Epigenetics and chromatin structure regulate var2csa expression and the placental-binding phenotype in Plasmodium falciparum
- Department of Molecular, Cell and Systems Biology, University of California, Riverside, United States
- Department of Microbiology, Tumor and Cell Biology (MTC), Karolinska Institutet, Sweden
- Department of Cell and Molecular Biology (ICM), Uppsala University, Sweden
Figures
Figure 1 with 1 supplement
Generation of parasites with a homogenous placental-binding phenotype and consequential dominant expression of var2csa.
(A) The presence of in vivo relevant knob structures on pRBCs was confirmed by SEM prior to phenotypic selection. Shown is a representative image of a pRBC with knobs appearing as white protrusions. (B) Flow cytometry-based antibody recognition of VAR2CSA surface expression on repeatedly selected pRBCs. NF54 ptefKO served as negative control, NF54 was the original and unselected parasite line, NF54CSAm was an intermediately selected line and NF54CSAh was panned enough to achieve a homogenous population of VAR2CSA-expressing pRBCs. (C) Representative image of a Giemsa-stained placental section with bound NF54CSAh pRBCs. Tissue boundaries are indicated by white dashed lines. (D) Quantification of relative placental binding of NF54CSAh from three biological replicates. pRBCs bound per mm2 were normalized to the VAR2CSA negative NF54 ptefKO parasite line (n=3). (E) Spearman correlation of var gene expression profiles determined by RNA-seq for three bio-replicates each of NF54CSAh and the original, unselected NF54. (F) Differential gene expression between NF54CSAh and NF54 identified a limited number of significant genes (FDR <0.05, Log2 fold change >1) of which the majority were var genes (var2csa in blue, all other var genes in green). (G) Normalized read counts for all var genes revealed NF54 to be phenotypically highly heterogeneous, whereas var2csa was the only var gene expressed by NF54CSAh.
Figure 1—figure supplement 1
Genome-wide expression profiles are highly similar among all samples.
(A) Correlation heatmap of regularized logarithmic read counts displaying strong similarity in expression profiles between replicates and samples. (B) Principal component analysis shows similar variance between replicates of each sample.
Figure 2 with 1 supplement
The var2csa locus is devoid of repressive H3K9me3 in placental-binding parasites.
(A) Chromosomal distribution of repressive H3K9me3 as determined by ChIP-seq from merged biological triplicates of NF54 and NF54CSAh. Regional clusters of antigenically variable genes are noted with green below each plot. The only noticeable difference between the parasite lines was located to the left arm of chromosome 12, which contains the var2csa locus (dashed box). (B) Heatmaps of H3K9me3 occupancy in NF54 and NF54CSAh for all var genes (0.5 kb upstream of the transcription start site (TSS) and 1.5 kb into the gene body) with var2csa displayed with bolded gene ID. (C, D) H3K9me3 enrichment in the antigenically variable multigene families var, rif, stevor, and pfmc-2tm partitioned into 5’ UTR, exons, and introns versus the rest of the genomes (all other genes and intergenic regions) of NF54 and NF54CSAh.
Figure 2—figure supplement 1
Differential peak calling of H3K9me3 reveals highly focused changes in var2csa region.
(A) Correlation analysis indicates similarity in global binding of H3K9me3 between replicates and samples. (B) Read counts are greater in NF54 than NF54CSAh within all peaks identified by differential binding (FDR<0.05) analysis. (C) There are only four significantly differentially bound (FDR <0.05) sites between NF54 and NF54CSAh with a log₂ fold change greater than 1 (black), all of which contain more reads in NF54 than NF54CSAh. (D) Read counts for the four differentially bound regions are consistent among replicates within each sample. (E) Heatmaps of input-normalized H3K9me3 occupancy for all genes separated by chromosome in NF54 and NF54CSAh (0.5 kb 5’ of the transcription start site (TSS) and 1.5 kb 3’ of the TSS). Genes within each chromosome are sorted by the mean read count within the 2 kb region surrounding the TSS. A majority of genes across all chromosomes contain low levels of H3K9me3 surrounding the TSS.
Figure 3 with 4 supplements
Changes in chromatin organization and perinuclear repositioning upon var2csa activation and silencing.
