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



Dynamin independent endocytosis is an alternative cell entry mechanism for multiple animal viruses [1]

['Ravi Ojha', 'Department Of Virology', 'Faculty Of Medicine', 'University Of Helsinki', 'Helsinki', 'Anmin Jiang', 'Clem Jones Centre For Ageing Dementia Research', 'Queensland Brain Institute', 'The University Of Queensland', 'Brisbane']

Date: 2024-12

Mammalian receptor-mediated endocytosis (RME) often involves at least one of three isoforms of the large GTPase dynamin (Dyn). Dyn pinches-off vesicles at the plasma membrane and mediates uptake of many viruses, although some viruses directly penetrate the plasma membrane. RME is classically interrogated by genetic and pharmacological interference, but this has been hampered by undesired effects. Here we studied virus entry in conditional genetic knock-out (KO) mouse embryonic fibroblasts lacking expression of all three dynamin isoforms (Dyn-KO-MEFs). The small canine parvovirus known to use a single receptor, transferrin receptor, strictly depended on dynamin. Larger viruses or viruses known to use multiple receptors, including alphaviruses, influenza, vesicular stomatitis, bunya, adeno, vaccinia, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and rhinoviruses infected Dyn-KO-MEFs, albeit at higher dosage than wild-type MEFs. In absence of the transmembrane protease serine subtype 2 (TMPRSS2), which normally activates the SARS-CoV-2 spike protein for plasma membrane fusion, SARS-CoV-2 infected angiotensin-converting enzyme 2 (ACE2)-expressing MEFs predominantly through dynamin- and actin-dependent endocytosis. In presence of TMPRSS2 the ancestral Wuhan-strain bypassed both dynamin-dependent and -independent endocytosis, and was less sensitive to endosome maturation inhibitors than the Omicron B1 and XBB variants, supporting the notion that the Omicron variants do not efficiently use TMPRSS2. Collectively, our study suggests that dynamin function at endocytic pits can be essential for infection with single-receptor viruses, while it is not essential but increases uptake and infection efficiency of multi-receptor viruses that otherwise rely on a functional actin network for infection.

To initiate their infection cycle, most viruses first need to enter their target cells, a process called endocytosis. In mammalian cells, endocytosis often involves a class of proteins called dynamins. While numerous viruses, including SARS-CoV-2, efficiently infect cells by dynamin-mediated endocytosis we found that in the absence of these proteins an alternative cell entry mechanism exists, allowing multiple pathogenic human viruses to enter and infect cells. Unlike the dynamin-mediated infection, the efficient internalization of viral particles via dynamin-independent endocytosis seems to always require functional actin fibers, which are structural components of the cell contractile cytoskeleton. Thus, multiple viruses can infect their target cells using at last two entry mechanisms, inhibiting both may provide effective antiviral therapies.

Funding: For the funding that supported this research we thank the University of Helsinki Graduate Program in Microbiology and Biotechnology for supporting R.O (2019-2023). The strategic Research Council of Finland grant 335527 (G.B), The Sigrid Juselius Foundation (G.B., 2024-2028 https://www.sigridjuselius.fi/en/ ), the Faculty of Medicine at the University of Helsinki (G.B.), The Helsinki Institute for Life Sciences at University of Helsinki (2023-2025, G.B.) the Jane and Aatos Erkko Foundation (O.P.V., https://jaes.fi/en/frontpage ), Helsinki University Hospital funds TYH2021343 (O.P.V.), European Union’s Horizon Europe Research and Innovation Program grant 101057553 (G.B., O.P.V.), The University of Queensland Amplify Fellowship (M.J.), the Australian National Health and Medical Research council (grant 2010917 G.B. and F.A.M.). This work was also supported by Australian Research Council (ARC) Discovery Early Career Researcher Award (DE190100565). The work was supported by an ARC Linkage Infrastructure Equipment and Facilities grant (LE130100078) and a National Health and Medical Research Council (NHMRC) Senior Research Fellowship (GNT1155794) to F.A.M.; P.Y.L is supported by the Agence Nationale de la Recherche (ANR) funding (grant numbers ANR-21-CE11-0012 and ANR-22-CE15-0034). Work in the Greber lab was supported by grants from Schweizerischer Nationalfonds (Swiss National Science Foundation 31003A_179256 and 310030_212802). Y.Y. was supported by ERC Synergy grant CHUbVi (ID: 856581). We are grateful to the Jane and Aatos Erkko Foundation for support to M.V.R. and the Research Council of Finland grant 330896 (to M.V.R.) and 332615 (to E.M.). J.M. was supported by core funding to MRC Laboratory for Molecular Cell Biology at University College London (MC_UU_00012/7). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, using genetic depletion of all three dynamin isoforms, we surveyed a range of animal viruses for their dynamin-dependency to infect cells. We show that while some viruses, including SARS-CoV-2, strongly depend on the presence of dynamins to productively infect cells, other animal viruses, including alphaviruses, influenza, and bunyaviruses, among others, can use dynamin-independent endocytosis (DIE) as an alternative and efficient entry mechanism. This yet uncharacterized endocytic pathway appears sensitive to perturbations of the actin cytoskeleton.

