Nonadditive gene expression contributing to heterosis in partially heterozygous maize hybrids is predominantly regulated from heterozygous regions

Abstract

Introduction

The term heterosis describes the observation that hybrid progeny of genetically distinct parents display superior agricultural performance (Shull, 1914). The introduction of hybrids in maize breeding in the 1930s is considered one of the landmark innovations of modern agriculture and has contributed to an enormous increase in yield (Duvick, 2005; Hochholdinger & Baldauf, 2018; Hochholdinger & Yu, 2024). It has been observed that the phylogenetic distance between the parental inbred lines is positively associated with heterosis (East, 1936). The observation that specific parent combinations result in especially high levels of heterosis has resulted in the definition of typical female and male heterotic groups (Reif et al., 2005). Other crops, such as rice, also benefit from the classification of genotypes into heterotic groups and their combination as heterotic patterns (Melchinger & Gumber, 1998; Beukert et al., 2017).

Heterosis is observed in all parts of the plant throughout development, but is typically investigated for aboveground traits related to yield (Paril et al., 2024). In maize roots, which play an important role in the overall performance of plants, heterosis becomes apparent 5–7 d after germination (Hoecker et al., 2006).

Classical genetic concepts to explain heterosis include the dominance and overdominance models. The dominance model postulates that heterosis is caused by complementation of slightly deleterious alleles at many loci in the hybrid by dominant or at least stronger alleles (Jones, 1917). The overdominance model postulates that two different alleles at the same locus cause heterosis by their interaction and that the heterozygous state itself is advantageous to the homozygous situation of the parents (East, 1936). Despite examples of single genes displaying overdominance (Krieger et al., 2010; Lin et al., 2020), none of these models alone can fully explain heterosis (Duvick, 2001; Chen & Birchler, 2013; Hochholdinger & Yu, 2024).

Genes with differential expression between two maize lines can show a variety of expression levels in the resulting hybrid. They can display additive expression, reflecting the average expression of their parents, or deviate from this pattern and display nonadditive expression (Hochholdinger & Hoecker, 2007). Depending on the surveyed tissues, developmental stages, and genotypes, maize displays a highly variable degree of nonadditive gene expression (Uzarowska et al., 2007; Hoecker et al., 2008; Paschold et al., 2012; Zhang et al., 2016). Reciprocal maize hybrids of B73 and Mo17 generally share the same nonadditive pattern (Stupar & Springer, 2006; Baldauf et al., 2016). In early primary roots and developing ear shoots of the same hybrids, a trend toward adoption of high parent expression, rather than low parent expression, was observed (Paschold et al., 2012; Qin et al., 2013; Baldauf et al., 2016), but not further investigated in detail. Both additive and nonadditive expression have been considered to contribute to heterosis (Guo et al., 2006; Stupar & Springer, 2006; Hoecker et al., 2008; Stupar et al., 2008; Baldauf et al., 2016). The observation that nonadditive genes are conserved under stress conditions and mostly belong to evolutionarily less conserved, nonsyntenic genes suggests that they are involved in adaptation to different environments or stress conditions (Baldauf et al., 2016; Marcon et al., 2017).

Gene expression differences are the result of alterations in gene regulation. Regulatory elements can be classified as cis if they are positioned close to the regulated gene, and trans if the element is located at a different position, often on a different chromosome (Jansen & Nap, 2001). A possible connection between transcriptional variation in the regulation of cis- and trans-acting factors and hybrid performance was discussed (Botet & Keurentjes, 2020) and an association of trans-regulated gene expression in hybrids with paternal alleles was shown in maize (Swanson-Wagner et al., 2009).

Recombinant inbred line (RIL) populations as well as backcross populations have been extensively used in genetics for quantitative trait locus (QTL) mapping, candidate gene identification, and heterosis studies (Rahman et al., 2011; Pan et al., 2017; Huo et al., 2023; Yang et al., 2024). They can further be used to identify expression quantitative trait loci (eQTL). These are genomic regions associated with variation in gene expression across the mapping population and provide direct insights into the regulation of gene expression (Jansen & Nap, 2001).

In this study, we analyzed the transcriptomes of the maize intermated B73 and Mo17 (IBM) RIL Syn. 4 population (Lee et al., 2002) and their partially homozygous and heterozygous backcross hybrid populations with the original parents B73 and Mo17 (Supporting Information Fig. S1). We demonstrated that nonadditive gene expression patterns influence the manifestation of heterosis in seedling root development. We further showed that regulatory elements of nonadditive genes are predominantly located in heterozygous regions, suggesting that heterozygosity at the regulatory level promotes a higher expression in the hybrid compared to the parental average. Depending on their parental genetic origin, these regulatory elements act predominantly in either cis or trans, possibly influencing the formation of heterotic patterns.