“Flash & Click”: Multifunctionalized Lipid Derivatives as Tools To Study Viral Infections

Abstract

In this perspective article, we describe the current status of lipid tools for studying host lipid–virus interactions at the cellular level. We discuss the potential lipidomic changes that viral infections impose on host cells and then outline the tools available and the resulting options to investigate the host cell lipid interactome. The future outcome will reveal new targets for treating virus infections.

Introduction

Lipids are a central part of the various cell membranes to which they give shape, fluidity, and polarity. Tens of thousands of individual lipid species constantly shift and interact to enable dynamic membrane processes such as fusion, vesiculation, and receptor clustering;¹⁻³ they are also often key players in initiating or continuing signaling cascades that govern many of the most essential cellular functions, such as secretion, endocytosis, growth, migration, mitosis, and apoptosis: lipids such as ceramide,^4,5 sphingosine,⁶⁻⁸ diacylglycerol,⁹ many lysolipids,¹⁰ and phosphoinositides,¹¹ to say nothing of the huge families of arachidonic acid derived second messengers,¹² all have been shown to have specific signaling roles, often carried out by interacting with specific receptors or other proteins. Descriptions of these lipid functions have lagged far beyond our understanding of protein function, in large part due to a dearth of powerful techniques to visualize, manipulate, and analyze the interactions of individual lipids. Of the four major classes of biomolecules, nucleotides and proteins can be genetically manipulated, which greatly facilitates their detailed molecular study; carbohydrates and lipids, on the other hand, are not directly genetically encoded. As such, lipids behave in many ways like small molecule drugs, which come with associated benefits and drawbacks for the experimentalist. They can, under the right circumstances, be taken up passively to the cell interior; however, there is a lack of genetic techniques to manipulate lipid composition in a controlled way.

As lipids behave like other small molecules when entering cells, major questions are what is the lipid binding to, where does it locate, and how is it metabolized? Answers to these questions are highly relevant for the development and understanding of therapeutics that are intended to work on lipid-binding and -metabolizing proteins. While tools to address these questions for drugs and drug candidates are quite common, they have been less often applied to the small molecules that cells make themselves, including lipids. One strategy that has gained traction in recent years is that of the “bifunctional” derivative of the molecule of interest, bearing (usually) a diazirine for photo-cross-linking (“Flash”) and an azide or alkyne for copper-catalyzed click chemistry (“Click”). This technique has been used extensively in the field of chemoproteomics to identify targets of small molecules in cells, both of endogenous small molecules and of small-molecule drugs.^13,14 In the lipid field, bi- (and tri)functionalized lipid derivatives have emerged recently, and these tools offer the opportunity to understand how individual lipid species are functioning in a variety of biological systems. Here, we present the perspective of applying these tools to study the biology of the life cycle of viruses during infection.

Viruses are obligate intracellular parasites. While they contain the information necessary to propagate themselves, they lack the environment, the energy, and, to a large degree, the enzymatic machinery to do it alone. Nucleic acid replication is as complicated and energetically costly an endeavor for a virus as for any other carrier of genetic material, the virus must acquire diverse resources from its host: not only nucleic acids and enzymes for replication but also metabolic energy, translational machinery for their own proteins, membranous platforms to replicate in and on, lipid components of the virion itself, and routes into and out of their host cells. Moreover, as they are often replicating in a hostile environment, they have the added burden of protecting themselves from host defenses: this can include antagonizing innate immune pathways, such as autophagy and the unfolded protein response, disrupting the synthesis of inflammatory lipid mediators such as eicosanoids and prostaglandins, and even shredding host membranes to refashion protected enclaves for themselves and evade detection altogether.

While much effort over many decades has gone into understanding how viruses manipulate and interact with host proteins, much less is known about how viruses affect and are affected by host lipids. In part, this is due to the relatively recent appreciation of the myriad roles that tens of thousands of individual lipids play in cell biology more generally, from membrane structure and function to complex cellular signaling. Intensive biochemical, chemical, and cell biological investigations have begun to reveal the absolutely critical role host lipids play in the life cycles of many viruses, and the impressive lengths viruses will go to remodel the lipid compositions and membrane structures of their hosts. However, many tantalizing questions remain unanswered, including how individual lipid species are organized subcellularly in an infected cell, how the metabolism of individual lipids changes during infection, and the molecular details of any protein–lipid interactions during infection. These questions require new chemical tools in order to be satisfactorily answered.

