Fig 1.
Detection of cbVGs at the single-cell level.
(A) A virus-specific RT primer (blue) with a 5′ overhang (red) was designed to bind 256 bp from the 3′ end of the viral genome to capture both stVGs and cbVG-546. (B) Diagram showing mRNA and viral RNA capture and reverse transcription (RT) inside 10x Genomics Chromium droplets. (C) Illustration of RT and library preparation for Illumina sequencing. (a) RT products of mRNA, viral genome, and cbVG-546 with the cbVG junction region labeled. A 278 bp of cbVG-546 that includes the junction region is transcribed. (b) Final library prep products with Illumina adaptors (green and purple) added to both ends. The cbVG-546 junction is 94 bp from the TSO. (D) Number of cells passing quality control (QC) and numbers of SeV-positive and cbVG-positive cells in each sample. (E) Total UMI counts, SeV reads, and cbVG reads per cell in each sample. Statistical differences in UMI counts, SeV reads or cbVG reads per cell across clusters were assessed using the Kruskal-Wallis test. Following a significant overall test, pairwise comparisons were performed using Dunn’s post-hoc test. Significance levels are indicated as ****p < 0.0001.
Fig 2.
Unbiased clustering of each sample and identification of IFN-producing clusters in SeV-infected cells.
(A) UMAPs showing unbiased clustering of mock and SeV-infected A549 cells at different infection conditions (5 MOI or 1.5-2 MOI) and time points (6, 12, and 24 hpi). Each color represents a distinct cell cluster in each sample. Clusters in different samples are not comparable based on their color. (B) UMAPs of SeV-, SeV+cbVG-, and SeV+cbVG⁺ cells in each sample. Cells are colored according to viral RNA detection status: SeV- (beige), SeV+cbVG⁻ (pink), and SeV+cbVG⁺ (slate blue). (C) UMAPs showing IFN-producing (blue) clusters across mock and SeV-infected samples under different infection conditions (5 MOI or 1.5-2 MOI) and time points (6, 12, and 24 hpi). (D) Expression levels of the top 10 expressed marker genes of the IFN-producing clusters in different infection conditions. (E-F) Percentages of SeV+ cells (E) and cbVG⁺ cells (F) among total cells in each cluster. (G) Abundance of SeV reads in each cluster represented as the ratio of SeV UMI counts over total UMI counts per cell. Each dot represents a single cell. Statistical differences across all clusters in each sample were analyzed using the Kruskal-Wallis test. Following a significant overall test, pairwise comparisons were performed using Dunn’s post-hoc test, in which each cluster was compared specifically against the IFN-producing cluster. Significance levels are indicated as ns for not significant and asterisks for significant differences (***p < 0.001 and ****p < 0.0001).
Fig 3.
SeV+ cells can be divided into five populations during the first 24 h of infection based on transcriptome signatures.
(A) Integrated analysis of all SeV⁺ cells from each sample. (B) Bubble plot showing the top 10 marker genes for each SeV⁺ cluster (C0-C4) identified relative to other clusters. Dot size represents log2 fold change, and color indicates statistical significance (-log10 adjusted p value). Functional categories of marker genes are labeled below. (C) Heatmap showing the expression of the same top 10 marker genes in each cluster relative to mock-infected cells, with the color scale indicating log2 fold change. White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds. (D) Heatmap showing the proportion of cells expressing each marker gene across clusters and mock (uninfected sample), with the color scale representing expression proportion. (E) Percentage of cells in each population across samples. (F) UMAPs showing the distribution of SeV+ cells under different infection conditions (5 MOI or 1.5-2 MOI) and time points (6, 12, and 24 hpi). Separate panels show SeV+cbVG− (pink) and SeV+cbVG⁺ (slate blue) populations individually.
Fig 4.
Viral RNA levels and cbVG-associated gene expression across infection conditions.
(A-B) Venn diagrams showing the overlap of DEGs between cbVG⁺ and cbVG⁻ cells compared with mock-infected cells at 6hpi (5MOI), 6hpi, 12hpi, and 24hpi. Numbers indicate unique and shared genes between the two groups. (A) Upregulated genes (pct.cbVG⁺ > 0.3, padj < 0.01, and log2FC > 1). (B) Downregulated genes (pct.mock > 0.3, padj < 0.01, and log2FC < -1). (C-F) Heatmaps showing the expression levels (C and E) and the proportion of expressing cells (D and F) for the top 20 genes significantly upregulated (C-D) or downregulated (E-F) in cbVG⁺ or cbVG⁻ cell populations in each sample. The color scale indicates the average log2 fold change (C and E) or the proportion of cells expressing each gene (D and F). White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds. (G) SeV NP, (H) IFNL1, and (I) IFIT1 expression in A549 cells infected with SeV cbVG clean stock or cbVG-high stocks (Cantell and rS2) at indicated MOIs (0.02, 0.2, 2, and 5) and time points (6, 12, and 24 hpi). Expression of mRNA was calculated relative to the housekeeping index with GAPDH and β-actin. IFNL1 and IFIT1 expression represents fold change relative to mock-infected controls. Data represent means ± SD of three biological replicates. Data were analyzed using two-way ANOVA, followed by Šídák’s multiple-comparisons test to compare Clean versus Cantell or rS2. Adjusted P values are shown. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.
