COMMENTARY | SPECIAL FEATURE
Genome engineering: the next genomic revolution Charles A Gersbach
npg
© 2014 Nature America, Inc. All rights reserved.
A decade of advances in genome engineering technologies has enabled the editing of genome sequences much like one edits computer code; many more applications for precisely manipulating genome structure and function are on the horizon. The genomic revolution of the last 15 years has led to tremendous advances in science, technology and medicine. The sequencing of the human genome, first announced in 2001, and the sequencing of many other species’ genomes, have provided a list of the DNA parts that cells use to perform the myriad functions necessary for life. Subsequent progress in genetics and genomics has further defined how these parts are organized and altered in disease and healthy states. Despite this wealth of valuable information, these data are largely limited to telling us what DNA parts exist in a genome. Without the tools to reach into the genome and manipulate each part individually, as an engineer would do with circuits in a computer or lines of code in a software program, it is not possible to clearly decipher how each of these parts functions in the genome. The ability to engineer the genomic code like one would a computer program makes it possible to determine the function of specific lines of code, create new code or repair broken code. Conventional genetic engineering methods have generally been limited to adding new pieces of DNA, without any control over where this DNA goes within the cell, and have thus not allowed the editing of specific DNA sequences in their natural context. A fundamental breakthrough that suggested it could be possible to target Charles A. Gersbach is in the Department of Biomedical Engineering and at the Center for Genomic and Computational Biology, Duke University, and in the Department of Orthopaedic Surgery, Duke University Medical Center, Durham, North Carolina, USA. e-mail: [email protected]
specific sequences among the billions of base pairs in a complex genome was the crystal structure of the zinc-finger protein ZIF268 in 1991 (ref. 1). The discovery of how this DNA-binding protein targets a particular DNA sequence (Fig. 1a) suggested that altering key residues would enable targeting of the protein to a new DNA sequence. In fact, this marked the beginning of the field of programmable synthetic zincfinger proteins. The fusion of the FokI endonuclease catalytic domain to these DNAbinding proteins made it possible to target the resulting zinc-finger nucleases (ZFNs) to a DNA sequence of choice 2. Because targeted double-strand DNA breaks in the genome can be used to guide DNA repair outcomes and catalyze gene targeting 3,4 (Fig. 1b), this collective work inspired the birth of the genome-editing field. In 2005, ZFN-mediated genome editing was used to correct genetic mutations that cause human disease5. This landmark achievement was the first of many applications of genome editing in science and medicine. Numerous innovations to improve the utility of ZFNs have occurred in the subsequent decade, but the technical challenges and nuances of engineering zinc-finger proteins targeted to new sites have limited their use beyond by experts in the field. In 2009, the process of designing programmable DNA-binding proteins was greatly simplified by the discovery of the DNA recognition code of transcription activator–like effectors (TALEs)6,7, proteins that exist naturally in plant pathogenic bacteria. Each unit of a TALE protein recognizes only a single base pair and appears to be more modular and independent than the complex interactions of each zinc-fin-
ger domain with three or four base pairs (Fig. 1a). Thus, the creation of TALE proteins targeted to new sequences is generally more straightforward and successful. Consequently, fusions of TALEs to the FokI domain were generated shortly thereafter to create TALE nucleases (TALENs) for genome editing8,9. This approach enabled much more widespread application of genome editing but still required specialized molecular biology expertise to design, assemble and validate the genes encoding the large TALE proteins targeted to userdefined sequences. The recent emergence of the CRISPRCas9 technology has transformed the genome engineering field by removing the need for any expertise in protein engineering. Clustered, regularly interspaced, short palindromic repeats (CRISPR) refers to an adaptive immunity system in which bacteria incorporate short sequences of invading viral DNA into their genome as a memory of infection. These arrays of repeats of foreign DNA are transcribed and processed into CRISPR RNAs (crRNAs), which anneal with another short RNA, called the tracrRNA, and form a complex with the Cas9 endonuclease. A cell reexposed to the same foreign DNA will recognize it by complementary base-pairing of the crRNA. Subsequent cleavage by the associated Cas9 enzyme leads to elimination of this invading DNA. In 2012, it was shown that the crRNA and tracrRNA can be combined into a single chimeric guide RNA (gRNA) and that Cas9 and the gRNA are the only components required for targeted nuclease activity 10. Targeting the Cas9-gRNA complex to any new sequence involves only altering a short (~20 base pair) sequence in the gRNA
NATURE METHODS | VOL.11 NO.10 | OCTOBER 2014 | 1009
SPECIAL FEATURE | COMMENTARY a
Zinc-finger protein
b
Transcription activator– like effector
Cas9-gRNA
Target gene
npg
Gene disruption
Genomic deletion
Gene addition or replacement
c Kim Caesar/Nature Publishing Group
© 2014 Nature America, Inc. All rights reserved.
