Analysis of randomly mutated AlSRKb genes reveals that most loss-of-function mutations cause defects in plasma membrane localization

Introduction

Errors in DNA replication that result in a change in the DNA sequence produce nucleotide variation in living organisms. Although nonsynonymous variants may cause genes to lose their function or develop new ones, some have no effect on the functions of genes. Advances in sequencing technology over the past two decades have revealed the full picture of DNA polymorphisms across a genome (Hu et al., 2021), enabling the construction of genome-wide association study (GWAS) platforms in many organisms (Alseekh et al., 2021). Identifying the causative single nucleotide variants (SNVs) within the set of genome-wide DNA polymorphisms is a crucial step in genetic analysis, but also the most difficult (Witte, 2010). Therefore, knowing the characteristics of nonsynonymous variants that affect gene function will be useful for finding causative SNVs in GWAS and quantitative trait locus (QTL) mapping analyses, and will also contribute to resolving the pressing challenges facing agriculture and human healthcare.

To analyze the characteristics of nonsynonymous variants (mutations) with large influences on gene function, we focused on the S-locus receptor kinase (SRK) gene, a gene whose protein product functions as the female determinant of self-incompatibility (SI) in the Brassicaceae (Stein et al., 1991; Takasaki et al., 2000). SI enables plants to avoid self-fertilization; by facilitating cross-fertilization, it avoids inbreeding depression and maintains genetic variation. About 40% of Angiosperm families show SI (McCubbin & Kao, 2000; Barrett, 2002; Igić & Kohn, 2006). SI is usually determined by a single locus, the S locus, which contains several genes and forms a distinctive haplotype, called the S haplotype (Silva & Goring, 2001). Population genetic theory predicts that many S haplotypes should be maintained, given that individuals possessing rare S haplotypes have more mating opportunities than those carrying common haplotypes (Wright, 1939; Schierup, 1998). Consistent with this, more than 50 S haplotypes, each of which shows different self-recognition activity, are known from cultivated Brassica species (Oikawa et al., 2011; Yamamoto et al., 2023).

SI in the Brassicaceae is genetically controlled by two tightly linked, highly polymorphic genes within the S locus. S-locus receptor kinase (SRK) encodes a plasma membrane-localized receptor kinase expressed in stigmatic papillae cells (Stein et al., 1991; Takasaki et al., 2000) and S-locus cysteine-rich protein/S-locus protein 11 (SCR/SP11, hereafter referred to as SCR) encodes a cysteine-rich peptide ligand of SRK displayed at the pollen surface (Schopfer et al., 1999; Takayama et al., 2000). SRK can only interact with SCR of the same S haplotype. This allele-specific SRK–SCR interaction activates the SI response, inhibiting pollen germination and growth of the pollen tube on the stigma surface. Functional diversification of S haplotypes is achieved by multiple amino acid polymorphisms in SRK and SCR. The mature region of SCR shows extraordinary diversity, including frequent insertions and deletions (Sato et al., 2002). Although similarity of full-length SRK is relatively high compared with SCR (Sato et al., 2002), three hypervariable regions (hv I, II, and III) are found in the receptor domains of SRK in the S haplotypes in Brassica and Arabidopsis species (Kusaba et al., 1997; Nishio & Kusaba, 2000; Castric & Vekemans, 2007). Structural analysis of the SRK-SCR complex suggests that amino acid residues in the hv regions of SRK are involved in the allele-specific SRK–SCR interaction (Ma et al., 2016; Murase et al., 2020). Since SRK shows high diversity in allelic sequences as well as a specific recognition activity, we considered it a suitable protein model for unraveling the characteristics of nonsynonymous variants with influences on gene function.

The model plant Arabidopsis thaliana is a useful experimental tool for analyzing the effects of mutations on gene function because it can be transformed simply and efficiently. Although A. thaliana shows self-compatibility (SC), as its S locus contains either a nonfunctional SCR gene or nonfunctional SRK and SCR genes (Kusaba et al., 2001; Shimizu et al., 2008; Tang et al., 2007; Tsuchimatsu et al., 2010, 2017), SI can be induced by introduction of the SRK and SCR genes of Sb, also known as the S20 haplotype, from Arabidopsis lyrata, a closely related SI species (Nasrallah et al., 2002, 2004). To characterize the mutations in A. lyrata SRKb (AlSRKb) that caused the defect in SI activity, we determined the SI phenotypes and genotypes of 300 A. thaliana transformants expressing AlSRKb containing randomly introduced mutations, and also measured the levels of AlSRKb expression and localization in the plasma membrane in 139 transformants. Almost all mutations in AlSRKb that caused SI defects also affected localization to the plasma membrane. Mutations in amino acid residues that were highly conserved across S haplotypes and located within the interior of the AlSRKb molecule resulted in significant changes to amino acid properties, which were associated with SI defects. Our analysis revealed that abnormalities in AlSRKb biosynthesis were more commonly found among mutations causing defects in SI than among mutations causing loss of self-recognition activity. In addition, we used the RandomForest and Extreme Gradient Boosting methods with 164 mutations to make preliminary predictions about the SI phenotypes of A. thaliana transformants.