Editing the Human Cytomegalovirus Genome with the CRISPR/Cas9 System

Establishment of CRISPR-mediated homology-directed HCMV mutagenesis.

To determine the feasibility of using CRISPR/Cas9 to make large (~1.5kb) changes to the viral genome, we created a construct encoding a GFP-blasticidin deaminase fusion protein (GFP-BSDR) flanked by homologous HCMV sequences (Figure 1A), which enables homologous recombination (HR) into the HCMV genome. A non-essential locus between US34 and TRS1 was selected to assess HR efficiencies. Briefly, MRC5 fibroblasts were transduced with a lentiviral construct expressing Cas9 and a gRNA _targeting this region. These cells were then infected with wild type virus at varying multiplicities of infection (MOIs) and transfected with the GFP-BSDR template under a variety of electroporation, infection, and culture conditions to elucidate whether HR was possible, and identify the treatments resulting in the highest recombination efficiency. After five days of infection, supernatants containing viral progeny were harvested, diluted, and plated (Figure 1B). Subsequent plaque formation and monitoring of GFP expression indicated successful HR. GFP positive plaques emerged with varying efficiencies, with the results summarized in Figure 1C. No recombination was evident when electroporation was performed prior to infection (Figure 1C). Electroporation at 24h post infection gave rise to successful recombination events (Figure 1C), which were subsequently verified by sequencing several isolates to confirm GFP-recombination in the proper locus. Of the various MOIs tested, the maximum recombination efficiency was observed at an MOI = 0.25 (Figure 1C). Blasticidin selection reduced the total amount of WT virus produced (data not shown); however, it substantially increased the HR efficiency, e.g., by 12- and 26-fold at MOIs of 0.25 and 1.0 respectively, with the maximum recombination efficiency observed under these conditions being 11.8%.

High efficiency mutation of the HCMV genome via CRISPR-mediated non-homologous end-joining (NHEJ).

Repair of CRISPR-mediated DNA cleavage can take place via homologous recombination. Alternatively, CRISPR-mediated cleavage can occur via the more imprecise non-homologous end-joining pathway (NHEJ), which introduces insertions or deletions (INDELs) that can disrupt amino acid coding. NHEJ repair requires the activity of DNA Ligase IV, and it has been reported that the efficiency of HR can be increased via pharmaceutical inhibition of DNA Ligase IV (Chu et al., 2015). Studies have also shown that the size of the repair template can influence HR efficiencies, with smaller single stranded oligonucleotide (ssODN) templates leading to higher efficiency (Song and Stieger, 2017). We therefore assessed the effect of a DNA Ligase IV inhibitor, SCR7, and smaller HR repair templates on HR efficiency. Using the GFP+ virus generated in Figure 1, we analyzed HR-mediated and NHEJ-mediated mutation of the GFP open reading frame (Figure 2A). Primary fibroblasts that co-express Cas9 and a GFP-_targeting gRNA, were infected in the presence or absence SCR7, and a small 200bp ssODN repair template, homologous to GFP, but containing a triple stop codon (Figure 2A) to ablate GFP expression. The efficiency of GFP KO was assessed via monitoring GFP+ plaque formation. High efficiency GFP knockout was observed under all conditions (~80–90% loss of GFP). This loss of functional GFP expression occurred irrespective of the presence of repair template (Figure 2B), suggesting that HR was not responsible for the loss of GFP, and that the SCR7 treatment was not effective at inhibiting NHEJ. Subsequent cloning and sequencing of 25 different clones confirmed that the loss of GFP was primarily due to INDELs and not due to recombination with the repair template (data not shown). These results suggest that neither treatment with SCR7 nor the utilization of a small ssODN repair template improved HR efficiencies. However, our results with INDEL-mediated inactivation of GFP expression suggest that NHEJ-mediated INDEL mutation of the HCMV genome occurs at very high efficiency, and could therefore be a helpful tool for INDEL-mediated _targeting of the HCMV genome.

