INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

There are multiple posttranscriptional processes that are essential for expression of eukaryotic cellular and viral genes. These include RNA capping (5), splicing (18), polyadenylation (7), and transport (12, 16). Pre-mRNA splicing is often complex in mammalian cells, since a majority of pre-mRNAs have multiple introns and these sometimes contain more than one 5' and/or 3' splice site. Alternative splicing of these pre-mRNAs involves the use of alternative 5' or 3' splice sites and exon skipping or inclusion, generating different pre-mRNAs potentially encoding multiple protein isoforms with distinct functions. Thus, the correct selection of splice sites has been a major focus of splicing research.

A number of exonic and intronic cis elements that affect splicing efficiency and splice site choice have been identified. One of these elements is the purine-rich exonic splicing enhancer (ESE). Through interactions with serine/arginine-rich (SR) proteins, an ESE recruits U2AF to suboptimal 3' splice sites and stimulates spliceosome assembly (14, 17, 23, 24, 28, 29). More recently exonic splicing suppressors or silencers (ESSs) have been identified in several eukaryotic cellular and viral genes (1, 2, 6, 8, 9, 19-21, 25, 28). These cis elements negatively regulate utilization of upstream 3' splice sites and are frequently located downstream of a juxtaposed ESE. However, in the human immunodeficiency virus type 1 (HIV-1) 6D exon, an ESS is located upstream of an ESE (25), and in the fibroblast growth factor receptor 2 K-SAM exon, an ESS is present without an adjacent ESE (8, 9). Unlike purine-rich ESEs, the sequences of the ESSs show little similarity. The mechanisms by which ESSs suppress pre-mRNA splicing remain largely unknown.

The expression of the late genes of bovine papillomavirus type 1 (BPV-1) is regulated in part through alternative splicing (3, 29). Splicing of the majority of mRNAs at early stages of the viral life cycle utilizes a common 3' splice site at nucleotide (nt) 3225 even though alternative 3' splice sites are present both upstream and downstream of this site. In contrast, use of an alternative 3' splice site at nt 3605 is required for the major capsid protein (L1) mRNA (3). This 3' splice site is used only at late stages of the viral life cycle in fully differentiated keratinocytes (4). We have recently demonstrated that the BPV-1 nt 3225 3' splice site is a weak 3' splice site with a nonconsensus branch point and a suboptimal polypyrimidine tract (28). In addition, we identified three exonic splicing elements (ESEs) between the two alternative 3' splice sites at nt 3225 and nt 3605 (Fig. 1) (28). Two of these cis elements are purine-rich ESEs, and they were named SE1 and SE2. SE1 enhances utilization of the nt 3225 3' splice site in vitro and in vivo, and this effect is mediated through binding of SR proteins (29). The third cis element is a pyrimidine-rich sequence (nt 3306 to 3353) immediately downstream of SE1, and it was named ESS because it functions as a splicing suppressor in vitro (28). The BPV-1 ESS is 48 nt long and can be divided into three regions based on sequence composition. The 5' U-rich region (nt 3306 to 3317) is 67% uridines, the central C-rich region (nt 3319 to 3345) is 56% cytosines, and the 3' AG-rich region (nt 3346 to 3353) consists of the sequence AGAGCAGG. The BPV-1 ESS was not further characterized in that study.

View larger version (9K): [in this window] [in a new window]   FIG. 1.   Position and sequence of the BPV-1 ESS in late pre-mRNAs. Only that portion of the pre-mRNA containing the nt 3225 and 3605 3' splice sites (SS) and the nt 3764 5' splice site is shown. Diagonal dotted lines indicate splicing.

