Nanoscopy reveals heparan sulfate clusters as docking sites for SARS-CoV-2 attachment and entry

eLife Assessment

The authors use solid high resolution microscopy techniques to present a model in which SARS-CoV-2 attachment and endocytosis are mediated by heparan sulfate, whereas ACE2 only functions downstream of these early processes. These findings serve as a potentially valuable starting point to examine SARS-CoV-2 entry models that challenge current paradigms. However, examination of the model in a clean heparan sulfate deficient background and in the context of TMPRSS2-dependent plasma membrane fusion are lacking. Thus, the broader impact remains uncertain in the absence of additional controls and orthogonal approaches.

https://doi.org/10.7554/eLife.108925.3.sa0

Significance of the findings:

Valuable: Findings that have theoretical or practical implications for a subfield

Landmark
Fundamental
Important
Valuable
Useful

Strength of evidence:

Solid: Methods, data and analyses broadly support the claims with only minor weaknesses

Exceptional
Compelling
Convincing
Solid
Incomplete
Inadequate

During the peer-review process the editor and reviewers write an eLife Assessment that summarises the significance of the findings reported in the article (on a scale ranging from landmark to useful) and the strength of the evidence (on a scale ranging from exceptional to inadequate). Learn more about eLife Assessments

Abstract
Introduction
Results
Discussion
Methods
Data availability
References
Article and author information
Metrics

Abstract

Virus entry is thought to involve binding a unique receptor for cell attachment and cytosolic entry. For SARS-CoV-2 underlying the COVID-19 pandemic, angiotensin-converting enzyme 2 (ACE2) is widely considered the receptor for cell-surface attachment and subsequent cell entry. Using advanced light microscopy to resolve individual virions and receptors, we found instead that heparan sulfate (HS), not ACE2, mediates SARS-CoV-2 cell-surface attachment, and subsequent endocytosis. ACE2 functions only downstream of HS to enable viral genome expression. Instead of binding single HS molecules that electrostatically interact with viral surface proteins weakly, SARS-CoV-2 binds clusters of ~6–137 HS molecules projecting 60–410 nm above the plasma membrane. These tall, HS-rich clusters, present at about one per 6 μm², act as docking sites for viral attachment. Blocking HS binding with the clinically used HS-binding agent pixantrone strongly inhibited an authentic pathogen, the SARS-CoV-2 Omicron JN.1 subvariant, from attaching to and infecting human airway cells. This work establishes a revised entry paradigm in which HS clusters mediate SARS-CoV-2 attachment and endocytosis, with ACE2 acting downstream, thereby identifying HS interactions as a key anti-COVID-19 strategy. This paradigm and its therapeutic implications may apply broadly beyond COVID-19 because, analogous to SARS-CoV-2, HS binds many other viruses but is only considered an attachment regulator.

Introduction

Understanding how viruses enter host cells is essential for developing effective vaccines and therapies. Viral entry is often initiated by engagement of host receptors and associated co-factors that together facilitate subsequent viral membrane penetration. Decades of research suggest that heparan sulfate (HS), a linear sulfated anionic polysaccharide covalently linked to cell-surface proteoglycans in many cell types, can facilitate attachment of viruses and bacteria by virtue of negative charge interaction with basic virion surface amino acids (Koehler et al., 2020; Lebsir et al., 2023). It is thought that viruses exploit weakly interacting with HS to augment the chance of contacting a more specific, high-affinity cell-surface entry receptor (Cagno et al., 2019).

A prominent contemporary example of this model is severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) entry. SARS-CoV-2 is thought to use human angiotensin-converting enzyme 2 (hACE2) as its receptor, based on spike protein (S) high-affinity binding to hACE2, and the requirement of hACE2 for infection (Jackson et al., 2022; Li et al., 2003; Shang et al., 2020; Lan et al., 2020; Hoffmann et al., 2020). HS binds S in vitro enhancing the open conformation of the receptor-binding domain (Shang et al., 2020; Lan et al., 2020). Perturbation of HS significantly reduces SARS-CoV-2 infection in many cell types (Clausen et al., 2020; Zhang et al., 2020; Bermejo-Jambrina et al., 2021). These results led to the widely accepted view that SARS-CoV-2 attachment is mediated by SARS-CoV-2 binding to cell surface hACE2, which can be significantly enhanced by the attachment factor HS (Jackson et al., 2022; Clausen et al., 2020; Zhang et al., 2020; Bermejo-Jambrina et al., 2021).

Based on similar experimental results, HS is thought to facilitate the binding of other CoVs to cell surface hACE2 (Lebsir et al., 2023; Cagno et al., 2019; Lang et al., 2011; Milewska et al., 2014). This includes SARS-CoV, responsible for the SARS outbreak two decades ago, and human coronavirus NL63, a cause of the common cold (Lebsir et al., 2023; Cagno et al., 2019; Lang et al., 2011; Milewska et al., 2014). HS is also considered an attachment factor for medically important non-CoVs including hepatitis C, HIV, Ebola, papilloma, and herpes viruses (Koehler et al., 2020; Lebsir et al., 2023). Together, these studies establish HS as an important factor for a broad range of viruses (and other pathogens) that augments pathogen binding with specific cognate cell surface receptors. However, to our knowledge, there is no direct, high-resolution microscopic evidence demonstrating that viruses, including SARS-CoV-2, bind their specific receptors at the cell surface, and not, for example, only after internalization into an endosome. Here, we address this decisive lacuna in the currently accepted model for SARS-CoV-2 entry.

To this end, we employed several advanced microscopic methods to achieve unprecedented precision in defining SARS-CoV-2 entry. Using confocal microscopy, stimulated emission depletion microscopy (STED) at ~60 nm resolution (Shin et al., 2018), and minimal photon flux nanoscopy (MINFLUX) at ~1–3 nm localization precision (Schmidt et al., 2021; Gwosch et al., 2020), and electron microscopy (EM), we show that HS, not hACE2 (Jackson et al., 2022), is the receptor for SARS-CoV-2 cell-surface attachment and endocytic internalization. Virions exclusively dock at HS clusters that may protrude hundreds of nanometers from the plasma membrane (PM). Internalized virions then bind endosomal hACE2 to deliver viral cores to the cytosol to initiate infection. Blocking HS binding with a clinically used HS-binding agent leads to substantial inhibition of authentic SARS-CoV-2 infection in primary human airway cells. These findings suggest perturbing S-HS binding as a broadly effective strategy for blocking infection by SARS-CoV-2 and other viruses that rely on HS clusters for internalization.

Results

Pseudotyped SARS-CoV-2 internalization is independent of ACE2

To observe the virus entering cells and replication, we first used a replicative recombinant vesicular stomatitis virus (VSV) expressing EGFP (infection indicator) and Wuhan-hu-1 SARS-CoV-2 S in place of the natural G receptor/fusion protein (Dieterle et al., 2020; Case et al., 2020) (VSV_EGFP-S, Figure 1A). VSV_EGFP-S virus was amplified in Vero or BHK_hACE2 cells. The amplified VSV_EGFP-S virus was concentrated for purification. The purified VSV_EGFP-S was labeled with Alexa or Atto dyes (e.g. Alexa Fluor 647 NHS ester, abbreviated as A647) via primary amines (Cureton et al., 2010; e.g. VSV_EGFP-S-A647, Figure 1A, see Methods) for seeing virions entering cells. VSV_EGFP-S-A647 virions attached to glass showed that nearly all viral A647 (V-A647) spots overlapped with immunolabeled S puncta (92.96 ± 0.02%, 606 particles, Figure 1A). This result suggests that the purified fluorescent V-A647 spots reflect VSV_EGFP-S-A647 virions, rather than non-S-containing fluorescent spots (e.g. VSV_EGFP-A647). We used similar methods to label other recombinant or authentic SARS-CoV-2 virions with various Alexa/Atto dyes (see Methods).

Figure 1 with 2 supplements see all

Download asset Open asset

hACE2-independent internalization of pseudo-typed SARS-CoV-2.

(A) *Upper*: Schematic drawing of a pseudo-typed SARS-CoV-2 virion, VSV_EGFP-S-A647, in which the *EGFP*-coding sequence was inserted into the genomic RNA of VSV, the G protein of VSV was replaced with S of SARS-CoV-2, and A647 was conjugated with primary amines of superficial envelope proteins. *Lower*: Confocal images of A647 of purified VSV_EGFP-S-A647 (V-A647, red) on coverslips and immunofluorescent staining with antibody against S2 subunit of S-protein (green). (B) Bright-field and confocal images of VSV_EGFP-S-A647’s A647 (V–A647) and *EGFP* expression (*V-EGFP*) in BHK21 cells stably expressing hACE2 (BHK_hACE2), BHK21 cells without ACE2 expression (BHK), and BHK cells transiently overexpressed with hACE2-BFP plasmid (BHK +hACE2). Top: Incubation protocol – cells were incubated with VSV_EGFP-S-A647 for 1 hr at 37 °C, followed by washout for 24 hr (applies to panels B–G). (C) The white dotted box in panel B enlarged to show V-A647 spots in EGFP-labeled cytosol. (D) Similar to panel B, except in 293T_hACE2 and 293T cells. (E) Similar to panel B, except in Vero (containing ACE2), Hela (no ACE2), and MRC5 (no ACE2) cells. (F) Similar to panel B, except VSV_EGFP-S-A647 was replaced with Lenti_EGFP-S_Omi-A647. (G) Fluorescence intensity of VSV_EGFP-S-A647’s V-A647 (F₆₄₇) and *V-EGFP* (F_EGFP) in BHK_hACE2 (471 cells, 3 independent experiments), BHK (370 cells, 3 independent experiments), BHK +hACE2 (515 cells, 3 independent experiments), 293T_hACE2 (404 cells, 3 independent experiments), 293T (234 cells, 3 independent experiments), Vero (295 cells, 3 independent experiments), Hela (305 cells, 3 independent experiments), and MRC5 (237 cells, 3 independent experiments) cells. Lenti_EGFP-S_Omi-A647’s F₆₄₇ and F_EGFP in BHK_hACE2 (364 cells, 3 independent experiments) and BHK (335 cells, 3 independent experiments) cells are also plotted. Virus incubation protocol is shown in panel B. F₆₄₇: mean ± s.e.m., reflecting viral uptake; F_EGFP: mean ± s.e.m., reflecting viral infection. F₆₄₇ and F_EGFP were measured per cell identified in the bright field and were normalized to cells expressing ACE2 in each subgroup (see gaps between subgroups). ***: p<0.001, t-test.

Figure 1—source data 1 Numerical values underlying the viral uptake and viral gene expression quantifications shown in Figure 1G.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig1-data1-v1.xlsx
Download elife-108925-fig1-data1-v1.xlsx

4–24 h after 1 h VSV_EGFP-S-A647 incubation with cells at 37 °C (protocol in Figure 1B), viral EGFP (V-EGFP) expression, which indicates viral genome expression, was observed in various ACE2-expressing cells, including hACE2-stably-expressed BHK21 (BHK_hACE2) or 293T cells (293T_hACE2), and Vero cells expressing ACE2 endogenously; in contrast, V-EGFP was not expressed in non-ACE2-expressing cells, including BHK21 (BHK), 293T, Hela, and MRC5 cells (Figure 1, Figure 1—figure supplement 1). Similarly, flow cytometric measurements of V-EGFP fluorescence from many thousand cells showed V-EGFP expression in BHK_hACE2 but not BHK cells (Figure 1—figure supplement 2). Transient hACE2 overexpression restored V-EGFP expression in BHK cells (Figure 1B-lower and G). These results confirmed the hACE2-dependence of SARS-CoV-2 infection (Shang et al., 2020; Lan et al., 2020; Zhou et al., 2020; Walls et al., 2020; Wang et al., 2020).

Surprisingly, V-A647 spot intensity and distribution were independent of ACE2 in the seven cell lines tested (Figure 1B–E, summarized in Figure 1G, Figure 1—figure supplement 1). Confocal sectioning revealed that these V-A647 spots were in the cytoplasm (Figure 1C, Figure 2—figure supplement 1), suggesting that VSV_EGFP-S-A647 internalization is independent of ACE2.

Extending these observations, incubating BHK_hACE2 or BHK cells with Lenti_EGFP-S_Omi-A647, an A647-labeled, EGFP-expressing lentivirus (Lenti_EGFP) pseudotyped with S-protein Omicron variant (S_omi), revealed similar hACE2-dependent EGFP expression, but hACE2-independent virion internalization (Figure 1F, G, Figure 1—figure supplement 1). While showing that VSV-S and Lenti-S_Omi infection is hACE2-dependent as expected, these results revealed surprisingly that VSV-S and Lenti-S_Omi internalization is hACE2-independent.

Endocytic internalization is the main viral infection pathway

Given that hACE2 is required for the fusogenic activity of pseudotyped or authentic SARS-CoV-2 (Jackson et al., 2022), the hACE2-independent internalization of V-A647 spots from VSV_EGFP-S-A647 or Lenti_EGFP-S_Omi-A647 (Figure 1) may reflect endocytic internalization. Four lines of evidence described below strengthen this suggestion.

First, to track endocytic vesicles, we included a fluorescent dye, Atto 490LS (A490), in the bath solution, which can be taken up by endocytosis into, and thus label, the endocytic vesicle (Figure 2—figure supplement 1, Figure 2A; Shin et al., 2021; Wu and Chan, 2024). BHK_hACE2 or BHK cells were incubated with VSV_EGFP-S-A647 plus A490 in the bath solution for 1 h at 37 °C, followed by washout. 24 h later, we found near complete colocalization of cytoplasmic V-A647 with A490 spots in either BHK_hACE2 or BHK cells (Figure 2—figure supplement 1A–C). This result suggests that endocytosis takes up both the virus and the bath A490 in an ACE2-independent manner.

Figure 2 with 3 supplements see all

Download asset Open asset

Endocytosis is the primary route for pseudotyped SARS-CoV-2 internalization and gene expression.

