A TALE nuclease architecture for efficient genome editing

(Jernej Mustar)

Summarized from Miller, J. C. et al. A TALE nuclease architecture for efficient genome editing. Nat. Biotechnol., vol. 29, pp. 143–148. Feb. 2011.

Introduction to genome editing

Genome editing is a type of genetic engineering, where an organisms DNA is manipulated by altering; inserting, removing or replacing nucleotides, therefore changing its genome. This process is performed by the use of immensely useful tool, t. i. engineered nucleases, which are capable of cutting both strands of DNA or nicking single strand. Resulting sites are detected and repaired by endogenous repair mechanisms, which include homologous recombination (HR) or non-homologous end joining (NHEJ). The latter is error prone, therefore it is likely for mutations to occur in repair site. These repair pathways are highly conserved in diverse cell types, which is beneficial for the use of genome editing technologies. As a result, various cell types have been engineered by the use of portable cleavage-based technologies, including human stem cells [1].

Recently, there has been tremendous development in the field of synthetic biology and genome editing technology. Virtually any gene can be manipulated in diverse range of organisms. Most common approach is the use of nuclease, which can be manipulated to target specific sequences of genome. This can be achieved by exploitation of sequence-specific DNA-binding domains (ZFNs, TALENs) or by the use of complementary guide RNA (CRISPR-Cas9 system). In this paper, the history of genome editing will first be briefly reviewed, further on genome editing will be discussed with emphasis on the TALEN’s family.

Brief history of genome editing

Past studies in the field of genome editing were perusing a common goal, t. i. to obtain high specificity, which would limit the target affects [2]. The first engineered technology, based on nuclease activity, was introduced in 1991 by Pavletich and Pabo in the journal Science [2]. This technology by the name of Zinc Finger nuclease was a great breakthrough and was later on frequently applied for genome targeting. It is based on the use of cleavage domain of restriction enzyme FokI, which is linked with designated zinc finger protein (ZFP), which is responsible for targeting. By the use of different combinations of ZFP domains the selectivity for single genomic cleavage event is achieved. With the use of nuclease domains with monomeric activity there is a high probability of off-target events. This can be avoided by the use of a distinct pair of ZFNs with FokI domains that are obligate heterodimers [3]. Overtime, reports of problems and drawbacks of this technology were reported, such as targeting problem for certain triplets and reduced specificity, suffered by interactions within a zinc finger array [2]. The next technology, which emerged in 2009, is called transcription activator-like effector nuclease (TALENs). These are fusion proteins, where nuclease domain is associated with TAL effector DNA binding domain. This approach may be better in certain aspects, for example it enables a larger spectrum of targets and is more practical than ZFNs. In 2011, TALENs was proclaimed as “Method of the Year” by the journal Nature methods. The details of this system and molecular mechanisms of function are described later on. We probably all at least heard if not read about latest breakthrough in the field of genome editing, which is CRISPR-Cas9 system. The discovery was somewhat fortuity, because researchers were studying how bacteria defend against phages and foreign plasmids. The paper was published in August 2012 in journal Science [4]. A bacterial nuclease called Cas9 is capable of altering genomes of invading viruses, consequently inactivating them. Clustered regularly interspaced short palindromic repeats (CRISPR) are DNA loci that contain short sequence repetitions, which can be transcribed, processed and then used to guide nuclease to a specific site [5]. This is in contrast with TALEN and Zinc Finger methods, where the need to prepare customized nuclease for each target remains.

A TALE nuclease architecture

As mentioned before, two studies from 2009 proposed the use of TALEs effectors for targeted genome engineering. These transcriptional activators originate from Xanthomonas plant pathogens. During pathogenesis, TALEs specifically bind and regulate plant gene expression, thus aiding bacterial infection [6]. A central repeat domain, which mediates DNA recognition, is located within TALE structure [1]. Each repeat unit comprises of 33 – 35 amino, these specify one target base. The units base preference is determined by two crucial adjacent amino acids, also called the “repeat variable di-residue” (RVD). Therefore, the sequence of specified units in central repeat domain determines the target sequence on genome. TALEs generated with new repeat combinations have been shown to recognize target sequences predicted by this code. In the light of aforementioned advances the interest in potential use of this system in combination with nuclease was initiated. Subsequently, a TALE-nuclease chimeras (TALENs) was proposed, which would allow for site-specific genome cleavage. In selected paper, the advances in the field of TALEN’s technology are discussed. TALE activity in a mammalian cell environment for targeted regulation of episomal reporters and an endogenous gene is presented, resuming with the research of the minimal TALE region required for high-affinity DNA binding. Paper also includes practical application, showing modification of endogenous human genes NTF3 and CCR5 by the use of TALEN approach.

