TFII is required for transcription of the naturally TATA-less but initiator-containing Vbeta promoter.

The proximal or core promoter of a typical eukaryotic protein coding gene comprises distinct elements, TATA and/or initiator (Inr). The existence of TATA or Inr at the core promoter suggests that the mechanism of transcription initiation mediated by these two genetic elements may be different. Accordingly, it has been demonstrated that the transcriptional requirements for the TATA-containing, Inr-less (TATAInr) promoters are different from the transcriptional requirements for the TATA-less, Inr-containing (TATAInr) promoters. Although both types of promoters require the transcription initiation factor (TFIID) in addition to other common initiation factors, a TATAInr promoter requires accessory component(s). Here we have employed in vitro analyses to address the transcription factor requirements for a TATAInr promoter. We demonstrate that in addition to TFIID, a naturally occurring TATAInr promoter requires TFII-I, an Inr element-dependent transcription factor. Consistent with its Inr element-dependent activities, TFII-I is dispensable for a TATAInr promoter. Furthermore, we demonstrate that both TFII-I and TFIID activities in nuclear extracts are temperature-sensitive. However, TFII-I is heat-inactivated at temperatures lower than that required to inactivate TFIID. Therefore, differential heat treatment of nuclear extracts provides an assay to discriminate between transcriptional requirements at TATAInr and TATAInr promoters.

The proximal or core promoter of a typical eukaryotic protein coding gene comprises distinct elements, TATA and/or initiator (Inr). The existence of TATA or Inr at the core promoter suggests that the mechanism of transcription initiation mediated by these two genetic elements may be different. Accordingly, it has been demonstrated that the transcriptional requirements for the TATA-containing, Inr-less (TATA ؉ Inr ؊ ) promoters are different from the transcriptional requirements for the TATA-less, Inr-containing (TATA ؊ Inr ؉ ) promoters. Although both types of promoters require the transcription initiation factor (TFIID) in addition to other common initiation factors, a TATA ؊ Inr ؉ promoter requires accessory component(s). Here we have employed in vitro analyses to address the transcription factor requirements for a TATA ؊ Inr ؉ promoter. We demonstrate that in addition to TFIID, a naturally occurring TATA ؊ Inr ؉ promoter requires TFII-I, an Inr element-dependent transcription factor. Consistent with its Inr element-dependent activities, TFII-I is dispensable for a TATA ؉ Inr ؊ promoter. Furthermore, we demonstrate that both TFII-I and TFIID activities in nuclear extracts are temperature-sensitive. However, TFII-I is heat-inactivated at temperatures lower than that required to inactivate TFIID. Therefore, differential heat treatment of nuclear extracts provides an assay to discriminate between transcriptional requirements at TATA ؉ Inr ؊ and TATA ؊ Inr ؉ promoters.
Transcription initiation of protein coding genes is brought about by RNA polymerase II and a set of general transcription initiation factors (1)(2)(3). These factors, in a combinatorial fashion, can direct transcription initiation of a variety of eukaryotic promoters in an in vitro assay (1)(2)(3). The core promoter region of a typical eukaryotic gene consists of a TATA box and/or an Inr 1 element (4,5). The presence of distinct core promoter elements in different genes suggests distinct transcriptional strategies. However, the differences in mechanism of transcriptional initiation mediated by TATA or Inr elements have yet to be elucidated. Biochemical complementation assays employing heat-treated nuclear extracts have demonstrated that the tran-scription factor requirements for a TATA ϩ Inr Ϫ promoter are different from a TATA Ϫ Inr ϩ promoter (6). The TATA-binding transcription factor complex TFIID is required for both promoters (7), and heat treatment of nuclear extracts (at 49°C for 15 min) renders the TFIID inactive (6,8). Hence, a heat-treated nuclear extract was incapable of transcribing either a TATAcontaining or a TATA-less promoter unless supplemented with exogenous TFIID (6). Exogenously added TFIID could restore only a TATA ϩ Inr Ϫ promoter activity but not a TATA Ϫ Inr ϩ promoter activity, suggesting that in addition to TFIID, another heat-sensitive component(s) was required for the TATA Ϫ Inr ϩ promoter (6). The mechanism of action of this component is unclear. Because TFIID is required for a TATAless promoter, it is possible that such a factor may serve to anchor TFIID to a TATA Ϫ Inr ϩ promoter in the absence of a TATA box (9).