(A) *Left upper*: incubation protocol for viral attachment and early endocytosis – incubating cells with VSV_EGFP-S-A647 for 1 hr at 4 °C, followed by 3 min washout with 500–750 μM A490 at 37 °C (applies to panel A-D). *Left lower*: strategy for labeling endocytic vesicles taking up PM-attached virions (VSV_EGFP-S-A647, red) with a fluorescent dye in the bath (A490, cyan). Right: STED images showing V-A647 spots (red dots) colocalization with A490 spots (cyan) in the cytoplasm (cell boundary beyond the plot). Representative images are from BHK cells. (B) STED XY-plane images of V-A647 and A490 at different Z sections as labeled showing a cytosolic A490 spot (endocytic vesicle) containing a single V-A647 spot that reflects a single VSV-S virion. Representative images are from BHK cells. (C) STED XY-plane images of V-A647 and A490 showing a cytosolic A490 spot may contain 1–4 V-A647 spots, each reflecting a single virion. Top: the V-A647 spot number (N_A647) is labeled. Representative images from BHK cells. (D) *Left*: W_H distribution of A490 spots (Spot_A490) containing V-A647 spots (55 Spot_A490, 3 independent experiments from BHK cells). Middle: N_A647 plotted versus the corresponding Spot_A490 W_H (55 Spot_A490, 3 independent experiments, from BHK cells). A linear regression fit (correlation coefficient: 0.53) showed that larger Spot_A490 contains more V-A647 spots. Right: V-A647-containing Spot_A490 W_H (mean ± s.e.m.) in BHK_hACE2 (70 Spot_A490, 3 independent experiments) and BHK cells (55 Spot_A490, 3 independent experiments). (E) *Left*: EM images showing that endosomes may contain one or more VSV_EGFP-S-A647 virions (arrows). Representative images are from BHK_hACE2 cells. Incubation protocol: VSV_EGFP-S-A647 for 1 hr at 4 °C, 15 min washout at 37 °C. *Right*: W_H (mean ± s.e.m.) of EM-observed virus (VSV_EGFP-S-A647)-containing endosome (38 endosomes, 2 independent experiments) and STED-observed virus (VSV_EGFP-S-A647)-containing Spot_A490 (70 Spot_A490, 3 independent experiments). (F) Bright field and confocal images of V-A647 (uptake) and *V-EGFP* (viral gene expression) in BHK_hACE2 cells in control (Ctrl) or treated with dynasore (DnS, 80 μM, bath). Top: VSV_EGFP-S-A647 incubation protocol – 1 hr incubation, 24 hr washout, all at 37 °C. (G) VSV_EGFP-S-A647’s F₆₄₇ (mean ± s.e.m., uptake) and F_EGFP (mean ± s.e.m., gene expression) in cells in Ctrl (789 cells, 3 independent experiments), treated with DnS (653 cells, 3 independent experiments), or overexpressed with dynamin 2 K44A (19 cells, 2 independent experiments). Virus incubation protocol shown in panel F. **: p<0.01; ***: p<0.001, t-test. (H) *Left*: bright field and confocal images of V-A647 (uptake) in BHK_hACE2 cells in Ctrl or treated with cathepsin inhibitor 1 (CPS Inh, 10 μM). Top: VSV_EGFP-S-A647 incubation protocol – 1 hr incubation at 37 °C. Right: bright field and confocal images of V-A647 (uptake) and *V-EGFP* (gene expression) in BHK_hACE2 cells in Ctrl or treated with CPS Inh. Top: VSV_EGFP-S-A647 incubation protocol – 1 hr incubation, 24 hr washout, all at 37 °C. (I) F₆₄₇ (mean ± s.e.m., uptake) or F_EGFP (mean ± s.e.m., gene expression) in cells in Ctrl (540 cells, 3 independent experiments) or treated with CPS Inh (742 cells, 3 independent experiments). F₆₄₇ and F_EGFP were measured after VSV_EGFP-S-A647 incubation in panel H-left and H-right, respectively. ***: p<0.001, t-test.

Figure 2—source data 1 Numerical values underlying the vesicle size, virion number, viral uptake, and viral gene expression analyses shown in Figure 2.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig2-data1-v1.xlsx
Download elife-108925-fig2-data1-v1.xlsx

Second, we adsorbed VSV_EGFP-S-A647 to BHK_hACE2 or BHK cells for 1 hr at 4 °C to prevent endocytosis (cold temperature prevents endocytosis) and then incubated cells with A490 at 37 °C for 3 min to allow for, and thus image, endocytic vesicle formation within 3 min (Figure 2A). STED microscopy showed that most V-A647 spots contained no A490, but colocalized with the PM (Figure 2—figure supplement 2, 14 cells, 4 cultures) and thus labeled the cell outline. Inside this outline, V-A647 spots colocalized with A490 spots in either BHK_hACE2 or BHK cells, indicating ACE2-independent endocytic uptake of the virus with the bath A490 as early as 3 min (11 cells, 3 independent experiments; Figure 2A–C). Consistent with this observation, prolonged washout (4–24 hr) completely depleted V-A647 at the cell-surface outline while increasing cytoplasmic V-A647 (Figure 1B–F), suggesting that surface-attached virions are taken up by endocytosis within 4 h.

Third, 3D-STED imaging showed that A490-labeled vesicle’s full-width-at-half-maximum (W_H) was 363±17 nm (n=55) in BHK cells, similar to that (333±13 nm, n=70) in BHK_hACE2 cells (Figure 2B–D), supporting a similar endocytic path regardless of hACE2 presence or not. As W_H increased, V-A647 spot number per vesicle increased from 1 to 5 (Figure 2B and C). V-A647 spots appeared cylinder-shaped (length: 135–224 nm) or round-shaped (W_H: 74–118 nm) (Figure 2B and C). The cylinder shape is consistent with ~180 × 70 nm bullet-shaped VSV virions (Cureton et al., 2010; Cureton et al., 2009), whereas the round shape may reflect the cross-section of the cylinder shape.

Thin-section (70 nm) EM showed endosome-like structures with a diameter of 365±28 nm (range: 115–789 nm; 38 structures) containing 1–3 virions with the expected width and length of VSV (Cureton et al., 2010; Cureton et al., 2009; Figure 2E). The EM-measured diameter of endosomes containing virions was similar to the STED-measured W_H of A490-labeled vesicles containing V-A647 spots (Figure 2E), strengthening the STED observation of endocytic vesicles. The EM-observed maximal virion number (Cagno et al., 2019) in an endosome was less than the STED-observed (Li et al., 2003), likely because we counted virions only in one thin section with EM, but the entire A490-labeled vesicle with 3D-STED microscopy.

The large size (115–789 nm) of the virus-containing vesicles observed with STED and EM implies macropinocytosis that may form large vesicles engulfing multiple virions. Our observation is consistent with a recent study showing that a macropinocytosis inhibitor 5-[N-ethyl-N-isopropyl] amiloride (EIPA; Mercer et al., 2010) reduces SARS-CoV-2 replication in Vero cells (Zhang et al., 2022).

Fourth, dynasore or dominant-negative dynamin 2-K44A overexpression, which inhibits fission of dynamin-dependent endocytosis (Wu et al., 2014; Chanaday et al., 2019; Antonny et al., 2016), substantially reduced V-A647 internalized 1–24 hr after viral incubation (Figure 2F-G, Figure 2—figure supplement 3). This result suggests that viruses are internalized via dynamin-dependent endocytosis.

In addition to inhibiting V-A647 endocytosis, dynasore or dynamin 2-K44A inhibited V-EGFP expression 8–24 hr after virus incubation by ~66–77% (Figure 2F-G, Figure 2—figure supplement 3), suggesting that endocytosis is the main route for viral genome expression. Further strengthening this suggestion, cathepsin inhibitor 1, which inhibits endosomal cathepsin L crucial for SARS-CoV-2 endosomal fusion and infection (Asaad et al., 2009; Huang et al., 2006; Simmons et al., 2005; Zhao et al., 2021), nearly abolished V-EGFP expression without affecting V-A647 uptake (Figure 2H and I). We concluded that endocytic internalization is the major pathway for pseudotyped SARS-CoV-2 genome expression.

Heparan sulfate, but not hACE2, is required for viral cell-surface attachment and internalization

Four sets of evidence suggest that HS, but not hACE2, is the receptor underlying VSV-S cell-surface attachment. First, VSV-S and Lenti-S_Omi internalization is hACE2-independent (Figure 1, Figure 1—figure supplement 1), implying that the cell-surface attachment of these viruses is independent of hACE2. Second, with 1 hr viral incubation at 4 °C for virus attachment at the cell surface, V-A647 intensity at the cell surface was similar between BHK_hACE2 cells that contained hACE2 at the cell surface and BHK cells that did not contain hACE2 (Figure 3A and B), revealing hACE2-independent viral cell-surface attachment. Further supporting this conclusion, flow cytometric measurements from many thousands of cells showed similar V-A647 intensity between BHK_hACE2 and BHK cells after 1 hr viral incubation at 4 °C for viral cell-surface attachment (Figure 3—figure supplement 1). Flow cytometry also showed that the V-A647 attachment at the cell surface of BHK cells was similar to that in BHK_hACE2 cells over a ~600-fold dilution of the virus concentration (Figure 3—figure supplement 2), indicating that the virus cell-surface attachment is independent of hACE2 across a wide range of virus concentrations. Third, 93–94% V-A647 spots colocalized with immunolabeled HS spots in BHK_hACE2 (94 ± 2%, 8 cells) and BHK (93 ± 2%, 8 cells) cells (Figure 3C), indicating hACE2-independent VSV-S colocalization with HS. HS spots may thus serve as the docking site for VSV-S attachment at the cell surface. Fourth, heparinase I/II/III mixture (HPRase), which nearly abolished cell-surface immunolabeled HS (Figure 3D; López-Muñoz et al., 2023), reduced cell-surface V-A647 by ~83–85% in both BHK_hACE2 and BHK cells (after 1 hr viral incubation at 4 °C for viral cell-surface attachment, Figure 3E), suggesting that HS is crucial for VSV-S cell-surface attachment. Application of HPRase also reduced intracellular V-A647 spots by ~81–82% in both BHK_hACE2 and BHK cells, suggesting that HS-mediated viral attachment is a critical step preceding viral endocytosis (Figure 3F; incubation protocol: 1 hr virus incubation at 4 °C for viral attachment, followed by 1 hr washout at 37 °C for viral uptake). Further, HPRase reduced V-EGFP expression in BHK_hACE2 cells by ~79%, suggesting that HS-mediated viral attachment is a critical step preceding viral genome expression (24 hr after virus incubation, Figure 3G). In summary, these results (Figure 3A–G) suggest that HS binding with VSV-S underlies viral cell-surface attachment, which is crucial for the subsequent virus endocytosis and genome expression.

Figure 3 with 2 supplements see all

Download asset Open asset

Heparan sulfate (HS) is the receptor for pseudo-typed SARS-CoV-2 attachment at the cell surface.

(A) Immunolabeling (A647) of hACE2 in a BHK_hACE2 (left) and a BHK (right) cell. (B) ACE2-independent viral cell-surface attachment. *Images*: STED images showing virus cell-surface attachment in a BHK_hACE2 (left) or BHK (right) cell. Middle: images in green (left) and yellow (right) boxes replotted on a larger scale. Top: incubation protocol – virus (VSV_EGFP-S-A490) 1 hr at 4 °C for viral attachment at the cell surface (applies to panels B, C, E, and H). *Bar graph*: Cell-surface virus fluorescence intensity (F_Virus) per BHK_hACE2 (mean ± s.e.m., 92 cells, 3 independent experiments) or BHK cell (mean ± s.e.m., 98 cells, 3 independent experiments). Data normalized to the mean of the BHK_hACE2 group. (C) Cell-surface-attached viruses colocalized with HS puncta. *Images*: STED images of viral spots (VSV_EGFP-S-A490’s V-A490) and immunolabeled HS (Alexa 488-conjugated antibody) at BHK_hACE2 or BHK cell surface. Top: incubation protocol – VSV_EGFP-S-A647 1 hr at 4 °C for cell-surface attachment. *Bar graph*: the percentage of viral spots overlapping with HS puncta in BHK_hACE2 (mean ± s.e.m., 8 cells, 4 independent experiments) or BHK cells (8 cells, 4 independent experiments). (D) Bright field and confocal images of immunolabeled HS (left) and HS fluorescence intensity (F_HS, right) in BHK_hACE2 cells in Ctrl (mean ± s.e.m., 188 cells, 3 independent experiments) or treated with Heparinases I/II/III (HPRase, mean ± s.e.m., 264 cells, 3 independent experiments). ***: p<0.001, t-test. (E) *Images*: bright field and confocal image of V-A647 in BHK_hACE2 cells in Ctrl or treated with HPRase. Top: incubation protocol – VSV_EGFP-S-A647 1 hr at 4 °C for cell-surface attachment. *Bar graph*: cell-surface F₆₄₇ per cell (mean ± s.e.m.) for BHK_hACE2 cells (left) in Ctrl (125 cells, 3 independent experiments) or treated with HPRase (150 cells, 3 independent experiments), and for BHK cells (right) in Ctrl (147 cells, 3 independent experiments) or treated with HPRase (127 cells, 3 independent experiments). ***: p<0.001, t-test. (F) Images: bright field and confocal image of V-A647 (uptake) in BHK_hACE2 cells in Ctrl or treated with HPRase. Top: incubation protocol – VSV_EGFP-S-A647 1 h rat 4 °C (for viral attachment), 1 hr washout at 37 °C (for endocytosis of cell-surface-attached viruses). ***: p<0.001, t-test. *Bar graph*: F₆₄₇ (mean ± s.e.m., uptake) for BHK_hACE2 cells in Ctrl (201 cells, 3 independent experiments) or treated with HPRase (197 cells, 3 independent experiments), and for BHK cells in Ctrl (202 cells, 3 independent experiments) or treated with HPRase (206 cells, 3 independent experiments). (G) *Images*: bright field and confocal image of *V-EGFP* expression in BHK_hACE2 cells in control or treated with HPRase. Top: incubation protocol – VSV_EGFP-S-A647 1 hr at 4 °C, 24 hr washout at 37 °C. ***: p<0.001, t-test. *Bar graph*: F_EGFP (mean ± s.e.m., viral gene expression) per BHK_hACE2 cell in Ctrl (482 cells, 3 independent experiments) or treated with HPRase (391 cells, 3 independent experiments). (H) STED images of viral spots (VSV_EGFP-S-A490’s A490) and immunolabeled hACE2 puncta (A647) at a BHK_hACE2 cell surface. (I) STED images of viral spots (VSV_EGFP-S-A490’s A490) and immunolabeled hACE2 puncta (A647) inside BHK_hACE2 cells – hACE2 is sometimes taken up with viruses in endosomes.

Figure 3—source data 1 Numerical values underlying the viral attachment, heparan sulfate abundance, viral uptake, viral gene expression shown in Figure 3.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig3-data1-v1.xlsx
Download elife-108925-fig3-data1-v1.xlsx

hACE2 acts downstream of endocytosis to promote viral infection

We showed that hACE2 is essential for viral genome expression but not cell-surface viral attachment or endocytosis (Figures 1—3), suggesting that hACE2 acts downstream of endocytosis. Unlike HS spots, only 22 ± 2% cell-surface V-A647 spots colocalized with hACE2 spots in BHK_hACE2 cells (Figure 3H; 13 cells, 2 independent experiments). However, 38 ± 3% cell-surface V-A647 spots were within 200 nm from neighboring hACE2 spots (Figure 3H, 13 BHK_hACE2 cells, 2 independent experiments). These virions and hACE2 may be packed into the same (large) vesicle, as we noted that 38 ± 5% intracellular V-A647 spots colocalized with or within 200 nm from hACE2 spots (13 BHK_hACE2 cells, 2 independent experiments; Figure 3I). These results suggest that hACE2’s infection role may be due to cell-surface hACE2 co-internalization with virions to facilitate cathepsin L-dependent (Figure 2H–I; Asaad et al., 2009; Huang et al., 2006; Simmons et al., 2005; Zhao et al., 2021) viral fusion at the endosomal membrane.

