SV40 Large T Antigen Binds to the TBP-TAF I Complex SL1 and Coactivates Ribosomal RNA Transcription

Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

SV40 large T antigen binds to the TBP-TAF I complex SL1 and coactivates ribosomal RNA transcription

Weiguo Zhai, JoAnn Tuan, and Lucio Comai 1 Department of Molecular Microbiology and Immunology, University of Southern California, School of Medicine, Los Angeles, California 90033 USA

SV40 large T antigen is a multifunctional regulatory protein that plays a key role in the viral life cycle and can stimulate cell proliferation. To accomplish this, large T antigen has to control the expression of cellular genes involved in cell cycle progression and cell growth, rRNA synthesis by RNA polymerase I (Pol I) is tightly associated with cell growth and proliferation, and previous studies indicated that large T antigen up-regulates RNA Pol I transcription in SV40-infected cells. How this process occurs is currently unclear. To investigate the mechanisms of large T antigen stimulation of RNA Pol I transcription, we have established an in vitro transcription system that is responsive to large T antigen. Here, we show that recombinant large T antigen stimulates Pol I transcription reconstituted with purified RNA Pol I, UBF, and the TBP/TAF complex SL1. Immunoprecipitation experiments revealed that large T antigen directly binds to SL1, in vitro, as well as in SV40-infected cells. In addition, our data indicate that this interaction occurs by direct association with three SL1 subunits, namely TBP, TAFI48 , and TAF~ll0. Transcription studies with large T antigen deletion mutants show that the 538-amino-acid amino-terminal domain is necessary for full stimulation of Pol I transcription. Importantly, mutants that no longer bind to SL1 are also unable to stimulate Pol I transcription. This indicates that recruitment of large T antigen to the rRNA promoter by SL1 constitutes a crucial step in the activation process. Taken together with recent studies on large T antigen activation of RNA Pol II transcription, these results suggest that viral modulation of genes involved in cell proliferation involves direct targeting of promoter-specific TBP/TAF complexes (i.e., SL1 or TFIID) by large T antigen. [Key Words: Transcription; SL1; TAFs; Pol I; large T antigen] Received January 24, 1997; revised version accepted May 6, 1997.

RNA polymerase I (Pol I) directs RNA synthesis from a box sequence-specific DNA-binding protein, binds to the single class of genes, the rRNA genes, which are found in rDNA promoter and recruits the selectivity factor SL1 multiple tandem arrayed copies in the nucleoli of eu- (Bell et al. 1988; Jantzen et al. 1990, 1992). SL1 is re- karyotic cells (Reeder 1994; Jacob 1995; Moss and quired to direct initiation from the rRNA promoter and Stefanovsky 1995). rRNA transcription plays a critical plays a crucial role in the promoter recognition proper- role in ribosome biogenesis, and changes in the tran- ties of the rRNA transcriptional apparatus (Learned et al. scription rate of RNA Pol I are associated with profound 1985; Closet al. 1986). Human SL1 is a species-specific alterations in the growth rate of the cell (Sollner-Webb factor, and when added to a mouse cell-flee transcription and Tower 1986; Sommerville 1986). In humans, tran- extract, it has the ability to reprogram that extract to scription by RNA Pol I requires at least two transcription recognize a human rRNA promoter (Mishima et al. 1982; factors in addition to RNA Pol I: the upstream binding Learned et al. 1985; Bell et al. 1990; Schnapp et al. 1991). factor (UBF), and the selectivity factor SL1 (Learned et al. SL1 is analogous to the TFIID and TFIIIB factors in the 1985, 1986). Biochemical studies have established that a Pol II and Pol III transcription systems and consists of complex set of interactions of UBF and SL1 with each TATA-binding protein (TBP), plus three distinct TBP- other and with the DNA template are necessary to form associated factors (TAFs) essential to reconstitute Pol I an active complex and to initiate promoter-specific tran- transcription (Comai et al. 1992; Eberhard et al. 1993; scription (Learned et al. 1986; Bell et al. 1988). In vitro Hernandez 1993). In the presence of UBF, the SL1 com- studies indicated that UBF, a high mobility group (HMG) plex binds the DNA template and extends the DNase I footprint to include the species-specific element (SSE) (Bell et al. 1988; Safrany et al. 1989). In response to physiological and environmental 1Corresponding author. changes, cells must regulate the expression level of E-MAIL [email protected];FAX (213) 342-1721. metabolically important genes such as the rRNA genes.

GENES & DEVELOPMENT11:1605-1617 © 1997 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/97 $5.00 1605 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Zhai et al.

Extracellular stimuli, as well as many pathological phe- with preimmune serum (Ide et al. 1977). Although these nomena, including amino acids starvation, toxic lesion, studies established the stimulatory effect of large T an- aging, cancer, and viral infections, dramatically modu- tigen on Pol I transcription, the mechanisms by which T late RNA Pol I transcription (Grummt et al. 1976; antigen affects rRNA synthesis, and the cellular factors Mishima and Muramatsu 1979; Cavanaugh and Thomp- involved in this phenomenon, remain unknown. son 1986; Hammond and Bowman 1988; Reeder 1994; In this study we investigate the molecular mecha- Jacob 1995). Early studies showed that infection of mam- nisms of stimulation of RNA Pol I transcription by large malian cells with SV40 induced rRNA synthesis (May et T antigen. First, we show that recombinant large T an- al. 1976; Pockl and Wintersburger 1980). Further analysis tigen can stimulate transcription from the rRNA pro- indicated that the product of the viral gene A, the large moter in transfection assays and in a cell-free system tumor antigen (large T antigen), was directly involved in with purified RNA Pol I, UBF, and SL1. Second, we pro- this process (Ide et al. 1977; Soprano et al. 1981; Learned vide evidence that large T antigen binds to the SL 1 com- et al. 1983). Large T antigen is a multifunctional protein plex through interactions with T/i~FI48, TAF~ll0, and that has been shown to induce cell proliferation, trans- TBP. Finally, we show that the region of large T antigen formation, and tumor formation in animals (for review, that binds to the TAF~s and TBP is also necessary for see Fanning 1992; Conzen and Cole 1994). Several stud- stimulation of transcription by RNA Pol I. In summary, ies have indicated that large T antigen can regulate cell our findings suggest that large T antigen is recruited to growth, in part, by binding and inactivating the tumor the rRNA promoter by binding to the TBP/TAF I com- suppressor gene products pRB and p53 (Hinds and Wein- plex (i.e., SL1), and this protein-protein interaction plays berg 1994; Nevins 1994). In addition, large T antigen can a crucial role in activation of transcription. transactivate cellular genes, and several studies have indicated that this process is mediated by specific protein- protein interactions with multiple components of the Results transcriptional machinery (Gruda et al. 1993; Berger et T antigen can stimulate rRNA synthesis in a defined al. 1996; Damania and Alwine 1996; Johnston et al. RNA Pol I transcription system 1996). These alterations in the transcriptional activity of specific cellular genes most likely play an important role Early studies indicated that large T antigen can stimu- in the growth-promoting activity induced by large T an- late RNA Pol I transcription. To further establish these tigen. findings we tested whether large T antigen could acti- Transcription of the rRNA genes is necessary for cell vate the rRNA gene in transfection assays in HeLa cells. growth and proliferation; therefore, it is not surprising A human rRNA promoter construct fused to the CAT that the stimulation of Pol I transcription is one of the gene was cotransfected into HeLa cells with a plasmid strategies adopted by large T antigen to induce and sus- expressing either a sense [pCMV-LT (large T)] or an an- tain cell growth. Experiments with isolated nuclei and tisense [pCMV-asT (antisense T)] large T antigen. Figure cell-free extracts suggested that the stimulation of rRNA 1 shows the primer extension analysis of the RNA iso- synthesis occurred primarily at the transcriptional level lated from the transfected cells. Transfection of the pro- (Ide et al. 1977; Learned et al. 1983). In addition, in iso- moter construct plasmid produced a 123-nucleotide frag- lated nuclei, stimulation of rRNA synthesis was abol- ment, which corresponds to an RNA molecule initiated ished by pretreatment with T antigen antibodies, but not at the bona fide RNA Pol I transcription initiation site.