Transcription factor requirements for TATA Ϫ Inr ϩ promoters are controversial. Several factors have been reported as Inr element-binding proteins including TFII-I (10 -12), USF (10,13), YY1 (14,15), RNA polymerase II (16), and a member of the TBP-associated factors (TAF; 6,[17][18][19][20], although other studies have suggested that a TAF may not bind directly to an Inr element (21). However, these observations may not be mutually exclusive. The differences possibly indicate redundancy in Inr element-mediated interactions, which may be condition-and/or promoter context-dependent. In addition, it is possibile that structurally (and perhaps functionally) different classes of Inr elements exist.
We wish to elucidate the molecular mechanisms of transcription initiation mediated via an Inr element in TATA Ϫ Inr ϩ promoters. Here, we report that TFII-I (10 -12) is necessary for transcription of a naturally occurring TATA Ϫ Inr ϩ but not for a TATA ϩ Inr Ϫ promoter. For our analyses, we have used the T cell receptor variable region-derived (V␤) promoter (22) as a model TATA Ϫ Inr ϩ promoter and the B cell immunoglobulin heavy chain-derived (IgH) promoter (23, 24) as a model TATA ϩ Inr Ϫ promoter. We provide three lines of evidence in support of a requirement of TFII-I for the TATA Ϫ Inr ϩ V␤ promoter. In each case we have selectively blocked the transcription of the V␤ promoter and subsequently restored its transcriptional activity by exogenous addition of purified TFII-I. We demonstrate that: 1) Immunodepletion of nuclear extracts with an anti-TFII-I antibody completely abrogates transcription of the TATA Ϫ Inr ϩ V␤ promoter, which is restored by addition of purified TFII-I. Importantly, these antibodies have no effect on the TATA ϩ Inr Ϫ IgH promoter. 2) TFII-I binds specifically to the V␤ Inr element. Thus, an oligonucleotide containing the wild type V␤ Inr element sequences efficiently inhibits V␤ transcription; exogenously added TFII-I restores V␤ transcription. A control oligonucleotide containing the mutant V␤ Inr sequence does not inhibit V␤ transcription. 3) In addition to TFIID, TFII-I is temperature-sensitive. Thus, heat treatment of nuclear extracts impairs both TFII-I and TFIID activities. However, the two activities are affected at different temperatures. Heat treatment of nuclear extracts at 42°C ablates TFII-I but not TFIID activity, whereas heat treatment at 49°C destroys TFIID activity as well. Transcriptional complementation assays using heat-treated nuclear extracts demonstrate that although TFIID is both necessary and sufficient for a TATA ϩ Inr Ϫ promoter, TFII-I is additionally required for the TATA Ϫ Inr ϩ V␤ promoter.

Nuclear Extracts
Jurkat and HeLa cells were grown in culture, and 3 liters (3 ϫ 10 10 cells) were harvested to prepare nuclear extracts as described (25). Protein concentrations for each nuclear extract were determined spectrophotometrically via Bio-Rad protein assay.

Heat Inactivation of Nuclear Extracts
Nuclear extracts (50 l) were aliquoted and incubated at 42°C for 6 or 15 min as specified in Fig. 4. Following heat treatment, the extracts were centrifuged for 1 min, and the supernatants were placed in fresh tubes for immediate use in transcription.

Immunodepletion of Nuclear Extracts
The anti-TFII-I antibody was raised in rabbits against a synthetic peptide corresponding to the putative DNA binding domain of TFII-I. 2 The polyclonal serum was obtained from a 10-week bleed. Immunodepletion of nuclear extracts was achieved in two ways. One method involved incubation of an extract with either the preimmune serum or with the immune serum (anti-TFII-I antibody) at 30°C for 10 min prior to starting the transcription reaction. The other method included the additional steps of applying the mixture of nuclear extract and serum to protein A-Sepharose beads and incubating at 0°C for 30 min; the mixture was centrifuged at low speed for 5 min, and the supernatant was retained for transcription. Differences were not observed between these methods. In addition, immunodepletion of TFII-I was accomplished by passing nuclear extracts over an immobilized anti-TFII-I antibody column, which yielded identical results (not shown). All of the above described treatments were done immediately preceding the transcription reactions.

