Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Glutamine supports human cytomegalovirus induced lipidome remodeling through reductive carboxylation

  • Rebekah L. Mokry,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing

    Affiliations BIO5 Institute, University of Arizona, Tucson, Arizona, United States of America, Department of Immunobiology, University of Arizona, Tucson, Arizona, United States of America

    ⨯
  • Caleb B. de Lacy,

    Roles Investigation, Writing – review & editing, Data curation

    Affiliations BIO5 Institute, University of Arizona, Tucson, Arizona, United States of America, Department of Immunobiology, University of Arizona, Tucson, Arizona, United States of America

    ⨯
  • John G. Purdy

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Visualization, Writing – original draft, Writing – review & editing

    jgpurdy@arizona.edu

    Affiliations BIO5 Institute, University of Arizona, Tucson, Arizona, United States of America, Department of Immunobiology, University of Arizona, Tucson, Arizona, United States of America, Cancer Biology Interdisciplinary Program, University of Arizona, Tucson, Arizona, United States of America

    ⨯

Abstract

Lipidome remodeling during human cytomegalovirus (HCMV) replication is a complex process that requires induction of lipogenic proteins and altered metabolite flow to support synthesis of fatty acids and lipids. HCMV infection increases the utilization of glucose and acetate to provide enough carbons to support increased demand for lipogenesis during virus replication, but other carbon contributors have not been studied. Here, we identify glutamine as a carbon source for lipogenesis during HCMV infection. Metabolic tracing with 13C-labeled glutamine revealed carbons from glutamine are enriched in phospholipids and neutral lipids during infection, including phosphatidylcholine, phosphatidylethanolamine, diacylglycerol, and triacylglycerol. Additional metabolic tracing demonstrates that HCMV infection promotes glutamine flow to fatty acid synthesis primarily through reductive carboxylation, i.e., conversion of glutamine to citrate through isocitrate. Through the use of two different 13C-labeled forms of glutamine, we found that ~70% of the carbons from glutamine are delivered to fatty acid synthesis through reductive carboxylation. Our current understanding of metabolite utilization during virus replication is based on cell culture models in which glucose is in excess, suggesting that HCMV may encode mechanisms to promote glutamine flow to lipids if glucose levels in vivo are insufficient. To determine if concentrations of glucose and glutamine change their contributions to fatty acid synthesis, we investigated lipogenesis when glucose and glutamine are at physiological levels (5 mM and 0.55 mM, respectively). We determined that physiological levels of glucose and glutamine are sufficient to support the increased demand for fatty acid synthesis caused by HCMV infection, despite a reduction in virus production. Using metabolic tracing with 13C-labeled forms of glucose or glutamine, we determined that both carbon sources contribute to fatty acid synthesis when present at physiological levels. Overall, our results identify viral activation of reductive carboxylation that increases glutamine flow to lipogenesis during infection. This work provides additional insight into metabolic reprogramming that supports HCMV-induced lipidome remodeling.

Author summary

Many viruses hijack cellular metabolic processing to obtain the components needed for replication. Human cytomegalovirus (HCMV) uses several mechanisms to reprogram lipid metabolism and remodel the lipidome of infected cells. HCMV promotes synthesis of very long-chain fatty acids that are found in phospholipids and triacylglycerol. Glucose and acetate contribute carbon to fatty acid synthesis and elongation following HCMV infection. In this work, we demonstrate that glutamine is an additional carbon source for fatty acid and lipid synthesis. Phospholipids and neutral lipids are enriched with carbons from glutamine during HCMV infection. Mechanistically, HCMV induces reductive carboxylation to increase glutamine flow to fatty acid synthesis and increased metabolite availability supports additional carbon flow to fatty acids. Overall, this study provides additional insight into virus-induced metabolic remodeling that supplies the molecular building blocks for virus replication.

Citation: Mokry RL, de Lacy CB, Purdy JG (2026) Glutamine supports human cytomegalovirus induced lipidome remodeling through reductive carboxylation. PLoS Pathog 22(8): e1014496. https://doi.org/10.1371/journal.ppat.1014496

Editor: Eain A. Murphy, State University of New York Upstate Medical University, UNITED STATES OF AMERICA

Received: March 17, 2026; Accepted: July 22, 2026; Published: August 12, 2026

Copyright: © 2026 Mokry et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The lipidomic mass spectrometry data files have been deposited in Metabolomics Workbench [63], where it has been assigned Project ID PR003210. The data can be accessed directly via its Project ID DOI: http://dx.doi.org/10.21228/M8RS11. Individual studies are accessible using Study ID: ST004999, ST005000, ST005002, ST005006, ST005007, and ST005008. All other data referred to in this study are presented in the figures or accompanying supporting information files.

Funding: This work was supported by the National Institute of Allergy and Infectious Diseases (NIAID) of the National Institutes of Health (NIH) under Award Number R01AI162671 (J.G.P.), R01AI155539 (J.G.P.), and F32AI178919 (R.L.M.). Additional support was provided by BIO5 Institute Postdoctoral Fellowship Program 004240 award (R.L.M.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Viruses depend on the metabolic environment of their host cell to successfully replicate. Many viruses manipulate metabolic pathways to support the synthesis of molecular building blocks for replication, including DNA, protein, and lipids. Nutrients in the environment play specific roles in viral reprogramming of metabolism during infection. Two major nutrients, glucose and glutamine, are crucial for replication of several viruses, including the beta-herpesvirus human cytomegalovirus (HCMV) [1]. HCMV is a prevalent virus that is the leading cause of viral-mediated congenital disorders, including microcephaly, developmental and cognitive disabilities, and hearing loss [2,3]. In immunocompromised individuals, such as those who have undergone solid-organ or stem cell transplant or those with secondary immune suppressive infections, HCMV can cause severe, life-threatening diseases [4]. HCMV has evolved an adept system to manipulate and reprogram cellular metabolism, providing a model to understand the mechanisms by which viruses hijack and rewire nutrient flow [1,5].

HCMV requires lipid synthesis for efficient replication and remodels the lipidome during infection [6–11]. Lipids support viral assembly and infectivity of the viral particle [11–13]. HCMV remodels the lipidome in part by promoting fatty acid (FA) synthesis and elongation to increase very long chain FA (VLCFA) with 26 or more carbons [10,11]. VLCFA are used to generate phospholipids and neutral lipids with very long-chain fatty acyl tails [7–9]. To promote lipid synthesis, HCMV induces lipogenic proteins involved in synthesis and elongation of FA and drives glucose carbons to citrate and acetyl-CoA to support FA synthesis [6,7,10,11,14,15]. Since glucose carbons are fueling FA production, glutamine uptake and glutaminolysis, i.e., catabolism of glutamine to support energy production and synthesis of other macromolecules, are induced to support the TCA cycle [16].

Glutamine has been largely overlooked as a potential carbon contributor for FA and lipid synthesis since glucose is the primary source of carbon for FA synthesis [6,10,11]. It is well established that glutamine maintains the TCA cycle during HCMV infection [6,16], but other roles of glutamine during infection are not well defined [17]. In this study, we identify glutamine as a carbon source for lipogenesis during HCMV replication even in conditions with supraphysiological levels of glucose. Using metabolic tracing coupled with liquid chromatography high-resolution tandem mass spectrometry (LC-MS/MS), we found that glutamine directly contributes to lipid synthesis. We determined that glutamine supports lipogenesis of most lipid classes included in this study. Additionally, glutamine supports synthesis of lipids across lipid tail lengths (number of carbons) and degrees of desaturations (number of double bonds). Moreover, we found that HCMV infection induces reductive carboxylation, i.e., “reverse” metabolism of glutamine to citrate, during which glutamine is metabolized to α-ketoglutarate, converted to isocitrate, and then to citrate, which supplies acetyl-CoA for FA synthesis. Although reductive carboxylation is the primary mechanism of glutamine support of FA synthesis, we identified additional glutamine carbon flow to FA production that is likely supported by malic enzyme pathway. Finally, to understand the potential effects of glucose and glutamine availability in the nutrient environment on HCMV lipogenesis, we reduced the high level of glucose and glutamine found in standard DMEM to physiological concentrations. When glucose and glutamine are provided at the levels found in plasma of healthy people, HCMV was able to promote FA synthesis and elongation but had a reduced amount of virus production. Overall, our data reveal that reductive carboxylation supports lipidome remodeling during HCMV replication, including in the presence of glucose and glutamine at physiological levels. The findings expand our understanding of the role of glutamine during viral infection and demonstrate another example of HCMV-induced metabolic reprogramming that supports successful virus replication.

Results

Glutamine supports lipid and fatty acid synthesis during HCMV replication

Glutamine is required for HCMV replication [16,17]. HCMV increases glutaminolysis to support TCA cycle continuation [6,16]. Since glutamine can feed lipid synthesis outside the context of virus infection [18–21], we hypothesized that glutamine could support lipid synthesis during HCMV infection. To determine if glutamine supports lipogenesis, we used an isotopic labeling approach with uniformly labeled 13C glutamine (U-13C-Gln). Experiments were performed on fully confluent human foreskin fibroblast (HFFs) cultured in serum-free DMEM to eliminate any contribution of unlabeled glutamine in the bovine serum. These conditions are widely used in studies of HCMV metabolism [6–8,10,11,14,22–29]. We infected or mock-infected HFFs at MOI 3. At 1 hpi, inoculum was removed, cells were washed with PBS, and DMEM containing 4 mM U-13C-Gln was provided to cells. At 48 and 72 hpi, lipids were extracted and identified using LC-MS/MS. These times were selected because HCMV induces lipidome remodeling by 48 hpi with more pronounced alterations by 72 hpi [7–9]. We first examined if HCMV infection promotes glutamine flow to synthesis of the major lipid classes. To do so, the percentage of lipids containing at least one 13C atom were quantified after correcting for natural isotopic abundance. Each class of lipids has several dozen lipid species. We calculated the percent of labeled lipids by averaging the amount of label in the observed species within a lipid class. We also examined lysophosphatidylcholine (LPC) and lysophosphatidylethanolamine (LPE) that contain one fatty acyl tail since interconversion between one-tailed LPC or LPE with two-tailed phosphatidylcholine (PC) or phosphatidylethanolamine (PE) is active in fibroblasts [9,30]. By 48 hpi, the percentage of PC, PE, phosphatidylinositol (PI), diacylglycerol (DG), and triacylglycerol (TG) labeled carbons were increased in HCMV-infected cells compared to mock (Fig 1A), and by 72 hpi, these lipids had more pronounced labeling (Fig 1B). At 72 hpi, glutamine carbon labeling of phosphatidylserine (PS), LPC, and LPE were also significantly increased during infection (Fig 1B). Most lipids from HCMV-infected cells had approximately two times more labeling from glutamine compared to mock by 72 hpi. For example, we observed 8% labeling in mock compared to 23% in HCMV-infected cells for PC, 6% compared to 16% for PE, 12% compared to 19% for PI, and 12% compared to 17% for PS. Labeled DG and TGs were a respective 11 and 10% in mock compared to 25 and 23% in infected cells. Labeled LPC and LPE were subtly, but significantly, increased during HCMV infection. LPC had 7% labeling in mock compared to 10% during infection and LPE had 1% labeled in mock compared to 2% during infection. Phosphatidylglycerol (PG) was the only lipid class that we examined where infection did not alter labeling from glutamine, suggesting that HCMV does not promote PG synthesis using glutamine carbons (Fig 1A-1B).

thumbnail
Download:
Fig 1. Glutamine supports lipid synthesis during HCMV replication. Human foreskin fibroblasts (HFFs) were infected with TB40/E-GFP or mock-infected.

