(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Expanded ACE2 dependencies of diverse SARS-like coronavirus receptor binding domains [1]

['Sarah M. Roelle', 'Department Of Pathology', 'Case Western Reserve University School Of Medicine', 'Cleveland', 'Ohio', 'United States Of America', 'Nidhi Shukla', 'Anh T. Pham', 'Anna M. Bruchez', 'Kenneth A. Matreyek']

Date: 2022-08 

Viral spillover from animal reservoirs can trigger public health crises and cripple the world economy. Knowing which viruses are primed for zoonotic transmission can focus surveillance efforts and mitigation strategies for future pandemics. Successful engagement of receptor protein orthologs is necessary during cross-species transmission. The clade 1 sarbecoviruses including Severe Acute Respiratory Syndrome-related Coronavirus (SARS-CoV) and SARS-CoV-2 enter cells via engagement of angiotensin converting enzyme-2 (ACE2), while the receptor for clade 2 and clade 3 remains largely uncharacterized. We developed a mixed cell pseudotyped virus infection assay to determine whether various clades 2 and 3 sarbecovirus spike proteins can enter HEK 293T cells expressing human or Rhinolophus horseshoe bat ACE2 proteins. The receptor binding domains from BtKY72 and Khosta-2 used human ACE2 for entry, while BtKY72 and Khosta-1 exhibited widespread use of diverse rhinolophid ACE2s. A lysine at ACE2 position 31 appeared to be a major determinant of the inability of these RBDs to use a certain ACE2 sequence. The ACE2 protein from Rhinolophus alcyone engaged all known clade 3 and clade 1 receptor binding domains. We observed little use of Rhinolophus ACE2 orthologs by the clade 2 viruses, supporting the likely use of a separate, unknown receptor. Our results suggest that clade 3 sarbecoviruses from Africa and Europe use Rhinolophus ACE2 for entry, and their spike proteins appear primed to contribute to zoonosis under the right conditions.

Data Availability: Numerical data for all plots shown in Figs 1 – 8 or S1 – S11 , as well as an R Markdown script capable of recreating the calculation and generation of all plots, are provided at https://github.com/MatreyekLab/ACE2_dependence . The aggregate infection data, as well as all summarized datasets used to make each plot, are published with the paper as Supporting Information files.

Here, we characterize the extent of ACE2 dependence across sarbecovirus clades. We utilized a single-copy HEK 293T genome modification platform to strongly overexpress multiple cell surface proteins proposed to serve as receptors for SARS-CoV-2, alongside the well-established receptor, ACE2 [ 12 ]. As the clade 2 and clade 3 sarbecoviruses were observed in samples collected from various Rhinolophus bats, we synthesized and expressed ACE2 orthologs from Rhinolophus ferrumequinum, Rhinolophus affinis, R. alyone, Rhinolophus landeri, Rhinolophus pearsonii, and various ACE2 alleles observed in Rhinolophus sinicus. We observed differing patterns of ACE2 ortholog usage by various clade 3 sarbecoviruses RBDs during cell entry, including human ACE2-dependent entry by the BtKY72 and Khosta-2 RBDs. We observed little to no ACE2-dependent infection with RBDs from clade 2 sarbecoviruses, including various alleles from R. sinicus and R. pearsonii from which these viruses were isolated. Thus, our study provides a new genetic approach for characterizing receptor utilization during viral entry and demonstrated that clade 3 sarbecoviruses likely utilize ACE2 as a cell-entry receptor during infection.

Multiple viral clades exist within the sarbecovirus subgenus, and the cell surface receptor dependencies of each clade are not well established [ 10 , 11 ]. Clade 1 sarbecoviruses including SARS-CoV and SARS-CoV-2 are known to utilize ACE2, while the receptors for clade 2 and clade 3 viruses are unknown [ 10 , 11 ]. The lack of observed ACE2-dependent enhancement to infection by clade 2 and clade 3 sarbecovirus spike proteins, such as YN2013 or BM48–31, can be explained in 3 ways: (1) these RBDs have weak but functionally relevant affinity for ACE2, below the limit of detection of commonly used assay; (2) these RBDs have affinity for certain orthologs of ACE2, but little or no affinity for human ACE2 or for any orthologs that have been tested so far; or (3) these RBDs primarily utilize an entry mechanism distinct from ACE2.

