Enhancing canola breeding by editing a glucosinolate transporter gene lacking natural variation

A low seed glucosinolate resource was developed in polyploid B. napus using a method that identifies the functions of genes with rare or no genetic variation.

Dear Editor, A major challenge in current genetics study is to identify functions of genes with rare or no genetic variation through forward genetics approaches, such as quantitative trait locus mapping and association study in germplasm, and particularly in polyploid crops, it is difficult to study functional differentiation of duplicated genes. Here, we report a causal gene with rare mutation in glucosinolate transport and creation of a genotype of low seed glucosinolate for quality and resistance breeding of polyploid Brassica napus canola, the second largest global source of edible oil and protein meal.
Glucosinolates are well-known secondary metabolites with important biological and economic roles in plant defense to disease and insects and human nutrition/health, such as anti-cancer effect (Sønderby et al., 2010). However, high seed meal glucosinolates can cause goiter and other harmful effects. Thus "double-low" (low seed glucosinolate and low erucic acid content) canola breeding was initiated in the mid-20th century, which has dramatically reduced seed glucosinolate content from .100 mmolÁg -1 to 530 mmolÁg -1 (Kondra and Stefansson, 1970). Unfortunately, glucosinolate content in vegetative tissues also decreased dramatically due to disruption of the biosynthesis pathway, which caused decreased plant disease resistance . Although the pathways of glucosinolate biosynthesis in vegetative tissues and subsequent transport to developing seeds have been well-characterized in Arabidopsis thaliana  and mutation of the genes encoding glucosinolate transporters (GTRs) substantially reduced seed glucosinolates in Brassica plants (Nour-Eldin et al., 2012, low seed glucosinolate germplasms already created by gene-editing in B. napus cannot be applied for breeding because of their negative effects on other traits (Tan et al., 2022), probably due to involvement of edited genes in other trait formation.
To address the above problems, first we comprehensively identified all homologs of A. thaliana GTRs from 12 de novo assembled B. napus genomes (Chalhoub et al., 2014;Zou et al., 2019;Lee et al., 2020;Song et al., 2020;Supplemental Figure S1) and those from one high glucosinolate cultivar ZY821 were divided into three subclades showing sequence differentiation of GTR1, GTR2, and GTR3 ( Figure 1A). Then we detected expression patterns of all 20 BnaGTRs in 20 stages/tissues by RNA-seq from the representative B. napus accessions with high (cv. ZY821) and low (cv. ZS11) seed glucosinolate content. The expression patterns of these Manes.15G180400 Each error bar is means ± SD (n 5 4 plants for each line). Student's t test was used for statistical analysis, single asterisk indicates significant differences at P 5 0.05, double asterisk indicates significant differences at P 5 0.01, all compared to the BnaA06.GTR2 WT .
BnaGTRs are very complicated and there was no clear cue to judge their contribution to glucosinolates except for BnaA06.GTR2 and BnaA02.GTR2 showing higher expression in some tissues ( Figure 1B; Supplemental Table S1). Subsequently, we performed a genome-wide association study on seed glucosinolate content of 312 diverse B. napus germplasm accessions collected worldwide (Supplemental Table S2). Based on single-nucleotide polymorphisms genotyped by whole-genome resequencing and aligned on ZY821 genome, three loci harboring five BnaGTR2s on chromosomes A02, A09, and C02 were significantly associated with seed glucosinolate content ( Figure 1C). Surprisingly, four BnaGTRs located in significantly associated loci lowly or moderately expressed, while BnaA06.GTR2 and BnaC03.GTR2, whose genomic regions did not show significant association, exhibited high expression in silique wall of late development stages ( Figure 1B) and BnaA06.GTR2 showed higher expression levels than those of BnaC03.GTR2.