Figure 1. Transactivation of a ribosomal gene construct by cotransfection with a large T antigen-expressing vector. Transfec- tion were performed as described in Mate- rial and Methods, and Pol I-specific tran- scripts synthesized from the rRNA reporter prHu3CAT 2ug 6ug 10ug gene prHu3CAT were detected by primer extension. When not indicated 3 ~lg (lanes 1,2) and 6 ~ag (lanes 3,4), of either pCMV- asT or pCMV-LT were used, respectively. The specific transcript yielded a 123-nucleotide cDNA fragment. To account for losses during the processing of the RNAs, a 5' end-labeled single-stranded DNA oligo- mer (60 nucleotides) was added to each reaction. The primer extension data were quantified using a PhosphorImager (Mo- pCMV-LT + + 3ug 6ug 1 2ug lecular Dynamics), and activity was ex- pCMV-asT + + + pressed as fold induction relative to control transfection (pCMV-asT). fold induction 4.8x 4.7x 1.5x 2.6x 4.4x

1606 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

T antigen stimulation of Pol I transcription

Deletion of the Pol I promoter region completely abol- lied by either conventional or affinity chromatography. ished transcription from the reporter plasmid (data not These results established that large T antigen, when shown). The results of this experiment indicated that at added to HeLa nuclear extract, stimulated RNA Pol I each amount of reporter tested, the addition of large T transcription to levels comparable to the ones that have antigen-expressing plasmid resulted in approximately been reported in vivo (this work; May et al. 1976). fourfold activation of the rRNA gene. In addition, co- We next investigated the set of factors necessary for transfection of increasing amounts of pCMV-LT plasmid large T antigen stimulation of rRNA transcription. For resulted in proportionally higher stimulatory activity. this purpose, RNA Pol I, SL1, and UBF were purified Therefore, these data provide further evidence that large from HeLa nuclear extracts. As expected, purified SL 1, in T antigen can activate RNA Pol I transcription in human addition to Pol I and UBF, was required to reconstitute cells. transcription, both in the absence and the presence of To study in more detail the mechanism of RNA Pol I large T antigen (Fig. 2B, lanes 1-3). More importantly, stimulation by large T antigen, we then established a when purified T antigen was added to the transcription cell-flee transcription system that could mimic the in reaction, we observed a strong stimulation of rRNA syn- vivo process. As a first step, rRNA was transcribed using thesis, and the level of stimulation was proportional to HeLa cell nuclear extract in the absence or presence of the amount of recombinant large T antigen (lanes 4-8). recombinant large T antigen purified from Sf9 insect As in the previous experiment, a similar stimulatory ef- cells. Large T antigen protein was purified to apparent fect on rRNA synthesis was observed with large T anti- homogeneity using either conventional chromatography gen purified by either conventional or affinity chroma- or antibody-affinity chromatography (Clark et al. 1983; tography. Taken together, these results suggest that pu- Simanis and Lane 1985). As shown in Figure 2A, addition rified SL1, UBF, and RNA Pol I are necessary and of purified T antigen to a HeLa cell nuclear extract re- sufficient to support the large T antigen-mediated stimu- sulted in an increase in the level of rRNA synthesis lation of rRNA synthesis. (lanes 2,3). Quantitation of the in vitro-synthesized RNA We also tested whether another viral transactivator indicates that large T antigen can stimulate RNA Pol I protein, the adenovirus 13S-E1A protein, could also ac- transcription up to three- to fourfold. Identical results tivate Pol I transcription. When increasing amounts of were obtained using recombinant large T antigen puff- purified glutathione S-transferase (GST)-E 1A- 13 S fusion

A nuclear extract B fractionated system

- + - ,~-/~] T antigen --. ~ T antigen ----+ + + + + + SL1

Figure 2. Recombinant large T antigen stimulates RNA Pol I transcription. (A) Increasing amounts of recombinant large T antigen were tested in an in vitro transcription system with HeLa nuclear extract. In vitro-synthesized tran- scripts were detected by S1 nuclease assay. The transcription reaction contains either no T an- 1 2 3 12345678 tigen protein (lane 1) or increasing amounts of large T antigen (lanes 2 and 3; 25 and 50 ng of purified large T antigen, respectively). (B) Tran- C nuclear extract D fractionated system scription was performed using purified RNA Pol I, SL1, and UBF either in the absence (lanes EIA 1,3) or in the presence (lanes 2,4-8; 100 and E1A 50-250 ng of purified large T antigen, respec- ----.++++ SL1 tively) of increasing amounts of large T antigen. Experimental conditions were as in A. In lanes 1 and 2, the SL 1 fraction was omitted from the transcription reaction. (C) Transcription was performed with HeLa nuclear extract and with increasing amounts of purified GST-E1A-13S, as described in A. (D) Transcription was performed as in B except that purified GST-E1A- 13S, instead of large T antigen, was added to the transcription reactions (lanes 1 and 3; no E1A; lanes 2,4-6, 150 and 50-150 ng of E1A, respec- 1 2 3 4 1 2 3 4 5 6 tively).

GENES & DEVELOPMENT 1607 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Zhai et al. protein were added to either a crude nuclear extract or a (lane 6), and TBP (lane 9) in the immunoprecipitation reconstituted transcription system with partially puri- product. On the other hand, SL1 was not immunopre- fied Pol I, UBF, and SL1, we did not observe any detect- cipitated from mock-infected HeLa extracts, indicating able change in the level of RNA Pol I transcription (Fig. that the monoclonal antibody used in this study did not 2C,D). Therefore, we conclude that the ability to stimu- cross-react with any of the SL1 subunits. In addition, late RNA Pol I transcription is an intrinsic property of UBF did not coimmunoprecipitate with large T antigen, SV40 large T antigen. indicating that these two proteins did not interact with each other (lane 12). Thus, our results indicate that large T antigen can associate with the SL1 complex in vivo Large T antigen associates with the SL1 complex in and provide the first evidence that a protein-protein in- SV40-infected cells teraction may play a critical role in the stimulatory pro- The finding that the stimulatory effect of T antigen on cess. rRNA synthesis could be reconstituted with purified To further establish the specificity of this protein-pro- RNA Pol I, SL1, and UBF strongly suggested that this tein interaction, the nuclear extracts from the SV40-in- regulatory process was primarily mediated at the level of fected cells were fractionated by heparin-agarose chro- transcription. Thus, one possibility would be that large T matography. Aliquots of the flowthrough, washes, and antigen is recruited to the rRNA promoter through pro- eluted fractions were then separated on an SDS-poly- tein-DNA or protein-protein interactions. The ability of acrylamide gel and the presence of large T antigen was T antigen to interact specifically with sequences in the determined by Western blot analysis. Individual frac- rRNA promoter has been tested by us and others. Foot- tions from the column were also assayed for SL 1 activity printing and DNA-binding assays indicated that there in a reconstituted transcription assay with partially pu- are no specific binding sites for T antigen in the pro- rified RNA Pol I and UBF (data not shown). As shown in moter region of the human rRNA gene (Learned et al. Figure 4A, large T antigen was found in the flowthrough 1983; L. Comai, unpubl.). Therefore, we postulated that (lane 2), which also includes the TFIID fraction (Tanese large T antigen may interact with one or more of the Pol et al. 1991; Comai et al. 1992; data not shown). In addi- I transcription factors. To determine whether large T an- tion, a significant portion of large T antigen eluted at tigen was associated with SL1 and/or UBF in vivo, HeLa higher salts and copurified with the SL1 activity (lane 4). cells were infected with SV40 at a m.o.i, of 20 PFU/cell. To determine whether the subpopulation of large T Forty-eight hours after infection, the cells were har- antigen that eluted at high salts from the heparin-aga- vested and nuclear extracts were prepared. The nuclear rose was associated with the SL1 complex, this fraction extracts from the infected cells and from mock-infected was incubated with monoclonal antibodies raised cells were then incubated with monoclonal antibodies against large T antigen and the resulting complex was raised against large T antigen, and the resulting immu- resolved by SDS-PAGE. After transfer to nitrocellulose, nocomplex was precipitated with protein A-Sepharose the blots were probed with antibody against each of the beads. The immunoprecipitation products were then SL1 subunits. As shown in Figure 4B, the SL1 complex analyzed by SDS-PAGE and subsequently immunoblot- (TBP and TAFs) from the SV40-infected cells (lanes ted with antibodies against TBP, TAFIS, and UBF (Fig. 3). 2,5,8), but not from the mock-infected cells (lanes 3,6,9), The results of this experiment revealed that in SV40- coprecipitated with large T antigen. Moreover, because infected cells, large T antigen was associated with SL1, DNase I was added to the extract prior to and during the as indicated by the presence of TAFI110 (lane 3), TAFt63 immunoprecipitation step, the association of large T an-

