Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of SV40 large T antigen

Blossom Damania and James C. Alwine^ Department of Microbiology, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6142 USA

The simian virus 40 (SV40) early gene product large T antigen promiscuously activates simple promoters containing a TATA box or initiator element and at least one upstream transcription factor-binding site. Previous studies have suggested that promoter activation requires that large T antigen interacts with both the basal transcription complex and the upstream-bound factor. This mechanism of activation is similar to that proposed for TBP-associated factors (TAFs). We report genetic and biochemical evidence suggesting that large T antigen performs a TAF-like function. In the tsl3 cell line, large T antigen can rescue the temperature-sensitive (ts) defect in TAFi,250. In contrast, neither El a, small t antigen, nor mutants of large T antigen defective in transcriptional activation were able to rescue the ts defect. These data suggest that transcriptional activation by large T antigen is attributable, at least in part, to an ability to augment or replace a function of TAFn250. In addition, we show that large T antigen interacts in vitro with the Drosophila TAFs (dTAFs) dTAFiilSO, dTAFnllO, and dTAF,i40, as well as TBP. The relevance of these in vitro results was established in coimmunoprecipitation experiments using extracts of SV40-infected a3 cells that express an epitope-tagged TBP. Large T antigen was coimmunoprecipitated by antibodies to epitope-tagged TBP, endogenous TBP, hTAFiilOO, hTAFnl30, and hTAFnISO, under conditions where holo-TFIID would be precipitated. In addition, large T antigen copurified and coimmunoprecipitated with phosphocellulose-purified TFIID from SV40-infected a3 cells. Large T antigen also coprecipitated with anti-TBP antibody from extracts of tsl3 cells expressing wild-type large T antigen under conditions where the ts defect in TAF„250 was rescued. In contrast, a trans-activation mutant of large T antigen, which was unable to rescue the ts defect, failed to coprecipitate. We conclude from these data that transcriptional activation of many promoters by large T antigen results from its performing a TAF-like function in a complex with TFIID. [Key Words: TAFs; TFIID; transcription; SV40 virus; large T antigen] Received January 22, 1996; accepted in revised form April 12, 1996.

The simian virus 40 (SV40) early gene product large T factor-1 (TEF-1) (Gruda et al. 1993) and Sp-1 (B. Damania antigen (T antigen) is known to be a promiscuous acti­ and J.C. Alwine, unpubl.). It is important to note that vator of many viral and cellular promoters. Such prom­ although large T antigen can interact with a component iscuity is suggested by the structural simplicity required of TFIID, it cannot activate a promoter containing only a of a promoter for activation by large T antigen; a TATA TATA element. The additional interaction with an up­ box or initiator element with at least one upstream tran­ stream-bound factor appears to be essential for activa­ scription factor-binding site (which can be variable) is tion. Such a mechanism of activation is similar to that of adequate (Gilinger and Alwine 1993; Gruda et. al 1993; the TBP-associated factors (TAFs). These components of Rice and Cole 1993). Although large T antigen is a TFIID cannot mediate transcriptional activation unless known DNA-binding protein, this function is not essen­ they interact with upstream-bound factors (Hoey et al. tial for transcriptional activation (Keller and Alwine 1993; Chen et al. 1994). These similarities in function 1985; Gallo et al. 1988, 1990; Beard and Bruggmann raise the question of whether large T antigen may per­ 1989; Zhu et al. 1991; Casaz et al. 1995). Our previous form a TAF-like function. studies have suggested that the activation of such pro­ Analysis of the Drosophila TFIID complex has shown moters requires large T antigen to interact, through pro­ that it consists of TBP and at least eight tightly bound tein-protein interactions, with both the basal transcrip­ subunits called TBP-associated factors or TAFs: tion complex and the upstream bound factors (Gilinger dTAFii250, dTAFiilSO, dTAF„I10, dTAF„80, dTAF„60, and Alwine 1993; Gruda et al. 1993). For example, we dTAFii40, dTAFii30a, and dTAFii30p (Chen et al. 1994). have shown that large T antigen can interact with The largest TAF in the TFIID complex, dTAF„250, is TATA-binding protein (TBP), transcriptional enhancer believed to provide a scaffolding function through inter­ actions with TBP and several other TAFs. The mamma­ * Corresponding author. lian TAF„250 is encoded by CCGl (Hisatake et al. 1993;

GENES & DEVELOPMENT 10:1369-1381 £ 1996 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/96 $5.00 1369 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

Damania and Alwine

Ruppert et al. 1993). In the temperature-sensitive (ts) Ela hamster cell line tsl3 (Talavera and Basilico \911] the ts defect results from mutation of CCGl (Sekiguchi et al. 1988; Hayashida et al. 1994; Noguchi et al. 1994). The ts defect in tsl3 cells was first characterized as a cell cycle defect because the cells arrest in Gj at the nonpermissive temperature (Talavera and Basihco 1977). How^ever, the ts defect in TAFulSO can be noted at the transcriptional level (Liu et al. 1985; Wang and Tjian 1994). At the non- permissive temperature, tsl3 cells do not exhibit a global defect in transcription, but transcription of specific genes is dramatically decreased (Wang and Tjian 1994). For example, the activity of the cyclin A (cycA) promoter is decreased by 8- to 10-fold at the nonpermissive tem­ 0.5 1 2 ^.g Effector Plasmid perature compared to its activity at the permissive tem­ perature; however, the activity of the jos promoter is relatively unaffected by the temperature (Wang and T Antigen Tjian 1994). Stable transfection of the TAFulSO gene into tsl3 cells rescues the ts defect in transcriptional activa­ tion (Wang and Tjian 1994). ^ 1.5 Previously, studies of the cell cycle phenotype of the ts defect in tsl3 cells indicated that the introduction of T antigen overcame the Gj arrest (Floros et al. 1981). This suggested to us that T antigen may be able to rescue the ts transcriptional defect in TAFulSO. In this study we 0.5 show that SV40 large T antigen can rescue the transcrip­ tional defect of TAF„250 in tsl3 cells at the nonpermis­ sive temperature. In addition, we show that large T an­ 0 12 3 tigen and TAFii250 share common in vitro interactions ^g Effector Plasmid with TAFs and TBP, and that large T antigen coimmu- Figure 1. Activation of the CycA promoter by Ela {top) and noprecipitates with TFllD from infected cell extracts. In large T antigen [bottom] in tsl3 cells at the permissive (32°C) addition, we show that large T antigen copurifies with and nonpermissive (39°C) temperatures. Two micrograms of the TFIID over phosphocellulose and coimmunoprecipitates CycA-luciferase reporter plasmid were transfected either alone with the purified TFIID. Mutants in large T antigen de­ (with filler plasmid) or with the indicated amounts of effector fective in transcriptional activation neither rescued the plasmids that expressed either Ela or large T antigen. Transfec- ts defect in TAFulSO nor coimmunoprecipitated with tions and assays were done as described in Materials and meth­ TFIID. We conclude from these findings that large T an­ ods. tigen performs a TAF-like function in a complex with TFIID. Tjian (1994). The addition of the Ela-expressing plasmid showed that Ela can activate the cycA promoter at the Results permissive temperature. Importantly, however, Ela failed to activate the promoter at the nonpermissive T antigen, like TAFjj250, can rescue the temperature. Conversely, T antigen caused very little ac­ transcriptional defect in tsl3 cells tivation at the permissive temperature; in numerous ex­ In previous studies the ts transcriptional defect in tsI3 periments the greatest activation mediated by large T cells was studied using the cycA promoter (Wang and antigen was <2.5-fold at 32°C. However, the more sig­ Tjian 1994). Hence we used the same cycA-luciferase nificant observation is that large T antigen activated the reporter plasmid in these studies. In the following trans­ cycA promoter at the nonpermissive temperature, in­ fection studies we asked whether T antigen, like wild- creasing its activity to a level approximately equal to type TAFii250, could rescue the ts defect of TAFulSO in that at the permissive temperature. These data suggest tsl3 cells. that T antigen and Ela are activating by very different Figure 1 shows the results of transfection of tsl3 cells mechanisms (see Discussion) and that T antigen, and not with the cycA reporter plasmid either alone or with in­ Ela, can rescue the ts defect in TAFii250. creasing amounts of a plasmid expressing T antigen (bot­ Figure 2 shows the results of experiments with tsl3 tom) or Ela (top). In both experiments it can be seen that cells at the nonpermissive temperature where the cycA- at the nonpermissive temperature (39°C) the cycA pro­ luciferase reporter plasmid was transfected with increas­ moter alone (the 0 effector plasmid point) showed -10% ing amounts of plasmids that express various activators of the activity seen at the permissive temperature (32°C). such as TAF,i250, T antigen, Ela, and SV40 small t an­ This is in agreement with similar studies by Wang and tigen. The results are expressed as the percentage of the

