SaCas9 and KKH SaCas9 have robust gene editing activity in X. tropicalis
SaCas9 and KKH SaCas9 have been proven effective in a number of species including zebrafish [4, 26]. To test their mutagenic effects in X. tropicalis embryos, for each of them, we designed 10 gRNAs targeting 9 different genes. Following the strategy illustrated in Fig. 1, we injected each individual Sa gRNA (50pg/egg) together with either SaCas9 mRNA (300pg/egg) or KKH SaCas9 mRNA (300pg/egg) into fertilized eggs before the first cleavage. For each injected group, 18-24 hours after injection, 60 healthy embryos were collected and divided into two replicates (each 30 embryos) for genome extraction, PCR amplicon library construction, and high-throughput DNA sequencing. The deep sequencing data indicated that vast majority of the target sites tested showed profound targeting efficiency (Fig. 2). For SaCas9, half of the targeted sites (5/10) displayed indel formation frequency above 90% and only 2 appeared inefficient (Fig. 2a, c and Additional file 1: Fig. S1). For KKH SaCas9, the indel efficiency of 8 target sites is more than 60%, and that of other 2 is about 40% and 25%, respectively (Fig. 2b, d and Additional file 1: Fig. S2). Most indels detected are small deletions rarely exceeding 10 bp (Additional file 1: Fig. S1 and Fig. S2). Taken together, these data indicate that both SaCas9 and KKH SaCas9 can effectively edit the genome in X. tropicalis embryos.
VQR Cas9, SpG Cas9, SpRY Cas9, and xCas9 3.7 showed poor mutagenic activity in X. tropicalis embryos
To test more CRISPR/Cas variants for indel induction in X. tropicalis embryos, we designed 8, 15, 4, and 4 gRNAs for SpG Cas9, SpRY Cas9, VQR Cas9, and iSpyMac Cas9, respectively, and checked their mutagenic activity with the T7EI assay. Weak T7EI signals were detected in 2 SpG Cas9 targeting sites, 5 SpRY Cas9 sites, and 3 VQR Cas9 sites. None of the 4 iSpyMac sites displayed discernible T7EI signals (Additional file 1: Fig. S3). For the 10 T7EI positive sites, we further carried out deep sequencing analysis, which confirmed 9 T7EI signals, while the SpRY-best1-NAT site appeared to be a T7EI false-positive one with no indels detected by deep sequencing (Fig. 3a-d and Additional file 1: Fig. S4, Fig. S5 and Fig. S6). Gray value analysis revealed that for each site tested, the low value of the T7EI signal matched well with the corresponding low indel frequency detected by deep sequencing (Fig. 3a-d and Additional file 1: Fig. S3, Fig. S4, Fig. S5 and Fig. S6). Large number of T7EI positive samples have routinely been confirmed by Sanger DNA sequencing in our laboratory. Thus, T7EI assay is a reliable method for detecting targeted gene disruption in our hands.
As xCas9 3.7 has a broad range of PAM compatibility including NG, NNG, NGG, GAA, GAT, and CAA, we designed 143 non-NGG PAM gRNAs targeting 6 genes (tyr, ptf1a/p48, ctcf, apc, kcnj2, and bmp4), 1 NGG PAM gRNA targeting bace2 to cover all these PAMs and assessed their targeting efficacy with the T7EI assay. For NGG PAM sites, we also included 3 gRNAs targeting apc-163, apc-1714, and tbx5-237, which had been proven effective for SpCas9 [24]. Weak T7EI signals were detected at 51 non-NGG sites and 2 NGG sites. Gray value analysis revealed that editing efficiency for most of the positive ones was about 5%. Only 3 sites displayed efficiency slightly over 50% (Fig. 3e and Additional file 1: Fig. S7, Fig. S8a). We selected 4 non-NGG sites and 2 NGG sites with different T7EI signal intensities and carried out a Sanger DNA sequencing verification. The indel frequency detected by Sanger sequencing confirmed the T7EI signals (Fig. 3f and Additional file 1: Fig. S7, Fig. S8c). For NGG sites, the editing efficiency of xCas9 3.7 is much lower than that of SpCas9 (Additional file 1: Fig. S8, and [24]). Collectively, these data indicate that VQR Cas9, SpG Cas9, SpRY Cas9, and xCas9 3.7 do have certain gene editing efficacy with low efficiency in X. tropicalis. But they cannot serve as robust gene editing tools in frogs.
LbCas12a/crRNA RNP complexes are effective for DNA fragment deletion in X. tropicalis
Targeted genome DNA fragment deletion is key to establish animal models of human genetic disorders and to study the function of lncRNA, miRNA, cis-regulatory elements, as well as functionally uncharacterized genome regions. To test the activity of the LbCas12a/crRNA RNP complexes in inducing segmental deletion in X. tropicalis, we designed two pairs of crRNAs targeting two fragments on the tyr locus. For both pairs, gray value analysis of the PCR products on the agarose gel showed that the quantity of the fragments with deletions was roughly the same as that of wild-type fragments, which was confirmed by Sanger DNA sequencing, revealing deletion efficiency of about 50% (Fig. 4a, b). We further designed 4 crRNAs targeting 4 genes (ptf1a/p48, sftpb, tbx5 and ctcf) and tested the mutagenic activity of the LbCas12a/crRNA RNP complexes with the T7EI assay. The data obtained indicate that 3 sites displayed editing efficiency above 50%, of which 2 were further confirmed by Sanger DNA sequencing (Fig. 4c, d). Most indels induced by the LbCas12a/crRNA RNP complexes are deletions that are obviously longer than those induced by SaCas9 or KKH SaCas9(Fig. 4d and Additional file 1: Fig. S1, Fig. S2). Together, these data indicate that LbCas12a/crRNA RNP complexes are highly effective for DNA fragment deletion and indeed robust for gene editing in X. tropicalis.
Phenotyping G0 embryos mutagenized by SaCas9 and KKH SaCas9
Given the high efficiency of gene disruption induced by SaCas9 and KKH SaCas9, knockout phenotypes could be expected in their mutagenized G0 embryos. We have shown that albinism and heart defects could be easily identified upon direct disruption of tyr and tbx5, respectively, in X. tropicalis embryos [21, 24]. Indeed, varying levels of albinism, heart defects, and heart failure related edema depending on the corresponding gene disruption efficiency (effective frameshifts) were observed in SaCas9- or KKH SaCas9-mediated tyr and tbx5 disruption, respectively (Fig. 5; Additional file 1: Fig. S9). These data indicate that SaCas9 and KKH SaCas9 can be used as tools for G0 phenotyping in X. tropicalis.
SaCas9, KKH SaCas9, and LbCas12/crRNA RNP complexes all have low off-target effects in X. tropicalis
To further test the specificity of SaCas9, KKH SaCcas9, and LbCas12/crRNA RNP complexes, we identified genome-wide potential off-target sites with less than five mismatches for 2 SaCas9 targeting sites, 2 KKH SaCas9 targeting sites and 3 LbCas12a targeting sites, with the online software CAS-OFFinder ([27]; Additional file 2: Table S1), of which we selected 23 for T7EI assay (Additional file 2:Table S2). No T7EI signals were detected in these samples (Additional file 2: Table S2, Fig. S10), suggesting low off-target effects of SaCas9, KKH SaCas9, and LbCas12a in X. tropicalis embryos, which is consistent with the SpCas9 activity in frog embryos [21].