Rewiring Host Lipid Metabolism by Large Viruses Determines the Fate of Emiliania huxleyi, a Bloom-Forming Alga in the Ocean

Abstract

Marine viruses are major ecological and evolutionary drivers of microbial food webs regulating the fate of carbon in the ocean. We combined transcriptomic and metabolomic analyses to explore the cellular pathways mediating the interaction between the bloom-forming coccolithophore Emiliania huxleyi and its specific coccolithoviruses (E. huxleyi virus [EhV]). We show that EhV induces profound transcriptome remodeling targeted toward fatty acid synthesis to support viral assembly. A metabolic shift toward production of viral-derived sphingolipids was detected during infection and coincided with downregulation of host de novo sphingolipid genes and induction of the viral-encoded homologous pathway. The depletion of host-specific sterols during lytic infection and their detection in purified virions revealed their novel role in viral life cycle. We identify an essential function of the mevalonate-isoprenoid branch of sterol biosynthesis during infection and propose its downregulation as an antiviral mechanism. We demonstrate how viral replication depends on the hijacking of host lipid metabolism during the chemical "arms race" in the ocean.

Figure 1.

Infection Dynamics and Ultrastructure Analyses of E. huxleyi and Its Virus EhV. (A) Cultures of E. huxleyi were monitored during infection by a lytic (EhV201) or nonlytic (EhV163) virus and compared with noninfected control cells. Images were taken at 72 hpi. (B) and (C) Host cell abundance (B) and host cell death (C) as assessed by Sytox Green fluorescence by flow cytometry (mean ± sd, n = 3, at least 6000 cells were measured at each time point). (D) to (F) Transmission electron micrographs of control cells (D) and cells infected by the lytic (E) and nonlytic (F) virus at 48 hpi. Viruses are only visible in lytic infected cells. (G) Intracellular viral abundance determined by quantitative PCR of the viral DNA within the cellular fraction, probing the major capsid protein (MCP) gene (mean ± sd, n = 3). (H) Abundance of extracellular viruses measured by flow cytometry (mean ± sd, n = 3). (I) Relative abundance of RNA reads mapped to host or virus genomes at 1 and 24 hpi. [See online article for color version of this figure.]

Figure 2.

Global Gene Expression Profiles of E. huxleyi and EhV during Infection. (A) Global gene expression profiles of host genes during infection by the lytic and nonlytic viruses. K-means clustering of genes with altered expression levels during the course of infection is presented. Red, high expression level; blue, low expression level. (B) Significantly enriched GO terms and KEGG pathways (hypergeometric test, P < 0.05) related to host gene clusters as displayed in (A). Colors refer to clusters as indicated in (A). For a full list of enriched biological pathways in each cluster, see Supplemental Data Set 2. (C) Global gene expression profiles of viral genes during lytic viral infection. K-means clustering of genes induced at the early and late phases of infection is shown. Red, high expression level; blue, low expression level. (D) Significantly enriched biological functions (hypergeometric test, P < 0.05) related to viral gene clusters as displayed in (C). Colors refer to clusters as indicated in (C).

Figure 3.

Viral-Induced Remodeling of Host Metabolism during Infection. (A) to (D) CAP of metabolic profiles derived from control cells compared with cells infected by the lytic or nonlytic virus 4, 24, 32, and 48 hpi (n = 3). (E) Correlation of transcriptome expression patterns (1 and 24 hpi) and metabolite abundance (4 and 24 hpi) based on weighted correlation network analysis. Symbols indicate the affiliation of metabolites to metabolic classes. Blue, low correlation between transcriptome and metabolome; red, high correlation between transcriptome and metabolome. The numerical correlation coefficient values are given in Supplemental Data Set 7.

Figure 4.

Viral-Induced Rewiring of Host Metabolism Leads to an Increase in Fatty Acid Biosynthesis. (A) Integrated metabolic map of significantly enriched biological pathways in host gene cluster 3. Genes associated with cluster 3 are marked in dark red. Inserted heat maps present normalized intensities (mean, n = 3) of specific metabolites in noninfected E. huxleyi cells and cells infected by the lytic or nonlytic virus at 4, 24, 32, and 48 hpi. Dark-red bar indicates inhibition of fatty acid biosynthesis by C75. Fold-change values of changes in metabolite concentration relative to the control are presented as Supplemental Data Set 4. Metabolites marked by an asterisk were detected in nonphosphorylated form and a minus sign indicates the absence of a metabolite at a specific time point. Dashed lines refer to connections via a known metabolic pathway. Green, low metabolite concentration; yellow, high metabolite concentration. For abbreviations of enzyme names, see Supplemental Data Set 12. (B) Gene expression pattern of host cluster 3. The average of the expression normalized values is presented as thick black line. (C) The effect of inhibition of fatty acid biosynthesis by various concentrations of C75 on extracellular virus abundance at 72 hpi relative to viruses released from infected cells without the addition of the inhibitor (percentage of control) (mean ± sd, n = 3), and the percentage of viable cells at 24 hpi as measured by flow cytometry (mean ± 3·sd, n = 3). FA, fatty acid; P, phosphate; RPM, reads per million.

