Expanding the genetic toolkit in Xenopus: Approaches and Opportunities for Human Disease Modeling

Abstract

The amphibian model Xenopus, has been used extensively over the past century to study multiple aspects of cell and developmental biology. Xenopus offers advantages of a non-mammalian system, including high fecundity, external development, and simple housing requirements, with additional advantages of large embryos, highly conserved developmental processes, and close evolutionary relationship to higher vertebrates. There are two main species of Xenopus used in biomedical research, Xenopus laevis and Xenopus tropicalis; the common perception is that both species are excellent models for embryological and cell biological studies, but only Xenopus tropicalis is useful as a genetic model. The recent completion of the Xenopus laevis genome sequence combined with implementation of genome editing tools, such as TALENs (transcription activator-like effector nucleases) and CRISPR-Cas (clustered regularly interspaced short palindromic repeats-CRISPR associated nucleases), greatly facilitates the use of both Xenopus laevis and Xenopus tropicalis for understanding gene function in development and disease. In this paper, we review recent advances made in Xenopus laevis and Xenopus tropicalis with TALENs and CRISPR-Cas and discuss the various approaches that have been used to generate knockout and knock-in animals in both species. These advances show that both Xenopus species are useful for genetic approaches and in particular counters the notion that Xenopus laevis is not amenable to genetic manipulations.

Introduction

With the advent of customizable genome editing technologies, such as TALENs (transcription activator-like effector nucleases) and CRISPR-Cas (clustered regularly interspaced short palindromic repeats-CRISPR associated), it is now possible to model human genetic disorders in any animal or cellular system, including previously non-genetic models such as Xenopus laevis (Harrison et al., 2014; Peng et al., 2014). These advances provide researchers with a broader range of animal models, allowing them to choose the most experimentally tractable and biologically relevant system in which to test the function of specific disease-associated genes. In this review, we discuss the application of new genome editing technologies in both Xenopus laevis and Xenopus tropicalis, and how these advances will enhance our understanding of the molecular mechanisms underlying human disease.

Xenopus has many experimental advantages that make it a well-suited model for the study and functional characterization of candidate genes involved in human disease. The Xenopus tropicalis genome contains orthologues of 79% of the identified human disease genes (Hellsten et al., 2010; Khokha, 2012). Both Xenopus species have been used to study all aspects of vertebrate embryology, such as gastrulation, axis development and organ formation, dating back to the nineteenth century. These studies have provided insight into highly conserved members of major signal transduction pathways, for instance, BMPs and Wnts (Bier and De Robertis, 2015; Hikasa and Sokol, 2013). Notably, Sir John Gurdon (2012) and Professor Tim Hunt (2001) were awarded Nobel prizes for their ground-breaking research performed in Xenopus laevis (Gurdon, 2013; Hunt, 2002). The ease with which X. laevis and X. tropicalis embryos can be cultured in simple buffers and raised to adulthood enables the use of this model in most laboratory settings. Due to the large number of progeny that can be obtained from a single mating, Xenopus provides ample embryonic tissue for a wide range of experimental assays, including phenotypic analyses, RNAseq, ChIPseq and proteomics (Chung et al., 2014; Onjiko et al., 2015; Peshkin et al., 2015; Wühr et al., 2014; Yanai et al., 2011). Because of their relatively large size and ability to develop in culture, Xenopus embryos are particularly amenable to studies focused on manipulating gene function via microinjection of morpholinos, DNA constructs, translated capped RNA, and protein. Overexpression or misexpression of wildtype or dominant negative proteins were facilitated by simple and efficient transgenesis approaches, by a variety of means, including restriction enzyme mediated integration (REMI), I-SceI meganuclease or transposon mediated insertion of constructs into the Xenopus genome, and are reviewed elsewhere (Buchholz, 2012; Ishibashi et al., 2012b; 2012c; Kelley et al., 2012). In addition, the ability to fate map cell lineages from various embryonic blastomeres and tissue regions in Xenopus has greatly enhanced our understanding of the developmental, genetic, and evolutionary origin of adult structures and organs, which is essential for determining disease etiology (Chalmers and Slack, 2000; M. C. Lane and Sheets, 2006). These experimental advantages, together with their rapid external development, detailed temporal staging atlas, and relative transparency facilitate gene function assessment, making both Xenopus species versatile model systems for disease research and phenotypic drug screening (Harland and Grainger, 2011; Schmitt et al., 2014).

