Multiplexed Genome Editing

Abstract

A method of modulating some or all copies of a gene in a cell is provided including introducing into a cell one or more ribonucleic acid (RNA) sequences that comprise a portion that is complementary to all or a portion of each of the one or more target nucleic acid sequences, and a nucleic acid sequence that encodes a Cas protein and maintaining the cells under conditions in which the Cas protein is expressed and the Cas protein binds and modulates the one or more target nucleic acid sequences in the cell.

Claims

1-43. (canceled)

44. An isolated porcine organ comprising a plurality of genetically engineered porcine cells, wherein each genetically engineered porcine cell of said plurality is characterized by a plurality of retrovirus genes and in that at least 20% of the retrovirus genes comprise a genetically engineered mutation.

45. The organ of claim 44, wherein at least 50% of the retrovirus genes comprise a genetically engineered mutation.

46. The organ of claim 44, wherein at least 70% of the retrovirus genes comprise a genetically engineered mutation.

47. The organ of claim 44, wherein at least 80% of the retrovirus genes comprise a genetically engineered mutation.

48. The organ of claim 44, wherein at least 85% of the retrovirus genes comprise a genetically engineered mutation.

49. The organ of claim 44, wherein at least 90% of the retrovirus genes comprise a genetically engineered mutation.

50. The organ of claim 44, wherein at least 95% of the retrovirus genes comprise a genetically engineered mutation.

51. The organ of claim 44, wherein at least 96% of the retrovirus genes comprise a genetically engineered mutation.

52. The organ of claim 44, wherein the retrovirus genes are endogenous retrovirus genes.

53. The organ of claim 44, wherein the retrovirus genes are PERV genes.

54. The organ of claim 44, wherein the retrovirus genes are retroviral pol genes.

55. The organ of claim 44, wherein the retrovirus genes are retroviral PERV pol genes.

56. The organ of claim 44, wherein the genetically engineered mutation is observed by sequencing.

57. The organ of claim 44, wherein the plurality of retrovirus genes includes 5 or more retrovirus genes.

58. The organ of claim 44, wherein the plurality of genetically engineered porcine cells contains intact PERV gag and/or env genes.

59. An isolated porcine organ comprising a plurality of genetically engineered porcine cells, wherein each genetically engineered porcine cell of said plurality comprises genetically engineered mutations in PERV genes and exhibits at least a 1000-fold reduction in PERV transmission co-culturing a human cell with the genetically engineered porcine cell as compared to co-culturing a human cell with a wild-type porcine cell.

60. The organ of claim 59, wherein the human cell is HEK293.

61. The organ of claim 59, wherein the PERV genes are PERV pol genes.

62. The organ of claim 59, wherein the genetically engineered mutations are observed by sequencing.

63. The organ of claim 59, wherein the plurality of genetically engineered porcine cells contains intact PERV gag and/or env genes.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0029] FIGS. 1A-1C illustrate CRISPR-Cas9 gRNAs were designed to specifically target the pol gene in 62 copies of PERVs in PK15 cells. (A) Phylogenetic tree representing endogenous retroviruses present in the pig genome. PERVs are highlighted in blue. (B) Copy number determination of PERVs in PK15 cells via digital droplet PCR. The copy number of pol elements was estimated to be 62 using three independent reference genes: ACTB, GAPDH, and EB2. N=3, mean+/−SEM. (C) Two CRISPR-Cas9 gRNAs were designed to target the catalytic region of the PERV pol gene. The two gRNA targeting sequences are shown below a schematic of PERV gene structure. Their PAM sequences are highlighted in red. (SEQ ID NO:27-28)

[0030] FIGS. 2A-2B illustrate clonal PK15 cells with inactivation of all copies of PREV pol genes after Cas9 treatment. (A) A bimodal distribution of pol targeting efficiencies was observed among the single-cell-derived PK15 clones after 17 days of Cas9 induction. 45/50 exhibited <16% targeting efficiency; 5/50 clones exhibited >93% targeting efficiency. (B) PK15 haplotypes at PERV pol loci after CRISPR-Cas9 treatment. In red, indel events in the PERV pol sequence are represented. Shades of purple indicate endogenous PERVs.

[0031] FIGS. 3A-3D illustrate: (A) Detection of PERV pol, gag, and env DNA in the genomes of HEK-293-GFP cells after co-culturing with PK15 cells for 5 days and 7 days (293G5D and 393G7D, respectively). A pig GGTA1 primer set was used to detect pig cell contamination of the purified human cells. (B) qPCR quantification of the number of PERV elements in 1000 293G cells derived from a population co-cultured with wild type PK15 cells using specific primer sets. (N=3, mean+/−SEM) (C) qPCR quantification of the number of PERV elements in PK15 Clones 15, 20, 29, and 38, with high levels of PERV pol modification, and minimally modified Clones 40 and 41. (N=3, mean+/−SEM) (D) Results of PCR on PERV pol on genomic DNA from various numbers of HEK 293-GFP cells (0.1, 1, 10, and 100) isolated from populations previously cultured with highly modified PK15 Clone 20 and minimally modified Clone 40. See Figure S18-21 for a full panel of PCR reactions.

