Genome Editing in Caenorhabditis briggsae using the CRISPR/Cas9 System

Author Summary The CRISPR/Cas9 system has recently emerged as a powerful tool to engineer the genome of an organism. The system is adopted from bacteria where it confers immunity against invading foreign DNA. This work reports the first successful use of the CRISPR/Cas9 system in C. briggsae, a cousin of the well-known nematode C. elegans. We used two plasmids, one expressing Cas9 endonuclease and the other an engineered CRISPR RNA corresponding to the DNA sequence to be cleaved. Our approach allows for the generation of loss-of-function mutations in C. briggsae genes thereby facilitating a comparative study of gene function between nematodes. Abstract The CRISPR/Cas9 system is an efficient technique for generating targeted alterations in an organism’s genome. Here we describe a methodology for using the CRISPR/Cas9 system to generate mutations via non-homologous end joining in the nematode Caenorhabditis briggsae, a sister species of C. elegans. Evidence for somatic mutations and off-target mutations are also reported. The use of the CRISPR/Cas9 system in C. briggsae will greatly facilitate comparative studies to C. elegans.

Linking genotype and phenotype is an important step in the characterization of a gene. Targeted   1" genome editing, defined as the creation of alterations at specific sites in an organism's genome, is 2" a powerful means to study the relationship between gene and phenotype. Genome editing 3" techniques are based on guiding an endonuclease to a specific target in the genome in order to 4" generate a double strand break (DSB) [1][2][3]. Breaks are subsequently repaired by either error 5" prone non-homologous end joining (NHEJ) or template-directed homologous recombination 6" (HR) [4]. While the former introduces random mutations at the point of cleavage, the latter can 7" be used to generate specific alterations based on the presence of a donor sequence. Although 8" several technologies currently exist for genome editing, such as zinc finger nucleases (ZFN) and 9" transcription activator-like effector nucleases (TALEN), these techniques leave room for 10" improvement in their ease of use, as each new sequence to be targeted requires the labor intensive 11" process of generating a new protein construct [2].

12"
Clustered, regularly interspaced, short palindromic repeats (CRISPR) and CRISPR-13" associated (Cas) systems are adaptive immune mechanisms evolved by archaea and bacteria to 14" defend against foreign plasmids and viral DNA [5]. Manipulation of the Streptococcus pyrogenes 15" type II CRISPR/Cas system has been used to develop an efficient genome editing technique.

16"
First, a 20 bp sequence in a gene of interest is selected to act as a guide for the S. pyrogenes 17" nuclease, Cas9. This sequence, termed the CRISPR RNA (crRNA), has the only requirement that RNA molecule, termed the trans-activating crRNA (tracrRNA), is used for binding to Cas9 [6].

20"
For the purpose of experimental simplification, the crRNA and tracrRNA sequences can be fused

21"
into a single guide RNA (sgRNA) [7]. By expressing this sgRNA along with Cas9 in germ line

22"
cells, heritable genome mutations can be created.

23"
The CRISPR/Cas9 system has been successfully established in two leading nematode 24" models -C. elegans and Pristionchus pacificus [2,8]. Friedland et al. [9] developed a simple 25" protocol for C. elegans that involved injecting plasmids into the gonad of adult hermaphrodites.

26"
The authors modified Cas9 to include a SV40 NLS to ensure nuclear localization and expressed

27"
under an eft-3 translation elongation factor promoter, chosen for its effectiveness in germ line 28" expression. The sgRNAs were expressed under a U6 small nuclear RNA polymerase III 29" 4" " promoter, chosen for its ability to drive expression of small RNAs. As the optimal expression 1" from this promoter requires the first base to be a purine, the sgRNA target sequence is restricted 2" to the form (G/A)(N) 19 NGG [9, 10].

3"
Adaptation of CRISPR/Cas9 to C. briggsae, a species that is closely related to C. elegans, 4" would provide a powerful tool to investigate the function of any given gene. C. briggsae is used 5" routinely by many laboratories in comparative evolutionary studies. The two animals diverged 6" less than 30 million years ago yet share similar morphology [11]. A comparison of their genome 7" sequences has revealed that roughly one-quarter of their genes lack clear orthologs including 8" many that are highly divergent and species-specific [12]. This suggests that underlying gene characterizing phenotypes, we can learn the functional relevance of genomic differences,

13"
including any alterations in genetic pathways and developmental mechanisms between the two 14" species. With this goal in mind, we set out to develop a method for using this system in C.

16"
The wild type AF16 strain was used as a reference strain in all experiments. Strains

19"
We first used the CRISPR/Cas9 system in C. briggsae in an attempt to generate targeted 20" loss-of-function mutations by employing NHEJ. For this, two conserved genes were chosen 21" based on visible phenotypes, Cbr-dpy-1, a cuticle protein causing a dumpy (Dpy) phenotype, and

10"
Following microinjection, F2 worms were screened for desired phenotypes. We

11"
successfully isolated mutants for both Cbr-dpy-1 and Cbr-unc-22 at comparable frequencies to

12"
those observed in C. elegans (Table 1) [9]. Sequencing of the alleles of each of these genes 13" revealed insertions and deletions at the sgRNA target sites ( Table 2). The phenotypes of mutant 14" animals are indistinguishable from those in C. elegans corresponding to orthologous genes,

15"
demonstrating conservation of gene function. Together, these results show that the CRISPR/Cas9

16"
system works in C. briggsae and can utilize conserved C. elegans promoters to express sgRNAs 17" and Cas9.