Donor DNA template
Figure 1 | Applications of genome-editing technologies. (a) Structures of a zinc-finger protein (PDB: 1AAY), TALE (3UGM) and CRISPR (4OO8) binding DNA. Left and center: gray, DNA; blue, proteins; red, amino acid side chains responsible for DNA targeting. Right: gray, DNA; blue, Cas9 enzyme; red, gRNA. (b) A nuclease-induced double-strand break can lead to gene knockout, deletion of genomic regions or gene modification. (c) Applications include reverse genetics, disease modeling, agriculture and gene therapy.
(Fig. 1a). Soon thereafter, the system was applied in human cells11–13, zebrafish14 and many other organisms. The simplicity of the CRISPR-Cas9 system has led to its rapid adoption by many laboratories in diverse areas of science. Its robustness across many cell types and species has been evidenced by diverse applications, which have generated numerous landmark achievements in less than 2 years since the first proof-ofprinciple work. Consequently, genome editing is becoming a commoditized technique available to any research laboratory. Collectively, these technologies have created a scientific paradigm that envisions the genome as an infinitely editable piece of software. This has obvious influence on the ability to study gene function: genes can be knocked out, specifically altered for gain of function, modified for conditional
control or tagged for tracking of the protein product. This has also had an immediate and significant impact on reverse genetics. Historically, deciphering the function of a particular human genetic variant on a cellular phenotype was convoluted by the millions of other base pairs that varied between individuals. Genome editing circumvents this problem by allowing the precise alteration of specific base pairs in the context of an unchanged genome15. Combined with genetic reprogramming, these approaches enable drug screening and other studies on diverse cell types that might otherwise be inaccessible from human patients, such as neurons, cardiomyocytes and hepatocytes16. These methods can also be used to directly manipulate DNA regulatory elements to decipher relationships between genome structure and function. Furthermore,
1010 | VOL.11 NO.10 | OCTOBER 2014 | NATURE METHODS
these tools can be implemented in highthroughput genetic screens17,18. In contrast to RNA interference screens that are often confounded by off-target effects and incomplete knockdown of target genes, libraries of nucleases for functional gene knockout may result in more robust and selectable phenotypes. Beyond the editing of cultured cells, these technologies have broad application in creating genetically modified organisms (Fig. 1c). Genome editing can be achieved in whole organisms by direct injection of editing nucleases, or of DNA or mRNA encoding the nucleases, into fertilized eggs. This process thus circumvents many of the steps for traditional generation of transgenic organisms. Many different species have already been genetically modified with these methods, including mice, rats, monkeys, pigs, cows, rabbits, frogs, zebrafish, fruit flies, worms, yeast, bacteria and others. These organisms have use in basic science for studies of gene function, genetics, genomics and disease modeling—but also in agriculture and in medicine, for example, by serving as hosts for allogeneic tissues and organs. Genome editing in plants to introduce a variety of agriculturally desirable properties is becoming a major industry. Interestingly, because editing can be used to manipulate plant genomes without the transfer of genes from other species, similarly to natural mechanisms of selective plant breeding, plants modified by these tools may not be classified as genetically modified organisms by some federal regulatory agencies. Finally, the greatest potential for genome editing to directly affect human health is arguably in gene therapy, where the benefits of targeted gene editing are manifold. These include the ability to knock out endogenous genes that promote disease progression or are otherwise detrimental, to target therapeutic transgenes to well-characterized ‘safe harbor’ loci to prevent disruption of endogenous gene regulation and to correct disease-causing mutations. Following the first proof-of-principle demonstration of correction of mutations that cause a human genetic disease5, ZFNs have now entered clinical trials to disrupt the CCR5 gene used by HIV to enter T cells19. Efforts for the clinical translation of TALENs and CRISPR-Cas9 are under way as well. All three genomeediting technologies have been applied to preclinical models of many diverse diseases,
npg
© 2014 Nature America, Inc. All rights reserved.
