitin via a covalent thioester linkage; (ii) ubiquitin-conjugating enzymes (E2), which mediate the transfer of ubiquitin from E1 to E2, conjugate ubiquitin via thioester bonds and, (iii) together with ubiquitin-protein ligase (E3), link ubiquitin to target proteins via isopeptide bonds (7). Polyubiquitination requires all three enzymes, whereas monoubiquitination of proteins requires E1 and E2 activities only (7). Unlike polyubiquitination, monoubiquitination of histones has been correlated with activation of gene expression (3, 6,8, 9). However, the functional connections between histone ubiquitination and activation of gene expression remain unknown. Thus, as a first step toward understanding the role of histone ubiquitination for transcriptional regulation, we sought to identify enzymes that ubiquitinate histones in Drosophila embryonic nuclear extract (10) using an activity gel assay (11, 12).

Nuclear extract was separated in SDS-polyacrylamide gels containing histones. After electrophoresis (SDS-PAGE), gel-bound proteins were subsequently denatured, renatured, and, to monitor enzymatic activities, incubated with 32P-labeled ubiquitin (13, 14). By using this assay, we identified a protein with a molecular mass of approximately 200 kD that mediates ubiquitination of histones (Fig. 1A). To purify the identified activity from nuclear extract, we used heparin-Sepharose, S300 gel filtration, and MonoQ chromatography (15). Gel assays revealed that the 200-kD protein was present in the 0.4 M KCl heparin-Sepharose eluate, the void-volume of S300 gel filtration and the 0.2 M KCl eluate of MonoQ chromatography [Web fig. 1 (16)]. This fractionation pattern resembles the pattern described for the general transcription factor TFIID (15). TFIID is composed of the TATA box–binding protein (TBP) and at least 10 different TBP-associated factors (17). To investigate the functional relation, if any, between the 200-kD protein and TFIID, we immunopurified TFIID from the 0.2 M KCl MonoQ fraction using antibodies to TBP (18). Gel assays revealed that the 200-kD activity is present in immunoprecipitates obtained with antibodies to TBP and suggest that the 200-kD protein either interacts withDrosophila TFIID or is an integral component of the TBA-TAFII complex (Fig. 1B). As the largest TAFII-subunit (TAFII250) within the TFIID-complex has a molecular mass of about 200 kD (17), the precipitated activity may correspond toDrosophila TAFII250 (dTAFII250). Western blot analyses of the immunoprecipitates from antibodies to TBP using antibodies to dTAFII250 (19) revealed that the 200-kD activity coincided with dTAFII250, suggesting that dTAFII250 may ubiquitinate histones (Fig. 1B).

Figure 1

A 200-kD Drosophila protein ubiquitinates. (A) Autoradiogram of activity gel assays.Drosophila nuclear extract was fractionated by SDS-PAGE using gels supplemented with histones or BSA. The 200-kD activity (arrow) was observed in gels containing histones (lane 1) but not in gels containing BSA (lane 2). (B) Proteins in the 0.2 M KCl MonoQ fraction [Web fig. 1 (16)] were immunoprecipitated using antibodies to TBP (lanes 1, 3, and 4) or HA (lanes 2 and 5) and were analyzed by SDS-PAGE using gels containing histones (lanes 1, 2, 4, and 5) or BSA (lane 3). Ubac activity (arrow) was detected by autoradiography (lanes 1 to 3). dTAFII250 (arrowhead) was detected using antibodies to dTAFII250: 2B2 (lanes 4 and 5) or 30H9 (22). The asterisk represents an unknown protein that interacts with TFIID and ubiquitinates histones. (C) Membrane assays of dTAFII250 separated by SDS-PAGE using gels containing H1 (lanes 1 and 2) or BSA (lane 3). Reaction products were detected by Western blot using antibodies to TAFII250 (lane 1) or ubiquitin (lanes 2 and 3). The positions of dTAFII250 (arrowhead) (lane 1) and the 200-kD protein having ubac activity (arrow) are indicated. The position and size (in kilodaltons) of protein standards are indicated to the left (A to C).

To support this hypothesis we used membrane assays (20). Purified, recombinant dTAFII250 [Web note 1 (16)] was loaded onto SDS-polyacrylamide gels supplemented with individual histones and separated by SDS-PAGE, transferred onto polyvinylidene difluoride (PVDF) membrane, denatured, renatured, and incubated with ubiquitin. After incubation, proteins were treated with reducing agents to disrupt thioester bonds. On membranes loaded with histone H1 (H1) (Fig. 1C), but not on membranes loaded with H2A, H2B, H3, or H4 [Web fig. 2A (16)], antibodies to ubiquitin detected ubiquitin at a position that coincides with dTAFII250, suggesting that dTAFII250 ubiquitinates H1. The membrane assay [which subjects proteins to boiling in the presence of SDS, size-fractionation by electrophoresis, and denaturation by 7 M guanidine HCl (GuHCl)] and the molecular weight of known enzymes in the ubiquitin pathway (7) suggest that membrane-bound dTAFII250 most likely does not interact with E1, E2, or E3 enzymes. As mono-ubiquitination requires at least E1 and E2 activities, our results imply that TAFII250 may have intrinsic E1 and E2 activities. The ubiquitin/H1 conjugates detected in the membrane assay resisted reducing agents, suggesting that dTAFII250 may mediate a covalent bond between ubiquitin and H1 by means of isopeptide linkages. As this enzymatic reaction is characteristic for E2 enzymes, dTAFII250 may have intrinsic E2 activity.

