DNA sequence evidence for the segmental allotetraploid origin of maize

Abstract

It has long been suspected that maize is the product of an historical tetraploid event. Several observations support this possibility, including the fact that the maize genome contains duplicated chromosomal segments with colinear gene arrangements. Some of the genes from these duplicated segments have been sequenced. In this study, we examine the pattern of sequence divergence among 14 pairs of duplicated genes. We compare the pattern of divergence to patterns predicted by four models of the evolution of the maize genome-autotetraploidy, genomic allotetraploidy, segmental allotetraploidy, and multiple segmental duplications. Our analyses indicate that coalescent times for duplicated sequences fall into two distinct groups, corresponding to roughly 20.5 and 11.4 million years. This observation strongly discounts the possibility that the maize genome is the product of a genomic allotetraploid event, and it is also difficult to reconcile with either autotetraploidy or multiple independent segmental duplications. However, the presence of two (and only two) coalescent times is predicted by the segmental allotetraploid model. If the maize genome is the product of a segmental allotetraploid event, as these data suggest, then its two diploid progenitors diverged roughly 20.5 million years ago (Mya), and the allotetraploid event probably occurred approximately 11.4 Mya. Comparison of maize and sorghum sequences suggests that one of the two ancestral diploids shares a more recent common ancestor with sorghum than it does with the other ancestral diploid.

Figure 1

Diagrammatic representation of the expected distribution of coalescent times under the four models of chromosomal duplication. Columns: A, The autotetraploid model under the assumption that the switch to disomy is coordinated among chromosomes; B, the genomic allotetraploid model; C, the segmental allotetraploid model; D, the model of segmental duplication.

Figure 2

A graph of synonymous distances between duplicated sequences and between interspecific sequences. Shaded boxes represent the standard deviation of distance estimates. The time scale is based on an assumed rate of synonymous nucleotide substitution (see text).

Figure 3

Graphical representation of k, the average synonymous distance between duplicated maize sequences and a rice outgroup sequence. Shaded boxes represent standard deviations.