Antiviral drugs can act at any stage of the viral life cycle: many block viral entry, as is the case for SARS-CoV-2 monoclonal antibody therapies.¹⁵ Others target the translation or replication machinery of the virus; these include viral proteases (prominently, of HIV¹⁶ and HepC,¹⁷ viral polymerases,¹⁷ and, recently, the SARS-CoV-2 nsp5 protease,¹⁸ which are often targeted with nucleoside derivatives that lead to incomplete replication, such as acyclovir, which is used against the Herpes simplex virus.¹⁹ As discussed above, multiple viral proteins interact with lipids and one could expect that disrupting lipid interaction will have a negative effect on the virus life cycle. Further, the formation of replication cisterns and their complex membrane structure seems to be crucial for the replication of many viruses including flaviviruses and coronaviruses. Because lipids are a central hub for host–virus interactions and are required at every stage of the life cycle, studying these intricate relationships would certainly not only provide insight into their basic biology but also help identify drug targets. In order to determine potential drug targets we first need to identify the lipid binding proteins involved in the course of infection. The “Flash & Click” toolbox developed at the chemistry bench will be key to obtain a comprehensive map of the lipid-specific interactomes and to gain mechanistic insights into the infection biology of viruses.

The Lipid Toolbox

Lipids as a class of biomolecule resist the central dogma of biology. As they are not genetically encoded, methods of tagging that have allowed detailed studies of protein structure and function are unavailable, and knockouts of lipid biosynthetic enzymes often have unexpected pleiotropic effects which makes the role of an individual lipid species difficult to tease out. A tool to study individual lipid function would ideally have several characteristics: (1) a way of getting into the target cell; (2) a means of protecting the lipid from rapid endogenous metabolism and/or preventing it from inducing unwanted signaling; (3) a tag to allow the lipid to be tracked, visualized, or pulled down; (4) a way to render permanent the transient and weak forces that characterize protein–lipid interactions (summarized in Figure 1).

Figure 1.

Multifunctional lipid probes contain up to four functional groups to facilitate analysis of their cellular function. Esters (1) are applied to phosphates, carboxylates, and hydroxy groups to mask their negative charge and hydrophilic functionality. These groups are removed after cell entry by endogenous esterases. Photocages (2), often based on coumarins or nitrobenzyl groups, prevent premature metabolism, until they are released by blue light. Terminal alkynes (3) are used as handles for click chemistry to derivatize lipid–protein complexes with fluorophores or affinity tags for downstream proteomic analysis. Diazirines (4) form reactive carbenes upon irradiation with UV light or isomerize into diazo intermediates, both of which can form covalent bonds with nearby proteins.

Cell Entry

Many naturally occurring lipid species are commercially available. However, when lipids need to be applied to living cells, the incompatibility of many lipids with the aqueous environment of a cell dish makes administration tricky. This is particularly true for charged and hence amphiphilic lipid species. Negatively charged lipids such as stearic acid or the phosphoinositides will rapidly form insoluble aggregates when added to cell dishes. In addition, the negative charge prevents cell entry partly due to the unfavorable negative potential of the plasma membrane. Here, chemists came into action and used esters previously employed to make cyclic nucleotides and inositol polyphosphates membrane-permeant²⁰⁻²² (Figure 1.1). Lipids were equipped with uncharged masking groups to prevent aggregation and to permit passive cell entry.^23,24 Once inside cells, the masking groups are removed by endogenous hydrolases and the delivered lipid becomes active.²⁵ However, at this point, a new challenge arises, as endogenous lipid metabolism is fast and efficient, and the newly introduced lipid species is often rapidly metabolized. If this metabolism is faster than the often multistep enzymatic bioactivation, there might be never enough of the lipid species to exhibit biological function.

Protection from Metabolism

To address this problem, we and others introduced caging groups that functioned as a light-removable protecting group²⁶⁻²⁸ (Figure 1.2). Such photocages function mostly through steric bulk to prevent biosynthetic machinery from accessing the lipid, and they also prevent signaling lipids from interacting with the protein intermediaries of their signaling cascades. When applied to phosphates of phospholipids, premature metabolism was mostly prevented and the cell had time to remove all other bioactivatable protecting groups, for the most part esters,²⁹ providing excellent tools to elevate endogenous lipid levels noninvasively. Caged lipids have proliferated recently, and many caged lipid derivatives have been used to study lipid signaling pathways,^8,30 lipid biophysics,³¹ and direct lipids to specific subcellular compartments,³² even without added functionalities. After removal of the often fluorescent cage by light, unfortunately the lipids are not directly trackable anymore. For the reasons discussed in the introduction, a tracking method geared toward small molecules was needed.

