Seeing red: the origin of grain pigmentation in US weedy rice

Abstract

Weedy forms of crop species infest agricultural fields worldwide and are a leading cause of crop losses, yet little is known about how these weeds evolve. Red rice (Oryza sativa), a major weed of cultivated rice fields in the US, is recognized by the dark-pigmented grain that gives it its common name. Studies using neutral molecular markers have indicated a close relationship between US red rice and domesticated rice, suggesting that the weed may have originated through reversion of domesticated rice to a feral form. We have tested this reversion hypothesis by examining molecular variation at Rc, the regulatory gene responsible for grain pigmentation differences between domesticated and wild rice. Loss-of-function mutations at Rc account for the absence of proanthocyanidin pigments in cultivated rice grains, and the major rc domestication allele has been shown to be capable of spontaneous reversion to a functional form through additional mutations at the Rc locus. Using a diverse sample of 156 weedy, domesticated and wild Oryzas, we analysed DNA sequence variation at Rc and its surrounding 4 Mb genomic region. We find that reversion of domestication alleles does not account for the pigmented grains of weed accessions; moreover, we find that haplotypes characterizing the weed are either absent or very rare in cultivated rice. Sequences from genomic regions flanking Rc are consistent with a genomic footprint of the rc selective sweep in cultivated rice, and they are compatible with a close relationship of red rice to Asian Oryzas that have never been cultivated in the US.

Figure 1. Relationships among Oryza species and varieties based on genome-wide variation and the Rc locus

Schematic representations of the relationships between different Oryza species and variety groups based on A) genome-wide surveys of genetic variation (Garris et al. 2005; Caicedo et al. 2007) and B) genetic variation at the Rc locus. In B, potential relationships of the US weed haplotypes to cultivated varieties are indicated by red stars. Dotted lines indicate potential origins of functional Rc alleles through reversion from domestication alleles.

Figure 2. Maximum likelihood (ML) tree of Rc haplotypes

Haplotypes carrying domestication alleles (rc, Rc-s) are grouped as triangles, with the size of the triangle roughly proportional to the number accessions carrying the allele. Colors indicate the identities of accessions carrying a haplotype, as indicated in the key. The BR and putative crop-weed hybrid (MIX) red rice accessions are not represented by separate colors, but are labeled where they occur on the tree. Numbers following labels indicate the number of individuals sharing a haplotype. Haplotypes represented by only one individual are labeled but not numbered, except for haplotypes unique to O. rufipogon, which are neither labeled nor numbered. Numbers at nodes represent percent bootstrap support; bootstrap values over 50% are shown. For simplicity, only O. meridionalis is used as the outgroup.

Figure 3. Schematic of sequenced regions

Locations of the Rc locus with up- and downstream sequenced regions and flanking fragments on chromosome 7. At the Rc locus (top), exons are indicated as rectangles and the length of the sequenced region is shown in light grey. Distance of flanking fragments from the sequenced region of the Rc locus and names of the fragments are indicated below. Distances are approximately to scale, and are based on the Nipponbare reference sequence.

Figure 4. Selection and weed-crop differentiation across the Rc genomic region

Solid black lines indicate genetic differentiation (γ_ST; left-hand Y axis) between red rice phenotypic variants (BHA and SH) and their closest crop relatives (aus and indica, respectively). The dashed gray line indicates an index of the rc selective sweep in cultivated rice (right-hand Y axis); the index is calculated as silent-site π for crop accessions carrying the rc domestication allele as a proportion of π for all crop accessions at that locus. Numbers below flanking fragment names indicate their approximate location in relation to the Rc locus (see Figure 3). Zero γ_ST values for rc5_003 indicate that all red rice, aus, and indica accessions share an identical sequence at this locus.