RNA polymerase II (pol II) transcribes mRNA and small nuclear RNA. The subunits of RNA polymerase II have been the final targets of variety of factors to regulators gene transcription1, 2, 3, 4. Pol II consists of 12 subunits of which RPB5 is the common subunit of pol I, II and III4. Evidence accumulated that RPB5 is one of active subunit of pol II during the gene transcription. HBx was found to target RPB5 to stimulate transactivation5; RMP or RPB5-mediating protein, interacts with RPB5 and counteracts transactivation by HBx6, 7; RPB5 binds general transcription factor IIB and together with HBx forms a ternary complex8, 9. Evidence revealed that RPB5 is directly involved in activated transcription10. TFIIF and RAP30 have been shown to associate with pol II and recruit pol II to the promoter in the transcription initiation11, 12, 13. Recent report demonstrates that RPB5 is the responsible subunit of pol II to associate with subunit RAP30 of TFIIF complex14.

TFIIF is a general transcription factor which functions in both transcription initiation and elongation11, 15. Eukaryotic TFIIF is a heteromeric tetramer of RAP30 and RAP74 which have been isolated as RNA polymerase II-associated proteins (RAP) by an affinity column containing immobilized pol II16. The structure and function of both RAP30 and RAP74 subunits of TFIIF have been well defined15, 17, 18, 19, 20. The N-terminus of RAP30 is proposed to associate with RAP74 to form the TFIIF complex and is necessary for transcription initiation16, 21, 22. The middle part of RAP30 is a sigma-homologous region that associates with pol II through binding to RPB5 and is essential for transcription elongation23. The C-terminus of RAP30 contains a cryptic DNA-binding domain, which is suggested to be homologous to the DNA template-binding domain of prokaryotic sigma factor13. TFIIF has been shown to be necessary for most, if not all, preinitiation complex formation and gene transcription24, 25, 26, 27.

TFIIF communicates with a number of factors to regulate gene transcription. It was reported that TFIIF directly bind to basal factors of TFIID, TFIIE and TFIIB21. Transcriptional activators, such as serum response factor (SRF), have been demonstrated to bind RAP74 in the middle of the molecule. On promoters such as adenovirus major late promoter, RAP74 subunit helps to wrap the DNA approximately one turn around the general transcription factors and RNA polymerase II. The RAP30 component of TFIIF can enhance the assembly of pol II into the initiation complex and RAP74 binding to the initiation complex will allow pol II to make promoter contact28, 29.

RMP was identified as a protein with 508 amino acids. Various mammalian tissues have been found to ubiquitously express RMP. RMP interacts with RPB5 and counteracts transactivation by HBx. Immunoprecipitation demonstrated that RMP associate with assembled pol II complex and probably is an integral component of pol II holoenzyme. RMP has also been found to suppress activated transcription by VP166.

As both RMP and TFIIF tightly associate with RPB5, RMP and TFIIF must be spatially close and an interaction between them is therefore possible during the preinitiation complex formation and gene transcription. Effort was made to elucidate the possible interaction between TFIIF and RMP with its effect on activated transcription.


Plasmids constructions

The plasmids pNKFLAG and pNKGST, derived from pSG5UTPL, are FLAG-tagged and GST-tagged mammalian expression vectors, respectively, as reported30. The plasmids pGENK1 and pGENKS are bacterial expression vectors for GST-fused proteins as reported previously9.

The construction of full-length and truncated RMP expression plasmids have been described6. The internal deletion mutant RMP/Id5 was generated by splicing PCR using a primer set of 5′-GCGAGCTCCATG AGG CTA GGA AAT GTA-3′ and 5′-ATT TTC TTG CTC AAC AGT ATG TGA AAA ATA TAT-3′ together with an other primer set of 5′-ACT GTT GAG CAA GAA AAT CAA AAG AAA CTT TTG-3′ and 5′-GCGGATCCGTC TTT CTG TTG CAA-3′. The PCR product containing an artificial SacI site and a BamHI site was digested and inserted into pGENK1 and pNKFLAG. The plasmids pGST-RAP30 and His-ET-TAP74 were a gift from R.G Roeder. The other RAP30 and RAP74 or their translation expression plasmids were reported14.

