Human Herpesvirus-6 Entry into Host Cells

Huamin Tang; Yasuko Mori


Future Microbiol. 2010;5(7):1015-1023. 

In This Article

HHV-6 Entry

Cell Tropism of HHV-6

Initially, HHV-6 was designated as human B lymphotropic virus;[3] however, both HHV-6A and HHV-6B replicate most efficiently in CD4+ T lymphocytes in vitro.[37] Other human cells have also been reported to be infected by HHV-6, including natural killer cells,[38] differentiated liver cells,[39] epithelial cells,[40] endothelial cells,[41] fetal astrocytes[42] and dendritic cells.[43]

All HHV-6A and HHV-6B isolates effectively infect activated peripheral or cord blood mononuclear cells, although their abilities to infect T-cell lines are different.[4,44] The GS strain, one of the two most widely used strains of HHV-6A, is commonly propagated in the HSB-2 T-cell line, and the other, the U1102 strain, is often propagated in the J JHAN T-cell line.[7] For HHV-6B, the Z29 and HST strains grow well in the Molt-3 and MT4 T-cell lines.[45] The host tissue range for HHV-6 in vivo is even broader than in vitro, and includes brain tissue,[46–48] liver tissue,[49–51] salivary glands,[52] tonsillar tissue[53] and endothelium.[41] Bone-marrow progenitors were reported to harbor latent HHV-6, which may be transmitted longitudinally to the differentiated blood cells of different lineages.[19] The species susceptible to HHV-6 infection are limited, and to date no rodent model for HHV-6 infection exists. Nevertheless, a serological survey of monkeys conducted by Higashi et al. showed antibodies to HHV-6 and neutralizing antibodies against HHV-6 infection.[54] In addition, experimental infections of different monkey species and macaque tissues with HHV-6 have been achieved.[55,56]

Endocytosis & HHV-6 Entry

In an immunoelectron microscopic analysis of the early events of HHV-6 binding and internalization (using the GS strain of HHV-6A and the BA92 strain of HHV-6B), Cirone et al. demonstrated that viral internalization occurs through smooth-surfaced pits and vesicles. They could not observe fusion of the virions with the cell plasma membrane. Moreover, no envelope proteins were seen on the cell plasma membrane at any time during internalization, and the internalized virions were still surrounded by their envelopes. The treatment of the target cells with chloroquine, a drug that disrupts the endocytic pathway, almost completely inhibited the viral infectivity. These findings suggest HHV-6 enters its target cells through the endocytic pathway and not by fusion with the plasma membrane.[57]

Cellular Receptor for HHV-6

As mentioned previously, receptor–ligand binding is a critical step for virus entry. Santoro et al. reported that the human CD46 molecule is a receptor for HHV-6, based on observations that CD46 is downregulated by HHV-6 infection, that HHV-6 infection is inhibited by an anti-CD46 antibody or by soluble CD46, and that nonhuman cells that express recombinant human CD46 become susceptible to HHV-6 infection.[32] CD46 is a ubiquitous immunoregulatory receptor that is expressed on all nucleated cells. Initially discovered as a widely expressed C3b- and C4b-binding protein, CD46 was subsequently shown to act as a cofactor for serine protease factor I to inactivate C3b and C4b, which are opsonins and components of convertases, by limited proteolysis, thereby protecting host cells from damage by complement factors.[58]

In recent years, CD46 has been reported to be a cellular receptor for a large variety of microorganisms, including measles virus, all species of B adenoviruses except 3 and 7, Streptococcus pyogenes, Neisseria gonorrhoea and Neisseria meningitides in addition to HHV-6, earning it the moniker, 'pathogen magnet'.[59] However, different binding domains of CD46 are used in measles compared with HHV-6 infection.[60] A typical CD46 molecule contains four short consensus repeats (SCRs), a serine–threonine–proline-rich domain, a sequence of unknown significance, a transmembrane region, and a cytoplasmic sequence. The CD46 gene has 14 exons, which are highly spliced to create alternative serine-threonine-proline-rich domain, transmembrane region, and cytoplasmic sequence domain sequences, but the SCR domains are invariant.[61] SCR1 and SCR2 are reported to form the binding site for the measles virus.[62] For HHV-6, using molecular chimeras of CD46 and CD55, Greenstone et al. reported that SCR2 and SCR3 are the binding domains for HHV-6,[60] and using deletion mutants of CD46, Mori et al. found that SCR2, SCR3 and SCR4 are essential for the cell fusion induced by HHV-6 infection.[63] HSV-1 is reported to use different receptors to infect different cell types.[64] It is necessary to determine whether other receptors also exist for HHV-6, because, as described below, much evidence supports the possibility of alternative receptor usage during HHV-6 infection.