(A–C) Hi-C generated and normalized contact counts representing intrachromosomal interactions for chromosomes 2, 4, and 12 in NF54 (left panels) and NF54CSAh (middle panels). Differential intrachromosomal interactions between the two parasite lines (right panels) displayed a general increase in interactions between subtelomeric and chromosome internal var gene clusters for NF54CSAh compared to NF54. Chromosome 12 (C) stood out as the only exception with decreased interactions in NF54CSAh for the subtelomere containing the var2csa locus. (D) 3D chromatin modeling for NF54 displays a polarized nucleus with clustering of centromeres (gray) and telomeres (red) in distinct regions. The majority of var genes (green), including var2csa (blue), are located in close vicinity to the telomeric cluster. (E) The 3D chromatin model for NF54CSAh displayed a similar topology to NF54 with the exception of decreased distance between var genes (green) and telomere ends (red) overall and an increased distance of var2csa (blue) from the telomeric cluster.
Figure 3—figure supplement 1
There is a strong correlation in genome-wide interactions between samples and replicates and between contact count probability and genomic distance.
(A) Distance normalized correlation heatmap showing similar correlation between replicates of both samples. (B) Scatter plot indicating a negative log-linear relationship between contact probability and genomic distance for NF54 and NF54CSAh.
Figure 3—figure supplement 2
Hi-C intrachromosomal and interchromosomal contact count heatmaps for NF54.
Heatmaps generated from ICED (26619908) and per-million read count normalized intrachromosomal contact count matrices binned at 10 kb resolution for all 14 chromosomes of NF54 with annotated centromeres (gray) and var gene containing bins (green). Whole Genome heatmap depicting the normalized genome-wide interchromosomal contact count matrix for NF54 with intrachromosomal interaction matrices removed.
Figure 3—figure supplement 3
Hi-C intrachromosomal and interchromosomal contact count heatmaps for NF54CSAh.
Heatmaps generated from ICED (26619908) and per-million read count normalized intrachromosomal contact count matrices binned at 10 kb resolution for all 14 chromosomes of NF54CSAh with annotated centromeres (gray) and var gene containing bins (green). Whole Genome heatmap depicting the normalized genome-wide inter-chromosomal contact count matrix for NF54CSAh with intrachromosomal interactions removed.
Figure 3—figure supplement 4
Differential interaction contact count heatmaps show higher interactions between telomeric and internal var gene-containing regions in NF54CSAh.
Heatmaps generated from differential interaction matrices for all 14 chromosomes, identifying regions with an increase (red) or decrease (blue) of intrachromosomal interactions in NF54CSAh over NF54. Centromeric bins (gray) and var gene-containing bins (green) are annotated for each chromosome. Whole Genome differential interchromosomal interaction heatmap with intrachromosomal interactions removed.
Figure 4 with 1 supplement
Regional differences in 5mC occupancy track with global transcriptional activity but are dissociated from var2csa regulation.
(A) Genome-wide distribution of input normalized MeDIP-seq counts for 5mC in NF54 and NF54CSAh reveal a high level of concordance between the analyzed parasite lines. Regional clusters of antigenically variable genes are denoted in green below each plot with an area of chromosome 12 marked by a dashed box and a zoomed-in insert displaying minor differences in 5mC occupancy in the var2csa locus between NF54 and NF54CSAh (var2csa in blue, other var genes in green and rif in black). (B) Levels of 5mC marks binned in 5’ flanks and first exon for the top 20% highest expressed genes and 20% lowest expressed genes in NF54 and NF54CSAh. (C) Distribution of 5mC across exon-intron boundaries for the 5mC enriched var, rif, and other two-exonic genes.
Figure 4—figure supplement 1
The P. falciparum genome contains very little 5hmC and 6mA compared to 5mC and no pattern between gene type or feature.
(A) Genome-wide distribution of input-normalized 5-methylcytosine (5mC, top), 5-hydroxymethylcytosine (5hmC, middle), and 6-methyladenosine (6mA, bottom), with centromeres (gray) and var genes (green) annotated within each track. (B, C) Bar graphs depicting per-million read count normalized 5mC and 6mA per cytosine partitioned into the 500 bp 5’ region, exons, and introns for four antigenically variable gene families (var, rifin, stevor, and pfmc-2tm), as well as the remaining protein coding genes and intergenic regions.
Additional files
-
Supplementary file 1
RNA-seq analysis.
- https://cdn.elifesciences.org/articles/93632/elife-93632-supp1-v1.xlsx
-
Supplementary file 2
ChiP-seq analysis.
- https://cdn.elifesciences.org/articles/93632/elife-93632-supp2-v1.xlsx
-
Supplementary file 3
MeDIP-seq analysis.
- https://cdn.elifesciences.org/articles/93632/elife-93632-supp3-v1.xlsx
-
MDAR checklist
- https://cdn.elifesciences.org/articles/93632/elife-93632-mdarchecklist1-v1.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Epigenetics and chromatin structure regulate var2csa expression and the placental-binding phenotype in Plasmodium falciparum
eLife 13:RP93632.
https://doi.org/10.7554/eLife.93632.3