The E1 enzyme requires ATP to conjugate with ubiquitin by means of thioester bonds (7). Therefore, to explore whether dTAFII250 has E1 activity, we investigated the capability of dTAFII250 for conjugating with ubiquitin by means of thioester bonds (21). dTAFII250 conjugated with ubiquitin in an ATP-dependent manner in the absence, but not in the presence, of reducing agents (Fig. 2A), suggesting that dTAFII250 and ubiquitin form a covalent bond by means of a thioester linkage. Thus, dTAFII250 may have both E1 and E2 activities and may therefore be a ubac.

Figure 2

The Drosophila coactivator dTAFII250 has H1-specific ubac activity. (A) Coomassie blue–stained gel (C) or autoradiogram (A) of solution assays containing dTAFII250 (double arrow) and32P-labeled ubiquitin and ATP, DTT, or β-mercaptoethanol (β-ME) in the reaction (RB) or SDS-PAGE sample buffer (SB). (B) Aliquots of solution assays containing H1 (arrowhead), 32P-labeled ubiquitin (asterisk), dTAFII250 (double arrowhead), and ATP, as indicated, were separated by SDS-PAGE, and proteins were detected by Coomassie blue staining. (C) Aliquots of solution assays described in (B) were separated by SDS-PAGE, and proteins were detected by autoradiography (lanes 1 to 5), antibodies to H1 (lanes 6 to 10) or ubiquitin (lanes 11 to 15). The positions of monoubiquitinated H1 (arrow) and H1 (arrowhead) are indicated. (A to C) The position and size (in kilodaltons) of protein standards are indicated to the left.

To provide supporting evidence that dTAFII250 mediates ubiquitination of H1, we used solution assays [Web note 2 (16)]. Reactions containing dTAFII250,32P-labeled ubiquitin, H1, and ATP mediated the formation of a 39-kD protein that was recognized by antibodies to both ubiquitin and H1 (Fig. 2, B and C). These results suggest that the 39-kD protein represents a conjugate composed of one ubiquitin moiety (7 kD) and H1 (32 kD). By contrast, dTAFII250 did not ubiquitinate other histones [Web fig. 2B (16)], H2A/H2B dimers, H3/H4 tetramers, or core nucleosomes (22). Thus, dTAFII250 mediates monoubiquitination of H1.

To determine the portion of dTAFII250 that mediates monoubiquitination of H1, we used dTAFII250 mutants truncated at the COOH-terminal [Web fig. 3 and note 1 (16)]. Membrane assays indicate that full-length dTAFII250 and 250(ΔC850), lacking the 850 amino acids closest to the COOH-terminal, but not 250(ΔC1300), lacking the 1300 amino acids closest to the COOH-terminal, ubiquitinated H1 (Fig. 3A). Thus, the H1-specific ubac activity is likely to reside between amino acids 768 and 1218.

Figure 3

Point mutations in the ubac domain of dTAFII250 abolish H1-specific ubac activity. (A) Coomassie blue-stained gel (lanes 1 to 3) or membrane assays (lanes 4 to 6) of dTAFII250 (arrow) (lanes 1 and 4), 250(ΔC850) (arrowhead) (lanes 2 and 5), and 250 (ΔC1300) (asterisk) (lanes 3 and 6), and Coomassie blue-stained gel (lanes 7 to 9) and membrane assays (lanes 10 to 12) of TAF250- M (lanes 7 and 10), TAF250-M-V1072D (lanes 8 and 11), and TAF250-M-R1096P (lanes 9 and 12). Proteins were separated in SDS-polyacrylamide gels containing H1 and detected using antibodies to ubiquitin (lanes 4 to 6 and 10 to 12). The position and size (in kilodaltons) of protein standards are indicated to the left. (B to K) Reduction of gene expression in embryos lacking dTAFII250 ubac activity. In situ hybridizations showing twist (B to F) orsnail (G to K) expression in dl/+ (B and G),TAF250XS-2232 (C and H),TAF250S-625 (D and I), dl/+;TAF250XS-2232 (E and J) and dl/+;TAF250S-625 (F and K) embryos. The ventral surface of blastoderm-stage embryos is shown with anterior pointing to the left. (L to P) Dark-field images of the cuticular body pattern of dl/+ (L),TAF250XS-2232 (M),TAF250S-625 (N), dl/+;TAF250XS-2232 (O), and dl/+;TAF250S-625 (P) embryos.