Molecular compatibility during viral entry is a key determinant of viral tropism and host switching [ 2 – 6 ]. The Betacoronavirus genus include known zoonotic viruses of pandemic potential including Middle East Respiratory Syndrome-related Coronavirus (MERS-CoV), SARS-CoV, and SARS-CoV-2. These viruses use the spike glycoprotein to catalyze entry into target cells upon binding to a compatible host cell receptor. Unlike MERS-CoV, which uses Dipeptidyl-peptidase 4 (DPP4) as the cell surface receptor [ 7 ], the lineage B viruses of the sarbecovirus subgenus SARS-CoV and SARS-CoV-2 utilize angiotensin converting enzyme-2 (ACE2) as the host cell entry receptor [ 8 , 9 ]. ACE2 binding from SARS-like CoVs is dictated by an independently folded domain of up to 223 residues in length, referred to as the receptor binding domain (RBD).

As shown by the ongoing Severe Acute Respiratory Syndrome-related Coronavirus 2 (SARS-CoV-2) pandemic, viral spillover from animal reservoirs can decimate public health systems and the global economy. The likelihoods of zoonotic spillovers are multifactorial, including both ecological and molecular factors. Human disruptions to world ecosystems are increasing the likelihood of future zoonotic events [ 1 ]. We still lack a clear understanding of the molecular factors that play key roles during zoonosis.