Enveloped viruses, i.e., viruses surrounded by lipid membranes, deliver the viral genome into the cytoplasm by the fusion of the viral and cellular membranes, a process driven by viral surface proteins (often referred to as ‘spikes’ on the virion). For most enveloped viruses the cue that triggers fusion is the drop in pH that occurs once the viral particle reaches the lumen of endosomal vesicles (e.g., Influenza virus, Vesicular stomatitis virus, Semliki Forest virus, among others) [ 15 ]. For other enveloped viruses (including coronaviruses), the fusion is triggered by proteolytic cleavage of the spike proteins which results in conformational changes, the insertion of the viral spike into the host membrane and the fusion of viral and cellular membranes [ 26 , 27 ]. In the case of coronaviruses, these proteolytic cleavages are catalyzed by cellular proteases present either in the endo-lysosomal compartment (e.g., the cysteine protease cathepsin-L) or at the PM (e.g., the transmembrane serine protease 2, TMPRSS2) [ 28 – 31 ]. Depending on the availability of these proteases, virus fusion and the delivery of the viral genome into the cytoplasm can either occur at the PM or early/recycling endosomes (i.e., in cells that express PM serine proteases) or from within endo-lysosomes (i.e., in cells that express active cathepsins in endo/lysosomes but not serine proteases at the cell surface and early endosomes) [ 32 ]. The virus requires endocytosis to reach endosomes and lysosomes. Unlike the cell entry in the cells of the respiratory mucosa, the endosomal route of cell entry has been proposed for SARS-CoV-2 infection of human neurons, which do not express TMPRSS2 [ 33 ].

Irrespective of the upstream steps, viruses in endocytic vesicles are sorted to intracellular membranous organelles, from which the free or capsid enclosed viral genome is released to the cytosol, transported to the site of replication or protein synthesis, or becomes subject to antiviral sensing and degradation (for reviews, see [ 21 – 25 ]).

For their replication in the cytoplasm or the nucleus, most animal viruses use receptor-mediated endocytosis (RME) pathways [ 1 ] through which the cells engulf particles and nutrients into vesicular compartments. Multiple endocytic mechanisms have been described and viruses have been shown to utilize one or more of these pathways [ 2 , 3 ]. In addition to endocytosis, some viruses infect cells by penetrating directly through the plasma membrane (PM) [ 4 ] demonstrating the remarkably fexibility of viruses and their interactions with a diverse range of host cells. How viruses utilise endocytic processes is still incompletely known, but it is key to understanding viral tropism for different cell types, and opening of potential cellular targets for infection interference. Viral endocytosis starts by the virus particle engaging a receptor at the plasma membrane [ 3 ], often times followed by viral motions driven by the host receptors and the underlying acto-myosin cytoskeleton, membrane deformation and pinching off of a vesicle containing the virus particle to the cytoplasm [ 5 ]. The fission process requires energy typically provided by one of the three dynamin isoforms (dyn1-3), all of which have GTPase activity [ 6 ]. Although not all the endocytic pathways require dynamin and some use the actin cytoskeleton, membrane deforming proteins, and motor proteins to generate endocytic vesicles [ 7 – 9 ], dyn-dependent pathways have been the predominant ones described for mammalian viruses [ 2 ]. While some viruses, such as Vaccinia virus (VACV) [ 10 , 11 ], Lymphotropic choromeningitis virus (LCMV) [ 12 , 13 ], and human papilloma virus (HPV) [ 14 ] have been shown to enter cells via dynamin-independent endocytic mechanisms, others such as Semliki Forest virus (SFV) [ 15 , 16 ] and Influenza virus [ 17 , 18 ] have been shown to enter cells via dynamin-dependent as well as dynamin independent endocytosis. The studies addressing the role of dynamins in virus infection have so far being carried using small molecule inhibitors of dynamins, overexpressing dominant-negative mutants of dynamin, or RNA interference (RNAi) approaches. These loss-of-function approaches, although easy to implement, can suffer from uncharacterized off-targets effects [ 19 , 20 ] or incomplete depletion of the dynamin isoforms.

