|Home | About | Journals | Submit | Contact Us | Français|
Gammaherpesviruses Kaposi's sarcoma-associated herpesvirus and Epstein-Barr virus are associated with multiple human cancers. Our goal was to develop a quantitative, high-throughput functional profiling system to identify viral cis-elements and protein subdomains critical for virus replication in the context of the herpesvirus genome. In gamma-2 herpesviruses, the transactivating factor RTA is essential for initiation of lytic gene expression and viral reactivation. We used the RTA locus as a model to develop the functional profiling approach. The mutant murine gammaherpesvirus 68 viral library, containing 15-bp random insertions in the RTA locus, was passaged in murine fibroblast cells for multiple rounds of selection. The effect of each 15-bp insertion was characterized using fluorescent-PCR profiling. We identified 1,229 insertions in the 3,845-bp RTA locus, of which 393, 282, and 554 were critically impaired, attenuated, and tolerated, respectively, for viral growth. The functional profiling phenotypes were verified by examining several individual RTA mutant clones for transactivating function of the RTA promoter and transcomplementing function of the RTA-null virus. Thus, the profiling approach enabled us to identify several novel functional domains in the RTA locus in the context of the herpesvirus genome. Importantly, our study has demonstrated a novel system to conduct high-density functional genetic mapping. The genome-scale expansion of the genetic profiling approach will expedite the functional genomics research on herpesvirus.
Human gammaherpesviruses Epstein-Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus (KSHV) are involved in the development of epithelial, mucosal, hematopoietic, and endothelial cell cancers (7, 28). Murine gammaherpesvirus 68 (MHV-68), a pathogen affecting wild rodents, has been used as an animal model to study the fundamental principles underlying replication of gammaherpesviruses (32, 35, 37). For gamma-2 herpesviruses, including KSHV and MHV-68, the balance between latent and lytic phases of the viral life cycle is controlled by the immediate-early viral protein, RTA (replication and transcription activator) (11, 25, 26, 34, 41); hence, RTA plays a pivotal role in the viral life cycle. RTA is a b-zipper protein consisting of N-terminal DNA binding (DBD) and dimerization domains and a C-terminal transactivation (TA) domain (9, 10, 13, 15, 36). Understanding both the regulation of RTA transcription and the mechanism of RTA function is important. The promoter of RTA contains many regulatory cis-elements that are bound by viral and cellular factors, resulting in transcriptional activation or repression of RTA. Protein kinase C, cyclic AMP/protein kinase A, AP-1, Ras/Raf/MEK/ERK/Ets-1, and Notch signaling pathways have been demonstrated to induce KSHV reactivation through direct or indirect activation of RTA expression (8, 23, 38, 43, 45). NF-κB and LANA negatively regulate the RTA expression (4, 19, 20).
The promoter elements of RTA have been studied extensively using transient transfection of reporter systems. Most of these reporter studies were done independently of the viral genome context. The organization of chromatin structure and regulation could differ between native promoters in the genome and transfected promoter-reporter constructs. The functional domains of the RTA promoter and protein have not been defined at a high resolution during viral de novo infection. The MHV-68 RTA locus under study is 3.8 kb in size (nucletotides [nt] 65570 to 69414) and is comprised of open reading frame 48 (ORF48), ORF49, and ORF50. By mutating each ORF, the functions of ORF48, ORF49, and ORF50 were defined as nonessential, less critical, and essential, respectively, for virus replication (27, 29, 33, 42). The RTA transcript contains two exons, with the first exon (amino acids [aa] 1 to 12) overlapping with the 3′ end of the ORF49 gene locus and the second (aa 13 to 571) mainly encoded by ORF50 (25, 41). The ORF48 coding region contains the RTA promoter cis-regulatory elements for RTA transcription (25, 41). ORF49 cooperates with RTA for activating downstream viral lytic genes (14, 22).
Herpesviruses contain large DNA genomes, with sizes ranging from 140 to 240 kb, which have the capacity to code for 70 to 162 genes. Using reverse genetics approaches, the functions of multitudes of genes of several species have been elucidated. Site-directed mutagenesis is specific for engineering mutations into the genome; however, it is labor-intensive and time-consuming for genome-scale studies. Systematic and transposon insertional mutagenesis of beta- and gammaherpesvirus bacterial artificial chromosome (BAC) plasmids allows for rapid identification of viral genes that are essential for replication in various cell types and in vivo conditions (5, 6, 12, 29, 33, 44). Most viral proteins have evolved to have multiple functions as a result of limited genome space; thus, transposon insertion cannot be used to elucidate the functions of multidomain proteins. Moreover, in compact genomes the inserted transposon could potentially affect the expression of neighboring genes. To overcome these shortcomings, a Mu-transposon mutagenesis method has been employed where the transposon segment is removed by restriction digestion, resulting in only a 15-bp insertion in the viral genome (18, 21). The location of the 15-bp insertion was identified by a genetic footprinting analysis. Random linker-insertion mutagenesis has been employed in structure-function studies of proteins, and the results indicate a minimal effect on protein conformation (18, 21, 24, 30). Therefore, to perform genome-scale functional profiling analysis of a herpesvirus, we have developed a high-throughput mutational analysis platform by combining a Mu transposon-mediated 15-bp random insertion mutagenesis system and a quantitative, genetic profiling method with capillary electrophoresis. We applied this method to dissect the functional domains of the RTA locus and obtained a high-resolution functional profile of the RTA locus in the context of the viral genome during de novo infection.
