Design of CRISPR/Cas9 system
It has reported that PLCD1 is required for skin stem cell lineage commitment and PLCD1 knockout mice with disruption of X and Y domains (corresponding to exons 7 and 8) by homologous recombination in mouse ES cells undergo progressive hair loss in the first postnatal hair cycle [34,35,36]. To generate PLCD1-modified mice by CRISPR/Cas9 technology, exons 2 (beginning of the translation PLCD1), 7 and 10 (corresponding to X and Y domains, respectively) of PLCD1 were scanned for potential sgRNA target sequences using Optimized CRISPR Design (http://crispr.mit.edu). Three sgRNAs were designed (Additional file 1: Table S1), and then the efficiency was estimated by SSA (Single-strand annealing) assay in vitro, as described previously [23]. Finally, sgRNA targeting exon 2 was chose to modify PLCD1 (Fig. 1).
Generation of PLCD1-modified mice
PLCD1-modified mice were generated via co-microinjection of PLCD1-sgRNA (targeting exon 2) and Cas9 mRNA into the cytoplasm of C57BL/6J zygotes. Three injection sessions yielded 28 pups, 26 of which survived until weaning (Additional file 1: Table S2). Among these survived pups, 24 pups began to shed hair at 2 weeks old, and showed bald abdomen and sparse hair on back and head at 3 weeks old (Fig. 2a, b). At 3 months, a variety of hair loss were observed, 5 (#2, #4, #11, #21 and #25) out of 24 mice remained nearly hairless on abdomen, and sparse hair on back and head (~ 19%) (Fig. 2a, b). The rest of two mice (#23 and #38) displayed normal hair.
All mice (n = 26) were genotyped by sequencing for CRISPR/Cas9-induced mutations in exon 2 of the PLCD1 locus. Amplified DNA fragments by PCR using one primer pair that flanks the sgRNA target sequence were subcloned, and subsequently sequenced. The mutations around the target site were identified in 11 of 26 mice (~ 42.3%), and four mice (~ 15.4%) were bi-allelic mutations (Fig. 2c).
Generation of PLCD1-deficient mice (F1)
To obtain mutant mice homozygous(−/−) for PLCD1 mutation, the conventional method is mating PLCD1-modified mice (F0) with wild-type mice to obtain mutant mice heterozygous (∓) for PLCD1 mutation, and then PLCD1 heterozygous mutant mice were intercrossed to produce the homozygous PLCD1-deficient mice. Thus, it is clear that the conventional method is time-consuming. In this study, PLCD1-modified mice (F0) (#11, #21 and #25) with similar mutation were intercrossed to produce PLCD1 homozygous mutant mice (F1). Finally, we obtained 7 PLCD1-deficient mice (F1), among which 4 PLCD1-deficient mice (#40, #41, #42 and #43) were produced by breeding mouse #11 with mouse #25, and three PLCD1-deficient mice (#44, #45 and #46) were generated by breeding mouse #11 with mouse #21. 6 out of 7 mutant mice began to shed hair at 2 weeks old which was observed in PLCD1-modified mice (F0), and demonstrated bald abdomen and sparse hair on back at 3 weeks old which is more obvious than that of PLCD1-modified mice (F0). Furthermore, the PL CD1-deficient mice (F1) showed bald head and neck at the age of 3 weeks, which was not observed in PLCD1-modified mice (F0). The distribution of abdominal and dorsal hair remained the same stage at the age of 6 weeks. The hair of PLCD1-deficient mice (F1) was softer and thinner than that of wild-type mice and PLCD1-modified mice (Fig. 3a, b).