Results and discussion

Characterization of dynamin 1,2 conditional knock-out cells to study virus entry Many animal viruses, including the alphaviruses Semliki Forest (SFV) and Sindbis (SINV) virus, influenza virus (IAV), and vesicular stomatitis virus (VSV), have been shown to infect cells using dynamin-dependent endocytosis [2]. Most of these studies have been performed by treating cells with small molecule dynamin inhibitors, by overexpression of dominant-negative inactive forms of dynamin, or by depletion of dynamin mRNAs using RNA interference methods. These loss-of-function approaches, although easy to implement, can suffer from uncharacterized off-targets effects [19,20] or incomplete depletion of the dynamin isoforms. To address the role of dynamins in virus infection, and to overcome the above-mentioned limitations, we used genetically engineered dynamin 1,2 conditional double knockout (KO) mouse embryonic fibroblasts (MEFDKO) [34]. In these cells, the two main isoforms of dynamin (dyn1,2) can be completely depleted within 6 days of cell treatment with 4OH-tamoxifen (4OH-TMX) [34] (Fig 1A). The third isoform of dynamin (dyn3) is not detectable in these cells [34]. To functionally monitor the specificity and level of inhibition of dynamin-dependent endocytosis in this model system, we used transferrin (Tf) as a positive control, which is internalized in a dynamin-dependent manner [35], and the cholera toxin subunit B (CTB), which can enter cells via both dynamin-dependent [36,37] (at low doses) or dynamin-independent (at high doses) endocytic mechanisms [38]. All the original data shown in the graphs presented in this work are available in supplementary data file “S1 Data”. As expected, 6 days after 4OH-TMX treatment, the uptake of fluorescently labelled Tf (5 μg/ml) was strongly inhibited in most cells (Fig 1B). Consistent with a previous report [34], in all of our experiments, approximately 3–5% of the MEFDKO cells did not respond to 4OH-TMX treatments and as a result maintained normal levels of Tf uptake (S1A–S1C Fig). Importantly, the block in Tf uptake was not due to decreased levels of the Tf receptor at the cell surface, as demonstrated by increased binding of fluorescent Tf at the cell surface of dynamin depleted cells (S1D–S1E Fig). Likewise, the internalization of CTB was also significantly reduced at low toxin concentration (i.e., 50 ng/ml) (Fig 1C), while at higher concentration (i.e., 1 μg/ml), the levels of CTB uptake in MEFDKO cells were comparable to those of vehicle control treated cells (Fig 1D). These experiments confirmed that the MEFDKO represent a suitable model system to study dynamin-dependent and -independent endocytosis. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Characterization of MEF dynamin 1,2 conditional KO cells to study virus infection. A) Western blot analysis of dynamin 1 and 2 levels in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days. Tubulin was used as a loading control. The Dyn1,2 antibody used recognizes both dynamin 1 and 2. B) Quantification of Tf Alexa-647 (Tf) (5 μg/ml) following a 30 min uptake in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days prior fixation and Hoechst DNA staining. Representative fluorescent images on the right show Tf (white) and Hoechst (red) in MEFDKO cells treated with vehicle control or 4OH-TMX. C-D) Quantification of CTB Alexa-647 (CTB) uptake in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days and incubated with indicated concentrations of CTB for 30 min prior to fixation and Hoechst DNA staining. Representative fluorescent images on right show CTB (white) and Hoechst (red) in MEFDKO cells treated with vehicle control or 4OH-TMX. E) Quantification of CPV infection in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days and infected with CPV for 24h, and immunostained for non-structural protein 1 (NS1). The inset schematic illustrates the entry mechanism of CPV. F) Quantification of VACV infection in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days and infected with VACS for 6 h. The inset schematic illustrates VACV entry by micropinocytosis. G) Effect of Latrunculin-B (LatB) on VACV infection in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days and infected with VACS for 6 h. H) Quantification of SFV infection in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days and infected with indicated MOIs of SFV-EGFP for 7 h. The inset illustrates the FACS analysis of virus-induced EGFP fluorescence in non-infected (red line) and infected (blue line) vehicle-control treated cells. Values represent the mean of 3 independent experiments. Error bars represent the standard deviation (STDEV). Statistical significance was calculated using a unpaired two-tailed t-test (*p<0,05; ****p<0,0001; n.s. = non-significant). https://doi.org/10.1371/journal.ppat.1012690.g001 To test whether the inducible dynamin depletion system was suitable to study virus infection, we first tested three well characterized viruses in the MEFDKO cells treated with vehicle control (EtOH) or 4OH-TMX for 6 days: i) canine parvovirus (CPV), a small single-stranded DNA virus that uses the Tf receptor (TfR) and dynamin-dependent endocytosis to enter cells [39,40] (here used as a positive control); ii) vaccinia virus (VACV-EGFP, EGFP expressed under viral early gene promoter [41]), a large DNA virus that enters cells via actin-dependent, dynamin-independent macropinocytosis [10,11] (negative control); and Semliki Forest virus (SFV-GFP, GFP expressed as a fusion with viral replicase protein nsP3 [42]), which has been shown to infect cells mainly via dynamin-dependent endocytosis [15], although alternative entry mechanisms have also been proposed [16]. Infection rates were determined by fluorescence-activated cell sorting (FACS) flow cytometry analysis of virus-induced expression of GFP (for VACV-GFP and SFV-GFP) or following immunofluorescence staining of viral proteins (for CPV). In the absence of the canine TfR MEFDK cells are not susceptible to CPV infection, indicating that the cell entry process of this virus relies on receptor-mediated endocytosis and no other murine receptors can be efficiently use to facilita virus entry (S2A Fig, non transfected). As expected, transient expression of the feline TfR [40] was sufficient to extablish CPV infection and viral protein synthesis (S2B Fig, Ctrl). 4OH-TMX treatment in MEFDKO cells transiently over-expressing the feline TfR [40] inhibited CPV infection by more than 90% (Figs 1E and S2B). The infectivity of VACV in dyn1,2 depleted MEFDKO cells was comparable to that of control cells, confirming that this virus enters cells via a dynamin-independent mechanism (Fig 1F). Disruption of the actin cytoskeleton using the actin depolymerizing drug Latrunculin-B (LatB), on the other hand, blocked VACV infection in both dyn1,2 depleted and control MEFDKO cells (Fig 1G). The actin-dependency of VACV infection was expected and consistent with viral entry through macropinocytosis, a process where actin polymerization is required to support the formation of PM protrusions that can engulf extracellular fluids and large particles such as VACV virions that have a size of approximately 250x250x350 nm [10,11,43]. Notably, the infection of SFV was only partially inhibited by dyn1,2 depletion, and increasing the virus dose (i.e., the multiplicity of infection, MOI) from MOI 1 to 10 fully restored infection (Fig 1H). In summary, the MEFDKO cells are a suitable model to study the role of dynamins in virus infection. Small viruses, e.g. CPV, that target a receptor (e.g. TfR) that is internalized exclusively via dynamin-dependent endocytic routes cannot efficiently infect cells in the absence of dynamins. Compensatory dynamin-independent endocytosis is not available for such single-receptor binding viruses. On the other hand, viruses such as SFV, which may target more than one protein receptor, including a variety of heparan-sulphate-containing glycoproteins [44], are internalized primarily via dynamin-dependent endocytosis. If this entry mechanism is not available, infection can occur through an alternative dynamin-independent pathway, albeit less effectively. Regardless of the mechanism of endocytosis, SFV infection in these cells appears to rely on negatively-charged receptors, such as the heparan-sulphate-containing glycoproteins [44], as shown by competition experiments with increasing concentrations of heparin blocking SFV infection in both control and dynamin depleted cells to a similar extent (S2C–S2D Fig).