[0032] FIG. 4 (S1) illustrates PERV pol consensus sequence and gRNA design.

[0033] FIG. 5 (S2) is a schematic of CRISPR/Cas9 construct targeting PERVs.

[0034] FIG. 6 (S3) illustrates measurement of Cas9-gRNAs activity.

[0035] FIG. 7 (S4) illustrates optimization of DOX concentration to induce Cas9 expression for PERV targeting.

[0036] FIG. 8 (S5) illustrates time series measurement of Piggybac-Cas9/gRNAs PERV targeting efficiencies.

[0037] FIG. 9 (S6) illustrates time series measurement of Lenti-Cas9/2gRNAs PERV targeting efficiency.

[0038] FIG. 10 (S7) illustrates Sanger sequencing validation of PERV targeting efficiency and indel patterning. (SEQ ID NO:29)

[0039] FIG. 11 (S8) illustrates repeated the gene editing experiment.

[0040] FIGS. 12A-B (S9) illustrates PERV pol targeting efficiency of single cells.

[0041] FIG. 13 (S10) illustrates phylogeny of PERV haplotypes.

[0042] FIG. 14 (S11) illustrates distribution of pol gene disruption.

[0043] FIGS. 15A-15B (S12) illustrate karyotype analysis of highly and lowly modified PK15 clones.

[0044] FIG. 16 (S13) illustrates Summary of karyotype analysis of PK15 clones.

[0045] FIG. 17 (S14) illustrates Karyotype nomenclature.

[0046] FIG. 18 (S15) illustrates Detection of PERV reverse transcriptase activity.

[0047] FIG. 19 (S16) illustrates an experimental design to detect the transmission of PERVs to human cells.

[0048] FIGS. 20A-20C (S17) illustrate quality control of the purified HEK293-GFP cells by FACS.

[0049] FIGS. 21A-21D (S18) illustrates detection of pig cell contamination in HEK293 cells using pig GGTA1primers.

[0050] FIGS. 22A-22D (S19) illustrates detection of PERV DNA elements in HEK293 cells using PERV pol primers.

[0051] FIGS. 23A-23D (S20) illustrates detection of PERV DNA elements in HEK293 cells using PERV env primers.

[0052] FIGS. 24A-24D (S21) illustrates detection of PERV DNA elements in HEK293 cells using PERV gag primers.

[0053] FIGS. 25A-25B (S22) illustrates Cas9/2gRNAs expression levels in highly and lowly modified clones.

[0054] FIG. 26 (S23) illustrates principle component analysis of highly and lowly modified PK15 clones.

[0055] FIGS. 27A-27B (S24) illustrate gene set enrichment analysis.

[0056] FIG. 28 (S25) illustrates indel composition analysis and comparison among highly modified clones.

[0057] FIGS. 29A-29D (S26) illustrates a Markov model analysis of DNA repair processes leading to Cas9 elimination of active PERV elements.

[0058] FIG. 30 (S27) illustrates off-target analysis using Whole Genome Sequencing (WGS).

DETAILED DESCRIPTION OF THE INVENTION

[0059] Aspects of the present invention are directed to the use of CRISPR/Cas9, for nucleic acid engineering. Described herein is the development of an efficient technology for the generation of animals (e.g., pigs) carrying multiple mutated genes. Specifically, the clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR associated genes (Cas genes), referred to herein as the CRISPR/Cas system, has been adapted as an efficient gene targeting technology e.g., for multiplexed genome editing. Demonstrated herein is that CRISPR/Cas mediated gene editing allows the simultaneous inactivation of 62 copies of the porcine endogenous retrovirus (PERV) pol gene in a porcine kidney epithelial cell line (e.g., PK15) with high efficiency. Co-injection or transfection of Cas9 mRNA and guide RNA (gRNA) targeting PERVs into cells generated a greater than 1000 fold reduction in PERV transmission to human cells with biallelic mutations in both genes with an efficiency of up to 100%. Shown herein is that the CRISPR/Cas system allows the one step generation of cells carrying inactivation of all copies of PERV. In certain embodiments a method described herein generates cell and animals, e.g., pigs, with inactivation of 1, 2, 3, 4, 5, or more genes with an efficiency of between 20% and 100%, e.g., at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 85%, 90%, 95%, or more, e.g., up to 96%, 97%, 98%, 99%, or more.

Exemplification

Example 1

Genome-Wide Inactivation of Porcine Endogenous Retroviruses (PERVs)

[0060] The shortage of organs for transplantation is a major barrier to the treatment of organ failure. While porcine organs are considered promising, their use has been checked by concerns about transmission of porcine endogenous retroviruses (PERVs) to humans. Here, the eradication of all PERVs in a porcine kidney epithelial cell line (PK15) was performed. It was first determined the PK15 PERV copy number to be 62. Using CRISPR-Cas9, all 62 copies of the PERV pol gene were disrupted and demonstrated a >1000-fold reduction in PERV transmission to human cells using our engineered cells. This study showed that CRISPR-Cas9 multiplexability can be as high as 62 and demonstrates the possibility that PERVs can be inactivated for clinical application to porcine-to-human xenotransplantation.

