CRISPR/Cas9 mediated genetic resource for unknown kinase and phosphatase genes in Drosophila

Kinases and phosphatases are crucial for cellular processes and animal development. Various sets of resources in Drosophila have contributed significantly to the identification of kinases, phosphatases and their regulators. However, there are still many kinases, phosphatases and associate genes with unknown functions in the Drosophila genome. In this study, we utilized a CRISPR/Cas9 strategy to generate stable mutants for these unknown kinases, phosphatases and associate factors in Drosophila. For all the 156 unknown gene loci, we totally obtained 385 mutant alleles of 105 candidates, with 18 failure due to low efficiency of selected gRNAs and other 33 failure due to few recovered F0, which indicated high probability of lethal genes. From all the 105 mutated genes, we observed 9 whose mutants were lethal and another 4 sterile, most of which with human orthologs referred in OMIM, representing their huge value for human disease research. Here, we deliver these mutants as an open resource for more interesting studies.

all indicated their important functions, while perhaps due to current limited resources, their functions have not been defined yet. Thus, we planned to generate stable mutants to define the function of all these 156 unknown kinases and phosphatases.

CRISPR/Cas9 mediated mutagenesis of the unknown candidates.
To ensure the mutagenesis efficiency and completely disrupt the gene function, we designed two gRNAs per gene around the translational start codon or at least before 1/3 of the coding sequence when carrying out mutagenesis according to the standard CRISPR/Cas9 strategy. In case that these unknown genes could be lethal when mutated, we used different FRT flies for microinjection according to their genomic distribution in order to perform tissue-specific mosaic analysis. Through separate operation on the five individual FRT groups, we finally obtained 385 mutant alleles for 105 genes from all the 156 candidates and another 33 s failed due to few F0, indicating high probability of lethal genes, which we called potential lethal genes. While for the remnant 18 failure, we found they were nothing with the gRNA working efficiency, length, sequence specificity or PAM sequence specificity (Supplementary Figure 1), which perhaps mainly caused by limited detected F0/F1. Consistent with results from other organisms 18 , we also detected insertions, deletions including frameshift and in-frame mutations among the 385 alleles (Fig. 2, Supplementary Note 1), in which case there were 11 genes of the 105 with only in-frame mutants in the resource. Of the 385 mutant alleles, we randomly chose 6 mutants to detect the mRNAs of the corresponding mutant genes using qRT-PCR. As expected, we observed reduced mRNAs to different degrees of detected genes (Supplementary Figure 2), demonstrating our resource with corresponding gene deficiency.
Obvious phenotypes in the CRISPR/Cas9 resource. Considering complicated functions of the kinases and phosphatases, we performed simple lethality and sterility screening for the resource only in this study. Taken together, we observed 9 genes with lethal mutants, 1 gene with male sterile mutants and 3 genes with female sterile www.nature.com/scientificreports www.nature.com/scientificreports/ mutants, these genes were predicted involved in various developmental processes (Table 1). To be recommended, of the 3 female sterile events, mutations of CG3608 were all in-frame styles, implicating the importance of the missed amino acids in the mutants, meanwhile, we could not exclude the possibility that there were off-targeting effects in the mutants. The same situation occurred in the rescue experiments for the phenotypic mutants. Of the 13 phenotypic mutants, we totally generate 4 transgenic lines including CG33671, CG17028, CG14305 and CG15743 to rescue the corresponding mutants, but we found the CG17028 transgenic lines could not rescue the lethality of CG17028 mutants, which might be caused by ectopic transgenes or off-targeting effects. Besides these, we also observed other abnormal lethal events in the resource, such as CG12229 and CG17027, with different lethality between similar or same genotypic mutants, even one line of CG17028, without any mutation detected near the gRNA recognition sites (Supplementary Figure 3, Supplementary Note 2). Anyway, despite uncertainty of the total resource, we could define more certain phenotypes of unknown candidates such as lethality and sterility. We look forward to more interesting phenotypes in our resource.

Discussion
Our work in this study delivered a valuable genetic resource for unknown kinase, phosphatase and associate genes in Drosophlia. Using this resource, we can carry out lots of meaningful and interesting screenings besides of lethality and fertility. Of the lethal and sterile genes identified in this study (Table 1), we re-searched for gene function and found some of them had been experimentally investigated, such as CG34380, proved to be required for normal thermal nociception in a RNAi-based screening study 19 ; CG6767, proved to affect olfactory behavior using P-element insertional mutagenesis together with targeted RNAi 20 ; and CG34455, prove to encode a www.nature.com/scientificreports www.nature.com/scientificreports/ pyridoxal kinase and contribute to chromosome integrity and glucose homeostasis through mutation isolated from EMS-induced late lethals 21 , our identification of which providing direct evidence for their important roles in multiple developmental processes and efficient genetic tools to address associated issues.
During the generation of this resource, we observed three insights into large-scale CRISPR/Cas9 application. First, the regular CRISPR/Cas9 system is not as well suited to screening lethal genes by high-throughput mutagenesis as it is to viable genes in vivo, and it was completely unable to outperform genetic screens as observed in cell lines 22 . For the 33 s failure due to few F0s in the resource, more efforts were needed to get stable mutant lines through regular CRISPR/Cas9.
Second, based on roughly calculation between the mutation yield and gRNA working efficiency, length, sequence, PAM sequence, we found these factors seem do nothing with the mutation yield, except for a 19-nt GGG gRNA pattern, which perhaps caused by the single 17-G (Supplementary Figure 1C-D), consistent with results previously described 23 .
Finally, we observed some potential off-targeting effects during the generation of our resource. In addition to the confirmed abnormality, the abnormal lethality that could not be rescued for CG17028 hinted at the increasing off-targeting possibility of the remaining viable alleles. Regretfully, we sequenced only about 200 bp near the gRNA recognition sites of these abnormal lethal lines, thus we cannot exclude the possibility that there were different mutations between the lethal lines and viable lines for the same gene in the target gene region far away from the gRNA recognition sites 24 , another side-effects caused by CRISPR/Cas9 treatment. Regardless, our results present there did exist off-targeting or side-effects during the CRISPR/Cas9 mediated mutagenesis, suggesting the need for more specific CRISPR/Cas9 25 or more efficient and precise genome editing methods for gene therapy of diseases. T7EI assay and mutation identification. Dead larvae or single flies were squashed in 30 uL of squashing buffer containing 10 mM Tris-Hcl (PH 8.0), 1 mM EDTA, 25 mM NaCl, 1 mg/mL proteinase K (Takara, Beijing, China), and incubated at 37 °C for 1 h, followed by heating to 95 °C for 2 min and used as PCR templates. PCR was performed in a 2 x Taq MasterMix (Aidlab Biotech, Beijing, China) under standard PCR procedure, with primers 100-200 bp away from the gRNA targeting site. The PCR products were then digested with T7 Endonuclease I (Viewsolid Biotech, Beijing, China), and T7EI positive F1 PCR products were sequenced for mutants. Fertility screening. Every three males or virgin-females of our resource with viable homozygous were crossed with w 1118 virgin-females or males. Parents were kept for 7 days and discarded. Then adult flies of each cross were counted for fertility of the viable resource. The ones with no offspring were defined as sterile.  Table 1. Obvious phenotypes in the CRISPR/Cas9 resource. a The lethality phenotype of CG33671 and male sterility phenotype of CG14305 both can be rescued by their own original genomic transcripts. b The lethality phenotype of CG17028 cannot be rescued by its original genomic transcript. # The female sterility phenotype of CG3608 was observed from the in-frame mutant alleles, because we did not obtain frameshift mutant alleles for CG3608 in our resource.