Dynamin-independent endocytosis is an alternative, efficient entry pathway for multiple animal viruses Endocytic pathways that require dynamins, such as clathrin-mediated endocytosis (CME), have been associated with the cell entry of numerous viruses, including members of the mosquito-delivered alphaviruses, such as SFV, SINV [45], and IAV [17], as well as members of Bunyaviridae such as Uukuniemi virus (UUKV) [46], vesiculoviruses such as vesicular stomatitis virus (VSV) [47,48], common cold human rhinovirus (RV) B14 and A89 [45,49], and species C AdV such as AdV-C2 or C5 [50–53] as well as AdV-B3 and AdV-B35, although the dynamin requirement of B3 and B35 has been shown to be variable between cell lines [54,55]. Influenza virus has been shown to enter cells also by a micropinocytosis-like mechanism [18]. The results obtained here with SFV (Fig 1H) prompted us to test if the dynamin-independent pathway could be used as a productive entry route for other viruses that are known to use dynamin-dependent endocytosis to enter host cells. To this end, we tested if the infectivity of SINV, VSV, IAV, UUKV, RV-A1 and AdV-C5 was inhibited in dyn1,2 depleted MEFDKO cells. VACV, which is known to infect cells via actin-dependent macropinocytosis irrespective of the presence of dynamins (Fig 1F–1G), was used as a comparison. Infection was monitored by FACS or immunofluorescence analysis. VACV [41], VSV [48], and AdV-C5 [56] were engineered to express the EGFP protein, and SINV to express the mCherry, as fluorescent reporters of infection. MEFDKO cells were treated with vehicle or 4OH-TMX for 6 days, and the infection rates for these recombinant viruses were quantified at 7 hours post infection (hpi) (VACV-EGFP, VSV-EGFP, SINV-mCherry) and 22 hpi (AdV-C5-EGFP) by monitoring the number of EGFP or mCherry expressing cells, or by using immunofluorescence staining against viral antigens with virus-specific antibodies at 8 hpi (IAV X31, UUKV) and 24 hpi (RV-A1). Interestingly, all the tested viruses, albeit with different efficiency, were able to infect cells in the absence of dynamins (Fig 2). At low viral doses (equivalent to an infection of approximately 10–30% of cells), the infections of SINV-mCherry, VSV-EGFP, and IAV X31 were significantly decreased in dyn1,2 depleted cells. In contrast, in the absence of dynamins, the infection of UUKV was not significantly blocked, and in the case of the common cold viruses RV-A1 and AdV-C5-EGFP, infection in dynamin-depleted cells was enhanced in comparison to controls (Figs 2 and S3). Thus, dynamin-independent endocytosis is an efficient, alternative virus entry portal that can be exploited by multiple animal viruses. The results also demonstrate that using a low viral dose is advisable to estimate the contribution of the two entry mechanisms to infection. A comparative analysis of all tested viruses, at a viral dose that corresponds to 20–40% infection rates in vehicle control treated cells (Figs 2 and S2), indicates that the dependence on dynamin-mediated endocytosis follows this qualitative order: VSV>SFV = SINV = IAV>UUNV. Infections with VACV, RV-A1 and AdV-C5 were not inhibited by dynamin depletion, rather increased, suggesting that the cells may have adapted to the depletion of dynamins by up-regulation of dynamin-independent endocytic pathways. Upond complete depletion of dynamins, MEFDKO cells stop dividing due to a block in cytodieresis, a process that requires dynamin function [34]. However, the infection observed by multiple viruses in dynamin-depleted cells demonstrates that the lack of cell division does not prevent the virus entry process and expression of viral genes, at least of the viruses used in this study. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Dynamin-independent endocytosis is an alternative, efficient virus entry pathway for multiple animal viruses. Infection of indicated viruses in MEFDKO cells treated with vehicle control or 4OH-TMX for 6 days and infected for 6 h (VACV), 7 h (SINV, VSV), and 8 h (IAV X31, UUKV). Virus infection was determined by FACS analysis of EGFP (VAVC and VSV), mCherry (SINV), or after immunofluorescence of viral antigens using virus-specific antibodies (IAV X31 and UUKV). Values indicate the mean of three independent experiments and the error bars represent the standard deviation (STDEV). n.i. = non infected. Statistical significance was calculated by unpaired two-tailed t-test (*p<0,05; ** p<0,01; n.s. = non-significant). https://doi.org/10.1371/journal.ppat.1012690.g002