[0061] Pig genomes contain from a few to several dozen copies of PERV elements. Unlike other zoonotic pathogens, PERVs cannot be eliminated by biosecure breeding. Prior strategies for reducing the risk of PERV transmission to humans have included small interfering RNAs (RNAi), vaccines, and PERV elimination using zinc finger nucleases and TAL effector nucleases, but these have had limited success. Here, the successful use of the CRISPR-Cas9 RNA-guided nuclease system can be used to inactivate all copies of the PERV pol gene and effect a 1000-fold reduction of PERV infectivity of human cells.

[0062] To design Cas9 guide RNAs (gRNAs) that specifically target PERVs, the sequences of publically available PERVs and other endogenous retroviruses in pigs (Methods) were analyzed. A distinct Glade of PERV elements (FIG. 1A) were identified and determined there to be 62 copies of PERVs in PK15 cells using droplet digital PCR (FIG. 1B). Two Cas9 guide RNAs (gRNAs) were designed that targeted the highly conserved catalytic center of the pol gene on PERVs (FIG. 1C, Fig. S1). The pol gene product functions as a reverse transcriptase (RT) and is thus essential for viral replication and infection. It was determined that these gRNAs targeted all PERVs but no other endogenous retrovirus or other sequences in the pig genome (Methods).

[0063] Initial experiments showed inefficient PERV editing when Cas9 and the gRNAs were transiently transfected (Fig. S2). Thus a PiggyBac transposon system was used to deliver a doxycycline-inducible Cas9 and the two gRNAs into the genome of PK15 cells (Fig. S2-3). Continuous induction of Cas9 led to increased targeting frequency of the PERVs (Fig. S5), with a maximum targeting frequency of 37% (˜23 PERV copies per genome) observed on day 17 (Fig. S5). Neither higher concentrations of doxycycline or prolonged incubation increased targeting efficiency (Fig. S4,5), possibly due to the toxicity of non-specific DNA damage by CRISPR-Cas9. Similar trends were observed when Cas9 was delivered using lentiviral constructs (Fig. S6). The cell lines that exhibited maximal PERV targeting efficiencies were genotyped. 455 different insertion and deletion (indel) events centered at the two gRNA target sites (FIG. 2B) was observed. Indel sizes ranged from 1 to 148 bp; 80% of indels were small deletions (<9 bp). The initial deep sequencing results was validated with Sanger Sequencing (Fig. S7).

[0064] Single cells from PK15 cells with high PERV targeting efficiency were sorted using flow cytometry and genotyped the pol locus of the resulting clones via deep sequencing. A repeatable bimodal (FIG. 2A, S8-9) distribution was observed with ˜10% of the clones exhibiting high levels of PERV disruption (97%-100%), and the remaining clones exhibiting low levels of editing (<10%). Individual indel events were examined in the genomes of these clones (FIG. 2B, Fig. S10-11). For the highly edited clones (Clone 20, 100%; Clone 15, 100%; Clone 29, 100%; Clone 38, 97.37%), only 16-20 unique indel patterns in each clone (FIG. 2B, S11) were observed. In addition, there was a much higher degree of repetition of indels within each clone than across the clones (Fig. S25), suggesting a mechanism of gene conversion in which previously mutated PERV copies were used as templates to repair wild-type PERVs cleaved by Cas9 (FIG. 2B, Fig. S25). Mathematical modeling of DNA repair during PERV elimination (Fig. S26) and analysis of expression data (Fig. S22-24) supported this hypothesis and suggested that highly edited clones were derived from cells in which Cas9 and the gRNAs were highly expressed.

[0065] Next, unexpected genomic rearrangements had occurred as a result of the multiplexed genome editing was examined. Karyotyping of individual modified clones (Fig. S12-S14) indicated that there were no observable genomic rearrangements. 11 independent genomic loci with at most 2 bp mismatches to each of the intended gRNA targets were examined and observed no non-specific mutations (Fig. S27). This suggests that our multiplexed Cas9-based genome engineering strategy did not cause catastrophic genomic instability.

[0066] Last, disruption of all copies of PERV pol in the pig genome could eliminate in vitro transmission of PERVs from pig to human cells was examined. No detection of RT activity in the cell culture supernatant of the highly modified PK15 clones (Fig. S15) was observed, suggesting that modified cells only produced minimal amounts of PERV particles. Co-culture of WT and highly modified PK15 cells with HEK 293 cells were tested directly for transmission of PERV DNA to human cells. After co-culturing PK15 WT and HEK 293 cells for 5 days and 7 days (Fig. S16-17), PERV pol, gag, and env sequences in the HEK 293 cells were detected (FIG. 3A). The estimated frequency of PERV infection was approximately 1000 PERVs/1000 human cells (FIG. 3B). However, PK15 clones with >97% PERV pol targeting exhibited up to 1000-fold reduction of PERV infection, similar to background levels (FIG. 3C). These results were validated with PCR amplification of serial dilutions of HEK293 cells that had a history of contact with PK15 clones (FIG. 3D, S18-21). PERVs in single HEK293 cells isolated from the population co-cultured with minimally modified Clone 40 was consistently detected, but could not distinctly detect PERVs in 100 human cells from the population co-cultured with highly modified Clone 20. Thus, PERV infectivity of the engineered PK15 cells had been reduced by up to 1000 fold.