Figure 3. Coimmunoprecipitation of SL1 and large T antigen from SV40-infected cells. Nuclear extracts from SV40-infected HeLa cell (lanes 3,6,9,12) and mock-in- nxt nxt nxt nxt fected HeLa cells (lanes 2,5,8,11) were im- ...... ! ! munoprecipitated with monoclonal anti- ! i ! .,~..,~. i ! .~- .,~- i ~-~. bodies against large T antigen. The immunoprecipitation products were separated MW by SDS-PAGE and transferred to nitrocellulose. The Western blots were then 104 -,-, probed with antibodies against TAFIll0 82 ,~, (lanes I-3), TAFI63 (lanes 4-6), TBP (lanes 7-9), and UBF (lanes 10-12). Purified SL1 fraction (lanes 1,4,7) and UBF (lane 10) 48 "- were used as positive controls. The arrows indicate the positions of each protein. The 33 --" cross-reacting band at -60 kD present in 1 2 3 4 S 6 7 8 9 10 11 12 all the immunoprecipitation reactions rep- resents the IgG heavy chain. a TAFI110 a TAFI63 a TBP a UBF

1608 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

T antigen stimulation of Pol I transcription

Figure 4. (A) A subpopulation of SV40 A Heparin-Agarose C large T antigen copurifies with the SL1 MW @ @~ complex. Crude nuclear extracts prepared SLI + - - + +/- from SV40-infected HeLa cells in TMI°/ 104 1 82 = 0.4 M KC1 were loaded onto a heparin-a- garose column. After extensive washes with TMI°/0.4 M KCI buffer, the SL1 com- 48~ 123 4 5 6 7 8 9 plex was eluted with TMI°/0.8 M KC1. The eluted fractions (lanes 4-9), one flow- i FT W El ,~E6 33~ 1 2 3 through fraction (lane 2), one fraction from T antigen the wash (lane 3), and an aliquot of the T antigen original nuclear extract (lane 1) were resolved on an SDS-polyacrylamide gel and B .,4" .,4" -,4".,4" analyzed by Western blot using monoclonal antibodies against large T antigen. The SL1 activity of each fraction was assessed using in vitro transcription assays as de- MW scribed previously. (B) Coimmunoprecipi- tation of large T antigen and SL1. SL1 was 105 1 purified from nuclear extracts prepared 79= 7 from either SV40-infected or mock-in- 491 ~ fected HeLa cells as described above. The 331 SL1 fractions were then dialyzed to 0.1 M KC1 in TM 1° buffer and used in immunoprecipitation reactions with monoclonal 1 2 3 4 5 6 7 8 9 antibody pAb 101 (anti-large T) cross- TAFI110 ~ TAFI63 ~ TBP linked to protein A-agarose beads. The resulting immune complex was split into three aliquots, which were then resolved by SDS-PAGE and analyzed by Western blotting using antibodies against TAFt110 (lanes 1-3), TAFt63 (lanes 4-6), and TBP (lanes 7-9). Arrows indicate the position of each protein. Immunoprecipitation from either SV40-infected cells (lanes 2,5,8) or mock-infected cells (lanes 3,6,9) are as indicated. An aliquot of the SL1 fraction (lanes 1,4, 7) was used as a positive control in the Western blot analysis. The asterisk indicates cross-reactivity with the antibody heavy chain. (C) An SL1 fraction prepared from SV40-infected cells was incubated with either affinity-purified TBP antibodies (lane 1) or affinity-purified GST antibodies (negative control, lane 2). The immunoprecipitation products were resolved by SDS-PAGE and analyzed by Western blotting using anti-large T antigen antibodies. Purified recombinant large T antigen was loaded in lane 3.

tigen with SL 1 was likely attributable to protein-protein complexes were resolved by SDS-PAGE and analyzed by interactions rather than to DNA tethering. The interac- autoradiography. Input and bound radiolabeled proteins tion of SL1 with large T antigen was verified by perform- were also quantified using a PhosphorImager. As shown ing the reverse immunoprecipitation. Affinity purified in Figure 5A, TAF~ll0, TAFt48, and TBP bound effi- antibodies raised against TBP were used to precipitate ciently to large T antigen (lanes 4,12,16). On the other SL1 from the high salt heparine-agarose eluate, and the hand, no significant interaction was observed with resulting product was analyzed by Western blotting with TAFt63 (lane 8). To ensure that the interactions were not large T antigen antibodies (Fig. 4C). These results con- mediated by nucleic acids, in a parallel experiment, firmed that large T antigen coimmunoprecipitated with ethidium bromide (200 pg/ml) was added to each binding SL1. In summary, our data indicate that upon SV40 in- reaction without any detectable effect on the results fection, a subpopulation of large T antigen binds to the (data not shown). In addition, we also showed that each RNA Pol I factor SL1. This strongly suggests that SL1 of the TAF~s interacted with GST-TBP (lanes 3,7,11) and plays a critical role in the recruitment of large T antigen that TBP bound to Flag-tagged TAFI48 (lane 15), as re- to the rRNA promoter. ported previously (Comai et al. 1994). As determined in the in vivo experiments, there was no significant interaction between large T antigen and UBF (lane 20), which, Interactions between SV40 large T antigen and TBP-TAFIs as shown previously, could bind efficiently to GST-TBP (lane 19; Beckmann et al. 1995). The results of these To better understand the association between large T protein-protein interaction assays were confirmed in antigen and SL1, we then proceeded to identify which coinfection experiments in Sf9 cells using recombinant subunit of SL1 makes direct contact with large T anti- baculoviruses expressing large T antigen, TBP, UBF, and gen. In vitro-translated [3SS]methionine-labeled TAFI48 , TAFIs (W. Zhai and L. Comai, unpubl.). TAFt63, TAF~110, and TBP were incubated with a GST- Although the previous experiments suggested that large T antigen fusion protein immobilized on affinity TAF~110, TAFI48, and TBP were important for the asso- resin. After extensive washing, the resulting protein ciation of large T antigen with SL1, we did not defini-

GENES & DEVELOPMENT 1609 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Zhai et al.

Figure 5. (A) Protein-protein interactions among large T antigen and TBP, TAFzs, and A UBF. TAFIll0, TAFI63, TAFI48, TBP, and UBF were in vitro translated and labeled MW i~ ~ i i i with [3SS]methionine. GST (lanes 2,6,10, 14,18), GST-large T antigen (lanes 4,8,12, 97 m 66 ~,. 16,20) and Flag-tagged TAFI48 (lane 15) were expressed in Sf9 cells using recombi- 45 --. nant baculoviruses. GST-TBP (lanes 3,7, I1,19) was expressed and purified from E. 31 -,. coli. GST and GST-fusion proteins were 1234 5 6 7 8 9 10 11 12 13 14 tS 16 17 18 19 Z0 immobilized on glutathione-Sepharose resin (Pharmacia), Flag-tagged TAFI48 was TAFI110 TAFI63 TAFI48 TBP UBF immobilized on anti-Flag M2 beads (Ko- dak), and subsequently incubated with the radiolabeled proteins. After several washes, B the bound complexes were analyzed by SDS-PAGE and autoradiography. Ten per- MW cent of the total 3SS-labeled protein used for each binding assay was loaded in the input i t T i t T i t T lanes (lanes labeled i). (B) Flag-tagged TAFs, 97 GST-large T antigen (GST-WT-T) and 66 GST-small t antigen (GST-small-t) were purified as described in Materials and 45 ~" // i- i- i- Methods. ( • ) The corresponding purified 31 o~ TAFI63 a TAFI110 a TAFI48 protein. The identity of each of the TAFs 1 2 3 4 5 was verified by Western blot analysis. High molecular weight bands in lanes I and 2 are Silver Stain Western Blot Sf9 contaminant protein, which copurify with TAFI48 and TAFt63 on DEAE-Sepharose and anti-Flag M2 affinity gel (see Materials and Methods). Some of the lower molecular weight bands in the TAFt110 lane cross-react with the antibodies and most likely represent TAF~110 breakdown products. The purified GST-large T and GST-small t antigen were mixed with purified TAFs as described in the text, and after several hours of incubation at 4°C, the complexes were collected by precipitation with glutathione-Sepharose beads. The bound products were resolved by SDS-PAGE and transferred to nitrocellulose, and blots were probed with antibodies against the individual TAFs. Lanes i show 20% of the TATs used in each reaction. Lanes t and T represent small t antigen and large T antigen binding reactions, respectively. tively establish that these interactions were mediated TAFlllO, TAFt48, and TBP interact with the solely by the two components being tested. To provide amino-terminal region of large T antigen more evidence for a direct protein-protein interaction, we purified recombinant Flag-tagged TAFI48 , TAFt63, Several of the biochemical activities of large T antigen and TAF,110 from Sf9 insect cells using DEAE-Sepha- have been assigned to discrete functional domains using rose and Flag affinity chromatography (Fig. 5B, lanes deletion and point mutant analysis (Fanning 1992). To 1-3). Similarly, large T antigen-and small t antigen-GST identify the regions of T antigen involved in the inter- fusion proteins were expressed in Sf9 insect cells and action with TBP and TAFs, we tested a set of T antigen purified on glutathione-Sepharose resin. (Fig. 5B, lanes deletion mutants in a series of protein-binding assays. 4,5). Purified GST-large T and GST-small t antigen were The large T antigen mutants used in this study, contain- mixed, individually, with a threefold molar excess of ing deletion at either the carboxyl terminus (TN1-3), the TAFI48, TAFI63, or TAF~ll0 and incubated at 4°C for amino terminus (TC 1-3), or both termini (TM1), are rep- several hours. Large T and small t antigen were then resented schematically in Figure 6A. The large T antigen precipitated from the respective reaction mixture using mutants were cloned into baculovirus expression vectors glutathione-Sepharose beads, and the precipitates were and transfected into Sf9 cells for the production of re- analyzed by Western blotting using antibodies against combinant viruses. Recombinant proteins were ex- the TAFs. As shown in Figure 5B, TAF,48 and TAFI110, pressed as GST fusions and affinity purified using gluta- but not TAFI63 , were coprecipitated with GST-large T thione-Sepharose resin. A typical silver-stained SDS- antigen (lanes T). On the other hand, none of the TAFs polyacrylamide gel of the various fusion proteins eluted associated with GST-small t (lanes t). Our laboratory from GST affinity beads with 20 mM reduced glutathione and others have also demonstrated a direct interaction is shown in Figure 6B. For the protein-protein interac- between TBP and large T antigen (Gruda et al. 1993; tion assays, TAFt48, TAFt63, TAFIll0 , and TBP were Johnston et al. 1996; data not shown). In summary, our translated in vitro and labeled with [3SS]methionine. The results indicate that the association of large T antigen radiolabeled proteins were then incubated with GST, with SL1 is mediated by specific protein-protein con- GST-large T antigen (GST-WT-T), each of the GST- tacts with TAFI48 , TAFI110, and TBP. large T antigen mutants, and GST-small t antigen bound