VSV-S recapitulates the SARS-CoV-2 entry pathway

To extend our findings to authentic SARS-CoV-2 virions, we repeated the above VSV-S experiments using SARS-CoV-2 Omicron variant (CoV2-S_omi) (López-Muñoz et al., 2023; Suzuki et al., 2022), containing inserted mCherry gene and A647- or A490-labeled envelope proteins (CoV2_mCherry-S_Omi-A647 or CoV2_mCherry-S_Omi-A490). In addition, we replaced S in VSV-S with S_omi for comparison between wild-type (VSV_EGFP-S-A647) and Omicron variants (VSV_EGFP-S_omi-A647). Three sets of evidence suggest that the authentic SARS-CoV-2 entry path is similar to VSV-S.

First, VSV-S (Figure 1B–E , and G), Lenti-S_omi (Figure 1F–G), VSV-S_omi (Figure 4A–B), and CoV2-S_Omi (Figure 4C–D) all exhibited hACE2-independent viral internalization, but hACE2-dependent viral gene (EGFP or mCherry) expression. Second, similar to VSV-S (Figure 2A–D), 1–5 CoV2-S_omi virions were found in ~150–700 nm vesicles in the cytoplasm (Figure 4E, 18 cells, 2 independent experiments). Third, analogous to VSV-S (Figure 3E), removing HS by HPRase nearly abolished cell-surface-attached CoV2-S_omi (Figure 4F). Based on these observations, we conclude that the Omicron variant of SARS-CoV-2 enters BHK cells indistinguishably to VSV-S.

Figure 4

Download asset Open asset

SARS-CoV-2 cell entry path is analogous to pseudo-typed SARS-CoV-2.

(A) Images (bright field and confocal images) of VSV_EGFP-S_omi-A647’s V-A647 (left) and F₆₄₇ (right, viral uptake) in BHK_hACE2 (mean ± s.e.m.,146 cells, 3 independent experiments) or BHK (mean ± s.e.m.,149 cells, 3 independent experiments) cells. Top: incubation protocol – VSV_EGFP-S_omi-A647 1 h at 37 °C (for virus endocytosis). (B) Images (bright field and confocal images) of VSV_EGFP-S_omi-A647’s *V-EGFP* (left) and F_EGFP (right, viral gene expression) in BHK_hACE2 (mean ± s.e.m., 273 cells, 3 independent experiments) or BHK (mean ± s.e.m., 200 cells, 3 independent experiments) cells. Top: incubation protocol – VSV_EGFP-S_omi-A647 1 hr, 24 hr washout, all at 37 °C (for viral gene expression). (C) Images (bright field and confocal images) of CoV2_mCh-S_omi-A647’s A647 fluorescence (V-A647, left) and F₆₄₇ (right, uptake) in BHK_hACE2 (mean ± s.e.m., 120 cells, 3 independent experiments) or BHK (mean ± s.e.m., 120 cells, 3 independent experiments) cells. Top: incubation protocol – CoV2_mCh-S_omi-A647 1 hr at 37 °C (for virus endocytosis). (D) Images (bright field and confocal images) of CoV2_mCh-S_omi-A647’s *mCherry* expression (*V-mCherry*, pseudo color, left) and F_mCherry (right, uptake) in BHK_hACE2 (mean ± s.e.m., 120 cells, 3 independent experiments) or BHK (mean ± s.e.m., 120 cells, 3 independent experiments) cells. Top: incubation protocol – CoV2_mCh-S_omi-A647 1 hr, 6 hr washout, all at 37 °C (for viral gene expression). (E) STED XY-plane images of CoV2_mCh-S_omi-A647’s V-A647 and A490 in cytosol showing that the number (N_A647) of V-A647 spots (virions) in an A490 spot (endocytic vesicle) may vary as labeled. Top: incubation protocol – CoV2_mCh-S_omi-A647 1 hr at 4 °C, 3 min washout with 500–750 μM A490 at 37 °C. (F) Images (bright field and confocal images) of CoV2_mCh-S_omi-A647’s V-A647 (left) and surface F₆₄₇ per cell (right) for BHK_hACE2 cells in Ctrl (mean ± s.e.m., 120 cells, 3 independent experiments) or treated with HPRase (mean ± s.e.m., 120 cells, 3 independent experiments). Top: incubation protocol – CoV2_mCh-S_omi-A647 1 hr at 4 °C for cell-surface attachment.

Figure 4—source data 1 Numerical values underlying the viral uptake, viral gene expression, and cell-surface attachment analyses shown in Figure 4.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig4-data1-v1.xlsx
Download elife-108925-fig4-data1-v1.xlsx

Structural arrangement of HS as the SARS-CoV-2 attachment receptor

We characterized HS topology at molecular resolution to better understand its function as the virus receptor. 3D confocal imaging of immunolabeled HS revealed 563±38 HS-spots per BHK_hACE2 cell-surface (range: 244–860 spots, 16 cells, 4 independent experiments, Figure 5A). Given an average cell-surface area of 3412±360 μm² (14 cells, Figure 5—figure supplement 1), this amounts to an average of 1 HS-spot per 6 μm² (or 1 per 2.5×2.5 μm²). 2-D MINFLUX imaging (~3 nm localization precision; Gwosch et al., 2020) revealed oval-shaped clusters of HS molecules labeled by Flux 647-conjugated anti-HS antibody (HS-Ab-F647); cluster width (parallel to PM) and height (vertical to PM) were 36–243 (133±6 nm) and 63–407 nm (202±10 nm, 53 clusters, 9 cells, Figure 5B-D, Figure 5—figure supplement 2A-C). Although individual molecules spread between clusters were also observed (Figure 5B), MINFLUX-detected HS-clusters corresponded to confocally resolved HS-spots at the PH-mNeonGreen (PH_mNG)-labeled PM (Figure 5B). Virus (CoV2-S_omi or VSV-S) incubation resulted in virion colocalization with MINFLUX-detected HS-clusters (Figure 5C) or confocally resolved HS-spots (Figure 3C), indicating that SARS-CoV-2 virions dock at HS-clusters rather than individual HS molecules alone.

Figure 5 with 2 supplements see all

Download asset Open asset

Structural arrangement of heparan sulfate (HS) as the viral receptor at the cell surface.

(A) Confocal (CONF) XY-plane images of immunolabeled HS (A647-conjugated secondary antibody) at various Z-axis locations as labeled (16 cells, 4 independent experiments). (B) MINFLUX (2D) pixel-based rendering of immunolabeled HS (HS-Ab-F647, see Methods), the corresponding confocal (CONF) image of immunolabeled HS, and confocal PH_mNG (labeling PM) image at a BHK_hACE2 cell surface (extracellular [Extra] compartment is above the PM). (C) MINFLUX image of immunolabeled HS and STED image of CoV2_mCh-S_omi-A490 (Virus) at a BHK_hACE2 cell surface (merged at the bottom). Cells were incubated with CoV2_mCh-S_omi-A490 for 1 hr at 4 °C for viral cell-surface attachment. Top is the extracellular side. (D) MINFLUX coordinate-based rendering of HS clusters at the cell surface with each trace ID (TID) labeled in different colors (four different regions are shown). Extra: extracellular side; PM: plasma membrane position. (E) *Left*: MINFLUX pixel-based rendering (upper) and confocal image (lower) of immunolabeled (HS-Ab-F647) single HS molecules in vitro (on the coverslip). *Right*: MINFLUX coordinate-based rendering of single HS molecules color-coded by trace IDs (TIDs; data I, II, III, IV from left panel). (F) Distribution of HS-molecule number per HS-cluster (N_HS/cluster, 51 clusters, 8 cells, 3 independent experiments). (G) MINFLUX pixel-based rendering (upper) and confocal image (lower) of CoV2-S_omi (CoV2_mCh-S_omi-A647’s A647) at the cell surface (extracellular side: above the confocal spots). Cells were incubated with CoV2_mCh-S_omi-A647 for 1 hr at 4 °C for viral cell-surface attachment. (H) MINFLUX coordinate-based rendering of CoV2_mCh-S_omi-A647 showing three individual virions from panel G (I, II, and III). (I) SARS-CoV-2 cell-entry model: HS molecules in clusters protruding tens to hundreds of nanometers from the PM serve as the receptor for viral cell-surface attachment; endocytosis is the primary route to internalize surface-attached virions. If endocytosis co-internalizes local hACE2 in the close vicinity to the HS cluster at the cell surface, viral RNA genome can be released from endosomes for viral genome expression to infect cells.

Figure 5—source data 1 Numerical values underlying the heparan sulfate cluster dimensions, molecule number estimates, and virion size analyses shown in Figure 5.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig5-data1-v1.xlsx
Download elife-108925-fig5-data1-v1.xlsx

The mean diameter [(width +length)/2] of the cell-surface HS-Ab-F647 cluster (168±8 nm, 53 clusters, 9 cells; Figure 5—figure supplement 2A, B, see also Figure 5D) was much larger than that of single HS-Ab-F647 molecules in vitro (26±1 nm, 67 molecules, Figure 5E; Figure 5—figure supplement 2D and E), suggesting that each cell-surface HS cluster is composed of many single HS molecules. Normalizing the number of Flux 647 emission bursts (trace ID) per cell-surface cluster (73±8, from 51 clusters, Figure 5—figure supplement 2C) onto that per single molecule (2.0±0.2, 67 molecules, Figure 5—figure supplement 2) yielded 38±4 HS-molecules per cell-surface cluster (range: 6–137, 51 clusters, 8 cells, Figure 5F). The number is likely higher due to steric limitations to antibody binding with individual HS molecules in the cluster.

Importantly, confocally resolved CoV2_mCh-S_omi-A647 spots corresponded to clusters of MINFLUX-resolved CoV2_mCh-S_omi-A647’s A647 localizations (Figure 5G). Many clusters measured ~70–130 nm in mean diameter (Figure 5G and H), agreeing with SARS-CoV-2 virion diameter of ~100 nm (Jackson et al., 2022); some were larger (Figure 5G), likely representing multiple virions. These MINFLUX-resolved SARS-CoV-2 images verified our virus labeling strategy (Figure 1A).

SARS-CoV-2 JN.1 variant binds to heparan sulfate, not hACE2, in primary human airway cells

To test whether our findings with the engineered Omicron variant (CoV2_mCh-S_omi) extend to an authentic strain that infected people, we examined the JN.1 subvariant (BA.2.86.1.1) virion binding in Chinese hamster ovary (CHO) cell lines with glycosaminoglycan (GAG) deficiencies, the CHO-pgs-745 and CHO-pgsB-618 cells, in both of which HS biosynthesis is blocked by loss of xylosyltransferase-2 (XYLT2) or β–1,4-galactosyltransferase-7 (B4GALT7) activity (Esko et al., 1987; Esko et al., 1985; López-Muñoz et al., 2022). Consistent with our previous observations, Alexa 647-labeled JN.1 variant (CoV2-S_JN.1-A647) showed robust binding to wild-type CHO-K1 cells, but a ~81–84% decrease in binding in both CHO-pgsA-745 and CHO-pgsB-618 cells lacking HS (Figure 6A–B). These results indicate that HS is critical for CoV2-S_JN.1-A647 attachment to the cell surface.

Figure 6

Download asset Open asset

Attachment of SARS-CoV-2 JN.1 variant at the cell surface depends on heparan sulfate (HS).

(A) CoV2-S_JN.1-A647 attachment at the cell surface: sampled bright field and confocal images of CoV2-S_JN.1-A647’s V-A647 in wild-type CHO cells (CHO-K1), GAGs-deficient CHO-pgsA-745 cells lacking HS, and GAGs-deficient CHO-pgsB-618 cells lacking HS (images merged in the right) after incubation of CoV2-S_JN.1 for 1 hr at 4 °C for labeling cell-surface attachment (incubation protocol shown on the top). (B) Quantification of CoV2-S_JN.1-A647 attachment at the cell surface: cell-surface fluorescence of CoV2-S_JN.1-A647 (F₆₄₇) per cell (mean + s.e.m.) in CHO-K1 (117 cells, 2 repeats), CHO-pgsA-745 (93 cells, 2 repeats), and CHO-pgsB-618 cells (90 cells, 2 repeats). ***: p<0.001 (ANOVA test).

Figure 6—source data 1 Numerical values underlying the quantification of SARS-CoV-2 JN.1 attachment to CHO-K1, CHO-pgsA-745, and CHO-pgsB-618 cells shown in Figure 6B.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig6-data1-v1.xlsx
Download elife-108925-fig6-data1-v1.xlsx

To further test the clinical relevance, we pre-treated normal human bronchial/tracheal epithelial (NHBE) cells with: (1) control solution, (2) pixantrone (PIX), a drug under clinical trial for cancer treatment due to its DNA intercalating activity, which can bind HS to inhibit HS binding with proteins (Zhang et al., 2020), or (3) anti-hACE2 antibody that blocks CoV2-S_JN.1-A647 binding with ACE2 (Jackson et al., 2022). Immunostaining of nucleocapsid protein (N protein) and spike showed that PIX nearly abolished the cell-surface attachment of CoV2-S_JN.1-A647, while anti-hACE2 antibody had no effect compared to control (Figure 7A), suggesting that HS, but not ACE2, mediates authentic SARS-CoV-2 cell-surface attachment. Confocal and STED images showed the colocalization of HS and CoV2-S_JN.1-A647 (Figure 7B). Analogous to HS colocalization with VSV-S (Figure 3C) or CoV2_mCh-S_omi (Figure 5C), this result (Figure 7B) suggests that HS binds CoV2-S_JN.1-A647, mediating CoV2-S_JN.1-A647 attachment on the NHBE cell surface.

Figure 7 with 1 supplement see all

Download asset Open asset

SARS-CoV-2 JN.1 variant attachment and infection in primary human airway cells depend on heparan sulfate (HS) and are inhibited by a clinically used HS-binding agent.