Ultrastructural analysis of dynamin-independent SFV entry Because SFV entered cells efficiently both in the presence and absence of dynamins, we used transmission electron microscopy (TEM) to gain ultrastructural information on the endocytic processes that mediate SFV entry in dyn1,2 depleted cells. MEFDKO cells were treated with 4OH-TMX or vehicle control for 6 days and, following virus (MOI = 1000) adsorption at 4°C for 1 h, cells were shifted to 37°C for 15 minutes to promote viral internalization. In MEFDKO cells treated with vehicle control, TEM analysis revealed numerous viruses at the outer surface of the cells (Fig 4A), as well as inside endocytic invaginations that were surrounded by an electron dense coat, consistent with the appearance of clathrin coated pits [63,64] (CCP) (Fig 4B). A sizable fraction of SFV particles were also found inside bulb-shaped non-coated pits (NCP) (Fig 4C). In addition to small PM invaginations, large (>250 nm in diameter) vacuoles containing viruses and located close to the PM were detected. These structures, here annotated as ‘large endocytic/endosomal profiles’ (LEP), could represent early endosomal vesicles or large invaginations of the PM that appear circular in TEM cross-sections (Fig 4D, Ctrl). In dyn1,2 depleted MEFDKO cells, SFV virions were also readily detected on the PM (Fig 4E). It has been previously reported that in the absence of dynamins the final step of CME does not occur, and the respective ‘stalled’ vesicles are pulled from the PM towards the cytoplasm by actin polymerisation followed by depolymerization and release back towards the PM [34]. In agreement with these reports [34], in dyn1,2 depleted MEFDKO cells, CCPs were often associated with long tubular membraneous structures (Fig 4F, white arrowhead). A fraction of the viruses was found in these stalled CCPs (Fig 4F), and the rest in NCPs (Fig 4G) and in LEPs (Fig 4D, 4OH-TMX). Consistent with the fast entry kinetics of SFV, quantitative analysis of the TEM images revealed that 10 min after entry in Ctrl cells, 60% of the virions were localized in LEP, which included endosomes, 10% in CCP, and approximately 30% in NCP (Fig 4H, Ctrl). Upon dynamin depletion, this ratio changed (Fig 4H, 4OH-TMX), and the number of virions in CCPs and NPCs at the PM increased at the expense of viruses found in the endosomal proceses. Cumulative analysis revealed that in the absence of dynains up to 70–90% of viral particles where in CCP and NCP and only 10–30% in large endocytic structures, indicating a strong inhibition of virus endocytosis (Fig 4I). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Ultrastructural analysis of Semliki Forest Virus entry in MEFDKO cells. Representative TEM images of MEFDKO cells treated with vehicle control (A-D) or 4OH-TMX (E-G, and D, 4OH-TMX) for 6 days and infected with SFV (MOI 1000) on ice for 2 h followed by shift to 37°C for 15 min before fixation and processing for TEM. The fractions of total viral particles found in each of the described endocytic processes are quantified in H and I. Scale bar 100 nm. CCP = clathrin-coated profile, Mv = microvilli, PM = plasma membrane. The asterisks (*) in panel A indicate an endocytic process. White arrowheads indicate elongated tubular membranous structures connected to CCP. Boxed areas are magnified at bottom right corner of each figure panel. All values represent the mean and standard deviation of three replicas. Quantification of each treatment (EtOH vehicle ctrl or 4OH-TMX) includes over 120 viral particles. Statistical analysis was performed using ordinary two-way ANOVA multiple comparisons test (* p<0.05; *** p<0.001; **** p<0.0001). Scale bars = 5 μm. https://doi.org/10.1371/journal.ppat.1012690.g004 This analysis is consistent with two entry mechanisms for SFV, one dynamin-dependent and another dynamin-independent. We speculate that upon depletion of dynamins, a fraction of the virus remains trapped in CCPs and NCPs. This stalled entry pathways might account for the partial inhibition of infection in dyn1,2 depleted MEFDKO cells observed in this study for SFV and possibly for other viruses. The fact that in unperturbed cells, SFV infection is not sensitive to actin depolymerizing drugs, suggests that the dynamin-dependent route is the predominant entry pathway. However, a sizable fraction of the virions can access an alternative entry pathway that, based on the effect of actin depolymerizing drugs, seems to rely more on the actin cytoskeleton. The precise role of actin is unknown but it could potentially help the final pinching step of the dynamin-independent endocytic process.

