Human Cytomegalovirus Use and Manipulation of Host Phospholipids
Item Type text; Electronic Thesis
Authors Harwood, Samuel John
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 27/09/2021 22:10:36
Link to Item http://hdl.handle.net/10150/632563 HUMAN CYTOMEGALOVIRUS USE AND MANIPULATION OF HOST PHOSPHOLIPIDS
by
Samuel Harwood
______
Copyright © Samuel Harwood 2019
A Thesis Submitted to the Faculty of the
DEPARTMENT OF MOLECULAR AND CELLULAR BIOLOGY
In Partial Fulfillment of the Requirements
For the Degree of
MASTER OF SCIENCE
In the Graduate College
THE UNIVERSITY OF ARIZONA
2019 2
THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE As members of the Master's Committee, we certify that we have read the thesis prepared by Samuel Harwood, titled Human Cytomegalovirus Use and Man.!.e.ulationof Host Phos holipids, and recommend that it be accepted as fulfilling the thesis requirement for the Master's Degree.
Z'i Date +/ I Z.OI � .
4/ 1 / 7 Date: 2 r
( Date: � /L Final approval and acceptance of this thesis is contingent upon the candidate's submission of the final copies of the thesis to the Graduate College. I hereby certify that I have read this thesis prepared under my direction and recommend that it be accepted as fulfilling the Master's requirement. 4/ 1 '��v Date: 2 I 2ol � Dr. John Purdy Assistant Professor Department of lmmunobiology 3 Acknowledgments I would like to thank Dr. John Purdy, Lisa Wise, Elizabeth Dahlmann, and Yuecheng Xi for their mentorship and assistance with my research. I would also like to thank Dr. Joyce Schroeder and Dr. Daniela Zarnescu for serving on my thesis committee. 4 Table of Contents List of Figures and Tables……………………………………………………………………………………………………………………….5 Abstract………………………………………………………………………………………………………………………………………………….6 Introduction……………………………………………………………………………………………………………………………………………7 Materials and Methods………………………………………………………………………………………………………………………….20 Results………………………………………………………………………………………………………………………………………………..…27 Discussion……………………………………………………………………………………………………………………………………………..49 References…………………………………………………………………………………………………………………………………………….57 5 List of Figures and Tables Table 1…………………………………………………………………………………………………………………………………………………..13 Figure 1………………………………………………………………………………………………………………………………………………….16 Figure 2………………………………………………………………………………………………………………………………………………….31 Figure 3………………………………………………………………………………………………………………………………………………….34 Figure 4………………………………………………………………………………………………………………………………………………….35 Figure 5………………………………………………………………………………………………………………………………………………….39 Figure 6……………………………………………………………………………………………………………………………………………….…40 Figure 7………………………………………………………………………………………………………………………………………………….44 Figure 8………………………………………………………………………………………………………………………………………………….46 Figure 9………………………………………………………………………………………………………………………………………………….48 6 Abstract Human cytomegalovirus (HCMV) is a widely-spread β-herpesvirus that causes a congenital infection that results in devastating disabilities in newborns. Infection also causes life-threatening disease in people with compromised immune systems. HCMV requires viral remodeling of host cell metabolism to obtain metabolites and lipids to support replication and contains an envelope made of lipids ‘stolen’ from the host. Envelopment is a crucial step in the HCMV life cycle and requires numerous host cell lipids, most prominently phospholipids. I hypothesized that HCMV upregulates the production of cellular phospholipids and uses host cell lipid transport proteins to transport those phospholipids to the viral envelope. I used tandem LC-MS/MS to quantitatively determine HCMV manipulation of the abundance of phospholipids. I found that HCMV significantly upregulates phosphatidylcholine (PC) lipids and identified a required viral protein: pUL37x1. PC lipids are transferred between membranes by phosphatidylcholine transfer protein (PC-TP). I found that HCMV upregulates PC-TP expression during infection, suggesting that it is important to infection. To determine the role of PC-TP in virus replication, I used a small molecule inhibitor and a CRISPR/Cas9 knockout of PC-TP. My preliminary data indicates that the loss of PC-TP activity reduces HCMV infection late in the viral replication cycle, consistent with my hypothesis. Additionally, I analyzed how the loss of PC-TP affects HCMV remodeling of host lipid metabolism. Overall, my findings demonstrate that HCMV upregulates cellular phospholipids and that PC-TP may play a role in transporting those lipids in viral replication 7 Introduction Human cytomegalovirus (HCMV) is a globally ubiquitous β-herpesvirus infecting over 60% of the world population [1] with prevalence varying based on location and socioeconomic status (i.e. <90% of preschool children are infected in the developing world compared to 20% in developed countries) [2]. Generally, HCMV infection is asymptomatic for healthy and immunocompetent children, adults, and newborns. However, HCMV is the most common congenital viral infection and can result in deafness, blindness, liver abnormalities, and severe learning disabilities in children. In the United States, HCMV is the leading cause of birth defects [3]. Infection has also been associated with glioblastoma [4, 5], cardiovascular disease [6], and the deterioration of the immune system [7]. Infection can lead to prolonged drug treatment, hospitalization and, in some cases, death. HCMV is also a life-threatening opportunistic infection in immunocompromised patients, including HIV/AIDS patients and stem- cell/solid-organ transplant recipients [1, 8]. HCMV is typically transmitted by the exchange of bodily fluids [9], including sexual activity [10]. HCMV infection can be limited by widespread sanitation (e.g. washing hands, not sharing food and drink-ware), but this practice is impractical on a large scale [11]. There is no cure or vaccine for HCMV. Like all herpesviruses, once a person is infected, they will harbor the HCMV for life since the immune system fails to clear the virus [12]. HCMV has a double-stranded DNA genome surrounded by a capsid, a proteinaceous layer called the tegument, and an envelope composed of host derived lipids and viral proteins. HCMV encodes at least 150 known proteins which are classified as either immediate early, early, or late. However, the virus may encode an additional 400 genes [13]. With a genome length of approximately 236 kbp, HCMV has the largest genome of any herpesvirus. To suppress HCMV infection, patients are typically give acyclovir-related drugs (nucleoside analogues) [14] that include drugs such as ganciclovir, valganciclovir, foscarnet, and cidofovir [15]. These drugs are all related by their competitive inhibition of the viral DNA polymerase [16, 17] and attempt to prevent viral gene synthesis. Although these drugs have improved 8 outcomes across the globe, particularly in immunocompromised hosts, there are several drawbacks to these therapies, particularly that of the associated toxicity of the drugs [15] (leading to the onset of conditions such as cytopenia [14]) as well as the frequent development of anti-viral resistance. The issues of toxicity and viral resistance are connected. Because of the toxicity of the drugs, only limited doses can be administered to patients. As a result, viral replication is incompletely suppressed, resulting in a higher chance of resistant progeny [16]. Interestingly, one of the most commonly prescribed drug combinations, ganciclovir and valganciclovir (GCV), requires a phosphorylation step catalyzed by the viral kinase UL97 [18], offering two points of resistance: mutation in UL97 (accounting for 90% of GCV resistance cases) [19] or the viral DNA polymerase. This leads GCV therapy to having more cases of anti- viral resistance than any other HCMV therapy. These drawbacks have created a need for novel therapies that are less prone to viral resistance as well as less toxic to the host [15]. HCMV modulation of host cell metabolism is one possible therapeutic target. HCMV manipulation of host cell metabolism likely plays a role HCMV virion assembly, particularly that of envelopment. Assembly starts in the nucleus and continues in a virally generated perinuclear organelle called the viral assembly compartment that contains various host and virus proteins and host membranes [1]. The formation of this compartment has been shown to not only increase the expression of organelle markers but also relocate them to the perinuclear site of virion assembly [20]. HCMV undergoes two distinct and independent envelopment phases: primary envelopment that occurs at the inner nuclear membrane of as the viral capsid exits the nucleus (this envelope is only temporary as it is lost at the outer nuclear membrane), and a final envelopment that likely happens in the assembly complex. During final envelopment, lipids made by host metabolic pathways are used to form an infectious HCMV virion [1]. HCMV upregulates the expression of host proteins, including fatty acid elongase 7 enzyme (ELOVL7), to meet the lipid requirements of infection [21]. ELOVL7 produces saturated very long chain 9 fatty acids that are found in the viral envelope [21] and are required for infection [22]. HCMV has also been shown to upregulate PKR-like endoplasmic reticulum kinase (PERK) to induce lipogenesis [23], and recent work has shown that PERK may be necessary for the synthesis of phospholipids with very long chain fatty acid tails (VLCFAs) (Purdy Lab, unpublished). HCMV infection of human fibroblasts also results in upregulation of key metabolites in glycolysis, the citric acid cycle, as well as de novo pyrimidine biosynthesis [24]. HMCV utilizes multiple mechanisms to remodel host cell metabolism. For example, HCMV upregulates estrogen-related receptor α to increase glucose metabolism [25]. When the receptor was knocked down in HCMV infected cells, levels of glycolysis enzymes including enolase, triosephosphate isomerase, and hexokinase were reduced [25], indicating that HCMV may utilize the receptor to modify cellular metabolism via transcriptional control. To upregulate the flux of the citric acid cycle (TCA), HCMV enhances glutamine uptake and its metabolism to the TCA intermediate alpha-ketoglutarate [26]. HCMV upregulates a diverse set of kinases to modulate metabolism [21, 23, 27-29]. Activation of AMPK induces increased flux through glycolysis while blocking AMPK-mediated inhibition of lipid metabolism [28, 29]. HCMV also utilizes calmodulin dependent kinase kinase to increase glycolytic flux [27]. Upregulation of host cell metabolism may be linked to the activity of certain HCMV viral proteins. Recent work, for example, has suggested that the viral protein UL38, a protein critical for the inhibition of cellular apoptosis [30], indirectly upregulates mTOR activity [31], a protein essential for the regulation of protein synthesis [32]. This upregulation occurs through inhibition of the tumor suppressor TSC2 [31]. When in complex with TSC1, TSC2 inhibits mTOR activity [33], thus limiting protein synthesis. Therefore, HCMV manipulation of host cell metabolism may also be dependent on the activity of certain viral proteins. 