Mechanism of Viral Entry Mediated by the S Protein
A coronavirus contains four structural proteins, namely spike (S), envelope (E), membrane (M), and nucleocapsid (N) proteins. These proteins assemble around a lipid bilayer to provide the shell enclosing the viral genome (Figure 4A; Tang et al., 2020). Homotrimers of S protrude from the viral surface, and are densely decorated by N-linked glycans, creating the “crown” (“Corona” in Latin) that christens this virus group (Walls et al., 2016). S is a ∼180 kDa glycoprotein anchored in the viral membrane, which plays the most important roles in viral attachment, fusion and entry (Ou et al., 2020). Sequence analysis has shown that SARS-CoV-2 S protein shares 76% of the primary sequence with the corresponding S of human SARS-CoV (Ou et al., 2020). Accordingly, it has been early proposed that SARS-CoV-2 utilizes a similar cell entry mechanism as SARS-CoV, pivoted on S protein. This hypothesis has been confirmed from an experimental point of view. By using pseudotyped virus bearing SARS-CoV S or SARS-CoV-2 S, it was shown that a large panel of cell lines allows comparable entry of SARS-CoV or SARS-CoV-2 viruses (Hoffmann et al., 2020b; Walls et al., 2020).
FIGURE 4  (A) Structure of a Coronavirus. (B) Functional motifs in the sequence of the S “spike” protein of SARS-CoV and SARS-CoV-2. The S protein consists of ∼1300 aminoacids and it is composed by a N-terminal “S1”subunit (∼700 aa) and a C-terminal “S2”subunit (∼600 aa); binding to the host receptor is mediated by S1, whereas S2 induces fusion of the viral envelope with cellular membranes (Walls et al., 2017). S1 and S2 can be further subdivided in functional segments with different roles in viral entry (Figure 4B; Tang et al., 2020). The S1 subunit contains two subdomains, the N-terminal domain (NTD) and the C-terminal domain (CTD). In SARS-CoV (Li, 2015) and SARS-CoV-2 (Wang et al., 2020) CTD encloses the receptor-binding domain (RBD), and the RBD section that directly contacts the receptor is named as receptor-binding motif (RBM). The N-region of S2 contains a fusion peptide (FP), two heptapeptide repeat domains (HR1, HR2), a transmembrane domain (TM), and cytoplasmic peptide (CP). FP is a short segment composed of mostly hydrophobic residues, such as glycine (G) or alanine (A), which inserts in the host cell membrane to trigger the fusion event. HR1 and HR2 are composed of a repetitive heptapeptide with HPPHCPC sequence, where H represents hydrophobic or bulky residues, P polar or hydrophilic residues, and C charged residues. HR regions typically fold into α-helices with a hydrophobic interface that drives membrane fusion.
On the basis of the strong similarity between the S proteins of SARS-CoV and SARS-CoV-2, many researchers early set-out to demonstrate whether both viruses target the same host cell receptor, ACE2. Zhou et al. (2020) highlighted that the virus was able to infect cell lines only when they expressed human, bat, civet, and pig (but not mouse) ACE2 receptor. Hoffmann et al. (2020b), Ou et al. (2020), and Walls et al. (2020) elegantly outlined that the BHK cell line could be infected by pseudotyped SARS-CoV-2 or SARS-CoV only upon ACE2 expression. Conversely, the expression of different human receptors used by other CoVs (hDPP4 and APN, used by MERS CoV and HCoV-229E, respectively) did not enable pseudovirus access to cells (Hoffmann et al., 2020b). Taken together, these findings are solid evidence that SARS-CoV-2 engages ACE2 for cell entry.
Nonetheless, the two viruses were demonstrated by Xia to share also the membrane fusion mechanism, as strongly suggested by the impressive 89.9% sequence identity of S2 between SARS-CoV and SARS-CoV-2 (Xia et al., 2020a,b).
