Vector construction for the CRISPR/Cas9 system
Initially, we aligned the full-length CDS of NtMLO1 and NtMLO2 to identify a site to design sgRNA targeting both genes simultaneously. The alignment showed 95.62% similarity between the NtMLO1 and NtMLO2 sequences (Supplementary Fig. 1). To increase the chances of obtaining complete loss-of-function mutations, a site targeting a conserved region, including a 20 bp nucleotide with the protospacer adjacent motif (PAM) at the 3’ region, was chosen within exon 1 of these genes (Fig. 1A). The sgRNA was cloned into the pORE-Cas vector to generate the pORE-Cas-gRNA-MT construct used to transform the susceptible tobacco cultivar ZY300 (Fig. 1B).
CRISPR/Cas9-mediated mutagenesis of NtMLO
In total, 20 transgenic plants (T0) were obtained from the leaf discs transformed with the Agrobacterium (Supplementary Fig. 2). The genomic DNA of these T0 individuals was amplified by PCR using NtMLO1-specific (MJC2F/MJC2R) and NtMLO2-specific (MJC8F/MJC8R) primers and sequenced to characterize the mutations. Sequence analysis showed that 13 out of the 20 T0 transgenic plants had heterozygous mutations in NtMLO1, 12 had heterozygous mutations in NtMLO2 (Table 1, Supplementary Fig. 3). Among these mutants, 10 plants had heterozygous mutations in both genes. The approach achieved a mutation frequency of 65.0% in NtMLO1 and 60.0% in NtMLO2. Approximately 66.7% of the mutants had both genes mutated (Table 1).
We further characterized the T1 progenies derived from the T0 plants to obtain the homozygous mutants. Plants homozygous for the mutant alleles were selected from the T1 segregating lines by sequencing the specific PCR fragments. Three homozygous mutation types were identified for NtMLO1, including a 1 bp deletion (-T), a 2 bp deletion (-GA), and a 1 bp insertion (+A) (Fig. 2A). For NtMLO2, two homozygous mutation types, including a 1 bp insertion (+A) and an 11 bp deletion (-ACGGTCGATGG), were identified (Fig. 2B). We identified individual plants with simultaneous homozygous mutations in both genes from the progenies of the 10 T0 plants (C1, C5, C7, C10, C11, C12, C13, C18, C19, C20). Four different mutation events were identified in these simultaneous homozygous mutants (Fig. 2C). Event 1 resulted in a 1 bp deletion in NtMLO1 and a 1 bp insertion in NtMLO2 in the T1 plants obtained from the T0 lines C5, C10, C11, C12, and C20. Event 2 resulted in a 2 bp deletion in NtMLO1 and a 1 bp insertion in NtMLO2 in the T1 plants from the T0 lines C1, C7, and C18. Event 3 generated a 1 bp insertion in NtMLO1 and an 11 bp deletion in NtMLO2 in the T1 plants obtained from the T0 line C13. Event 4 created 1 bp insertions in NtMLO1 and NtMLO2 in the T1 plants obtained from the T0 line C19. We then obtained the self-pollinated seeds from the homozygous T1 mutants for the further analysis.
Off-target analysis
Furthermore, we predicted the putative off-target mutations induced by the sgRNA in the whole genome using Cas-OFFinder (Bae et al. 2014), allowing a mismatch of three or fewer nucleotides between the sgRNA and the potential off-target regions to assess the specificity of this CRISPR/Cas9 system. Four potential off-targets with a PAM sequence were identified in the whole genome (Supplementary Table 2), of which one off-target, located on chromosome AYMY01S000835.1, had two mismatches, and the other three off-targets, located on chromosomes AYMY01S001145.1, AYMY01S001369.1, and AYMY01S024885.1, had three mismatches. We then checked for these potential off-target sites in three plants randomly selected from the homozygous T2 lines C1-7 and C13-21 by PCR using different primer combinations (Supplementary Table 1). Sequencing showed no differences between the wild type and the mutant on these four potential sites (Supplementary Fig. 4), suggesting that no off-target mutation occurred in the potential sites.
Resistance of mutants to powdery mildew
Finally, to assess the impact of CRISPR/Cas9-induced mutations on powdery mildew resistance in tobacco, the T2 lines with homozygous mutations in only NtMLO1 (C2-8 and C8-10), only NtMLO2 (C3-1 and C1-15), and both NtMLO1 and NtMLO2 (C1-7 and C13-21) (Supplementary Fig. 5) were inoculated with Gc. No apparent disease symptom was observed on the leaves of the C1-7 and C13-21 lines with homozygous mutations in both NtMLO1 and NtMLO2 (Fig. 3A, Supplementary Fig. 6). Meanwhile, fungal growth was obvious in the wild-type plants and the mutants with only one edited gene (Fig. 3A). Consistent with these findings, the DI and relative Gc biomass on the leaves of the C1-7 and C13-21 lines were significantly lower than the wild-type plants. However, no significant differences in the disease index or the relative Gc biomass were observed among the wild-type plants and single mutants (Fig. 3B, C). These results indicated the role of both NtMLO1 and NtMLO2 genes in responding to powdery mildew infection, and mutating both genes could confer resistance to powdery mildew in tobacco.