Skip to main content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Virology. Author manuscript; available in PMC 2010 Mar 30.
Published in final edited form as:
Virology. 2009 Mar 30; 386(1): 32–43.
Published online 2009 Feb 1. doi: 10.1016/j.virol.2008.12.037
PMCID: PMC2683163
NIHMSID: NIHMS106988
PMID: 19187948

Human papillomavirus type 16 E2 and E6 are RNA-binding proteins and inhibit in vitro splicing of pre-mRNAs with suboptimal splice sites

Sohrab Bodaghi, Rong Jia, and Zhi-Ming Zheng*
Author information Copyright and License information PMC Disclaimer

Abstract

Human papillomavirus type 16 (HPV16) genome expresses six regulatory proteins (E1, E2, E4, E5, E6, and E7) which regulate viral DNA replication, gene expression, and cell function. We expressed HPV16 E2, E4, E6, and E7 from bacteria as GST fusion proteins and examined their possible functions in RNA splicing. Both HPV16 E2, a viral transactivator protein, and E6, a viral oncoprotein, inhibited splicing of pre-mRNAs containing an intron with suboptimal splice sites, whereas HPV5 E2 did not. The N-terminal half and the hinge region of HPV16 E2 as well as the N-terminal and central portions of HPV16 E6 are responsible for the suppression. HPV16 E2 interacts with pre-mRNAs through its C-terminal DNA-binding domain. HPV16 E6 binds pre-mRNAs via nuclear localization signal (NLS3) in its C-terminal half. Low-risk HPV6 E6, a cytoplasmic protein, does not bind RNA. Notably, both HPV16 E2 and E6 selectively bind to the intron region of pre-mRNAs and interact with a subset of cellular SR proteins. Together, these findings suggest that HPV16 E2 and E6 are RNA binding proteins and might play roles in posttranscriptional regulation during virus infection.

Keywords: Human papillomavirus type 16, RNA splicing, RNA-protein interaction, SR proteins, Protein-protein interaction, viral proteins

Introduction

RNA splicing is an essential step in the control of viral and mammalian gene expression. It occurs immediately after a nascent primary message is transcribed and consists of a series of cascaded biochemical reactions that take place in a spliceosome. Spliceosome-mediated pre-mRNA splicing involves five small U RNAs (U1, U2, U4, U5, and U6) and many splicing factors. The first step in the accurate recognition of intron splice sites involves interaction of the 5′ splice site (5′ ss) with U1, of the branch site with U2, and of the 3′ splice site (3′ ss) with U2AF (U2 auxiliary factor). These interactions are modulated by many cellular splicing factors including SR proteins (Graveley, 2000). SR proteins are a growing family of structurally related and highly conserved cellular splicing factors that are characterized by the presence of an RNA-recognition motif (RRM) and RS dipeptides. A group of classical SR proteins contains extensive phosphorylated RS domains and can be recognized by the monoclonal antibody mAb104 (Zahler et al., 1993). SR proteins are essential splicing factors that modulate the selection of a suboptimal splice site (Zheng, 2004). Many of them have redundant functions.

Human papillomaviruses (HPVs), a group of small DNA tumor viruses, usually infect keratinocytes of skin or epithelial cells of mucosa and cause benign warts or occasionally malignancies (Lowy and Howley, 2001; zur Hausen H., 2002). Viral gene expression in infected cells depends on cell differentiation and usually leads to the expression of six nonstructural viral regulatory proteins (E1, E2, E4, E5, E6 and E7) from early regions of the virus genome and two structural viral capsid proteins (L1 and L2) from late regions of the genome. Expression of each of these genes requires extensive RNA splicing. However, the factors that determine when a specific splicing pathway will be used to express a specific gene remains largely unexplored.

The papillomavirus E2 protein is a 42-kDa nuclear protein containing two defined functional domains that are relatively conserved among all papillomaviruses. The N-terminal domain, consisting of ~200 amino acid (aa) residues, is crucial for transcriptional activation, whereas the C-terminal domain, consisting of ~100 aa residues, possesses the DNA binding and dimerization properties of the protein. These two domains are linked by a hinge region that lacks a conserved aa sequence and varies in length among papillomaviruses (Hegde, 2002). The hinge in epidermodysplasia verruciformis-associated HPV E2 contains ~200 aa residues and multiple RS dipeptide repeats, but the hinge in anogenital HPV E2 is ~40–80 aa residues and lacks RS repeats (Sakai et al., 1996; Hegde, 2002). Besides its involvement in papillomavirus DNA replication (Hughes and Romanos, 1993; Sakai et al., 1996; Frattini and Laimins, 1994), E2 is also a transcriptional activator or repressor that regulates the E6 promoter through four consensus E2-binding sites (E2-BSs), ACC(N6)GGT (Androphy et al., 1987; Hawley-Nelson et al., 1988; Sousa et al., 1990; Romanczuk et al., 1990), upstream of the viral E6 promoter. However, E2 functions as a repressor at steps after TBP or TFIID binding (Hou et al., 2000) and its transcriptional repression occurs only in cells harboring integrated, but not episomal HPV16 DNA (Bechtold et al., 2003), raising a question about how E2 might function. In general, HPV16 E2 (16E2) by transient transfection is toxic to mammalian cells and is usually under detection level, suggesting other functions of E2 in the induction of cell toxicity. The finding that HPV5 E2 (5E2) facilitates RNA splicing and interacts with SR proteins (Lai et al., 1999) suggests that HPV E2 might play a role at post-transcriptional level. RNA transcription and splicing are coupled processes and many transcription factors have unexpected roles in this coupled network (Maniatis and Reed, 2002).

HPV E4 is expressed as an E1^E4 protein of 92 aa residues in which the N-terminal 5 aa residues are derived from the E1 ORF spliced to the E4 ORF. The E4 protein which is ~10 kDa in size is the most abundantly expressed HPV protein and accumulates in differentiating cells of the upper epithelial layers. E4 expression collapses the cytokeratin network (Doorbar et al., 1991) and mediates cell cycle arrest in G2 (Davy et al., 2002; Nakahara et al., 2002). The E1^E4 protein of HPV16 also binds to a DEAD-box containing RNA helicase, but the function of this association remains unknown (Doorbar et al., 2000). Since DEAD-box proteins regulate gene expression mostly at post-transcriptional levels, this observation implies that HPV16 E1^E4 might be involved in functions other than its effect on cytokeratin.

HPV16 E6 (16E6) and E7 (16E7) are viral oncoproteins that inactivate, respectively, cellular p53 and pRB, two tumor suppressor proteins essential for cell cycle control. 16E6 is an ~18 kDa nuclear protein and is composed of 151 aa residues. 16E6 contains two hypothetical zinc fingers involved in zinc binding (Kanda et al., 1991) and three nuclear localization signals (NLS) (Tao et al., 2003) as well as a PDZ-binding site in the N-terminus (Kiyono et al., 1997; Lee et al., 1997). Besides its ability to immortalize and transform cells and induce p53 degradation, 16E6 is also functionally involved in the regulation of gene transcription (Desaintes et al., 1992; Klingelhutz et al., 1996) through interaction with other transcription factors and coactivators (Patel et al., 1999; Ronco et al., 1998; Kumar et al., 2002; Veldman et al., 2001; Veldman et al., 2003; Thomas and Chiang, 2005). However, it remains to be determined what part of 16E6 is responsible for these protein-protein interactions because 16E6, like16E2, is usually under detection level in cells after transient transfection. 16E7 is a nuclear protein with 98 aa residues in size. The N-terminal 37 aa residues of 16E7 have been characterized as an important portion of the protein that contributes to pRB binding and degradation as well as cell transformation. Similar to 16E6, 16E7 also interacts with cellular transcription factors and coactivators (Massimi et al., 1997; Avvakumov et al., 2003; Bernat et al., 2003; Huang and McCance, 2002). Many of these interactions appear to involve the C-terminal half of 16E7, but their biological relevance remains to be understood. In addition to the protein-protein interactions of both 16E6 and 16E7, 16E6 has been shown to be a DNA binding protein (Imai et al., 1989; Ristriani et al., 2000; Ristriani et al., 2001; Nomine et al., 2003); the function of this DNA binding also remains unknown.

We recently showed that GFP-16E6 fusion is a nuclear and specifically, a nucleolar protein (Tao et al., 2003), and its nuclear or nucleolar localization is controlled by its three NLSs. Since the nucleoli are enriched with various small RNAs and NLS3 seems to strengthen, rather than to promote, the nuclear localization of the protein, we proposed that the major function of NLS3 might be to retain the protein in the nucleus, perhaps through an interaction with a nucleic acid (DNA or RNA) (Tao et al., 2003). Based on these observations and difficulties to express detectable HPV16 E2 and E6 in mammalian cells, we expressed E2, E4, E6 and E7 as GST-fusion proteins from individual HPV16 ORFs in E. coli and analyzed their possible interactions with RNAs and involvements in post-transcriptional regulation. We report here that both 16E2 and 16E6 are RNA binding proteins. 16E2 binds RNA through its characteristic DNA-binding domain and 16E6 binds RNA via its C-terminal half. Perhaps most importantly, both 16E2 and 16E6 were found to suppress RNA splicing and to interact with SR proteins.