Preparation of recombinant proteins

GST-fused proteins were expressed in Escherichia coli by induction with 0.4 mM isopropyl-D-thio-galactopyranoside at 30 °C for 3 h. The cells were harvested and sonicated in PBST buffer (phosphate-buffered saline containing 1% Triton X-100) (4, 5). After centrifugation, the extracts (supernatants) were collected and stored at −80 °C. For purification, the extracts were incubated with glutathione-Sepharose 4B (Amersham Pharmacia Biotech) at room temperature for 1 h. The beads were precipitated, washed four times with an excess amount of PBST buffer, and then eluted with 10 mM reduced glutathione in 50 mM Tris-HCI (pH 8.0). The eluted proteins were divided into aliquots and stored at −80°C.

FLAG-tagged proteins were expressed in BL21 by induction with 0.4 mM isopropyl-D-thiogalactopyranoside at 30 °C for 3-6 h. The cells were harvested and sonicated in 50 mM Tris-HCI, pH 8.0, 150 mM NaCI, and 0.1% Triton X-100. After centrifugation, the supernatant was collected and stored at −80 °C. FLAG-tagged proteins were purified by incubating the sonication supernatant with anti-FLAG M2 resin (Kodak Scientific Imaging Systems) followed by several washes. The bound proteins were eluted with buffer containing FLAG peptide (0.2 mg/ml FLAG peptide, 50 mM Tris-HCI, pH 8.0, 150 mM NaCl).

In vitro GST resin pull-down assays

GST resin pull-down assay was carried out as reported6,8, 9, 14. Approximately 1 μg of GST or GST-fused protein immobilized on 20 μl of glutathione-Sepharose 4B preblocked in 0.5% nonfat milk and 0.05% bovine serum albumin was incubated with 0.1 μg of FLAG-tagged proteins in 0.5 ml of modified GBT buffer for 1-2 h on a rotator at 4°C. After being washed four times with modified GBT buffer, the bound proteins were eluted, fractionated by 12.5% SDS-PAGE, and subjected to Western blot analysis using anti-FLAG monoclonal antibody (M2).

Immunoprecipitation and western blot analysis

Transient transfection of COS1 cells was carried out as reported previously1, 7, 8. The cells were harvested, washed, and sonicated in LAC buffer (10% glycerol, 20 mM HEPES, pH 7.9, 50 mM KCI, 0.4 M NaCI, 1 mM MgCl, 0.1 mM dithiothreitol, 0.1 mM EDTA, 9 mM CHAPS, 0.5 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin and leupeptin) and then centrifuged. The total cell lysates were stored at −80 °C. Approximately 1.5 mg of protein of total cell lysates was diluted with 4 times the volume of TBST buffer (50 mM Tris-HCI, pH 7.5, 50 mM NaCl, 0.05% Tween 20) and incubated with 50 μl of packed Sepharose 4B for 30 min. The supernatant was then obtained by centrifugation. Supernatants with FLAG-tagged proteins were immunoprecipitated with 20 μl of 50% anti-FLAG M2 resin, rotated for 2 h at 4oC , and washed four times with washing buffer TBS (50 mM Tris-HCI, pH 7.5, 150 mM NaCl). The bound proteins were eluted, fractionated by 12.5% SDS-PAGE, transferred onto nitrocellulose membranes, and subjected to Western blot analysis with the antibody. The proteins were visualized by enhanced chemiluminescence (ECL), according to the manufacturer's instructions (Amersham Pharmacia Biotech).

Far-Western detection

The far-Western blotting was performed as previously described11. The target protein (200 ng each) were fractionated on SDS-PAGE gels and then electrotransfered onto nitrocellulose membranes, which were denatured, renatured, and blocked with 5% skim milk in modified GBT buffer before subjected to far-Western blot7. The binding reaction proceeded in modified GBT buffer containing 32P-labeled probe (40-100 ng/ml of protein with 2×10 cpm/μg of protein), 1%BSA, 2 mM of unlabeled ATP, and the sonicated supernatant of E. Coli JM109 transformed by pGENK1, which contained a final GST protein concentration of 1 mg/ml. The membranes were washed five times with modified GBT buffer and exposed to X-ray films (XAR Omat, Kodak) or imaging plates (Fuji).

Transfection and luciferase assay

Cell culture and transient transfection were carried out as reported previously9, 30. Luciferase assay was conducted using dual-luciferase reporter system according to the manufacturer's instructions (Promega). Briefly 100ng of experiment reporter pFRu-luc and 10 ng of control reporter pRu-luc were co-transfected with 5-20 ng of Gal-VP16 and 0-2 μg of RMP expression plasmids. Luciferase values were corrected for differences in apparent transfection efficiency by expression as a ratio with Renilla luciferase signals in the corresponding samples to normalize the transfection efficiency.