In the experiments presented in Figure 2, measurement of HR efficiency required extensive clonal sequencing. To more quickly determine HR efficiency, we assessed the HR-dependent repair of a mutant GFP allele. We isolated a viral clone created by the experiment described in Figure 2 in which GFP was inactivated by an INDEL. The clone selected possessed a 21 base-pair deletion near the N-terminus of GFP. This mutation can subsequently be repaired by a newly designed gRNA, and repair template to restore GFP function, thereby facilitating better estimates of HR efficiencies (Figure 3A). For these experiments, we assessed the effect of SCR7 treatment, repair template volume, and electroporation timing, on HR efficiency. We found that electroporation earlier during infection (12hpi vs. 24hpi) increased HR efficiencies by approximately 10-fold (Figure 3B). Further, increasing the amount of HR template (10uL vs. 20uL of a 10uM oligo solution) also increased HR efficiencies by approximately 3–4-fold. In contrast, treatment with 10uM SCR7 solution during the viral incubation cycle did not appear to significantly increase recombination efficiency. For ssODN-mediated repair maximum efficiency (14/83 GFP+ plaques or ~17%) was achieved with 20uL of oligo solution, electroporation at 12 hours post infection, and no SCR7 treatment (Figure 3B).

In Figure 1, we found the recombination templates encoding blasticidin resistance increased HR efficiencies by approximately 10-fold (Figure 1C). We therefore assessed whether blasticidin selection could improve HR efficiencies with the optimized parameters employed in Figure 3B, i.e., electroporation at 12 hpi, and increased HR template volume. We also assessed the impact of increased electroporation voltage, i.e. 150V versus 100V, and examined the frequency of recombination at a different locus, in between US6 and US7. As shown in Figure 3C, these modifications resulted in a maximum HR efficiency of 42.3%, i.e. with 30uL HR electroporation at 150 volts, 12 hours after infection at MOI = 0.25, and with blasticidin selection. These results suggest that CRISPR-mediated HR could be an effective method for homologous site-directed mutation of the HCMV genome.

Cell culture and infection.

HEK293T cancer cells and immortalized MRC5-hTERT fibroblasts (referred to as MRC5 cells in the text) were cultured in Dulbecco’s modified Eagle medium (DMEM, Invitrogen), supplemented with 10% fetal bovine serum, 4.5g/L glucose, and 1% penicillin-streptomycin. MRC5s were seeded at 3.4 × 10⁵ in a 10cm dish 72 hours prior to viral infection with the HCMV strain AD169. Cells were exposed to a 4mL viral inoculum at the specified MOI for 2 hours, after which, the virus was aspirated and replaced with fresh media.

Pharmacological Inhibitors.

SCR7 Pyrazine and Blasticidin S Hydrochloride were purchased from Sigma Aldrich and solubilized in DMSO. Cells were treated with 10μM SCR7 in DMEM for the full 120 hours of initial infection. Cells were treated with 10ug/mL Blasticidin from 48 to 120 hours post infection.

Homologous recombination templates.

For GFP-BSD insertion, upstream and downstream homology arms (see homology arm sequences and PCR primers below) were PCR amplified from 750 base-pair gene blocks (IDT). The SV40 promoter and polyadenylation sequences were PCR amplified from pEPkan-S, and EGFP/BSDR was amplified from LeGO/BSD (Addgene). These individual components were then assembled through Gibson Assembly into the PstI site of pBluescript KS (−) (Addgene). Correct clones were identified by sequencing using the T7F and M13R promoter primers. For the Stopx3-HindIII and GFP-deletion repair templates (see sequences below), oligos were ordered as 200 base-pair Ultramers from IDT and dissolved to 100uM in water.

Guide RNA Cloning of US34/TRS1 _targeted CRISPR/Cas9 lentiviral vector.