Tagging

The small molecular size of lipids adds significant level of difficulty for the chemist who would simply like to append a small-molecule fluorophore to the structure of a lipid, as fluorophores frequently cause mislocation of the tagged lipid. A caged version of the signaling lipid PIP₃ is predominantly located to endoplasmic reticulum and Golgi membranes,²⁷ while its endogenous counterpart is known to reside at the plasma membrane. The reason for this mislocation is not well understood, but it has been frequently observed that lipid derivatives that have an aromatic entity attached end up in endomembranes, seemingly independent of the lipid headgroup and the location of the aromatic attachment.³³ Thus, it is critical in the design of lipid-based probes that any modifications to the lipid structure are as minimal as possible, and particularly avoid the inclusion of aromatic groups.

One of the most minimal chemical modifications to date are the azides and alkyne groups used for “click” chemistry based on the Huisgen [3 + 2] cycloaddition. This commonly used reaction is bioorthogonal, rapid, and quantitative, allowing selective visualization of tagged moieties even in the complex chemical milieu that is a living cell. There are numerous click chemistries available;^34,35 however, the click labeling of lipids has almost exclusively involved azides and alkynes, as they represent the smallest alterations to the overall molecular interaction potential. Azides have been used both at the fatty acid termini and at head groups.^36,37 The first alkyne lipids for click chemistry were introduced by Salic and co-workers.³⁸ Soon after, we introduced cyclic alkynes that permitted intracellular labeling in living cells.³⁹ In the meantime, a large number of lipids have been equipped with terminal alkyne groups, some of them located at the fatty acids, some of them at the lipid headgroup⁴⁰ (Figure 1.3). “Clickable” lipids have been used predominantly to track metabolism in cells⁴¹⁻⁴³ and in enzymatic assays,⁴⁴ since, as will be discussed below, the simple inclusion of a click handle is not sufficient for all downstream applications. However, it is still a powerful tool for pulse-chase labeling of lipids in cells; previous tools for this usually relied on autoradiography, which in this context has been shown to be orders of magnitude less sensitive than fluorescence.³⁸

Photo-cross-linking

A tagging group alone, however, is unfortunately not enough to identify a lipid’s location or its binding partners. Visualizing the tagged lipid is hampered by the fact that fixation and permeabilization for microscopy involve organic solvents such as methanol and detergents, all of which can perturb lipid localization or wash them out of a fixed cell entirely. Proteins and lipids interact via weak, transient van der Waals forces, which are not robust enough to survive a technique such as coimmunoprecipitation which is the gold standard in identifying protein–protein interactions. To overcome these problems, a functional group that can transform a transient, electrostatic interaction into a robust covalent bond was added. Initially, photo-cross-linking benzophenone groups were attached to phosphoinositides by Prestwich and others.^45,46 As discussed above, however, aromatic groups lead to a mis-location of lipids. As an alternative, Per Haberkant and others used nonaromatic diazirine groups as minimal, effective light-triggered cross-linkers^47,48 (Figure 1.4). Similarly, the Cravatt lab in their seminal work also introduced diazirines for studying the cholesterol interactome⁴⁹ and the interaction partners of various fatty acids and fatty acid amides.⁵⁰ Diazirines form covalent linkages with proteins either via a reactive carbene intermediate, which nonspecifically inserts into the bonds of any amino acid, or via the diazo isomer, which specifically reacts with acidic amino acids in a pH-dependent manner; indeed, for aliphatic amino acids, the diazo route has been shown to dominate over the carbene route.⁵¹ The combination of an alkyne and a diazirine in a “bifunctional” lipid derivative then allowed “Flash & Click”: to photo-cross-link the lipid to its target proteins, fix the cells, and tag the conjugate with a fluorophore to determine the lipid location. “Flash & Click” was successfully applied to cholesterol,⁴⁹ sphingosines,⁵² and ceramides.⁵³ In addition to demonstrating the subcellular location, the interactomes were determined by fishing out proteins covalently connected to the lipid derivatives. Specificity was determined by outcompeting the functionalized with nonfunctionalized lipid derivatives. It needs to be stressed that the location of the photo-cross-linkable group on the lipid will likely determine which interacting proteins can be captured. For instance, while using a multifunctional PI(3,4,5)P₃ derivative,⁵⁶ we never found proteins with a pleckstrin homology (PH) domain although these are well-known to bind to this lipid. The reason was likely that the diazirine was located on C11 of the sn1 fatty acid, which presumably resulted in a location fairly deep in the membrane. Additional derivatives with different locations of the photo-cross-linkable groups are required to more completely map lipid interactomes.

