Systems Biology of Cold Adaptation in the Polyextremophilic Red Alga Galdieria sulphuraria

Abstract

Rapid fluctuation of environmental conditions can impose severe stress upon living organisms. Surviving such episodes of stress requires a rapid acclimation response, e.g., by transcriptional and post-transcriptional mechanisms. Persistent change of the environmental context, however, requires longer-term adaptation at the genetic level. Fast-growing unicellular aquatic eukaryotes enable analysis of adaptive responses at the genetic level in a laboratory setting. In this study, we applied continuous cold stress (28°C) to the thermoacidophile red alga G. sulphuraria, which is 14°C below its optimal growth temperature of 42°C. Cold stress was applied for more than 100 generations to identify components that are critical for conferring thermal adaptation. After cold exposure for more than 100 generations, the cold-adapted samples grew ∼30% faster than the starting population. Whole-genome sequencing revealed 757 variants located on 429 genes (6.1% of the transcriptome) encoding molecular functions involved in cell cycle regulation, gene regulation, signaling, morphogenesis, microtubule nucleation, and transmembrane transport. CpG islands located in the intergenic region accumulated a significant number of variants, which is likely a sign of epigenetic remodeling. We present 20 candidate genes and three putative cis-regulatory elements with various functions most affected by temperature. Our work shows that natural selection toward temperature tolerance is a complex systems biology problem that involves gradual reprogramming of an intricate gene network and deeply nested regulators.

Keywords: Cyanidiales; cold stress; extremophile; microevolution; red algae; temperature adaptation.

FIGURE 1

Growth parameters of Galdieria sulphuraria RT22. (A) Cultures were grown heterotrophically at 28°C (blue) and at 42°C (red) on plates made of 1.5% Gelrite mixed 1:1 with 2× Allen Medium containing 50 mM glucose. The fastest growing colonies were iteratively selected and re-plated over a period of ∼7 months until >100 generations were achieved under both conditions. Propagation occurred through picking the five biggest colonies from each plate and transferring them to a new plate. (B) The doubling time at 42°C was 1.32 days on average. The doubling time at 28°C was 2.70 days on average. The differences in growth between 42°C and 28°C were significant (Wilcoxon rank sum test, p = 0.0002). Cultures grown at 42°C were re-plated 10 times due to faster growth. In comparison, cultures growing at 28°C were re-plated only six times. (C) Samples grown at 42°C grew slightly slower over time. By contrast, samples grown at 28°C appeared to decrease their doubling time. While no statistically significant trend could be detected at 42°C, Jonckheere’s test for trends reported a significant trend toward faster growth for the populations grown at 28°C.

FIGURE 2

“Random” and “Non-random” variant acquisition patterns. “Non-random” variants were defined as mutations gained at some point during growth at 42°C or 28°C, and fixed in the genome of G. sulphuraria RT22 during the remaining time points. All variants were translated into binary code according to their haplotype relative to the reference genome. “Cold”: variant was obtained and fixed at 28°C. “Hot”: variant was obtained and fixed at 42°C. Left: Evolutionary patterns and their frequencies. In the specific case of patterns “111111| 0000” and “000000| 1111,” a variant was already gained before the first sampling time point. Hence, it was not possible to determine the condition at which the variant was gained. “Background” mutations represent the cases where the sequence of all samples was in disagreement with the reference genome “111111| 1111.” The remaining combinations were considered as “random” evolution patterns. Here, variants were gained but not fixated in the subsequent samples of the same growth condition. The numbers in the boxes indicate the count of a specific pattern. Right: Count by variant type. The right column of each category indicates the number of variants located in the intergenic space. The left column counts the number of variants located in the coding sequence.

FIGURE 3

(A) GO-Term (GO) enrichment analysis revealed the cellular functions most affected by mutations. Each GO was manually revised and attributed to one of the nine categories contained in the legend. (B) Differential gene expression in G. sulphuraria RT22 orthologs in G. sulphuraria 074W (reciprocal best blast hit), here measured as log -fold change (logFC) vs. transcription rate (logCPM). Differentially expressed genes are colored red (quasi-likelihood F-test, Benjamini-Hochberg, p <= 0.01). Genes affected by variants are shown by large circles. Genes without significant differential expression are represented by triangles. The blue dashes indicate the average logCPM of the dataset. The orthologs in G. sulphuraria 074W of genes affected by variance in G. sulphuraria RT22 did not show more, or less, differential expression under fluctuating temperature.