Population-level variation of the preproricin gene contradicts expectation of neutral equilibrium for generalist plant defense toxins

Abstract

The preproricin gene encodes ricin, the highly toxic, type II ribosome-inactivating protein of castor bean ( Ricinus communis L.). As a generalist plant defense gene, preproricin is expected to exhibit population-level variation consistent with the neutral equilibrium model and to comprise few functionally different alleles. We first test the hypothesis that the preproricin gene family should comprise six to eight members by searching the publicly available draft genome sequence of R. communis and analyzing its ricin-like loci. We then test the neutral equilibrium expectation for the preproricin gene by characterizing its allelic variation among 25 geographically diverse castor bean plants. We confirm the presence of six ricin-like loci that share with the preproricin gene 62.9–96.3% nucleotide identity and intact A-chains. DNA sequence variation among the preproricin haplotypes significantly rejects tests of the neutral equilibrium model. Replacement mutations preserve the 12 amino acids known to affect catalytic and electrostatic interactions of the native protein toxin, which suggests functional divergence among alleles has been minimal. Nucleotide polymorphism is maintained by purifying selection ( ω < 0.3) yet includes an excess of rare silent mutations greater than predicted by the neutral equilibrium model. Development of robust detection methods for ricin contamination must account for the presence of these other ricin-like molecules and should leverage the specificity provided by the numerous single nucleotide polymorphisms in the preproricin gene.