Results

Expression of HPV proteins as GST-fusion proteins in E. coli

Various expression conditions were used to efficiently express each viral protein as a GST fusion in two bacteria strains, BL21(DE3)pLysS and BL21 Codon Plus (DE3)-RP, a bacteria strain harboring extra copies of tRNA genes that are rare in E. coli, but common in humans. Optimized expression conditions were obtained empirically for each protein and are summarized in Table 1. HPV16 E2 and E6, as well as some of their truncated mutants, were difficult proteins to express in BL21(DE3)pLysS. When expressed in this bacteria strain at room temperature or at 37°C, they tended to form inclusion bodies and were toxic to the bacteria. Especially when 16E2 was expressed, the transformed bacteria had difficulty reaching the OD600 value (0.6) required for the initiation of IPTG induction and frequently collapsed during early exponential growth. However, 16E2, 16E6, and some of their mutants were expressed relatively more efficiently, with fewer inclusion bodies, in the BL21 Codon Plus (DE3)-RP strain. Other proteins, such as E4 and E7 were relatively easy to express in either bacteria strains at room temperature or at 37°C with a few hours of induction.

Table 1

Optimized conditions for expressing HPV-GST fusion proteins in E. coli a

GST fusionsIPTG (mM)bInduction at °CInduction time (h)
16E2 wt0.71820
16E2ΔC2271252
16E2ΔN2200.51820
16E2ΔN25911820
16E2ΔC2601252
16E2ΔN220+C2601252
16E6 wt0.25253
16E6 mtc0.25253
16E6ΔN410.7252
16E6ΔN1021252
16E6ΔN102 mtd1252
16E6ΔC1031252
16E6ΔC421252
16E6ΔN41+ΔC10311820
16E4 wt1252
16E7 wt1372
5E2 wt0.71820
6E6 wt0.5253
GST1372
aBL21 Codon Plus (DE3)-RP strain.
bFinal concentration.
cpoint mutations in both NLS1 and NLS3 motifs (Tao et al., 2003).
dPoint mutations in NLS3 motif (Tao et al., 2003).

Effects of HPV16 E2, E4, E6 and E7 on RNA splicing

HPV16 E2-, E4-, E6- and E7-GST fusion proteins extracted to near homogeneity (Fig. 1A) were examined for their possible role in RNA splicing. We utilized an RNA substrate, HPV16 E6E7 pre-mRNA(Zheng et al., 2004), containing an exon 2 that was short but had a U1 binding site attached as a splicing enhancer (Fig. 1B). The viral fusion proteins (at 0.1, 1.0 or 5.0 μg) (5 μg of semipurified GST fusions were equivalent to approximately 500 ng of an expected pure viral fusion protein, see Materials and Methods) were pre-mixed with HeLa or 293 nuclear extracts (NE), followed by the addition of the RNA substrate to be tested in a splicing condition. As shown in Fig. 1C, HPV16 E6E7 RNA was spliced efficiently in the presence (lanes 2–4) or absence (lane 1) of 0.1, 1.0 or 5 μg of GST, but its splicing efficiency was reduced in the presence of 16E2- and 16E6-GST fusion proteins in a dose-dependent manner, reaching up to 3-fold reduction at 5 μg (equal to only 4% of total NE protein in a splicing reaction) (lanes 7 and 13), indicating that the 16E2- and 16E6-GST fusions are suppressive. HPV16 E4- and E7-GST fusions at 5 μg exhibited a smaller suppressive effect (less than 2-fold, lanes 10 and 16). We therefore focused our investigation on 16E2 and 16E6 for the remaining studies in this report.

An external file that holds a picture, illustration, etc. Object name is nihms106988f1.jpg

Inhibitory effect of the GST-HPV16 E2, E4, E6, and E7 fusion proteins on HPV16 E6E7 pre-mRNA splicing. (A) SDS-PAGE analysis of purified GST and GST-HPV16 E2, E4, E6 and E7 fusion proteins (marked with *). Five μl of GST and each GST-fusion protein expressed in E. coli was resolved on a 10% SDS-PAGE gel and stained with Coomassie blue. (B) The structure of HPV16 E6E7 pre-mRNA. The nucleotide positions of the exons (boxes) and an intron (line) in the virus genome are indicated below the structure. The black box on the RNA 3′ end indicates a U1 binding site attached as a splicing enhancer. (C) Effect of GST and the GST-HPV16 E2, E4, E6 and E7 fusion proteins on splicing of HPV16 E6E7 pre-mRNA. In vitro RNA splicing reactions were performed in the presence or absence of 0.1, 1.0, or 5.0 μg of each fusion protein for 2 h at 30°C. The splicing products were analyzed by electrophoresis on a 6% denaturing PAGE gel. The splicing efficiency (% spliced) of each reaction is shown at the bottom of the gel. The identities of the spliced products and splicing intermediates are shown to the right of the gel. M, size markers.

We further examined whether 16E2 and 16E6 with or without GST (Fig. 2A) would inhibit the splicing of BPV-1 late pre-mRNA (Fig. 2B) as described (Zheng et al., 1996). When mixed with HeLa NE, both 16E2 and 16E6 with or without GST suppressed the splicing. 16E2 and 16E6 lacking the GST fusion appeared to be more suppressive (Fig. 2C, compare lanes 3 and 4 to lanes 7 and 8). The same was true for 16E2 and 16E6 when used for in vitro splicing of Drosophila doublesex (D. dsx) exon 3 and exon 4 pre-mRNA (data not shown). Data indicate that 16E2 and 16E6 suppress splicing of all tested pre-mRNAs, with little change by GST fusion.

An external file that holds a picture, illustration, etc. Object name is nihms106988f2.jpg

Suppression of BPV-1 late pre-mRNA splicing by HPV16 E2 and E6 with or without GST fusion. (A) Electrophoretic profile of purified proteins without GST fusion. GST was removed by digestion with PreScission protease, and the proteins were resolved on 4%–12% SDS-PAGE gels stained with Coomassie Blue. (B) Structure of BPV-1 late pre-mRNA containing an exonic splicing enhancer, SE1, in its exon 2. The sizes (nts) of two exons (boxes) and an intron (line) are indicated below the diagram. (C) Splicing gels. The RNA splicing reactions were performed in the presence of 5 μg of 16E2 or 16E6 protein with (gel on the right) or without (gel on the left) GST fusion. Lanes 1 and 5 were unspliced pre-mRNA controls. The splicing products were resolved on 6% denaturing PAGE gels and are identified between the two gels. Splicing inhibition (% inh.) by each protein is shown below each gel. M, size markers.

Mapping of the HPV16 E2 and E6 regions involved in splicing suppression of pre-mRNAs containing a suboptimal splice site

To define which regions of 16E2 and 16E6 are involved in suppressing the splicing of pre-mRNAs, we constructed a series of 16E2 and 16E6 deletion mutants that were linked in frame to the C-terminus of GST. Since three functional domains within the 365-aa 16E2 protein have been proposed and analyzed partially (Sakai et al., 1996; Hegde and Androphy, 1998; Antson et al., 2000), the deletion mutants of 16E2 were constructed based on the study by Sakai et al. (Sakai et al., 1996) by deleting the C-terminal DNA binding domain, the N-terminal transactivation domain, or both, with or without deleting the hinge region (Fig. 3A). Because little is known about the functional domains of 16E6, we made successive deletions beginning at the N-terminus and progressing to the C-terminus of the 151-aa protein, or vice versa (Fig. 3B). All mutants (mt) with deletions were expressed as GST fusion proteins and examined for their effects on splicing of BPV-1 pre-mRNA.

An external file that holds a picture, illustration, etc. Object name is nihms106988f3.jpg

Mapping of HPV16 E2 and E6 regions affecting splicing of BPV-1 late pre-mRNA. (A) Schematic diagrams of full-length and truncated 16E2 (left) and their expression from E. coli (right) as well as their effect on RNA splicing. The numbers above each line diagram indicate the first and last aa residue positions in each E2 protein. Expression profiles of each GST-E2 fusion protein were resolved on a 10% SDS-PAGE gel stained with Coomassie blue, with (+) or without (−) inhibitory effect of the individual E2 protein on RNA splicing indicated below. (B) Schematic diagram of full-length and truncated 16E6 (left) and their expression from E. coli (right) as well as their effect on RNA splicing. See other details in panel A. (C, D) Inhibitory effects of the individual E2 (C) or E6 (D) fusion protein (5 μg) on RNA splicing of BPV-1 late pre-mRNA (see Fig. 2B for RNA structure). Lane 1 in (C) or (D) was unspliced pre-mRNA controls. See other details in Fig. 1C. Splicing inhibition (% inh.) by each protein is shown below each gel.