Retargeting a natural TALE to an endogenous mammalian sequence

Firstly, a suitable TAL effector for this study was chosen. Criteria for TALE selection was based on specificity and empirical evidence. The group was searching for natural TALE with high specificity for target sequence and with confirmed activity in mammalian cells. To achive this goal, PCR amplifications for several TALEs using Xanthomonas axonopodis pathovar citri genomic DNA were performed. Native TALE-associated sequences which allow transport into plant cells were excluded. Protein products of acquired coding regions were characterized by a SELEX assay using target of interest. Outcome of this selection yielded one highly selective candidate, identified as TALE13. When combined with VP-16 activation domain, a 70-fold induction of reporter gene expression rate was observed in HEK293 cells. Based on these results, TALE13 was chosen for further work and design. First goal was to demonstrate that chosen candidate was appropriate for mammalian gene regulation. Chosen target was NTF3 gene, which encodes a secreted nerve growth factor that has therapeutic potential for neurodegenerative diseases. Wild type TALE usually contain 18 repeats, by which they target specific sequence and bind to proximal promoter elements. TALE13 was modified to the extent of recognising proximal promoter region of NTF3 by replacing the 18 wild-type repeat units with appropriate alternatives. This altered TALE was named NT-L (meaning NTF3 Left), which highlights its binding position on target. In order to test its activity, the truncated version of NT-L was fused with VP-16 and expressed in human HEK293 cells. By measuring the activity of endogenous NTF3 locus, a strong induction (over 20 fold) in both NTF3 transcript and protein product was detected. The subsequent SELEX analysis revealed high specificity. These results revealed, that NT-L variant is a viable candidate for endogenous mammalian gene regulation.

Development of a TALE-nuclease architecture

The next step was to design TALENs chimera protein using NT-L and nuclease, which would allow implementation of targeted cleavage of its genome target. The process was realized through three stages. First, the minimal region required for high affinity bonding was determined using gene regulation studies. This was beneficial for optimization of nuclease (FokI) cleavage domain attachment with its target. Desired outcome of this step was improved catalytic activity of chimeric tool. The truncated variant named NT-L+95 was shown to retain high specificity and at the same time contribute to improved nuclease attachment. In order to examine this, a C-terminal truncated series of TALE13 were generated and subsequently linked with FokI. The resulting candidates were screened as homodimers for cleavage of specific targets. The results have shown several interesting occurrences. For TALENs comprising of NT-L+95 no detectable nuclease activity was observed. The cause for this is unclear; it is possible that this is due to a certain effect of NT-L+95 to nuclease domain positioning. Efficient DNA cleavage was observed for certain shorter specimens, providing candidates for cellular studies. The minimal region between TALEN binding sites for optimal activity is over 12 bp, otherwise reduce cleavage may occur. These observations are useful for design of TALENs used for in vivo studies. In the last stage, homodimer TALEN partner for NT-L was designed and named NT-R (NTR3 Right), indicating the direction of binding. This pair was expressed in human K562 cells as fusions with catalytic domain of FokI. Target locus was later on tested for gene modification. By using two different C-terminal subregions (+28 and +63) the designed TALENs were tested. Significant activity in in vitro cleavage study was detected. With expression at 37°C, up to 3% modification was observed by the Surveyor assay [7]. This assay reveals gene modification by the appearance of digestion products. The percentage of gene modification can be obtained by quantification of “cut” bands. Limit of detection of this assay is about 1%. The activity was higher (9% of modification) in conditions of transient hypothermia, t. i. at 30°C. Sequences from latter were analyzed by Sanger sequencing. Out of 84 analyzed alleles there were 7 mutated by the occurrence of short deletions. This observation is consistent with expectation of non-homologous end joining (NHEJ) activity. Results have also shown a NHEJ-mediated capture of an oligonucleotide duplex at the endogenous NTF3 locus. These results confirmed that NHEJ-mediated genome modification by the use of TALEN architecture is possible in mammalian environment.