[0067] In summary, it was successfully targeted the 62 copies of PERV pol in PK15 cells and demonstrated greatly reduced in vitro transmission of PERVs to human cells. While in vivo PERV transmission to humans has not been demonstrated, PERVs are still considered risky and our strategy could completely eliminate this. As no porcine embryonic stem cells exist, this system will need to be recapitulated in primary porcine cells and cloned into animals using somatic cell nuclear transfer. Moreover, simultaneous Cas9 targeting of 62 loci in single pig cells without salient genomic rearrangement was achieved. To our knowledge, the maximum number of genomic sites previously reported to be simultaneously edited has been six. Our methods thus open the possibility of editing other repetitive regions of biological significance.

Example 2

Methods

[0068] PERV copy number quantification: Droplet Digital PCR™ PCR (ddPCR™) was usd to quantify the copy number of PERVs according to the manufacturer's instructions (Bio-Rad). Briefly, genomic DNA (DNeasy Blood & Tissue Kit, Qiagen) from cultured cells was purified, digested 50 ng genomic DNA with MseI (10 U) at 37° C. for 1 hour, and prepared the ddPCR reaction with 10 μl 2× ddPCR Master mix, 1 μl of 18 μM target primers & 5 μM target probe (VIC), 1 μl of 18 μM reference primers & 5 μM reference probe (FAM), 5 ng digested DNA, and water to total volume of 20 μl. The sequence of the primers and the probe information can be found in Extended Data Table 1.

Methods

[0069]

TABLE-US-00001 TABLE 1 Primers used in ddPCR assay Name Sequence PrimerPol1-FW CGACTGCCCCAAGGGTTCAA (SEQ ID NO: 1) PrimerPo12-FW CCGACTGCCCCAAGAGTTCAA (SEQ ID NO: 2) PrimerPol-RV TCTCTCCTGCAAATCTGGGCC (SEQ ID NO: 3) ProbePol /56FAM/CACGTACTGGAGGAGGGTCACCTG (SEQ ID NO: 4) Primerpig_actin_F Taaccgatcctttcaagcattt (SEQ ID NO: 5) Primerpig_actin_R Tggtttcaaagcttgcatcata (SEQ ID NO: 6) Probepig_actin /5Hex/cgtggggatgcttcctgagaaag (SEQ ID NO: 7) Primerpig_GAPDH_F Ccgcgatctaatgttctctttc (SEQ ID NO: 8) Primerpig_GAPDH_R Ttcactccgaccttcaccat (SEQ ID NO: 9) Probepig_GAPDH /5Hex/cagccgcgtccctgagacac (SEQ ID NO: 10)

[0070] CRISPR-Cas9 gRNAs design: MUSCLE was used to carry out a multiple sequence alignment of 245 endogenous retrovirus found in the porcine genome. A phylogenetic tree of the sequences was built and identified a Glade that included the PERVs (see FIG. 1a). The R library DECIPHER was used to design specific gRNAs that target all PERVs but no other endogenous retroviral sequences.

[0071] Cell culture: PK15 were maintained in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) high glucose supplemented with 10% fetal bovine serum (Invitrogen), and 1% penicillin/streptomycin (Pen/Strep, Invitrogen). All cells were maintained in a humidified incubator at 37° C. and 5% CO.sub.2.

[0072] PiggyBac-Cas9/2gRNAs construction and cell line establishment: PiggyBac-Cas9/2gRNAs construct is derived from a plasmid previously reported in Wang et al (2). Briefly, a DNA fragment encoding U6-gRNA1-U6-gRNA2 was synthesized (Genewiz) and incorporated it into a PiggBac-Cas9 construct. To establish PK15 cell lines with PiggyBac-Cas9/2gRNAs integration, 5.Math.10.sup.5 PK15 cells was transfected with 4 μg PiggyBac-Cas9/2gRNAs plasmid and 1 μg Super PiggyBac Transposase plasmid (System Biosciences) using Lipofectamine 2000 (Invitrogen). To enrich for the cells carrying the integrated construct, 2 μg/mL puromycin was added to the transfected cells. Based on the negative control, puromycin was applied to wild type PK15 cells, it was determined that the selection completed in 3 days. The PK15-PiggyBac cell lines were maintained with 2 μg/mL puromycin hereafter. 2 μg/ml doxycycline was applied to induce Cas9 expression.

[0073] Lentivirus-Cas9/2gRNAs construction and cell line establishment: Lenti-Cas9/2gRNAs constructs were derived from a plasmid previously reported (3). A DNA fragment encoding U6-gRNA1-U6-gRNA2 was synthesized (Genewiz) and incorporated it into a Lenti-Cas9-V2. To generate lentivirus carrying Lenti-Cas9/2gRNAs, ˜5.Math.10.sup.6 293FT HEK cells was transfected with 3 μg Lenti-Cas9-gRNAs and 12 μg ViraPower Lentiviral Packaging Mix (Invitrogen) using Lipofectamine 2000. The lentiviral particles were collected 72 hours after transfection, and the viral titer was measured using Lenti-X GoStix (Takara Clonetech). ˜10.sup.5 lentiviral particles to ˜1.Math.10.sup.6 PK15 cells were transduced and conducted selection by puromycin to enrich transduced cells 5 days after transduction. The PK15-Lenti cell lines were maintained with 2 μg/mL puromycin thereafter.

