Human Cytomegalovirus uses a Host Stress Response to Balance the Elongation of Saturated/Monounsaturated and Polyunsaturated Very Long Chain Fatty Acids

TLDR: The role of stress in HCMV reprogramming of lipid metabolism is poorly understood as discussed by the authors, and the role of ER stress mediator, namely PKR-like ER kinase (PERK) in the ER stress pathway and measured lipid changes using lipidomics to determine if PERK is needed for lipid changes associated with HCMVs infection.

Abstract: Stress and virus infection are known to regulate lipid metabolism in cells. Human cytomegalovirus infection induces fatty acid (FA) elongation and increases the cellular abundance of lipids with very long-chain FA tails (VLCFAs). While reprogramming of metabolism can be stress-related, the role of stress in HCMV reprogramming of lipid metabolism is poorly understood. In this study, we engineered cells to knockout PKR-like ER kinase (PERK) in the ER stress pathway and measured lipid changes using lipidomics to determine if PERK is needed for lipid changes associated with HCMV infection. We found that in HCMV-infected cells, PERK promotes the increase in the levels of phospholipids with saturated FA (SFA) and monounsaturated FA (MUFA) VLCFAs tails. Consistent with the SFA/MUFA lipidome changes, PERK enhances the protein levels of FA elongase 7 (ELOVL7), which elongates SFA and MUFA VLCFAs. Additionally, we found that increases in the elongation of polyunsaturated fatty acids (PUFAs) associated with HCMV infection was independent of PERK and that lipids with PUFA tails accumulated in HCMV-infected PERK knockout cells. Consistent with the PUFA lipidome changes, the protein levels of ELOVL5, which elongates PUFAs, are increased by HCMV infection through a PERK-independent mechanism. These observations show that PERK differentially regulates ELOVL7 and ELOVL5, creating a balance between the synthesis of lipids with SFA/MUFA tails and PUFA tails. Additionally, we found that PERK was necessary for virus replication and the infectivity of released viral progeny. Overall, our findings indicate that PERK--and more broadly, ER stress--may be necessary for membrane biogenesis needed to generate infectious HCMV virions. IMPORTANCEHCMV is a common herpesvirus that establishes lifelong persistent infections. While infection is asymptomatic in most people, HCMV causes life-threatening illnesses in immunocompromised people, including transplant recipients and cancer patients. Additionally, HCMV infection is a leading cause of congenital disabilities. HCMV replication relies on lipid synthesis. Here, we demonstrated that the ER stress mediator, PERK, controls fatty acid (FA) elongation and cellular abundance of several types of lipids following HCMV infection. Specifically, we found that PERK promotes FA elongase 7 synthesis of lipids with saturated/monounsaturated very long-chain FA tails which are important for building the viral membrane of infectious HCMV virions. Overall, our study shows that PERK is an essential host factor that supports HCMV replication and promotes lipidome changes caused by HCMV infection.

Figures

Figure 9. HCMV infection elevates PLs with SFA and MUFA VLCFA tails that is in part dependent on PERK. (A) Relative abundance of PLs that were most prominently changed in TB40/E-infected cells relative to uninfected cells at 96 hpi. These are the top six elevated PLs shown in the heatmap of Fig. 8B. Three independent experiments were performed, each with duplicated samples for a total of six data points. Each data point is graphed relative to the levels observed in uninfected NT cells. For comparison, the relative levels at 72 hpi are also shown. *p<0.5, **p<0.01, ***p<0.001. One-way ANOVA, Turkey test. (B) Tail composition for lipids shown in part A. SFA and MUFA VLCFA tails are in bold text.

Figure 1. HCMV infection upregulates PERK and ATF4 protein levels. (A)

Figure 6. HCMV induces TGs with very long-chain saturated (SFA) or monounsaturated fatty acid (MUFA) tails independent of PERK. (A The levels of TGs that were most prominently changed in TB40/E-infected cells relative to uninfected cells at 96 hpi. These are the top six elevated TGs shown in the heatmap of Fig. 3C. Three independent experiments were performed, each with duplicated samples for a total of six data points. Each data point is graphed relative to the levels observed in uninfected NT cells. For comparison, the relative levels at 72 hpi are also shown. *p<0.5, **p<0.01, ***p<0.001. One-way ANOVA, Turkey test. (B) The fatty acid tails for DGs shown in part A were identified by LC-MS/MS. SFA and MUFA VLCFA tails are in bold text.

