Fusarium ear rot (FER), a fungal disease caused by Fusarium spp., leads to great yield losses to maize worldwide and to mycotoxin contamination, which seriously threatens human and animal health (Yao et al., 2020). FER has become an especially serious problem in maize–wheat rotation systems because the major causal agent can also lead to Fusarium head blight (FHB) in wheat (Li et al., 2019). Employing resistant cultivars is the most effective method for controlling FER but is currently limited to only a few resistance genes. Fhb1, a major quantitative trait locus (QTL) for FHB resistance, was cloned recently (Li et al., 2019; Su et al., 2019). Although the reported resistance mechanisms were controversial, TaHRC underlying Fhb1 was confirmed to be a susceptibility gene; a 752-bp deletion in this gene led to a loss of susceptibility (Li et al., 2019; Su et al., 2019). Knockout of the susceptible TaHRC allele improved FHB resistance (Chen et al., 2022; Su et al., 2019). Because FER and FHB could be caused by the same fungus (Li et al., 2019; Yao et al., 2020), we hypothesized that engineering TaHRC homologues in maize could represent a novel approach to FER resistance breeding.

Two TaHRC homologues have been identified in maize: Zm00001d040118 (designated as ZmFER1) and Zm00001d008500 (Li et al., 2019). Several QTLs underlying resistance to FER have also been identified to be adjacent to or overlap with ZmFER1 (Ding et al., 2008; Kebede et al., 2016; Yao et al., 2020; Figure 1a). Using the resistant maize inbred lines Shen137 and Qi319 (Hu et al., 2021) and the susceptible line B73 (Kebede et al., 2016) and wild-type ZC01 (WT), we identified 12 polymorphisms in the coding sequence of ZmFER1, resulting in two haplotypes (Figure 1b). A 6-bp deletion led to the absence of two amino acids in the ZmFER1 histidine-rich region in the resistant lines (Figure 1b). By contrast, no polymorphism was detected in the coding sequence of Zm00001d008500 among the four inbred lines (data not shown). Based on these results, we reasoned that ZmFER1 might participate in FER resistance in maize.

To test this hypothesis, we constructed a CRISPR/Cas9 vector (Figure 1c) targeting ZmFER1 ~25 bp downstream of the translation start codon (Figure 1d) and stably transformed WT with this CRISPR/Cas9 vector. Using targeted PCR and sequencing, we selected three T₀ homozygous events: E1 (1-bp insertion), E2 (1-bp insertion) and E3 (5-bp deletion), which were predicted to produce truncated proteins of 35, 14 and 33 amino acids, respectively (Figure 1b,e). We selected transgene-free individuals from T₁ plants lacking Cas9 and Bar, as described previously (Liu et al., 2022) and used the T₁ and transgene-free T₂ and T₃ plants for pathogen resistance assays.

Fusarium verticillioides was used as the pathogen (Figure 1f). The sequence of a 654-bp fragment on chromosome 1 in the gene (GeneID: 30067583) of the pathogen was 100% identical to the Fusarium verticillioides sequence at NCBI (accession no. NC_031675.1) (Figure 1f). We mixed seeds from the WT and each mutant separately at a 1:1 ratio for blind artificial inoculation. Eight days after self-pollination, we performed inoculation of the ears through the silk channel using a 2 mL conidial suspension (1 × 10⁶ spores/mL) and covered each ear with a bag to reduce natural infection (Ding et al., 2008). We evaluated FER resistance by counting diseased kernels per ear (DKPE) at 40 days after inoculation. We performed artificial inoculation in three environments: Beijing (39.9°N, 116.3°E) in the summer of 2020, Beijing in the summer of 2021, and Hainan (18.1°N, 109.3°E) in the winter of 2021. The non-inoculated WT and E1 ears showed no symptoms of FER (Figure 1g). By contrast, inoculated WT, E1, E2 and E3 ears showed obvious symptoms of FER, including ‘starburst’ kernels, and scattered or groups of kernels with white mould (Figure 1g). These results indicated the effectiveness of artificial inoculation.

To evaluate the resistance of mutants to FER, we sequenced the target site of ZmFER1 in all inoculated ears and counted the DKPE values. The ears were categorized into two groups (WT and mutant) based on the genotype of ZmFER1. We determined the difference in DKPE between the two groups using Student's t-test. In Beijing in the summer of 2020, the differences in DKPE between the WT and mutants were all statistically significant at P = 0.01 (Figure 1h), with average DKPE values of 28.65 and 9.54 for the WT and mutants, respectively. In Beijing in the summer of 2021, the DKPE values for E1, E2 and E3 were also significantly different from the WT at P = 0.01 (Figure 1h), with average DKPE values of 8.70 and 4.08 for the WT and mutants, respectively. In Hainan in the winter of 2021, the DKPE values for E2 and E3 were also significantly different from the WT, but the DKPE for E1 was not. This is consistent with the finding that FER resistance is significantly influenced by the environment (Yao et al., 2020). As determined using the National Standards of the People's Republic of China (GB5009.240-2016), diseased kernels had much higher fumonisin contamination than normal kernels (Figure 1i,j), indicating the serious infection of Fusarium verticillioides. We also investigated the agronomic traits of the lines without inoculation in the field using a random-block design with two repeats. Although some traits of the mutants were different from that of the WT, no obvious agronomic penalty was observed (Figure 1k).

In summary, we demonstrated that the targeted generation of null mutants in ZmFER1 could improve resistance against Fusarium verticillioides without agronomic penalty. These results provide a new approach for breeding FER resistant cultivars, which could have huge benefits, especially in intensive maize–wheat rotation systems.