In this study we have utilized both homologous and heterologous pre-mRNAs in vitro to further characterize the structure and function of the BPV-1 ESS. We have found that the BPV-1 ESS requires an upstream suboptimal 3' splice site for splicing suppression. Furthermore, the ESS can function when located either upstream or downstream of SE1. However, an adjacent ESE is not essential for its function. Finally, mutational analysis of the ESS demonstrated that only the central C-rich region of the ESS is essential for suppression of splicing in all pre-mRNAs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

BPV-1 DNA templates. BPV-1 DNA templates were generated from plasmid pZMZ19-1 (28) by PCR with a 5' T7 primer (oZMZ79; 5'-ATTAATACGACTCACTATAG-3') combined with an antisense 3' primer with a 5' end at nt 3245 (oMD6; 5'-CAGACTCCGTCTGGGCGATC-3'), nt 3305 (oZMZ84; 5'-GGCTGGGCTGGCTCGGCTTCTTTT-3'), nt 3345 (oZMZ76; 5'-GATGGGACCGCAGGCGGGGGAGCCGAG-3'), or nt 3353 (oZMZ102; 5'-CCTGCTCTGATGGGACCGCAGGC-3') in the BPV-1 genome. Alternatively, antisense chimeric BPV-1-Rous sarcoma virus (RSV) src (oZMZ126 [5'-GCCTGGCCA CAGTGGTACGCGAGGCCACCAGCAGAGTCAGCTTAGCTC/CAGACT CCGTCTGGGCGATC-3'] or oZMZ156 [5'-GGCGCTGGCTGGGGTCCTTA GGCTTGCTCTTGCTGCTCCCCATGG/GGATGCGACCCAGACTCCGTC TGGGCGATC-3']) or BPV-1-(Py3)₂ (oZMZ127; 5'-TAGCTTCTAGTCTTAGCTTCTAGTTAGCTTCTAGTCTTAGCTTCTAGT/CAGACTCCGTCTGGG CGATC-3') 3' primers were used for template preparation. In the preceding sequences, the "/" indicates the junction between two sequences in the chimera. Each pre-mRNA prepared by transcription of PCR products contains a fixed-size exon 1 (187 nt), a truncated intron 1 (333 nt), and a variable-size exon 2, ranging from 21 to 129 nt depending upon which antisense primer was used to generate the PCR product.

Chimeric HIV-1 tat-rev DNA templates. Plasmid pHS2 (2), which contains a purine-rich ESE upstream of an ESS within tat-rev exon 3 of HIV-1, was used for introduction of the BPV-1 ESS by PCR. The plasmid pHS2 was first digested with HpaI and PvuII, and a restriction fragment of about 1.1 kb was gel purified and used as a template for PCR with a 5' sense T3 primer (oFD122; 5'-ATTAACCCTCACTAAAG-3') combined with a 3' antisense primer for HIV ESE (oZMZ94; 5'-TGTCTCTGTCTCTCTCTCC-3'), HIV tat-rev ESS (oZMZ95; 5'-CGTTCACTAATCGAATGGATCT-3'), or chimeric HIV ESE-BPV-1 full-length ESS (oZMZ93; 5'-CCTGCTCTGATGGGACCGCAGGCGGGGGAGCCGAGCAAAGAAGAGACA/TGTCTCTGTCTCTCTCTCCACCTTCTTCTT C-3'). The chimeric HIV-BPV-1 ESS PCR product was then reamplified by PCR with the same 5' T3 primer combined with the 3' primer oZMZ76 to prepare a truncated BPV-1 ESS lacking the sequence AGAGCAGG at the 3' end. T3 RNA polymerase was used for in vitro transcription of HIV-1 tat-rev pre-mRNA from these templates.