Results

Developing a robust genetic assay for viral entry Knowing that sarbecovirus spike proteins may exhibit weak affinity for ACE2 proteins from mismatched hosts, we designed an assay for measuring biochemically weak but functionally important interactions promoting viral entry. We previously developed a Bxb1 recombinase-based transgenic expression system, wherein human ACE2 or its coding variants could be stably and precisely expressed by a Tet-inducible promoter already engineered into the cell genome, upon integration of a single promoterless plasmid [12] (Fig 1A). We found that the human ACE2 cDNA, when encoded behind a consensus Kozak sequence permitting frequent ribosomal translation of the mRNA, yielded high ACE2 cell surface abundance, roughly 10-fold greater than ACE2 protein observed in Vero-E6 cells, commonly used to propagate SARS-CoV or SARS-CoV-2 in cell culture [12]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Duplex pseudovirus infection assay. (A) Schematic of the Bxb1 recombinase “landing pad” engineered in HEK 293T cells and the elements of the attB recombination vectors used to stably express transgenic DNA upon plasmid integration. Inverted red triangles are 2A translational stop-start sequences. (B) Cartoons describing the traditional singleplex and new duplex pseudovirus infection formats. (C and D) Representative microscopy images (C) and flow cytometry profiles (D) of ACE2-dependent and independent pseudovirus infection in the duplex infection assay. The scale bar denoting 50 μm distance shown in the top-left images apply to all image panels. (E) Comparison of pseudovirus infection results obtained using the singleplex and duplex formats. Error bars denote 95% confidence intervals from 7 replicate experiments. Fold increase to infection in ACE2 overexpressing cells over nonexpressing cells is shown on the bottom, while the CV of the replicate results are shown on the top. The underlying data can be found in S2 Data, and the source code can be found at https://github.com/MatreyekLab/ACE2_dependence. ACE2, angiotensin converting enzyme-2; BFP, blue fluorescent protein; Bsd, blasticidin resistance gene; CV, coefficient of variation; GFP, green fluorescent protein; H2A, histone 2A; iCasp9, inducible caspase 9; Int, Bxb1 integrase; MFI, xxxxx; Pac, puromycin resistance gene; SARS-CoV, Severe Acute Respiratory Syndrome-related Coronavirus; SARS-CoV-2, Severe Acute Respiratory Syndrome-related Coronavirus 2; VSV, vesicular stomatitis virus; VSV-G, vesicular stomatitis virus glycoprotein. https://doi.org/10.1371/journal.pbio.3001738.g001 We previously discovered that our pseudovirus infection system was more sensitive than traditional in vitro binding assays utilizing soluble proteins. For example, expression of ACE2 mutants K31D or K353D, which reduced binding to soluble monomeric SARS-CoV RBD in vitro [13], had little to no effect for SARS-CoV spike pseudovirus infection when translated from a consensus Kozak sequence in our expression platform [12]. Instead, we only observed reduced pseudovirus infection when the K31D or K353D ACE2 mutant protein levels were reduced 30-fold, suggesting that avidity effects conferred by high cell-surface ACE2 abundances can compensate for reductions to binding affinity. Thus, we focused on further developing a flexible pseudovirus infection assay capable of detecting weak but specific protein interactions enabling infection. To increase throughput, we converted the traditional singleplex pseudovirus infection format into a duplex assay configuration. Traditionally, control and experimental cells are plated separately into different wells (Fig 1B, left). All wells are then exposed to the same volume of viral inoculum and infectivity is quantitated by taking a ratio of the amount of infection present in the experimental wells divided by the amount of infection present in the control wells. While ensemble measurements such as luciferase activity require a traditional singleplex format, fluorescent reporters for infection, such as GFP positivity, are single-cell assays and easier to multiplex. Thus, we developed an approach wherein the experimental and control cells are marked by different fluorescent proteins, allowing the 2 cell types to be mixed together and infected by the same inoculum of GFP-reporter pseudovirus within the same well (Fig 1B, right). Instead of calculating the ratio of GFP positivity from 2 different wells, we take the ratio of GFP positivity in mCherry-negative control cells or putative receptor overexpressing iRFP670-positive cells, all from a single well. When testing 2 nearly isogenic cell lines differing solely by their expression of a putative receptor transgene, this ratio quantifies the amount of receptor-dependent enhancement to infection that has occurred. To validate this approach, we created ACE2(dEcto)-negative control HEK 293T cells encoding human ACE2 lacking its entire ectodomain, and thus incapable of serving as a cell surface receptor for SARS-CoV or SARS-CoV-2 spike. We marked these cells with red nuclei using mCherry-fused histone H2A (Fig 1B). Notably, HEK 293T cells naturally express a low but detectable amount of endogenous ACE2 from the X chromosome [12], thus accounting for the low, background level of infection in the assay. We next created ACE2 HEK 293T cells encoding full-length human ACE2 and marked these cells with near-infrared fluorescent nuclei using iRFP670-fused histone H2A. These cells exhibit more than 100-fold increased ACE2 protein than unmodified HEK 293T cells [12]. These cells were mixed into the same well and exposed to GFP encoding lentiviral particles coated with the ACE2-dependent envelope glycoprotein SARS-CoV spike (Fig 1C, left), or an ACE2-independent envelope glycoprotein such as vesicular stomatitis virus glycoprotein (VSV-G; Fig 1C, right), which uses LDLR as the viral entry receptor [14]. After 2 or more days, the entire well of cells can be analyzed with multicolor flow cytometry to simultaneously measure the infection rates in ACE2-expressing or control cells. We observed that ACE2-dependent viruses, such as those with SARS-CoV spike, exhibited preferential infection of the ACE2-expressing and iRFP670-fluorescent cells, whereas pseudoviruses coated with VSV-G infected the mCherry and iRFP670 expressing cells equally (Fig 1D). We will heretofore refer to this as the duplex infection assay. We next performed a systematic analysis of how the duplex infection assay performed when the 2 cells were mixed at different ratios and compared these results with data obtained using the traditional singleplex assay format. We observed the greatest ACE2-dependent infection when the ACE2-expressing cells were a 10th of the total cells in the well (Fig 1E), with the coefficient of variation similar to the traditional singleplex assay format. As the proportion of ACE2-expressing cells increased, the amount of ACE2-dependent SARS-CoV-2 spike mediated infection reduced from approximately 52-fold at 10% ACE2-expressing cells to approximately 16-fold at 40% ACE2-expressing cells. There was a concomitant increase in the coefficient of variation suggesting a loss of data precision, at least partially due to insufficient sampling of the background level of infection in the control cells. While we generally tried to keep the ACE2-expressing cells a minor fraction of the mixed cultures, some experiments were performed before we had characterized this phenomenon. We observed that the SARS-CoV ACE2 dependencies in experiments where the ACE2-expressing cells were a larger fraction yielded lower magnitudes of ACE2 dependence, consistent with our above results and explaining some of the heterogeneity in effect sizes between experiments (S1 Fig). Thus, when the receptor-expressing cells were infrequent in the mixed pool of cells, the mixed-cell infection assay was capable of producing data of comparable magnitude and precision to the traditional singleplex assay format, while reducing the number of total samples and requiring fewer physical manipulations. Due to these largely favorable characteristics, we used the duplex infection assay for all subsequent experiments.