Fig 5.
Functional classification and validation of genes upregulated exclusively in cbVG⁺ cells.
(A) Heatmap showing genes specifically upregulated in cbVG⁺ but not cbVG⁻ cells at both 6hpi samples. Genes were grouped into seven functional categories based on their known or predicted biological roles. The red-blue color scale indicates the log2 fold change (log2FC), while the orange-green color scale represents the proportion of cells expressing each gene. White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds. (B-G) Representative genes validated by qPCR. A549 cells were infected with SeV Cantell stocks containing different cbVG compositions: cbVG clean stock, Cantell: high level of cbVG-546, rS2: high levels of cbVGs other than cbVG-546. Each gene mRNA levels were quantified by qPCR and are presented as fold change relative to mock-infected samples. Cells collected at 6, 12, and 24 hpi. Ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test comparing each infection group with the mock control (n = 3). Adjusted P values are shown. Significance P values are indicated as follows: P < 0.05 (*), < 0.01 (**), < 0.001 (***), < 0.0001 (****). (H) Heatmap showing genes specifically upregulated in cbVG⁺ but not cbVG− cells at both 12hpi and 24hpi samples. Genes were grouped into six functional categories based on their known or predicted biological functions. The red-blue color scale indicates the log2 fold change (log2FC), while the orange-green color scale represents the proportion of cells expressing each gene. White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds. (I-J) Representative genes validated by qPCR. A549 cells were infected with SeV Cantell stocks containing different cbVG compositions: cbVG clean stock, Cantell, and rS2. mRNA levels were quantified by qPCR and are presented as fold change relative to mock-infected samples. Cells were collected at 6, 12, and 24 hpi. Data represent mean ± SD from three biological replicates. Ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test comparing each infection group with the mock control (n = 3). Adjusted P values are shown. Significance P values are indicated as follows: P < 0.05 (*), < 0.01 (**), < 0.001 (***), < 0.0001 (****).
Fig 6.
Functional classification and validation of genes upregulated in cbVG⁺ and cbVG⁻ cells.
(A) Heatmap showing genes specifically upregulated in cbVG⁺ and cbVG⁻ cells at both 6hpi samples. Genes were grouped into five functional categories based on their transcriptional induction mechanisms. The red-blue color scale indicates the log2 fold change (log2FC), while the orange-green color scale represents the proportion of cells expressing each gene. White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds. (B-E) Representative genes validated by qPCR. A549 cells were infected with SeV Cantell or rS2. mRNA levels were quantified by qPCR and are presented as fold change relative to mock-infected samples. Cells were collected at 6, 12, and 24 hpi. Data represent mean ± SD from three biological replicates. Ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test comparing each infection group with the mock control (n = 3). Adjusted P values are shown. Significance P values are indicated as follows: P < 0.05 (*), < 0.01 (**), < 0.001 (***), < 0.0001 (****). (F) Heatmap showing genes specifically upregulated in cbVG⁺ and cbVG− cells at both 12 hpi and 24 hpi, but not at 6 hpi. Genes were grouped into seven functional categories based on their known or predicted biological functions. The red-blue color scale indicates the log2 fold change (log2FC), while the orange-green color scale represents the proportion of cells expressing each gene. White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds. (G-I) Representative genes validated by qPCR. A549 cells were infected with SeV Cantell or rS2. mRNA levels were quantified by qPCR and are presented as fold change relative to mock-infected samples. Cells were collected at 6, 12, and 24 hpi. Data represent mean ± SD from three biological replicates. Ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test comparing each infection group with the mock control (n = 3). Adjusted P values are shown. Significance P values are indicated as follows: P < 0.05 (*), < 0.01 (**), < 0.0001 (****).
Fig 7.
Functional classification and validation of genes downregulated at late infection.
(A) Heatmap showing genes specifically downregulated in cbVG⁺ but not cbVG⁻ cells at both 12hpi and 24hpi samples. Genes were grouped into seven functional categories based on their known or predicted biological functions. The red-blue color scale indicates the log2 fold change (log2FC), while the orange-green color scale represents the proportion of cells expressing each gene. (B) Heatmap showing genes specifically downregulated in cbVG⁺ and cbVG⁻ cells at both 12hpi and 24hpi samples but not common at 6hpi. Genes were grouped into seven functional categories based on their known or predicted biological functions. The red-blue color scale indicates the log2 fold change (log2FC), while the orange-green color scale represents the proportion of cells expressing each gene. White tiles with diagonal lines indicate features that did not meet the predefined adjusted P-value and proportion thresholds.
Fig 8.
Model summarizing the role of cbVGs in driving antiviral responses during infection.
At initial infection, there are two types of infected cells: cbVG⁺ cells (a) and cbVG⁻ cells (b). cbVG⁺ cells act as early initiators of innate immune signaling, inducing strong interferon (IFN) and interferon-stimulated gene (ISG) responses (c) and transmitting antiviral signals to neighboring cbVG⁻ cells through paracrine signaling (d). As infection progresses, IFN and ISG expression becomes broadly activated across the infected cell population (e, f). At later stages, cbVG⁺ cells additionally initiate unique transcriptional programs (e), while global transcriptional repression is observed in all infected cells and is stronger in cbVG⁺ cells (e) than in cbVG⁻ cells (f).