Chimeric RSV src DNA templates. Plasmids pRSV-7169 and pRSV-7169SRE⁺ have been described in a study of the HIV ESS (1). Plasmid pRSV-7169SRE⁺ contains an HIV-1 tat exon 2 ESS (HIV-1 nt 5821 to 5860) replacing RSV nt 7098 to 7127 in the RSV src exon. Plasmids pRSV-7169 and pRSV-7169SRE⁺ were both linearized at the NaeI site at nt 7171 for in vitro transcription by SP6 RNA polymerase. To replace the HIV-1 ESS with the BPV-1 ESS, the plasmid pRSV-7169 was first amplified by PCR with a 5' SP6 primer (oJR3; 5'-ATTTAGGTGACACTATAG-3') combined with a 3' antisense RSV-BPV-1 wt ESS primer (oZMZ85; 5'-CGCCATGG/CCTGCTCTGATGGGACCGCAGGCGGGGGA GCCGAGCAAAGAAGAGACA/TGGCCACAGTGGTACGCGAG-3'). The PCR product containing the BPV-1 wt ESS was then reamplified by PCR with the same 5' SP6 primer combined with either a 3' antisense primer oZMZ104 (5'-GGCGCTGGCTGGGGTCCTTAGGCTTGCTCTTGCTGCTCCCCATG G/CCTGCTCTGATGGGA-3') or oZMZ131 (5'-GGCGCTGGCTGGGGTCC TTAGGCTTGCTCTTGCTGCTCCCCATGG/GATGGGACCGCAGGCGG GGGAG-3'). The resulting PCR templates contain either the wt BPV-1 ESS or a truncated ESS (lacking AGAGCAGG at the 3' end) followed by the rest of the RSV src exon 2. SP6 RNA polymerase was used to generate chimeric RSV src pre-mRNAs from these templates.

Chimeric human -globin DNA templates. The BPV-1 ESS was cloned at a distance of 209 nt downstream of the human -globin 3' splice site between the BamHI and EcoRI sites of pSP64-H6, obtained from Promega (13). The resulting plasmid (p3063) was linearized with EcoRI to generate a template with the BPV-1 wt ESS. The parent plasmid (pSP64-H6) was linearized with BamHI to generate a template without the BPV-1 ESS. In addition, a 5' SP6 primer (oJR3) combined with either a 3' antisense -globin primer (oZMZ106; 5'-GGGTTGCCCATAACAGCATCAGG-3') or a chimeric 3' antisense -globin-BPV-1 ESS primer (oZMZ103; 5'-CCTGCTCTGATGGGACCGCAGGCGGG GGAGCCGAGCAAAGAAGAGACA/GGGTTGCCCATAACAGCATCAG G-3') and pSP64-H6 DNA were used to generate templates containing only 84 nt of the -globin exon 2 with or without the BPV-1 ESS. Each template was transcribed in vitro with SP6 RNA polymerase.

Chimeric dsx DNA templates. The chimeric dsx plasmid p3013 (29) was used for insertion of synthetic oligonucleotides containing BPV-1 wt or mutant ESS sequences between the HindIII and XhoI sites downstream of BPV-1 SE1. Templates for in vitro transcription were prepared by linearization of the plasmids with XhoI and then transcribed in vitro with T7 RNA polymerase.

In vitro splicing and image analysis. Pre-mRNAs were spliced in vitro with HeLa nuclear extracts (Promega) and analyzed as described previously (28, 29). The splicing efficiency for each pre-mRNA was calculated as the percentage of the total splicing products (intermediate and fully spliced) divided by the sum of the total splicing products plus the remaining pre-mRNA (29). The figures were prepared with Adobe Photoshop and Micrografx Designer.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

