African Swine Fever Virus Gets Undressed: New Insights on the Entry Pathway

ABSTRACT

African swine fever virus (ASFV) is a large, multienveloped DNA virus composed of a genome-containing core successively wrapped by an inner lipid envelope, an icosahedral protein capsid, and an outer lipid envelope. In keeping with this structural complexity, recent studies have revealed an intricate entry program. This Gem highlights how ASFV uses two alternative pathways, macropinocytosis and clathrin-mediated endocytosis, to enter into the host macrophage and how the endocytosed particles undergo a stepwise, low pH-driven disassembly leading to inner envelope fusion and core delivery in the cytoplasm.

INTRODUCTION

Most viruses enter host cells by exploiting one or more of the multiple cellular endocytic mechanisms, such as clathrin-mediated endocytosis (CME), caveola- and lipid raft-dependent endocytosis, macropinocytosis, or phagocytosis (1). Endocytosis allows the incoming viruses to overcome physical barriers, like the plasma membrane and the underlying cortical cytoskeleton, besides bringing them closer to their replication site. Thus, the endocytosed particles traffic through a complex and dynamic vesicular network that crosses the cytoplasm from the periphery to the perinuclear area of the host cell. In order to transfer the viral genome and associated proteins to the cytoplasm, the endocytosed particles must pass over a second barrier, the vesicular membrane. To this aim, nonenveloped viruses use certain proteins capable of inducing membrane lysis or pore formation. Conversely, enveloped viruses invariably fuse their viral envelopes with the endosomal membranes by using fusogenic surface proteins. For both kinds of viruses, the activation of these penetration mechanisms usually involves structural rearrangements and partial virus disassembly triggered by specific cellular cues, such as the host receptor engagement, the exposure to the acidic endosomal pH, or the activity of cellular proteases (1–3). After membrane penetration, the released nucleoprotein cores are transported to their replication sites, either at the cytoplasm or at the nucleus, where they usually become uncoated to release a replication-competent viral genome. Overall, virus uncoating represents not merely a final phase of the entry pathway but also the result of a stepwise disassembly program triggered by specific protein, chemical, and mechanical cues acting at different entry stages, from the cell surface to the replication sites (2, 3).

In this Gem, I describe recent advances in understanding the entry pathway of African swine fever virus (ASFV), a complex, multienveloped DNA virus that uses an intricate uncoating program to gain access to the replication site. ASFV is the causative agent for a highly lethal hemorrhagic disease of domestic pigs for which there is no vaccine or antiviral strategy available (4). ASFV is the sole member of the family Asfarviridae, which belongs to the group of nucleocytoplasmic large DNA viruses (NCLDV). It infects mainly swine monocytes and macrophages and replicates within discrete perinuclear viral factories. The ASFV genome, a double-stranded DNA molecule of 170 to 190 kbp, encodes more than 150 polypeptides. The virus particle, of icosahedral morphology with a diameter of about 200 nm, contains more than 50 proteins. The viral architecture consists of a genome-containing nucleoid surrounded by a thick protein layer referred to as the core shell, which is successively wrapped by an inner lipid envelope, a protein icosahedral capsid, and an outer lipid membrane (5). The assembling particle acquires its inner lipid envelope at the virus factories from membrane fragments derived from the endoplasmic reticulum. Conversely, the outer membrane is obtained by budding at the plasma membrane during virus egress. Both intracellular and extracellular ASFV forms are infectious (5).

According to early studies, ASFV enters host cells by receptor-mediated endocytosis, transits along the endolysosomal pathway, and penetrates in a low pH-dependent manner after a presumed fusion event (6, 7). In keeping with these results, later studies identified classical clathrin- and dynamin-dependent endocytosis as the primary entry route for ASFV (8, 9). On the other hand, it has also been established that ASFV induces its own uptake by macropinocytosis (10), a nonselective endocytosis involving fluid-phase uptake mediated by actin-driven membrane ruffles (1). Notably, each study excluded the alternative mechanism. These contradictory pieces of evidence have been related to the employment of different and, in many cases, indirect methodologies to assess virus uptake and to the usage of different target cell types and non-highly purified virus preparations (11).