Trifunctional Strategy

In a final step, our group synthesized many multifunctional lipid derivatives that are membrane-permeant and caged and bear a diazirine and an alkyne group, including diacylglycerol, palmitic acid, sphingosine,⁵⁴ and several phosphoinositides.^55,56 These probes were used in localization experiments via cross-linking and fluorescent tagging which showed that the transfer of the lipid from the ER membrane to the plasma membrane took only 30 s. Furthermore, we could demonstrate that several proteins found in the interactome screen served as transport proteins for PIP₃. These results highlight the great potential these fully synthetic tools have for discovering lipid behavior in cells. Such findings would be very difficult to make any other way. This opens the possibility to now investigate other phenomena involving changes in lipid location, concentration, and metabolism. Observations such as the fact that phosphoinositides reside inside the nucleus despite the absence of a membrane system have hitherto remained tantalizing mysteries; with a trifunctional phosphoinositide, nuclear binding proteins of these lipids could be directly identified, leading to a more thorough understanding of their function. Mysteries such as this are persistent in cell biology and, pressingly, virology.

Lipids and the Virus Life Cycle

While there are as many variations on intracellular pathogenesis as there are viruses, the overarching steps of the viral life cycle are largely conserved, and the vast majority of them rely on lipids in some way (Figure 2). All viruses must find some way to enter the host cell, by either hijacking host receptors for endocytosis or directly fusing with the plasma membrane. Once inside, the virus must establish itself at a place where it can replicate its genome and express its proteins. After this, it must assemble the virus particle, which includes at minimum the genetic material and structural proteins, and often a host-derived lipid envelope as well. Finally, assembled virions must leave the host cell, in some cases undergoing some sort of maturation process along the secretory pathway. At this stage of our understanding, we have a collection of interesting anecdotes describing how individual viruses require lipids at various stages of their life cycles, but the mechanistic details still largely elude us.

Figure 2.

A generalized RNA virus life cycle and the cell membranes involved. Viruses enter their host cells typically by either receptor-mediated endocytosis (1a) or direct attachment to the plasma membrane (1b). Fusion with either the endosomal membrane (2a) or the plasma membrane (2b) releases genetic material from the virus into to cytosol, where host ribosomes are engaged for protein translation (3). Viral nonstructural proteins manipulate the cellular machinery to create membranous compartments (4), often referred to replication complexes, where the virus replicates its genetic material and produces structural proteins (5); a micrograph showing this phenomenon is taken from ref (63). Enveloped viruses derive their envelopes from host membranes (6a), and then often leave the cell using the host exocytic pathway (7a). Nonenveloped viruses typically leave their host cells either by packaging themselves in PS-rich vesicles (6b and 7b), or by causing lysis of the host cell and leaving as free virions (6c and 7c). ER = endoplasmic reticulum, Ve = virus-induced vesicles, Vi = single virus particles, CM = convoluted membranes, T = membrane tubules.

Virus Entry

Any virus, DNA or RNA, enveloped or not, must cross at least one host lipid bilayer to access the compartment in which it replicates (Figure 2.1 and 2.2). Specific lipids are critical to this step in several ways, and different viruses have different requirements for individual lipids. For example, alphaviruses and flaviviruses both rely on Class II fusion proteins for membrane access, but have different lipid requirements of their target membranes: the alphavirus E1 fusion protein requires cholesterol for efficient fusion, while the flavivirus E fusion protein does not;⁵⁷ however, the flavivirus E protein does require lipids of negative curvature, while lipids of positive curvature inhibit fusion.⁵⁸ Many viruses require cholesterol at the site of entry (vaccinia virus⁵⁹ and Ebola virus,⁶⁰ among others), or require cholesterol in order to bind to their entry receptor (in the case of the LASV GP protein and its LAMP1 receptor.⁶¹ Lipids can also play a role on the other side of the equation: some viruses, including vaccinia virus⁶² and Ebola virus,⁶³ display PS on the surface of the envelopes (a strategy called “apoptotic mimicry”) and can use PS receptors to gain entry into their target cells.

