Genome-scale analysis of Arabidopsis splicing-related protein kinase families reveals roles in abiotic stress adaptation

Abstract

Nearly 60 - 80 % of intron-containing plant genes undergo alternative splicing in response to either stress or plant developmental cues. RNA splicing is performed by a large ribonucleoprotein complex called the spliceosome in conjunction with associated subunits such as serine arginine (SR) proteins, all of which undergo extensive phosphorylation. In plants, there are three main protein kinase families suggested to phosphorylate core spliceosome subunits and related splicing factors based on orthology to human splicing-related kinases: the SERINE/ARGININE PROTEIN KINASES (SRPK), ARABIDOPSIS FUS3 COMPLEMENT (AFC), and Pre-mRNA PROCESSING FACTOR 4 (PRP4K) protein kinases. To better define the conservation and role(s) of these kinases in plants, we performed a genome-scale analysis of the three families across photosynthetic eukaryotes, followed by extensive transcriptomic and bioinformatic analysis of all Arabidopsis thaliana SRPK, AFC, and PRP4K protein kinases to elucidate their biological functions. Unexpectedly, this revealed the existence of SRPK and AFC phylogenetic groups with distinct promoter elements and patterns of transcriptional response to abiotic stress, while PRP4Ks possess no phylogenetic sub-divisions, suggestive of functional redundancy. We also reveal splicing-related kinase families are both diel and photoperiod regulated, implicating different orthologs as discrete time-of-day RNA splicing regulators. This foundational work establishes a number of new hypotheses regarding how reversible spliceosome phosphorylation contributes to both diel plant cell regulation and abiotic stress adaptation in plants.

Keywords: Abiotic Stress; Diel cycle; Evolution; Protein kinases; mRNA splicing.

Fig. 1

Maximum likelihood phylogenetic tree of SRPK kinases across unicellular and multicellular photosynthetic and select non-photosynthetic organisms. Key nodes are labelled with branch support values from maximum likelihood interference (IQTree), bayesian (Mr. Bayes) and an additional maximum likelihood interference (PhyML), respectively. Node A: (0.97, 1.00, 0.95); Node B (0.98, 0.99, 0.97); Node C: (1.00,1.00, 0.99); Node D: (1.00, 1.00, 1.00); Node E: (1.00, 1.00, 0.80); Node F: (1.00, 0.99, 0.99)

Fig. 2

Maximum likelihood phylogenetic tree of AFC kinases across unicellular and multicellular photosynthetic and select non-photosynthetic organisms. Key nodes are labelled with branch support values from maximum likelihood interference (IQTree), bayesian (Mr. Bayes) and an additional maximum likelihood interference (PhyML), respectively. Node A: (1.00, 1.00, 1.00); Node B (1.00, 1.00, 0.99); Node C: (0.97, 1.00, 0.96); Node D: (0.88, 0.93, 0.78); Node E: (0.99, 1.00, 0.99)

Fig. 3

Maximum likelihood phylogenetic tree of PRP4K kinases across unicellular and multicellular photosynthetic and select non-photosynthetic organisms. Key nodes are labelled with branch support values from maximum likelihood interference (IQTree), bayesian (Mr. Bayes) and an additional maximum likelihood interference (PhyML), respectively. Node A: (1.00, 1.00, 1.00); Node B: (0.77, 0.78, 0.80); Node C: (0.86, 1.00, 0.90); Node D: (0.98,0.86, 0.95); Node E (0.98, 0.56, 0.96)

Fig. 4

Comparative phylogenetic analysis of conserved domain across phylogenetics groups: opisthokonts, bryophytes, gymnosperms, monocots, and eudicots. Positionality and length of domain is displayed on the x-axis, while the y-axis represents number of organisms whose peptide sequence contains the identified domain. Domain prediction was acquired through PFAM using DomainViz (http://uhriglabdev.cirrus.ualberta.ca/domainviz; [57])

Fig. 5

Identification of putative cis regulatory elements (CREs) on the gene sequence of SRPKs, AFCs, and PRP4Ks. Presence of CREs is denoted by dark blue while absence is denoted by light blue. Data was acquired by mining AtCisDB database (https://agris-knowledgebase.org/AtcisDB/; [68])

Fig. 6

Relative transcript abundance at various stages of Arabidopsis development, from seed to senescence. Values were acquired from BAR ePlant and absolute values were log₂ transformed (https://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi; [94])

Fig. 7

Relative Log₂ fold change of splicing-related kinases transcript abundance under abiotic stresses: osmotic, salt, heat, and cold stress. Each stress was induced with parallel parameters from Kilian et al., 2007. 4 replicates were averaged and the comparison to control were FDR adjusted p-values

Fig. 8

Relative transcript abundance under various photoperiods: 8 h: 16 h, 12 h: 12 h, 16 h: 8 h, 24 h: 0 h. Normalized values were log₂ transformed and averaged across replicates