RMP associates with TFIIF in vitro

As RPB5 associates with both RMP and TFIIF, we examined the possible interaction between the full-sized RMP and each subunits of TFIIF in vitro. GST-fused RMP was expressed in E. Coli, purified by glutathione Sepharose 4B and fractionated by SDS-PAGE (Fig 1B, lane 1). Flag-tagged RAP74 and RAP30 were expressed in E.coli and purified as described in Material and Methods. Pull-down assay was carried out with GST-RMP and FLAG-RAP30 or FLAG-RAP74. The result demonstrated that both RAP74 and RAP30 in Flag-fused form associated with full-sized RMP in GST-fused form. Although as a negative control, GST bound neither FLAG-RAP30 nor FLAG-RAP74, the FLAG-RPB5, as a positive control, strongly bound GST-RMP (Fig 2, lane 1 and lane 10), confirming the previous result by Dorjisuren et al.

Figure 1
figure 1

Delineation and purification of RMP proteins. A. Schematic map of various RMP deletion constructs. The expression plasmids were constructed as described in “Materials and Methods”. B. Various GST-fused RMP truncation proteins were expressed in E. Coli, purified, fractionated by 12.5% SDS-PAGE as described under “Materials and Methods” and visualized by Coomassie Brilliant Blue staining. The positions of molecular mass makers are indicated on the left.

Figure 2
figure 2

GST resin pull-down assay. Approximately 1 μg of GST (lane 10) or GST-fused RMP truncation proteins purified from bacterial expression was immobilized on glutathione resin and incubated with 0.1 μg of bacterially expressed FLAG-RAP74 (A), or FLAG-RAP30 (B), or FLAG-RPB5 (C) in GBT buffer. Pull-down assay and Western blot analysis were carried out with anti-FLAG M2 antibody as described under “Materials and Methods”. Lane 11 shows 5% of the input of FLAG-tagged proteins.

Both subunits of TFIIF interact with the same region of RMP

As the full RMP contains several domains which have been elucidated for the bindings with RPB5 and TFIIB6, 7, we tried to delineate binding regions of RMP for the association with TFIIF subunits. A series of constructions were made to express various trancation mutants of RMP in GST-fused form. The bacterially expressed proteins of RMP mutants were purified and fractionated by SDS-PAGE as shown in Fig 1.

To map the binding region of RMP, pull-down assay was carried out with various trancation mutants of RMP in GST-fused form and RAP74 or RAP30 in Flag-fused form. The results were shown in Fig 2A and 2B. Interestingly RAP74 and RAP30 showed a similar binding pattern with RMP. In another word, D5 and D10 of RMP bound both Flag-RAP74 and Flag-RAP30 while other trancation mutants of RMP in GST-fused form or GST alone bound neither Flag-RAP74 nor Flag-RAP30. D5 is the minimal binding region in RMP for the interaction with both RAP74 and RAP30. To confirm the result we also set a positive control of binding between RPB5 and RMP which has been well defined6. There was a very strong binding between Flag-RPB5 and full-sized RMP in GST-fused form (Fig 2C, lane 1). Both D6 and D1 of RMP trancation mutants bound Flag-RPB5 although D1 was the minimal region for RPB5-binding, which is consistent with the previous report6. The binding pattern of RMP with RPB5 is different from that of RMP with TFIIF subunits in which RAP30 and RAP74 bound to the same region of RMP.

D5 region is required and sufficient for the association between RMP and TFIIF

To further confirm that D5 region of RMP bound both subunits of TFIIF, we also did Far-Western blotting. In addition to other trancation mutants of RMP used in Fig 2, we constructed a mutant with an internal deletion of D5 region (from residual 315 to 413) in RMP. Bacterially expressed RMP or its mutants in GST-fused form was purified, fractionated and transferred to nitrocellulose membrane. After being denatured and renatured, the membrane was subjected to Far-western detection by GST-RAP30 or GST-RAP74 probe. Although full RMP in GST-fused form bound both GST-RAP30 and GST-RAP74 probes, the internal deletion of D5 region abolished its ability to bind GST-RAP30 and GST-RAP74. The D5 region alone was sufficient to associate with GST-RAP30 and GST-RAP74. The other trancation mutants of GST-RMP together with GST alone bound neither RAP30 nor RAP74 in the Far-Western detection, consistent with the result of pull-down assay (Fig 3). The results suggest that D5 region is necessary and sufficient for the association between RMP and RAP74 or RAP30.