As shown in Fig. 3C, full-length wt 16E2 suppressed splicing (> 2-fold) of BPV-1 late pre-mRNA. Similarly, mt 16E2 without the C-terminal DNA binding domain and with or without a hinge region (E2ΔC227 and E2ΔC260) was also suppressive by ~ 40–42%, suggesting that the N-terminal transactivation domain is important for the suppression. Deletion of the N-terminal transactivation domain of 16E2 (E2ΔN220 and E2ΔN259) was detrimental to the suppression, further supporting this conclusion. However, the E2 hinge region itself, when expressed as GST-fusion protein (E2ΔN220+ΔC260), also inhibited the splicing by 44%, but when the hinge was combined with the C-terminal DNA binding domain (E2ΔN220), the suppression was only minimal (Fig. 3A and 3C), implying that the hinge in E2ΔN220 fusion may have a different conformation than the hinge alone in the E2ΔN220+ΔC260 fusion.

E6 with a deletion of the N-terminal 41 aa residues (E6ΔN41) or of the C-terminal 49 aa residues (E6ΔC103) was able to suppress BPV-1 RNA splicing by 46% or 34%, respectively (Fig. 3D), but the protein without the central portion from aa 42 to 102 (E6ΔN102 and E6ΔC42) had only a mild inhibitory effect (20% and 27%, respectively), indicating that the central region (aa 42–102) of 16E6 is more suppressive. The finding of that the fusion protein with the central region of 16E6 alone (E6ΔN41+ΔC103) was even more suppressive (65%) than any other E6 fusions, including wt full-length E6 (38%) (Fig. 3B and 3D), supports this conclusion.

To confirm our mapping results, 16E2, 16E6, and the mutants that retained the ability to suppress splicing were further examined on splicing of HPV16 E6E7 pre-mRNA. The results for individual proteins were very similar to the results for BPV-1 (data not shown). Similar results were also obtained for splicing of D. dsx pre-mRNA (data not shown). We also tested the suppressive proteins without GST (Fig. 4A), and found that all wt and mt forms of 16E2 (Fig. 4B, lanes 5–7) and 16E6 tested (data not shown) were suppressive. However, 5E2 with (lane 1, Fig. 4C and 4D) or without GST (lane 4, Fig. 4B), which was reported to enhance RNA splicing in vivo (Lai et al., 1999), did not have much suppressive activity.

An external file that holds a picture, illustration, etc. Object name is nihms106988f4.jpg

Effect of full-length or truncated HPV16 E2 without a GST tag and full-length HPV5 E2 on splicing of pre-mRNAs. (A) Purified wt and truncated 16E2 (lanes 2–4) without GST and 5E2 with GST (lane 6). All proteins were digested at 4°C overnight with PreScission protease, and each protein was resolved on a 4%–20% SDS-PAGE gel followed by Coomassie blue staining. (B) Effect of full-length and truncated 16E2 proteins and full-length 5E2 protein without GST on 16E6E7 RNA splicing. The splicing products were analyzed on a 6% denaturing PAGE gel. Lane 2 in (B) was unspliced pre-mRNA controls. (C, D) Effect of GST-5E2 on splicing of 16E6E7 (C) and BPV-1 late (D) pre-mRNAs. The identities of the spliced products and splicing intermediates are shown on the right of the gel. See other details in Fig. 2C. Splicing inhibition (% inh.) by each protein is shown below each gel.

Although both 16E2 and 16E6 inhibited the splicing of all three pre-mRNAs (HPV16 E6E7, BPV-1ate, and D. dsx) bearing a suboptimal intron, these proteins and most of their suppressive mutants showed only a little effect on human β-globin pre-mRNA which contains a strong intron 1 (data not shown), suggesting that the observed suppressive effects of 16E2 and 16E6 on RNA splicing are dependent upon the features of the pre-mRNAs. One of the 16E6 mutants, 16E6ΔN41+ΔC103, did have some suppressive effect (~30%) on β-globin pre-mRNA (Data not shown); this mutant was also the strongest inhibitor of splicing for all of the other pre-mRNAs.

HPV16 E2 and E6 are RNA-binding proteins and contain a protein-RNA interaction domain in their C-terminal regions

To understand how 16E2 and 16E6 affect RNA splicing, we carried out a series of protein-RNA interaction experiments. We hypothesized that 16E2 and 16E6 could affect RNA splicing by binding to the RNA. Although both 16E2 and 16E6 have been demonstrated to be DNA-binding proteins (Thain et al., 1997; Ristriani et al., 2000; Ristriani et al., 2001), it was not known whether the two proteins could also bind to RNAs. Since both 16E2 and 16E6 induced degradation of HPV16 E6E7 pre-mRNA in our gel-shift conditions (data not shown), we therefore analyzed full-length 16E2, 16E6, and their truncated mutants for RNA-binding (Fig. 5A and 5B) by UV cross-linking and RNase digestion assays. In these assays, a protein that binds to a 32P-labeled RNA in close contact can be covalently crosslinked to the RNA by UV irradiation and consequently prevents the bound RNA region from RNase digestion. As shown in Fig. 5C and 5D, both full-length 16E2 and 16E6 fusion proteins exhibited RNA-binding activities for HPV16 E6E7 pre-mRNA (lane 1 in C and D), whereas GST itself did not (lanes 7 and 10 in C and lane 7 in D). All of the fusion proteins and GST itself sometimes displayed a non-specific binding band (arrows in C and D) which overlaps the position of full-length 16E2 (lane 1 in C). However, a wt, full-length E2-RNA binding reaction gave a much higher density band at this position, representing the binding of 16E2 to the RNA (lane 1 in C). In separate experiments using a gradient gel, the 16E2-specific band could be separated from the non-specific band (lane 8 in C) which was not a problem for the much smaller, wt 16E6 fusion protein (lane 1 in D). Thus, we conclude that both 16E2 and 16E6 are RNA-binding proteins.

An external file that holds a picture, illustration, etc. Object name is nihms106988f5.jpg

Mapping of HPV16 E2 and E6 domains involved in protein-RNA interaction using UV cross-linking and RNase digestion. (A, B) Diagrams of various 16E2 (A) and 16E6 (B) fusions and their binding activities to HPV16 E6E7 pre-mRNA. (C) 16E2 RNA-binding gel. (D) 16E6 RNA-binding gel. Each protein (see protein profiles in Fig. 3A and B) was incubated first with α-32P-labeled HPV16 E6E7 pre-mRNAs (see RNA structures in Fig. 1B) before UV cross-linking and RNase A/T1 digestion. The protein-RNA complexes were then heated to 95°C for 10 min in 2X SDS sample buffer and resolved on 10% [left gel in (C) and gel in (D)] or 4%–12% gradient [right gel in (C)] SDS-PAGE gels. One representative experiment of five is shown. Arrows, non-specific (NS) binding; *, specific binding.

Fig. 5C also shows that the truncated 16E2 fusion proteins retaining the N-terminal transactivation domain with or without a hinge region (E2ΔC227 and E2ΔC260, lanes 2 and 5) exhibited no RNA-binding activity, whereas the truncated 16E2 proteins containing the C-terminal DNA-binding domain with (E2ΔN220, lane 3) or without (E2ΔN259, lanes 4 and 9) the hinge region bound to HPV16 E6E7 pre-mRNA. The hinge region itself (lane 6) had no RNA-binding activity and neither of its extended version [20 aa more on both ends (Antson et al., 2000)] could render such binding (data not shown).

Fig. 5D shows mapping analysis of 16E6 to identify the region in 16E6 responsible for the RNA binding, using all six 16E6 fusions. Three of the six proteins, wt E6, E6ΔN41, and E6ΔN102, bound to RNA (lanes 1–3), but the other three fusion proteins, E6ΔC103, E6ΔC42, and E6ΔN41+ΔC103, did not (lanes 4–6). The truncated 16E6 without the N-terminal 102 aa residues retained the RNA binding activity (E6ΔN102, lane 3), but a 16E6 mutant containing only the N-terminal 102 aa residues was deficient in the binding (E6ΔC103, lane 4), indicating that the C-terminal 48 aa residues of 16E6 is responsible for the protein-RNA interaction.

In addition to their binding to HPV16 E6E7 pre-mRNA, both HPV16 E2 and E6 were also shown to bind BPV-1 late and human β-globin pre-mRNAs in separate experiments (data not shown).