(A) *Images for virus attachment*: bright field, confocal, and merged images of CoV2-S_JN.1-A647 in normal human bronchial/tracheal epithelial cells (NHBE) pre-treated with control (Ctrl), pixantrone (PIX), or anti-ACE2 antibody (ACE2 Ab). Top: incubation protocol – CoV2-S_JN.1-A647 1 hr at 4 °C for cell-surface attachment. *Bar graph for virus attachment*: cell-surface fluorescence of CoV2-S_JN.1-A647 (F₆₄₇) per cell (mean ± s.e.m.) for normal human bronchial/tracheal epithelial (NHBE) cells pre-treated with control (Ctrl, 16 cells, 2 independent experiments), pixantrone (PIX, 18 cells, 2 independent experiments), or anti-ACE2 antibody (ACE2 Ab, 24 cells, 2 independent experiments). ***: p<0.001 (ANOVA test). (B) Cell-surface-attached viruses colocalized with HS puncta: confocal images of viral spots (CoV2-S_JN.1-A647’s V-647, red) and immunolabeled HS (Alexa 488-conjugated antibody, green) at NHBE cell surface (images merged in the bottom). Top: incubation protocol – CoV2-S_JN.1-A647 1 hr at 4 °C for cell-surface attachment. (C) PIX, but not ACE2 antibody, inhibits CoV2-S_JN.1 attachment in human primary small airway epithelial cells in air-liquid interface (ALI) format. *Images*: sampled confocal images of Hoechst (gray, labeling cell nucleus), antibody-labeled nucleocapsid protein (N protein of CoV2-S_JN.1, green), and spike protein (anti-S2P6 of CoV2-S_JN.1, red) in ALI cells were incubated with CoV2-S_JN.1 for 1 hr at 4 °C (for viral cell-surface attachment, protocol shown at the top). Before virus incubation, cells were pretreated with (1) control solution (Ctrl), (2) pixantrone (PIX), or (3) anti-ACE2 antibody (ACE2 Ab). Images were projection images. *Bar graphs*: quantification of the fluorescence (mean ± s.e.m.) of immunolabeled nucleocapsid protein (F_{N protein}, upper) and spike (F_Spike, lower) for ALI cells incubated with CoV2-S_JN.1 for 1 hr at 4 °C in three pre-treated conditions stated above (Ctrl: 2 independent experiments; PIX: 2 independent experiments; ACE2 Ab: 2 independent experiments). ***: p<0.001 (ANOVA test). (D) Both PIX and ACE2 antibody inhibit CoV2-S_JN.1 infection of ALI cells. *Images*: immunolabeled N-protein images from ALI cells being incubated with viruses with a protocol for viral genome expression in three conditions, including (1) control (Ctrl), (2) pretreated with PIX (before virus incubation), and (3) pretreated with ACE2 Ab. The images are from the entire ALI culture using 20 x objective (scale bar: 1 mm). Incubation protocol (shown at the top): CoV2-S_JN.1 incubation for 1 hr at 4 °C, followed by incubation at 37 °C for 24 hr. *Bar graph*: quantification of the fluorescence (mean ± s.e.m.) of immunolabeled N protein (anti-nucleocapsid protein antibody) for ALI cells being incubated with CoV2-S_JN.1 with a protocol for viral genome expression (protocol shown in the left) in three conditions stated above (Ctrl: 3 repeats; PIX: 3 repeats; ACE2 Ab: 3 repeats). ***: p<0.001 (ANOVA test).

Figure 7—source data 1 Numerical values underlying the SARS-CoV-2 JN.1 attachment and infection analyses in primary human airway cells shown in Figure 7.: https://cdn.elifesciences.org/articles/108925/elife-108925-fig7-data1-v1.xlsx
Download elife-108925-fig7-data1-v1.xlsx

Next, we validated our findings in a physiologically relevant human model with primary human airway epithelial cells cultured at air-liquid interface (ALI) format. Similar to results observed in NHBE cells (Figure 7A and B), immunostaining for nucleocapsid protein and spike confirmed that PIX, but not anti-hACE2 antibody, substantially reduced surface-associated viral signals after the incubation protocol (4 °C for 1 hr) for virus cell-surface attachment (Figure 7C, Figure 7—figure supplement 1). Cell outlines were not clear at ALI format because cells are densely packed in many layers (Figure 7—figure supplement 1), making imaging difficult to resolve each cell’s outline and thus the virus distribution along the cell PM (Figure 7C). However, this result reflects virus cell-surface attachment because (1) the 4°C-for-1-hr incubation protocol allowed for largely virus attachment at the cell surface as shown in Figures 3 and 6, (2) inhibition of cell-surface HS-binding (Figure 6), but not hACE2, substantially reduced the virus signal after the incubation for surface attachment (Figure 7C), and (3) inhibition of either HS-binding or hACE2 substantially reduced the virus signal after an incubation protocol for virus infection (Figure 7D).

For Figure 7D, we assessed CoV2-S_JN.1 infection in primary human small airway epithelial cells cultured at ALI format, which were incubated with CoV2-S_JN.1 for 1 hr at 4 °C, followed by incubation at 37 °C for 24 hr for viral genome expression. Immunostaining for nucleocapsid protein revealed robust JN.1 infection in control cultures, whereas pre-treatment of PIX or anti-hACE2 antibody significantly reduced N protein signal (Figure 7D). This result (Figure 7D), together with the attachment data (Figures 6 and 7A-C), suggests that HS mediates initial cell-surface binding with CoV2-S_JN.1, while hACE2 contributes at a downstream step required for infection.

Discussion

A two-receptor entry model – HS for attachment/endocytosis and ACE2 for downstream

We used a variety of advanced microscopic methods, including confocal, STED, MINFLUX, and EM imaging of pseudotyped (Figures 1—3) and authentic SARS-CoV-2 (Figures 4 and 5), HS, hACE2, as well as SARS-CoV-2-containing endocytic vesicles. This revealed HS as the SARS-CoV-2 receptor underlying virus cell-surface attachment and subsequent endocytic internalization, whereas hACE2, the generally believed SARS-CoV-2 receptor (Jackson et al., 2022), is only required after virus attachment and endocytic internalization to promote viral genome expression (summarized in Figure 5I). Clusters of ~6–137 HS molecules, not single HS molecules, are the docking site for SARS-CoV-2 attachment (Figures 3 and 5). The cell surface is covered by extremely dense clusters (~1 cluster per 6 μm²) protruding from the PM by up to ~410 nm (Figure 5). This nanoscale arrangement – the tall, anionic HS-concentrated clusters densely covering the cell surface – may significantly enhance the chance that HS molecules in clusters bind multivalently to basic virion surface amino acids. We propose that HS outcompetes ACE2, which extends less than 10 nm from the PM (Yan et al., 2020), for virion attachment (Clausen et al., 2020; Walls et al., 2020; Wrapp et al., 2020; Liu et al., 2021).

Although hACE2 is not required for SARS-CoV-2 attachment and endocytosis, it is required for viral genome expression (Figure 1). It has been suggested that hACE2 binding changes S conformation to enable the protease cleavage of S needed for viral fusion (Jackson et al., 2022). Cell-surface hACE2 co-internalized with SARS-CoV-2 in endosomes (Figure 3I) may thus promote infection by changing S conformation, allowing endosomal cathepsin L to cleave S (Figure 5I). Taken together, our results suggest a SARS-CoV-2 entry model with two receptors: HS clusters serve as the docking site for virus attachment and subsequent endocytic internalization, while hACE2 promotes downstream genome expression (Figure 5I). We verified this model for authentic SARS-CoV-2 JN.1 subvariant binding on the cell-surface with HS and subsequent infection in primary human airway cells (Figures 6 and 7). Such cell surface attachment and infection were inhibited by a clinically used HS-binding agent, pixantrone (Figure 7), suggesting that interfering with HS-binding is a key therapeutic strategy for COVID-19.

Our data suggest that, under physiological conditions in which HS is present at the cell surface, hACE2 is not essential for viral cell-surface attachment or endocytosis. We do not know whether hACE2 expression alone, in the absence of HS, can support viral cell-surface attachment and endocytosis. Although our data suggest that hACE2 functions downstream of endocytosis to facilitate viral fusion at the endosome for genome delivery to the cytosol, we do not know whether hACE2 binding with the virus occurs at the cell surface or endosomes. The binding may occur in both places, but not essential for virus attachment and endocytosis. Consistent with this possibility, HS may promote spike-dependent ACE2 super-cluster assembly at the cell surface and enhance SARS-CoV-2-associated syncytium formation, suggesting that HS may organize ACE2 nanoscale architecture in a cell–cell fusion context (Zhang et al., 2023).

Our model (Figure 5I) differs in three aspects from the generally accepted current model (Jackson et al., 2022; Clausen et al., 2020; Zhang et al., 2020). First, the current model considers hACE2 essential in mediating viral attachment and initiation of endocytosis, whereas our model does not. Second, the current model considers HS as an attachment factor augmenting hACE2’s role in binding viruses for attachment and initiation of subsequent endocytosis, whereas our model considers HS as the receptor for viral attachment and subsequent endocytosis. Third, the current model considers virus-hACE2 binding and its facilitation by HS at the cell surface before endocytosis, whereas our model moves these actions downstream of endocytosis, likely in the endosomal membrane, the principal site of virion entry.

Although our SARS-CoV-2 entry model (Figure 5I) differs from the widely accepted model, our findings do not necessarily contradict previous findings underpinning the dominant model. hACE2 binding with S in vitro (Li et al., 2003; Shang et al., 2020; Lan et al., 2020) and its pivotal role in infection (Hoffmann et al., 2020) led to the current view that hACE2 is the receptor for SARS-CoV-2 cell-surface attachment and subsequent entry into cells (Jackson et al., 2022), despite the lack of direct observation of SARS-CoV-2 binding with hACE2 at the cell surfaces. Although light microscopic imaging showed S and hACE2 at the PM (Wang et al., 2020; Bayati et al., 2021), limited resolution precluded demonstrating a direct association. We clearly showed the cell-surface association of VSV-S and SARS-CoV-2 with HS, but not ACE2 (Figures 3 and 5). This finding is consistent with previous studies showing HS binding with S in vitro and inhibition of SARS-CoV-2 infection by perturbing this binding with existing drugs or heparins, despite these results leading to proposing HS as an attachment factor facilitating S binding to the cell-surface ACE2 (Jackson et al., 2022; Clausen et al., 2020; Zhang et al., 2020; Bermejo-Jambrina et al., 2021). While some studies using genome-wide CRISPR screening to identify genes involved in SARS-CoV-2 reveal genes for HS biosynthesis, others do not (Chan et al., 2023; Wang et al., 2021; Baggen et al., 2021; Rebendenne et al., 2022; Wei et al., 2021; Zhu et al., 2021). The negative result, which might reflect screen sensitivity, incomplete knockout, pathway redundancy, or viral dose/stringency, needs to be verified with specific gene knockout.

Our observation of the large size (115–789 nm) of vesicles containing one to five virions suggests that the endocytosis involved may form large vesicles. This form of endocytosis could be macropinocytosis, which can form large vesicles to engulf multiple virions. Consistent with this suggestion, a recent study showed that a macropinocytosis inhibitor, EIPA, reduces SARS-CoV-2 replication in Vero cells (Zhang et al., 2022). Our finding that viral genome (V-EGFP) expression was largely inhibited (by ~80%) or nearly abolished by inhibition of endocytosis or endosomal cathepsin L crucial for SARS-CoV-2 endosomal fusion (Figure 2) suggests endocytic internalization as the major pathway for SARS-CoV-2 genome expression. While we cannot completely rule out a minor role for viral fusion at the PM in viral genome expression, our data indicate endocytosis following HS-mediated viral cell-surface attachment as the primary pathway for SARS-CoV-2 infection in the cells we studied. For other cells not examined in the present work, if TMPRSS2 is highly expressed, we could not rule out the possibility that the fusion pathway could also be dominant.

The two-receptor model’s potential clinical relevance and broader significance

Interfering with HS binding has been suggested as a therapeutic strategy to prevent and treat many viral infections that depend on HS for entry, including COVID-19 (Koehler et al., 2020; Lebsir et al., 2023; Clausen et al., 2020; Lang et al., 2011). Supporting this strategy, disrupting Spike–HS interactions, including inhibition by heparin and related glycans, reduces SARS-CoV-2 attachment/entry (Mycroft-West et al., 2020). Clinical evaluation of inhaled/nebulized unfractionated heparin has reported improved clinical outcomes without major bleeding signals, supporting the feasibility of targeting airway-surface HS interactions (van Haren et al., 2025). HS mimetics, such as pixatimod (PG545), inhibit SARS-CoV-2 infection and exhibit greater potency than heparin in assays measuring inhibition of Spike/ACE2 engagement and viral infectivity (Guimond et al., 2022). These reports support the translational potential of therapeutically interfering with virion–HS binding. However, this strategy has not been the focus for developing methods to prevent and treat COVID-19, likely because HS is considered only a regulator that is not essential for SARS-CoV-2 entry. Our finding that HS is the attachment receptor re-emphasizes the importance of perturbing virus-HS binding, the first step of the viral entry, to efficiently block SARS-CoV-2 infection. Further supporting this view, inhibition of HS binding with a clinically used HS-binding agent, pixantrone, inhibits authentic SARS-CoV-2 JN.1 subvariant binding with HS on the cell-surface and infection in primary human airway cells (Figures 6 and 7). These results suggest a combinatorial anti-SARS-CoV-2 strategy: early HS blockade to prevent attachment combined with ACE2 targeting to inhibit post-attachment steps.

Analogous to SARS-CoV-2 (Jackson et al., 2022; Clausen et al., 2020; Zhang et al., 2020; Bermejo-Jambrina et al., 2021), without direct supporting evidence, HS has been considered a cell-surface attachment factor regulating other coronaviruses binding with their receptor hACE2 for attachment and subsequent cell entry (Lebsir et al., 2023; Cagno et al., 2019; Lang et al., 2011; Milewska et al., 2014). This includes SARS-CoV, responsible for the SARS pandemic two decades ago, and human common cold coronavirus NL63 (Lebsir et al., 2023; Cagno et al., 2019; Lang et al., 2011; Milewska et al., 2014). Our findings suggest that HS may also serve as the receptor for attachment and subsequent endocytosis of these CoVs. Perturbing virus-HS binding may be a broad anti-human CoV strategy.

Besides coronaviruses, HS is considered an attachment factor for and can bind to hepatitis C virus, HIV, Ebola virus, human papillomavirus, and herpesviruses (Koehler et al., 2020). For Ebola, HS has been proposed to deliver virions into endosomes (O’Hearn et al., 2015). Our results implicate that HS may be a common cell-surface attachment receptor rather than a regulator for these viruses. This is because the anionic HS-concentrated clusters are densely packed at the cell surface and may extend hundreds of nanometers above the PM (Figure 5) for easy access and strong binding with basic residues of approaching viruses’ surface proteins. Docking at the HS cluster may thus be a general strategy for many viruses to attach to the cell surface. Interfering with HS binding (Lebsir et al., 2023; Jackson et al., 2022; Clausen et al., 2020; Lang et al., 2011) may constitute a general antiviral strategy for not only HS/hACE2-dependent coronaviruses, but also many other HS-dependent viruses mentioned above. For SARS-CoV-2 and other HS-binding respiratory viruses, nasal spray delivery of HS competitors or heparinase may be an inexpensive but effective general antiviral strategy.