Figure 3
figure 3

RMP binding with TFIIF detected by Far-western blotting. Equal amounts of GST (lane 9) or GST-fused RMP truncation proteins (200 ng each) was fractionated by 12.5% SDS-PAGE and electrotransferred onto nitrocellulose membranes. After denature, renature and being blocked by 5% skim milk in modified GBT buffer, the membranes were subjected to Far-western detection with 32P-labed probes of GST-RAP74 (A) or GST-RAP30 (B).

RMP associates with TFIIF in vivo

Although the result of pull-down assay and Far-western detection were consistent, both are in vitro binding assay which may not reflect the physiological situation in vivo. Therefore we performed immunoprecipitation to observe the association between RMP and RAP74 or RAP30 in vivo. COS1 cells were cotransfected with mammalian expression plasmids of Flag-RMP or its trancation mutants and GST-RAP74 or GST-RAP30. The proteins were expressed approximately equally in the transfection as shown in Fig 4A. Immunoprecipitation was carried out with anti-Flag M2 antibody-bound resin and detected by anti-GST antibody. The results matched with that of in vitro (Fig 4B). Full-sized RMP and its D5 mutant in Flag-tagged form recovered both GST-RAP74 and GST-RAP30 while other mutants including Id5 in which D5 region was deleted recovered neither GST-RAP74 nor GST-RAP30. And apparently the immunoprecipitated RAP74 and RAP30 were in approximately equal molar ration, suggesting that RMP associated with RAP74 and RAP30 which were in the form of TFIIF complex rather than individual subunits.

Figure 4
figure 4

RMP associates with TFIIF in vivo. COS 1 cells were cotransfected with mammalian expression plasmids pNKFLAG-RMP and pNKGST-RAP74 or pNKGST-RAP30. Cell lysates were prepared as described under “Materials and Methods”. Approximately 1.5 mg proteins of total lysate was immunoprecipitated with 20 μl of packed anti-FLAG M2 antibody-bound resin. After being washed, the bound proteins were eluted and then fractionated by 12.5% SDS-PAGE and subjected to Western blot analysis with anti-FLAG M5 antibody (A) or anti-GST antibody (B and C). Lane 8 of panel B and C shows 5% of the input of total lysate used in lane 4.

Interaction with TFIIF is required for the RMP-mediated suppression of activated transcription

As RMP was reported to functions as a corep- ressor of the activated transcription, we wonder if the interaction between RMP and TFIIF is necessary for the suppression function of RMP. Therefore we examined the effects of RMP on activated transcription by Gal-VP16, a chimeric activator with Gal4 DNA binding domain fused to the VP16 activation domain. pFRu-luc luciferase reporter was driven by five Gal4 binding sites. COS1 cells were cotransfected with expression plasmids of reporter, Gal-VP16 and RMP or its mutants. The results of the luciferase assay demonstrated that the activated transcription by Gal-VP16 was suppressed by RMP, which is consistent with the results by CAT assay6. However RMP-Id5 (in which TFIIF-binding region was deleted) lost the suppression function for the activated transcription by Gal-VP16 (Fig 5). The results show that the association with TFIIF is necessary for the RMP suppression function of activated transcription.

Figure 5
figure 5

D5 region of RMP is required to suppress activated transcription. COS1 cells were cotransfected with pGalLuc and Gal-VP16 together with RMP or RMP-Id5 constructs. The transfected plasmid DNA was 0.1 μg of pGalLuc together with the following: bars 1 to 4, 0, 5, 10, and 15 ng of pGal-VP16, respectively; bar 5 to 7, 0, 1, and 2 μg of pSG5UTPL-RMP, respectively; bar 8 to 10, 0, 1, and 2 μg of pSG5UTPL-RMP, respectively, plus 15 ng of pGal-VP16; bar 11 to 13, 0, 1, and 2 μg of pSG5UTPL-RMP-Id5, respectively; bar 14 to 16, 0, 1, and 2 μg of pSG5UTPL-RMP-Id5, respectively, plus 15 ng of pGal-VP16. The total amount of DNA added per transfection was adjusted to 10 μg with the control vector, pSG5UTPL. Error bars shows standard deviations.


Most RNA polymerase subunits are unique to their respective RNA polymerases, while some subunits are common among RNA polymerses I, II, and III. With the exception of RPB5, all of the common subunits in human RNA polymerases I, II, and III can substitute for their yeast homologues, although RPB5 is highly conserved among humans, Saccharomyces cerevisiae, and Schizosaccharomyces pombe, and is essential in yeast2, 31, 32. Yeast RPB5 has been reported to interact with RPB3 in vitro, and both RPB5 and RPB3 are present in two molar amounts in RNA polymerase II and seem to play an important role in subunit assembly, as does human RPB5.

