Infection of Glial Cells by the Human Polyomavirus JC Is Mediated by an N-Linked Glycoprotein Containing Terminal α(2-6)-Linked Sialic Acids

The human JC polyomavirus (JCV) is the etiologic agent of the fatal central nervous system (CNS) demyelinating disease progressive multifocal leukoencephalopathy (PML). PML typically occurs in immunosuppressed patients and is the direct result of JCV infection of oligodendrocytes. The initial event in infection of cells by JCV is attachment of the virus to receptors present on the surface of a susceptible cell. Our laboratory has been studying this critical event in the life cycle of JCV, and we have found that JCV binds to a limited number of cell surface receptors on human glial cells that are not shared by the related polyomavirus simian virus 40 (C. K. Liu, A. P. Hope, and W. J. Atwood, J. Neurovirol. 4:49–58, 1998). To further characterize specific JCV receptors on human glial cells, we tested specific neuraminidases, proteases, and phospholipases for the ability to inhibit JCV binding to and infection of glial cells. Several of the enzymes tested were capable of inhibiting virus binding to cells, but only neuraminidase was capable of inhibiting infection. The ability of neuraminidase to inhibit infection correlated with its ability to remove both α(2-3)- and α(2-6)-linked sialic acids from glial cells. A recombinant neuraminidase that specifically removes the α(2-3) linkage of sialic acid had no effect on virus binding or infection. A competition assay between virus and sialic acid-specific lectins that recognize either the α(2-3) or the α(2-6) linkage revealed that JCV preferentially interacts with α(2-6)-linked sialic acids on glial cells. Treatment of glial cells with tunicamycin, but not with benzyl N-acetyl-α-d-galactosaminide, inhibited infection by JCV, indicating that the sialylated JCV receptor is an N-linked glycoprotein. As sialic acid containing glycoproteins play a fundamental role in mediating many virus-cell and cell-cell recognition processes, it will be of interest to determine what role these receptors play in the pathogenesis of PML.Approximately 70% of the human population worldwide is seropositive for JC virus (JCV). Like other polyomaviruses, JCV establishes a lifelong latent or persistent infection in its natural host (40, 49, 50, 68, 72). Reactivation of JCV in the setting of an underlying immunosuppressive illness, such as AIDS, is thought to lead to virus dissemination to the central nervous system (CNS) and subsequent infection of oligodendrocytes (37, 40, 66, 68). Reactivation of latent JCV genomes already present in the CNS has also been postulated to contribute to the development of progressive multifocal leukoencephalopathy (PML) following immunosuppression (19, 48, 55, 70, 75). Approximately 4 to 6% of AIDS patients will develop PML during the course of their illness (10). In the CNS, JCV specifically infects oligodendrocytes and astrocytes. Outside the CNS, JCV genomes have been identified in the urogenital system, in the lymphoid system, and in B lymphocytes (2, 17, 18, 30, 47, 59). In vitro, JCV infects human glial cells and, to a limited extent, human B lymphocytes (3, 4, 39, 41, 42). Recently, JCV infection of tonsillar stromal cells and CD34⁺ B-cell precursors has been described (47). These observations have led to the suggestion that JCV may persist in a lymphoid compartment and that B cells may play a role in trafficking of JCV to the CNS (4, 30, 47).Virus-receptor interactions play a major role in determining virus tropism and tissue-specific pathology associated with virus infection. Viruses that have a very narrow host range and tissue tropism, such as JCV, are often shown to interact with high affinity to a limited number of specific receptors present on susceptible cells (26, 44). In some instances, virus tropism is strictly determined by the presence of specific receptors that mediate binding and entry (7, 16, 27, 35, 46, 53, 56, 67, 73, 74, 76). In other instances, however, successful entry into a cell is necessary but not sufficient for virus growth (5, 8, 45, 57). In these cases, additional permissive factors that interact with viral regulatory elements are required.The receptor binding characteristics of several polyomaviruses have been described. The mouse polyomavirus (PyV) receptor is an N-linked glycoprotein containing terminal α(2-3)-linked sialic acid (12–14, 22, 28). Both the large and