SDS-PAGE, Western Blot Analyses, and Peptide Block
A purified preparation of TFII-I (100 ng) and either undepleted, TFII-I-depleted, or mock-depleted Jurkat nuclear extract (10 g in each case) was subjected to SDS-PAGE (7.5%) and subsequently transferred to nitrocellulose by Western blot technique. The blotted proteins were probed with the anti-TFII-I antibody (1:2500 dilution) and visualized using ECL technique (Amersham Corp.). Similar methods were employed for visualization of immune precipitates.
For peptide block experiments, the anti-TFII-I antibody was preincubated for 30 min at 0°C prior to probing with 2.5, 0.25, or 0.025 mg/ml of the synthetic peptide derived from the putative DNA binding region of TFII-I.

Purification of Transcription Factors
TFIID-All procedures were performed at 4°C. HeLa nuclear extract was chromatographed over a heparin-Sepharose (Pharmacia Biotech Inc.) column equilibrated in buffer A containing 20 mM Tris, pH 7.9, 0.2 mM EDTA, 5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 10% (v/v) glycerol, and 100 mM KCl; the TFIID activity was eluted in buffer A containing 500 mM KCl. These fractions were pooled together and dialyzed against buffer A (100 mM KCl). The dialyzed pool was loaded onto a DEAE-52 (Whatman) column equilibrated in buffer A (100 mM). The column was developed with a linear gradient in buffer A containing 100 -500 mM KCl. TFIID activity was recovered at about 250 mM KCl. The fractions containing TFIID activity were pooled once again, dialyzed against buffer A (100 mM KCl), and loaded onto a phosphocellulose (P-11, Whatman) column equilibrated in buffer A (100 mM). TFIID activity was recovered at 850 mM KCl. The active fractions were pooled, dialyzed against buffer A (100 mM), aliquoted, and frozen at Ϫ80°C.
TFII-I-TFII-I (p120) polypeptide was visualized at each chromato-graphic step by Western blot analyses using an anti-TFII-I antibody. All chromatographic steps were done at 4°C in buffer A containing 20 mM Tris, pH 7.9, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10% (v/v) glycerol, 0.2 mM EDTA, and KCl as indicated below. HeLaderived nuclear extract dialyzed against 100 mM buffer A was subjected to chromatography on phosphocellulose column (Whatman, P-11). TFII-I (p120) eluted in the 300 mM KCl fraction. The 300 mM fraction was dialyzed against 40 mM buffer A, loaded onto a DEAE-52 (Whatman) column, and subjected to a linear gradient elution with 40 -500 mM KCl in buffer A. TFII-I-containing fractions eluted between 100 -120 mM salt. These fractions were pooled together and loaded on to a Mono-S fast protein liquid chromatography (HR 5/5, Pharmacia) column without dialysis. The column was developed with a 100 -500 mM linear salt gradient in buffer A. TFII-I-containing fractions were eluted between 300 and 320 mM salt. These fractions were pooled, dialyzed to 100 mM salt in buffer A, and used for EMSA and transcriptional assays.

EMSA
EMSA was done with either the adenovirus major late (AdML, Fig.  1a) Inr element containing probe or the V␤ Inr element containing probe (Figs. 2a and 3a). The AdML probe contained sequences from Ϫ22 to ϩ43 (10), whereas the V␤ probe contained sequences from Ϫ27 to ϩ 13 (V␤5. 2,22). Both probes were labeled with [␣-32 P]dCTP (3000 Ci/ mmol) using Klenow fragment. Where indicated in the figures, the nuclear extracts were treated with an anti-TFII-I antibody (or control antibodies) for 10 min at 0°C prior to the addition of the probe. Nuclear extracts (1 g) or purified TFII-I (100 ng) were incubated at 30°C for 30 min after addition of the probe (approximately 10 -15 fmol in each reaction). The final reaction volume was 20 l in buffer A with 80 mM KCl and 500 ng poly(dA:dT) as a carrier. All reactions were subjected to electrophoresis through a 5% native polyacrylamide gel containing 5% glycerol in 0.5 ϫ TBE (40 mM Tris, pH 7.6, 40 mM boric acid, 2 mM EDTA) for 3 h at 140 V.