Among aquatic vertebrate model animals, Xenopus excels by having comparable organ development and morphology to mammalian systems, but with the added benefit of being able to regenerate adult tissues, such as optic nerve, lens, spinal cord and limb tissue (Blitz et al., 2006; Muñoz et al., 2015; Slack et al., 2008). Xenopus animals and oocytes are used extensively to understand normal organ function and disease in humans (Labonne and Zorn, 2015), including cardiac congenital heart disorders and heterotaxy (Boskovski et al., 2013; Duncan and Khokha, 2016; Fakhro et al., 2011; Kaltenbrun et al., 2011; Langdon et al., 2012; 2007), gastrointestinal and pancreatic diseases (Kofent and Spagnoli, 2016; Pearl et al., 2009; 2011; Salanga and Horb, 2015; Womble et al., 2016), endocrine functions and disorders (Buchholz, 2015), kidney disease (Lienkamp, 2016), lung development (Rankin et al., 2011; 2015; Wallmeier et al., 2014), cancer (Chernet and Levin, 2013; Cross and Powers, 2009; Hardwick and Philpott, 2015; Haynes-Gilmore et al., 2014; Van Nieuwenhuysen et al., 2015; Wylie et al., 2015), ciliopathies (Kim et al., 2010; Klos Dehring et al., 2013; Ma et al., 2014), orofacial defects (Dickinson, 2016), and neurodevelopmental disorders (Erdogan et al., 2016; Pratt and Khakhalin, 2013). Looking forward, Xenopus is poised to take advantage of the new developments in genomics and genome engineering to better understand the molecular mechanisms underlying human disease (Harland and Grainger, 2011; Labonne and Zorn, 2015).

While genetics is available in fish and mice as surrogate systems for understanding human biology and disease, the development of Xenopus genetics offers a number of advantages not found in other organisms. Unlike the mouse, Xenopus embryos can be produced in large numbers and are accessible throughout their development, simplifying phenotypic screening at embryonic stages. Xenopus shares surprising similarities with humans both at the level of its genome and its anatomy. The frog genome has long regions in which genes exhibit remarkably similar syntenic relationships to those found in the human genome (Amodeo et al., 2015; Blitz, 2012; Blitz et al., 2013; Bodart and Duesbery, 2006; Davidson, 1973; Grant et al., 2015; Hellsten et al., 2010; Paranjpe et al., 2013; Roe et al., 1985; Showell et al., 2011; Uno et al., 2013). In many cases, orthologous genes are found in equivalent regions of the human and frog genomes, in which the order of genes along the chromosomes is largely conserved. In fact, the vast majority of the breaks in synteny are from single exon genes identified by automated ORF prediction algorithms but not supported by EST evidence (Blitz, 2012; Geach et al., 2012; Krylov and Tlapakova, 2015; Macha et al., 2012; Pollet and Mazabraud, 2006; Showell and Conlon, 2007). Therefore, shared synteny may be even more prevalent than is currently thought. Moreover, while it is well established that mice are genetically tractable, mice are difficult for live imaging or biochemistry. Thus by combining genome editing with the advantages of both Xenopus species, researchers have the unique opportunity to integrate systems level genomic and proteomic analyses with quantitative live imaging of cell behaviors in genetically approachable vertebrate model systems.

Xenopus laevis versus Xenopus tropicalis

Currently, there are two main Xenopus species used in biomedical research, Xenopus laevis and Xenopus tropicalis. Historically, X. laevis has been the predominant Xenopus species studied since the 1950s due to its large size, ability to ovulate year round, and experimental robustness. These animals can be grown between 18–23°C in basic salt solutions, in relatively simple aquatic housing. They are the animal of choice for many biomedical researchers, including those using the Xenopus oocyte model to study ion channel electrophysiology, cell protein biochemistry, cell cycle biology, and cytoskeletal dynamics, as well as those employing the Xenopus embryo model to study development (Dubaissi and Papalopulu, 2011; Kay and Peng, 1991; Maksaev and Haswell, 2015; Mitchison et al., 2015). One of the drawbacks often cited with X. laevis is its poor genetic tractability, which was thought due to its allotetraploid genome and relatively long generation time of 10–12 months; thus, most researchers have had little interest in the genetic background of X. laevis and have purchased outbred genetically heterogeneous frogs from commercial vendors. However, in the past 5 years, an inbred strain of X. laevis, known as the J strain, has become available through the National Xenopus Resource and the European Xenopus Resource Centre (Gantress et al., 2003; Pearl et al., 2012; Tochinai and Katagiri, 1975), and the J strain draft genome sequence is now available on Xenbase (http://www.xenbase.org; X. laevis Genome Project Consortium). This newly sequenced strain will allow for more genetic analyses in X. laevis and when combined with recent advances in genome editing technologies, has made it possible for essentially any lab to make mutants through targeted reverse genetics. In more recent years, X. tropicalis has become increasingly used as a genetic model because it offers the same embryological benefits as X. laevis, but has a shorter generation time of 5–7 months, a smaller size, and a diploid genome. However, water conditions are slightly different for X. laevis and X. tropicalis, and hence they require separate housing systems, making it more expensive for individual labs to maintain colonies of both species. In light of these considerations, most researchers focus on one model species, but both species offer unique advantages that are being exploited by the wider Xenopus community.

Inbred J strain X. laevis and Nigerian X. tropicalis frogs can be purchased from the National Xenopus Resource (NXR) in the US or the European Xenopus Resource Center (EXRC) in the UK. These resource centers were established in the past 10 years to serve as the main stock centers for the Xenopus community for inbred, transgenic and mutant lines (Pearl et al., 2012). Their recent advances in husbandry have optimized X. laevis maintenance, and it is now possible to perform in vitro fertilization from males as young as 4 months old, which has decreased the time required to raise lines of X. laevis (Horb, M.E. unpublished). In addition, these stock centers have worked to improve the cryopreservation of X. laevis and X. tropicalis sperm, allowing subsequent matings and the sharing of colony lines among the community. These improvements bring Xenopus more in line with other model systems and enable the proper maintenance and distribution of different strains and lines from both species.