To date, the cell entry mechanism of SARS-CoV and SARS-CoV-2 has been understood in its general details and it is based on a concerted action of receptor binding and proteolysis of the S protein (Figure 5; Tang et al., 2020). Ultrastructural studies showed a metastable “prefusion” V-shaped trimer composed by three S1 heads sitting on top of a trimeric S2 stalk anchored into the virus membrane (Walls et al., 2016). The RBD constantly switches between a standing-up (“open”) position for receptor binding and a lying-down (“closed”) configuration, the latter allowing immune evasion (Figure 6; Song et al., 2018; Wrapp et al., 2020). Yet only one of the three RBD in trimeric S can flip up at a time and interact with the receptor (Wrapp et al., 2020). The second key feature of the fusion mechanism is “priming” by host proteases, which recognize and cleave a short peptide motif at the S1/S2 boundary (Figure 4B). This cleavage does not disassemble S1 from S2 in pre-fusion conditions (Belouzard et al., 2009), but the binding interaction of RBD with its receptor, accompanied by a further cleavage in a second site in S2 (S2’site, upstream of FP, Figure 4B), triggers the possible dissociation of S1 and the irreversible refolding of S2 into a “post-fusion” state (Figure 4B). In detail, HR1 undergoes a dramatic “jack-knife” conformational change, converting four helical stretches that run in an antiparallel fashion into a single long (∼130 aa) α-helix (Heald-Sargent and Gallagher, 2012; Walls et al., 2017). At first, three of these helices assemble into a homotrimeric bundle and stick the FP into the host cell membrane. Then, HR2 (one for each S2 chain) fold backward and bind to HR1, yielding the “six-helix bundle fusion core” (6-HB) of post-fusion S2 (Song et al., 2018). This conformational foldback brings the FP (at N-terminus of HR1) and the TM (at the C terminus of HR2) close to each other, so that the viral and host cell membranes approach until their outer leaflets merge (hemifusion, Figure 5). Eventually the inner leaflets merge (pore formation), enabling a connection between the interior of the virus and the host cell cytoplasm, that allows the delivery of viral genome (Figure 5; Tang et al., 2020).
FIGURE 5  Coronavirus viral fusion pathway model. Initially, the S protein is in the pre-fusion native state (1). Then S undergoes priming of the S1 subunit at S1/S2 by local proteases yielding the pre-fusion metastable state (2); note that priming at S1/S2 could also happen upon virus formation in releasing cell: in such a case the virus attaches to a host cell already in the pre-fusion metastable state (2). Subsequent triggering by a protease on S2’ enables the FP to insert in the host membrane upon the “jack-knife” transition of HR1 and HR2 yielding the pre-hairpin intermediate (3). The pre-hairpin folds back on itself due to HR1 and HR2 interactions eventually forming the post-fusion (6) state. During the S protein foldback, the two membranes approach each other until the outer leaflets merge (hemifusion) and eventually the inner leaflets merge (pore formation). Note that cell membrane may refer to plasma membrane (direct fusion) or endosomal membrane (fusion in endocytic vesicle). Adapted from Tang et al. (2020).
FIGURE 6  Trimeric S protein of SARS-CoV-2 in the ”Closed” and “Open” forms. Note the single RBD protruding out of the V-shaped conformation of the protein assembly. The structures have been drawn from PDB 6X2C (R. Henderson, 10.1101/2020.05.18.10208) by Mol on the PDB website. Although not directly related to ACE2, the role of S “priming” by host cell proteases deserves particular attention in the context of SARS-CoV-2 virus entry and tropism. Possibly, the most notable feature of SARS-CoV-2 genome, as compared to SARS-CoV and some related bat coronaviruses, is a four basic aminoacid insert (PRRA) at the S1/S2 junction (Figure 4B; Jaimes et al., 2020). This site is potentially cleavable by the protease furin, a proprotein convertase widely recognized to activate the fusion machinery of viral glycoprotein. Indeed, many authors showed that pseudoviruses bearing SARS-CoV-2 S were already “primed” (i.e., cleaved) at the S1/S2 boundary by furin upon assembly in the cell, at odds with pseudoviruses bearing SARS-CoV S (Shang et al., 2020a). SARS-COV-2 shows a large flexibility with regard to protease priming, which may independently occur by a) furin and furin-like proteases intracellularly, b) trypsin-like proteases such as TMPRSS2 that are present on the host cell membrane (particularly on airway epithelial cells), and 3) endosomal cathepsins activated by a drop in pH (e.g., cathepsin L) (Figure 7; Hoffmann et al., 2020a,b). This flexibility could be the crucial factor that explain SARS-CoV-2 cell tropism and the peculiar features of COVID-19 symptoms (Jaimes et al., 2020). Additionally, the kind of protease “priming” may determine whether the membrane fusion process occur directly at the plasma membrane or at endosomal level (Tang et al., 2020; Figure 7).
FIGURE 7  Relevance of S S1/S2 “priming” by host proteases for viral fusion to cells. The left cells produce viruses that can be “primed” by endogenous proteases such as furin (blue scissors); other viruses are not primed when they exit the cell. The primed viruses (marked by a yellow internal shadow) reach another cell (pathway A), where a membrane protease (e.g., TMPRSS2) may cleave the S2’ site (see Figure RB1b) leading to membrane fusion and delivery of viral RNA. Non-primed viruses can deliver their genome by two routes: in B, the virus reaches the cell, is primed on the membrane at both S1/S2 and S2’ by a local protease and then fuse with the plasma membrane; alternatively, in C the virus is internalized by endocytosis and priming/fusion occurs in endocytic vesicles. Note that also “primed” viruses may undergo pathway C, depending on their interaction with the recipient cell.