Fig 1.
CVB3 infection of triggers increases in purine and pyrimidine metabolites in HeLa cells.
13C-glucose isotope tracing study in mock- and CVB3-infected HeLa R19 cells (MOI 5, three replicates; one of three representative experiments is shown). Cells were infected, lysed at 2,4,6 or 8 hpi and measured by LC-MS to identify metabolites and quantify the different isotopologues. The different isotopologues are not distinguished in this Figure. A) Heatmap of log2 fold changes of the indicated metabolites between CVB3- and mock-infected cells. Log2 fold changes are calculated based on the mean of three replicates. B) Absolute peak areas of representative nucleotide monophosphates, nucleosides and nucleobases in mock and CVB3-infected cells. C) Absolute peak areas of dihydroorotate and N-carbamoylaspartate in mock and CVB3-infected cells. The p-values were calculated using linear mixed effect models with an interaction of time and treatment and a random effect of replicate. A rank transformation on the data was performed to ensure a normal distribution of the residuals. For hypoxanthine and guanosine, a normal distribution of the residuals could not be assumed and therefore a non-parametric linear mixed effect model with an interaction of time and treatment and a random effect of replicate was performed. Afterwards, a contrast analysis was done to calculate the p-values between specific groups. *p < 0.05, **p < 0.01, ***p < 0.001.
Fig 2.
CVB3 infection increases purine and pyrimidine metabolites through nucleic acid degradation and nucleotide salvage.
13C-glucose isotope tracing study in mock- and CVB3-infected HeLa R19 cells (MOI 5, three replicates; one of three representative experiments is shown). Cells were infected, lysed at 2, 4, 6 or 8 hpi and measured by LC-MS to identify metabolites and quantify the different isotopologues. A) Schematic representation of nucleotide metabolism. Nucleotides can be synthesized de novo, released during degradation of nucleic acids, or recycled (i.e. called the salvage pathway). PRPP, phosphoribosyl pyrophosphate; R5P, ribose-5-phosphate; R1P, ribose-1-phosphate. B) Schematic representation of nucleotide metabolism and nucleotide labeling in 13C-glucose tracing studies. Orange = phosphate group; Blue = pentose sugar; Green = nucleobase. C) Isotope distribution of AMP, ADP, ATP, UMP, UDP and UTP in mock and CVB3-infected cells. D) Isotope distribution of dihydroorotate and N-carbamoylaspartate in mock and CVB3-infected cells. The p-values of C) and D) were calculated using linear mixed effect models with an interaction of time and treatment and a random effect of replicate. Afterwards, a contrast analysis was done to calculate the p-values between specific groups. For this analysis, the fractions of all labels were added together and tested whether the total amount of labeling differed between mock and CVB3 infection. *p < 0.05, **p < 0.01, ***p < 0.001.
Fig 3.
Inhibition of nucleotide salvage constraints CVB3 replication.
A CVB3 luciferase reporter virus carrying a Renilla luciferase (Rluc CVB3) was used to study the replication of CVB3 in the presence or absence of compounds inhibiting the salvage pathways. A, E) Schematic representation of the purine (A) and pyrimidine (E) salvage pathway with the mode action of 6-MP (A) and CPU (E). B, F) Luciferase levels in cells infected with Rluc CVB3 (MOI 0.1) in the presence of different concentrations of 6-MP (B) or CPU (F). Cells were lysed at 2,4 or 6 hpi. Representative data of seven (B) or four (F) independent experiment are depicted (mean ± SD of 3 technical replicates). C, G) Percentage luciferase levels in 6-MP treated and CVB3-infected cells compared to control. Data represent mean ± SEM of seven (C) or four (G) independent experiments. D, H) Metabolomic analysis of hypoxanthine (D) or uridine (H) levels in 6-MP (D) or CPU (H) treated cells. Statistical analysis was done using an ANOVA. For 6-MP, not every concentration was included in all seven independent experiments, but every concentration is measured in at least three independent experiments with three replicates. ***p < 0.001, ****p < 0.0001.
Fig 4.
Integrative analysis of metabolomics and phosphoproteomics confirms the importance of nucleotide metabolism in CVB3 infection.
A, B) DIABLO analysis (supervised analysis of multiple omics datasets) of the metabolomics and the total phosphoproteomics dataset (A) or the phosphoproteomics dataset filtered for metabolic proteins (B) of mock and CVB3-infected HeLa R19 cells. C) Enrichment analysis of proteins extracted from panel A using inBio discoverer (featuring Gene Ontology enrichment analysis). Fold enrichment of the major Gene Ontology (GO) terms is plotted, D) Regulated phosphosites on enzymes directly involved in nucleotide metabolism. Plotted are the log2 LFQ intensities of four replicates (one experiment).The p-values are calculated using ANOVA and subsequent Welch’s T-tests *p < 0.05, **p < 0.01, ***p < 0.001.
Fig 5.
Metabolic changes during EMCV infection resemble those observed in CVB3 infection.
A) Metabolomics study of mock- and CVB3- or EMCV-infected HeLa R19 cells (MOI 5). Cells were lysed at 2, 4, 6 or 8 hpi and subjected to LC-MS for the measurement of metabolites. Heatmap showing log2 fold changes in metabolite levels in CVB3- and EMCV-infected cells compared to mock infection (three replicates; one experiment for each virus). One sample was removed from analysis (one replicate of 2 hpi CVB3), because of a technical defect. B) A Pearson correlation test was performed to determine the level of correlation between metabolic changes induced by CVB3 and EMCV. C) 13C-glucose isotope tracing study in mock- and EMCV-infected HeLa R19 cells (three replicates; one experiment; MOI 5). Cells were infected, lysed at 2, 4, 6 or 8 hpi and measured by LC-MS to identify metabolites and quantify the different isotopologues. Isotope patterns in ADP, ATP, UDP and UTP in mock and EMCV infected HeLa R19 cells are depicted. One sample was removed from analysis (one replicate of 8 hpi EMCV), because of a technical defect. The p-values were calculated using linear mixed effect models with an interaction of time and treatment and a random effect of replicate. A rank transformation on the data was performed to ensure a normal distribution of the residuals. Afterwards, a contrast analysis was done to calculate the p-values between specific groups. For this analysis, the fractions of all labels were added together and tested whether the total amount of labeling differed between mock and CVB3 infection. *p < 0.05, **p < 0.01, ***p < 0.001.