10 Upregulation of cellular lipid metabolism, specifically phospholipid metabolism, likely plays a role in HCMV envelopment. HCMV envelopes are composed of several types of phospholipids, with phosphatidylcholine (PC) lipids composing about 37% of the phospholipids. Phosphatidylethanolamines compose 45.2% of the envelope, phosphatidylserines, 8.1%, phosphatidylglycerols, 5.1%, phosphatidylinositols, 2.9%, and phosphatidic acid, 1.7%. [34]. Because of HCMV’s reliance on cellular phospholipids, the synthesis and transport of phospholipids may be a possible drug target to attenuate HCMV replication. All phospholipids consist of a head group bound to a phosphate molecule which is covalently linked to the SN3 position of a glycerol molecule. Two fatty acid molecules of varying chain lengths (i.e. the total number of carbons in the fatty acid hydrocarbon tails) and degree of saturation (i.e. the number of double bonds in the hydrocarbon tails) are bonded to the SN1 and SN2 positions of the glycerol molecule. Both the headgroup and the fatty acid tails impart unique properties onto each phospholipid, making them useful in a variety of cellular functions including signaling, protein recruitment, and membrane structure [35]. Phosphatidic acid, in addition to being a major phospholipid class, acts as an important precursor to the synthesis of the other phospholipids [35] and thus will be the first lipid discussed. PA is the most basic of all phospholipids, consisting of only the phosphate group, glycerol, and fatty acid tails [36]. PA plays an important role in cellular signaling, G-protein regulation [37], vesicular trafficking (including both vesicle fission [38] and fusion [39]), and metabolism [37]. PA synthesis begins with the addition of long chain fatty acids to CoA-esters by acyl-CoA synthetase. PA is then formed by the addition of the acyl-CoA molecules to glycerol-3-phosphate (derived from glycolysis) by glycerophosphate acyltransferases (GPATs). PA is a precursor for both DAG (essential for the synthesis of PC, PE, and, indirectly, PS) and CDP-DAG (essential for the synthesis of Pi and possibly PG). The 11 dephosphorylation of PA to DAG occurs in the cytoplasm via PA phosphatases called LIPINs. DAG can also be phosphorylated back to PA through the action of DAG kinases [35]. PCs are a type of glycerophospholipid consisting of a phosphorylcholine moiety as the headgroup [36] (Figure 1a). In mammalian cells, PCs are essential for the formation of most membranes, lipoproteins, and lipid droplets [40]. PC is synthesized by three pathways: CDP-choline (also called the Kennedy Pathway [41]), lyso-phosphatidylcholine to PC (also called the Lands cycle) [42], and phosphatidylethanolamine to PC [43]. In the Kennedy pathway, choline is phosphorylated by choline kinase to become phosphocholine. Phosphocholine cytidylyltransferase then catalyzes the formation of cytidine-disphosphocholine (CDP-choline). Finally, choline phosphotransferase transfers a phosphocholine group from CDP-choline to DAG lipids forming PC lipids [41]. The Lands cycle involves the addition of a fatty acid chain to lysophosphatidylcholine—a PC lipid that has only a single tail—by lysophosphatidylcholine acyltransferases to form the two-tailed PC lipid [42]. The third pathway involves converting phosphatidylethanolamine lipids to PCs via three successive methyl transfer reactions [43, 44]. This third pathway is oftentimes discussed as occurring only in the liver [43], however it can also take place in the fat [45] and a few other tissues including heart, brain, thymus, and lungs [46, 47]. This pathway produces about 30% of PCs and can be upregulated in the event that choline consumption is reduced [35]. PCs are synthesized in the ER and must be transported to the site of final viral envelopment. Like all other phospholipids, PCs do not spontaneously diffuse across membranes, and PC is primarily transported by membrane imbedded transport proteins [48]. PC can be transferred by several proteins, including phosphatidylinositol transfer protein (PI-TP) alpha and beta [49], phospholipid transfer protein (PL-TP) [50], and phosphatidylcholine transfer protein (PC-TP) [48]. PC-TP is the only one of these proteins that exclusively transfers PCs [48]. 12 PC-TP is a 25 kDa member of the START super family domain (STARD family) of proteins (Table 1) [9], all of which are related by their lipid binding motifs. Mammalian genomes encode 15 members of the STARD family. The START domain is composed of approximately 210 amino acids which forms an α/β helix-grip structure. This structure results in a hydrophobic lipid binding pocket that is common across many lipid binding domains [51]. Despite this similarity among STARD family proteins, the only proteins within this family that can transfer PC are PC-TP, StarD7, and StarD10. PC-TP is classified as a STARD family minimal protein as it consists only of the lipid binding motif. PC-TP interacts with PCs via Agr74 (which binds the negatively charged phosphate) and Trp 101, Tyr 114, and Tyr 155, all of which interact with the positively charged choline group. PC-TP transfers PC one molecule at a time [48] and is found mostly in the nucleus and cytoplasm but is also present in the mitochondria. Because of the known specificity of PC-TP and PCs abundance in the cell, PC-TP (and other STARD proteins) may play a role in HCMV replication by transporting lipids to the viral envelope. 13 Table 1: STARD family proteins and their functions. Notably, in addition to PC-TP, STARD10 can also transfer PCs across membranes. Name Function STAR Metabolism of cholesterol into pregnenolone; transfer of pregnenolone from inner mitochondrial membrane to outer mitochondrial membrane PC-TP Transfer of PC between membranes STARD3 Various, but mainly involved in cholesterol transport STARD4 Intracellular transport of sterols and lipids STARD5 Intracellular transport of sterols and lipids STARD6 Intracellular transport of sterols and lipids STARD7 Prevents allergic reactions in the mucosal tissue STARD8 Promotes GTPase activity of RHOA and CDC44’ stimulates PLCD1 STARD9 Stabilizes the pericentriolar material during spindle pole assembly STARD10 Phospholipid transport protein able to transport PC and phosphatidylethanolamine COL4A3BP Intracellular trafficking of ceramides and diacylglycerol lipids DLC1 GTPase activating protein (GAP) of various proteins STARD13 GTPase activating protein (GAP) of various proteins ACOT11 Metabolism of medium and long chain fatty acids ACOT12 Hydrolyzes acetyl-CoA to acetate and CoA 14 PE’s are another type of glycerophospholipid abundant in the cell and viral envelope. PE’s head group is phosphoethanolamine and therefore does not include the three methyl groups on the terminal amine group found in PC [36] (Figure 1a). This structural similarity, however, allows for rapid conversion of PE to PC through the addition of the three methyl groups via phosphatidylethanolamine N- methyltransferase [43, 44]. PE takes on a conical shape, largely due to the difference in volume between its relatively small headgroup and long fatty acid tails. This shape causes this lipid to induce negative curvature in membranes, enhancing membrane fusion and folding of membrane proteins [35]. PE is synthesized via the CDP-ethanolamine pathway which is analogous to the Kennedy pathway in PC synthesis [35]. In this pathway, ethanolamine is imported into the cell via CTL1 and phosphorylated by either ETNK1 or ETNK2 [52]. CTP:phosphoethanolamine cytidylytransferase then catalyzes the formation of CDP-ethanolamine from CTP and phosphoethanolamine. Finally, choline/ethanolamine phosphotransferase transfers a phosphoethanolamine molecule to DAG, forming phosphatidylethanolamine. As the name suggests, this enzyme can also catalyze the same reaction in PC synthesis. PE is also synthesized by a second pathway called the phosphatidylserine decarboxylase pathways. PS is imported into the mitochondria almost immediately after synthesis via a contact site between the ER and mitochondria called the MAM [53]. In yeast, this site is formed by the ER/Mitochondrial encounter structure (ERMES), however this has been disputed in mammalian cells as mammals do not have the same structural proteins that make up the yeast ERMES structure [54]. Despite that, this complex has been proposed to be required for the PS to PE conversion [55]. PS is then decarboxylated by phatidylserine decarboxylase to form PE [53]. PE intracellular transport is much less well defined than PC transport within the cell, but it is speculated that PE can be transported by both PI- T and PL-TP [48], [56]. PS lipids contain a phosphoserine head group, and is identical to that of PE with the exception that PS includes a carboxylic acid group on the terminal carbon [36] (Figure 1a). The carboxylic acid 15 group, at physiologic pH, carries a negative charge, imparting that charge to whatever membrane in which it’s present. This feature makes PS an important recognition site for a variety of membrane proteins [35]. PS is synthesized by the serine base exchange pathway which involves either PS synthase 1 or PS synthase 2 catalyzing the exchange of choline with serine on PC and ethanolamine with serine on PE, respectively. This reaction occurs at the MAM site and is a calcium dependent reaction [53]. Interestingly, despite synthesis and transport at the MAM site, PS is highly localized to the plasma membrane (specifically the cytosolic side) when compared to the endoplasmic reticulum and mitochondria [35]. Like PE, PS transport within the cell, occurs primarily by the non-specific transfer protein PL-TP [56]. PS has also been known to be exchanged at membrane contact points, most prominently at the MAM site during PS to PE synthesis. Because PS synthesis primarily occurs at the MAM site, PS transport into the mitochondria is thought to be coupled to PS synthesis [54]. Because of their structural similarity, PC, PE, and PS are all intimately connected via several synthesis pathways. PS is synthesized directly from PC or PE through the PS synthase enzymes and then can be imported into the mitochondria to be synthesized back to PE through the phatidylserine decarboxylase pathway. PE can then be exported back out to the endoplasmic reticulum and synthesized into PC via the methyltransferase pathway [35] (Figure 1b). These synthesis pathways, in turn, highlight the importance of both the ER and mitochondria in phospholipid production and their potential importance in HCMV infection. 16 A B ER Mitochondria PC, PE PSS1/2 PS PS PSD Other organelles PE PE PEMT PC PC PC Figure 1: Structure and synthesis of PC, PE, and PS lipids. (A) Structure of the choline head group (left), ethanolamine head group (middle), and serine head group (right). Note that these head groups are shown without the glycerol-3-phosphate molecule. Images were sourced from PubChem [36]. (B) Diagram showing the synthesis pathways and spatial relationship between PC, PE, and PS. Diagram adapted from Neale Ridgway [35]. 