1610 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

T antigen stimulation of Pol I transcription

region from amino acid 362 to 538), TC1, and TC3 A (amino-terminal deletion mutants), showed no signifi- i 708 cant interaction with each of the TAFIS and TBP. Inter- WT-T: 1-708 NH2 I COOH estingly, small t antigen did not bind to TBP or TAF~s. TN1:1-538 Because large T and small t antigen share the amino- TN2:1-436 terminal 82 amino acids, our results suggest that this TN3:1-361 region of large T is not directly involved in the interactions with TBP and the TAF~s. As discussed previously, TC1:362-708 large T antigen did not bind to TAFt63. TC2:411-708 I ' In summary, these experiments indicate that the in- TC3:538-708 teractions with TAFt110, TAFI48, and TBP are mediated TM1:362-538 by the amino terminus of large T antigen. In addition, the different affinities of TAFIS and TBP to the individual B deletion mutants suggest, most likely, that distinct surfaces within the amino-terminal region of large T antigen mediate the interaction with different SL1 subunits. MW 1 2 3 4 5 6 7 8 9 Additional binding experiments between this set of large T antigen mutants and an integral SL1 complex con- 97~ 66--- firmed that the amino-terminal 436 amino acids of large T antigen were required for efficient binding to SL1 (data 45m not shown).

31 m The amino-terminal portion of large T antigen is also 21m required for stimulation of Pol I transcription Figure 6. (A) Schematic representation of recombinant large T The previous protein-protein interaction studies sug- antigen (WT-T) and large T antigen mutants (TN1-3; TC1-3; gested, most likely, that large T antigen was brought to TM1). Mutants representing an amino-terminal (TN1-3) and a the rRNA promoter via interactions with three subunits carboxy-terminal (TC1-3) deletion, or the middle region of large T antigen (TM1) were generated by deleting the carboxy-termi- of the SL1 complex. However, the role of these interac- nal, amino-terminal, or both ends from wild-type T antigen, as tions in the stimulatory process remained unclear. To described in Materials and Methods. The amino acids included establish the functional significance of large T antigen- in each GST-fusion are indicated next to the name of the con- SL1 interaction and to identify regions of large T antigen struct. (B) Silver-stained SDS-polyacrylamide gel showing wild- involved in the stimulation of rRNA synthesis, the type and mutant forms of GST-large T antigen eluted from aforementioned set of mutants was tested in the in vitro glutathione-Sepharose resin. All proteins were expressed in Sf9 transcription assays. GST fusion proteins were purified cells using recombinant baculovirus as described in Materials by affinity chromatography using glutathione-Sepharose and Methods. Proteins shown are GST (lane 1), GST-WT-T resin and used in transcription reactions with purified (lane 2), GST-TN1 (lane 3), GST-TN2 (lane 4), GST-TC2 (lane RNA Pol I, SL1, and UBF. As shown in Figure 8, addition 5), GST-TN3 (lane 6), GST-TC1 (lane 7), GST-TC3 (lane 8), and GST-TM1 (lane 9). Molecular weight markers are indicated. of increasing amounts of wild-type large T antigen stimulates rRNA synthesis (lanes 6-8), as compared to a reaction that contained no large T antigen (lane 5) or GST alone (lanes 2-4). Moreover, GST-TN1, which con- to glutathione-Sepharose affinity resin. All of the reac- tains the first 538 amino acids of large T antigen, re- tions were carried out in the presence of 200 pg/ml of tained the ability to activate rRNA synthesis as effi- ethidium bromide. After several hours of incubation at ciently as the wild-type protein (lanes 10-12). However, 4°C, the beads were washed extensively and the bound further deletion of the carboxy-terminal region (TN2, material was resolved by SDS-PAGE. The amount of ra- TN3) caused a sharp decrease in the stimulation of RNA diolabeled protein was then quantified using a Phospho- Pol I transcription (lanes 14-16 and 18-20). Addition of rImager and plotted as the percentage of bound protein large T antigen amino-terminal mutants (TC1, TC2, relative to the protein input. As shown in Figure 7, large TC3), a mutant containing the internal region from T antigen mutant TN1, which contains the amino-ter- amino acid 362 to 538 (TM1), or small t antigen did not minal 1-538 amino acids, bound as efficiently as the cause any detectable stimulation of transcription (lanes wild-type protein to TAFIll0 , TAFt48, and TBP. When 22-24, 26-28, 30-32, 34-36, and 38-40, respectively). the amino-terminal region was further deleted (TN2), These results suggested that the amino-terminal portion there was a significant decrease in the binding to TBP, of large T antigen, from amino acid 1 to 538, has an whereas binding to TAFI110 was not affected and bind- important role in the stimulation of Pol I transcription. ing to TAFI48 was diminished only slightly. The re- The observation that a shorter portion of large T antigen, moval of an additional 75 amino acids (TN3) virtually which contains the first 361 amino acids, did not show abolished the binding to the TAFs and TBP. The other any Pol I stimulatory activity, suggesting that the region large T antigen mutants, TM1 (containing the internal between amino acids 361 and 538 was necessary for this

GENES & DEVELOPMENT 1611 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Zhai et al.

25 Figure 7. Interactions of large T antigen A mutants with TAFIll0 , TAFI63 , TAFI48 , • TAFI110 and TBP. GST, GST-large T antigen, GST- 20 large T antigen mutants, and GST-small t [] TA FI48 antigen were expressed in Sf9 cells using re- [] TAFI63 combinant baculovirus and immobilized on -= glutathione-affinity resin (see Fig. 6 for no- ¢: [] TBP menclature). Radiolabeled TAF~110, TAFt63, I TAFI48 , and TBP were incubated with the various GST fusion proteins, and after washing the beads extensively, the result- 5 ing protein complexes were resolved by SDS-PAGE. Radiolabeled bound proteins were quantified using a PhosphorImager and plotted as percentage bound (%) protein relative to the input for each reaction. The type of bar representing each factor is shown on the side legend. Numbers next to each large T antigen mutant refer to the amino acid residues delimiting each mutant. Large T antigen Mutants

regulatory process. However, because the mutant TM1 results demonstrate further that the association of large T (amino acids 362-538) was transcriptionally inactive, antigen with SL1 plays an important role in this process, as this portion of large T antigen was clearly not sufficient deletion mutants TC 1, TC2, TC3, and TM 1, which did not to enhance Pol I activity (lanes 34-36). Moreover, these bind to SL1, were unable to stimulate Pol I transcription.