A NLS3 sequence motif in the C-terminus of HPV16 E6 is responsible for protein-RNA interactions

The C-terminus of 16E6 contains a NLS, NLS3. NLS3 is unable to convert the cytoplasmic E6 of low-risk HPV6 (6E6) into a nuclear protein, but appears to strengthen nuclear localization of 16E6; we have therefore proposed that the major function of NLS3 might be to retain the protein in the nucleus, probably through a nucleic acid (DNA or RNA) interaction (Tao et al., 2003). To examine whether the RNA binding of 16E6 could be a function of NLS3, the C-terminal 48-aa region with or without an NLS3 mutation (Fig. 6A)(Tao et al., 2003) was examined for its ability to bind RNA. The C-terminal region containing a mutant NLS3 lost its RNA binding activity (compare E6ΔN102 in lanes 1 and 3 to E6ΔN102 mt in lane 2, Fig. 6B), indicating that the NLS3 sequence motif is responsible for the 16E6-RNA interaction. This result was further confirmed in the context of wt, full-length 16E6 fusion protein with or without mutations in both NLS1 and NLS3 (compare lane 4 to lane 5, Fig. 6B). Since the N-terminal region containing wt NLS1 (Tao et al., 2003) (see E6ΔC42 in Fig. 5D, lane 5) had no RNA binding activity, the loss of the RNA binding activity in the full-length 16E6 with mutations in both the NLS1 and NLS3 had to be a result solely from mutation in the NLS3.

An external file that holds a picture, illustration, etc. Object name is nihms106988f6.jpg

High-risk HPV16 E6 containing NLS3 mutation and low-risk HPV6 E6 lack RNA binding activity. (A) Diagrams of wt or mt 16E6 fusion proteins with (black boxes) or without (crossed black boxes) a wt NLS3 and their RNA binding activities in comparison with 6E6. 6E6 has a non-functional NLS3-like structure, with few aa residue variations, in the region corresponding to 16E6 NLS3 (Tao et al., 2003) and is a cytoplasmic protein. (B) A 16E6 NLS3 motif is responsible for protein-RNA interaction. GST-16E6 fusion protein with or without mutations in both NLS1 and NLS3 (Tao et al., 2003) and a truncated 16E6ΔN102 fusion protein with or without mutations in its NLS3 motif were examined for their binding and UV cross-linking to 16E6E7 pre-mRNA. (C) HPV6 E6 is not an RNA binding protein. GST-6E6 was compared with wt or truncated 16E6 for its RNA binding. HPV16 E6E7 pre-mRNA (see Fig. 1B) was used for each assay. One representative experiment of three is shown in (B) and (C). Arrows, non-specific (NS) binding; *, specific binding.

We also compared HPV16 E6 with the low-risk 6E6 for their ability to bind RNA since the 6E6 is a cytoplasmic protein and its corresponding region to the 16E6 NLS3 has variations in aa residues. Data in Fig. 6C shows, as expected, that the low-risk cytoplasmic 6E6 lacked this RNA-binding activity (compare lane 4 to lane 2).

HPV16 E2 and E6 preferentially interact with the intron region of a pre-mRNA

We next wished to identify the RNA sequences targeted by the C-terminal binding domain of 16E2 and the NLS3 of 16E6. HPV16 E6E7 pre-mRNA was chosen because both 16E2 and 16E6 bind to this RNA which is available for the binding during virus infection. Various sizes of 16E6E7 pre-mRNA transcribed in vitro and uniformly labeled with 32P (Fig. 7A) were utilized for the mapping in the presence of the 16E2ΔN259 (Fig. 7B) or 16E6ΔN102 (Fig. 7C) fusions. The 16E2ΔN259 and 16E6ΔN102 fusions were selected for the assay because they are smallest proteins to confer the same amount of RNA binding activity as their full-length counterparts (Fig. 5C and D). As shown in Fig. 7, almost all the RNA fragments used for the assay, except for RNA C, interacted efficiently with the two proteins. RNA C is composed solely of the exon 1 region of the 16E6E7 pre-mRNA and exhibited no binding to 16E2ΔN259 (Fig. 7B) or 2–3-fold less binding affinity to 16E6ΔN102 (Fig. 7C) than that of the other RNAs. All of the RNAs that did bind (A, B, D, E, and F) contained all or part of intron 1, and RNA E and RNA F contained only intron 1. We therefore concluded that intron 1 of 16E6E7 pre-mRNA is likely to be the region interacting with the 16E2ΔN259 and 16E6ΔN102 fusions.

An external file that holds a picture, illustration, etc. Object name is nihms106988f7.jpg

Intron regions of HPV16 E6E7 pre-mRNA have high affinity for binding of a truncated 16E2 or 16E6 fusion protein. (A) Schematic diagrams of HPV16 E6E7 pre-mRNAs used for in vitro protein binding and UV cross-linking. On the top of the panel is a diagram of the HPV16 E6E7 pre-mRNA including exon 1, intron 1, and exon 2, showing various 5′ ss and 3′ ss. The numbers at each RNA end are the nt positions in the virus genome. (B, C) Mapping of HPV16 E6E7 RNA regions that interact with a truncated 16E2ΔN259 (B) or 16E6ΔN102 (C) fusion protein. See Fig. 5 for other details. Protein-RNA complexes from each reaction were resolved on a 10% SDS-PAGE gel. The letters at the tops of each gel indicate the individual RNA substrate in (A) used for the binding and UV cross-linking. One representative experiment of two is shown. NS arrows, non-specific binding; S arrowheads, specific RNA-protein interactions.

Since 16E6E7 pre-mRNA contains three alternative 3′ splice sites in intron 1, the size of intron 1 in the pre-mRNA would vary depending on which 3′ splice site is selected for the splicing. Thus, the exon sequences upstream of an alternative 3′ splice site could be part of intron 1. If the intron region binds 16E2ΔN259 and 16E6ΔN102, the part of the intron upstream of an alternative 3′ splice site should have the same binding affinity. This prediction was confirmed using RNAs G, H, and I and further verified using RNA J, which was transcribed from an E6E7 cDNA lacking intron 1 from nt 227 to nt 408, but remaining the entire portion of RNA G (Fig. 7B and 7C). Altogether, these data indicate that the entire intron 1 region of 16E6E7 pre-mRNA contains multiple sequence motifs that interact directly with both 16E2ΔN259 and 16E6ΔN102 and prevent the bound RNA from RNase digestion.

HPV16 E2 and E6 interact with multiple cellular SR proteins

To further explore how 16E2 and 16E6 could suppress pre-mRNA splicing by interacting with RNA, we investigated protein-protein interactions using a GST pull-down assay, under the assumption that the two viral proteins might interfere with the cellular splicing machinery in addition to their RNA binding activities. Full-length and several truncated 16E2 fusions were chosen for GST pull-down assays using HeLa NE. The proteins that were pulled down were examined by Western blotting using mAb104, a monoclonal anti-SR protein antibody that recognizes a phosphoepitope in the RS domain of all classical SR proteins (Zahler et al., 1993). As shown in Fig. 8A, the full-length 16E2 was found to interact with SRp30, SRp40 and SRp75 (lane 3), but only weakly with SRp55. The truncated 16E2ΔN259 which binds RNA but has little effect on RNA splicing interacted with SRp30, SRp40 and very little SRp55, but not SRp75 (lane 4). A protein with a size little larger than SRp75 that could be stained by mAb104 but was absent at the same position from NE (compare lane 4 to lane 1) appeared to be non-specific. GST (lane 6) and the E2ΔN220+ΔC260 fusion containing the E2 hinge region only (lane 5) did not interact with any SR protein in the assay. Several other forms of truncated 16E2, including E2ΔC227 and E2ΔC260, also showed no SR protein interaction (data not shown). These data suggest that the 16E2-SR protein interaction domain (s) is distributed most likely across the C-terminus of the protein.

An external file that holds a picture, illustration, etc. Object name is nihms106988f8.jpg

HPV16 E2 and E6 fusion proteins interact with SR proteins. Full-length or truncated 16E2 (A) or 16E6 (B) fusions as well as GST protein freshly expressed from E. coli were immobilized on glutathione Sepharose 4B beads in microspin column and then incubated overnight at 4°C with HeLa NE. After extensive washes, the captured proteins were resolved on a 4%–12% SDS-PAGE gel, transferred onto a nitrocellulose membrane, and blotted with mAb104. HeLa NE was loaded as an SR protein control in lane 1 (2.5 μg or ~2% of the input) in (A) and in lane 3 (5.8 μg or ~5% of the input) in (B). See E2 and E6 protein diagrams in Fig. 3A and 3B. Shown on the left in each panel are four major SR proteins in HeLa NE that are recognized by mAb104. One representative experiment of three is shown. (C) An NLS3 motif in HPV16 E6 interacts with SR proteins. Proteins in the pull-downs were analyzed by Western blotting using mAb104. Lanes 3–7 in (C) also show some background staining for each input fusion protein. One representative experiment of two is shown. M, pre-staining size markers (lane 2 in A, B, and C).