Replication

One of the fundamental characteristics that distinguish different viral life cycles is the location of replication. DNA viruses such as herpesviruses usually remain in the nucleus,⁶⁴ where they use host polymerases without much interference, and retroviruses such as HIV can integrate into the host genome⁶⁵ and remain undetected for years. Other viruses, meanwhile, are restricted to the cytosol, where they are much more vulnerable to antiviral defenses. In order to safely replicate in this hostile environment, RNA viruses engineer vast and complicated membranous structures where they can concentrate their own enzymes and mask themselves from detection by the host⁶⁶ (Figure 2.5). These structures can be observed by electron microscopy in cells infected with flaviviruses and coronavirus,⁶⁷⁻⁷¹ and some of the virus-induced structures seem to be the site of viral RNA replication, suggested by the presence of the replication intermediate dsRNA and many viral nonstructural proteins at these sites. The scale of the membrane rearrangement is truly astounding and, perhaps unsurprisingly, is accompanied by global remodeling of the host cell’s lipid composition, as has been shown comprehensively for ZIKV⁷² and SARS-CoV-2.⁷³ Some specific lipid species have also been implicated in the formation and maintenance of the replication organelles of many (+)-sense RNA viruses, notably PI(4)P and cholesterol. Many picornaviruses, including coxsackie virus B3,⁷⁴ Aichi virus,⁷⁵ and rhinovirus⁷⁶ recruit PI4KIIIβ to enrich PI(4)P, and in some cases also cholesterol, in their replication organelles; hepatitis C virus⁷⁷ and EMCV⁷⁸ use PI4KIIIα to achieve the same function. These observations offer the tantalizing hypothesis that (+)-sense RNA viruses in general remodel the host lipid composition in order to enrich their hosts with lipids with the biophysical properties they need to produce highly vesicularized and intensely curved membrane structures. Such a hypothesis has remained difficult to test directly in the absence of tools that would enable the visualization of specific lipids in infected cells. In the future, functional lipid probes will be employed to investigate lipid compositions in infected cells (for example, ceramide in the case of ZIKV and SARS-CoV-2, and PI(4)P in the case of various picornaviruses), and to investigate the association of these lipids with viral replication centers using superresolution fluorescence microscopy or correlative light and electron microscopy (CLEM).

Assembly and Egress

After (and in many cases during) replication, viruses must assemble the virions that will go on to infect other cells. A key distinction here is between enveloped and nonenveloped viruses, as enveloped viruses must acquire an appropriate lipid envelope with correctly inserted and/or peripherally associated membrane proteins. For example, Ebola virus hijacks host scramblases to ensure that the plasma membranes it buds from contains PS at the outer leaflet to produce sufficiently infectious virions.⁶³ Some viruses use other lipid structures as a platform for assembly: the hepatitis C virus has been shown to associate strongly with lipid droplets during assembly,⁷⁹ where viral RNA and the Core structural protein colocalize. Again, however, details of the molecular interactions between lipids and viral proteins remain elusive; as above, localization of viral proteins to lipid-based structures can be observed, but the involvement of individual lipid species cannot. Probes based on lipids of interest could be used to directly implicate specific lipids in the assembly of viral particles by microscopy. Egress can occur either lytically or nonlytically. Enveloped viruses often hijack the host secretory pathway to leave via exocytosis⁸⁰ (Figure 2.7a); nonenveloped viruses in some cases require cell lysis to release capsidated viral particles into the extracellular milieu^81,82 (Figure 2.7c). However, it has also been shown that some enteroviruses can be transmitted as collections of viral particles in autophagosome-derived extracellular vesicles⁸³ (Figure 2.7b).

