Functional Dissection of an Alternatively Spliced Herpesvirus Gene by Splice Site Mutagenesis

Abstract

Herpesviruses have large and complex DNA genomes. The largest among the herpesviruses, those of the cytomegaloviruses, include over 170 genes. Although most herpesvirus gene products are expressed from unspliced transcripts, a substantial number of viral transcripts are spliced. Some viral transcripts are subject to alternative splicing, which leads to the expression of several proteins from a single gene. Functional analysis of individual proteins derived from an alternatively spliced gene is difficult, as deletion and nonsense mutagenesis, both common methods used in the generation of viral gene knockout mutants, affect several or all gene products at the same time. Here, we show that individual gene products of an alternatively spliced herpesvirus gene can be inactivated selectively by mutagenesis of the splice donor or acceptor site and by intron deletion or substitution mutagenesis. We used this strategy to dissect the essential M112/113 gene of murine cytomegalovirus (MCMV), which encodes the MCMV Early 1 (E1) proteins. The expression of each of the four E1 protein isoforms was inactivated individually, and the requirement for each isoform in MCMV replication was analyzed in fibroblasts, endothelial cells, and macrophages. We show that the E1 p87 isoform, but not the p33, p36, and p38 isoforms, is essential for viral replication in cell culture. Moreover, the presence of one of the two medium-size isoforms (p36 or p38) and the presence of intron 1, but not its specific sequence, are required for viral replication. This study demonstrates the usefulness of splice site mutagenesis for the functional analysis of alternatively spliced herpesvirus genes.

Importance: Herpesviruses include up to 170 genes in their DNA genomes. The functions of most viral gene products remain poorly defined. The construction of viral gene knockout mutants has thus been an important tool for functional analysis of viral proteins. However, this strategy is of limited use when viral gene transcripts are alternatively spliced, leading to the expression of several proteins from a single gene. In this study, we showed, as a proof of principle, that each protein product of an alternatively spliced gene can be eliminated individually by splice site mutagenesis. Mutant viruses lacking individual protein products displayed different phenotypes, demonstrating that the products of alternatively spliced genes have nonredundant functions.

FIG 1

Analysis of the intron-exon structure of the MCMV M112/113 gene. (A) Schematic representation of the M112/113 locus and the four E1 proteins according to Ciocco-Schmitt et al. Splice donor and acceptor sites are marked with D and A, respectively. Nucleotide positions of exon boundaries are indicated. (B) Splice junctions were detected by RNA-Seq of cells infected with MCMV Smith or MCMV K181. Splice junctions are depicted with the last position of the upstream exon and the first position of the downstream exon separated by a circumflex. (C) Schematic representation of the revised M112/113 locus based on RNA-Seq data. (D and E) Relative abundances of spliced and unspliced transcripts for the M112/113 intron 1 (D) and intron 2 (E).

FIG 2

Mutagenesis of the M112/113 locus. (A) Schematic view of the MCMV M112/113 locus and the four proteins expressed from the locus by alternative splicing. Mutations introduced to inactivate the expression of individual protein isoforms are shown. Splice donor and acceptor sites are indicated by arrowheads, and the minimal consensus sequences are shown in boldface. Nucleotide and amino acid changes are underlined. (B) NIH 3T3 cells were transfected with empty vector or expression plasmids containing the wt or mutant M112/113 sequence. E1 protein expression was detected by immunoblotting using an E1-specific antiserum. Actin was detected as a loading control.

FIG 3

Replication kinetics and E1 protein expression of MCMV M112/113 mutants. (A) NIH 3T3 cells were infected at an MOI of 0.02 TCID₅₀/cell with wt MCMV and Rev p87. Virus released into the supernatant was quantified by titration. (B and C) Replication kinetics of wt and Δp36 (B) and wt, Δp38, and Rev p38 (C) were determined as described for panel A. Means ± standard errors of the mean (SEM) of the results of experiments done in triplicate are shown. DL, detection limit. (D) NIH 3T3 cells were infected at an MOI of 3 TCID₅₀/cell with wt and mutant MCMVs. E1 protein expression was determined by immunoblotting. IE1 was used as infection control and β-actin as a loading control. (E) NIH 3T3 cells were infected with MCMV wt, Δp38, or Rev p38 at an MOI of 0.5 TCID₅₀/cell. Cell lysates were harvested at different times postinfection. Expression levels of the viral proteins IE1, E1, M44, and gB were analyzed by immunoblotting. The upper part of the E1 immunoblot is shown as a longer exposure for better visualization of the p87 isoform.

FIG 4

Characterization of a spontaneously adapted Δp36 Δp38 mutant. (A) NIH 3T3 cells were infected at an MOI of 0.02 TCID₅₀/cell with wt MCMV or Rev p36p38. Viral replication kinetics were determined as described in the legend to Fig. 3. Means ± SEM of the results of experiments done in triplicate are shown. DL, detection limit. (B) NIH 3T3 cells were infected at an MOI of 3 TCID₅₀/cell with wt MCMV or an MCMV mutant that spontaneously emerged in fibroblasts transfected with the MCMV Δp36 Δp38 BAC (Δp36Δp38mut). E1 protein expression was analyzed by immunoblotting. The asterisk marks a new E1 isoform expressed by the spontaneously occurring MCMV mutant. (C) Identification of a mutation by sequencing the exon 2 to exon 3 region. The nucleotide position of the mutation is shown above. (D) Viral replication kinetics of wt MCMV and Δp36Δp38mut were determined as described in the legend to Fig. 3.

FIG 5

Deletion and substitution mutagenesis of M112/113 intron 1. (A) Schematic showing the deletion of M112/113 intron 1 and insertion of the p33-coding ORF into the nonessential m02-m06 region of the MCMV genome. (B) Schematic of the MCMV SynI1 mutant carrying a synthetic intron in place of the authentic intron 1. An unspliced transcript is predicted to encode an E1 protein significantly larger than p33 (termed p33*). (C) NIH 3T3 cells were infected at an MOI of 3 TCID₅₀/cell with wt MCMV or the SynI1 mutant. E1 protein expression was analyzed by immunoblotting. IE1 was used as an infection control and β-actin as a loading control. (D) NIH 3T3 cells were infected at an MOI of 0.02 TCID₅₀/cell with wt MCMV and the SynI1 mutant. Virus release into the supernatant was determined by titration. Means ± SEM of the results of experiments done in triplicate are shown. DL, detection limit. (E) An M112/113 fragment was PCR amplified from cDNA of wt-MCMV- or SynI1-infected cells. The same PCR amplification was done using different MCMV BACs as templates. The sizes of unspliced (us) and spliced (s) transcripts are shown.

FIG 6

Replication of MCMV M112/113 mutants in endothelial cells and macrophages. (A and B) SVEC4-10 endothelial cells were infected at an MOI of 0.1 TCID₅₀/cell (A) and RAW264.7 macrophages were infected at an MOI of 0.5 TCID₅₀/cell (B) with MCMV mutants. Viral replication kinetics were determined as described in the legend to Fig. 3. (C and D) SVEC4-10 (C) or RAW264.7 (D) cells were infected at an MOI of 3 TCID₅₀/cell with the indicated viruses. E1 protein expression was determined by immunoblotting. IE1 was used as an infection control, and β-actin was used as a loading control.