Efficient modification by NHEJ and homology directed repair (HDR)

To expand the practical application for this developed system, modification at an additional locus was proposed. CCR5 gene was targeted for modification. First, a set of TALENs with a range of target lengths from 13 to 17 bp and of separation distance from 5 to 27 bp were designed and tested to identify values compatible with endogenous activity. To this end, a cluster of four left and four right binding sites were chosen for targeting. This defined a matrix of 16 heterodimer targets. For each binding site, two alternative proteins were generated, with 28 or 63 residues on C-terminal (as before). Altogether, this yielded 8x8=64 combinations of left and right proteins. These were expressed by pairs in K562 cells under standard conditions. Once again, Surveyor assay was used for evaluation of allele modification. Obtained results indicate that introduction of some +63/+63 pairs yielded over 20% allele modification. TALEN treated cells seem to retain a stable modification lever after 10 day growth, suggesting that TALENs are generally well tolerated. The effect of spacer length on activity is observed once again. At least 5% modification is achieved by the use of 12 – 19 bp spacer length in the +63/+63 TALENs section. In contrast, +28/+28 section show maximal activity in narrow spacer range of 12 – 13 bp. The next challenge was to show that it is possible to induce gene through HDR using this system. For this, a part of CCR5 region was selected for targeting, also mentioned as safe harbor previously used for transgene integration. 24 TALEN pairs were then screened for their ability to induce NHEJ. Two highly active combinations were identified, first with 18% mutation rate and second with 21%. These TALEN were introduced into K562 cells simultaneously with donor DNA fragment carrying 46 bp with restriction site BglI. The success of integration was proven by PCR and BglI digestion. About 16% of inspected alleles were modified. This is an example how TALEN architecture can be used to activate HDR.

Conclusion

In presented study it is shown how TALEN architecture can be designed and generated for efficient genome editing in mammalian environment. In this case, the system consists of TALE truncation variants, linked to catalytic domain of FokI. Engineered TALENs can be used for gene modification of human cells by activating either NHEJ or HDR repair pathways. Furthermore, three distinct applications are demonstrated; NHEJ mediated gene disruption, NHEJ-mediated insertion by capture of a DNA fragment and HDR mediated gene editing. Aforementioned suggests that TALEN architecture is a compelling tool in synthetic biology.

Discussion

A step to facilitate the use of this architecture would undoubtedly be the linkage of minimal central TALE repeat domain to the catalytic domain of FokI. This was previously attempted by several groups without success, whereas the lack of activity of such constructs [8]. An alternative idea is to use a full length TALE for targeting. Disadvantage of the latter is the size of linkage peptide connecting two domains, which would affect the nuclease activity. Third approach is to generate truncated variants of TALEs and link them with catalytic domain of FokI, as shown in this paper. This approach has proven affective when using truncated variants with 28 or 63 of the 278 original C-terminal residues. During the experimental work of this study, two papers were submitted where they demonstrated the use of TALEN architecture with the use of unmodified TALE or TALE with slightly reduced C-terminus [8], [9]. These systems were used for episome cleavage in yeast, an approach that is hardly translated to more complex genomes. As expected, these proteins have extended range of cleavage compared to truncated variants. Presented TALEN architecture is in some characteristics similar to ZFNs. One example is the high cleavage activity of TALENs, which enables high gene modification efficiency, which is in some cases superior to the one of ZFNs. A shortcoming of TALEN architecture compared to ZFNs is its relative size. TALE repeats are 3 – 4 times larger than ZFPs, which may cause difficulty in delivery methods. Also, limitations due to high levels of TALE repeat homology may complicate the assembly of constructs. Future experiments will provide a fuller picture of pros and cons of the two platforms. Targeted endogenous gene regulation by TALE proteins is a useful tool in research and for biotechnology applications. We can use this technology in a broad sense, such as design and generation of transcriptional activators or repressors. Although TALEs were discovered as such, previous studies did not inspect this type of activity outside of their native plant context. The experiments shown in this paper suggest that TALEs can be used as transcriptional regulators. The use of SELEX assay for identification and characterization of TALEs has previously not been reported. By using this method, it was possible to assess the DNA binding preferences of natural and designed TALE proteins. Binding sites for TALE13 and four engineered TALEs were examined. On this basis, an observation that 73 out of 76 TALE repeats facilitate base preference of RVD. An interpretation of rare exceptions is proposed, facilitating further design and use of this system. Overall data collected can be used to examine previously acquired information on RVDs and thereby gaining more insight on mechanisms of action and selectivity. Findings that common RVDs such as NI, HD, NN and NG selectively bound selectively bound adenine, cytosine, guanine and thymine, respectively, are in accordance with previous knowledge. Pattern NN has shown relaxed specificity with substantial binding to a second base, adenine. NK is therefore more suitable for specificity, due to much stronger preference for guanine. This is a potential enhancement, proposed in this paper.