ASFV ENTERS HOST MACROPHAGES BY TWO ALTERNATIVE ENDOCYTIC PATHWAYS

We have recently revisited the entry pathway of extracellular ASFV into the swine macrophage, its main target cell, by using highly purified virus stocks and a combination of flow cytometry, fluorescence microscopy, electron microscopy (EM), live-cell imaging, targeted small interfering RNA (siRNA) interference, and pharmacological inhibitors of virus entry (12). Using a direct flow cytometry-based assay, we found that the uptake of fluorescent extracellular virions was significantly reduced in the presence of inhibitors of macropinocytosis {5-(N-ethyl-N-isopropyl) amiloride [EIPA], p21-activated kinase inhibitor III [IPA-3], and cytochalasin D} and clathrin-mediated endocytosis (chlorpromazine, pitstop2, and dynasore). Similar inhibitory effects were obtained when viral gene expression was analyzed, thus indicating that both endocytic pathways significantly contribute to a productive infection. Consistent with these results, EM analysis detected ASFV particles within clathrin-coated pits (Fig. 1A and B) and coated vesicles or engulfed by large membrane protrusions (Fig. 1A and C) and inside macropinosome-like vesicles. That ASFV uses macropinocytosis is not surprising since this is a highly activated, constitutive process in macrophages related to antigen presentation. However, our results do not support ASFV triggering macropinocytosis as previously described (10). In fact, classical cues for this mechanism, such as the induction of membrane ruffling, the activation of Pak 1 kinase, or the increase in fluid-phase uptake, were not significantly detected (12). The ASFV uptake by clathrin-dependent endocytosis is somewhat unexpected, considering that this pathway is usually restricted to small- and intermediate-sized viruses (1), which can fit well in coated vesicles, whose diameter usually ranges from 50 to 100 nm. Interestingly, extracellular ASFV particles, with a 200-nm diameter, are detected in enlarged (∼250 nm) clathrin-coated vesicles, which indicates that these can be deformed to adapt to larger cargo sizes, as previously described for the bullet-shaped vesicular stomatitis virus (180 by 70 nm) (13). The usage of multiple endocytic mechanisms, such as CME and macropinocytosis, is intriguing, but not exclusive for ASFV, as it has been reported for Ebola, influenza, and foot-and-mouth disease viruses. Besides macrophages and monocytes, ASFV also infects secondary target cells like vascular endothelial cells, hepatocytes, or epithelial cells. Thus, it is possible that, by exploiting different entry alternatives, ASFV increases its ability to infect different target cells and to adapt to the changing conditions of the infection process.

FIG 1.

Working model for ASFV entry and uncoating. (A) ASFV can enter swine macrophages by clathrin-mediated endocytosis and actin-driven macropinocytosis. Once endocytosed, incoming particles transit from early endosomes and macropinosomes to late multivesicular endosomes, where they undergo a low pH-driven, stepwise disassembly process. It involves disruption of the capsid (light blue), dissociation of the outer membrane (purple), and fusion of the inner viral envelope (pink) with the limiting endosomal membrane (dark blue). As a consequence, naked cores are delivered into the cytosol, where they presumably undergo further uncoating to release replication-competent genomes. Membrane fusion and core delivery depend on inner envelope protein pE248R. Electron micrographs illustrate clathrin-mediated endocytosis (B) and macropinocytosis (C) of extracellular ASFV particles, an intact virion within an early endosome (D) a disrupted particle (without capsid and with a fragmented, partially detached outer envelope) within a late endosome (E), and a cytosolic naked core (F). Bars, 100 nm.

ASFV ENDOCYTIC TRAFFICKING AND DISASSEMBLY

Upon endocytosis, incoming ASFV particles move along the entire endolysosomal pathway, from peripheral early endocytic/macropinocytic vesicles, containing Rab5 and EEA1 markers, to perinuclear late endocytic compartments and lysosomes (Rab7⁺, CD63⁺, Lamp1⁺, cathepsin L⁺) (12, 14). Interestingly, virus trafficking from early to late endocytic vesicles seems to be essential for a productive infection in swine macrophages. Thus, inhibitors of endosome maturation like wortmannin, a phosphatidylinositol 3 (PI3)-kinase inhibitor that blocks early endosome fusion, nocodazole, which disturbs microtubule-dependent endosomal transport, and bafilomycin A1 (Baf A1), an inhibitor of the vacuolar H⁺/ATPase pump that prevents endosomal acidification, impair ASFV infection (12). Furthermore, dominant-negative overexpression and knockdown approaches indicate that the small GTPase Rab7, a key regulator of late endosome maturation, is required for ASFV infection (12, 14).

Interestingly, during the endocytic transport, ASFV undergoes a disassembly process that is part of its uncoating program (12). Thus, whereas the incoming particles within early endocytic vesicles appear nearly intact (Fig. 1D), most of the particles within multivesicular late endosomes lack the protein capsid. Moreover, a significant proportion of these particles also lack the outer lipid envelope or it appears broken and detached (Fig. 1A and E). Importantly, cell treatment with Baf A1 prevents virus disassembly, whereas the in vitro exposure of purified viruses to low pH (5.0) mimics the observed disassembly of the endocytosed virions. Taken together, these results strongly indicate that the acidic pH of late endocytic endosomes/macropinosomes provides the uncoating chemical cue that triggers ASFV disassembly. In this respect, low pH has been described to induce directly the desencapsidation of foot-and-mouth disease virus during entry, probably due to the repulsive electrostatic interactions raised across the interface of their capsid subunits (15). Whether capsid protonation might induce a similar effect in ASFV awaits a detailed structural elucidation.

