Distinct Roles for Carbohydrate and Protein Receptors in Coronavirus Infection

Loyola University Chicago Loyola eCommons

Dissertations Theses and Dissertations

2020

Enya Qing

Follow this and additional works at: https://ecommons.luc.edu/luc_diss

Part of the Virology Commons

Recommended Citation Qing, Enya, "Distinct Roles for Carbohydrate and Protein Receptors in Coronavirus Infection" (2020). Dissertations. 3818. https://ecommons.luc.edu/luc_diss/3818

This Dissertation is brought to you for free and open access by the Theses and Dissertations at Loyola eCommons. It has been accepted for inclusion in Dissertations by an authorized administrator of Loyola eCommons. For more information, please contact [email protected].

LOYOLA UNIVERSITY CHICAGO

DISTINCT ROLES FOR CARBOHYDRATE AND PROTEIN RECEPTORS IN

CORONAVIRUS INFECTION

A DISSERTATION SUBMITTED TO

THE FACULTY OF THE GRADUATE SCHOOL

IN CANDIDACY FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

PROGRAM IN MICROBIOLOGY AND IMMUNOLOGY

ENYA QING

CHICAGO, ILLINOIS

MAY 2020

ACKNOWLEDGEMENTS

I would like to thank my mentor Dr. Tom Gallagher (TG) for his invaluable patience and mentorship. I had no interest in viruses until TG introduced them to me, and for that I count myself among the luckiest. His pure scientific enthusiasm and intellectual cognizance have not only matured me into the scientist I am today, but its remanence will guide me still in the years to come. I would emphasize my gratitude for my former lab mates Drs. Jung-Eun Park, Jim Earnest, and Mike Hantak for their insightful discussions, as well as current members Dr. Binod Kumar, Grant Hawkins,

Tom Kicmal, and Emily Timm for carrying the torch forward. I would also like to thank

Drs. Susan Uprichard, Ed Campbell, Karen Visick, and Seth Robia for their guidance

while on my dissertation committee. I also thank Drs. Stanley Perlman, Paul McCray,

and Luis Enjuanes for their persisting enthusiasm with my project. Furthermore, I owe

my gratitude to everyone in the Department of Microbiology and Immunology, for they

all contribute to this nurturing environment for future superlative scientific minds.

Most importantly, I would like to thank those whom I hold dearest. I thank my

parents for raising and supporting me to follow my heart, my in-laws for treating me as

one of their own, my children for being the purpose of my life and dreams, and my wife

especially, whose mere presence embodies a home I can always return to, whose

unwavering support is the single most crucial foundation for all that I do, and to whom I

will forever be in debt. I thank all for making this experience one that I have enjoyed

through and through, truly.

III

To Colleen, Felix, Hazel, Hugo, and Evander

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... III

LIST OF FIGURES ...... V

ABSTRACT ...... X

CHAPTER I: REVIEW OF LITERATURE ...... 1 Coronaviruses ...... 1 Coronaviruses Threatens Global Health ...... 3 Coronaviruses are Zoonotic Pathogens ...... 4 Bats are the Most Prominent CoV Reservoirs ...... 7 Enveloped Virus Entry Requires Membrane Fusion...... 9 Coronavirus Entry...... 14 Spike Protein is the Major Determinant of CoV Tropism ...... 17 CoV Spike Receptor-binding Domains ...... 17 The utilization of CoV spike receptor-binding domain A ...... 19 The utilization of CoV spike receptor-binding domain B ...... 22 The dual-utilization of CoV spike receptor-binding domains ...... 24 CoV spike receptor-binding domains enable viral zoonosis ...... 27 CoV Spike Membrane Fusion Mechanism ...... 30 CoV spike membrane fusion is triggered by proteases ...... 30 CoV spike proteolytic sensitivity dictates entry kinetics ...... 33 CoV tropism depends on protease utilization ...... 36 Coronavirus Spread ...... 37 Purpose of Dissertation ...... 41

CHAPTER II: MATERIALS AND METHODS ...... 43 Cells ...... 43 Viruses ...... 43 Pseudoviruses ...... 44 Virus-like Particles (VLPs) ...... 44 Particle Concentration ...... 45 In vitro Fusion Assay ...... 46 Hemagglutination Assay ...... 46 Neuraminidase Treatments ...... 47 VLP Binding Assay ...... 47 VLP Entry Assay ...... 48 Virus and Pseudovirus Particle Entry Assays ...... 49 Cell-cell Fusion Assay ...... 49 Fc Constructs ...... 50 Fluorescence Staining ...... 51 Western Blotting ...... 51 Microscopy and Image Acquisition ...... 52

Statistical Analysis ...... 52

CHAPTER III: RESULTS ...... 53 SECTION 1: CoV S, Receptor, and Protease Form the Basic Elements for Membrane Fusion ...... 53 CoV Entry is Spike-directed, Receptor-facilitated, and Protease-triggered ...... 53 FRET-based in vitro Characterization of Viral Fusion ...... 55 MERS-CoV S, hDPP4, S15-GFP, and S15-mCherry are Incorporated into HIV pps ...... 56 Virion Fusion Minimally Requires CoV S, Protein Receptor, and Protease ...... 58 SECTION 2: MHV-JHM Spike Utilizes Host Sialosides During Viral Entry and Spread ...... 67 MHV-JHM CoV Agglutinates RBCs ...... 67 JHM Spike Protein Engages Sialosides on Target Cells ...... 70 JHM Spike Requires Protein Receptor for Membrane Fusion During Viral Entry ...... 75 Virus Binding to Sialic Acids Facilitates Protein Receptor-dependent Entry ...... 78 Sialic Acid Receptors Promote JHM Viral Spread without Requiring mCEACAM Protein Receptors ...... 79 A JHM-CoV Spike Mutation Increases mCEACAM-independent Cell Binding ...... 82 A JHM-CoV Spike Mutation Increases mCEACAM-independent Membrane Fusion ...... 85 SECTION 3: MERS Spike Utilizes Host Sialosides During Viral Entry and Spread ...... 87 MERS-CoV Spikes Mediate Cell Fusion without Requiring hDPP4 Protein Receptors ...... 87 MERS Spike-mediated, Protein Receptor-independent Membrane Fusion Depends on S Domain A and Host Sialosides ...... 90 An Analogous MERS-CoV Spike Mutation Similarly Increases hDPP4-independent Cell Binding ...... 92 An Analogous MERS-CoV Spike Mutation Similarly Increases hDPP4-independent Membrane Fusion ...... 95

CHAPTER IV: DISCUSSION ...... 99 Summary of Data ...... 99 In Vitro Particle Fusion Identifies Minimal Fusion Requirements ...... 102 JHM Spike Utilizes Host Sialosides for Viral Attachment ...... 104 Differential Host Sialoside Utilizations for Viral Attachment and Cell-Cell Fusion ...... 106 Host Surface Protease Utilization and CoV Pathogenisis ...... 108 The Evolution of CoV S Protein Domain A ...... 109 The Co-Evolutionary Relationship between CoV S Protein and HE ...... 111 The Mechanism of the S Protein Domain A Gain-of-Function Mutations ...... 112 Pathogenesis of Sialoside-utilizing CoVs ...... 114

CoV S Protein Domain A and Zoonosis ...... 115

REFERENCE LIST...... 118

VITA…………………………………………………………………………………… ...... 141

VII

LIST OF FIGURES

Figure 1. Murine Hepatitis Virus (MHV) Genomic and Virion Structure ...... 2

Figure 2. The Progression of Membrane Fusion ...... 11

Figure 3. Coronavirus Spike Protein...... 16

Figure 4. Coronavirus S1 Subunit ...... 21

Figure 5. Coronavirus S2 Subunit ...... 32

Figure 6. Coronavirus Entry Routes Depend on Protease Utilization ...... 35

Figure 7. Coronavirus Infection Spread ...... 38

Figure 8. The Role of S, Protein Receptor, and Protease During CoV Entry ...... 54

Figure 9. FRET-based In Vitro Characterization of Viral Fusion ...... 57

Figure 10. HIV pps Incorporate MERS S, hDPP4, S15-GFP, and S15-mCherry ...... 59

Figure 11. Functional Assessment of Factors Used in In Vitro Fusion Assay ...... 61

Figure 12. In Vitro Viral Membrane Fusion Assay ...... 63

Figure 13. FRET Induction Reflects S15-GFP – S15-mCherry Compartmentalization ...... 65

Figure 14. In vitro Fusion Assesses the Sufficiency of Proteases in S-mediated Membrane Fusion ...... 66

Figure 15. MHV-JHM CoV Agglutinates RBCs ...... 69

Figure 16. MHV-JHM VLPs Resemble Authentic Viruses ...... 72

Figure 17. JHM S Protein Engages Sialosides on Target Cells ...... 74

VIII

Figure 18. Viral Binding to Sialosides Facilitates Protein Receptor- dependent Entry ...... 77

Figure 19. JHM-CoV Cell-to-cell Spread Depends on Host Sialic Acids ...... 81

Figure 20. JHM S G176E Mutation Elevates Target Cell Binding ...... 84

Figure 21. JHM S G176E Mutation Increases mCEACAM-independent Cell-cell Fusion ...... 86

Figure 22. MERS S Mediates Cell Fusion without Requiring hDPP4 Protein Receptors ...... 89

Figure 23. MERS hDPP4-independent Cell-cell Fusion Depends on Host Sialosides and S Domain A ...... 91

Figure 24. MERS S N222 is a Hypervariable Residue within the Sialoside-binding Distal Site ...... 93

Figure 25. MERS-CoV VLPs Resemble Authentic Viruses ...... 94

Figure 26. MERS S N222D Elevates Target Cell Binding ...... 96

Figure 27. MERS S N222D Increases hDPP4-independent Cell-cell Fusion ...... 98

Figure 28. Sialic Acid and Protein Receptor Binding Events During CoV Infection ...... 106

ABSTRACT

Coronaviruses (CoVs) are common human and animal pathogens causing a range of organ pathologies including respiratory, gastrointestinal, and brain diseases.

In humans, four endemic CoV species together account for one third of mild respiratory infections worldwide. More severe and frequently fatal respiratory pathologies are caused by recent CoV outbreaks that resulted from occasional zoonotic spillover from animal CoV reservoirs. Within the past 18 years, we humans have experienced three of such outbreaks. The causative agents of these outbreaks are: SARS-CoV in 2002,

MERS-CoV in 2012, and SARS-CoV-2 in 2019, with the outbreaks caused by the latter

two still ongoing. These CoVs collectively have demonstrated the spread of CoVs in

human populations, as well as highlighted the threat CoVs impose on global health as

an emerging viral pathogen. Any chance of relieving CoV’s threat on human populations

would rely heavily on our understanding of the mechanistic requirements for CoV

tropism. The major determinant for CoV tropism happens during viral entry. CoVs have

evolved to use a single viral protein for entry: spike (S). The S protein binds viruses to

host cell receptors and catalyzes virus-cell membrane fusion, an entry requirement for

all enveloped viruses. Uniquely, CoV S proteins contain at least two receptor binding

domains, a domain A that generally engages host sialic acids, and a domain B that

recognizes host transmembrane proteins. A putative advantage of this bivalent binding

is elevated CoV zoonotic potential, for each binding domain can, theoretically,

independently evolve affinity to distinct host factors. To test this hypothesis, we aimed to

identify roles for each receptor for the S proteins of two beta-coronaviruses, the prototypic MHV-CoV strain JHM and human MERS-CoV. We focused on three distinct stages of the CoV life cycle: (1) CoV particle-cell binding; (2) CoV particle-cell entry; (3)

CoV cell-to-cell spread, where the infection passes directly from an infected cell to a neighboring healthy cell via cell-cell fusion, or syncytia formation. For MHV-JHM S protein, we identified its interaction with host sialic acids, using a virus-like particle

(VLP) system. The interaction with sialic acids, and not with the protein receptor, is responsible for the majority of the particle binding mediated by the S protein.

Additionally, the S-sialic acid interaction assists in JHM-CoV entry into various target cells, as well as the subsequent cell-to-cell spread, both in the presence and absence of the protein receptor. For the S protein of MERS-CoV, we first found that viral particles minimally require S, protein receptor, and protease for fusion. Additionally, similar to

JHM, MERS S-sialic acid interaction is sufficient for cell-cell fusion. Lastly, we identified single-nucleotide changes in both S protein-sialic acid binding S domain As that conferred elevated cell-binding and cell-cell fusion capabilities. Overall, these findings reveal roles for sialic acids in virus-cell binding, viral S protein-directed cell-cell fusion, and resultant spread of CoV infections. This study highlights distinctive CoV receptor binding domain activities in the infection process and raise the possibility that the lectin- like properties of many CoVs are responsible for facile CoV zoonotic transmission and intercellular spread within infected organisms.

CHAPTER I

REVIEW OF LITERATURE

Coronaviruses

Coronaviruses (CoVs) are positive-sense, single-stranded RNA (ssRNA) viruses that contain the largest genome of all RNA viruses, which ranges from 24 to 32 kilobases (kb). The 5’ two-thirds of the CoV genome encodes the various non-structural proteins, and is expressed as a polyprotein. The 3’ one third of the CoV genome encodes the structural and accessory proteins, whose expressions result from the generation of viral subgenomic RNAs (Figure 1A). CoVs are phylogenetically categorized under the order Nidovirales and are subcategorized into four groups (alpha-

, beta-, gamma-, and delta-).

CoVs are enveloped viruses, meaning the mature virion consists of an outer- most layer made up of a lipid bilayer derived from the host membrane, enriched with transmembrane viral proteins. Indeed, three out of four CoV major structural proteins and sometimes a non-major structural protein can be found on the CoV envelope.

These are the spike (S) protein, membrane (M) protein, envelope (E) protein, and hemagglutinin esterase (HE). The nucleocapsid (N) proteins associates with the CoV genome and, through its interactions with the self-assembled M proteins, resides inside the viral particle (1–6). Together with the CoV genomic RNA, these structural proteins form a complete, infectious virion (Figure 1B).

Figure 1. Murine Hepatitis Virus (MHV) Genomic and Virion Structure. (A) CoV genomic (top) and subgenomic (bottom) structures, using murine hepatitis virus (MHV) as an example. CoV subgenomic RNAs are generated by joining of the 5’ leader (red box) sequence to various downstream sites. CoV non-structural genes consists of the 5’ two-thirds of the RNA genome, whereas structural genes are located within the 3’ one- third. During viral replication, non-structural genes are expressed as polyproteins, while structural genes are expressed by each subgenomic RNAs. Accessory genes are numbered. (B) Schematic for CoV virion composition. Viral structural proteins S, E, M, and sometimes HE, are incorporated in the viral envelope. N proteins associate with the genomic RNA (gRNA) and are together incorporated into the virion interior. The color codes are consistent between panel A and B.

CoVs are highly adaptive, in part due to their error-prone RNA-dependent RNA polymerase (RdRp), whose substitution rate has been averaged to ~10-4 per year per

site, with even higher rates for certain hypervariable regions (i.e. S protein), at 0.6 x 10-2

per site per year (7, 8). Another contributor for CoV adaptability is the RNA recombination during viral replication, via the process of RdRp template-switching (9–

12). Together with a large genome size, which tolerates higher genome variations, and the generation of subgenomic RNAs during viral replication, which elevates the rate of

homologous recombination between different viral lineages, CoVs are capable of

quickly adapting to a wide range of biological niches.

Coronaviruses Threatens Global Health

CoVs were initially isolated in 1930s from chickens (13), and 1940s from swine

(14) and mice (15). Human CoVs were first identified in 1967 from tissue samples of the

upper respiratory track (16). To this day, there are four CoVs that cause mild respiratory

track infections in humans, with occasional fatalities with immunocompromised

individuals. These are human coronavirus 229E (229E-CoV (16)), human coronavirus

OC43 (OC43-CoV (16)), human coronavirus NL63 (NL63-CoV (17, 18)), and human

coronavirus HKU1 (HKU1-CoV (19)). They together account for about one-third of

common colds worldwide (20, 21), highlighting the spread of CoVs in human

populations.

CoVs responsible for more recent human outbreaks, however, cause severe,

often fatal acute respiratory syndromes (22–29). In November 2002, a CoV epidemic

started in the Guangdong province of southern China. The virus was moderately apt at

human-human transmission, and had spread to 26 countries and infected over 8000

individuals, with 15% fatality (28). This epidemic was fortunately contained and

subsequently eliminated via effective epidemiologic approaches (30). The CoV in

question was later identified and designated Severe Acute Respiratory Syndrome

(SARS)-CoV. This initial demonstration of the extent to which CoVs can threaten global health was echoed only a decade later in the Middle East. In 2012, another CoV epidemic broke out in and around Saudi Arabia, causing lethal pneumonia (26, 31).

Since then, this epidemic has mainly been contained in the Middle East, around the

Arabian Peninsula, with Saudi Arabia containing the most cases. Since the onset of the

outbreak in 2012, however, the virus has been occasionally exported to other countries

via infected individuals, sometimes causing clusters of secondary outbreaks, the largest

of which happening in South Korea in 2015. The causative agent for this epidemic was

identified and designated Middle-East Respiratory Syndrome (MERS)-CoV, and as of this writing, the epidemic is still ongoing. As of November 2019, MERS-CoV has caused a total of 2494 laboratory-confirmed cases, with 34% fatality (32). Currently (as of Feb.

16, 2020), a pneumonia outbreak originated from Wuhan, China, in ongoing, infecting over 71,000 individuals, over 11,000 in serious or critical conditions, and over 1,700

fatalities. This outbreak is identified to be caused by a novel SARS-like CoV (2019-

nCoV, or SARS-CoV-2), which has been sequence-identified in a large number of

patients (33). Together with SARS-CoV and MERS-CoV, these recent human outbreaks

demonstrate the threat CoVs impose on global health as an emerging viral pathogen.

Coronaviruses are Zoonotic Pathogens

CoVs primarily infect the respiratory and gastrointestinal tract and cause endemic

infections in a wide range of mammalian and avian species. For any individual CoV, it

mostly appears to be restricted to a narrow host range, species, or even tissue. This is

likely because all current viruses are results of highly-specific adaptations, and therefore

do not actively display their zoonotic potentials, or the ability to jump from one animal

species to another. However, viral genome sequencing and subsequent phylogenetic

analyses reveal that CoVs have crossed the zoonotic barrier frequently (34, 35).

In fact, all human CoVs share a common ancestor with bat CoVs, suggesting

that, throughout history, bat CoVs were transmitted to humans. These transmissions

could be direct, or indirectly through an intermediate animal species. For example,

ancestral 229E-CoV likely resided in bats up until around 200 years ago, when it

transmitted and became endemic in an intermediate host, possibly camelids, before it

eventually spread to humans (36). OC43-CoV, on the other hand, is the result of a

recent (~1890s) cross-species transmission event of bovine (B-) CoV, which has been endemic in cows (37, 38). Similarly, ancestors of NL63-CoV and HKU1-CoV both have successfully breached the species barrier and are maintained in the human populations worldwide (21). These four human CoVs, which cause mild respiratory tract infections and are distributed in human populations globally seemingly without animal reservoirs, are considered well-adapted to humans.

On the other hand, those CoVs that have only recently entered into human populations tend to cause more severe or fatal pneumonia, and are considered not-well- adapted. Such CoVs are likely maintained in zoonotic reservoirs, and via intermediate host species, occasionally spillover into the human population. In the past 18 years, the world witnessed three such zoonotic events. For example, epidemiology data had traced the 2002 SARS-CoV outbreak to the masked palm civets (Paguma larvata) from local live animal markets (39–41). Interestingly, while the civets in the markets were infected, wild civets were largely SARS-CoV-negative (40). This suggests that the virus was not endemic in the masked palm civets, who are merely the intermediate host, and not the reservoir, for the human transmission. Conversely, it was identified that SARS-

CoV and a number of SARS-like CoVs were present in the region’s Chinese horseshoe

bats (Rhinolophidae family), suggesting their role as the animal reservoir for the virus

(12, 42, 43). The current zoonosis model for SARS-CoV describes a transmission path

from bats to civets, and then civets to human.

A similar zoonotic trajectory was proposed for MERS-CoV. Immediately after the onset of the MERS outbreak in 2012 and the identification of the virus, dromedary camels were found to contain neutralizing antibodies against MERS-CoV. This is the case for both camels in the outbreak zone of the Arabian Peninsula and camels that had originated there but had since been exported around Africa (44, 45). Unlike the case with SARS-CoV and civets, this finding suggests that dromedary camels are likely the animal reservoir. Additionally, live MERS-CoV can be readily isolated from the nasal swabs of these camels (46–48), suggesting that they are the likely source of the viral infection, possibly via contaminated body fluids of the animal’s respiratory tract. The reason the viral outbreak happened in the Middle East and not in Africa is potentially due to the differences in the regional concentrations of camels and their proximity to large human populations. In support of the claim that camels are the animal reservoirs for MERS-CoV, a recent survey of camel CoVs in Saudi Arabia revealed that multiple lineages of MERS-CoVs co-circulate within these animals. Evidently, these co- circulating MERS-CoV strains had recombined and resulted in a MERS-CoV strain that had dominated the camel population by 2014. Most intriguingly, the same recombinant

MERS-CoV was responsible for the viral outbreaks in the year of 2015 (49).

Interestingly, camelids might not be the only animal in the outbreak zone that can carry

7 the virus. While MERS-CoV has yet to be found in bats to date, viral genome fragments identical to MERS-CoV have been identified in bats around the Arabian Peninsula

(Taphozous perforatus (50)) and South Africa (Neoromicia zuluensis (51)). Additionally,

CoVs that display the closest phylogenetic relationship with MERS-CoV are bat-CoVs

HKU4 (52) and HKU5 (53). Taken together, these findings support a MERS-CoV zoonosis model that is analogous to SARS-CoV, where the ancestral MERS-CoV spread from bats to camelids, and then camelids to human.

As of this writing, more information on the SARS-CoV-2 outbreak is being updated daily. While there has not been conclusive analyses on the source(s) and reservoir(s) of the virus, several patients were associated with a local seafood market

(33). Additionally, the SARS-CoV-2 genome is most closely related to one SARS-like

CoV in Bats (54). The specific epidemiological relationship between SARS-CoV-2, the seafood market, and bats is an area of great importance. Taken together, human CoVs, including four endemic stains and three outbreak-causing animal-spillovers, have demonstrated the level of viral prevalence, facile zoonotic transmission, and potential to cause severe respiratory disease. This brings urgency to research aimed at discovering

CoV infection mechanisms, and specifically, CoV features that prime zoonosis.

Bats are the Most Prominent CoV Reservoirs

All human CoVs can be genetically traced back to bats, which has been estimated to be the largest reservoir of alpha- and beta-CoVs (34). Indeed, out of the numerous animal species that have been surveyed, bats have risen out to be one of the most abundant source for novel viral sequences (55). There may be several key

8 characteristics that contribute to this understanding. First of all, bats are among the oldest and most diverse mammals. Being of the order Chiroptera, bats comprise about

20% of all classified mammalian diversity, with over 1,200 species (56). This diversity offers encompassing biological niches for viral adaptation. Secondly, bat ecology ranges from isolated individuals to mega-colonies occupying diverse niches that can span thousands of miles, greatly enabling viral spread and co-circulation. Third, for reasons unclear, despite being identified with such a diverse assortment of viruses, bats rarely manifest disease. One attractive explanation links this phenomenon to the ability of flight (57). Being the only flying mammal, bats produce a large amount of reactive oxygen species (ROS), and hence have evolved to limit oxidative stress (58), which could, inadvertently, limit viral replication and pathogenesis (59). Additionally, several bat species lack or repress components of the inflammasome and natural killer pathways, which could limit disease and post-infection damage (58, 60). Furthermore, unlike most other mammals, bat interferon (IFN) is less responsive to canonical inflammatory signals (61), but instead is constitutively expressed at high levels, which could modulate disease but maintain low-level viral infections (62). Lastly, most bat-

CoVs were isolated from bat gut, a very different immunological environment than the respiratory tract, which could be more tolerant for viral maintenance. These factors likely work in unisons to shape bats into an organism highly suitable for viral adaptation and evolution.

In addition to the high tolerance to CoV presence, the immunological environment in bats may also diversify viral quasi-species gene pools. Flight-induced

ROS may act as a mutagen for CoVs by overwhelming CoV proofreading repair, or

altering RNA-dependent RNA-polymerase (RdRp) fidelity (63). Increases in the quasi-

species pool size potentiate viral adaptations into new hosts. Furthermore, the

constitutive, instead of inducible, expression of IFN in bats may select for CoVs that

have adapted for enhanced resistance to innate immunity (62). Viral adaptions such as

these may encounter little resistance, due to the bat’s lack of key inflammatory

mediators (60, 61). This may result in some viral variants with a replication advantage

when they infect a new host, possibly associated with massive and pathogenic

inflammatory responses. These phenotypes have been observed with both SARS-CoV

and MERS-CoV infections in humans (64–66). Taking together, bats host CoVs and may enhance CoV zoonotic potential by enriching viral diversity, potentially selecting for innate immunity-adaptive variants.

Enveloped Virus Entry Requires Membrane Fusion

The viral genome of enveloped viruses, including CoVs, are encased in a lipid bilayer derived from the infected host. Therefore, to deliver the viral genome into the target cell cytoplasm to establish infection, the virus and target cell membranes need to fuse together (Figure 2A). Hence, viral fusion follows the underlining principles of

membrane fusion.

Biological membranes are consist of amphipathic phospholipids, each of which

are composed of a hydrophilic head group and two hydrophobic tails. In aqueous (i.e.

H2O) environments, phospholipids aggregate via the cis-hydrophilic and cis-

hydrophobic interactions between the head groups and tails, respectively. The resulting

10 macromolecular structures position the hydrophilic head groups against water molecules, while shielding the hydrophobic tails. The most stable of these structures is a membrane vesicle, compartmentalized by a phospholipid bilayer. Due to this architecture, most hydrophilic molecules cannot penetrate the membrane bilayer.

Therefore, one approach to achieve the delivery or exchange of internal molecules of two particular vesicles, e.g. an enveloped virus and its target cell, is the melding of the lipid bilayers, or membrane fusion.

The current model describes that membrane fusion is completed in several steps (67).

First, the two membranes in question have to come into close proximity, close enough that the transfer of a few phospholipids from one membrane to another would not compromise the bilayer architecture. This is facilitated by the “positive” curvatures

(curves extending toward the target membrane), such as the membrane curvature on the virus against the membrane on the target cell (68). At this proximity, the phospholipids on the proximal leaflets of the two membranes are capable of interacting and merging, forming a continuous lipid layer, while the distal leaflets remain distinct

(69, 70). This is the state of hemifusion (Figure 2A), where the exchange of membrane lipids, but not lumen contents, is permitted. If the fusion process does not stall on hemifusion, eventually the distal lipid leaflets will merge, and form a fusion pore (Figure

2A). At this stage, the two compartments are joined and their lumen contents are mixed.

For enveloped viruses, fusion pore formation enables its viral genome to enter the host cytoplasm and hence concludes the viral entry process.

Figure 2. The Progression of Membrane Fusion. (A) Membrane fusion takes place in several steps. First, opposing membranes (made up of lipids with brown a head group and black hydrophobic tails) are brought into close proximity, likely aided by positive membrane curvatures away from compartment lumens (orange and blue). Then, outer leaflets of the membrane merge, forming a hemifusion intermediate. Finally, distal leaflets merge to complete fusion pore formation, allowing for the mixing of compartment lumens (now green). (B) During viral entry, the viral envelope is localized to the host membrane by the interactions between the viral glycoprotein (for CoVs, spike) and the host receptor. Upon exposure to subsequent host triggering factors (for CoVs, cleavage(s) of spike by host proteases), the viral glycoprotein goes through extensive conformational changes, resulting in its extension and insertion into the target cell membrane, physically bridging the two membranes together. Then, the glycoprotein completes its conformational changes by folding back onto itself, while bringing the two membranes into close proximity by creating positive membrane curvatures (see panel A). If successful, the process will proceed to hemifusion and fusion pore formation, concluding viral entry.

While a fused membrane is at a lower free energy state than two unfused ones, meaning membrane fusion is thermodynamically favorable, the energy barrier for membrane fusion is high, making it kinetically unfavorable. As the result, biological membrane fusions evolved fusion catalysts that lower this kinetic barrier (67, 71–73), resulting in highly-regulated fusion processes. The most well-studied membrane fusion catalysts are the viral glycoproteins on enveloped viruses. Viral glycoproteins are transmembrane proteins synthesized in the endoplasmic reticulum (ER). These proteins are folded in ways such that intramolecular energy is stored and maintained (71) either through internal interactions (74) or by the actions of associated viral proteins (75).

When the virus becomes associated with the target cell membrane prior to viral entry, in the presence of appropriate host triggering factors, the viral glycoprotein will go through a dramatic conformational unfolding and refolding, exhausting the stored energy in an attempt to pull the viral and target cell membrane together to enable fusion (74, 76, 77).

Additionally, because the structural transition endpoint into which the viral glycoprotein refolds is highly thermodynamically favorable, the protein will be “locked” in that endpoint conformation and becomes inactive (78). In other words, a viral glycoprotein generally can go through the conformational changes that are required for membrane fusion only once. It is therefore crucial for these glycoproteins to remain inactive on the virion until the virus reaches the environment that is ideal for membrane fusion and infection. To acquire optimal entry efficiency, enveloped viruses have evolved their glycoproteins to recognize specific cellular and environmental factors as fusion triggers.

Several host factors can be recognized by enveloped viruses as fusion triggers.

These include host cell receptors, local proton concentrations, and proteases. Some viruses only require a single trigger for fusion. For example, influenza A virus (IAV),

West-Nile virus (WNV), and vesicular stomatitis virus (VSV) glycoproteins are triggered by elevated local proton concentration, a feature in the host endosome (74, 79, 80).

Other viruses may require multiple triggers for fusion. Human immunodeficiency virus type 1 (HIV-1) glycoprotein, via sequential engagements with its receptor and coreceptors, undergo dramatic conformational changes to facilitate fusion (81–84).

Ebola virus (EBOV) glycoprotein requires endosomal protease cleavage before it can bind to its endosomal receptor, and fusion is only triggered after a subsequent, uncharacterized step (85–90).

