CRISPR/Cas9-Mediated Genome Editing of Herpesviruses Limits Productive and Latent Infections

Abstract

Herpesviruses infect the majority of the human population and can cause significant morbidity and mortality. Herpes simplex virus (HSV) type 1 causes cold sores and herpes simplex keratitis, whereas HSV-2 is responsible for genital herpes. Human cytomegalovirus (HCMV) is the most common viral cause of congenital defects and is responsible for serious disease in immuno-compromised individuals. Epstein-Barr virus (EBV) is associated with infectious mononucleosis and a broad range of malignancies, including Burkitt's lymphoma, nasopharyngeal carcinoma, Hodgkin's disease, and post-transplant lymphomas. Herpesviruses persist in their host for life by establishing a latent infection that is interrupted by periodic reactivation events during which replication occurs. Current antiviral drug treatments target the clinical manifestations of this productive stage, but they are ineffective at eliminating these viruses from the infected host. Here, we set out to combat both productive and latent herpesvirus infections by exploiting the CRISPR/Cas9 system to target viral genetic elements important for virus fitness. We show effective abrogation of HCMV and HSV-1 replication by targeting gRNAs to essential viral genes. Simultaneous targeting of HSV-1 with multiple gRNAs completely abolished the production of infectious particles from human cells. Using the same approach, EBV can be almost completely cleared from latently infected EBV-transformed human tumor cells. Our studies indicate that the CRISPR/Cas9 system can be effectively targeted to herpesvirus genomes as a potent prophylactic and therapeutic anti-viral strategy that may be used to impair viral replication and clear latent virus infection.

Fig 1. Editing of the EBV genome in latently infected tumor cells using CRISPR/Cas9.

a) Latently infected gastric carcinoma SNU719 cells were transduced with lentiviral CRISPR/Cas vectors targeting the indicated EBV miRNA genes. The lines were subsequently selected with puromycin for 2 days and allowed to recover for 12 days. The activity of the targeted miRNAs was subsequently monitored by introduction of the indicated miRNA sensor vectors and assessment of the mCherry reporter expression after 4 days. Increased sensor expression indicates a loss of EBV miRNAs. The activities for the EBV miRNAs in the absence of gRNAs is presented in S1 Fig. b) Sequencing of CRISPR-targeted EBV genomes indicates editing at the target sites. The EBV genomic locus of BART5 and BART16 were amplified by PCR, cloned in a DNA cloning vector, and subjected to Sanger DNA sequencing. The miRNA sequence is presented in yellow, the gRNA-target sites are displayed in bold, the PAM sequence as red, underlined text, and the cleavage site as a triangle. The number of times each variant has been sequenced is indicated.

Fig 2. CRISPR/Cas9-mediated clearance of EBV genomes from latently infected Burkitt’s lymphoma cells.

a) Anti-EBV gRNAs induce a potent loss of EBV genomes from latently infected cells. Burkitt’s lymphoma Akata-Bx1 cells latently infected with eGFP-EBV (endogenously driving eGFP expression) were transduced with anti-EBV gRNAs targeting EBNA1, OriP, or control genes and selected with puromycin for 2 days. Subsequently, the cells were monitored for the presence of EBV-eGFP by flow cytometry 21 days post transduction. The percentage of eGFP negative cells as measure for EBV-eGFP loss is indicated. b) Combinatorial anti-EBV gRNA treatment of Akata-Bx1 cells causes increased loss of EBV genomes from latently infected cells. Similar experimental set-up as in a), but with a larger set of anti-EBV gRNAs and combinations thereof introduced through sequential application of two separate CRISPR vectors. The percentage of EBV-eGFP negative cells is presented. c) Samples from b) were subjected to qPCR to quantify the relative EBV genome content in the indicated gRNA-expressing Akata-Bx1 cells. Since the amplified region in the qPCR lies outside the genomic region that is targeted by the gRNAs, the qPCR can also detect mutated, yet repaired EBV variants. For all bar diagrams, measurements for (at least) triplicate experiments + STD are presented.

Fig 3. Anti-HCMV gRNAs efficiently abrogate HCMV replication in human cells.

a) Targeting essential HCMV genes with CRISPR/Cas9 impairs HCMV replication. MRC5 cells transduced with the indicated gRNAs were infected with HCMV-eGFP strain TB40/E at an MOI of 0.05 and subjected to flow cytometry at 2, 5, 8, and 11 dpi to assess the percentage of eGFP-positive infected cells. For each essential gene, four different gRNAs were monitored. Besides targeting human genes as controls, gRNAs targeting nonessential HCMV genes US6, US7, and US11 were included. b) Anti-HCMV gRNAs impair replication of both TB40/E and AD169 strains with the exception of anti-UL84 gRNAs. gRNA-expressing MRC5 cells were infected with eGFP-tagged TB40/E or AD169 and the percentage of eGFP-expressing cells was monitored at 2, 5, and 8 dpi. For bar diagrams in a) triplicate measurements + STD are presented. Bar diagrams in b) are from (at least) duplicate measurements.