ACE2-mediated endocytosis of SARS-CoV-2 trimeric spike is mainly dynamin-dependent The exact mechanism of SARS-CoV-2 cell entry is not fully understood, and evidence suggests that both clathrin-dependent and -independent endocytosis could be involved in cell culture models, including human neurons that do not express the cell surface protease TMPRSS2 [33]. A study from Bayati et al. [65] used chemical inhibitors of dynamins and siRNA-mediated depletion of clathrin to show that the internalization of both the SARS-CoV-2 trimeric soluble Spike (S) protein and the infection of lentiviruses pseudotyped with SARS-CoV-2 S were decreased. Another study indicated that cell entry of SARS-CoV-2 Spike protein occurs via clathrin independent endocytosis in cells devoid of the human angiotensin converting enzyme 2 (ACE2) [66]. Both studies implied entry via dynamin-dependent mechanisms. To address the role of endocytosis in SARS-CoV-2 infection, and to identify the potential endocytic mechanisms that leads to productive entry, we started our investigation by following the uptake and intracellular trafficking of fluorescently labelled, recombinant soluble trimeric SARS-CoV-2S protein. This soluble version of the viral S protein is held as a trimer by a molecular clamp [67,68] that replaces the trans-membrane domain of the glycoprotein. In addition, for stabilization purposes, the polybasic furin-cleavage site, required to activate the viral spike for fusion, was rendered uncleavable in each monomer by mutagenesis [49]. This likely also blinded the S-trimer to neuropilin-1 (NRP1) binding [69]. MEFDKO cells transiently expressing the main viral receptor ACE2 [70], and an EGFP-tagged version of the early endosome protein Rab5 [71], were used in these studies. Time-course uptake assays followed by confocal fluorescence microscopy were performed to distinguish the fraction of Alexa Fluor-555-labelled S localized at the PM from those internalized into endosomal vesicles labelled by Rab5-EGFP (early endosomes) (Fig 5). Tf labelled with Alexa Fluor-647 was used as an internal positive control to monitor the extent of inhibition of dynamin-dependent endocytosis in each cell analysed. In the absence of ACE2, MEF cells did not significantky bind SARS-COV-2 S (Fig 5A and 5B). In vehicle control treated MEFDKO cells transiently expressing ACE2 and Rab5-EGFP, the uptake of S was efficient, and the internalized S protein colocalized with Rab5-EGFP positive vesicles (Fig 5C and 5D, Ctrl). In dyn1,2 depleted MEFDKO cells most of the S fluorescent signal remained associated with the PM of the cells, even after 3h, indicating a strong, albeit not complete, inhibition of endocytosis (Fig 5C and 5D, 4OH-TMX). Consequently, the colocalization of the intracellular Rab5 compartment with the S protein remained low throughout the time course (Fig 5A and 5B, 4OH-TMX). Similar results were obtained for fluorescently labelled Tf, which also accumulated on the PM in dyn1,2 depleted MEFDKO cells at the expenses of intracellular compartments (Fig 5A, 4OH-TMX). Taken together, these results demonstrate that, similarly to Tf, the ACE2-mediated endocytosis of SARS-CoV-2 S trimeric proteins is mainly dynamin-dependent. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. ACE2-mediated endocytosis of SARS-CoV-2 trimeric spike proteins is dynamin-dependent. A) Representative confocal fluorescence images of MEFDKO cells non transfected, transiently expressing hACE2, or hACE2 and Rab5-EGFP. B) Quantification of the mean spike fluorescence per cell. A.u. = arbitrary units. C) Representative confocal fluorescence images of MEFDKO cells transiently expressing hACE2 and Rab5-EGFP and treated with vehicle control or 4OH-TMX for 6 days and incubated for 3h with Alexa Fluor-555-labelled transferrin (Tf, magenta) and Alexa Fluor-647 labeled soluble trimeric SARS-CoV-2 Spike protein (red). The insets in A) and C) show magnified images from the indicated white boxed areas. The images represent a single optical slice of the imaged cell. B) Quantification of colocalization between indicated proteins in cells treated as described in C). The mean ±STDEV of 16 Ctrl cells and 14 4OH-TMX treated cells for 15 min; 16 Ctrl cells and 15 4OH-TMX cells for 45 min; and 17 Ctrl cells and 18 4OH-TMX cells for 3 h are shown. The statistical significance was calculated using a non parametric Mann-Whitney U test (****P<0,0001). https://doi.org/10.1371/journal.ppat.1012690.g005