Viral glycoproteins are structurally categorized into three classes (71). While the structure and triggering mechanisms for these viral glycoproteins can vary greatly, their fusion processes, once triggered, share many common features. First, in their pre- fusion, ready-to-be-triggered conformation, viral glycoproteins bury a short stretch of hydrophobic residues called a fusion peptide (91–93). Upon triggering, the protein initiates dramatic structural transitions, which unveil the fusion peptide and extend it towards the target cell membrane. If the target cell membrane is within range, the fusion peptide will insert into this membrane and, due to its hydrophobicity, be anchored among the hydrophobic tails of the phospholipids. At this stage, the viral glycoprotein assumes a conformation known as the extended fusion intermediate. This conformation is putatively short-lived, because to complete the structural transition the glycoprotein

14 requires its transmembrane domain to coincide with the fusion peptide (94–96).

Therefore, the glycoprotein then refolds onto itself. Since the transmembrane domain is on the viral envelope and the fusion peptide is in the target cell membrane, this refolding pulls the two membranes into close proximity. Then, the outer leaflets of the lipid bilayers meld to establish hemifusion (69), and if this progresses, the inner leaflets also meld to form a fusion pore (97). Due to the high energy barrier for membrane fusion, it is likely that most if not all enveloped viruses require the simultaneous refolding of multiple glycoprotein complexes to establish the fusion pore (74, 79, 80). Once the fusion pore is formed and maintained, the viral genome can invade the host cell cytoplasm, and hence concludes the process of viral entry.

Coronavirus Entry

As an enveloped virus, CoV entry requires membrane fusion. The mechanism by which CoVs enter host cells may explain, in part, the remarkable expansion of these viruses into new hosts, including humans. For any particular virus to achieve cross- species transmission, it is generally accepted that four criteria need to be met (98).

First, to achieve a specific cellular tropism, the target cell need to be susceptible, in other words, expressing the appropriate receptor needed for entry. Second, the target cell needs to be permissive, that is to say, allowing the virus to replicate so as to complete its life cycle. Third, these susceptible and permissive cells need to be accessible, in a biologically meaningful way, to the virus. Lastly, the innate immune response of a particular host must tolerate the early rounds of viral growth. For CoVs, while their non-structural gene products govern viral replication and the accessory

proteins assist in host tropism and adaptation (Figure 1A), the CoV entry process (i.e.

host susceptibility) appears to be the main determinant for the success of initial cross-

species infection events.

CoVs have evolved to use a single viral protein for entry: the spike (S) protein

(Figure 2B). The S proteins adhere viruses to host-cell receptors, and they

subsequently function as catalysts of virus-cell membrane fusion (99, 100). S protein is a class I viral fusogen (101), sharing the category with IAV hemagglunin (HA), HIV gp160, and EBOV GP (71). Like other class I fusogens, CoV S folds into a metastable pre-fusion conformation in a homotrimeric structure (Figure 3). S proteins are large N- glycosylated glycoproteins, and among different CoV species it can vary greatly in size, ranging from 1300 to 1600 amino acid residues, and 160 to 220 kDa (99, 100). Native S proteins are trimeric, and form 18-23 nm long, club-shaped spike protrusions out of the membrane. This can be clearly observed under the electron microscope (EM) as bulbous surface projections on the viral envelope (102). CoV S is a type I transmembrane protein: a large N-terminus ectodomain, a single-pass transmembrane domain, and a short C-terminus tail. Functionally, the S protein can be divided into two distinct subunits: A globular membrane-distal S1 subunit (Figure 3, right) that is involved in host attachment and receptor recognition, and the membrane-associated S2 subunit

(Figure 3, right) that facilitates membrane fusion. This conceptual separation of S1 and

S2 is a reflection of the natural arrangement. All identified CoV S proteins contain a cleavage site between S1 and S2, which is recognized by host proteases.

Figure 3. Coronavirus Spike Protein. The CoV virion structure is depicted as shown in Figure 1 (top). One single S protein functional unit (red box) is a trimer of S monomers. Each S monomer is depicted in surface representation and colored green, cyan, or magenta. S trimers are shown both from profile (center) and top-down (left) vantage points. The S trimer is formed by S monomers interlacing onto each other. The magenta monomer is re-depicted as cartoons of secondary structures (right). The receptor- engaging S1 subunit is colored yellow, and the membrane-fusing S2 subunit is colored gray. The S protein of MHV (PDB: 3JCL) is used as an example for generalizing CoV S structures.

The S proteins cannot facilitate membrane fusion with S1 covalently linked to S2, and only after the liberation of the two domains that is the S1/S2 cleavage, either during virion assembly or viral entry, can viral fusion be initiated (Figure 2B). The CoV membrane fusion process is initiated by a further S2’ cleavage. Traditionally, the study of S protein structures was limited to stable small protein fragments that accommodates

X-ray crystallography (103–107). In recent years, the advancement of single-particle

cryo-electron microscopy has enabled the elucidation of the native S protein trimeric

structure in several CoV species (Figure 3), providing ample novel insights into the

structural mechanism of S protein functions (108–114), including ones that may enable

CoV zoonosis.

Spike Protein is the Major Determinant of CoV Tropism

CoV Spike Receptor-binding Domains

The key contributor to the high propensity CoVs have for zoonotic transmissions

could lie within the S1 subunit of the S protein. Recent structural studies have

determined that S protein S1 subunit displays a multidomain architecture (Figure 4A

and B). For beta-, gamma-, and delta-CoVs, their S1 subunit is structurally organized in

four distinct domains A, B, C, and D (110, 115, 116). Alpha-CoVs appear to have an

additional domain N-terminus of A designated as domain 0 (109, 117). Based on the

trimeric S structure, the solvent-exposed domains A and B (for alpha-CoV, domains 0,

A, and B) may serve as independent receptor-binding domains (RBDs). All identified domain A to date resemble a galectin-like structure (104, 105, 116, 117, 108–115),

consisting of a beta-sandwich core structure with additional components (a mixture of

secondary structures and flexible loops) around it. It has been proposed that ancestral

CoVs likely have acquired a host galectin via gene capture, which served as the

structural basis for domain A (118). Alpha-CoVs domain 0 is structurally analogous to

domain A (109, 117), hence could have originated from a domain duplication event.

Domain B, on the other hand, contains a structurally conserved core of antiparallel beta-

sheets, decorated with an extended loop on the viral envelope-distal side (103, 106–

108, 110, 115, 116). This loop may differ greatly both in size and in structure across the spectrum of CoVs, and hence is also referred to as the hypervariable region (HVR). In contrast to domains A and B, domains C and D consist of much smaller beta structures and are positioned away from the target cell during entry. When in the native trimeric form, however, they are in close contact with the S2 subunit (Figure 4A), including the fact that domain D ends with the S1/S2 cleavage site. It is therefore proposed that domains C and D could play crucial roles in relating the structural transitions, which are initiated by the host receptors interacting with the target cell-facing domains A and B, to

S2 to enable membrane fusion.

S1 atomic structures suggest that domains A and B (and 0 for alpha-CoVs) may function as independent RBDs. While having such independently-positioned RBDs is a unique CoV feature, if and how this feature is related to the apparently high propensity for CoV zoonosis has been unexplored. One reason could be the technical difficulties.

All present-day CoVs have narrow host ranges, but this is likely due to years of viral adaptations to a specific host, which would result in a loss of zoonotic characteristics.

One intuitive advantage of having multiple RBDs is the associated expanded surveying capability of potential receptors upon transmitting to a new host. Interestingly, we do observe characteristics on present-day CoVs that support this putative function. First, different CoVs have adapted to use different RBDs as their primary receptor-engaging domains. CoVs such as murine hepatitis virus (MHV), bovine (B-) CoV, OC43-CoV, and infectious bronchitis virus (IBV) have evolved to use domain A for receptor-binding (104,

105, 113, 119). Conversely, CoVs including NL63-CoV, SARS-CoV, MERS-CoV, and

porcine deltacoronavirus (PDCoV) have opted to engage their primary receptors with

domain B (103, 106, 107, 120). It is clear that, depending on the CoV, different RBDs

can be relied on to develop optimal interactions with receptors. Therefore, even though

contemporary CoV tropisms are more restricted to their existing RBD-receptor interactions, their ancestors may have benefited from having multiple RBDs to engage the optimal receptor upon first entering the new host.

The utilization of CoV spike receptor-binding domain A. Another reasonable expectation for the likelihood of expanded receptor-surveying with multiple RBDs is that each RBD may have different predispositions toward different host ligand types. This has been, as expected, observed to be the case. Presumably due to its close resemblance to human galectins (121), domain A, when adapted as the primary CoV

RBD, tends to bind host carbohydrates, specifically sialic acids (105, 113, 119). B-CoV and OC43-CoV use their domain As to engage host 9-O-acetylated sialic acids as receptors (105, 113), whereas IBV domain A binds host receptor alpha-2,3-linked sialic acids (122, 123). One interesting point is that the sialic acid binding site on domain A is different from the beta-galactoside binding site on galectins, as a recent OC43 S-sialic acid cryo-EM co-structure has shown (113). Presumably, after genetically capturing the galectin structure from its host, ancestral CoVs have modified its ligand specificity for their own purposes.

When CoV S protein domain A is used to engage the primary receptor, the receptor is usually sialic acid, except in the case of MHV. Instead, MHV S domain A engages a protein, murine carcinoembryonic antigen-related cell adhesion molecule 1a

(mCEACAM1a), as the principle receptor for the virus. Remarkably, the mCEACAM

binding site on MHV domain A structurally overlaps with the sialic acid binding site on

OC43 domain A (104, 113). This could suggest a possible evolutionary trajectory where

the ancestral MHV had adapted its domain A for a protein receptor and through this

adaptation, had gradually lost its sialic acid binding capabilities (105). Indeed, the S proteins of several MHV strains lack detectable sialic acid interactions (124), and are commonly used as negative-controls in research (105, 125, 126). On the other hand, if we accept that the protein-binding MHV domain A had evolved from a sialic acid-binding ancestor, then there should be an evolutionary intermediate, where the domain A binds to sialic acids as well as a protein (Figure 4C). A CoV with a domain A of this

“intermediate phenotype” may have an expanded tropism, or rather, the tropism of this

CoV cannot be fully accounted for by the availability of its protein receptor. Surprisingly, this description coincides with the phenotype of MHV strain JHM. Like other MHV strains, JHM engages its principle receptor, mCEACAM1a, via domain A. Unlike other strains, where the virus kills the host due to liver damage (hence the naming of the virus), JHM eliminates the host via brain damage. It has been further elucidated that their S proteins are the main determinants for these phenotypic difference (127–129).

Interestingly, compared to the liver, mCEACAM expression in the mouse brain is very low (130). Conversely, neuronal cells are known for their abundance in sialic acids

(131). Additionally, the presence of JHM S protein alone is sufficient for the virus to infect mCEACAM1a knock-out mice (129).

Figure 4. Coronavirus S1 Subunit. (A) MHV (PDB: 3JCL) S is depicted as in Figure 3, with additional color coding of the S1 sub-domains. (B) S1 extract from panel A, with sub-domains color coded (red, domain A; orange, domain B; green, domain C; blue, domain D). (C) Proposed CoV S1 domain A evolution. Ancestral CoV might have acquired a galectin structure from its host, and adapted to bind different sugar moieties. Ancestral MHV may have further diverged and evolved protein-binding potentials. An evolutionary intermediate was proposed to interact with both host sugar and protein factors. Galectin-3 crystal structure (orange) was obtained from PDB: 1A3K. BCoV S domain A crystal structure (blue) was obtained from PDB: 4H14. MHV-A59 S domain A cryo-EM structure (red) was obtained from PDB: 3JCL. The mock structure for the intermediate was a recolored version of MHV-A59. (D) MERS S (PDB: 5X5F) S1 is shown in surface representation, oriented similar to panel B. The S domain B is depicted in both down (green) and up (cyan) conformations.

Even though the exact mechanism JHM S utilizes to infect the brain has not been

resolved, it is possible that JHM S domain A could engage host mCEACAM1a as well

as sialic acids.

The utilization of CoV spike receptor-binding domain B. Different from S protein domain A’s general predisposition towards host sialic acids, domain B, when used to engage the primary receptor, the receptor is always a host protein. 229E-CoV, transmissible gastroenteritis virus (TGEV), and feline infectious peritonitis virus (FIPV) use their domain Bs to engage host aminopeptidases N (APN) as receptors (132–135).

NL63-CoV, SARS-CoV, and likely Wuhan-CoV domain Bs bind to host receptor angiotensin-converting enzyme 2 (ACE2) (103, 107, 136). The domain B of MERS-CoV

engages host receptor dipeptidyl peptidase IV (DPP4) (106, 137). Interestingly, domain

B appears to have a bias towards recognizing these peptide-processing exopeptidases

as receptors, without relying on their enzymatic activities. This bias evidently was

prominent enough to result in convergent evolution, where both NL63-CoV and SARS-

CoV have adapted to use ACE2 as the receptor, but they achieved this by evolving

unique contact motifs on their domain Bs (103, 107). While we do not know all the

advantages of CoVs using these exopeptidases as receptors, they may be related to

their proximities to other, required, co-factors for CoV entry. Indeed, one study has

shown that hDPP4 is likely in the same host plasma membrane microdomain as some

type II transmembrane serine proteases (TTSPs), whose cleavage activities drive facile

MERS-CoV entry kinetics (138).

As a prerequisite to infection, viral glycoproteins engages their receptors. This

receptor-ligand interaction is believed to impose certain structural properties on the glycoprotein, which primes the glycoprotein for subsequent steps toward membrane fusion. The documentation of these effects, however, can be challenging. One of the few cases where these structural changes were well characterized is in the case of HIV-

1 gp160 (139), and more recently for EBOV GP (140). It was revealed via single-

molecule Förster resonance energy transfer (FRET) that the gp160 native trimer is

structurally dynamic, transitioning through multiple prefusion conformations. Importantly,

receptor engagement stabilized gp160 to particular conformations that favor

downstream events (139), which has been verified through recent structural studies

(83). Interestingly, we see a similar theme in CoVs. CoV S protein domain B, as well as

domain A, engages host receptors and potentiates the S protein for proteolytic

cleavage, which facilitates viral fusion. How this signal is relayed structurally has been

unclear, but recent cryo-EM advances may offer some insight. It has been noted that

the domain B-receptor affinity is higher when soluble domains are used than the native

S trimers. This is because, in the native state, the domain B receptor-binding motif is

shielded from interactions with the protein receptor (109, 112). Therefore, by extension,

domain B would need to go through structural transitions to reveal this binding motif

before receptor engagement. Indeed, the domain B of MERS S and SARS S have been

observed to spontaneously transition between the receptor-unresponsive “down” state and the receptor-responsive “up” state (Figure 4D). Interestingly, this domain B structural transition appears to be linked to the overall stability of the spike trimer. While

cryo-EM readily obtained native S trimer structures with zero, one, or two domain Bs in

the “up” state (111, 112), the structure with all three domain Bs in the “up” state could

only be obtained using an unfusogenic S that has been stabilized in the S2 region (141).

These results support a model where the native S trimer is maintained stable as long as

some of the domain Bs are in the “down” state. When a domain B engages with the

protein receptor, their high affinity interaction holds the domain in the “up” state. When

all three domain Bs are “up”, there is a significant reduction in the contact residues

between S monomers, which may result in S1 timer destabilization. The S proteins in

this metastable state may be prone to proteolytic cleavages, which unleashes the series

of structural transitions that facilitate membrane fusion, and finally land in the hyper-

stable post-fusion conformation. Indeed, the highest proportion of S post-fusion

conformations were obtained when the native S trimers where coincubated with an

antibody that maintains the SARS S domain B “up” state, and exogenous proteases

(141). While the specific structural transitions domain A may elicit upon receptor

engagement is currently unclear, it is possible that both domains converge on the

domain B “up” state to prime S for viral fusion.

The dual-utilization of CoV spike receptor-binding domains. Even though

having multiple RBDs may offer major advantages in receptor surveying during the

initial stages of CoV cross-species transmissions, other benefits may exist as well. Most

contemporary CoVs, which have fully adapted to their biological niches, still retain all of

their RBDs. 229E-CoV appears to have lost its domain 0, the putative sialic acid-binding domain for alpha-CoVs, but this is a very rare phenomenon for CoVs (133). Because

CoVs generally evolve one RBD to engage with the primary receptor, it is possible that

the other RBDs can be adapted to acquire secondary benefits. Indeed, two such

examples exist: MHV and MERS-CoV. All MHVs use their S protein domain As to bind

to the protein receptor mCEACAM, while to this day no factor was ever identified to

engage their domain Bs. Instead, MHV domain B has been observed to modulate the

stability of the S trimer and modulate the energy thresholds for S-mediated membrane fusion activation (142–145). MHV-JHM has a very unstable S protein, in that its S1 can be found to dissociate from S2 from incubation at physiological conditions (pH=7, 37°C)

(146). Because S1-S2 dissociation is considered a prerequisite for S-facilitated

membrane fusion, this S protein instability has been viewed as a requirement for the

unique entry properties of JHM-CoV. Interestingly, JHM has the largest domain B of all

MHVs. While JHM maintains its domain B size when grown in the mouse brain, large

domain B deletions (up to ~150 residues) arises upon in vitro passaging of the virus in

cell cultures (142, 145, 147). The decrease in domain B size corrected with the

decrease in JHM neurovirulence, S1-S2 stability, and the mCEACAM-independent

activities (146). These results suggest that, when a domain B is not used to engage

receptors, it can nonetheless be adapted to provide certain structural segues for

membrane fusion. These structural segues could be the very same ones observed for

MERS-CoV and SARS-CoV, where domain B spontaneously transitions between “up”

and “down” conformations (Figure 4D). Intriguingly, even though JHM receptor-binding

occurs on the domain A, which is distally-positioned from domain B, its receptor-

engagement may still be connected to the domain B-mediated S instability. The support

for this came from early mutagenesis studies, where the linker region between domain

A and B was found to affect the same JHM properties as its domain B sizes (148). It

was, therefore, suggested that the signal of domain A-receptor-engagement is

transmitted, structurally, to other parts of the S protein (presumably to domain B).

Another example where CoVs utilize both RBDs is the MERS-CoV. MERS-CoV

engages its protein receptor, DPP4, with the S protein domain B, the virus has also

evolved to utilize its non-receptor-binding domain, domain A, to aid in the viral infection.

Perhaps the strongest evidences that MERS domain A is important for MERS-CoV infection are the demonstrations that antibodies against this domain can be neutralizing

(149–151), and protected model mice from lethality (151). The mechanistic function of

MERS domain A was revealed in a recent study. Specifically, MERS-CoV was found to bind host sialic acids, and this interaction was narrowed to its S protein domain A.

Interestingly, unlike the strong sialic acid-binding domain As of B-CoV and OC43, the weak interactions between MERS domain A and sialic acids were not detectable using soluble domain, and only after multimerically appended onto nanoparticles can they be quantified (126). The report also identified the specific ligand to be α-2,3-linked, unacetylated sialic acids (126). Furthermore, the MERS domain A sialic acid binding pocket was visualized in a cryo-EM study to be at the same location on domain A as

OC43 (113, 114). Despite being a relatively weak interaction, the engagement between

MERS S domain A and host sialic acids appears to be a mechanistically relevant one, for sialic acid depletion significantly reduced MERS S-facilitated pseudovirus and authentic virus entry (114, 126), even though the target cells expressed sufficient levels

27 of hDPP4. These results support an infection model where CoVs, such as MERS, can differentially utilize strong interactions (e.g. domain B and hDPP4) for infection, and weak interactions (e.g. domain A and sialic acids) for host attachment. In this model, the viral tropism is determined by the host interactions with both RBDs. While it is widely accepted that strong receptor interactions enable viral infection, it is not clear, however, to what extent weak S-host interactions can achieve. More targeted studies where the host interactions with each CoV RBD can be investigated in isolation would offer additional enlightenment with respect to the contribution of each interaction to the viral tropism on the whole.

CoV spike receptor-binding domains enable viral zoonosis. CoV infections are generally restricted to their specific host species. As zoonotic pathogens, however,

CoVs may occasionally cross the species barrier and expand their tropism within a new host. Often, the virus will replicate but fail to transmit to new members of the species, and eventually be eliminated. Sometimes, the virus will evolve within this host and become transmissible to other members of the same species. Overtime, the virus will become endemic to the new species. The host requirements for CoV zoonosis, thus far, have been centered around the utilization of orthologous receptors. SARS-CoV uses its

S domain B to engage its primary receptor, hACE2. One strong supporting result for the viral zoonosis trajectory being from bats to civets and then to humans is that the virus can bind to the ACE2 from any of these species to enable infection. Remarkably, structural studies allowed us a glimpse into this zoonotic event. Among the contact residues between human (h) SARS-CoV and hACE2, civet (c) ACE2 differed at one

residue, which reduced its binding to hSARS-CoV. Conversely, this variance was compensated by a point mutation in cSARS-CoV, restoring the S-ACE2 affinity.

Additionally, mouse (m) ACE2 contains a residue change that eliminates a critical interaction with hSARS-CoV. Through forced viral adaptation, mSARS-CoV mutated a residue which partially-compensated for the missing interaction to mACE2 (reviewed in

(152)). These results demonstrate that the interactions between orthologous protein receptors and S protein domain B can be sufficient to enable CoV cross-species transmission. On the other hand, sialic acid-S domain A interactions have also been implicated to be sufficient for zoonosis. B-CoV had been endemic only in cattle, until around the 1890s, when the virus crossed the zoonotic barrier and spilled over into humans (38). After adaptations in humans (153), the resulting CoV, OC43, eventual became endemic in humans (16). Current B-CoV and OC43-CoV share remarkable similarities of up to 96% nucleotide identity (38), suggesting that they likely also share common mechanisms of transmission. Indeed, both B-CoV and OC43 engage their common receptor, 9-O-acetylated sialic acids, with their S protein domain As to facilitate membrane fusion (105, 113, 125). Therefore, common sialoside receptors may enable

CoV zoonosis, via domain A interactions. The third scenario where zoonosis may be achieved appears to require both appropriate domain A and domain B receptor interactions. MERS-CoV is considered to have originated from bats (52), which passed to camels (49), which then passed to humans in 2012 (26, 31). MERS-CoV S protein domain B engages its principle protein receptor, hDPP4 (106), whose orthologue utilization appears to be sufficient to convey susceptibility. Indeed, in cell cultures,

MERS S engages DPP4 orthologues from human, camel, bat, goat, sheep, horse, rabbit, and pigs, and but not mouse, hamster, and ferret, for infection (154, 155).

Remarkably, among these putatively susceptible species, horses, goats, sheep, pigs, and rabbits are generally found to be sero-negative for the virus (156–158). These differences suggest that there are factors other than DPP4 that could function as zoonotic barriers. In addition to its domain B interactions with DPP4, MERS S domain A also engages target cell α-2,3-linked sialic acids (126). The presence of this glycotope in the gastrointestinal/respiratory tract indeed correlates well with the susceptible animal species of bats, camels, and humans, while it is absent in the non-susceptible horses, pigs, rabbits, and sheep ((159), and reviewed in (160)). Further evidence for host sialic acids potentially functioning as zoonotic barriers is provided by forward genetic studies.

During the adaptations of MERS-CoV to two independently-developed humanized mice models, the virus fixed an adaptive mutation at Asn222, located on the apical side of the sialic acid-binding S domain A (161, 162). Even though Asn222 is located outside of the identified sialic acid binding site on MERS S domain A (114), it does fall within a cluster of residues that have been shown to allosterically affect the sialic acid binding of related viruses (105, 125). Studies that aim directly at the sialic acid interactions of the mutant

MERS S domain A would shed light on the relationship between host sialic acids and

CoV cross-species transmissions, in addition to requiring appropriate S domain B-DPP4 interactions. Without a doubt, MERS S domain A is under active selection, as Asn222 is among a hand full of domain A variants whose nucleotide polymorphisms were detected

in a recent survey of circulating MERS-CoV strains in human populations of Saudi

Arabia (163).

Through the utilization of multiple RBDs, CoVs acquire specific zoonotic niches.

These utilizations may involve specific sialic acid receptors (e.g. B-CoV, OC43-CoV),

protein receptors (e.g. 229E-CoV, SARS-CoV), or both (e.g. MERS-CoV), as well as

resulting in fine-tuned S protein stabilities (e.g. MHV). Exploring answers to remaining

questions, such as whether bi-valent RBD exists, or a possible sequential functioning of

the RBDs during CoV entry, or whether some RBDs are truly vestigial, can greatly

enrich our understanding of CoV biology, and bringing us closer to constructing

appropriate strategies to combat the threat of CoVs to global health, as the ongoing

SARS-CoV-2 outbreak escalates.

CoV Spike Membrane Fusion Mechanism

CoV spike membrane fusion is triggered by proteases. Viral entry is a highly

regulated process, for it governs viral tropism. Therefore, it is not uncommon for viruses

to utilize additional host factors as triggers for membrane fusion, even after receptor

engagement. IAV HA, after engagement with its sialic acid receptors, requires a low pH

environment to initiate membrane fusion (71, 74). VSV glycoprotein (G) engages low-

density lipoprotein receptors (LDLRs) for entry, but membrane fusion is not initiated until

the low pH stimulation following endocytosis (80, 164). While CoV S protein S1 subunit

governs host attachment, the S2 subunit facilitates virus-target membrane fusion.

Structurally, S2 is shielded by S1, therefore, this restraint would need to be lifted for

membrane fusion. S1 receptor engagement destabilizes S1-S2 interactions (141, 143,

165, 166), but to dissociate S1 from S2, CoVs further require host proteases to cleave

at the S1/S2 junction (Figure 5). Depending on the specific amino acid sequences, the region downstream of S1 domain D may have different susceptibilities to the various host proteases. For MHV-JHM and MERS-CoV, their S1/S2 site is so sensitive to proteases, that the majority of their S proteins are cleaved during viral assembly in the infected cell and so S1 is only non-covalently associating with S2 on the mature virion

(146, 167). After these virions engage with receptors on the new host, there is only one additional proteolytic cleavage event required to trigger fusion. The S proteins of MHV-

A59 and SARS-CoV, on the other hand, are much more resistant to host proteases, so that the S1/S2 junction remains largely intact on the mature virion (167, 168). To infect, these virions would require the S1/S2 cleavage after receptor engagement on the new host, before advancing to the next stage.

After the S1-S2 dissociation, it is expected that all CoV S proteins require a final cleavage event to enable membrane fusion. This second cleavage site in located immediately upstream of the fusion peptide, termed the S2’. Unlike the S1/S2 cleavage site, which is putatively accessible to host factors including proteases, the S2’ site is buried within an α-helix and not exposed in the pre-fusion conformation (110, 141). This suggests the requirement of extensive S conformation changes before S2’ cleavage, presumably the structural destabilization from receptor engagement, and the S1 dissociation from S1/S2 cleavage.

Figure 5. Coronavirus S2 Subunit. S2 structural transitions during CoV entry. CoV entry is depicted as shown in Figure 2B. During host attachment, CoV S protein subunit S2 (gray) is shielded by S1 (yellow) in the pre-fusion conformation (top-left). Cryo-EM structure of MHV S (PDB: 3JCL) pre-fusion conformation is used as an example. Receptor engagement potentiates the dissociation of S1 from S2, revealing S2 for proteolytic processing (top-middle). The unshielded S2 Structure is modeled by extracting S2 from the trimeric S structure in panel A. Proteolytic cleavage triggers drastic S2 conformational changes, which facilitated membrane fusion and results in a stable helix-helix bundle that is the post-fusion conformation (top-right). MHV S post- fusion structure (PDB: 6B3O) is used as an example.

Even though the specific structural transitions after the S2’ cleavage have not been observed, it is generally believed that this cleavage would release the tension holding four short α-helices in place, and they instead would join and refold into a long trimeric coiled coil, with the fusion peptide positioned at the distal end, projected towards the target membrane (99, 100). Thereon, with success, the fusion peptide will be embedded

33 into the target membrane, physically bridging the two membranes (Figure 5). Next, a part of the long coiled coil, termed the heptad repeat region 1 (HR1), through extensive hydrophobic interactions with the downstream helical HR2 (169), pulling the two membranes into close apposition to catalyze fusion (Figure 5).

CoV spike proteolytic sensitivity dictates entry kinetics. Because of the absolute requirements of the S1/S2 and the S2’ cleavages, host proteases are obligated factors for CoV entry. And because of the intrinsic viral sequence differences at these cleavage sites, as well as the inherent variations in the availability of the various host protease types, it is only natural to conclude that these proteases form another class of entry factors that, potentially, dictates CoV tropism, pathology, and zoonosis. Indeed, the MHV-2 and SARS-CoV S1/S2 cleavage sites are insensitive to furin, the major and most well-characterized serine proprotein convertase (170). As the result, when these viruses bind to receptors on target cells, their S proteins still require two cleavage events before entry can be completed. Presumably because two cleavages per S is beyond the capability of cell surface proteases at physiological levels, these CoVs are obligatory endocytic viruses, where endocytosed virions encounter high levels of endosomal cysteine proteases in the cathepsin family (167, 171, 172). After cleavages at both S1/S2 and S2’ by endosomal cathepsins, these particles complete viral entry.

Conversely, the S1/S2-cleaved S proteins on MHV-JHM and MERS-CoV only need one additional cleavage (S2’) before membrane fusion. In these cases, cell surface proteases such as the transmembrane protease, serine 2 (TMPRSS2), a representative

TTSP, is sufficient to cleave at S2’ and allow these viruses compete entry at or near the

34 cell surface (138, 161, 173, 174). While entry in the endosome and at the cell surface both are explored by CoVs (Figure 6), cell surface entry does offer a few advantages.

First, surface entry is more kinetically favorable than endosomal entry. Studies have determined that MERS-CoV can entry target cells four times faster when using the cell surface route, with 50% viral entry at ~15 min vs. ~60 min for endosomal entry (138,

161, 174). With a more facile entry kinetics, these CoVs presumably establishes infection in a more competitive manner. Second, surface entry potentially avoids innate antiviral factors. These include the endosomal RNA sensors, which could potentially alert the host innate immune system, which compromises CoV infections (175). Another class of innate antiviral factors are the interferon-induced transmembrane proteins

(IFITMs). These small (~140 residues) single-pass transmembrane proteins somewhat non-specifically inhibit viral-induced membrane fusions, including IAV, VSV, WNV,

EBOV, HIV, SARS-, and MERS-CoV (176–181). Third, endosomal entry is time- sensitive. Because the endocytic pathway is used for digestion, it is an intuitive dead end for viral entry. In an event where a virion moves too far down the endocytic process, the late-endocytic / lysosomal digestive enzymes will excessively proteolyze and destroy the infectious agent. Hence, cell surface entry may be relatively protective to the virion. Conversely, the utilization of the endocytic entry route may be a byproduct from specific CoVs developing the S protein stability needed for efficient viral transmission.

