key: cord-0808471-9mkabkc6
authors: Pearson, Guy J; Broncel, Malgorzata; Snijders, Ambrosius P; Carlton, Jeremy G
title: Exploitation of the Secretory Pathway by SARS-CoV-2 Envelope
date: 2021-06-30
journal: bioRxiv
DOI: 10.1101/2021.06.30.450614
sha: 4f936e4572247a0f409efcaec66043b8c78580c6
doc_id: 808471
cord_uid: 9mkabkc6

The beta-coronavirus SARS-CoV-2 is the causative agent of the current global COVID-19 pandemic. Coronaviruses are enveloped RNA viruses. Assembly and budding of coronavirus particles occur at the Endoplasmic Reticulum-Golgi Intermediate Compartment (ERGIC), with the structural proteins Nucleocapsid, Spike, Membrane and Envelope facilitating budding and release of virions into the secretory pathway lumen. This allows viral release which can occur through delivery of virus particles to deacidified lysosomes and subsequent lysosomal secretion. Coronaviral Envelope proteins are necessary for coronavirus assembly, play important roles in replication and can form oligomeric cation channels. Whilst synthesised in the ER, the mechanism by which Envelope achieves its steady state localisation to the ERGIC remains unclear. Here, we used fluorescent reporters to illuminate the Envelope protein from SARS-CoV-2. We discovered that internal tagging of this protein is necessary to preserve the functionality of a C-terminal ER-export motif and to allow localisation of Envelope to the ERGIC. Using this non-disruptive form of tagging, we used proximity biotinylation to define the vicinal proteome of wild type and ER-restricted versions of Envelope. We show that both Envelope and the presence of its ER-export motif contribute to the packaging of nucleocapsid into virus like particles. Finally, using our labelled versions of Envelope, we discovered that a minor pool of this protein is delivered to lysosomes. We show that lysosomal Envelope is oligomeric and can contribute to pH neutralisation in these organelles.

The Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) is the causative agent of the global COVID-19 pandemic. SARS-CoV-2 is an enveloped, beta-coronavirus with a positive-sense, single-stranded RNA genome encoding 29 proteins 1 . Late events in the beta-coronaviral lifecycle are orchestrated by 4 of these proteins, the RNA-binding protein Nucleocapsid (N) and three transmembrane proteins Spike (S), Membrane (M) and Envelope (E). S is a class I transmembrane glycoprotein that is necessary for viral entry 2, 3 . M contains three-transmembrane domains and is the most abundant transmembrane in the viral particle 4 . E is a small single pass transmembrane protein that is found as a minor component of viral particles and can assemble as a pentameric cation channel 5, 6 . Viral assembly occurs at the Endoplasmic Reticulum/Golgi Intermediate Compartment (ERGIC) and

involves the budding of nascent S, M and E-containing particles into the ERGIC lumen 7, 8 . During this process, newly replicated viral RNA is incorporated into budding virions in complex with N. Virus like particle (VLP) systems represent a minimal system in which to study virion assembly, and have demonstrated that M and E are both necessary and sufficient for particle assembly [9] [10] [11] [12] and can package RNAs when N is present 10 . Once virus particles have been released into the secretory pathway, exposed sequences can be post-translationally modified and virions can egress from cells as luminal content. Whilst the canonical secretory pathway was assumed to be the egress route, recent data suggest that beta-coronaviruses can be delivered to deacidified lysosomes for atypical secretion via lysosomal exocytosis 13 . The mechanism by which lysosomal function is compromised in these infected cells is unclear, but may involve manipulation of HOPS-dependent lysosomal fusion by SARS-CoV-2 accessory proteins such as ORF3a 14, 15 or the presence of viroporins in these membranes. Beyond roles in assembly, what role, if any, SARS-CoV-2 E plays in modulating organellar pH remains to be determined.