COMMENTARY | SPECIAL FEATURE including viral infection, cancer immunotherapy, sickle-cell disease, hemophilia, Duchenne muscular dystrophy and various immunodeficiencies, just to name a few examples. Notably, the recent demonstrations of efficient gene editing in vivo20,21 and in human hematopoietic stem cells22 are indicative of the widespread therapeutic applicability of genome editing. Importantly, there are both common and unique challenges to each of the genomeediting platforms. A primary concern is their specificity and precision in making exact sequence changes. Off-target activity at unknown genomic locations has the potential to convolute scientific results or cause adverse effects. Tremendous advances have been made in improving the specificity of each technology, but the field is still in dire need of methods that are more sensitive and comprehensive in determining gene-editing specificity for each individual cell of a modified population. Typically, only a minority of the treated cells are actually edited, and thus there is also a major emphasis on increasing nuclease activity and developing methods to select for correctly modified cells. Delivery of the nucleases is another challenge, particularly for in vivo applications, and the large size of the genes encoding Cas9 and TALENs is a hurdle that is currently being addressed. Advances in engineering of viral vectors and nanoparticles, as well as strategies for direct delivery of nuclease proteins, are likely to play a significant role in this work moving forward. Finally, issues such as immunogenicity of gene-editing tools have largely been unreported thus far and will be a subject of future interest for clinical applications. Looking ahead, there are numerous nascent technologies that are likely to have a major impact on the genome engineering field in the next few years. Alternative enzy-
matic domains to the FokI- and Cas9-based nucleases have been slower to develop, but the engineering of meganucleases, recombinases, transposases and hybrids of these enzymes fused to zinc-finger proteins, TALEs and the catalytically inactive Cas9 (dCas9) have all seen recent breakthroughs. Building on the effectiveness of the CRISPRCas9 system, other technologies that exploit RNA-DNA and DNA-DNA interactions for target-site recognition may also emerge23. In addition to the editing of DNA sequences, applications of these genome engineering tools that manipulate transcriptional regulation 24–27 and edit the epigenome28,29 will also have a significant impact in the near future. These technologies will be greatly valuable in decoupling the complexity of eukaryotic gene regulation and for applications such as reprogramming gene networks to direct changes in cell phenotype. The availability of orthogonal Cas9 enzymes30 may facilitate this work by enabling independent control of multiple epigenetic modifiers. Other applications of these tools for imaging chromosomal loci, controlling three-dimensional genome structure and generating complex libraries of DNA sequences that have emerged only in the last year are likely to find extensive use. Finally, technologies for dynamically regulating the activity of these genome engineering tools with precise spatiotemporal control, such as optogenetic methods29, will be critical for achieving the ultimate goal of programming any property of the genome, in any cell, at any time. ACKNOWLEDGMENTS I thank the members of the Gersbach laboratory and our collaborators that have contributed to our work in this area. C.A.G. is supported by US National Institutes of Health (NIH) Director’s New Innovator Award (DP2OD008586), US National Science Foundation (NSF) Faculty Early Career