17 Phosphatidylinositol (PI) is a glycerophospholipid with phosphoinositol as it’s a headgroup. Inositol consists of cyclohexane group with hydroxyl groups bonded to each carbon [36]. PI is unique in that it is often phosphorylated at the 3, 4, or 5 position on the inositol head group, resulting in seven different PI species, each with diverse signaling activity. The phosphorylation pattern on the headgroup dictates what proteins bind to PI in membranes, making PI an important player in membrane protein recruitment [35]. PI synthesis is dependent on the synthesis of CDP-DAG, which is synthesized by CDP- DAG synthase (CDS) 1 or 2. CDP-DAG is then combined with inositol by phosphatidylinositol synthase (PIS) to produce PI. Both PIS and CDS1 are integral membrane proteins in the ER, again highlighting the importance of the ER in mammalian phospholipid synthesis [57]. PI, like PC, is primarily transported via intermembrane transport proteins, in this case, by PI-TP. PI-TP differs from PC-TP in that it has the ability to transfer both PC and PI between membranes. Mammalian cells have two isoforms of PI-TP: PI-TPα and PI-TPβ [54], both of which have a 16 fold affinity for PI over PC [58]. Phosphatidylglycerol (PG) has a phosphoglycerol head group [36]. PG is one of the least abundant phospholipids in the cell, composing only about 1% of all lipids. However, it has important physiological functions, the most prominent being its role as the precursor for cardiolipin, a lipid with roles in mitochondrial dynamics and apoptosis. PG is synthesized in the mitochondria and, like PI synthesis, requires CDP-DAG as precursor. Unlike PI, however, CDP-DAG in this pathway is synthesized by Tam41 which utilizes imported PA as a precursor. CDP-DAG is then combined with glycerol-3- phosphate by phosphatidylglycerol phosphate synthase to form phosphatidylglycerol phosphate (PGP). PGP is then dephosphorylated by PTPM1 to form PG [35]. PG is primarily transported by the non-specific PL-TP [48]. My thesis takes a broad look at HCMV’s use and manipulation of cellular phospholipids. In order to look at the potential mechanism behind HCMV’s manipulation of cellular phospholipids, I also investigated the role of the HCMV protein pUL37x1. UL37x1 is an immediate early gene and is one of the 18 first genes to be expressed during infection [1]. Classically, pUL37x1 has been linked to inhibition of apoptosis by localizing to the mitochondria [59] and is thus also known as vMIA (viral mitochondria- localized inhibitor of apoptosis). vMIA has been shown to localize to the mitochondria and bind GADD45 protein family members [60], proteins involved in DNA damage and stress response [61], and the BCL-2 family proteins Bcl-xL and Bax. This complex prevents the release of cytochrome C from the mitochondria, ultimately preventing the activity of caspasae-3 and cellular apoptosis [62]. pUL37x1 has been also been known to induce Ca2+ release in the ER. During the early stages of infection, pUL37x1 localizes to the ER and colocalizes with ER proteins and calcium stores, resulting in a regulated release of Ca2+. Interestingly, when infected cells underwent drug-induced Ca2 release, infection was halted, demonstrating that pUL37x1 promotes a regulated release of calcium that is not mimicked by other calcium releasing drugs [63]. This calcium release was also shown to be necessary and sufficient for the early stages of HCMV cytopathology such as cellular swelling, rounding, and membrane re-organization [63, 64], and it may be linked to pUL37x1’s anti-apoptotic activities [63]. Based on pUL37x1 control of Ca2+ flux and its localization to both the ER and mitochondria, the Purdy Lab hypothesized that pUL37x1 may play a role in mediating HCMV induced lipid metabolism (Purdy Lab, unpublished). Previous work has shown that HCMV mediates aspects of its metabolism through Ca2+ release, including activation of AMPK through Ca2+-calmodulin-dependent kinase kinase activity [27]. Recent work from the Purdy Lab has shown that pUL37x1 is required for the efficient replication of HCMV, its elongation of fatty acids through ELOVL7, and the upregulation of metabolic enzymes such as PERK and ACC-1. My Master’s Thesis research project tests the overall hypothesis that HCMV upregulates the production of cellular phospholipids and uses host cell lipid transport proteins in order to transport those phospholipids to the viral envelope. I first test the sub-hypothesis that HCMV utilizes pUL37x1 to upregulate all cellular phospholipids, specifically VLCFAs. I used tandem HPLC MS/MS to analyze lipids in 19 infected cells, and I found that HCMV, through pUL37x1, primarily upregulates PC lipids and downregulates PS lipids. Based on those results, I hypothesized that HCMV utilizes PC-TP in order to transport PC lipid to its envelope, thereby making PC-TP essential for HCMV replication. To determine the role of PC-TP, I blocked PC-TP activity using a known small molecule inhibitor and a CRISPR/Cas9 knockout and analyzed any lipid changes through HPLC tandem MS/MS. My results demonstrated that PC-TP was upregulated during infection and required for HCMV replication depending on the infection conditions. Overall, my results show that HCMV upregulates cellular phospholipids and that PC-TP may play a role in viral replication, potentially by transporting host cell phospholipids to the viral envelope. 20 Materials and Methods Infection Conditions Most experiments were performed in human foreskin fibroblast cells (HFF cells) in serum-free DMEM. Cells were initially grown to confluency in DMEM containing 10% FBS. Once the cells had been held for four days at confluency, the media was switched from 10% FBS treated DMEM to SF DMEM [21, 24, 29]. SF media removes exogenous metabolites and lipids that may confound experiments [24]. In some treatments, SF media resulted in increased, in which cases cells are infected at 80% confluency in 10% FBS DMEM. HCMV AD169-GFP I used the lab-adapted HCMV strain AD169-GFP throughout all my experiments. As a lab- adapted strain, AD169-GFP strongly replicates in fibroblast cells. It lacks the ULb’ genomic region [65], blocking its ability to replicate in epithelial cells and latently infect cells [66]. The virus strain used expresses GFP from an exogenous SV40 promoter, providing a tool to visualize infected cells. To generate the virus, HFF cells were grown in 15 cm plates. At 85 % confluency the plates were transferred to roller bottles. Cells were grown to 90% confluency and then infected with AD169-GFP HCMV at an MOI of 0.1 in 10% fetal bovine serum (FBS) treated Dulbecco’s Modified Eagle’s Medium (DMEM). The virus used for this infection was generated by electroporating a BAC (bacteria artificial chromosome) containing a sequenced AD169-GFP genome into fibroblasts, ensuring that little to no o genetic changes have occurred [67]. Each roller bottle was incubated at 37 C at 5% CO2. After 14 days, or when the cells showed 100 % cytopathic effect (CPE), the media was removed and centrifuged at 1,000 x g for 10 minutes at room temperature (RT) to remove cells and dense cellular debris. Viruses were concentrated by ultracentrifugation through a 20% sorbitol solution to further remove cellular 21 debris. Concentrated viral stocks were stored at -80o C in serum free (SF) DMEM. A sample of the virus was tittered as described below to determine infectious viral titer. Infection and infectious viral titers Infection The following protocol was used for all infectious viral titer experiments. HFF cells were plated in a 12 well plate at a density of 0.8 x 105 cells/ml. Cells were infected at either 80% confluency in FBS DMEM or 100 % confluency in SF DMEM as described above. The infectious dose was calculated by multiplying the number of cells by the experiment’s multiplicity of infection (MOI). AD169-GFP was used for all viral infections unless otherwise noted. Cells were infected with appropriate amount of virus at a o volume of 300 µl in either FBS or SF DMEM, incubated at 37 C in 5% CO2 for 1 hour, and agitated every 5-10 minutes. Cells were then removed from the incubator and the infectious inoculum was removed. The cells were washed with phosphate buffered saline (PBS) two times to remove any unbound virus o and 1 ml of the appropriate media was added to the cells. The cells were incubated at 37 C in 5% CO2. The media was changed at 48 hours post infection (hpi). At 96 hpi and 120 hpi, 400 µl of media was removed and stored for future analysis to determine the quantity of infectious virus produced by the cells using the tissue culture infectious dose 50 (TCID50). Infectious viral titer I used TCID50 to quantitatively measure infectious virus particles. Each sample of infectious media was diluted from 10-1 to 10-8. Each dilution was plated into two 96 well plates containing HFF cells. After 14 days, each well was examined for infectious plaques using the GFP reporter cloned into AD169-GFP. The resulting score was used to determine the infectious titer of each sample. 22 Lipidomics HFF cells were plated in 6 well plates at a density of 2 x 105 cells per well. Cells were infected at a MOI of 3 infectious units (IU)/cell in a volume of 500 µl using the protocol described below. Lipids were extracted at 72 hpi. For lipid extraction, cells were washed three times with PBS prior to the addition of cold 50% methanol containing 0.05 M HCL. The 50% methanol rapidly quenches the metabolic activity of the cells. The cells were transferred to a glass vials and lipids were extracted twice using chloroform. The chloroform was removed via nitrogen evaporation. The lipids were resuspended in a 1:1:1 solution of chloroform, methanol, and isopropanol. All samples were normalized by total cell number. Lipids were identified and quantified via liquid-chromatography high-resolution tandem mass spectrometry (LC- MS/MS) [22, 68]. Lipids were separated based on their hydrophobicity during the LC step and were identified and quantified by mass spectrometry. A gradient starting with 25% buffer A (60 % acetonitrile, 1% methanol, 39% water, and 0.1% formic acid) and ending with 100% Buffer B (90% isopropyl alcohol, 9% acetonitrile, 1% methanol, 0.1% formic acid, 10 µM ammonium formate) at a flow rate of 0.25 ml/min was used for LC. Each run lasted 30 minutes, and the C18 column was washed between each run. Following the LC separation, lipids were ionized and analyzed using a Q Exactive Plus orbitrap mass spectrometer (Thermo). MS1 data was collected at 140,000 resolution (at 200 m/z) and spectra ranged from 200-1600 m/z. Compounds were identified through MS1 data based on their mass and retention time using the MAVEN: Metabolomic Analysis and Visualization Engine software [69, 70]. MS1 analysis allowed preliminary determination of the head group and carbon number for each lipid. 23 MS2 analysis was performed using ThermoFisher Xcalibur software. MS2 analysis was used to both confirm the head group of each lipid as well as to determine the identity of the lipid’s fatty acid tails using characteristic fragmentation patterns. Depending on the chemical nature of the lipid’s head group, each phospholipid was identified using either positive or negative ions. MS2 fragmentation allows for identification of lipid tails. Fatty acids from glycerophospholipids were fragmented from the parent lipid using collision-induced dissociation. Because of their carboxylic acid functional group these molecules easily generate negative ions, as such, FA tails were identified in negative mode mass [M-H]- [71]. PC lipids were primarily analyzed in positive mode using an [M+H] mass to identify the parent. Positive mode was used to identify the phosphocholine moiety present in the PC headgroup. The fragment measured 184.0732 m/z and is the characteristic fragment used to identify PC lipids. PC lipids presented an interesting case in negative mode analysis because their inherent positive charge. The LC conditions contain formic acid which neutralizes the positive charge allowing PCs to be detected in negative after the phosphate becomes negatively charged during electrospray ionization. These PC adducts were identified at a mass of [M+44.9908]- [72]. PEs were identified both in positive and negative mode. In positive mode, the most abundant ion used to identify PEs was identified at [M-H-141.018]. This corresponded to the mass of the parent with a loss of the phosphoethanolamine headgroup [72]. In negative mode, the fragments of mass 141.018 m/z (corresponding to the head group) and 196.038 m/z (corresponding to a prominent intermediate