GST GST-WT-T GST-TN1 GST-TN2 GST-TN3 Figure 8. Transcriptional stimulation in vitro of RNA Pol I by large T antigen mutant proteins. Wild-type or mutant forms of large T antigen and small t antigen were tested in an in vitro transcription system with partially purified RNA Pol I, UBF, and II SL1. GST-fusion proteins were expressed in Sf9 cells using recombinant baculovirus and purified using glutathione-Sepharose resin as described in Material and Methods. In- creasing amounts of each protein (from -0.15 to 1.2 pmoles), as judged from silver- 1 2 3 4 5 6 7 8 9 10 11 12 i13 14 15 16 17 18 19 20 stained gels of the purified proteins, were added to the respective reaction mixtures. Transcription reactions were carried out in the presence of 100 ng of DNA plasmid G ST-TC 1 G ST-TC 2 G ST-TC 3 GST-TM1 GST-small-t prHu3, and the in vitro-synthesized tran- scripts were detected by S 1 nuclease protec- tion assay as described previously (Bell et al. 1990; Comai et al. 1992). The transcription reaction contained either no added protein (lanes 1,5,9,13,17,21,25,29,33,37), GST pro- ..~ tein only (lanes 2-4), GST-large T antigen (lanes 6-8), GST-large T antigen mutant proteins (lanes 10-12, 14-16, 18-20, 22-24, 26-28, 30-32, 34-36) or GST-small t antigen (lanes 38-40). Arrows indicate the positions of the protected oligonucleotide frag- ments. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

1612 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

T antigen stimulation of Pol I transcription

Discussion activity. Moreover, analysis of large T antigen deletion mutants indicated that the amino acid region of large T Transcription of the rRNA genes by RNA Pol I is an antigen between residues 1 and 538 was necessary for essential process for cell homeostasis, growth, and pro- full stimulation of Pol I transcription. A weak transcrip- liferation (Jacob 1995; Moss and Stefanovsky 1995). A tional stimulatory activity was retained by a T antigen variety of extracellular stimuli have been shown to mutant comprising amino acids 1-436. Further deletion modulate the rate of rRNA synthesis (Jacob 1995). Fol- into the amino terminus completely abolished the abil- lowing viral infection, viruses such as SV40 up-regulate ity of large T antigen to stimulate Pol I transcription. the expression of the rRNA genes, and early studies have This analysis suggests that the region between amino identified large T antigen as the viral protein primarily acids 361 and 538 is important for the stimulation of responsible for the stimulatory effect (Ide et al. 1977; rRNA synthesis. This region of large T antigen overlaps Learned et al. 1983). Although well documented, rela- with the ATPase domain and has also been shown to be tively little was known about the mechanisms underly- involved in p53 binding (Fanning 1992; Conzen and Cole ing this process. Most of these studies have been ham- 1994)(Fig. 9). By binding to p53, large T antigen inacti- pered by the lack of a well-defined RNA Pol I transcrip- vates the antiproliferative properties of this tumor sup- tion system. With the molecular cloning of two essential pressor protein (Lin and Simmons 1991; Mietz et al. Pol I transcription factors, UBF and the multiprotein 1992). Thus, it is quite interesting that an overlapping complex SL1, we can now address more directly how domain is also required for up-regulation of Pol I tran- viruses affect Pol I transcription and elucidate the mo- scription, and it suggests that this region is important for lecular mechanisms of action. the proliferative program initiated by large T antigen. In this report we focused on the biochemical analysis The amino-terminal 362 amino acids of large T anti- of the stimulation of RNA Pol I transcription by the gen, in addition to amino acids 362-538, were also nec- DNA tumor virus SV40. Using an in vitro transcription essary to reconstitute full stimulation of RNA Pol I tran- system, we have reconstituted this process using recom- scription. Protein-protein interaction studies indicated binant large T antigen and purified Pol I transcription that large T antigen amino acids 1-436 are required for factors. In vitro, recombinant large T antigen stimulated efficient binding to TAFt48, TAFIll0 , and TBP. These RNA Pol I transcription up to fivefold, which is similar results provide strong evidence that the interaction be- to what has been observed in vivo (this work; May et al. tween large T antigen and SL1 plays a critical role in the 1976; Pockl and Wintersburger 1980). These results sug- stimulatory process. On the other hand, the region be- gest that the in vitro system faithfully reproduces the tween amino acids 436 and 538 was also necessary for transcriptional activation function of large T antigen and efficient stimulation of Pol I transcription, suggesting therefore is an important tool for the identification and that the recruitment of large T antigen to the rRNA pro- biochemical analysis of the cellular proteins that medi- moter is most likely the initial step in the activation ate this process. process, and other biochemical modifications and/or Protein-protein interaction and immunoprecipitation changes, mediated by large T antigen, are also required. experiments showed that large T antigen binds to the Therefore, it remains to be determined how the binding selectivity factor SL1, both in vivo and in vitro. This of large T antigen to SL1 influences the network of pro- finding provides the first evidence that a viral protein tein-protein and protein-DNA interactions between such as large T antigen, through direct protein-protein SL1, UBF, the RNA Pol I holoenzyme, and the rRNA interactions with SL1, modulates Pol I transcriptional promoter.

Helicase/Origin Unwinding 131 I II 627

Rb/pl07 p53 Binding

102 115 272 517

Pol c~ DNA Binding ATPase/ATP Binding IIIII 1 82 131 259 418 627 Figure 9. Schematic representation of SV40 large T antigen functional domains. I (Top) Localization of the sequences re- 708 quired for the various activities, as defined SL1 binding by the analysis of a variety of large T anti- 1 436 gen mutants (Fanning 1992; Conzen and Cole 1994). (Bottom) Region of large T an- Stimulation of RNA pol I Transcription tigen required for SL1 binding and stimulation of Pol I transcription (as described in 538 this work).

GENES & DEVELOPMENT 1613 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Zhai et al.

