Global spatial analysis of Arabidopsis natural variants implicates 5'UTR splicing of LATE ELONGATED HYPOCOTYL in responses to temperature

Abstract

How plants perceive and respond to temperature remains an important question in the plant sciences. Temperature perception and signal transduction may occur through temperature-sensitive intramolecular folding of primary mRNA transcripts. Recent studies suggested a role for retention of the first intron in the 5'UTR of the clock component LATE ELONGATED HYPOCOTYL (LHY) in response to changes in temperature. Here, we identified a set of haplotypes in the LHY 5'UTR, examined their global spatial distribution, and obtained evidence that haplotype can affect temperature-dependent splicing of LHY transcripts. Correlations of haplotype spatial distributions with global bioclimatic variables and altitude point to associations with annual mean temperature and temperature fluctuation. Relatively rare relict type accessions correlate with lower mean temperature and greater temperature fluctuation and the spatial distribution of other haplotypes may be informative of evolutionary processes driving colonization of ecosystems. We propose that haplotypes may possess distinct 5'UTR pre-mRNA folding thermodynamics and/or specific biological stabilities based around the binding of trans-acting RNA splicing factors, a consequence of which is scalable splicing sensitivity of a central clock component that is likely tuned to specific temperature environments.

Keywords: 5′UTR; Arabidopsis; RNA; alternative splicing; circadian clock; natural variation; temperature; thermosensor.

Figure 1

LHY 5′UTR haplotype prevalence and potential influence on pre‐mRNA secondary structure. (a) Single nucleotide polymorphism (SNP) coordinates and prevalence (%, vertical bars) for a set of 1,131 natural Arabidopsis variants. SNP bars align with ENSEMBL transcript models for the 5′UTR of LHY (At1g01060; orange assemblies) and the constitutively spliced model (At1g01060.1, pale blue model). Horizontal bars and lines; exons and introns, respectively. (b) R4RNA arc diagrams for predicted secondary structure comparison of the (upper) G/G/U/G/C and (lower) A/U/G/C/A haplotypes. Predictions are based on the folding of 782 nt (transcriptional start site to the first AUG start codon) of LHY pre‐mRNA. Vertical arrows and dotted lines map the coordinates of SNPs 37437, 37245, and 37138 from Panel (a) onto the arcs; coloured arcs highlight three SNP‐associated arcs. (c) Regions of secondary structure divergence for the three haplotypes in Panel (b) projected (vertical dotted lines) back onto the LHY 5′UTR model featuring red symbols, putative pY regions containing UCUU/UUCU (circles; regions <15 nt, rectangles regions >15 < 30 nt) and green rectangles, potential SUA consensus binding elements UCUUCUUC, including four tandem repeats (red outline). LHY = LATE ELONGATED HYPOCOTYL

Figure 2

In vitro binding of PTB1 with LHY pre‐mRNA. (a) Predicted secondary structure (Vienna RNAfold) of intron 2 with location of SNPs 37245 and 37268 of the G/G/U/G/C haplotype highlighted. (b) In vitro RNA‐EMSAs (long and short exposures) showing the binding of recombinant PTB1 at the denoted concentrations with 3′‐biotin labelled LHY RNA (approximately 10 fg, spanning exons 1 to 5 plus the intervening introns), the three tracks on the right show competition with unlabelled RNA at the denoted fold abundance relative to labelled probe). LHY = LATE ELONGATED HYPOCOTYL [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 3

Global distribution of haplotypes. Frequency of haplotypes at the ADMIXTURE group level for the (a) WRLD and (b) WRLD_ran50 datasets. (c) Frequency of haplotypes at the latitude level. Colour shading categorization of haplotypes in Panels (b) and (c) is as Panel (a). For Panels (a) and (c), the number of accessions in each group is denoted above the bars [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 4

Haplotypes correlate with bioclimatic variables. (a) Projection of haplotypes from the WRLD_ran50 dataset onto the global index of annual mean diurnal range (BIO2, http://worldclim.org, index scale; °C × 10). Accessions are plotted across Europe, North Africa, and Central Asia. A global projection of the entire WRLD_ran50 dataset is presented in Figure S6a. (b) Principal component analysis (PC1 vs. PC2) of 20 continuous variables (11 temperature [T], 8 precipitation [P], and 1 altitude [A] variable) and categorized for the four haplotypes for the WRLD_ran50 dataset. Means and ±SEM of (c) BIO2: annual mean diurnal range (°C × 10) and (d) BIO7: temperature annual range (°C × 10) for each haplotype cohort. Pairs of means grouped by a horizontal bracket are not significantly different from each other (Tukey–Kramer method, p > .05; see Section 2) [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 5

Haplotypes correlate with elevation. (a) Projection of haplotypes from the WRLD dataset onto the global elevation profile (index; metres). Accessions are plotted across Europe, North Africa, and Central Asia. (b) Means and ±SEM of elevations for each haplotype group for the left; WRLD_ran50 and right; WRLD datasets. Pairs of means grouped by a horizontal bracket are not significantly different from each other (Tukey–Kramer method, p > .05). (c) Left, projection of haplotypes onto the Iberian Peninsula elevation profile (index; metres), and right, means of elevations for each haplotype group for the Spanish cohort; all means not significantly different from each other (Tukey–Kramer method, p > .05), except for G/G/U/G/A versus G/G/U/G/C (Tukey–Kramer method, p < .05; see Section 2) [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 6

Distinct LHY 5′UTR splicing sensitivity for two haplotypes. Expression levels, at dawn for the denoted temperature conditions, for the individual accessions (Panels (a), (c), and (e)) and means and ±SEM for the grouped haplotypes (Panels (b), (d), and (f); n = 3) for ((a) and (b)) fully spliced (FS) 5′UTR, ((c) and (d)) intron 1 retained (I1R) transcripts, and ((e) and (f)) the splice ratio (FS transcripts as a fraction of total transcripts). Expression levels for individual accessions are derived from pooled tissue (9–13 plants per temperature condition) and from two technical repeats of the qPCR assay (see Section 2) [Colour figure can be viewed at http://wileyonlinelibrary.com]