produced in the fragmentation of the head group) were used for identification [73]. PS lipids were identified exclusively in negative mode. PS lipids were identified by the neutral loss of the serine head group corresponding to the loss of a mass equal to 87.032 da. PI lipids were 24 identified in negative mode via the fragment 241.012 which corresponded to a cyclic anion of phosphoinostiol [72]. PG lipids were identified in negative mode using three requirements. The first was identification of a glycerol group 152.9960 m/z. Since this fragment is common to all glycerophospholipid [73] I used additional information to confirm PGs. Secondly, fatty acid tails were identified that corresponded to the tails predicted by the PGs mass identified by MS1. Finally, the molecule’s mass was cross-referenced to a publicly available database managed by LIPID MAPS partnership [74, 75]. PGs were reported only if they passed all three verification steps. LDN193188 LDN193188 (Cayman Chemicals) stocks were made in DMSO and stored at -20o C in single use aliquots [76]. If the cells were infected, LDN193188 was added at 1 hpi and replaced at 48 HPI. CRISPR/Cas9 CRISPR/Cas9 [77] was used to functionally knockout the PC-TP gene. Three independent guide RNAs (gRNAs) were used to target PC-TP. The gRNA oligos were cloned into a LentiCRISPR plasmid [78] and confirmed by sequencing. The plasmids were transfected into 293 T cells and the resulting media that contained gRNA/Cas9 lentiviral pseudo-particles was used to treat the target human foreskin fibroblast (HFF) cells. Since the gRNA/Cas9 plasmid also contained the puromycin resistance gene, I selected successfully transfected cells using two rounds of puromycin treatment. Next, I performed single cell cloning to isolate clones that contain homozygous knockouts of PC-TP. This was done by diluting the cells and transferring them to 96 well plates so that one well will contain one cell. To support their growth, 100 25 wildtype HFF cells were added to each well. Over the next six weeks, the wells were grown and treated with puromycin to remove the wildtype HFF feeder cells. Once the wells reached sufficient confluency (>90%), the potential PC-TP knockout cells were screened for loss of PC-TP. I verified clones using DNA sequencing. Four clones were chosen for sequencing based on their ease of growth. I isolated the DNA, amplified the DNA via PCR and verified PCR efficiency via a DNA gel. I then purified the PCR product and used colony screening to sub-clone the DNA population to verify homogenous knockouts. Some colony plasmids appeared to have re-ligated without DNA, so no DNA was sequenced. To prevent this, the colony screening was repeated on ampicillin/X-Gal plates, allowing me to visualize if the DNA was successfully cloned in each colony. I selected one PC-TP knockout clone, C10B, for subsequent studies. Cytotoxicity Assay Cells were grown in 96 well plates to either 80% confluency or 100% confluency. The 80% confluent cells remained in FBS growth media while the 100% confluent cells were switched to SF media at day 4 of confluency. At day 1 of 80% confluency or day 5 of 100% confluency, three wells for each condition were treated with one of five concentrations of LDN193188: 100 µM, 50 µM, 25 µM, 10 µM, and 1µM. For all LDN concentrations tested, the final DMSO concentration was 0.2 %. Cytotoxicity was measured using the CellTiter96 AQueous One Solution Cell Proliferation Assay (Promega). Cytotoxicity was measured at 48 and 72 hours post treatment (hpt). Western Blot PC-TP Cells were lysed in a SDS lysis buffer and immediately kept cold until used for Western Blot. The lysate was boiled at 95o C for ten minutes and then vortexed and centrifuged for 2 minutes at 21.2 g. 26 Proteins were separated on a Bio-Rad Any KD SDS-PAGE gel. The proteins were transferred to a nitrocellulose membrane at 4 oC. The membrane was briefly washed with tris-buffered saline with Tween-20 (TBS-T) and then blocked overnight in 5% bovine serum albumin (BSA) diluted in TBS-T at 4 oC with gentle agitation. The membrane was incubated with Santa Cruz Biotechnology anti-PC-TP mouse monoclonal antibody at a 1:200 dilution in 5% BSA overnight at 4 oC with gentle agitation. A florescent anti-mouse secondary antibody (1:10,000 dilution in 5% BSA) was used for imaging on a Li-COR system. Actin was blotted using a LI-COR primary antibody at a 1:2000 dilution in 3% BSA in TBS-T for 1 hour at RT with gentle agitation and then incubated with an anti-rabbit secondary antibody at a 1:10,000 dilution in 3% BSA for 1 hour at RT with gentle agitation. Viral proteins Western blots were done as described above. The nitrocellulose membrane was incubated with primary antibodies for 1 hour at RT. The following antibodies and conditions were used: • Anti-IE1 at 1:100 in 5% BSA • Anti-pUL44 at 1:2000 in 3% BSA • Anti-pUL99 at 1:100 in 1% BSA 27 Results Lytic replication of HCMV requires remodeling of host metabolism, including lipid metabolism [21-26, 79, 80]. The HCMV viral envelope is composed of host cell derived phospholipids, with PC and PE lipids being the most abundant [34]. Because phospholipids do not diffuse across membranes spontaneously, HCMV may use existing host cell mechanisms to transport lipids from their site of synthesis to the HCMV envelope. I hypothesize that that HCMV upregulates the production of cellular phospholipids and uses host cell lipid transport proteins in order to transport those phospholipids to the viral envelope. I first investigated whether HCMV infection results in the upregulation of host cell phospholipids. My results showed that WT HCMV infection substantially upregulated the production of PC lipids, specifically those with tails that have 40 or more total carbons (Figure 2a, Figure 2c). PC lipids were the most heavily upregulated and the most abundant. PC 46:1, for example was increased between 270 and 1,400-fold. Other classes of phospholipids, including PE, PS, PI, and PG lipids, were also upregulated, although not to same extent as PC lipids with 40 or more total carbons. HCMV infection also sporadically downregulated some lipids, including LPI 20:3 and PC 34:0. Next, I examined possible mechanisms used by HCMV to remodel lipid metabolism. While there is understanding of some host cell factors the HCMV uses to manipulate metabolism, little is known about the role viral factors in the hijacking of host cell metabolism. The viral protein pUL37x1 has been implicated in Ca2 release from the endoplasmic reticulum [63, 64] and therefore may also play a role in Ca2+ dependent metabolic reactions. Furthermore, pUL37x1 localizes to the metabolically active membranes of the mitochondria [63], indicating it may play a role in manipulating host cell metabolism. Recent work from the Purdy Lab showed that pUL37x1 was required for efficient HCMV replication and elongation of VLCFAs (Purdy Lab, unpublished). Therefore, I analyzed whether pUL37x1 has a role in 28 HCMV upregulation of phospholipids. I performed the same analysis on HFF cells infected with subUL37x1 HCMV. This mutant virus contains a deletion of UL37 exon 1. I found that PC lipids were greatly downregulated when compared to WT HCMV infection. However, the PC lipids were still elevated above mock infected cells, indicating that pUL37x1 may not be the only viral protein utilized in hijacking host cell metabolism. (Figure 2b, Figure 2c). Furthermore, I found that PS lipids were slightly upregulated (up to 4-fold) in the mutant virus, indicating that pUL37x1 may downregulate their production during infection (Figure 2b, 2c). Additionally, I analyzed the corresponding MS2 data to identify the tail lengths of the PC and PS lipids modified by HCMV. The six most upregulated PC lipids consisted of total carbon numbers of 44 or greater, indicating that they may contain VLCFAs, lipid tails with carbon lengths greater than 26 and identified in previous studies as essential to HCMV replication [21]. While a previous analysis of PC lipids found that PCs with long chain fatty acid tails were increased during infection [34], the analysis only examined lipids with 30-42 carbons in their tail, making it unlikely that those lipids contained VLCFAs. Through MS2 analysis, I identified VLCFAs in the six most upregulated PC lipids (Figure 3a). I found a large amount of diversity in tail lengths and degrees of saturation, with PC (44:1) and PC (46:2) having the largest amount of diversity with five tail combinations each (Figure 3a). Together, this data suggests that pUL37x1 mediates HCMV remodeling of the cellular lipidome and that PC lipids (especially those with VLCFAs) are particularly upregulated during HCMV infection. Many lipids are isomeric; that is, they have the same mass. For example, PCs with odd-chain FA tails will have the same mass as PEs with even-chain tails. Although, humans only synthesize even-chain FAs, odd-chain FAs are present in the diet. Importantly, odd-chained FAs are present in bovine serum and may be present in HFF cells. To address this possibility, I used LIPID MAPS [74, 75] to identify odd chain PC lipids that may have been misidentified as PE lipids. This protocol was used when two PE peaks of the same mass were found at different retention times. Unfortunately, most of these lipids were not 29 selected for MS2 fragmentation at both retention times given the current lipidomic method. However, I do have MS2 data for one such lipid, PC 45:2, which was identified as a possible odd chain PC lipid. PC 45:2 was upregulated 17-fold in infected cells compared to mock-uninfected cells. I also found that it was upregulated 5.5-fold in WT HCMV infected cells compared to subUL37x1 HCMV infected cells (Figure 3c). I identified it as a PC using a fragment for its head group in positive mode (Figure 3d). Unfortunately, I do not have any MS2 data for this lipid in negative mode, so at this point I am unable to confirm the identity of its FA tails. Therefore, my conclusions are limited, but the preliminary data suggest that HCMV infection maybe altering lipids with odd-chain tails. Further investigation is needed to confirm this observation and to understand its biological implications for infection. I also identified the potential upregulation of plasmalogens and ether phospholipids. These are glycerophospholipids that contain a vinyl either or an alkyl ether (respectively) at the SN1 position (as opposed to a carboxyl group) [72] (Figure 4a). The biochemistry of these lipids gives them unique functions in the cell, including maintaining membrane structure and rigidity and acting as a precursor for second messengers [81, 82]. Therefore, these lipids may play a role in the both the composition of the viral envelope as well as manipulation of host cell signaling. I found that several classes of either ethers (denoted by a “O-“) or plasmalogens (denoted by a “P-“) were upregulated, most of which contained total carbon numbers above 44, with the exception being O-28:1/P-28:0. I also found that several lipids were downregulated during infection, however, they were significantly outnumbered by the number of upregulated lipids (Figure 4b). Unfortunately, those lipids were not consistently identified in the MS2 analysis across three independent experiments, so I could not collect data on their tails or verify their head group identity. Furthermore, I could not find any literature describing how to distinguish between ethers and plasmalogens in an MS2 analysis, a limitation likely due to the structural similarity of the two molecules (Figure 3a). Thus, each lipid is identified by both its carbon and degree of saturation as either 30 an ether or plasmalogen. Therefore, while my data cannot draw any conclusions on the role of these lipids in HCMV infection, the lipids may warrant further investigation in the future. 