BPV-1 ESS inhibits pre-mRNA splicing when located either upstream or downstream of SE1. We have previously reported that a BPV-1 late pre-mRNA containing the nt 3225 3' splice site and only 20 nt of exon 2 is spliced very inefficiently in vitro (28), suggesting that this splice site is suboptimal. However, it has been reported that a human -globin transcript with less than 50 nt of exon 2 is not spliced well in vitro and that the efficiency of its splicing depends on the length of exon 2 and not on a specific sequence (11). To determine if the apparently suboptimal nature of the nt 3225 3' splice site was an artifact of exon size, 48 nt of unrelated sequence was added 20 nt downstream of the 3' splice site. In in vitro splicing reactions, BPV-1 pre-mRNAs containing either a (Py3)₂ sequence or the first 48 nt of the RSV src exon 2 were spliced as inefficiently as a pre-mRNA containing only 20 nt of exon 2 (Fig. 2, compare pre-mRNAs 2 and 3 with pre-mRNA 1). In contrast, the addition of SE1 strongly stimulated splicing (Fig. 2, pre-mRNA 4), consistent with our previous report (28). These data indicate that the nt 3225 3' splice site is indeed suboptimal and requires an exonic splicing enhancer for efficient utilization. This observation is consistent with our earlier observation that the nt 3225 3' splice site has a nonconsensus branch point sequence (28).

View larger version (44K): [in this window] [in a new window]   FIG. 2.   BPV-1 ESS functions upstream of SE1. (A) Structures of BPV-1 late pre-mRNAs used for splicing in vitro. The pre-mRNAs were transcribed in vitro from templates prepared by PCR following the strategy described in Materials and Methods. The following pre-mRNA-specific 3' primers were used: oMD6 (pre-mRNA 1); oZMZ127 (pre-mRNA 2); oZMZ126 (pre-mRNA 3); oZMZ84 (pre-mRNA 4); oZMZ102 (pre-mRNA 5); oZMZ84 and oZMZ148 (pre-mRNA 6); oZMZ84, oZMZ148, and oZMZ149 (pre-mRNA 7); oZMZ102 (pre-mRNA 8), and oZMZ130 (pre-mRNA 9). The sequences of the primers are given in Materials and Methods. The nucleotide positions of the ends of each element are shown above each pre-mRNA, while the size of each element is shown below the pre-mRNA. Pre-mRNAs 2 and 3 each have an exon 2 with either two tandem copies of the Py3 sequence (24) or the src sequence (nt 7054 to 7101) downstream of BPV-1 nt 3245. The box labeled ESS with a leftward-facing arrow indicates a reverse antisense ESS, in which each base of the ESS from nt 3316 to 3353 was replaced by its Watson-Crick complement. (B and C) Electrophoretic analysis of the spliced products of BPV-1 late pre-mRNAs on 8% polyacrylamide-8 M urea gels. The corresponding spliced products are diagrammed between the two gels, and 100-bp DNA ladders are shown on the left (B) or right (C) of each gel. The numbers at the tops of the gels correspond to the pre-mRNAs shown in panel A.

Function of BPV-1 ESS is sequence dependent and does not require a specific ESE. We have previously utilized a Drosophila melanogaster dsx exon 3-exon 4 pre-mRNA to characterize BPV-1 SE1 and SE2 structure and function, since this pre-mRNA has a suboptimal 3' splice site and therefore does not splice well without an ESE in exon 4 (29). In order to determine if a specific ESE is required for the function of the BPV-1 ESS, the ESS was connected downstream of either BPV-1 SE1 or SE2 in dsx pre-mRNAs and the chimeric pre-mRNAs were then examined for their splicing efficiency in vitro. The BPV-1 ESS gave a five- to sixfold suppression of the splicing of the dsx pre-mRNAs stimulated by either SE1 or SE2, indicating that a specific ESE is not a prerequisite for the ESS's function as a splicing suppressor (Fig. 3, compare pre-mRNAs 2 and 6 with pre-mRNAs 1 and 5).