NIH 3T3, NIH 3T12, BHK-21, Vero, and 293T cells were cultured at 37°C with 5% CO2 in complete Dulbecco's modified Eagle medium containing 10% fetal bovine serum and supplemented with penicillin and streptomycin.
Sequences of the primers used for plasmid constructions are available in Table Table1.1. The genomic coordinates of MHV-68 described in this study are based on the NCBI accession number U97553 sequence. The MHV-68 RTA locus (corresponding to the genome position of nt 65570 to 69414) with flanking Flp recognition target (FRT) sequences (5′-GAAGTTCCTATTCTCTAGAAAGTATAGGAACTTC-3′) was PCR amplified using primers P1 and P2 (Table (Table1).1). The PCR product was cloned into AscI-NheI sites of pSMART VC variable copy vector (Lucigen) and the sequence was verified.
The shuttle vector pGS284 was a generous gift from G. Smith and L. Enquist (Princeton University). To generate a pGS284 FRT/ΔRTA locus shuttle vector, the MHV-68 fragment (nt 64777 to 70212) with a single FRT site, but lacking the RTA locus, was two-step PCR amplified using primers P3/P4 and P5/P6. The final PCR product (generated using primers P3 and P6) was cloned into the pGS284 vector.
The pFLAG-CMV2 (Sigma)-based wild-type (WT) MHV-68 RTA mammalian expression vector was described previously (31). The alanine substitutions in the RTA coding region were introduced by two-step PCR. Each construct required identical outside primers, P7 or P8 and P9, and unique internal primers containing desired mutations as follows: RTA-QQ (aa 37 to 38), primers P10 and P11; RTA-KD (aa 159 to 160), P12 and P13; RTA-REVE (aa 202 to 205), P14 and P15; RTA-QMD (aa 214 to 216), P16 and P17; RTA-Ins1 (aa 389-TPCGR-390), P18 and P19; RTA-Ins2 (aa 390-NCGRT-391), P20 and P21; RTA-EYTG (aa 400 to 403), P22-P23; RTA-SLYD (aa 564 to 567), P24 and P25. The final PCR products were inserted into the NotI-XbaI or EcoRI-XbaI sites of pFLAG-CMV2. The insert sequences for all plasmids were verified.
The WT MHV-68 (BAC) was used for construction of a mutant MHV-68 BAC plasmid by allelic exchange using recA+ Escherichia coli strain GS500 harboring the target MHV-68 (BAC) plasmid and conjugation-competent E. coli GS111 containing the donor suicide shuttle vector pGS284 as previously described (2). To generate the MBAC FRT/ΔRTA locus, the WT MHV-68 BAC plasmid harboring strain GS500 and pGS284 FRT/ΔRTA locus carrying GS111 E. coli were used. The consequent mutation in the recombinant viral genome was confirmed by DNA sequencing, and the genomic integrity of mutated BAC MHV-68 was investigated by restriction enzyme digestion.
The plasmid pSMART FRT-RTA locus was subjected to in vitro Mu transposon-mediated mutagenesis (MGS kit; Finnzymes). A total of 1.2 × 105 individual bacterial colonies were obtained, and the mutant plasmids were isolated from the pooled bacterial colonies. To remove the transposon DNA fragment, 22 μg of the pooled mutant plasmids was subjected to NotI digestion, self-ligation, and selection in bacteria. This resulted in a library of mutants having a 15-bp sequence, 5′-NNNNNTGCGGCCGCA-3′ (with N representing five duplicated nucleotides from target DNA), that were inserted randomly in the RTA locus.