[0074] Genotyping of colonized and single PK15 cells: PK15 cultures were dissociated using TrypLE (Invitrogen) and resuspended in PK15 medium with the viability dye ToPro-3 (Invitrogen) at a concentration of 1-2.Math.10.sup.5 cells/ml. Live PK15 cells were single-cell sorted using a BD FACSAria II SORP UV (BD Biosciences) with 100 mm nozzle under sterile conditions. SSC-H versus SSC-W and FSC-H versus FSC-W doublet discrimination gates and a stringent ‘0/32/16 single-cell’ sorting mask were used to ensure that one and only one cell was sorted per well. Cells were sorted in 96-well plates with each well containing 100 μl PK15 medium. After sorting, plates were centrifuged at 70 g for 3 min. Colony formation was seen 7 days after sorting and genotyping experiment was performed 2 weeks after FACS.

[0075] To genotype single PK15 cells without clonal expansion, the PERV locus was directly amplified from sorted single cells according to a previously reported single cell genotyping protocol (4). Briefly, prior to sorting, all plastics and non-biologic buffers were treated with UV radiation for 30 min. Single cells were sorted into 96-well PCR plates with each well carrying 0.5 μl 10× KAPA express extract buffer (KAPA Biosystems), 0.1 μl 1 U/μl KAPA Express Extract Enzyme and 4.6 μl water. The lysis reaction was incubated at 75° C. for 15 min and inactivated the reaction at 95° C. for 5 min. All reactions were then added to 25 μl PCR reactions containing 12.5 μl 2× KAPA 2G fast (KAPA Biosystems), 100 nM PERV illumina primers (Methods Table 2), and 7.5 μl water. Reactions were incubated at 95° C. for 3 min followed by 25 cycles of 95° C., 10 s; 65° C., 20 s and 72° C., 20 s. To add the Illumina sequence adaptors, 5 μl of reaction products were then added to 20 μl of PCR mix containing 12.5 ml of 2 KAPA HIFI Hotstart Readymix (KAPA Biosystems), 100 nM primers carrying Illumina sequence adaptors and 7 μl water. Reactions were incubated at 95° C. for 5 min followed by 15-25 cycles of 98° C., 20 s; 65° C., 20 s and 72° C., 20 s. PCR products were checked on EX 2% gels (Invitrogen), followed by the recovery of 300-400 bp products from the gel. These products were then mixed at roughly the same amount, purified (QIAquick Gel Extraction Kit), and sequenced with MiSeq Personal Sequencer (Illumina). Deep sequencing data was analyzed and determined the PERV editing efficiency using CRISPR-GA (5).

TABLE-US-00002 TABLE 2 Primers used in the PERV pol genotyping Name Sequence illumina_primerPol1 ACACTCTTTCCCTACACGACGCTCTTCCGATCTCGACTGCCCC AAGGGTTCAA (SEQ ID NO: 11) illumina_primerPol2 ACACTCTTTCCCTACACGACGCTCTTCCGATCTCCGACTGCCC CAAGAGTTCAA (SEQ ID NO: 12) illumina_primerPo3 GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCTTCTCTCCTG CAAATCTGGGCC (SEQ ID NO: 13)

[0076] Targeting efficiency estimation: a custom pipeline was built to estimate the efficiency of PERV inactivation. Briefly, the pol gene was amplified and sequenced via Illumina Next Generation Sequencing using PE250 or PE300. First, the two overlapping reads were combined using PEAR (6) and mapped to the reference region using BLAT. After mapping, the reads were grouped into sets containing specific combinations of haplotypes (see Extended Data FIG. 7), and indel types. Read sets with representation lower than 0.5% of the total number of mapped reads were discarded. Finally, the mapping output was parsed to call the different insertions and deletions as described in Güell et al (5).

[0077] RNA-seq analysis: The susScr3 pig genome and Ensembl transcripts were obtained from the UCSC Genome Brower Database. RNA-Seq reads were mapped to the reference genome using the STAR software (7) and the RPKM of the transcripts were quantified using BEDTools (8). Differential expression analysis was performed in R using the DESeq2 package (9), and gene set enrichment analysis was carried out by the GSEA software (10), with gene set definitions obtained from the software's website.

[0078] Reverse transcriptase (RT) assay: To test the RT activity of the PK15 cells and modified PK15 clones (4 highly and 1 lowly modified clones), 5.Math.10.sup.5 cells were plated in T75 cm.sup.2 flasks, and collected the supernatant 4 days after seeding. The media was filtered using a 0.45 μM Millex-HV Syringe Filter (EMD Millipore Corporation), and the filtered supernatant was concentrated at 4000 g for 30 min using Amicon Ultra-15 Centrifugal Filter Unit (EMD Millipore Corporation). The concentrated supernatant was ultra-centrifuged at 50,000 rpm for 60 min. The supernatant was carefully removed, and the virus pellet was collected and lysed with 20 μl of 10% NP40 at 37° C. for 60 min.

