Divergence of the Dof Gene Families in Poplar, Arabidopsis, and Rice Suggests Multiple Modes of Gene Evolution after Duplication^{<a href="#fn1" rid="fn1" class=" fn">1</a>,}^{<a href="#fn2" rid="fn2" class=" fn">[W]</a>}

Abstract

Darwin's positive selection theory cannot adequately explain the rapid rise and early diversification of more than 250,000 flowering plant species (Darwin and Seward, 1903; Davies et al., 2004; De Bodt et al., 2005). The theory that gene duplication events are the primary source of genetic novelty leading to speciation, first postulated by Ohno (1970), has gained wide acceptance (Lynch and Conery, 2000; Gu et al., 2003; Moore and Purugganan, 2003, 2005; Blanc and Wolfe, 2004; Li et al., 2005). However, the specific evolutionary route(s) of duplicated genes has remained largely unresolved (Force et al., 1999; Lynch and Conery, 2000; He and Zhang, 2005b; Moore and Purugganan, 2005). According to Ohno (1970), after a duplication event, one daughter gene retains the preduplication function, while the other one, in the majority of cases, accumulates deleterious mutations and is eliminated, or, in the minority of cases, survives by gaining a new function. This hypothesis, referred to here as Ohno's Duplication-Retention-Non/Neofunctionalization (DRNNF), has been the subject of intensive debate (Taylor and Raes, 2004). Hughes (1994) proposed the subfunctionalization (SF) model for proteins, under which duplicated genes share the same functions for a period of time and then evolve into functionally distinct proteins with each daughter gene specialized in a subset of functions of the ancestral gene. Force et al. (1999) summarized three observations on genome-wide duplication events that are contradictory to DRNNF, including (1) a higher proportion of the duplicated genes retained than expected by chance alone, (2) nucleotide substitution patterns reflective of purifying selection on both copies of the duplicated genes, and (3) a relative paucity of null allele for loci that have avoided nonfunctionalization. They then proposed a model similar to SF, called duplication-degeneration-complementation (DDC), to also include regulatory element functions (Force et al., 1999). Under this model, the majority of duplicated genes accumulates degenerative mutations for a period of time and then undergoes functional specialization by complementary partition of ancestral functions. Therefore, preservation of duplicated genes is through complementary subfunctionalization of the progenitor gene rather than the evolution of new functions. Recently, He and Zhang (2005b) further extended the DDC model, termed subneofunctionalization (SNF), under which a large proportion of duplicate genes undergo rapid subfunctionalization, and the subfunctionalized genes may later also evolve new functions not in the ancestral gene. A fundamental assumption of the SF/DDC model is that each ancestral gene had at least two functions that could subsequently be partitioned between two daughter genes. However, this assumption could be valid only if the genes of ancestral species had almost all the functions that extant relatives contain, and, to our knowledge, this is unlikely the case. Moreover, if the ancestral genes possessed multiple functions, where were their origins? To answer this question, a subfunction co-option concept has also been put forth (Raff, 1996; Carroll, 2001; Cameron et al., 2005), which suggests that a gene evolves by co-opting a new function not found in the ancestral gene. Recently, it was suggested that epigenetic changes might play important roles in the evolution of duplicated genes (Rodin and Riggs, 2003; Rapp and Wendel, 2005; Rodin et al., 2005). The term epigenetics can be applied to mean alteration of phenotype, morphological or molecular, without change in either gene coding sequence or promoter region (Rapp and Wendel, 2005). Epigenetic changes in gene expression include promoter methylation, DNA packaging, repositioning, microRNA, and small interfering RNA (Bender, 2002; Rapp and Wendel, 2005; Rodin et al., 2005). Epigenetic changes in proteins include posttranslational modification of proteins such as methylation (Chen et al., 2006).

Based on the above information, we hypothesized that both genetic and epigenetic changes are involved in the evolution of duplicated genes (Fig. 1). Genetic changes in proteins include retention (R), indicating that a copy retains the original motif organization and function; degeneration (D), indicating that a copy degenerates or loses one or more motifs and functions; and neofunctionalization (N), indicating that a copy acquires one or more motifs and functions. There are six possible combinations of these three types of genetic changes in coding regions for duplicated genes: RR, RD, RN, DD, NN, and ND. Epigenetic changes in proteins (i.e. protein methylation) or in promoter regions (i.e. DNA methylation) can cause functional diversification in duplicated genes that share the same motif structure and change them from RR into RD or DD type. RD and RN correspond to Ohno's hypothesis (Ohno, 1970), DD corresponds to the SF or DDC model (Hughes, 1994; Force et al., 1999), and NN corresponds to the SNF model (He and Zhang, 2005b). We consider that gene function consists of both protein function conferred by the coding region, as suggested by Hughes (1994), and the pattern of gene expression mostly controlled by regulatory elements, as suggested by Force et al. (1999).

Figure 1.

Possible evolutionary modes of paralogs in the protein coding region after gene duplication.

In this study, we compared genes of a plant-specific gene family, the DNA binding with one finger domain (Dof) transcription factor, in three angiosperm plants, poplar (Populus trichocarpa) Torr. &amp; Gray ex Brayshaw, Arabidopsis (Arabidopsis thaliana), and rice (Oryza sativa), all of which have been completely sequenced and have undergone at least one round of genome-wide duplication (Bowers et al., 2003; Raes et al., 2003; Sterck et al., 2005; Tuskan et al., 2006). We selected the Dof gene family because of its diverse biological functions (Yanagisawa, 2004). Dof proteins typically consist of multiple domains, including a highly conserved N-terminal DNA-binding domain and a C-terminal domain for transcriptional regulation (Yanagisawa, 2002). The high degree of conservation of the Dof domain and the diversity of the remaining portion of the protein provide rich materials for studying the fates of gene diversification. Our purpose was to search for footprints of genes evolution; to determine the validity of DRNNF, SF, DDC, or SNF in describing the fates of duplicated genes; and to examine the forces that have driven the divergence of duplicated genes. We first constructed a maximum-likelihood phylogenetic tree using full-length protein sequences of the Dof genes. From the phylogenetic tree, we identified 27 pairs of paralogous Dof genes. Then we predicted the ancestral protein sequences for the paralogs. Analysis of protein motif structure of the Dof paralogs and their ancestors revealed six different genetic changes in coding region after gene duplication. We also investigated potential epigenetic changes in the proteins of duplicated genes that have the same protein motif structure. Prediction of methylation in protein sequences indicates that epigenetics is also involved in gene evolution. We also examined expression of the paralogs by reverse transcription (RT)-PCR, massively parallel signature sequencing (MPSS), and microarray data analysis. To examine the driving force for the gene evolution, we performed nonsynonymous and synonymous substitution rate (k_a and k_s) analysis of the duplicated genes. We also searched the non-Dof genes in Arabidopsis, poplar, and rice for the motifs in Dof proteins and revealed that motif co-option may be one of the forces driving gene diversification. We provided evidence that the previously proposed DRNNF, SF/DDC, and SNF hypotheses are complementary with, not alternative to, each other. Our study also suggested that epigenetic changes might be involved in evolution of duplicated genes.

Acknowledgments

Notes

^{This work was supported by the National Science Foundation (grant no. 0421743 to G.A.T. and Z.-M.C.), by the U.S. Department of Energy/Oak Ridge National Laboratory (subcontract to Z.-M.C.), and by the Tennessee Agricultural Experiment Station.}

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantphysiol.org) is: (Max) Zong-Ming Cheng (ude.ktu@gnehcz).

^{The online version of this article contains Web-only data.}

www.plantphysiol.org/cgi/doi/10.1104/pp.106.083642