A total of 144 μg of the pSMART FRT-RTA locus library was subjected to AscI-SphI restriction digestion to release the FRT-RTA locus-FRT mutant fragments from the vector backbone. These mutant fragments were recombined into the MBAC FRT/ΔRTA locus in BHK-21 cells using Flp recombinase expression plasmid (pOG44; Invitrogen). BHK-21 cells support efficient Flp recombination compared to the murine fibroblast cell lines tested. For reconstitution, a total of 30 μg, 36 μg, or 24 μg of MBAC FRT/ΔRTA locus, RTA locus mutant fragments, or pOG plasmid, respectively, was transfected into 1.2 × 107 BHK-21 cells (10 12-well plates) using Lipofectamine-Plus. At 2 days posttransfection (dpt) the cells were replated onto 10 15-cm dishes. At 4 dpt viral plaques were observed, indicating cytopathic effects; cell-free culture supernatant and the cells were separately harvested. The virus titer was measured using a plaque assay. Approximately one-third of the cell pellet was used for DNA isolation. The remaining cell pellet and aliquots of infectious supernatants were stored at −80°C. The harvested DNAs were subjected to DpnI restriction digestion to remove nonrecombined and nonreplicated RTA locus fragments.
The reconstituted RTA locus mutant virus library was passaged in NIH 3T12 murine fibroblast cells for an additional five rounds. To avoid transcomplementation during selection, the NIH 3T12 cells (6 × 106 cells/dish; total of 10 15-cm dishes) were infected with a multiplicity of infection (MOI) of 0.05 of the mutant virus library for each passage. At the completion of each round of selection, the virus supernatant and cell pellets were harvested and the virus titers were measured using plaque assay. The infectious supernatant obtained from the previous round of selection was used for the inoculum for the subsequent round of selection. The total DNA was isolated from the cells using phenol-chloroform extraction and was used for functional profiling analysis.
A total of 12 μg of DNA from each of the nonselected and cell culture-selected RTA locus mutant DNAs was used as template for PCR amplification of two overlapping fragments (F1 and F2) using RTA locus-specific primers P26/P47 and P32/P48 (Table (Table2).2). Fifty nanograms of purified PCR product was used as a template for a second PCR with an insertion-specific mini-primer (5′-TGCGGCCGCA-3′), the 5′ end of which is labeled with a fluorescent dye, VIC (Applied Biosystems), and one of the RTA locus-specific primers P26 to P46 (Table (Table2).2). A total of 21 RTA locus-specific primers, designed at approximately 200-nt intervals, were used. Each of the RTA locus-specific primer and mini-primer combinations tested negative for generating any spurious PCR products using wild-type MHV-68 genome as a template. For each primer, the PCRs were done in duplicate. The conditions used for the second, nested PCR were 95°C for 5 min (1 cycle); 95°C for 1 min, 52°C for 1 min, and 72°C for 2 min (35 cycles); and 72°C for 20 min (1 cycle). The fluorescent-labeled PCR products were analyzed in duplicate with a Liz-500 size standard set (Applied Biosystems) by using a 96-capillary DNA analyzer (3730xl DNA Analyzer; Applied Biosystems) at the UCLA Genotyping and Sequencing core facility.
The data generated by the capillary genotyper were processed by using GeneMapper software (Applied Biosystems) using the amplified fragment length polymorphism analysis tool. The normalized data were visualized with an electropherogram and exported as a data file. For each PCR sample, the exported data contained information regarding the PCR product size at the nucleotide level and peak area. The exact position of an insertion in the genome, for each of the 21 RTA locus-specific primer-generated PCR products, was calculated by subtracting 15 nt from the size of a particular PCR product and adding the MHV-68 genome position of the locus-specific primer. Comparison of the 15-nt insertion sites identified in the mutated MHV-68 genome by PCR profiling and sequencing revealed that the accuracy of PCR profiling was within 1 to 2 nucleotides. For each sample, the PCR profiles were consistent among duplicates, and representative data were used for obtaining a final assembly. To assemble the locations of insertion sites for the entire RTA locus, the insertion profiles obtained between 50 and ~250 nt for each specific primer were taken into account. To assign a phenotype for each insertion mutant, the ratio of peak areas between selected and nonselected pools was calculated.
The 10-fold-serially diluted virus samples were inoculated in duplicate onto monolayers of Vero or BHK-21 cells in 12-well plates. The infected cells were overlaid with 1% methylcellulose-containing growth medium. At 4 to 6 dpi the cells were fixed and stained with 2% crystal violet in 20% ethanol. Plaques were counted at various dilutions to determine the titers.