Figure 5. Loss of PERK leads to an accumulation of DGs with saturated and polyunsaturated fatty acid tails in HCMV-infected cells. (A) The levels of DGs that were most prominently changed in TB40/E-infected PERK-KO cells relative to infected NT cells at 96 hpi. These represent the top six elevated DGs shown in the heatmap of Fig. 3H. Each data point represents a sample from three independent experiments and is graphed relative to the levels observed in uninfected NT cells. For comparison, the relative levels at 72 hpi are also shown. *p<0.5, **p<0.01, ***p<0.001. One-way ANOVA, Turkey test. (B) The fatty acid tails for DGs shown in part A were identified by LC-MS/MS.

Figure 8. HCMV infection elevates phospholipids (PLs), which are reduced by the loss of PERK. (A-B) Relative levels of PLs in TB40/E-infected and uninfected cells at 96 hpi. PLs were measured by LC-MS/MS following normalization by cell number. Each dot in the plot represents the level of a PL in TB40/E-infected cells relative to the its level in uninfected cells. The dash line represents a relative level of 1 (e.g., the level in infected cells is equal to the level in uninfected cells). (C) Relative levels of PLs in uninfected PERK-KO and NT cells. (D-E) Relative levels of PLs in TB40/E-infected PERK-KO and NT cells at 96 hpi. Abbreviation: PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PI, phosphatidylinositol; and PS, phosphatidylserine. MOI=3. N=3.

Figure 4. HCMV infection elevates DG lipids with long-chain and very long-chain polyunsaturated fatty acid tails. (A) The levels of DGs that were most prominently changed in TB40/E-infected cells relative to uninfected cells at 96 hpi. These are the top six elevated DGs shown in the heatmap of Fig. 3C. Three independent experiments were performed, each with duplicated samples for a total of six data points. Each data point is graphed relative to the levels observed in uninfected NT cells. For comparison, the relative levels at 72 hpi are also shown. *p<0.5, **p<0.01, ***p<0.001. One-way ANOVA, Turkey test. (B) The fatty acid tails for DGs shown in part A were identified by LC-MS/MS. Long-chain and very long-chain polyunsaturated tails (LCFA-PUFAs and VLCFA-PUFAs) are in bold text.

Full text

1
Human Cytomegalovirus uses a Host Stress Response to Balance the Elongation
of Saturated/Monounsaturated and Polyunsaturated Very Long Chain Fatty Acids
Yuecheng Xi
a
, Lena Lindenmayer
b
, Ian Kline
a
, Jens von Einem
b
, and John G.
Purdy
a,c,d,#
a
Department of Immunobiology, University of Arizona, Tucson, Arizona, USA
b
Institute of Virology, Ulm University Medical Center, Ulm, Germany
c
BIO5 Institute, University of Arizona, Tucson, Arizona, USA
d
Cancer Biology Interdisciplinary Program, University of Arizona, Tucson, Arizona, USA
Running title: PERK Balances FA Elongation Following HCMV Infection
#
Corresponding Author: John G. Purdy, jgpurdy@arizona.edu
Word counts
Abstract: 256 words
Text: 7,281 words
.CC-BY-NC-ND 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted January 17, 2021. ; https://doi.org/10.1101/2021.01.16.426974doi: bioRxiv preprint