At 1 hpi, virus inoculum was removed and replaced with media containing uniformly labeled 13C glutamine (U-13C-Gln). At 48 and 72 hpi, lipids were extracted and measured using LC-MS/MS. (A-B) Percent of 13C-labeled lipids were calculated for each lipid class by averaging the percent label for each lipid species at 48 (A) and 72 hpi (B). Percent of 13C-labeled lipids by tail desaturations for PC (C), PE (E), DG (E), and TG (F) at 72 hpi. Each dot represents a lipid from one replicate. Error bars represent standard deviation (SD). Unpaired t-test were used to determine significance. P < 0.05, *; P < 0.01, **; P < 0.001, ***. n = 6.

https://doi.org/10.1371/journal.ppat.1014496.g001

Glutamine carbon contribution to lipid synthesis could be directed to fatty acyl tails of the lipids. To determine if glutamine supports synthesis of different tail types during infection, we quantified the percent of labeled lipids containing saturated (no double bond), monounsaturated (one double bond), diunsaturated (two double bonds), or polyunsaturated (three or more double bonds) fatty acyl tails at 72 hpi (Fig 1C-1F). For those that have two fatty acyl tails, saturated lipids contain two saturated tails, while monounsaturated lipids contain one saturated and one monounsaturated tail. Diunsaturated lipids contain one saturated tail and one diunsaturated tail or two monounsaturated tails, and polyunsaturated lipids contain one saturated or monounsaturated tail with a polyunsaturated tail or two polyunsaturated tails or two diunsaturated tails. We focused on lipid classes with the highest levels of labeling from glutamine in HCMV-infected cells compared to mock, i.e., PC, PE, and DG. For labeled PC and PE, HCMV significantly increased the percent 13C in all tail types, with greatest enrichment for saturated and monounsaturated (Figs 1C-1D and S1A and S1B). For labeled PC in HCMV-infected cells, we observed that most diunsaturated tail types had a 2-fold or greater level of labeling, while a small cluster had a lower level of labeling (Fig 1C). This distribution was the result of an enrichment in labeled diunsaturated PC species as the carbon chain length increased, e.g., PC(48:2) had a greater level of labeling than PC(34:2) (S1A Fig). Although less robust than diunsaturated PC, diunsaturated PE labeling was also enriched in HCMV-infected cells, including a higher level of labeling in the longest PE species we measured, PE(44:2) (Figs 1D and S1B). Labeled DG tails were also increased during HCMV infection, with the greatest enrichment for monounsaturated tails (Figs 1E and S1C). Saturated DG were not identified in this experiment. We also quantified 13C-labeling of TG tail types, which contain three fatty acyl tails. Monounsaturated TGs contain two saturated and one monounsaturated tail, diunsaturated TGs could contain one saturated and two monounsaturated tails or two saturated and one diunsaturated tail. We observed the highest labeling for saturated TG tails during infection; mono- and diunsaturated are also increased compared to mock-infected (Figs 1F and S1D). Polyunsaturated TGs could contain several different tail combinations, such as three monounsaturated tails or two unsaturated and one polyunsaturated tail. We observed wide variability in the percentage of labeling for polyunsaturated TGs, which may be due to the broad range of tail combinations that comprise TG lipids (Figs 1F and S1D). Overall, these data demonstrate that HCMV promotes glutamine use for lipid synthesis in infected cells.

The results of labeling lipids using U-13C-Gln suggest that glutamine is utilized for FA synthesis during infection (Fig 1). To directly test if glutamine is a carbon source for FA synthesis, we examined FA after feeding cells U-13C-Gln. In these experiments, esterified FA were liberated from lipids via saponification before measurement by LC-MS. This method enables us to measure all esterified tails independent of lipid class to gain a better understanding of how glutamine contributes to the building blocks used in lipid synthesis. We infected or mock-infected HFFs and cultured in media containing U-13C-Gln at 1 hpi. FA were extracted at 48 and 72 hpi and labeled carbons were measured using LC-MS. FA with at least one 13C were quantified after correcting for natural isotopic abundance. Since glutamine primarily contributes to FA synthesis through its conversion to citrate which then yields acetyl-CoA, the U-13C-Gln labeling strategy is expected to yield two 13C atoms per two-carbon unit used for FA synthesis and elongation (Fig 2A). We focused our analysis on saturated (SFA) and monounsaturated FA (MUFA) since HCMV strongly induces their production [10,11], and our lipid labeling indicates that those with saturated and monounsaturated fatty acyl tails have the highest percentage of label (Fig 1C-1F). In mock-infected cells, low levels of labeled SFA and MUFA were observed at both 48 and 72 hpi (Figs 2B-2C and S2). In HCMV-infected cells, labeled FA were increased by 48 hpi compared to mock (S2 Fig), and labeling was further enhanced by 72 hpi (Fig 2B-2C). By 72 hpi, labeled FA were significantly increased in HCMV-infected cells for all identified SFA and very long chain MUFA beginning with C24:1.

thumbnail
Download:
Fig 2. Glutamine supports fatty acid synthesis during HCMV replication.

(A) Schematic of predicted fatty acid (FA) labeling from U-13C-Gln. Blue circles indicate 13C-labeled carbons, starting at the bottom with U-13C-Gln. One unlabeled carbon (black circle) is incorporated in the conversion of α-ketoglutarate to isocitrate. (B-D) HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing U-13C-Gln. At 72 hpi, lipids were extracted and FA were saponified. Labeled FA were measured using LC-MS. Percent 13C-labeled SFA (B) and MUFA (C) were quantified. (D) Percent 13C-labeled C26:0. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.001, ***. n = 6.

https://doi.org/10.1371/journal.ppat.1014496.g002

Our analyses to this point focused on lipids or FA that contain one or more 13C atoms from glutamine. Next, we quantified the number of labeled carbons incorporated into FA to better understand the extent of glutamine utilization in lipogenesis. We focused on C26:0 since it showed a high level of labeling from glutamine (Fig 2B) and has been previously examined for glucose and acetate labeling [7,11]. By 72 hpi, C26:0 in HCMV-infected cells had up to 8-labeled atoms derived from glutamine, whereas mock-infected cells contained only up to 4 labeled atoms (Fig 2D). Of note, we observed a small amount of 1-labeled and 3-labeled C26:0, suggesting that labeled carbons may be entering FA synthesis through additional steps. Overall, our data demonstrate that glutamine contributes to FA synthesis and elongation during HCMV replication.

Reductive carboxylation supports FA synthesis during HCMV replication

Glutamine contributes to FA synthesis via reductive carboxylation, also called reductive glutamine metabolism, which metabolizes glutamine to citrate through the intermediate α-ketoglutarate in a series of enzymatic reactions (Fig 2A) [18–21]. This pathway requires the activity of isocitrate dehydrogenase 1 (IDH1) or isocitrate dehydrogenase 2 (IDH2) that catalyze the conversion of α-ketoglutarate to isocitrate in a NADPH-dependent manner (Fig 3A). Both enzymes can also support the reverse reaction. This α-ketoglutarate to isocitrate reaction can occur in the cytosol via IDH1 or mitochondria via IDH2 [18–20,31,32]. We investigated IDH1 and IDH2 levels to determine if these enzymes are increased to support glutamine flow to citrate during HCMV infection. Neither IDH1 nor IDH2 were increased during HCMV replication despite increased glutamine utilization for FA and lipid synthesis (Fig 3B-3D). These data suggest increased enzyme levels are not required for the induction of glutamine flow to lipids in HCMV-infected cells.

thumbnail
Download:
Fig 3. HCMV induces IDH3A but not IDH1 or IDH2.

(A) Schematic of predicted FA labeling from U-13C-Gln involving isocitrate dehydrogenase (IDH). Blue circles indicate 13C-labeled carbons, and black circles indicate unlabeled carbons. (B-D) HFFs were infected with TB40/E-GFP or mock-infected. Whole-cell lysates were collected at the indicated times points and analyzed by western blot. A representative western blot from three biological replicates is shown in B. Levels of IDH1 (C), IDH2 (D), or IDH3A (E) were normalized to tubulin and quantified relative to mock at 4 hpi. α-Ketoglutarate (αKG). Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.01, **; P < 0.001, ***. n = 3.

https://doi.org/10.1371/journal.ppat.1014496.g003

Isocitrate dehydrogenase 3 (IDH3) supports oxidative turning of the TCA cycle and isocitrate conversion to α-ketoglutarate in an NAD+ dependent manner. We investigated IDH3 levels by blotting for the subunit IDH3A. IDH3A is increased by 72 hpi during HCMV replication (Fig 3B and 3E), suggesting IDH3 is induced to support TCA cycle maintenance [6,27]. These data indicate that glutamine contribution to lipid synthesis does not require increased IDH1 and IDH2 levels, while increased IDH3 supports HCMV-induced TCA cycle activity [6,27].