293T or NIH 3T3 cells (1 × 105 cells per well) were seeded onto 24-well plates 16 h prior to transfection. Ten ng each of MHV-68 RTA promoter reporter plasmid, pRpluc (1-kb sequence upstream of MHV-68 RTA exon 1 translational initiation codon driving a firefly luciferase gene), and WT or mutant RTA expression plasmids, 1 ng of pCMV-Renilla Luc and 380 ng of filler DNA (pFLAG-CMV2 vector) were cotransfected using Lipofectamine Plus reagent (Invitrogen). At 24 h posttransfection (hpt), cells were harvested and both firefly and Renilla luciferase activities were assayed using the dual-luciferase reporter assay system (Promega). The firefly luciferase activities were first normalized against the corresponding internal control Renilla luciferase activities. Activation was calculated by comparing normalized firefly luciferase activities of cells cotransfected with an RTA expression plasmid versus those transfected with a control plasmid.
A total of 300 ng of RTA expression plasmid (wild type or mutant) or pFLAG-CMV2 vector alone and 100 ng of RTA-null MHV-68 BAC (contains a mini-Mu transposon insertion at ORF50 nt 68766) were cotransfected into approximately 2.0 × 105 293T cells using Lipofectamine Plus reagent (Invitrogen). At 48 hpt, cells were harvested for quantitative PCR and Western blotting.
The DNA from cell culture was purified using phenol-chloroform extraction. Ten nanograms of genomic DNA template and MHV-68 ORF56 primers (F, five′-GTAACTCGAGACTGAAACCTCGCAGAGGTCC-3′; R, 5′-CCGAAGCTTGCACGGTGCAATGTGTCACAG-3′) was combined with 10× PCR buffer, Taq, and SYBR green. MHV-68 BAC DNA (100 to 107 copies) was included as a standard for copy number determination. The reaction was run at 95°C for 3 min, followed by 55 cycles at 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s. The results were analyzed with an Opticon monitor (MJ Research).
Harvested cells were lysed with radioimmunoprecipitation assay buffer containing 1 mM phenylmethylsulfonyl fluoride protease inhibitor. Equal volumes of cell lysates were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. The membrane was blocked with phosphate-buffered saline-0.1% Tween with 5% milk and probed first with mouse monoclonal antibodies to FLAG (1:10,000) and to β-actin (1:5,000). Subsequently, the membrane was stripped and reprobed with polyclonal antibodies to MHV68 lytic antigen ORF26 or ORF65 (generated in our laboratory). Secondary antibodies, goat anti-rabbit or goat anti-mouse, conjugated to horseradish peroxidase (Amersham Pharmacia Biotech) were detected by chemiluminescence (ECL Plus; Amersham Pharmacia Biotech).
The plasmid containing the 3.8-kb MHV-68 RTA locus (nt 65570 to 69414) with flanking FRT sites was mutagenized in vitro using a mini-Mu transposon (Fig. (Fig.1).1). Following selection in E. coli, a total of 1.2 × 105 individual kanamycin- and chloramphenicol-resistant colonies were obtained. Sequencing 20 randomly chosen clones revealed that 75% and 25% of the insertions were present in the RTA locus and vector sequence, respectively. All of these clones harbored a single transposon insertion. The lower frequency of insertion in the vector sequence is consistent with the fact that some vector sequences are required for plasmid replication in E. coli. From the pooled plasmid library, the body of the transposon was removed by NotI restriction digestion and was self-ligated, leaving only a 15-bp insertion in the target plasmid. Without any stop codons, each 15-bp insertion codes for different amino acid sequences depending on the reading frame (Fig. (Fig.2).2). This RTA locus mutant plasmid library was used as a nonselected input pool for genetic profiling PCR analysis and for reconstitution of the mutant RTA locus into the viral genome (Fig. (Fig.11).
To generate the mutant viral library, the RTA locus plasmid library was reconstituted with a recombinant MHV-68 BAC that had a single FRT site in place of the RTA locus (MBAC FRT/ΔRTA locus) (Fig. (Fig.3A).3A). The reconstitution was performed by cotransfecting Flp recombinase in BHK-21 cells (Fig. (Fig.1B1B and and3A).3A). At the reconstitution step, mutant viruses harboring deleterious insertions were subjected to negative selection. At 4 dpt the viral plaques appeared and the infected cell culture supernatants and total DNA were harvested. After DpnI enzyme digestion to remove input RTA DNA, the isolated total DNA was used for genetic profiling PCR analysis. A total of 5.1 ×107 PFU of reconstituted viruses were recovered. The analysis of the genomic structure of individual reconstituted viruses by Southern blotting confirmed the FRT-mediated recombination events (Fig. (Fig.3B).3B). The FRT sequence contains an XbaI restriction site, and thus the reconstituted virus had two additional XbaI sites. The XbaI restriction pattern was used for distinguishing the reconstituted viruses from the wild-type virus. The FRT sites in the reconstituted virus were located in the intergenic regions of ORF47-48 and ORF50-M7, and the inserted FRT sequences had no deleterious effect on viral replication. The mutant virus library was subsequently selected in mouse fibroblast NIH 3T12 cells for five rounds (passages) at an MOI of 0.05. The low MOI and multiple-passage selection conditions were used to reduce the possibility of transcomplementation among mutant viruses. During each passage of mutant virus selection, the cell culture supernatants and viral DNA from infected cells were harvested, generating selected pools (Fig. (Fig.1B).1B). Comparison of viral titers during the selection process showed an upward trend as the selection progressed (Fig. (Fig.4),4), indicating that the incompetent and attenuated mutant populations were being negatively selected, thus allowing the replication-competent population to increase in proportion.