All of PLCD1-deficient mice (F1) (n = 7) were genotyped by sequencing for CRISPR/Cas9-induced mutations as mentioned above. The anticipated mutations were identified in 6 out of 7 mice (Fig. 3c). Among 6 PLCD1 mutant mice (F1), there are three homozygous PLCD1-deficient mice (F1), including 2 mice (#40 and #43) with deletion of 7 bp and one mouse with deletion of 11 bp and insertion of 1 bp. In summary, homozygous PLCD1-deficient mice were obtained in first-filial generation, indicating that intercrossing of genetically modified mice (founder; F0) produced by CRISPR/Cas9 technology to generate homozygote is feasible and time-saving.
The histological analysis in the skin of PLCD1-deficient mice
Almost all PLCD1-modified mice displayed the defect in the fur development, whereas control mice had well-developed coats (Figs. 2a, b and 3a, b), which prompted us to investigate skin histological structure of PLCD1-deficient mice (Figs. 4 and 5). We found that the number of hair shaft decreased in the skin of PLCD1-deficient mice (Figs. 4C–F and 5C–F), compared with that of wild-type mice (Figs. 4A, B and 5A, B). Additionally, most of hair shaft in PLCD1-deficient mice are abnormal and fail to penetrate the epidermis (Figs. 4C–F and 5C–F), whereas hair shaft in wild-type mice penetrate the epidermis (Figs. 4A, B and 5A, B). In PLCD1-deficient mice, some hair canals were occluded by differentiated keratinocytes (Figs. 4D, F and 5D, F). Moreover, interfollicular epidermis (IFE) of PLCD1-deficient mice (Figs. 4C–F and 5C–F) was thicker than that of wild-type mice. Together, these findings demonstrate that PLCD1-deficient mice display abnormal histology, in the skin which is almost similar to those of nude mice.
PLCD1 and other related gene expression in the skin of PLCD1-deficient mice
Because abnormal epidermal and hair follicle morphologies were observed in skin from PLCD1-deficient mice, we investigated how the lack of PLCD1 influences the cell proliferation and differentiation in these structures. Thus, the expression levels of PLCD1 and other members of PLC family/delta family (i.e., PLCB1, PLCG1 and PLCE1), and these genes (i.e. Ki67, Krt1, Krt5, Krt13, loricrin and involucrin) involved in the growth and differentiation of hair follicle and epithelial tissues [37,38,39,40,41] were examined in the skin tissues of PLCD1-deficient mice by IHC (Fig. 6) or qRT-PCR (Fig. 7). The expression level of PLCD1 was very significantly decreased (Figs. 6A, B and 7), suggesting that CRISPR/Cas9-induced mutation in exon 2 of PLCD1 locus results in PLCD1 deficiency in homozygous PLCD1-deficient mice. There were no significant differences in the expression levels of PLCB1, PLCG1 and PLCE1 between wild-type mice and PLCD1-deficient mice (Fig. 7). More importantly, the expression levels of Krt1, Krt5, Krt13, loricrin and involucrin involved in the differentiation of hair follicle and epithelial tissues were remarkerably increased in skin tissues of PLCD1-deficient mice (Figs. 6E, F and 7), while the number of hyperproliferative Ki67-positive cells in skin tissues of PLCD1-deficient mice were significantly increased compared with control (Fig. 6C, D). Summarily, our results reveal the epidermal hyperplasia and disturbed differentiation of epidermis in PLCD1-deficient mice.
Off-target analysis of PLCD1-deficient mice
Off-target effect is a major drawback concern of the CRISPR/Cas9 system [42, 43]. To examine whether off-target occurred in these genetically modified mice, possible off-target sequences within mouse genome were predicted using the CRISPR Design Tool (http://crispr.genome-engineering.org). The sgRNA targeting PLCD1 can potentially recognize 40 putative off-target sequences that have variable numbers of base mismatches. Only three of these are exonic sites (Additional file 1: Table S4). Approximately 500–800 bp genomic fragments containing off-target site were amplified by PCR using the primers listed in Additional file 1: Table S5, and subject to sequencing analysis. As a result, none of the sequencing reads exhibited any mutations, suggesting that no off-target occurred in any of 6 PLCD1-deficient mice (F1).