We also examined full-length 16E6 and two truncated 16E6 fusions, 16E6ΔN102 and 16E6ΔN41+ΔC103, for SR protein interactions in GST pull-down assays using HeLa NE. 16E6ΔN102 and 16E6ΔN41+ΔC103 were chosen because 16E6ΔN102 binds RNA, but has little effect on RNA splicing, and 16E6ΔN41+ΔC103 does not bind RNA, but suppresses RNA splicing. As shown in Fig. 8B, wt 16E6 interacted with SRp30, SRp55, and SRp75 (lane 4). 16E6ΔN102 interacted weakly with SRp30, SRp40, and SRp55, but not SRp75 (lanes 5 and 7). A protein with a size little smaller than SRp40 in the 16E6 pull-downs (compare lane 4 to lane 3) or a little larger than SRp75 in the 16E6ΔN102 pull-downs (compare lanes 5 and 7 to lane 3) that could be stained by mAb104 but was absent at the corresponding positions from NE appeared to be non-specific. In contrast, 16E6ΔN41+ΔC103 interacted much less with SRp30 and SRp75 (lane 1). Thus, as with 16E2, the C-terminus of 16E6 contains most of the SR protein interaction domains.

To identify which portion of the C-terminus of 16E6 interacts with SR proteins, we compared the SR protein interactions of 16E6ΔN102 and 16E6ΔN102 mt. 16E6ΔN102 mt differs from 16E6ΔN102 by mutation of five aa residues in its NLS3 motif (see NLS3 mutations, Fig. 6A). The introduction of point mutations in the NLS3 motif disrupted RNA binding (Fig. 6B) and was detrimental to the protein-protein interaction (Fig. 8C, compare lane 3 to lane 5). Despite of some background staining of the input fusion proteins, Fig. 8C also shows that the N-terminal 41 aa residues of 16E6 had no activity for binding SR proteins (lanes 4 and 7). Therefore, the C-terminal NLS3 of 16E6 that binds RNA also binds SR proteins.

Discussion

In this study, we have demonstrated that both 16E2 and 16E6 proteins inhibit splicing, and this inhibition of RNA splicing may be associated with their RNA and protein-protein interactions. We have also demonstrated that in both 16E2 and 16E6, suppression of RNA splicing and interaction directly with RNA are two separate functions that are performed by different domains: the N-terminal half of each protein participates in the suppression of RNA splicing, and the C-terminal half contributes to RNA binding (Fig. 9), despite the fact that the C-terminal half does not contain any well characterized RNA recognition motifs (Burd and Dreyfuss, 1994). 16E2 and 16E6 must be in close contact directly with the bound RNAs and therefore could be covalently cross-linked to the bound RNAs to prevent from RNase digestion. Several reports have indicated that NLS motifs overlap DNA- or RNA- binding domains in nuclear acid-binding proteins, and zinc fingers in these proteins are also possible RNA-binding motifs (LaCasse and Lefebvre, 1995). HPV16 E6 is known to contain these structures (Tao et al., 2003), but they have not been identified in the 16E2 protein. More importantly, we have demonstrated that both 16E2 and 16E6 interact with a subset of classical SR proteins, cellular splicing factors that are pivotal components of the cellular splicing machinery. Unlike 5E2, which contains multiple RS dipeptide repeats in a long, 200-aa hinge region and interacts with SR proteins and other splicing factors, such as SRp30 (ASF/SF2 and SC35), U1-70K, and U5-100K, through its RS repeats (Lai et al., 1999), 16E2 has a short, 40-aa hinge (Sakai et al., 1996) and lacks repeated RS sequences. 16E6 lacks all of those structures. Although 5E2 has been suggested to facilitate RNA splicing in vivo of CAT pre-mRNAs containing a β-globin intron (Lai et al., 1999), we were unable to observe this activity in splicing of any pre-mRNAs examined in our in vitro assays.

An external file that holds a picture, illustration, etc. Object name is nihms106988f9.jpg

Functional domains of HPV16 E2 and E6 involved in RNA binding, RNA splicing, and SR protein interactions. (A) HPV16 E2. (B) HPV16 E6. (C) A proposed model showing how HPV16 E2 and E6 might interfere with SR protein-mediated intron removal through interacting with SR proteins or other unknown factors and/or with an intron region of pre-mRNA.

Functional analysis of each protein domain showed that the region of the C-terminus with RNA binding activity overlaps with the region responsible for SR protein interactions in both 16E2 and 16E6 (Fig. 9A and 9B), as exemplified by the NLS3 motif of 16E6. This overlapping functions are not necessarily associated with the suppressive effects of 16E2 and 16E6 on RNA splicing, since truncated proteins containing the C-terminal portion alone had no significant effect on RNA splicing, yet retained interactions with RNA and weakly with a subset of SR proteins. However, that only the wt, full-length 16E2 or 16E6, but not the truncated ones, were able to interact strongly with SRp75 suggests that another region of 16E2 or 16E6 or conformational structures of these two proteins might interact with SRp75 to mediate splicing suppression. Together with the finding that both 16E2 and 16E6 preferentially bind to a pre-mRNA intron, the interactions of 16E2 and 16E6 with RNA and SRp75 may affect RNA splicing by interfering recruitment and cross-talk of the splicing machinery over the intron during RNA splicing (Fig. 9C). It has been known that binding of SR proteins to an intronic region negatively regulates RNA splicing (Kanopka et al., 1996) and many intronic elements are regulatory elements in RNA splicing (Ladd and Cooper, 2002)

Other unidentified cellular factors or mechanisms might also be involved in 16E2- and 16E6-mediated splicing suppression. For example, several of our truncated proteins with C-terminal deletions, such as 16E2ΔN220+ΔC260, 16E2ΔC260 and 16E6ΔC103, had no activity in binding of RNAs or classical SR proteins, but suppressed RNA splicing. How these proteins block RNA splicing remains unknown. However, it is important to note that in addition to interact with SR proteins, HPV5 E2 is also capable to interact with other splicing components, such as U1-70K and U5-100K (Lai et al., 1999). Since the spliceosome comprises up to 150 different proteins (Zhou et al., 2002) and the mAb104 in our pull-down western blotting assays detects only a subset of the classical SR proteins (Zahler et al., 1993; Graveley, 2000). It is possible that these truncated proteins might interact with other unknown cellular factors to disrupt functions mediated by the normal splicing machinery.

Nevertheless, both wt 16E2 and wt 16E6 appeared to inhibit splicing of all 3 pre-mRNAs containing a suboptimal intron in our assays (16E6E7, BPV-1 late, and D. dsx), but had very little effect on splicing of a human β-globin pre-mRNA, which contains an optimal intron. The difference in RNA splicing among these pre-mRNAs could therefore be due to the intrinsic structures of the pre-mRNAs. A pre-mRNA containing a suboptimal intron or weak splice sites would not be spliced efficiently in the absence of an exonic splicing enhancer that is necessary, through interaction with SR proteins, for recruitment of U2AF to a suboptimal 3′ splice site. In contrast, splicing of a pre-mRNA containing an optimal intron or strong splice sites usually does not rely on an exonic splicing enhancer (Zheng et al., 2000; Zheng, 2004), and the 3′ splice site itself in this type of pre-mRNA is strong enough for U2AF and U2 interaction. Therefore, the intrinsic structure of a pre-mRNA may determine the consequences of the interactions among the mRNA, 16E2, 16E6, and SR proteins, as characterized in the present study.

However, these findings leave many questions to be addressed: First, how do the protein-RNA and protein-protein interactions take place in cells and if so, in what circumstances? Since RNAs are bound with all kinds of cellular RNA-binding proteins along transcription in progression, 16E2 and 16E6 must be in a competitive position for the binding of their targeted RNAs during HPV infection. Second, are the RNA and SR protein binding of 16E2 and 16E6 biologically significant in the virus life cycle and in viral oncogenesis? Unfortunately, many attempts to detect the RNA binding and SR protein interactions of 16E2 and 16E6 in vivo were unsuccessful due to our difficulties with expressing a detectable level of either 16E2 or 16E6 from mammalian expression vectors in transient transfection assays or in the established 16E2 or inducible 16E6 stable cell lines. Nevertheless, the findings that both 16E2 and 16E6 are RNA binding proteins, interact with cellular splicing factor SR proteins, and inhibit RNA splicing of pre-mRNAs with a suboptimal intron provides compelling evidence that some HPV regulatory proteins may also regulate gene expression at posttranscriptional level.

