Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2013 Mar;23(3):486-96.
doi: 10.1101/gr.148361.112. Epub 2012 Dec 3.

A chromatin link to caste identity in the carpenter ant Camponotus floridanus

Daniel F Simola  1 , Chaoyang Ye, Navdeep S Mutti, Kelly Dolezal, Roberto Bonasio, Jürgen Liebig, Danny Reinberg, Shelley L Berger
Affiliations

A chromatin link to caste identity in the carpenter ant Camponotus floridanus

Daniel F Simola et al. Genome Res. 2013 Mar.

Abstract

In many ant species, sibling larvae follow alternative ontogenetic trajectories that generate striking variation in morphology and behavior among adults. These organism-level outcomes are often determined by environmental rather than genetic factors. Therefore, epigenetic mechanisms may mediate the expression of adult polyphenisms. We produced the first genome-wide maps of chromatin structure in a eusocial insect and found that gene-proximal changes in histone modifications, notably H3K27 acetylation, discriminate two female worker and male castes in Camponotus floridanus ants and partially explain differential gene expression between castes. Genes showing coordinated changes in H3K27ac and RNA implicate muscle development, neuronal regulation, and sensory responses in modulating caste identity. Binding sites of the acetyltransferase CBP harbor the greatest caste variation in H3K27ac, are enriched with motifs for conserved transcription factors, and show evolutionary expansion near developmental and neuronal genes. These results suggest that environmental effects on caste identity may be mediated by differential recruitment of CBP to chromatin. We propose that epigenetic mechanisms that modify chromatin structure may help orchestrate the generation and maintenance of polyphenic caste morphology and social behavior in ants.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Genome-wide patterns of histone PTMs in Camponotus floridanus. (A) Prevalence of histone PTMs and RNA Pol II, partitioned into genic (Exon, Intron, 5′ and 3′ UTRs, 2-kb promoter), Noncoding, and Intergenic groups. Prevalence was computed by counting nucleotides covered by significant regions of interest (ROIs; P < 0.05) and averaging over major, minor, and male estimates. Prevalence over repetitive DNA, defined here as sequences appearing multiple times in the genome, is shown below (see Supplemental Fig. 6C for specific classes of repetitive elements); by this definition, ∼22% of the genome is repetitive (Supplemental Methods). An ROI is considered repetitive if the majority of nucleotide loci delimited by this ROI are repetitive. Error bars denote 1 standard error (SE) over castes. (B) Quantitative profiles of ChIP enrichment across gene loci, formula image,…, formula image for ChIP samples p, averaged over castes and all protein-coding genes that contain significant PTM prevalence (14,368 genes). Error bars denote SE over genes. (TSS) Transcription start site; (Exon) first (1), internal (i), and last (n) exons; (TTS) transcription termination site; (kb) kilobase. (C) Visualization of the 351-dimensional ChIP-seq data set summarized for 9861 protein-coding genes, grouped by k-means clustering into k = 4 significant clusters (chromatin domains) using AIC model selection and a Euclidean distance measure. Columns were grouped by hierarchical clustering with a Euclidean distance measure. (Middle) ChIP enrichment profiles averaged over gene loci from each chromatin domain. On right, average expression levels for genes in each chromatin domain, separated by caste. Error bars denote SE over genes. (FPKM) Fragments per kilobase per million (mapped) reads (proxy for mRNA expression level).
Figure 2.
Figure 2.
Histone PTMs discriminate ant castes. (A) High-resolution profiles for H3K36me3 and H3K27ac spanning 600–700 kb on scaffold107. Asterisks denote significant differentially marked regions of interest (dmROIs) between majors and minors (P < 0.01) and are colored by enriched caste (major, green; minor, blue). (B) Example dmROIs for H3K27ac (yellow) with 1-kb flanking sequence. Error bars denote SE over two biological replicates. (C) Comparison of major (top) and minor (bottom) ChIP enrichment profiles for H3K27ac over a single gene locus (Cflo_05605), including 2-kb flanking intergenic sequence. (TSS) Transcription start site; (Exon) first (1), internal (i), and last (n) exons; (TTS) transcription termination site; (kb) kilobase. (D) Numbers of genes with TSS-proximal dmROIs (50 bp–2 kb). The expected number of genes for each distance cutoff is shown on right, using randomly sampled coordinates (average of 100 replicates shown). (Dashed line) Expected cutoff for dmROIs within 150 bp of a gene TSS. (E) Two-dimensional linear discriminant (LD) analysis of caste, using ChIP enrichment for 9944 protein-coding genes (13 genic regions, nine ChIP samples per gene). Similar analyses for H3K27ac and H3K9ac also shown on right; Supplemental Figure 10D shows a similar plot for H3K27me3. Data points are labeled by genic region and/or ChIP sample and are colored by caste as indicated. (F) Caste separation formula image was assessed near genes as shown by randomly sampling 10,000 loci and performing LD analysis, as done in E. Each distribution shows n > 30 replicates. (Boxes) 25th–75th percentiles; (whiskers) 5th–95th percentiles; (points) outliers. (bp) Base pairs.
