Genome-wide analysis of histone modifiers in tomato: gaining an insight into their developmental roles

Abstract

Background: Histone post-translational modifications (HPTMs) including acetylation and methylation have been recognized as playing a crucial role in epigenetic regulation of plant growth and development. Although Solanum lycopersicum is a dicot model plant as well as an important crop, systematic analysis and expression profiling of histone modifier genes (HMs) in tomato are sketchy.

Results: Based on recently released tomato whole-genome sequences, we identified in silico 32 histone acetyltransferases (HATs), 15 histone deacetylases (HDACs), 52 histone methytransferases (HMTs) and 26 histone demethylases (HDMs), and compared them with those detected in Arabidopsis (Arabidopsis thaliana), maize (Zea mays) and rice (Oryza sativa) orthologs. Comprehensive analysis of the protein domain architecture and phylogeny revealed the presence of non-canonical motifs and new domain combinations, thereby suggesting for HATs the existence of a new family in plants. Due to species-specific diversification during evolutionary history tomato has fewer HMs than Arabidopsis. The transcription profiles of HMs within tomato organs revealed a broad functional role for some HMs and a more specific activity for others, suggesting key HM regulators in tomato development. Finally, we explored S. pennellii introgression lines (ILs) and integrated the map position of HMs, their expression profiles and the phenotype of ILs. We thereby proved that the strategy was useful to identify HM candidates involved in carotenoid biosynthesis in tomato fruits.

Conclusions: In this study, we reveal the structure, phylogeny and spatial expression of members belonging to the classical families of HMs in tomato. We provide a framework for gene discovery and functional investigation of HMs in other Solanaceae species.

Figure 1

Phylogenetic tree and domain composition of HAG proteins. Maximum likelihood phylogenetic tree of HAG predicted proteins from Arabidopsis thaliana (At), Oryza sativa (Os), Solanum lycopersicon (Sl) and Zea mays (Zm). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. AT1 (PF00583) and C-terminal BrD (PF00439) are conserved domains of GCN5-like members; N-terminal ELP (IPR006638) and C-terminal AT1 are domains of ELP3-like; N-terminal Hat1_N (PF10394) and C-terminal AT1 are motifs of HAT1-like members while the only AT1 domain is of HPA2-like proteins. Overlapping domains are hyphenated and represented with a continuous dotted line.

Figure 2

Expression profiles of tomatoHATs. Heat map of RNA-seq expression data from root, leaf, bud, flower, 1cm_fruit, 2cm_fruit, 3cm_fruit, mature green fruit (MG), berry at breaker stage (B) and berry ten days after breaking (B10). HAGs with low, middle and high expression values are reported in A, B and C, respectively. The expression values are measured as reads per kilobase of exon model per million mapped reads (RPKM).

Figure 3

Phylogenetic tree and domain composition of HAC and HAF proteins. Bayesian phylogenetic tree of HAC (A) and HAF (B) predicted proteins from Arabidopsis thaliana (At), Oryza sativa (Os), Solanum lycopersicon (Sl) and Zea mays (Zm). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. (A) KAT11 (PF08214), PHD-finger (PF00628) and zf-TAZ (PF02135) are conserved domains of HAC proteins. (B) N-terminal kinase (PF09247) (TBPb), ubiquitin (PF00240), zinc-finger C2HC (PF01530), and C-terminal bromo BrD (PF00439) are conserved domains of HAFs. A domain of unknown function DUF3591 is also shown.

Figure 4

Phylogenetic tree and domain composition of HDA proteins. Maximum likelihood phylogenetic tree of HDA predicted proteins from Arabidopsis thaliana (At), Oryza sativa (Os), Solanum lycopersicon (Sl) and Zea mays (Zm). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Class I and Class II include the homologs of RPD3 and of HDA1 from yeast, respectively; Class III includes the homologs of human HDAC11. The hist_deacetyl domain (PF00850) is the conserved domain of HDA proteins.