In Vitro Transcription
Each reaction contained 30 g of Jurkat nuclear extract (in buffer A) and 500 ng of either linearized (KpnI) V␤5.2 template (kind gift from Dr. Loh) or pMU-(Ϫ47)-IgH G-less cassette template (22,23). The nuclear extracts, where indicated, were immunodepleted of TFII-I with the anti-TFII-I antibody or mock-depleted with the preimmune serum, and, as indicated, also preincubated with 100 ng of partially purified TFII-I for 10 min at 30°C in buffer A preceding transcription. For oligonucleotide competition assays, approximately 125 ng of competitor oligonucleotides were added to the nuclear extract, along with purified TFII-I where indicated, and preincubated for 15 min at 30°C preceding transcription. The competitor oligonucleotides contained the following sequences: V␤ Inr (wt ACTCTTCTTC; mut AGCCGGACGG), E-box (CACGTG) (9), and AdML TATA element (wt TATAAAA; mut GC-TATTT). All transcription reactions were done in a final volume of 20 l and included 15 mM Tris, pH 7.9, 7. . All reactions were incubated for 60 min, including the specified preincubation times, at 30°C and then stopped by the addition of 400 l of Stop Mix (8 M urea, 10 mM Tris, pH 7.8, 10 mM EDTA, 0.5% SDS, 100 mM LiCl, 100 g/ml tRNA, 300 mM NaOAc). The transcripts were then extracted with phenol-chloroform, precipitated with 2-propanol, and washed with 70% (v/v) ethanol. The pellets were resuspended in formamide loading buffer and run on a 6% polyacrylamide/8 M urea gel in 1 ϫ TBE for 2 h at 250 V. The gel was fixed with 10% (v/v) acetic acid, dried, and subjected to autoradiography.

V␤ Transcription Requires TFII-I: Antibody-mediated Block of TFII-I Activity-
To test the effects of anti-TFII-I antibody, we employed the antibody in an EMSA. Two different nuclear extracts (HeLa and Jurkat) were tested for their ability to bind an Inr element (Fig. 1a). Both nuclear extracts gave a predominant band (lanes 1 and 2) that comigrated with a purified preparation of TFII-I (data not shown). Importantly, the HeLa and Jurkat nuclear extract-derived TFII-I band was abrogated by the anti-TFII-I antibody (␣I, lanes 3 and 5). But a control antibody (anti-TBP antibody, ␣T, lanes 4 and 6) had no effect on TFII-I binding.
Sequence comparison between different promoters reveals that the V␤ promoter contains a consensus Inr element (Fig.  1b). The V␤ promoter (kind gift from Dr. D. Loh) is typically expressed in T cells (22). Thus, we employed a T cell-derived nuclear extract (Jurkat) for all of the in vitro transcriptional assays. We used the anti-TFII-I antibody to deplete TFII-I from a transcriptionally competent Jurkat nuclear extract (Fig. 1c). Employing an undepleted nuclear extract (lane 1), the run-off assay from the linearized V␤ promoter (containing wild type sequences from Ϫ480 to ϩ260) produced an accurately initiated 260-nucleotide major transcript. Mock depletion of Jurkat nuclear extract with a control antibody (preimmune serum) had no appreciable effects on transcription (lane 2). However, immunodepletion with an anti-TFII-I antibody severely impaired the V␤ transcription (lane 3).
To demonstrate that the antibody treatment causes depletion of only TFII-I, we added back TFII-I, exogenously, to an immunodepleted Jurkat nuclear extract (Fig. 1d). Antibody- Therefore, an anti-TFII-I antibody inhibits transcription from a TATA Ϫ Inr ϩ promoter, which can be restored by exogenous addition of TFII-I.
As a control for promoter specificity, we employed the TATA ϩ Inr Ϫ IgH (23, 24) promoter (Fig. 1e). The minimal (Ϫ47 to ϩ1) IgH promoter does not exhibit any tissue type specificity in vitro and therefore can be transcribed by a T cell (Jurkat) nuclear extract. Most importantly, TFII-I depletion of a Jurkat nuclear extract did not affect the TATA ϩ Inr Ϫ IgH promoter. Thus, the level of the 400-nucleotide transcript produced from the IgH promoter remains unaltered in undepleted (lane 1), mock-depleted (lane 2), and immunodepleted (lane 3) Jurkat nuclear extracts. These data clearly demonstrate that TFII-I is required for a TATA Ϫ Inr ϩ promoter but not for a TATA ϩ Inr Ϫ promoter.