With S1/S2 uncleaved, the S protein of SARS-CoV is more stable than the S1/S2- cleaved MERS-CoV.

Figure 6. Coronavirus Entry Routes Depend on Protease Utilization. To facilitate fusion between the CoV envelope and the target cell membrane that is the pinnacle of viral entry, CoV S proteins requires activation by proteases. Because of this, the specific subcellular location at which membrane fusion is completed during CoV entry is directly related to the level of suitable proteolytic activities at these locations. Hence, CoV entry, depends on the proteases present, can either be completed at or near the plasma membrane or within the endosome after endocytosis. If the target cell expresses high levels of cell surface serine proteases such as hTMPRSS2, sufficient cleavage events can happen at the S1/S2 and/or S2’ sites on the S protein. In this scenario, the utilization of hTMPRSS2 skews the CoV entry site towards plasma membrane. However, if the target cell expresses insufficient levels of hTMPRSS2, or the S protein in question does not contain an optimal peptide sequence for hTMPRSS2 cleavages, the entry process may not be facile enough to be completed at or near the plasma membrane. Because the virus would most likely remain bound to the cell due it the high- affinity interactions between S and the receptor, the virion would mostly likely be endocytosed and trafficked into the endosome. As the virus progresses through the endocytic pathway, it would encounter the resident cysteine protease cathepsins. For most CoV S proteins, endosomal cathepsins are also able to cleave and trigger S conformational changes needed for membrane fusion. Because cell surface entry is kinetically more facile than endosomal entry, these entry pathways are also referred to as “early” or “late”, respectively.

This correlated with the epidemiological determination of their transmission rate and

basic reproduction numbers (R0), where R0(SARS) > 1, a highly-transmissible virus, and

R0(MERS) < 1, a low-transmission virus (28, 32).

CoV tropism depends on protease utilization. While entry route preferences

may lead to entry kinetic differences, it is the variation in the requirement of different

host protease types that drives CoV tropism diversity. Additionally, there may be more

CoVs that utilize the cell surface entry route than currently characterized, because the

laboratory environment tend to skew CoVs towards the endocytic entry route.

Laboratory-adapted 229E-CoV reportedly uses the endocytic route, whereas Shirato et

al. demonstrated that the 229E clinical isolates preferred the cell surface route, and it is

the cleavage site mutations arose during laboratory passages of the virus that

accounted for this phenotypic difference (182). Importantly, the 229E clinical isolates

preferentially utilized cell surface proteases, such as TMPRSS2, for entry, and

replicated to higher titers in human airway epithelial (HAE) cell cultures, the best

mimicry to the human bronchial environment to date. Conversely, the laboratory-

adapted virus yielded higher titers in cell cultures with low levels of surface TMPRSS2,

and instead preferred to use endosomal cathepsins for entry (182). Therefore, the reliance on cell surface proteases correlated with the natural tissue tropism of 229E-

CoV. Moreover, one of the results supporting the bat origin of MERS-CoV is the finding that the virus can use bat (b) DPP4 and infect bat cells (52, 183). Interestingly, however, the most closely-related MERS-like CoVs in bats, HKU4-CoV, cannot infect human cells, even though HKU4 can bind to hDPP4 (52, 183). Remarkably, two residue

substitutions that made the HKU4 S1/S2 cleavage site more MERS-like enabled HKU4

S-facilitated viral entry into human cells (184). Taken together, these results suggest that the diversity in the utilization of different host proteases during viral entry may not only affect specific CoV tissue tropism, but also dictate the likelihood CoVs can achieve cross-species transmissions, or zoonosis.

Coronavirus Spread

Existing CoV infections can be spread through two different mechanisms. First of all, new viruses assemble and bud into the endoplasmic reticulum—Golgi intermediate compartment (ERGIC) of the infected cell (185, 186), facilitated by the retention signal sequences in the M protein (1, 6), as well as the interactions between M and N (1, 3).

Mature virions are eventually released out of the secretory pathway and into the extracellular space (Figure 7), where they are passively brought to close proximity to naïve cells, and, if successful, another round of infection is initiated via canonical CoV entry. Alternatively, established CoV infections can be transmitted directly into neighboring cells, without the need for manufacturing de novo virions. This is achieved through the process of cell-cell fusion, or syncytia formation (Figure 7). This process is, mechanistically, facilitated by the S proteins decorating the plasma membrane of the infected cell. Through the interactions with factors on the neighboring cell plasma membrane (e.g. receptor engagement, protease cleavage), these cell-bound S proteins go through structural transitions and facilitate membrane fusions analogous to canonical

CoV entry. Once the fusion pore is formed, the CoV genome can diffuse into the naïve cell cytoplasm and expand the infection (Figure 7).

Figure 7. Coronavirus Infection Spread. CoVs use two methods to spread existing infections, either through the canonical viral assembly and release, which initiates de novo infections using receptors and proteases, or the same host factors can be engaged by S proteins on the infected cell surface and spread the infected directly into neighboring cells, via the process of cell-cell fusion, or syncytia formation.

Perhaps because the fusion requirements do not include any obligatory endocytic

elements such as low pH stimulations, this ability to facilitate cell-cell fusion appears to

be a universal trait for CoVs, as it has been documented for a wide range of S proteins

(143, 187–195).

The mechanisms by which CoVs spread existing infections intercellularly, may

explain, in part, the remarkable expansion of these viruses into new hosts, including

humans. It is generally accepted that cell-cell fusion requires the same set of triggers as virus-cell fusion, which includes RBD-receptor interactions, followed by proteolytic cleavage(s). However, whether these factors are needed to the same extent remains an open question. Intuitively, unlike in virus entry, where virions are diffused in the extracellular space and hence first require membrane anchorage, cell-cell fusion do not have the requirement for this cell attachment. Instead, the membrane of the S- expressing cells are naturally in close proximity to the neighboring cell membranes.

Additionally, an individual CoV virion has a finite number of S proteins to facilitate entry, whereas an infected cell replenishes its surface S proteins continuously. Due to these characteristic differences, one may expect the way CoV cell-cell spread utilize host factors to be relatively different. Perhaps certain modest interactions between S and the target cell (i.e. with sialic acids) may be sufficient to advance to cell-cell membrane fusion.

Amongst the CoV cell-cell fusions, the ones caused by the S protein of MHV-

JHM stands out. Unlike the S protein from other CoVs, or even from other MHV stains, the S protein of JHM has been documented to facilitate cell-cell fusion independent of

40 its principle protein receptor, mCEACAM1a, resulting in the fusion between mCEACAM+ and mCEACAM- cells, even if these cells originated from different animal species (143,

144, 188, 196, 197). Additionally, JHM-CoV is the only MHV strain known for causing lethal brain infection, even in mice that lack CEACAM1a (129). JHM spike was identified to be the major contributor to this phenotype (129), and JHM-CoV, but not the related A59-CoV, spread inter-neuronally both in vivo and within in vitro cultures of

CNS-derived cells (198, 199). Moreover, these cell-to-cell syncytial spread correlates with pathogenesis in several infection models (129, 188, 200–203). However, the mechanism for this protein receptor-independent JHM spread is unclear. Because JHM

S domain A engages mCEACAM1a while its domain B has been not been shown to engage any host factors, but instead functions to lower the energy barrier for S conformational changes (143, 146), it was proposed that JHM S does not need to engage with receptors to fuse cells. An alternate hypothesis is that JHM S has weak interactions with a host factor that is more promiscuous across different animal species.

Sialic acids are good candidates, especially since the structure of MHV domain A highly resembles those of the sialic acid-binding relatives (104, 105, 113). Notably, neural cell membranes are known for their abundant sialic acid content (131). Therefore, a JHM S- sialic acid binding can potential accounts for an inter-neuronal syncytial spread that is rapidly lethal. A prediction is that variants of JHM-CoV exhibiting enhanced sialic acid affinity will have unusually high neurovirulence. On the other hand, the pathological impacts from cell-cell fusion have not been well studied for other CoVs, mainly due to the limitations in non-mouse animal studies and the lack of human biopsies. Because

cell-cell fusion potentially amplifies the effects of weak S interactions, it may be in this

context can the link between certain host factors and pathogenesis be characterized.

MERS-CoV strain causes lethal pneumonia and here it is significant that antibodies

specific for the MERS-CoV S domain A both neutralize the virus and reduce infection

and pathogenesis in a mouse MERS-CoV model system (149, 151, 204). Because

MERS S domain A has weak interactions with host sialic acids, conceivably, these

antibodies interfere with sialic acid binding, reducing expansion of MERS-CoV that may take place via cell-cell fusion. Variants of MERS-CoV with enhanced cell-binding may be useful in assessing the in vivo significance of these interactions.

Purpose of Dissertation

A review of the literature shows that CoVs utilize multiple host factors to spread their infections. These include viral attachment factors, receptors, and host proteases.

While the involvement of these factors have been characterized, our knowledge is lacking when it comes to the extent any particular factor contributes to the fusion process as a whole. Additionally, past studies have focused on CoV virion entry, leaving the cell-cell fusion mode of infection transmission largely unexplored. Our goal was to use reductionist approaches to investigate the involvement of each host factor in isolation, at each step of CoV entry and spread.

We hypothesis that distinct RBDs operate at different CoV infection stages. Initial infection by virus particles likely requires weak associations with the target cell, before durable interactions with cell receptors are needed for entry. At later infection stages,

CoVs produce S proteins in abundance, far exceeding the amounts that are

42 incorporated into secreted progeny virus particles. These unincorporated S proteins accumulate on infected-cell plasma membranes, where, depending on the CoV strain, they mediate cell-cell fusions, i.e., syncytial developments, which rapidly expand infection. Close cell-cell contacts may obviate the need for high-affinity adherence of S proteins to adjacent, uninfected cell surfaces. Conceivably, S domain A RBDs with relatively low affinity for sialic acids might be sufficient to promote syncytial expansion of infection, without requiring high-affinity domain B protein receptor interactions. This putative role for sialic acids has yet to be tested. This study provides results that implicate sialic acids in facilitating CoV infection and CoV cell-cell spread, bringing insights concerning CoV dissemination in nature.

CHAPTER II

MATERIALS AND METHODS

Cells

HEK293T (ATCC), HeLa (ATCC) and HeLa-mCEACAM (205, 206) cells were

maintained in DMEM-10% FBS media [Dulbecco’s Modified Eagle Media (DMEM)

containing 10 mM HEPES, 100 nM sodium pyruvate, 0.1 mM non-essential amino

acids, 100 U/ml penicillin G, and 100 µg/ml streptomycin, and supplemented with 10%

fetal bovine serum (FBS, Atlanta Biologicals)]. BHK-21 cells (ATCC) were maintained in

DMEM-5% FBS media. LET-1 cells (BEI Resources; (207)) were maintained in DMEM-

10% FBS media lacking HEPES, sodium pyruvate and non-essential amino acids.

Calu3 cells (ATCC) were maintained in MEM-20% FBS media [Minimum Essential

Media (MEM) supplemented with 20% FBS, 100 U/ml penicillin G, and 100 µg/ml streptomycin]. DBT cells (208, 209) were maintained in MEM-5% FBS media [MEM supplemented with 5% FBS, 10% tryptose-phosphate broth, 26.8 mM sodium bicarbonate, 2 mM L-glutamine, 100 U/ml penicillin G, and 100 µg/ml streptomycin]. All cell lines were cultured in a 5% CO2 incubator at 37°C.

Viruses

Recombinant MHV strains JHM.SD (210) and A59 (211), both containing a firefly

luciferase (Fluc) reporter between the viral E (envelope) and M (matrix) genes, were

44 grown in DBT cells. Media were collected at 24- to 48-h post-infection. JHMHE- arose during previous laboratory passaging of JHM.SD.

Pseudoviruses

HIV-based pseudovirus particles (pps) were constructed as described in (167).

Briefly, HEK293T cells were transfected with pNL4.3-luc R- E- obtained from the NIH

AIDS Research and Reference Protram, cat. #3418, and viral or host protein expression plasmids, including pHEF-VSV G-Indiana (BEI Resources), pcDNA3.1-EMC-S-C9

(MERS S) as described in (154), and pCMV6-Entry-hDPP4FLAG (accession #

NM_001935; purchased from OriGene), and pCAGGS-TMPRSS2FLAG (212). To produce fluorescently-labeled pps, expression plasmids for either S15-GFP or S15- mCherry was included as described in (213). 6 h later, cells were replenished with

DMEM-10% FBS. Media were collected after a 48-h incubation period.

Vsvpps were constructed as described in (214). Briefly, HEK293T cells were transfected with viral glycoprotein expression plasmids, including pHEF-VSV G-Indiana

(BEI Resources), pcDNA3.1-HA5-QH and pcDNA3.1-PR8 NA1 (obtained from Lijun

Rong, University of Illinois—Chicago). 24 h later, cells were inoculated for 2 h with

VSVdeltaG/Junin GP-luciferase (214, 215). Cells were rinsed twice with FBS-free

DMEM media and replenished with DMEM-10% FBS. Media were collected after a 48- h incubation period.

Virus-like Particles (VLPs)

CoV VLPs were constructed by co-transfection with equimolar amounts of CoV

S, E (envelope), M (matrix) and N (nucleocapsid) – encoding plasmids. Coding

sequences for A59-CoV S, E, M and N genes are presented in accession #

AY910861.1; for JHM-CoV, accession # AC_000192.1; and for MERS-CoV [EMC 2012

strain (216), accession # JX869059.2, where only S gene is codon-optimized (154)].

The A59 / JHM-CoV and the MERS-CoV genes were inserted into pCAGGS and

pcDNA3.1 expression vector plasmids, respectively. Recombinant pCAGGS-DSP1-7-

N and pCAGGS-DSP8-11-N were constructed by fusing the DSP1-7 or DSP8-11 coding sequences [pDSP1-7 and pDSP8-11 (217, 218) provided by Zene Matsuda (University

of Tokyo)], followed by a 2x SGGS linker, to the 5’ end of the coding sequences for N

genes.

Expression plasmids (1 ug) were complexed with polyethylenimine (PEI;

Polysciences) (6 ug) in 0.2-ml of Opti-MEM (Life Technologies) for 15 min at room

temperature, then added dropwise to 5 x 105 HEK293T cells in 1-ml DMEM-10% FBS.

For spikeless VLPs, S expression plasmids were replaced with empty vector plasmid

DNAs. 6 h later, cells were replenished with fresh DMEM-10% FBS. Media were

collected at 48 h post-transfection.

Particle Concentration

Media containing viruses, pseudovirus particles, and VLPs were clarified by

differential centrifugation (300xg, 4°C, 10 min; 3000xg, 4°C, 10 min). Particles were

concentrated from clarified media by overnight centrifugation (SW28, 6500 rpm, 4°C, 18

hrs) through cushion comprised of 20% w/w sucrose in FBS-free DMEM. The resulting pellets were resuspended in FBS-free DMEM, and the resulting 100x concentrated particle stocks were stored at -80°C.

For particle concentration using ultrafiltration, pps were collected in SFM. After differential centrifugation (300xg, 4°C, 10 min; 3000xg, 4°C, 10 min), clarified supernatant was loaded into 10 kDa filter units (Millipore, UFC901024) and concentrated 100-fold, and subsequently stored at -80°C.

In Vitro Fusion Assay

S15-GFP – labeled MERS-S-HIVpps were mixed with the S15-mCherry – labeled hDPP4-HIVpps at 1:1 ratio. Particle binding was allowed for 30 min at 37°C.

Subsequently, 5 µg/ml trypsin (EMC Millipore) and/or a final concentration of 1 µM

MERS HR2 peptides (174) was added to the particle mixture, and incubated for 30 min at 37°C. Upon trypsin inactivation via the addition of soybean trypsin inhibitor (STI;

Sigma), the mixture was subjected to spinoculation (1200 x g, 2 h, 4°C) to pellet virions onto glass coverslips (219). After fixation, GFP and mCherry fluorescent signals were assessed using confocal microscopy.

To quantify FRET induction, GFP+ mCherry+ puncta was analyzed. mCherry was photobleached. GFP intensities for each pucta was measured both before and after mCherry-photobleaching, and difference was calculated. FRET induction was signified by an increase in GFP emission after mCherry-photobleaching.

Hemagglutination Assay

Serial 2-fold dilutions of 100x concentrated virus and pseudovirus particle stocks were prepared using FBS-free DMEM as diluent. Diluted samples were placed into V- bottom 96-well plates and incubated for 2 h at 4°C. 0.5% washed adult human

erythrocytes (in PBS) were then added and scoring was performed after 2-12 hours

incubation at 4°C.

Neuraminidase Treatments

Adherent, confluent cells in 24-well plates were rinsed once with PBS before

fixation with 3.7% paraformaldehyde in 0.159 M PIPES for 30 min at room temperature.

Following three PBS rinses, cells were treated with vehicle or neuraminidase (1 U/ml,

Sigma N2876) in neuraminidase buffer [200 mM NaOAc, 2mM CaCl2, pH 5.0] for 3 hrs

at 37°C. The cells were then rinsed three times with PBS, and then blocked with PBS +

2% FCS for 30 min at room temperature, and then used in virus binding assays.

For live cell assays, adherent, confluent cells in 24- or 96-well plates were

pretreated with vehicle or neuraminidase (1 U/ml, Sigma N2876) in FBS-free DMEM for

2 hrs at 37°C. The cells were rinsed three times with PBS, and then used in virus transduction assays.

VLP Binding Assay

VLPs used for binding assays were prepared by cotranfection with CoV S, E

(envelope), M (matrix) and DSP1-N + DSP8-11-N encoding plasmids. The resulting

VLPs contained complemented DSP1-7/8-11 Rluc-positive interiors. To quantify VLP-

associated Rluc, VLPs were serially diluted into Passive Lysis Buffer (PLB; Promega #

E194A) and introduced into opaque microplate wells (50 ul per well). Renilla luciferase

substrate (1.1 M NaCl, 2.2 mM Na2EDTA, 0.22 M KH2PO4, 1.3 mM NaN3, .44 mg/mL

Bovine Serum Albumin, 2.5 µM Coelenterazine [pH 5]) was added (100 µl per well) and

luminescence read using a Veritas microplate luminometer (Turner BioSystems,

Sunnyvale, CA).

VLPs were incubated for 2 h at 4oC to adherent, confluent cells, either with or without prior neuraminidase treatments. Incubations were in FBS-free DMEM, and at known Rluc VLP / cell ratios. After incubation, unbound VLPs were removed, and cells were rinsed variably with PBS. To measure cell-bound VLPs, adherent cells were dissolved into PLB and Rluc activity quantified.

To quantify S protein-mediated cell binding, equivalent Rluc levels of spike-less

(No S) VLPS were inoculated onto target cells. The resulting cell-associated Rluc signal in the spike-less (No S) VLP conditions were subtracted from the S-positive VLP conditions. For data depiction, the cell-associated Rluc signal in the spike-less (No S)

VLP conditions were set to zero for each biological repeat of the experiment.

VLP Entry Assay

VLPs used for entry assays were prepared by cotranfection with CoV S, E

(envelope), M (matrix) and DSP8-11-N encoding plasmids. The resulting VLPs contained

DSP8-11 fragments in their interiors. To quantify VLP-associated DSP8-11 levels, VLPs

were mixed with excess DSP1-7, in Passive Lysis Buffer, for 30 min at 37oC, to allow for

DSP1-7-DSP8-11 complementation. The excess DSP1-7 was obtained from HEK293 cells

overexpressing DSP1-7 fragments. Following post-lysis complementation, samples were

introduced into opaque microplate wells and luminescence readings used to infer DSP8-

11 levels.

VLPs in FBS-free DMEM were inoculated onto target cells that were transfected

2 days earlier with pDSP1-7. Indicated experiments included cotransfection of target

cells with pDSP1-7 and pCAGGS-TMPRSS2FLAG (212). For VLP inoculations, input

multiplicities were normalized to VLP DSP8-11 levels. After 6 h at 37°C, cells were

rinsed three times with PBS, dissolved into Passive Lysis Buffer and Rluc activities

quantified.

Virus and Pseudovirus Particle Entry Assays

Virus particles were incubated with cells, either with or without prior neuraminidase treatments, for 2 h at 4°C. Cells were then rinsed three times with PBS and replenished with FBS-containing DMEM / MEM. After 16 h at 37°C, cells were dissolved in lysis buffer [25 mM Tris-phosphate (pH 7.8), 2 mM DTT, 2mM 1,2- diaminocyclohexane-N,N,N’-tetraacetic acid, 10% (vol/vol) glycerol, 1% Triton X-100] and Fluc levels quantified with Fluc substrate [1 mM D-luciferin, 3 mM ATP, 15 mM

MgSO4·H2O, 30 mM Hepes (pH 7.8)] and a Veritas microplate luminometer.

For HIVpps entry, particles were inoculated onto target cells, and viral entry was allowed for 18 h. Unbound particles were rinsed away before intracellular Fluc levels were quantified.

Cell-cell Fusion Assay

Effector and target cells were prepared by introducing expression plasmids, using the PEI transfection method. Effector cells were co-transfected with pDSP1-7 and

the indicated S-expressing plasmids. S-expressing plasmids included pcDNA3.1-229E

SC9 (accession # AB691763.1, obtained from Dr. Fang Li, University of Minnesota).

Control effector cells received pDSP1-7 and empty vector plasmids. Target cells were

co-transfected with pDSP8-11 and the indicated S receptor-expressing plasmids. These included pCMV6-Entry-hDPP4FLAG and pcDNA3.1-hAPN (accession # M22324; obtained from Dr. Fang Li). Indicated experiments included additional cotransfection of target cells with pCAGGS-TMPRSS2FLAG.

At 6 hrs post-transfection, effector and target cells were rinsed with PBS, lifted with 0.05% trypsin, and mixed at 1:1 ratios. The co-cultures were incubated for 2 to 18 h at 37°C. Fused cells were visualized microscopically as GFP+ cells, and extents of

cell-cell fusion were quantified as Rluc activities present in PLB cell lysates. Samples

were chilled and maintained at 4°C during PLB with PLB and Rluc quantifications to

eliminate DSP post-lysis complementation.

To detect JHM-CoV cell-to-cell spread, DBT cell were inoculated with JHM-CoV

at moi = 10 for 5 h. Cells were rinsed extensively to remove excess viruses. Cells were

lifted via trypsinization, and were added to target HeLa / HeLa-mCEACAM cells, in the

presence of titrated levels of NA, and the live-cell Rluc substrate, EnduRen (promega,

E6482). The target cells were co-cultures of DSP1-7 and DSP8-11- expressing cells.

JHM-CoV infected cells fuse with multiple target cells, which enables DSP1-7:DSP8-11

complementation into active Rluc, which was quantified.

Fc Constructs

pCEP4-mCEACAM-Fc was constructed previously (146). Splice overlap

extension PCR was used to insert unique MreI and thrombin cleavage site and a

GSGGGG linker sequences between the mCEACAM (codons 1-142) and the human

IgG1 splice donor site. Using this modified construct, the mCEACAM coding region was

removed and replaced with MERS-S1 (codons 1-751), MERS-S1A (codons 1-357),

MERS S1A (N222D), and MERS S1B (codons 1-24 from hCD5 signal followed by

MERS S codons 358-588).

HEK293T cells were transfected using the PEI method and then incubated in

FBS-free DMEM containing 2% (wt/vol) Cell Boost 5 (Hyclone). Supernatants were collected on days 4 and 7, and clarified by sequential centrifugation (300xg, 4°C, 10 min; 4500xg, 4°C, 10 min). The Fc-tagged proteins were then purified using HiTrap

Protein A High Performance Columns (GE Healthcare) according to the manufacturer’s instruction. The resulting purified proteins were quantified spectrophotometrically, verified using western blotting with a known quantity of human IgG as the control, and stored at -20°C until use.

Fluorescence Staining

HeLa or HeLa-mCEACAM cells were pretreated with NA at 0, 0.5, 1, or 2 U/ml at

37°C. Following rinses, cell surface sialate levels were assessed by co-incubating cells with AlexaFluor 647-conjugated wheat germ agglutinin (WGA-Alexa647, Molecular

Probes) for 1 h at 4°C. Following fixation and Hoechst staining, immunofluorescence was assessed using confocal microscopy.

Western Blotting

Proteins in SDS solubilizer [0.0625 M Tris·HCl (pH 6.8), 10% glycerol, 0.01%

bromophenol blue, 2% (wt/vol) SDS, +/- 2% 2-mercaptoethanol] were heated at 95°C

for 5 min, electrophoresed through 8% (wt/vol) polyacrylamide-SDS gels, transferred to

nitrocellulose membranes (Bio-Rad), and incubated with mouse monoclonal anti-MHV-

HR2 10G (obtained from Fumihiro Taguchi, Nippon Veterinary and Life Sciences,

Tokyo, Japan), mouse anti-C9 (EMD MIlliopore), mouse monoclonal anti-MHV HE, clone 5A11, goat anti-human IgG (sc-2453, Santa Cruz Biotechnologies), or rabbit anti-

FLAG (Sigma F7425). After incubation with appropriate HRP-tagged secondary antibodies and chemiluminescent substrate (Thermo Fisher), the blots were imaged and processed with a FlourChem E (Protein Simple).

Microscopy and Image Acquisition

Live and formalin-fixed cell images were captured in a z series on an electron multiplied charge coupled device digital camera (EMCCDCascade 2; photometrics) and deconvolved using SoftWoRx. Identical conditions were applied to all acquisitions.

Deconvolved images were analyzed by an identical algorithm in Imaris 8.3.1 (Bitplane).

Statistical Analysis

Unless stated otherwise, all experiments were repeated independently at least three times. Each data point graphed represents the mean of an independent repeat

(N), generated from three or four technical repetitions (n = 3 or 4). Statistical comparisons were made by unpaired Student’s t test. Error bars indicate the standard error of the data. P-values lesser than 0.05 were considered statistically significant.

CHAPTER III

RESULTS

SECTION 1: CoV S, Receptor, and Protease Form the Basic Elements for

Membrane Fusion

CoV Entry is Spike-directed, Receptor-facilitated, and Protease-triggered

As previously mentioned, the major determinant for CoV permissibility is at the initial step of viral entry (99, 100). CoVs evolved to utilize a single viral surface glycoprotein, S, for viral entry and to establish de novo infections. Biologically, CoV S has two main functions: First, S is responsible for locating CoV particles to permissive target cells. Second, once tethered, S is also responsible for the delivery of CoV genome into the target cell cytoplasm, through the fusion between the viral envelope and the target membrane. The mechanistic details entailed within and between these two steps have been best characterized in the context of a virion and a cell (Figure 8A).

To infect a cell, CoVs require S proteins to engage with receptors (sialoside or protein).

This interaction putatively induces S to go through a structural transition that promotes the subsequent proteolytic cleavages. Cellular proteases, soluble or membrane- associated, cleaves to separate and enable the membrane fusion machinery. These cleavages trigger further S conformational changes that involve connecting and fusing the viral and cell membrane, which completes CoV entry.

Figure 8. The Role of S, Protein Receptor, and Protease During CoV Entry. (A) To infect a cell, CoV requires S : receptor interactions, Protease cleavages of S, and S- mediated membrane fusion mechanisms. (B) HEK293T cells were inoculated with HIV- pps bearing either MERS S or VSVG. Target cells overexpressed MERS protein hDPP4 receptor either with or without hTMPRSS2. After inoculation the excess virions were rinsed away. The inoculated cell culture was incubated overnight for Fluc reporter expression. The activity of Fluc, whose coding sequence is within the HIV-core plasmid, was quantified to estimate viral entry. The results are representative of three biological repeats.

To validate the current knowledge regarding CoV virus-cell entry, we co-

transfected HEK293T cells with expression plasmids for HIV-core-Fluc, MERS S or

VSVG and harvested cell supernatant 2 d post-transfection. Clarified supernatant, which

contained HIV pseudoparticles (HIVpp) bearing MERS S or VSVG, was laid onto

HEK293T cells overexpressing the protein receptor hDPP4 and/or the transmembrane

protease hTMPRSS2. The experiment was terminated at 48 hpi and viral entry was

inferred by quantifying Fluc activities, whose open reading frame (ORF) is within the

HIV-core sequences. In agreement with the field, MERS S-directed viral entry signal was elevated ~20-fold with sufficient level of hDPP4 (Figure 8B). Viral entry was further increased ~50-fold when target cell surface protease, hTMPRSS2, was enriched. This

entry characteristic is MERS S-specific, because viral entry mediated by VSVG, which

uses LDLRs as receptors and endocytic protonation as the fusion trigger (80, 164),

were consistent regardless of hDPP4 and hTMPRSS2 expression levels (Figure 8B).

FRET-based In Vitro Characterization of Viral Fusion

The most common ways to characterize viral entry is to quantify certain effects

that take place downstream of viral entry, such as reporter expressions (e.g. luciferase,

GFP), specific target cell factor alterations (e.g. IFN), or cell death (e.g. plaques). The

HIVpp-based assay we used in figure 8B is one of them. While these traditional ways to

detect viral entry are efficient, they do have a few limitations. First, their readouts are

not the actual viral entry event, but derivatives a few steps downstream of it, which

potentially adds many other variables. Second, viral fusion targets in these systems are

intact cells, whose complexity make it difficult to exclude the effect of other factors from

the entry process. Third, cell maintenance have specific requirements, which limit the

extent to which viral fusion environments can be manipulated. In order to characterize

CoV viral entry through direct detection of membrane fusion, as well as to have flexible

manipulations of potential factors for this fusion, we developed a FRET-based in vitro

viral membrane fusion system to characterize CoV viral fusion. It was demonstrated that

fusing the first 15 amino acid sequences of c-Src protein (S15) to certain fluorescent proteins would, via the post-translational modification of myristoylation, anchor these proteins to cellular membranes (220). Furthermore, these membrane-associated fluorescent proteins (S15-GFP or S15-mCherry) are readily incorporated into HIV-1 virions (213), enabling the tracking of viral particles (Figure 9A). We hypothesized that, if CoV S, receptor, and protease are the minimum requirement for membrane fusion, then a mixture of MERS S-bearing HIVpps, hDPP4-bearing HIVpps, and trypsin, should cause the two particles fusing together. Furthermore, if the S-bearing particle contained

S15-GFP, and the receptor-bearing particle contained S15-mCherry, then fusion could be quantified by detecting the FRET activity induced by the two fluorescent proteins being within close proximities of each other (Figure 9B).