Materials and Methods

Construction of expression vectors

ORFs of E2 (nt 2756 to 3853), E4 (nt 865–880/3358 to 3620), E6 (nt 104 to 559) and E7 (nt 562 to 858) were amplified from pBR322-HPV 16R DNA (a gift from Dr. E.-M. de Villiers and H. zur Hausen) by individual PCRs. Each was then cloned in frame into the C terminus of GST at the EcoRI and XhoI sites of the polylinker region of the expression vector pGEX-6P-1 (Amersham Pharmacia, Piscataway, NJ), producing GST-E2, E4, E6 and E7 as individual GST fusion genes in plasmid pSB-1, pSB-2, pSB-3 and pSB-4, respectively. The same cloning strategy was used to create the following plasmids to express the truncated E2 and E6 fusion proteins described in Fig. 3A and B: plasmid pSB-5 for 16E2ΔC227 with a C terminal deletion from aa 227 to 365, plasmid pSB-6 for 16E2ΔN220 with an N-terminal deletion from aa 1 to 220, plasmid pSB-7 for 16E2ΔN259 with an N-terminal deletion from aa 1 to 259, plasmid pSB-8 for 16E2ΔC260 with a C-terminal deletion from aa 260 to 365, plasmid pSB-9 for 16E2ΔN220+ΔC260 with both an N-terminal deletion from aa 1 to 220 and a C-terminal deletion from aa 260 to 365 (thus, the fusion protein expressed from plasmid pSB-9 has only the 16E2 hinge [40 aa] (Sakai et al., 1996)), plasmid pSB-10 for 16 E6ΔN41 with an N-terminal deletion from aa 1 to 41, plasmid pSB-11 for 16E6ΔN102 with an N-terminal deletion from aa 1 to 102, plasmid pSB-12 for 16E6ΔC103 with a C-terminal deletion from aa 103 to 151, plasmid pSB-13 for 16E6ΔC42 with a C-terminal deletion from aa 42–151 and plasmid pST-14 for 16E6ΔN41+ΔC103 with both an N-terminal deletion from aa 1 to 41 and a C-terminal deletion from aa 103 to 151 (thus, the fusion protein expressed from pSB-14 contains only the central part [from aa 42 to 102] of E6). To express 16E6 with mutations in the NLS3 motif, two plasmids derived from plasmid pTMF45 (Tao et al., 2003) were constructed: plasmid pSB-16 constructed encodes a full-length E6 GST fusion with mutations in both the NLS1 and NLS3 motifs, plasmid pSB-17 constructed as described for pSB-11 encodes a 16E6ΔN102 fusion with mutations in the NLS3 motif and the protein expressed from this plasmid was designated 16E6ΔN102 mt. The same strategy was also used to construct plasmids to express wild-type (wt) HPV 5E2-GST and HPV6b E6-GST fusions; plasmid pSB-15 for 5E2 expression originated from plasmid pSP5E2 (a gift from Dr. Tarn) (Lai et al., 1999) and plasmid pSB-20 for HPV6b E6 (6E6) expression was derived from plasmid pZMZ66-1 (Tao et al., 2003).

Preparation of DNA templates and in vitro transcription

Various DNA templates were prepared for in vitro RNA transcription. The transcribed RNAs were then used for in vitro RNA splicing or RNA-protein interaction. To generate various pre-mRNAs for in vitro RNA splicing, four different DNA templates were prepared: HPV16 E6 DNA templates were generated from pBR322-HPV16R by PCR with a chimeric 5′ primer T7-HPV16 Pr107 (oZMZ 208, 5′-TAATACGACTCACTATAGGGA/TTTCAGGACCCACAGGAGCGA-3′) combined with an antisense 3′ primer with a 5′ end at nt 477 (oZMZ 219, 5′-GTACTCACCCC/AATCTTTGCTTTTTGTCC-3′); BPV-1 late gene DNA templates were generated by PCR from plasmid pZMZ19-1 (Zheng et al., 1996) with a chimeric 5′ primer T7-BPV-1 Pr7276 (oFD 127, 5′-TAATACGACTCACTATAGGGA/GCGCCTGGCACCGAATCC-3′) combined with an antisense primer (oZMZ 84, 5′-GGCTGGGCTGGCTCGGCTTCTTTT-3′) with a 5′ end at nt 3305; D. dsx DNA templates were obtained by linerization of pZMZ1-1 with Hind III (Zheng et al., 1996); and the human β-globin minigene in plasmid pSP64-HβΔ6 linearized with Bam HI was originally purchased from Promega (Madison, WI) and was described in our previous publication (Zheng et al., 1999).

To generate various sizes of RNA transcripts from HPV16 E6E7 DNA templates for the RNA binding assays described in Fig. 7, various 5′ chimeric T7-HPV16 primers at different locations were combined with various antisense 3′ primers in different positions for PCR amplification from pBR322-HPV16 DNA. The PCR DNA templates were then subjected to in vitro run-off transcription assays for the production of individual RNAs with desired regions. The positions and sizes of individual templates used for in vitro run-off transcription are reflected at the ends of each RNA described in Fig. 7A.

Capped, 32P-labeled run-off transcripts were synthesized from each DNA template by in vitro run-off transcription using T7 RNA polymerase in the presence of [α-32P]GTP (Zheng and Baker, 2000). However, SP6 RNA polymerase was used to transcribe the human β-globin transcript as suggested by Promega. All RNA transcripts were gel-purified and quantified with a liquid scintillation counter (Beckman Instruments, Palo Alto, CA). The labeled transcripts were stored at −80°C until use.

In vitro RNA splicing

Splicing reactions were carried out in the presence of HeLa or 293 NE at 30°C for 2 h. The detailed protocol for in vitro splicing and calculation of the splicing efficiency have been described elsewhere (Zheng and Baker, 2000). Splicing inhibition (%) by each protein was calculated as [1−(a/b)] ×100, where the a equals to the percentage of splicing efficiency in the presence of a testing protein and the b equals to the splicing efficiency in the absence of a testing protein. A splicing gel in each figure is a representative of at least 2–3 separate experiments.

Expression and purification of GST fusion proteins

The individual viral gene constructs in pGEX-6P-1 expression vectors described above were expressed as GST fusion proteins in E. coli either in strain BL21(DE3)pLysS (Promega) or in BL21 Codon Plus (DE3)-RP (Stratagene, La Jolla, CA). Codon Plus RP bacteria cells are engineered to encode extra copies of specific tRNA genes that are rare in E. coli, but common in humans. Overnight pre-cultures were diluted 100-fold in fresh 2X YTA medium (yeast extract, 10 g/l; tryptone, 16 g/l; sodium chloride, 5 g/l; pH 7.0) containing 100 μg/ml ampicillin and were incubated at 37°C until an OD600 of 0.6–0.8 was reached. The expression of the fusion proteins was induced with 0.25–1.0 mM IPTG at 18°–37°C for 2–20 h (see details in Table 1). Cells were collected by centrifugation and resuspended in 50 μl (for each milliliter of the culture centrifuged) of ice-cold 1X PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.3) in the presence of 1% (v/v) Triton X-100 and Complete Mini EDTA-free Protease Inhibitor Cocktail (1X, Cat.# 1836170, Roche, Indianapolis, IN). Resuspended cells were aliquoted into a fresh Eppendorf tube and stored at −20°C until protein purification. Bacteria cells were lysed in the presence of 100 μg/ml lysozyme and 3–5 U DNase I by 10 cycles of freeze/thaw, in combination with homogenization using a syringe. Subsequently, cell debris was removed from the lysate by centrifugation at 13,000 rpm for 10 min. Finally 750–1500 μl of supernatant was loaded onto a microspin column containing glutathione Sepharose 4B beads (50 μl bed volume, Amersham Pharmacia) and incubated with gentle rocking on a shaker for 20 min at room temperature. The beads were then washed 6 times with 150 gel-volumes of ice-cold PBS (1X). GST or recombinant fusion proteins were eluted at least 4 times from each column; each elution was performed by low speed centrifugation (735 × g) with 50 μl of 10 mM glutathione in the presence of 1 mM DTT and Complete Mini EDTA-free Protease Inhibitor Cocktail as described above. To remove the GST tag from selected GST fusion proteins, a total of 50 μl of cleavage buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM DTT, pH 7.0) and 4 U of PreScission protease (Amersham Pharmacia) was applied to the column after the last wash and before elution of the fusion proteins from the column. The column was then incubated at 4°C overnight. The digested proteins without GST were eluted four times, each with 50 μl of cleavage buffer. All protein samples were analyzed using SDS-PAGE gels stained with Coomassie Blue. The concentration of proteins was measured either by UV absorbance at 280 nm or by comparing the intensity of the band of the protein of interest to that of the standard protein marker (5 μg) or BSA (2 μg). The intensity comparison was used to calibrate the UV absorbance curve and 5 μg of purified GST fusions quantified from UV absorbance was found to contain approximately 500 ng of an expected pure protein. For our convenience, the concentration of GST fusion proteins measured with UV is described in the text for all experiments.

All purified proteins were freshly prepared from 1.5 ml aliquots of the transformed and induced bacteria suspension less than 2 weeks after induction and were analyzed immediately in each experiment. Occasionally, glycerol was added to individual protein preparation at a final concentration of 20% for a short storage at −80°C.