31 A Fold Change (relative to Mock) 1/8X 8X B Fold Change (relative to Mock) 1/8X 8X 32 C 33 Figure 2: HCMV increases phospholipid levels through pUL37x1. (A) Relative quantities of phospholipids in infected cells as compared to mock infected cells. HFF cells were infected with HCMV AD169 at a MOI of three. Lipids were identified and quantified using high resolution LC-MS/MS as described. (B) Relative quantities of phospholipids in HCMV subUL37x1 infected cells as compared to WT HCMV infected cells. Cells were infected and lipids were analyzed as described in Figure 2a. (C) Abundance of PC and PS lipids with VLCFAs in mock infected cells, WT HCMV infected cells, and subUL37x1 HCMV infected cells. Cells were infected and lipids were analyzed as described in Figure 2a (Figure prepared by Dr. John Purdy). 34 A Tail Information PC(44:0) C16:0+C28:0 PC(44:1) C18:1+C26:0 / C18:0ˆ+C26:1 / C16:0+C28:1 / C14:0ˆ+C30:1 PC(46:1) C18:1+C28:0 / C18:0ˆ+C28:1 PC(48:4) C22:4+C26:0ˆ / C22:3+C26:1ˆ /C20:3+C28:1 PC(48:5) C22:5+C26:0 / C20:4+C28:1 PC(48:7) C22:6+C26:1 ˆ unobserved calculated tail Tail Information PS(32:0) C16:0+C16:0 / C14:0ˆ+C18:0 PS(34:2) C16:1+C18:0ˆ / C16:1+C18:1 PS(36:2) C18:0+C18:2 / C18:1+C18:1 / C16:0+C20:2 PS(36:3) C18:1+C18:2 / C16:0+C20:3 PS(36:4) C16:0+C20:4 PS(38:4) C18:0+C20:4 / C18:1+C20:3 / C16:0+C22:4ˆ ˆ unobserved calculated tail B 35 C Relative Abundance of PC 45:2 100 90 80 70 60 50 40 30 Relative Abundance 20 10 0 WT subUL37x1 Mock Infection Condition D Figure 3: MS2 analysis identifies structural details of HCMV upregulated lipids. (A) Representative PC and PS lipids and their tails as revealed through MS2 analysis (figure prepared by Dr. John Purdy). (B) Representative MS2 spectra of PC 44:1 showing the fragments used to identify fatty acid tails. (C) Relative abundance of PC 45:2 in WT infected cells, subUL37x1 infected cells, and mock infected cells. (D) Positive mode spectra of PC 45:2 showing the presence of the PC head group. 36 A R2 R2 Sn2 X Sn2 X R1 Sn3 R1 Sn1 Sn1 Sn3 B Fold Change (relative to Mock) 1/8X 8X Figure 4: HCMV manipulates host cell plasmalogen and ether abundance. (A) Structure of an ether phospholipid (left) and plasmalogen phospholipid (right). X denotes the phospholipid head group. Structures were designed on MolView (B) Relative quantities of ethers (denoted by “O-“) and plasmalogens (denoted by “P-“) in infected cells as compared to uninfected cells. Cells were infected and lipids were analyzed as described in Figure 2a. 37 PC lipids are a major component of the viral envelope and are likely related to HCMV infectivity [34]. My data showing that PC lipids were greatly upregulated by HCMV infection therefore led me to focus on those lipids exclusively. Recent lipidomic studies in the Purdy lab substantiated the data shown in Figure 2a and showed a broader range of PC lipids increased by HCMV at 72 hpi in primary human fibroblasts (Figure 5a). As PC lipids are an important constituent in the viral envelope, they must be properly localized from their site of synthesis to the location where final envelopment occurs. Most PC lipids are transferred by PC lipid transport protein (PC-TP). Thus, I tested the hypothesis that PC-TP is required for HCMV infection. First, I determined if PC-TP was altered by HCMV, specifically if infection altered the intracellular protein concentration of PC-TP, I used immunoblotting to determine if HCMV upregulates PC-TP. I found that the PC-TP protein level is increased in HCMV infected HFF cells starting at 24 hpi and peaking at 72 hpi and 96 hpi (Figure 5b). Importantly, PC-TP levels are highest late in infection at times when final envelopment is occurring. Next, I sought to determine if PC-TP is required for viral replication by blocking its activity using a pharmacological inhibitor [76]. Infected cells were treated with the small molecule inhibitor LDN193188 which has been demonstrated to inhibit PC-TP [76]. I infected HFF cells with HCMV strain AD169-GFP at 100% confluency in SF DMEM media at an MOI of three. Immediately following the infection, I treated the cells with LDN193188 at six concentrations: 100 µM, 50 µM, 25 µM, 10 µM, 1 µM and performed an infectious viral titer at 96 hpi and 120 hpi. Virus produced by LDN193188 and DMSO- control treated cells was measured by TCID50. LDN193188 reduced HCMV replication at concentrations of 100 µM, 50 µM, 25 µM, 10 µM compared to DMSO-treated control cells (Figure 6a). I tested LDN193188 effects on cytotoxicity using a cell viability assay as described above. LDN193188 was non- 38 cytotoxic under all the conditions tested, expect when at 100µM (Figure 6b). The cytotoxicity of LDN193188 at 100µM was rescued when the cells were provided FBS, suggesting that lipids or some other component of serum provides a product of PC-TP activity that is necessary for cell survival. Since replication is blocked at non-cytotoxic levels of LDN193188 treatment, I conclude that PC-TP activity is required for HCMV replication. These results show that PC-TP may have a role in HCMV infection, however, they may be confounded by off-target effects of LDN193188. To further understand the role of PC-TP, I performed a lipidomic analysis on HFF cells infected at an MOI of three treated either with 100 µM LDN193188 or 0.2% DMSO. I found there was a decrease in several classes of lipids, most notably PCs (Figure 6c). 39 A Fold Change (relative to Mock) 1/8X 8X B H.P.I 4 24 48 72 96 PC-TP Actin Figure 5: HCMV increases PC-TP protein and PC lipid levels. (A) Relative quantities of PC lipids in infected cells as compared to mock infected cells. HFF cells were infected with HCMV AD169 at an MOI of three. Lipids were identified and quantified using high resolution LC-MS/MS as described. (B) Western blot showing relative quantity of PC-TP protein over 96 hours post infection. HFF cells were infected with AD169-GFP at an MOI of 3. Protein was prepared and analyzed as described. Actin is shown as a control. 40 A - LDN TCID50 200000 180000 160000 140000 120000 100000 80000 TCID50/ml 60000 40000 20000 0 DMSO 1 10 25 50 100 LDN Concentration (uM) 96 HPI 120 HPI B LDN Treated Cell Viability 120 100 80 60 40 Cell Viability (%) 20 0 DMSO DMEM 1 10 25 50 100 LDN Concentration (uM) 80% FBS 100% SF 41 C Fold Change (relative to Mock) 1/8X 8X Figure 6: LDN193188 reduces viral replication and PC synthesis. (A) Median infectivity of HFF cells infected with AD169-GFP at an MOI of three and treated with various concentrations of LDN19388 and a DMSO control. (B) Cytotoxicity of HFF cells at various concentrations of LDN19388 and DMEM and DMSO controls as measured by cell viability. (C) Relative change of PC and LPC lipids in HFF cells infected with AD169-GFP at an MOI of three treated with 100 µM of LDN19388 compared to infected cells treated with a DMSO control. 42 To confirm my initial conclusion that PC-TP supports HCMV replication, I performed a CRISPR/Cas9 knockout of PC-TP. I decided to use a genetic approach to test PC-TP to limit possible off- target and cytotoxic effects that may be caused by LDN19388. Briefly, gRNAs targeting PC-TP were used to generate multiple PC-TP knockout clones. I selected clone C10B for further study. C10B had a 31- nucleotide deletion in exon 1 of PC-TP (Figure 7a), causing a frameshift mutation in exon 1 and a premature stop codon in exon 2 (Figure 7b). Loss of PC-TP protein was confirmed by western blot (Figure 7c). I tested the ability of HCMV to replicate in the knockout cells. Both C10B knockout cells and CRISPR non-targeting (NT) control cells were infected in the presence of FBS. I found that that loss of PC- TP resulted in a lower level of viral replication at 120 hpi, but not 96 hpi, compared to control NT cells. (Figure 7d). HCMV lytic replication can be divided into three stages: immediate-early (viral gene expression that occurs following viral entry), early (genes expressed by immediate-early proteins and prior to viral DNA replication), and late (steps that require viral DNA replication) [1]. To determine if the loss of PC-TP caused a shift in the kinetics of virus replication, I infected C10B and NT cells with HCMV at an MOI of three at 80% confluency in FBS media. At various timepoints, I lysed the cells to isolate viral proteins to examine proteins that mark each of the three stages: immediate early (IE1 protein, also known as pUL122), early (pUL44) and late (pUL99). I found no significant difference between the control CRISPR non-target cells and PC-TP knockout cells in the expression of any viral proteins (Figure 8a). These observations suggest that PC-TP functions during the very late stages of HCMV replication. PC-TP has also been noted to have a possible role in lipid metabolism [83], specifically the production of highly unsaturated PC species. I performed a lipidomic analysis of the CRISPR cells to determine if the late block in HCMV replication was due to a change in the lipidome. PC-TP knockout cells were compared to the NT control cells under the same HCMV infection conditions as in Figure 7d. I 43 found no significant effect on lipid levels when PC-TP is knocked out (Figure 8b), demonstrating that PC lipid synthesis is not disrupted. Since high MOI infections, such as the MOI of 3 infectious units per cell that I used in my studies, can mask phenotypes that are observed at lower MOI, I conducted the same TCID50 experiment at MOIs of 0.5 and 1. This method has previously shown to be effective in TCID50 experiments performed in EVOLV7 depleted cells [21].My results showed that at a reduced MOI, the PC-TP knockout did not reduce viral replication (Figure 9a and Figure 9b). 44 A B C10B:MELAAGSFSEEQFWEACAELQQPALAGADWQLLVETRRLDFMSIKSLVF WRTAHQLYWQTSIWTQITENNGTSMLKNSMNKNATERWSTGKStop WT:MELAAGSFSEEQFWEACAELQQPALAGADWQLLVETSGISIYRLLDKKTGL YEYKVFGVL... C NT C10B H.P.I 4 24 48 72 96 4 24 48 72 96 PC-TP Actin 45 PC-TP Knockout TCID50 MOI 3 D 500000 450000 400000 350000 300000 250000 200000 TCID50/ml 150000 100000 50000 0 96 120 H.P.I NT C10B Figure 7: CRISPR/Cas9 knockout of PC-TP reduces viral replication (A) DNA sequencing data of PC-TP exon 1 in C10B cell line as compared to WT PC-TP exon 1. (B) Protein sequence in exon 1 and exon 2 of PC-TP in C10B as compared to WT PC-TP exon 1 and exon 2. (C) Western blot showing the absence of PC-TP in C10B cells as compared to NT cells. Cells are infected with HCMV AD169-GFP at an MOI of three. Actin is present as a control. (D) Median infectivity of viral progeny generated in C10B cells as compared to NT cells at both 96 hpi and 120 hpi. Cells are infected with HCMV AD169-GFP at an MOI of three. 46 A NT C10B H.P.I 4 24 48 72 96 4 24 48 72 96 IE1 UL44 UL99 Actin B Fold Change (relative to Mock) 1/8X 8X 47 Figure 8: CRISPR/Cas9 Knockout of PC-TP does not affect viral protein synthesis or PC synthesis. Western Blot showing the abundance of IE1, pUL44, and pUL99 in C10B cells infected with HCMV AD169-GFP at an MOI of three as compared to NT cells infected under the same conditions. Actin is present as a control. (F) Relative PC lipid abundances in C10B cells infected with HCMV AD169-GFP at an MOI of three as compared to NT cells infected under the same conditions. 