MERS-CoV S, hDPP4, S15-GFP, and S15-mCherry are Incorporated into HIV pps

The viability of the proposed in vitro viral fusion system requires the correct incorporations of MERS S, hDPP4, S15-GFP, and S15-mCherry into HIV pps. To test this, HIVpps were produced from HEK293T cells expressing either MERS S and S15-

GFP, or hDPP4 and S15-mCherry. HIVpp-containing cell supernatants were collected and clarified at 2 d post-transfection.

Figure 9. FRET-based In Vitro Characterization of Viral Fusion. (A) Schematic for HIV pps incorporating GFP that is N-terminally fused to the first 15 amino acid sequences of c-Src protein (S15), followed by a linker sequence. S15 is subjected to myristoylation, and together with internal positive charges, locates GFP to cell membrane, where it can be incorporated into the budding HIV pps. (B) Schematic for in vitro fusion system. Using the methodology from panel A, HIV pps could be produced to incorporation MERS S proteins, hDPP4, S15-GFP, and S15-mCherry. If S15-GFP and S15-mCherry are incorporated into MERS S and hDPP4 virions, respectively, then only after inter-particle membrane fusion can the two fluorescent proteins come within ~10 nm of each other, the distance required for FRET activity.

To assess the incorporation of MERS S and hDPP4 onto HIVpps, particles were

concentrated down via ultrafiltration, and 100x samples were used for immunoblotting.

MERS S detection was mediated by antibody against C9-tag, which was fused to the C- terminus of the S protein. hDPP4 was detected by antibody against flag-tag, which was added to the C-terminus of hDPP4 protein. Both proteins were readily secreted with

HIVpps (Figure 10A and B). To test the secretion of S15-GFP and S15-mCherry, we pelleted MERS S-HIVpps and hDPP4-HIVpps onto separate coverslips, utilizing an established spinoculation method (197, 213). After fixation, coverslips were evaluated using fluorescent microscopy. As expected, GFP+ particles were detected in the MERS

S condition, and mCherry+ particles were detected in the hDPP4 condition (Figure 10C).

We further reasoned that, if native, functional, MERS S and S15-GFP, as well as

hDPP4 and S15-mCherry, were co-incorporated onto their respective HIVpps, then

these particles should bind to each other upon mixing, which could be detected as GFP-

mCherry co-localization. To test this, these two HIVpp populations were mixed one-to-

one before spinoculation, and visualized via fluorescent microscopy. Remarkably, we

observed a near-total co-localization between GFP+ and mCherry+ particles (Figure

10D), suggesting that all components were co-secreted as expected.

Virion Fusion Minimally Requires CoV S, Protein Receptor, and Protease

We aimed to test whether CoV S, protein receptor, and protease are the minimal

requirements for viral membrane fusion that is required for CoV entry.

Figure 10. HIV pps Incorporate MERS S, hDPP4, S15-GFP, and S15-mCherry. (A) Western blot detection of MERS-S-C9 protein incorporation into HIV pps. HIV pps were collected at 48 h post-transfection, and concentrated/purified via ultrafiltration. The protein content of HIV pps bearing either MERS S or No S (Bald) were resolved via SDS-PAGE and blotted onto nitrocellulose membrane. The presence of MERS S was detected by mouse α-C9. (B) Wester blot detection of hDPP4-FLAG. Procedures were as described in panel A. hDPP4 presence was detected by rabbit α-FLAG. (C) HIV pps were generated by the co-transfection of plasmids for either S15-GFP and MERS S, or S15-mCherry and hDPP4. Particles were spinoculated onto glass coverslips and their fluorescence were assessed after fixation. (D) MERS-S-S15-GFP pps were mixed with hDPP4-S15-mCherry pps. Mixture was spinoculated and particle fluorescence and GFP-mCherry colocalization assessed.

Trypsin has been demonstrated to be an adequate triggering protease for CoV S-

mediated membrane fusions (167, 170). To determine the effective trypsin dose needed

for the in vitro fusion system, we incubated MERS-S-HIVpps with titrated concentrations of trypsin for 30 min. The S proteins on these virions were then subjected to immunoblotting, because trypsin-induced S cleavage patterns correlated well with infectivity (167). Because we saw a dose-dependent gradual decrease of S oligomer upon trypsin treatment, and a dramatic reduction of intact S proteins upon high trypsin presence (Figure 11A), we chose an intermediate trypsin dose of 5 µg/ml to trigger in vitro membrane fusions in our subsequent experiments. Incidentally, 5 µg/ml of trypsin is also within the effective dose window determined previously for MERS-S-HIVpps

(167).

After deciding on using trypsin as the triggering protease for MERS S-induced in vitro fusion, we next searched for a fusion inhibitory agent. Soluble HR2 peptides were developed to be CoV specific fusion inhibitors (195, 221–223), and their effectiveness against MERS S has been well characterized (174). Because HR2 peptides are derived from an S protein C-terminal α-helix region whose interaction with the HR1 region is required for the S protein fold-back action that pulls the viral and target membrane into close proximity, HR2 peptides, mechanistically, inhibit S-mediated membrane fusion at the very last stage. Because its mechanism of action is at a step subsequent to receptor engagement and proteolytic cleavage, HR2 peptides were our fusion inhibitor of choice.

Figure 11. Functional Assessment of Factors Used in In Vitro Fusion Assay. (A) Western blot detection of dose-dependent MERS S proteolysis. MERS S pps were incubated with a titration of trypsin for 30 min at 37°C. Following trypsin inhibition, samples were resolved and S cleavage products detected by mouse α-C9. (B) S15- GFP+ pps were mixed with S15-mCherry+ pps and their % co-localization was assessed via fluorescent microscopy. Specific conditions are as indicated. In “Bald” conditions, no membrane protein was overexpressed. The results are representative of three biological repeats.

Because variations in the co-localization between GFP+ and mCherry+ particles,

facilitated by the S : hDPP4 interactions, could greatly impact the efficiency of the

downstream viral fusion event, we next tested whether trypsin or HR2 peptides had any

unexpected effects on S : hDPP4 interactions. To this end, MERS-S-GFP pps were

mixed with hDPP4-mCherry pps either with or without trypsin or HR2 peptides, and the

percent co-localization of GFP+ and mCherry+ particles were microscopically quantified.

As a negative control, (Bald)-GFP particles were mixed with (Bald)-mCherry particles.

As expected, the S : hDPP4 interaction resulted in the co-localization of the majority of the particles (Figure 11B). We noted that the addition of neither trypsin nor HR2 peptides had affected this co-localization. This co-localization, however, was reduced by soluble S1 subunits (S1-Fc) of the S protein in a dose-dependent manner. The co- localization was further brought down to the level of the negative control (Bald x Bald) in the absence of hDPP4 (S x Bald). Both of these observations suggested that particle co-localization is a specific reflection of S : hDPP4 interactions. MERS S proteins have been found to bind to sialic acid moieties on the target cell membrane (126), but this interaction was evidently not sufficient to co-localize particles. We reasoned that our in vitro particle binding system might not be ideal for MERS S : sialic acid interactions, either because S proteins are not appropriately-assembled on HIVpps, as demonstrated that correct oligomerization of MERS S is required for detectable sialoside-interactions

(126), or the sialoside moieties on HIVpps are insufficient for the avidity-dependent

MERS S interactions. Additionally, the affinity between MERS S and sialic acids might simply be too low for detection in this context (114, 126), because HIVpps bearing IAV

HAs resulted in a detectable, albeit reduced particle co-localization (IAV HA x Bald)

(Figure 11B). Taken together, these results suggest that, in our system, particle co- localization is a direct result of MERS S : hDPP4 interaction, and neither trypsin nor

HR2 peptides affects this interaction.

To induce in vitro particle fusion, MERS-S-GFP pps were incubated with hDPP4- mCherry pps. Subsequently, trypsin was added either with or without HR2 peptides.

Trypsin activity was inhibited 30 min later by the addition of soybean trypsin inhibitor.

The mixture was then spinoculated onto a coverslip, fixed, and GFP+ mCherry+ particles, including particles either merely bound together or fused together, were analyzed for FRET induction (Figure 12).

Figure 12. In Vitro Viral Membrane Fusion Assay. MERS-S-S15-GFP pps were incubated with hDPP4-S15-mCherry pps for 30min at 37°C. Trypsin either with or without HR2 peptides were added into the mixture and further incubated for 30min before inhibitor was added to neutralize trypsin activity. Particles in the mixture was subsequently spinoculated onto a glass coverslip. After fixation, FRET activity was measured for GFP+ mCherry+ puncta. The representative micrograph was a result from mixing Bald-S15-GFP pps with Bald-S15-mCherry pps.

To detect FRET induction, we photobleached the FRET acceptor (mCherry), and quantified the fluorescence change in the FRET donor (GFP). If particles were bound but not fused, the distance between GFP and mCherry would be too great (>10 nm) for energy transfer. In this case, photobleaching mCherry would result in no GFP fluorescence change. If particles were fused, mCherry can come close enough (<10 nm) to GFP and absorb some of the GFP fluorescence. In this case, photobleaching mCherry would result in a net gain of detectable GFP signals (Figure 13A). To quantify relative FRET induction, we first set the GFP emission pre-bleach to 1. In the absence of FRET, the post-bleach GFP emissions should be evenly distributed around 1. In other words, the % to the right of 1 minus the % to the left of 1 equals zero. Therefore, if

FRET occurs, the resulting value should be greater than zero (Figure 13A). To define the experimental FRET detection window, we first tested the controls. As a positive

FRET control, we made Bald pps that incorporated both S15-GFP and S15-mCherry.

As expected, we detected high relative FRET in these particles. As a negative FRET control, we mixed MERS-S-GFP pps with hDPP4-mCherry pps without trypsin, and found a negligible FRET induction (Figure 13B).

Using FRET as a direct readout for S-induced membrane fusion, we found that the addition of trypsin alone was sufficient to induce FRET among the co-localizing

GFP+ mCherry+ particles (Figure 14A). It is known that HR2 peptides block CoV S- mediation fusions by physically interfering directly with the fusion machinery within S2

(169, 223).

Figure 13. FRET Induction Reflects S15-GFP – S15-mCherry Compartmentalization. (A) Schematic for FRET quantification. When GFP is in separate membrane compartments from mCherry, their distance is too great so that mCherry photobleaching would not impact GFP emission. Conversely, when GFP and mCherry share a membrane compartment, their distance can be within 10 nm, so that mCherry photobleaching would disable FRET, and resulted in an apparent gain of GFP emission. If we set the GFP emission before mCherry photobleaching to 1.0, then when there is no FRET, GFP emission should stay constant, normally-distributed around relative emission of 1.0. When FRET is present, the puncta GFP emission should be greater than 1.0. In other words, if % Right - % left = 0, then there is no FRET. If % Right - % Left > 0, then FRET is induced. Hypothetical data are depicted. (B) As a proof-of-concept, as well as assessing control conditions, FRET induction was calculated, as described in panel A, for conditions where S15-GFP and S15-mCherry were co-incorporated into pps, and S15-GFP pps were merely bound to S15-mCherry pps via S : hDPP4 interactions. For the former, % Right - % Left = 96.8, suggesting strong FRET induction. For the latter, % Right - % Left = 6, implying little to no FRET induction.

Figure 14. In Vitro Fusion Assesses the Sufficiency of Proteases in S-mediated Membrane Fusion. (A) The assessment of the effect of exogenous trypsin on MERS S- mediated in vitro membrane fusion. Experimental procedure was as described in Figure 12. Each data point represents the average of an independent experiment, where >300 GFP+mCherry+ puncta were analyzed. Error bars represent standard error (SE) from the mean. Statistically significant deviations from MERS S pps mixed with hDPP4 pps without additives (S x hDPP4) were assessed by unpaired Student’s t test; ns, not significant. * p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001. (B) Western blot detection of hTMPRSS2-FLAG. hDPP4- hTMPRSS2-S15-mCherry pps were produced by adding the expression plasmid for hTMPRSS2-FLAG during HIV pp production stage. Samples were processed as described in Figure 10B. (C) The assessment of the effect of hTMPRSS2 on MERS S-mediated in vitro membrane fusion. Experimental procedure was identical to panel A, only that trypsin was not added to the hTMPRSS2 condition. Data processing and statistical analysis were done as described in panel A.

Therefore, the potent inhibitory effect of HR2 peptides observed here validated that the

detected FRET was due to the induction of MERS S : hDPP4 – dependent membrane

fusions, and not alternative mechanisms, such as proteolysis-induced membrane

damages. Furthermore, because MERS-CoVs have been identified to preferentially use cell surface TTSPs for entry (138, 167, 174), we assessed the role of hTMPRSS2 on triggering S-mediated membrane fusion. To this end, HEK293T cells producing hDPP4-

HIVpps were co-transfected with hTMPRSS2 plasmid, which resulted in co- incorporation of hDPP4 and hTMPRSS2 onto HIV pps (Figure 14B). When hDPP4- hTMPRSS2-HIVpps were used in the in vitro fusion system, we observed a significant

FRET elevation comparing to hDPP4-HIVpps, comparable to the induction level of exogenous trypsin (Figure 14C). Taken together, we concluded that MERS S requires only its protein receptor and the triggering protease to induce membrane fusion. The triggering proteases include trypsin and a more biologically relevant hTMPRSS2.

Furthermore, we established this FRET-based in vitro fusion assay using MERS S as a proof-of-concept, but we believe this system can be readily adapted to other fusion contexts, given that the fusogen and receptor can be enriched on the pseudoparticles.

SECTION 2: MHV-JHM Spike Utilizes Host Sialosides During Viral Entry and

Spread

MHV-JHM CoV Agglutinates RBCs

It was proposed that CoV S domain As evolved from carbohydrate-binding

galectins (118). This proposal is supported by the findings that these domain As are

mostly predisposed to interact with host carbohydrates, specifically sialic acids (105,

106, 113, 125). The domain A of MHVs appears to be an evolutionary divergent, where they have evolved protein (mCECACAM) binding determinants on their galectin-like domain As (104, 121). We considered it likely that some MHV CoV strains also retain sialic acid binding competence, making these domain As evolutionary intermediates

(Figure 4A). This presumption came from studies of the JHM-CoV strain of MHV. JHM-

CoV can infect mCEACAM knockout mice (129), and can mediate cell-cell fusion of several mCEACAM-negative cell types (143, 146, 196, 197, 224, 225). Amongst the

CoVs, only the JHM-CoV strain has these documented activities that are apparently independent of a protein receptor, making it a sensible virus to use in addressing the hypothesis that sialate- and protein receptor-binding activities coexist on CoV S domain

As.

We were first interested in determining whether the S proteins on JHM-CoV interacts with host sialic acids. To narrow our interpretation on any potential results, we needed to eliminate the effect of other possible sialic acid-binding agents on the JHM virion. JHM is one of a few MHVs that can express and incorporate the non-major structural protein, hemagglutinin-esterase (HE), into virus particles (226, 227). CoV HEs enzymatically cleave the acetyl group on sialic acids, and hence their putative biological role is to destroy RBD-sialic acid interactions to facilitate viral release, analogous to the role of neuraminidase (NA) during IAV release, as it was documented for HKU1-CoV

(228). As expected, CoV HEs are capable of binding to host sialic acids (228, 229). It has also been documented that JHM-CoV isolates can vary greatly in their HE expressions, especially after laboratory passaging (226, 227).

Figure 15. MHV-JHM CoV Agglutinates RBCs. (A) Western blot detection of virion hemagglutinin esterase (HE) proteins. Purified recombinant JHM-CoV particles, normalized based on viral RNA abundance, were subjected to western immunoblotting. Virion HE proteins were detected using mouse α-MHV-HE 5A11. (B) Western blot detection of virion S proteins. Purified A59-CoV and JHMHE--CoV were subjected to western immunoblotting. Virion spike (S) proteins were detected using mouse α-MHV-S (10G). The relative band intensities were used to normalize viral input in the hemagglutination assays. (C) Hemagglutination assay. JHMHE--CoV, A59-CoV, or IAV pps were serially 2-fold diluted from left to right, placed into V-bottom wells, and incubated for 2 h at 4°C. Human adult erythrocytes were then added, to a final 0.25% v/v. Hemagglutination was scored (positivity highlighted in blue circles) after 2 to 12 h at 4°C. The experiment was performed four times.

Therefore, we selected a JHM-CoV isolate that lacks HE expression (Figure 15A), so that any observable sialic acid interactions are likely attributed to S proteins.

Sialic acid-binding viruses will agglutinate erythrocytes. To determine whether

sialic acid binding has a critical role in JHM-CoV entry, we evaluated hemagglutination

of human erythrocytes. Influenza A virus hemagglutinin (HA)-bearing pseudoparticles

(IAVpps) were used as a positive control. As a comparison, another MHV strain, A59-

CoV, was used, whose S protein has not been found to interact with sialic acids (125).

Because JHM and A59 vary in their S protein stabilities, we normalized particle input based on their spike contents (Figure 15B). Viral particles, or mock, were placed in 2- fold dilutions across the V-bottom plate and incubated for 2 h. Subsequently, human erythrocytes were mixed in and hemagglutination titers were scored after 2-12 h.

Intriguingly, JHMHE--CoV virus particles agglutinated human erythrocytes, while the

related A59-CoV strain did not (Figure 15C), even at particle-associated S protein input

levels slightly exceeding the corresponding JHMHE—CoV. These findings suggested that

JHM uniquely engages host sialic acids, and this interaction is likely facilitated by its S

protein.

JHM Spike Protein Engages Sialosides on Target Cells

To validate the presumed JHM virus spike – host cell sialic acid interaction, we

needed a system where we could compare sialic acid interactions down to a single

variable, the S protein. The feasibility of any potential system was further complicated

by the finding that weak sialic acid interactions are not detectable unless the S proteins

are oligomerized similarly to the authentic virus. Indeed, while the soluble domain As

elicited no detectable readout, sialic acid interactions were readily detected when the

domain As of MERS-CoV and HKU1-CoV were oligomerized onto nanoparticles (125,

126). Additionally, we noted that the aforementioned in vitro fusion system (see

SECTION 1) was likely not suitable to detect weak sialic acid interactions (Figure 11B).

Because we expected a weak JHM S-sialic acid interaction, we developed an

analogous system with virus-like particles (VLPs), where S proteins are oligomerized

similarly to the authentic virus. It has been shown that CoV major structural proteins

self-assemble into VLPs (1, 2, 4, 5, 230). We developed a MHV-based VLP system, in

which VLPs contained reporter proteins that can be tracked during virus-cell binding and entry (Figure 16A). To produce MHV-VLPs, HEK293T cells were co-transfected with the expression plasmids for the structural genes E, M, N, and either S or vector control (1,

3, 4). The N genes plasmids were modified by in-frame fusion with dual-split protein

(DSP) encoding sequences. While devoid of signals in separation, when DSP1-7 encounters DSP8-11, the two molecules complement to form a complete Renilla

luciferase (Rluc) and green fluorescent protein (GFP) chimeric reporter (217, 218).

Therefore, the incorporation of DSP1-7-N and DSP8-11-N proteins into VLPs allowed for

the release of particles that were intrinsically Rluc- and GFP-positive. These MHV-VLPs were further purified from cellular debris through a 20% sucrose cushion.

Next we needed to analyze the VLPs, and test whether they resembled authentic

MHVs, specifically, whether the S proteins were incorporated similar to the authentic virus. To this end, we analyzed the state of the incorporated JHM S proteins on these

VLPs, via immunoblotting.

Figure 16. MHV-JHM VLPs Resemble Authentic Viruses. (A) Schematic for VLP production and purification. Supernatants containing MHV-CoV VLPs were collected from HEK293T cells expressing viral structural genes S, E, M, and DSP1-7-N / DSP8-11- N. VLPs were purified through 20% sucrose cushions and quantified by assessing Rluc levels. Both DSP reporter halves were used to generate VLPs that contained internal Rluc and GFP. (B) Western blot detection of JHM S proteins in JHM pp (left), JHM VLP (middle), and authentic JHM-CoV (right). Particle-incorporated spike (S) proteins were detected using mouse α-MHV-S (10G), which binds to the C-terminus of the S2 subunit. The peptide molecular weight associated with the S2 subunit is an indicator of whether the S protein has been uncleaved at S1/S2 site (~200 kDa.), cleaved at S1/S2 (~110 kDa), or degraded beyond a functional size (~40 kDa.). Particles were produced in parallel and purified identically.

We noted that MHV VLPs incorporated intact JHM spike proteins at levels that were

similar to authentic viruses (JHM CoV), indicated by the prominent S2 band at ~100

kDa. This was in contrast to JHM spikes that were incorporated into VSV-based pseudoparticles (JHM pp), a widely utilized approach to assess CoV S functions, where most of which were degraded, as indicated by a prominent S2 band at ~40 kDa (Figure

16B). These findings indicated that the VLPs structurally resemble authentic virus

structures, at least with respect to spike stability and particle incorporation. As spike

stability and multivalent presentation is frequently required to detect sialate interactions

(125, 126), we considered the VLPs to be well-suited to reflect S-sialioside interactions of the authentic virus. Additionally, VLPs are equally produced either with or without incorporated S proteins (2), making spike-less (No S) VLPS the ideal control for any

non-S-dependent signals, making it possible to quantify S protein – dependent virus-cell binding processes. Furthermore, we reasoned that, with the spike-less VLP as background control, we could minimize / avoid cell-rinsing and be well-equipped to detect weak viral associations with target cells.

To this end, we first generated VLPs either with or without JHM or A59 S proteins. Via immunoblotting, we noted that the S protein levels on the A59 VLPs exceeded those of JHM VLPs (Figure 17A), even though the VLP quantities were normalized based on their internal Rluc signals (Figure 17B). This was expected, since it was known that A59 S is more stable than JHM S (147, 192). Interestingly, when these VLPs were applied to mCEACAM-negative HeLa cells at equivalent Rluc multiplicities, in agreement with the hemagglutinating effects of the viruses, the JHM S proteins increased VLP : HeLa cell binding above background levels (No S, subtracted from each data point), while the A59 S proteins did not (Figure 17C).

To assess whether the observed, JHM S-facilitated, VLP-binding is mediated by target cell sialiosides, we compared VLP-binding to cells either with intact sialosides or with sialosides removed by neuraminidase (NA) from Clostridium perfringens prior to

VLP innoculation.

Figure 17. JHM S Protein Engages Sialosides on Target Cells. (A) Western blot detection of A59 and JHM S proteins in VLPs. Particle-incorporated spike (S) proteins were detected using mouse α-MHV-S (10G). (B) Bald (No S), A59 or JHM S-bearing VLPs contained equal internal Rluc activities. (C) VLP binding assay. HeLa cells were formalin-fixed, rinsed, and incubated for 2h at 4°C with VLPs bearing A59 or JHM S. Rluc input multiplicities were equal for all VLP inoculations. Media were removed, cells lysed, and cell-associated Rluc activities quantified. Data are presented after subtracting background (“No S”) Rluc+ VLP levels, with each data point representing averages from independent experiments (N = 8 (A59) and 7 (JHM); n = 4 technical replicates per experiment). Error bars present standard error (SE) from the mean. Statistically significant deviations from background (“No S”) binding were assessed by unpaired Student’s t test; ns, not significant. * p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001. (D) HeLa or HeLa-mCEACAM cells were fixed and treated with vehicle (PBS) or neuraminidase (NA) for 3 h at 37°C. JHM-CoV VLPs were then added for 2h at 4°C, cell-associated Rluc activities quantified, and data presented after subtracting background (“No S”) Rluc+ VLP levels. Error bars present standard error (SE) from the mean. Statistically significant deviations were assessed by unpaired Student’s t test; ns, not significant. * p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001.

Notably, the JHM VLPs did not bind to HeLa cells that were pre-treated with NA (Figure

17D, left), suggesting that the observed VLP binding was dependent on cellular sialic

acids, which serves as an attachment factor for JHM S.

This result appeared to contradict with the general belief that the principle

receptor, mCEACAM in the case of JHM, is responsible for viral attachment. Therefore,

we tested the role of mCEACAM in JHM VLP binding. Surprisingly, VLP binding to cells

was not increased by the expression of the MHV protein receptor mCEACAM in HeLa

cells, and more importantly, all observed VLP-bindings to mCEACAM-expressing HeLa cells remain completely sensitive to the pre-removal of sialic acids by NA (Figure 17D, right). Together these data suggest that, during MHV-JHM infection, the initial viral

attachment process is mediated by S protein:sialic acid interactions, and not by the

protein receptor mCEACAM.

JHM Spike Requires Protein Receptor for Membrane Fusion During Viral Entry

Since our VLP-binding results suggested that the protein receptor, mCEACAM, is not the major contributor to viral binding (Figure 17D, right), we wanted to test its involvement during viral entry. In addition, since we have identified the prominent role of sialic acids on JHM viral binding, we next aimed to identify the biological importance of the JHM spike-sialate interaction. We reasoned that this interaction could either be sufficient for viral entry or may instead be an early step preceding viral engagement with protein (mCEACAM) receptors. We modified our VLP system to turn it from a binding assay to an entry assay. Instead of transfecting in plasmids for both DSP1-7-N and

DSP8-11-N proteins, which generates Rluc- and GFP-positive VLPs, only the DSP8-11-N

plasmid was included. This produced DSP8-11-positive, but Rluc- and GFP-negative

VLPs. To quantify the relative numbers of these DSP8-11-positive VLPs, purified VLP samples were lysed, and subsequently mixed with cell lysates containing a surplus of

DSP1-7 proteins. The Rluc signals generated from the post-lysis complementation of

DSPs correlated with the number of VLPs supplied. To assess virus entry, we

inoculated equal Rluc multiplicities of either spike-less or JHM S, DSP8-11-positive, VLPs

onto DSP1-7-expressing target cells. Because the delivery of viral content into target cell

cytoplasm is the direct outcome of viral entry, VLP entry into target cells could be

measured by reporter complementation into active Rluc or GFP moieties (Figure 18A).

S-dependent VLP entry events could be quantified by the elevated signals of JHM S

VLPs over the background signals of spike-less VLPs. We compared VLP entries in

HeLa and HeLa-mCEACAM cells and found that, even though JHM S : sialic acid, but not JHM S : mCEACAM, interactions facilitated significant VLP binding (Figure 17D, right), Rluc entry signals required mCEACAM presence (Figure 18B), measured by the

~5-fold signal of JHM S VLPs over spike-less (So S) VLP background. The entry

signals-to-background ratio were further elevated ~10-fold by the presence of

TMPRSS2 (Figure 18B, hTMPRSS2), a known CoV fusion-triggering protease (99, 100,

167, 170, 173, 182, 190, 194, 212, 231). This facilitating effect of TMPRSS2 supported the contention that the VLP-based entry assay faithfully reflected authentic CoV entry processes.

Figure 18. Viral Binding to Sialosides Facilitates Protein Receptor-dependent Entry. (A) Schematic for VLP entry assay. VLPs containing DSP8-11-N were collected from transfected HEK293T cells, purified as described in Figure 16A, and inoculated onto target cells expressing DSP1-7. Rluc signals arise after the fusion between VLP and target cell membranes, which enables the entry of DSP8-11-N into cells and subsequent complementation with DSP1-7. (B) HeLa target cells were transfected with DSP1-7, mCEACAM, and hTMPRSS2, as indicated. At 2 d posttransfection, DSP8-11- VLPs were collected, purified, and inoculated for 6 h at 37°C, then cells were lysed and Rluc activities measured. (C) The indicated target cells were pretreated with vehicle or NA for 2 h at 37°C, followed by inoculation with the indicated virus particles for 2 h at 4°C. Unbound particles were removed by rinsing, and after 16 h at 37°C, cells were lysed and Fluc quantified to approximate viral entry. Data are presented as % viral entry, normalized to vehicle controls. Statistical analyses were performed as described in the Figure 17 legend.

In conjunction with VLP binding data, these findings suggest that JHM-CoV cell entry is

a multi-step process, with initial viral attachment facilitated by S : sialate interactions,

and subsequent viral fusion initiated by S : mCEACAM engagements.

Virus Binding to Sialic Acids Facilitates Protein Receptor-dependent Entry

Using our VLP system, we have identified that JHM S binds to host sialic acids.

We next aimed to determine whether this JHM S : sialic acids interaction has a detectable biological role. To this end, we tested whether NA pre-treatment of target cells affected protein-receptor-dependent JHM-CoV entry. Here authentic JHMHE--CoV particles were bound for 2h to different target cells, with or without prior cell exposures to NA. We included a small collection of different susceptible target cells: Mouse brain- derived DBT cells were a common murine cell type used to grow MHVs (232), and we included this cell type to reflect common JHM entry conditions; Human HeLa cells were engineered to stably express the MHV murine protein receptor, mCEACAM (205). We included this cell type to represent a scenario where an originally non-permissive cell type were made susceptible by the protein receptor alone; Hamster-derived BHK cells were included to represent zoonotic cell types that, even though lacking mCEACAM, are nonetheless susceptible to JHM-CoV infections (197). Because host sialic acids are the receptors for IAV HA (233–235), IAVpps were used as a positive control for neuraminidase-mediated de-sialylation. VSVGpps were utilized as the negative control, because VSVG engages LDLR, and not sialic acids, to facilitate viral entry (164). After viral binding at 4°C, unbound particles were removed via extensive rinsing. Viral replication was continued for 18 h, before it was measured by quantifying viral Fluc reporter expression. Interestingly, NA pre-treatments universally decreased JHMHE--CoV entry into DBT (38%), HeLa-mCEACAM (38%), and BHK (54%) cells (Figure 18C).

After NA pre-treatments, the viral entry to these cells were similarly reduced for IAV pp

79 transduction, (Figure 18C). Partial resistance to IAV pp is best explained by incomplete removal of sialic acids by either the type or the dose of they neuraminidase used. We therefore infer that the partial NA resistance of JHMHE--CoV is similarly explained by incomplete sialic acid removal. On the other hand, the NA treatment deployed here likely did not cause any relevant secondary effects, because the transduction efficiency of VSVG pp was not hindered. These results support a model in which the sialic acid moieties on the cells from a wide range of animal species universally function as the attachment factor for the initial JHM virus binding to target cells, and this attachment aids the downstream protein receptor-mediated entry process.