Next, we determined if reductive carboxylation is the primary pathway by which glutamine contributes to FA synthesis during HCMV replication. We infected or mock-infected HFFs and cultured them in media containing glutamine with 13C at position 5 (5-13C-Gln). Metabolism of 5-13C-Gln through reductive carboxylation retains the labeled carbon on one atom of acetyl-CoA, thereby incorporating one 13C atom per two-carbon unit added to the FA chain (Fig 4A). Metabolism of 5-13C-Gln through oxidative TCA cycle yields 1-labeled α-ketoglutarate that upon oxidative turning of the TCA cycle will result in unlabeled acetyl-CoA, since the labeled carbon is retained in OAA following the conversion of citrate to acetyl-CoA and OAA. If the acetyl-CoA produced through this pathway were to be used for FA synthesis, it would result in unlabeled FA (Fig 4B). In mock-infected cells, low levels of labeling were observed for SFA and MUFA at 48 and 72 hpi (Figs 4C-4D and S3). At 48 hpi, HCMV-infected cells showed increased labeling for C14:0, C16:0, and VLCFA compared to mock (S3 Fig). By 72 hpi, labeled FA were significantly increased for all identified SFA and very long chain MUFA beginning with C26:1 (Fig 4C-4D). For the labeled carbons of C26:0, HCMV-infected cells showed increased levels for 1–3-labeled atoms compared to mock-infected cells at 72 hpi (Fig 4E). These data support our conclusion that HCMV infection promotes glutamine use for FA synthesis and demonstrate that glutamine contributes to this process via reductive carboxylation.

thumbnail
Download:
Fig 4. Glutamine is metabolized via reductive carboxylation during HCMV replication.

(A-B) Schematic of predicted FA labeling from positionally labeled 5-13C-glutamine (5-13C-Gln) for reductive carboxylation (A) and oxidative flow (B). Glucose (Glc); oxaloacetate (OAA); α-ketoglutarate (αKG). Blue circles indicate 13C-labeled carbons, black circles indicate unlabeled carbons from glutamine, and grey circles indicate unlabeled carbons from glucose. (C-E) HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing 5-13C-Gln. At 72 hpi, lipids were extracted and FA were saponified. Labeled FA were measured using LC-MS. Percent 13C-labeled SFA (C) and MUFA (D) were quantified. (E) Percent 13C-labeled C26:0. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.001, ***. n = 5-6.

https://doi.org/10.1371/journal.ppat.1014496.g004

Additional metabolic activity supports glutamine feeding FA synthesis during HCMV replication

Our glutamine labeling strategies provide information on the contribution of glutamine from reductive carboxylation (5-13C-Gln) versus the total contribution of glutamine (U-13C-Gln) to FA synthesis [18]. Similar percentages of labeled FA would indicate that glutamine contributes to FA synthesis solely through reductive carboxylation. However, if the percentage of labeled FA is higher for U-13C-Gln labeling than 5-13C-Gln, it indicates glutamine contributes through additional pathways. To determine if additional pathways contribute to glutamine use for FA synthesis in infected cells, we compared percentages of labeled SFA and MUFA from cells cultured in media containing U-13C-Gln to those in 5-13C-Gln. FA in mock-infected cells showed similar percentage of labeled SFA and MUFA between the two forms of labeled glutamine, except for C28:1 at 72 hpi, which showed a subtle increase in 5-13C-Gln fed cells (Fig 5A-5B). These results indicate that the low level of FA labeling in mock-infected cells occurs through reductive carboxylation with minimal contribution from an additional pathway. In contrast, HCMV-infected cells showed significantly higher labeling from U-13C-Gln for SFA C14:0, C24:0, C26:0, and C28:0 (Fig 5C). Similar results were observed for MUFA with increased labeling of C26:1, C28:1, C30:1, and C32:1 (Fig 5D). These data indicate that reductive carboxylation accounts for the majority of FA labeling in HCMV-infected cells, but an additional pathway also contributes to glutamine flow to FA synthesis. To determine the percent of label from reductive carboxylation versus other pathways, we calculated the percent of FA label from 5-13C-Gln relative to U-13C-Gln. Labeling from U-13C-Gln was set to 100% and considered total FA labeling. For significantly altered SFA and MUFA between the two labeled glutamine forms, we found ~70% of FA label occurs through reductive carboxylation (Fig 5E-5F). These data demonstrate that glutamine supports FA synthesis primarily through reductive carboxylation during infection. However, ~ 30% of FA label from glutamine is shuttled from additional metabolic steps during HCMV replication.

thumbnail
Download:
Fig 5. Reductive carboxylation is the primary pathway through which glutamine is metabolized to support FA synthesis.

Percent 13C-labeled SFA (A & C) and MUFA (B & D) in mock or HCMV-infected cells at 72 hpi from Figs 2B-2C and 4B-4C were compared. (B & D) Percent labeled FA in HCMV-infected cells from reductive carboxylation versus other pathways for SFA (C) and MUFA (D) were calculated by quantifying the ratio of 5-13C to U-13C-Gln. Octothorpe (#) indicates statistically significant differences of U-13C-Gln compared to 5-13C-Gln shown in C and D. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.01, **; P < 0.001, ***. n = 5-6.

https://doi.org/10.1371/journal.ppat.1014496.g005

Reductive carboxylation of glutamine yields cytosolic oxaloacetate when citrate is converted to acetyl-CoA via ATP Citrate Lyase (ACLY) (Fig 6A). Several fates are possible for oxaloacetate, including conversion to malate followed by conversion to pyruvate. The resulting pyruvate can feed into the TCA cycle to generate additional citrate that can support FA synthesis. In the U-13C-Gln labeling approach, this metabolism would result in 2-labeled and 3-labeled forms of pyruvate that can feed the TCA cycle by forming 2-labeled acetyl-CoA. The labeled acetyl-CoA can produce 2-labeled citrate via the TCA cycle followed by the generation of 2-labled acetyl-CoA in the cytosol that could be used for FA synthesis (Fig 6A) [7]. Additionally, our U-13C-Gln labeling strategy would also label metabolites in oxidative TCA cycle during HCMV infection (Fig 6B) [6,16]. In this case, U-13C-Gln feeding oxidative TCA would produce 5-labeled α-ketoglutarate, which can be converted to 4-labeled oxaloacetate in the mitochondria. The 4-labeled oxaloacetate could generate 6-labeled citrate if combined with 2-labeled acetyl-CoA from the pathway described in Fig 6A, resulting in labeled acetyl-CoA in the cytosol available for FA synthesis (Fig 6B). Alternatively, the 4-labeled oxaloacetate can be combined with unlabeled acetyl-CoA from glucose or other non-glutamine sources to generate 4-labeled citrate. The 4-labeled citrate generated in this manner would yield unlabeled acetyl-CoA in the cytosol. However, if the 4-labeled citrate generated through oxidative TCA cycle activity is converted to acetyl-CoA and oxaloacetate in the cytosol then the resulting 4-labeled oxaloacetate can be converted to 4-labeled malate and 3-labeled pyruvate which could be used to generate 2-labeled carbons in acetyl-coA that would supply FA synthesis (Fig 6B). Cytosolic oxaloacetate produced from oxidative TCA activity or reductive carboxylation would account for the additional labeling observed in U-13C-Gln compared to 5-13C-Gln in Fig 5C-5F. If this is the case, it would require the production of pyruvate from malate (Fig 6A-6B).

thumbnail
Download:
Fig 6. HCMV induces malic enzymes.

(A-B) Schematic of predicted FA labeling from malic enzyme (ME) pathway using U-13C-Gln. Blue circles indicate 13C-labeled carbons, and black circles indicate unlabeled carbons. Reductive carboxylation is shown in A and glutaminolysis is shown in B. (C-F) HFFs were infected with TB40/E-GFP or mock-infected. Whole-cell lysates were collected at the indicated times and analyzed by western blot. A representative western blot from three biological replicates is shown in B. Levels of ME1 (C), ME2 (D), or ME3 (E) were normalized to tubulin and quantified relative to mock at 4 hpi. Asterisk (*) indicates protein band based on reported protein size. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.01, **; P < 0.001, ***. n = 3.

https://doi.org/10.1371/journal.ppat.1014496.g006

Malate conversion to pyruvate is catalyzed by malic enzyme (ME). There are three ME isoforms: ME1 and ME3 are NADP+-dependent and found in the cytosol and the mitochondria, respectively, whereas ME2 is localized to the mitochondria with dual NAD+ and NADP+ specificity but a preference for NAD+ [33–37]. We measured ME1, ME2, and ME3 to determine if HCMV infection alters their levels. ME1 levels are increased by 72 hpi, and ME2 and ME3 levels are increased by 48 hpi during HCMV replication (Fig 6C-6F), suggesting HCMV induces these enzyme levels to support malate conversion to pyruvate. Collectively, increased ME levels coupled with U-13C-Gln versus 5-13C-Gln FA labeling results suggest that increased cytosolic oxaloacetate availability produced from citrate derived from oxidative TCA cycle and reductive carboxylation could feed the ME pathway and supply additional carbons to support HCMV-induced FA synthesis (Figs 5C-5F and 6).