1370 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of T antigen

300 Figure 2. Effects of several activators on the activity of tlae CycA promoter in tsl3 cells at the nonpermissive temperature. Two micrograms of the CycA-luciferase T Antigen reporter piasmid were transfected either alone (with filler piasmid) or with increas­ t Antigen ing amounts of plasmids that express var­ ious activators such as hTAFii250, large T antigen (T antigen), Ela, and SV40 small t T+t Sep. antigen jt antigen). In addition, the combi­ nation of both large T and small t antigens T+tER was tested in two ways: T + t Sep indicates that large T antigen and small t antigen Ela were expressed from separate plasmids,- T + t ER indicates that a single piasmid was transfected encoding the entire early TAF250 region (ER) that can produce both large T antigen and small t antigen by alterna­ tively splicing the early transcript. The ac­ 1 3 tivity of the CycA-luciferase reporter pias­ mid alone (with filler piasmid) was deter­ \i g Effector Piasmid mined at the permissive temperature (32°C) and compared to the activity at the nonpermissive temperature (39°C) in the presence or absence of effectors. Thus the activities at the nonpermissive temperature are expressed as the percentage of the eye A promoter activity at 32°C, which is set as 100%. This provides a comparative display of the ability of the activators to restore cycA promoter activity at the nonpermissive temperature. Transfections and assays were done as described in Materials and methods.

cycA promoter activity at 32°C where the activity at p53 and Rb binding were not essential for transcriptional 32°C is set as 100%. This provides a comparative display activation (Zhu et al. 1991) and data that showed that of the ability of the activators to restore cycA promoter transcriptional activation by large T antigen was not al­ activity at the nonpermissive temperature. As noted in tered in cells that lacked both Rb and p53 (Trifillis et al. Figure 1, at the nonpermissive temperature, the activity 1990). of the cycA promoter alone fell to 10%-15% of the ac­ As mentioned in the introductory section, the results tivity of the promoter at the permissive temperature. of Wang and Tjian (1994) indicated that the ts transcrip­ The addition of the TAFulSO-expressing piasmid rescued tional defect in TAFii250 in tsl3 cells was promoter the activity of the cycA promoter to the level seen at the specificity, that is, the cycA promoter was affected permissive temperature. This agrees well with the data whereas the fos promoter was not. To test whether T of Wang and Tjian (1994) showing that TAFulSO restored antigen showed a similar specificity in tsl3 cells we the activity of the eye A promoter in tsl3 cells at the asked whether the presence of large T antigen would nonpermissive temperature. Similarly, the addition of affect the activity of the fos promoter at the nonpermis­ large T antigen rescued the promoter activity to -90% of sive temperature. In transfection experiments similar to the activity at the permissive temperature. However, those described above, we noted that the activity of the neither Ela nor small t antigen affected the activity of fos promoter alone increased approximately twofold at the cycA promoter at the nonpermissive temperature. the nonpermissive temperature relative to the permis­ Interestingly, the combination of both large T antigen sive temperature. Cotransfection of increasing amounts and small t antigen produced a synergistic activation. of the large T antigen expression piasmid (1,3,5 jxg) pro­ This is shown in two experiments in Figure 2 where (1) vided no additional activation (data not shown). This large T antigen and small t antigen were introduced on suggests similar promoter specificity between large T an­ separate plasmids (T + t Sep.); and (2) a single piasmid tigen and TAFii250, in agreement with the interpretation was transfected encoding the entire early region (ER), that large T antigen provides a function in tsl3 cells which can produce both large T antigen and small t an­ similar to TAFii250. tigen by alternatively splicing the early transcript (T + t ER). Trsins-activation mutants of large T antigen fail to In a similar experiment, mutants of large T antigen rescue the ts defect in TAF}j250 that fail to bind either p53 (Bentivoglio et al. 1992) or retinoblastoma gene product (Rb) (Kaelin et al. 1990) To establish that large T antigen's ability to rescue the ts were tested (data not shown). The results suggested that defect in TAFii250 correlated with defined trdns-activa- failure of large T antigen to bind either p53 or Rb had no tion functions of large T antigen we tested a number of dramatic affect on the ability of large T antigen to acti­ mutants that have been characterized previously for vate the cycA promoter at the nonpermissive tempera­ trans-activation (Zhu et al. 1991). The mutants tested ture. This agrees with previous results that showed that (small in-frame deletions and insertions; Table 1) all

GENES & DEVELOPMENT 1371 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

Damania and Alwine make full-length or near full-length large T antigen. Zhu et al. (1991) determined the ability of each mutant to activate the SV40 late promoter in CV-1 cells. In Figure 3A these late promoter trflns-activation data (TA LP) are plotted along with our determination of the effect of each mutant (5 |xg of transfected plasmid) on cycA pro­ moter activity in tsl3 cells at 39°C (TA CYC39); pro­ moter activity in all cases is expressed as the percentage of the activation by wild-type T antigen where wild type is 100%. Each promoter was similarly affected by the mutants suggesting that the activation of the cycA pro­ moter in tsl3 cells at 39°C correlates with a defined trans-activation function of large T antigen. In Figure 3B the activation of the CycA promoter in Basal 2803 2811 2815 2817 2831 2835 WT tsl3 cells at 39°C is expressed as the percentage of the cycA promoter activity at 32°C, similar to the compari­ ITALP ITACYC39 son in Figure 2. As seen previously, the basal activity of the cycA promoter at 39°C fell to a level of 10%-15% of 100 the activity at 32°C. In this set of experiments wild-type large T antigen restored the activity of the promoter to 80% of the activity at 32°C. Mutants inA2815 and inA2835, previously shown to be the most defective in transcriptional activation, are the most defective for res­ cue of the TAFii250 ts mutation.

Interactions of TAFs with SV40 T antigen The above data show that T antigen can rescue the ts defect of TAFu250 in tsl3 cells at the nonpermissive temperature. This suggests that large T antigen may function in a manner similar to a TAF, specifically Basal 2803 2811 2815 2817 2831 2835 WT TAF„250. One of the characteristics of TAFii250 is its Figure 3. [A] Mutants of T antigen characterized previously for ability to interact with TBP and other TAFs, acting as a transcriptional activation (Zhu et al. 1991) were tested for acti­ scaffolding protein for the formation of TFIID. To deter­ vation of the cycA promoter in tsl3 cells at 39°C. Five micro­ mine whether large T antigen can interact with compo­ grams of tfie respective plasmids jTable 1) were transfected and nents of TFIID, we used a group of glutathione-S-trans- luciferase activity assayed as described in Materials and meth­ ferase fusion proteins (Fig. 4) with full-length large T ods. The trans-activation of the cycA promoter (TA CYC39; antigen (GST-FLT) and portions of it (T1-T5) to assay hatched bars) is shown in comparison with the data of Zhu et al. binding to in vitro transcribed and translated ^^S-labeled (1991) for the trans-activation of the SV40 late promoter (TA LP; dTAFn250, dTAFiilSO, dTAF„I10, dTAF„80, dTAFii60, solid bars). Activation of each promoter by wild-type large T antigen is set at 100%. [B] The activation of the cycA promoter dTAFn40 (Hoey et al. 1993). It should be noted that we at 39°C is expressed as the percentage of the activity of the cycA have tried to make larger portions of the carboxy-termi- promoter alone at 32°C, similar to the representation of the data in Fig. 2. The shaded bars on top indicate standard deviation.