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Function of the BPV-1 ESS is sequence specific. (A) Structures of Drosophila dsx pre-mRNAs used for splicing. The pre-mRNAs were transcribed in vitro from templates prepared by PCR following the strategy described in Materials and Methods. The following pre-mRNA-specific 3' primers were used: oZMZ84 (pre-mRNA 1), oZMZ102 (pre-mRNA 2), oZMZ132 (pre-mRNA 3), oZMZ140 (pre-mRNA 4), and oZMZ133 (pre-mRNA 7). The sequences of the primers are given in Materials and Methods. Pre-mRNAs 5 and 6 were transcribed from the plasmid p3062 linearized with HindIII or XhoI, respectively. The box labeled ESS+ refers to a sense orientation ESS. The box labeled ESS, with a leftward-facing arrow, refers to a reverse antisense ESS (in which each base of the ESS was replaced by its Watson-Crick complement), while ESS+ with a leftward-facing arrow refers to a reverse sense strand ESS. The sizes of the introns and exon segments are shown below the maps. (B and C) Electrophoretic analysis of spliced dsx pre-mRNA products on 5% polyacrylamide-8 M urea gels. The corresponding spliced products are diagrammed between the gels, and 100-bp DNA ladders are shown on the left (B) and right (C). The numbers on the top of each gel correspond to the pre-mRNAs in panel A used for splicing. The splicing efficiency for each pre-mRNA was calculated from the gels (see Materials and Methods) and is indicated at the bottom of each gel.

BPV-1 ESS inhibits splicing of heterologous pre-mRNAs containing a suboptimal 3' splice site but not a constitutive splice site. It has been reported that an HIV-1 ESS can function as a splicing suppressor in heterologous substrates (1). Suppression of dsx pre-mRNA splicing by the BPV-1 ESS indicates that it too can work in heterologous pre-mRNAs. To further examine the context in which the BPV-1 ESS can suppress splicing, the ESS was connected downstream of either a suboptimal (weak) or constitutive (strong) 3' splice site in three additional pre-mRNAs: HIV-1 tat-rev exon 3 (2), RSV src (1, 27), and human -globin (13). The first two pre-mRNAs feature a suboptimal 3' splice site, as do the BPV-1 late pre-mRNA and dsx pre-mRNA, whereas the human -globin pre-mRNA has a constitutive 3' splice site. In addition, the HIV tat-rev pre-mRNA contains its own ESE downstream of the 3' splice site (2, 21) whereas the RSV src and human -globin pre-mRNAs have no known ESE in exon 2 and appear not to require an ESE for in vitro splicing. These experiments were designed to address the following questions. First, is the BPV-1 ESS functionally equivalent to other ESSs? Second, does ESS function require a suboptimal 3' splice site? And third, does ESS function require an adjacent upstream ESE?

View larger version (23K): [in this window] [in a new window]   FIG. 4.   Inhibition of HIV-1 tat-rev pre-mRNA splicing by BPV-1 ESS. (A) Structures of HIV-1 tat-rev pre-mRNAs used for splicing in vitro. Every pre-mRNA has a native HIV tat-rev ESE in exon 2. HIV tat-rev exon 3 ESS (pre-mRNA 2) and BPV-1 wt ESS (pre-mRNA 3) and a truncated BPV-1 ESS (pre-mRNA 4) were connected downstream of the HIV-1 tat-rev ESE (see Materials and Methods). The splicing efficiency was calculated as described in Materials and Methods. (B) Electrophoretic analysis of spliced HIV-1 tat-rev pre-mRNA products on an 8% polyacrylamide-8 M urea gel. Each number on the top of the gel indicates the pre-mRNA in panel A used for in vitro splicing. The corresponding spliced products are diagrammed on the left, and a 100-bp DNA ladder is shown on the right of the gel.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   Inhibition of RSV src pre-mRNA splicing by the BPV-1 and HIV ESSs. (A) Structures of RSV src pre-mRNAs used for splicing. The src pre-mRNA (pre-mRNA 1) without an ESS in exon 2 was included as a control. BPV-1 wt ESS (pre-mRNA 2) and truncated ESS (pre-mRNA 3) were inserted by PCR between nt 7098 and 7127 in RSV src exon 2 (see Materials and Methods). The HIV-1 tat exon 2 ESS (nt 5821 to 5860 in the HIV-1 viral genome) in pre-mRNA 4 was cloned at the same location (1) and used as a control for ESS function. The template DNAs were transcribed in vitro with SP6 RNA polymerase. (B and C) Electrophoretic analysis of spliced RSV src pre-mRNA products on 8% polyacrylamide-8 M urea gels. The corresponding spliced products are diagrammed between the gels, and a 100-bp DNA ladder is shown on the left of panel B. The numbers on the top of each gel correspond to the pre-mRNAs in panel A used for splicing. The splicing efficiency for each pre-mRNA was calculated as described in Materials and Methods and is indicated at the bottom of each gel. (D and E) Analysis of possible splicing enhancer activity of the 3' sequence in src exon 2 in a BPV-1 late pre-mRNA. (D) Structures of BPV-1 late pre-mRNAs. Pre-mRNAs 1 and 2 are identical to pre-mRNAs 4 and 6 in Fig. 2 and were included as a positive and negative control, respectively. Pre-mRNA 3 contains the src sequence from nt 7127 to 7171, which was connected to BPV-1 exon 2 by PCR with 5' T7 primer oZMZ79 combined with 3' antisense primer oZMZ156, as described in Materials and Methods. (E) Splicing gel showing the corresponding splicing products on the right. The products from each splicing reaction were analyzed by electrophoresis on an 8% polyacrylamide-8 M urea gel. The numbers at the top of the gel indicate the pre-mRNAs in panel D used for splicing. The splicing efficiency for each pre-mRNA was calculated from the gel as described in Materials and Methods and is indicated at the bottom of the gel.