Methods

Cell lines, primary cells, and air-liquid interface airway epithelial cell culture

We used baby hamster kidney fibroblast BHK21 (BHK, ATCC, #CCL-10), human embryonic kidney 293T (293T, Thermo Fisher Scientific, R70007), African green monkey kidney Vero E6 (Vero, ATCC, #C1008 [clone E6]), HeLa (ATCC, #CCL-2), and human lung fibroblast MRC5 (ATCC, #CCL-171) cells. We generated BHK21 cells stably expressing human ACE2 (BHK_hACE2) and 293T cells stably expressing human ACE2 (293T_hACE2) using the Sleeping Beauty transposon plasmid expression system (Kowarz et al., 2015). All cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, 11885084) supplemented with 10% FBS (Gibco, 10082147) and incubated at 37 °C and 5% CO₂ in a humidified incubator. Wild-type CHO-K1 (ATCC, #CCL-61) and GAG-deficient mutants CHO-pgsA-745 (ATCC, #CCL-2242) and CHO-pgsB-618 (ATCC, #CCL-2241) were maintained in F-12K medium (Gibco, 21127022) supplemented with 10% FBS and 1% MEM NEAA (Gibco, 11140050) and incubated at 37 °C and 5% CO₂ in a humidified incubator.

We generated 293T cells stably expressing T7 polymerase and VSVG (293T-T7-VSVG) for VSV virion production and Vero cells stably expressing human transmembrane serine protease 2 (Vero-TMPRSS2) for authentic SARS-CoV-2 virion production, using the Sleeping Beauty transposon plasmid expression system (Kowarz et al., 2015). Cell lines were confirmed to be mycoplasma-free using MycoStrip detection kit (InvivoGen, rep-mys-50).

Normal human bronchial/tracheal epithelial cells (Lonza, CC-2540) were expanded in PneumaCult-Ex Plus Basal medium (Stemcell, #05041) in flasks and incubated at 37 °C and 5% CO₂ in a humidified incubator. Differentiated primary human airway epithelial cultures were purchased from Epithelix (MucilAir) and maintained per the manufacturer’s instructions. Upon receipt, inserts were equilibrated for 24–48 hr at 37 °C, 5% CO₂ with MucilAir medium (MA medium, Epithelix) in the basolateral chamber (typical volume 500–600 µL per insert), with no liquid on the apical side. MA medium was replaced two to three times per week. Before assays, the apical surface was gently rinsed twice with warm DPBS to remove accumulated mucus. Only well-differentiated tissues with robust ciliary beating and intact morphology were used.

Generation of pseudo-typed SARS-CoV-2 virions: VSV_EGFP-S-A647, VSV_EGFP-S-A490, VSV_EGFP-S_Omi-A647, and Lenti_EGFP-S_Omi-A647

Generation of recombinant VSV (containing EGFP-coding RNA in the genome) pseudo-typed with SARS-CoV-2’s wild-type S-protein (S) of Wuhan-hu-1 strain with two modifications (VSV_EGFP-S), the C-terminal 21 amino acids deletion and a furin cleavage site mutation R685S, was performed as previously described (Ye et al., 2020). Its use in tissue culture at biosafety level 2 was approved by the Institutional Biosafety Committee (IBC) at The National Institute of Allergy and Infectious Diseases (NIAID). Plasmid-based rescue of VSV_EGFP-S was carried out as described previously with some modifications. Briefly, 293T-T7-VSVG cells in a six-well plate were transfected with the VSV_EGFP-S antigenome plasmid, along with plasmids expressing codon-optimized T7 polymerase (T7opt in pCAGGS, Addgene, 65974), and VSV N, P, L, and G-expressing plasmids (Addgene, 64087, 64088, 64085, and 8454), using TransIT-LT1 transfection reagent. The medium was exchanged with fresh DMEM containing doxycycline (1 μg/ml, for VSVG expression). EGFP expression was monitored every day for 2–3 days. Transfected cells with EGFP-positive clusters and typical cytopathic effect (CPE) were collected as passage 0 and then inoculated to fresh BHK_hACE2 cells to propagate as viral stock P1, which underwent further propagations to generate the VSV_EGFP-S virions in Vero or BHK_hACE2 cells. Viral passages were subjected to Sanger and next-generation sequencing to verify the S gene or the viral genome.

VSV_EGFP-S virus was amplified in Vero or BHK_hACE2 cells. The amplified VSV_EGFP-S virus was either concentrated with an ultrafast centrifuge or Lenti-X concentrator (Takara, #631232) according to the manufacturer’s protocol. The protein level of VSV_EGFP-S virus was ~6 μg/μl, which was measured with a NanoDrop spectrophotometer (ND-1000).

Purified VSV_EGFP-S was incubated with Alexa Fluor 647 NHS ester (A647, 7.7 μM, Invitrogen, #A20006) with gentle agitation for ~1.5 hr at 22–24°C, which led to A647 conjugation with primary amines of superficial envelope proteins of virus (VSV_EGFP-S-A647). Free A647 dyes were removed with 7 k molecular weight cutoff desalting columns (7 k MWCO, Thermo Fisher, #89882). Viruses were aliquoted and stored at –80 °C for further experiments. The resulting virion, VSV_EGFP-S-A647, could be observed with fluorescent imaging of A647 (Figure 1A).

In some experiments, we replaced wild-type S protein in VSV_EGFP-S-A647 with the Omicron variant to produce the VSV_EGFP-S_Omi-A647 virion; we replaced A647 in VSV_EGFP-S-A647 with Atto 490LS NHS ester (A490, 63 μM, Atto-Tec GmbH, #AD 490LS-25) to produce the VSV_EGFP-S-A490 virion.

Lentivirus (with EGFP-coding sequence in the genome) pseudo-typed with S protein Omicron variant (Lenti_EGFP-S_Omi) was purchased from BPS Bioscience (#78624–2). Lenti_EGFP-S_Omi was incubated with A647 (7.7 μM) for ~1.5 hr at 22–24°C, and purified with desalting columns, which led to A647 conjugation with primary amines of superficial envelope proteins. The resulting virion was termed Lenti_EGFP-S_Omi-A647.

Generation of recombinant SARS-CoV-2 Omicron variant

The generation of recombinant SARS-CoV-2 Omicron variant (CoV2-S_Omi) and its use in tissue culture at biosafety level 3 were approved by the IBC and the Dual Use Research of Concern Institutional Review Entity (DURC-IRE) at NIAID. Genetic modification inserted a mCherry gene to the nucleoprotein via a 2 A linker, derived from a bacterial artificial chromosome-based SARS-CoV-2 reverse genetic system (referred to as CoV2_mCh-S_omi; Ye et al., 2020). Virus rescue experiments were performed as previously described (Ye et al., 2020). Briefly, BHK_ACE2 cells were transfected with pBAC-SARS-CoV-2 and media were changed with DMEM containing 2% FBS 6 hr later. mCherry-positive cells were detached and collected along with supernatant during 48–72 hr and stored at –80 °C as the first stock. The stock was centrifuged to remove cell debris and used to infect fresh Vero-TMPRSS2 cells for 48–72 hr. The supernatant was collected and stored at –80 °C as the second stock. After the rescued virus in the second stock was confirmed by Sanger sequencing, the virus in the first stock was allowed to undergo two rounds of propagation to generate the CoV2_mCh-S_Omi virions. The viral genome sequence was verified by next-generation sequencing before virus labeling with fluorescent dyes.

Purified CoV2_mCh-S_omi was incubated with Alexa Fluor 647 NHS ester (A647, 7.7 μM, Invitrogen, #A20006) for ~1.5 hr at 22–24°C, which led to A647 conjugation with primary amines of superficial envelope proteins. The resulting virion, CoV2_mCh-S_omi-A647, was purified with desalting columns (7 k MWCO, Thermo Fisher, #89882). CoV2_mCh-S_omi-A647 could be observed with confocal, STED, and MINFLUX imaging of A647 (Figures 4 and 5). In some experiments, we replaced A647 in CoV2_mCh-S_omi-A647 with Atto 490LS NHS ester (A490, Atto-Tec GmbH, #AD 490LS) to produce CoV2_mCh-S_omi-A490.

Preparation of SARS-CoV-2 JN.1 variant and fluorescent labeling

A clinically isolated JN.1 subvariant (BA.2.86.1.1) of SARS-CoV-2 (CoV2-S_JN.1) was propagated under BSL-3 conditions. Purified CoV2-S_JN.1 was incubated with Alexa Fluor 647 NHS ester (A647, 7.7 μM, Invitrogen, #A20006) for ~1.5 hr at 22–24°C, which led to A647 conjugation with primary amines of superficial envelope proteins. The resulting virion, CoV2-S_JN.1-A647, was purified with desalting columns (7 k MWCO, Thermo Fisher, #89882). For some binding and infection assays, unlabeled JN.1 stocks were used.

Plasmids, transfection, cell plating, and fluorescent dyes

The hACE2-TagBFP plasmid was purchased from Addgene (#164219). The mutant Dynamin 2-K44A-mCherry plasmid was created from Dynamin 2-mCherry (Addgene, #27689). The PH_mNG (phospholipase C delta PH domain attached with mNeonGreen) construct was created by replacing the EGFP tag of PH-EGFP (obtained from Dr. Tamas Balla) with mNeonGreen (Allele Biotechnology; Shaner et al., 2013).

Cells were transfected with Lipofectamine LTX reagent (Thermo Fisher Scientific, #A12621) according to the manufacturer’s protocol and plated on the collagen-coated glass bottom of dishes (MatTek, #P35GCol-1.5–14 C). Images were taken 48 hr after transfection.

We incubated cells with 50 μM A490 for 1 hr to identify internalized vesicles with confocal imaging. We incubated cells with 500–750 μM A490 for 3 min for the measurement of the internalized vesicle size with STED microscopy.

Immunostaining for confocal or STED imaging

Cells were fixed with 4% paraformaldehyde, blocked with 2% BSA in PBS solution for 1 hr at 22–24°C, and subsequently incubated with primary and secondary antibodies. Primary antibodies were diluted in PBS containing 2% BSA and incubated with cells at 4 °C for 16–20 hr. After three times wash in PBS, cells were incubated with fluorescence-conjugated secondary antibodies for 1 hr at 22–24°C. After washing three times with PBS, cells were mounted on confocal and STED microscopes to acquire images. Primary antibodies used here included goat anti-ACE2 (1:500, R&D, #AF933), mouse anti-HS (1:500, Amsbio, 10E4), human anti-S2P6 (1:500), and rabbit anti-nucleocapsid (1:500, GeneTex, GTX135357); secondary antibodies included Alexa Fluor 647-labeled donkey anti-goat IgG (Thermo Fisher, #A21447) and Alexa Fluor 568-labeled goat anti-human IgG (Thermo Fisher, #A21090), Alexa Fluor 488-labeled goat anti-mouse IgG (Thermo Fisher, #A11029), Alexa Fluor 488-labeled goat anti-rabbit IgG (Thermo Fisher, #A11028), Alexa Fluor 488-labeled donkey anti-mouse IgM (Jackson, #715-545-140).

Confocal imaging

Confocal images were acquired with Leica TCS SP8 inverted confocal microscope (Leica, Germany) equipped with a×63/1.40 oil immersion objective. Alexa 647 (A647), EGFP, A490, and mCherry were excited by solid-state lasers with wavelengths of 638 nm (30 mW output), 488 nm (20 mW output), 514 nm (20 mW output), and 552 nm (20 mW output) respectively, and their fluorescence signals were collected with Leica HyD detectors at 650–795 nm, 500–550 nm, 560–630 nm, and 650–795 nm, respectively. To visualize the viruses labeled by A647 or A490, 638 nm or 514 nm laser was set at 0.2–2% or 5% of the maximum power; to detect virus EGFP or mCherry expression, 488 nm or 552 nm excitation laser was set at 0.5–2% or 50% of the maximum power, respectively.

To measure virus attachment or endocytosis, or virus genome (EGFP/mCherry) expression, XY-plane scanning along the Z-axis every 1 μm was performed (XY/Z_stack scanning), and the projected fluorescence was quantified with Fiji software for each cell to quantify all the particles at cell surface or inside the cell, or the total fluorescence.

To measure HS puncta number at cell surface or total area of cell surface, XY/Z_stack scanning with a Z-axis interval of 0.5–1 µm was performed to obtain the images of the entire cell. 3D reconstruction was generated and analyzed using Spots or Surfaces function with Imaris software (Oxford Instruments).

STED imaging

The STED images were acquired with XY-plane scanning mode at a fixed Z-axis focal plane. STED images were acquired with the Abberior Expertline system based on an inverted Olympus IX83 microscope that is equipped with a UPlanSApo 100x1.4 NA oil immersion objective. To measure virus (VSV_EGFP-S-A647 or CoV2_mCherry-S_Omi-A647) endocytosis with bath A490 (Figures 2A-D and 4E), A647, and A490 were excited at 640 nm and 512 nm respectively, and both collected at 650–765 nm. STED depletion was conducted with a 775 nm depletion beam at 5–10% of the maximum depletion power. XY/Z_stack imaging with a Z-axis interval of 200 nm was performed.

To measure viral colocalization with HS and hACE2 on the cell surface, cells were incubated with VSV_EGFP-S-A490 at 4 °C for 1 hr, fixed with 4% paraformaldehyde, and stained with primary and secondary antibodies. For STED imaging of VSV_EGFP-S-A490 and HS (Figure 4C), mouse anti-HS (1:500, Amsbio, 10E4) and Alexa fluor 488-labeled donkey anti-mouse IgM (Jackson, #715-545-140) were used to label HS. We first acquired VSV_EGFP-S-A490 STED images with a 775 nm depletion laser together with HS-Alexa 488 (A488) confocal images, and then both STED and confocal images of HS-Alexa 488 were acquired with a 595 nm depletion laser. VSV_EGFP-S-A490 and HS-A488 were excited by 640 nm and 512 nm lasers, respectively, and their fluorescence signals were collected at 650–765 nm and 528–585 nm, respectively. The power of the 640 nm laser was set at 3–5% of the maximum (maximum power: 1 mW); the corresponding 775 nm depletion laser power was set at 5–10% of the maximum (maximum power: 3 W). The power of the 512 nm laser was set at 3–5% of the maximum (maximum power: 1 mW); the corresponding 595 nm depletion laser power was set at 5–10% of the maximum (maximum power: 700 mW).

For imaging of VSV_EGFP-S-A490 and hACE2 (Figure 4H–I), goat anti-ACE2 (1:500, R&D, #AF933) and Alexa fluor 647-labeled donkey anti-goat IgG (Thermo Fisher, #A-21447) were used to label hACE2. We acquired VSV_EGFP-S-A490 and ACE2-Alexa 647 (A647) STED images with a 775 nm depletion laser. VSV_EGFP-S-A490 and ACE2-A647 were sequentially excited by 512 nm and 640 nm lasers, respectively, and their fluorescence signals were sequentially collected at 650–765 nm. The power of the 512 nm and 640 nm lasers was set at 3–5% of the maximum, and the 775 nm depletion laser power was set at 5–10% of the maximum.