Protein–Lipid Interactions

The genomes of viruses are littered with (putative) lipid-interacting proteins. For example, transmembrane proteins represent 7 out of the 10 flavivirus proteins (and an eighth contains a lipid-binding channel)⁸⁴ and at least 12 of the ∼29 coronavirus proteins.⁸⁵ Most of the lipid–protein interactions are inferred, however, or nonspecific; and in cases where specific lipid–protein interactions have been documented (as between the SARS-CoV-2 S protein with linoleic acid,⁸⁶ the Marburg Virus VP40 protein⁸⁷ and the HIV envelope protein gag with phosphatidylinositol 4,5-bisphosphate⁸⁸), it has required intense, in vitro biochemical analytical techniques such as cryoelectron microscopy and hydrogen–deuterium exchange mass spectrometry.

Trifunctional lipid probes have the potential to completely revolutionize this type of study. For one thing, they can be used to document lipid–protein interactions in situ, that is, in a living, infected cell. This offers the opportunity not only to validate the interactions of specific lipids with specific viral proteins but also to demonstrate how lipid interactions with host proteins might change during the course of an infection. Furthermore, they can be used as a screening tool: a way to reveal hitherto unsuspected lipid–protein interactions, especially with host cell proteins. Of the transmembrane viral proteins alluded to above, very few have well-elucidated functions, and many are completely mysterious.

Trifunctional probes could be used in two main types of experiment in the context of viral infection. First, they can be used to track subcellular lipid location in a living, infected cell. After infection, probe treatment, followed by uncaging, photo-cross-linking, cell fixation, and click chemistry with a dye, tags the individual lipids in a time-resolved manner. Visualization of lipids this way is compatible with downstream antibody-based immunofluorescence, allowing the lipid location to be compared to other host or viral markers. Alternatively, probe-based fluorescence can be paired with electron microscopy (in correlative light and electron microscopy, CLEM), to compare lipid location to the virus-induced membrane structures of the flaviviruses and coronaviruses. A second type of experiment involves identifying the protein binding partners of lipids during infection. Here, after infection cells are treated with probe, uncaged, photo-cross-linked, and then clicked to some sort of substrate for pulldown, either biotin azide or azide-agarose beads. The enriched lysate can then be aggressively washed to remove non-cross-linked proteins prior to reduction, derivatization, and trypsinization for LC-MS/MS analysis. A workflow describing both types of experiments is depicted in Figure 3.

Figure 3.

Workflow for a “Flash & Click” experiment to interrogate lipid-virus interactions, including infection, probe treatment, subsequent photoreactions to activate the probe (“uncaging”), photo-cross-linking, and click chemistry-based derivatization for subsequent microscopy or proteomics-based analyses.

Conclusion and Outlook

“Flash & Click” has developed into one of the most useful techniques for studying small molecule–protein interactions in the intact cell environment. It reveals the location of the tool at the time of photo-cross-linking when combined with fluorescent tagging, and it helps identifying specific binding proteins by mass spectrometry after extraction of the cross-linked conjugates. “Flash & Click” is particularly useful for studying lipid location and interactomes because of the intrinsic difficulties in tracking lipids in intact cells. There are at least 770 proteins featuring a lipid binding domain⁸⁹ and about 5,400 transmembrane proteins in the human proteome (27%), not counting proteins that only indirectly associate with membranes. Each of these proteins directly interacts with lipid molecules, in most cases in a highly dynamic fashion. Hence, lipid–protein interactions are essential for the proper function of receptors, ion channels, the formation, direction, and fusion of vesicles, and the many intracellular signaling events. As such, there is a need for a “Flash & Click” tool for every known lipid species in order to determine its function. Once the tools have been prepared, they will be particularly useful for comparing healthy and diseased cells. Due to the large coverage of proteins, the differential interactomes will provide a highly useful basis for better understanding disease consequences and for targeting pharmacological intervention. These data sets will complement transcriptome, proteome, and lipidome data.

For the virologist, the type of information that could be gleaned from this type of experiment is rich in many ways. Looking at the virus life cycle from the point of view of a specific lipid centers a largely unregarded aspect of cell biology, which is nonetheless fundamental to viral survival, and, critically, encompasses an array of potentially druggable enzymatic processes. A comprehensive understanding of lipid–protein interactions could lead the way to the design or application of novel host-targeted therapeutics, or the repurposing of existing molecules for an antiviral effect. Host-targeted therapies in general are less susceptible to the development of resistance, especially when they target host pathways that are fundamental to viral survival. The same therapeutic molecule could even potentially be applied to different viruses with similar infection strategies.

The authors declare no competing financial interest.