Interestingly, our experimental results correlate well structural studies of TFIID and SL1 should provide the with previous work on the reactivation of rRNA genes in necessary framework to address these issues. mouse-human hybrid cells (Soprano et al. 1979, 1981, 1983). In these studies, wild-type and mutant large T antigens were either microinjected or transfected into a Materials and methods mouse-human hybrid cell line that contained both Plasmids and DNA constructs mouse and human rRNA genes, but only expressed the Large T antigen eDNA was released from pNT2 (Lanford 1988) mouse rRNA. When expressed in these cells, large T an- by cleavage with BamHI and ligated into BamHI-digested tigen could reactivate the human rRNA gene(s). In addi- pBluescriptIIKS(+) in T7 orientation. A 2.3-kb HindIII partially tion, this study indicated that T antigen polypeptides digested fragment released from this plasmid was then ligated extending from the amino terminus to amino acid 509 or into pBluescriptIIKS(+)/HindIII to generate pBSKS-T(ATAG) further, were active in rRNA reactivation, whereas those with T antigen in T7 orientation. Baculovirus-expressing vec- shorter than 420 amino acids were inactive (Soprano et tors pVL2T and pVL3X were generated by religating SmaI/ al. 1983). Thus, it appears that the region of large T an- EcoRI-digested pVL1393 (Pharmingen) with a 0.75-kb HincII- tigen, which is involved in the stimulation of rRNA syn- EcoRI fragment released from pGEX2T and pGEX3X (Pharma- cia), respectively, pVL3X-WT-T was generated by ligating a 2.3- thesis in vitro, is also necessary for reactivation of hu- kb EcoRI fragment released from pBSKS-T(ATAG) into pVL3X/ man rRNA genes in the mouse-human hybrid cells. A EcoRI. pVL3X-TN1 was generated by ligating a 1.7-kb EcoRI- question is then raised by the observation that the hu- PstI fragment from pBSKS-T(ATAG) into PstI/EcoRI-digested man species-specific factor SL1 was either inactivated or pVL3X, pVL2T-TN2 was constructed by ligating a 1.4-kb Sinai~ absent in three of these hybrid cell lines (Miesfeld et al. PvuII fragment from pBSKS-T(ATAG) into SmaI digested 1984). It is possible that large T antigen, by binding to pVL2T, pVL2T-TN3 was generated by ligating a 1.1-kb SmaI- human or mouse SL1, can either reactivate the human HincII fragment from pBSKS-T(ATAG) into SmaI-digested factor or reprogram the mouse SL1 to transcribe the hu- pVL2T, pVL2T-TC1 was generated by ligating a 1.3-kb SmaI- man rRNA gene(s). To this end, using recombinant large HincII fragment from pBSKS-T(ATAG) into Sinai-digested T antigen in a cell-free system, we have not been able to pVL2T, pBSKS-TC2 was constructed by ligating a 1.0-kb NsiI- EcoRI fragment from pBSKS-T(ATAG) into PstI/EcoRI-digested reprogram a mouse nuclear extract to transcribe the hu- pBluscriptIIKS(+), pVL2T-TC2 was generated by ligating a 1.0- man rRNA gene. The cloning of the mouse SL1 subunits kb Sinai fragment from pBSKS-TC2 into SmaI-digested pVL2T. and a more detailed characterization of the mouse-hu- pVL2T-TC3 was generated by ligating a 0.7-kb PstI fragment man hybrid cell lines will be necessary to further vali- from pBSKS-T(ATAG) into PstI-digested pVL2T, pVL2T-TM1 date our hypothesis. was generated by self-ligating the bigger fragment from PstI- Recently, it has been shown that T antigen activation digested pVL2T-TC1. For small t antigen constructs, small t of specific genes transcribed by RNA Pol II is mediated cDNA was released from pSV010, a plasmid containing the by specific protein-protein interactions with the tran- SV40 genome. A 1.9-kb PvuII fragment released from pSV010 scription factor TFIID (Gruda et al. 1993; Damania and (gift from B. Stillman, Cold Spring Harbor Laboratory) was di- Alwine 1996; Johnston et al. 1996). Both TFIID and SL1 gested with HindIII to produce a 1.1-kb fragment containing small t eDNA. The fragment containing small t cDNA was are multiprotein complexes comprised of TBP and poly- ligated into HindIII digested pBluescriptIIKS(+) to generate pB- merase-specific TAFs (Hernandez 1993; Roeder 1996; SKS-small t. pVL2T-small t was generated by ligating a 1.2-kb Verrijzer and Tjian 1996). Thus, we postulate that the SmaI-HincII fragment from pBSKS-small t into SmaI-digested TBP/TAF complexes could serve as the common cellular pVL2T. The plasmids pCMV-LT and pCMV-asT were generated target used by small DNA viruses such as SV40 to modu- by insertion of BamHI large T antigen DNA fragment into late the transcriptional activity of RNA Pol I, Pol II, and, pcDNA3 (Invitrogen), in both orientations. possibly, Pol III genes. Whether or not the interactions of Plasmid prHu3 containing the human rRNA promoter and large T antigen with TBP-TAFs are conserved between the plasmids containing the human TAFIs and TBP have been the TFIID and SL1 complex is currently unclear. We described previously (Bell et al. 1988; Comai et al. 1992, 1994). have shown that large T antigen, in addition to binding prHu3CAT was constructed by inserting the human rDNA promoter (-500 to +78) into pBSMCAT (CAT eDNA inserted in to TBP, makes direct contacts with TAFI48 and TAFI110. pBS; gift of Dr. S. Tahara, University of Southern California, Los The study by Damania and Alwine (1996) indicated that Angeles) at the StuI site. The contruct was sequenced to ensure TBP, TAFII150, TAFHll0 , and TAFII40 are involved in that the promoter was placed in the correct orientation. the interactions between TFIID and large T antigen. However, in both studies, the specific protein-protein Ceil culture, transfections, and SV40 infections interaction domains in TBP and TAFs have not been mapped yet. TBP appears to be the only common target HeLa cells (strain $3) were propagated and maintained in mini- in the two complexes. Lack of homology between any of mal essential medium (MEM) supplemented with 5 % newborn the Pol I- and Pol II-specific TAFs suggests that each of calf serum at 37°C in 5% CO2. Cells were kept in a spinner at a density of 2 x l0 S to 5 x l0 s cells/ml. For infection with SV40, the TAFs in SL1 and TFIID interacts with large T antigen 1 liter of HeLa cells was collected by spinning at 1000g for 5 min through different surfaces or that large T antigen recog- and cell pellets were resuspended in 10 ml of MEM and 10 ml of nizes a tertiary structure that is conserved between SL1 SV40 virus. Infection was done at a m.o.i of 20 PFU/cell and and TFIID. Therefore, in future work, it will be impor- incubated for 2 hr at 37°C with periodic rocking. Then fresh tant to define the contact points between large T antigen medium was added and cells were incubated for 48 hr at 37°C. and TAFs in the TFIID and SL1 complexes. Ultimately, Infected and mock-infected cells were harvested, and nuclear

1614 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

T antigen stimulation of Pol I transcription extracts were prepared as described in Comai et al. (1992). The 0.5% NP-40 followed by two times with elution buffer [100 mM protein concentration of the extracts was determined by the Tris (pH 8.0), 50 mM NaC1, 0.1% NP-40]. GST-fusion proteins Bradford assay. were eluted from the beads using 20 mM reduced glutathione For the transfection assays, HeLa cells were maintained in dissolved in elution buffer. Eluted proteins were dialyzed DMEM supplemented with 10% FBS at 37°C with 5% CO2. against TMI°/0.1 M KC1. The amount of eluted protein was Cells were seeded at 7.5 x 105 per 100-mm plate 1 day prior to estimated by Coomassie and silver staining of SDS-polyacryl- transfection. Cells were transfected by the calcium phosphate amide gels. method with the indicated amounts of plasmid DNA. Filler DNA (pBS) was used to normalize the amount of DNA used per Purification of Flag-tagged TAFs Flag-tagged TAFI110, TAFI63, transfection. Cells were harvested 55-60 hr after transfection, and TAFI48 were expressed individually in insect cells as de- and total RNA was isolated using the guanidinium-thiocynate scribed previously (Comai et al. 1994). Each one of the TAFs was method. Thirty micrograms of total RNA was then used in a puirified from the respective cytoplasmic fraction as follows. standard primer extension assay with a CAT-specific primer. Cytoplasmic extracts were prepared by lysing the cells in buffer The cDNA product was resolved on an 8 % polyacrylamide-urea H [10 mM Tris-HC1 (pH 7.9), 10 mM KC1, 1.5 mM MgC12, 1 mM gel and subjected to autoradiography. Quantitation was per- DTT] using a Dounce homogenizer. The extract was then ad- formed using a PhosphorImanger (Molecular Dynamics). A 60- justed to 0.1 M KC1, loaded on a DEAE-Sepharose column, and nucleotide 32p-end-labeled oligonucletide was included in all of step-eluted with TM buffer containing 0.3, 0.5, 0.7, and 1.0 M the reaction mixtures to account for sample losses during the KC1. TAFI63 and TAFI48 eluted at 0.3 M KC1, and TAFIll0 experimental procedure. eluted at 0.5 M KC1. Fractions containing TAF were then loaded Sf9 cells were propagated and maintained in Hink's TNM-FH onto an anti-Flag M2-affinity resin (Kodak), washed extensively insect medium (JRH Biosciences) supplemented with 10% heat- with TM buffer containing 0.4 M KC1, and the TAFs were eluted inactivated fetal bovine serum at 28°C. To generate recombi- with TM buffer/0.1 M KC1 containing 0.4 mg/ml of Flag pep- nant baculoviruses, Sf9 cells were seeded on a six-well tissue tide. The eluted proteins were then dialyzed in TM/0.1 M KC1, culture plate to -80% confluency and allowed to attach for 30 and the concentration of TAFs was estimated from silver- min. Cells were then transfected with 1.3 ~g of baculovirus stained SDS-polyacrylamide gels. All of the buffers were supple- expression construct together with 0.17 ~g of BaculoGold DNA mented with 1 mM PMSF, 1 mM sodium metabisulfite, 5 ~g/ml according to the manufacturer's instructions (PharMingen). Re- of leupeptin, and 5 ~g/ml of aprotinin. combinant viruses were harvested 6 days later and amplified. For virus amplification, Sf9 cells were infected at a m.o.i of 0.1 Purification of GST-E1A-13S GST fusion protein was ex- and virus was harvested 4 days later. For protein expression, Sf9 pressed in Escherichia coli and affinity purified on glutathione- cells were infected at a m.o.i of 5 and harvested 42-48 hr after Sepharose beads as described above. For transcription assays, infection. For coinfection experiments, Sf9 cells were infected from 50 to 150 ng of purified protein was added to each reaction, with two viruses, each at a m.o.i, of 10. as estimated by silver-stained SDS-polyacrylamide gels.