Specificity of the Anti-TFII-I Antibody: Peptide-mediated Block of the Antibody-We further established the specificity of the anti-TFII-I antibody. Western blot analysis of Jurkat nuclear extract was carried out and probed with the antibody (Fig.  2a). A purified preparation of TFII-I (120 kDa) was used as a control. This experiment revealed that the antibody, under the assay conditions, recognizes only TFII-I in nuclear extracts. The other cross-reactive species (125 kDa) is a modified (and inactive) form of the 120-kDa form. 3 Most significantly, the antibody reactivity was completely abolished upon treatment with either 2 or 0.2 mg/ml but not with 0.02 mg/ml of the TFII-I-derived synthetic peptide (Fig. 2a). These data suggest that the antibody reacts specifically with TFII-I. Finally, we analyzed the nuclear extract before and after immunodepletion (Fig. 2b) with either anti-TFII-I or preimmune sera together with the corresponding immune precipitates (Fig. 2c). These analyses revealed that immunodepletion of Jurkat nuclear extract led to a substantial decrease in TFII-I, whereas the mock depletion had no appreciable effect compared with an untreated extract. Furthermore, the immune precipitate obtained from TFII-I depletion had a significant amount of TFII-I, whereas the precipitate obtained from mock depletion had negligible TFII-I. Taken together, these data revealed that the anti-TFII-I antibody primarily recognizes TFII-I and that the 3 C. D. Novina and A. L. Roy, submitted for publication. 1 and 2) demonstrates a major complex that is blocked by an anti-TFII-I antibody (␣I; lanes 3 and 5) but not by an anti-TBP antibody (␣T; lanes 4 and 6). A preimmune serum had no effect on TFII-I mobility shift (not shown). For this EMSA we used an AdML promoter-derived Inr element. Identical results were obtained with all Inr elements tested under our conditions. b, sequence comparison of different Inr elements. The initiating nucleotide is indicated by the arrow. c, in vitro transcription using a linearized V␤ template. The run-off transcript from the V␤ template was not affected by mock depletion of Jurkat nuclear extract with a preimmune (pI) serum (com- pare lanes 1 and 2). The transcription was abolished completely upon immune depletion with the anti-TFII-I antibody (I, lane 3). d, the anti-TFII-I antibody (I) severely decreased the V␤ transcription (com- pare lanes 1 and 2). However, the transcription was restored completely upon exogenous addition of a purified preparation of TFII-I (lane 3). As before, a preimmune (pI) serum had negligible effects on the V␤ transcription (lane 4). e, immunodepletion of TFII-I from Jurkat nuclear extract had no significant effect on IgH transcription (compare lanes 1  and 3). Similarly, the control antibody (preimmune serum, pI) had no effects on IgH transcription (lane 2) .   FIG. 2. Specificity of the anti-TFII-I antibody. a, the anti-TFII-I antibody specifically recognized p120/TFII-I in Jurkat nuclear extract (lane 1) and in purified TFII-I (lane 2). The cross-reactive 125-kDa band in Jurkat nuclear extract is a modified form of TFII-I (not shown). Importantly, the antibody reactivity can be blocked by pretreatment with the antigenic peptide, either 2.5 (lanes 3 and 4) or 0.25 mg/ml (lanes 5 and 6) but not with 0.025 mg/ml (lanes 7 and 8). b, antibody depletion (TFII-I dep.) leads to removal of TFII-I from a Jurkat nuclear extract but mock depletion (mock dep.) has no appreciable effect compared with undepleted extract (undep.). c, the immune precipitate (␣I ppt) obtained from antibody depletion shows the presence of substantial amount of TFII-I, whereas the precipitate from mock depletion (pI ppt) has no significant amount of TFII-I.

FIG. 1. Immunodepletion of TFII-I affects a TATA -Inr ؉ (V␤) but not a TATA ؉ Inr -(IgH) promoter. a, EMSA of HeLa and Jurkat nuclear extracts (lanes
immunodepletion specifically leads to removal of TFII-I from the extract.