We sought to create tagged versions of SARS-CoV-2 E to understand how this protein achieves its steady-state localisation at sites of particle assembly. Given the localisation of SARS-CoV E to membranes of the ERGIC and cis-Golgi 16, 17 and the degree of identity between E from SARS-CoV and SARS-CoV-2 ( Figure 1A ), we were surprised to find SARS-CoV-2 E bearing N-or C-terminal HaloTag (HT) fusions was restricted to the Endoplasmic Reticulum (ER) ( Figure 1A, 1B) . Positioning within the secretory pathway is governed by a cargo's interaction with cytoplasmic anterograde and retrograde trafficking machineries, most notably coatomer complex-1 and -2. We wondered if addition of tags to SARS-CoV-2 E's extreme termini altered folding or obscured these signals, preventing ER export to the ERGIC and cis-Golgi. We inserted HT at three additional positions in the coding sequence and discovered that tag insertion immediately after the transmembrane domain (Site3), or in a region of E's cytoplasmic tail (Site4) allowed steady state localisation of HTversions of SARS-CoV-2 E to perinuclear structures with Golgi morphology ( Figure   1A and 1B). In addition to the predominate perinuclear localisation, these versions of HT-E also decorated vesicular structures suggesting that they had fully cleared the ER and accessed the secretory pathway ( Figure 1A and 1B) . Placement of the HT at Site2 also prevented anterograde traffic. We confirmed these localisations using mEmerald (Em) in place of HT ( Figure S1A and S1B) and given the cell-to-cell variability observed when assessing membrane trafficking phenotypes, we devised a quantitative reporting system to visually depict E's position within the secretory pathway ( Figure 1C and Figure S1B ). We used a panel of antibodies and fusion proteins to confirm location of E-HT Site3 to the ERGIC and Golgi ( Figure 1D) suggesting that our internally tagged versions of E recapitulate the known localisation of coronaviral E proteins. Intriguingly, we also detected colocalisation of the peripheral puncta of E-HT Site3 with LAMP1, suggesting that a small pool of E is trafficked to lysosomes ( Figure 1D ). We also observed limited colocalization with large, but not small, EEA1-decorated puncta ( Figure 1D ), suggesting that at steady state, a small pool of E engages with the endocytic pathway and can be delivered to lysosomes.

The C-terminus of SARS-CoV E encodes a PDZ-ligand able to interact with PDZdomain containing proteins including Syntenin and PALS1. This sequence is conserved in SARS-CoV-2 E ( Figure 1A ) and we were surprised to find that its deletion restricted both Em and HT versions of SARS-CoV-2 E to the ER ( Figure 1E and 1F, Figure 1 Supplement 1C-E). Grafting this PDZ-sequence onto E-HT Site5 restored its anterograde traffic ( Figure 1G-1I) , indicating that this sequence acts as a dominant ER-export motif allowing SARS-CoV-2 E to gain access to the secretory pathway.

C-terminal hydrophobic sequences can act as COP-II binding sequences for ER export. Whilst the C-terminal Valine provided most of the activity relating to anterograde traffic, some SARS-CoV-2 E-Em Site3 Δ V still reached the Golgi and full truncation of the PDZ-domain was needed to restrict ER-export ( Figure 1 Supplement 1C-1E). Likewise, replacement of the DLLV PDZ-ligand with AAAV or AALL sequences could only partially recover ER-export ( Figure 1 Supplement 1C -1E) suggesting that the context of this hydrophobic terminal residue is important for ER export. Exploration of different classes of PDZ-ligands in this context revealed that a C-terminal hydrophobic residue was not dominant for ER-export, that a variety of class-I and class-II (but not class-III) PDZ domains could substitute for the DLLV sequence, although none were as effective in allowing ER-export as replacement of the DLLV sequence with chimeric C-termini from MHV (strain S) or MERS ( Figure 1 Supplement 1F -1H). These isolated sequences could similarly rescue ER-export when grafted onto the extreme C-terminus of SARS-CoV-2 E-HT Site5 ( Figure 1H and 1I). The beta-variant of SARS-CoV-2 encodes a P71L mutation in E, proximal to this sequence. We wondered if this mutation influenced ER-export and discovered that whilst SARS-CoV-2 E-Em Site3 P71L displayed steady state localisation to the Golgi, a fraction was retained in the ER and its ability to reach post-Golgi vesicles was limited (Figure 1 Supplement 1I and 1J). These data suggesting that the steady state distribution of E from beta variants of SARS-CoV-2 is shifted earlier in the secretory pathway.