View larger version (26K): [in this window] [in a new window]   FIG. 6.   BPV-1 ESS does not suppress splicing of a human -globin pre-mRNA. (A) Structures of human -globin pre-mRNAs used for in vitro splicing. The BPV-1 wt ESS was connected downstream of the constitutive 3' splice site at a distance of 84 or 209 nt (see Materials and Methods). (B and C) Electrophoretic analysis of spliced human -globin pre-mRNA products on 8% polyacrylamide-8 M urea gels. The corresponding spliced products are diagrammed on the right, and a 100-bp DNA ladder is shown on the left of each gel. The numbers on the top of each gel correspond to the pre-mRNAs in panel A used for splicing. The splicing efficiency for each pre-mRNA was calculated from the gel as described in Materials and Methods and is indicated at the bottom of each gel.

Functional analysis of BPV-1 ESS mutants with heterologous pre-mRNAs. To determine what structural features of the BPV-1 ESS are required for splicing suppression, several deletion and point mutations were made in the ESS and the resulting mutant ESSs were assayed for their ability to inhibit splicing of the heterologous pre-mRNAs presented above. Deletion of the 8-nt sequence AGAGCAGG from the 3' end of the ESS resulted in more than a twofold decrease in splicing suppression in vitro in the context of the BPV-1 late pre-mRNA (Fig. 7, compare pre-mRNAs 2 and 3). The same deletion in the context of the HIV tat-rev pre-mRNA had a similar effect (Fig. 4, pre-mRNAs 3 and 4). These results suggest that the 3' AG-rich region is essential for full suppression of splicing. We were unable to make a similar comparison in the context of the RSV src pre-mRNA because of the overall low levels of splicing (Fig. 5A, B, and C, pre-mRNAs 2 and 3). In contrast, deletion of the 3' AG-rich region of the ESS in the context of a dsx pre-mRNA containing SE1 had no deleterious effect on ESS function (Fig. 8, pre-mRNAs 2 and 3). Similar results were obtained whether the dsx pre-mRNAs were transcribed from restriction endonuclease-digested plasmids retaining a 5-nt polylinker sequence at the 3' end or from PCR-generated templates lacking these extra polylinker sequences (data not shown). The reason for this discrepancy is unknown, but it may be related to differences in splice site and/or ESE strengths in the different pre-mRNAs.