Physiological levels of glutamine and glucose support increased FA levels during HCMV replication

Our results demonstrate that glutamine feeds lipid and FA synthesis during HCMV replication, including through reductive carboxylation and potentially through the ME pathway. Thus far, we used glucose and glutamine concentrations at levels found in most DMEM compositions, as this is a common medium used for studies investigating HCMV interactions with host metabolism in fibroblasts [6,11,14,16,22,24,38,39]. Since glucose and glutamine are at supraphysiological concentrations in most cell culture media, including DMEM, we sought to determine if physiological levels of glutamine alter its overall contribution to FA synthesis. In these experiments, we also reduced the level of glucose to human serum concentrations, since glucose was found to be the primary carbon source for FA synthesis in HCMV infection [6,7,10,11]. First, we determined if infection altered the levels of FA in DMEM with physiological glucose (5 mM) and glutamine (0.55 mM) compared to standard DMEM (25 mM glucose and 4 mM glutamine). We chose these physiological concentrations from levels reported in healthy human serum and are used in cell culture media that mimic nutrient levels in human plasma [40,41]. We infected or mock-infected HFF-1 cells and changed the media at 1 hpi to standard DMEM with 25 mM glucose and 4 mM glutamine (25/4) or DMEM containing 5 mM glucose and 0.55 mM glutamine (5/0.55). FA were measured at 72 hpi. In mock-infected cells, reducing glucose and glutamine resulted in little change to FA levels with only small differences in VLCFA that are at low abundance in uninfected cells (Fig 7A). HCMV infection in standard 25/4 DMEM increased the levels of VLCFA relative to either mock condition. Similarly, HCMV infection in 5/0.55 DMEM increased VLCFA levels, indicating that physiological levels of glucose and glutamine are sufficient to support increased FA synthesis and elongation caused by infection. Since our earlier experiments showed an increase in glutamine labeling of saturated and monounsaturated VLCFA, we examined the levels of C26:0, C26:1, C28:0, and C28:1 in cells cultured in 25/4 and 5/0.55. In mock-infected cells, these VLCFA levels in 5/0.55 remained similar to those cultured in 25/4 (Fig 7B-7E). As anticipated, these four VLCFA were increased in HCMV-infected cells cultured in 25/4. Levels in infected cells cultured in 5/0.55 were comparable to those in infected cells cultured in high glucose and glutamine media. To determine if physiological glucose and glutamine levels alter HCMV production, we measured infectious virus production at 120 and 144 hpi from cells infected at MOI 1 or 3. Virus production was reduced by 0.8-1-log in DMEM with physiological levels of glutamine and glucose (Fig 7F). Given this finding, we measured viral protein levels from the canonical kinetic classes of HCMV lytic replication. We found no difference in viral protein levels of any kinetic class, i.e., immediate-early, early, or late (Fig 7G-7J). We next asked if reduced HCMV production was the result of altered autophagic activation in 5/0.55 media. HCMV hijacks autophagic machinery for secondary envelopment and particle maturation [42], while nutritional stress activates autophagy [43]. We reasoned that nutrient stress from reduced levels of glucose and glutamine in 5/0.55 conditions could alter the autophagic response in HCMV-infected cells. We compared the levels of a common marker of autophagosomes, LC3-I and its lipidated form LC3-II, and found no difference in 5/0.55 DMEM conditions (S4 Fig). Overall, these results demonstrate that despite reduced infectious virus production, physiological glucose and glutamine levels are sufficient to support increased FA abundance and viral protein production that occurs during HCMV replication in supraphysiological glucose and glutamine conditions.

thumbnail
Download:
Fig 7. Physiological concentrations of glucose and glutamine support increased FA abundance during HCMV replication.

(A-E) HFF-1 were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing 25 mM glucose and 4 mM glutamine (25/4) or 5 mM glucose and 0.55 mM glutamine (5/0.55). At 72 hpi, lipids were extracted and FA were saponified and measured using LC-MS. (A) Heatmap of FA abundance relative to mock 25/4. Each column is a replicate for the indicated condition. Grey bars indicate FA not identified in the sample. FA abundance for C26:0 (B), C28:0 (C), C26:1 (D), and C28:1 (E) relative to mock 25/4. (F) HFFs were infected with TB40/E-GFP or mock-infected at MOI 1 or 3 and cultured in 25/4 or 5/0.55 mM glucose/glutamine. Culture supernatant was collected at 120 and 144 hpi. Viral titers were measured by 50% tissue culture infection dose assay (TCID50). (G-J) HFFs were infected and treated as described for F at MOI 3. Whole-cell lysates were collected at the indicated times and analyzed by western blot. A representative western blot from three biological replicates is shown in G. Levels of IE1 (H), UL38 (I), and pp28 (J) were normalized to tubulin and quantified relative to 25/4 at 120 hpi. Error bars represent SD. Statistics were performed on transformed data for F. One-way ANOVA with Dunnett’s test (B-E) and two-way ANOVA with Šídák’s test (F, H-J) were used to determine significance. P < 0.05, *; P < 0.001, ***. Error bars represent SD. n = 3-6.

https://doi.org/10.1371/journal.ppat.1014496.g007

Since physiological glucose and glutamine concentrations support HCMV-induced demand for FA synthesis, we evaluated the contribution of glutamine for FA synthesis in 5/0.55 DMEM. Specifically, we sought to understand if reducing glutamine and glucose would alter the use of glutamine in FA metabolism. We investigated this possibility with an isotopic labeling approach using U-13C-Gln in 5/0.55 DMEM compared to 25/4 DMEM. We infected or mock-infected HFFs and changed the culture media at 1 hpi to media containing either 4 mM or 0.55 mM U-13C-Gln and unlabeled glucose at 25 mM or 5 mM, respectively. First, we examined if altering carbon source concentrations altered glutamine labeling of SFA and MUFA in mock-infected cells. We found the level of glutamine carbon use for FA in mock-infected cells was reduced in 5/0.55 compared to 25/4 for C14:0, C16:0, C20:1, and ≥C22 saturated and monounsaturated VLCFA (Fig 8A-8B). However, HCMV infection was able to increase the amount of labeling in both 25/4 and 5/0.55 DMEM (Fig 8A-8B). When comparing 5/0.55 to 25/4 DMEM, we observed reduced glutamine feeding of FA synthesis in HCMV-infected cells (Fig 8A-8B). Further, the number of labeled carbons incorporated into C26:0 during infection was highest in 25/4 DMEM (Fig 8C). Together, these observations indicate that the level of glutamine incorporation into FA depends on the concentration of glutamine and glucose. Moreover, the results demonstrate that HCMV infection promotes the flow of glutamine to FA synthesis independent of supraphysiological levels of glutamine and glucose.

thumbnail
Download:
Fig 8. Glutamine and glucose contribute to FA synthesis when present at physiological levels during infection.

HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing U-13C-Gln (A-C) or U-13C-Glc (D-F) at 25 mM glucose and 4 mM glutamine (25/4) or 5 mM glucose and 0.55 mM glutamine (5/0.55). At 72 hpi, lipids were extracted and FA were saponified and measured using LC-MS. Percent 13C-labeled SFA or MUFA were quantified for glutamine (A-B) or glucose (D-E) labeling. Percent 13C-labeled C26:0 in HCMV-infected cells for glutamine (C) or glucose labeling (F). Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.05, *; P < 0.01, **; P < 0.001, ***. n = 6.

https://doi.org/10.1371/journal.ppat.1014496.g008

Given our findings that reducing the concentration of glutamine and glucose in DMEM to human plasma-like concentrations resulted in a lower level of glutamine incorporation into FA, we asked if the flow of glucose to FA synthesis is altered in 5/0.55 DMEM. For these experiments, we used uniformly labeled glucose (U-13C-Glc) and unlabeled glutamine in 25/4 and 5/0.55 DMEM. We quantified the percent of FA containing at least one labeled carbon from glucose after correcting for natural isotopic abundance. We focused on glucose labeling of SFA and MUFA. In mock-infected cells, glucose labeling of C14:0, C20:1, and ≥C22 saturated and monounsaturated VLCFA was reduced in 5/0.55 compared to 25/4 (Fig 8D-8E). HCMV increased the flow of glucose to FA in both 25/4 and 5/0.55 DMEM. Further, HCMV-infected cells showed no difference in FA labeling from glucose in 5/.055 compared to 25/4, except for a subtle but statistically significant reduction in C16:0 labeling (Fig 8D-8E). We counted the number of labeled carbons in C26:0 and found that ~8–10 labeled carbons from glucose were incorporated at a similar level in HCMV-infected cells cultured in 25/4 or 5/0.55 (Fig 8F). These results indicate that the flow of glucose to FA synthesis at supraphysiological levels is maintained at physiological levels of glucose and glutamine during HCMV infection. Since we observed that glutamine flow to FA, but not glucose flow, was impacted by reducing these two carbon sources to plasma-like levels in 5/0.55 DMEM, it is possible that their roles in lipogenesis are, at least in part, differentially regulated in HCMV-infected cells.

Discussion

Many viruses reprogram host metabolism to generate a metabolic environment favorable for viral replication. HCMV uses host factors to alter nutrient flow and remodel the lipidome for efficient virus production. Prior FA and lipidomic studies determined that HCMV reprograms glucose flow to supply carbon for lipogenesis [7,10,11,27]. Previous research demonstrated that HCMV replication is dependent on glutamine [16]. Though it was hypothesized that glutamine could contribute to FA synthesis [5], the only demonstrated role of glutamine during HCMV replication was maintenance of the TCA cycle (Fig 9, left side) [6,16]. In this study, we determined that glutamine is metabolized via reductive carboxylation to feed FA and lipid synthesis; thereby, contributing to lipidome remodeling necessary for efficient virus production (Fig 9, right side) [7,10,11]. Moreover, our results suggest malic enzyme pathway may be active to support additional carbon flow to FA synthesis (Figs 5, 6, and 9) [16]. The contribution of glutamine to FA synthesis was likely overlooked until now since glucose is the primary carbon contributor to this process and is at supraphysiological levels in most metabolic studies of HCMV (Fig 8) [6,7]. Based on the findings reported here, we now know of three carbon suppliers for FA synthesis during HCMV infection: glutamine, glucose, and acetate [6,7,10,11]. One interesting question is: why does HCMV promote redundant carbon suppliers for the same process? The answer may lie in HCMV’s broad tissue distribution and cell tropism. The virus replicates in various nutrient niches in the body with vastly different metabolite availability than what is modeled in current cell culture approaches [40,41]. In the human body, it may be advantageous for HCMV to promote reprogramming of metabolic processes using different metabolite contributors to support the same molecular building blocks, e.g., FA and lipids. In this case, if a metabolite like glucose becomes limited at the site of infection and can no longer provide for its varied needs in virus replication, another metabolite like glutamine can compensate for the loss. In further support of this idea, uridine partially compensates for a loss of glucose to support viral genome replication and viral protein glycosylation during HCMV replication [44]. Our current study highlights the importance of uncovering metabolic mechanisms that are overlooked in cell culture models but could be required for HCMV replication in the human body.

thumbnail
Download:
Fig 9. Model of glutamine carbon flow during HCMV replication.