The transmembrane glycoproteins of coronaviruses are heavily post translationally modified 8 . We wondered if SARS-CoV-2 E's positioning within the secretory pathway influenced the degree of these modifications. By immunoprecipitating SARS-CoV-2 E-HT Site3 , we discovered that versions of E that were retained in the ER were O- 

VLP systems have proved an excellent tool by which to analyse late events in the coronavirus lifecycles and have demonstrated that E and M are necessary and sufficient for particle assembly. Using a 4-component (E, S, M, N) VLP system in 293T cells, we discovered that SARS-CoV-2 M was essential for particle generation, as its omission resulted in a failure to recover N or S in VLP fractions ( Figure 2A ). Whilst omission of E had only minor effects on the biogenesis of VLPs, we noticed that particles produced in the absence of E contained less N ( Figure 2A ). We next wondered how restricting E to the ER would impact VLP generation. As with omission of E, we found that E Δ DLLV did not disrupt VLP generation, as M could be similarly recovered from the VLP fraction. However, N incorporation into these VLPs was similarly impaired ( Figure 2A ). These data suggest that E's C-terminal DLLV motif or the presence of E at the site of particle biogenesis supports N incorporation.

We next inserted HA-tagged TurboID, a fast-acting promiscuous biotin ligase 18 

To question how this minor pool of E reached lysosomes, we disrupted internalisation using a dominant-negative version of the key endocytic GTPase, Dynamin 22 . Whilst this prevented transferrin internalisation, it had no impact on the delivery of E-Em Site3 to lysosomes ( Figure 4A -4C). Although we cannot discount dynamin-independent internalisation, these data suggest that E is trafficked directly to lysosomes independently of the PM. Whilst b-coronaviruses assemble by budding into the ERGIC lumen, they are trafficked directly to lysosomes for secretion by lysosomal exocytosis 13 . Importantly, during infection, these secretory lysosomes are deacidified, which may limit proteolytic destruction of egressing viruses 13 . SARS-CoV-2 E is predicted to form a pentameric cation channel 23 and given the predicted parallels between the route toward lysosomes taken by both assembled virions and SARS-CoV-2 E, we wondered whether SARS-CoV-2 E contributed to pH neutralisation in lysosomes. To function as a cation channel, SARS-CoV-2 E must oligomerise. We used Fluorescence Lifetime Imaging-Forster Radius Energy Transfer (FLIM-FRET) in cells expressing both Em Site3 -and HT Site3 -versions of SARS-CoV-2 E to assess its oligomeric status. In this assay, protein-protein interactions are revealed as a reduction in the donor fluorophore's fluorescence lifetime. We discovered that when combined with JF646-labelled E-HT Site3 , the lifetime of E-Em Site3 in lysosomes was reduced, suggesting that this pool of E is oligomeric ( Figure 4D and 4E). E-Em Site3 's lifetime reduction in the Golgi was less influenced by the presence of JF646-labelled E-HT Site3 ( Figure 4D and 4E). We also used Δ DLLV versions of E and discovered that when restricted to the ER, the lifetime of E-Em Site3 Δ DLLV was not influenced by the presence of JF646-labelled E-HT Site3 Δ DLLV ( Figure 4F and 4G). These data suggest that a progressive increase in the oligomeric status of E as it moves through the secretory pathway. We next employed a recently described reporter of lysosomal pH based upon a modified version of LAMP1 fused to luminal pH-sensitive superfolder-GFP and cytosolic pH-insensitive mCherry 24 (Figure 4 Supplement 1). Here, we discovered that pHLARE-positive lysosomes containing higher levels of E-HT Site3 were less acidic than pHLAREpositive lysosomes in the same cell containing lower levels of E-HT Site3 ( Figure 4H -I).

These data suggest that SARS-CoV-2 E is trafficked from Golgi to lysosomes and can both oligomerise and contribute to pH neutralisation in these organelles.

Here, we developed a labelling strategy to allow visualisation of SARS-CoV-2 E. Importantly, we found that canonical tagging mechanisms are incompatible with preservation of E's biology resulting in their restriction to the ER. For C-terminal tags, we suggest that this is likely due the occlusion of a C-terminal DLLV sequence that acts as an ER-export signal. This may influence the interpretation of previously described localisations 25 , vicinal [26] [27] [28] or physical 29 interactomes reported for SARS-CoV-2 E. This DLLV sequence has been described previously to function as a PDZligand, and consistent with this, we observed a significant enrichment of PDZ-domain containing proteins in our vicinal proteomes from internally TurboID versions of E.