Development (CAREER) Award (CBET-1151035), NIH R01DA036865, NIH R21AR065956 and the Muscular Dystrophy Association (MDA277360). COMPETING FINANCIAL INTERESTS The author declares competing financial interests: details are available in the online version of the paper (doi:10.1038/nmeth.3113). 1. Pavletich, N.P. & Pabo, C.O. Science 252, 809– 817 (1991). 2. Smith, J. et al. Nucleic Acids Res. 28, 3361–3369 (2000). 3. Bibikova, M., Beumer, K., Trautman, J.K. & Carroll, D. Science 300, 764 (2003). 4. Porteus, M.H. & Baltimore, D. Science 300, 763 (2003). 5. Urnov, F.D. et al. Nature 435, 646–651 (2005). 6. Boch, J. et al. Science 326, 1509–1512 (2009). 7. Moscou, M.J. & Bogdanove, A.J. Science 326, 1501 (2009). 8. Christian, M. et al. Genetics 186, 757–761 (2010). 9. Miller, J.C. et al. Nat. Biotechnol. 29, 143–148 (2011). 10. Jinek, M. et al. Science 337, 816–821 (2012). 11. Mali, P. et al. Science 339, 823–826 (2013). 12. Cong, L. et al. Science 339, 819–823 (2013). 13. Cho, S.W., Kim, S., Kim, J.M. & Kim, J.S. Nat. Biotechnol. 31, 230–232 (2013). 14. Hwang, W.Y. et al. Nat. Biotechnol. 31, 227–229 (2013). 15. Soldner, F. et al. Cell 146, 318–331 (2011). 16. Ding, Q. et al. Cell Stem Cell 12, 238–251 (2013). 17. Wang, T., Wei, J.J., Sabatini, D.M. & Lander, E.S. Science 343, 80–84 (2014). 18. Shalem, O. et al. Science 343, 84–87 (2014). 19. Tebas, P. et al. N. Engl. J. Med. 370, 901–910 (2014). 20. Li, H. et al. Nature 475, 217–221 (2011). 21. Yin, H. et al. Nat. Biotechnol. 32, 551–553 (2014). 22. Genovese, P. et al. Nature 510, 235–240 (2014). 23. Swarts, D.C. et al. Nature 507, 258–261 (2014). 24. Zhang, F. et al. Nat. Biotechnol. 29, 149–153 (2011). 25. Gilbert, L.A. et al. Cell 154, 442–451 (2013). 26. Perez-Pinera, P. et al. Nat. Methods 10, 973–976 (2013). 27. Maeder, M.L. et al. Nat. Methods 10, 977–979 (2013). 28. Maeder, M.L. et al. Nat. Biotechnol. 31, 1137– 1142 (2013). 29. Konermann, S. et al. Nature 500, 472–476 (2013). 30. Esvelt, K.M. et al. Nat. Methods 10, 1116–1121 (2013).
NATURE METHODS | VOL.11 NO.10 | OCTOBER 2014 | 1011
Genome engineering: the next genomic revolution Charles A Gersbach
npg
© 2014 Nature America, Inc. All rights reserved.
A decade of advances in genome engineering technologies has enabled the editing of genome sequences much like one edits computer code; many more applications for precisely manipulating genome structure and function are on the horizon. The genomic revolution of the last 15 years has led to tremendous advances in science, technology and medicine. The sequencing of the human genome, first announced in 2001, and the sequencing of many other species’ genomes, have provided a list of the DNA parts that cells use to perform the myriad functions necessary for life. Subsequent progress in genetics and genomics has further defined how these parts are organized and altered in disease and healthy states. Despite this wealth of valuable information, these data are largely limited to telling us what DNA parts exist in a genome. Without the tools to reach into the genome and manipulate each part individually, as an engineer would do with circuits in a computer or lines of code in a software program, it is not possible to clearly decipher how each of these parts functions in the genome. The ability to engineer the genomic code like one would a computer program makes it possible to determine the function of specific lines of code, create new code or repair broken code. Conventional genetic engineering methods have generally been limited to adding new pieces of DNA, without any control over where this DNA goes within the cell, and have thus not allowed the editing of specific DNA sequences in their natural context. A fundamental breakthrough that suggested it could be possible to target Charles A. Gersbach is in the Department of Biomedical Engineering and at the Center for Genomic and Computational Biology, Duke University, and in the Department of Orthopaedic Surgery, Duke University Medical Center, Durham, North Carolina, USA. e-mail: [email protected]