Figure 5.

Sphingolipid Metabolic Network Is Modulated toward Production of Viral-Derived Sphingolipids during Lytic Infection. (A) RNA-seq–based gene expression profiles of host (gray) and viral (yellow) genes encoding enzymes in sphingolipid metabolism during infection with the lytic or nonlytic virus at 1 and 24 hpi. Host de novo synthesis (top) and catabolic reactions (bottom) are indicated. Blue, low relative expression; red, high relative expression. Reads per million (RPM) values for the associated genes are presented as Supplemental Data Set 8. (B) Relative abundance of vGSL and two host glycosphingolipids (hGSL and sGSL) are shown at 1, 4, 24, 32, and 48 hpi. vGSL relative abundance was normalized to its level in the nonlytic infection at 1 hpi and was not detected in control cells. Host sphingolipid abundance was normalized to noninfected cells. For abbreviations of enzyme names, see Supplemental Data Set 12. 3KSa, 3-ketosphinganine; DH, dihydro; GSL, glycosphingolipid; hGSL, host glycosphingolipids; P, phosphate; sGSL, sialic acid glycosphingolipids. Data are presented as mean ± se; n = 3.

Figure 6.

Terpenoid and Sterol Biosynthesis via the Mevalonate Pathway Is Required for Viral Replication. (A) Enzymatic and metabolic patterns of terpenoid and sterol biosynthesis as comprised in host gene cluster 1. Genes associated with cluster 1 are marked in dark red. Inserted heat maps present normalized intensities (mean, n = 3) of specific metabolites in E. huxleyi control cells and cells infected by the lytic or nonlytic virus at 4, 24, 32, and 48 hpi. Dark-red bars indicate inhibition of the MVA pathway by cerivastatin and of the MEP pathway by fosmidomycin. Sterols 1 to 3 are isomers of ergostatriene, sterol 4 is a putative ergostadienone, and sterols 5 to 7 remain unidentified. Dashed lines refer to connections via a known metabolic pathway. A minus sign indicates the absence of a metabolite at a specific time point. Green, low metabolite concentration; yellow, high metabolite concentration. Fold-change values of changes in metabolite concentration relative to the control are presented in Supplemental Data Set 4. For abbreviations of enzyme names, see Supplemental Data Set 12. (B) Gene expression pattern of host cluster 1. The average of the expression normalized values is presented as a thick black line. (C) The effect of inhibition of the MVA pathway by cerivastatin on extracellular virus abundance at 72 hpi relative to viruses released from infected cells without the addition of the inhibitor (percentage of control) (mean ± sd, n = 3) and the percentage of viable cells at 24 hpi as measured by flow cytometry (mean ± 3·sd, n = 3). (D) Mass spectrum of epibrassicasterol, the main sterol in E. huxleyi cells. (E) Mass spectrum of epibrassicasterol isolated from concentrated, purified EhV201 virion membranes. The insert displays the structure of the compound. P, phosphate; PP, pyrophosphate; RPM, reads per million.

Figure 7.

Rewiring of Host Metabolism during Lytic Viral Infection. Viral replication in E. huxleyi depends on the host metabolic machinery to provide building blocks for viral progeny formation (e.g., sphingolipids, fatty acids, and sterols in the lipid membrane). During the early stage of viral infection, upregulated glycolysis shuffles energy to fatty acid biosynthesis, bypassing the TCA. Concomitantly, overexpression of viral encoded sphingolipid genes induces de novo sphingolipid biosynthesis. During the onset of lytic infection demands for nucleotide biosynthesis are met via upregulation of the pentose phosphate pathway (PPP). These processes result in assembly of progeny viruses and finally lead to host cell death. (Inset: graph of virion based on Hurst [2011]). Downregulation of fatty acid, terpenoid, and sterol biosynthesis in addition to induction of host-derived sphingolipid catabolic reactions in the late phase of viral infection may facilitate host resistance responses.