V␤ Transcription Requires Binding of TFII-I to the Inr Element: Oligonucleotide-mediated Block of TFII-I Activity-
In order to correlate the DNA binding and transcriptional activities of TFII-I at the V␤ Inr element, transcription-coupled competitor challenge assays were done. First we demonstrated that a purified preparation of TFII-I specifically binds to the V␤ Inr element (Fig. 3a, lane 2). A wt V␤ Inr sequence containing oligonucleotide could compete for TFII-I binding (lane 3). A mut oligonucleotide could not compete for TFII-I binding (lane 4). These oligonucleotides were then employed in transcriptional assays.
Consistent with the DNA binding analyses, the wild type Inr oligonucleotide competitively inhibited transcription from the V␤ promoter in nuclear extracts (Fig. 3b, lanes 3 and 8), whereas the mutant oligonucleotide failed to inhibit transcription (lane 4). Most importantly, the Inr oligonucleotide-mediated inhibition of transcription was restored upon the addition of TFII-I (lane 9). As an additional control of DNA binding specificity, an E-box containing oligonucleotide was used (lane 2). Although TFII-I can bind to both an Inr element and an E-box (dual specificity), an E-box containing oligonucleotide cannot block TFII-I binding to the Inr element and vice versa (10). Accordingly, the E-box containing oligonucleotide failed to abrogate Inr-dependent transcription. Taken together, these results demonstrate that TFII-I binds specifically to the V␤ Inr element and that this binding is necessary for the V␤ transcription initiation.
TFIID has been shown to be required for transcriptional activity of TATA-containing as well as TATA-less promoters (7,9). Consistent with this notion, an oligonucleotide containing only a wild type (lane 5) but not a mutant TATA box (lane 6) competitively inhibited the V␤ transcriptional activity. It is important to note that TFIID is necessary but not sufficient to direct Inr-dependent V␤ transcription (see below).
Inr Element-dependent Activity of TFII-I Is Heat-labile: Temperature-mediated Block of TFII-I Activity-Heat treatment of nuclear extracts interferes with TFIID activity leading to transcriptional inhibition of both TATA-containing and Inr-containing promoters (6). The inhibition of a TATA ϩ Inr Ϫ promoter can be rescued by exogenous addition of TFIID (6). However, inhibition of a TATA Ϫ Inr ϩ promoter cannot be rescued by the addition of TFIID (6), suggesting that these promoters require additional heat-sensitive component(s). Because TFII-I is required specifically for a TATA Ϫ Inr ϩ promoter, we tested whether TFII-I is heat-sensitive. We heat treated nuclear extracts at various temperatures and for varying periods of time to monitor the heat sensitivity of TFII-I (data not shown). Our analyses demonstrated that the TFII-I binding activity (confirmed by an anti TFII-I antibody, Fig. 4a, compare lanes 1 and  2) in a nuclear extract was abrogated by heat treatment of the nuclear extract minimally at 42°C for 6 min (lane 3). Similarly heat treatment at 42°C for 15 min also abolished TFII-I binding (lane 4).
Next, we employed the heat-treated nuclear extracts in in vitro transcriptional assays with the V␤ (Fig. 4b) and IgH (Fig.  4c) promoters. Heat treatment of a Jurkat nuclear extract at 42°C for 6 min led to abrogation of the V␤ transcription (Fig.  3a, lane 2). Surprisingly, the addition of TFII-I did not alleviate the transcriptional block (lane 3), suggesting that this heat treatment interfered with additional component(s). The Addition of a partially purified TFIID fraction in the absence of exogenous TFII-I did not rescue the transcriptional block either (lane 4). However, the addition of both TFII-I and TFIID simultaneously rescued completely the heat-induced block of V␤ transcription (lane 5), suggesting that the additional component was present in the TFIID fraction. Therefore, in addition to TFII-I and TFIID, a third component, which is present in the TFIID fraction, is required for a TATA Ϫ Inr ϩ promoter. Although the exact identity of this component is presently un -FIG. 3. Binding of TFII-I to the Inr element is essential for TATA -Inr ؉ transcription. a, a purified preparation of TFII-I (HeLaderived) binds to the V␤ Inr element (lane 2). The binding was specific because it was competitively inhibited by a wt V␤ Inr element containing oligonucleotide (lane 3) but not by a mutant oligonucleotide (mut, lane 4). b, V␤ transcription by Jurkat nuclear extract (lanes 1 and 7) was competitively inhibited by a wt Inr oligonucleotide (lanes 3 and 8). known, the existence of such a component (activity) has been described before (6). Similar results were obtained when nuclear extracts were heat treated at 42°C for 15 min (not shown). Furthermore, only TFIID (and not TBP) was effective in these complementation assays (not shown).