Complete sequencing of both X. laevis and X. tropicalis genomes has revealed larger regions of synteny, and presumably the sharing of a longer common evolutionary history, with humans than those of other popular model systems (Hellsten et al., 2010). X. tropicalis is the only diploid species in the Xenopus genus, and its genome is comprised of twenty chromosomes, whereas X. laevis is an allotetraploid species that arose from the interspecific hybridization of two diploid species, and its genome consists of 36 chromosomes (Matsuda et al., 2015; Uno et al., 2013). There are nine pairs of homeologous chromosomes in X. laevis, which are named Long (L) and Short (S) that functionally segregate as two separate diploid genomes (Matsuda et al., 2015; Tymowska, 1991). A new nomenclature for the X. laevis chromosomes was recently established based on their phylogenetic relationship and length, such that the homeologous chromosomes are named Xla1L and Xla1S through Xla9L and Xla9S (Matsuda et al., 2015). The first eight pairs of X. laevis chromosomes correspond to those of X. tropicalis, whereas the ninth chromosome pair contains fusions of chromosomes 9 and 10 from X. tropicalis and thus, they are also named Xla9_10L and Xla9_10S to emphasize their phylogenetic relationship. As the name implies, S chromosomes are shorter than L chromosomes, and this is due to loss of genes on the S chromosomes; it has been postulated that at least 17%, but perhaps as much as 50%, of the genes in X. laevis remain in a diploid/singleton state due to loss of genes on the S chromosomes (Hellsten et al., 2007; Uno et al., 2013), making the generation of mutants relatively simple for such target genes. These genome annotation advances, together with the recently completed Xenopus ORFeome project (Grant et al., 2015) that identified orthologues of 2,724 human genes associated with an Online Mendelian Inheritance in Man (OMIM)-recognized disease, makes X. laevis an ideal model to study vertebrate development and human disorders.

All of the benefits discussed above demonstrate that both Xenopus species are excellent model systems that will be greatly enhanced with emerging genome-editing methodology. These revolutionizing technologies will allow for the rapid creation of mutant frogs to model human diseases, providing an abundant source of material for functional studies. Utilizing genome editing methods in Xenopus laevis and Xenopus tropicalis will provide a cost effective platform to rapidly identify, validate, and characterize genes involved in human diseases, which will ultimately provide more detailed mechanistic insight to guide new therapeutic strategies. Below, we describe the use of TALENs and CRISPR-Cas based methods in Xenopus, and highlight some of the emerging applications of these methods to understand human diseases.

Genome Editing Tools in Xenopus

In the past few years numerous methods have been generated to efficiently edit the genomes of almost any cell type or organism, including the amphibian models Xenopus laevis and tropicalis, through the use of Zinc Finger Nucleases (ZFNs), Transcription Activator-Like Effector Nucleases (TALENs) and Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR-Cas) nuclease systems (Harrison et al., 2014; Peng et al., 2014). The nucleases are targeted to a specific region of interest in the genome through the microinjection of sequence-specific RNA constructs, and induce double/single-stranded breaks in the DNA sequence. Upon recognizing this damage, the cell uses non-homologous end joining (NHEJ) or homology-directed repair (HDR) to repair the break. NHEJ produces random insertions or deletions (indels) into the target site, thereby causing potential DNA coding frame shifts and altered protein translation. HDR, on the other hand, uses a template sequence to repair the broken DNA strands; this enables the insertion of exogenous DNA sequences, such as fluorescent reporter transgenes to track gene expression and cell lineage, or point mutations mimicking known human single nucleotide polymorphisms (SNP) associated with a disease trait. As both TALENs and CRISPR-Cas systems have proven to be more versatile than ZFNs, we will focus on these more recently described gene-editing methods that have revolutionized basic biology as well as biomedical and other translational research.

TALENs in Xenopus

Originally identified in the plant pathogen Xanthomonas, TALE proteins are transcriptional activators that specifically bind and regulate plant gene expression upon infection (Joung and Sander, 2013; Kim and Kim, 2014). The TALE structure comprises a central domain harboring specific repeating units of 33–35 amino acids that target individual DNA bases (Bogdanove and Voytas, 2011). These repeat variable di-residue (RVD) domains enable TALE proteins to target almost any genomic region of interest. Site-specific TALE proteins can be tethered to endonucleases to modify genome sequence, or to transcriptional effector proteins, such as VP16 and KRAB, to regulate gene activity in most eukaryotic organisms tested to date. To induce DNA breaks and modify genome sequences, two TALEN arms are required to recognize and bind the DNA sequence with a small 15–25 base spacer region in between. This spacer allows the dimerization of the tethered Fok1 endonuclease catalytic domains, thereby enabling its enzymatic function. Cloning individual TALENs is a multi-day process that involves piecing together individual RVDs to create custom repeat arrays targeted to a specific DNA sequence (Cermak et al., 2011).