Pyruvate (Pyr); oxaloacetate (OAA); α-ketoglutarate (αKG). Dashed lines indicate metabolic intermediates not shown. (Left) Prior model of glutamine flow during infection. Glutamine is converted through glutaminolysis to αKG that feeds oxidative TCA cycle. (Right) Revised model of glutamine flow during infection. Glutamine is metabolized to glutamate and then to αKG in the cytosol. Alternatively, glutamate can enter the mitochondria and be metabolized to αKG. αKG is metabolized to isocitrate in the cytosol via IDH1 or the mitochondria via IDH2. Isocitrate is converted to citrate. Citrate produced through reductive carboxylation or oxidative TCA is cleaved via ATP citrate lyase to produce acetyl-CoA and OAA. OAA can be metabolized to malate and then converted to pyruvate via ME1 resulting in cytosolic pyruvate which can then be transported into the mitochondria for acetyl-CoA production. Alternatively, the malate to pyruvate reaction can occur in the mitochondria via ME2/3 resulting in mitochondrial pyruvate production and subsequent conversion to acetyl-CoA.

https://doi.org/10.1371/journal.ppat.1014496.g009

Our previous understanding of glutamine during HCMV replication was to maintain the TCA cycle through glutaminolysis, i.e., catabolism of glutamine to α-ketoglutarate that feeds the TCA cycle [6,16]. In addition to its role in TCA cycle maintenance, our data demonstrate that glutamine is also metabolized through reductive carboxylation to support lipid synthesis (Fig 9). Together, these results indicate that HCMV promotes both oxidative TCA cycle and reductive carboxylation during replication. Our data cannot elucidate the ratio of glutamine contributing to oxidative TCA cycle versus reductive carboxylation. Both metabolic uses of glutamine may be essential for HCMV replication. Glutamine deprivation inhibits HCMV replication but starves both oxidative TCA cycle and reductive carboxylation of glutamine in infected cells [16,17]. Chambers et al. demonstrated that α-ketoglutarate, oxaloacetate, and pyruvate supplementation are each sufficient to restore virus production during glutamine deprivation [16]. Each supplement supports both metabolic pathways. α-Ketoglutarate can feed reductive carboxylation and oxidative TCA cycle, while oxaloacetate is a product of both pathways via cytosolic citrate cleavage (Fig 9). Pyruvate can be produced from glucose through glycolysis or from glutamine via the malic enzyme pathway (Figs 5-6) and can also be converted into oxaloacetate via pyruvate carboxylase if it is active. One possibility is that the virus has evolved to promote glutaminolysis that supports either oxidative TCA cycle or FA synthesis in specific metabolic niches where nutrients may be limiting [16]. It remains to be investigated how these two processes are regulated, and if differing nutrient environments alter their regulation and contribution to the metabolic needs of HCMV-infected cells. Although our study and the study by Chambers et al. cannot elucidate the essentiality of oxidative TCA cycle versus reductive carboxylation, they highlight the remarkable metabolic adaptability occurring in HCMV-infected cells.

Reductive carboxylation is widely recognized as the primary pathway by which glutamine contributes to FA and lipid synthesis [18,31,45,46]. However, its contribution to the pro-viral metabolic environment is only beginning to be interrogated. Our data demonstrate that HCMV promotes reductive carboxylation during replication (Figs 4, 5, 9, and S3). In the context of virus infection, reductive carboxylation activity has been reported in four other viruses, including two other mammalian virus infections: Adenovirus type 5 (Ad5) (non-enveloped, DNA virus) and murine norovirus (MNV) (non-enveloped, RNA virus) infection [47–50]. These studies identified an increase in 3-labeled TCA cycle metabolites during infection after culture in media containing U-13C-Gln. Malate, fumarate, or aspartate with 3-labeled carbons are only obtained through reductive carboxylation of U-13C-Gln after citrate cleavage to acetyl-CoA and oxaloacetate. However, these studies did not directly measure glutamine carbon incorporation into FA and lipids. Our data establish that reductive carboxylation supports FA synthesis during virus replication and remains the primary pathway by which glutamine contributes to FA synthesis (Figs 2, 4, 5, 9, S2, and S3). Moreover, by comparing FA labeling from uniformly and positionally labeled 13C-glutamine tracers, we demonstrated that additional carbons from glutamine are used for FA synthesis (Fig 5). Our results from HCMV, coupled with studies on Ad5 and MNV, indicate that reductive carboxylation is likely a shared metabolic mechanism in a diverse range of enveloped and non-enveloped DNA and RNA viruses.

Reductive carboxylation is active in cancer cells during hypoxia or impaired cellular respiration to sustain FA synthesis for rapid cellular proliferation [18–20,31,51]. Conversion of α-ketoglutarate to isocitrate during reduction carboxylation is catalyzed by IDH1 (cytosolic) or IDH2 (mitochondrial) (Fig 3A). In cancer cells, both IDH1 and IDH2 are shown to support reductive carboxylation [18,20]. We found that IDH1 and IDH2 are expressed in HFFs but do not increase during HCMV infection (Fig 3B-3D). These results suggest that reductive carboxylation during HCMV replication is regulated by other means, such as increased substrate availability (i.e., α-ketoglutarate) or IDH1 or IDH2 post-translational modification rather than increased enzyme expression [52,53]. Our findings indicate HCMV-induced reductive carboxylation occurs through either cytosolic IDH1 or mitochondrial IDH2 activity (Fig 9). Though it is possible this activity can be active in both locations, our understanding of metabolic flux during HCMV replication suggests that reductive carboxylation occurs in the cytosol through IDH1 activity. Munger et al. and others demonstrated increased flux through oxidative TCA cycle during infection, along with glucose carbons primarily shuttled to citrate production for FA synthesis [6,7,10,11]. Chamber et al. showed glutamine supports oxidative turning of the TCA, which allows glucose shuttling to citrate production [6,16]. In support of these studies, we observed increased IDH3A levels (Fig 3B and 3E), suggesting that oxidative TCA cycle activity is supported by increased IDH3 that converts isocitrate to α-ketoglutarate. Together, these results indicate HCMV replication drives oxidative turning of the TCA cycle. This conclusion suggests that IDH2 in the mitochondria is unlikely to support the reverse reaction, i.e., α-ketoglutarate to isocitrate (Fig 9). If IDH1 is active in the cytosol, it would allow both mechanisms of glutamine use during infection: oxidative TCA cycle maintenance and reductive carboxylation.

By using positionally labeled 5-13C-Gln, we found that reductive carboxylation is the primary pathway through which glutamine contributes to FA synthesis (Figs 2, 4, 5, S2 and S3). Comparison of U-13C-Gln and 5-13C-Gln FA labeling indicate that ~70% is derived from reductive carboxylation while 30% percent of FA labeling comes from other metabolic means (Fig 5C-5F). Cytosolic oxaloacetate produced from reductive carboxylation can feed malic enzyme pathway, where it is converted to malate and pyruvate through malate dehydrogenase and malic enzyme, respectively (Fig 6A-6B). The resulting pyruvate could feed oxidative TCA cycle to produce citrate to be used for FA synthesis. ME pathway would support the increased FA labeling we observed in U-13C-Gln relative to 5-13C-Gln (Fig 5C-5D). In support of this mechanism, ME activity is increased during HCMV infection [16], and we found all three isoforms of malic enzyme are increased during HCMV replication (Fig 6C-6F). ME activity could occur in the cytosol (ME1) or the mitochondria (ME2/ME3) (Fig 9). All three ME isoforms also produce NADPH, a cofactor for lipid synthesis, during malate to pyruvate conversion. HCMV could promote ME expression to support carbon flow to FA or to provide additional NADPH for lipid synthesis. Understanding the dynamic activity of enzymes that catalyze the same reaction in different cellular compartments is crucial for understanding metabolic reprogramming during infection. Our future studies will focus on elucidating the regulation and activity of ME isoforms and their role in HCMV replication.

HCMV infection remodels the lipid composition of the cell. We have previously demonstrated that several phospholipids and neutral lipid classes are increased during infection [7–9,54]. Our data indicate that glutamine carbons contribute to synthesis of all the lipid classes we examined except PG (Fig 1A-1B). PG is primarily synthesized in the mitochondria unlike other phospholipids that are synthesized at the endoplasmic reticulum [55]. Some PG with long-chain fatty acyl tails with two or fewer double-bonds are enriched in HCMV-infected cells, while most others are unchanged or reduced [7,8]. Further research is needed to understand the role of PG in HCMV-infected cells and the impact infection has on their metabolism. Nonetheless, our data show that HCMV infection does not promote the contribution of glutamine to the synthesis of PG lipids. Glutamine carbons were enriched in various tail lengths and degrees of saturation for PC, PE, DG, and TG (Fig 1C-1F). Of the different phospholipid classes, PC with VLCFA tails (PC-VLCFA) are among the most increased by infection [7,9]. Consistent with these findings, PC show the greatest increase in labeling compared to mock by 72 hpi, and carbons from glutamine are enriched in VLCFA (Figs 1A-1B, 2B-2C, S1A). PC with saturated and monounsaturated fatty acyl tails showed the highest amount of labeling in HCMV-infected cultures compared to other phospholipid classes (Figs 1C-1F, and S1). These data suggest that glutamine supports SFA and MUFA production to support increased PC-VLCFA synthesis [9–11]. Overall, our results indicate glutamine contributes to lipidome remodeling by supporting FA and lipid synthesis in ways that were not previously identified.

HCMV infection promotes reductive carboxylation to supply glutamine carbons for FA and lipid synthesis (Figs 1, 2, 4, S1, S2, and S3), indicating a viral protein drives glutamine metabolic reprogramming. HCMV UL38 is heavily involved in metabolic reprogramming during HCMV infection [11,23,39]. UL38 supports fatty acid elongation and drives consumption of glucose and glutamine [11,39]. Moreover, UL38 is necessary and sufficient for glutamine consumption [39]. Raymonda et al. demonstrated that cells expressing UL38 increased glutamine consumption in response to glucose deprivation [23]. Notably, acetyl-CoA and malonyl-CoA, carbon suppliers for FA synthesis, were increased in UL38-expressing cells deprived of glucose [23]. It is plausible that UL38 could drive glutamine carbon flow to FA synthesis via reductive carboxylation. UL37x1 is another potential candidate that may induce glutamine metabolic reprogramming, since it supports lipidome remodeling during infection [7]. While UL37x1 is not sufficient to induce lipid synthesis on its own [7], it could function with UL38 or UL13 [56] to regulate metabolism, including glutamine feeding reductive carboxylation for FA and lipid synthesis.