small plaque strains of PyV recognize α(2-3)-linked sialic acid. The small-plaque strain also recognizes a branched disialyl structure containing α(2-3)- and α(2-6)-linked sialic acids. Neither strain recognizes straight-chain α(2-6)-linked sialic acid. The ability of the large- and small-plaque strains of PyV to differentially recognize these sialic acid structures has been precisely mapped to a single amino acid in the major virus capsid protein VP1 (21). The large-plaque strains all contain a glycine at amino acid position 92 in VP1, and the small-plaque strains all contain a negatively charged glutamic acid at this position (21). In addition to forming small or large plaques, these strains also differ in the ability to induce tumors in mice (20). This finding suggests that receptor recognition plays an important role in the pathogenesis of PyV.The cell surface receptor for lymphotropic papovavirus (LPV) is an O-linked glycoprotein containing terminal α(2-6)-linked sialic acid (26, 33, 34). Infection with LPV is restricted to a subset of human B-cell lines, and recognition of specific receptors is a major determinant of the tropism of LPV for these cells (26).Unlike the other members of the polyomavirus family, infection of cells by simian virus 40 (SV40) is independent of cell surface sialic acids. Instead, SV40 infection is mediated by major histocompatibility complex (MHC)-encoded class I proteins (5, 11). MHC class I proteins also play a role in mediating the association of SV40 with caveolae, a prerequisite for successful targeting of the SV40 genome to the nucleus of a cell (1, 63). Not surprisingly, SV40 has been shown not to compete with the sialic acid-dependent polyomaviruses for binding to host cells (15, 26, 38, 58).Very little is known about the early steps of JCV binding to and infection of glial cells. Like other members of the polyomavirus family, JCV is known to interact with cell surface sialic acids (51, 52). A role for sialic acids in mediating infection of glial cells has not been described. It is also not known whether the sialic acid is linked to a glycoprotein or a glycolipid. In a previous report, we demonstrated that JCV bound to a limited number of cell surface receptors on SVG cells that were not shared by the related polyomavirus SV40 (38). In this report, we demonstrate that virus binding to and infection of SVG cells is dependent on an N-linked glycoprotein containing terminal α(2-3)- and α(2-6)-linked sialic acids. Competitive binding assays with sialic acid-specific lectins suggest that the virus preferentially interacts with α(2-6)-linked sialic acids. We are currently evaluating the role of this receptor in determining the tropism of JCV for glial cells and B cells.     

A Putative α-Helical Structure Which Overlaps the Capsid-p2 Boundary in the Human Immunodeficiency Virus Type 1 Gag Precursor Is Crucial for Viral Particle Assembly

The capsid (CA) and nucleocapsid domains of the human immunodeficiency virus type 1 Gag polyprotein are separated by the p2 spacer peptide, which is essential for virus replication. Previous studies have revealed that p2 has an important role in virus morphogenesis. In this paper, we show that a crucial assembly determinant maps to the highly conserved N terminus of p2, which is predicted to form part of an α-helix that begins in CA. A mutational analysis indicates that the ability of the N terminus of p2 to adopt an α-helical structure is essential for its function during virus assembly. To prevent CA-p2 processing, it was necessary to mutate both the CA-p2 cleavage site and an internal cleavage site within p2. Virions produced by the double mutant lacked a conical core shell and instead contained a thin electron-dense shell about 10 nm underneath the virion membrane. These results suggest that p2 is transiently required for proper assembly, but needs to be removed from the C terminus of CA to weaken CA-CA interactions and allow the rearrangement of the virion core shell during virus maturation.The internal structural proteins of the human immunodeficiency virus type 1 (HIV-1) virion are synthesized in the form of a polyprotein (Pr55^gag) which can efficiently form enveloped virus-like particles even when expressed alone (17). Pr55^gag is modified by N-terminal myristylation, which is required for its stable association with the inner leaflet of the plasma membrane, where virus assembly occurs (4, 21). During or after the