Protein pull-down assays and Western blotting

Each freshly expressed fusion protein or GST was first immobilized on glutathione Sepharose 4B beads and washed 7 times with 150 gel-volumes of ice-cold PBS (1X) as described above. The protein-immobilized beads was placed in a fresh microspin column (Amersham Pharmacia) and washed once at 800 × g for 5 min with 750 μl of buffer A [PBS (1X) with a final concentration of 1 mM each of EGTA, EDTA, KF, Na3VO4, β-glycerophosphate, 1% (v/v) Triton X-100 and Complete Mini EDTA-free Protease Inhibitor Cocktail (1 tablet/10 ml) as described above]. Twenty microliters (7.5 μg/μl) of HeLa NE (ProteinOne, College Park, MD) in 500 μl of buffer A were added to the microspin column containing the protein-immobilized beads, which was tumbled overnight at 4°C. After washing the beads 7 times with 750 μl of ice-cold buffer A, 50 or 120 μl of 4X SDS sample buffer, depending on the individual protein were added to the beads, which were then heated to 95°C for 10 min. The protein samples were resolved on a 4%–12% SDS-PAGE gel, transferred onto a nitrocellulose membrane, probed with a monoclonal anti-SR protein antibody mAb104 followed by a horseradish peroxidase-conjugated goat anti-mouse IgM (Sigma), and visualized using a SuperSignal West Pico chemiluminescence detection system (Pierce Biotech, Rockford, IL).

Protein-RNA interaction and UV crosslinking

The same procedure was followed as described previously (Zheng et al., 1997). Briefly, each purified fusion protein (~5 μg) was incubated with 8 ng (~2.8 × 105 cpm) of 32P-labeled RNA substrates in 5X splicing buffer containing 3% polyvinyl alcohol, 2 mM ATP, 100 mM creatine phosphate, and 25 mM HEPES (pH 7.9) in a total volume of 20 μl/well in an ice-cold, flat-bottom 96-well plate. After 15 min incubation on ice, heparin was added to each well at a final concentration of 0.5 mg/ml, and the reaction was immediately subjected to UV irradiation (total 4800 mJ/cm2) in a UV Stratalinker (Stratagene) on ice. The irradiated sample was mixed with RNase A (20 μg) and RNase T1 (200 U), incubated for 30 min at 37°C, and then subjected to 10% or 4%–12% SDS-PAGE gel electrophoresis. The cross-linked RNA-protein resolved on the gel was captured using a Molecular Dynamic PhosphorImager Storm 860 and analyzed with Image-Quant software. The dried SDS-PAGE gel was then re-hydrated overnight and stained with Coomassie Blue for protein sampling.