2
ABSTRACT
Stress and virus infection are known to regulate lipid metabolism in cells. Human
cytomegalovirus infection induces fatty acid (FA) elongation and increases the cellular
abundance of lipids with very long-chain FA tails (VLCFAs). While reprogramming of
metabolism can be stress-related, the role of stress in HCMV reprogramming of lipid
metabolism is poorly understood. In this study, we engineered cells to knockout PKR-
like ER kinase (PERK) in the ER stress pathway and measured lipid changes using
lipidomics to determine if PERK is needed for lipid changes associated with HCMV
infection. We found that in HCMV-infected cells, PERK promotes the increase in the
levels of phospholipids with saturated FA (SFA) and monounsaturated FA (MUFA)
VLCFAs tails. Consistent with the SFA/MUFA lipidome changes, PERK enhances the
protein levels of FA elongase 7 (ELOVL7), which elongates SFA and MUFA VLCFAs.
Additionally, we found that increases in the elongation of polyunsaturated fatty acids
(PUFAs) associated with HCMV infection was independent of PERK and that lipids with
PUFA tails accumulated in HCMV-infected PERK knockout cells. Consistent with the
PUFA lipidome changes, the protein levels of ELOVL5, which elongates PUFAs, are
increased by HCMV infection through a PERK-independent mechanism. These
observations show that PERK differentially regulates ELOVL7 and ELOVL5, creating a
balance between the synthesis of lipids with SFA/MUFA tails and PUFA tails.
Additionally, we found that PERK was necessary for virus replication and the infectivity
of released viral progeny. Overall, our findings indicate that PERK—and more broadly,
ER stress—may be necessary for membrane biogenesis needed to generate infectious
HCMV virions.
.CC-BY-NC-ND 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted January 17, 2021. ; https://doi.org/10.1101/2021.01.16.426974doi: bioRxiv preprint

3
IMPORTANCE
HCMV is a common herpesvirus that establishes lifelong persistent infections. While
infection is asymptomatic in most people, HCMV causes life-threatening illnesses in
immunocompromised people, including transplant recipients and cancer patients.
Additionally, HCMV infection is a leading cause of congenital disabilities. HCMV
replication relies on lipid synthesis. Here, we demonstrated that the ER stress mediator,
PERK, controls fatty acid (FA) elongation and cellular abundance of several types of
lipids following HCMV infection. Specifically, we found that PERK promotes FA
elongase 7 synthesis of lipids with saturated/monounsaturated very long-chain FA tails
which are important for building the viral membrane of infectious HCMV virions. Overall,
our study shows that PERK is an essential host factor that supports HCMV replication
and promotes lipidome changes caused by HCMV infection.
.CC-BY-NC-ND 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted January 17, 2021. ; https://doi.org/10.1101/2021.01.16.426974doi: bioRxiv preprint

4
INTRODUCTION
Regulation of metabolism in response to acute and chronic external or internal
stressors allows cells to adapt to stressful situations, including viral infections (1-6).
Human cytomegalovirus (HCMV) reprograms metabolism, in part by inducing stress
responses, to create a metabolic state that supports virus replication (7, 8). The
endoplasmic reticulum (ER) has roles in both lipid metabolism and stress. The
integrated stress response, lipid synthesis, and protein translation and folding take
place in the ER. Disruption of ER homeostasis triggers a stress response that can
include activating PKR-like ER kinase (PERK, also known as eukaryotic translation
initiation factor 2α kinase 3, EIF2AK3). PERK reduces ER-stress related to the
accumulation of misfolded proteins by inactivating EIF2α through phosphorylation to
decrease translation. PERK has an emerging role in lipid metabolism as well. PERK-
mediated lipogenesis supports the proper development of lipid-generating tissues, such
as adipocytes and mammary glands (9). In the context of viral infection, PERK is
needed for HCMV-induced lipid synthesis (7, 10). However, the effects of PERK on
specific classes or types of lipids in HCMV-infected cells remains unknown.
HCMV infection induces several lipogenic enzymes, including those involved in
fatty acid (FA) synthesis and elongation (8, 10-14), which are required for HCMV
replication (10, 12-14). FAs have several functions as free fatty acids or as hydrophobic
tails in lipids, including establishing and maintaining membrane integrity. The function of
FAs depends on their length and the number and placement of double bonds in the
hydrocarbon chain. HCMV infection promotes the synthesis of lipids with very long-
chain fatty acid (VLCFA) tails containing 24 or more carbons with no double bonds (i.e.,
.CC-BY-NC-ND 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted January 17, 2021. ; https://doi.org/10.1101/2021.01.16.426974doi: bioRxiv preprint