Table 1. Large T antigen mutants tested Region/domain nal region of large T antigen (e.g., a combination of T3 Mutant Type of mutation affected and T4); however, these have consistently failed to be produced in Escherichia coli. In addition, we tested bind­ inA2803 insertion at amino amino terminus acid 34 ing to a GST fusion with small t antigen (GST-t). inA2811 insertion at amino ATP-binding and Figure 5A shows the results of the binding analyses acid 424 ATPase domain (lanes; In = input; G = GST; 1 = GST-Tl; 2 = GST-T2; inA2815 insertion at amino DNA-binding domain 3 = GST-T3; 4 = GST-T4; 5 = GST-T5; T = GST-FLT; acid 168 t = GST-t). Because exposure times are variable between inA2817 insertion at amino DNA-binding domain the various experiments, the data were quantitated by acid 219 Phosphorlmager analysis and presented in Figure 5B as dl2831 deletion of amino amino terminus the percentage input TAF bound after subtraction of the acids 4-34 nonspecific binding to the GST moiety. The data suggest inA2835 insertion at amino amino terminus that full-length large T antigen bound significantly to acids 85 and 86 dTAF„150, dTAFiillO, and dTAFii40. In experiments us-

1372 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of T antigen

too 200 300 400 500 600 700 gen may interact with a number of the components of J I J I J TFIID. To confirm that such interactions occur in vivo, we used a3 cells, a HeLa cell line that expresses consti- GST-FLT AAs 5-708 GST-TI AAs 5-172 A GST-T2 AAs 168-383 250-B GST-T3 AAs 379-561 GST-T4 In G 1 2 3 4 5 T t AAs 557-707 GST-T5 f _ |f AAs 5-383 150-1 Figure 4. Diagram of full-length T antigen and regions of large T antigen that were fused to the glutathione binding site of In G 1 2 34 5 T t glutathione-S-transferase (GST). The amino acids included in each GST fusion are indicated along with the regions name; FLT (full length T antigen), Tl, T2, T3, T4, and T5. no-f In G 1 2 3 4 5 T t ing human TAFs (not shown) we have noted very good 80^ binding to hTAF„130, the counterpart of dXAFnl 10, and hTAFii32 but not to hTAFjilOO, the counterpart of In G12345Tt dTAFijSO. An interaction with dTAFii60 was also indi­ cated at 5% input bound; however, the significance of 60 binding levels of ^5% may be questionable. Little to no 1 interaction was detected with dXAFulSO and dXAFuSO. In G 1 2 3 4 5 T t Using GST fusions to regions Tl through T5, we found that dTAFiillO bound fairly well to regions Tl and T5 40-| that overlap, both containing the amino terminal 172 amino acids of T antigen. Previously, Tl and T5 have In G12345Tt been shown to interact with TBP (Gruda et al. 1993). Small t antigen also showed a modest interaction with dTAFiillO. The amino terminal 82 amino acids of large 30 T and small t antigens are shared. In an experiment not shown, we determined that dTAFullO can interact with 25 a GST fusion with the amino terminal 82 amino acids of T antigen. However, additional amino acids substan­ 20 tially increased binding, and full-length large T antigen is clearly the best binding substrate. In addition, full- 1 length large T antigen was the only substrate that bound significantly with dTAFnlSO and dTAFii40. This sug­ gests that different TAFs may use different regions of , large T antigen for binding, and that the structures nec­ essary for optimal binding of TAFs may be formed from separate domains of large T antigen brought together 1 only in the folded structure of the full-length protein. TAF250 TAF150 TAFllO TAF80 TAF60 TAF40 The GST-binding data were not altered when 200 |JLg/ml Figure 5. [A] The results of binding of dTAFi,250, dTAFnlSO, ethidium bromide was added to the binding reactions dTAF„110, dTAF„80, dTAF„60, and dTAFii40, to GST fusions of (data not shown). This indicates that the binding was large T antigen and small t antigen. Lane "In" of each lane attributable to protein-protein interactions rather than shows the total input of in vitro transcribed and translated ^^S- tethering attributable to binding to a common piece of labeled TAFs used in each binding reaction. Each labeled TAF DNA (Lai and Herr 1992). was bound to the GST moiety alone (lane G) as well as to the various GST fusions with T antigen shown in Fig. 4: (lane 1] GST-Tl; (lane 2) GST-T2; (lane 3) GST-T3; (lane 4] GST-T4; Coimmunoprecipitation of T antigen with complexes (lane 5) GST-T5; (lane T) GST-full length T. In addition, lane "t" containing TBP shows the binding to a GST fusion with small t antigen. [B] The binding data in A were quantitated using a Molecular Dynamics Previously, we have shown that TBP can bind to GST- Phosphorlmager. After subtraction of background binding to the FLT, GST-Tl, and GST-T5 in vitro (Gruda et al. 1993). GST moiety alone, the percentage of input TAF bound was cal­ These data and the data above suggest that large T anti­ culated and plotted.

GENES & DEVELOPMENT 1373 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

Damania and Alwine tutively an influenza hemagglutinin (HA) epitope-tagged TBP (eTBP; Zhou et al. 1992). Immunoprecipitation of ,#• 5" v«r> a3 cell nuclear extracts with anti-HA antibody has been shown to precipitate holo-TFIID (Zhou et al. 1992). To test whether T antigen is associated with TFIID in vivo we used SV40 (strain 776) to infect normal HeLa cells (HeLa/SV40) or a3 cells (a3/SV40) at an m.o.i. of 20 pfu/ cell. Forty-eight hours after infection, the cells were har­ vested and nuclear extracts were prepared and analyzed in a variety of immunoprecipitation experiments de­ 105- scribed below (see also Material and methods). Figure 6 shows the coimmunoprecipitation of eTBP 70- and large T antigen using either the anti-HA antibody or an anti-TBP antibody. Nuclear extract was mixed with sTBP antibodies and precipitated with protein A-Sepharose 43- beads. The immunoprecipitates were separated by SDS- PAGE and transferred to nitrocellulose. The resulting 25- Western blot was probed with anti-T antigen antibody to detect the presence of T antigen in the immunoprecipi­ tates. Lanes 6 and 7 show 50 |JLg of the HeLa/SV40 and a3/SV40 unprecipitated cell extracts. The presence of SV40 large T antigen can be seen migrating at —92 kD, Figure 7. Coimmunoprecipitation of eTBP using an antibody similar extracts from noninfected cells showed no band against large T antigen. The extracts and experiments are the same as shown in Fig. 6 except the precipitating antibody was anti-large T antigen and the blot was probed for eTBP. The .^ strong band migrating just above eTBP is caused by the anti­ .^ bodies used for the immunoprecipitation that, because of their large amounts, cross-react with the secondary antibody. Lanes 1 P and 3 show one-tenth of the HeLa/SV40 and a3/SV40 unprec­ / ipitated cell extracts used in these experiments.

at this position (not shown). Lanes 4 and 5 show that both the anti-HA and anti-TBP immunoprecipitates /// from the a3/SV40 extracts coprecipitated T antigen. I I I I I I 105. Conversely, control immunoprecipitates from extracts of noninfected a3 cells (a3) showed no T antigen (lanes 2 70- and 3). An additional control in lane 1 shows that the anti-HA antibody precipitated no T antigen from the HeLa/SV40 extracts indicating no cross-reactivity be­ tween the anti-HA antibody and large T antigen. The 43 ^Wi ^^ results confirm that the large T antigen detected in the 6 7 immunoprecipitates from the a3/SV40 extracts was present because of coprecipitation in a complex with Figure 6. Coimmunoprecipitation of large T antigen using an­ TBP or TFIID. The bands at -50 kD are the heavy chains tibodies against eTBP and TBP. SV40 strain 776 was used to of the antibodies used for the immunoprecipitation. infect HeLa cells (HeLa/SV40) or a3 cells (a3/SV40) or a3 cells These are present in large amounts and cross-react with were mock infected (aS) for 48 hr and nuclear extracts prepared (see Material and methods). As indicated, the extracts were im- the secondary antibody used to probe the blot. munoprecipitated with anti-HA antibody to precipitate the To verify the coprecipitation of large T antigen and eTBP present in a3 cells, or with anti-TBP antibody. The im­ TBP or TFIID, we performed the reverse immunoprecip­ munoprecipitates were separated by electrophoresis on a 9% itation where anti-T antigen antibody was used to pre­ SDS-polyacrylamide gel and transferred to nitrocellulose. The cipitate the cell extracts and the resulting Western blots resulting Western blot was probed with anti-T antigen antibody were examined for coprecipitated eTBP using the anti- to detect the presence of T antigen in the immunoprecipitates. HA antibody. Figure 7 shows the results of such analysis Lanes 6 and 7 show one-tenth of the HeLa/SV40 and a3/SV40 on uninfected a3 cell extract, a3/SV40 extract, and unprecipitated cell extracts used in these experiments. The presence of SV40 large T antigen can be seen migrating at —92 HeLa/SV40 extract. Comparisons of lanes 4, 5, and 6 kD, similar extracts from noninfected cells showed no band at show that the anti-T antigen antibody precipitated no this position (not shown). The bands at —50 kD are the heavy eTBP from uninfected a3 cell extracts but it clearly pre­ chains of the antibodies used for the immunoprecipitation that, cipitated eTBP from extracts of SV40-infected a3 cells. because of their large amounts, cross-react with the secondary The control immunoprecipitation of HeLa/SV40 ex­ antibody used to probe the blot. tracts showed no eTBP as expected. A similar experi-

