Genomic basis for coral resilience to climate change

Abstract

Recent advances in DNA-sequencing technologies now allow for in-depth characterization of the genomic stress responses of many organisms beyond model taxa. They are especially appropriate for organisms such as reef-building corals, for which dramatic declines in abundance are expected to worsen as anthropogenic climate change intensifies. Different corals differ substantially in physiological resilience to environmental stress, but the molecular mechanisms behind enhanced coral resilience remain unclear. Here, we compare transcriptome-wide gene expression (via RNA-Seq using Illumina sequencing) among conspecific thermally sensitive and thermally resilient corals to identify the molecular pathways contributing to coral resilience. Under simulated bleaching stress, sensitive and resilient corals change expression of hundreds of genes, but the resilient corals had higher expression under control conditions across 60 of these genes. These "frontloaded" transcripts were less up-regulated in resilient corals during heat stress and included thermal tolerance genes such as heat shock proteins and antioxidant enzymes, as well as a broad array of genes involved in apoptosis regulation, tumor suppression, innate immune response, and cell adhesion. We propose that constitutive frontloading enables an individual to maintain physiological resilience during frequently encountered environmental stress, an idea that has strong parallels in model systems such as yeast. Our study provides broad insight into the fundamental cellular processes responsible for enhanced stress tolerances that may enable some organisms to better persist into the future in an era of global climate change.

Fig. 1.

PCA components 1 and 2 (x and y axis, respectively) of expression values for all 33,496 contigs in the reference assembly for all samples (A) and control coral samples (B). The numbers in parentheses represent the proportion of variance explained by that principal component. Specific colors reflect treatments and shapes reflect sample populations as shown in each legend (mixSym represents those colonies where <95% of a single Symbiodinium clade type was found). PCA was computed in R using the princomp function and a correlation matrix.

Fig. 2.

Venn diagram showing the number of differentially expressed genes detected during analysis based on temperature, location, within-location temperature response, and within-treatment location differences. Bold numbers in parentheses represent totals and respective shades of gray denote up- vs. down-regulated or higher in HV vs. MV, respectively.

Fig. 3.

Scatterplot of the log2 fold changes in gene expression in response to heat stress in the MV corals vs. the HV corals for the 169 genes that were unique to the MV control vs. heated comparison. Each open circle represents an individual contig, the dashed line is a 1:1 line, ** denotes a highly significant departure (P < 1e⁻¹⁵ and 1e⁻⁵ for up- and down-regulated contigs, respectively) from a 50/50 null expectation of distribution around the 1:1 line (χ² test for goodness of fit).

Fig. 4.

Scatterplot comparing the relative ratio of heat-to-control fold changes in expression between HV and MV corals (on the x axis) to the HV to MV control expression ratio (on the y axis) across the 135 up-regulated genes unique to the MV comparison set to examine whether HV controls show higher constitutive expression (points that are >1 on y axis) relative to MV controls for those genes with reduced response to heat stress in HV corals (points that are <1 on x axis). The lighter and darker portions of the graph represent the genes that are potentially frontloaded or stress indicators in expression, respectively. The trend line was calculated using a logarithmic regression, and associated R² and P values are shown in the plot.