View larger version (20K): [in this window] [in a new window]   FIG. 7.   Functions of the BPV-1 wt ESS and a truncated ESS lacking the 3' AG-rich region. (A) Structures of BPV-1 late pre-mRNAs used for in vitro splicing (B) All of the pre-mRNAs have BPV-1 SE1 in exon 2, but pre-mRNAs 2 and 3 contain a truncated ESS (40 nt, with deletion of the AGAGCAGG sequence at the 3' end) and wt ESS (48 nt, full length), respectively. The pre-mRNAs were transcribed with T7 RNA polymerase from pZMZ19-1 DNA templates amplified by PCR with the 5' T7 primer oZMZ79 combined with the antisense 3' primer oZMZ84 (pre-mRNA 1), oZMZ76 (pre-mRNA 2), or oZMZ102 (pre-mRNA 3). The splicing efficiency was calculated as described in Materials and Methods. (B) Splicing gel, showing the corresponding splicing products on the right. The products from each splicing reaction were analyzed by electrophoresis on an 8% polyacrylamide-8 M urea gel. The numbers at the top of the gel indicate the pre-mRNAs in panel A used for splicing.

View larger version (32K): [in this window] [in a new window]   FIG. 8.   Functional analysis of BPV-1 ESS mutants in a Drosophila dsx pre-mRNA. (A) Sequences of the 3' ends of Drosophila dsx pre-mRNAs used for splicing. BPV-1 ESS with or without U-to-A and/or C-to-A mutations was cloned downstream of BPV-1 SE1 between HindIII and XhoI sites in plasmid p3013 (29), yielding the following chimeric dsx plasmids: 1, p3013; 2, p3057; 3, p3058; 4, p3059; 5, p3060; and 6, p3061. Capital letters indicate wt ESS sequence, and lowercase letters are polylinker sequence; underlined capital letters represent mutated nucleotides; and dots indicate the sequences upstream of the ESS. The splicing efficiency for each pre-mRNA was calculated from the gel shown in panel C (see Materials and Methods) and is shown on the right. (B) Schematic diagram of dsx pre-mRNAs containing BPV-1 SE1 and the ESS. The numbers below the map indicate sizes in nucleotides. (C) Electrophoretic analysis of spliced dsx pre-mRNA products on a 5% polyacrylamide-8 M urea gel. The corresponding spliced products are diagrammed on the right, and a 100-bp DNA ladder is shown on the left. The numbers at the top of the gel correspond to the pre-mRNAs in panel A used for splicing.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we extensively assayed the BPV-1 ESSs in five different pre-mRNAs. Four (BPV-1 late pre-mRNA, HIV-1 tat-rev pre-mRNA, RSV src pre-mRNA, and Drosophila dsx pre-mRNA) have suboptimal (weak) 3' splice sites, and one (human -globin pre-mRNA) has a constitutive (strong) 3' splice site. The BPV-1 ESS suppressed splicing of only those pre-mRNAs that contained a weak 3' splice site (summarized in Table 1). Although three of the pre-mRNAs contain ESEs, we were unable to identify a functional ESE in the RSV src exon 2. The splicing of this pre-mRNA was still suppressed by both the BPV-1 ESS and the HIV-1 tat ESS (1), suggesting that an upstream ESE is not required for ESS function. Therefore, it is likely that the BPV-1 ESS directly inhibits the function of a suboptimal upstream 3' splice site rather than inhibiting the function of an upstream or downstream ESE. This conclusion is consistent with reports that the HIV-1 tat and tat-rev ESSs are able to act independently on the 3' splice sites in both RSV src pre-mRNA (1) and human fibronectin pre-mRNA (21).