1374 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of T antigen ment using anti-T antigen to immunoprecipitate a3 and a3/SV40 extracts showed that the expected results were obtained when the Western blot was probed with anti- TBP antibody (not shown). Treatment of extracts with DNase followed by immu- # i^ i^ # noprecipitation in the presence of DNase resulted in no ^ ?5 J if if if if if if alterations in the above data, which suggest that T anti­ I I I I I gen and eTBP coimmunoprecipitated (not shown). This indicates that coimmunoprecipitation resulted from pro­ tein-protein interactions and not by tethering attribut­ able to binding to a common piece of DNA.

Coimmunoprecipitation of T antigen with anti-human TAF antibodies The above data suggest strongly that a stable complex exists in vivo containing large T antigen and TBP or TFIID. To determine whether large T antigen coimmu­ noprecipitated with other components of TFIID we per­ formed similar immunoprecipitations with monoclonal 23 45 678 9 10 11 antibodies to three different human TAFs: hTAFjjIOO Anti Anti Anti Anti Anti (analog of dTAF„80), hTAFnlSO (analog of dTAF„110), 100 130 250 HA hnRNPCl and hTAFii250 (analog of dTAFnlSO). The a3 and a3/ SV40 extracts were precipitated with the various anti­ Figure 8. Coimmunoprecipitation of large T antigen by anti- TAF antibodies. a3 and a3/SV40 nuclear extracts were immu- bodies and the precipitates were analyzed by Western noprecipitated with anti-hTAF„100 (Anti 100, lanes 2,3], anti- blotting using anti-T antibody to probe the blots. Figure hTAF„130 (Anti 130, lanes 4,5], anti-hTAF„250 (Anti 250, lanes 8 shows that antibodies to each hTAF (Anti 100, Anti 6,7); anti-HA, a repeat of the data in Fig. 7 (lanes 8,9], and a 130, and Anti 250) coimmunoprecipitated large T anti­ control antibody against the hnRNPCl protein (lanes 10,11]. gen from a3/SV40 extracts in a fashion similar to the The immunoprecipitates were analyzed by Western blotting as precipitation of T antigen with eTBP using anti-FiA an­ in Figs. 6 and 7 using anti-large T antigen antibody to probe the tibody. A control monoclonal antibody against the hu­ blot. Lane 1 shows one-tenth of the a3/SV40 unprecipitated cell man hnRNPCl protein showed no precipitation of T an­ extracts used in these experiments. tigen. As in the other immunoprecipitation experiments, the coprecipitation of T antigen with anti-TAF antibod­ ies was not altered when the extracts were treated with periments to confirm that the copurification of large T DNase and the immunoprecipitations were done in the antigen with TFIID was attributable to an interaction presence of DNase (data not shown). In addition, West­ between large T antigen and TFIID. The purified TFIID ern blot analysis has shown that the anti-TAF antibodies fraction was immunoprecipitated with either anti-HA do not cross-react with T antigen (not shown). The over­ antibody (aHA), to precipitate the eTBP and holoTFIID, all data indicate that large T antigen coimmunoprecipi- or a control antibody, anti-hnRNPCl antibody (aCl). tates because it is in complex with TFIID. The precipitates were subjected to Western blot analysis and probed with anti-large T antigen antibody. Figure 9B clearly shows that large T antigen coimmunoprecipi­ Large T antigen purifies as a complex with TFIID tated suggesting that copurification over phosphocellu­ Nuclear extract from SV40-infected a3 cells was frac­ lose results from a physical interaction with TFIID. tionated on phosphocellulose to produce the TFIID frac­ tion as described by Dignam et al. (1983b) (see Materials Trans-activation mutants of large T antigen that fail and methods). Figure 9A shows a Western blot of the 1.0 to rescue the ts defect in TAFij250 also fail to interact M KCl eluate that contains TFIID. The same blot appears with TFIID in each lane, it has been probed consecutively, first with anti-large T antigen (lane 1), then with anti-HA to detect In Figure 3A,B we showed that inA2815 was the large T the eTBP (lane 2), then with anti-hTAF„I30 (lane 3), and antigen mutant most defective in its ability to rescue the finally with anti-hTAFii250. To retain bound proteins ts defect in TAFn250. This correlated with a severe de­ the blot was not stripped between probings. T antigen fect in trans-activation of the SV40 late promoter (Zhu et and each component of TFIID were detected readily in al. 1991). If these functions of large T antigen are medi­ the 1 M fraction. Thus, in agreement with the data sug­ ated through a complex with TFIID then it would be gesting an association between T antigen and TFIID, predicted that a mutant like inA2815 may be unable to these data indicate that some large T antigen copurifies form such complexes. To test this we transfected tsl3 with TFIID. cells at both 32°C and 39°C with plasmids expressing In Figure 9B we performed immunoprecipitation ex- wild-type large T antigen, inA2815 and inA2817 (which

GENES & DEVELOPMENT 1375 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

Damania and Alwine

these lanes was one-twentieth of the amount used in the anti-TBP coimmunoprecipitation experiments. Thus we •^o^ can make a semiquantitative estimate that under wild- A if if if B I I I ^I I^ type conditions at least 5% of the total large T antigen is 205— ^' (F •*• <-TAF„250 in TBP-containing complexes which are stable enough to 105 — be retained through the immunoprecipitation and wash­ •"*^ # <-TAr,130 TAg ing. The results of several other similar immunoprecip­ 105— itation experiments (not shown) have indicated that as 70 — ■^ ' ff* 7"" 7*- ^^ As much as 15%-20% of the wild-type large T antigen may 70 — be in complexes containing TBP. I." H 43 — 43 — Discussion A TAF-like function foi large T antigen Our previous studies (Gruda and Alwine 1991; Gilinger and Alwine 1993; Gruda et al. 1993) and studies of others Figure 9. Large T antigen copurifies in complex with TFIID. (Casaz et al. 1991, 1995; Kelly and Wildeman 1991; Rice Nuclear extract from SV40-infected a3 cells was fractionated on and Cole 1993) have suggested that transcriptional acti­ phosphocellulose (see Materials and methods). [A] The 1.0 M vation by SV40 large T antigen requires protein-protein eluate containing TFIID was analyzed by Western blotting. The same blot appears in each lane [1-4] and has been probed con­ secutively, with anti-large T antigen (lane 1], anti-HA to detect the eTBP (lane 2), anti-hTAFnlSO (lane 3], and anti-hTAFii250. Input aTBP Ppt. To retain bound proteins the blot was not stripped between probings. [B] The same 1.0 M fraction was immunoprecipitated anti-HA antibody (aHA) or a control antibody, anti-hnRNPCl antibody (aCl). The precipitates were subjected to Western blot analysis and probed with anti-large T antigen antibody.

