Research ArticleECOLOGY

Epigenome-associated phenotypic acclimatization to ocean acidification in a reef-building coral

See allHide authors and affiliations

Science Advances  06 Jun 2018:
Vol. 4, no. 6, eaar8028
DOI: 10.1126/sciadv.aar8028
  • Fig. 1 Epigenetic landscape of S. pistillata.

    (A) More than half of all methylated positions in S. pistillata are located in annotated introns. (B) Introns have proportionally more methylated positions (11.3%) than exons (8.6%) or intergenic regions (3.3%) even when accounting for the different amounts of CpG dinucleotides in the respective regions. (C) Distribution of methylated positions across a standardized gene model with flanking 4-kb regions. Solid lines depict transcription start site (left) and transcription termination site (right). Exons and introns in the plot have normalized lengths that correspond to their respective mean lengths in S. pistillata (from left to right: 373, 1258, 363, 1114, 302; 253, 967, 312, 1037, and 669 bp). arb. units, arbitrary units.

  • Fig. 2 Effect of gene body methylation on genic expression.

    (A) The expression of methylated genes is significantly higher than that of unmethylated genes (P < 10−300, Student’s t test). (B) Expression values for methylated genes are exponentially proportional to methylation level; however, the relationship of expression to methylation density is nonlinear. Methylation density at low levels is exponentially proportional to expression values, but it plateaus at ~40%. (C) Expression levels of the first six exons were calculated as natural logarithms of fold changes (FC) relative to the expression of the first exon. The difference in expression levels is likely driven by the reduction of cryptic transcription initiation in methylated genes. The difference is greater in highly methylated genes (median methylation level >80%). Asterisks represent P values from t tests of methylated (orange) or highly methylated genes (red) against unmethylated genes (peach), and colored accordingly. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (D) There is a linear relationship between the inverse of the coefficient of variation and the log10-transformed mean expression values from all samples. The coefficient of variation, defined as the SD of measured expression values divided by their mean, is consistently lower in methylated genes than unmethylated genes.

  • Fig. 3 Differential methylation of genes in the MAPK/JNK signaling and cell growth pathways in the coral S. pistillata under pH stress.

    (A) Most of the general growth– and stress response–related Gene Ontology (GO) terms are significantly enriched at pH 7.2. GTPase, guanosine triphosphatase. (B) A composite pathway diagram consisting of KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways “ko04010”: “MAPK signaling pathway” and “ko04110”: “cell cycle” was produced in Cytoscape. At pH 7.2, more genes respond with a significant increase (red) than with a decrease (blue) in methylation; unshaded genes are methylated but not differentially regulated. (C) The methylation of genes that negatively regulate MAPK and JNK (red lines) increases, whereas the methylation of genes that positively regulate the same kinases (blue lines) decreases in a reciprocal manner (error bars denote ±1 SE).

  • Fig. 4 Effect of lower pH levels on cell growth.

    (A) Mean methylation levels were significantly increased at pH 7.2. (B) Cell sizes were significantly larger in nubbins grown in the pH 7.2 tank. (C) Larger cell sizes were associated with skeletal structure that contained larger calyxes. (D) The skeletal porosity was significantly higher at lower pH. (E) Representative longitudinal sections of S. pistillata skeletons under pHs of 7.2 and 8.0. Error bars represent 1 SE. Asterisks denote significance of t test P values. **P < 0.01; ***P < 0.001.

  • Table 1 Carbonate chemistry parameters in the four experimental seawater pH treatments.

    Parameters of carbonate seawater chemistry were calculated from total scale pH, total alkalinity (TA), temperature, and salinity using the open-access CO2SYS package (33) using constants from the study of Mehrbach et al. (34) as refit by Dickson and Millero (35). Values shown are means ± SD. pHT, pH in treatment tanks; Ωar, saturation state of aragonite.

    Treatment namepHTTA
    (mmol/kgSW)
    pCO2
    (μatm)
    HCO3
    (μmol/kgSW)
    CO32−
    (μmol/kgSW)
    Total carbon
    (μmol/kgSW)
    Ωar
    7.27.222496.623513.152384.1346.252528.270.72
    ±0.01±6.99±67.20±4.42±1.08±3.62±0.02
    7.47.432474.162109.32299.4371.662429.861.11
    ±0.01±5.89±59.63±0.76±2.13±1.23±0.03
    7.87.812461.77798.772082.22155.172259.642.41
    ±0.01±5.97±14.36±0.60±2.75±1.75±0.04
    87.952447.48537.661951.14202.412168.533.14
    ±0.00±7.16±4.88±1.90±2.25±4.02±0.03

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/6/eaar8028/DC1

    discussion S1. Comparisons with previous methylation studies in corals.

    discussion S2. S. pistillata promoter methylation has no effect on gene expression.

    discussion S3. Correlation of genic expression to methylation.

    discussion S4. Validation of methylation patterns.

    discussion S5. Differential expression and methylation of biomineralization genes.