Depletion of dynamins blocks endocytosis of authentic SARS-CoV-2 virions In addition to the MEFDKO cells, a triple dynamin 1,2,3 conditional KO cell line (here referred to as MEFTKO) [72] became available for this study. The dyn1,2,3-depletion in MEFTKO cells following a 6-day treatment with 4OH-TMX, blocks Tf uptake similarly to the dyn1,2-depletion in MEFDKO cells [72]. An advantage of this triple dyn1,2,3 KO model system is that in addition to depletion of all three dynamin isoforms 1, 2, and 3, these cells adhere better to culture surfaces compared to the MEFDKO cell clone. In addition, the responsiveness to 4OH-TMX treatment is more stable over cell passaging than in MEFDKO cells (not shown). The dynamin inhibitor OH-dynasore (also called Dingo-4A) showed off-target effects in both MEFDKO and MEFTKO cells, bloking the infection of the alphavirus SFV in control and dynamin-depleted cells to a similar extent, when tested both at high and moderate drug concentrations (S5A and S5B Fig). Because murine cells are not infectable by SARS-CoV-2, to study how a clinical isolate of the ancestral SARS-CoV-2 Wuhan strain (here referred to as Wuhan [73]) enters MEFTKO cells, we used a lentiviral expression vector to stably express the human ACE2 receptor (MEFTKO-ACE2). To confirm that the inhibition of S internalization observed in dynamin depleted cells corresponded to a proportional block in endocytosis of SARS-CoV-2 virions, we directly measured the extent of virus internalization in dyn1,2,3-depleted and control MEFTKO-ACE2 cells. Viruses (Wuhan, equivalent MOI of 10) were added to cells at 37°C for 60 min in the presence of 50 μM cycloheximide to prevent viral protein synthesis. After fixation, cells were processed for sequential immunofluorescence staining as described above for SFV to distinguish viral particles outside of the PM, i.e. particles not yet internalized (virus out), from internalized virions (virus in). Hence, in this assay, non-internalized virions are stained either with the first (i.e. magenta spots) or with both (i.e. white colocalizing spots) fluorophores (Fig 6A, Ctrl, magenta or white dots). Internalized virions are stained only with the second fluorophore (Fig 6A, 4OH-TMX, green dots). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Dynamin-dependent and -independent entry of SARS-CoV-2 Wuh and Delta infection in ACE2-expressing MEFTKO cells is actin dependent. B) Representative confocal images of virus entry in MEFTKOACE2 cells (4OH-TMX) or control cells (Ctrl) fixed at 60 min after virus inoculation and processed for sequential immunofluorescence as described in A. Inset images show a magnification of the area indicated by the white dashed boxes, with merged as well as separated fluorescence images. Virus out = non internalized viruses; virus in = internalized viruses. C) Quantification of non-internalized (virus out) and internalized (virus in) virions using automated image analysis. Values represent the mean of 15 cells from 3 independent experiments, and error bars represent the STDEV. Scale bars = 10 μm Statistical significance was calculated by unpaired two-tailed t-test (**p<0,01; n.s. = non-significant). D) Schematic description of the Lat-B treatment in MEFTKOACE2. E) Representative fluorescence images of MEFTKOACE2 treated with indicated compounds 15 minutes before infection and infected with Wuhan or the Delta variant of SARS-CoV-2 for 20 h at 6 days after vehicle control (Ctrl) or 4OH-TMX treatment. Scale bars = 200 μm F-H) Quantification by image analysis of SARS-COV-2 Wuhan or Delta infection in MEFTKOACE2 cells treated with vehicle control (Ctrl) or 4OH-TMX for 6 days and infected with indicated MOIs for 20 h. Values indicate the mean of at least three independent experiments and the error bars represent the STDEV. Statistical significance was calculated using a non parametric Mann-Whitney U test (*p<0,05; **p<0,01; ***p<0,001). https://doi.org/10.1371/journal.ppat.1012690.g006 In the unperturbed control cells, at 45 min p.i., image analysis after confocal imaging revealed that a sizable fraction of viruses was found inside the cells (Fig 6B, Ctrl). In contrast, depletion of dyn1,2,3 blocked virus endocytosis almost completely (Fig 6B, 4OH-TMX). This reduction of virus internalization did not correlate with proportional changes in the cell surface levels of the receptor ACE2, which decreased less than 20% upon dynamin depletion (S5C and S5D Fig). This experiment demonstrates that the strong block of S internalization observed in dynamin depleted cells corresponds to a similar block in virus endocytosis.

SARS-CoV-2 infection in ACE2-expressing MEF cells requires endosome maturation and is insensitive to TMPRSS2 inhibitors To study how a clinical isolate of the ancestral SARS-CoV-2 Wuhan strain [73] infects MEFTKO cells, in addition to the MEFTKO-ACE2 cells, we generated a cell line stably expressing ACE2 together with a N-terminally GFP-tagged version of TMPRSS2 (here referred to as MEFTKO-AT). We used fluorescence activated cell sorting to isolate cells that expressed high, medium, and low levels of TMPRSS2-GFP. However, only the cells that expressed the lowels levels of GFP reattached to the culture flasks (se materials and methods), indicating that in MEF cells a high surface expression of the trypsin-like serine protease TMPRSS2 interferes with cell adhesion. We first determined to which extent in these cell lines the cell entry of SARS-CoV-2 Wuhan depended on endosomal proteases or cell surface serine proteases such as TMPRSS2. To this end, we tested the sensitivity of Wuhan infection to nafamostat, an inhibitor of TMPRSS2 [74], and apilimod, an inhibitor of the phosphoinositol-5 kinase (PIP5K) required for efficient early to late endosome maturation and, therefore, delivery of endocytosed cargo to the lysosomal compartment [75,76]. Drug or vehicle-control pretreated infected cells were fixed at 20 hpi and the percentage of Wuhan infection was monitored by immunofluorescence using antibodies against the viral nucleoprotein, followed by automated high-content imaging and image analysis (S5E Fig). As expected, in cells that did not express TMPRSS2, Wuhan infection was strongly inhibited by apilimod but not by Nafamostat (S5E and S5F Fig, MEFTKO-ACE2). The over-expression of TMPRSS2-GFP, even if at low levels, had two main effects, firstly, it slightly increased the overall infectivity of the virus in DMSO-control treated cells compared to the values obtained in cells that only expressed ACE2 (S5E and S5G Fig, MEFTKO-AT). Secondly, it rendered infection partially resistant to apilimod and sensitive to nafamostat (S5E and S5F Fig, MEFTKO-AT). These results indicate that the MEF-ACE2 cells either do not express TMPRSS2 or, if they do, the protein is not available to the virus. Hence, infection depends on virus endocytosis and delivery to endo/lysososmes. In MEFTKO-AT, the low levels of TMPRSS2 allow the virus to bypass the need for endosome maturation.

