General and target-specific RNA binding properties of Epstein-Barr virus SM posttranscriptional regulatory protein

Abstract

Epstein-Barr virus (EBV) SM protein is an essential nuclear shuttling protein expressed by EBV early during the lytic phase of replication. SM acts to increase EBV lytic gene expression by binding EBV mRNAs and enhancing accumulation of the majority of EBV lytic cycle mRNAs. SM increases target mRNA stability and nuclear export, in addition to modulating RNA splicing. SM and its homologs in other herpesvirus have been hypothesized to function in part by binding viral RNAs and recruiting cellular export factors. Although activation of gene expression by SM is gene specific, it is unknown whether SM binds to mRNA in a specific manner or whether its RNA binding is target independent. SM-mRNA complexes were isolated from EBV-infected B-lymphocyte cell lines induced to permit lytic EBV replication, and a quantitative measurement of mRNAs corresponding to all known EBV open reading frames was performed by real-time quantitative reverse transcription-PCR. The results showed that although SM has broad RNA binding properties, there is a clear hierarchy of affinities among EBV mRNAs with respect to SM complex formation. In vitro binding assays with two of the most highly SM-associated transcripts suggested that SM binds preferentially to specific sequences or structures present in noncoding regions of some EBV mRNAs. Furthermore, the presence of these sequences conferred responsiveness to SM. These data are consistent with a mechanism of action similar to that of hnRNPs, which exert sequence-specific effects on gene expression despite having multiple degenerate consensus binding sites common to a large number of RNAs.

FIG. 1.

Experimental design for quantitative analysis of RNA association with SM in EBV-infected cells. Lymphoma cells were either mock induced or induced to permit lytic EBV replication, and the cells were divided into fractions for immediate RNA isolation (gray boxes) or IP. After IP with either control or SM-specific antibody, RNA was isolated from the immunoprecipitates. All RNA was reverse transcribed, and the relative amount of each EBV transcript was measured by qRT-PCR.

FIG. 2.

Relative abundances of EBV mRNAs associated with SM. C_T values from qRT-PCR for each EBV transcript in the RNA isolated from the SM IP were converted to the increase above or below the mean enrichment for all transcripts (y axis). Enrichment for each transcript was determined by comparison of its C_T value in the SM IP versus its C_T value in the control immunoprecipitate. Each EBV mRNA is represented by a bar on the x axis, and its enrichment relative to the mean amount of all RNAs is represented on the y axis.

FIG. 3.

Relationship of SM association to transcript abundance and level of gene induction. (A) The C_T value for each EBV transcript in the total RNA from induced P3HR1-ZHT cells is shown on the y axis, with each EBV transcript plotted on the x axis, demonstrating the range and distribution of transcript abundances. (B) Scatter plot of the relative enrichment of each transcript in the SM IP relative to the mean enrichment for all transcripts (ΔC_T; y axis) versus the abundance of each transcript (C_T; x axis). The area highlighted in gray contains those transcripts highly enriched in the SM IP. (C) The level of induction of each EBV transcript during lytic replication (ΔC_T in induced versus uninduced P3HR1-ZHT RNA) is represented on a logarithmic y axis, and each EBV transcript is plotted on the x axis. (D) Scatter plot of the relative enrichment of each transcript in the SM IP versus the control IP (ΔC_T; y axis) versus the level of induction of each transcript on a logarithmic x axis.

FIG. 4.

Northern blot analysis of SM-enriched and unenriched EBV RNAs. (A) RNA from mock-induced P3HR1-ZHT cells, induced cells, PI IP, and SM IP were analyzed by Northern blotting for BFRF3 and BBRF3 (identified as SM associated by qRT-PCR) or BALF2 (non-SM enriched). The relative amounts of RNA were quantitated by phosphorimager measurement and are shown below the lanes. IgG, immunoglobulin G. (B) The amounts of SM present in the total cell lysate from uninduced (U) and induced (I) P3HR1-ZHT cells, input lysate (Input), the control IP (IgG), and SM IP (SM) were measured by immunoblotting with anti-SM antibody; 1% of input lysate and 15% of the IP from cells are shown.

FIG. 5.

EBV RNAs associated with SM in B95-8 cells. (A) Enrichment of each EBV transcript (the amount in SM IP versus PI IP) compared to the mean enrichment for all transcripts in the SM IP was calculated and plotted on the logarithmic y axis. Each EBV mRNA is represented by a bar on the x axis. (B) Scatter plot of the relative enrichment of each transcript in the SM IP versus the mean enrichment for all transcripts (ΔC_T; y axis) versus the abundance of each transcript (C_T; x axis). The area highlighted in gray contains those transcripts most highly enriched in the SM IP. (C) RNAs from mock-induced B95-8ZHT cells, induced cells, PI IP, and SM IP were analyzed by Northern blotting for BFRF3 and BBRF3 (identified as SM associated by qRT-PCR) or BALF2 (non-SM enriched). The relative amounts of RNA were quantitated by phosphorimager measurement and are shown below the lanes.

FIG. 6.

In vitro binding of SM to BFRF3 RNA. (A) In vitro-transcribed and [³²P]UTP-labeled RNA corresponding to four quarters (qtr) of the BFRF3 mRNA, full-length mRNA, and antisense mRNA were purified by urea-PAGE electrophoresis. (B) Radiolabeled BFRF3 transcripts or antisense BFRF3 transcripts (AS-BFRF3) were incubated with lysates from cells transfected with empty vector (C) or SM as shown above each panel and cross-linked with UV light. RNase-treated samples were immunoprecipitated with either PI serum (PI) or anti-SM antibody (Ab), electrophoresed, and analyzed by autoradiography. (C) Transcripts from each quarter of BFRF3 were incubated with SM lysate and analyzed as for panel B.

FIG. 7.

In vitro binding of SM to BDLF3 RNA. (A) Full-length BDLF3 RNA, antisense BDLF3, or transcripts from four quarters (qtr) of the BDLF3 gene were analyzed for binding to SM by in vitro photo-cross-linking and label transfer assay. A control reaction performed in parallel without UV cross-linking (no X-link) is also shown. (B) Transcripts from Q1 of BFRF3 or the corresponding region of BALF2 were incubated with SM lysate (SM) in duplicate or mock-transfected lysate (C) and analyzed as for panel A.

FIG. 8.

Differential binding of BFRF3 subclones to SM in vivo and activity as an SM response element. (A) Plasmids in which BFRF3 Q1 or BFRF3 Q4 was fused upstream of the luciferase (luc) coding sequence in a reporter vector were cotransfected with either vector or SM expression plasmid. RNA was isolated from both the nuclei (N) and cytoplasm (C) of transfected cells and analyzed by Northern blotting with luciferase probe. (B) BFRF3 Q1 and BFRF3 Q4 plasmids were transfected with either vector or SM plasmid, and RNA was analyzed by Northern blotting with BFRF3 probe. (C) Cells were transfected with SM and either BFRF3 Q1, BFRF3 Q4, or full-length BFRF3 plasmid (F.L.). RNA was isolated from total cell lysate (input RNA) or from SM immunoprecipitates (SM-IP-RNA) and analyzed by Northern blotting with BFRF3 probe. The amounts of SM in the input lysate and in each immunoprecipitate (SM IP) were measured by immunoblotting and are shown below.