Seven ESSs (three cellular and four viral) have been identified in eukaryotic pre-mRNAs (Table 2). Each ESS is located downstream of a suboptimal 3' splice site in a regulated exon of its pre-mRNA and plays a negative role in the splicing of the upstream intron. These ESSs are frequently positioned adjacent to an ESE and form a bipartite splicing regulatory element in that exon. Although initial examples of bipartite splicing regulatory elements contained the ESS downstream of the ESE, this is not always the case. In the human fibronectin EDA exon, deletion or mutation of an ESS sequence upstream of a non-purine-rich ESE results in a 20-fold increase in the amount of spliced RNA (20). In the HIV-1 6D exon, a single-point mutation (U to C) within the ESS element upstream of an ESE activates exon 6D inclusion (25). Although the BPV-1 ESS is normally located downstream of an ESE (SE1), we have demonstrated that the ESS still functions when placed upstream of SE1 (Fig. 2). This provides further evidence that an ESS can function either downstream or upstream of an ESE. Although each ESS has its own functional motif that is required for splicing suppression, no obvious consensus sequences have been found among these ESSs. Thus, splicing of a suboptimal 3' splice site in different pre-mRNAs may be regulated by different sets of ESEs and ESSs with very different mechanisms.

Since comparison of the sequences of the different ESS elements failed to give an obvious consensus sequence that would have given us insight into the mechanism of ESS function, we carried out a mutational analysis of the ESS to determine what ESS sequences are required for its function. The reverse antisense strand of the ESS was unable to suppress splicing, indicating that the function of the ESS is sequence dependent. Extensive mutational analysis of the BPV-1 ESS by using the Drosophila dsx pre-mRNA as an assay system demonstrated that the central C-rich region of the ESS is critical for efficient suppression of splicing (Fig. 8). This is consistent with the observation that reversing the order of nucleotides in the sense strand of the ESS does not destroy function (Fig. 3). These data suggest that the sequence CCCCC may be an important determinant of ESS function (Table 2). In contrast, the 5' U-rich and 3' AG-rich parts contribute very little to splicing suppression in the dsx system. However, deletion of the 3' AG-rich end (AGAGCAGG) gave contradictory results in other pre-mRNAs. The ESS containing the 3' AG-rich region always inhibited splicing of BPV-1 late pre-mRNA and HIV-1 tat-rev pre-mRNA more strongly than did a truncated ESS lacking this region, suggesting that the AG-rich region plays a functional role in splicing suppression. However, mutation of all Gs to Us in the AG-rich region of the ESS in the BPV-1 late pre-mRNA did not significantly affect splicing suppression, suggesting that the requirement for the AG-rich region in some pre-mRNAs may be a nonspecific size effect (data not shown). This may be an in vitro artifact, since the ESS is normally located in the middle of an exon in BPV-1 pre-mRNAs in vivo. Alternatively, the different effects of the AG-rich region in different pre-mRNAs may be due to differences in the relative strengths of the 3' splice sites and/or ESEs in each pre-mRNA.

Like ESEs, the BPV-1 ESS has also been shown to bind SR proteins (30). The ESS was unable to function as an ESE in the BPV-1 late pre-mRNA, however (Fig. 2). This suggests that an additional protein(s) binds to the ESS, converting an SR-binding element into a negative splicing element. The ESS also binds U2AF and PTB, but the binding of these proteins appears not to be essential for splicing suppression (30). Therefore, the identity of this additional factor(s) remains to be determined. However, one candidate is a new family of splicing repressors that includes the Drosophila RSF1 protein (Rox21), which has an RNA binding specificity similar to that of SR proteins. RSF1 has been shown to inhibit in vitro splicing of several pre-mRNAs at the level of spliceosome assembly (24a).

There are several possible models for the function of the BPV-1 ESS. The ESS, by binding splicing factors or inhibitory factors, could suppress pre-mRNA splicing by competing with the 3' splice site for splicing factors, by interfering with the binding of splicing factors at the 3' splice site, or by interfering with normal bridging interactions between 5' and 3' splice sites (10, 15, 22, 26). Experiments are currently in progress to identify the factors that bind to the BPV-1 ESS and to determine how these factors inhibit splicing.