Sialic Acid Receptors Promote JHM Viral Spread without Requiring mCEACAM

Protein Receptors

To spread existing infections, CoVs utilize two distinct pathways. De novo CoV virions are produced from infected cells and released into the extracellular space, where they can be passively transported into close proximity to other naïve, but susceptible cells to initiate a new round of viral infection, via the process of canonical CoV entry.

Alternatively, a CoV-infected cell can directly spread its infection into neighboring cells through the process of cell-cell fusion, or syncytia formation, mediated by CoV S proteins on the infected cell surface interacting with the cellular factors on the neighboring cell plasma membrane (99). For most CoV S proteins, cell-cell fusion requires the engagement with their principle receptors (194, 195, 212, 236). JHM S proteins can uniquely mediate syncytia in several conditions, even when mCEACAM protein receptors are absent (196, 197). Previously, this mechanism for this

mCEACAM-independent fusion activity was proposed to not require any receptor

engagements.

Because we have identified a weak JHM S : sialic acid interaction, we wanted to determine whether sialic acids operate independent of mCEACAM receptors in JHM cell-cell membrane fusion and syncytial development. Because our results suggested that we were inefficient at removing sialic acids from live cells (Figure 18C), we deployed an independent approach to assess the efficiency of live-cell de-sialylation.

Wheat germ agglutinin (WGA) is known to bind various sialic acid species on the cell

membrane (237). We incubated live HeLa or HeLa-mCEACAM cells with titrated doses of NA. Subsequently, cells were incubated with fluorescently-labeled WGA at 4°C, to eliminate the internalization of WGA. Confocal micrographs of the subsequently fixed cells revealed that, WGA-binding was reduced by NA in a dose-dependent manner

(Figure 19A). WGA-binding, however, was not eliminated even at the highest NA dose

tested. Cell morphology was not visibly altered at any of the NA doses tested, in

agreement with the absence of detectable cytotoxic effects during VSVGpp transduction

(Figure 18C).

To determine the effect of target cell sialic acids on the JHM-CoV cell-to-cell

spread, we utilized the DSP-complementation system and developed an assay to

directly quantify JHM-CoV-induced viral expansion via cell-cell fusion. First, we

inoculate JHMHE--CoV with high multiplicity of infection (MOI = 10) onto permissive DBT

cells.

Figure 19. JHM-CoV Cell-to-cell Spread Depends on Host Sialic Acids. (A) Immunofluorescence micrographs of target cells after NA treatments. HeLa or HeLa-mCEACAM cells were pretreated with NA at the indicated doses at 37°C. Following rinses, cell surface sialate levels were assessed by co-incubating cells with AlexaFluor 647-conjugated wheat germ agglutinin (WGA-Alexa647, Molecular Probes) at 4°C. Following fixation and Hoechst staining, images were captured by EMCCDCascade 2 camera and processed in Imaris 8.3.1. Images are representative of three technical repeats. (B) Schematic for JHM S protein-mediated cell-cell fusion measurements. JHM-CoV infected DBT cells (moi = 10; 5 hpi) were added to target HeLa / HeLa-mCEACAM cells. The target cells were co-cultures of DSP1-7 and DSP8-11- expressing cells. JHM-CoV infected cells fuse with multiple target cells, which enables DSP1- 7:DSP8-11 complementation into active Rluc. (C and D) Quantification of Rluc over time in mCEACAM-negative (C) or mCEACAM-positive (D) target cells, in the presence of the indicated NA concentrations. Means (data points), SE (error bars), and the polynomial trendlines (R2 > 0.9) are shown. The results are representative of two biological repeats.

At 5 hpi, we rinse away unbound particles, lift the DBT cells, and overlay them onto target cells consisting a one-to-one mix of DSP1-7-expressing cells and DSP8-11-

expressing cells. Overtime, if the cytoplasms of the DSP-expressing cells are bridged

together by JHM S-mediated cell-cell fusions, then DSP1-7 would complement with

DSP8-11, and Rluc and GFP signals would arise and be quantified (Figure 19B). Using a

stable live-cell Rluc substrate, the development of active Rluc was measured over time.

To analyze the role of target cell sialic acids on JHM S-mediated cell-cell fusion, a titration of neuraminidase doses were present throughout the time-course experiment.

For JHM-CoV spread into mCEACAM-negative HeLa cells, the results (Figure 19C) revealed a NA dose-dependent reductions in JHM CoV-induced syncytia, where syncytia formation was completely eliminated at the highest 2 U/ml NA dose. This result suggested that JHM S-mediated, mCEACAM-independent cell-cell fusion heavily depends on target cell sialic acid moieties. Similar, but less pronounced reductions were observed in parallel cultures of mCEACAM-positive HeLa cells (Figure 19D), suggesting that sialic acid interactions may assist mCEACAM-dependent JHM S cell-cell fusion.

From these findings, we conclude that sialic acids can enable a CoV S-mediated cell- cell fusion process without requiring a prototypic proteinaceous receptor.

A JHM-CoV Spike Mutation Increases mCEACAM-independent Cell Binding

To date, analyses of the multi-RBD CoV S proteins have revealed that all identified sialic acid interactions exist on S protein domain As (105, 113, 125, 126), while most protein receptor interactions point to S protein domain Bs (103, 106, 107).

The exceptions are the MHV beta-CoVs, which have their domain As binding to protein

(CEACAM) receptors (104). From these findings, and our findings that JHM S does interact with host sialic acids, we inferred that the JHM-CoV S domain A could contain a novel dual-receptor binding capability, able to bind both sialic acid and CEACAM receptors. If this were true, then we expected to be able to manipulate the sialic acid- dependent JHM S phenotypes via specific domain A mutations. Sialic acid binding sites on beta-CoV S proteins were originally inferred from mutagenesis studies (105), and most recently, from structural resolution of S proteins in complex with sialosides (113,

114). Mutations in the inferred site did decrease sialoside binding (105, 125), even though they are distal from the structurally resolved sialoside binding grooves (113,

114). Upon comparings the sialoside binding site and the distal site between MHV and the sialic acid-binding B-CoV domain As, we noted that, while the sialoside binding sites were relatively divergent, the distal sites were surprisingly conserved, based on both the crystalized structures and the amino acid sequences. Here we noted a single variation in the distal site. Among a stretch of 4 residues important for sialate-binding activity of

B-CoV S domain A, JHM differed only at residue 176, where there is a Gly on a divergent loop in place of the orthologous B-CoV Glu170 (Figure 20A).

To evaluate the importance of this divergence to sialic acid binding, we constructed a G176E – mutant JHM-CoV S protein. We hypothesized that this single- residue change could enhance the mCEACAM-independent activities of JHM S. We first verified that G176E S protein expression, proteolytic processing, and VLP-incorporation were all comparable to WT S (Figure 20B).

Figure 20. JHM S G176E Mutation Elevates Target Cell Binding. (A) Structural superimposition and sequence alignment of B-CoV S domain A (PDB: 4H14, green) and A59 S domain A (PDB: 3JCL, magenta) using PyMOL. Side chains in stick representation are included for residues of interest. The sialic acid binding site is circled with dash line, and the distal site is boxed. (B) Western blot detection of VLP-associated S proteins. Particle-incorporated spike (S) proteins were detected using mouse α-MHV- S (10G). (C) VLPs lacking S proteins, or bearing JHM (WT) or JHM (G176E) S proteins, were applied to HeLa cells pretreated with either vehicle or NA, at equivalent Rluc levels. Experimental procedures, data acquisition and processing were performed as described in the Figure 17 legend.

Notably, when VLPs bearing no S, WT S, or G176E S were inoculated onto HeLa cells at equivalent Rluc multiplicities, about twice more G176E-VLPs bound to HeLa cells than WT. Importantly, and G176E spike-containing VLP binding remained very sensitive

to NA pre-treatment of the cells (Figure 20C). Because a single residue mutation on

JHM S domain A conferred a gain-of-function to cell binding, our results suggested that

JHM S engages target cell sialic acids via its S domain A, supporting the notion that

JHM S domain A is an evolutionary intermediate between sialic acid binding and protein receptor engagement.

A JHM-CoV Spike Mutation Increases mCEACAM-independent Membrane Fusion

If sialic acid engagement enables JHM S mCEACAM-independent cell-cell fusion, then a gain-of-function mutation such as G176E should enhance this fusion signature. To correlate the increased cell binding of G176E mutant with membrane fusion, we compared the syncytial potency of JHM wild type and G176E-mutant S proteins. The DSP-complementation-based cell-cell fusion system was modified slightly.

Cell-cell fusion effector cells were co-transfected with expression plasmids for DSP1-7,

with JHM WT S, G176E S, or vector control. Target cells were co-transfected with

plasmids for DSP8-11, with mCEACAM or vector control. Effector and target cells were

mixed and cell-cell fusion measured by DSP 1-7 : DSP 8-11 complementation (Figure

21A), visualized as GFP-positive syncytia via confocal fluorescent microscopy (Figure

21B), and quantified by measuring Rluc signals (Figure 21C), where relative signals

over background (No S) were depicted. Of note, when mCEACAM receptors were

present, synyctia induced by wild type and mutant S proteins developed comparably,

implying that both utilized mCEACAM for membrane fusion comparably.

Figure 21. JHM S G176E Mutation Increases mCEACAM-independent Cell-cell Fusion. (A) Schematic for JHM S protein-mediated cell-cell fusion measurements. GFP and Rluc signals arise only after S-expressing effector cells fuse with target cells, which enables DSP1-7:DSP8-11 complementation. (B) Live-cell images of the syncytia formed by S-expressing cells, without (left) or with (right) mCEACAM on target cells. Images were captured by EMCCDCascade 2 camera and compiled using Imaris 8.3.1. (C) Quantification of syncytial developments. After imaging, cells were lysed and their Rluc activities quantified. Data are presented as fold-increases above background (“No S”) values. Statistical analyses were performed as described in the Figure 17 legend.

In contrast, in the absence of mCEACAM, the G176E spikes were significantly more

robust in their syncytium-inducing activity (Figure 21C), in agreement with the enhanced

VLP cell binding capabilities (Figure 20C). We concluded that a single point mutation in

the JHM-CoV S domain A increases the cell binding and membrane fusion activities of

S proteins, independent of the prototype JHM protein receptor. Furthermore, our data

supported the JHM neuropathogenesis model where JHM interneuronal spread is

facilitate by S : sialic acid interactions.

SECTION 3: MERS Spike Utilizes Host Sialosides During Viral Entry and Spread

MERS-CoV Spikes Mediate Cell Fusion without Requiring hDPP4 Protein

Receptors

MERS-CoV S domain Bs bind to host protein (hDPP4) receptors (106, 137, 154),

and MERS-CoV S domain As bind to host sialic acids (114, 126). While the principle

receptor hDPP4 confers cells susceptible to MERS-CoV infections (137, 154, 161),

recent studies have revealed some biological importance for MERS S domain A :

sialoside interactions on viral infections. First, de-sialylation significantly reduced MERS

pp and MERS-CoV entry into susceptible cells (114, 126). Also, antibodies against

MERS S domain A can be neutralizing (149, 150), and protect mice from dying (151). In

constrast, how and whether host sialosides can be utilized for MERS S-mediated cell- cell fusion have not been explored. Our recent data suggested that weak sialic acid interactions can be sufficient for protein receptor-independent viral spread, which could be a major contributing factor to specific CoV tropism and pathogenesis. This raised the question of whether MERS-CoV spikes can catalyze cell fusions after the binding of their S domain As to cellular sialosides, and independent of high-affinity protein receptors, similar to the JHM-CoV spikes. To address this question, MERS-CoV S-

mediated cell-cell fusions between hDPP4-negative DBT cells were measured by DSP

1-7 : DSP 8-11 complementation. Similar to the JHM S cell-cell fusion analysis in the previous section, effector cells expressing DSP1-7, MERS S or vector control, were

mixed with target cells expressing DSP8-11 and various cellular factors (Figure 22A).

Cell-cell fusions were quantified by measuring Rluc signals from DSP-

complementations, and relative signals over background (No S in effector cells and no

exogenous factors in target cells) were depicted (Figure 22B). First, comparing to the

No S background control, MERS S facilitated cell-cell fusion in the presence of hDPP4,

and the Rluc signal further increased ~6-fold upon introduction of the triggering protease

hTMPRSS2. This is in agreement with the current understanding of MERS S fusion

being S-dependent, hDPP4-facilitated, and protease-triggered (160, 170, 238).

Remarkably, our data suggested that syncytial development mediated by MERS-CoV S

proteins did not require hDPP4, but did require sufficient levels of the S protein-

activating protease hTMPRSS2 (Figure 22B). Specifically, this fusion was not due to

MERS S using hTMPRSS2 as a surrogate receptor, because the catalytically-inactive

TMPRSS2 (S441A) did not facilitate any cell-cell fusion (Figure 22B). The MERS S-

mediated cell fusions contrasted with those generated by 229E-CoV S proteins. We

noted that 229E S-mediated fusions absolutely required target cell hAPN protein

receptors, whose fusion was further elevated ~50-fold with added hTMPRSS2

proteases (Figure 22B). Interestingly, 229E S do not bind sialic acids (126), suggesting

the observed MERS S-mediated, hDPP4-independent cell-cell fusion could be enabled

by target cell sialic acids.

Figure 22. MERS S Mediates Cell Fusion without Requiring hDPP4 Protein Receptors. (A) Schematic for MERS S protein-mediated cell-cell fusion measurements, where murine DBT effector cells either with or without MERS S were laid onto DBT target cells either with or without hDPP4 or hTMPRSS2. (B) Effector cells expressing the indicated S proteins were cocultured with target cells expressing hTMPRSS2 or the catalytically inactive mutant hTMPRSS2 (S441A), and/or protein receptors (hDPP4 for MERS, and hAPN for 229E). The dotted line represents background Rluc signals. Data are presented as fold-increases above background (“No S vs. Vector”) values. Data acquisition and processing were performed as described in the Figure 17 legend.

Thus, with sufficient cell-surface protease activities, the primary hDPP4 receptor was dispensable, conceivably because the secondary sialate-binding RBDs can operate at a stage in the process. This is the first documentation of hDPP4-independent fusion by

MERS S proteins.

MERS Spike-mediated, Protein Receptor-independent Membrane Fusion Depends on S Domain A and Host Sialosides

To directly determine whether the observed hDPP4-independent MERS S cell-cell fusion is facilitated by S : sialoside interactions, secreted, soluble S domain A (S1A-Fc) proteins were constructed and added in titrated levels during the cell-cell fusion assays.

Notably, exogenous domain As interfered with hDPP4-independent MERS S-mediated fusion in a dose-dependent manner, but had no effect on fusion when hDPP4 was present (Figure 23A). This result suggested that the observed fusion was facilitated by

S domain A engagement with target cell factors. To determine whether target cell sialic acids are the domain ligand that operate in the cell fusion process, NA was added to cell-cell fusion time-course experiments (Figure 23B). Comparing to the MERS S- mediated DBT cell-cell fusion in the absence of hDPP4 (Figure 23C), whose fusion signal was inhibited by exogenous NA, exogenous NA had no effect on the cell fusion when hDPP4 was present (Figure 23D), implying a role of host sialosides in hDPP4- independent fusion. Taken together, these results further suggested that either of the two S RBDs, domain A or domain B, can tether S proteins to target cells for subsequent cell-cell fusions, with distinct host receptors for each domain.

Figure 23. MERS hDPP4-independent Cell-cell Fusion Depends on Host Sialosides and S Domain A. (A) MERS S-expressing effector cells were mixed with hTMPRSS2 or hTMPRSS2/hDPP4 target cells, in the presence of a titration of soluble MERS-S1A-Fc proteins. Data acquisition and processing were performed as described in the Figure 17 legend. (B) Schematic for MERS S protein-mediated cell-cell fusion measurements. Rluc signals arise only after S-expressing effector cells fuse with target cells, which enables DSP1-7:DSP8-11 complementation. Syncytial development was quantified by measuring Rluc signals over time. (C and D) The kinetics of syncytial developments of hDPP4-negative (C) or hDPP4-positive (D) target cells, in the presence of the indicated NA concentrations. Means (data points), SE (error bars), and the polynomial trendlines (R2 > 0.9) are shown. The results are representative of two biological repeats.

An Analogous MERS-CoV Spike Mutation Similarly Increases hDPP4-independent

Cell Binding

We next determined whether analogous changes in S domain A of the related

MERS-CoV also modified viral attachment. The MERS S domain A interaction with

sialic acids has biological significance, as antibodies against S1A protects model mice

from lethal MERS-CoV infections (151). Additionally, domain A functions may be closely

linked to CoV zoonotic potentials. This is supported by findings that MERS S domain A

polymorphisms exist in humans (163) and adaptive changes occur in mouse models of human MERS-CoV lung infection (161, 162). In mice, this selectivity was at Asn222

(161, 162), a known glycan addition site (112, 141, 204). The specific phenotype for the

adaptation at residue 222 has been unclear. Furthermore, evolutionary speaking,

Asn222 is a documented hypervariable residue. Among the two most MERS-related bat

CoVs, HKU4 and HKU5, the former group has maintained this glycosylation site while

the same site is absent in the latter group (Figure 24A). Remarkably, upon

superimposing the domain A structures for B-CoV, MHV, and MERS-CoV, we noted

that the MERS Asn222 position overlaps closely with B-CoV residue Glu170 and MHV-

CoV residue Gly176, in the putative sialioside-binding distal site (Figure 24B). Due to

this profound locational similarity, we constructed MERS VLPs to determine whether

this S proteins with the Asn222 change impacted cell binding. Because of the weak-

interaction nature between MERS S and sialosides, the VLP system was adapted to

make MERS S-bearing particles where the S proteins are oligomerized similar to

authentic MERS-CoVs.

Figure 24. MERS S N222 is a Hypervariable Residue within the Sialoside-binding Distal Site. (A) Comparison of selected MERS S sequences relative to bat MERS-like HKU4 and HKU5. MERS S residue N222, as well as hDPP4-binding loops 1 and 2, are delineated by red boxes. MERS S1 structure is from PDB: 5X59. (B) Structural superimposition of B-CoV S domain A (PDB: 4H14, green), A59 S domain A (PDB: 3JCL, magenta), and MERS S domain A (PDB: 5X4R, yellow) using PyMOL. MERS S1A residue N222 (stick representation) is located proximally to B-CoV S1A residue E170, in the sialic acid distal site (boxed).

To test this, we produced MERS-CoV VLPs and evaluated their sialate binding

properties. Comparing to spike-less (No S) MERS VLPs, S-bearing VLPs, when inoculated at equivalent Rluc multiplicities, enabled agglutination of human erythrocytes, and this agglutination was eliminated via NA pre-treatment of erythrocytes (Figure 25).

This result implied that the MERS VLPs did incorporate and display S proteins the same way as authentic viruses, which enabled detectable sialoside-binding activities. We then constructed wild type (WT) and N222D VLPs containing internal Rluc. We then verified that WT and mutant S proteins were processed and incorporated comparably (Figure

26A).

Figure 25. MERS-CoV VLPs Resemble Authentic Viruses. (A) The indicated IAV pps or MERS VLPs were placed into V-bottom wells and incubated for 1 h at room temperature. The VLPs (No S or MERS S) were present at equivalent levels, as measured by their internal Rluc contents. Pre-warmed (37°C) human adult erythrocytes, either with or without prior NA treatment, were added to a final 0.25% v/v. Hemagglutination was scored after 2 to 12 h at room temperature. The experiment was performed two times. (B) MERS VLPs (No S or MERS S) were serially 2-fold diluted, placed into V-bottom wells, and incubated for 1 h at room temperature. Pre-warmed (37°C) human adult erythrocytes were added to a final 0.25% v/v. Hemagglutination was scored after 2 to 12 h at room temperature. The experiment was performed two times. Positivity was highlighted in blue circles.

After normalizing particle input of VLP bearing no S, WT S, and N222D S based on their internal Rluc signals (Figure 26B), VLPs were inoculated onto various target cells, and binding was assessed by quantifying cell-associated Rluc. Intriguingly, we saw the effect of G176E mutation on JHM S mirrored by N222D mutation on MERS S. First,

N222D did not alter VLP binding to permissive human lung derived Calu3 cells. This was in stark contrast to the VLP binding signatures for the non-permissive mouse lung derived LET-1 and mouse brain derived DBT cells, whereas WT S mediated very modest levels of VLP binding, N222D S elevated that binding dramatically (Figure 26C).

While the involvement of MERS S domain A in VLP binding to Calu3 cells was difficult to assess, mainly due to the fact that these human cells express hDPP4 protein receptor for MERS-CoV, the mouse-adaptive mutation at Asn222 clearly enhanced the ability of S domain A to bind to murine cells (Figure 26C).

An Analogous MERS-CoV Spike Mutation Similarly Increases hDPP4-independent

Membrane Fusion

Since we had observed remarkable similarities on the effect of mutations in the sialoside-binding distal site on viral binding for both JHM and MERS S proteins, we next considered whether this MERS-CoV adaptation for increased cell binding had consequences in MERS S – mediated cell-cell membrane fusion, similar to the JHM-

CoV G176E mutant. To this end, cell-cell fusion assays were set up to observe the effect of N222D change either in the presence or absence of the hDPP4 protein receptor, both in the presence of sufficient levels of the hTMPRSS2 protease.

Figure 26. MERS S N222D Elevates Target Cell Binding. (A) Western blot detection of MERS WT and N222D S proteins in VLPs. Particle-incorporated spike (S) proteins were detected using mouse α-C9. (B) Bald (No S), WT or N222D S-bearing VLPs contained equal internal Rluc activities. (C) VLPs were added at equivalent Rluc input multiplicities to human Calu3, murine LET-1 cells, and murine DBT cells. After 2h at 4°C, cell-associated Rluc activities were quantified, and data were presented after subtracting background (“No S”) Rluc+ VLP levels. Data acquisition and processing were performed as described in the Figure 17 legend.

Again mirroring the effects we saw for JHM S, we found that the N222D adaptive mutant S proteins did indeed exhibit enhanced cell fusion, relative to wild type S proteins (Figure 27A), but this was only observed by eliminating hDPP4 (Figure 27A, left), which removes the dominant domain B : protein receptor interaction and demands domain A utilization. To further probe the relationships between domain A-cell binding

and cell fusion, specifically to verify whether the adaptive effects of N222D is indeed at

the level of domain A : host engagement, we determined whether soluble N222D

domain A (N222D S1A-Fc) proteins would suppress cell fusion more effectively than

soluble WT domain A (WT S1A-Fc) proteins. We therefore constructed N222D S1A-Fc

and added it at equivalent molar ratios with WT S1A-Fc into the cell-cell fusion assay

(Figure 27B). Soluble mCEACAM (mCEACAM-Fc) proteins were added as the negative

control. Indeed we found that exogenously added N222D S1A-Fc proteins suppressed

hDPP4-independent cell-cell fusions more potently than the corresponding WT S1A-Fc

(Figure 27C), further reinforcing the direct relationship between domain A binding to

cells and resultant cell-cell fusion and the adaptive benefits of N222D. Altogether, these

findings support the hypothesis that MERS-CoV and JHM-CoV can acquire increased cell binding through S protein domain A mutations, which consequently allow for more robust cell-cell fusion capability. Both of these CoVs can procure these cell binding and cell fusion properties through similarly-localized S1A mutations, suggesting a common

CoV evolution strategy for expanded tropisms, including zoonosis.

Figure 27. MERS S N222D Increases hDPP4-independent Cell-cell Fusion. (A) Effector cells expressing the indicated S proteins were cocultured with target cells expressing hTMPRSS2 and/or hDPP4. (B) Western blot detection of various purified Fc- tagged proteins. Protein concentration was assessed using Nanodrop (Thermo Fisher), and hIgG1 (Sigma) was loaded as a quantification control. (C) MERS S-expressing effector cells were mixed with hTMPRSS2-expressing target cells, in the presence of the indicated Fc proteins. All Fc proteins were at 10 uM concentration throughout the coculture period. Data acquisition and processing were performed as described in the Figure 17 legend.

CHAPTER IV

DISCUSSION

Summary of Data

The central goal of our study was to identify and characterize specific ways CoVs

utilize their S proteins during viral entry and spread. Our methodology was to develop

and utilize reductionist approaches where different steps and individual factors can be

assessed in isolation. Our results indicate that CoVs that have developed high-affinity

interactions with protein receptors can nonetheless utilize weak interactions with host

sialic acids to assist in the establishment of de novo CoV infections, as well as

spreading existing ones. Interestingly, sialic acid interactions contribute to different

stages of S-mediated membrane fusion in viral entry than in viral spread. Remarkably,

CoVs readily fine tune these interactions to adapt to specific biological niches.

First, we sought to identify the requirements for CoV S-mediated membrane fusion, in the context of viral particles. Using an in vitro fusion detection system, we determined the minimum requirements for CoV viral particle fusion to be the CoV S protein, its protein receptor, and a triggering protease. This conclusion was drawn from the direct detection of membrane fusion events in the absence of live cells (Figure 9B), hence this system is devoid of other complex cellular factors, as well as the constant replenishment of these factors. In this system, we observed the role of hDPP4 to be durably tethering MERS S-bearing particles to the target membrane (Figure 11B), and

99 100

in this context, the presence of exogenous protease was sufficient to trigger membrane

fusion (Figure 14A and C). To our knowledge, this is the first characterization of MERS-

CoV S-mediated membrane fusion in a cell-free context, and the second for CoV entry

in general (239).

Next, we identified MHV-JHM elicits sialic acid interactions. To narrow the viral

binding agent of sialic acid, we developed an MHV VLP system and identified the

binding agent to be the JHM S protein. We demonstrated for the first time that sialic acid

interactions are responsible for most of the JHM viral attachment to target cells (Figure

17D), whereas its protein receptor mCEACAM is specifically required for viral entry

(Figure 18B). Sialic acid interactions were subsequently identified to assist JHM-CoV mCEACAM-dependent entry (Figure 18C). We also found that sialic acid is the cryptic receptor JHM S engages to mediate mCEACAM-independent cell-cell spread of infections into cells of various animal species (Figure 19C), a widely-recognized feat whose mechanism had been elusive until now.

We further identified a point mutation in JHM S domain A that specifically elevates the S-mediated, mCEACAM-independent, viral binding (Figure 20C), as well as enhancing mCEACAM-independent cell-cell fusion. The correlation of these findings suggest that, in addition to its mCEACAM engagement, JHM S protein domain A also engages host sialic acids. The dual-valent JHM S domain A therefore potentially qualifies as an evolutionary intermediate representing a switch between sugar-binding and protein-binding S proteins (105, 118).

101

Energized by our findings on the mechanism of JHM S-mediated mCEACAM-

independent cell-cell fusion, we were the first to observe an analogous, hDPP4-

independent cell-cell fusion mediated by MERS S proteins, in the presence of sufficient

levels of target cell surface proteases (Figure 22B). Similar to JHM S, the alternative

receptor MERS S utilized to mediate this fusion is likely sialic acids, because (1) this

fusion was not elicited by the sialic acid-insensitive 229E-CoV S (Figure 22B); (2) the

fusion depended on the binding activity of the sialic acid-binding S domain A (Figure

23A); (3) the fusion was reduced by the introduction of NA (Figure 23C).

Finally, because residue 222 of MERS S domain A is a hyper-variable region

(Figure 24A), and is under the selective pressure during viral zoonosis (161–163), we

decided to characterize its adaptive effects. We first noted that the structural position of

residue 222 overlaps with the gain-of-function mutation we identified in JHM S (Figure

24B). Remarkably, a change in the 222 position functionally resonated our observations for JHM S, where the mutant enhance viral attachment to target cells (Figure 26C), as well as elevated MERS S fusion capabilities against the hDPP4-negative DBT cells

(Figure 27A). These results are the first to identify functional advantages between wild type and adapted MERS S domain As, and they suggest that fine-tuning weak sialic acid interactions is an integral element for the zoonotic potential of MERS-CoV.

Overall, these results indicate that CoV S proteins utilize host factors in ways more extensive and intricate than previously appreciated. The same host factor can be differentially utilized during viral entry and spread. Weak interactions, which are difficult to identify in a laboratory setting, can be responsible for major entry steps. Additionally,

102

CoVs readily evolve modification to these modest interactions to adapt to new biological

niches.

In Vitro Particle Fusion Identifies Minimal Fusion Requirements

Fluorescence have been used to analyze CoV membrane fusion events (239).

Here, we used FRET induction as a direct readout for membrane fusion. In our in vitro

fusion system, the interior of one particle is labeled with a fluorescent donor (GFP) and

another an acceptor (mCherry). Because the particle interiors are separated by two lipid

bilayers and their associated membrane proteins (e.g. S, hDPP4; Figure 9B), only

through membrane fusion can the two fluorescent proteins be in close enough proximity

for FRET (Figure 13A). It has been previously demonstrated that MERS-CoV entry is spike-directed, receptor-facilitated, and protease-triggered (Figure 8B, and reviewed in

(160)). However, because traditional studies used live cells as viral entry targets, it has been difficult to assess the sufficiency of these entry factors in mediating MERS-CoV entry. The in vitro particle fusion system we developed avoids some major limitations of using live cells to analyze entry factors. First, cells maintain complex biological processes, involving a large number of factors that are generally difficult to assess in isolation from one another. This may leading to incorrect interpretations of factors for viral entry. Porcine epidemic diarrhea virus (PEDV) allegedly uses APN as its protein receptor (240), but studies ten years later have demonstrated that APN is not a functional receptor for PEDV, but rather, its enzymatic activity promotes PEDV viral infection (241, 242). Another familiar scenario is that the viral entry characteristics observed for one cell type may vary greatly in another cell type. These variations are

103

generally resulted from differences in the quantity and/or the quality of the various entry

factors in different cell types. The particles in our in vitro fusion assay have greatly-

reduced levels of complexity in terms of unaccounted cell factors. Additionally, these

particles are biologically dead, in that they contain no live signaling pathways that may

complicate the viral fusion process. Hence, when factors are characterized in this

reductionistic approach, they are very likely to operate independently from other cellular

processes, at least in the context of S-mediated viral fusion. Taken together, our data

suggest that, to fuse a viral envelope with a target membrane, MERS S requires only

hDPP4 engagement and proteolytic processing.