[0079] The RT reaction was conducted using the Omniscript RT Kit (Qiagen). The total volume of the reaction was 20 μl, which contained 1□ RT buffer, 0.5 mM dNTPs, 0.5 μM Influenza reverse primer (5′ CTGCATGACCAGGGTTTATG 3′) (SEQ ID NO:14), 100 units of RnaseOUT (Life Technology, Invitrogen), 100 units of SuperRnase Inhibitor (Life Technologies), 5 μl of sample lysis and 40 ng of IDT-synthesized Influenza RNA template which was rnase resistant in both 5′ and 3′ end. The RNA template sequence was 5′ rA*rA*rC*rA*rU*rGrGrArArCrCrUrUrUrGrGrCrCrCrUrGrUrUrCrArUrUrUrUrArGrArArAr UrCrArArGrUrCrArArGrArUrArCrGrCrArGrArArGrArGrUrArGrArCrArUrArArArCrCrCrUr GrGrUrCrArUrGrCrArGrArCrCrU*rC*rA*rG*rU*rG 3′ (* phosphodiester bond) (SEQ ID NO:15). After the RT reaction was completed, the RT product was examined by PCR using Influenza forward (5′ ACCTTTGGCCCTGTTCATTT 3′) (SEQ ID NO:16) and Influenza reverse primers (sequence shown as above). The expected size of the amplicon was 72 bp.

Infectivity Assay

[0080] HEK293-GFP cell line establishment: The Lenti-GFP construct was derived from the plasmid pLVX-IRES-ZsGreen1 (Clontech. Catalog No. 632187; PT4064-5). To generate the lentivirus carrying Lenti-GFP, ˜5.Math.10.sup.6 293FT HEK cells were transfected with 3 μg of pVX-ZsGreen plasmid and 12 μg of ViraPower Lentiviral Packaging Mix (Invitrogen) using Lipofectamine 2000 (Invitrogen). Lentiviral particles were collected 72 hours after transfection, and the viral titer was measured using Lenti-X GoStix (Takara Clonetech). ˜10.sup.5 lentivirus particles to ˜1.Math.10.sup.6 HEK293 cells were transfected and conducted selection by puromycin to enrich the transduced cells 5 days after transduction. The 293-GFP-Lenti cell lines were maintained with 0.5 μg/mL puromycin thereafter.

[0081] Infectivity test of PK15 WT to HEK293-GFP: 1.Math.10.sup.5 cells of Lenti-GFP-293FT HEK cells and 1.Math.10.sup.5 PK15 WT cells were cultured together in a 6-well plate. In parallel, 2.Math.10.sup.5 PK15 WT cells were cultured alone in another well as a control. The puromycin selection experiment was done by adding 5 μg/ml of the antibiotic for 7 days. The time point was determined when no viable cells in the control well and approximately 100% GFP positive cells in the experimental well as the time point when the puromycin selection was completed to purify lenti-GFP-293FT human cells. Cells from the 293FT HEK/PK15 WT co-culture were collected at different time periods. The genomic DNA was extracted using (DNeasy Blood & Tissue Kit, Qiagen) from cultured cells of the 293-GFP WT, PK15 WT and the co-cultured cells. The genomic DNA concentration was measured using a Qubit 2.0 Fluorometer (Invitrogen), and 3 ng from each sample was used as DNA template for PCR. In all, 1 μL of the genomic DNA were added to 25 μL of a PCR mix containing 12.5 μL 2× KAPA Hifi Hotstart Readymix (KAPA Biosystems) and 100 μM of primers as listed in Methods Table 3. Reactions were incubated at 95° C. for 5 min followed by 35 cycles of 98° C., 20 s; 65° C., 20 s and 72° C., 20 s. PCR products were visualized on EX 2% gels (Invitrogen) and observed for bands of 300-400 base pairs.

TABLE-US-00003 TABLE 3 A table exhibiting the primers used in the  infectivity assay Name Sequence PERV pol-Forward GGG AGT GGG ACG GGT AAC CCA (SEQ ID NO: 17) PERV pol-Reverse GCC CAG GCT TGG GGA AAC TG (SEQ ID NO: 18) PERV env-Forward ACC TCT TCT TGT TGG CTT TG (SEQ ID NO: 19) PERV env-Reverse CAA AGG TGT TGG TGG GAT GG (SEQ ID NO: 20) PERV gag-Forward CGC ACA CTG GTC CTT GTC GAG (SEQ ID NO: 21) PERV gag-Reverse TGA TCT AGT GAG AGA GGC AGA G (SEQ ID NO: 22) Pig GGTA1-Forward GGA GCC CTT AGG GAC CAT TA (SEQ ID NO: 23) Pig GGTA1-Reverse GCG CTA AGG AGT GCG TTC TA (SEQ ID NO: 24) Human ACTB-Forward GCC TTC CTT CCT GGG CAT GG (SEQ ID NO: 25) Human ACTB-Reverse GAG TAC TTG CGC TCA GGA GG (SEQ ID NO: 26)

[0082] Quantification of PERV copy numbers infected in HEK293-GFP cells: qPCR was performed to quantify the PERV copy number in HEK293-GFP cells. Genomic DNA of PK15 WT cells of different amounts was used as the template for the qPCR reactions. Reactions were conducted in triplicate using KAPA SYBR FAST qPCR Master Mix Universal (KAPA Biosystems). PERV pol, env, gag primers, human ACTB and pig GGTA1 primers (Methods Table 3) were added to a final concentration of 1 μM. Reactions were incubated at 95° C. for 3 min (enzyme activation) followed by 50 cycles of 95° C., 5 s (denaturation); 60° C., 60 s (annealing/extension). The logarithm of the genomic DNA amount linearizes with the quantification cycle (Cq). pol, gag, env primers were used to examine for presence of PERVs. Pig GGTA1 primers served to control for potential porcine genome contaminants in human cells after infection. All experiments were conducted in triplicate.