release of an immature particle from the plasma membrane, Pr55^gag is cleaved by the viral protease. The major Gag cleavage products are matrix (MA), capsid (CA), nucleocapsid (NC), and p6 (25, 34). MA, which has a crucial role in the incorporation of the viral surface glycoproteins (10, 52), remains associated with the host cell-derived lipid envelope of the virion (16). CA forms the shell of the characteristic cone-shaped core of the mature virion which encloses the viral genomic RNA (16, 27). NC is essential for the encapsidation of the viral genome and is believed to coat the viral RNA within the core of the virion (2, 19, 30). The C-terminal p6 domain of Pr55^gag facilitates the release of assembled viral particles from the cell surface (20) and is also needed for the incorporation of the regulatory viral protein Vpr (31, 39).Within the context of Pr55^gag, two spacer peptides, p2 and p1, are located between CA and NC and between NC and p6, respectively (24, 25). Cleavage between CA and p2 is much slower than that between p2 and NC or between MA and CA (41). As a consequence, a CA-p2 protein (p25) accumulates in virus-producing cells (34). However, CA-p2 is normally found only in trace amounts in virions. In addition to p2, which comprises 14 amino acids (Ala-363 through Met-376) of the HIV-1_HXB2 Gag precursor, a 10-amino-acid p2 fragment which extends from Ser-367 through Met-376 has been isolated from HIV-1 virions, indicating that the viral protease can also cleave within p2 (24, 25).Genetic analyses indicate that the region surrounding the CA-p2 boundary has an important role in particle assembly (21, 28, 50). Within CA, the N-terminal two-thirds forms a domain which appears dispensable for particle assembly but is required for the formation of the cone-shaped core of the mature virion (8, 44, 51). Recent structure determinations have revealed that the N-terminal HIV-1 CA domain is largely α-helical (18, 35). An exposed loop region between two α-helices interacts with the prolyl isomerase cyclophilin A (14), which leads to the incorporation of the cellular enzyme into virions (13, 48). The C-terminal third of CA forms a distinct domain which is essential for Gag oligomerization and particle assembly (8, 12, 44). While genetic and structural studies indicate that the N-terminal boundary of the CA assembly domain coincides with a uniquely conserved sequence, termed the major homology region (8, 15, 18, 32), its C-terminal boundary remains less well defined.The replacement of the scissile dipeptide Leu-Ala at the CA-p2 boundary with Ser-Arg in a mutant designated SVC-C2 led to the formation of grossly distorted capsid structures and caused a significant reduction in particle yield, indicating that the very C terminus of CA and/or p2 is crucial for HIV-1 morphogenesis (21). The possibility that the CA assembly domain extends into p2 is also suggested by the finding that the precise deletion of p2 from Pr55^gag markedly reduced particle production (28). Electron microscopy revealed an accumulation of large electron-dense plaques underneath the plasma membrane in the absence of p2 (28), a phenotype which is similar to that observed for the SVC-C2 cleavage site mutant (21). However, the role of p2 in virus assembly remains controversial, because its removal appeared to have no effect on particle release in another study (41).In the present study, we focused on the N-terminal portion of p2, since it is considerably more conserved than the C terminus and because it is predicted to be part of an α-helix which begins in CA. The analysis of a panel of single-amino-acid changes shows that the conserved N terminus of p2 is essential for virus replication and indicates that its predicted α-helical conformation is crucial for virus assembly. In contrast, a deletion which removed 5 out of 10 amino acids between a previously reported cleavage site within p2 and NC delayed but did not abolish virus replication, demonstrating that this relatively variable region of p2 has no essential function in the viral life cycle. We also show that processing of CA-p2 can be essentially prevented by disrupting both the CA-p2 cleavage site and the reported Met-Ser site (25) within p2. Interestingly, the mutant particles often contained a prominent circular structure underneath the viral membrane, indicating that the presence of p2 at the C terminus of CA prevented the rearrangement of the core into a conical tube.