Purification of recombinant proteins In vitro transcription assay In vitro transcription reactions were carried out as de- Large T antigen Conventional chromatography. Lysates from scribed previously in the presence of 100 lug/ml of c~-am- Sf9 cells infected with recombinant baculovirus-expressing anitin (Comai et al. 1992). In vitro-synthesized RNAs SV40 large T antigen were prepared as described (Lanford 1988). were detected by S1 nuclease analysis using 5'-end la- The lysates were fractionated on a 1.2-ml poros-HQ column (PerSeptive Biosystems) and eluted with a linear gradient of 0.2- beled single-strand DNA oligonucleotides. In the pres- 1.0 M KC1 in TM buffer. Fractions containing T antigen were ence of wild-type and mutant large T antigen proteins, pooled and dialyzed against TM m containing 0.1 M KC1 [50 transcription was carried out by preincubating the large mMM Tris (pH 7.9), 100 mM KC1, 12.5 mM MgC1, 1 mM EDTA, T antigen protein with a transcription mixture without 10% glycerol, 1 mM DTT, 1 mM phenylmethylsulfonylfluoride the four ribonucleotides on ice for 20 min. After the in- (PMSF)]. After dialysis, the T antigen pool was loaded to a hepa- cubation, all four ribonucleotides were added to initiate rin-agarose column and eluted with a 0.1-0.7 M KC1 linear gra- the transcription reaction. dient. Fractions were subjected to silver staining and Western blotting using T antigen monoclonal antibody to assess the pu- rity and integrity of large T antigen. Protein-protein interaction assays Affinity purification. Lysates from Sf9 cells infected with re- Sf9 cells infected with recombinant baculovirus were harvested combinant baculovirus-expressing SV40 large T antigen were 42-48 hr postinfection and washed with 1x phosphate-buffered prepared as described above. An affinity column using either saline (PBS). Cell pellets were resuspended in TMI°/0.1 M KC1 monoclonal antibody 101 (pAb 101) (against the carboxyl termi- buffer and lysed by sonication. NP-40 was then added to cell nus of T antigen) or pAb 108 (against the amino terminus of T lysates to a final concentration of 0.1%. Cell lysates were antigen) were prepared by cross-linking the antibodies to pro- cleared by centrifugation at 30,000g for 30 min. The amount of tein A-agarose beads (Harlow and Lane 1988). Purification of GST-fusion proteins used in the binding reactions was esti- large T antigen was carried out as described previously ISimanis mated by Bradford assays and Coomassie staining of SDS-poly- and Lane 1985). acrylamide gels. Equal molar amounts of each GST-fusion protein were used in the interaction assays. All binding reactions Purification of GST fusion proteins GST-large T antigen and were performed at 4°C with constant mixing on a nutator. Glu- GST-large T antigen mutant proteins were expressed in Sf9 tathione beads were allowed to bind the GST-fusion proteins cells using recombinant viruses. Cell lysates were incubated for 1 hr in TMI°/0.1 M KCL plus 0.1% NP-40. The beads were with glutathione-affinity resin (Pharmacia) for 1 hr at 4°C. The then washed four times with the binding buffer. Human TAFs beads were then washed five times with TMI°/0.2 M KC1 plus were synthesized using the TNT-coupled in vitro transcription

GENES & DEVELOPMENT 1615 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Zhai et al. and translation system (Promega). In vitro-translated [3SS]me- expression vector pNT2, Bruce Stillman for SV40 virus and plas- thionine-labeled TAFs were then added to the beads, and the mid pSV010, Brian Dynlacht for the E1A-13S expression con- reaction mixture was incubated overnight at 4°C. The beads struct, and Stanley Tahara for pBSMCAT plasmid. We are grate- were then washed five times with TM~°/0.1 M KC1 plus 0.1% ful to M. Lai, A. Lee, A. Schonthal, D. Johnson, and members of NP-40 buffer, boiled in SDS sample buffer, and resolved by SDS- their laboratories for helpful advice. We thank Luca Comai, P. PAGE. Gels were dried, and radiolabeled proteins were detected Verrjizer, R. Wilds, and J. Sarabia for critically reading the by autoradiography. Bound radiolabeled proteins were also manuscript. This work was supported, in part, by an American quantified using a Molecular Dynamics PhosphorImager. Cancer Society Institutional Research grant (IRG-21-35) and a James H. Zumberge Faculty Research and Innovation Fund Purification of protein from HeLa cells (L.C.). W.Z. was supported, in part, by a Kenneth Norris, Jr., Comprehensive Cancer Center graduate fellowship. Nuclear extracts from HeLa cells were prepared as described The publication costs of this article were defrayed in part by (Comai et al. 1992). They were loaded onto a heparin-agarose payment of page charges. This article must therefore be hereby column and RNA Pol I, UBF, and SL1 were eluted with a 0.1-1.0 marked "advertisement" in accordance with 18 USC section M KC1 gradient. Fractions containing SL1 activity, as deter- 1734 solely to indicate this fact. mined by in vitro transcription assays, were pooled and dialyzed against TMI°/0.2 M KC1. The SL1 pool was then loaded onto a SP-Sepharose column (Pharmacia) preequilibrated in TMI°/0.2 M KC1, and after extensive washes with TMI°/0.2 M KC1, SL1 was References eluted with TMI°/0.8 M KC1 and dialyzed against TMI°/0.1 M Beckmann, H., J-L. Chen, T. O'Brien, and R. Tjian. 1995. Coac- KC1. RNA Pol I used in the transcription assays was purified as tivator and promoter-selective properties of RNA polymer- described previously (Comai et al. 1992). UBF was either par- ase I TAFs. Science 270: 1506-1509. tially purified by chromatography on heparin-agarose, DEAE, Bell, S.P., R.M. Learned, H.-M. Jantzen, and R. Tjian. 1988. and Q-Sepharose, or purified to homogeneity (Bell et al. 1988). Functional cooperativity between transcription factors The two preparations of UBF were used interchangeably in the UBF1 and SL1 mediates human ribosomal RNA synthesis. transcription assays, with identical results. Although RNA Pol Science 241:1192-1197. I and SL1 were purified partially, they were not cross-contami- Bell, S.P., H.-M. Jantzen, and R. Tjian. 1990. Assembly of alter- nated with each other or UBF. native multiprotein complex directs rRNA promoter selec- For the in vivo co-immunoprecipitation analysis, nuclear ex- tivity. Genes & Dev. 4: 943-954. tracts from SV40-infected and mock-infected HeLa nuclear ex- Berger, L., D.B. Smith, I. Davidson, J.-J. Hwang, E. Fanning, and tracts were prepared as described above with the exception that A.G. Wildeman. 1996. Interaction between T antigen and the KC1 concentration in the nuclear extracts was brought to 0.4 TEA domain of the factor TEF-1 derepresses Simian virus M before loading onto a heparin-agarose column pre-equili- 40 late promoter in vitro: Identification of T-antigen do- brated in TM~°/0.4 M KC1. After washing extensively with mains important for transcriptional control. [. Virol. 70: TMI°/0.4 M KC1 buffer, The SL1 fraction was step eluted with 1203-1212. TMI°/0.8 M KC1 buffer and dialyzed against TMI°/0.1 M KC1. Cavanaugh, H.A. and E.A. Thompson, Jr. 1986. Hormonal regulation of transcription of rDNA. L Biol. Chem. 261: 12738- Co-immunoprecipitation from SV40-infected HeLa cells and 12744. Western blot analysis Clark, R., K. Peden, J.M. Pipas, D. Nathans, and R. Tjian. 1983. SL1 from either SV40-infected or mock-infected HeLa cell Biochemical activities of T-antigen proteins encoded by sim- nuclear extracts was prepared as described previously. The SL1 ian virus 40 A gene deletion mutants. Mol. Ceil Biol. 3: 220- purified from 100 ml of either SV40-infected or mock-infected 228. HeLa cells was used in each assay. For the co-immunoprecipi- Clos, J., D. Buttgereit, and I. Grummt. 1986. A purified tran- tation, anti-large T antigen monoclonal antibody (pAb 101) was scription factor (TIF-IB) binds to essential sequences of the incubated with 30 ~1 of a 50% (vol/vol) slurry of protein mouse rDNA promoter. Proc. Natl. Acad. Sci. 83: 604-608. A-Sepharose beads for 1 hr at 4°C. The beads were washed five Comai, L., N. Tanese, and R. Tjian. 1992. The TATA-binding times with TMl°/0.1 M KC1 containing 0.1% NP-40 and then protein and associated factors are integral components of the mixed with the SL1 fraction for 8 hr at 4°C with constant mix- RNA polymerase I transcription factor, SLI. Cell 68: 965- ing on a nutator. After washing with the binding buffer five 976. times, the protein was eluted with 1.0 M guanidine-hydrochlo- Comai, L., J.C.B.M. Zomerdijk, H. Beckmann, S. Zhou, A. Ad- ride in TMI°/0.1 M KC1. Elution was performed twice for 10 min mon, and R. Tjian. 1994. Reconstitution of transcription fac- each in 50 lal at 4°C. Eluate was combined and diluted with 3 tor SLI: Exclusive binding of TBP by SL1 or TFIID subunits. volumes of TMI°/0.1 M KC1. Proteins were then precipitated Science 266: 1966-1972. with TCA for 20 min on ice and collected by centrifugation at Conzen, S.D. and C.N. Cole. 1994. The transforming proteins of 30,000g for 20 min. Pellets were washed twice with cold ac- simian virus 40. Semin. Virol. 5: 349-356. etone and air-dried. Samples were boiled in SDS sample buffer, Damania, B. and J.C. Alwine. 1996. TAF-like function of SV40 resolved by SDS-PAGE, and transferred to nitrocellulose mem- large-T antigen. Genes & Dev. 10: 1369-1381. brane. The SL1 subunits were detected by incubation with the Eberhard, D., L. Tora, J.M. Egly, and I. Grummt. 1993. A TBP- appropriate antibody (either ~-TAFI48 , R-TAFI63, c~-TAFI110, or containing multiprotein complex (TIF-IB) mediates tran- ~-TBP) and visualized by using either the ECL chemilumines- scription specificity of murine RNA polymerase I. Nucleic cence procedure (Pierce) or the alkaline-phosphatase methods Acids Res. 21: 4180-4186. following reaction with NBT and BCIP (Harlow and Lane 1988). Fanning, E. 1992. Simian virus 40 large T antigen: The puzzle, the pieces, and the emerging picture. J. Virol. 66" 1289-1293. Acknowledgments Gruda, M.C., J.M. Zablolotny, J.H. Xiao, I. Davidson, and J.C. Alwine. 1993. Transcriptional activation by simian virus 40 We thank Robert Lanford for the large T antigen baculovirus large T antigen: Interaction with multiple components of the