The first use of TALEN-induced mutations in Xenopus were performed in X. tropicalis, demonstrating efficient somatic and germline mutagenesis (Ishibashi et al., 2012a; Lei et al., 2012). The initial study targeted eight Xenopus genes involved in human disease, and showed that TALENs induced mutations in F0 embryos with a high efficiency at all eight loci (Lei et al., 2012). In particular, they demonstrated that injection of TALENs targeting the pancreatic transcription factor ptf1a resulted in pancreatic agenesis in F0 tadpoles, mimicking the phenotype seen in humans and confirming previous morpholino knockdown phenotypes observed in Xenopus (Afelik et al., 2006; Jarikji et al., 2007; Sellick et al., 2004), with efficient germline transmission of mutations. In addition, they showed that TALEN mutagenesis is more efficient and less toxic when compared directly to ZFNs, with no detectable off-target effects (Lei et al., 2012). Another study by the Chen lab showed that TALEN-mediated disruption of the n-myc downstream regulated 1 (ndgr1) gene in X. tropicalis displayed a similar phenotype to that observed using morpholino knockdown of the gene in X. laevis (Zhang et al., 2013). More recently, a comprehensive study by the Grainger lab demonstrated that X. tropicalis is a useful model for understanding the developmental basis of human eye disorders. They performed an extensive characterization of different TALEN-induced mutations in the pax6 gene and identified several different phenotypes in both F0 and F1 frogs (Nakayama et al., 2015). This group found that partial loss of function of pax6 in F1 animals resulted in froglets with an underdeveloped iris, a phenotype similar to that observed in human aniridia. Collectively, these initial studies demonstrated that TALEN-mediated genome editing works efficiently in F0 X. tropicalis, and that these mutations are transmitted through the germline.

Xenopus not only provides a platform for the study of genes involved in congenital malformations, but also for the study of genes involved in developmental processes and diseases at later stages. Recent studies by the Shi and Buchholz labs demonstrated that TALEN-mediated genome editing is an effective method to induce mutations in genes involved in hormonal control of metamorphosis in F0 X. tropicalis tadpoles, specifically in thyroid hormone receptor, alpha (thra) gene (Choi et al., 2015; Wen et al., 2015; Wen and Shi, 2015). Choi et al. showed that F0 phenotype analysis and germline transmission of the TALEN-induced mutations are facilitated by injection into one cell of two-cell embryos. The Vleminckx lab used TALENs to create mutations in the X. tropicalis tumor suppressor gene adenomatous polyposis coli (apc), which is implicated as the initiating mutation in many colorectal cancers, including familial adenomatous polyposis (FAP) syndrome (Van Nieuwenhuysen et al., 2015). These studies demonstrated F0 tadpoles derived from embryos injected with apc-specific TALENs develop intestinal hyperplasia and other neoplasms commonly observed in FAP patients within 6 weeks providing a useful model for the rapid testing of chemical compounds to treat FAP. These results therefore demonstrate that it is possible to utilize the mosaic nature of TALEN-induced mutagenesis for studying advanced developmental events as well as tumorigenesis in Xenopus tadpoles.

Circumventing embryonic lethality in F0 animals

One of the potential problems in generating mutant Xenopus lines using TALENs is that, in some instances, the mutations result in embryonic lethality, thus limiting analysis to the F0 generation. Ideally, the best way to overcome this issue would be to induce mutations only in germ cells and not in somatic tissue. In Xenopus, there are two methods used to create such germ cell-specific mutants. The first is to limit the translation of injected mRNA to germ cells. Several maternal mRNAs have specific germ cell restricted translation due to the 3’UTR region, such as the gene ddx25 (Kataoka et al., 2006). Recent work from the Yaoita lab showed that the ddx25 3’UTR is sufficient for largely restricting translation of TALEN-injected mRNAs to the germ cells (Nakajima and Yaoita, 2015a). They further demonstrated that bi-allelic TALEN-induced mutations are present in the F0 adult germ cells and transmitted to the F1 generation, with limited mutations found in other organs, thus defining a method for creating germ cell-specific mutants using TALENs. Another method to produce germ cell-specific mutations is to transplant primordial germ cells from a mutant embryo into a wild-type host embryo. In Xenopus, this can be done by transplanting a portion of the vegetal hemisphere, which contains the germ plasm of the early developing embryo, at the blastula stage (Yang et al., 2015). Recent work from the Cho lab showed that this method works well in X. tropicalis to produce F0 adults that are wild-type in the soma, but contain bi-allelic mutations in the germ cells (Blitz and Cho personal communication); adults produced using this method can be mated to create F1 null progeny. These two methods will help speed up the process of analyzing null mutations and allow for more detailed studies of mutations that are embryonic lethal. In addition, because Xenopus have long life spans (over 10 years), the mutant adults can be used for many years.