TFII-I is an Inr element-dependent factor and therefore is not required for an Inr-less promoter. Accordingly, mild heat treatment (42°C for 6 min) of nuclear extracts did not abolish the TATA ϩ Inr Ϫ IgH promoter activity (Fig. 4c). In fact, we observed a reproducible increase in the IgH promoter activity upon mild heat treatment (compare lanes 1 and 2). Because TFIID activity was not affected at 42°C, the addition of TFIID had no effect on IgH transcription at this temperature (lane 3). Similarly, heat treatment of nuclear extracts at 42°C for 15 min had no negative effect on IgH transcription (the background was reduced under these conditions), and subsequently added TFIID had no appreciable effect on transcription (lanes 4 and 5). Under similar conditions, nuclear extracts heat treated at 42°C for 15 min failed to transcribe the V␤ promoter (data not shown). Taken together, our data demonstrate that transcription factor requirements between the TATA Ϫ Inr ϩ and TATA ϩ Inr Ϫ promoters are different. TATA Ϫ Inr ϩ promoters require TFII-I, whereas the TATA ϩ Inr Ϫ promoters do not. DISCUSSION The control region of typical eukaryotic messenger RNA coding genes is comprised of proximal (core) and distal (enhancer) promoter regions (1). The core promoter region consists predominantly of two elements: the TATA box and/or the Inr element, which can be present either alternately (TATA ϩ Inr Ϫ or TATA Ϫ Inr ϩ ) or in limited cases simultaneously (TATA ϩ Inr ϩ ) (26). To understand the various transcriptional strategies that exist in nature, it is important to elucidate why different genes have adopted different core promoter elements and how these elements mediate transcription.
Transcription initiation in eukaryotes is mediated by a set of general transcription factors that assemble at the core promoter to form the preinitiation complex (27)(28)(29). The core promoter structures are different for different genes. Consequently, the preinitiation complexes (containing general transcription initiation factors), which assemble at different core promoter elements (TATA or Inr) are different (10,11). These experiments, however, employed a composite TATA ϩ Inr ϩ core promoter and were reconstituted with purified and/or recombinant proteins (10,11). To distinguish be-tween the mechanisms of transcription initiation mediated by TATA and Inr, we employed nuclear extracts to transcribe the naturally occurring TATA Ϫ Inr ϩ (V␤) and TATA ϩ Inr Ϫ (IgH) promoters. Our analyses demonstrate that the mechanisms of promoter utilization and the requirement of transcription factors are distinct for the two classes of promoters.
We present multiple approaches that were undertaken to demonstrate differences in promoter utilization. First, we depleted various nuclear extracts for the transcription factor TFII-I, which is important for Inr element-containing promoters (10 -12). Depletion of TFII-I by an anti-TFII-I antibody led to complete inhibition of transcription of a TATA Ϫ Inr ϩ promoter, whereas TFII-I depletion did not have a negative effect on TATA ϩ Inr Ϫ transcription. Furthermore, the addition of TFII-I relieved the antibody-mediated inhibition of the V␤ promoter, suggesting that the active component was indeed TFII-I. This conclusion is supported by the fact that the antibody predominantly recognizes TFII-I in nuclear extracts (as evidenced by Western blot analysis) and can be effectively blocked by the antigenic peptide derived from TFII-I (Fig. 2a). However, because the preparation of TFII-I used to reconstitute the transcriptional activity is partially pure, involvement of additional components cannot be ruled out completely.