Most recently, two Drosophila TAF250 alleles,TAF250XS-2232 andTAF250S-625 , have been described, which contain single–amino acid point mutations that reside within the putative dTAFII250 ubac domain (23).TAF250XS-2232 contains a valine-1072 to aspartic acid change, and TAF250S-625 an arginine-1096 to proline change [Web fig. 3 (16), (23)]. To investigate the effect of these mutations on dTAFII250 ubac activity, the middle region of TAFII250 containing amino acids 612 to 1140 (TAF250-M), TAFII250-M-V1072D containing the (V1072 to D) mutation, and TAFII250-M-R1096P containing the (R1096 to P) mutation (23), were subjected to membrane assays. Although TAF250-M ubiquitinated H1, the mutants did not (Fig. 3A). Wild-type and mutant TAFII250-M proteins have histone acetyltransferase activity (23). The TAF250-M proteins used for the membrane assays acetylated histones (22), suggesting that the lack of ubac activity seen with TAF250-M-V1072D and TAFII250-M-R1096P is most likely not due to a general functional inactivity of the mutant proteins.

In Drosophila, TFIID mediates transcriptional activation by the maternal activator Dorsal (24). Dorsal activates the expression of the mesoderm-determining genestwist (twi) and snail (sna), which are transcribed in 20 and 18 of the ventral-most cells of cellularizing embryos, respectively (24). To investigate the functional relevance of TAFII250 ubac activity for Dorsal-dependent transcriptional activation in vivo, we used in situ hybridization (25) to monitor twi andsna expression in Drosophila embryos containing reduced levels of Dorsal and expressing TAFII250XS-2232 or TAFII250S-625, which lack ubac activity in vitro. Both twi and sna expression were severely reduced in dl-sensitized,TAF250XS-2232 embryos (Fig. 3, E and J) and dl-sensitized,TAF250S-625 embryos (Fig. 3, F and K) but not in control embryos (Fig. 3, B to D and G to I). Weaktwi mRNA levels were detectable in 10 to 12 cells (Fig. 3, E and F), and sna expression was restricted to 4 to 12 ventral-most cells and disrupted by gaps (Fig. 3, J and K). Analyses of cuticular preparations (26) revealed thatdl-sensitized TAF250XS-2232 mutants (Fig. 3O) or dl-sensitizedTAF250S-625 mutants (Fig. 3P), but not control embryos (Fig. 3, L to N) exhibited a dorsalized and twisted body pattern. These results indicate that Dorsal-dependent activation of transcription is impaired in embryos lacking TAFII250 ubac activity.

To investigate whether H1 may represent a target for dTAFII250 ubac activity in Drosophila, we purified H1 from nuclei prepared from 0- to 3-hour-old wild-type and TAFII250 mutant embryos (27, 28). Western blot analyses indicate that antibodies to both H1 and ubiquitin detected a monoubiquitin/H1 conjugate (Fig. 4). This result indicates that at least a fraction of H1 present in earlyDrosophila embryos is monoubiquitinated. Moreover, Western blot analyses indicate that compared with wild-type embryos, mutant embryos that lack TAFII250 ubac activity contained a significantly reduced level of monoubiquitinated H1 (Fig. 4). These results suggest that dTAFII250 ubac activities may contribute to monoubiquitination of H1 in Drosophila.

Figure 4

Histone-1 is ubiquitinated in Drosophila.Western blot analysis of H1 isolated from wild-type (lanes 1 and 4), homozygous mutant TAFXS-2232 embryos (lanes 2 and 5), or homozygous mutant TAFS-625 embryos (lanes 3 and 6). Proteins were detected using antibodies to H1 (lane 1 to 3) or ubiquitin (lanes 4 to 6). The positions of H1 and monoubiquitinated H1 (uH1) are indicated. The position and size (in kilodaltons) of protein standards are indicated to the left.

How coactivators convert activation signals from activation domains of transcription factors into enhanced levels of mRNA synthesis lies at the heart of transcriptional regulation. Our results suggest that one coactivator, TAFII250, may use intrinsic ubiquitin-activating/conjugating activities to mediate activation of transcription. Multiple-alignment analysis and comparison with protein database sequences revealed that TAF250-M exhibits similarities to E1 and E2 enzymes (29). Thus, our result that TAFII250 mediates monoubiquitination of H1 in vitro is in agreement with other results suggesting that E1 and E2 activities are sufficient to mediate monoubiquitination of proteins (7). As point mutations that abrogate TAFII250 ubac activity in vitro also reduce gene expression in the Drosophila embryo, TAFII250 ubac activity may play an important role for the activation of gene expression inDrosophila. Although the in vivo targets of TAFII250 ubac activity remain unknown, our results that H1 is monoubiquitinated in Drosophila and that the level of monoubiquitinated H1 is significantly reduced in embryos lacking dTAFII250 ubac activity imply that H1 may represent one in vivo target of dTAFII250. Thus, ubiquitination of H1 or other proteins within the transcription machinery, or both, by TAFII250 may constitute an important coactivator function of TAFII250 and, hence, may allow TFIID to direct events during the processes of transcriptional activation.

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


View Abstract

Navigate This Article