Although inducing mutations in both homeologs in the allotetraploid X. laevis is, in principal, more difficult to achieve, several recent studies have revealed that it is possible to generate highly efficient gene knockouts in both homeologs in X. laevis. The complication with X. laevis is that, if the alloalleles are thought to be functionally redundant, mutations must be induced in both homeologs on the S and L chromosomes. The first reports of the effectiveness of TALENs in X. laevis showed that a single TALEN pair can be designed to target both homeologous genes, and they revealed that mutations can be detected as early as the morula stage of development (Sakane et al., 2013; Suzuki et al., 2013). Because sequence differences do exist between the alloalleles, it is also possible to design TALENs to target only one of the homeologous genes. In one study, injection of two different pairs of TALENs, targeting each alloallele, into X. laevis embryos was found to induce mutations in each homeologous gene (Nakade et al., 2015). This study also demonstrated mutation of the individual homeolog was not possible when the specific TALEN binding site contained three mismatches to the endogenous sequence, highlighting the specificity of TALENs. Although these results showed that it is possible to induce mutations in both alloalleles in X. laevis, either separately or together, they also revealed that generating mutations in all four copies is not efficient enough to produce null F0 frogs due to the delay in initiating NHEJ in all cells.

Using a well-established Xenopus method called the oocyte-host transfer technique, which was developed to knockdown maternal transcripts, it is possible to induce mutations more rapidly and efficiently than traditional embryo injections. Therefore, using the oocyte-host transfer (OHT) technique combined with genome editing will result in less mosaic F0 animals and more efficient germ line transmission of mutations (Fig. 1). In this method, heterologous mRNAs are injected into Xenopus oocytes, which are then cultured in vitro for 24–48 hours, transferred into an ovulating host female, and the laid eggs are subsequently fertilized in vitro (Olson et al., 2012). Using this technique, the Yaoita lab demonstrated that TALEN-mediated gene disruption is more efficient when injected into oocytes rather than embryos (Nakajima and Yaoita, 2015b). Similarly, we found that a single TALEN pair targeting both tyrosinase homeologs was able to generate almost complete albinism in F0 animals when injected into X. laevis oocytes (Ratzan et al., 2016); conversely, injection of these same TALENs into embryos resulted in mosaic F0 frogs, with only small patches of albinism (Fig. 1). Furthermore, the F0 oocyte-injected adults laid albino eggs, and when mated with sibling males, produced 50–75% albinism in the F1 generation; in contrast, all offspring from sibling-mated F0 embryo-injected adults failed to produce albino offspring, and very few mutations were recovered in the F1 generation (Ratzan et al., 2016).

Figure 1. Genome editing using Xenopus embryos or oocytes.

A comparison of the steps required to generate Xenopus mutants using either embryo injection or oocyte-host transfer methods. (A) For the embryo injection method, TALEN or CRISPR-Cas9 capped mRNA and/or protein is microinjected into fertilized embryos at the one-cell stage. The embryos are genotyped to confirm editing efficiency using PCR-sequencing, T7 endonuclease assays, or high-resolution melt analysis. The F0 mosaic embryos are allowed to develop, and gene function is analyzed using a host of established assays, including in situ hybridization (ISH) and immunohistochemistry (IHC). These embryos, once grown to adulthood, can be tested for germline transmission to generate subsequent mutant lines. (B) An image of a mutant frog generated from embryos injected at the one-cell stage with TALEN mRNAs targeting the tyrosinase gene shows mosaic pigmentation throughout the skin. (C) For the oocyte-host transfer method, stage VI oocytes are surgically removed from an adult female frog, manually defollicated, and microinjected with TALENs or CRISPR-Cas9 capped mRNA and/or protein. The oocytes are then matured using progesterone and colored with vital dyes for visualization; the coloring of oocytes is not necessary if implanted into an albino female. The oocytes are then transferred into pre-primed host females and subsequently laid to incorporate the jelly coat that is essential for in vitro fertilization with sperm. The resulting embryos are genotyped and phenotyped as previously described. (D) An image of a mutant frog generated from oocytes injected with the same TALENs as in panel B, targeting the tyrosinase gene, shows more dramatic levels of albinism than the embryo-injected frog, thereby confirming more efficient mutagenesis.

An alternative method to induce genetic mutations prior to fertilization has recently been published by the Gurdon lab (Miyamoto et al., 2015a; 2015b; 2013). To forego oocyte transfer back into a host female and therefore abolish further surgery, they used intracytoplasmic sperm injection (Miyamoto et al., 2015a). In this procedure, X. laevis oocytes are extracted, enzymatically defolliculated and subsequently injected with TALEN mRNA (Fig. 1); these late stage oocytes are then matured by addition of progesterone and subsequently injected with a sperm mixture (to negate the need of a jelly coat for normal in vitro fertilization methods). This method has the benefits of being less technically challenging by eliminating manual oocyte defolliculation and reintroduction of injected oocytes in the host female. In their surviving embryos Miyamoto observed a similar striking mutation efficiency for the tyrosinase and pax6 genes (between 80–90%) with all four alloalleles being targeted. Both oocyte methods are highly efficient at inducing mutations early in development, and thus allows researchers to study gene function in an F0 generation. This thereby enables the rapid assessment of disease-causing gene mutations without the need to establish mutant lines whilst also allowing the study of those genes that are essential for embryo or sexual maturation. In conclusion, these results show that TALENs work efficiently in X. tropicalis and X. laevis, and that whilst homeologous X. laevis genes can be independently targeted with two different pairs of TALENs, OHT is the most efficient method to generate mutations in all alloalleles in F0 animals.