Figure 5

Phylogenetic trees and domain composition of Class I and Class II SDG proteins. Maximum likelihood phylogenetic trees of Class I (A) and Class II (B) SDG predicted proteins from Arabidopsis thaliana (At) and Solanum lycopersicon (Sl). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Two EZD, SANT (SM00717), CXC (PF03638), and SET (PF00856) are conserved domains of Class I. N-terminal AWS (SM00570), SET and Post-SET (SM00508) are conserved domains of Class II. The PHD domain (PF00628) and CW domain (SM00605) are also found in Class II proteins.

Figure 6

Phylogenetic trees and domain composition of Class III and Class IV SDG proteins. Maximum likelihood phylogenetic trees of Class III (A) and Class IV (B) SDG predicted proteins from Arabidopsis thaliana (At) and Solanum lycopersicon (Sl). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. N-terminal PWWP (PF00855), FYRN (PF05964), FYRC (PF05965), two PHD, SET, and Post-SET are conserved domains of Class III. Some Class III proteins lack the FYRN (PF05964) and FYRC (PF05965) domains. Other Class III have N-terminal GYF (PF02213), SET and Post-SET domains. N-terminal PHD and C-terminal SET are conserved domains of Class IV.

Figure 7

Phylogenetic trees and domain composition of Class V and Class VI/VII SDG proteins. Maximum likelihood phylogenetic trees of class V and class VI/VII SDG predicted proteins from Arabidopsis thaliana (At) and Solanum lycopersicon (Sl). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. N-terminal SRA-YDG (PF02182), Pre-SET, SET and Post-SET are conserved domains of the group I of Class V; N-terminal WIYLD (PF10440), or C2H2 (PF00096) or absence of domain, Pre-SET, SET and Post-SET are conserved domains of the group II of Class V (so-called SUVH-related). Proteins with an interrupted SET domain or SET-related proteins are indicated as Class VI and VII.

Figure 8

Expression profiles of tomatoSDGs. Heat map of RNA-seq expression data from root, leaf, bud, flower, 1cm_fruit, 2cm_fruit, 3cm_fruit, mature green fruit (MG), berry at breaker stage (B) and berry ten days after breaking (B10). The expression values are measured as reads per kilobase of the exon model per million mapped reads (RPKM).

Figure 9

Phylogenetic tree and domain composition of JMJ-only and KDM4 proteins. Bayesian phylogenetic tree of JMJ-only (A) and KDM4 (B) predicted proteins from Arabidopsis thaliana (At), Oryza sativa (Os), Solanum lycopersicon (Sl) and Zea mays (Zm). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. JmjC domain (PF02373) is the conserved domain of the JMJs-only class; N-terminal JmjN (PF02375) and JmjC, and C-terminal C5HC2 (PF02928) (subgroup I) or C2H2 (PF00096) (subgroup II) domains are conserved domains of KDM4 proteins.

Figure 10

Phylogenetic tree and domain composition of JMJD6 and KDM5 proteins. Bayesian phylogenetic tree of JMJD6 (A) and KDM5 (B) predicted proteins from Arabidopsis thaliana (At), Oryza sativa (Os), Solanum lycopersicon (Sl) and Zea mays (Zm). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. N-terminal F-box (PF00646), and C-terminal JmjC are conserved domains of JMJD6s demethylases; JmjN, BRIGHT/ARID (PF01388), two PHD (PF00628), JmjC, and C5HC2 are conserved domains of the subgroup I of KDM5; JmjN, JmJC, C5HC2, FYRN (PF05964) and FYRC (PF05965) are conserved domains of the subgroup II of class KDM5.

Figure 11

Phylogenetic tree and domain composition of KDM3 proteins. Bayesian phylogenetic tree of KDM3 predicted proteins from Arabidopsis thaliana (At), Oryza sativa (Os), Solanum lycopersicon (Sl) and Zea mays (Zm). Bootstrap values higher than 50% are shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. RING-finger (IPR001841) and JmjC are conserved domains of KDM3 demethylases.