is only partially defective in transcriptional activation and rescue; see Fig. 3A,B). In Figure 10 (Input lanes) 20 jjLg of the 32°C nuclear extracts or 10 jjtg of the 39°C nuclear 32 extracts were analyzed by Western blotting for large T antigen. It can be seen that steady-state levels of wild- type and both mutant large T antigens are similar at each respective temperature. Each nuclear extract (400 ixg of fit the 32°C nuclear extracts and 200 |xg of the 39°C nuclear extracts) was then precipitated with anti-TBP and the coprecipitated large T antigen was detected by Western blotting (aTBP Ppt. panels). Both wild-type and inA2817 large T antigens coprecipitated from the 32°C and 39°C nuclear extracts. The coimmunoprecipitation of - -T inA2817 large T antigen from the 32°C nuclear extracts 39 appears to be low in this experiment; in additional ex­ periments (not presented) we have detected coimmuno­ precipitation of inA2817 at levels similar to that of wild- type large T antigen. In comparison, little if any inA2815 III large T antigen coprecipitated from either the 32°C or Figure 10. Coimmunoprecipitation of large T antigen and TBP the 39°C nuclear extracts, suggesting that its defects in from transfected tsl3 cells. The tsl3 cell line was transfected at trans-activation and rescue result from an inability to 32°C and 39°C with 5 p.g of plasmid encoding either wild-type form a complex with TFIID. In control immunoprecipi­ large T antigen (p6-lAL) or mutants inA2815 or inA2817. Forty- tation experiments, performed identically except for the eight hr after transfection nuclear extracts were prepared and use of anti-HA antibody, we detected no precipitation of immunoprecipitated using anti-TBP antibody. The precipitates large T antigen (not shown). were analyzed for large T antigen by Western blotting. The In­ put lanes show the large T antigen detected in 20 (xg of the crude In Figure 10 we note that at both 32°C and 39°C the nuclear extracts prepared at 32°C or 10 ^.g of the crude nuclear intensities of the bands of wild-type large T antigen re­ extracts prepared at 39°C. The aTBP precipitate (Ppt.) lanes sulting from coimmunoprecipitation with TBP are sim­ show the T antigen coimmunoprecipitated using anti-TBP an­ ilar to the intensities of the bands in the nuclear extract tibody with 400 \xg of the crude nuclear extracts prepared at (input) lanes. The amount of nuclear extract examined in 32°C or 200 |xg of the crude nuclear extracts prepared at 39°C.

1376 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of T antigen interactions between T antigen and both the basal tran­ transcriptional activation mechanisms mediated by scription complex and upstream bound transcription fac­ these tumor suppressor proteins are not associated with tors. These data indicated that although T antigen inter­ the ability of large T antigen to rescue ts TAFji250. How­ acts with the basal transcription complex it could not ever, characterized trans-activation mutants inA2835 activate a promoter containing only a TATA element; and inA2815 (Table 1) were found to be defective in res­ activation was dependent on the additional interaction cue. Our data indicate that this defect results from an with the upstream bound factor. Such requirements for inability of the mutant large T antigen to complex with activation are similar to those of the TAFs that, as inte­ TFIID. The mutation in inA2835 maps to the amino ter­ gral components of TFIID, cannot mediate transcrip­ minus. This correlates with both our in vitro binding tional activation unless they interact with upstream data that indicated that dTAFnHO interacts with the bound factors (Hoey et al. 1993; Chen et al. 1994). This amino-terminal 172 amino acids of T antigen and our similarity in mechanisms raised the question of whether previous data that indicated that TBP interacts within large T antigen may perform a TAF-like function. this same region (Gruda et al. 1993). Thus the amino- If large T antigen performs as a TAF it would be pre­ terminal mutation may be defective because of a loss of dicted that both genetic and biochemical evidence could the ability to bind to TAFul 10 or TBP. The mutation in be generated to show that large T antigen is associated inA2815 maps to the DNA-binding domain; however, it with TFIID and affects its function. In the present stud­ is unlikely that the loss of ability to bind to DNA is the ies we have provided evidence showing that large T an­ cause of defects in trans-activation or rescue. First, the tigen can rescue the temperature-sensitive defect in mutation that affects mutant inA2835 does not affect TAFij250 in tsI3 cells and that mutants in transcrip­ the DNA-binding domain. In addition, several previous tional activation cannot mediate rescue. This result sug­ studies have demonstrated that specific and nonspecific gests that large T antigen can function like a component DNA binding by large T antigen is not essential for tran­ of TFIID by augmenting, or replacing, a function of scriptional activation (Keller and Alwine 1985; Gallo et TAFii250. In addition, it correlates this effect with a al. 1988, 1990; Beard and Bruggmann 1989; Zhu et al. known transcriptional activation function of large T an­ 1991; Casaz et al. 1995). It has been proposed (Casaz et tigen. al. 1995) that sequences within the DNA-binding do­ That large T antigen is a component of TFIID was main are also critical for correct protein folding. Thus indicated by in vitro binding studies and coimmunopre- mutations in this region may result in an aberrantly cipitation experiments. The in vitro binding studies in­ folded molecule that is unable to trans-activate. This dicated that large T antigen interacts significantly with suggests that the structure of large T antigen necessary dTAFnISO, dTAF„110, and dTAFii40, as well as for optimal binding of TAFs, and related proteins, may be hTAFniaO, and hTAFnSl (data not shown). These data formed from separate domains of large T antigen brought compliment our previous binding studies that indicated together only in the case of the full-length wild-type pro­ an interaction with TBP (Gruda et al. 1993). Hence large tein. Our observation that a GST fusion of full-length T antigen is able to interact with at least four compo­ large T antigen was the only GST substrate that bound nents of TFIID. The in vivo significance of interactions dTAFiil50 and dTAF[i40 agrees with this conclusion. with multiple components of TFIID was demonstrated by the finding that large T antigen in SV40-infected cell extracts was coimmunoprecipitated using antibodies Contributions of small t antigen specific for eTBP, TBP, hTAFnlOO, hTAFniaO, and We have tested the effects of small t antigen in transcrip­ hTAFu250, under conditions where holo-TFIID would be tional activation of the cycA promoter in tsl3 cells at the precipitated. In addition, we demonstrate that wild-type nonpermissive temperature. Using this hamster cell line T antigen can be coprecipitated with TBP from tsl3 cell we observed that small t antigen could not rescue the ts extracts and that mutants that fail to rescue the ts defect defect in TAFii250. However, the presence of small t in TAFii250 fail to coprecipitate. Finally, we have pro­ antigen with large T antigen increased synergistically vided data showing that large T antigen purifies as a the effects of large T antigen. Previously, small t antigen component of TFIID. The combined data strongly sug­ has been shown to transcriptionally activate a subset of gest a TAF-like function where large T antigen is closely promoters activated by large T antigen (Loeken et al. associated with TFIID to mediate transcriptional activa­ 1988; Loeken 1992). We have noted in the in vitro bind­ tion and rescue. ing studies that small t antigen does interact with dTAFjillO, an interaction that may relate to the synergy with large T antigen. However, it has been shown that Mutants of large T antigen affecting trans-activation small t antigen binds to protein phosphatase type 2A also affect rescue of the ts defect of TAF^ZSO (PP2A) and inhibits its ability to dephosphorylate a va­ The examination of large T antigen mutants indicates riety of phosphoproteins (Pallas et al. 1990; Scheidtmann that interaction with either p53 or Rb (and Rb-related et al. 1991; Ruediger et al. 1992; Sontag et al. 1993). This family members) is not essential for the rescue of ts has been shown to affect the activity of several transcrip­ TAFii250. This is in agreement with other studies of tion factors (Frost et al. 1994; Wheat et al. 1994). Hence, transcriptional activation by large T antigen (Trifillis et the synergy caused by small t antigen may relate to its al. 1990; Zhu et al. 1991). These findings indicate that effects on the phosphorylation state of transcription fac-