STED images were deconvolved using Huygens software (Scientific Volume Imaging) as described previously (Guo et al., 2022) and analyzed with Image J and LAS X (Leica).

Electron microscopy

BHK_hACE2 cells were incubated with VSV_EGFP-S at 4 °C for 1 hr, then transferred to 37 °C for 15 min (to track VSV_EGFP-S in intracellular vesicles, Figure 2E). BHK_hACE2 cells were fixed with 4% glutaraldehyde (freshly prepared, Electron microscopy sciences, Hatfield, PA) in 0.1 N Na-cacodylate buffer for 1 hr at 22–24°C, and stored in a 4 °C refrigerator till processing. Cells were washed with 0.1 N cacodylate buffer and treated with 1% OsO4 in cacodylate buffer at 4 °C for 1 hr, then with 0.25% uranyl acetate (pH 5.0) at 4 °C for 16–20 hr. Next, cells were dehydrated with ethanol and embedded in epoxy resin. Thin (70 nm) sections were counterstained with uranyl acetate and lead citrate and then imaged in a JEOL 200 CX TEM. Images were captured with a CCD digital camera system (XR-100 from AMT, Danvers, MA) at a primary magnification of ×10,000–30,000.

Reagents and their application procedures

Dynasore (DnS, 80 μM, Millipore Sigma; Macia et al., 2006; Zhao et al., 2016) was applied to the cell media 30 min before the virus was applied and was continuously present for another 1 hr during which the virus was added to the cell media (see Figure 2F and G for the viral incubation protocol). Cathepsin inhibitor 1 (CPS Inh, 10 μM, Selleckchem, S2847; Asaad et al., 2009; Huang et al., 2006; Simmons et al., 2005; Zhao et al., 2021) was applied to the cell media 30 min before the virus was applied; CPS Inh was continuously present for another 1 hr (Figure 2H, left) during which the virus was added to the cell media or for another 24 hr (Figure 2H–I, right) during which the virus was added for 1 hr and then washed out.

HPRase (López-Muñoz et al., 2023), the heparinase I/II/III mixture containing heparinase I (3.2 units, NEB, #P0735S), heparinase II (1 unit, NEB, #P0736S), and heparinase III (0.19 units, NEB, #P0737S), was applied to cells in FBS-free DMEM at 30 °C for 1 hr before the virus was applied. After washing out, cells were either fixed for immunostaining of HS (Figure 3D) and hACE2 or incubated with viruses at 4 °C for 1 hr. Virus attachment was measured after 1 hr incubation at 4 °C (Figures 3E and 4E). Cells were then put back to 37 °C incubator for another 1 hr to measure the virus uptake (Figure 3F) or for another 24 hr to measure the EGFP gene expression (Figure 3G).

Pixantrone (PIX, 20 µM, Selleck), a HS-binding agent under clinical trial (Ye et al., 2020), or anti-human ACE2 monoclonal antibody (ACE2 Ab, 20 µg/ml, R&D, #AF933) was applied to the cells 30 min before the virus was applied and was continuously present for another 1 hr during which the virus was added to the cell media (see Figure 7A and C for the viral incubation protocol). For ALI tissues, excess apical mucus was gently removed with warm PBS (1–2 rinses) prior to incubation. Plates with inserts were pre-chilled and PIX or ACE2 Ab was applied to the apical surface and maintained during attachment or infection. After incubation, cells were washed with ice-cold medium to remove unbound virions. Inserts were immediately fixed with 4% PFA for immunostaining for binding assay (Figure 7C) or changed to culture at 37 °C for 24 hr, then fixed with 4% PFA for immunostaining for infection assay (Figure 7D).

Flow cytometry

To measure viral attachment, BHK_hACE2 or BHK cells were incubated with or without VSV_EGFP-S-A647 at 4 °C for 1 hr, washed three times with PBS, incubated with ice-cold DPBS (Gibco) supplemented with EDTA solution (5 mM, Quality Biologicals) for 30 min. Cells were rigorously pipetted to generate single-cell suspensions (Figure 3—figure supplement 1). To measure viral infection on linked samples, BHK_hACE2 or BHK cells were then incubated with or without VSV_EGFP-S-A647 at 37 °C for 24 hr. After incubation, cells were washed with PBS three times and resuspended in DPBS (Gibco) supplemented with BSA (0.1%; Figure 1—figure supplement 2). Flow cytometry was performed on BD FACSCelesta Cell Analyzer with BD FACSDiva software version 9.0. Initial data processing was done with FlowJo version 10.10. For all flow cytometry experiments, we did not include trypsin for digestion to avoid the potential impact of trypsin in cleaving proteins attached to HS (dela Paz et al., 2014).

Verification of VSV_EGFP-S-A647 in vitro

To visualize VSV_EGFP-S (Figure 1A), a glass-bottomed cell culture dish was coated with lectin (Sigma) at 37 °C for 1 hr, and then 10 µl VSV_EGFP-S-A647 was spread onto glass and followed by fixation with 2% paraformaldehyde at 22–24°C for 30 min. After washing, 2% BSA/PBS was used for block and antibody staining. VSV_EGFP-S-A647 were incubated with anti-spike antibody (1A9, 5 µg/ml, Cat. No. GTX632604, GeneTex) targeting S2 subunit at 4 °C for 16–20 hr, followed by wash and 1 hr of incubation with Alexa Fluor 488-labeled donkey anti-mouse IgG (Abcam). After the final wash, images (e.g. Figure 1A) were acquired under a Leica TCS SP8 inverted confocal microscope (Leica, Germany) equipped with a×63/1.40 oil immersion objective.

MINFLUX nanoscopy

MINFLUX data acquisition

MINFLUX imaging was performed on a commercial 3D MINFLUX microscope that was operated by the Imspector software with MINFLUX drivers (Abberior Instruments; Schmidt et al., 2021). Generally, fields of view with multiple gold beads were chosen and locked in a 3D set position for active sample stabilization using the near-infrared scattering from gold beads and active feedback correction via the piezo stage. It was ensured that the standard deviation of the sample position relative to the stabilization set point was less than 2 nm in all directions during measurements. Fields of view were chosen close to the coverslip surface at the bottom of cells. Before starting the MINFLUX data acquisition, the fluorophores were driven into the dark state (cyanine-thiol adduct) using iterative confocal scans with the 640 nm excitation laser and a power between 8–15% (maximum power at periscope: 1.94 mW). The sample was imaged with the standard MINFLUX imaging sequence provided by the manufacturer using 10% fixed laser power. During the MINFLUX measurement, the 405 nm activation laser power was ramped up slowly from 0% to 50% over several hours (maximum power at periscope: 27 μW). Samples were usually imaged for ~2–6 hr.

MINFLUX data analysis

The raw final valid molecule position estimates were exported directly from the MINFLUX Imspector interface as a .mat file. Custom MATLAB analysis software was then used to identify and segregate clusters of localizations. The data were filtered to remove trace IDs (TIDs, groups of localizations originating from the same fluorophore emission burst) with a standard deviationSD of more than 10 nm and less than three localizations per trace (Mulhall et al., 2023; Pape et al., 2020).

For the estimate of the copy number of HS labeled with HS-antibody-A647, the TIDs per purified HS molecule in vitro were counted and used to normalize the number of TIDs per HS cluster at the cell-surface, assuming an unchanged A647 emission behavior between measurements in vitro and on the cell surface. Here, we did not count the molecules directly using the number of TIDs per cell-surface HS cluster, due to the reversibility of A647 photoswitching which would lead to overcounting. The diameter of each cluster of localizations was estimated by averaging the height and the width of an ellipse fit [(height +width)/2, Figure 5—figure supplement 2A and D]. The diameter may refer to localizations for a single HS molecule in vitro, a cell-surface HS cluster, or a virion.

Sample preparation for MINFLUX imaging

Preparing heparan sulfate on coverslips

For the attachment of HS to coverslips, 400 µl HS (Sigma-Aldrich, H7640) solution (1 µg/µl) was added to the glass-bottom dishes coated with collagen (MatTek, P35GCol-1.5–14 C) and allowed to stay at 4 ^oC for 16–20 hr to adhere to the glass surface. After washing with PBS, the HS on coverslips was blocked with 2% BSA in PBS for 1 hr at 22-24 ^oC, followed by 1 hr incubation of mouse anti-HS (1:500, Amsbio, 10E4). Dishes with HS were washed with PBS and incubated with a secondary goat anti-mouse IgG antibody conjugated with Flux 647 (1:500, Abberior) for 1 hr. The HS dishes were ready for the next step after washing three times with PBS.

Immediately prior to the confocal and MINFLUX imaging, the dish with HS on the glass bottom was incubated with 200 µl undiluted gold bead solution (Gold nanoparticles 150 nm, BBI solutions, SKU:EM.GC150) for 5 min. After rinsing off the gold bead solution, the dish was washed with PBS several times. An isotonic imaging buffer containing 200 mM Tris-HCl (pH 8.0), 10 mM NaCl, 10% (w/v) glucose, 63 µg/ml bovine liver catalase (Sigma-Aldrich, C1345), 393 µg/ml glucose oxidase from Aspergillus niger (type VII, Sigma-Aldrich, G2133), and 12.6 mM cysteamine (Sigma-Aldrich, 30070) was added to the glass bottom of the dish. A 25 mm coverslip (Neuvitro) was placed above the glass bottom to generate a reaction cavity; excessive imaging buffer was removed to seal the cavity before mounting the dish onto the microscope. The coverslips (above the glass bottom dishes) were sealed with Elite double 22 dental epoxy (Zhermack).

Preparing cells for MINFLUX imaging

BHK_hACE2 cells were transfected with PH_mNG plasmid and plated on glass bottom dishes (MatTek, P35GCol-1.5–14 C). The transfected cells were allowed to express PH_mNG for 24–48 hr. For MINFLUX imaging of HS (Figure 5B–D), cells were incubated with CoV2_mCh-S_Omi-A490 at 4 ^oC for 1 hr (for confocal imaging), fixed with 4% paraformaldehyde, and then stained with mouse anti-HS antibody and Flux 647-conjugated goat anti-mouse secondary antibody (HS-Ab-F647, for MINFLUX imaging). For MINFLUX imaging of CoV2_mCh-S_Omi-A647 (Figure 5G–H), cells were incubated with CoV2_mCh-S_Omi-A647 at 4 ^oC for 1 hr, and then fixed with 4% paraformaldehyde. Immediately prior to MINFLUX imaging, the dish was incubated with 200 µl undiluted gold bead solution (Gold nanoparticles 150 nm, BBI solutions, SKU:EM.GC150) for 5 min. After rinsing off the gold bead solution, the dish was washed with PBS several times. An isotonic imaging buffer containing 200 mM Tris-HCl (pH 8.0), 10 mM NaCl, 10% (w/v) glucose, 63 µg/ml bovine liver catalase (Sigma-Aldrich, C1345), 393 µg/ml glucose oxidase from Aspergillus niger (type VII, Sigma-Aldrich, G2133), and 12.6 mM cysteamine (Sigma-Aldrich, 30070) was added to the glass bottom to generate a reaction cavity. The coverslips (above the glass bottom dishes) were sealed with Elite double 22 dental epoxy (Zhermack).

Statistics

Data were presented as mean ± s.e.m. Replicates are indicated in results and figure legends. N represents the number of cells, A490-labeled vesicle spots, endosomes, viral spots, ACE2 or HS spots, HS clusters, or experiments as indicated in results and figure legends. The statistical significance was tested with the unpaired Student’s t-test.

Data availability

All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Information.

References

1. Antonny B
2. Burd C
3. De Camilli P
4. Chen E
5. Daumke O
6. Faelber K
7. Ford M
8. Frolov VA
9. Frost A
10. Hinshaw JE
11. Kirchhausen T
12. Kozlov MM
13. Lenz M
14. Low HH
15. McMahon H
16. Merrifield C
17. Pollard TD
18. Robinson PJ
19. Roux A
20. Schmid S
(2016) Membrane fission by dynamin: what we know and what we need to know
The EMBO Journal 35:2270–2284.

https://doi.org/10.15252/embj.201694613
- PubMed
- Google Scholar
1. Asaad N
2. Bethel PA
3. Coulson MD
4. Dawson JE
5. Ford SJ
6. Gerhardt S
7. Grist M
8. Hamlin GA
9. James MJ
10. Jones EV
11. Karoutchi GI
12. Kenny PW
13. Morley AD
14. Oldham K
15. Rankine N
16. Ryan D
17. Wells SL
18. Wood L
19. Augustin M
20. Krapp S
21. Simader H
22. Steinbacher S
(2009) Dipeptidyl nitrile inhibitors of Cathepsin L
Bioorganic & Medicinal Chemistry Letters 19:4280–4283.

https://doi.org/10.1016/j.bmcl.2009.05.071
- Google Scholar
1. Baggen J
2. Persoons L
3. Vanstreels E
4. Jansen S
5. Van Looveren D
6. Boeckx B
7. Geudens V
8. De Man J
9. Jochmans D
10. Wauters J
11. Wauters E
12. Vanaudenaerde BM
13. Lambrechts D
14. Neyts J
15. Dallmeier K
16. Thibaut HJ
17. Jacquemyn M
18. Maes P
19. Daelemans D
(2021) Genome-wide CRISPR screening identifies TMEM106B as a proviral host factor for SARS-CoV-2
Nature Genetics 53:435–444.

https://doi.org/10.1038/s41588-021-00805-2
- PubMed
- Google Scholar
(2021) SARS-CoV-2 infects cells after viral entry via clathrin-mediated endocytosis
The Journal of Biological Chemistry 296:100306.

https://doi.org/10.1016/j.jbc.2021.100306
- PubMed
- Google Scholar
(2021) Infection and transmission of SARS-CoV-2 depend on heparan sulfate proteoglycans
The EMBO Journal 40:e106765.

https://doi.org/10.15252/embj.2020106765
- PubMed
- Google Scholar
(2019) Heparan sulfate proteoglycans and viral attachment: true receptors or adaptation bias?
Viruses 11:596.

https://doi.org/10.3390/v11070596
- PubMed
- Google Scholar
1. Case JB
2. Rothlauf PW
3. Chen RE
4. Kafai NM
5. Fox JM
6. Smith BK
7. Shrihari S
8. McCune BT
9. Harvey IB
10. Keeler SP
11. Bloyet LM
12. Zhao H
13. Ma M
14. Adams LJ
15. Winkler ES
16. Holtzman MJ
17. Fremont DH
18. Whelan SPJ
19. Diamond MS
(2020) Replication-competent vesicular stomatitis virus vaccine vector protects against SARS-CoV-2-mediated pathogenesis in mice
Cell Host & Microbe 28:465–474.