Acknowledgments

This study was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. We greatly acknowledge Drs. E.-M. de Villiers and H. zur Hausen for providing HPV 16 plasmid and Dr. W.Y. Tarn for HPV5 E2 expression vector. We thank Alison McBride for inducible HPV16 E2 cell line and Ian Morgan for monoclonal anti-E2 antibody and HPV16 E2 stable U2OS cells for our in vivo attempts. We also thank Drs. Douglas Lowy, Carl Baker, and Alison McBride for critical reading of the manuscript. S.B. was supported by an NCI intramural grant 8340201 (to Z.M.Z).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Androphy EJ, Lowy DR, Schiller JT. Bovine papillomavirus E2 trans-activating gene product binds to specific sites in papillomavirus DNA. Nature. 1987;325:70–73. [PubMed] [Google Scholar]
  • Antson AA, Burns JE, Moroz OV, Scott DJ, Sanders CM, Bronstein IB, Dodson GG, Wilson KS, Maitland NJ. Structure of the intact transactivation domain of the human papillomavirus E2 protein. Nature. 2000;403:805–809. [PubMed] [Google Scholar]
  • Avvakumov N, Torchia J, Mymryk JS. Interaction of the HPV E7 proteins with the pCAF acetyltransferase. Oncogene. 2003;22:3833–3841. [PubMed] [Google Scholar]
  • Bechtold V, Beard P, Raj K. Human papillomavirus type 16 E2 protein has no effect on transcription from episomal viral DNA. J Virol. 2003;77:2021–2028. [PMC free article] [PubMed] [Google Scholar]
  • Bernat A, Avvakumov N, Mymryk JS, Banks L. Interaction between the HPV E7 oncoprotein and the transcriptional coactivator p300. Oncogene. 2003;22:7871–7881. [PubMed] [Google Scholar]
  • Burd CG, Dreyfuss G. Conserved structures and diversity of functions of RNA-binding proteins. Science. 1994;265:615–621. [PubMed] [Google Scholar]
  • Davy CE, Jackson DJ, Wang Q, Raj K, Masterson PJ, Fenner NF, Southern S, Cuthill S, Millar JB, Doorbar J. Identification of a G(2) arrest domain in the E1 wedge E4 protein of human papillomavirus type 16. J Virol. 2002;76:9806–9818. [PMC free article] [PubMed] [Google Scholar]
  • Desaintes C, Hallez S, Van Alphen P, Burny A. Transcriptional activation of several heterologous promoters by the E6 protein of human papillomavirus type 16. J Virol. 1992;66:325–333. [PMC free article] [PubMed] [Google Scholar]
  • Doorbar J, Elston RC, Napthine S, Raj K, Medcalf E, Jackson D, Coleman N, Griffin HM, Masterson P, Stacey S, Mengistu Y, Dunlop J. The E1E4 protein of human papillomavirus type 16 associates with a putative RNA helicase through sequences in its C terminus. J Virol. 2000;74:10081–10095. [PMC free article] [PubMed] [Google Scholar]
  • Doorbar J, Ely S, Sterling J, McLean C, Crawford L. Specific interaction between HPV-16 E1–E4 and cytokeratins results in collapse of the epithelial cell intermediate filament network. Nature. 1991;352:824–827. [PubMed] [Google Scholar]
  • Frattini MG, Laimins LA. Binding of the human papillomavirus E1 origin-recognition protein is regulated through complex formation with the E2 enhancer-binding protein. Proc Natl Acad Sci U S A. 1994;91:12398–12402. [PMC free article] [PubMed] [Google Scholar]
  • Graveley BR. Sorting out the complexity of SR protein functions. RNA. 2000;6:1197–1211. [PMC free article] [PubMed] [Google Scholar]
  • Hawley-Nelson P, Androphy EJ, Lowy DR, Schiller JT. The specific DNA recognition sequence of the bovine papillomavirus E2 protein is an E2-dependent enhancer. EMBO J. 1988;7:525–531. [PMC free article] [PubMed] [Google Scholar]
  • Hegde RS. The papillomavirus E2 proteins: structure, function, and biology. Annu Rev Biophys Biomol Struct. 2002;31:343–360. [PubMed] [Google Scholar]
  • Hegde RS, Androphy EJ. Crystal structure of the E2 DNA-binding domain from human papillomavirus type 16: implications for its DNA binding-site selection mechanism. J Mol Biol. 1998;284:1479–1489. [PubMed] [Google Scholar]
  • Hou SY, Wu SY, Zhou T, Thomas MC, Chiang CM. Alleviation of human papillomavirus E2-mediated transcriptional repression via formation of a TATA binding protein (or TFIID)-TFIIB-RNA polymerase II-TFIIF preinitiation complex. Mol Cell Biol. 2000;20:113–125. [PMC free article] [PubMed] [Google Scholar]
  • Huang SM, McCance DJ. Down regulation of the interleukin-8 promoter by human papillomavirus type 16 E6 and E7 through effects on CREB binding protein/p300 and P/CAF. J Virol. 2002;76:8710–8721. [PMC free article] [PubMed] [Google Scholar]
  • Hughes FJ, Romanos MA. E1 protein of human papillomavirus is a DNA helicase/ATPase. Nucleic Acids Res. 1993;21:5817–5823. [PMC free article] [PubMed] [Google Scholar]
  • Imai Y, Tsunokawa Y, Sugimura T, Terada M. Purification and DNA-binding properties of human papillomavirus type 16 E6 protein expressed in Escherichia coli. Biochem Biophys Res Commun. 1989;164:1402–1410. [PubMed] [Google Scholar]
  • Kanda T, Watanabe S, Zanma S, Sato H, Furuno A, Yoshiike K. Human papillomavirus type 16 E6 proteins with glycine substitution for cysteine in the metal-binding motif. Virology. 1991;185:536–543. [PubMed] [Google Scholar]
  • Kanopka A, Muhlemann O, Akusjarvi G. Inhibition by SR proteins of splicing of a regulated adenovirus pre-mRNA. Nature. 1996;381:535–538. [PubMed] [Google Scholar]
  • Kiyono T, Hiraiwa A, Fujita M, Hayashi Y, Akiyama T, Ishibashi M. Binding of high-risk human papillomavirus E6 oncoproteins to the human homologue of the Drosophila discs large tumor suppressor protein. Proc Natl Acad Sci U S A. 1997;94:11612–11616. [PMC free article] [PubMed] [Google Scholar]
  • Klingelhutz AJ, Foster SA, McDougall JK. Telomerase activation by the E6 gene product of human papillomavirus type 16. Nature. 1996;380:79–82. [PubMed] [Google Scholar]
  • Kumar A, Zhao Y, Meng G, Zeng M, Srinivasan S, Delmolino LM, Gao Q, Dimri G, Weber GF, Wazer DE, Band H, Band V. Human papillomavirus oncoprotein E6 inactivates the transcriptional coactivator human ADA3. Mol Cell Biol. 2002;22:5801–5812. [PMC free article] [PubMed] [Google Scholar]
  • LaCasse EC, Lefebvre YA. Nuclear localization signals overlap DNA- or RNA-binding domains in nucleic acid-binding proteins. Nucleic Acids Res. 1995;23:1647–1656. [PMC free article] [PubMed] [Google Scholar]
  • Ladd AN, Cooper TA. Finding signals that regulate alternative splicing in the post-genomic era. Genome Biol. 2002;3 reviews0008. [PMC free article] [PubMed] [Google Scholar]
  • Lai MC, Teh BH, Tarn WY. A human papillomavirus E2 transcriptional activator. The interactions with cellular splicing factors and potential function in pre-mRNA processing. J Biol Chem. 1999;274:11832–11841. [PubMed] [Google Scholar]
  • Lee SS, Weiss RS, Javier RT. Binding of human virus oncoproteins to hDlg/SAP97, a mammalian homolog of the Drosophila discs large tumor suppressor protein. Proc Natl Acad Sci U S A. 1997;94:6670–6675. [PMC free article] [PubMed] [Google Scholar]
  • Lowy DR, Howley PM. Papillomaviruses. In: Knipe DM, Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B, Straus SE, editors. Fields Virology. Lippincott Williams & Wilkins: Philadelphia; 2001. pp. 2231–2264. [Google Scholar]
  • Maniatis T, Reed R. An extensive network of coupling among gene expression machines. Nature. 2002;416:499–506. [PubMed] [Google Scholar]
  • Massimi P, Pim D, Banks L. Human papillomavirus type 16 E7 binds to the conserved carboxy-terminal region of the TATA box binding protein and this contributes to E7 transforming activity. J Gen Virol. 1997;78:2607–2613. [PubMed] [Google Scholar]
  • Nakahara T, Nishimura A, Tanaka M, Ueno T, Ishimoto A, Sakai H. Modulation of the cell division cycle by human papillomavirus type 18 E4. J Virol. 2002;76:10914–10920. [PMC free article] [PubMed] [Google Scholar]
  • Nomine Y, Charbonnier S, Ristriani T, Stier G, Masson M, Cavusoglu N, Van Dorsselaer A, Weiss E, Kieffer B, Trave G. Domain substructure of HPV E6 oncoprotein: biophysical characterization of the E6 C-terminal DNA-binding domain. Biochemistry. 2003;42:4909–4917. [PubMed] [Google Scholar]
  • Patel D, Huang SM, Baglia LA, McCance DJ. The E6 protein of human papillomavirus type 16 binds to and inhibits co- activation by CBP and p300. EMBO J. 1999;18:5061–5072. [PMC free article] [PubMed] [Google Scholar]
  • Ristriani T, Masson M, Nomine Y, Laurent C, Lefevre JF, Weiss E, Trave G. HPV oncoprotein E6 is a structure-dependent DNA-binding protein that recognizes four-way junctions. J Mol Biol. 2000;296:1189–1203. [PubMed] [Google Scholar]
  • Ristriani T, Nomine Y, Masson M, Weiss E, Trave G. Specific recognition of four-way DNA junctions by the C-terminal zinc-binding domain of HPV oncoprotein E6. J Mol Biol. 2001;305:729–739. [PubMed] [Google Scholar]
  • Romanczuk H, Thierry F, Howley PM. Mutational analysis of cis elements involved in E2 modulation of human papillomavirus type 16 P97 and type 18 P105 promoters. J Virol. 1990;64:2849–2859. [PMC free article] [PubMed] [Google Scholar]
  • Ronco LV, Karpova AY, Vidal M, Howley PM. Human papillomavirus 16 E6 oncoprotein binds to interferon regulatory factor-3 and inhibits its transcriptional activity. Genes Dev. 1998;12:2061–2072. [PMC free article] [PubMed] [Google Scholar]
  • Sakai H, Yasugi T, Benson JD, Dowhanick JJ, Howley PM. Targeted mutagenesis of the human papillomavirus type 16 E2 transactivation domain reveals separable transcriptional activation and DNA replication functions. J Virol. 1996;70:1602–1611. [PMC free article] [PubMed] [Google Scholar]
  • Sousa R, Dostatni N, Yaniv M. Control of papillomavirus gene expression. Biochim Biophys Acta. 1990;1032:19–37. [PubMed] [Google Scholar]
  • Tao M, Kruhlak M, Xia S, Androphy E, Zheng ZM. Signals That Dictate Nuclear Localization of Human Papillomavirus Type 16 Oncoprotein E6 in Living Cells. J Virol. 2003;77:13232–13247. [PMC free article] [PubMed] [Google Scholar]
  • Thain A, Webster K, Emery D, Clarke AR, Gaston K. DNA binding and bending by the human papillomavirus type 16 E2 protein. Recognition of an extended binding site. J Biol Chem. 1997;272:8236–8242. [PubMed] [Google Scholar]
  • Thomas MC, Chiang CM. E6 oncoprotein represses p53-dependent gene activation via inhibition of protein acetylation independently of inducing p53 degradation. Mol Cell. 2005;17:251–264. [PubMed] [Google Scholar]
  • Veldman T, Horikawa I, Barrett JC, Schlegel R. Transcriptional activation of the telomerase hTERT gene by human papillomavirus type 16 E6 oncoprotein. J Virol. 2001;75:4467–4472. [PMC free article] [PubMed] [Google Scholar]
  • Veldman T, Liu X, Yuan H, Schlegel R. Human papillomavirus E6 and Myc proteins associate in vivo and bind to and cooperatively activate the telomerase reverse transcriptase promoter. Proc Natl Acad Sci U S A. 2003;100:8211–8216. [PMC free article] [PubMed] [Google Scholar]
  • Zahler AM, Neugebauer KM, Stolk JA, Roth MB. Human SR proteins and isolation of a cDNA encoding SRp75. Mol Cell Biol. 1993;13:4023–4028. [PMC free article] [PubMed] [Google Scholar]
  • Zheng ZM. Regulation of alternative RNA splicing by exon definition and exon sequences in viral and Mammalian gene expression. J Biomed Sci. 2004;11:278–294. [PMC free article] [PubMed] [Google Scholar]
  • Zheng ZM, Baker CC. Parameters that affect in vitro splicing of bovine papillomavirus type 1 late pre-mRNAs. J Virol Methods. 2000;85:203–214. [PubMed] [Google Scholar]
  • Zheng ZM, He P, Baker CC. Selection of the bovine papillomavirus type 1 nucleotide 3225 3′ splice site is regulated through an exonic splicing enhancer and its juxtaposed exonic splicing suppressor. J Virol. 1996;70:4691–4699. [PMC free article] [PubMed] [Google Scholar]
  • Zheng ZM, He PJ, Baker CC. Structural, functional, and protein binding analyses of bovine papillomavirus type 1 exonic splicing enhancers. J Virol. 1997;71:9096–9107. [PMC free article] [PubMed] [Google Scholar]
  • Zheng ZM, He PJ, Baker CC. Function of a bovine papillomavirus type 1 exonic splicing suppressor requires a suboptimal upstream 3′ splice site. J Virol. 1999;73:29–36. [PMC free article] [PubMed] [Google Scholar]
  • Zheng ZM, Quintero J, Reid ES, Gocke C, Baker CC. Optimization of a weak 3′ splice site counteracts the function of a bovine papillomavirus type 1 exonic splicing suppressor in vitro and in vivo. J Virol. 2000;74:5902–5910. [PMC free article] [PubMed] [Google Scholar]
  • Zheng ZM, Tao M, Yamanegi K, Bodaghi S, Xiao W. Splicing of a Cap-proximal Human Papillomavirus 16 E6E7 Intron Promotes E7 Expression, but can be Restrained by Distance of the Intron from its RNA 5′ Cap. J Mol Biol. 2004;337:1091–1108. [PubMed] [Google Scholar]
  • Zhou Z, Licklider LJ, Gygi SP, Reed R. Comprehensive proteomic analysis of the human spliceosome. Nature. 2002;419:182–185. [PubMed] [Google Scholar]
  • zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical application. Nat Rev Cancer. 2002;2:342–350. [PubMed] [Google Scholar]
-