In the course of investigating the metabolic role of glutamine in lipogenesis, we observed that supraphysiological levels of 25 mM glucose and 4 mM glutamine support a higher level of HCMV production than plasma-like levels of 5 mM glucose and 0.55 mM glutamine (Fig 7F). Since viral protein levels were similar in both media conditions (Fig 7G-7J), we investigated if autophagy was altered by the media with reduced glucose and glutamine. Previous studies showed that HCMV infection induces autophagy, including increasing the levels of LC3-I and LC3-II autophagosome markers [57–60]. We found similar levels of LC3-I and LC3-II in HCMV-infected cells in both standard DMEM and 5/0.55 DMEM (S4 Fig), indicating that the defect in HCMV production is not due to an altered autophagic response.

Physiological glucose and glutamine levels did not alter increased FA abundance during HCMV infection but did impact carbon suppliers for FA synthesis. Glucose labeling of FA was reduced in mock-infected cells, but HCMV infection continues to promote the flow of carbons from glucose to FA synthesis (Fig 8D-8E). In contrast, glutamine labeling of FA was reduced in both mock and HCMV-infected cells in physiological glucose and glutamine concentrations (Fig 8A-8B). This observation suggests that cells reduce the flow of glutamine carbons to FA synthesis in response to the lower concentrations of glucose and glutamine. Further, this result suggests that while HCMV infection increases the flow of glutamine to FA, infection cannot completely overcome the impact that reduced glucose and glutamine have on the cell. There are several other cases in which the metabolic environment of infected cells affects HCMV replication. As discussed earlier, glutamine deprivation hinders replication [16]. Likewise, removal of glucose from the environment blocks HCMV replication [16,23,44]. However, even when glutamine is removed, adding α-ketoglutarate, oxaloacetate, and pyruvate back to the environment provides a partial rescue [16]. Similarly, the loss in replicative steps of virus replication when glucose is removed can be restored by the addition of uridine or ribose to the environment [44]. Recently, we found that the addition of a FA to the environment can reduce HCMV replication [61]. Together, these studies demonstrate that the metabolic environment impacts the outcomes of infection and is a critical determinant in HCMV’s success. These findings highlight the need to further study the influence of nutrients in the metabolic environment of infection.

Collectively, our study adds to our current model of the metabolic flow of nutrients following HCMV infection (Fig 9). First, HCMV infection promotes glutamine carbon flow through reductive carboxylation to FA synthesis, and second, infection increases the activity in metabolic steps that feed additional glutamine carbons to FA production, potentially through increased malic enzyme levels. Moreover, these results suggest that HCMV-induces metabolic reprogramming to support FA synthesis from several metabolic pools; this may allow metabolic flexibility in different nutrient niches at sites of infection to ensure infected cells can support virus replication. Overall, the findings expand our understanding of the role of glutamine in infected cells and metabolic reprogramming driven by HCMV.

Materials and methods

Cells, viruses, and reagents

Human foreskin fibroblasts (HFFs) or HFF-1 (American Type Culture Collection; ATCC) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 4.5 g/L (25 mM) glucose (Gibco 11965), 10% fetal bovine serum (FBS), 10 mM HEPES, and penicillin/streptomycin (P/S). Prior to infection, cells were held at confluency for three days in full growth medium and then another day in serum-free (SF) medium (DMEM, 10 mM HEPES, P/S). DMEM containing 13C-labeled glucose or glutamine were prepared using DMEM powder (Gibco) and adding 44 mM sodium bicarbonate with the indicated concentration of 13C-labeled glucose or glutamine and the indicated concentration of unlabeled glucose or glutamine.

Cells were infected with HCMV strain TB40/E encoding free GFP (TB40/E-GFP) at MOI 3 IU/cell (unless noted otherwise) for 1 h and washed two times with PBS before adding fresh media. Mock-infected cells were treated with inoculum that did not contain virus. At 48 hpi, media was changed to maintain nutrient levels.

Virus stocks were produced by propagating BAC-derived HCMV in fibroblasts and concentrating infected cell culture supernatant by pelleting through a sorbitol (20% sorbitol, 50 mM Tris pH 7.2, 1 mM MgCl2) cushion for 80 min at 52,931xg using ultracentrifugation. Viruses were resuspended in SF DMEM. Stocks and experimental samples were titered using 50% tissue culture infectious dose assay (TCID50) beginning with a 1:10 dilution of concentrated virus. GFP-positive wells were counted at two-weeks post infection on duplicate plates. A well was considered positive if it contained a foci of ≥3 GFP-positive cells. All TCID50 assays were completed in DMEM (Gibco 11965) containing 25 mM glucose and 4 mM glutamine supplemented with 10% fetal bovine serum (FBS), 10 mM HEPES, and penicillin/streptomycin (P/S).

Lipid analysis

Lipid analysis was performed using LC-MS/MS. Duplicate wells for each condition were analyzed, and a third well was used to determine a cell count for normalization. Additional wells that did not contain cells were used as a control to determine contaminants from the extraction or LC-MS/MS steps. Cells were washed three times in cold PBS and fixed in cold 50% methanol (MeOH) with 0.1 M HCl. Samples were scraped and transferred to glass vials. Lipids were extracted two times in chloroform and dried under nitrogen gas. Lipids were resuspended in 1:1:1 solution of MeOH:isopropanol:chloroform at 100 µl per 200,000 cells, stored at 7°C in an autosampler, and separated using reverse-phase chromatography on a Vanquish UHPLC system using a Kinetex C18 reverse-phase column (Phenomenex) at 60°C using solvent A (40:60 water:MeOH, 10 mM ammonium formate, 0.1% formic acid) and solvent B (10:90 MeOH:Isopropanol, 10 mM ammonium formate, 0.1% formic acid). UHPLC was performed at a flow rate of 0.25 mL/min with a total run time of 30 min. The gradient was established at 25% solvent B for 2 min, 65% solvent B at curve value 4 from 2-4 min, 65% held for 4–5 min, 100% solvent B at a curve of 4 from 5-16 min, and held at 100% solvent B from 16-20 min. The column was briefly washed and equilibrated after each sample. Lipids were analyzed using a Thermo Fisher Q-Exactive Plus mass spectrometer operating in full MS1/dd-MS2 with TopN mode set to 5 scans. MS1 spectra were collected from 200 to 1,600 m/z at 70,000 resolution using an automatic gain control (AGC) target of 1e6 with transient time of 250 ms. Lipids were ionized using HESI in positive and negative mode. In positive mode, 8 µl of sample was analyzed using sheath gas of 45, auxiliary gas of 16, sweep gas of 2, spray voltage of 3.5 kV, S-lens RF of 65, ion transfer tube temperature of 320°C, and vaporization temperature of 220°C. In negative mode, 10 µl of sample was analyzed using sheath gas of 44, auxiliary gas of 14, sweep gas of 2, spray voltage of 3.3 kV, S-lens RF of 76, ion transfer tube temperature of 320°C, and vaporization temperature of 220°C. MS2 spectra were collected at 35,000 resolution with an AGC target of 1e5. In positive mode, NCE of 30 was used. In negative mode, NCE value of 20 was used. The instrument was calibrated weekly and immediately prior to analysis. For all lipid analysis, blank samples were run before, after, and interspersed between samples. Data were analyzed using Maven 2 software [62] and corrected for the natural 13C abundance using MATLAB [6]. The percent of labeled carbons were determined by subtracting the unlabeled fraction from one and multiplying by 100 to obtain percent 13C label.

Fatty acid (FA) analysis

FA analysis was performed using LC-MS. Duplicate wells for each condition were extracted and analyzed in parallel. Additional wells that did not contain cells were used as a control to determine contaminants from the extraction or LC-MS steps. Cells were collected and lipids extracted as detailed in the lipid analysis section. After drying under nitrogen gas, FA tails were chemically cleaved by saponifying in 90% MeOH and 0.3 M potassium hydroxide at 80°C for 1 h followed by neutralization using formic acid. FA were extracted two times in hexane and dried under nitrogen gas. FA were resuspended in 600 µl of 1:1:1 solution of MeOH:isopropanol:chloroform and stored at 7°C in an autosampler. FA were separated by reverse-phase chromatography on a Vanquish UHPLC system using a Luna C8 reverse-phase column (Phenomenex) at 60° and solvent A and solvent B described in the lipids analysis section. UHPLC was performed at a flow rate of 0.2 mL/min with a total run time of 30 min. The gradient was established at 10% solvent B for 2.5 min, 100% solvent B at curve value 5 from 2.5-10 min, and held at 100% solvent B from 10-22 min. The column was briefly washed and equilibrated after each sample. FA were analyzed using a Thermo Scientific Orbitrap Exploris 240 mass spectrometer at a resolution of 120,000 for unlabeled FA analysis or 240,000 for 13C-labeled FA analysis. Spectra were collected at a range of 200–650 m/z. FA were ionized by HESI in negative mode using 10 µl of sample. The following settings were used: sheath gas of 50, auxiliary gas of 25, sweep gas of 10, spray voltage of 3 kV, RF lens of 110%, ion transfer tube temperature of 300°C, and vaporization temperature of 320°. The internal calibration source “Easy-IC” was used in scan-to-scan mode. For labeling experiments, the instrument was calibrated immediately prior to the start of the analysis; otherwise, the instrument was calibrated weekly. Blank samples were run before, after, and interspersed between samples. Data were analyzed using Maven 2 software [62]. For labeling experiments, the data were corrected for the natural 13C abundance using MATLAB [6]. Unlabeled FA abundances were determined by calculating the total ion count for each sample and setting each FA in the sample relative to the total ion count. The percent of labeled carbons were determined by subtracting the unlabeled fraction from one and multiplying by 100 to obtain percent 13C label.

Protein analysis

Protein levels were analyzed by western blot. Whole-cell lysates were collected by in-well lysis followed by scraping. Proteins were resolved by SDS-PAGE using Mini-Protean 4–20% gradient gels (BioRad) and transferred to nitrocellulose membrane (LI-COR). Membranes were blocked in 5% milk in tris buffered saline with 0.1% Tween 20 (TBST). Primary antibodies were diluted in 1–5% milk in TBST and incubated with rocking for 1 h at room temperature or overnight at 4°C. Secondary antibodies were diluted in 5% milk in TBST and incubated for 1 h with rocking in the dark at room temperature. Images were taken on a LI-COR Odessey CLx imager, and quantitation was performed using Image Studio software.