Viral Glycoproteins & HHV-6 Entry

Unlike most other enveloped viruses, for example HIV-1, which use only one or two proteins for viral entry, herpesviruses use a series of proteins for successful entry. The details of the entry process have largely been elucidated for HSV-1. HSV-1 uses gC and gB to attach to cell-surface glycosaminoglycans and gD as a viral ligand that binds to cellular receptors; finally, this binding mediates envelope–membrane fusion using the additional proteins gH/gL and gB.[65–67]

A number of glycoproteins are embedded in or associated with the envelope of HHV-6.[68,69] Among them, gH, gL, gM, gN and gB are conserved in all herpesviruses, and the functions of these five glycoproteins during herpesvirus entry have recently been investigated. Anderson et al. reported that, in HHV-6, gH and gL form a heterodimeric glycoprotein complex using disulfide bridges,[70] and Santoro et al. identified gH as the viral component responsible for binding CD46, since an anti-gH antibody can immunoprecipitate CD46 and an anti-CD46 antibody can coimmunoprecipitate gH from HHV-6-infected cells.[71] A neutralizing antibody against the HHV-6A gB has been reported.[72] gB is the most highly conserved glycoprotein in the herpesvirus family. HHV-6 gB may function in the attachment step and virus–cell fusion during viral entry, as it does in HSV-1.[73] Little is known about the roles played by gM and gN during HHV-6 infection.

In HHV-6, gO is the gene product of U47, the positional homolog of the HCMV gO gene. gO is conserved only in β-herpesviruses. HCMV gO interacts with the gH/gL complex and appears to be important for HCMV entry into fibroblasts or its spreading between them,[74] and a pH-independent cell-surface fusion pathway was reported when HCMV entered fibroblasts.[75] Disruption of HCMV gO results in a small-plaque phenotype,[76] which could be explained by the accumulation of viral capsids in the cytoplasm and the reduced virus release when gO gene was deleted.[77] The function of gO in murine cytomegalovirus (MCMV) entry has been reported. gO knockout mutants of MCMV enter the fibroblasts via an energy-dependent and pH-sensitive pathway however, entry is neither energy dependent or pH-sensitive in wild-type.[78] The chaperone function of gO has also been reported in the TR strain of HCMV, which promotes gH/gL incorporation into HCMV virions.[79,80] HHV-6 gO also interacts with the gH/gL complex, but the functions of this complex in HHV-6 infection still need to be elucidated, because the gH/gL/gO complex does not bind the HHV-6 receptor, CD46.[81] It may bind another cellular molecule to contribute to HHV-6 infection (Figure 1) or function in the virion maturation process, as it does in HCMV.

Figure 1.

Human herpesvirus-6 entry into target cells.
Two glycoprotein complexes exist on the envelope of the human herpesvirus-6 virion. Only the gH/gL/gQ1/gQ2 complex of human herpesvirus-6A binds the cellular receptor, CD46. The gH/gL/gO complex may bind an unknown factor that also contributes to viral entry. The binding of the virus to a target cell recruits CD46 to lipid rafts, where the virus enters the cell via endocytosis.