We were surprised too at the degeneracy of this C-terminal sequence. Common with many secretory cargos, a C-terminal hydrophobic residue provided the majority, although not all, of the ER-export activity. This is consistent with previously described roles for these sequences in binding to COP-II components 30, 31 , but suggests that the context of the hydrophobic residue is important for export.

Additionally, a variety of C-terminal PDZ-ligands of different classes and sequences could license at least limited export, albeit none approaching that of wildtype E, or that obtained when using versions of E bearing chimeric C-termini from MHV or MERS.

The importance of this C-terminal sequence has been demonstrated in the context of SARS-CoV infection, whereby mutation or deletion of this sequence generated revertants or reacquisition of these sequences by alternate proteins after serial passage in culture or infection in mice 32 suggesting that this sequence acts to enhance pathogenesis of coronaviral infection 21 

CoV Δ E replicated with 2-to 3-log lower titres than wildtype virus in animal models 33 , but the mechanism by which E contributes to productive infection is unknown. Using a VLP system with SARS-CoV-2 proteins, we found that N packaging was reduced in VLPs generated in the absence of E, or in VLPs generated with versions of E that were restricted to the ER. These data suggest that E plays an important role in particle assembly, not necessarily by sculpting the virion, but by allowing the incorporation of N, and thus genomic RNA into the assembling particle. It may be that this defect underlies the crippled replication dynamics of recombinant Δ E coronaviruses, or explains their aberrant morphologies 8 

CoV proteins has demonstrated that N is packaged into VLPs in an E-dependent manner, but only in a minimal system in which N-M interactions had been disrupted 11 . Similarly, requirements for E's C-terminal hydrophobic residue in releasing extracellular nucleocapsid were exposed in an E-only system but were not observed when M was present 11 . As such, we suggest there exist differences between N packaging mechanisms between SARS-CoV and SARS-CoV-2, with E playing a more important role in this process for SARS-CoV-2. Alternatively, it is possible that the presence of S, which was absent from the minimal systems described in 11 , renders N packaging more reliant upon E.

Although E is thought to contribute to the virus assembly process, coronavirus particles contain only limited amounts of this protein and the majority of E remains in the host cell 6 . We discovered that beyond the Golgi, a minor pool of E was delivered to lysosomes in a Dynamin-independent manner, suggesting that it is trafficked internally to these organelles. Coronaviral E proteins oligomerise into pentameric cation channels 5, 23 . We used FLIM-FRET using our internally tagged versions of E to explore at what stage of the secretory pathway oligomerisation was observed. Whilst we could detect FRET between labelled versions of E in lysosomes, we could not detect this FRET between the same proteins when they were localised to the ER.

Limited FRET was detected in the Golgi, indicating that a mixed population of oligomers was observed here and suggesting that channel formation occurs late along E's secretory journey. Using a newly developed ratiometric sensor of lysosomal pH, we discovered that E's presence resulted in lysosomal pH neutralisation, suggesting that an assembly-independent role of E is to compromise the degradative environment in these organelles, which may be important for preserving virus shed through secretory lysosomes. from Sigma-Aldrich for 180 minutes, or were untreated, and imaged in green and red channels using the system described. Images were analysed in FIJI using a customwritten script that removed background in all channels and then identified the lysosomes by their presence in the mCherry red channel, and then measured the Integrated Density of each of these lysosomes in both green and red channels. Data Methionine oxidation, Acetyl (N-term), Acetyl (K) and Deamidation (NQ) were selected as variable modifications. The enzyme specificity was set to Trypsin with a maximum of 2 missed cleavages. The precursor mass tolerance was set to 20 ppm for the first search (used for mass re-calibration) and to 4.5 ppm for the main search.

The datasets were filtered on posterior error probability (PEP) to achieve a 1% false discovery rate on protein, peptide and site level. Other parameters were used as preset in the software. 'Unique and razor peptides' mode was selected to allow identification and quantification of proteins in groups (razor peptides are uniquely assigned to protein groups and not to individual proteins). Intensity based absolute quantification (iBAQ) in MaxQuant was performed using a built-in quantification algorithm 36 

All data from label-free quantification of proximity biotinylation proteomics of HA-TurboID tagged SARS-CoV-2 E, SARS-CoV-2 E mutants, and HA-TurboID cytoplasmic controls.