To precisely identify the locations of the 15-bp insertion sites in the RTA locus, a quantitative high-throughput, functional profiling method was employed (1). DNA from the nonselected input pool and selected pools (reconstituted and fibroblast-selected pools: first, third, and fifth rounds) were used for genetic profiling as shown in Fig. 1C and I. Genetic profiling was not performed for the second and fourth rounds of fibroblast-selected pools. For genetic profiling, two overlapping fragments were PCR amplified from the RTA locus of the nonselected and cell culture-selected mutant DNAs (Fig. (Fig.1C).1C). The purified PCR product was subjected to a second PCR with an insertion-specific primer (fluorescent dye labeled) and 1 of the 21 RTA locus-specific primers (Fig. 1D to G). The fluorescent-labeled PCR products were analyzed by using a capillary DNA analyzer. The profile generated is presented as an electropherogram (Fig. 1H and I) and data file. The electropherogram of the RTA locus profile was useful for visual quantitative analysis of both the replication kinetics and the abundance of each insertion mutant (Fig. (Fig.5).5). During the selection process, the insertion mutants that presented at a below-detectable level in the input pool could be subjected to positive selection; thus, some of the insertion mutants were detected after the third round of selection. The functional profiling data were processed by using GeneMapper software (Applied Biosystems) and the amplified fragment length polymorphism analysis program as described in Materials and Methods. The locations of 15-bp insertions in the RTA locus were precisely determined with respect to the primer location in the MHV-68 genome.
During genetic selection, the 15-bp insertion resulted in maintenance (neutral selection or toleration), reduction (attenuation), and loss (negative selection or critical impairment) of viral replication fitness. To define a phenotype for each insertion mutant, the ratio of the peak areas between selected (fibroblast fifth-round selection) and nonselected pools was calculated. Moreover, the reduction kinetics of insertion mutants during multiple rounds of selection were taken into account while assigning the phenotypes. If an insertion mutant was present in the input nonselected pool, but not present in the reconstituted pool and/or in the first round of fibroblast selection, it was assigned a “critically impaired” phenotype. An insertion mutant that was not recovered after the first round of fibroblast selection was also considered critically impaired. An insertion mutant present in nearly all rounds of selection but absent or at least reduced fivefold at the fifth round of selection when compared to input was assigned an “attenuated” phenotype. An insertion mutant that was present throughout the selection process was assigned a “tolerated” phenotype. The final assembly containing the location and phenotype of each insertion in the RTA locus was obtained (Fig. (Fig.6).6). For a detailed map of RTA locus functional profile see Fig. S1 in the supplemental material.
Functional profiling revealed a total of 1,229 independent 15-bp insertions distributed across the MHV-68 RTA locus (Table (Table3),3), of which 393, 282, and 554 insertions were critically impaired, attenuated, and tolerated, respectively, for de novo lytic virus replication in fibroblasts. The numbers of insertions profiled in various coding and noncoding regions of the RTA locus are listed in Table Table3.3. The RTA coding region insertions had the most deleterious outcomes on replication kinetics, consistent with the critical role of RTA in initiation of viral replication.
The promoter cis-elements controlling the transcription of RTA, a critical gene, overlap with the ORF48 coding region in a reverse orientation (25). MHV-68 virus deficient in ORF48 protein is replication competent both in vitro and in vivo (T. T. Wu, unpublished observations) (27). We observed that insertions at stretches of the RTA promoter/ORF48 region at nucleotides 65582 to 65677, 65777 to 65812, 65863 to 65973, 66144 to 66199, and 66255 to 66439 were subjected to strong negative selection (see Fig. S1 in the supplemental material). Because ORF48 protein function is dispensable for virus replication, the critically impaired and attenuated phenotypes observed during multiple rounds of selection in ORF48 region could be a result of deleterious effects on RTA expression. More than 70% of insertions at the ORF48-49 intergenic region were tolerated for virus replication.