CRISPR-Cas in Xenopus

Another genome editing technique that works well in Xenopus is CRISPR-Cas, and it has become more widely used due to its simpler design. In prokaryotes there are chromosomal loci that harbor repetitive DNA sequences, termed CRISPR elements, and adjacent to these elements there are endonuclease gene coding regions, termed CRISPR-associated genes (Cas) (Bolotin et al., 2005; Jansen et al., 2002; Jinek et al., 2012; Makarova et al., 2015; Wright et al., 2016). This combination provides the bacteria with an acquired adaptive immune system capable of specifically targeting nucleic acid sequences of invading viruses or plasmids (Bhaya et al., 2011; Horvath and Barrangou, 2010; Karginov and Hannon, 2010). Researchers have exploited the programmable nature of this DNA targeting nuclease to mutate specific genetic loci of interest in many model systems including Xenopus (Blitz et al., 2013; Cho et al., 2013; Cong et al., 2013; Doudna and Charpentier, 2014; Guo et al., 2014; Harrison et al., 2014; Jinek et al., 2012; Mali et al., 2013a; Nakayama et al., 2013; Peng et al., 2014). To induce double strand breaks in Xenopus using CRISPR-Cas, one or more single guide RNAs (sgRNA) targeting a gene of interest is injected into Xenopus embryos together with the Cas endonuclease, either as mRNA or protein. This method is especially time saving when targeting multiple loci in the same embryo, because the sgRNAs can be generated by PCR in a single day, in contrast to the multi-day process for cloning TALENs. However, unlike TALENs, sgRNAs can only be designed to target regions containing a protospacer adjacent motif (PAM) site, limiting the regions that can be targeted.

Several recent reports in the past couple of years have shown that CRISPR-Cas is highly efficient in producing mutations in both X. tropicalis and X. laevis (Blitz et al., 2013; Guo et al., 2014; Nakayama et al., 2013; Wang et al., 2015). Initial studies attempted to optimize the amounts of sgRNA and Cas9 mRNA required to produce efficient mutagenesis in the F0 generation with limited developmental defects. Generally, the amount of sgRNA required to produce efficient indels varies with each locus. For X. tropicalis, most loci require a range of 50–200 pg sgRNA, whereas other loci require up to 400 pg sgRNA; for X. laevis, 300–500 pg sgRNA is optimal, but greater for some loci. Interestingly, the amount of Cas9 mRNA required to induce indels varied among these studies. Two of the studies revealed that a relatively high amount (2.2–3 ng) of Cas9 mRNA is required to induce efficient mutagenesis (Blitz et al., 2013; Nakayama et al., 2013), whereas the third study showed that a much lower amount (300–500 pg) is sufficient (Guo et al., 2014).

These reported Cas9 dosage discrepancies may be due to the different Cas9 versions used in each study, which are identical at the amino acid level, but are only 80% identical at the nucleotide level (Cong et al., 2013; Mali et al., 2013b). As Xenopus codon usage slightly differs from that of humans, the use of rare codons may impact the translational efficiency of each Cas9 transcript, especially with the large size of the Cas9 protein. In addition, the two versions differ at their N- and C-termini: the Cong Cas9, used by the Chen lab, contains two nuclear localization sequences (NLS), one at each end, and a 3× FLAG tag at the N-terminus, whereas the Mali Cas9, used by Cho and Grainger labs, contains a single NLS at the C-terminus. Comparison of the codon usage showed that the Cong Cas9 used slightly different codons, some of which were more optimal for X. laevis. These differences may explain why a lower dose of the Cong Cas9 was required in Xenopus. As both versions were tested in the standard Xenopus vector, pCS2, they contain identical 5’ and 3’UTRs. These initial reports illustrated that CRISPR-Cas works efficiently in both X. laevis and X. tropicalis, but that care must be taken in selecting codon-optimized transcripts when performing mRNA injections of Cas9. An alternative to Cas9 mRNA is to use Cas9 protein, which eliminates the concerns of codon differences. The Khokha laboratory recently compared the efficacy of protein versus mRNA in Xenopus tropicalis and found that Cas9 protein was more effective than mRNA (Bhattacharya et al., 2015). They also reported Cas9 protein was able to induce mutations earlier in development than Cas9 mRNA and was less toxic. However, in light of the discussion of Cas9 codon usage, it is relevant to point out that they used Cas9 protein from PNA Bio, which contains only one NLS.

In addition to the technical importance, the Khokha group’s study also clearly demonstrated that X. tropicalis can be used to rapidly produce F0 phenotypes for human disease genes using CRISPR-Cas. As we have discussed, many candidate genes for human diseases have been identified, but little functional analysis of them has been possible, particularly outside a few model organisms. In this study, they designed sgRNAs to six human disease genes and showed that the F0 CRISPR mutants largely reproduced the expected disease phenotypes. Thus, with the increased efficiency afforded by Cas9 protein injection, F0 analysis was sufficient for functional screening of candidate genes involved in human disease, expanding the utility of the X. tropicalis model for this important application.