18"
Next, we targeted six other conserved genes of the Wnt and Ras pathways (Cbr-lin-2,  were allowed to lay eggs for 24-36 hours, and then picked and lysed in pools of two. A region of 21" the genomic DNA spanning the sgRNA site (~200 bp) was amplified and examined on a 4%

22"
high-resolution agarose gel (Invitrogen UltraPure Agarose-1000, Catalog #16550-100) for 23" changes in band sizes (Figure S2). In some cases we recovered mutations as determined by 24" phenotypic as well as PCR-based screening approaches but none were found to be heritable 25" (Table 1). It is unclear to us whether it was due to sgRNAs being non-functional, less efficient or 26" requiring much larger F1s to be screened. Similar results were previously reported in C. elegans 27" [21]. In one case, Cbr-lin-17, we sequenced the animal that showed bi-vulva phenotype and 28" found possible evidence for a somatic mutation (T/A transversion causing M482L substitution).

29"
6" " The bi-vulva phenotype in this line was lost in subsequent generations. Evidence of somatic  Table 1. Phenotypes of transgenic animals generated using the CRISPR/Cas9 technique. The 3' target bases are those at positions 19 and 20 in the sgRNA target sequence. * One F2 showed Dpy phenotype. # 3 bivulva worms were recovered in F3 but the phenotype was not heritable. $ One F2 showed protruding vulva (Pvl) phenotype. @ wild type based on the C. elegans vit-2 mutant phenotype.

4"
Interestingly, our screens also recovered worms with unexpected phenotypes, e.g., Dpy in

2"
GG motif did not give rise to any mutation (Table 1). In the case of Cbr-bar-1, a β-catenin 3" homolog [26], the 3'GG motif sgRNA resulted in a disruption efficiency of 9.5% (Tables 1 and   4" 2). The enhanced efficiency of the 3'GG motif sgRNA sites for these two genes suggests that 5" such an approach in C. briggsae could improve the frequency of targeted mutations in genes of 6" interest.  Table 2. Alleles generated by the CRISPR/Cas9 approach. The DNA sequence includes the sgRNA target. The PAM site is bolded. Insertion and deletion sequences are underlined (dotted underline: insertion, solid underline: deletion). For clarity the 147 base pair inserted sequence in bh36 allele has been omitted. This long sequence matches with the E. coli gene EF-Tu. * The allele was recovered in a separate screen along with another allele bh32 that has small deletion. The exact base change in bh32 has not been determined.

9"
In addition to the CRISPR-mediated NHEJ approach we also attempted the HR method of 10" gene editing in C. briggsae. For this donor templates were designed to either disrupt a gene (by

11"
inserting a single-stranded oligonucleotide) or tag genes using double-stranded linear PCR

5"
short microhomology arms were generated by PCR to create translational fusions with Cbr-bar-1

12"
none of these HR approaches were successful, in some cases we did observe expected genomic

13"
changes in F1 and F2 animals (as determined by sequencing), which were not inherited in

14"
subsequent generations (Table 3).  Table 3. Genome editing events detected using CRISPR-mediated HR. The sgRNA efficiency shows all genome editing events, including those repaired by NHEJ and HR, based on phenotypic and PCR-based screens. HR efficiency indicates the number of HR events detected in F2 out of the total F1s screened. # Wild type based on the phenotype of C. elegans orthologs.

17"
In conclusion, we have shown that the CRISPR/Cas9 system can be effectively employed 18" in C. briggsae to alter a gene of interest. Similar to C. elegans the 3' GG motif appears to

19"
increase the frequency of NHEJ events. Interestingly, we observed a significant bias towards 20" 9" " insertion NHEJ events in C. briggsae. Of the total of 8 alleles recovered, for 4 different genes,

3"
work is needed to ascertain if such a bias in C. briggsae holds true in a larger sample size.

4"
Together with the recently developed TALEN-based genome editing approach [3], the 5" CRISPR/Cas9 approach described here provides a powerful means to investigate the functions of 6" conserved as well as divergent genes in C. briggsae. This promises to accelerate comparative 7" studies with C. elegans thereby leading to a greater understanding of the flexibility of genetic and 8" molecular mechanisms during animal development.

11"
We thank all members of the Gupta lab, particularly Scott Amon, Ayush Ranawade and Anand

12"
Adhikari, for their input and assistance throughout this project. We also thank John Calarco for   Figure S1. Donor sequence approaches generated as templates for HR for Cbr-bar-1.

1"
Templates may take the form of a donor vector (A), ssODN (B) or PCR amplicons (C). Blue

2"
letters represent the sgRNA target sequence while red letters represent the PAM site.