INTRODUCTION

RNA polymerase II in the arrested conformation cannot extend an RNA chain in the presence of all four NTPs. Arrested complexes retain transcription potential, however, since they can be reactivated for RNA chain extension by elongation factor SII. SII can bind RNA polymerase II in vitro, and SII binding is thought to be the activating event for a ribonuclease activity within RNA polymerase II (reviewed in Ref. 1). This ribonuclease removes a small number of nucleotides from the 3-end of the nascent RNA prior to its re-extension through the blockage (2, 3, 4) in a reiterative cleavage and resynthesis process, allowing an RNA polymerase II molecule multiple attempts at chain extension through an obstacle (1, 5, 6).

Genetic evidence suggests that SII has an in vivo role related to RNA polymerase II function which involves the binding of SII to RNA polymerase II's largest subunit (7, 8). Although SII binds to purified RNA polymerase II, it has been difficult to isolate elongation complex-associated SII, presumably because the lifetime of the interaction is very short (9, 10).

Full-length SII does not bind nucleic acids at physiological ionic strength (11). Specific internal and extensive N-terminal deletion mutants of SII bind DNA, RNA, and DNA:RNA hybrids in gel mobility shift assays, suggesting that cryptic nucleic acid binding activity is a property of SII which may be revealed only after SII associates with an elongation complex (11). However, the functional significance of SII's nucleic acid binding activity in the context of RNA cleavage and the stimulation of elongation is unknown.

Contacts between RNA and the largest two RNA polymerase II subunits have been elucidated by cross-linking radiolabeled RNAs synthesized in the presence of photoactive nucleotide analogues (12, 13, 14, 15). Similarly, the and  subunits of Escherichia coli RNA polymerase can be cross-linked to nascent RNA (see Ref. 16 and references therein).

Since full-length SII may bind nucleic acids preferentially in the context of an elongation complex, we set up a system to trap SII on the complex and detect an interaction between SII and the nascent RNA. We report here that full-length SII, but not transcriptionally inactive SII, can be cross-linked to the nascent RNA via a photoactive nucleotide situated at the 3-end of the transcript. This contact may be important in the elongation function of SII.

MATERIALS AND METHODS

RNA polymerase II and basal transcription factors were isolated from rat liver (17) or purified from E. coli (for TFIIB) (18) as described. Recombinant human and yeast SII were expressed from pT7-7Met (provided by Dr. C Kane, University of California, Berkeley) and pYSE1 (provided by Dr. S. Natori, Tokyo University), respectively, and purified as described (19). SII was expressed from pET22b(+)/TFIIS (provided by Dr. C. Kane, University of California, Berkeley) and purified by nickel affinity chromatography as described (20).

Fast protein liquid chromatography-purified nucleoside triphosphates were purchased from Pharmacia Biotech Inc. [-³²P]CTP was obtained from Amersham Corp. RNase A and T1 were obtained from United States Biochemical Corp. Proteinase K was from Promega (Madison, WI).

4-Thio-UTP was synthesized and purified by DEAE-cellulose or HiTrap Q (Pharmacia) chromatography as described (12).

3-End Labeling of RNA and Incorporation of 4-Thio-UMP in the Elongation Complex

Transcription was reconstituted with partially purified initiation factors and RNA polymerase II on pAdTerm-2 as described (21). Ternary complexes (approximately 25 fmol/reaction) containing a 14-nucleotide transcript were synthesized in the presence of 7 mM MgCl₂ and ATP, UTP, and CTP (20 µM each). Transcripts were extended by adding heparin to 10 µg/ml and all four NTPs to 800 µM (each). After 20 min, elongation complexes were immunoprecipitated with anti-RNA IgG (D44) and fixed Staphylococcus aureus (21, 22) and washed three times with reaction buffer (60 mM KCl, 20 mM Tris-HCl, pH 7.9, 3 mM Hepes-NaOH, pH 7.9, 0.5 mM EDTA, 2.2% (w/v) polyvinyl alcohol, 0.2 mg/ml acetylated bovine serum albumin, 3% (v/v) glycerol). SII and MgCl₂ were added, and the reaction was incubated for 1.5 min at 28 °C. After washing three times with reaction buffer, [-³²P]CTP (20 µCi, 3000 Ci/mmol), 400 µM GTP, 100 µM UTP or 4-thio-UTP, and 7 mM MgCl₂ were added and incubated for 1 h at 28 °C. Reactions were stopped with SDS and treated with proteinase K. RNA was precipitated with ethanol and electrophoresed on 5% polyacrylamide, 50% urea gels.