Beyond inducing genetic mutations, CRISPR-Cas can also be used to label specific chromosomal regions (Chen et al., 2013) and recent work by the Heald lab demonstrated how this can be applied to Xenopus (Lane et al., 2015). In this study they developed a technique that allows for the creation of a library of sgRNAs targeting a defined genomic region, which they call CRISPR EATING (everything available turned into new guides). The library is made by PCR amplification of specific genomic region followed by restriction digestion with an enzyme that cuts immediately 5’ to a PAM sequence resulting in various sized fragments; subsequent steps of adaptor ligations and restriction digests results in a mixture of 136 nt sgRNA fragments that contain a T7 RNA polymerase promoter for RNA transcription. The authors then complexed the sgRNAs with a recombinant catalytically inactive Cas9 fused to mNeon-Green (dCas9-Neon) to label specific chromosomal regions using Xenopus sperm nuclei in vitro. One of the main benefits of this approach is the ability to use Xenopus egg extracts and sperm nuclei to image genomic loci throughout the cell cycle. In addition to being useful for visualization of genomic loci, this approach can be adapted to generate mutations in larger regions of DNA that will be of benefit to making mutants in genes with many small exons.

Knock-in strategies in Xenopus

Several different methods have been used for many years to promote integration of exogenous DNA into the genome of Xenopus laevis or Xenopus tropicalis for the production of transgenic lines. These methods rely on random integration into the genome, and as such will not be discussed here, but we refer the reader to several excellent reviews that cover these methods in detail (Allen and Weeks, 2006; Chesneau et al., 2008; Love et al., 2011; Ogino et al., 2006; Takagi et al., 2013; Yergeau et al., 2009). In contrast, CRISPR-Cas and TALENs can be used for site-specific integration of exogenous DNA and three recent reports demonstrated that this can be accomplished in Xenopus. In the first study, they used TALENs to create F0 knock-in X. laevis tadpoles through a microhomology-mediated end joining (MMEJ) strategy, which they refer to as TALEN-mediated precise integration into target chromosome, or TAL-PITCh (Nakade et al., 2014). They used a single homology arm, containing the TALEN target site (with an inverted spacer sequence) to insert exogenous DNA into two different loci, no29 and fgk (Fig. 2A). In the first instance, they targeted the start codon of no29 to knock-in a no29-GFP fusion template, whereas for fgk they targeted the 3’ end to insert EGFP just before the stop codon to make an endogenous fusion. In both instances, they found integration at the 5’ junction was precise in most cases, but at the 3’ junction there were often deletions and insertions. In the second study, they used CRISPR-Cas to mediate integration of exogenous DNA in X. tropicalis and showed efficient germline transmission (Z. Shi et al., 2015). In contrast to traditional knock-in strategies that use homology-dependent integration, this study showed that targeted integration could be achieved independent of homology arms, as long as the sgRNA target site is included in the donor DNA (Fig. 2B). Although both methods showed that integration is imprecise and results in deletions and insertions around the target site, they illustrate that it is feasible to use TALENs and CRISPR-Cas for insertional mutagenesis.

Fig. 2. Integrating exogenous DNA into Xenopus using genetic editing tools.

Outline of the various knock-in strategies that have been employed to insert DNA into a targeted genomic locus in Xenopus. (A) Nakade et al. described the use of TALENs and microhomology-mediated end joining (MMEJ, TAL-PITCh) to integrate a fluorescent protein (eg. GFP) at the end of the coding region 5’ to the endogenous stop codon. (B) Shi et al. utilized CRISPR-Cas editing to insert plasmid DNA harboring a known pancreas tissue enhancer element (Elastase promoter) driving GFP, into the intron of their target gene. (C) Jaffe et al, used targeting constructs containing allele-specific homology arms to insert fluorescent proteins into an sgRNA-targeted exon, thereby visualizing cells in which specific gene function was abrogated. TAA; stop codon, FokI; Fok1 nuclease, GFP; green fluorescent protein, pA; poly-A tail, sgRNA; guide RNA for CRISPR.

While knock-in of exogenous DNA in F0 embryos is mosaic, this does not preclude all F0 studies as illustrated by a recent study by the Mitchell lab (Jaffe et al., 2016). In that study, they used an elegant knock-in approach to create mosaic mutant F0 X. laevis in c21orf59, a gene involved in cilia polarization in multi-ciliated cells (MCC) in Xenopus (Fig. 2C). As there are two c21orf59 alloalleles on chromosomes 2L and 2S, the Mitchell lab developed a strategy that allowed them to independently verify if one or both alloalleles were mutated in individual cells. The injection of a single sgRNA to target both gene homeologs was coupled with the incorporation of two different donor vectors harboring either BFP or RFP reporter genes. These vectors differed in their homology arm DNA sequence that would target them specifically to either the Xla2L or Xla2S alloallele. Using this approach, they identified those cells where one or both alloalleles was disrupted. Cells that expressed both BFP and RFP displayed a complete loss of cilia polarity, whereas cells that expressed only BFP or RFP displayed an intermediate phenotype. Such an approach takes good advantage of the different alloalleles in Xenopus laevis and illustrates a useful general method for knock-ins in Xenopus.

