Frequent changes in expression profile and accelerated sequence evolution of duplicated imprinted genes in arabidopsis

Abstract

Eukaryotic genomes have large numbers of duplicated genes that can evolve new functions or expression patterns by changes in coding and regulatory sequences, referred to as neofunctionalization. In flowering plants, some duplicated genes are imprinted in the endosperm, where only one allele is expressed depending on its parental origin. We found that 125 imprinted genes in Arabidopsis arose from gene duplication events during the evolution of the Brassicales. Analyses of 46 gene pairs duplicated by an ancient whole-genome duplication (alpha WGD) indicated that many imprinted genes show an accelerated rate of amino acid changes compared with their paralogs. Analyses of microarray expression data from 63 organ types and developmental stages indicated that many imprinted genes have expression patterns restricted to flowers and/or seeds in contrast to their broadly expressed paralogs. Assays of expression in orthologs from outgroup species revealed that some imprinted genes have acquired an organ-specific expression pattern restricted to flowers and/or seeds. The changes in expression pattern and the accelerated sequence evolution in the imprinted genes suggest that some of them may have undergone neofunctionalization. The imprinted genes MPC, HOMEODOMAIN GLABROUS6 (HDG6), and HDG3 are particularly interesting cases that have different functions from their paralogs. This study indicates that a large number of imprinted genes in Arabidopsis are evolutionarily recent duplicates and that many of them show changes in expression profiles and accelerated sequence evolution. Acquisition of imprinting is a mode of duplicate gene divergence in plants that is more common than previously thought.

Fig. 1.—

Asymmetric sequence rate evolution in duplicated gene pairs with imprinted genes. Pie charts indicating the number and percentage of asymmetrically evolving gene pairs (dark blue), the number and percentage of asymmetrically evolving gene pairs where the imprinted gene show faster sequence evolution than the paralog (light blue), and the number and percentage of symmetrically evolving gene pairs (gray). (a) Gene pairs with imprinted genes identified in Wolff et al. (2011). (b) Gene pairs with imprinted genes identified in Hsieh et al. (2011).

Fig. 2.—

Many imprinted genes have a more restricted expression pattern compared with their paralogs. (a) Expression width, based on microarray data. Dark blue: Number and percentage of imprinted genes with a smaller expression width. Gray: Number and percentage of imprinted genes with a larger expression width. (b) Organ specificity of expression, based on microarray data. Dark blue: Number and percentage of imprinted genes with a higher organ specificity of expression. Gray: Imprinted genes with a lower organ specificity of expression. In panels (a) and (b), gene pairs with imprinted genes identified in Wolff et al. (2011) are on the left, and gene pairs with imprinted genes identified in Hsieh et al. (2011) are on the right.

Fig. 3.—

RT-PCR results of imprinted genes, their paralogs, and orthologs from outgroup species. RT-PCR expression assays were performed using five organ types: Root, stem, leaf, flower, and silique (Arabidopsis)/seed (Carica and Vitis), listed above the corresponding columns. Plus signs indicate the presence of reverse transcriptase in the reaction, and minus signs indicate the absence of reverse transcriptase as negative controls. Gene pairs in Arabidopsis thaliana (At), and their orthologs from outgroup species, Carica papaya (Cp), and Vitis vinifera (Vv) are listed beside the corresponding panels. ACTIN in each species was used as a positive control for RT-PCR, and the results are shown in the bottom right graph.

Fig. 4.—

Analyses of MPC, FWA, and HDG3. (a) RT-PCR expression assays were performed using five organ types: Root, stem, leaf, flower, and silique (Arabidopsis)/seed (Carica and Vitis), listed above the corresponding columns. Plus signs indicate the presence of reverse transcriptase in the reaction, and minus signs indicate the absence of reverse transcriptase as negative controls gDNA, genomic DNA. Gene pairs in Arabidopsis thaliana (At), and their orthologs of MPC/PAB8, FWA/HDG7, and HDG3/HDG2 from outgroup species, Carica papaya (Cp) and Vitis vinifera (Vv) are listed beside the corresponding panels. (b) Sequence rate analyses. Phylogenetic trees of each gene have sequences from A. thaliana (At), C. papaya (Cp), Populus trichocarpa (Pt), Ricinus communis (Rc), Manihot esculenta (Me), and V. vinifera (Vv). Trees are unrooted. Branch lengths were generated by Codeml in PAML, and the scale bar indicates nucleotide substitutions per codon. Branch-wise Ka/Ks ratios are indicated above the branches. Imprinted genes and their larger Ka/Ks ratios are shown in bold. Circles indicate the gene duplication events that gave rise to the imprinted genes and their paralogs. (c) Site-specific selection analysis on FWA HD-ZIP domain to detect positively selected amino acid sites. An amino acid alignment of the HD-ZIP domain in FWA and HDG7 in A. thaliana (At), and orthologous genes from C. papaya (Cp), P. trichocarpa (Pt), R. communis (Rc), M. esculenta (Me), and V. vinifera (Vv) showing the amino acid substitution in FWA. Residues with more than 50% shared identity are black shaded, and similar amino acids are gray shaded. Motifs are labeled under the bar below the alignment. Stars indicate potential DNA sequence-specific contact residues that are conserved in all AtHDG family members but FWA. Positively selected sites on FWA were marked by triangles, where black triangles indicate posterior probabilities (PP) greater than 0.99 and gray triangles indicate PPs greater than 0.95.

Fig. 5.—

Sequence rate analyses of HDG9/HDG10 and SUVH7/SUVH8. Phylogenetic trees of each gene have sequences from Arabidopsis thaliana (At), Carica papaya (Cp), Populus trichocarpa (Pt), Ricinus communis (Rc), Manihot esculenta (Me), and Vitis vinifera (Vv). Trees are unrooted. Branch lengths were generated by Codeml in PAML, and the scale bar indicates nucleotide substitutions per codon. Branch-wise Ka/Ks ratios are indicated above the branches. Imprinted genes are shown in bold. The suffix of Arabidopsis genes indicates the imprinting status: meg versus peg. High Ka/Ks ratios of imprinted genes or their inferred precursor are shown also in bold. Circles indicate the gene duplication events that gave rise to the imprinted genes and their paralogs.