specific sequences among the billions of base pairs in a complex genome was the crystal structure of the zinc-finger protein ZIF268 in 1991 (ref. 1). The discovery of how this DNA-binding protein targets a particular DNA sequence (Fig. 1a) suggested that altering key residues would enable targeting of the protein to a new DNA sequence. In fact, this marked the beginning of the field of programmable synthetic zincfinger proteins. The fusion of the FokI endonuclease catalytic domain to these DNAbinding proteins made it possible to target the resulting zinc-finger nucleases (ZFNs) to a DNA sequence of choice 2. Because targeted double-strand DNA breaks in the genome can be used to guide DNA repair outcomes and catalyze gene targeting 3,4 (Fig. 1b), this collective work inspired the birth of the genome-editing field. In 2005, ZFN-mediated genome editing was used to correct genetic mutations that cause human disease5. This landmark achievement was the first of many applications of genome editing in science and medicine. Numerous innovations to improve the utility of ZFNs have occurred in the subsequent decade, but the technical challenges and nuances of engineering zinc-finger proteins targeted to new sites have limited their use beyond by experts in the field. In 2009, the process of designing programmable DNA-binding proteins was greatly simplified by the discovery of the DNA recognition code of transcription activator–like effectors (TALEs)6,7, proteins that exist naturally in plant pathogenic bacteria. Each unit of a TALE protein recognizes only a single base pair and appears to be more modular and independent than the complex interactions of each zinc-fin-
ger domain with three or four base pairs (Fig. 1a). Thus, the creation of TALE proteins targeted to new sequences is generally more straightforward and successful. Consequently, fusions of TALEs to the FokI domain were generated shortly thereafter to create TALE nucleases (TALENs) for genome editing8,9. This approach enabled much more widespread application of genome editing but still required specialized molecular biology expertise to design, assemble and validate the genes encoding the large TALE proteins targeted to userdefined sequences. The recent emergence of the CRISPRCas9 technology has transformed the genome engineering field by removing the need for any expertise in protein engineering. Clustered, regularly interspaced, short palindromic repeats (CRISPR) refers to an adaptive immunity system in which bacteria incorporate short sequences of invading viral DNA into their genome as a memory of infection. These arrays of repeats of foreign DNA are transcribed and processed into CRISPR RNAs (crRNAs), which anneal with another short RNA, called the tracrRNA, and form a complex with the Cas9 endonuclease. A cell reexposed to the same foreign DNA will recognize it by complementary base-pairing of the crRNA. Subsequent cleavage by the associated Cas9 enzyme leads to elimination of this invading DNA. In 2012, it was shown that the crRNA and tracrRNA can be combined into a single chimeric guide RNA (gRNA) and that Cas9 and the gRNA are the only components required for targeted nuclease activity 10. Targeting the Cas9-gRNA complex to any new sequence involves only altering a short (~20 base pair) sequence in the gRNA
NATURE METHODS | VOL.11 NO.10 | OCTOBER 2014 | 1009
SPECIAL FEATURE | COMMENTARY a
Zinc-finger protein
b
Transcription activator– like effector
Cas9-gRNA
Target gene
npg
Gene disruption
Genomic deletion
Gene addition or replacement
c Kim Caesar/Nature Publishing Group
© 2014 Nature America, Inc. All rights reserved.
Donor DNA template
Figure 1 | Applications of genome-editing technologies. (a) Structures of a zinc-finger protein (PDB: 1AAY), TALE (3UGM) and CRISPR (4OO8) binding DNA. Left and center: gray, DNA; blue, proteins; red, amino acid side chains responsible for DNA targeting. Right: gray, DNA; blue, Cas9 enzyme; red, gRNA. (b) A nuclease-induced double-strand break can lead to gene knockout, deletion of genomic regions or gene modification. (c) Applications include reverse genetics, disease modeling, agriculture and gene therapy.
