A horizontally transferred nuclear gene is associated with microhabitat variation in a natural plant population

Abstract

Horizontal gene transfer involves the non-sexual interspecific transmission of genetic material. Even if they are initially functional, horizontally transferred genes are expected to deteriorate into non-expressed pseudogenes, unless they become adaptively relevant in the recipient organism. However, little is known about the distributions of natural transgenes within wild species or the adaptive significance of natural transgenes within wild populations. Here, we examine the distribution of a natural plant-to-plant nuclear transgene in relation to environmental variation within a wild population. Festuca ovina is polymorphic for an extra (second) expressed copy of the nuclear gene (PgiC) encoding cytosolic phosphoglucose isomerase, with the extra PgiC locus having been acquired horizontally from the distantly related grass genus Poa. We investigated variation at PgiC in samples of F. ovina from a fine-scale, repeating patchwork of grassland microhabitats, replicated within spatially separated sites. Even after accounting for spatial effects, the distributions of F. ovina individuals carrying the additional PgiC locus, and one of the enzyme products encoded by the locus, are significantly associated with fine-scale habitat variation. Our results suggest that the PgiC transgene contributes, together with the unlinked 'native' PgiC locus, to local adaptation to a fine-scale mosaic of edaphic and biotic grassland microhabitats.

Keywords: Festuca ovina; habitat mosaic; horizontal gene transfer; local adaptation; natural plant transgene; phosphoglucose isomerase.

Figure 1.

Multivariate analyses show relationships between environmental variables, grassland community composition, and variation at PgiC in Festuca ovina. The four microhabitats are indicated by: red circles, dry/high pH, red squares, dry/low pH, blue circles, moist/high pH, blue squares, moist/low pH. (a) Axes 1 and 2 of a correspondence analysis showing differentiation in plant community composition among the 1 × 1 m plots from which F. ovina was sampled (n = 40). Arrows indicate relationships between the additional environmental variables characterizing the plots and the first two axes of community composition (accounting for 17% and 11% of the variation, respectively). Cover values are within-plot percentage covers of the respective plant groups; ‘bare ground’ is the percentage of the plot without vegetation cover. (b) Alvar grasslands on Öland illustrating three of the four microhabitats included in the study (photo by B. Widén). (c) Principal components analysis showing among-plot (n = 40) differentiation in environmental variables: components 1 and 2 account for 58% and 17% of the variation, respectively. Arrows indicate increasing within-plot proportions of significant (tables 1 and 2) PgiC variables: C2(f), the transgenic PgiC2(f) locus; Em 1, Em 2, Em 4, PgiC electromorphs 1, 2, and 4.