48 A PC-TP Knockout TCID50 MOI 0.5 1.20E+05 1.00E+05 8.00E+04 6.00E+04 TCID50/ml 4.00E+04 2.00E+04 0.00E+00 120 144 H.P.I. NT C10B B PC-TP Knockout TCID50 MOI 1 4.00E+05 3.50E+05 3.00E+05 2.50E+05 2.00E+05 1.50E+05 TCID50/ml 1.00E+05 5.00E+04 0.00E+00 120 144 H.P.I NT C10B Figure 9: PC-TP knockout does not reduce viral replication at MOIs of 0.5 and 1. (A) Median infectivity of viral progeny generated in C10B cells as compared to NT cells at 120 hpi and 144 hpi. Cells are infected with HCMV AD169-GFP at an MOI of 0.5. (B) Median infectivity of viral progeny generated in C10B cells as compared to NT cells at both 120 hpi and 144 hpi. Cells are infected with HCMV AD169- GFP at an MOI of one. 49 Discussion Viral hijacking of host cell metabolism has been demonstrated as a requirement for viral replication in a diverse set of viruses [80, 84], including all three classes of herpesviruses: α [85, 86], γ [87-89], and β-herpesviruses [21, 26, 79]. This remains true in HCMV, with studies documenting HCMV upregulation of glycolysis and the TCA cycle [24]. HCMV has also been demonstrated to upregulate proteins related to lipid synthesis, such as PERK [23]and EVOLV7 [21]. The viral envelope is a major determinant of infectivity in all herpesviruses and is derived from lipids in host cell membranes [90]. Therefore, disruption of HCMV manipulation of host cell metabolism may reduce infectivity. To this point, murine gamma herpesvirus 68, another enveloped virus, replication was recently found to be limited by type I interferon which targets fatty acid and cholesterol synthesis [91]. This data all suggests understanding HCMV’s manipulation of host cell lipid metabolism (both synthesis and transport) can be used to identify targets to reduce viral replication. HCMV Upregulation of Phospholipids PC is extensively upregulated over the course of a WT infection (Figure 2a and Figure 5a). This is consistent with HCMV’s incorporation of PC into the viral envelope [34]. I also found that HCMV upregulates all other types of phospholipids, just not to the same extent as PC lipids. This finding is most surprising for PE, which is incorporated into the viral envelope at a greater proportion than PC [34]. This finding may be due to either pre-existing PE lipids being incorporated into the viral envelope or PE lipids just being upregulated by the virus at a lower, more controlled rate. The possibility of pre-existing PE lipids being used can be confirmed via a pulse chase experiment, in which labeled ethanolamine is added to the cell media several hours before infection. After infection, the virus would be isolated and examined via mass spectrometry for the labeled ethanolamine. If the viral envelope contains a large portion of the labeled ethanolamine, then it is likely that it incorporated existing PE lipids. More 50 investigation is also needed into the mechanism by which HCMV upregulates phospholipids. While my data suggests that pUL37x1 helps modulate host cell metabolism, there may be possible upregulation of enzymes essential to phospholipid synthesis by HCMV, such as Choline Kinase or ETNK 1 or 2. Because of the dependence of phospholipids on DAG, the dephosphorylation of PA to DAG by LIPINs may warrant investigation. Finally, PE conversion to PC by phosphatidylethanolamine N-methyltransferase, which may also explain the discrepancy between PC and PE abundance upon infection, should also be investigated. The upregulation of these proteins can be investigated both by western blot and qRT-PCR. Analysis of metabolic flux through the different phospholipid synthesis pathways with metabolic tracer experiments may also reveal which synthesis pathways are upregulated during infection. Throughout the MS1 analysis, lipids with the same mass as PE lipids were identified, but at different retention times than expected. However, upon MS2 analysis, the identity of these lipids could not be confirmed. These could possibly be odd-chained PC lipids. I did confirm that one of these contained the PC headgroup fragment (Figure 3d). Because of their differences in headgroup structure, it is not possible for canonical PC lipids to have the same mass as PE lipids [36]. One explanation is that these PC lipids have odd chain fatty acid tails. Odd chain fatty acids have typically only been identified in plants and bacteria [92], however, it is possible that they can be present in FBS serum due to the diet of the animals. Interestingly, however, these lipids were found to be upregulated in infected cells. Because human fatty acid synthesis requires the successive addition of malonyl-CoA, a two-carbon unit [93], it is unknown how HCMV would increase the abundance of these odd chain lipids. One possibility is that HCMV uses the α-oxidation pathway, a pathway that involves the loss of a CO2 molecule from an even chain fatty acid to produce an odd chain fatty acid [94]. While odd chain fatty acids have traditionally been thought to be derived from the diet, a study has shown that they can be synthesized endogenously in human adipose tissues [95]. A separate study found that C17:0 is substantially biosynthesized in human tissue and suggested that it could be derived from the elongation of propionyl-CoA as well as the 51 α-oxidation of C18:0 [96]. Little research has been done on the biological significance of odd chain fatty acids; however, studies have shown an inverse correlation between concentration of odd chain fatty acids in blood plasma and risk of coronary heart disease [94]. While not directly relevant to HCMV infection, this data suggests that odd chain fatty acids may have a role in modulating fatty acid and cholesterol metabolism, implying that these lipids may play a larger role in HCMV infection than acting as a structural component of the viral envelope. pUL37x1 My data shows that pUL37x1 plays a role in HCMV upregulation of host phospholipids. Infections with UL37x1 knockout HCMV show a clear decrease in PC abundance and an increase in PS abundance (Figure 2b, Figure 2c). This data implies that pUL37x1 increases PC production and decreases PS production. pUL37x1 performs several functions, including Ca2+ release, ER localization [63], and inhibition of apoptosis [62]. pUL37x1 may increase PC lipid synthesis indirectly by releasing Ca2+ into the mitochondria, as calcium release has been linked to increased activation of mitochondrial metabolism [97]. Alternatively, pUL37x1 induced calcium efflux from the ER may result in ER stress [98], activating PERK via ER stress [99]. This mechanism is consistent with PERK’s role in mediating HCMV induced lipogenesis [23]. The fact that pUL37x1 induces fatty acid elongation and synthesis, two processes that occur in the ER, supports the idea that pUL37x1 manipulates metabolism through a change of ER physiology. pUL37x1 has also been noted to cause the accumulation of large vesicles in the cytoplasm [64], a phenomenon that may be explained by pUL37x1’s dramatic increase of FA synthesis. My lipidomic analysis also reveals that pUL37x1 has a role in decreasing PS synthesis. PS lipids can act as an apoptotic signal when on the outer leaflet of plasma membrane [100, 101]. Given pUL37x1’s anti-apoptotic role [62], this may be a potential mechanism of action to inhibit host cell apoptosis. The PS lipid decrease may, like the increase of PC synthesis, be explained by ER calcium 52 release. However, because the decreased PS lipids did not contain VLCFAs (Figure 4a), it is likely that pUL37x1’s mechanism for decreasing PS lipids is independent of its mechanism for increasing VLCFA abundance. PS synthesis is a calcium dependent reaction that occurs in the ER [53]. Therefore, calcium efflux from the ER would likely decrease PS synthesis. However, because of pUL37x1’s localization to the outer mitochondrial membrane, the site of ER contact and PS synthesis, it is possible that it interacts with and directly inhibits PS synthesis enzymes. My data established that pUL37x1 plays an important role in the interaction between HCMV and host cell metabolism. pUL37x1 both increases the abundance of lipids essential for the HCMV envelope as well as decreases the abundance of PS lipids that may contribute to apoptotic signaling. Recent work from the Purdy lab has shown that loss of pUL37x1 reduced HCMV infectivity. Therefore, it is likely that pUL37x1 contributes to HCMV infectivity by modulating host cell lipid metabolism. Further work should focus on determining pUL37x1’s mechanism of action. Focus should be placed on the interplay between PS, PC, and PE (Figure 1a), as pUL37x1 affects the synthesis of each lipid differently. pUL37x1, for example, may be affecting a conversion pathway between the three lipids, such as the conversion of PE and PC to PS by PSS 1 or 2. HCMV Upregulation of Ethers and Plasmalogens HCMV was found to upregulate ether and plasmalogen lipids (Figure 4b). These lipids, which compose about 15-20% of human phospholipids have several important roles, including maintaining structural integrity and rigidity of membranes, promoting integral membrane function, and acting as a precursor for second messengers [81]. PE plasmalogens specifically have been noted to make membranes more compressed and rigid [82]. While there have not been studies specifically analyzing the role of rigidity in the HCMV envelope or HCMV’s upregulation of host second messengers, it is possible that HCMV is manipulating these lipids for functions other than for the structural composition 53 of its envelope. Recent work has shown that HCMV increases peroxisome biogenesis, resulting in an increase in plasmalogen lipids. Interestingly, the study also found that, while HCMV and another enveloped virus, HSV-1, were dependent on peroxisome biogenesis for efficient replication, adenovirus type 5, a non-enveloped virus, was not [102]. While more work is needed, this suggests that HCMV may be incorporating host-derived plasmalogens into its envelope. Future work should focus on a methodology to differentiate between ether and plasmalogen lipids in MS2 analysis and analyze whether these lipids are incorporated into the viral envelope. PC-TP HCMV has been shown utilize host cell proteins, such as the ESCRT machinery [103] and the SNARE protein syntaxin 3 [104] in final envelopment, showing that HMCV requires host cell proteins during envelopment. Based on HCMV’s upregulation of PC synthesis (Figure 2a, Figure 5a), I decided to further investigate PC transfer. I focused on PC-TP, as it is the only PC transport protein that transfers PC lipids exclusively. My data shows that PC-TP activity is not required for HCMV replication under all infection conditions but may suggest that it in combination with other lipid transport proteins are. Although I have yet to fully define the role of PC-TP in HCMV replication, my findings suggest that PC-TP is utilized for replication during a late step in virus replication cycle such as final envelopment, consistent with my hypothesis. This proposed mechanism fits with my data showing that PC-TP is highly expressed at 72 and 96 hpi (Figure 5b), the role of the HCMV envelope in virion infectivity [90], and PC- TP’s mechanism in normal human biology [48]. CRISPR/Cas9 Knockout of PC-TP and LDN193188 The CRISPR/Cas9 knockout of PC-TP showed that viral replication was reduced at 120 hpi but not 96 hpi at an MOI of 3 (Figure 7d), suggesting that another STARD family protein may support viral replication in the absence of PC-TP activity, although less efficiently than PC-TP. This observation further 54 suggests that at 120 hpi, when viral replication is occurring at a faster rate, the STARD proteins are no longer able to compensate for PC-TP, and the effect of the loss of PC-TP results in an inhibition of replication (Figure 