ORF49 is a cotransactivating factor, and its absence greatly reduces the fitness of virus replication (22, 33). During the competition selection in murine fibroblasts, we observed that most of the insertions at aa 40 to 67, aa 93 to 108, and aa 193 to 224 were deleterious for virus replication. Insertions between aa 145 and 169 were mostly attenuating. Many of the insertions at the ORF49 C-terminal 75 aa were tolerated. About 87% of the insertions at the noncoding ORF49-RTA exon 2 intergenic region were tolerated.
The RTA exon 1 coding region overlaps with the ORF49 C terminus in the opposite direction. Seven of the 11 insertions profiled in RTA exon 1 were detrimental for virus replication. RTA exon 2 is comprised of 571 amino acid residues. The RTA N-terminal DBD and dimerization domains are conserved across gammaherpesviruses, while the C-terminal transactivating domain (TA) is less conserved (10). A total of 595 insertions were profiled at RTA (exons 1 and 2), of which 208, 134, and 253 insertions were critically impaired, attenuated, and tolerated, respectively, for MHV-68 replication. Most of the insertions at subdomains aa 1 to 195, aa 266 to 378, and aa 504 to 583 were both deleterious and attenuating for virus replication (Fig. (Fig.6B;6B; see also Fig. S1 in the supplemental material). The remaining subdomains, aa 196 to 264 and aa 379 to 494, had many tolerated insertions. The 3′ end of the RTA noncoding region contained the most insertions that were deleterious for virus replication. These insertions could affect the RTA mRNA stability, poly(A) signal, or enhancer cis-elements.
Since RTA is the master regulator of the lytic gene expression program and viral reactivation, we validated the phenotypes of several potential RTA protein subdomains identified by our functional profiling analysis. Insertions at N-terminal DNA binding and dimerization domains, such as aa 37 to 38 (RTA-QQ), aa 159 to 160 (RTA-KD), and aa 214 to 216 (RTA-QMD), resulted in critically impaired phenotypes (see Fig. S1 in the supplemental material). Insertions at RTA C-terminal transactivation domain aa 564 to 567 (RTA-SLYD) led to attenuation of virus replication. Insertions at aa 202 to 205 (RTA-REVE), aa 390 (RTA-Ins1, 389-TPCGR-390), aa 391 (RTA-Ins2, 390-NCGRT-391), and aa 400 to 403 (RTA-EYTG) were tolerated. According to these results, alanine substitutions or 5-aa insertions were introduced into RTA protein expression plasmids (Fig. (Fig.7A).7A). In a functional profiling study of hepatitis C virus, we observed that the phenotypes of functional subdomain insertion mutations were similar to that of alanine mutations (1). Thus, in this study we have verified the RTA functional profiling phenotypes by using alanine substitution and insertion mutagenesis approaches. Each RTA mutant was tested individually for its ability to both transactivate the RTA promoter and transcomplement an RTA-null virus in 293T cells (Fig. (Fig.7).7). Our previous studies have demonstrated that the transcomplementation assay is efficient for recapitulating the phenotypes of mutant viruses (2, 17, 31, 39, 40).
For the transactivation assay, the cells were transfected with MHV-68 RTA promoter reporter plasmid (1-kb sequence upstream of the MHV-68 RTA exon 1 translational initiation codon driving a firefly luciferase gene) and vector plasmid or wild-type or mutant RTA expression plasmids. Renilla luciferase expression plasmid was included as an internal control. At 24 hpt, the cell lysates were harvested for measuring luciferase activity. Activation was calculated and compared among mutants (Fig. (Fig.7B7B).
For transcomplementation studies, an RTA-null MHV-68 (BAC) having a Mu-transposon insertion at ORF50 genomic position nt 68766 was used (33). The vector or wild-type or mutant RTA expression plasmids were cotransfected with RTA-null MHV-68 (BAC) and at 48 hpt the cells were harvested to examine viral replication. Viral lytic antigen expression and viral genome replication were analyzed by Western blotting detection of viral antigens and quantitative PCR measurement of genome copies, respectively. The expression of ORF26 (triplex-2/capsid protein) (37) and ORF65 (small capsid protein) (37) proteins has been used for assessing MHV-68 viral replication (2, 3, 16, 39). Mutations of the RTA residues QQ, KD, and QMD were detrimental for transactivating function (Fig. (Fig.7B)7B) and failed to transcomplement RTA-null virus replication (Fig. 7C and D). The RTA-SLYD mutant had reduction in both transactivating and transcomplementing functions. Although the N-terminal DBD mutant KD and C-terminal TA domain mutant SLYD had reduced transactivating function, the mutant KD had no detectable transcomplementing function. This observation highlights the functional differences among various RTA subdomains for de novo viral replication. The plasmids containing mutant residues that exhibited the tolerated phenotype in the profiling retained both of the functions tested. The Western blot analysis of MHV-68 lytic protein (ORF26 and ORF65) production by the transcomplemented MHV-68 viruses was consistent with each mutant's phenotype (Fig. (Fig.7C).7C). Thus, the functional profiling phenotypes were verified by alanine substitution and insertion mutagenesis approaches. The transactivation and transcomplementation study performed in NIH 3T3 cells was comparable to that in 293T cells (data not shown). The results indicate that the insertions most likely affect the function of the RTA protein, but not the overlapping promoter cis-elements controlling the expression of neighboring genes, such as the downstream essential gp150 (M7) gene. Furthermore, the verification of functional profiling phenotypes using individual RTA mutant proteins revealed the effectiveness of this high-throughput functional profiling system.