MEMBRANE FUSION AND CORE DELIVERY

Once disrupted, the virus particles expose the inner envelope, which allows its interaction and subsequent fusion with the limiting membrane of late endosomes. As a result, naked cores are delivered (Fig. 1A and F) in perinuclear cytoplasmic areas (12).

Interestingly, core deposition may be dependent on a major structural change linked not to endocytosis but to virus maturation. During ASFV morphogenesis, the assembling core is tightly attached to the inner viral membrane, whereas in extracellular mature virions and incoming particles it is clearly detached (Fig. 1D and E) (5, 12). This alteration would enable the fusion of a “free” inner viral membrane and the consequent release of the viral core. It is thought that core attachment depends on N-myristoylation of viral polyprotein pp220, a precursor of four major structural proteins that form the core shell (16). Following the proteolytic processing of pp220 precursor by ASFV protease, the mature core would be separated from the inner membrane. In support of this, core detachment is not detected when the expression of the viral protease is prevented (17). Therefore, a proteolytic event linked to the virus maturation would be significant for the fusion of the endocytosed viruses.

Importantly, core delivery depends on viral protein pE248R, an N-myristoylated transmembrane protein located at the inner membrane (18). With use of a conditional lethal mutant, it has been shown that virus assembly and egress occur normally in the absence of pE248R expression. However, the pE248R-defective virus progeny is noninfectious due to a postentry blockage (18). When analyzed at the electron microscope level, incoming pE248R-defective particles become disrupted as normal virions but fail to penetrate into the cytosol and accumulate in lysosomes (12). It is worth noting that pE248R polypeptide shares a number of features, including sequence similarity, N-myristoylation, and type II membrane topology, with the vaccinia virus (VACV) protein L1, a component of the poxviral multiprotein entry/fusion complex (EFC), which is required for membrane fusion and core deposition in the cytoplasm (19).

CONCLUSIONS AND FUTURE DIRECTIONS

At present, cytosolic cores represent the last morphologically detectable stage in ASFV entry. Although they look significantly larger (by 25%) than those within intact virions, they keep the same overall structure: a dense, genome-containing nucleoid wrapped by a thick protein core shell (Fig. 1F). Therefore, it is expected that they undergo further disassembly to uncoat the viral DNA, a process that, as occurs with other viruses, might be triggered by cellular cues like the ubiquitin/proteasome system (2, 3).

Despite the major differences in virus structure, the extracellular forms of ASFV and VACV share remarkable similarities in relation to the uptake and uncoating pathways. The enveloped VACV (EV) particles consist of an inner nucleoprotein core wrapped by two consecutive lipoprotein membranes. EV particles lose their outermost membrane upon interaction with cell surface glycosaminoglycans or by low acid-driven rupture after macropinocytic uptake. Then, incoming particles fuse their internal membranes at late macropinosomes to release cores in the cytoplasm (20). Interestingly, the analogies between asfiviruses and poxviruses might also be extended to the fusion machinery. As mentioned above, ASFV membrane protein pE248R features functional and sequence similarities to those of VACV L1R protein, one of the 11 nonglycosylated proteins of the EFC, an inner membrane-embedded macromolecular assembly required for fusion and core penetration (12, 19). Moreover, a second ASFV transmembrane protein, pE199L, displays sequence similarity with the poxviral EFC components G9, A16, and J5 proteins. Thus, it is tempting to speculate that an entry/fusion multiprotein complex may also be present in the inner membrane of ASFV. Remarkably, viral orthologs of the poxviral EFC components have been found in other NCLDVs such as iridoviruses, phycodnaviruses, ascoviruses, and mimiviruses (21). Since NCLDVs are thought to be derived from a common ancestor, it will be relevant to address to what extent the complex fusion machinery described in poxviruses might constitute a common element among NCLDVs.

In summary, the available data indicate that extracellular ASFV particles may use clathrin-dependent endocytosis and macropinocytosis to initiate productive infections in host macrophages. Upon endocytosis, ASFV particles undergo a stepwise, low pH-driven uncoating process involving the loss of the two outermost layers, the subsequent fusion of the inner envelope with the limiting late endosomal membrane, and the delivery of naked cores in the cytosol (Fig. 1A). This general picture awaits further research to address the molecular mechanisms of key events as membrane fusion and genome uncoating. Also, it will be important to investigate the entry pathway of the intracellular, single-enveloped ASFV form, which is infectious and may contribute to virus spread. Future insights may also provide new cell- and virus-specific targets for the design of antiviral strategies.