GENES & DEVELOPMENT 1377 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

Damania and Alwine tors, maintaining them in a phosphorylated state that is functions in this complex as a TAF. It seems likely that better used by large T antigen for transcriptional activa­ this function of large T antigen would mediate transcrip­ tion. tional effects. However, large T antigen influences the cell in many ways, including affecting the cell cycle, hence the interaction with TFIID may also alter other Comparison of trsms-activation by large T antigen and nuclear and cellular functions of TAFs. Large T antigen's adenovirus El a interactions with TFIID may be mediated by binding to Like large T antigen, adenovirus Ela is a promiscuous one or more components of TFIID; for example, we have activator of many viral and cellular promoters and has shown interaction with TBP and several TAFs been shown to interact with TBP and TFIID (Boyer and (dTAFiilSO, dTAF„110, and dTAFii40). TBP, dTAFuISO, Berk 1993; Geisberg et al. 1994). However, previously it and dTAFjilIO also interact with dTAFii250, the major has been proposed that large T antigen and Ela use dif­ scaffolding protein of TFIID. This similarity of interac­ ferent mechanisms for transcriptional activation (Loeken tions further suggests that large T antigen may provide a et al. 1986; Gruda et al. 1993). In the present data we have function analogous to TAFji250. shown that the ability of large T antigen to rescue the ts defect in TAFQISO was not shared by Ela, again suggesting that these two viral activators use different mechanisms Materials and methods for the activation of transcription. One striking difference between Ela and large T antigen comes from a comparison Plasmids of their abilities to activate transcription as Gal4 fusion Plasmids pRSV-Tex, pRSV-t, and pRSV-Ela contain cDNA proteins. Ela contains an activating region that functions copies of SV40 large T, small t antigen, and adenovirus Ela well as a Gal4 fusion protein; however, neither large T cDNA, respectively, under the control of the Rous sarcoma vi­ antigen nor portions of it (portions that normally activate rus long terminal repeat (Loeken et al. 1988). Plasmid pRSV3- transcription of test promoters), contain such activation BglU, the control plasmid for the above constructs was gener­ regions and will not function as Gal-4 fusion proteins ated by removing the T antigen cDNA from pRSV-Tex by cleav­ age with Bgill and religation of the vector (Gruda and Alwine (Gruda et al. 1993). It has been proposed that the cellular 1991). Plasmid p6-lAL encoded the entire early coding region of target of the activating region of the Ela is the basic tran­ SV40 under the control of the SV40 early promoter (Keller and scriptional apparatus, most likely TFHD (Martin et al. Alwine 1985) and thus, is able to produce both large T and small 1990; Boyer and Berk 1993; Liu and Green 1994). These t antigen. Like p6-lAL, pT2811 (Bentivoglio et al. 1992) and data suggest that the interaction of Ela with the basal com­ pT(Rb")(same as pSG5-Kl; Kaelin et al. 1990) each encode the plex may be similar to that of an upstream bound transcrip­ entire early region of SV40; however, the large T antigens pro­ tion factor (activator protein) by providing an activation duced by each are mutant. The pT2811 large T antigen caimot domain for interaction with the basal transcription com­ bind p53 and the pT(Rb^) large T antigen cannot bind the Rb. plex, for example, TAFs. Conversely, our data suggests that The control plasmid for p6-lAL, pT2811, and pT(Rb") was the interaction of large T antigen with TFIID provides a pL16HX, which contains only the early SV40 promoter. A num­ ber of plasmids encoding mutants of T antigen, which have been new, TAF-like surface with which the activation domains characterized for transcriptional activation (Zhu et al. 1991) of many activator proteins may interact. It has been pro­ were used in transfection studies in tsl3 cells, these are listed in posed that the transcriptional effect of eukaryotic activator Table 1. Each mutation is in the context of the early region and proteins is to increase the assembly of the preinitiation therefore, the wild-type control plasmid was p6-lAL and the complex (Lin and Green 1991). Hence Ela and upstream null control was pL16HX. The mutant large T antigens pro­ boimd transcription factors would accomplish this by pro­ duced are all full or near full length and, similar to p6-lAL, each viding activation domains that interact with the basal tran­ plasmid has the SV40 origin of replication mutated so that am­ scription complex. However, large T antigen would accom­ plification cannot occur (Zhu et al. 1991). plish this by increasing the ability of other protein's acti­ Plasmid pCMV-TAFi,250 expressed human TAFi,250 from vation domains to interact with the basal transcription the human cytomegalovirus immediate early promoter (Wang complex. This model explains the results shown in Figure and Tjian 1994). The reporter plasmids, pCyclin A, containing 1. Ela activated the cycA promoter at the permissive tem­ the human cyclin A promoter upstream of the luciferase gene perature because it provided an activation domain. How­ and pcfos-CAT, containing the human c-/os promoter with the CAT gene have been described previously (Wang and Tjian ever, large T antigen failed to activate because it has no 1994). activation domain and its TAF-like function is redundant The six plasmids encoding the Drosophila TAFs have been at the permissive temperature in these cells. At the non- described previously (Hoey et al. 1993; Chen et al. 1994). They permissive temperature the ts defect in TAFQISO resulted contain cDNAs for either Drosophila TAFjj40, TAFn60, in conditions where the activation domain of Ela no longer TAFi,80, TAF„110, TAF„150, or TAF,i250, all under the control functioned; however, large T antigen activated transcrip­ of the bacteriophage T7 promoter (Hoey et al. 1993) for use in in tion because its TAF-Hke function rescued the ts defect. vitro transcription and translation systems. The plasmid encod­ ing Drosophila TAF,j250 contains the carboxy-terminal 180 amino acids and produces a protein migrating at 180 kD. Hu­ Model man TAF encoding plasmids pT3hTAFjil30 and phTAF100N13 were generously provided by Naoko Tanese (New York Univer­ The above data and discussion lead to a model where sity, NY) and pBSKShTAF32 was kindly provided by Robert large T antigen forms a stable complex with TFIID and Tjian (University of California, Berkeley).

