Evolutionary proteomics identifies amino acids essential for ligand-binding of the cytokinin receptor CHASE domain

Abstract

Background: In plants the hormone cytokinin is perceived by members of a small cytokinin receptor family, which are hybrid sensor histidine kinases. While the immediate downstream signaling pathway is well characterized, the domain of the receptor responsible for ligand binding and which residues are involved in this process has not been determined experimentally.

Results: Using a live cell hormone-binding assay, we show that cytokinin is bound by a receptor domain predicted to be extracellular, the so called CHASE (cyclases, histidine kinase associated sensory extracellular) domain. The CHASE domain occurs not only in plant cytokinin receptors but also in numerous orphan receptors in lower eukaryotes and bacteria. Taking advantage of this fact, we used an evolutionary proteomics approach to identify amino acids important for cytokinin binding by looking for residues conserved in cytokinin receptors, but not in other receptors. By comparing differences in evolutionary rates, we predicted five amino acids within the plant CHASE domains to be crucial for cytokinin binding. Mutagenesis of the predicted sites and subsequent binding assays confirmed the relevance of four of the selected amino acids, showing the biological significance of site-specific evolutionary rate differences.

Conclusion: This work demonstrates the use of a bioinformatic analysis to mine the huge set of genomic data from different taxa in order to generate a testable hypothesis. We verified the hypothesis experimentally and identified four amino acids which are to a different degree required for ligand-binding of a plant hormone receptor.

Figure 1

Domain structure of CRE1/AHK4 and secondary structure prediction of the CHASE domain. (A) The domain structure of the full-length protein and different truncated versions of CRE1/AHK4 used in this study. (B) The predicted secondary structure of the CHASE domain with the amino acid substitutions tested in the cytokinin binding assay marked in red. White tubes represent α-helices and grey areas β-sheets. Secondary structure prediction was done by PSIPred v2.4 [43]. Abbreviations: CHASE, cyclases, histidine kinase associated sensory extracellular; HisKA, histidine kinase A domain; HATPase, histidine kinase-like ATPase; Rec, receiver domain; Tmds, transmembrane domains.

Figure 2

The CHASE domain of CRE1/AHK4 is necessary and sufficient for cytokinin binding. In vitro binding of trans-[2-3H]zeatin to full length CRE1/AHK4 protein or different domains of the receptor overexpressed in E. coli BL21. Bacterial cells were assayed for specific trans-[2-3H]zeatin binding. Data are mean ± s.d.; n = 4.

Figure 3

Phylogenetic tree and alignment of CHASE domains. (A) Phylogenetic tree of CHASE domains from five different subgroups (see Additional file 1 for all subgroups). Sequences of CHASE domains used for tree building are labelled with gi numbers and the start and end position of the respective CHASE domain. (B) Section of CHASE family alignment containing sequences used for the evolutionary rate analysis of the individual amino acids. The evolutionary site rate categories for each subgroup are given below the alignment, ranging from 1 (slow rate of evolution) to 8 (fast rate of evolution). Amino acid positions selected for experimental analysis (W244, K297, F304, R305, T317) are highlighted in orange, positions that are conserved throughout all CHASE sequences in grey. Interruptions of the alignment are indicated by two dots, gaps in the alignment by dashes. (See Additional file 1 for full sequence alignment). The selected blocks relate to the CRE1/AHK4 sequence 238–249, 292–308 and 313–321, respectively.

Figure 4

Identification of amino acid residues of the CRE1/AHK4 CHASE domain important for ligand binding. (A) Effect of point mutations in the CRE1/AHK4 CHASE domain on the specific binding of trans-[2-3H]zeatin. For the localization of the mutated sites see Fig. 1B. H482Q is a control carrying a mutated histidine residue of the cytoplasmic domain. Data are means ± S.D. from measurements with two different E. coli clones for each construct. (B) Western blot with whole protein extracts of an aliquot of E. coli cells used for the binding assay shown in (A). For protein detection a mouse-anti-GST antibody (GST B-14, Santa Cruz Biotechnology) was used.