Figure 3.
Figure 3.
Changes in H3K27ac reveal caste-specific transcriptome states. (A) Meta-gene average ChIP enrichment profiles for five PTMs, pooling caste data. ChIP enrichment profiles are grouped into five expression categories (20th percentiles of the genome-wide distribution of gene expression levels). R2 values indicate ordinary least-squares regression fit of gene-proximal (gene body ± 2 kb) ChIP enrichment data to log2(FPKM+1) expression (Supplemental Table 4). Error bars denote SE over genes. (TSS) Transcription start site; (Exon) first (E1), internal (Ei), and last (En) exons; (TTS) transcription termination site; (kb) kilobase. (B) Relationship between the difference in log2(FPKM+1) expression levels (y-axis) and the difference in H3K27ac ChIP enrichment (x-axis) between majors and minors. ChIP enrichment data were fitted by linear regression to mRNA expression using 11,459 genes. “High-confidence” genes contain no missing data requiring imputation across gene loci. “Diff. RNA” genes have significant differential expression (FDR < 0.15) and are significant in a χ2 multivariate outlier analysis. “Caste-specific” genes also exhibit significant ChIP enrichment and are grouped by false discovery rate (0.25 < FDR < 0.01). See Supplemental Methods for further details. Numbers of significant GO categories enriched using caste-biased genes (i.e., passing multivariate outlier analysis but not necessarily with significant changes in expression or H3K27ac) from each quadrant are shown (FDR < 0.01). (C) ChIP enrichment profiles for H3K27ac over caste-specific gene loci (FDR < 0.25) for 40 major-specific genes (top) and 79 minor-specific genes (bottom). Error bars reflect 1 SE over genes. (D) Validation of caste-specific genes from B. RNA levels (black bars) were measured by RT-qPCR and normalized to GAPDH levels, which shows similar expression both among castes and among biological replicates within caste (i.e., similar standard errors; data not shown). H3K27ac enrichment (gray bars) were measured in 5′ promoters by ChIP-qPCR and normalized to total H3. See Supplemental Table 10 for primer sequences. Average and standard deviation over five independently founded ant colonies are shown. Significance assessed by two-sample, unequal variance t-test: (*) P < 0.05; (**) P < 0.01; (***) P < 0.001. (E) Pearson correlations of between-caste differences in ChIP enrichment between pairs of PTMs, RNA Pol II, and mRNA expression using either all genes (n = 11,459; bottom right diagonal) or caste-specific genes (n = 119; top left diagonal). Reported values indicate the averages from major vs. minor and male vs. female comparisons. (Yellow) R = +1; (dark blue) R = −1.
Figure 4.
Figure 4.
Ant brains exhibit caste-specific variability. (A) mRNA expression and ChIP enrichment tracks from brain and head+thorax (H+T) samples for two gene loci, vitellogenin and mGluR2. Asterisks denote significant differential expression (FDR < 0.25). Total H3 brain samples are shown for vitellogenin for reference. Gene length is indicated above each gene model. Average ChIP enrichment across each gene locus is reported on right. (TSS) Transcription start site; (Exon) first (1), internal (i), and last (n) exons; (TTS) transcription termination site; (kb) kilobase. (B) Scatterplot comparisons of H3K27ac ChIP enrichment between brain and H+T tissues, shown separately for major and minor data. Caste-specific genes identified from H+T data are highlighted in blue to show their enrichment in brain tissue. Pearson correlation coefficients are reported for each caste. (C) Mean changes in RNA (left) and H3K27ac ChIP enrichment (right) for H+T caste-specific genes (over TSS, CDS, Exon, Intron, and TTS), measured in brains and in H+T. Error bars indicate SE over genes. P-values estimated using a two-tailed Mann-Whitney U-test. (D) Scatterplot comparisons of major vs. minor differences in RNA (left) and H3K27ac (right) between H+T and brain tissue samples, for H+T caste-specific genes. Concordance statistics report the percentage of H+T caste-specific genes that exhibit the same direction of change (major or minor) in brain and H+T samples. (FPKM) Fragments per kilobase per million (mapped) reads (proxy for mRNA expression level).
Figure 5.
Figure 5.