Fig 4. Targeting of HCMV by single gRNAs directed against essential genes selects for viral escape mutants after prolonged culture.

a) Sequence analysis of HCMV-eGFP variants present after 21 days of culturing in anti-UL57 and UL70 gRNA-expressing cells. The sequences of mutant variants were assessed and the percentage of these that contain in frame-shift mutations at the target site are presented. Genomic DNA was amplified by PCR and subjected to 454 sequencing. Sequences of HCMV variants upon CRISPR/Cas9 targeting of the nonessential HCMV genes US7 and US11 were derived at two days post infection. Number of sequences analyzed: 333 for US7, 331 for US11, 319 for UL57, and 541 for UL7. Bar diagrams are derived from a single 454 sequencing experiment. b and c) Commonly sequenced variants are depicted for UL57 #1 and UL70 #4. The gRNA-target site is displayed in bold, the PAM sequence as red underlined text, and the cleavage site as a triangle. The number of reads per variant is indicated.

Fig 5. Anti-HSV-1 gRNAs impair HSV-1 replication.

a) Vero cells were transduced with the indicated gRNAs and subsequently infected with HSV-1-eGFP at an MOI of 0.05. To assess the percentage of virus-infected cells, eGFP-expression was analyzed by flow cytometry at 2 dpi. Four gRNAs/gene targeting twelve essential HSV-1 genes and two non-essential genes (US3 and US8) were assessed. As controls, empty vector-transduced cells and gRNAs targeting the human genes HLA-A and B2M are presented. b) Prolonged inhibition of HSV-1 replication by gRNAs targeting essential genes UL8, UL29, and UL52. Select gRNA-expressing cells from a) were monitored for HSV-1-eGFP presence at 2 and 3 dpi. c) Indicated gRNA-expressing cells were infected with HSV-1-eGFP at MOI 0.5 or 0.05 and the amount of HSV-1 viral genomes present in the supernatant was assessed by qPCR at 4 dpi. The relative HSV-1 genome content was normalized to supernatant harvested from cells transduced with vector control. For all bar diagrams, measurements for triplicate experiments are presented + STD.

Fig 6. Simultaneous targeting of HSV-1 with two gRNAs completely impairs virus replication.

a) Vero cells were transduced with the indicated single or double gRNAs and subsequently infected with HSV-1-eGFP at an MOI of 0.5. Cells were analyzed for eGFP-expression by flow cytometry at 1, 2, and 3 dpi to assess the percentage of virus-infected cells. b) Similar experiment as in a) but now performed in human MRC5 cells and at an MOI of 0.005, as MRC5 cells are more susceptible to HSV-1 infection. Cells were analyzed for eGFP-expression by flow cytometry at 1 dpi and 3 dpi to assess the percentage of successfully infected cells. c) Supernatants from b) were subjected to plaque assays to quantify the infectious HSV-1 titer produced by gRNA-expressing MRC5 cells that had been infected with HSV-1-eGFP three days earlier. Plaques were scored if visible by eye. d) Plaques from c) obtained after infection of cells with HSV-1 harvested from control or anti-UL8 gRNA-expressing MRC5-cells were analyzed by light-microscopy. Whereas large plaques (‘P’) are observed in infected cells not carrying any gRNAs, virus harvested from UL8 gRNA- expressing cells induced microplaques. In double gRNA-treated cells, no signs of infection were observed. For all bar diagrams, measurements of triplicate experiments are presented + STD.

Fig 7. Anti-HSV-1 CRISPR gRNAs are ineffective in targeting quiescent HSV-1 but do abrogate replication of reactivated HSV-1.

a) MRC5 cells exogenously expressing Cas9 were infected with quiescent HSV-1-eGFP. Anti-HSV-1 gRNAs were introduced into the MRC5-Cas9 cells via lentiviral transduction and cells were subsequently superinfected with HCMV to reactivate latent HSV-1. The percentage of cells with replicating HSV-1 were assessed 3 days post HCMV superinfection by flow cytometry. Two quiescent control (‘cells alone’) samples are presented: one not superinfected with HCMV to assess spontaneous HSV-1 reactivation levels, and one superinfected with HCMV to assess the reactivation potential and subsequent HSV-1 replication in these cells. Control vector corresponds to empty gRNA-vector. b) Relative amount of HSV-1 genomes in quiescent HSV-1-eGFP cells transduced with the indicated gRNAs as assessed by TaqMan qPCR. DNA input was normalized to RNAseP levels. The relative HSV-1 content was compared to quiescent MRC5 cells transduced with empty vector control lentivirus. c) Analysis of anti-HSV-1 genome editing by next generation sequencing. The percentage of WT sequences at the indicated target sites is presented upon introduction of the corresponding gRNA. For UL8 and UL52, two samples were analyzed. In control vector treated quiescent HSV-1 cells, no CRISPR/Cas9 editing was observed at these target sites (gRNA specific indels <0.1%). CRISPR/Cas9 genome editing was only observed for UL52 #2 and UL8 #2 where CRISPR/Cas9 editing was apparent in ±6 and ±1% of sequences. The nature of these mutations is presented in S4 Fig.