Diel rewiring and positive selection of ancient plant proteins enabled evolution of CAM photosynthesis in Agave

Abstract

Background: Crassulacean acid metabolism (CAM) enhances plant water-use efficiency through an inverse day/night pattern of stomatal closure/opening that facilitates nocturnal CO₂ uptake. CAM has evolved independently in over 35 plant lineages, accounting for ~ 6% of all higher plants. Agave species are highly heat- and drought-tolerant, and have been domesticated as model CAM crops for beverage, fiber, and biofuel production in semi-arid and arid regions. However, the genomic basis of evolutionary innovation of CAM in genus Agave is largely unknown.

Results: Using an approach that integrated genomics, gene co-expression networks, comparative genomics and protein structure analyses, we investigated the molecular evolution of CAM as exemplified in Agave. Comparative genomics analyses among C₃, C₄ and CAM species revealed that core metabolic components required for CAM have ancient genomic origins traceable to non-vascular plants while regulatory proteins required for diel re-programming of metabolism have a more recent origin shared among C₃, C₄ and CAM species. We showed that accelerated evolution of key functional domains in proteins responsible for primary metabolism and signaling, together with a diel re-programming of the transcription of genes involved in carbon fixation, carbohydrate processing, redox homeostasis, and circadian control is required for the evolution of CAM in Agave. Furthermore, we highlighted the potential candidates contributing to the adaptation of CAM functional modules.

Conclusions: This work provides evidence of adaptive evolution of CAM related pathways. We showed that the core metabolic components required for CAM are shared by non-vascular plants, but regulatory proteins involved in re-reprogramming of carbon fixation and metabolite transportation appeared more recently. We propose that the accelerated evolution of key proteins together with a diel re-programming of gene expression were required for CAM evolution from C₃ ancestors in Agave.

Keywords: Circadian rhythm; Comparative genomics; Crassulacean acid metabolism; Photosynthesis; Positive selection; Transcriptome.

Fig. 1

Temporal expression of CAM and gene co-expression modules in Agave americana. a Diel expression pattern of selected co-expression modules in mature leaf, as identified from network analysis of RNA-Seq data with relevance to CAM physiology. The black and white bars indicate nighttime and daytime, respectively. b Diel expression pattern of other modules in mature leaf with distinct profiles. c Expression pattern in young leaves sampled at 3 time points and non-leaf tissues sampled at one time point (9 am). The number in the parentheses is the number of transcripts in each individual module

Fig. 2

Comparative analysis of protein sequences among CAM and non-CAM plant species. a Plant species used in comparative genomics analysis. b Ortholog groups in 15 plant species as identified by OrthoMCL. Number of ortholog groups is listed in each of the ortholog clades. c Percent of ortholog clade were predicted to be transcription factors in Agave americana. “a” and “b” indicate that the transcription factors are over-represented (p < 0.05) and under-represented (p < 0.05), respectively. d Percent of ortholog clade undergoing positive selection (i.e., nonsynonymous to synonymous substitution ratio (Ka/Ks) > 1, as calculated from Agave-Arabidopsis gene pairs with a sliding window of 50 amino acids). “*” indicates that the ortholog clade was over-represented (p < 0.0001) by Agave genes with Ka/Ks ratio greater than 1. Clade NVP:C₃:CAM:C₄ is shared by NVP, C₃, CAM, and C₄; NVP:C₃:CAM shared only by NVP, C₃, and CAM; C₃:CAM:C₄ shared only by C₃, CAM, and C₄; C₃:CAM shared only by C₃ and CAM; and CAM-only is specific to CAM species

Fig. 3

Positive selection region in phosphoenolpyruvate carboxylase kinase (PPCK1). a Ka/Ks profile of Agave americana (Aa) versus Arabidopsis thaliana (At); b superimposed structures in Aa and At, with the positive selection region highlighted; (c) Ka/Ks profile of Aa versus Oryza sativa (Os); d superimposed structures in Aa and Os, with the positive selection region highlighted; e Ka/Ks profile of Aa versus Zea mays (Zm); f superimposed structures in Aa and Zm. An ATP substrate that may bind to the Aa PPCK1 is marked by an arrow. The proteins are colored in grey for Aa, blue for At, green for Os and red for Zm. g A snapshot of PPCK1 (Aam048341) structure model revealing the positive selected sites. The PPCK1 model bound with an ATP substrate (blue surface) is after a 1-us MD simulation. K42 (codon 124, at the C-end of β3) and N73 (codon219, at the N-end of β4) are located at the two strands connecting to the αC helix, the orientation of which is known to be involved in the activation of the kinase

Fig. 4

Diel shift in gene expression pattern between Agave americana and Arabidopsis thaliana. a Morning-to-night shift with peak expression during morning in Arabidopsis and during night in Agave. b Afternoon-to-night shift with peak expression during afternoon in Arabidopsis and during late night in Agave. See gene annotation in Additional file 14: Table S11 and Additional file 15: Table S12

Fig. 5

Diel gene expression pattern of circadian system genes in Agave americana and Arabidopsis thaliana. In the circular heatmaps, the outer and inner rings represent Agave americana and Arabidopsis thaliana, respectively. The black and white half-circles inside the circular heatmaps indicate night-time and day-time, respectively. Full gene names are listed in Additional file 18: Table S13

Fig. 6

Functionally annotated Agave genes showing positive selection and diel rewiring of expression pattern relative to C₃ plants. a Genes involved in circadian clock. A gene involved in stomatal opening. c Genes involved in carboxylation, malate transport, decarboxylation, and starch/sugar metabolism. d Genes involved in photosynthetic electron transport chain. Red circles indicate positive selection. Green circles indicate morning-to-night shift in peak gene expression. Yellow circles indicate afternoon-to-night shift in peak gene expression. AGP16, Arabinogalactan protein 16; AKT2, Arabidopsis Shaker family K⁺ channels 2/3; CT-BMY, Chloroplast Beta-Amylase; ELF3, Early Flowering 3; GAUT7, Galacturonosyltransferase 7; LHY, Late Elongated Hypocotyl; LUX, Phytoclock 1; PPCK1, Phosphoenlpyruvate Carboxylase Kinase 1; PPDK-RP, Pyruvate orthophosphate dikinase regulatory protein; SS2, Starch Synthase 2; TDT, Tonoplast Dicarboxylate Transporter