Oligos with the following sequences: 5’CACCG GAC GCG AAA GCA ACG TGT AT 3’ and 5’AAAC AT ACA CGT TGC TTT CGC GTC C 3’ were phosphorylated and annealed in a thermocycler (Applied Biosystems 2720 Thermo Cycler) under the following conditions: 37°C for 30 min, 95°C for 5 min then ramp down to 25°C at 5°C/min. Oligos were then cloned into the BsmB1 site of pLentiCRISPRv2 (Addgene), amplified via transformation into stbl3 E. coli, and confirmed by sequencing using a h-U6F promoter primer. Similar cloning approaches where employed for guide RNA cloning constructs _targeting US6/7, the Stopx3-HindIII insertion, and the GFP deletion repair: US6/7 oligos: 5’ CACCG CCC CGA TGA ACG CTC TCG TC 3’; 5’ AAAC GA CGA GAG CGT TCA TCG GGG 3’;, Stopx3-HindIII insertions oligos: 5’ CACCG GGG CGA GGA GCT GTT CAC CG 3’, 5’ AAAC CG GTG AAC AGC TCC TCG CCC C 3’; GFP-deletion repair oligos:5’CACCG GAG CAA GGT GGT GCC CAT CC 3’, 5’ AAAC GG ATG GGC ACC ACC TTG CTC C 3’.

Lentiviral transfection/transduction to introduce viral-_targeted Cas9.

293T cells were seeded at 6.8 × 10⁵ cells in a 10 cm dish and grown for 24 hours. Using the Fugene 6 transfection reagent (Promega), each 10 cm dish was transfected with 2.6 μg pLentiCRISPRv2 (Addgene), 0.25 μg pVSVg, and 2.5μg PAX2. After 24 hours, the media were replaced by 4 mL fresh media. After another 24 hours, lentiviral-containing media were taken from each plate, filtered through a 0.45μm filter, and applied to MRC5 fibroblasts with polybrene over the course of 24 hours. The media were then replaced with 10 mL fresh media and, after another 48 hours, cells were placed under 1μg/mL puromycin selection for 4 days.

Electroporation of GFP-BSD HR template into Fibroblasts.

Ten μg of plasmid HR template were linearized with HindIII in a 30uL reaction and purified using a Qiaquick PCR Purification Kit (Qiagen). A 10 cm dish of 90% confluent MRC5 fibroblasts (~1×10⁶ cells) was trypsinized, resuspended in 100 μL Nucleofection Solution #6 (140mM sodium phosphate buffer, 5 mM MgCl₂, 5 mM KCl, pH 7.2) and electroporated with 30uL linearized HR template using the exponential decay pulse setting at 100V or 150V, and 960 μF. Similar results were obtained when electroporating cells from a single well of a six well dish (~3×10⁵ cells) using identical electroporation conditions as indicated above. Cells were then re-plated into a single well of a 6-well plate. For the STOPx3HindIII insertion and GFP-deletion repair, cells were treated as described above, except 10uL of 100uM HR template solution was added per electroporation.

Plaque Assays.

Forty-eight hours prior to infection, 5 × 10⁴ MRC5-hTERT cells were plated into each well of a 12-well plate. Prior to infection, using a deep (1mL) 96-well plate, each viral stock was serially diluted from 10⁻¹ to 10⁻⁶ to a volume of 1mL per well. For a given viral stock, 450 ul of each dilution was added to 1 well of cells and incubated overnight. Twenty-four hours post infection, viral-inocula were aspirated and replaced with pre-warmed gel overlay solution (50% 2X DMEM with 1% PenStrep, 40% of 2.5% low melting NuSieve agarose in H20, 10% FBS). Plaques were typically counted 10 days post infection, and viral clones isolated 15 days post infection.

Viral/Cellular DNA Extraction.

Seventy-two hours post infection with viral supernatant, cells were scraped in their media, pelleted by centrifugation at 2000 rpm for 5 min, washed with PBS, and centrifuged once more. Cells were resuspended in 200–500uL lysis buffer (100mM NaCl, 100mM Tris at pH 8, 25mM EDTA, 0.5% SDS, 0.1 mg/mL Proteinase K, 40ug/mL RNAse A) and incubated at 55°C overnight. Samples were extracted once with an equal volume amount of Phenol/Chloroform/Isoamyl Alcohol (25:24:1). A second extraction was performed with Chloroform/Isoamyl Alcohol (24:1). DNA was precipitated by addition of 0.1 volume of 3M sodium acetate and 2 volumes of 95% ethanol at −20 degrees. DNA was pelleted by centrifugation for 15 min at 14,000 rpm, and washed once with 70% ethanol. The DNA pellet was resuspended in 60uL H20.

Homology Arm Sequences.