The gQ proteins are unique to HHV-6. The mRNAs are transcribed from the U97, 98, 99, and 100 genes of HHV-6A, a highly spliced region that encodes a series of 80- and 74-kDa glycoproteins.[82,83] Only the 80-kDa product of these genes is coimmunoprecipitated with an anti-gH antibody, and was designated gQ1–80K.[84] In HHV-6A-infected cells, another gQ gene product, of 74 kDa, exists, but it does not interact with the gH/gL complex and is not incorporated into the HHV-6 virion. The functions of this product during HHV-6 infection remain unknown. Interestingly, the gQ gene was found to encode another, smaller glycoprotein of 37 kDa, called gQ2 (gQ2–37K). Therefore, the larger gQ proteins were named gQ1–80K and gQ1–74K, respectively. As with gQ1, a second gQ2 gene product, of 34 kDa (gQ2–34K), is found in HHV-6-infected cells, but not in HHV-6 virions.[83] Endoglycosidase digestion showed that all the gQs are sensitive to PNGase F (an amidase that removes high mannose, hybrid and complex oligosaccharides from N-linked glycoproteins) digestion and therefore are glycosylated with N-linked glycans. However, only gQ1–80K and gQ2–37K are glycosylated with complex N-linked glycans, which means that only these two forms of gQ are glycosylated in the Golgi apparatus. Using an anti-gH antibody, Mori et al. demonstrated that only gQ1–80K and gQ2–37K form a complex with gH/gL in HHV-6-infected cells and HHV-6 virions and that this gH/gL/gQ1/gQ2 complex interacts with the cellular receptor, CD46, to initiate virus entry (Figure 1).[83] Interestingly, there is also an unique glycoprotein complex, gH/gL/UL128–131 in HCMV, which is required for the entry into epithelial and endothelial cells through a pH-dependent endocytosis pathway.[85,86] The prevailing model of HCMV is that the gH/gL/gO complex promotes the entry into fibroblasts by cell-surface fusion, and gH/gL/UL128–131 promotes the entry into epithelial and endothelial cells by an endocytosis pathway. As for HHV-6, it still needs to be elucidated whether it also uses either of two glycoprotein complexes, gH/gL/gO and gH/gL/gQ1/gQ2, for different cell type entry. However, a novel HCMV entry model has also been suggested by Wille et al. As described above, gO functioned as a chaperone for the incorporation of gH/gL into HCMV virions in the clinical strain, TR. Simultaneously, the same group reported that gO-null mutant virus exhibited defects in entering not only into fibroblasts, but also epithelial and endothelial cells. These data suggest that the gH/gL complex but not the gH/gL/gO complex, mediated entry into fibroblasts and that the gH/gL complex, is also required for the entry into epithelial and endothelial cells. A similar function may also exist in HHV-6, but this remains to be investigated. In addition, a number of questions regarding the gH/gL/gQ1/gQ2 complex remain to be answered: which part of the complex contains the epitope that binds CD46? Is complex formation really required for CD46 binding? Do the N-linked glycans on gQ1 and gQ2 contribute to the formation of the complex? Why can only gQ1 and gQ2, which are glycosylated with complex N-linked glycans, interact with the gH/gL complex? Do the N-linked glycans contribute to the receptor binding? In addition, although the HHV-6A gH/gL/gQ1/gQ2 complex is clearly a viral ligand for CD46, the role of this complex in HHV-6B still needs to be investigated. The complex of HHV-6B (HST strain) cannot bind to recombinant CD46 [Mori Y, Unpublished Data]. However, the other strains of HHV-6B (PL-1 strain) can use CD46,[87] indicating that there may be differences in affinity towards CD46 between variants, and even between strains within a variant. Another point of note is that there may be additional receptors specific for variants or strains.[87] The distinct receptor usage between both variants may affect the different tropism. Therefore, it will be interesting to research the additional cellular factor(s) and their roles.

Lipid Rafts & HHV-6 Entry

Lipid rafts are unique microdomains in the plasma membrane of the cell surface. They are enriched with sphingolipids and cholesterol.[88] Under physiological conditions, they are small and associated with numerous cellular processes, such as membrane trafficking and signaling.[88,89] In addition, essential roles of lipid rafts for viral entry have been reported for several viruses.[90,91] Huang et al. found that when the HHV-6 envelope was depleted of cholesterol, viral infection was significantly reduced, as was virus-induced cell fusion.[92] These results suggest that viral cholesterol plays an important role in HHV-6 infection. Cholesterol depletion of the HHV-6 envelope did not affect viral binding, although ligand–receptor interaction was reduced.[92] Similarly, when cholesterol was removed from the target-cell surface, reduced infection was seen, and viral binding was only slightly affected. Interestingly, the cellular receptor (CD46) and viral ligand protein (gQ1) accumulated in lipid rafts after virus attachment. Collectively, these findings show that both cellular and viral lipid rafts play an important role in HHV-6 infection, and suggest that HHV-6 may enter its target cells through a lipid raft-associated mechanism.[93] Interestingly, Kawabata et al. have recently shown that a significant proportion of envelope glycoproteins are enriched in raft fractions from HHV-6-infected cells and also demonstrated that the raft component is expressed in HHV-6 virions, suggesting that the lipid rafts function during the HHV-6 virion maturation process.[94]

Fusion from without & HHV-6 Infection

Fusion from without (FFWO) is a specific phenomenon of some virus infections.[95,96] In this process, cell fusion is induced by a virus in the absence of viral protein synthesis. The detailed analysis of this mechanism has been performed in HSV-1. gB is a determinant of FFWO during HSV-1 infection.[97] HHV-6 can also induce FFWO.[63,98] Interestingly, HHV-6A (strain U1102) induces marked FFWO in target cells, but FFWO is quite rare in HHV-6B (strain HST) infections. That is, although HHV-6B (HST) can induce cell fusion when it infects MT4 cells, in which HHV-6B can replicate, its ability to induce cell fusion in other cell lines, such as SupT1 and 293T cells, is very poor even at 103–105 50% tissue culture infective doses (TCID50).[63] Surprisingly, Pedersen et al. reported that a different HHV-6B strain, PL-1, can effectively induce cell fusion in HEK293 and SupT1 cells at 182 and 23 TCID50, respectively.[98] The detailed mechanisms underlying such different inter- and intrastrain characteristics still need to be elucidated. The functions of the glycoproteins of HHV-6 in FFWO have also been analyzed. Monoclonal antibodies to gB and gH of HHV-6A inhibit this event, indicating that the FFWO induced by HHV-6A requires these two glycoproteins. In addition, CD46, the receptor for HHV-6, is required for this cell–cell fusion.[63]