To verify BnaA06.GTR2 as a crucial player in seed glucosinolate accumulation and create variation for low seed glucosinolate breeding, we conducted clustered regularly interspaced short palindromic repeats/CRISPR-associated system 9 (CRISPR/Cas9) genome editing against BnaA06.GTR2. Two single-guide RNAs (sgRNAs) were designed to specifically target the third exon of BnaA06.GTR2 (Figure 1, D and E; Supplemental Figure S2). Then the pHSE401-based vector Cas9-BnaA06.GTR2 was transformed into ZY821 through Agrobacterium tumefaciens-mediated hypocotyl transformation (Xing et al., 2014), and 45 independent positive T 0 transgenic plants were obtained. Among them, three plants (named BnaA06.gtr2 CR-1 , BnaA06.gtr2 CR-2 , and BnaA06.gtr2 CR-3 ) in which homozygous small deletions occurred at both target sites for gRNA1 and gRNA2 ( Figure 1E), produced three putative truncated proteins with respective sizes of 174, 168, and 173 amino acids. Then the three genome-edited plants were selfpollinated, and genome editing events were successfully transmitted to the T 1 generation and no additional mutation was identified in the progenies of the three transgenic plants (Supplemental Table S3). High-performance liquid chromatography measurement showed a sharp reduction (76.05% ± 0.76% reduction on average) in total seed glucosinolate content with an average content of 32.6 ± 11.9, 31.5 ± 6.8, 30.6 ± 13.4 mmolÁg -1 in the three transgenic lines, respectively, while the transgene-negative control (BnaA06.GTR2 WT ) was 131.8 ± 8.3 mmolÁg -1 in mature seeds ( Figure 1G; Supplemental Table S4; "Methods"). Predominant aliphatic glucosinolates existing in seed also showed significant reductions in three edited lines. The content of the indolic glucosinolates, accounting for a relatively small proportion in the total glucosinolate content, was not affected by BnaA06.GTR2 knockout or even showed a slight increase in seeds. In developing seeds and silique walls at 40 d after pollination, the content of aliphatic, indolic, and total glucosinolates showed similar results as those in mature seeds ( Figure 1H; Supplemental Figure S3 and Supplemental Table S4) although relative amount of aliphatic and indolic glucosinolates showed a slight difference from that of mature seeds. Furthermore, the loss-of-function mutants of BnaA06.GTR2 showed no apparent alteration in morphological and yield-related traits ( Figure 1F; Supplemental Figure  S4 and Supplemental Table S5) in normal growth condition, and could serve as a good low seed glucosinolate germplasm. These results contrast with published data (Tan et al., 2022) that showed a $33.96% decline in seed glucosinolate content in a line (only one line) of CRISPR/Cas9edited BnaA06.GTR2 which was accompanied by seed weight decrease. The reason for the difference is probably due to the mutation difference: that line harbors one degenerate base of synonymous substitution at the 3 0 -end of coding sequence (CDS) of BnaA06.GTR2, but the three lines in the present study have small deletions resulting in three putative truncated proteins; unlike editing of multiple BnaGTR2 genes (Tan et al. 2022), mutation of the major player of aliphatic glucosinolate transporter BnaA06.GTR2 did not have negative effect on yield-related traits, suggesting that BnaA06.GTR2 may not be responsible for other trait formation.
In summary, we provided a method to identify functions of genes with rare or no genetic variation in natural populations, we used genome-editing to create a low seed glucosinolate resource that is rare in the germplasm, and we proved BnaA06.GTR2 with rare natural mutation is crucial in seed glucosinolate accumulation in polyploid B. napus. The combination of our results and the published ones (Tan et al. 202) suggests that, probably due to subfunctionalization or neofunctionalization of these duplicated genes, simultaneous editing of polyploidy multiple duplicated genes or a gene family may not be an effective strategy for mutation of a target trait.

Supplemental data
The following materials are available in the online version of this article.
Supplemental Figure S1. Neighbor-joining tree of all GTRs homologs in A. thaliana and 12 B. napus genomes.
Supplemental Figure S3. Glucosinolate content of BnaA06.GTR2 knockout mutants and BnaA06.GTR2 WT control in developing silique walls at 40 d after pollination.
Supplemental Table S1. Gene expression level of all BnaGTRs in 20 stages/tissues from ZY821 and ZS11.
Supplemental Table S2. The list of 312 B. napus germplasm accessions used for genome-wide association studies in this study.
Supplemental Table S3. Detection of mutations at potential CRISPR/Cas9 off-target sites in the T 1 generation. Table S4. Glucosinolate content (lmol/g) in different tissues between BnaA06.gtr2 mutants and their control in the T 2 generation.

Supplemental
Supplemental Table S5. Thousand seed weight (g) of BnaA06.gtr2 mutants and their control in the T 2 generation.
Supplemental Table S6. Primers used in this study.