1378 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of T antigen

Plasmids encoding fusion proteins between the glutathione- with NETN -I- and then incubated for another hour in 1 ml of binding site of glutathione-S-transferase and large T antigens NETN + containing 3% BSA to block nonspecific protein bind­ (full-length and pieces) or small t antigen have been described ing to the beads. TCA precipitable counts of in vitro translated previously (Gruda et al. 1993). TAFs (5x 10'^) were then added to the beads and incubated for 1 hr at 4°C. The beads were then washed five times with NETN -I-, boiled in SDS sample buffer, and separated by elec­ Cell culture, transfections, and infections trophoresis in 9% SDS-polyacrylamide gels. Proteins were vi­ The tsl3 cell line (Talavera and Basilico 1977) was propagated sualized by fluorography using Amplify (Amersham) and data and maintained in Dulbecco's minimal essential medium were quantitated using a Molecular Dynamics Phosphorlmager. (DMEM) supplemented with 10% fetal calf serum at 32°C in 5% COj. Cells (3x10^1 were seeded on 60-mm plates and grown overnight. Monolayers at ~70%-80% confluency were trans- Coimmunoprecipitations and Western analyses fected with 7 |xg of DNA by the calcium phosphate precipitation Five hundred micrograms of HeLa or a3 cell nuclear extracts procedure as described by Gruda et al. (1993). Two micrograms were used in each immunoprecipitation or coimmunoprecipi- of the cycA-luciferase reporter plasmid were used with varying tation experiment. Antibodies used included anti-T antigen amounts of effector plasmids. The respective fillers were used to monoclonal antibody Pab419; anti-HA antibody mAbl2CA5 normalize the amount of DNA used per transfection . After the from Boehringer Mannheim; and anti-TBP, anti-hTAFulOO, addition of DNA, cells were placed at the appropriate temper­ anti-hTAFijl30, anti-hTAFi,250 all from Santa Cruz Biotechnol­ atures, either 32°C for permissive conditions or 39°C for non- ogy. The control hnRNPCl antibody was a gift from Gideon permissive conditions. Cells were harvested 42-46 hr after Dreyfuss (University of Pennsylvania). Immunoprecipitations transfection. For chloramphenicol acetyltransferase (CAT) anal­ were performed using the method of Zhou et al. (1992) except ysis, cells were processed as described previously (Gruda et al. that binding was performed in 0.1 M KCl buffer D. Nuclear 1993). For luciferase analysis, cells were processed using the extract was allowed to incubate with the respective antibody (1 Luciferase Assay System (Promega) and procedures supplied by M.g) and 10 M-I of 50% vol/vol protein A-Sepharose beads (Phar­ the manufacturer. Mini-nuclear extracts from tsl3 cells were macia) for 8 hr at 4°C with constant mixing on a Nutator. The made according to the protocol by Lee et al. (1988). Cells were beads were then washed four times with RIPA buffer [150 mM transfected with either wild-type or mutant T antigens (Table 1) KCl, 1% NP-40, 50 mM Tris (pH 7.8), 0.5% deoxycholate, 0.1% and placed in the incubators at the respective temperatures. SDS], and then boiled in SDS-PAGE loading buffer. The eluted Cells were harvested 48 hr after transfection. For the cells at proteins were then separated by electrophoresis on a 12% SDS- 39°C, cells were scraped off the plates as well as pelleted from polyacrylamide gel and transferred to nitrocellulose. Specific the media in the plates. The two pellets were combined and proteins were detected by incubation with the appropriate an­ used in the mini-nuclear extract protocol. tibody (either the anti-T antigen, anti-HA or anti-TBP antibod­ HeLa cells were propagated and maintained in Iscove's me­ ies followed by visualization using the ECL luminescence kit; dium supplemented with 5% fetal calf serum at 37°C in 5% Amersham). CO2. The a3 cell line, a HeLa line that constitutively produces For coimmunoprecipitations from tsl3 cells at 32°C and an influenza HA epitope-tagged TBP (Zhou et al. 1992), was 39°C, the same procedure was used except in this case only 100 maintained in DMEM supplemented with 5% fetal calf and 200 \xg of nuclear extract was used in each coimmunoprecipitation |xg/ml G418 at 37°C in 5% COj. For infection with SV40, 9x 10^ reaction, and the immunoprecipitates were washed with 0.1 M HeLa or a3 cells were plated on 100-mm plates and grown over­ KCl buffer D four times, followed by two washes with RIPA night. The medium was then replaced with 1 ml of medium buffer. containing SV40 (WT strain 776) at an m.o.i. of 20 pfu/cell and incubated for 1 hr at 37°C with periodic rocking. Then fresh medium (9 ml) was added to each plate and the cells were in­ Phosphocellulose purification of TFIID cubated for 48 hr at 37°C. Infected and mock infected cells were harvested and nuclear extracts prepared by the procedure of Dig- Nuclear extract from SV40-infected a3 cells was prepared as nam et al. (1983a). The protein concentration of the extracts was described above and the nuclear extract was loaded onto a pre- determined using the Bradford assay. equilibrated phosphocellulose column at a concentration of 15 mg of nuclear extract per milliliter of column volume. Chro­ matography was performed as described by Dignam et al. Protein binding assay (1983b). The 1.0 M KCl fractions collected from the column Drosophila TAFs (dTAFs) 40, 60, 80, 110, 150, and 250(180) were TCA precipitated using 5 ixg of BSA as carrier protein, and were synthesized using the coupled in vitro transcription and the precipitates were separated on an 8% SDS-PAGE gel, which translation system (Promega). The proteins were labeled with was then transferred to nitrocellulose. The resulting Western [^^S]methionine and normalized for incorporation using TCA blot was then probed for the presence of SV40 large T antigen, precipitatable counts. eTBP, hTAFiiI30, and hTAFji250 sequentially without stripping Expression and purification of GST fusion proteins were done between probing. as described previously (Gruda et al. 1993). The amounts of GST To coimmunoprecipitate TBP and SV40 large T antigen from proteins used in the binding reactions were normalized for ex­ the 1.0 M KCl fractions, the fractions were first dialyzed into 0.1 pression using the Bradford protein assay and visually by silver M KCl buffer D and then coimmunoprecipitations were per­ staining of the gel. formed on each fraction using 2 |xg of anti-HA antibody and 50 All binding reactions were performed at 4°C with constant |xg of 50% slurry of protein A beads in PBS. The complexes were mixing on a Nutator. Glutathione agarose beads were bound allowed to incubate for 8 hr and were washed three times with with GST fusion proteins for 1 hr in NETN-(- [20 mM Tris (pH 0.1 M KCl buffer D plus 0.1 % NP40, twice with 0.5 M KCl buffer 7.5), 100 mM NaCl, 1 mM EDTA, 0.5% NP-40, 1 mM DTT, 1 mM D and twice with 0.7 M KCl buffer D. The beads were then phenylmethylsulfonyl fluoride (PMSF), 1 mM tosyl-lysine chlo- resuspended in SDS sample buffer and separated on an 8% SDS- romethyl ketone (TLCK)] The beads were washed three times PAGE gel that was transferred to nitrocellulose. The resulting