The following primary antibodies were used for western blot analysis: Mouse-anti tubulin (Sigma; 1:1,000), mouse anti-IE1 (clone 1B12; 1:100) [11], mouse anti-UL38 (clone 8D6; 1:100), mouse anti-pp28 (clone 10B4; 1:100), rabbit anti-IDH1 (Proteintech; 1:1,000), rabbit anti-IDH2 (Proteintech; 1:1,000), rabbit anti-IDH3A (Proteintech;1:1,000), rabbit anti-ME1 (Proteintech; 1:1,000), rabbit anti-ME2 (Proteintech; 1:1,000), rabbit anti-ME3 (ThermoFisher; 1:1,000), and rabbit anti-LC3A/B (D3U4C; Cell Signaling Technology; 1:500). HCMV antibodies were a gift from Dr. Thomas Shenk (Princeton University). Goat anti-mouse DyLight 800 (1:10,000) or goat anti-rabbit DyLight 680 (1:10,000) were used for secondary antibodies.

Graphics and statistics

Schematics and models were created with BioRender.com. Graphs were designed and statistical testing were performed using GraphPad Prism. The appropriate statistical tests for each experiment are noted in the figure legends. Heatmaps were designed using R.

Supporting information

S1 Fig. Glutamine supports lipid synthesis across tail desaturations during HCMV replication.

HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing uniformly labeled 13C glutamine (U-13C-Gln). At 72 hpi, lipids were extracted and measured using LC-MS/MS. Percent labeled lipids are shown for PC (A), PE (B), DG (C), and TG (D). Error bars represent SD. n = 6.

https://doi.org/10.1371/journal.ppat.1014496.s001

(TIF)

S2 Fig. Glutamine supports FA synthesis by 48 hpi during HCMV replication.

(A-C) HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing U-13C-Gln. At 48 hpi, lipids were extracted and FA were saponified. Labeled FA were measured using LC-MS. Percent 13C-labeled SFA (A) and MUFA (B) were quantified. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.05, *; P < 0.01, **; P < 0.001, ***. n = 6.

https://doi.org/10.1371/journal.ppat.1014496.s002

(TIF)

S3 Fig. Reductive carboxylation activity increases by 48 hpi during HCMV replication.

(A-C) HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing 5-13C-Gln. At 48 hpi, lipids were extracted and FA were saponified. Labeled FA were measured using LC-MS. Percent 13C-labeled SFA (A) and MUFA (B) were quantified. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. P < 0.001, ***. n = 5–6.

https://doi.org/10.1371/journal.ppat.1014496.s003

(TIF)

S4 Fig. Physiological concentrations of glucose and glutamine do not alter autophagy during HCMV replication.

(A-C) HFFs were infected with TB40/E-GFP or mock-infected. At 1 hpi, virus inoculum was removed and replaced with media containing 25 mM glucose and 4 mM glutamine (25/4) or 5 mM glucose and 0.55 mM glutamine (5/0.55). Whole-cell lysates were collected at the indicated times and analyzed by western blot. A representative western blot from three biological replicates is shown in A. Levels of LC3-1 (B) and LC3-II (C) were normalized to tubulin and quantified relative to 25/4 at 120 hpi. Error bars represent SD. Two-way ANOVA with Šídák’s test were used to determine significance. n = 3.

https://doi.org/10.1371/journal.ppat.1014496.s004

(TIF)

S1 File. Raw Gel. Raw western blot images.

Original uncropped images of all blots presented in this study. Images are labeled according to the corresponding figure.

https://doi.org/10.1371/journal.ppat.1014496.s005

(PDF)

Acknowledgments

We thank Dr. Jim Alwine for helpful feedback. We thank additional members of the Purdy Lab for helpful feedback on this project, including Ian Kline, Nate Schmidt, and Magi Passols Manzano. We thank the members of Dr. Felica Goodrum and Dr. Koenraad Van Doorslaer laboratories for reagents and their input on the project.