7d). This explains why the change in titer in the PC-TP knockout cells from 96 hpi to 120 hpi is much smaller than the NT control change over the same period. Interestingly, this phenotype was not seen at 120 hpi and 144 hpi in cells infected at an MOI of 0.5 and 1 (Figure 9a and Figure 9b). It was expected that a lower MOI would produce a more pronounced phenotype, as higher MOIs have been shown to mask phenotypes [21]. However, this data can still be explained by the notion that viral replication, at these lower MOIs, never reached the threshold where PC-TP was necessary for viral replication. Treatment with LDN193188 showed a decrease in replication at both 96 hpi and 120 hpi, even at non-cytotoxic concentrations (Figure 6a, Figure 6b). LDN193188 outcompetes PC in the PC-TP lipid biding pocket, therefore, LDN193188 may be inhibiting other STARD proteins that also transfer PC by outcompeting PC in their binding pocket [76]. Based on the non-specific effects of LDN193188 and HCMV’s possible upregulation of STARD family proteins, it is necessary to continue my study to fully determine their role in supporting HCMV replication. In future studies, an siRNA screen of all STARD family proteins to determine which ones are required for HCMV replication should be performed. HCMV Upregulation of PC-TP HCMV infection was shown to upregulate PC-TP protein (Figure 5b). HCMV may be upregulating the transcription of PC-TP, ultimately increasing the level of PC-TP protein. Alternatively, HCMV could be preventing the degradation of PC-TP. A qPCR experiment measuring the amount of PC-TP mRNA in infected and mock infected cells will be able to determine if HCMV is upregulating transcription of PC- TP. PC-TP is upregulated by peroxisome proliferated activator receptor alpha (PPAR-α) at the transcriptional level in response to increased glucose and lipid concentration in the cell. [48]. If HCMV 55 were to upregulate PC-TP transcription, the protein and transcript level of PPAR-α should also be measured during infection. Mechanism of PC-TP in HCMV Replication PC-TP likely transfers PC lipids into the HCMV envelope. To delve further into the mechanism of PC-TP in HCMV replication, work should be done to test if the composition of the HCMV envelope has reduced PCs upon PC-TP knockout. If it does, then it can be inferred that PC-TP transfers PCs into the HCMV envelope. HCMV virions generated in both NT and PC-TP knockout cells should be isolated and purified and analyzed via LC-MS/MS, identifying and quantifying their lipid composition. HCMV envelopment occurs in the assembly complex, a novel organelle induced by HCMV which consists of the host cell’s nucleus, endoplasmic reticulum, and Golgi apparatus [1]. Therefore, if my hypothesis that PC-TP is involved envelopment is correct, I would expect that PC-TP localizes to the assembly compartment. While there are several ways to determine this, a GFP tag is likely the best method. The alternative method would be to perform immunofluorescent studies using an anti PC-TP antibody if one can be found to be effective for microscopy studies. PC-TP may also be involved in HCMV remodeling of host cell metabolism, specifically glucose and fatty acids. PC-TP has a documented regulatory role in which its loss results in the upregulation of PPAR-α and hepatocyte nuclear factor 4 alpha (HNF4α) [105], a transcription factor involved in hepatic liver metabolism [106]. PC-TP knockout mice show large decreases in plasma glucose and free fatty acids [107], consistent with the role of PC-TP in glucose and fatty acid metabolism in the liver. To explore this, a broad metabolomics [80] experiment measuring the change in metabolite concentration from infected NT CRIPSR cells to infected PC-TP knockout cells will be valuable. As mentioned earlier in the discussion, this approach can also be used to understand how HCMV upregulates phospholipid synthesis by analyzing metabolic flux through phospholipid synthesis pathways. By incorporating isotope tracers into 56 precursor metabolites and using LC-MS/MS, it can be determined what metabolic pathway HCMV is inducing to generate PC lipids required for infection [80]. Overall, my project has shown that HCMV upregulates host cell lipid metabolism through infection, particularly the synthesis of PCs with VLCFAs. My data also reveals a role for PC-TP in HCMV replication as well as its potential role in envelopment. This data agrees well with my overall hypothesis that HCMV upregulates the production of cellular phospholipids and uses host cell lipid transport proteins in order to transport those phospholipids to the viral envelope. Although further study is needed to elucidate the true mechanism of PC-TP in envelopment and replication, as well as the role of other STARD family proteins, this work further demonstrates the importance of host proteins in viral replication. My work also helped reveal the role of the viral protein pUL37x1 in HCMV manipulation of lipid metabolism. While my analysis of pUL37x1’s role in phospholipid production revealed potential mechanisms by which pUL37x1 may act (such as ER calcium release) further study is needed to understand pUL37x1’s complete role and mechanism in HCMV infection. Perhaps more importantly, future work is needed on the role of viral proteins in metabolism as whole, as a complete understanding of the interplay between viral and host factors is needed to develop therapeutics for HMCV infection. My work begins to shed light on that interplay, showing that both viral and host proteins are required for replication at different stages of lipid metabolism and transport. My data and conclusions therefore may lead to new therapeutic options targeting the mechanisms that enveloped viruses like HCMV use to generate their lipid membrane envelope. 57 References 1. Edward Mocarski, T.S., Paul Griffiths, and Robert Pass, Cytomegaloviruses, in Fields Virology. 2013, Wolters Kluwer Health/Lippincott Williams & Wilkins: Philadelphia. 2. Boppana, S.B. and K.B. Fowler, Persistence in the population: epidemiology and transmisson, in Human Herpesviruses: Biology, Therapy, and Immunoprophylaxis, A. Arvin, et al., Editors. 2007: Cambridge. 3. Cheeran, M.C., J.R. Lokensgard, and M.R. Schleiss, Neuropathogenesis of congenital cytomegalovirus infection: disease mechanisms and prospects for intervention. Clin Microbiol Rev, 2009. 22(1): p. 99-126, Table of Contents. 4. Cobbs, C.S., et al., Human cytomegalovirus infection and expression in human malignant glioma. Cancer Res, 2002. 62(12): p. 3347-50. 5. Cobbs, C.S., Cytomegalovirus and brain tumor: epidemiology, biology and therapeutic aspects. Curr Opin Oncol, 2013. 25(6): p. 682-8. 6. Streblow, D.N., et al., Mechanisms of cytomegalovirus-accelerated vascular disease: induction of paracrine factors that promote angiogenesis and wound healing. Curr Top Microbiol Immunol, 2008. 325: p. 397-415. 7. Moss, P., The emerging role of cytomegalovirus in driving immune senescence: a novel therapeutic opportunity for improving health in the elderly. Curr Opin Immunol, 2010. 22(4): p. 529-34. 8. Britt, W., Manifestations of human cytomegalovirus infection: proposed mechanisms of acute and chronic disease. Curr Top Microbiol Immunol, 2008. 325: p. 417-70. 9. McCluskey, R., R. Sandin, and J. Greene, Detection of airborne cytomegalovirus in hospital rooms of immunocompromised patients. J Virol Methods, 1996. 56(1): p. 115-8. 10. Staras, S.A., et al., Influence of sexual activity on cytomegalovirus seroprevalence in the United States, 1988-1994. Sex Transm Dis, 2008. 35(5): p. 472-9. 11. Swanson, E.C. and M.R. Schleiss, Congenital cytomegalovirus infection: new prospects for prevention and therapy. Pediatr Clin North Am, 2013. 60(2): p. 335-49. 12. Wills, M.R., et al., The immunology of human cytomegalovirus latency: could latent infection be cleared by novel immunotherapeutic strategies? Cell Mol Immunol, 2015. 12(2): p. 128-38. 13. Stern-Ginossar, N., et al., Decoding human cytomegalovirus. Science, 2012. 338(6110): p. 1088- 93. 14. Matsumoto, K., et al., Risk factors for ganciclovir-induced thrombocytopenia and leukopenia. Biol Pharm Bull, 2015. 38(2): p. 235-8. 15. Ahmed, A., Antiviral treatment of cytomegalovirus infection. Infect Disord Drug Targets, 2011. 11(5): p. 475-503. 16. Prichard, M.N. and E.R. Kern, The search for new therapies for human cytomegalovirus infections. Virus Res, 2011. 157(2): p. 212-21. 17. Biron, K.K., et al., Metabolic activation of the nucleoside analog 9-[( 2-hydroxy-1- (hydroxymethyl)ethoxy]methyl)guanine in human diploid fibroblasts infected with human cytomegalovirus. Proc Natl Acad Sci U S A, 1985. 82(8): p. 2473-7. 18. Sullivan, V., et al., A protein kinase homologue controls phosphorylation of ganciclovir in human cytomegalovirus-infected cells. Nature, 1992. 359(6390): p. 85. 19. Erice, A., Resistance of human cytomegalovirus to antiviral drugs. Clin Microbiol Rev, 1999. 12(2): p. 286-97. 20. Das, S., A. Vasanji, and P.E. Pellett, Three-dimensional structure of the human cytomegalovirus cytoplasmic virion assembly complex includes a reoriented secretory apparatus. J Virol, 2007. 81(21): p. 11861-9. 58 21. Purdy, J.G., T. Shenk, and J.D. Rabinowitz, Fatty acid elongase 7 catalyzes lipidome remodeling essential for human cytomegalovirus replication. Cell Rep, 2015. 10(8): p. 1375-85. 22. Koyuncu, E., et al., Saturated very long chain fatty acids are required for the production of infectious human cytomegalovirus progeny. PLoS Pathog, 2013. 9(5): p. e1003333. 23. Yu, Y., et al., PKR-like endoplasmic reticulum kinase is necessary for lipogenic activation during HCMV infection. PLoS Pathog, 2013. 9(4): p. e1003266. 24. Munger, J., et al., Dynamics of the cellular metabolome during human cytomegalovirus infection. PLoS Pathog, 2006. 2(12): p. e132. 25. Hwang, J., et al., Estrogen-related receptor alpha is required for efficient human cytomegalovirus replication. Proc Natl Acad Sci U S A, 2014. 111(52): p. E5706-15. 26. Chambers, J.W., T.G. Maguire, and J.C. Alwine, Glutamine metabolism is essential for human cytomegalovirus infection. J Virol, 2010. 84(4): p. 1867-73. 27. McArdle, J., X.L. Schafer, and J. Munger, Inhibition of calmodulin-dependent kinase kinase blocks human cytomegalovirus-induced glycolytic activation and severely attenuates production of viral progeny. J Virol, 2011. 85(2): p. 705-14. 28. McArdle, J., N.J. Moorman, and J. Munger, HCMV targets the metabolic stress response through activation of AMPK whose activity is important for viral replication. PLoS Pathog, 2012. 8(1): p. e1002502. 29. Terry, L.J., et al., Human kinome profiling identifies a requirement for AMP-activated protein kinase during human cytomegalovirus infection. Proc Natl Acad Sci U S A, 2012. 109(8): p. 3071- 6. 30. Terhune, S., et al., Human cytomegalovirus UL38 protein blocks apoptosis. J Virol, 2007. 81(7): p. 3109-23. 31. Rodriguez-Sanchez, I., et al., The Human Cytomegalovirus UL38 protein drives mTOR- independent metabolic flux reprogramming by inhibiting TSC2. PLoS Pathog, 2019. 15(1): p. e1007569. 32. Hay, N. and N. Sonenberg, Upstream and downstream of mTOR. Genes Dev, 2004. 18(16): p. 1926-45. 33. Inoki, K., et al., TSC2 is phosphorylated and inhibited by Akt and suppresses mTOR signalling. Nat Cell Biol, 2002. 4(9): p. 648-57. 34. Liu, S.T., 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): p. 12869-74. 35. Ridgway, N.D., Phospholipid Synthesis in Mammalian Cells, in Biochemistry of Lipids, Lipoproteins, and Membranes, N.D.R.a.R.S. Mcleod, Editor. 2016, Elsevier. 