Xenopus resources for genome editing (Xenbase and NXR)

Xenbase, the Xenopus model organism bioinformatics database, is an invaluable resource for the design and application of reverse genetics in Xenopus (James-Zorn et al., 2015). Xenbase (http://www.xenbase.org) is a unique resource that serves as the central repository for all things related to Xenopus genomics. It provides a user-friendly interface to interrogate data related to a specific gene, and in particular, it provides essential genomic sequences and web-based tools for genome editing. Individual gene pages in Xenbase provide a wealth of information for a particular gene, including functional descriptions and expression profiles, associations with human disease, and links to other model organism databases. Individual gene pages also have information about the genes in both X. tropicalis and X. laevis, including both L and S homeologs in X. laevis. Each gene page provides a direct link to GBrowse, which allows visualization of the gene structure (exons and introns) and comparison of synteny with human and mouse genomes to confirm that the correct gene is being targeted. For an unannotated gene, a BLAST search can be performed directly in Xenbase to interrogate the respective genome and determine if the gene sequence is present. As annotation of the Xenopus genomes continues to be updated, often times a gene may not appear as annotated on the main gene page, but when viewed within GBrowse, the exon and intron information is properly annotated. Thus, it is critical to use the Xenbase BLAST function to identify the chromosomal location of each gene. This information is extremely useful when searching the X. laevis genome for both homeologs, because Xenbase will identify both genes on the L and S chromosomes. From GBrowse, one can then download the exon and intron sequences for an individual gene. Xenbase also provides useful links to several web-based genome editing tools for the identification of sgRNA or TALEN target sites. Thus, all of the information on Xenbase provides an essential platform to identify chromosomal locations of genes, and to design sgRNA and TALEN target sites.

Due to cost and space constraints, one of the difficulties within the Xenopus community is the raising and breeding of specific lines. In fact, most Xenopus laboratories are not experienced in breeding and maintaining mutant lines, particularly with the diploid X. tropicalis species that is ideal for genetic studies. For those unable to generate or breed their own frog lines, the National Xenopus Resource (NXR) offers a custom mutant service that will design, inject, and breed F0 or F1 X. tropicalis or X. laevis frogs. The NXR works closely with individual researchers to identify the specific region of a gene that should be mutated, and they offer multiple choices for genome editing, including TALEN and CRISPR sgRNA design and injection. In addition, for those researchers unable to maintain colonies of both X. laevis and X. tropicalis, or for those wishing to work with multiple mutant lines, the NXR provides a unique service called research facility service, which allows researchers to come to the NXR facility at the Marine Biological Laboratory (MBL) for short-term visits. Here, they can access the resources at the NXR, including the large number of different X. laevis and X. tropicalis strains and lines, and take advantage of the genome editing expertise at the NXR. The NXR also provides a service to raise and maintain animal lines at the MBL, thus allowing researchers to enhance the scope of their research. Lastly, a recent project funded by the NIH at the MBL, in collaboration with the NXR, is focused on creating 100–200 mutants for the Xenopus community. This new project works with individual researchers to generate CRISPR-Cas- and TALEN-mediated mutations in specific genes to create new Xenopus models of human disease.

Conclusions

There is substantial and growing evidence that TALENs and CRISPR-Cas genome editing tools can now be used to manipulate endogenous genes in both X. tropicalis and X. laevis, providing researchers with the powerful ability to model a host of human disorders. In Figure 3, we outline the steps required to generate mutants in Xenopus using TALENs or CRISPR-Cas systems; however, a more detailed discussion of CRISPR-Cas protocols in Xenopus can be found in a recent review (Nakayama et al., 2014). The design and application of genome editing tools for Xenopus has been streamlined by incorporating online software to design target oligonucleotides and assess potential off-target effects in the Xenopus genome. These tools, combined with simple DNA extraction and PCR techniques that are already employed by most laboratories, will enable the generation and identification of mutant embryos in a fast and efficient manner, allowing the establishment of mutant lines (Fig. 3). TALENs and CRISPR-Cas can be used in a variety of ways in Xenopus to modify specific protein domains, rearrange chromosomal organization, or to introduce the equivalent human point mutations identified through genome-wide association studies (GWAS). In addition, unbiased studies of organ formation and function in Xenopus have been shown to reveal phenotypes similar to those observed in human diseases (Iwasaki and Thomsen, 2014; Pearl et al., 2011; Sojka et al., 2014); thus, the generation of new mutants via genome editing may lead to the identification of new disease candidates. Therefore, the genome-edited Xenopus model will be instrumental as an initial tool for understanding the components and pathways affected by genetic disorders in a highly conserved vertebrate in vivo environment, which is not yet achievable with primary cell cultures or mammalian models. Furthermore, when coupled with high-throughput assays that require tissue explants and/or large numbers of embryonic samples, the genome-edited Xenopus model should aid the discovery of chemical or gene therapeutics that may serve to treat human diseases.

Figure 3. Workflow for generating Xenopus mutants using TALENs and CRISPR-Cas9.

A schematic depicts the steps required to generate the gene editing tools to target a gene of interest, induce mutations in Xenopus embryos, perform subsequent assays to phenotype mosaic F0 embryos, and generate mutant lines. For more detailed information including web URLs we refer the reader to Xenbase (http://www.xenbase.org/other/static/CRISPr.jsp).