Data subset of label free quantification data of proximity biotinylation proteomics of HA-TurboID tagged SARS-CoV-2 E compared to HA-TurboID cytoplasmic control.

Data subset of label free quantification data of proximity biotinylation proteomics of HA-TurboID tagged SARS-CoV-2 E compared to SARS-CoV-2 E Δ DLLV.

Characteristics of SARS-CoV-2 and COVID-19

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor

Genotype and phenotype of COVID-19: Their roles in pathogenesis

SARS coronavirus E protein forms cation-selective ion channels

Coronavirus envelope (E) protein remains at the site of assembly

The intracellular sites of early replication and budding of SARScoronavirus

Coronavirus envelope protein: Current knowledge

Assembly of human severe acute respiratory syndrome coronavirus-like particles

Assembly of Severe Acute Respiratory Syndrome Coronavirus RNA Packaging Signal into Virus-Like Particles Is Nucleocapsid Dependent

SARS-CoV envelope protein palmitoylation or nucleocapid association is not required for promoting virus-like particle production

The M, E, and N Structural Proteins of the Severe Acute Respiratory Syndrome Coronavirus Are Required for Efficient Assembly, Trafficking, and Release of Virus-Like Particles

Coronaviruses Use Lysosomes for Egress

ORF3a of the COVID-19 virus SARS-CoV-2 blocks HOPS complex-mediated assembly of the SNARE complex required for autolysosome formation

Structural basis for SARS-CoV-2 envelope protein recognition of human cell junction protein PALS1

Subcellular location and topology of severe acute respiratory syndrome coronavirus envelope protein

Identification of a Golgi Complex-Targeting Signal in the Cytoplasmic Tail of the Severe Acute Respiratory Syndrome Coronavirus Envelope Protein

Efficient proximity labeling in living cells and organisms with TurboID

Sars-cov-2 envelope (e) protein interacts with pdz-domain-2 of host tight junction protein zo1

Structural basis of coronavirus E protein interactions with human PALS1 PDZ domain

The PDZ-Binding Motif of Severe Acute Respiratory Syndrome Coronavirus Envelope Protein Is a Determinant of Viral Pathogenesis

Membrane fission by dynamin: what we know and what we need to know

Structure and drug binding of the SARS-CoV-2 envelope protein transmembrane domain in lipid bilayers

pHLARE: A new biosensor reveals decreased lysosome pH in cancer cells

A comprehensive library of fluorescent constructs of SARS-CoV-2 proteins and their initial characterisation in different cell types

Global BioID-based SARS-CoV-2 proteins proximal interactome unveils novel ties between viral polypeptides and host factors involved in multiple COVID19-associated mechanisms

A SARS-CoV-2 BioID-based virus-host membrane protein interactome and virus peptide compendium: new proteomics resources for COVID-19

A SARS-CoV-2 -host proximity interactome

A SARS-CoV-2 protein interaction map reveals targets for drug repurposing

Role of cytoplasmic C-terminal amino acids of membrane proteins in ER export

Signals for COPII-dependent export from the ER: What's the ticket out?

Identification of the Mechanisms Causing Reversion to Virulence in an Attenuated SARS-CoV for the Design of a Genetically Stable Vaccine

A Severe Acute Respiratory Syndrome Coronavirus That Lacks the E Gene Is Attenuated In Vitro and In Vivo

Parallels between cytokinesis and retroviral budding: A role for the ESCRT machinery

Rapid Global Fitting of Large Fluorescence Lifetime Imaging Microscopy Datasets

MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification

Andromeda: A peptide search engine integrated into the MaxQuant environment

The Perseus computational platform for comprehensive analysis of (prote)omics data

The CRAPome: A contaminant repository for affinity purification-mass spectrometry data

The EMBL-EBI search and sequence analysis tools APIs in 2019

Jalview Version 2-A multiple sequence alignment editor and analysis workbench

PDZ domains -Common players in the cell signaling

JCG is a Wellcome Trust Senior Research Fellow (206346/Z/17/Z). This work was supported by the Francis Crick Institute which receives its core funding from Cancer Research UK (FC001002, FC001999) , the UK Medical Research Council FC001002, FC001999), and the Wellcome Trust (FC001002, FC001999) . We thank the Tooze