36. Kim, S., et al., PubChem Substance and Compound databases. Nucleic Acids Res, 2016. 44(D1): p. D1202-13. 37. Wang, X., et al., Signaling functions of phosphatidic acid. Prog Lipid Res, 2006. 45(3): p. 250-78. 38. Weigert, R., et al., CtBP/BARS induces fission of Golgi membranes by acylating lysophosphatidic acid. Nature, 1999. 402(6760): p. 429-33. 39. Blackwood, R.A., et al., Phospholipase D activity facilitates Ca2+-induced aggregation and fusion of complex liposomes. Am J Physiol, 1997. 272(4 Pt 1): p. C1279-85. 40. van der Veen, J.N., et al., The critical role of phosphatidylcholine and phosphatidylethanolamine metabolism in health and disease. Biochim Biophys Acta, 2017. 1859(9 Pt B): p. 1558-1572. 41. Kennedy, E.P. and S.B. Weiss, The function of cytidine coenzymes in the biosynthesis of phospholipides. J Biol Chem, 1956. 222(1): p. 193-214. 42. Lands, W.E., Metabolism of glycerolipides; a comparison of lecithin and triglyceride synthesis. J Biol Chem, 1958. 231(2): p. 883-8. 59 43. Schneider, W.J. and D.E. Vance, Conversion of phosphatidylethanolamine to phosphatidylcholine in rat liver. Partial purification and characterization of the enzymatic activities. J Biol Chem, 1979. 254(10): p. 3886-91. 44. Ridgway, N.D. and D.E. Vance, Purification of phosphatidylethanolamine N-methyltransferase from rat liver. J Biol Chem, 1987. 262(35): p. 17231-9. 45. Jacobs, R.L., et al., Impaired de novo choline synthesis explains why phosphatidylethanolamine N-methyltransferase-deficient mice are protected from diet-induced obesity. J Biol Chem, 2010. 285(29): p. 22403-13. 46. Vance, D.E. and N.D. Ridgway, The methylation of phosphatidylethanolamine. Prog Lipid Res, 1988. 27(1): p. 61-79. 47. Mozzi, R. and G. Porcellati, Conversion of phosphatidylethanolamine to phosphatidylcholine in rat brain by the methylation pathway. FEBS Lett, 1979. 100(2): p. 363-6. 48. Kang, H.W., J. Wei, and D.E. Cohen, PC-TP/StARD2: Of membranes and metabolism. Trends Endocrinol Metab, 2010. 21(7): p. 449-56. 49. Cockcroft, S. and N. Carvou, Biochemical and biological functions of class I phosphatidylinositol transfer proteins. Biochim Biophys Acta, 2007. 1771(6): p. 677-91. 50. Jiang, X.C., The effect of phospholipid transfer protein on lipoprotein metabolism and atherosclerosis. Front Biosci, 2002. 7: p. d1634-41. 51. Clark, B.J., The mammalian START domain protein family in lipid transport in health and disease. J Endocrinol, 2012. 212(3): p. 257-75. 52. Gustin, S.E., et al., Testis development, fertility, and survival in Ethanolamine kinase 2-deficient mice. Endocrinology, 2008. 149(12): p. 6176-86. 53. Kuge, O. and M. Nishijima, Phosphatidylserine synthase I and II of mammalian cells. Biochim Biophys Acta, 1997. 1348(1-2): p. 151-6. 54. Vance, J.E., Phospholipid synthesis and transport in mammalian cells. Traffic, 2015. 16(1): p. 1- 18. 55. Kornmann, B., et al., An ER-mitochondria tethering complex revealed by a synthetic biology screen. Science, 2009. 325(5939): p. 477-81. 56. Wirtz, K.W., Phospholipid transfer proteins. Annu Rev Biochem, 1991. 60: p. 73-99. 57. Paulus, H. and E.P. Kennedy, The enzymatic synthesis of inositol monophosphatide. J Biol Chem, 1960. 235: p. 1303-11. 58. Hsuan, J. and S. Cockcroft, The PITP family of phosphatidylinositol transfer proteins. Genome Biol, 2001. 2(9): p. REVIEWS3011. 59. Goldmacher, V.S., et al., A cytomegalovirus-encoded mitochondria-localized inhibitor of apoptosis structurally unrelated to Bcl-2. Proc Natl Acad Sci U S A, 1999. 96(22): p. 12536-41. 60. Smith, G.B. and E.S. Mocarski, Contribution of GADD45 family members to cell death suppression by cellular Bcl-xL and cytomegalovirus vMIA. J Virol, 2005. 79(23): p. 14923-32. 61. Salvador, J.M., J.D. Brown-Clay, and A.J. Fornace, Jr., Gadd45 in stress signaling, cell cycle control, and apoptosis. Adv Exp Med Biol, 2013. 793: p. 1-19. 62. McCormick, A.L., Control of Apoptosis by Human Cytomegalovirus in Human Cytomegalovirus, T.S.a.M.F. Stinski, Editor. 2008, Springer: Berlin. 63. Sharon-Friling, R., et al., Human cytomegalovirus pUL37x1 induces the release of endoplasmic reticulum calcium stores. Proc Natl Acad Sci U S A, 2006. 103(50): p. 19117-22. 64. Sharon-Friling, R. and T. Shenk, Human cytomegalovirus pUL37x1-induced calcium flux activates PKCalpha, inducing altered cell shape and accumulation of cytoplasmic vesicles. Proc Natl Acad Sci U S A, 2014. 111(12): p. E1140-8. 65. Wang, D., et al., The ULb' region of the human cytomegalovirus genome confers an increased requirement for the viral protein kinase UL97. J Virol, 2013. 87(11): p. 6359-76. 60 66. Woodroffe, S.B., J. Hamilton, and H.M. Garnett, Comparison of the infectivity of the laboratory strain AD169 and a clinical isolate of human cytomegalovirus to human smooth muscle cells. J Virol Methods, 1997. 63(1-2): p. 181-91. 67. Yu, D., et al., Construction of a self-excisable bacterial artificial chromosome containing the human cytomegalovirus genome and mutagenesis of the diploid TRL/IRL13 gene. J Virol, 2002. 76(5): p. 2316-28. 68. Kamphorst, J.J., et al., Liquid chromatography-high resolution mass spectrometry analysis of fatty acid metabolism. Anal Chem, 2011. 83(23): p. 9114-22. 69. Clasquin, M.F., E. Melamud, and J.D. Rabinowitz, LC-MS data processing with MAVEN: a metabolomic analysis and visualization engine. Curr Protoc Bioinformatics, 2012. Chapter 14: p. Unit14 11. 70. Melamud, E., L. Vastag, and J.D. Rabinowitz, Metabolomic analysis and visualization engine for LC-MS data. Anal Chem, 2010. 82(23): p. 9818-26. 71. Murphy, R.E., Fatty Acids, in Tandem Mass Spectrometry of Lipids, S.J. Gaskell, Editor. 2015: Cambridge. 72. Murphy, R.E., Glycerophospholipids, in Tandem Mass Spectrometry of Lipids, S.J. Caskell, Editor. 2015, Royal Society of Chemistry: Cambridge. 73. Pi, J., X. Wu, and Y. Feng, Fragmentation patterns of five types of phospholipids by ultra-high- performance liquid chromatography electrospray ionization quadrupole time-of-flight tandem mass spectrometry. Analytical Methods, 2016. 8(6): p. 1319-1332. 74. Sud, M., et al., LMSD: LIPID MAPS structure database. Nucleic Acids Res, 2007. 35(Database issue): p. D527-32. 75. Fahy, E., et al., Update of the LIPID MAPS comprehensive classification system for lipids. J Lipid Res, 2009. 50 Suppl: p. S9-14. 76. Shishova, E.Y., et al., Genetic ablation or chemical inhibition of phosphatidylcholine transfer protein attenuates diet-induced hepatic glucose production. Hepatology, 2011. 54(2): p. 664-74. 77. Ran, F.A., et al., Genome engineering using the CRISPR-Cas9 system. Nat Protoc, 2013. 8(11): p. 2281-2308. 78. Shalem, O., et al., Genome-scale CRISPR-Cas9 knockout screening in human cells. Science, 2014. 343(6166): p. 84-87. 79. Munger, J., et al., Systems-level metabolic flux profiling identifies fatty acid synthesis as a target for antiviral therapy. Nat Biotechnol, 2008. 26(10): p. 1179-86. 80. Rabinowitz, J.D., et al., Metabolomics in drug target discovery. Cold Spring Harb Symp Quant Biol, 2011. 76: p. 235-46. 81. Braverman, N.E. and A.B. Moser, Functions of plasmalogen lipids in health and disease. Biochim Biophys Acta, 2012. 1822(9): p. 1442-52. 82. Rog, T. and A. Koivuniemi, The biophysical properties of ethanolamine plasmalogens revealed by atomistic molecular dynamics simulations. Biochim Biophys Acta, 2016. 1858(1): p. 97-103. 83. Geijtenbeek, T.B., et al., Phosphatidylcholine transfer protein from bovine liver contains highly unsaturated phosphatidylcholine species. FEBS Lett, 1996. 391(3): p. 333-5. 84. Goodwin, C.M., S. Xu, and J. Munger, Stealing the Keys to the Kitchen: Viral Manipulation of the Host Cell Metabolic Network. Trends Microbiol, 2015. 23(12): p. 789-798. 85. Vastag, L., et al., Divergent effects of human cytomegalovirus and herpes simplex virus-1 on cellular metabolism. PLoS Pathog, 2011. 7(7): p. e1002124. 86. Grady, S.L., et al., Argininosuccinate synthetase 1 depletion produces a metabolic state conducive to herpes simplex virus 1 infection. Proc Natl Acad Sci U S A, 2013. 110(51): p. E5006- 15. 61 87. Delgado, T., et al., Induction of the Warburg effect by Kaposi's sarcoma herpesvirus is required for the maintenance of latently infected endothelial cells. Proc Natl Acad Sci U S A, 2010. 107(23): p. 10696-701. 88. Sanchez, E.L., et al., Glycolysis, Glutaminolysis, and Fatty Acid Synthesis Are Required for Distinct Stages of Kaposi's Sarcoma-Associated Herpesvirus Lytic Replication. J Virol, 2017. 91(10). 89. McFadden, K., et al., Metabolic stress is a barrier to Epstein-Barr virus-mediated B-cell immortalization. Proc Natl Acad Sci U S A, 2016. 113(6): p. E782-90. 90. Tandon, R. and E.S. Mocarski, Viral and host control of cytomegalovirus maturation. Trends Microbiol, 2012. 20(8): p. 392-401. 91. Lange, P.T., et al., Liver X Receptors Suppress Activity of Cholesterol and Fatty Acid Synthesis Pathways To Oppose Gammaherpesvirus Replication. MBio, 2018. 9(4). 92. Rezanka, T. and K. Sigler, Odd-numbered very-long-chain fatty acids from the microbial, animal and plant kingdoms. Prog Lipid Res, 2009. 48(3-4): p. 206-38. 93. Wakil, S.J., J.K. Stoops, and V.C. Joshi, Fatty acid synthesis and its regulation. Annu Rev Biochem, 1983. 52: p. 537-79. 94. Jenkins, B., J.A. West, and A. Koulman, A review of odd-chain fatty acid metabolism and the role of pentadecanoic Acid (c15:0) and heptadecanoic Acid (c17:0) in health and disease. Molecules, 2015. 20(2): p. 2425-44. 95. Roberts, L.D., et al., Metabolic phenotyping of a model of adipocyte differentiation. Physiol Genomics, 2009. 39(2): p. 109-19. 96. Jenkins, B.J., et al., Odd Chain Fatty Acids; New Insights of the Relationship Between the Gut Microbiota, Dietary Intake, Biosynthesis and Glucose Intolerance. Sci Rep, 2017. 7: p. 44845. 97. Gaspers, L.D. and A.P. Thomas, Calcium-dependent activation of mitochondrial metabolism in mammalian cells. Methods, 2008. 46(3): p. 224-32. 98. Dejeans, N., et al., Endoplasmic reticulum calcium release potentiates the ER stress and cell death caused by an oxidative stress in MCF-7 cells. Biochem Pharmacol, 2010. 79(9): p. 1221-30. 99. Ron, D. and P. Walter, Signal integration in the endoplasmic reticulum unfolded protein response. Nat Rev Mol Cell Biol, 2007. 8(7): p. 519-29. 100. Fadok, V.A., et al., The role of phosphatidylserine in recognition of apoptotic cells by phagocytes. Cell Death Differ, 1998. 5(7): p. 551-62. 101. Fadok, V.A., et al., Particle digestibility is required for induction of the phosphatidylserine recognition mechanism used by murine macrophages to phagocytose apoptotic cells. J Immunol, 1993. 151(8): p. 4274-85. 102. Jean Beltran, P.M., et al., Infection-Induced Peroxisome Biogenesis Is a Metabolic Strategy for Herpesvirus Replication. Cell Host Microbe, 2018. 24(4): p. 526-541 e7. 103. Tandon, R., D.P. AuCoin, and E.S. Mocarski, Human cytomegalovirus exploits ESCRT machinery in the process of virion maturation. J Virol, 2009. 83(20): p. 10797-807. 104. Cepeda, V. and A. Fraile-Ramos, A role for the SNARE protein syntaxin 3 in human cytomegalovirus morphogenesis. Cell Microbiol, 2011. 13(6): p. 846-58. 105. Kang, H.W., et al., Regulatory role for phosphatidylcholine transfer protein/StarD2 in the metabolic response to peroxisome proliferator activated receptor alpha (PPARalpha). Biochim Biophys Acta, 2010. 1801(4): p. 496-502. 106. Coordinators, N.R., Database resources of the National Center for Biotechnology Information. Nucleic Acids Res, 2013. 41(Database issue): p. D8-D20. 107. Scapa, E.F., et al., Regulation of energy substrate utilization and hepatic insulin sensitivity by phosphatidylcholine transfer protein/StarD2. FASEB J, 2008. 22(7): p. 2579-90.