We have noted a particular limitation of the in vitro fusion system. MERS S has been observed to associate with host sialic acid species (126). This association, albeit

weak, is responsible for the initial cell attachment step prior to viral entry (114, 126).

Since we did not detect any particle co-localizations between S-bearing pps and Bald

pps (Figure 11B, S x Bald), our system does not appear to facilitate efficient MERS S :

sialic acid interactions. This may be explained by the differences in the three-

dimensional structures between a viral envelope and the cell plasma membrane. When

a virus lands on a cell, its MERS S proteins are able to interact with the various sialic

acid moieties on the relatively-flat plasma membrane. Overtime, what lacks in the

MERS S : sialic acid binding strength can be made up in quantity. Indeed, avidity is

required for efficient MERS S – sialic acid interactions (126). On the contrary, when two

virions interact, the contact zone is very limited. Together with a finite quantity of sialic

acids, viral envelopes likely do not enable weak sialic acid interactions. Interestingly, we

104

do see IAV HA-bearing pps co-localizing with Bald pps (Figure 11B, IAV HA x Bald),

mirroring a previously documented comparison (126). Therefore, the in vitro fusion

system may yet be sufficient to characterize viral glycoprotein : sialic acid-mediated

membrane fusions, if this interaction is of high affinity. An alternative interpretation of

this apparent limitation to detect sialic acid binding could be simply that sialic acids were

not overexpressed, and hence this in vitro fusion system requires the enrichment of

entry factors onto the HIVpp envelope.

JHM Spike Utilizes Host Sialosides for Viral Attachment

Viruses frequently begin infection by attaching to cellular sialic acids (243). Such

viruses include several CoVs, which attach to sialates via S domain As (99, 100), at low

affinity (113), but relatively high multivalent avidity (125, 126). Atomic resolution structures of 9-O-acetylated sialic acid in complex with the human OC43-CoV domain A

(113) as well as several sialosides in complex with the MERS-CoV domain A (114) have

revealed architectures of CoV-sialate binding sites. Yet even with this detailed

understanding, it is not clear whether sialic acids confer susceptibility to CoV infection

on their own, or whether proteinaceous CoV receptors are also required. In considering

this question, we first focused on the murine MHV JHM-CoV strain. This strain can

infect cells and mice that lack the proteinaceous MHV receptor, mCEACAM1a (129,

143, 188, 196, 197, 200, 224, 225). This strain is also unusually sensitive to proteolytic

activation of spike-mediated membrane fusion (173), and therefore, we hypothesized

that a low-affinity cell binding event, conceivably to cellular sialoglycans, might be

sufficient for subsequent JHM-CoV protease-triggered fusion activation and cell entry.

105

We found that JHM-CoV binding to cells depended on host sialic acids. This binding accounted for initial virus attachment to cells (Figure 17D), which facilitated subsequent virus engagement with proteinaceous CEACAM receptors (Figure 18B and

C). The identification of this interaction was most likely achieved by the VLP system.

Traditionally, soluble RBDs are made to assess their interactions to various cellular factors. This enables any phenotype to be attributed to that specific RBD. Soluble MHV

S domain A proteins have been routinely used in this context as the negative control for sialic acid interations (125). Recently, the S domain A of MERS-CoV and HKU1-CoV have been demonstrated to bind sialic acids, but this binding was only detected when the soluble domains were oligomerized onto nanoparticles at a density comparable to authentic viruses (125, 126), implying that weak sialic acid interactions with CoV spike depends on avidity, not affinity. We therefore deployed a VLP system to identify and characterize the putative weak JHM S : sialoside interactions. We verified that JHM S proteins are assembled on VLPs similar to authentic viruses (Figure 16B). We decided to use VLPs over viruses for another advantage they offer. Traditionally, to characterize viral particle binding to host cells, cells were extensively rinsed to remove unbound particles, which minimizes background signals. We reasoned that these rinses likely also remove weakly-associated viral particles, and end up retaining only strong binders.

By using spike-less (No S) VLPs as the negative control, we theoretically can control for all background binding that is not contributed by the S protein, and therefore minimizes the need for rinsing. These advantages together may have enabled the detection of modest interactions S protein has with target cells. We reason that this VLP system can

106

be widely-adapted to assess previously underappreciated modest, but nonetheless

biologically meaningful, interactions between CoVs and their hosts.

Differential Host Sialoside Utilizations for Viral Attachment and Cell-Cell Fusion

Later in the infection cycle, JHM-CoV spikes facilitated cell fusions and further

virus dissemination in a sialic acid – dependent manner (Figure 19C). MERS-CoV spike

proteins could similarly fuse cells together without requiring prototype protein (hDPP4)

receptors, presumably by utilizing sialoside receptors to engage neighboring cells

(Figure 23C). These findings, summarized in figure 28, suggest that CoV-cell entry and intercellular spread involves the lectin-like activities of spike proteins during both virus- cell attachment and infected-cell expansion into syncytia.

Figure 28. Sialic Acid and Protein Receptor Binding Events During CoV Infection. In entry, CoV S proteins mediate weak interactions with abundant host surface sialates, keeping viruses concentrated on cells yet potentially diffusible across plasma membranes. S proteins subsequently engage protein receptors and are proteolytically activated into membrane fusion-inducing conformations. In spread, canonical virus release is concomitant with cell-cell fusion. Cell-cell fusion involves S binding to sialic acids, and does not require protein receptors, allowing infection to spread beyond the restricted distributions of protein receptors.

107

While sialic acids affected virus-cell binding (Figure 17D), infection by JHM-CoV

(Figure 18C) and cell entry by JHM-CoV VLPs (Figure 18B), on the other hand, required the proteinaceous mCEACAM receptors. This suggests that, prior to protein receptor engagement, CoV binding to target cells may be a physiologically separate step, requiring unique cellular factors. Therefore, one possibility is that multivalent CoV- sialate interactions are not sufficient to hold viruses throughout a virus particle-cell membrane fusion process, and that sialic acids may instead only tether viruses, transiently, onto host-cell surfaces, as they diffuse quasi-two-dimensionally along the target cell plasma membrane. Hence, this locally-elevated viral concentration at or near the plasma membrane may promote the likelihood CoVs can assume more stable interactions with high-affinity protein receptors. Yet, while insufficient for de novo viral entry, sialic acid binding may advance the spike-mediated cell-cell membrane fusions that are observed in several CoV infection processes (187, 189, 191, 196). This is likely due to the intrinsic differences between the fusion entities. In cell-cell fusion, the plasma membrane of the infected cell is juxtaposed against the plasm membrane of neighboring uninfected cells. Therefore, spikes on infected-cell surfaces are steadily directed at uninfected (target) cells in close proximity. In this scenario, there is no functional equivalence to viral attachment, since these cells are not liable to diffuse away as would cell-bound viruses. However, even though close, the spike proteins on the infected cell membrane may not be close enough to be processed by target cell surface proteases. Therefore, even a modest interaction with the target cell, which could elicit a transient spike conformational shift, may be enough to bring spike into the

108 effective range of surface proteases, which triggers membrane fusion. In this way, CoV cell-cell fusion still may require attachment to target cell sialates in order to effect membrane fusion. In support of this contention, we found correlations between sialic acid abundance and membrane fusion activity (Figure 19C and Figure 23C). This was observed for both JHM- and MERS-CoV spikes, and notably, both the JHM- and MERS-

CoV spike-directed cell fusions took place independent of any proteinaceous receptors.

These correlations suggest that sialic acids on uninfected cells can arrange spikes such that they will then operate cooperatively to pull opposing plasma membranes into coalescence.

Host Surface Protease Utilization and CoV Pathogenesis

Traditionally, the determination of proteolysis being an entry requirement for

CoVs have been assessed by adding various exogenous soluble proteases (e.g. trypsin, furin, and cathepsins) during the entry process (Figure 11A and 12, and reviewed in (170)). While effective, this approach has led to inaccurate interpretations on CoV protease preferences, particularly overlooking the importance of type II transmembrane serine proteases (TTSPs). In recent years, utilizing specific small molecule inhibitors and targeted overexpression, these cell surface proteases have been revealed to be preferentially utilized by many CoVs (138, 167, 173, 182, 194,

212). To date, the roles of TTSPs have been well characterized to facilitate facile CoV entry kinetics (138, 161, 174, 182). We have identified a new role for TTSP late in the

CoV viral lifecycle. We observed for the first time that MERS S mediates hDPP4- independent cell-cell fusion, and this remarkable activity specifically requires sufficient

109

level of the surface protease hTMPRSS2 (Figure 22B). Because hTMPRSS2 is highly

expressed on type II pneumocytes and resident alveolar macrophages in the human

lower respiratory tract, the physiological site of MERS-CoV pathogenesis (244), our

observation hence suggests that established MERS-CoV infections can be readily

spread into neighboring cells, skipping de novo viral assembly, even if the neighboring

cells have low hDPP4 expression levels. This facile spread of infection may contribute

to MERS-CoV reaching high enough viral replication to delay the onset of IFN response,

the main cause of MERS-induced pneumonia and death (66). Additionally, this

protease-dependent facile cell-to-cell spread may be an evolutionarily-conserved CoV strategy, since it has been demonstrated that exogenous trypsin is required for successful syncytia-dependent propagations of bat MERS-like CoVs, PDF2180-CoV and HKU5 (245). Interestingly, while SARS-CoV has reduced sensitivity to TTSPs, the closely-related, pandemic-ongoing, 2019-nCoV sequence suggested elevated sensitivity to TTSPs. It would be enlightening to assess whether SARS-CoV-2 has

gained the ability to spread into hTMPRSS2+ hACE2- target cells.

The Evolution of CoV S Protein Domain A

This differential utilization of sialoside and protein receptors raises questions

about CoV evolution. The CoV S domain As fold into a galectin-like structure (104, 105,

108, 109, 112, 113, 116) and several domain As demonstrably bind carbohydrate

ligands (113, 114, 125, 126). The domain As of MHVs, however, appear to be the

exception where, prior to this report, they were considered to bind only protein

(mCEACAM) receptors (104). Because of the putative shift in ligand specificity acquired

110

by MHV domain As, Li et al (105) proposed that ancestral MHV-CoVs bound

carbohydrates, with adaptive evolution generating the present-day CEACAM-binding

sites. This proposal infers the existence, past or present, of evolutionary intermediates

capable of binding both mCEACAM and sialic acid (Figure 4C). Contrary to previous

generalizations, we have observed sialic acid binding activities on MHV-JHM S proteins

(Figure 17C). Using targeted mutagenesis, this sialic acid interaction was attributed to

JHM S protein domain A (Figure 20). Hence, based on our findings, the MHV JHM-CoV

strain meets the criteria for such an intermediate – a virus in which its S domain A

retains two “receptor” binding activities, one for carbohydrate and the other for protein.

Interestingly, the mCEACAM-binding motif on JHM S domain A is close and may be

partially overlapping with its putative sialic acid binding pocket, based on studies of

related, known sialic acid-binding domain As (104, 113, 114). Therefore, conceivably

these two receptor capabilities interact, such that increased affinity for one ligand

reduces affinity for the other. Mutants with relatively high affinity for sialic acid may bind

poorly to mCEACAM, while strains lacking sialic acid binding may demonstrate strong

mCEACAM affinity. An inverse relationship would be consistent with Li et al., who

proposed that acquisition of a CEACAM binding site concomitantly destroyed a sialate

site (105). And by extension, these ligand affinity “trade-offs” would result in CoV

tropism changes. In support of this, while both the liver-tropic MHV-A59 and the neural-

tropic MHV-JHM engage protein mCEACAM receptors, JHM differs in a few positions

among all the residues facilitating mCEACAM binding as determined by the A59 domain

A – mCEACAM co-crystal structure (104).

111

The Co-Evolutionary Relationship between CoV S Protein and HE

Yet perhaps more intriguing relationships are between CoV S and hemagglutinin-

esterase (HE) proteins, a second, smaller viral glycoprotein that is expressed by some

beta-CoVs. While both S and HE bind sialosides (246, 247), HE possesses additional

esterase activity (228, 229). Often, the expression of HE correlates with the ability of

CoV S proteins to bind to host sialic acids. The S proteins of BCoV, OC43-CoV, and

HKU1-CoV all recognize 9-O-acetylated sialic acids (113, 125), and these CoVs also

incorporate HEs that recognize the same sialic acid species (246). Esterase activities

within HE proteins de-acetylate sialosides, and it has been documented, in HKU1-CoV

infections, that HE activity destroys an S-specific sialoside receptor, keeping viruses

detached from cells and allowing for virus dissemination (228), a relationship analogous to IAV HA and NA (248). Several MHV strains including JHM express a viral HE, which binds and deacetylates 4-O-acetyl sialiate (229). Similar to HKU-1-CoV HE, MHV HE

could destroy MHV sialoside receptors and facilitate dissemination. In our study, MHV

strain A59 did not bind sialic acids, which correlated well with it not retaining an intact

HE open reading frame (226, 227). Conversely, we determined that MHV strain JHM did

bind sialic acids and this coincided with its expression of HE. These observations

collectively suggest that the presence of HE correlates with the viral reliance on host

sialic acids, specifically, S : sialic acid – mediated CoV entry processes. For MHVs, their

abilities to utilize sialic acids may be the determinant for their tropisms. Indeed, MHV-

A59 bearing S and HE proteins from JHM has resulted in increased neurotropism and

112

elevated neurovirulence (249). Our identification of the JHM S : sialic acid interaction is in agreement with these views.

While the expression of HE generally implies S : sialic acid interactions, CoV S may develop modest sialic acid interactions in the absence of HE expression. MERS-

CoV genome does not encode HE, but its S protein does bind sialic acids, albeit weakly

(126). While evolutionary selection does impose on MERS S domain A (161–163), and

we were first to observe that this selection resulted in enhanced viral binding (Figure

26C), the increases in MERS S : sialic acid affinity for sialates may be limited by the

absence of a MERS-CoV HE gene. Conversely, for reasons yet to be determined, CoVs

adapted to human airways may not benefit from proficient acetyl-removing activities, as

demonstrated in OC43-CoV and HKU1-CoV, whose HE proteins are going through

progressive losses in their lectin and esterase properties (246).

The Mechanism of the S Protein Domain A Gain-of-Function Mutations

This study revealed S protein mutations that may endow CoVs with expanded

tropisms, beyond that determined by prototype proteinaceous CoV receptors. The JHM-

CoV mutation, G176E, engineered to reflect sialate-utilizing BCoV (Figure 20A),

increased viral S protein binding to cells (Figure 20C), as well as S protein-mediated

membrane fusion (Figure 23C), independent of mCEACAM. The MERS-CoV mutation,

N222D, an adaptation for virus growth in mouse lungs (161), operated remarkably

similar to the JHM-CoV G176E change, with mutant S proteins showing increased

hDPP4-independent cell binding (Figure 26C) and cell fusion (Figure 27A). The

underlying mechanism of this enhanced cell binding remains to be identified. N222D

113

may enhance MERS S binding to its native sialoside moieties, but VLPs bearing the

mutant S did not change in their binding to human lung-derived Calu3 cells, suggesting

the mutation did not affect human sialoside binding. A potential caveat to this

interpretation is that Calu3 cells express hDPP4, hence this condition was not a pure

assessment of protein receptor-independent binding (Figure 26C, Calu3), even though

our JHM results suggest that the protein receptor does not account for most of the viral

binding to target cells (Figure 17D). Conversely, N222D may enhance the association

between MERS S and mouse-specific sialosides, reflecting mouse adaptation. In

support of this, the mutant S significantly elevated VLP binding to cells derived from

both the mouse lung and brain (Figure 26C, LET-1 and DBT). It would be enlightening

to determine whether the mutant has changed the S : sialic acid specificity, as it was documented that minor changes are sufficient to shift sialic acid specificity in HE proteins (229). In general, our findings fit with the hypothesis that viruses with these mutations enhanced S protein binding to certain cellular sialotopes.

However, it is clear that the mutations characterized in this study (JHM G176E and MERS N222D) are not present in structurally-identified sialoside binding sites (113,

114), but rather in the sialic acid-binding distal site. This raises alternative hypotheses to the mutant-induced elevation in cell attachment, including mutation-induced allosteric restructuring of sialoside binding sites. Allosteric remodeling has been demonstrated in

BCoV, OC43-CoV, and porcine hemagglutinating encephalomyelitis virus (PHEV) as the putative role for the residues in this site (125). Alternatively, these mutations may restructure the yet-to-be-determined distinct receptor or coreceptor-binding sites, such

114 as presumed binding sites for orthologous CEACAMs on MHV S proteins (250), or sites for CEACAM5 (251) or GRP78 (252) on MERS-CoV S proteins.

Pathogenesis of Sialoside-utilizing CoVs

Furthermore, this report illuminates understanding of CoV pathogenesis.

Amongst the CoVs, the JHM-CoV strain is known for causing lethal brain infection, even in mice that lack the principal MHV receptor, CEACAM1a (129). JHM spike was identified to be the major contributor to this phenotype (129), and JHM-CoV, but not the related A59-CoV, spread inter-neuronally both in vivo and within in vitro cultures of

CNS-derived cells (198, 199). Notably, neural cell membranes are known for their abundant sialic acid content (131). These findings, combined with evidence that cell-to- cell syncytial spread correlates with pathogenesis in several infection models (129, 188,

200–203), prompts a hypothesis that JHM-CoV sialic acid binding potential accounts for an inter-neuronal syncytial spread that is rapidly lethal. A prediction is that variants of

JHM-CoV exhibiting enhanced sialic acid affinity will have unusually high neurovirulence. Similarly, the MERS-CoV strain causes lethal pneumonia and here it is significant that antibodies specific for the MERS-CoV S domain A both neutralize the virus and reduce infection and pathogenesis in a mouse MERS-CoV model system

(149, 151, 204). Conceivably, these antibodies interfere with sialic acid binding, reducing expansion of MERS-CoV that may take place via cell-cell fusion. Variants of

MERS-CoV with enhanced cell-binding may be useful in assessing the in vivo significance of the findings presented in this report.

115

CoV S Protein Domain A and Zoonosis

Traditionally, studies on CoV cross-species transmissions focus on the role of the primary receptors and their corresponding CoV RBDs. MHV can infect non-murine cells by adapting its S domain A to engage the orthologous CEACAM proteins on these cells

(250). SARS-CoV has been theorized to acquire specific S domain B modifications to efficiently engage human ACE2 after its zoonotic jump from civets to human (152).

MERS-CoV has been widely-accepted to have evolved from an ancestral bat virus that had adapted its S domain B to engage bat DPP4 receptors (52, 253). These compelling observations helped promote the notion that, for CoVs that utilize a primary protein receptor, their utilization of orthologous forms of this receptor in other animal species is the zoonotic barrier. Recent studies on MERS-CoV, however, have started to challenge this dogma. First, the 2015 MERS-CoV outbreak in South Korean has resulted in human-adapted variants whose adaptive mutations actually reduced the affinity between S domain B and hDPP4, debunking the previous belief of a linear association between ortholog receptor affinity and viral transmissibility and fitness (254, 255).

Second, upon artificially adapting MERS-CoV to humanized-mouse, S domain A adaptations on the same residue were identified for two independently-developed systems (161, 162). Concurrently, MERS S domain A was identified to engage specific sialosides for host attachment (126). Together, these observations argue that, in addition to DPP4 orthologs, host sialic acid species are also functional zoonotic barriers, and MERS-CoV may need to overcome both to achieve zoonosis. This understanding helped explained the reason some animals (bats, camels) could be infected by MERS-

116

CoV and other (horses, pigs, rabbits, sheep) not, while all express appropriate DPP4

orthologs, which can elicit susceptibility in vitro (160). In support of this, we observed

that adaptive S domain A changes enhance viral attachment (Figure 26C), directly

highlighting a role host sialosides can serve as a promoter of efficient cross-species

transmissions.

Our study also revealed a grossly underappreciated role for host sialoside

utilization by CoVs. MHV-JHM has a profound zoonotic potential, best characterized as

its ability to cause syncytia of cells from various animal origins that lack mCEACAM. We

identified its underlying mechanism to be S domain A : sialic acid engagement (Figure

19C). Surprisingly, we observed the same mechanism enabling MERS S to induce the

analogous hDPP4-independent syncytia formation (Figure 22 and 23). Together these

observations suggest that host sialoside-induced direct cell-to-cell spread of existing infections may be a general strategy CoVs utilize to boost their zoonotic potential, especially during early stages, when the virus has yet to develop optimal receptor recognitions. In light of the current severe SARS-CoV-2 outbreak in China, we analyzed

its S sequence and found that, comparing to the closely-related pathogen of the 2002

outbreak, SARS-CoV, SARS-CoV-2 is divergent in both the putative sialoside-binding

site and the distal site on their S domain As (54). It is therefore likely that, compared to

SARS-CoV, SARS-CoV-2 has acquired a more potent viral attachment to human lungs,

as well as an elevated efficiency of hACE2-independent cell-to-cell spread. This

prediction is in line with the epidemiological characterization of SARS-CoV-2 having a

higher (i.e. more infectious) R0 value than SARS-CoV. A thorough assessment of the

117

SARS-CoV-2 S domain A functionality would greatly advance our understanding of crucial factors that govern CoV zoonosis.

REFERENCE LIST

1. de Haan CA, Kuo L, Masters PS, Vennema H, Rottier PJ. 1998. Coronavirus particle assembly: primary structure requirements of the membrane protein. J Virol 72:6838–50. 2. Godeke GJ, de Haan CA, Rossen JW, Vennema H, Rottier PJ. 2000. Assembly of spikes into coronavirus particles is mediated by the carboxy-terminal domain of the spike protein. J Virol 74:1566–71. 3. Narayanan K, Maeda a, Maeda J, Makino S. 2000. Characterization of the coronavirus M protein and nucleocapsid interaction in infected cells. J Virol 74:8127–8134. 4. Vennema H, Godeke GJ, Rossen JW, Voorhout WF, Horzinek MC, Opstelten DJ, Rottier PJ. 1996. Nucleocapsid-independent assembly of coronavirus-like particles by co-expression of viral envelope protein genes. EMBO J 15:2020– 2028. 5. Boscarino JA, Logan HL, Lacny JJ, Gallagher TM. 2008. Envelope Protein Palmitoylations Are Crucial for Murine Coronavirus Assembly. J Virol 82:2989– 2999. 6. Perrier A, Bonnin A, Desmarets L, Danneels A, Goffard A, Rouillé Y, Dubuisson J, Belouzard S. 2019. The C-terminal domain of the MERS coronavirus M protein contains a trans-Golgi network localization signal. J Biol Chem 294:14406–14421. 7. Pyrc K, Dijkman R, Deng L, Jebbink MF, Ross HA, Berkhout B, van der Hoek L. 2006. Mosaic structure of human coronavirus NL63, one thousand years of evolution. J Mol Biol 364:964–73. 8. Cavanagh D, Mawditt K, Adzhar A, Gough RE, Picault JP, Naylor CJ, Haydon D, Shaw K, Britton P. 1998. Does IBV change slowly despite the capacity of the spike protein to vary greatly? Adv Exp Med Biol 440:729–34. 9. Lau SKP, Woo PCY, Yip CCY, Tse H, Tsoi H, Cheng VCC, Lee P, Tang BSF, Cheung CHY, Lee RA, So L, Lau Y, Chan K, Yuen K. 2006. Coronavirus HKU1 and other coronavirus infections in Hong Kong. J Clin Microbiol 44:2063–71. 10. Lai MM, Baric RS, Makino S, Keck JG, Egbert J, Leibowitz JL, Stohlman SA. 1985. Recombination between nonsegmented RNA genomes of murine coronaviruses. J Virol 56:449–56. 11. Tian P-F, Jin Y-L, Xing G, Qv L-L, Huang Y-W, Zhou J-Y. 2014. Evidence of recombinant strains of porcine epidemic diarrhea virus, United States, 2013. Emerg Infect Dis 20:1735–8.

118 119

12. Lau SKP, Li KSM, Huang Y, Shek C-T, Tse H, Wang M, Choi GKY, Xu H, Lam CSF, Guo R, Chan K-H, Zheng B-J, Woo PCY, Yuen K-Y. 2010. Ecoepidemiology and complete genome comparison of different strains of severe acute respiratory syndrome-related Rhinolophus bat coronavirus in China reveal bats as a reservoir for acute, self-limiting infection that allows recombination events. J Virol 84:2808– 19. 13. Schalk AF, Hawn MC. 1931. An Apparently New Respiratory Disease of Baby Chicks. J Amer Vet Med Ass 78:413–423. 14. DOYLE LP, HUTCHINGS LM. 1946. A transmissible gastroenteritis in pigs. J Am Vet Med Assoc 108:257–259. 15. CHEEVER FS, DANIELS JB. 1949. A murine virus (JHM) causing disseminated encephalomyelitis with extensive destruction of myelin. J Exp Med 90:181–210. 16. Almeida JD, Tyrrell DA. 1967. The morphology of three previously uncharacterized human respiratory viruses that grow in organ culture. J Gen Virol 1:175–8. 17. van der Hoek L, Pyrc K, Jebbink MF, Vermeulen-Oost W, Berkhout RJM, Wolthers KC, Wertheim-van Dillen PME, Kaandorp J, Spaargaren J, Berkhout B. 2004. Identification of a new human coronavirus. Nat Med 10:368–73. 18. Fouchier RAM, Hartwig NG, Bestebroer TM, Niemeyer B, de Jong JC, Simon JH, Osterhaus ADME. 2004. A previously undescribed coronavirus associated with respiratory disease in humans. Proc Natl Acad Sci U S A 101:6212–6. 19. Woo PCY, Lau SKP, Chu C, Chan K, Tsoi H, Huang Y, Wong BHL, Poon RWS, Cai JJ, Luk W, Poon LLM, Wong SSY, Guan Y, Peiris JSM, Yuen K. 2005. Characterization and complete genome sequence of a novel coronavirus, coronavirus HKU1, from patients with pneumonia. J Virol 79:884–95. 20. Isaacs D, Flowers D, Clarke JR, Valman HB, MacNaughton MR. 1983. Epidemiology of coronavirus respiratory infections. Arch Dis Child 58:500–503. 21. Su S, Wong G, Shi W, Liu J, Lai ACK, Zhou J, Liu W, Bi Y, Gao GF. 2016. Epidemiology, Genetic Recombination, and Pathogenesis of Coronaviruses. Trends Microbiol 24:490–502. 22. Ksiazek TG, Erdman D, Goldsmith CS, Zaki SR, Peret T, Emery S, Tong S, Urbani C, Comer JA, Lim W, Rollin PE, Dowell SF, Ling A-E, Humphrey CD, Shieh W-J, Guarner J, Paddock CD, Rota P, Fields B, DeRisi J, Yang J-Y, Cox N, Hughes JM, LeDuc JW, Bellini WJ, Anderson LJ, SARS Working Group. 2003. A novel coronavirus associated with severe acute respiratory syndrome. N Engl J Med 348:1953–66.

120

23. Peiris JSM, Lai ST, Poon LLM, Guan Y, Yam LYC, Lim W, Nicholls J, Yee WKS, Yan WW, Cheung MT, Cheng VCC, Chan KH, Tsang DNC, Yung RWH, Ng TK, Yuen KY, SARS study group. 2003. Coronavirus as a possible cause of severe acute respiratory syndrome. Lancet (London, England) 361:1319–25. 24. Rota PA, Oberste MS, Monroe SS, Nix WA, Campagnoli R, Icenogle JP, Peñaranda S, Bankamp B, Maher K, Chen M-H, Tong S, Tamin A, Lowe L, Frace M, DeRisi JL, Chen Q, Wang D, Erdman DD, Peret TCT, Burns C, Ksiazek TG, Rollin PE, Sanchez A, Liffick S, Holloway B, Limor J, McCaustland K, Olsen- Rasmussen M, Fouchier R, Günther S, Osterhaus ADME, Drosten C, Pallansch MA, Anderson LJ, Bellini WJ. 2003. Characterization of a novel coronavirus associated with severe acute respiratory syndrome. Science 300:1394–9. 25. Marra MA, Jones SJM, Astell CR, Holt RA, Brooks-Wilson A, Butterfield YSN, Khattra J, Asano JK, Barber SA, Chan SY, Cloutier A, Coughlin SM, Freeman D, Girn N, Griffith OL, Leach SR, Mayo M, McDonald H, Montgomery SB, Pandoh PK, Petrescu AS, Robertson AG, Schein JE, Siddiqui A, Smailus DE, Stott JM, Yang GS, Plummer F, Andonov A, Artsob H, Bastien N, Bernard K, Booth TF, Bowness D, Czub M, Drebot M, Fernando L, Flick R, Garbutt M, Gray M, Grolla A, Jones S, Feldmann H, Meyers A, Kabani A, Li Y, Normand S, Stroher U, Tipples GA, Tyler S, Vogrig R, Ward D, Watson B, Brunham RC, Krajden M, Petric M, Skowronski DM, Upton C, Roper RL. 2003. The Genome sequence of the SARS- associated coronavirus. Science 300:1399–404. 26. Fouchier RAM, van Boheemen S, Osterhaus ADME, Zaki AM, Bestebroer TM. 2012. Isolation of a Novel Coronavirus from a Man with Pneumonia in Saudi Arabia. N Engl J Med 367:1814–1820. 27. Coleman CM, Frieman MB. 2014. Coronaviruses: important emerging human pathogens. J Virol 88:5209–5212. 28. World Health Organization(WHO). 2008. WHO | Update 49 - SARS case fatality ratio , incubation period. Update 1–2. 29. World Health Organization(WHO). 2017. Middle East Respiratory Syndrome Coronavirus (MERS-CoV). WHO. World Health Organization. 30. Lipsitch M, Cohen T, Cooper B, Robins JM, Ma S, James L, Gopalakrishna G, Chew SK, Tan CC, Samore MH, Fisman D, Murray M. 2003. Transmission dynamics and control of severe acute respiratory syndrome. Science 300:1966– 70. 31. Hijawi B, Abdallat M, Sayaydeh A, Alqasrawi S, Haddadin A, Jaarour N, Alsheikh S, Alsanouri T. 2013. Novel coronavirus infections in Jordan, April 2012: epidemiological findings from a retrospective investigation. East Mediterr Health J 19 Suppl 1:S12-8.