Here we describe a high-resolution functional profile of the MHV-68 RTA locus in the context of the viral genome. Using a high-throughput mutational analysis system, potential RTA promoter cis-elements and RTA protein subdomains critical for virus replication were identified. The high-resolution profiling analysis of herpesvirus, however, has been challenging due to the large viral genome size (140 to 240 kb). Advances in recombination technology, capillary electrophoresis, optical detection, and bioinformatics made this large-scale mutational analysis study possible. We have identified a total of 1,229 independent insertions in the RTA locus, of which 393, 282, and 554 insertions were critically impaired, attenuated, and tolerated, respectively, for viral replication. The RTA locus profiling was conducted by selecting the mutant library for several rounds of replication in cell culture. During the selection, the insertion mutants with low replication fitness were out-competed by viruses with normal levels of replication fitness. This resulted in critically impaired or attenuated phenotypes.
Transcomplementation among mutants during population selection is an issue that can potentially affect the outcome of the functional profiling. During the reconstitution step, the possibility of transcomplementation was high, as the mutant RTA locus fragments and RTA locus null MHV-68 genomes were transfected into the cells. Thus, most of the mutant viruses can be generated. To avoid transcomplementation during subsequent selection steps, however, we used a low MOI of 0.05 to infect fibroblast cells. Furthermore, we subjected the mutant viral library to an additional five rounds of selection after reconstitution. During each round of selection, the chance of two transcomplementing mutants in a population of over 1,000 different mutant viruses coinfecting the same cell is very remote. Thus, these selection conditions greatly minimized the effect of transcomplementation in the final observed mutant phenotypes. Moreover, we were able to reproduce the functional profiling phenotypes by using individual mutants.
We have introduced 15-bp random insertions in the RTA locus using a mini-Mu transposon by in vitro mutagenesis. We have used the optimum concentrations of transposon donor DNA and RTA locus target plasmids to have a single insertion per RTA locus plasmid. Sequencing and restriction digestion analysis of randomly picked RTA locus plasmids revealed only a single transposon insertion. Since the insertions are random, the frequency of the mutant plasmids with two insertions in identical sites is extremely low. If both or one of the insertions is lethal, that mutant will be negatively selected. If both of the insertions are tolerated, that mutant will be subjected to neutral selection. The frequencies of double insertions that can significantly change the phenotypes are negligible. Therefore, mutants with more than one insertion have a low possibility of influencing the phenotypes observed in the functional profiling.
Recently, we reported a functional profiling study of the hepatitis C virus genome (9.6 kb) (1). The large genome size of herpesviruses (ranging from 120 to 240 kb) poses unique challenges for a genome-scale functional profiling study. To introduce a 15-bp insertion, the drug-resistant gene of the transposon that is inserted into the viral genome has to be removed by restriction digestion and religation. This step is technically less difficult for viruses with small genomes. We have observed that after digesting the 140- to 150-kb MHV-68 BAC plasmid with the restriction enzyme, it was very difficult to religate the BAC plasmid ends together. The intramolecular ligation was extremely inefficient. We have tested several approaches and selected a piecemeal approach in which the herpesvirus genome can be profiled as segments. Each segment can be individually subjected to transposon insertion, restriction digestion, and religation to introduce the 15-bp insertions. Subsequently, the mutated segment can be recombined into a virus lacking that particular segment using Flp recombination. We constructed MHV-68 BAC clones with a FRT site replacing each segment of the genome for the herpesviral genome-scale functional profiling. We observed that the Flp recombination step was very efficient in BHK-21 cells for MHV-68 viral reconstitution. We utilized the Flp recombination strategy for RTA locus functional profiling. Under cell culture selection conditions, about 45% of 15-bp insertions in the herpesviral RTA locus did not affect the virus replication, whereas only about 16% of insertions did not affect replication of hepatitis C virus (1). These results suggest that larger-genome-containing organisms can accommodate the genetic insertions without deleterious effects on survival fitness.
