The Arabidopsis Kinome: phylogeny and evolutionary insights into functional diversification

Abstract

Background: Protein kinases constitute a particularly large protein family in Arabidopsis with important functions in cellular signal transduction networks. At the same time Arabidopsis is a model plant with high frequencies of gene duplications. Here, we have conducted a systematic analysis of the Arabidopsis kinase complement, the kinome, with particular focus on gene duplication events. We matched Arabidopsis proteins to a Hidden-Markov Model of eukaryotic kinases and computed a phylogeny of 942 Arabidopsis protein kinase domains and mapped their origin by gene duplication.

Results: The phylogeny showed two major clades of receptor kinases and soluble kinases, each of which was divided into functional subclades. Based on this phylogeny, association of yet uncharacterized kinases to families was possible which extended functional annotation of unknowns. Classification of gene duplications within these protein kinases revealed that representatives of cytosolic subfamilies showed a tendency to maintain segmentally duplicated genes, while some subfamilies of the receptor kinases were enriched for tandem duplicates. Although functional diversification is observed throughout most subfamilies, some instances of functional conservation among genes transposed from the same ancestor were observed. In general, a significant enrichment of essential genes was found among genes encoding for protein kinases.

Conclusions: The inferred phylogeny allowed classification and annotation of yet uncharacterized kinases. The prediction and analysis of syntenic blocks and duplication events within gene families of interest can be used to link functional biology to insights from an evolutionary viewpoint. The approach undertaken here can be applied to any gene family in any organism with an annotated genome.

Figure 1

Maximum Likelihood phylogeny of 942 kinase domains in Arabidopsis thaliana. (A) Color coding according to functional families separating soluble kinases (blue) from receptor kinases (green). (B) Color coding according to subcellular location from SUBA3.

Figure 2

Visualization of results as provided by the MCScanX utility. (A) Dot plot indicating segmentally duplicated regions in the Arabidopsis thaliana genome. The axes indicate the genes on the five chromosomes (at1 to at5) and mitochondria (M). Colored dots denote different duplication events. (B) Bar plot showing the estimated proportion of segmental duplications on each chromosome. The corresponding origin for each chromosome is indicated in color. (C) Circle plot of kinase family-specific collinear regions (red curves) between chromosomes against the background of collinear regions in other genes (grey).

Figure 3

Duplication events in Arabidopsis kinases. (A) Phylogenetic tree of kinase families indicating gene duplication events. Small subfamilies were collapsed into artificial monophyletic clades as discussed earlier and indicated in the figure. Duplications are marked by asterisks at branching points corresponding to the most recent common ancestor of duplicates. The different types of duplication events are color-coded. red: segmental, blue: tandem, green: proximal, and orange: transposed. (B) Box plot of ratios of observed to expected tandem and segmental duplication frequencies for each family. The boxed region is centred on the median ratios (indicated by a black square) for tandem and segmental duplications. The mean of ratios is marked by a black triangle. The boxed region and additional lines refer to one and two standard deviations above and below the median ratios. Pie charts for each family indicate the localization of genes according to SUBA3 as the proportion of all genes in that family. (C) Results of the enrichment analysis. For each family and duplication type, the Pearson residuals from Chi-square tests are plotted together with the corresponding significance levels from Fisher’s exact tests. Duplication types are coded as W/D (whole-genome or segmental duplication), D (dispersed), T (tandem), P (proximal) and S (singleton). Red and green color gradients correspond to negative and positive Pearson residuals indicating depleted and enriched counts for each type of duplication. The significance is coded as follows: *corresponds to p < 0.05, ***corresponds to p < 0.001. P-values were adjusted for multiple testing by applying the Benjamini-Hochberg correction.

Figure 4

Relationship between phylogeny and gene coexpression patterns. (A) Coexpression analysis for subfamilies LRR_8B, LRR_3, and LRCK_9. Phylogenetic information was incorporated by reordering the rows according to the topology in the phylogenetic tree (left) or disregarded (right). Green color indicates high coexpression values, and red color indicates dissimilarity in expression context. Square patches of similar color denote agreements between the clustering of columns and rows. (B) Phylogenetic relationship and duplication annotation correspond for members of the LRR_8B, LRR_3, and LRCK_9 subfamily. Genes are connected by curves if they are identified as duplicated pairs by proximal duplication (green), segmental duplications (red) and transpositions (orange). In the case of transposition events, the ancestral gene is indicated by a vertical line.

Figure 5

Interaction network of protein kinases and their first interaction neighbours based in AI1. Node shape: Square: MAP-Kinases including MAP2K and MAP3K; Hexagon: AGC kinases; Parallelogram: CDK; Diamond: CDPK and SnRK; Triangle: Receptor kinases; Vee: Receptor-like cytoplasmic kinases (RLCK); Circle: other soluble kinases. Node colors indicate subcellular location of kinases according to SUBA3: blue = cytosol; pink = endoplasmic reticulum; purple = extracellular; yellow = mitochondrion; orange = nucleus; green = plastid; red = plasma membrane. Node border color indicates duplication type: cyan = WGD/segmental; beige = dispersed; dark green = proximal; black = singleton; dark red = tandem duplications.

Figure 6

Frequencies of correlation values for kinase gene pairs. Colored lines indicate different types of duplications in comparison to randomly sampled kinase-specific gene pairs.