121

32. World Health Organization(WHO). 2019. WHO MERS Global Summary and Assessment of Risk. Geneva, Switzerland. 33. World Health Organization(WHO). WHO Statement Regarding Cluster of Pneumonia Cases in Wuhan, China. 34. Woo PCY, Lau SKP, Lam CSF, Lau CCY, Tsang AKL, Lau JHN, Bai R, Teng JLL, Tsang CCC, Wang M, Zheng B-J, Chan K-H, Yuen K-Y. 2012. Discovery of seven novel Mammalian and avian coronaviruses in the genus deltacoronavirus supports bat coronaviruses as the gene source of alphacoronavirus and betacoronavirus and avian coronaviruses as the gene source of gammacoronavirus and deltacoronavi. J Virol 86:3995–4008. 35. Chan JF-W, To KK-W, Tse H, Jin D-Y, Yuen K-Y. 2013. Interspecies transmission and emergence of novel viruses: lessons from bats and birds. Trends Microbiol 21:544–55. 36. Corman VM, Baldwin HJ, Tateno AF, Zerbinati RM, Annan A, Owusu M, Nkrumah EE, Maganga GD, Oppong S, Adu-Sarkodie Y, Vallo P, da Silva Filho LVRF, Leroy EM, Thiel V, van der Hoek L, Poon LLM, Tschapka M, Drosten C, Drexler JF. 2015. Evidence for an Ancestral Association of Human Coronavirus 229E with Bats. J Virol 89:11858–70. 37. Vijgen L, Keyaerts E, Lemey P, Maes P, Van Reeth K, Nauwynck H, Pensaert M, Van Ranst M. 2006. Evolutionary history of the closely related group 2 coronaviruses: porcine hemagglutinating encephalomyelitis virus, bovine coronavirus, and human coronavirus OC43. J Virol 80:7270–4. 38. Vijgen L, Keyaerts E, Moës E, Thoelen I, Wollants E, Lemey P, Vandamme A-M, Van Ranst M. 2005. Complete genomic sequence of human coronavirus OC43: molecular clock analysis suggests a relatively recent zoonotic coronavirus transmission event. J Virol 79:1595–604. 39. Guan Y, Zheng BJ, He YQ, Liu XL, Zhuang ZX, Cheung CL, Luo SW, Li PH, Zhang LJ, Guan YJ, Butt KM, Wong KL, Chan KW, Lim W, Shortridge KF, Yuen KY, Peiris JSM, Poon LLM. 2003. Isolation and characterization of viruses related to the SARS coronavirus from animals in Southern China. Science (80- ) 302:276–278. 40. Poon LLM, Chu DKW, Chan KH, Wong OK, Ellis TM, Leung YHC, Lau SKP, Woo PCY, Suen KY, Yuen KY, Guan Y, Peiris JSM. 2005. Identification of a novel coronavirus in bats. J Virol 79:2001–9. 41. Perlman S, Netland J. 2009. Coronaviruses post-SARS: Update on replication and pathogenesis. Nat Rev Microbiol 7:439–450.

122

42. Ge XY, Li JL, Yang X Lou, Chmura AA, Zhu G, Epstein JH, Mazet JK, Hu B, Zhang W, Peng C, Zhang YJ, Luo CM, Tan B, Wang N, Zhu Y, Crameri G, Zhang SY, Wang LF, Daszak P, Shi ZL. 2013. Isolation and characterization of a bat SARS-like coronavirus that uses the ACE2 receptor. Nature 503:535–538. 43. Wu Z, Yang L, Ren X, Zhang J, Yang F, Zhang S, Jin Q. 2016. ORF8-related genetic evidence for Chinese horseshoe bats as the source of human severe acute respiratory syndrome coronavirus. J Infect Dis 213:579–583. 44. Reusken CBEM, Haagmans BL, Müller MA, Gutierrez C, Godeke G-J, Meyer B, Muth D, Raj VS, Smits-De Vries L, Corman VM, Drexler J-F, Smits SL, El Tahir YE, De Sousa R, van Beek J, Nowotny N, van Maanen K, Hidalgo-Hermoso E, Bosch B-J, Rottier P, Osterhaus A, Gortázar-Schmidt C, Drosten C, Koopmans MPG. 2013. Middle East respiratory syndrome coronavirus neutralising serum antibodies in dromedary camels: a comparative serological study. Lancet Infect Dis 13:859–66. 45. Chan JFW, Lau SKP, To KKW, Cheng VCC, Woo PCY, Yuen K-Y. 2015. Middle East respiratory syndrome coronavirus: another zoonotic betacoronavirus causing SARS-like disease. Clin Microbiol Rev 28:465–522. 46. Hemida MG, Chu DKW, Poon LLM, Perera RAPM, Alhammadi MA, Ng H-Y, Siu LY, Guan Y, Alnaeem A, Peiris M. 2014. MERS coronavirus in dromedary camel herd, Saudi Arabia. Emerg Infect Dis 20:1231–4. 47. Raj VS, Farag EABA, Reusken CBEM, Lamers MM, Pas SD, Voermans J, Smits SL, Osterhaus ADME, Al-Mawlawi N, Al-Romaihi HE, AlHajri MM, El-Sayed AM, Mohran KA, Ghobashy H, Alhajri F, Al-Thani M, Al-Marri SA, El-Maghraby MM, Koopmans MPG, Haagmans BL. 2014. Isolation of MERS coronavirus from a dromedary camel, Qatar, 2014. Emerg Infect Dis 20:1339–42. 48. Chu DKW, Poon LLM, Gomaa MM, Shehata MM, Perera RAPM, Abu Zeid D, El Rifay AS, Siu LY, Guan Y, Webby RJ, Ali MA, Peiris M, Kayali G. 2014. MERS coronaviruses in dromedary camels, Egypt. Emerg Infect Dis 20:1049–53. 49. Sabir JSM, Lam TTY, Ahmed MMM, Li L, Shen Y, Abo-Aba SEM, Qureshi MI, Abu-Zeid M, Zhang Y, Khiyami MA, Alharbi NS, Hajrah NH, Sabir MJ, Mutwakil MHZ, Kabli SA, Alsulaimany FAS, Obaid AY, Zhou B, Smith DK, Holmes EC, Zhu H, Guan Y. 2016. Co-circulation of three camel coronavirus species and recombination of MERS-CoVs in Saudi Arabia. Science (80- ) 351:81–84. 50. Memish ZA, Mishra N, Olival KJ, Fagbo SF, Kapoor V, Epstein JH, Alhakeem R, Durosinloun A, Al Asmari M, Islam A, Kapoor A, Briese T, Daszak P, Al Rabeeah AA, Lipkin WI. 2013. Middle East respiratory syndrome coronavirus in bats, Saudi Arabia. Emerg Infect Dis 19:1819–23.

123

51. Ithete NL, Stoffberg S, Corman VM, Cottontail VM, Richards LR, Schoeman MC, Drosten C, Drexler JF, Preiser W. 2013. Close relative of human Middle East respiratory syndrome coronavirus in bat, South Africa. Emerg Infect Dis 19:1697– 9. 52. Wang Q, Qi J, Yuan Y, Xuan Y, Han P, Wan Y, Ji W, Li Y, Wu Y, Wang J, Iwamoto A, Woo PCY, Yuen KY, Yan J, Lu G, Gao GF. 2014. Bat origins of MERS-CoV supported by bat Coronavirus HKU4 usage of human receptor CD26. Cell Host Microbe 16:328–337. 53. Lau SKP, Li KSM, Tsang AKL, Lam CSF, Ahmed S, Chen H, Chan K-H, Woo PCY, Yuen K-Y. 2013. Genetic Characterization of Betacoronavirus Lineage C Viruses in Bats Reveals Marked Sequence Divergence in the Spike Protein of Pipistrellus Bat Coronavirus HKU5 in Japanese Pipistrelle: Implications for the Origin of the Novel Middle East Respiratory Sy. J Virol 87:8638–8650. 54. Zhang, Y.-Z., Wu, F., Chen, Y.-M., Pei, Y.-Y., Xu, L., Wang, W., Zhao S, Yu, B., Hu, Y., Tao, Z.-W., Song, Z.-G., Tian, J.-H., Zhang, Y.-L. Liu, Y., Zheng, J.-J., Dai, F.-H., Wang, Q.-M., She, J.-L. and Zhu T-Y. Wuhan seafood market pneumonia virus isolate Wuhan-Hu-1, complete geno - Nucleotide - NCBI. 55. Anthony SJ, Epstein JH, Murray KA, Navarrete-Macias I, Zambrana-Torrelio CM, Solovyov A, Ojeda-Flores R, Arrigo NC, Islam A, Ali Khan S, Hosseini P, Bogich TL, Olival KJ, Sanchez-Leon MD, Karesh WB, Goldstein T, Luby SP, Morse SS, Mazet JAK, Daszak P, Lipkin WI. 2013. A Strategy To Estimate Unknown Viral Diversity in Mammals. MBio 4:e00598-13. 56. Teeling EC, Springer MS, Madsen O, Bates P, O’brien SJ, Murphy WJ. 2005. A molecular phylogeny for bats illuminates biogeography and the fossil record. Science 307:580–4. 57. Menachery VD, Graham RL, Baric RS. 2017. Jumping species—a mechanism for coronavirus persistence and survival. Curr Opin Virol 23:1–7. 58. Zhang G, Cowled C, Shi Z, Huang Z, Bishop-Lilly K a, Fang X, Wynne JW, Xiong Z, Baker ML, Zhao W, Tachedjian M, Zhu Y, Zhou P, Jiang X, Ng J, Yang L, Wu L, Xiao J, Feng Y, Chen Y, Sun X, Zhang Y, Marsh G a, Crameri G, Broder CC, Frey KG, Wang L-F, Wang J. 2013. Comparative analysis of bat genomes provides insight into the evolution of flight and immunity. Science 339:456–60. 59. Reshi ML, Su Y-C, Hong J-R. 2014. RNA Viruses: ROS-Mediated Cell Death. Int J Cell Biol 2014:467452. 60. Ahn M, Cui J, Irving AT, Wang L-F. 2016. Unique Loss of the PYHIN Gene Family in Bats Amongst Mammals: Implications for Inflammasome Sensing. Sci Rep 6:21722.

124

61. Xie J, Li Y, Shen X, Goh G, Zhu Y, Cui J, Wang L-F, Shi Z-L, Zhou P. 2018. Dampened STING-Dependent Interferon Activation in Bats. Cell Host Microbe 23:297-301.e4. 62. Zhou P, Tachedjian M, Wynne JW, Boyd V, Cui J, Smith I, Cowled C, Ng JHJ, Mok L, Michalski WP, Mendenhall IH, Tachedjian G, Wang L-F, Baker ML. 2016. Contraction of the type I IFN locus and unusual constitutive expression of IFN-α in bats. Proc Natl Acad Sci U S A 113:2696–701. 63. Seronello S, Montanez J, Presleigh K, Barlow M, Park SB, Choi J. 2011. Ethanol and reactive species increase basal sequence heterogeneity of hepatitis C virus and produce variants with reduced susceptibility to antivirals. PLoS One 6:e27436. 64. Zhao J, Zhao J, Van Rooijen N, Perlman S. 2009. Evasion by stealth: inefficient immune activation underlies poor T cell response and severe disease in SARS- CoV-infected mice. PLoS Pathog 5:e1000636. 65. de Wit E, van Doremalen N, Falzarano D, Munster VJ. 2016. SARS and MERS: recent insights into emerging coronaviruses. Nat Rev Microbiol 14:523–34. 66. Channappanavar R, Fehr AR, Zheng J, Wohlford-Lenane C, Abrahante JE, Mack M, Sompallae R, McCray PB, Meyerholz DK, Perlman S. 2019. IFN-I response timing relative to virus replication determines MERS coronavirus infection outcomes. J Clin Invest 130:3625–3639. 67. Chernomordik L V., Kozlov MM. 2008. Mechanics of membrane fusion. Nat Struct Mol Biol. 68. Efrat A, Chernomordik L V., Kozlov MM. 2007. Point-like protrusion as a prestalk intermediate in membrane fusion pathway. Biophys J 92:L61-3. 69. Yang L, Huang HW. 2002. Observation of a membrane fusion intermediate structure. Science 297:1877–9. 70. Zhao W-D, Hamid E, Shin W, Wen PJ, Krystofiak ES, Villarreal SA, Chiang H-C, Kachar B, Wu L-G. 2016. Hemi-fused structure mediates and controls fusion and fission in live cells. Nature 534:548–52. 71. Harrison SC. 2015. Viral membrane fusion. Virology 479–480:498–507. 72. Mi S, Lee X, Li X, Veldman GM, Finnerty H, Racie L, LaVallie E, Tang XY, Edouard P, Howes S, Keith JC, McCoy JM. 2000. Syncytin is a captive retroviral envelope protein involved in human placental morphogenesis. Nature 403:785–9. 73. Fédry J, Liu Y, Péhau-Arnaudet G, Pei J, Li W, Tortorici MA, Traincard F, Meola A, Bricogne G, Grishin N V., Snell WJ, Rey FA, Krey T. 2017. The Ancient Gamete Fusogen HAP2 Is a Eukaryotic Class II Fusion Protein. Cell 168:904- 915.e10.

125

74. Ivanovic T, Choi JL, Whelan SP, van Oijen AM, Harrison SC. 2013. Influenza- virus membrane fusion by cooperative fold-back of stochastically induced hemagglutinin intermediates. Elife 2:e00333. 75. Plemper RK, Brindley MA, Iorio RM. 2011. Structural and mechanistic studies of measles virus illuminate paramyxovirus entry. PLoS Pathog 7:e1002058. 76. Kuzmin PI, Zimmerberg J, Chizmadzhev YA, Cohen FS. 2001. A quantitative model for membrane fusion based on low-energy intermediates. Proc Natl Acad Sci U S A 98:7235–40. 77. Chakraborty H, Tarafdar PK, Lentz BR. 2013. A novel assay for detecting fusion pore formation: implications for the fusion mechanism. Biochemistry 52:8510–7. 78. Bullough PA, Hughson FM, Skehel JJ, Wiley DC. 1994. Structure of influenza haemagglutinin at the pH of membrane fusion. Nature 371:37–43. 79. Chao LH, Klein DE, Schmidt AG, Peña JM, Harrison SC. 2014. Sequential conformational rearrangements in flavivirus membrane fusion. Elife 3:e04389. 80. Kim IS, Jenni S, Stanifer ML, Roth E, Whelan SPJ, van Oijen AM, Harrison SC. 2017. Mechanism of membrane fusion induced by vesicular stomatitis virus G protein. Proc Natl Acad Sci U S A 114:E28–E36. 81. Wang H, Barnes CO, Yang Z, Nussenzweig MC, Bjorkman PJ. 2018. Partially Open HIV-1 Envelope Structures Exhibit Conformational Changes Relevant for Coreceptor Binding and Fusion. Cell Host Microbe 24:579-592.e4. 82. Lasky LA, Nakamura G, Smith DH, Fennie C, Shimasaki C, Patzer E, Berman P, Gregory T, Capon DJ. 1987. Delineation of a region of the human immunodeficiency virus type 1 gp120 glycoprotein critical for interaction with the CD4 receptor. Cell 50:975–85. 83. Shaik MM, Peng H, Lu J, Rits-Volloch S, Xu C, Liao M, Chen B. 2019. Structural basis of coreceptor recognition by HIV-1 envelope spike. Nature 565:318–323. 84. Iliopoulou M, Nolan R, Alvarez L, Watanabe Y, Coomer CA, Jakobsdottir GM, Bowden TA, Padilla-Parra S. 2018. A dynamic three-step mechanism drives the HIV-1 pre-fusion reaction. Nat Struct Mol Biol 25:814–822. 85. Chandran K, Sullivan NJ, Felbor U, Whelan SP, Cunningham JM. 2005. Endosomal proteolysis of the Ebola virus glycoprotein is necessary for infection. Science 308:1643–5. 86. Schornberg K, Matsuyama S, Kabsch K, Delos S, Bouton A, White J. 2006. Role of endosomal cathepsins in entry mediated by the Ebola virus glycoprotein. J Virol 80:4174–8.

126

87. Carette JE, Raaben M, Wong AC, Herbert AS, Obernosterer G, Mulherkar N, Kuehne AI, Kranzusch PJ, Griffin AM, Ruthel G, Dal Cin P, Dye JM, Whelan SP, Chandran K, Brummelkamp TR. 2011. Ebola virus entry requires the cholesterol transporter Niemann-Pick C1. Nature 477:340–3. 88. Côté M, Misasi J, Ren T, Bruchez A, Lee K, Filone CM, Hensley L, Li Q, Ory D, Chandran K, Cunningham J. 2011. Small molecule inhibitors reveal Niemann-Pick C1 is essential for Ebola virus infection. Nature 477:344–8. 89. Spence JS, Krause TB, Mittler E, Jangra RK, Chandran K. 2016. Direct Visualization of Ebola Virus Fusion Triggering in the Endocytic Pathway. MBio 7:e01857-15. 90. Simmons JA, D’Souza RS, Ruas M, Galione A, Casanova JE, White JM. 2016. Ebolavirus Glycoprotein Directs Fusion through NPC1+ Endolysosomes. J Virol 90:605–10. 91. Lin YP, Xiong X, Wharton SA, Martin SR, Coombs PJ, Vachieri SG, Christodoulou E, Walker PA, Liu J, Skehel JJ, Gamblin SJ, Hay AJ, Daniels RS, McCauley JW. 2012. Evolution of the receptor binding properties of the influenza A(H3N2) hemagglutinin. Proc Natl Acad Sci U S A 109:21474–9. 92. Zhang X, Ge P, Yu X, Brannan JM, Bi G, Zhang Q, Schein S, Zhou ZH. 2013. Cryo-EM structure of the mature dengue virus at 3.5-Å resolution. Nat Struct Mol Biol 20:105–10. 93. Roche S, Rey FA, Gaudin Y, Bressanelli S. 2007. Structure of the prefusion form of the vesicular stomatitis virus glycoprotein G. Science 315:843–8. 94. Chen J, Skehel JJ, Wiley DC. 1999. N- and C-terminal residues combine in the fusion-pH influenza hemagglutinin HA(2) subunit to form an N cap that terminates the triple-stranded coiled coil. Proc Natl Acad Sci U S A 96:8967–72. 95. Modis Y, Ogata S, Clements D, Harrison SC. 2004. Structure of the dengue virus envelope protein after membrane fusion. Nature 427:313–9. 96. Roche S, Bressanelli S, Rey FA, Gaudin Y. 2006. Crystal structure of the low-pH form of the vesicular stomatitis virus glycoprotein G. Science 313:187–91. 97. Lee KK. 2010. Architecture of a nascent viral fusion pore. EMBO J 29:1299–311. 98. Racaniello VR, Skalka AM, Flint J, Rall GF. 2015. Principles of Virology, Bundle. American Society of Microbiology. 99. Li F. 2016. Structure, Function, and Evolution of Coronavirus Spike Proteins. Annu Rev Virol 3:237–261. 100. Hulswit RJG, de Haan CAM, Bosch B-J. 2016. Coronavirus Spike Protein and Tropism Changes. Adv Virus Res 96:29–57.

127

101. Bosch BJ, van der Zee R, de Haan CAM, Rottier PJM. 2003. The coronavirus spike protein is a class I virus fusion protein: structural and functional characterization of the fusion core complex. J Virol 77:8801–11. 102. Davies HA, Macnaughton MR. 1979. Comparison of the morphology of three coronaviruses. Arch Virol 59:25–33. 103. Li F, Li W, Farzan M, Harrison SC. 2005. Structure of SARS coronavirus spike receptor-binding domain complexed with receptor. Science 309:1864–8. 104. Peng G, Sun D, Rajashankar KR, Qian Z, Holmes K V, Li F. 2011. Crystal structure of mouse coronavirus receptor-binding domain complexed with its murine receptor. Proc Natl Acad Sci U S A 108:10696–701. 105. Peng G, Xu L, Lin Y, Chen L, Pasquarella JR, Holmes K V, Li F. 2012. Crystal structure of bovine coronavirus spike protein lectin domain. J Biol Chem 287:41931–8. 106. Wang N, Shi X, Jiang L, Zhang S, Wang D, Tong P, Guo D, Fu L, Cui Y, Liu X, Arledge KC, Chen Y, Zhang L, Wang X. 2013. Structure of MERS-CoV spike receptor-binding domain complexed with human receptor DPP4. Cell Res 23:986–993. 107. Wu K, Li W, Peng G, Li F. 2009. Crystal structure of NL63 respiratory coronavirus receptor-binding domain complexed with its human receptor. Proc Natl Acad Sci U S A 106:19970–4. 108. Kirchdoerfer RN, Cottrell CA, Wang N, Pallesen J, Yassine HM, Turner HL, Corbett KS, Graham BS, McLellan JS, Ward AB. 2016. Pre-fusion structure of a human coronavirus spike protein. Nature 531:118–21. 109. Walls AC, Tortorici MA, Frenz B, Snijder J, Li W, Rey FA, DiMaio F, Bosch B, Veesler D. 2016. Glycan shield and epitope masking of a coronavirus spike protein observed by cryo-electron microscopy. Nat Struct Mol Biol 23:899–905. 110. Walls AC, Tortorici MA, Bosch B, Frenz B, Rottier PJM, DiMaio F, Rey FA, Veesler D. 2016. Cryo-electron microscopy structure of a coronavirus spike glycoprotein trimer. Nature 531:114–117. 111. Gui M, Song W, Zhou H, Xu J, Chen S, Xiang Y, Wang X. 2017. Cryo-electron microscopy structures of the SARS-CoV spike glycoprotein reveal a prerequisite conformational state for receptor binding. Cell Res 119–129. 112. Yuan Y, Cao D, Zhang Y, Qi J, Wang Q, Lu G, Wu Y, Yan J, Shi Y, Zhang X, Gao GF. 2017. Cryo-EM structures of MERS-CoV and SARS-CoV spike glycoproteins reveal the dynamic receptor binding domains. Nat Commun. 113. Alejandra Tortorici M, Walls AC, Lang Y, Wang C, Li Z, Koerhuis D, Boons G-J, Bosch B-J, Rey FA, de Groot RJ, Veesler D. 2019. Structural basis for human coronavirus attachment to sialic acid receptors. Nat Struct Mol Biol 26:481–489.

128

114. Park Y, Walls AC, Wang Z, Sauer MM, Li W, Tortorici MA, Bosch B, DiMaio F, Veesler D. 2019. Structures of MERS-CoV spike glycoprotein in complex with sialoside attachment receptors. Nat Struct Mol Biol 26:1151–1157. 115. Shang J, Zheng Y, Yang Y, Liu C, Geng Q, Luo C, Zhang W, Li F. 2018. Cryo-EM structure of infectious bronchitis coronavirus spike protein reveals structural and functional evolution of coronavirus spike proteins. PLoS Pathog 14:e1007009. 116. Xiong X, Tortorici MA, Snijder J, Yoshioka C, Walls AC, Li W, McGuire AT, Rey FA, Bosch B-J, Veesler D. 2017. Glycan shield and fusion activation of a deltacoronavirus spike glycoprotein fine-tuned for enteric infections. J Virol JVI.01628-17. 117. Wrapp D, McLellan JS. 2019. The 3.1-Angstrom Cryo-electron Microscopy Structure of the Porcine Epidemic Diarrhea Virus Spike Protein in the Prefusion Conformation. J Virol 93:1–13. 118. Li F. 2015. Receptor recognition mechanisms of coronaviruses: a decade of structural studies. J Virol 89:1954–64. 119. Promkuntod N, van Eijndhoven REW, de Vrieze G, Gröne A, Verheije MH. 2014. Mapping of the receptor-binding domain and amino acids critical for attachment in the spike protein of avian coronavirus infectious bronchitis virus. Virology 448:26– 32. 120. Li W, Hulswit RJG, Kenney SP, Widjaja I, Jung K, Alhamo MA, van Dieren B, van Kuppeveld FJM, Saif LJ, Bosch B-J. 2018. Broad receptor engagement of an emerging global coronavirus may potentiate its diverse cross-species transmissibility. Proc Natl Acad Sci 115:E5135–E5143. 121. Seetharaman J, Kfanigsberg A, Slaaby R, Leffler H, Barondes SH, Rini JM. 1998. X-ray crystal structure of the human galectin-3 carbohydrate recognition domain at 2.1-Å resolution. J Biol Chem 273:13047–13052. 122. Schultze B, Cavanagh D, Herrler G. 1992. Neuraminidase treatment of avian infectious bronchitis coronavirus reveals a hemagglutinating activity that is dependent on sialic acid-containing receptors on erythrocytes. Virology 189:792– 4. 123. Wickramasinghe INA, de Vries RP, Gröne A, de Haan CAM, Verheije MH. 2011. Binding of avian coronavirus spike proteins to host factors reflects virus tropism and pathogenicity. J Virol 85:8903–12. 124. Langereis MA, van Vliet ALW, Boot W, de Groot RJ. 2010. Attachment of mouse hepatitis virus to O-acetylated sialic acid is mediated by hemagglutinin-esterase and not by the spike protein. J Virol 84:8970–4.

129

125. Hulswit RJG, Lang Y, Bakkers MJG, Li W, Li Z, Schouten A, Ophorst B, van Kuppeveld FJM, Boons G-J, Bosch B-J, Huizinga EG, de Groot RJ. 2019. Human coronaviruses OC43 and HKU1 bind to 9- O -acetylated sialic acids via a conserved receptor-binding site in spike protein domain A. Proc Natl Acad Sci 201809667. 126. Li W, Hulswit RJG, Widjaja I, Raj VS, McBride R, Peng W, Widagdo W, Tortorici MA, van Dieren B, Lang Y, van Lent JWM, Paulson JC, de Haan CAM, de Groot RJ, van Kuppeveld FJM, Haagmans BL, Bosch B-J. 2017. Identification of sialic acid-binding function for the Middle East respiratory syndrome coronavirus spike glycoprotein. Proc Natl Acad Sci 201712592. 127. Navas S, Weiss SR. 2003. Murine coronavirus-induced hepatitis: JHM genetic background eliminates A59 spike-determined hepatotropism. J Virol 77:4972–8. 128. Navas-Martin S, Brom M, Chua M-M, Watson R, Qiu Z, Weiss SR. 2007. Replicase genes of murine coronavirus strains A59 and JHM are interchangeable: differences in pathogenesis map to the 3’ one-third of the genome. J Virol 81:1022–6. 129. Miura TA, Travanty EA, Oko L, Bielefeldt-Ohmann H, Weiss SR, Beauchemin N, Holmes K V. 2008. The Spike Glycoprotein of Murine Coronavirus MHV-JHM Mediates Receptor-Independent Infection and Spread in the Central Nervous Systems of Ceacam1a / Mice. J Virol 82:755–763. 130. Godfraind C, Langreth SG, Cardellichio CB, Knobler R, Coutelier JP, Dubois- Dalcq M, Holmes K V. 1995. Tissue and cellular distribution of an adhesion molecule in the carcinoembryonic antigen family that serves as a receptor for mouse hepatitis virus. Lab Invest 73:615–27. 131. Schnaar RL, Gerardy-Schahn R, Hildebrandt H. 2014. Sialic acids in the brain: gangliosides and polysialic acid in nervous system development, stability, disease, and regeneration. Physiol Rev 94:461–518. 132. Yeager CL, Ashmun RA, Williams RK, Cardellichio CB, Shapiro LH, Look AT, Holmes K V. 1992. Human aminopeptidase N is a receptor for human coronavirus 229E. Nature 357:420–422. 133. Li Z, Tomlinson ACA, Wong AHM, Zhou D, Desforges M, Talbot PJ, Benlekbir S, Rubinstein JL, Rini JM. 2019. The human coronavirus HCoV-229E S-protein structure and receptor binding. Elife 8:1–22. 134. Reguera J, Santiago C, Mudgal G, Ordoño D, Enjuanes L, Casasnovas JM. 2012. Structural bases of coronavirus attachment to host aminopeptidase N and its inhibition by neutralizing antibodies. PLoS Pathog 8:e1002859. 135. Tresnan DB, Levis R, Holmes K V. 1996. Feline aminopeptidase N serves as a receptor for feline, canine, porcine, and human coronaviruses in serogroup I. J Virol 70:8669–74.

130

136. Letko M, Munster V. 2020. Functional assessment of cell entry and receptor usage for lineage B β-coronaviruses, including 2019-nCoV. bioRxiv 2020.01.22.915660. 137. Raj VS, Mou H, Smits SL, Dekkers DHW, Müller MA, Dijkman R, Muth D, Demmers JAA, Zaki A, Fouchier RAM, Thiel V, Drosten C, Rottier PJM, Osterhaus ADME, Bosch BJ, Haagmans BL. 2013. Dipeptidyl peptidase 4 is a functional receptor for the emerging human coronavirus-EMC. Nature 495:251– 254. 138. Earnest JT, Hantak MP, Li K, McCray PB, Perlman S, Gallagher T. 2017. The tetraspanin CD9 facilitates MERS-coronavirus entry by scaffolding host cell receptors and proteases. PLOS Pathog 13:e1006546. 139. Munro JB, Gorman J, Ma X, Zhou Z, Arthos J, Burton DR, Koff WC, Courter JR, Smith AB, Kwong PD, Blanchard SC, Mothes W. 2014. Conformational dynamics of single HIV-1 envelope trimers on the surface of native virions. Science (80- ) 346:759–763. 140. Durham ND, Howard AR, Govindan R, Senjobe F, Fels JM, Diehl WE, Luban J, Chandran K, Munro JB. 2020. Real-Time Analysis of Individual Ebola Virus Glycoproteins Reveals Pre-Fusion, Entry-Relevant Conformational Dynamics. Viruses 12:1–15. 141. Walls AC, Xiong X, Park Y-J, Tortorici MA, Snijder J, Quispe J, Cameroni E, Gopal R, Dai M, Lanzavecchia A, Zambon M, Rey FA, Corti D, Veesler D. 2019. Unexpected Receptor Functional Mimicry Elucidates Activation of Coronavirus Fusion. Cell 176:1026-1039.e15. 142. Gallagher TM, Buchmeier MJ. 1990. Monoclonal Antibody-Selected Variants of MHV-4 Contain Substitutions and Deletions in the E2 Spike Glycoprotein, p. 385– 393. In Advances in experimental medicine and biology. 143. Krueger DK, Kelly SM, Lewicki DN, Ruffolo R, Gallagher TM. 2001. Variations in Disparate Regions of the Murine Coronavirus Spike Protein Impact the Initiation of Membrane Fusion. J Virol 75:2792–2802. 144. Tsai CW, Chang SC, Chang MF. 1999. A 12-amino acid stretch in the hypervariable region of the spike protein S1 subunit is critical for cell fusion activity of mouse hepatitis virus. J Biol Chem 274:26085–26090. 145. Parker SE, Gallagher TM, Buchmeier MJ. 1989. Sequence analysis reveals extensive polymorphism and evidence of deletions within the E2 glycoprotein gene of several strains of murine hepatitis virus. Virology 173:664–73. 146. Gallagher TM. 1997. A role for naturally occurring variation of the murine coronavirus spike protein in stabilizing association with the cellular receptor. J Virol 71:3129–37.