Previously we identified that HBx, the multifunctional viral regulator protein of hepatitis B virus, directly target RPB5 of RNA polymerase II4, 5. And both RPB5 and HBx communicate with TFIIB. The trimeric interaction of these three factors may facilitate transcription and HBx acts as coactivator in activated transcription8, 9. Based on these results, we proposed that RPB5 is a communicating subunit of pol II that interacts with transcriptional regulators. In support of the notion, we identified a novel protein, RPB5-mediating protein (RMP), which counteracts the coactivator function of HBx by competitively binding RPB56, 7. The crystal structure of pol II revealed that RPB5 is composed of the exposed domain of N-terminal and the embedded C-terminal domain. The exposed domain is close to the DNA template and responsible for transcription regulation by interaction with transcription factors, such as TFIIB and RMP33, 34.

RMP was originally identified by its association with RPB5. RMP has been shown to counteract HBx transactivation and the increased amount of HBx reduced the suppression effect of RMP. In the in vitro and in vivo binding assay RPB5 and HBx competed with each other to associate with RMP. The activated transcription by Gal-VP16 was suppressed by RMP in mammalian cells. Based upon these observations RMP has been suggested to function as a corepressor of transcription regulator6.

In this report D5 region of C-terminal RMP has been shown to interact with both subunits of TFIIF, which has been confirmed by pull-down assay, Far-western blot and immunoprecipitation. It is possible that RAP30 and RAP74 contact the different surfaces of D5 region of RMP. Or the RMP-interacting surfaces of RAP30 and RAP74 are very closely positioned within the D5 region so that further fine delineation could possibly separate the RAP30 and RAP74-binding domains. The fact that D5 region associates with both subunits of TFIIF may suggest a function of RMP in the regulation of formation of TFIIF complex by associating with both subunits of RAP30 and RAP74.

It was reported that RMP also interact with TFIIB by two independent domains of the N-terminal region D8 and the C-terminal domain D10 of RMP which overlaps with TFIIF binding region (7 and unpublished data). D5 region of RMP seems to be the minimal region of C-terminal RMP for TFIIB-binding as the D9 region from residual 434 to 508 was negative of TFIIB-binding and D10 region is positive for the binding. It could be deduced that the region of RMP from residual 316 to 434 is responsible for the association between C-terminal RMP and TFIIB. This is an interesting phenomenon as this D5 region of RMP from residual 316 to 434 is exactly the common domain to interact with TFIIB and TFIIF. The relationship and mechanism among RMP, TFIIB and TFIIF remain to be addressed. One possibility is that the two general transcription factors of TFIIF and TFIIB compete with each other to bind RMP and positively or negatively regulate the suppression function of RMP. RMP may also serve as a scaffold for TFIIF and TFIIB to communicate with RPB5 of pol II.

In addition to associating with RPB5, TFIIB and TFIIF, RMP also bind with pol II complex6, suggesting that RMP may be a necessary component of pol II holoenzyme complex. So it is possible that RMP regulates transcription from pol II holoenzyme complex by interacting with many different factors except RPB5 and TFIIB. In this report we demonstrated that both subunits of TFIIF interacted with RMP, which may be an additional way of transcription regulation by RMP. Alternatively TFIIF may be a necessary component for the suppression function of RMP through interaction with RPB5 and/or TFIIB. Actually as RMP, TFIIF, TFIIB and RPB5 bind with each other, they may form a complex and cooperatively regulate the transcription.

Accumulated evidences appear that transcription is regulated by factors through targeting TFIIF. A number of activators enhance the transcription by its association with RAP74 or RAP3026, 27, 35. Recently the interaction between heat shock factor-4, a transcriptional repressor, with the basal transcription factor TFIIF was firstly identified48.

We report here that RMP interacts with TFIIF and suppresses the activated transcription by Gal-VP16. Disassociation of RMP and TFIIF by the internal deletion of TFIIF-binding region in RMP abolished suppression function of RMP. A similar phenomenon was observed in the interaction between the RMP and RPB5. Disruption of the association between RMP and RPB5 results in loss of suppression function of RMP, which suggests that contacting with both RPB5 and TFIIF are necessary for the suppression function of RMP. It has been shown that pol II is associated with both RMP and TFIIF6, 14. Destruction of the conformation of the holoenzyme complex by disassociation of RMP with TFIIF or RPB5 leads to the loss of suppression function of RMP. The interaction between TFIIF and RMP could be either an independent passway for RMP to regulate the transcription or a necessary part of network with TFIIB, RPB5 and pol II for the corepressor function of RMP, which remain to be addressed.