(Fig. 1a). Soon thereafter, the system was applied in human cells11–13, zebrafish14 and many other organisms. The simplicity of the CRISPR-Cas9 system has led to its rapid adoption by many laboratories in diverse areas of science. Its robustness across many cell types and species has been evidenced by diverse applications, which have generated numerous landmark achievements in less than 2 years since the first proof-ofprinciple work. Consequently, genome editing is becoming a commoditized technique available to any research laboratory. Collectively, these technologies have created a scientific paradigm that envisions the genome as an infinitely editable piece of software. This has obvious influence on the ability to study gene function: genes can be knocked out, specifically altered for gain of function, modified for conditional
control or tagged for tracking of the protein product. This has also had an immediate and significant impact on reverse genetics. Historically, deciphering the function of a particular human genetic variant on a cellular phenotype was convoluted by the millions of other base pairs that varied between individuals. Genome editing circumvents this problem by allowing the precise alteration of specific base pairs in the context of an unchanged genome15. Combined with genetic reprogramming, these approaches enable drug screening and other studies on diverse cell types that might otherwise be inaccessible from human patients, such as neurons, cardiomyocytes and hepatocytes16. These methods can also be used to directly manipulate DNA regulatory elements to decipher relationships between genome structure and function. Furthermore,
1010 | VOL.11 NO.10 | OCTOBER 2014 | NATURE METHODS
these tools can be implemented in highthroughput genetic screens17,18. In contrast to RNA interference screens that are often confounded by off-target effects and incomplete knockdown of target genes, libraries of nucleases for functional gene knockout may result in more robust and selectable phenotypes. Beyond the editing of cultured cells, these technologies have broad application in creating genetically modified organisms (Fig. 1c). Genome editing can be achieved in whole organisms by direct injection of editing nucleases, or of DNA or mRNA encoding the nucleases, into fertilized eggs. This process thus circumvents many of the steps for traditional generation of transgenic organisms. Many different species have already been genetically modified with these methods, including mice, rats, monkeys, pigs, cows, rabbits, frogs, zebrafish, fruit flies, worms, yeast, bacteria and others. These organisms have use in basic science for studies of gene function, genetics, genomics and disease modeling—but also in agriculture and in medicine, for example, by serving as hosts for allogeneic tissues and organs. Genome editing in plants to introduce a variety of agriculturally desirable properties is becoming a major industry. Interestingly, because editing can be used to manipulate plant genomes without the transfer of genes from other species, similarly to natural mechanisms of selective plant breeding, plants modified by these tools may not be classified as genetically modified organisms by some federal regulatory agencies. Finally, the greatest potential for genome editing to directly affect human health is arguably in gene therapy, where the benefits of targeted gene editing are manifold. These include the ability to knock out endogenous genes that promote disease progression or are otherwise detrimental, to target therapeutic transgenes to well-characterized ‘safe harbor’ loci to prevent disruption of endogenous gene regulation and to correct disease-causing mutations. Following the first proof-of-principle demonstration of correction of mutations that cause a human genetic disease5, ZFNs have now entered clinical trials to disrupt the CCR5 gene used by HIV to enter T cells19. Efforts for the clinical translation of TALENs and CRISPR-Cas9 are under way as well. All three genomeediting technologies have been applied to preclinical models of many diverse diseases,
npg
© 2014 Nature America, Inc. All rights reserved.
COMMENTARY | SPECIAL FEATURE including viral infection, cancer immunotherapy, sickle-cell disease, hemophilia, Duchenne muscular dystrophy and various immunodeficiencies, just to name a few examples. Notably, the recent demonstrations of efficient gene editing in vivo20,21 and in human hematopoietic stem cells22 are indicative of the widespread therapeutic applicability of genome editing. Importantly, there are both common and unique challenges to each of the genomeediting platforms. A primary concern is their specificity and precision in making exact sequence changes. Off-target activity at unknown genomic locations has the potential to convolute scientific results or cause adverse effects. Tremendous advances have been made in improving the specificity of each technology, but the field is still