PCR Primers.

Day 1:

One day prior to transfection, split a confluent 10 cm dish of 293Ts 1:8 in 10 mL DMEM. Ideally cells should be 30–40% confluent before transfection. Three 10 cm dishes of 293Ts will be required for each 10 cm dish of MRC5s to be transduced.

Day 2:

For each plate of 293Ts to be transfected, combine the following in a 15 mL conical tube:

• 600 uL PBS

• 2.4 ug psPax2

• 0.25 ug pVSVg

• 2.6 ug pLentiCRISPRv2 / gRNA ligation mixture

Incubate mixture at room temperature for 5 minutes.

Add 18 ul FuGene (Promega) to each conical and incubate at room temperature for 20–30 minutes.

Finally, at the edge of each plate, add 600uL of Fugene/DNA/PBS complex dropwise to avoid dislodging cells from dish. Gently rock the plates back and forth, and incubate overnight at 37C, 5% CO2.

Day 3 (BSL2+):

Aspirate media from each plate of transfected 293Ts, and replace with 4 mL fresh DMEM.

Additionally, seed 3–4e5 MRC5s in the desired number of 10 cm plates to be transduced. Be sure to include an extra dish of MRC5 for a non-transduced control.

Day 4 (BSL2+):

For each 10 cm plate of MRC5s to be transduced, aspirate medium and immediately add lentivirus collected from one dish of 293Ts that had been previously filtered through a 0.45 um filter. Then, add polybrene to 10 ug/mL. Incubate MRC5s with lentivirus at at 37C, 5% CO2 for 3–4 hours, aspirate media and repeat incubation with filtered lentivirus collected from the second dish of transfected 293Ts, omitting polybrene this time. Allow the plate to sit for an additional 3–4 hours, then repeat transduction a third time with filtered lentivirus collected from the last plate of 293Ts. Incubate transduced cells with lentivirus overnight.

Day 5 (BSL2+):

From each dish of transduced MRC5s, aspirate lentiviral inoculum, and add fresh DMEM.

Day 7:

If sufficiently confluent (>80%), passage cells 1:3 into three 10 cm dishes with DMEM and 1ug/mL puromycin. The non-transduced control cells should be passaged and placed under selection at this step as well. Puromycin selection is usually complete after 3–4 days, as assessed by complete toxicity of non- transduced MRC5s.

If cells are insufficiently confluent, they can be split at lower dilutions or be allowed to fill in for several days before splitting.

Day 10+:

Once selected for, transduced cells can be used immediately or frozen and cryogenic stored.

Day 1:

Cells should ideally be 50–80% confluent prior to infection. Incubate cells with WT HCMV in DMEM for 2 hours at an MOI of 0.25 (approximately 6e4 pfu in a confluent dish). Replace viral inoculum with fresh medium and incubate overnight.

Day 2:

Approximately 12 hours after infection, trypsinize, inactivate serum and resuspend cells as usual. Centrifuge cells at 1500 RPM for 5 min to pellet. Resuspend cells in 80 uL of the following solution: 140mM sodium phosphate buffer, 5 mM MgCl₂, 5 mM KCl, pH 7.2. Transfer cells to a 4mm electroporation cuvette and add 30uL 10uM ssODN or 5–10ug plasmid HR template. Tap to mix the DNA/cell suspension.

Electroporate with the following settings using an exponential decay pulse: 150V, 960uF. We’ve found 150 volts to be an effective middle ground between maximizing transfection and minimizing cell death, although specific electroporation conditions may need to be empirically optimized.

Immediately transfer the electroporated cells into 1 well of a 6-well dish with 2mL fresh pre-warmed DMEM. Swirl the plate to disperse the cells, and incubate overnight.

Day 3:

Change medium on the electroporated cells with fresh warmed DMEM.

Day 6:

To harvest virus, scrape cells in their medium and transfer to a 15 ml conical. Sonicate 3 times, for 30 seconds each time, vortex gently, then centrifuge at 2500 rpm for 5 minutes. Transfer the viral supernatants to cryovials and discard the pellet. For storage, viral supernatants should be first frozen on dry ice, then transferred to liquid nitrogen (not −80C?).