1616 GENES& DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

T antigen stimulation of Pol I transcription

transcription complex. MoI. Cell. Biol. 13: 961-969. quired for accurate transcription of mammalian ribosomal Grummt, I., A. Smith, and F. Grummt. 1976. Amino acid star- RNA genes: Identification of a species-dependent factor. vation affects the initiation frequency of nucleolar RNA Nucleic Acids Res. 10: 6659-6670. polymerase. Cell 7: 439-445. Moss, T. and V.Y. Stefanovsky. 1995. Promotion and regulation Hammond, L.M. and H.L. Bowman. 1988. Insulin stimulates of ribosomal transcription in eukaryotes by RNA polymer- the translation of ribosomal proteins and the transcription of ase I. Progr. Nucleic Acids Res. Mol. Biol. 50" 25-66. rRNA in mouse myoblast. J. Biol. Chem. 263" 17785-17791. Nevins, J.R. 1994. Cell cycle targets of the DNA tumor viruses Harlow, E. and D. Lane. 1988. Antibodies: A laboratory Curr. Opin. Genet. Dev. 4" 130-134. manual. Cold Spring Harbor Laboratory, Cold Spring Har- Pockl, E. and E. Wintersburger. 1980. Increased rate of RNA bor, NY. synthesis: Early reaction of primary mouse kidney cells to Hernandez, N. 1993. TBP, a universal eukaryotic transcription infection with polyoma virus or simian virus 40. J. Virol. factor? Genes & Dev. 7" 1291-1308. 35" 8-19. Hinds, P.W. and R.A. Weinberg. i994. Tumor suppressor genes. Reeder, R.H. 1994. Regulation of transcription by RNA poly- Curr. Opin. Genet. Dev. 4: 135-141. merase I. Cold Spring Harbor Laboratory Press, Cold Spring Ide, T., S. Whelly, and R. Baserga. 1977. Stimulation of RNA Harbor, NY. synthesis in isolated nuclei by partially purified preparation Roeder, R.G. 1996. The role of general initiation factors in tran- of simian virus 40 T antigen. Proc. Natl. Acad. Sci. 74: 3189- scription by RNA polymerase II. Trends Biochem. Sci. 3192. 21" 327-334. Jacob, S.T. 1995. Regulation of ribosomal gene transcription. Safrany, G., N. Tanaka, T. Kishimoto, Y. Ishikawa, H. Kato, R. Biochem. J. 306" 617-626. Kominami, and M. Muramatsu. 1989. Structural determi- Jantzen, H.-M., A. Admon, S.P. Bell, and R. Tjian. 1990. Nuclear nant of the species-specific transcription of the mouse rRNA transcription factor hUBF contains a DNA-binding motif gene promoter. Mol. Cell Biol. 9: 349-353. with homology to HMG proteins. Nature 344: 830-836. Schnapp, A., H. Rosenbauer, and I. Grummt. 1991. Trans-acting Jantzen, H.-M., A.M. Chow, D.S. King, and R. Tjian. 1992. Mul- factors involved in species-specificity and control of mouse tiple domains of the RNA polymerase I activator hUBF in- ribosomal gene transcription. Mol. Cell. Biol. 104: 137-147. teracts with the TATA-binding protein complex hSL1 to me- Simanis, V. and D.P. Lane. 1985. An immunopurification pro- diate transcription. Genes & Dev. 6: 1950-1963. cedure for SV40 Large-T antigen. Virology 144: 88-100. Johnston, S.D., X.-M. Yu, and J.E. Mertz. 1996. The major tran- Sollner-Webb, B. and J. Tower. 1986. Transcription of cloned scriptional transactivation domain of Simian virus 40 large T eukaryotic ribosomal RNA genes. Annu. Rev. Biochem. antigen associates nonconcurrently with multiple compo- 55: 801-830. nents of the transcriptional preinitiation complex. J. Virol. Sommerville, J. 1986. Nucleolar structure and ribosome biogen- 70:1191-1202. esis. Trends Biochem. Sci. 11" 438-442. Lanford, R.E. 1988. Expression of simian virus 40 T antigen in Soprano, K.J., V.G. Dev, C.M. Croce, and R. Baserga. 1979. Re- insect cells using a baculovirus expression vector. Virology activation of silent rRNA genes by simian virus 40 in hu- 167" 72-81. man-mouse hybrid cells. Proc. Natl. Acad. Sci. 76: 3885- Learned, R.M., S.T. Smale, M.M. Haltiner, and R. Tjian. 1983. 3889. Regulation of human ribosomal RNA transcription. Proc. Soprano, K.J., G.J. Jonak, N. Galanti, J. Floros, and R. Baserga. Natl. Acad. Sci. 80: 3558-3562. 1981. Identification of an SV40 DNA sequence related to the Learned, R.M., S. Cordes, and R. Tjian. 1985. Purification and reactivation of silent rRNA genes in human > mouse hybrid characterization of a transcription factor that confers pro- cells. Virology 109: 127-136. moter specificity to human RNA polymerase I. MoI. Cell. Soprano, K.J., N. Galanti, G.J. Jonak, S. McKercher, J.M. Pipas, Biol. 5" 1358-1369. K.W.C. Peden, and R. Baserga. 1983. Mutational analysis of Learned, R.M., T.K. Learned, M.M. Haltiner, and R. Tjian. 1986. simian virus 40 T antigen: Stimulation of cellular DNA syn- Human rRNA transcription is modulated by the coordinate thesis and activation of rRNA genes by mutants with dele- binding of two factors to an upstream control element. Cell tions in the T-antigen gene. Mol. Cell. Biol. 3: 214-219. 45" 847-857. Tanese, N., B.F. Pugh, and R. Tijan. 1991. Coactivators for a Lin, J.-Y. and D.T. Simmons. 1991. The ability of large T antigen proline-rich activator purified from the multisubunit human to complex with p53 is necessary for the increased life span TFIID complex. Genes & Dev. 5: 2212-2224. and partial transformation of humna cells by simian virus Verrijzer, C.P. and R. Tjian. 1996. TAFs mediate transcriptional 40. J. Virol. 65" 6447-6453. activation and promoter selectivity. Trends Biochem. Sci. May, P., E. May, and J. Borde. 1976. Stimulation of cellular RNA 21" 338-342. synthesis in mouse-kidney cell cultures infected with SV40 virus. Exp. Cell Res. 100" 433-439. Miesfeld, R., B. Sollner-Webb, C. Croce, and N. Arnheim. 1984. The absence of a human-specific ribosomal DNA transcription factor leads to nucleolar dominance in mouse greater than human hybrid cells. Mol. Cell Biol. 4: 1306-1312. Mietz, J.A., T. Unger, J.M. Huibregtse, and P.M. Howley. 1992. The transcriptional transactivation function of wild-type p53 is inhibited by SV40 large T antigen and by HPV-16 E6 oncoprotein. EMBO J. 11: 5013-5020. Mishima, Y.M. and M. Muramatsu. 1979. The mechanism of decrease in nucleolar RNA synthesis by protein synthesis inhibitor. J. Biochem. 85: 807-818. Mishima, Y.M., L. Financsek, R. Dominami, and M. Mura- matsu. 1982. Fractionation and reconstitution of factors re-

GENES & DEVELOPMENT 1617 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

SV40 large T antigen binds to the TBP-TAF(I) complex SL1 and coactivates ribosomal RNA transcription.

W Zhai, J A Tuan and L Comai

Genes Dev. 1997, 11: Access the most recent version at doi:10.1101/gad.11.12.1605

References This article cites 47 articles, 26 of which can be accessed free at: http://genesdev.cshlp.org/content/11/12/1605.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.