Estimates of the percentage of tissue-specific caste variation present in H+T data when comparing majors and minors. Shown are proportions of variation (in the distribution of ChIP enrichment differences for all protein-coding genes) explained by a linear regression model allowing allometry in 50 independent tissue types. Results for 250 simulated tissue-specific ChIP enrichment matrices are shown for each ChIP sample. See Supplemental Figure 18 for complete results and Supplemental Methods for more details. (Boxes) 25th–75th percentiles; (whiskers) 5th–95th percentiles; (points) outliers. Adjusted R2 values estimated by weighted least squares, with the same weights used for caste differential regression of RNA and ChIP data in Figure 3B.
Figure 6.
Figure 6.
CBP colocalizes with dmROIs and identifies putative transcription factor binding partners. (A) Cumulative distribution of distance (kb) from a CBP ROI to the nearest dmROI. “All” curve computed using the pool of all dmROIs for PTMs and Pol II (n = 161,401). “Random” curve computed using the average distribution obtained by randomly generating 161,401 ROIs with the observed ROI length distribution (n = 100 replicates). Error bars for “Random” curve denote SE over replicates. Dashed line marks 2-kb distance, which accounts for 92% of CBP ROIs. (B) Linear discriminant (LD) analysis for H3K27ac ChIP enrichment using all 22,353 CBP ROIs (diamonds in cartoon). F-value denotes observed caste separation of H3K27ac at CBP ROIs. Boxplot shows distribution of 250 random F-values obtained by LD analysis using the same number of ROIs drawn randomly genome-wide with lengths from Supplemental Figure 21A but excluding CBP ROIs. (C) Mean difference in ChIP enrichment between majors and minors for sets of major-specific, minor-specific, and remaining (nonspecific) genes. Error bars denote SE over genes. Asterisks denote significant deviation from 0 (one sample t-test, P < 0.01). R-values denote Pearson correlation of CBP versus H3K27ac or H3K9ac across caste-specific gene loci (gene body ±2kb), using ChIP enrichment values fitted by linear regression with RNA. (D) Nine TF motifs significantly enriched in promoters (2 kb) of caste-specific genes from worker and sex comparisons (FDR < 0.1). Seven of these motifs (denoted by asterisks) are also enriched in DNA sequences denoted by all CBP ROIs (FDR < 0.05). Four nonenriched TFs are shown as controls.
Figure 7.
Figure 7.
Genes with evolutionarily increased CBP binding exhibit increased caste variability. (A) Comparison of significant CBP ROIs identified by the same method and parameters using adult female ChIP-seq data for D. melanogaster and C. floridanus. Total significant ROIs are indicated on left (P < 0.05) and total assembled genome sizes are indicated (in megabase pairs, Mb) on right. (Uextra and mitochondrial chromosomes are excluded for Drosophila). (B) Distributions of CBP ROIs occurring within open reading frames (ORFs; exons+introns) of orthologous genes (n = 7042 orthologs). (C) Comparison of mRNA expression (left) and ChIP enrichment (right) between majors and minors for 601 genes showing more than five CBP ROIs in ants compared with flies; similar results found using different cutoffs (data not shown). Error bars denote SE over genes. ChIP enrichment profiles are shown for CBP, H3K4me1, and H3K27ac averaged over the CBP ROIs located within 50 kb of these ant CBP-enriched genes. (FPKM) Fragments per kilobase per million (mapped) reads (proxy for mRNA expression level). (D) Absolute difference in log2(FPKM+1) mRNA expression between majors and minors for 7042 orthologs (gray), 1219 genes with at least one more CBP ROI in ants than flies (green), and 5562 genes with at least one more CBP ROI in flies than ants (red). RNA-seq data from head+thorax and brain tissues are shown. P-values estimated using a two-tailed Mann-Whitney U-test.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Alarcón JM, Malleret G, Touzani K, Vronskaya S, Ishii S, Kandel ER, Barco A 2004. Chromatin acetylation, memory, and LTP are impaired in CBP+/− mice: A model for the cognitive deficit in Rubinstein-Taybi syndrome and its amelioration. Neuron 42: 947–959 - PubMed
    1. Amberg DC, Burke DJ, Strathern JN 2000. Methods in yeast genetics: A Cold Spring Harbor laboratory course manual, pp. 149–154. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY
    1. Ament SA, Corona M, Pollock HS, Robinson GE 2008. Insulin signaling is involved in the regulation of worker division of labor in honey bee colonies. Proc Natl Acad Sci 105: 4226–4231 - PMC - PubMed
    1. Bantignies F, Goodman RH, Smolik SM 2000. Functional interaction between the coactivator Drosophila CREB-binding protein and ASH1, a member of the trithorax group of chromatin modifiers. Mol Cell Biol 20: 9317–9330 - PMC - PubMed
    1. Bejarano F, Busturia A 2004. Function of the trithorax-like gene during Drosophila development. Dev Biol 268: 327–341 - PubMed

Publication types

MeSH terms