131

147. Gallagher TM, Parker SE, Buchmeier MJ. 1990. Neutralization-resistant variants of a neurotropic coronavirus are generated by deletions within the amino-terminal half of the spike glycoprotein. J Virol 64:731–41. 148. Ontiveros E, Kim TS, Gallagher TM, Perlman S. 2003. Enhanced Virulence Mediated by the Murine Coronavirus, Mouse Hepatitis Virus Strain JHM, Is Associated with a Glycine at Residue 310 of the Spike Glycoprotein. J Virol 77:10260–10269. 149. Chen Y, Lu S, Jia H, Deng Y, Zhou J, Huang B, Yu Y, Lan J, Wang W, Lou Y, Qin K, Tan W. 2017. A novel neutralizing monoclonal antibody targeting the N- terminal domain of the MERS-CoV spike protein. Emerg Microbes Infect 6. 150. Wang N, Rosen O, Wang L, Turner HL, Stevens LJ, Corbett KS, Bowman CA, Pallesen J, Shi W, Zhang Y, Leung K, Kirchdoerfer RN, Becker MM, Denison MR, Chappell JD, Ward AB, Graham BS, McLellan JS. 2019. Structural Definition of a Neutralization-Sensitive Epitope on the MERS-CoV S1-NTD. Cell Rep 28:3395- 3405.e6. 151. Widjaja I, Wang C, van Haperen R, Gutiérrez-Álvarez J, van Dieren B, Okba NMA, Raj VS, Li W, Fernandez-Delgado R, Grosveld F, van Kuppeveld FJM, Haagmans BL, Enjuanes L, Drabek D, Bosch B-J. 2019. Towards a solution to MERS: protective human monoclonal antibodies targeting different domains and functions of the MERS-coronavirus spike glycoprotein. Emerg Microbes Infect 8:516–530. 152. Li F. 2013. Receptor recognition and cross-species infections of SARS coronavirus. Antiviral Res 100:246–54. 153. Brandão PE. 2018. Could human coronavirus OC43 have co-evolved with early humans? Genet Mol Biol 41:692–698. 154. Barlan A, Zhao J, Sarkar MK, Li K, McCray PB, Perlman S, Gallagher T. 2014. Receptor Variation and Susceptibility to Middle East Respiratory Syndrome Coronavirus Infection. J Virol 88:4953–4961. 155. van Doremalen N, Miazgowicz KL, Milne-Price S, Bushmaker T, Robertson S, Scott D, Kinne J, McLellan JS, Zhu J, Munster VJ. 2014. Host species restriction of Middle East respiratory syndrome coronavirus through its receptor, dipeptidyl peptidase 4. J Virol 88:9220–32. 156. Meyer B, García-Bocanegra I, Wernery U, Wernery R, Sieberg A, Müller MA, Drexler JF, Drosten C, Eckerle I. 2015. Serologic assessment of possibility for MERS-CoV infection in equids. Emerg Infect Dis 21:181–2. 157. Hemida MG, Chu DKW, Perera RAPM, Ko RLW, So RTY, Ng BCY, Chan SMS, Chu S, Alnaeem AA, Alhammadi MA, Webby RJ, Poon LLM, Balasuriya UBR, Peiris M. 2017. Coronavirus infections in horses in Saudi Arabia and Oman. Transbound Emerg Dis 64:2093–2103.

132

158. Ali M, El-Shesheny R, Kandeil A, Shehata M, Elsokary B, Gomaa M, Hassan N, El Sayed A, El-Taweel A, Sobhy H, Fasina FO, Dauphin G, El Masry I, Wolde AW, Daszak P, Miller M, VonDobschuetz S, Morzaria S, Lubroth J, Makonnen YJ. 2017. Cross-sectional surveillance of Middle East respiratory syndrome coronavirus (MERS-CoV) in dromedary camels and other mammals in Egypt, August 2015 to January 2016. Euro Surveill 22:30487. 159. Widagdo W, Okba NMA, Li W, de Jong A, de Swart RL, Begeman L, van den Brand JMA, Bosch B-J, Haagmans BL. 2019. Species-Specific Colocalization of Middle East Respiratory Syndrome Coronavirus Attachment and Entry Receptors. J Virol 93:1–12. 160. Widagdo W, Sooksawasdi Na Ayudhya S, Hundie GB, Haagmans BL. 2019. Host Determinants of MERS-CoV Transmission and Pathogenesis. Viruses 11. 161. Li K, Wohlford-Lenane CL, Channappanavar R, Park J, Earnest JT, Bair TB, Bates AM, Brogden KA, Flaherty HA, Gallagher T, Meyerholz DK, Perlman S, McCray PB. 2017. Mouse-adapted MERS coronavirus causes lethal lung disease in human DPP4 knockin mice. Proc Natl Acad Sci 114:E3119–E3128. 162. Douglas MG, Kocher JF, Scobey T, Baric RS, Cockrell AS. 2017. Adaptive evolution influences the infectious dose of MERS-CoV necessary to achieve severe respiratory disease. Virology 0–1. 163. Sohrab SS, Azhar EI. 2019. Genetic diversity of MERS-CoV spike protein gene in Saudi Arabia. J Infect Public Health. 164. Finkelshtein D, Werman A, Novick D, Barak S, Rubinstein M. 2013. LDL receptor and its family members serve as the cellular receptors for vesicular stomatitis virus. Proc Natl Acad Sci U S A 110:7306–11. 165. Matsuyama S, Taguchi F. 2002. Receptor-Induced Conformational Changes of Murine Coronavirus Spike Protein. J Virol 76:11819–11826. 166. Kawase M, Kataoka M, Shirato K, Matsuyama S. 2019. Biochemical Analysis of Coronavirus Spike Glycoprotein Conformational Intermediates during Membrane Fusion. J Virol 93:1–19. 167. Park J, Li K, Barlan A, Fehr AR, Perlman S, McCray PB, Gallagher T. 2016. Proteolytic processing of Middle East respiratory syndrome coronavirus spikes expands virus tropism. Proc Natl Acad Sci U S A 113:12262–12267. 168. Mi D, Ou X, Li P, Peng G, Liu Y, Guo R, Mu Z, Li F, Holmes K, Qian Z. 2019. Glycine 29 is critical for conformational changes of spike glycoprotein of MHV-A59 triggered by either receptor binding or high pH. J Virol. 169. Walls AC, Tortorici MA, Snijder J, Xiong X, Bosch B-J, Rey FA, Veesler D. 2017. Tectonic conformational changes of a coronavirus spike glycoprotein promote membrane fusion. Proc Natl Acad Sci U S A 114:11157–11162.

133

170. Millet JK, Whittaker GR. 2015. Host cell proteases: Critical determinants of coronavirus tropism and pathogenesis. Virus Res 202:120–34. 171. Qiu Z, Hingley ST, Simmons G, Yu C, Das Sarma J, Bates P, Weiss SR. 2006. Endosomal proteolysis by cathepsins is necessary for murine coronavirus mouse hepatitis virus type 2 spike-mediated entry. J Virol 80:5768–76. 172. Matsuyama S, Taguchi F. 2009. Two-step conformational changes in a coronavirus envelope glycoprotein mediated by receptor binding and proteolysis. J Virol 83:11133–41. 173. Phillips JM, Gallagher T, Weiss SR. 2017. Neurovirulent Murine Coronavirus JHM.SD Uses Cellular Zinc Metalloproteases for Virus Entry and Cell-Cell Fusion. J Virol 91:76–84. 174. Park JE, Gallagher T. 2017. Lipidation increases antiviral activities of coronavirus fusion-inhibiting peptides. Virology 511:9–18. 175. Jensen S, Thomsen AR. 2012. Sensing of RNA viruses: a review of innate immune receptors involved in recognizing RNA virus invasion. J Virol 86:2900–10. 176. Suddala KC, Lee CC, Meraner P, Marin M, Markosyan RM, Desai TM, Cohen FS, Brass AL, Melikyan GB. 2019. Interferon-induced transmembrane protein 3 blocks fusion of sensitive but not resistant viruses by partitioning into virus-carrying endosomes. PLoS Pathog 15:e1007532. 177. Brass AL, Huang I-C, Benita Y, John SP, Krishnan MN, Feeley EM, Ryan BJ, Weyer JL, van der Weyden L, Fikrig E, Adams DJ, Xavier RJ, Farzan M, Elledge SJ. 2009. The IFITM Proteins Mediate Cellular Resistance to Influenza A H1N1 Virus, West Nile Virus, and Dengue Virus. Cell 139:1243–1254. 178. Diamond MS, Farzan M. 2013. The broad-spectrum antiviral functions of IFIT and IFITM proteins. Nat Rev Immunol 13:46–57. 179. Wrensch F, Winkler M, Pöhlmann S. 2014. IFITM Proteins Inhibit Entry Driven by the MERS-Coronavirus Spike Protein: Evidence for Cholesterol-Independent Mechanisms. Viruses 6:3683–3698. 180. Yu J, Li M, Wilkins J, Ding S, Swartz TH, Esposito AM, Zheng YM, Freed EO, Liang C, Chen BK, Liu SL. 2015. IFITM Proteins Restrict HIV-1 Infection by Antagonizing the Envelope Glycoprotein. Cell Rep 13:145–156. 181. Bailey CC, Zhong G, Huang I-C, Farzan M. 2014. IFITM-Family Proteins: The Cell’s First Line of Antiviral Defense. Annu Rev Virol 1:261–283. 182. Shirato K, Kanou K, Kawase M, Matsuyama S. 2017. Clinical Isolates of Human Coronavirus 229E Bypass the Endosome for Cell Entry. J Virol 91.

134

183. Yang Y, Du L, Liu C, Wang L, Ma C, Tang J, Baric RS, Jiang S, Li F. 2014. Receptor usage and cell entry of bat coronavirus HKU4 provide insight into bat-to- human transmission of MERS coronavirus. Proc Natl Acad Sci 111:12516–12521. 184. Yang Y, Liu C, Du L, Jiang S, Shi Z, Baric RS, Li F. 2015. Two Mutations Were Critical for Bat-to-Human Transmission of Middle East Respiratory Syndrome Coronavirus. J Virol 89:9119–23. 185. Klumperman J, Locker JK, Meijer A, Horzinek MC, Geuze HJ, Rottier PJ. 1994. Coronavirus M proteins accumulate in the Golgi complex beyond the site of virion budding. J Virol 68:6523–34. 186. Salanueva IJ, Carrascosa JL, Risco C. 1999. Structural maturation of the transmissible gastroenteritis coronavirus. J Virol 73:7952–64. 187. Xu Z, Lin Y, Zou C, Peng P, Wu Y, Wei Y, Liu Y, Gong L, Cao Y, Xue C. 2019. Attenuation and characterization of porcine enteric alphacoronavirus strain GDS04 via serial cell passage. Vet Microbiol 239:108489. 188. Nakagaki K, Nakagaki K, Taguchi F. 2005. Receptor-Independent Spread of a Highly Neurotropic Murine Coronavirus JHMV Strain from Initially Infected Microglial Cells in Mixed Neural Cultures. J Virol 79:6102–6110. 189. Kawase M, Shirato K, Matsuyama S, Taguchi F. 2009. Protease-mediated entry via the endosome of human coronavirus 229E. J Virol 83:712–21. 190. Matsuyama S, Nagata N, Shirato K, Kawase M, Takeda M, Taguchi F. 2010. Efficient activation of the severe acute respiratory syndrome coronavirus spike protein by the transmembrane protease TMPRSS2. J Virol 84:12658–64. 191. Taguchi F, Matsuyama S. 2009. [Cell entry mechanisms of coronaviruses]. Uirusu 59:215–22. 192. Bos ECW, Heijnen L, Luytjes W, Spaan WJM. 1995. Mutational analysis of the murine coronavirus spike protein: Effect on cell-to-cell fusion. Virology 214:453– 463. 193. Yamada Y, Liu XB, Fang SG, Tay FPL, Liu DX. 2009. Acquisition of cell-cell fusion activity by amino acid substitutions in spike protein determines the infectivity of a coronavirus in cultured cells. PLoS One 4:e6130. 194. Yamamoto M, Matsuyama S, Li X, Takeda M, Kawaguchi Y, Inoue J, Matsuda Z. 2016. Identification of Nafamostat as a Potent Inhibitor of Middle East Respiratory Syndrome Coronavirus S Protein-Mediated Membrane Fusion Using the Split- Protein-Based Cell-Cell Fusion Assay. Antimicrob Agents Chemother 60:6532– 6539. 195. Xia S, Xu W, Wang Q, Wang C, Hua C, Li W, Lu L, Jiang S. 2018. Peptide-Based Membrane Fusion Inhibitors Targeting HCoV-229E Spike Protein HR1 and HR2 Domains. Int J Mol Sci 19:8–11.

135

196. Gallagher TM, Buchmeier MJ, Perlman S. 1994. Dissemination of MHV4 (Strain JHM) Infection Does Not Require Specific Coronavirus Receptors, p. 279–284. In Advances in experimental medicine and biology. 197. Watanabe R, Matsuyama S, Taguchi F. 2006. Receptor-independent infection of murine coronavirus: analysis by spinoculation. J Virol 80:4901–4908. 198. Bender SJ, Phillips JM, Scott EP, Weiss SR. 2010. Murine Coronavirus Receptors Are Differentially Expressed in the Central Nervous System and Play Virus Strain- Dependent Roles in Neuronal Spread. J Virol 84:11030–11044. 199. Phillips JJ, Chua MM, Rall GF, Weiss SR. 2002. Murine Coronavirus Spike Glycoprotein Mediates Degree of Viral Spread, Inflammation, and Virus-Induced Immunopathology in the Central Nervous System. Virology 301:109–120. 200. Nash TC, Buchmeier MJ. 1996. Spike glycoprotein-mediated fusion in biliary glycoprotein-independent cell-associated spread of mouse hepatitis virus infection. Virology 223:68–78. 201. Massa PT, Wege H, ter Meulen V. 1986. Analysis of murine hepatitis virus (JHM strain) tropism toward Lewis rat glial cells in vitro. Type I astrocytes and brain macrophages (microglia) as primary glial cell targets. Lab Invest 55:318–27. 202. Massa PT, Wege H, ter Meulen V. 1988. Growth pattern of various JHM coronavirus isolates in primary rat glial cell cultures correlates with differing neurotropism in vivo. Virus Res 9:133–44. 203. van Berlo MF, Warringa R, Wolswijk G, Lopes-Cardozo M. 1989. Vulnerability of rat and mouse brain cells to murine hepatitis virus (JHM-strain): Studies in vivo and in vitro. Glia 2:85–93. 204. Zhou H, Chen Y, Zhang S, Niu P, Qin K, Jia W, Huang B, Zhang S, Lan J, Zhang L, Tan W, Wang X. 2019. Structural definition of a neutralization epitope on the N- terminal domain of MERS-CoV spike glycoprotein. Nat Commun 10:3068. 205. Rao P V., Gallagher TM. 1998. Intracellular complexes of viral spike and cellular receptor accumulate during cytopathic murine coronavirus infections. J Virol 72:3278–88. 206. Gallagher TM. 1996. Murine coronavirus membrane fusion is blocked by modification of thiols buried within the spike protein. J Virol 70:4683–90. 207. Rosenberger CM, Podyminogin RL, Askovich PS, Navarro G, Kaiser SM, Sanders CJ, McClaren JL, Tam VC, Dash P, Noonan JG, Jones BG, Surman SL, Peschon JJ, Diercks AH, Hurwitz JL, Doherty PC, Thomas PG, Aderem A. 2014. Characterization of innate responses to influenza virus infection in a novel lung type I epithelial cell model. J Gen Virol 95:350–362.

136

208. Hirano N, Fujiwara K, Hino S, Matumoto M. 1974. Replication and plaque formation of mouse hepatitis virus (MHV-2) in mouse cell line DBT culture. Arch Gesamte Virusforsch 44:298–302. 209. Hirano N, Fujiwara K, Matumoto M. 1976. Mouse hepatitis virus (MHV-2). Plaque assay and propagation in mouse cell line DBT cells. Jpn J Microbiol 20:219–25. 210. Zhang R, Li Y, Cowley TJ, Steinbrenner AD, Phillips JM, Yount BL, Baric RS, Weiss SR. 2015. The nsp1, nsp13, and M proteins contribute to the hepatotropism of murine coronavirus JHM.WU. J Virol 89:3598–609. 211. Shulla A, Gallagher T. 2009. Role of spike protein endodomains in regulating coronavirus entry. J Biol Chem 284:32725–32734. 212. Shulla A, Heald-Sargent T, Subramanya G, Zhao J, Perlman S, Gallagher T. 2011. A Transmembrane Serine Protease Is Linked to the Severe Acute Respiratory Syndrome Coronavirus Receptor and Activates Virus Entry. J Virol 85:873–882. 213. Campbell EM, Perez O, Melar M, Hope TJ. 2007. Labeling HIV-1 virions with two fluorescent proteins allows identification of virions that have productively entered the target cell. Virology 360:286–93. 214. Whitt MA. 2010. Generation of VSV pseudotypes using recombinant ΔG-VSV for studies on virus entry, identification of entry inhibitors, and immune responses to vaccines. J Virol Methods 169:365–374. 215. Brouillette RB, Phillips EK, Ayithan N, Maury W. 2017. Differences in Glycoprotein Complex Receptor Binding Site Accessibility Prompt Poor Cross-Reactivity of Neutralizing Antibodies between Closely Related Arenaviruses. J Virol 91:1–15. 216. van Boheemen S, de Graaf M, Lauber C, Bestebroer TM, Raj VS, Zaki AM, Osterhaus ADME, Haagmans BL, Gorbalenya AE, Snijder EJ, Fouchier RAM. 2012. Genomic characterization of a newly discovered coronavirus associated with acute respiratory distress syndrome in humans. MBio 3:1–9. 217. Nakane S, Matsuda Z. 2015. Dual Split Protein (DSP) Assay to Monitor Cell-Cell Membrane Fusion. Methods Mol Biol 1313:229–36. 218. Teeranaipong P, Hosoya N, Kawana-Tachikawa A, Fujii T, Koibuchi T, Nakamura H, Koga M, Kondo N, Gao GF, Hoshino H, Matsuda Z, Iwamoto A. 2013. Development of a rapid cell-fusion-based phenotypic HIV-1 tropism assay. J Int AIDS Soc 16:18723. 219. O’Doherty U, Swiggard WJ, Malim MH. 2000. Human immunodeficiency virus type 1 spinoculation enhances infection through virus binding. J Virol 74:10074– 80.

137

220. Rodgers W. 2002. Making membranes green: construction and characterization of GFP-fusion proteins targeted to discrete plasma membrane domains. Biotechniques 32:1044–6, 1048, 1050–1. 221. Sha Y, Wu Y, Cao Z, Xu X, Wu W, Jiang D, Mao X, Liu H, Zhu Y, Gong R, Li W. 2006. A convenient cell fusion assay for the study of SARS-CoV entry and inhibition. IUBMB Life 58:480–6. 222. Bosch BJ, Martina BEE, Van Der Zee R, Lepault J, Haijema BJ, Versluis C, Heck AJR, De Groot R, Osterhaus ADME, Rottier PJM. 2004. Severe acute respiratory syndrome coronavirus (SARS-CoV) infection inhibition using spike protein heptad repeat-derived peptides. Proc Natl Acad Sci U S A 101:8455–60. 223. Bosch BJ, Rossen JWA, Bartelink W, Zuurveen SJ, de Haan CAM, Duquerroy S, Boucher CAB, Rottier PJM. 2008. Coronavirus escape from heptad repeat 2 (HR2)-derived peptide entry inhibition as a result of mutations in the HR1 domain of the spike fusion protein. J Virol 82:2580–5. 224. Gallagher TM, Buchmeier MJ, Perlman S. 1992. Cell receptor-independent infection by a neurotropic murine coronavirus. Virology 191:517–22. 225. Taguchi F, Matsuyama S, Saeki K. 1999. Difference in Bgp-independent fusion activity among mouse hepatitis viruses. Arch Virol 144:2041–2049. 226. Yokomori K, Banner LR, Lai MM. 1991. Heterogeneity of gene expression of the hemagglutinin-esterase (HE) protein of murine coronaviruses. Virology 183:647– 57. 227. Yokomori K, Stohlman SA, Lai MMC. 1993. The Detection and Characterization of Multiple Hemagglutinin-Esterase (HE)-Defective Viruses in the Mouse Brain during Subacute Demyelination Induced by Mouse Hepatitis Virus. Virology 192:170–178. 228. Huang X, Dong W, Milewska A, Golda A, Qi Y, Zhu QK, Marasco WA, Baric RS, Sims AC, Pyrc K, Li W, Sui J. 2015. Human Coronavirus HKU1 Spike Protein Uses O -Acetylated Sialic Acid as an Attachment Receptor Determinant and Employs Hemagglutinin-Esterase Protein as a Receptor-Destroying Enzyme. J Virol 89:7202–7213. 229. Langereis MA, Zeng Q, Heesters B, Huizinga EG, de Groot RJ. 2012. The murine coronavirus hemagglutinin-esterase receptor-binding site: A major shift in ligand specificity through modest changes in architecture. PLoS Pathog 8:1–8. 230. Naskalska A, Dabrowska A, Szczepanski A, Milewska A, Jasik KP, Pyrc K. 2019. Membrane protein of HCoV-NL63 is responsible for interaction with the adhesion receptor. J Virol. 231. Belouzard S, Millet JK, Licitra BN, Whittaker GR. 2012. Mechanisms of coronavirus cell entry mediated by the viral spike protein. Viruses 4:1011–1033.

138

232. Leibowitz J, Kaufman G, Liu P. 2011. Coronaviruses: propagation, quantification, storage, and construction of recombinant mouse hepatitis virus. Curr Protoc Microbiol Chapter 15:Unit 15E.1. 233. Gamblin SJ, Haire LF, Russell RJ, Stevens DJ, Xiao B, Ha Y, Vasisht N, Steinhauer DA, Daniels RS, Elliot A, Wiley DC, Skehel JJ. 2004. The structure and receptor binding properties of the 1918 influenza hemagglutinin. Science 303:1838–42. 234. Stevens J, Blixt O, Glaser L, Taubenberger JK, Palese P, Paulson JC, Wilson IA. 2006. Glycan microarray analysis of the hemagglutinins from modern and pandemic influenza viruses reveals different receptor specificities. J Mol Biol 355:1143–1155. 235. Weis W, Brown JH, Cusack S, Paulson JC, Skehel JJ, Wiley DC. 1988. Structure of the influenza virus haemagglutinin complexed with its receptor, sialic acid. Nature 333:426–431. 236. Wanitchang A, Saenboonrueng J, Kaewborisuth C, Srisutthisamphan K, Jongkaewwattana A. 2019. A single V672F substitution in the spike protein of field-isolated PEDV Promotes Cell–Cell fusion and replication in Vero E6 cells. Viruses 11:1–15. 237. Damme EJMV. 1998. Handbook of plant lectins : properties and biomedical applications. Wiley. 238. Tortorici MA, Veesler D. 2019. Structural insights into coronavirus entry, p. 93– 116. In Advances in virus research. 239. Costello DA, Millet JK, Hsia C-Y, Whittaker GR, Daniel S. 2013. Single particle assay of coronavirus membrane fusion with proteinaceous receptor-embedded supported bilayers. Biomaterials 34:7895–904. 240. Li BX, Ge JW, Li YJ. 2007. Porcine aminopeptidase N is a functional receptor for the PEDV coronavirus. Virology 365:166–72. 241. Shirato K, Maejima M, Islam MT, Miyazaki A, Kawase M, Matsuyama S, Taguchi F. 2016. Porcine aminopeptidase N is not a cellular receptor of porcine epidemic diarrhea virus, but promotes its infectivity via aminopeptidase activity. J Gen Virol 97:2528–2539. 242. Li W, Luo R, He Q, van Kuppeveld FJM, Rottier PJM, Bosch B-J. 2017. Aminopeptidase N is not required for porcine epidemic diarrhea virus cell entry. Virus Res 235:6–13. 243. Stencel-Baerenwald JE, Reiss K, Reiter DM, Stehle T, Dermody TS. 2014. The sweet spot: Defining virus-sialic acid interactions. Nat Rev Microbiol 12:739–749.

139

244. Bertram S, Glowacka I, Blazejewska P, Soilleux E, Allen P, Danisch S, Steffen I, Choi S, Park Y, Schneider H, Schughart K, Pöhlmann S. 2010. TMPRSS2 and TMPRSS4 facilitate trypsin-independent spread of influenza virus in Caco-2 cells. J Virol 84:10016–25. 245. Menachery VD, Dinnon KH, Yount BL, McAnarney ET, Gralinski LE, Hale A, Graham RL, Scobey T, Anthony SJ, Wang L, Graham B, Randell SH, Lipkin WI, Baric RS. 2019. Trypsin treatment unlocks barrier for zoonotic bat coronaviruses infection. J Virol. 246. Bakkers MJG, Lang Y, Feitsma LJ, Hulswit RJG, de Poot SAH, van Vliet ALW, Margine I, de Groot-Mijnes JDF, van Kuppeveld FJM, Langereis MA, Huizinga EG, de Groot RJ. 2017. Betacoronavirus Adaptation to Humans Involved Progressive Loss of Hemagglutinin-Esterase Lectin Activity. Cell Host Microbe 21:356–366. 247. Wasik BR, Barnard KN, Ossiboff RJ, Khedri Z, Feng KH, Yu H, Chen X, Perez DR, Varki A, Parrish CR. 2017. Distribution of O-Acetylated Sialic Acids among Target Host Tissues for Influenza Virus. mSphere 2:1–17. 248. Byrd-Leotis L, Cummings RD, Steinhauer DA. 2017. The interplay between the host receptor and influenza virus hemagglutinin and neuraminidase. Int J Mol Sci 18. 249. Kazi L, Lissenberg A, Watson R, de Groot RJ, Weiss SR. 2005. Expression of hemagglutinin esterase protein from recombinant mouse hepatitis virus enhances neurovirulence. J Virol 79:15064–73. 250. Hensley LE, Holmes K V., Beauchemin N, Baric RS. 1998. Virus-Receptor Interactions and Interspecies Transfer of a Mouse Hepatitis Virus, p. 33–41. In Advances in experimental medicine and biology. 251. Chan C-M, Chu H, Wang Y, Wong BH-Y, Zhao X, Zhou J, Yang D, Leung SP, Chan JF-W, Yeung M-L, Yan J, Lu G, Gao GF, Yuen K-Y. 2016. Carcinoembryonic Antigen-Related Cell Adhesion Molecule 5 Is an Important Surface Attachment Factor That Facilitates Entry of Middle East Respiratory Syndrome Coronavirus. J Virol 90:9114–9127. 252. Chu H, Chan C-M, Zhang X, Wang Y, Yuan S, Zhou J, Au-Yeung RK-H, Sze K-H, Yang D, Shuai H, Hou Y, Li C, Zhao X, Poon VK-M, Leung SP, Yeung M-L, Yan J, Lu G, Jin D-Y, Gao GF, Chan JF-W, Yuen K-Y. 2018. Middle East respiratory syndrome coronavirus and bat coronavirus HKU9 both can utilize GRP78 for attachment onto host cells. J Biol Chem 293:11709–11726. 253. Han X, Qi J, Song H, Wang Q, Zhang Y, Wu Y, Lu G, Yuen K-Y, Shi Y, Gao GF. 2017. Structure of the S1 subunit C-terminal domain from bat-derived coronavirus HKU5 spike protein. Virology 507:101–109.

140

254. Kim Y, Cheon S, Min C-K, Sohn KM, Kang YJ, Cha Y-J, Kang J-I, Han SK, Ha N- Y, Kim G, Aigerim A, Shin HM, Choi M-S, Kim S, Cho H-S, Kim Y-S, Cho N-H. 2016. Spread of Mutant Middle East Respiratory Syndrome Coronavirus with Reduced Affinity to Human CD26 during the South Korean Outbreak. MBio 7:1–8. 255. Kleine-Weber H, Elzayat MT, Wang L, Graham BS, Müller MA, Drosten C, Pöhlmann S, Hoffmann M. 2019. Mutations in the Spike Protein of Middle East Respiratory Syndrome Coronavirus Transmitted in Korea Increase Resistance to Antibody-Mediated Neutralization. J Virol 93:1–14.

VITA

The author, Enya Qing was born in Chengdu, Sichuan, People’s Republic of

China on June 4, 1986 to Guangya Qing and Guanglan Luo. He moved to the United

States of America for his undergraduate studies in 2004, at Beloit College, WI. In 2007,

he received his first laboratory experience at Loyola University Chicago, where he

worked with Dr. Alan Wolfe and Dr. Karen Visick, on determining the mechanism

underlying the flagellation inhibition by cyclic-di-GMP in Vibrio fischeri. In 2008 he received his Bachelor’s of Science in Cellular & Molecular Biology, double-majored in

Studio Art. After graduation, Enya returned to China and worked in education.

In August 2013 Enya returned to the US and entered the Interdisciplinary

Program in Biomedical Sciences at Loyola University Chicago, and joined the

Department of Microbiology and Immunology shortly after. He completed his doctoral work in the laboratory of Tom Gallagher, Ph.D., where he focused on the mechanism of cellular factors that promote coronavirus entry. This work was partially supported by a

T32 Immunology Training Grant awarded to Dr. Katherine Knight. After completion of his graduate studies, Enya will continue his postdoctoral research in the Gallagher Lab, focusing on determining the mechanism of cellular entry factors for the newly-emerged novo 2019 coronavirus, SARS-CoV-2.

141