[0083] Infectivity Assay of the Modified PK15 clones to HEK293-GFP: 1.Math.10.sup.5 cells of HEK293-GFP cells and 1.Math.10.sup.5 cells of the high modified (15, 20, 29, 38) clones and low modified clones (40, 41) were co-cultured in a 6-well plate for 7 days. To isolate the HEK293-GFP cells in order to examine for PERV elements, the GFP positive cells were double sorted to purify the human cell populations.

[0084] To quantify the PERV infectivity of different clones to HEK293-GFP cells, both qPCR assays and PCR assays were conducted on series diluted HEK293-GFP cells after sorting. For the qPCR assays, the genomic DNA (DNeasy Blood & Tissue Kit, Qiagen) was extracted from double sorted HEK293-GFP cells. The genomic DNA concentration was measured using Qubit 2.0 fluorometer (Invitrogen). In all, 3 ng of the genomic DNA was added to 20 μL of KAPA SYBR FAST qPCR reaction (KAPA Biosystems) using PERV pol, env, gag and pig GGTA primers respectively (Extended Data Table 2). The qPCR procedure was performed as described above. For the series dilution assay, purified HEK293-GFP cells were sorted (1 cell/ well, 10 cells/well, 100 cells/well, 1000 cells/well) into 96-well PCR plates for direct genomic DNA extraction and PCR reactions. Briefly, cells were sorted into 20 μL lysis reaction including 2 μL of 10× KAPA Express Extract Buffer, 0.4 μL of 1 U/μl KAPA Express Extract Enzyme and 17.6 μL of PCR-grade water (KAPA Biosystems). The reactions were then incubated at 55° C. for 10 min (lysis), then at 95° C. for 5 min (enzyme inactivation). Subsequently, the PCR master mix was prepared. In all, 2 μL of the genomic DNA lysis was added to 4 different 25 μL of KAPA Hifi Hotstart PCR reactions (KAPA Biosystems) using 1 μM PERV pol, env, gag primers, and pig GGTA primers, respectively (Extended Data Table 2). The reactions were incubated at 95° C. for 3 min (initial denaturation) followed by 35 cycles of 95° C., 15 s (denaturation); 60° C., 15 s (annealing), 72° C., 15 sec/kb, then 75° C., 1 min/kb (final extension). (KAPA Biosystems). The PCR products were visualized on 96 well E-Gel® Agarose Gels, SYBR® Safe DNA Gel (Invitrogen).

[0085] CRISPR-Cas9 off-target analysis: whole genome sequencing (WGS) data was obtained for PK15 (untreated cell line) and clone 20 (highly edited clone). To investigate potential off-target effects of the Cas9/2gRNAs, the reference sequence (Sus Scrofa 10.2) was searched for sites that differed from the 20 bp sequences targeted by the two gRNAs by only 1 or 2 bp. 11 such sites were identified and extracted them, together with 200 bp of their neighboring regions (Fig. S1). BLAT was used to map the WGS reads to the extracted reference sequences and searched for potential indel patterns that had emerged in Clone 20 as a result of off-target effects. An average coverage of 7-8 X per loci was obtained. Reads with <50 bp matches with the reference sequence were excluded. In case of reads that mapped to the reference sequence with multiple alignment blocks, which could indicate the presence of indels, reads whose alignment blocks contained <20 bp matches were excluded with the reference sequence. After inspecting the remaining mapped reads, there was no detection any off-target indel patterns present in clone 20. Another challenge was to comprehensive searches for off-targets here is that the Sus Scrofa genome is still neither complete nor completely assembled, limiting the ability to do whole-genome analysis.

[0086] Mathematical model of DNA repair process interaction during cumulative PERV inactivation: In this study PERV elements were inactivated by mutations generated by DNA repair processes in response to dsDNA cuts created by Cas9. It is generally understood that dsDNA cuts may be repaired either by non-homologous end joining (NHEJ) or Homologous Repair (HR), and that while HR can create precise copies of a DNA template sequence at the cut site given the presence of a template with suitable homology arms, NHEJ can generate mutations (especially indels) and is often considered “error prone.” However, there is also evidence that NHEJ can also repair dsDNA cuts highly accurately (11, 12), and the relative rates of mutated vs. perfect repair by NHEJ have never been precisely measured. Especially when efficient targeted nucleases such as Cas9 are expressed for protracted time periods, perfect repair of a cut site by either NHEJ or HR would regenerate a target site that could be cut again. A plausible hypothesis is that the process of perfect repair and re-cutting would occur repeatedly until a mutation arose that destroyed the nuclease's ability to recognize the target site. To explore the way these repair modalities might work together during the course of PERV elimination, their interactions as a Markov process was modeled. Specifically, it was assumed: [0087] There are N identical copies of the nuclease target in a cell. [0088] Only wild-type targets are recognized and cut, and only one target is cut and repaired at a time. [0089] DNA repair is either [0090] perfect restoration of the target site by NHEJ (with probability n) [0091] NHEJ that results in generation of a mutation that ablates target recognition (with probability m) [0092] repair by HR using any one of the other N−1 target sequences in the cell (with probability h)
Thus, n+m+h=1.