in dire need of methods that are more sensitive and comprehensive in determining gene-editing specificity for each individual cell of a modified population. Typically, only a minority of the treated cells are actually edited, and thus there is also a major emphasis on increasing nuclease activity and developing methods to select for correctly modified cells. Delivery of the nucleases is another challenge, particularly for in vivo applications, and the large size of the genes encoding Cas9 and TALENs is a hurdle that is currently being addressed. Advances in engineering of viral vectors and nanoparticles, as well as strategies for direct delivery of nuclease proteins, are likely to play a significant role in this work moving forward. Finally, issues such as immunogenicity of gene-editing tools have largely been unreported thus far and will be a subject of future interest for clinical applications. Looking ahead, there are numerous nascent technologies that are likely to have a major impact on the genome engineering field in the next few years. Alternative enzy-
matic domains to the FokI- and Cas9-based nucleases have been slower to develop, but the engineering of meganucleases, recombinases, transposases and hybrids of these enzymes fused to zinc-finger proteins, TALEs and the catalytically inactive Cas9 (dCas9) have all seen recent breakthroughs. Building on the effectiveness of the CRISPRCas9 system, other technologies that exploit RNA-DNA and DNA-DNA interactions for target-site recognition may also emerge23. In addition to the editing of DNA sequences, applications of these genome engineering tools that manipulate transcriptional regulation 24–27 and edit the epigenome28,29 will also have a significant impact in the near future. These technologies will be greatly valuable in decoupling the complexity of eukaryotic gene regulation and for applications such as reprogramming gene networks to direct changes in cell phenotype. The availability of orthogonal Cas9 enzymes30 may facilitate this work by enabling independent control of multiple epigenetic modifiers. Other applications of these tools for imaging chromosomal loci, controlling three-dimensional genome structure and generating complex libraries of DNA sequences that have emerged only in the last year are likely to find extensive use. Finally, technologies for dynamically regulating the activity of these genome engineering tools with precise spatiotemporal control, such as optogenetic methods29, will be critical for achieving the ultimate goal of programming any property of the genome, in any cell, at any time. ACKNOWLEDGMENTS I thank the members of the Gersbach laboratory and our collaborators that have contributed to our work in this area. C.A.G. is supported by US National Institutes of Health (NIH) Director’s New Innovator Award (DP2OD008586), US National Science Foundation (NSF) Faculty Early Career
Development (CAREER) Award (CBET-1151035), NIH R01DA036865, NIH R21AR065956 and the Muscular Dystrophy Association (MDA277360). COMPETING FINANCIAL INTERESTS The author declares competing financial interests: details are available in the online version of the paper (doi:10.1038/nmeth.3113). 1. Pavletich, N.P. & Pabo, C.O. Science 252, 809– 817 (1991). 2. Smith, J. et al. Nucleic Acids Res. 28, 3361–3369 (2000). 3. Bibikova, M., Beumer, K., Trautman, J.K. & Carroll, D. Science 300, 764 (2003). 4. Porteus, M.H. & Baltimore, D. Science 300, 763 (2003). 5. Urnov, F.D. et al. Nature 435, 646–651 (2005). 6. Boch, J. et al. Science 326, 1509–1512 (2009). 7. Moscou, M.J. & Bogdanove, A.J. Science 326, 1501 (2009). 8. Christian, M. et al. Genetics 186, 757–761 (2010). 9. Miller, J.C. et al. Nat. Biotechnol. 29, 143–148 (2011). 10. Jinek, M. et al. Science 337, 816–821 (2012). 11. Mali, P. et al. Science 339, 823–826 (2013). 12. Cong, L. et al. Science 339, 819–823 (2013). 13. Cho, S.W., Kim, S., Kim, J.M. & Kim, J.S. Nat. Biotechnol. 31, 230–232 (2013). 14. Hwang, W.Y. et al. Nat. Biotechnol. 31, 227–229 (2013). 15. Soldner, F. et al. Cell 146, 318–331 (2011). 16. Ding, Q. et al. Cell Stem Cell 12, 238–251 (2013). 17. Wang, T., Wei, J.J., Sabatini, D.M. & Lander, E.S. Science 343, 80–84 (2014). 18. Shalem, O. et al. Science 343, 84–87 (2014). 19. Tebas, P. et al. N. Engl. J. Med. 370, 901–910 (2014). 20. Li, H. et al. Nature 475, 217–221 (2011). 21. Yin, H. et al. Nat. Biotechnol. 32, 551–553 (2014). 22. Genovese, P. et al. Nature 510, 235–240 (2014). 23. Swarts, D.C. et al. Nature 507, 258–261 (2014). 24. Zhang, F. et al. Nat. Biotechnol. 29, 149–153 (2011). 25. Gilbert, L.A. et al. Cell 154, 442–451 (2013). 26. Perez-Pinera, P. et al. Nat. Methods 10, 973–976 (2013). 27. Maeder, M.L. et al. Nat. Methods 10, 977–979 (2013). 28. Maeder, M.L. et al. Nat. Biotechnol. 31, 1137– 1142 (2013). 29. Konermann, S. et al. Nature 500, 472–476 (2013). 30. Esvelt, K.M. et al. Nat. Methods 10, 1116–1121 (2013).
NATURE METHODS | VOL.11 NO.10 | OCTOBER 2014 | 1011