GENES & DEVELOPMENT 1379 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

Damania and Alwine

Western blot was probed for the presence of SV40 large T anti­ main complexes with the TATA box binding protein. Proc. gen. Natl. Acad. Sci. 91: 2488-2492. Gilinger, G. and J.C. Alwine. 1993. Transcriptional activation by simian virus 40 large T antigen: Requirement for simple Acknowledgments promoter structures containing either TATA or Initiator el­ ements with variable upstream factor binding sites. /. Virol. We thank Edith Wang and Robert Tjian for plasmids, the tsl3 67: 6682-6688. cell line and valuable advice; Naoko Tanese for the hTAF„100 Gruda, M. and J.C. Alwine. 1991. Simian virus 40 T-antigen and hTAFijlSO expressing plasmids; Arnold Berk for the a3 cell transcriptional activation mediate through the Oct/SPH re­ line and helpful discussions; Charles Cole for plasmids and gion of the SV40 late promoter. /. Virol. 65: 3553-3558. helpful discussions; David Livingston for plasmids. In addition, we thank David Lukac for his assistance and the other members Gruda, M., J. Zabolotny, J. Xiao, I. Davidson, and J. Alwine. of the Alwine Laboratory for their help and comments on the 1993. Transcriptional activation by SV40 large T antigen: manuscript. Cheers to all. Interactions with multiple components of the transcription complex. Mol. Cell. Biol. 13: 961-969. The publication costs of this article were defrayed m part by Hayashida, T., T. Sekiguchi, E. Noguchi, H. Sunamuto, T. payment of page charges. This article must therefore be hereby Ohba, and T. Nishimoto. 1994. The CCG1/TAF„250 gene is marked "advertisement" in accordance with 18 USC section mutated in thermosensitive Gl mutants of BHK21 cell line 1734 solely to indicate this fact. derived from golden hamsters. Gene 141: 267-270. Hisatahe, K., S. Hasegawa, R. Takada, Y. Nakatani, M. Horiko- shi, and R.G. Roeder. 1993. The p250 subunit of native References TATA box-binding factor TFIID is the cell-cycle regulatory Beard, P. and H. Bruggmann. 1989. Control of transcription in protein CCGl. Nature 362: 179-181. vitio from simian virus 40 promoters by proteins from viral Hoey, T., R. Weinzieri, G. Gill, J. Chen, B. Dynlacht, and R. minichromosomes. Curr. Top. Micro. Immunol. 144: 47-54. Tjian. 1993. Molecular cloning and functional analysis of Bentivoglio, C, J. Zhu, and C.N. Cole. 1992. Mechanisms of drosophila TAFllO reveal properties expected of coactiva- interference with simian virus 40 (SV40) DNA replication by tors. Cell 72: 247-260. trans-dominant mutants of SV40 large T antigen. /. Virol. Kaelin, W.G. Jr., M. Ewen, and D.M. Livingston. 1990. Defini­ 66:2551-2555. tion of the minimal simian virus 40 large T antigen- and Boyer, T.G. and A.J. Berk. 1993. Functional interaction of ade­ adenovirus ElA-binding domain in the retinoblastoma gene novirus ElA with holo-TFIID. Genes & Dev. 7: 1810-1823. product. Mol. Cell. Biol. 10: 3761-3769. Casaz, P., R. Sundseth, and U. Hansen. 1991. Trans-activation Keller, J.M. and J.C. Alwine. 1985. Analysis of an activatable of the simian virus 40 late promoter by large T antigen re­ promoter: Sequences in the SV40 late promoter required for quires binding sites for the cellular transcription factor TEF- T-antigen-mediated trans activation. Mol. Cell. Biol. 5: 1859- 1. /. Virol. 65:6535-6543. 1869. Casaz, P., P.W. Rice, C.N. Cole, and U. Hansen. 1995. A TEF- Kelly, J. and A.G. Wildeman. 1991. Role of the SV40 enhancer in 1-independent mechanism for activation of the simian virus the early to late shift in viral transcription. Nucleic Acids 40 (SV40) late promoter by mutant SV40 large T antigen. /. Res. 19: 6799-6804. Virol. 69: 3501-3509. Lai, J. and W. Herr. 1992. Ethidium bromide provides a simple Chen, J-L., L. Attardi, C.P. Verrijzer, K. Yokomori, and R. Tjian. tool for establishing genuine DNA-independent protein as­ 1994. Assembly of recombinant TFIID reveals different co- sociations. Proc. Natl. Acad. Sci. 89: 6958-6962. activator requirements for distinct transcriptional activa­ Lee, K.W., A. Bindereif, and M. Green. 1988. A small-scale pro­ tors. Cell 79: 93-105. cedure for preparation of nuclear extracts that support effi­ Dignam, f., R. Lebovitz, and R. Roeder. 1983a. Accurate tran­ cient transcription and pre-mRNA splicing. Gene. Anal. scription initiation by RNA polymerase II in a soluble ex­ Technol. 5: 22-31. tract from isolated mammalian nuclei. Nucleic Acids Res. Lin, Y.S. and M.R. Green. 1991. Mechanism of action of an 11:11475-1489. acidic transcriptional activator in vitro. Cell 64: 971-981. Dignam, J.D., P.L. Martin, B.S. Shastry, and R. Roeder. 1983b. Liu, H.T., C. Gibson, R. Hirschhom, S. Rittling, R. Baserga, and Eucaryotic gene transcription with purified components. W. Mercer. 1985. Expression of thymidine kinase and dihy- Methods Enzymol. 101: 582-598. drofolate reductase genes in mammalian ts mutants of the Floros, J., G. fonak, N. Galanti, and R. Baserga. 1981. Induction cell cycle. /. Biol. Chem. 260: 3269-3274. of cell DNA replication in Gl-specific ts mutants by micro­ Liu, F. and M.R. Green. 1994. Promoter targeting by adenovirus injection of SV40 DNA. Exp. Cell Res. 132: 215-223. Ela through interaction with different cellular DNA-binding Frost, J.A., A.S. Alberts, E. Sontag, K. Guan, M.C. Mumby, and domains. Nature 368: 520-525. J.R. Feramisco. 1994. Simian virus 40 small t antigen coop­ Loeken, M. 1992. Simian virus 40 small t antigen trans activates erates with mitogen-activated kinases to stimulate AP-1 ac­ the adenovirus E2A promoter by using mechanisms distinct tivity. Mol. Cell. Biol. 14: 6244-6252. from those used by adenovirus ElA. /. Virol. 66: 2551-2555. Gallo, G.J., G. Gilinger, and J.C. Alwine. 1988. Simian virus 40 Loeken, M., G. Khoury, and J. Brady. 1986. Stimulation of the T antigen alters the binding characteristics of specific sim­ adenovirus E2 promoter by simian virus 40 T antigen and ian DNA-binding factors. Mol. Cell. Biol. 8: 1648-1656. Ela occurs by different mechanisms. Mol. Cell. Biol. Gallo, G.J., M.C. Gruda, J.R. Manuppello, and J.C. Alwine. 1990. 6: 2020-2026. Activity of simian DNA-binding factors is altered in the Loeken, M., I. Bikel, D.M. Livingston, and J. Brady. 1988. Trans- presence of simian virus 40 (SV40) early proteins: Character­ activation of RNA polymerase II and III promoters by SV40 ization of factors binding to elements involved in activation small t antigen. Cell 55: 1171-1177. of the SV40 late promoter. /. Virol. 64: 173-184. Martin, K.J., J.W. LiUie, and M.R. Green. 1990. Evidence for Geisberg, J.V., W.S. Lee, A.J. Berk, and R.P. Ricciardi. 1994. The interaction of different eukaryotic transcriptional activators zinc finger region of the adenovirus ElA transactivating do­ with distinct cellular targets. Nature 346: 147-152.

1380 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAP-like function of T antigen

Noguchi, E., T. Sekiguchi, Y. Nohiro, T. Hayashida, E. Hirose, N. Hayashi, and T. Nishimoto. 1994. Minimal essential re­ gion of CCGl/TAFn250 required for complementing the ts cell cycle mutants, tsBN462 and tsl3 cells, of hamster BHK21 cells. Som. Cell Mol. Genet. 20: 505-513. Pallas, D., L. Shahrik, D. Martin, S. Jaspers, T. Miller, D. Brauti- gan, and T. Roberts. 1990. Polyoma small and middle T an­ tigens and SV40 small t antigen form stable complexes with protein phosphatase 2A. Cell 60: 167-176. Rice, P.W. and C.N. Cole. 1993. Efficient transcriptional acti­ vation of many simple modular promoters by SV40 large T antigen. /. Virol. 67: 6689-6697. Ruediger R., D. Roeckel, J. Fait, A. Bergquist, G. Magnusson, and G. Walter. 1992. Identification of binding sites on the regulatory A subunit of protein phosphatase 2A for the cat­ alytic C subunit and for tumor anitgens of simian virus 40 and polyomavirus. Mol. Cell. Biol. 12: 4872-4882. Ruppert, S., E.H. Wang, and R. Tjian. 1993. Cloning and expres­ sion of human TAFii250: A TBP-associated factor implicated in cell-cycle regulation. Nature 362: 175-179. Scheidtraann, K., M. Mumby, K. Rundell, and G. Walter. 1991. Dephosphorylation of simian virus 40 large T antigen and p53 protein by protein phosphatase 2A inhibition by small t anitgen. Mol. Cell. Biol. 11: 1996-2003. Sekiguchi, T., T. Miyata, and T. Nishimoto. 1988. Molecular cloning of the cDNA of human X chromosomal gene (CCGl) w^hich complements the temperature-sensitive Gl mutants, tsBN62 and tsl3 of the BHK cell line. EMBO J. 7: 1683-1687. Sontag E., S. Fedorov, S. Kamibayashi, D. Robbins, M. Cobb, and M. Mumby. 1993. The interaction of SV40 small tumor an­ tigen with protein phosphatase 2A stimulates the MAP ki­ nase pathway and induces cell proliferation. Cell 75: 887- 897. Talavera, A. and C. Basilico. 1977. Temperature sensitive mu­ tants of BFIK cells affected in cell cycle progression. /. Cell. Physiol. 92: 425-436. Trifillis, P., J. Picardi, and J.C. Alwine. 1990. SV40 T antigen can transcriptionally activate and mediate DNA replication in cells which lack retinoblastoma susceptibility gene product. /. Virol. 64: 1345-1347. Wang, E. and R. Tjian. 1994. Promoter-selective transcriptional defect in cell cycle mutant tsl3 rescued by hTAFii250. Sci­ ence 263: 811-814. Wheat, W.H., W.J. Roesler, and D.J. Klemm. 1994. Simian virus 40 small tumor antigen inhibits dephosphorylation of pro­ tein kinase A-phosphorylated CREB and regulates CREB transcriptional stimulation. Mol. Cell. Biol. 14: 5881-5890. Zhou, Q., P. Lieberman, T. Boyer, and A. Berk. 1992. Holo- TFIID supports transcriptional stimulation by diverse acti­ vators and from a TATA-less promoter. Genes & Dev. 6:1964-1974. Zhu, J., P.W. Rice, M. Chamberlain, and C.N. Cole. 1991. Map­ ping the transcriptional transactivation function of simian vims 40 large T antigen. /. Virol. 65: 2778-2790.

GENES & DEVELOPMENT 1381 Downloaded from genesdev.cshlp.org on September 24, 2021 - Published by Cold Spring Harbor Laboratory Press

TAF-like function of SV40 large T antigen.

B Damania and J C Alwine

Genes Dev. 1996, 10: Access the most recent version at doi:10.1101/gad.10.11.1369

References This article cites 45 articles, 25 of which can be accessed free at: http://genesdev.cshlp.org/content/10/11/1369.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.

Copyright © Cold Spring Harbor Laboratory Press