5
saturated fatty acids, SFAs) or one double bond (i.e., monounsaturated fatty acids;
MUFAs) (13, 14). In HCMV infected cells, SFAs are made by fatty acid elongase 7
(ELOVL7) (13). Humans encode seven ELOVLs that elongate FAs based on the chain
length and double bond content (15-19). The expression of ELOVLs is regulated by
sterol regulatory-element binding proteins (SREBPs), liver X receptor α (LXRα), and
peroxisome proliferation-activated receptor α (PPARα) (20). However, little is known
concerning mechanisms involved in the differential expression of ELOVLs (i.e., how a
cell may regulate ELOVL7 differently than ELOVL5).
We recently demonstrated that HCMV UL37x1 protein (pUL37x1, also known as
viral inhibitor of apoptosis, vMIA) helps to support HCMV enhancement of SFA
elongation and synthesis of phospholipids (PLs) with VLCFA tails (PL-VLCFAs) (8).
Additionally, pUL37x1 helps to increase PERK protein levels following infection (8). In
this study, we investigated the role of PERK in the biogenesis of different lipids classes
in response to HCMV infection. We engineered PERK knockout (KO) cells using
CRISPR/Cas9 and performed several lipidomic analyses to define lipids regulated by
PERK following HCMV infection. We expanded on our previous PL lipidomic work by
including the analysis of diglycerides (DGs) and triglycerides (TGs). HCMV infection
increases the relative levels of most DGs and TGs with SFA/MUFA VLCFA tails. We
found that PERK affects the levels of PLs, DGs, and TGs, specifically in the double
bond content of their tails. Our findings reveal that PERK balances the ratio of lipids with
SFA/MUFA VLCFA tails and lipids with PUFA tails. Moreover, we show that in HCMV-
infected cells PERK differentially regulates FA elongases (ELOVLs) involved in
SFA/MUFA and PUFA elongation. Our study provides a new perspective on how HCMV
.CC-BY-NC-ND 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted January 17, 2021. ; https://doi.org/10.1101/2021.01.16.426974doi: bioRxiv preprint

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Human cytomegalovirus uses a host stress response to balance the elongation of saturated/monounsaturated and polyunsaturated very long chain fatty acids" ?

In this paper, the authors investigated the role of PERK in the biogenesis of different lipids classes in response to HCMV infection. 

Q2. What is the genome-to-IU ratio in PERK-KO cells?

Since the genome-to-IU ratio was 8- to 36-fold greater in the PERK-KO than control cells, their findings indicate that PERK-KO cells release virus particles with reduced infectivity. 

Q3. How did the authors measure the amount of infectious virus released at 96 and 120 hpi?

The authors used TCID50 to measure the amount of infectious virus released at 96 and 120 hpi by infecting NT and PERK-KO cells using a MOI of 1. 

Q4. What has been shown to increase the concentration of phospholipids?

The authors have previously shown that HCMV increases the concentration of phospholipids, particularly those with VLCFA tails (PLVLCFAs) (8). 

Q5. How many ELOVL7 protein levels are upregulated in HCMV cells?

Protein levels of ELOVL7 were also upregulated by 17-20-fold in HCMV-infected HFF cells compared to uninfected at 72 and 96 hpi, respectively (Fig. 11B-C). 

Q6. What is the effect of ELOVL5 on the synthesis of PUFAs?

This loss in the balance of ELOVL5 and ELOVL7 proteins likely reduces the available pool of SFA/MUFA VLCFAs relative to the pool of PUFA VLCFAs. 

Q7. What is the role of PERK in regulating ELOVLs?

This observatoin could indicate that in addition to regulating ELOVLs, PERK may also promote the flow of DGs to PLs rather than TGs. 

Q8. How many DGs were more abundant in PERK-KO cells at 96 ?

At 96 hpi, the relative abundance of these six DGs were 2- to 6- fold greater in infected PERK-KO cells than infected NT cells and 2.5- to 60-fold more abundant when compared to uninfected cells (Fig. 5A). 

Q9. What is the role of ER-stress in HCMV reprogramming?

While both HCMV infection and ER-stress are known to regulate lipid metabolism (3, 7, 8, 10, 12, 13, 32, 33), the role of ER-stress in HCMV reprogramming of lipid synthesis is poorly defined. 

Q10. How did the authors determine if TB40/E increased the levels of PLs?

Since TB40/E-infection increases the levels of PLs at 96 hpi, the authors examined earlier timepoints to establish if TB40/E alters PLs levels at 48 and 72 hpi. 

Q11. How did the authors determine if HCMV infection alters the levels of phospholipids?

the authors determined if HCMV infection alters the levels of PLs by comparing therelative abundance of PLs in TB40/E-infected cells to uninfected cells at 96 hpi.