The MHV-68 RTA locus contains ORF48, ORF49, and ORF50. Due to the compact nature of the viral genome, the promoter and coding regions of genes have a great deal of overlap. Thus, we are aware that some of the insertions may disrupt the function of protein domains and/or the cis-elements that control the expression of neighboring genes. The promoter cis-elements located in protein coding regions will need to be precisely identified by comparing the profiles of the mutant library that has been selected in a parental cell line and a cell line expressing the disrupted protein. Nevertheless, because ORF48 is not essential for viral replication in vitro, the critical regions identified in the ORF48 coding region most likely contain cis-elements that regulate RTA transcription. The transcription initiation sites of RTA were mapped between nt 66469 and 66503 using an RNase protection assay (25). We observed that most of the insertions upstream of the RTA transcription start site (between nt 66255 and 66439) exhibited the critically impaired phenotypes (see Fig. S1 in the supplemental material). These insertions can possibly interrupt the cis-element sequences essential for the binding of transcription factors involved in RTA mRNA synthesis, thus impairing the viral replication. Hence, in a separate study, we are exploring the mechanism of activation of these RTA promoter cis-elements by identifying and characterizing the trans-factors and the upstream signaling pathways involved.
RTA, as the master regulator for viral lytic replication, executes its function by interacting with various cellular and viral proteins. The MHV-68 RTA has 43% amino acid similarity with KSHV RTA (10). The N terminal of RTA is conserved between these two viruses with limited amino acid conservation in the C-terminal TA domain. We have identified many deleterious insertions in the regions corresponding to the RTA N-terminal DBD and C-terminal TA domain. These insertions may result in RTA mutant proteins that fail to bind DNA or interact with other critical proteins. The functional profiling phenotypes of the RTA protein were verified by examining several individual mutants having amino acid substitutions and insertions. Substituting RTA N-terminal DBD residues KD (Fig. (Fig.7)7) and C-terminal TA domain residue SLYD resulted in a reduction in transactivating function; however, the KD mutant was incompetent in transcomplementing RTA-null virus. This interesting observation could be a result of the failure of the KD mutant to bind to the RTA target promoter elements, thus rendering it defective in transactivating the lytic genes and completing lytic viral replication, whereas the SLYD mutant could bind to the RTA target DNA elements and could partially bind to the transactivating factors, resulting in initiation of the viral lytic gene expression cascade despite a lower level compared to that of wild-type virus. These results suggest that the DNA binding domain is absolutely critical for RTA function. The different results could also be due to the different ratios of RTA protein to the viral promoter DNA in the reporter assay and in the transcomplementation assay. Insertions in RTA amino acid residue 37 resulted in lethal and attenuated phenotypes, and insertions at residue 38 resulted in tolerated phenotypes. Analysis of the mutant RTA with alanine substitutions of residues 37 and 38 (RTA-QQ) exhibited a lethal phenotype. This result suggests a critical role for RTA residue 37 during viral replication.
The present study is the first comprehensive high-resolution mutational analysis of a 3.8-kb viral genome locus in the context of the viral genome. We have identified many of the RTA subdomains that were nonessential for virus replication in cell culture; however, these subdomains might play critical roles during in vivo lytic and latent infections. Thus, profiling the mutant library in wild-type and knockout mice, as well as in mice with various genetic backgrounds, would provide greater insight into the role of RTA in virus-host interactions. We have obtained a functional profile of the RTA locus during infection in BALB/c mice (unpublished data). We have shown that the mutant viral library can be recovered from the lung tissues to generate a profile, which sets the stage for our future in vivo studies. The viral promoter and protein are both positively and negatively regulated by viral and cellular factors. Selecting the mutant viral library in the presence or absence of these regulatory factors in cell culture would enable the identification of the viral subdomains that interact with these factors.
In the future, this approach can be expanded to genome-scale profiling. To demonstrate the feasibility, we have successfully reconstituted up to 21-kb viral fragments into the viral genome and efficiently recovered infectious viruses. Thus, the whole herpesvirus genome can be profiled by mutating and reconstituting overlapping viral fragments to cover the entire length of the genome. Functional domain mapping can complement structural biology studies of viral proteins. The whole genome library can be used to elucidate the function of viral subdomains involved in tissue tropism, immune regulation, autophagy, apoptosis, cell survival, signal transduction, and other cellular processes. This approach will greatly expedite the functional genomics studies of herpesviruses.
We thank Le-Ming Tong for technical assistance and Stacy Hu for helpful discussions and critical review of the manuscript.
This work was supported by the California HIV/AIDS Research Program training award (F06-LA-232) to V.A. and was partially supported by grants CA091791, DE019085, DE018337, and DE015612.
Published ahead of print on 10 December 2008.
†Supplemental material for this article may be found at http://jvi.asm.org/.