Expectations for the future

As already mentioned, targeting the designated loci with ZFNs or TALENs is often difficult and time consuming, due to design and generation of new custom ZF or TALE array for each locus. This is not the case in most recent approach in genome editing, that is the use of CRISPR-Cas9 system. Expression of distinct gRNAs alongside Cas9 results in Cas9–gRNA complexes with distinct cleavage specificities [10]. This approach is exploited in multiplex genome editing. Still, there are features of this system which are inferior to elderly time-demanding ZFNs or TALENs. One of those features is size of gene constructs, which is a limitation in field of gene delivery. Cas9 is the size of about 160 kDa, whereas a typical 17,5 repeat TALEN corresponds to 105 kDa. By far the smallest are ZFNs, wherein typical four finger ZFN has about 40 kDa [10]. It is likely for this compact nature to be beneficial for certain experiments. As for specificity, the number of bases for modification in Cas9 system is restricted to 20 and is somewhat limited by the rule of three guanine bases flanking the target site (GN19NGG) [10]. The main constriction of ZF arrays is a scarce set of fingers, resulting in inability to target certain triplets. In addition, some ZFs targeting GNN seem to be superior to other ZFs, contributing to limitations. TALE target specificity therefore seems superior, with capabilities of targeting up to 24 bases. For Cas9 system, off target studies in mammalian environment are yet to be conducted [10]. It is likely that distinct features will lead to differential use of genome editing systems for distinct applications in the fields of plant science, therapeutic sciences and biotechnology applications.

Literature

[1] J. C. Miller, S. Tan, G. Qiao, K. A. Barlow, J. Wang, D. F. Xia, X. Meng, D. E. Paschon, E. Leung, S. J. Hinkley, G. P. Dulay, K. L. Hua, I. Ankoudinova, G. J. Cost, F. D. Urnov, H. S. Zhang, M. C. Holmes, L. Zhang, P. D. Gregory, and E. J. Rebar, “A TALE nuclease architecture for efficient genome editing,” Nat. Biotechnol., vol. 29, no. 2, pp. 143–148, Feb. 2011.

[2] “Genome Editing.” [Online]. Available: http://www.allelebiotech.com/genome-editing/. [Accessed: 06-Jan-2015].

[3] R. M. Gupta and K. Musunuru, “Expanding the genetic editing tool kit: ZFNs, TALENs, and CRISPR-Cas9,” J. Clin. Invest., vol. 124, no. 10, pp. 4154–4161, Oct. 2014.

[4] M. Jinek, K. Chylinski, I. Fonfara, M. Hauer, J. A. Doudna, and E. Charpentier, “A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity,” Science, vol. 337, no. 6096, pp. 816–821, Aug. 2012.

[5] “CRISPR,” Wikipedia, the free encyclopedia. 09-Jan-2015.

[6] “TAL effector,” Wikipedia, the free encyclopedia. 06-Jan-2015.

[7] J. C. Miller, M. C. Holmes, J. Wang, D. Y. Guschin, Y.-L. Lee, I. Rupniewski, C. M. Beausejour, A. J. Waite, N. S. Wang, K. A. Kim, P. D. Gregory, C. O. Pabo, and E. J. Rebar, “An improved zinc-finger nuclease architecture for highly specific genome editing,” Nat. Biotechnol., vol. 25, no. 7, pp. 778–785, Jul. 2007.

[8] M. Christian, T. Cermak, E. L. Doyle, C. Schmidt, F. Zhang, A. Hummel, A. J. Bogdanove, and D. F. Voytas, “Targeting DNA double-strand breaks with TAL effector nucleases,” Genetics, vol. 186, no. 2, pp. 757–761, Oct. 2010.

[9] T. Li, S. Huang, W. Z. Jiang, D. Wright, M. H. Spalding, D. P. Weeks, and B. Yang, “TAL nucleases (TALNs): hybrid proteins composed of TAL effectors and FokI DNA-cleavage domain,” Nucleic Acids Res., vol. 39, no. 1, pp. 359–372, Jan. 2011.

[10] A. Strauβ and T. Lahaye, “Zinc Fingers, TAL Effectors, or Cas9-Based DNA Binding Proteins: What’s Best for Targeting Desired Genome Loci?,” Mol. Plant, vol. 6, no. 5, pp. 1384–1387, Sep. 2013.