Analysis of functional redundancies within the Arabidopsis TCP transcription factor family

Abstract

Analyses of the functions of TEOSINTE-LIKE1, CYCLOIDEA, and PROLIFERATING CELL FACTOR1 (TCP) transcription factors have been hampered by functional redundancy between its individual members. In general, putative functionally redundant genes are predicted based on sequence similarity and confirmed by genetic analysis. In the TCP family, however, identification is impeded by relatively low overall sequence similarity. In a search for functionally redundant TCP pairs that control Arabidopsis leaf development, this work performed an integrative bioinformatics analysis, combining protein sequence similarities, gene expression data, and results of pair-wise protein-protein interaction studies for the 24 members of the Arabidopsis TCP transcription factor family. For this, the work completed any lacking gene expression and protein-protein interaction data experimentally and then performed a comprehensive prediction of potential functional redundant TCP pairs. Subsequently, redundant functions could be confirmed for selected predicted TCP pairs by genetic and molecular analyses. It is demonstrated that the previously uncharacterized class I TCP19 gene plays a role in the control of leaf senescence in a redundant fashion with TCP20. Altogether, this work shows the power of combining classical genetic and molecular approaches with bioinformatics predictions to unravel functional redundancies in the TCP transcription factor family.

Keywords: Bioinformatics; TCP transcription factor.; gene regulation; leaf development; redundancy; senescence.

Fig. 1.

Relationships between TCP transcription factors. Relationship is based on (A) protein sequence, (B) AtGenExpress data, and (C) quantitative real-time PCR data on leaf development. The phylogenetic tree for TCP protein sequences was generated using PhyML. Trees representing expression data were generated by first converting expression patterns to distances between pairs of genes and then applying the neighbour-joining algorithm. Expression data for (B) were from the AtGenExpress microarray expression compendium by Schmid et al. 2005; expression data for (C) are from this study. class I TCPs are marked in grey.

Fig. 2.

Expression analysis of all 24 Arabidopsis TCP transcription factor genes during leaf development. Quantitative real-time PCR was performed on seedlings harvested at days 4 and 7 after germination and on the first rosette leaf harvested at days 11, 14, 16, 21, and 28 after germination. Analysis was done in triplicate and bars indicate SE. X-axis, time in days; Y-axis, normalized expression.

Fig. 3.

Results of the matrix-based yeast two-hybrid analysis of TCP–TCP protein–protein interactions. Cytoscape version 2.6.2 (Shannon et al., 2003) was used to visualize protein–protein interactions. Nodes represent the TCPs, edges represent the protein–protein interaction between these. White nodes are class II TCPs, black nodes are class I TCPs. As the graphical layout is spring embedded, groups of nodes are placed closer to each other equivalent to the number of edges between them. The representation reveals that TCPs prefer protein–protein interactions within their own class.

Fig. 4.

Functional complementation of the jaw-D phenotype by dexamethason (DEX) induction of TCP10m-GR. All material was grown for 3 weeks on half-strength MS medium with or without DEX treatment prior to phenotypic analyses. (A–C) Continuous induction of jaw-D/TCP10m-GR seedling by DEX leads to overcompensation; note that no new leaf primordia are formed. (D–G) When induced 6 days after germination and subsequently kept for the following 15 days on DEX, the first leaves of jaw-D/TCP10m-GR plants appeared normal, but plants remained small and eventually died. However, in these plants, various true leaf primordia were formed (G). (H) A representative 3-week-old untreated jaw-D/TCP10m-GR plant. (I) An untreated jaw-D control plant. (J) A jaw-D control plant continuously treated with DEX; note that, in contrast to the seedlings shown in (A–C), no effect of DEX is seen on the shoot apical meristem and the formation of leaf primordia. (K, L) Representative 3-week-old Col0 wild-type plants, with (K) and without (L) DEX treatment. Bars = 1cm (A, B, D–L), 0.3cm (C).

Fig. 5.

Phenotypic evidence for redundant functions between TCP19 and TCP20. The lines tcp8, tcp19, tcp20, and the double mutants tcp8tcp20 and tcp19tcp20 were subjected to a wound-induced senescence analysis together with the wild-type (WT) control. (A) A representative leaf for each analysed line after 4 days of incubation at room temperature in the dark. (B) In an assay involving 16 individual plants per line, tcp19tcp20 leaves showed earlier senescence in two independent experiments. Leaves of the various plant lines were categorized into four different classes based on appearance. The differences between tcp19tcp20 and the other lines are significant (P < 0.01, chi-squared test).