https://doi.org/10.1016/j.chom.2020.07.018
- PubMed
- Google Scholar
1. Chan K
2. Farias AG
3. Lee H
4. Guvenc F
5. Mero P
6. Brown KR
7. Ward H
8. Billmann M
9. Aulakh K
10. Astori A
11. Haider S
12. Marcon E
13. Braunschweig U
14. Pu S
15. Habsid A
16. Yan Tong AH
17. Christie-Holmes N
18. Budylowski P
19. Ghalami A
20. Mubareka S
21. Maguire F
22. Banerjee A
23. Mossman KL
24. Greenblatt J
25. Gray-Owen SD
26. Raught B
27. Blencowe BJ
28. Taipale M
29. Myers C
30. Moffat J
(2023) Survival-based CRISPR genetic screens across a panel of permissive cell lines identify common and cell-specific SARS-CoV-2 host factors
Heliyon 9:e12744.

https://doi.org/10.1016/j.heliyon.2022.e12744
- PubMed
- Google Scholar
(2019) The synaptic vesicle cycle revisited: new insights into the modes and mechanisms
The Journal of Neuroscience 39:8209–8216.

https://doi.org/10.1523/JNEUROSCI.1158-19.2019
- PubMed
- Google Scholar
1. Clausen TM
2. Sandoval DR
3. Spliid CB
4. Pihl J
5. Perrett HR
6. Painter CD
7. Narayanan A
8. Majowicz SA
9. Kwong EM
10. McVicar RN
11. Thacker BE
12. Glass CA
13. Yang Z
14. Torres JL
15. Golden GJ
16. Bartels PL
17. Porell RN
18. Garretson AF
19. Laubach L
20. Feldman J
21. Yin X
22. Pu Y
23. Hauser BM
24. Caradonna TM
25. Kellman BP
26. Martino C
27. Gordts PLSM
28. Chanda SK
29. Schmidt AG
30. Godula K
31. Leibel SL
32. Jose J
33. Corbett KD
34. Ward AB
35. Carlin AF
36. Esko JD
(2020) SARS-CoV-2 infection depends on cellular heparan sulfate and ACE2
Cell 183:1043–1057.

https://doi.org/10.1016/j.cell.2020.09.033
- PubMed
- Google Scholar
(2009) Vesicular stomatitis virus enters cells through vesicles incompletely coated with clathrin that depend upon actin for internalization
PLOS Pathogens 5:e1000394.

https://doi.org/10.1371/journal.ppat.1000394
- PubMed
- Google Scholar
(2010) The length of vesicular stomatitis virus particles dictates a need for actin assembly during clathrin-dependent endocytosis
PLOS Pathogens 6:e1001127.

https://doi.org/10.1371/journal.ppat.1001127
- PubMed
- Google Scholar
(2014) Heparan sulfates mediate the interaction between platelet endothelial cell adhesion molecule-1 (PECAM-1) and the Gαq/11 subunits of heterotrimeric G proteins
The Journal of Biological Chemistry 289:7413–7424.

https://doi.org/10.1074/jbc.M113.542514
- PubMed
- Google Scholar
1. Dieterle ME
2. Haslwanter D
3. Bortz RH III
4. Wirchnianski AS
5. Lasso G
6. Vergnolle O
7. Abbasi SA
8. Fels JM
9. Laudermilch E
10. Florez C
11. Mengotto A
12. Kimmel D
13. Malonis RJ
14. Georgiev G
15. Quiroz J
16. Barnhill J
17. Pirofski L
18. Daily JP
19. Dye JM
20. Lai JR
21. Herbert AS
22. Chandran K
23. Jangra RK
(2020) A Replication-competent vesicular stomatitis virus for studies of SARS-CoV-2 spike-mediated cell entry and its inhibition
Cell Host & Microbe 28:486–496.

https://doi.org/10.1016/j.chom.2020.06.020
- Google Scholar
(1985) Animal cell mutants defective in glycosaminoglycan biosynthesis
PNAS 82:3197–3201.

https://doi.org/10.1073/pnas.82.10.3197
- PubMed
- Google Scholar
1. Esko JD
2. Weinke JL
3. Taylor WH
4. Ekborg G
5. Rodén L
6. Anantharamaiah G
7. Gawish A
(1987)
Inhibition of chondroitin and heparan sulfate biosynthesis in Chinese hamster ovary cell mutants defective in galactosyltransferase I

The Journal of Biological Chemistry 262:12189–12195.
- PubMed
- Google Scholar
1. Guimond SE
2. Mycroft-West CJ
3. Gandhi NS
4. Tree JA
5. Le TT
6. Spalluto CM
7. Humbert MV
8. Buttigieg KR
9. Coombes N
10. Elmore MJ
11. Wand M
12. Nyström K
13. Said J
14. Setoh YX
15. Amarilla AA
16. Modhiran N
17. Sng JDJ
18. Chhabra M
19. Young PR
20. Rawle DJ
21. Lima MA
22. Yates EA
23. Karlsson R
24. Miller RL
25. Chen Y-H
26. Bagdonaite I
27. Yang Z
28. Stewart J
29. Nguyen D
30. Laidlaw S
31. Hammond E
32. Dredge K
33. Wilkinson TMA
34. Watterson D
35. Khromykh AA
36. Suhrbier A
37. Carroll MW
38. Trybala E
39. Bergström T
40. Ferro V
41. Skidmore MA
42. Turnbull JE
(2022) Synthetic heparan sulfate mimetic pixatimod (PG545) potently inhibits SARS-CoV-2 by disrupting the spike-ACE2 interaction
ACS Central Science 8:527–545.

https://doi.org/10.1021/acscentsci.1c01293
- PubMed
- Google Scholar
1. Guo X
2. Han S
3. Wei L
4. Arpino G
5. Shin W
6. Wang X
7. Wu LG
(2022) Real-time visualization of exo- and endocytosis membrane dynamics with confocal and super-resolution microscopy
STAR Protocols 3:101404.

https://doi.org/10.1016/j.xpro.2022.101404
- PubMed
- Google Scholar
1. Gwosch KC
2. Pape JK
3. Balzarotti F
4. Hoess P
5. Ellenberg J
6. Ries J
7. Hell SW
(2020) MINFLUX nanoscopy delivers 3D multicolor nanometer resolution in cells
Nature Methods 17:217–224.

https://doi.org/10.1038/s41592-019-0688-0
- PubMed
- Google Scholar
1. Hoffmann M
2. Kleine-Weber H
3. Schroeder S
4. Krüger N
5. Herrler T
6. Erichsen S
7. Schiergens TS
8. Herrler G
9. Wu NH
10. Nitsche A
11. Müller MA
12. Drosten C
13. Pöhlmann S
(2020) SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by a clinically proven protease inhibitor
Cell 181:271–280.

https://doi.org/10.1016/j.cell.2020.02.052
- PubMed
- Google Scholar
1. Huang I-C
2. Bosch BJ
3. Li F
4. Li W
5. Lee KH
6. Ghiran S
7. Vasilieva N
8. Dermody TS
9. Harrison SC
10. Dormitzer PR
11. Farzan M
12. Rottier PJM
13. Choe H
(2006) SARS coronavirus, but not human coronavirus NL63, utilizes cathepsin L to infect ACE2-expressing cells
The Journal of Biological Chemistry 281:3198–3203.

https://doi.org/10.1074/jbc.M508381200
- PubMed
- Google Scholar
1. Jackson CB
2. Farzan M
3. Chen B
4. Choe H
(2022) Mechanisms of SARS-CoV-2 entry into cells
Nature Reviews. Molecular Cell Biology 23:3–20.

https://doi.org/10.1038/s41580-021-00418-x
- PubMed
- Google Scholar
1. Koehler M
2. Delguste M
3. Sieben C
4. Gillet L
5. Alsteens D
(2020) Initial step of virus entry: virion binding to cell-surface glycans
Annual Review of Virology 7:143–165.

https://doi.org/10.1146/annurev-virology-122019-070025
- PubMed
- Google Scholar
(2015) Optimized sleeping beauty transposons rapidly generate stable transgenic cell lines
Biotechnology Journal 10:647–653.

https://doi.org/10.1002/biot.201400821
- PubMed
- Google Scholar
1. Lan J
2. Ge J
3. Yu J
4. Shan S
5. Zhou H
6. Fan S
7. Zhang Q
8. Shi X
9. Wang Q
10. Zhang L
11. Wang X
(2020) Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor
Nature 581:215–220.

https://doi.org/10.1038/s41586-020-2180-5
- PubMed
- Google Scholar
1. Lang J
2. Yang N
3. Deng J
4. Liu K
5. Yang P
6. Zhang G
7. Jiang C
(2011) Inhibition of SARS pseudovirus cell entry by lactoferrin binding to heparan sulfate proteoglycans
PLOS ONE 6:e23710.

https://doi.org/10.1371/journal.pone.0023710
- PubMed
- Google Scholar
(2023) Heparanase-1: from cancer biology to a future antiviral target
Viruses 15:237.

https://doi.org/10.3390/v15010237
- PubMed
- Google Scholar
1. Li W
2. Moore MJ
3. Vasilieva N
4. Sui J
5. Wong SK
6. Berne MA
7. Somasundaran M
8. Sullivan JL
9. Luzuriaga K
10. Greenough TC
11. Choe H
12. Farzan M
(2003) Angiotensin-converting enzyme 2 is a functional receptor for the SARS coronavirus
Nature 426:450–454.

https://doi.org/10.1038/nature02145
- PubMed
- Google Scholar
1. Liu L
2. Chopra P
3. Li X
4. Bouwman KM
5. Tompkins SM
6. Wolfert MA
7. de Vries RP
8. Boons GJ
(2021) Heparan sulfate proteoglycans as attachment factor for SARS-CoV-2
ACS Central Science 7:1009–1018.

https://doi.org/10.1021/acscentsci.1c00010
- PubMed
- Google Scholar
(2022) Cell surface SARS-CoV-2 nucleocapsid protein modulates innate and adaptive immunity
Science Advances 8:eabp9770.

https://doi.org/10.1126/sciadv.abp9770
- PubMed
- Google Scholar
(2023) Cell surface nucleocapsid protein expression: a betacoronavirus immunomodulatory strategy
PNAS 120:e2304087120.

https://doi.org/10.1073/pnas.2304087120
- PubMed
- Google Scholar
1. Macia E
2. Ehrlich M
3. Massol R
4. Boucrot E
5. Brunner C
6. Kirchhausen T
(2006) Dynasore, a cell-permeable inhibitor of dynamin
Developmental Cell 10:839–850.

https://doi.org/10.1016/j.devcel.2006.04.002
- PubMed
- Google Scholar
(2010) Virus entry by endocytosis
Annual Review of Biochemistry 79:803–833.

https://doi.org/10.1146/annurev-biochem-060208-104626
- PubMed
- Google Scholar
1. Milewska A
2. Zarebski M
3. Nowak P
4. Stozek K
5. Potempa J
6. Pyrc K
(2014) Human coronavirus NL63 utilizes heparan sulfate proteoglycans for attachment to target cells
Journal of Virology 88:13221–13230.

https://doi.org/10.1128/JVI.02078-14
- PubMed
- Google Scholar
1. Mulhall EM
2. Gharpure A
3. Lee RM
4. Dubin AE
5. Aaron JS
6. Marshall KL
7. Spencer KR
8. Reiche MA
9. Henderson SC
10. Chew TL
11. Patapoutian A
(2023) Direct observation of the conformational states of PIEZO1
Nature 620:1117–1125.

https://doi.org/10.1038/s41586-023-06427-4
- PubMed
- Google Scholar
1. Mycroft-West CJ
2. Su D
3. Pagani I
4. Rudd TR
5. Elli S
6. Gandhi NS
7. Guimond SE
8. Miller GJ
9. Meneghetti MCZ
10. Nader HB
11. Li Y
12. Nunes QM
13. Procter P
14. Mancini N
15. Clementi M
16. Bisio A
17. Forsyth NR
18. Ferro V
19. Turnbull JE
20. Guerrini M
21. Fernig DG
22. Vicenzi E
23. Yates EA
24. Lima MA
25. Skidmore MA
(2020) Heparin inhibits cellular invasion by SARS-CoV-2: structural dependence of the interaction of the spike S1 receptor-binding domain with heparin
Thrombosis and Haemostasis 120:1700–1715.

https://doi.org/10.1055/s-0040-1721319
- PubMed
- Google Scholar
1. O’Hearn A
2. Wang M
3. Cheng H
4. Lear-Rooney CM
5. Koning K
6. Rumschlag-Booms E
7. Varhegyi E
8. Olinger G
9. Rong L
(2015) Role of EXT1 and glycosaminoglycans in the early stage of filovirus entry
Journal of Virology 89:5441–5449.

https://doi.org/10.1128/JVI.03689-14
- PubMed
- Google Scholar
1. Pape JK
2. Stephan T
3. Balzarotti F
4. Büchner R
5. Lange F
6. Riedel D
7. Jakobs S
8. Hell SW
(2020) Multicolor 3D MINFLUX nanoscopy of mitochondrial MICOS proteins
PNAS 117:20607–20614.

https://doi.org/10.1073/pnas.2009364117
- PubMed
- Google Scholar
1. Rebendenne A
2. Roy P
3. Bonaventure B
4. Chaves Valadão AL
5. Desmarets L
6. Arnaud-Arnould M
7. Rouillé Y
8. Tauziet M
9. Giovannini D
10. Touhami J
11. Lee Y
12. DeWeirdt P
13. Hegde M
14. Urbach S
15. Koulali KE
16. de Gracia FG
17. McKellar J
18. Dubuisson J
19. Wencker M
20. Belouzard S
21. Moncorgé O
22. Doench JG
23. Goujon C
(2022) Bidirectional genome-wide CRISPR screens reveal host factors regulating SARS-CoV-2, MERS-CoV and seasonal HCoVs
Nature Genetics 54:1090–1102.

https://doi.org/10.1038/s41588-022-01110-2
- PubMed
- Google Scholar
1. Schmidt R
2. Weihs T
3. Wurm CA
4. Jansen I
5. Rehman J
6. Sahl SJ
7. Hell SW
(2021) MINFLUX nanometer-scale 3D imaging and microsecond-range tracking on a common fluorescence microscope
Nature Communications 12:1478.

https://doi.org/10.1038/s41467-021-21652-z
- PubMed
- Google Scholar
1. Shaner NC
2. Lambert GG
3. Chammas A
4. Ni Y
5. Cranfill PJ
6. Baird MA
7. Sell BR
8. Allen JR
9. Day RN
10. Israelsson M
11. Davidson MW
12. Wang J
(2013) A bright monomeric green fluorescent protein derived from Branchiostoma lanceolatum
Nature Methods 10:407–409.

https://doi.org/10.1038/nmeth.2413
- PubMed
- Google Scholar
1. Shang J
2. Ye G
3. Shi K
4. Wan Y
5. Luo C
6. Aihara H
7. Geng Q
8. Auerbach A
9. Li F
(2020) Structural basis of receptor recognition by SARS-CoV-2
Nature 581:221–224.