    fig. S1. Amplicon-specific bisulfite sequencing can accurately assay methylation levels of amplicons of interest.

    fig. S2. Down-regulation of Ras and Ras guanine nucleotide exchange factors and up-regulation of Ras GTPase-activating proteins suggest a reduction in active Ras.

    fig. S3. Differential expression of key genes was corroborated using RT-qPCR.

    fig. S4. Distribution of CpGO/E (“CpG bias”) for genes in S. pistillata.

    fig. S5. Multiple lines of evidence suggesting that promoter methylation does not influence expression patterns in S. pistillata.

    fig. S6. Methylation patterns are strongly tissue-specific.

    fig. S7. Effect of long-term pH stress on selected biomineralization-related genes.

    table S1. Modeling transcriptional noise as a function of expression level and methylation state.

    data S1. Trimming, mapping, and coverage calculations for WGBS reads.

    data S2A. Differentially methylated genes at pH 7.2, relative to control.

    data S2B. Differentially methylated genes at pH 7.4, relative to control.

    data S2C. Differentially methylated genes at pH 7.8, relative to control.

    data S3A. Enriched GO terms for differentially methylated genes at pH 7.2.

    data S3B. Enriched GO terms for differentially methylated genes at pH 7.4.

    data S3C. Enriched GO terms for differentially methylated genes at pH 7.8.

    data S4A. Sequences and miscellaneous details of primers used in qPCR.

    data S4B. Raw Ct values of qPCR experiment.

    data S4C. ddCt approach for fold change calculations.

    data S5A. Raw values of cell size measurements.

    data S5B. Raw values of calyx size measurements.

    data S5C. Raw values of porosity measurements.

    data S6A. Enriched GO terms for all methylated genes.

    data S6B. Enriched GO terms for highly methylated genes.

    data S7. Methylation levels measured via amplicon sequencing and WGBS.

    data S8. Results from tissue-specific amplicon sequencing.

    data S9. List of putative biomineralization-related genes in S. pistillata.

    References (4958)

  • Supplementary Materials

    This PDF file includes:

    • discussion S1. Comparisons with previous methylation studies in corals.
    • discussion S2. S. pistillata promoter methylation has no effect on gene expression.
    • discussion S3. Correlation of genic expression to methylation.
    • discussion S4. Validation of methylation patterns.
    • discussion S5. Differential expression and methylation of biomineralization genes.
    • fig. S1. Amplicon-specific bisulfite sequencing can accurately assay methylation levels of amplicons of interest.
    • fig. S2. Down-regulation of Ras and Ras guanine nucleotide exchange factors and up-regulation of Ras GTPase-activating proteins suggest a reduction in active Ras.
    • fig. S3. Differential expression of key genes was corroborated using RT-qPCR.
    • fig. S4. Distribution of CpGO/E (“CpG bias”) for genes in S. pistillata.
    • fig. S5. Multiple lines of evidence suggesting that promoter methylation does not influence expression patterns in S. pistillata.
    • fig. S6. Methylation patterns are strongly tissue-specific.
    • fig. S7. Effect of long-term pH stress on selected biomineralization-related genes.
    • table S1. Modeling transcriptional noise as a function of expression level and methylation state.
    • References (49–58)

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • data S1 (Microsoft Excel format). Trimming, mapping, and coverage calculations for WGBS reads.
    • data S2A (Microsoft Excel format). Differentially methylated genes at pH 7.2, relative to control.
    • data S2B (Microsoft Excel format). Differentially methylated genes at pH 7.4, relative to control.
    • data S2C (Microsoft Excel format). Differentially methylated genes at pH 7.8, relative to control.
    • data S3A (Microsoft Excel format). Enriched GO terms for differentially methylated genes at pH 7.2.
    • data S3B (Microsoft Excel format). Enriched GO terms for differentially methylated genes at pH 7.4.
    • data S3C (Microsoft Excel format). Enriched GO terms for differentially methylated genes at pH 7.8.
    • data S4A (Microsoft Excel format). Sequences and miscellaneous details of primers used in qPCR.
    • data S4B (Microsoft Excel format). Raw Ct values of qPCR experiment.
    • data S4C (Microsoft Excel format). ddCt approach for fold change calculations.
    • data S5A (Microsoft Excel format). Raw values of cell size measurements.
    • data S5B (Microsoft Excel format). Raw values of calyx size measurements.
    • data S5C (Microsoft Excel format). Raw values of porosity measurements.
    • data S6A (Microsoft Excel format). Enriched GO terms for all methylated genes.
    • data S6B (Microsoft Excel format). Enriched GO terms for highly methylated genes.
    • data S7 (Microsoft Excel format). Methylation levels measured via amplicon sequencing and WGBS.
    • data S8 (Microsoft Excel format). Results from tissue-specific amplicon sequencing.
    • data S9 (Microsoft Excel format). List of putative biomineralization-related genes in S. pistillata.

    Download data S1 to S9

    Files in this Data Supplement:

Navigate This Article