Elongation complexes bearing labeled RNA were washed three times and resuspended in cross-linking buffer (40 mM Tris-HCl, pH 7.9, 60 mM KCl, 0.5 mM EDTA). These complexes were liberated from S. aureus by digestion with ribonuclease T1 (200 units/reaction) at 28 °C for 15 min, followed by centrifugation for 2 min at 4 °C to remove the cells. This treatment removed bases 1-180 from the nascent RNA (23). SII or SII was added to the supernatant, and the reaction was incubated at 28 °C for 15 min. Reactions were transferred to a polystyrene microtiter plate (Corning), which was placed in a Pyrex dish filled with water and UV-illuminated for 30 min at 4 °C on a Photodyne model 3-3500 transilluminator (12). Proteins were precipitated in 10% trichloroacetic acid and electrophoresed on polyacrylamide-SDS gels as indicated. Gels were fixed in 0.56% (v/v) methanol, 1% (v/v) acetic acid, 10% (v/v) glycerol for 30 min, dried, and subjected to autoradiography and analysis on a Fuji phosphoimaging system.

RESULTS AND DISCUSSION

The template used in these reactions contains the core adenovirus major late promoter and a segment of the human histone H3.3 gene containing an intrinsic arrest site called Ia. Transcription in the absence of SII results in approximately half of the RNA polymerase II molecules becoming arrested at site Ia (positions 205, 206, and 207, where 1 is the transcription start site) (Fig. 1A; Ref. 5). SII treatment results in full read-through of site Ia. An intermediate in read-through is an elongation complex bearing an RNA shortened by 7-9 nucleotides at the 3-end (5). When arrested complexes are washed free of NTPs, the shortened transcript and the cleaved oligonucleotide can be observed (4). We took advantage of this intermediate to incorporate a radioactive marker and a photoactive nucleotide analogue, 4-thio-UMP, into RNA. In the presence of [-³²P]CTP and GTP, complexes bearing the 198-base cleaved transcript could be extended to G²⁰⁰ with a single ³²P atom in the transcript between bases 198 and 199 (Fig. 1, A and B, lane 1). In the presence of 100 µM UTP, the RNA is extended to positions U²⁰⁴, U²⁰⁵, and U²⁰⁶ (Fig. 1B, lane 2; Ref. 4). Although 4-thio-UTP can serve as a substrate in this reaction, it is utilized less efficiently than UTP and results in slightly shorter transcripts (Fig. 1B, lane 4 versus lane 2). Comparison with a ladder of UMP-containing transcripts differing in length by one nucleotide suggests that the 4-thio-UMP-containing transcripts are comprised primarily of RNAs of 203, 204, and 205 nucleotides in length, containing three, four, and five 4-thio-UMP residues, respectively (data not shown). 4-Thio-UMP-containing transcripts were also less effectively extended in the presence of all four natural NTPs (33 versus 10% read-through of site Ia; Fig. 1B, lane 3 versus lane 5). The decreased read-through efficiency could result from slower addition of nucleotides to the 4-thio-UMP-containing 3-end of the nascent transcript and/or an increasing likelihood that these complexes become arrested.

UV irradiation of these 4-thio-UMP-containing complexes resulted in the cross-linking of nascent RNA to the two largest subunits of RNA polymerase II as seen previously (Ref. 12; Fig. 2, lanes 1 and 2). When a fraction containing recombinant, full-length, human SII was allowed to bind the complexes in the absence of Mg²⁺ prior to irradiation, a polypeptide corresponding to the size of SII (48 kDa, including mass of cross-linked RNA) was also photoaffinity-labeled (Fig. 2, lane 1). Neither SII nor RNA polymerase II subunits were labeled if the complexes were not subjected to UV exposure (Fig. 2, lane 5). Similarly, when UTP was substituted for 4-thio-UTP in the reaction, photoaffinity labeling was not observed for either SII or RNA polymerase II subunits (Fig. 2, lane 6). The cross-linked products were sensitive to proteinase K (10 µg for 10 min; Fig. 2, lane 4), indicating that the radioactive bands contained a protein component. When photolabeled RNA polymerase II complexes were disrupted with SDS and treated with a mixture of RNase A and RNase T1, free RNA was digested and the ³²P label was removed from the end of the cross-linked RNA (Fig. 2, lane 3), demonstrating the presence of an RNA component in the labeled species. When SII and all basal transcription factors were mixed with 4-thio-UMP-containing RNA, UV irradiation failed to produce photoaffinity-labeled polypeptides.¹ In summary, SII-RNA cross-linking was dependent on the presence of SII, a transcription elongation complex, incorporation of 4-thio-UMP, and irradiation with UV light.

Quantitation of free and cross-linked RNA showed that approximately 1-2% of the RNA was cross-linked to SII, while 45% was cross-linked to high molecular weight polypeptides that co-migrate with the two largest RNA polymerase II subunits. RNA polymerase II was 1-2 orders of magnitude more reactive with derivatized RNA than was SII. We cannot exclude the possibility that additional SII molecules were co-cross-linked to RNA with RNA polymerase II subunits, precluding visualization as a high mobility band. That RNA polymerase II cross-links with greater efficiency than SII is consistent with SII's rapid partitioning between the complex and solution compared with the stoichiometric association of RNA polymerase II with RNA.