https://doi.org/10.1038/s41586-020-2179-y
- PubMed
- Google Scholar
1. Shin W
2. Ge L
3. Arpino G
4. Villarreal SA
5. Hamid E
6. Liu H
7. Zhao WD
8. Wen PJ
9. Chiang HC
10. Wu LG
(2018) Visualization of membrane pore in live cells reveals a dynamic-pore theory governing fusion and endocytosis
Cell 173:934–945.

https://doi.org/10.1016/j.cell.2018.02.062
- PubMed
- Google Scholar
1. Shin W
2. Wei L
3. Arpino G
4. Ge L
5. Guo X
6. Chan CY
7. Hamid E
8. Shupliakov O
9. Bleck CKE
10. Wu LG
(2021) Preformed Ω-profile closure and kiss-and-run mediate endocytosis and diverse endocytic modes in neuroendocrine chromaffin cells
Neuron 109:3119–3134.

https://doi.org/10.1016/j.neuron.2021.07.019
- PubMed
- Google Scholar
(2005) Inhibitors of cathepsin L prevent severe acute respiratory syndrome coronavirus entry
PNAS 102:11876–11881.

https://doi.org/10.1073/pnas.0505577102
- PubMed
- Google Scholar
1. Suzuki R
2. Yamasoba D
3. Kimura I
4. Wang L
5. Kishimoto M
6. Ito J
7. Morioka Y
8. Nao N
9. Nasser H
10. Uriu K
11. Kosugi Y
12. Tsuda M
13. Orba Y
14. Sasaki M
15. Shimizu R
16. Kawabata R
17. Yoshimatsu K
18. Asakura H
19. Nagashima M
20. Sadamasu K
21. Yoshimura K
22. Sawa H
23. Ikeda T
24. Irie T
25. Matsuno K
26. Tanaka S
27. Fukuhara T
28. Sato K
29. Genotype to Phenotype Japan (G2P-Japan) Consortium
(2022) Attenuated fusogenicity and pathogenicity of SARS-CoV-2 Omicron variant
Nature 603:700–705.

https://doi.org/10.1038/s41586-022-04462-1
- PubMed
- Google Scholar
1. van Haren FMP
2. Valle SJ
3. Neto AS
4. Schultz MJ
5. Laffey JG
6. Artigas A
7. Dixon B
8. Vilaseca AB
9. Barbera RA
10. Ismail TI
11. Mahrous RS
12. Badr M
13. DeNucci G
14. Sverdloff C
15. Camprubi-Rimblas M
16. Cosgrave DW
17. McNicholas B
18. Cody C
19. Curley G
20. Smoot TL
21. Staas S
22. Sann K
23. Sas C
24. Belani A
25. Hillman C
26. Manggala SK
27. Aditianingsih D
28. Sugiarto A
29. Yulianti M
30. Herikurniawan H
31. Sinto R
32. Auerkari AN
33. Permana SA
34. Woodcock A
35. Carroll M
36. Wilkinson T
37. Singh D
38. Shute JK
39. Carroll M
40. Page C
(2025) Efficacy of inhaled nebulised unfractionated heparin to prevent intubation or death in hospitalised patients with COVID-19: an investigator-initiated international meta-trial of randomised clinical studies
EClinicalMedicine 88:103339.

https://doi.org/10.1016/j.eclinm.2025.103339
- PubMed
- Google Scholar
1. Walls AC
2. Park YJ
3. Tortorici MA
4. Wall A
5. McGuire AT
6. Veesler D
(2020) Structure, function, and antigenicity of the SARS-CoV-2 spike glycoprotein
Cell 183:1735.

https://doi.org/10.1016/j.cell.2020.11.032
- Google Scholar
1. Wang Q
2. Zhang Y
3. Wu L
4. Niu S
5. Song C
6. Zhang Z
7. Lu G
8. Qiao C
9. Hu Y
10. Yuen KY
11. Wang Q
12. Zhou H
13. Yan J
14. Qi J
(2020) Structural and functional basis of SARS-CoV-2 entry by using human ACE2
Cell 181:894–904.

https://doi.org/10.1016/j.cell.2020.03.045
- PubMed
- Google Scholar
1. Wang R
2. Simoneau CR
3. Kulsuptrakul J
4. Bouhaddou M
5. Travisano KA
6. Hayashi JM
7. Carlson-Stevermer J
8. Zengel JR
9. Richards CM
10. Fozouni P
11. Oki J
12. Rodriguez L
13. Joehnk B
14. Walcott K
15. Holden K
16. Sil A
17. Carette JE
18. Krogan NJ
19. Ott M
20. Puschnik AS
(2021) Genetic screens identify host factors for SARS-CoV-2 and common cold coronaviruses
Cell 184:106–119.

https://doi.org/10.1016/j.cell.2020.12.004
- PubMed
- Google Scholar
1. Wei J
2. Alfajaro MM
3. DeWeirdt PC
4. Hanna RE
5. Lu-Culligan WJ
6. Cai WL
7. Strine MS
8. Zhang SM
9. Graziano VR
10. Schmitz CO
11. Chen JS
12. Mankowski MC
13. Filler RB
14. Ravindra NG
15. Gasque V
16. de Miguel FJ
17. Patil A
18. Chen H
19. Oguntuyo KY
20. Abriola L
21. Surovtseva YV
22. Orchard RC
23. Lee B
24. Lindenbach BD
25. Politi K
26. van Dijk D
27. Kadoch C
28. Simon MD
29. Yan Q
30. Doench JG
31. Wilen CB
(2021) Genome-wide CRISPR screens reveal host factors critical for SARS-CoV-2 infection
Cell 184:76–91.

https://doi.org/10.1016/j.cell.2020.10.028
- PubMed
- Google Scholar
1. Wrapp D
2. Wang N
3. Corbett KS
4. Goldsmith JA
5. Hsieh CL
6. Abiona O
7. Graham BS
8. McLellan JS
(2020) Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation
Science 367:1260–1263.

https://doi.org/10.1126/science.abb2507
- PubMed
- Google Scholar
1. Wu LG
2. Hamid E
3. Shin W
4. Chiang HC
(2014) Exocytosis and endocytosis: modes, functions, and coupling mechanisms
Annual Review of Physiology 76:301–331.

https://doi.org/10.1146/annurev-physiol-021113-170305
- PubMed
- Google Scholar
1. Wu LG
2. Chan CY
(2024) Membrane transformations of fusion and budding
Nature Communications 15:21.

https://doi.org/10.1038/s41467-023-44539-7
- PubMed
- Google Scholar
1. Yan R
2. Zhang Y
3. Li Y
4. Xia L
5. Guo Y
6. Zhou Q
(2020) Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2
Science 367:1444–1448.

https://doi.org/10.1126/science.abb2762
- PubMed
- Google Scholar
1. Ye C
2. Chiem K
3. Park JG
4. Oladunni F
5. Platt RN II
6. Anderson T
7. Almazan F
8. de la Torre JC
9. Martinez-Sobrido L
(2020) Rescue of SARS-CoV-2 from a single bacterial artificial chromosome
mBio 11:e02168-20.

https://doi.org/10.1128/mBio.02168-20
- PubMed
- Google Scholar
1. Zhang Q
2. Chen CZ
3. Swaroop M
4. Xu M
5. Wang L
6. Lee J
7. Wang AQ
8. Pradhan M
9. Hagen N
10. Chen L
11. Shen M
12. Luo Z
13. Xu X
14. Xu Y
15. Huang W
16. Zheng W
17. Ye Y
(2020) Heparan sulfate assists SARS-CoV-2 in cell entry and can be targeted by approved drugs in vitro
Cell Discovery 6:80.

https://doi.org/10.1038/s41421-020-00222-5
- PubMed
- Google Scholar
1. Zhang YY
2. Liang R
3. Wang SJ
4. Ye ZW
5. Wang TY
6. Chen M
7. Liu J
8. Na L
9. Yang YL
10. Yang YB
11. Yuan S
12. Yin X
13. Cai XH
14. Tang YD
(2022) SARS-CoV-2 hijacks macropinocytosis to facilitate its entry and promote viral spike-mediated cell-to-cell fusion
The Journal of Biological Chemistry 298:102511.

https://doi.org/10.1016/j.jbc.2022.102511
- PubMed
- Google Scholar
1. Zhang Q
2. Tang W
3. Stancanelli E
4. Jung E
5. Syed Z
6. Pagadala V
7. Saidi L
8. Chen CZ
9. Gao P
10. Xu M
11. Pavlinov I
12. Li B
13. Huang W
14. Chen L
15. Liu J
16. Xie H
17. Zheng W
18. Ye Y
(2023) Host heparan sulfate promotes ACE2 super-cluster assembly and enhances SARS-CoV-2-associated syncytium formation
Nature Communications 14:e41453-w.

https://doi.org/10.1038/s41467-023-41453-w
- Google Scholar
1. Zhao WD
2. Hamid E
3. Shin W
4. Wen PJ
5. Krystofiak ES
6. Villarreal SA
7. Chiang HC
8. Kachar B
9. Wu LG
(2016) Hemi-fused structure mediates and controls fusion and fission in live cells
Nature 534:548–552.

https://doi.org/10.1038/nature18598
- PubMed
- Google Scholar
1. Zhao MM
2. Yang WL
3. Yang FY
4. Zhang L
5. Huang WJ
6. Hou W
7. Fan CF
8. Jin RH
9. Feng YM
10. Wang YC
11. Yang JK
(2021) Cathepsin L plays a key role in SARS-CoV-2 infection in humans and humanized mice and is a promising target for new drug development
Signal Transduction and Targeted Therapy 6:134.

https://doi.org/10.1038/s41392-021-00558-8
- PubMed
- Google Scholar
1. Zhou P
2. Yang XL
3. Wang XG
4. Hu B
5. Zhang L
6. Zhang W
7. Si HR
8. Zhu Y
9. Li B
10. Huang CL
11. Chen HD
12. Chen J
13. Luo Y
14. Guo H
15. Jiang RD
16. Liu MQ
17. Chen Y
18. Shen XR
19. Wang X
20. Zheng XS
21. Zhao K
22. Chen QJ
23. Deng F
24. Liu LL
25. Yan B
26. Zhan FX
27. Wang YY
28. Xiao GF
29. Shi ZL
(2020) A pneumonia outbreak associated with a new coronavirus of probable bat origin
Nature 579:270–273.

https://doi.org/10.1038/s41586-020-2012-7
- PubMed
- Google Scholar
1. Zhu Y
2. Feng F
3. Hu G
4. Wang Y
5. Yu Y
6. Zhu Y
7. Xu W
8. Cai X
9. Sun Z
10. Han W
11. Ye R
12. Qu D
13. Ding Q
14. Huang X
15. Chen H
16. Xu W
17. Xie Y
18. Cai Q
19. Yuan Z
20. Zhang R
(2021) A genome-wide CRISPR screen identifies host factors that regulate SARS-CoV-2 entry
Nature Communications 12:961.

https://doi.org/10.1038/s41467-021-21213-4
- Google Scholar

Article and author information

Author details

Sue Han

National Institute of Neurological Disorders and Stroke, Bethesda, United States

Contribution
Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing

Contributed equally with
Xin Wang and Tiansheng Li

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-8451-8592
Xin Wang

National Institute of Neurological Disorders and Stroke, Bethesda, United States

Contribution
Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing

Contributed equally with
Sue Han and Tiansheng Li

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-6250-910X
Tiansheng Li

Cellular Biology Section, Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, United States

Contribution
Conceptualization, Data curation, Formal analysis, Validation, Investigation, Methodology, Writing – review and editing

Contributed equally with
Sue Han and Xin Wang

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-7972-7031
Ammar Mohseni

National Institute of Neurological Disorders and Stroke, Bethesda, United States

Contribution
Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing – review and editing

Competing interests
No competing interests declared
Ivan Kosik

Cellular Biology Section, Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, United States

Contribution
Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-5756-5182
Chung Yu Chan

National Institute of Neurological Disorders and Stroke, Bethesda, United States

Contribution
Data curation, Formal analysis, Investigation, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-6726-5087
Alberto Domingo López-Muñoz

Cellular Biology Section, Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, United States

Contribution
Investigation, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-1542-2153
Jessica Matthias

Abberior Instruments America LLC, Bethesda, United States

Contribution
Resources, Methodology, Writing – review and editing

Competing interests
Jessica Matthias works for Abberior Instruments America LLC, Bethesda, MD, United States, which commercializes the MINFLUX microscope
Reid Suddaby

National Institute of Neurological Disorders and Stroke, Bethesda, United States

Contribution
Investigation, Writing – review and editing

Competing interests
No competing interests declared
Zhixiong Wang

Intramural Research Program, National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health, Bethesda, United States

Contribution
Investigation, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-0879-5728
Albert J Jin

Intramural Research Program, National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health, Bethesda, United States

Contribution
Investigation, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-3826-1081
Christian A Wurm

Abberior Instruments America LLC, Bethesda, United States

Contribution
Resources, Methodology

Competing interests
Christian A. Wurm works for Abberior Instruments America LLC, Bethesda, MD, United States, which commercializes the MINFLUX microscope
Jonathan W Yewdell

Cellular Biology Section, Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, United States

Contribution
Conceptualization, Resources, Supervision, Funding acquisition, Methodology, Project administration, Writing – review and editing

For correspondence
jyewdell@nih.gov

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-3826-1906
Ling-Gang Wu

National Institute of Neurological Disorders and Stroke, Bethesda, United States

Contribution
Conceptualization, Resources, Supervision, Funding acquisition, Methodology, Writing – original draft, Project administration, Writing – review and editing

For correspondence
wul@ninds.nih.gov

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-4261-7079

Funding

National Institute of Neurological Disorders and Stroke (ZIA NS003009-15)

Ling-Gang Wu

National Institute of Neurological Disorders and Stroke (ZIA NS003105-10)

Ling-Gang Wu

National Institute of Allergy and Infectious Diseases (J.W.Y.)

Jonathan W Yewdell

National Institute of Biomedical Imaging and Bioengineering (A.J.J.)

Albert J Jin

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Susan Cheng and Sandra Lara Moreira for EM technical support. This work was supported by NINDS Research Program (ZIA NS003009-15 and ZIA NS003105-10 to LGW) and the Intramural Research Program of the National Institute of Allergy and Infectious Diseases National Institutes of Health (JWY) and the NIBIB (to AJJ), National Institutes of Health. The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

Version history

Sent for peer review: September 8, 2025
Preprint posted: September 11, 2025
Reviewed Preprint version 1: November 14, 2025
Reviewed Preprint version 2: July 15, 2026
Version of Record published: August 6, 2026

Cite all versions

You can cite all versions using the DOI https://doi.org/10.7554/eLife.108925. This DOI represents all versions, and will always resolve to the latest one.

Copyright

This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.