[0093] The Markov model computes the probability distribution P.sup.(c)=(p.sub.0.sup.(c),p.sub.1.sup.(c), . . . , p.sub.N.sup.(c)), where p.sub.i.sup.(c) is the probability that there are i target-ablating mutations at cut c, where c=0, 1, 2 . . . . It is assumed that the initial condition P.sup.(0)=(1,0, . . . , 0), i.e., that all targets begin as wild-type. The N+1-by-N+1 transition matrix M is given as

[00001]$\begin{array}{c}\begin{array}{c}M\left(i,i\right)=n+h.Math.\frac{N-i-1}{N-1}\\ M\left(i,i+1\right)=m+h.Math.\frac{i}{N-1}\end{array}\}.Math..Math.\mathrm{for}.Math..Math.0\le i\le N.Math.\\ M\left(N,N\right)=1.Math..Math.\mathrm{for}.Math..Math.i=N\\ M\left(i,j\right)=0.Math..Math.\mathrm{for}.Math..Math.\mathrm{all}.Math..Math.\mathrm{other}.Math..Math.0\le i,j\le N\end{array}$

Finally, P.sup.(c+1)=P.sup.(c)M for c=0,1,2, . . .

[0094] The formulas for M assume proposition ii above and state in mathematical terms that the number of mutated sites in a cell remains unchanged whenever a cut at a wild-type site is repaired perfectly by NHEJ or by HR using another copy of the wild-type template (formula for M(i, i)), but increases by one if the cut is repaired by mutagenic NHEJ or by HR using a previously mutated site (formula for M(i, i+1)).

[0095] The model incorporates two notable simplifications to actual biology: (i) Target recognition is assumed to be binary—either the nuclease recognizes a target or it does not. This is tantamount to assuming that small mutations that still support target recognition do not substantially alter wild-type cutting rates and therefore can be effectively lumped together with wild-type sites. (ii) HR repairs using mutated vs. wild-type templates are assumed to be equally efficient. Modifications could be made to the model to address these simplifications, but this is not considered here. It is also worth noting that, formally, given assumption ii above, the Markov process should actually stop should the condition p.sub.N.sup.(c)=1 be reached for some value of c, since at this point no wild-type sites remain to be cut, whereas what happens instead mathematically is that cuts continue but the model remains in a fixed state. Finally, the model effectively represents the mutation count distribution as a function of independent variable c (number of cuts) and not as a function of time. No prediction is made regarding the time rates of DNA repair or PERV site elimination, although time can be assumed to increase monotonically with c.

[0096] To analyze PERV elimination through the Markov model, N was always set to 62. However, since the relative efficiencies of perfect vs. mutagenic NHEJ repair are unknown (as noted above), and because relative rates of mutagenic NHEJ vs. HR repair can vary widely depending on cell state and type, the mutation count distributions for a discrete grid covering the complete two-dimensional space of all possible parameter values for n, m, and h, (2500 parameter combinations in all) was computed. The model was implemented both as a MatLab (Mathworks, Waltham) script and as an R script using the library markovchain (available as Supplemental Files modelMarkov.m, modelMarkov.R, respectively).

[0097] In addition to computing the mutation count distribution via the Markov model for particular parameter values, the MatLab script performed random simulations of the NHEJ and HR repair processes throughout a series of K cuts, allowing bivariate distributions of the numbers of total mutations vs. distinct NHEJ events to be estimated, illustrated in FIG. 27 B-C. The R script was used to estimate the most likely state of the system over the grid of n, m, and h combinations described above. K was varied depending on the computation. As illustrated in Fig. S27, the invariable result of the model was a unimodal distribution of mutation counts whose mean advanced towards fixation at N mutations with c, and in Figures S27 B.C, K was set to a value high enough to demonstrate fixation. For the calculation of the most likely state of the system over the n, m, and h grid, K was set to 50, 100, 200, or 500, and 100 simulations were conducted for each parameter combination.

Data Deposition

[0098] Illumina Miseq data with PERVs elements genotyping data has been uploaded to the European Nucleotide Archive (ENA) hosted by the European Bioinformatics Institute (EBI) with the submission reference PRJEB11222.

[0099] Appendix A provide further information regarding various aspects of the present teachings, which is herein incorporated by reference in its entirety.

[0100] The DNA sequence listing further includes genome sequences of multiple endogenous retroviral elements extracted from pig genome sequence and from public sequence databases. (SEQ ID NO:30-280)

[0101] The teachings of all patents, published applications and references cited herein are incorporated by reference in their entirety.

[0102] While this invention has been particularly shown and described with references to example embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.