Captivity induces a sweeping and sustained genomic response in a starfish

Abstract

Marine animals in the wild are often difficult to access, so they are studied in captivity. However, the implicit assumption that physiological processes of animals in artificial environments are not different from those in the wild has rarely been tested. Here, we investigate the extent to which an animal is impacted by captivity by comparing global gene expression in wild and captive crown‐of‐thorns starfish (COTS). In a preliminary analysis, we compared transcriptomes of three external tissues obtained from multiple wild COTS with a single captive COTS maintained in aquaria for at least 1 week. On average, an astonishingly large 24% of the coding sequences in the genome were differentially expressed. This led us to conduct a replicated experiment to test more comprehensively the impact of captivity on gene expression. Specifically, a comparison of 13 wild with 8 captive COTS coelomocyte transcriptomes revealed significant differences in the expression of 20% of coding sequences. Coelomocyte transcriptomes in captive COTS remain different from those in wild COTS for more than 30 days and show no indication of reverting back to a wild state (i.e. no evidence of acclimation). Genes upregulated in captivity include those involved in oxidative stress and energy metabolism, whereas genes downregulated are involved in cell signalling. These changes in gene expression indicate that being translocated and maintained in captivity has a marked impact on the physiology and health of these echinoderms. This study suggests that caution should be exercised when extrapolating results from captive aquatic invertebrates to their wild counterparts.

Seasonal tissue-specific gene expression reveals reproductive and stress-related transcriptional systems in wild crown-of-thorns starfish

“…By comparing gene expression in seven somatic tissues from wild COTS in the summer and winter, we have revealed new insights into seasonal changes in this coral reef pest. The procurement of RNA from small, targeted biopsies from multiple starfish immediately after their removal from the reef yielded consistent gene expression profiles within a given tissue, season and sex, supporting the proposition that this replicated transcriptomic dataset accurately reflects natural gene activity [29]. This is further evidenced by each tissue transcriptome being enriched in protein coding sequences whose roles reflect the known biological function of that tissue.…”

“…On average, 69.7% of the CEL-Seq2 reads from the 134 tissue transcriptomes that passed quality control filtering mapped to the GBR v1.1 gene models (see Materials and Methods) [29], with each tissue expressing on average 15,490 protein-coding mRNAs (Table 1 and S1, S2 Tables). Principal component analysis (PCA) and hierarchical clustering analysis revealed that gene expression in each tissue is very similar between individuals, regardless of sex or season (Fig 1C, D).…”

“…Raw reads were assessed for quality and adaptor contents using FastQC [58] and analysed using the CEL-Seq2 pipeline (https://github.com/yanailab/CEL-Seq-pipeline; version 1.0) [57]. Reads were trimmed to 35 bp, demultiplexed and mapped to the GBR v1.1 genome using Bowtie2 [29]. Transcript counts of each coding sequences were generated using HTSeq [59].…”

“…On average, 69.7% of the CEL-Seq2 reads from the 134 tissue transcriptomes that passed quality control filtering mapped to the GBR v1.1 gene models (see Materials and Methods) [29], with each tissue expressing on average 15,490 protein-coding mRNAs (Table 1 and S1, S2 Tables).…”

“…purpuratus [64], sea cucumber Apostichopus japonicus [65], and starfish Asterias rubens, Patiria miniata [66] and Acanthaster planci cf. solaris (COTS) [21,29]. COTS has 506 unique gene groups.…”