We estimate the purity of recombinant human SII (TSK-phenyl-5PW) fraction used in the experiment depicted in Fig. 2 to be approximately 30% based on silver staining of SDS gels. The failure of the other proteins in this mixture to become photolabeled argues for the specificity of cross-linking. To confirm that the cross-linked species was indeed SII, and not a contaminating bacterial protein, we performed similar experiments using recombinant SII further purified by TSK-SP-5PW chromatography to apparent homogeneity (4). We found that a polypeptide of identical mobility cross-linked with comparable efficiency when added in the same proportion relative to the RNA polymerase II complexes (approximately 150-fold molar excess) (Fig. 3, lanes 1 and 2). When an identical TSK-phenyl fraction containing an even greater excess of yeast SII (Fig. 3, lane 4), which was unable to activate RNA cleavage in the mammalian RNA polymerase II complex,¹ was added, no SII cross-linking was observed, as was the case for a buffer-only control (Fig. 3, lane 3). We further confirmed the identity of the putative SII polypeptide by testing as a substrate a smaller version of SII lacking the first 130 amino acid residues (SII). This deletion mutant can activate nascent RNA cleavage and stimulate read-through of the Ia site (20). This truncated form of the protein cross-linked to the nascent RNA, and the protein:RNA conjugate displayed a correspondingly faster electrophoretic mobility (27 kDa, including the mass of the RNA) on an SDS-polyacrylamide gel (Fig. 4, lane 1 versus lane 3). SII was cross-linked to RNA with approximately one-half the efficiency of the full-length molecule.

Current models for SII function suggest that this elongation factor binds the RNA polymerase II complex and activates nascent RNA cleavage and arrest site read-through in a Mg²⁺-dependent fashion. Thereafter, SII is probably released from the complex (9, 10). Here, we have trapped SII on the elongation complex by withholding MgCl₂, thereby preventing the cleavage and removal of the ³²P- and 4-thio-UMP-containing 3-end of the nascent RNA. Under cleavage conditions, i.e. when Mg²⁺ was included in the binding reaction prior to UV irradiation, photoaffinity labeling of RNA polymerase II subunits decreased by greater than 90%, and photoaffinity labeling of SII was no longer observed (Fig. 4, lane 4). Thus, the detection of the RNA:SII contact is a function of the presence of the RNA in an elongation complex prior to its removal by the cleavage reaction and again demonstrates the specificity of cross-linking.

To our knowledge, this is the first report of the direct detection of a biochemical interaction between full-length SII and RNA or a specific, active RNA polymerase II elongation complex. Earlier studies of RNA contacts with RNA polymerase II subunits employed a similar technique with HeLa nuclear extracts, which presumably contained SII (e.g. Ref. 12). The failure to detect this interaction could be related to the low efficiency of SII cross-linking, the low in vivo abundance of SII compared with these reactions, the presence of Mg²⁺ in the reaction, and/or the more random placement of 4-thio-UMP nucleotides at positions distant from the 3-end of the molecule. The experiments presented here were also performed using RNA polymerase II complexes located at or near an arrest site at which SII-activated cleavage is favored over most template positions.

GreA (24) and GreB (25) are structurally distinct elongation factors from E. coli that are functionally analogous to SII. Both factors promote read-through of arrest sites by RNA polymerase and stimulate nascent RNA cleavage by the enzyme (25). Like SII, GreB relieves the arrested condition by inducing cleavage of large oligonucleotides from nascent RNA (25). GreA functions differently since it must be present before arrest to stimulate read-through and the RNA is shortened by a dinucleotide increment (25). Similar to what we observe here, GreA has recently been shown to contact the nascent transcript in an E. coli elongation complex (26).

The contact of SII and RNA in the RNA polymerase II elongation complex is likely to be functionally important because: 1) it is specific for complex-associated RNA, 2) GreA makes a similar contact in the E. coli RNA polymerase complex, and 3) previous work using SII fragments suggests the existence of a cryptic nucleic acid binding capability in the full-length molecule (11). The close proximity or perhaps identity of the active site for SII-activated RNA cleavage and that for nucleotide addition (27) suggests that the SII:RNA contact may be of some functional or architectural importance to the catalytic center of the RNA polymerase II elongation complex. In addition to a possible allosteric rearrangement of RNA polymerase II upon binding, SII may be directly involved in the positioning of the RNA within the active site before cleavage. This may help explain how SII dramatically enhances the rate of RNA polymerase II's RNA cleavage reaction. It is tempting to speculate that SII's contact with the region of the transcript to be removed by cleavage may also facilitate the removal of the oligonucleotide from the RNA product site. The possibility remains, however, that the contact between SII and the 3-end of the RNA detected here may simply reflect a close relationship between SII's docking site on RNA polymerase II and the enzyme's active site. Nonetheless, this work places SII in close proximity to the catalytic center of the enzyme, and it is an important step toward defining the geometry of the active site of elongating RNA polymerase II.