SARS-CoV-2 uses dynamin-dependent and -independent entry to infect ACE2-expressing MEF cells and both pathways are actin-dependent To confirm that the dynamin-dependent endocytosis of the S protein and SARS-CoV-2 virions reflected infectious entry pathway of the virus, infection assays were first performed in vehicle control and 4OH-TMX treated MEFTKO-ACE2 cells. Two clinical isolates of SARS-CoV-2 were tested, the ancestral Wuhan [73] virus and the more infectious Delta variant [77] (here referred to as Delta). In these experiments, we also investigated the role of the actin cytoskeleton using the actin depolymerising drug Lat-B. Because prolonged treatments with this drug result in loss of fibroblasts cell morphology and lead to cell detachment, we implemented a procedure to interfere with actin dynamics only during virus entry, without compromising cell attachment (Fig 6C). Six days after 4OH-TMX or vehicle control treatments, MEFTKO-ACE2 cells were treated with 3 μg/ml Lat-B for 15 min before infection. Lat-B was also present in the virus inoculum for 2 additional hours. The medium containing unbound virus and Lat-B was then removed and replenished with new medium containing 2 μM apilimod, to limit further infection after Lat-B removal (Fig 6C). The fraction of infected cells was determined by immunofluorescence imaging and image analysis at 20 hpi. In cells treated with 4OH-TMX for 6 days and DMSO, infection with Wuhan or Delta variants was reduced by up to 80% (Fig 6D–6F, DMSO), indicating that in these cells, dynamin-dependent endocytosis represents the main productive viral entry route. Actin depolymerization inhibited infection by more than 60% in control cells and this inhibitory effect was even stronger in dynamin depleted cells for both tested viruses (Fig 6D–6F, Lat-B). Thus, in ACE2-expressing MEFTKO cells, where TMPRSS2 is ether not expressed or not accessible to the virus, the infection of SARS-CoV-2 is mainly dynamin-dependent, and it is facilitated by the actin cytoskeleton. A similar dynamin- and actin-dependent entry mechanism has been described for other viruses, such as VSV [47], that have a similar size to coronaviruses (i.e. 100–120 nm in diameter including the spikes), and may not completely fit into dynamin-accessible endocytic invaginations. Actin polymerization facilitates the maturation of the endocytic cup, allowing the formation of the narrow membranous neck where dynamin binds and cleaves off the nascent vesicle [47]. Interestingly, similarly to other tested viruses but to a lower extent, increasing the virus load of SARS-CoV-2 Wuhan to an amount sufficient to infect 25% of the cells restored infection up to 12% in a virus-dose dependent manner (Fig 6G). Considering that 3–5% of these cells are not responsive to 4OH-TMX-induced depletion of dynamins, the remaining infection observed in dyn1,2,3-depleted cells indicated virus infection via dynamin-independent endocytosis. Thus, albeit with much lower efficiency compared to the dynamin-dependent entry, at high doses SARS-CoV-2 can also enter cells by an alternative endocytic mechanism.

Ultrastructural analysis of SARS-CoV-2 virus entry Ultrastructural TEM analysis of SARS-CoV-2 Wuhan entry in MEFTKO-ACE2 was performed by quantitative transmission electron microscopy (Fig 7). Viral particles were first allowed to bind at the cell surface for two hours on ice, at an equivalent MOI of 100. Cells were then shifted at 37°C for 15 minutes to allow endocytosis before fixation and TEM processing. Virions were readily identified at the cell surface of both vehicle control and 4OH-TMX pretreated cells (Fig 7A and 7E). Occasionally, electrondense filaments, with lengths consistent with the 20-nm size of the viral spikes, were observed connecting the virions and the PM (Fig 7A and 7E, white arrow heads, and inset in A). Similar to SFV, in vehicle- and 4OH-TMX-treated cells SARS-CoV-2 viruses were found distributed between CCPs and NCPs, as well as in large endosomal vesicles (Fig 7B, 7C, 7F, 7G and 7D). Quantitative analysis confirmed that upon dynamin depletion the number of viral particles inside CCPs and NCPs increased to more than 80% of the total endocytosed viruses, at the expenses of the virions found inside large endosomal structures (Fig 7H and 7I). Confirming our previous results, this analysis indicated that dynamin-dependent endocytosis is the main endocytic pathway for SARS-CoV-2. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. Ultrastructural analysis of SARS-CoV-2 entry in MEFTKOACE2 cells. Representative TEM images of MEFTKOACE2 cells treated with vehicle control (A-D) or 4OH-TMX (E-G, and D, 4OH-TMX) for 6 days and infected with SARS-CoV-2 Wuhan (MOI 100) on ice for 2 h followed by shift to 37°C for 15 min before fixation and processing for TEM. The fraction of total viral particles found in each of the described endocytic processes is quantified in H and I. Scale bar 100 nm. CCP = clathrin-coated profile, Mv = microvilli, PM = plasma membrane, asterisks (*) in panel A indicate an endocytic process. White arrowheads in panel A and E indicate electrondense filaments connecting the virions to the cell membrane. Boxed areas are magnified at bottom right corner of each figure panel. All values represent the mean and standard deviation of three replicas. Quantification of each treatment (EtOH vehicle ctrl or 4OH-TMX) includes 120 viral particles per condition. Statistical analysis was performed using ordinary two-way ANOVA multiple comparisons test (* p<0.05; ** p<0.01; ns = non significant). https://doi.org/10.1371/journal.ppat.1012690.g007

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1012690

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

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

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.