Second, we demonstrate that an Inr element-containing oligonucleotide, which was competent in TFII-I binding, competitively inhibited transcription of a TATA Ϫ Inr ϩ promoter. This observation suggests that interactions of TFII-I to the Inr element is necessary for TATA Ϫ Inr ϩ transcription. Consistent with this suggestion, the Inr oligonucleotide-mediated inhibition of transcription was relieved by exogenous addition of TFII-I. Other factors (13)(14)(15)(16)(17)(18)(19)(20) have been implicated in Inr element binding. However, under the conditions tested, TFII-I is the predominant factor present in various nuclear extracts that is responsible for Inr-dependent binding and transcriptional activities via the V␤ promoter. This is consistent with our preliminary data, which indicate that TFII-I is also required for the V␤ promoter function in vivo.
TFIID is required for both TATA ϩ Inr Ϫ and TATA Ϫ Inr ϩ promoters (6,7). However, it is questionable whether or not TATA binding activity of TFIID is required for the TATA Ϫ Inr ϩ promoters (6). Accordingly, it has been shown that TATA binding activity is required for some TATA Ϫ Inr ϩ promoters but not for others ("true" TATA-less promoters) (6). The definition of a true TATA-less promoter is confusing and thus a TATA-less promoter should be defined by the lack of a consensus TATA sequence and not on the basis of mechanisms of TFIID binding. The promoter employed here (V␤) is a naturally occurring TATA-less promoter (lacking a consensus TATA box), a notion further supported by model building studies. 4 However, V␤ transcription requires the TATA binding activity of TFIID. Although we do not know the exact mechanism of TFIID recruitment to the V␤ promoter, it is possible that the binding of TFIID to the promoter may be mediated by TFII-I interactions because TFII-I interacts with the TATA binding subunit (TBP) of TFIID (11).
Finally, we demonstrate that heat treatment of nuclear extracts affects TATA ϩ Inr Ϫ and TATA Ϫ Inr ϩ promoters differently. It has been shown that heat treatment of a nuclear extract, normally competent for transcription, rendered the extract inactive for transcription of both types of promoters (6). The transcriptional activity of the extract for a TATA ϩ Inr Ϫ promoter could be restored upon exogenous addition of TFIID (6). However, for a TATA Ϫ Inr ϩ promoter, the addition of TFIID was insufficient, suggesting that an additional heat-labile com-4 S. K. Burley, personal communication.
FIG. 5. Factor requirements for TATA -Inr ؉ and TATA ؉ Inrpromoters. A TATA Ϫ Inr ϩ promoter requires binding of TFII-I to the Inr element. In the absence of a cognate TATA box, this initial interaction of TFII-I with the Inr element may be necessary to recruit TFIID and an additional component (?) to the promoter. The exact order of entry of different factors into the preinitiation complex is unclear at present. TFII-I is dispensable for a TATA ϩ Inr Ϫ promoter, which, however, requires TFIID. ponent(s) was necessary for TATA Ϫ Inr ϩ promoters (6). Here we demonstrate that TFII-I is heat-labile and is required in addition to TFIID for TATA Ϫ Inr ϩ promoter function. Our analyses also suggest the existence of a third component that is required for TATA Ϫ Inr ϩ promoters. This component is present in a partially purified TFIID fraction and is heat-labile (Fig. 5). It is unclear at present whether this component is directly associated or merely copurifies with TFIID.
Our data reveal that different transcription factor activities can be targeted by heat treating nuclear extracts at different temperatures. Thus, although heat treatment of nuclear extracts at 42°C for 6 min completely ablates TATA Ϫ Inr ϩ transcription, similar treatment does not ablate TATA ϩ Inr Ϫ transcription. This mild heat treatment does not affect TFIID activity but destroys other activities (including TFII-I) that are required for TATA Ϫ Inr ϩ transcription. Therefore, differential heat treatment of nuclear extracts at different temperatures can be used as an assay to distinguish between TATA ϩ Inr Ϫ and TATA Ϫ Inr ϩ -dependent transcriptional activities.
In conclusion, we clearly demonstrate the Inr-dependent function of TFII-I via the TATA Ϫ Inr ϩ V␤ promoter. It is possible that TFII-I may be necessary for transcription of other TATA Ϫ Inr ϩ promoters as well (30). Finally, although it appears at present that different Inr-dependent factors may function through different promoters or under different conditions, it is likely that multiple Inr-dependent factors may work in concert for some TATA Ϫ Inr ϩ promoters.