Research ArticleMICROBIOLOGY

A single-cell view on alga-virus interactions reveals sequential transcriptional programs and infection states

See allHide authors and affiliations

Science Advances  20 May 2020:
Vol. 6, no. 21, eaba4137
DOI: 10.1126/sciadv.aba4137
  • Fig. 1 Transcriptome profiling of EhV infection in E. huxleyi at single-cell resolution.

    (A) The experimental setup for using scRNA-seq to study EhV infection of E. huxleyi. Individual cells were isolated using FACS during a time course of 0 to 24 hours post infection (hpi). The single-cell transcriptomes were sequenced using the MARS-seq protocol, which allows absolute quantification of viral and host transcripts by tagging each individual cell and each transcript with cellular barcodes and UMIs, respectively. (B) Highly heterogeneous total viral UMI counts among single cells across samples. Each red dot represents an individual cell, and each blue diamond represents the mean of all cells in each column. A threshold of 10 UMIs (fig. S3) highlights the bimodality of the overall viral expression and excludes cells with low- or noise-level viral UMIs, as in mock infection by UV-deactivated EhV (UV; 4 and 6 hpi) and control (Ctrl) samples. (C) The proportion of cells with at least 10 viral UMIs (active viral expression) increased rapidly during early hours of infection and approached 90% at 9.5 hpi. (D) Cliff diagrams represent the relative mRNA levels of host and virus across the infection time course. The cells (columns) are sorted by the relative proportions of host and viral mRNA, showing the sharp decrease in host mRNA relative to viral mRNA (“cliff”) and the within-population heterogeneity through time.

  • Fig. 2 The continuum of infection states across individual cells.

    (A) A 2D projection of 2072 single infected cells (dots) with active viral expression (≥10 viral UMIs) that constitute seven metacells (in different colors). The cells form a continuum of infection states that can be quantified by the relative distance to the initial infection state. (B) Distribution of cells from each sampling time point along the infection continuum, with light gray dots representing cells from the other time points. Numbers are means ± SD of the infection states as defined in (A). CHX, cycloheximide treatment before infection. The infection index, a proxy for infection states, (A) of an individual cell was calculated as the ratio of the distance of this cell from the upper-left origin to the distance between the origin and the lower-right end. A color gradient was painted on the basis of the scale of the infection indices between 0 and 1.

  • Fig. 3 Resolving viral genes into kinetic classes across infection states of single cells.

    (A) Cells (columns) are ordered by the pseudotemporal scale (infection indices) as defined in Fig. 2, with hpi of each culture sample marked for each cell. (B) Viral genes (rows) are divided into kinetic classes based on the hierarchical clustering of their expression profiles across the individual cells, which are ordered by their infection indices the same way as in (A). (C) Presence in the EhV201 virion proteome, as well as categories of predicted functions, is indicated for each gene ordered the same way as in (B). See table S2 for protein identifiers (IDs) and detailed annotations of the viral genes. PA, packing ATPase; MCP, major capsid protein; undet., undetermined.

  • Fig. 4 Predicted promoter elements and genomic organizations of viral kinetic classes.

    (A) Enriched motifs in the promoter regions of viral genes in the newly defined kinetic classes. (B) Distribution of genes in each kinetic class (circle) on the assembled genome of EhV201. The outermost circle indicates the direction of transcription. The heights of genes in each kinetic class are proportional to their log2-transformed expression levels. The innermost circle marks the presence in the proteome of EhV201 virions. Immediate-early genes are mostly clustered in two genomic regions (gray arcs). Numbers indicate sequence lengths in kilobase.

  • Fig. 5 Virus-induced host mRNA shutdown as a function of infection states.

    (A) Levels of host transcripts across treatments and infection states. Single E. huxleyi cells from cultures infected with EhV are divided on the basis of overall viral UMI counts and metacell grouping (Fig. 2A). Protein-encoding genes with at least 100 UMIs were included. For each gene, the average UMI count was calculated for each category, and the sum of the average values of all categories was normalized to 100. Transcripts were then hierarchically clustered on the basis of the normalized values. Organelle-encoded transcripts are indicated on the left. (B and C) Abundance of organellar transcripts along the infection progression. Cells with at least 10 viral UMIs are ordered according to the infection states as determined by the MetaCell analysis (Figs. 2 and 3). Sliding window averages were calculated for each transcript (B) or mean of all transcripts (C) with a window size of 50 cells and steps of 10. Polycistronic transcripts in mitochondria (blue) and chloroplasts (green) (B) are named with their first and last genes, as well as some genes in between, according to the organellar genome annotations (40, 41).

Supplementary Materials

Stay Connected to Science Advances

Navigate This Article