Known and proposed receptors for SARS-CoV-2 spike-mediated infection While ACE2 is the primary SARS-CoV-2 spike receptor, numerous other proteins have been suggested to serve as alternative receptors. BSG encodes CD147 / Basigin, which was proposed to be a novel host cell receptor for SARS-CoV-2 [15], though this has since been refuted [16,17]. CLEC4M encodes L-SIGN / CD209L, which, along with the related DC-SIGN / CD209, was proposed to be a receptor [18,19], and glycomimetic antagonists can block this interaction and inhibit SARS-CoV-2 infection [20]. Lectins are generally regarded as attachment factors rather than bona fide receptors as they have been implicated in enhancing entry for over 30 different viral glycoproteins [21]. NRP1 and NRP2 encode Neuropilin-1 / NRP1 and Neuropilin-2 / NRP2, which were proposed to be receptors since they bind peptides formed upon furin cleavage [22,23], and the SARS-CoV-2 spike has a furin cleavage site that is important for its transmission [24], while SARS-CoV spike does not. We used our duplex infection assay platform to compare the functional impacts of these proposed alternative receptors with ACE2 during SARS-CoV and SARS-CoV-2 spike pseudotyped virus infection. We created plasmid constructs encoding both untagged and cytoplasmically HA-tagged cDNAs of each proposed receptor protein (Fig 2A) and genomically integrated these DNAs alongside an IRES-iRFP670-H2A cassette into HEK 293T cells. Cells encoding the HA-tagged versions were immunoblotted to confirm the expression of each protein (Fig 2B). The predominant bands in the ACE2, CD147, and L-SIGN lysates corresponded to the electrophoretic migration sizes of the full-length, glycosylated proteins. In contrast to these 3 proteins, the bands corresponding to NRP1 and NRP2 were less abundant, although bands consistent with full-length protein were seen upon longer exposure (Fig 2B, top). Thus, all 5 proteins were expressed from their corresponding cDNAs, albeit to varying steady-state abundances. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Established and proposed SARS-CoV-2 receptor proteins. (A) Schematic showing relative lengths, tag locations, and overall domain topologies of proteins. (B) Representative immunoblot of cells stably expressing HA-tagged versions of each protein, including longer (top) and shorter (bottom) anti-HA exposures. Beta actin was blotted as a loading control. (C) Scatter plot comparing normalized pseudovirus infection rates of cells stably expressing various HA-tagged and untagged proteins. Shapes correspond to receptors. Colors correspond to viruses. (D) Compiled normalized infection data for the established and proposed SARS-CoV-2 receptor proteins. Error bars denote 95% confidence intervals for at least 10 replicates. Asterisks denote samples with p < 0.01. The underlying data can be found in S2 Data, and the source code can be found at https://github.com/MatreyekLab/ACE2_dependence. ACE2, angiotensin converting enzyme-2; SARS-CoV, Severe Acute Respiratory Syndrome-related Coronavirus; SARS-CoV-2, Severe Acute Respiratory Syndrome-related Coronavirus 2; VSV-G, vesicular stomatitis virus glycoprotein. https://doi.org/10.1371/journal.pbio.3001738.g002 We next determined how the presence of each protein enhanced SARS-CoV and SARS-CoV-2 entry. Pseudovirus infection of cells expressing the HA-tagged or untagged forms of the protein were nearly identical (Pearson’s r2: 0.98, n = 15; Fig 2C). Thus, to improve statistical power for weak effect sizes, we merged the 2 datasets. Consistent with the known importance of ACE2, its expression increased infection with SARS-CoV 10-fold and SARS-CoV-2 spike 12-fold, while pseudoviruses with VSV-G were unaffected. L-SIGN increased SARS-CoV-2 spike infection 2.8-fold, while it had a more modest 1.7-fold effect on SARS-CoV spike (Fig 2D). The next strongest effect was a 1.5-fold increase to SARS-CoV-2 spike-mediated infection conferred by NRP2, though it similarly simulated infection by VSV-G, which is not processed by furin. Thus, our assay revealed that ectopic ACE2 expression conferred the strongest increase to SARS-CoV and SARS-CoV-2 spike-mediated pseudovirus infection while L-SIGN conferred a milder but still significant increase.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001738

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/