References

  1. 1. Yu Y, Clippinger AJ, Alwine JC. Viral effects on metabolism: changes in glucose and glutamine utilization during human cytomegalovirus infection. Trends Microbiol. 2011;19(7):360–7. pmid:21570293
  2. 2. Fulkerson HL, Nogalski MT, Collins-McMillen D, Yurochko AD. Overview of Human Cytomegalovirus Pathogenesis. Methods Mol Biol. 2021;2244:1–18. pmid:33555579
  3. 3. Lazzarotto T, Guerra B, Gabrielli L, Lanari M, Landini MP. Update on the prevention, diagnosis and management of cytomegalovirus infection during pregnancy. Clin Microbiol Infect. 2011;17(9):1285–93. pmid:21631642
  4. 4. Griffiths P, Reeves M. Pathogenesis of human cytomegalovirus in the immunocompromised host. Nat Rev Microbiol. 2021;19(12):759–73. pmid:34168328
  5. 5. Rodríguez-Sánchez I, Munger J. Meal for Two: Human Cytomegalovirus-Induced Activation of Cellular Metabolism. Viruses. 2019;11(3):273. pmid:30893762
  6. 6. Munger J, Bennett BD, Parikh A, Feng X-J, McArdle J, Rabitz HA, et al. Systems-level metabolic flux profiling identifies fatty acid synthesis as a target for antiviral therapy. Nat Biotechnol. 2008;26(10):1179–86. pmid:18820684
  7. 7. Xi Y, Harwood S, Wise LM, Purdy JG. Human Cytomegalovirus pUL37x1 Is Important for Remodeling of Host Lipid Metabolism. J Virol. 2019;93(21):e00843-19. pmid:31391267
  8. 8. Xi Y, Lindenmayer L, Kline I, von Einem J, Purdy JG. Human Cytomegalovirus Uses a Host Stress Response To Balance the Elongation of Saturated/Monounsaturated and Polyunsaturated Very-Long-Chain Fatty Acids. mBio 2021;12.
  9. 9. Kline I, Mokry RL, Xi Y, Passols Manzano M, Layesa S, Sohrab Ali N, et al. Human cytomegalovirus promotes de novo PC synthesis during early virus replication. J Virol. 2025;99(9):e0057925. pmid:40827916
  10. 10. Koyuncu E, Purdy JG, Rabinowitz JD, Shenk T. Saturated very long chain fatty acids are required for the production of infectious human cytomegalovirus progeny. PLoS Pathog. 2013;9(5):e1003333. pmid:23696731
  11. 11. Purdy JG, Shenk T, Rabinowitz JD. Fatty acid elongase 7 catalyzes lipidome remodeling essential for human cytomegalovirus replication. Cell Rep. 2015;10(8):1375–85. pmid:25732827
  12. 12. Liu STH, Sharon-Friling R, Ivanova P, Milne SB, Myers DS, Rabinowitz JD, et al. Synaptic vesicle-like lipidome of human cytomegalovirus virions reveals a role for SNARE machinery in virion egress. Proc Natl Acad Sci U S A. 2011;108(31):12869–74. pmid:21768361
  13. 13. Gudleski-O’Regan N, Greco TM, Cristea IM, Shenk T. Increased expression of LDL receptor-related protein 1 during human cytomegalovirus infection reduces virion cholesterol and infectivity. Cell Host Microbe. 2012;12(1):86–96. pmid:22817990
  14. 14. Yu Y, Maguire TG, Alwine JC. Human cytomegalovirus infection induces adipocyte-like lipogenesis through activation of sterol regulatory element binding protein 1. J Virol. 2012;86(6):2942–9. pmid:22258239
  15. 15. Spencer CM, Schafer XL, Moorman NJ, Munger J. Human cytomegalovirus induces the activity and expression of acetyl-coenzyme A carboxylase, a fatty acid biosynthetic enzyme whose inhibition attenuates viral replication. J Virol. 2011;85(12):5814–24. pmid:21471234
  16. 16. Chambers JW, Maguire TG, Alwine JC. Glutamine metabolism is essential for human cytomegalovirus infection. J Virol. 2010;84(4):1867–73. pmid:19939921
  17. 17. Mokry RL, Schumacher ML, Hogg N, Terhune SS. Nitric oxide circumvents virus-mediated metabolic regulation during human cytomegalovirus infection. mBio. 2020;11.
  18. 18. Metallo CM, Gameiro PA, Bell EL, Mattaini KR, Yang J, Hiller K, et al. Reductive glutamine metabolism by IDH1 mediates lipogenesis under hypoxia. Nature. 2011;481(7381):380–4. pmid:22101433
  19. 19. Fendt S-M, Bell EL, Keibler MA, Olenchock BA, Mayers JR, Wasylenko TM, et al. Reductive glutamine metabolism is a function of the α-ketoglutarate to citrate ratio in cells. Nat Commun. 2013;4:2236. pmid:23900562
  20. 20. Wise DR, Ward PS, Shay JES, Cross JR, Gruber JJ, Sachdeva UM, et al. Hypoxia promotes isocitrate dehydrogenase-dependent carboxylation of α-ketoglutarate to citrate to support cell growth and viability. Proc Natl Acad Sci U S A. 2011;108(49):19611–6. pmid:22106302
  21. 21. Sun RC, Denko NC. Hypoxic regulation of glutamine metabolism through HIF1 and SIAH2 supports lipid synthesis that is necessary for tumor growth. Cell Metab. 2014;19(2):285–92. pmid:24506869
  22. 22. Vastag L, Koyuncu E, Grady SL, Shenk TE, Rabinowitz JD. Divergent effects of human cytomegalovirus and herpes simplex virus-1 on cellular metabolism. PLoS Pathog. 2011;7(7):e1002124. pmid:21779165
  23. 23. Raymonda MH, Rodríguez-Sánchez I, Schafer XL, Smorodintsev-Schiller L, Harris IS, Munger J. Cytomegalovirus-induced inactivation of TSC2 disrupts the coupling of fatty acid biosynthesis to glucose availability resulting in a vulnerability to glucose starvation. mBio. 2024;15(1):e0303123. pmid:38117060
  24. 24. Yu Y, Maguire TG, Alwine JC. ChREBP, a glucose-responsive transcriptional factor, enhances glucose metabolism to support biosynthesis in human cytomegalovirus-infected cells. Proc Natl Acad Sci U S A. 2014;111(5):1951–6. pmid:24449882
  25. 25. DeVito SR, Ortiz-Riaño E, Martínez-Sobrido L, Munger J. Cytomegalovirus-mediated activation of pyrimidine biosynthesis drives UDP-sugar synthesis to support viral protein glycosylation. Proc Natl Acad Sci U S A. 2014;111(50):18019–24. pmid:25472841
  26. 26. Yu Y, Maguire TG, Alwine JC. Human cytomegalovirus activates glucose transporter 4 expression to increase glucose uptake during infection. J Virol. 2011;85(4):1573–80. pmid:21147915
  27. 27. Munger J, Bajad SU, Coller HA, Shenk T, Rabinowitz JD. Dynamics of the cellular metabolome during human cytomegalovirus infection. PLoS Pathog. 2006;2(12):e132. pmid:17173481
  28. 28. Moreno I Jr, Rodríguez-Sánchez I, Schafer X, Munger J. Human cytomegalovirus induces neuronal enolase to support virally mediated metabolic remodeling. Proc Natl Acad Sci U S A. 2022;119(49):e2205789119. pmid:36459650
  29. 29. Yu Y, Pierciey FJ Jr, Maguire TG, Alwine JC. PKR-like endoplasmic reticulum kinase is necessary for lipogenic activation during HCMV infection. PLoS Pathog. 2013;9(4):e1003266. pmid:23592989
  30. 30. Shon JC, Lee SM, Jung J-H, Wu Z, Kwon YS, Sim H-J, et al. Integrated metabolomics and lipidomics reveals high accumulation of polyunsaturated lysoglycerophospholipids in human lung fibroblasts exposed to fine particulate matter. Ecotoxicol Environ Saf. 2020;202:110896. pmid:32622306
  31. 31. Mullen AR, Wheaton WW, Jin ES, Chen P-H, Sullivan LB, Cheng T, et al. Reductive carboxylation supports growth in tumour cells with defective mitochondria. Nature. 2011;481(7381):385–8. pmid:22101431
  32. 32. Gaude E, Schmidt C, Gammage PA, Dugourd A, Blacker T, Chew SP, et al. NADH Shuttling Couples Cytosolic Reductive Carboxylation of Glutamine with Glycolysis in Cells with Mitochondrial Dysfunction. Mol Cell. 2018;69(4):581-593.e7. pmid:29452638
  33. 33. Rutter WJ, Lardy HA. Purification and properties of pigeon liver malic enzyme. J Biol Chem. 1958;233(2):374–82. pmid:13563505
  34. 34. Ochoa S, Mehler A, Kornberg A. Reversible oxidative decarboxylation of malic acid. J Biol Chem. 1947;167(3):871. pmid:20287922
  35. 35. Simpson E, Estabrook RW. The “malate shuttle” and control of steroid hydroxylation in the adrenal cortex. Adv Enzyme Regul. 1969;7:259–79. pmid:4391642
  36. 36. Simpson ER, Estabrook RW. Mitochondrial malic enzyme: the source of reduced nicotinamide adenine dinucleotide phosphate for steroid hydroxylation in bovine adrenal cortex mitochondria. Arch Biochem Biophys. 1969;129(1):384–95. pmid:4178715
  37. 37. Pongratz RL, Kibbey RG, Shulman GI, Cline GW. Cytosolic and mitochondrial malic enzyme isoforms differentially control insulin secretion. J Biol Chem. 2007;282(1):200–7. pmid:17102138
  38. 38. Landini MP. Early enhanced glucose uptake in human cytomegalovirus-infected cells. J Gen Virol. 1984;65 (Pt 7):1229–32. pmid:6086816
  39. 39. Rodríguez-Sánchez I, Schafer XL, Monaghan M, Munger J. The Human Cytomegalovirus UL38 protein drives mTOR-independent metabolic flux reprogramming by inhibiting TSC2. PLoS Pathog. 2019;15(1):e1007569. pmid:30677091
  40. 40. Cantor JR, Abu-Remaileh M, Kanarek N, Freinkman E, Gao X, Louissaint A Jr, et al. Physiologic Medium Rewires Cellular Metabolism and Reveals Uric Acid as an Endogenous Inhibitor of UMP Synthase. Cell. 2017;169(2):258-272.e17. pmid:28388410
  41. 41. Vande Voorde J, Ackermann T, Pfetzer N, Sumpton D, Mackay G, Kalna G, et al. Improving the metabolic fidelity of cancer models with a physiological cell culture medium. Sci Adv. 2019;5(1):eaau7314. pmid:30613774
  42. 42. Taisne C, Lussignol M, Hernandez E, Moris A, Mouna L, Esclatine A. Human cytomegalovirus hijacks the autophagic machinery and LC3 homologs in order to optimize cytoplasmic envelopment of mature infectious particles. Sci Rep. 2019;9(1):4560. pmid:30872707
  43. 43. Kroemer G, Mariño G, Levine B. Autophagy and the integrated stress response. Mol Cell. 2010;40(2):280–93. pmid:20965422
  44. 44. Mokry RL, Purdy JG. Glucose-independent human cytomegalovirus replication is supported by metabolites that feed upper glycolytic branches. Proc Natl Acad Sci U S A. 2024;121(48):e2412966121. pmid:39560652
  45. 45. Holleran AL, Briscoe DA, Fiskum G, Kelleher JK. Glutamine metabolism in AS-30D hepatoma cells. Evidence for its conversion into lipids via reductive carboxylation. Mol Cell Biochem. 1995;152(2):95–101. pmid:8751155
  46. 46. Yoo H, Antoniewicz MR, Stephanopoulos G, Kelleher JK. Quantifying reductive carboxylation flux of glutamine to lipid in a brown adipocyte cell line. J Biol Chem. 2008;283(30):20621–7. pmid:18364355
  47. 47. Thai M, Thaker SK, Feng J, Du Y, Hu H, Ting Wu T, et al. MYC-induced reprogramming of glutamine catabolism supports optimal virus replication. Nat Commun. 2015;6:8873. pmid:26561297
  48. 48. Hafner A, Meurs N, Garner A, Azar E, Kannan A, Passalacqua KD, et al. Norovirus NS1/2 protein increases glutaminolysis for efficient viral replication. PLoS Pathog. 2024;20(7):e1011909. pmid:38976719
  49. 49. He S-T, Lee D-Y, Tung C-Y, Li C-Y, Wang H-C. Glutamine Metabolism in Both the Oxidative and Reductive Directions Is Triggered in Shrimp Immune Cells (Hemocytes) at the WSSV Genome Replication Stage to Benefit Virus Replication. Front Immunol. 2019;10:2102. pmid:31555294
  50. 50. Fu X, Li K, Niu Y, Lin Q, Liang H, Luo X, et al. The mTOR/PGC-1α/SIRT3 Pathway Drives Reductive Glutamine Metabolism to Reduce Oxidative Stress Caused by ISKNV in CPB Cells. Microbiol Spectr. 2022;10(1):e0231021. pmid:35019690
  51. 51. Filipp FV, Scott DA, Ronai ZA, Osterman AL, Smith JW. Reverse TCA cycle flux through isocitrate dehydrogenases 1 and 2 is required for lipogenesis in hypoxic melanoma cells. Pigment Cell Melanoma Res. 2012;25(3):375–83. pmid:22360810
  52. 52. Weeks J, Strom AI, Widjaja V, Alexander S, Pucher DK, Sohl CD. Evaluating Mechanisms of IDH1 Regulation through Site-Specific Acetylation Mimics. Biomolecules. 2021;11(5):740. pmid:34065652
  53. 53. Yu Y, Chen Y, Liu K, Cheng J, Tu J. SUMOylation enhances the activity of IDH2 under oxidative stress. Biochem Biophys Res Commun. 2020;532(4):591–7. pmid:32900482
  54. 54. Griffante G, Hewelt-Belka W, Albano C, Gugliesi F, Pasquero S, Castillo Pacheco SF, et al. IFI16 Impacts Metabolic Reprogramming during Human Cytomegalovirus Infection. mBio. 2022;13(3):e0043522. pmid:35420480
  55. 55. Horvath SE, Daum G. Lipids of mitochondria. Prog Lipid Res. 2013;52(4):590–614. pmid:24007978
  56. 56. Betsinger CN, Jankowski CSR, Hofstadter WA, Federspiel JD, Otter CJ, Jean Beltran PM, et al. The human cytomegalovirus protein pUL13 targets mitochondrial cristae architecture to increase cellular respiration during infection. Proc Natl Acad Sci U S A. 2021;118(32):e2101675118. pmid:34344827
  57. 57. Chaumorcel M, Lussignol M, Mouna L, Cavignac Y, Fahie K, Cotte-Laffitte J, et al. The human cytomegalovirus protein TRS1 inhibits autophagy via its interaction with Beclin 1. J Virol. 2012;86(5):2571–84. pmid:22205736
  58. 58. Belzile J-P, Sabalza M, Craig M, Clark AE, Morello CS, Spector DH. Trehalose, an mTOR-Independent Inducer of Autophagy, Inhibits Human Cytomegalovirus Infection in Multiple Cell Types. J Virol. 2015;90(3):1259–77. pmid:26559848
  59. 59. McFarlane S, Aitken J, Sutherland JS, Nicholl MJ, Preston VG, Preston CM. Early induction of autophagy in human fibroblasts after infection with human cytomegalovirus or herpes simplex virus 1. J Virol. 2011;85(9):4212–21. pmid:21325419
  60. 60. Altman AM, Miller MJ, Mahmud J, Smith NA, Chan GC. Human Cytomegalovirus-Induced Autophagy Prevents Necroptosis of Infected Monocytes. J Virol. 2020;94(22):e01022-20. pmid:32878887
  61. 61. Xi Y, Mokry RL, Armas ND, Kline I, Wegner M, Purdy JG. Human Cytomegalovirus Infection Reduces an Endogenous Antiviral Fatty Acid by Promoting Host Metabolism. bioRxiv. 2025.
  62. 62. Seitzer P, Bennett B, Melamud E. MAVEN2: An Updated Open-Source Mass Spectrometry Exploration Platform. Metabolites. 2022;12(8):684. pmid:35893250
  63. 63. Sud M, Fahy E, Cotter D, Azam K, Vadivelu I, Burant C, et al. Metabolomics Workbench: An international repository for metabolomics data and metadata, metabolite standards, protocols, tutorials and training, and analysis tools. Nucleic Acids Res. 2016;44(D1):D463-70. pmid:26467476