http://www.blackwell-synergy.com/links/doi/10.1111/j.1462-5822.2004.00389.x/full/#h5

Herpes simplex virus: receptors and ligands for cell entry
Patricia G. Spear

Summary

Entry of herpes simplex virus (HSV) into cells depends upon multiple
cell surface receptors and multiple proteins on the surface of the
virion. The cell surface receptors include heparan sulphate chains on
cell surface proteoglycans, a member of the tumor necrosis factor
(TNF) receptor family and two members of the immunoglobulin
superfamily related to the poliovirus receptor. The HSV ligands for
these receptors are the envelope glycoproteins gB and gC for heparan
sulphate and gD for the protein receptors and specific sites in
heparan sulphate generated by certain 3-O-sulfotransferases. HSV gC
also binds to the C3b component of complement and can block
complement-mediated neutralization of virus. The purposes of this
review are to summarize available information about these cell surface
receptors and the viral ligands, gC and gD, and to discuss roles of
these viral glycoproteins in immune evasion and cellular responses as
well as in viral entry.


HSV and disease Go to:  Choose Top of page HSV and disease <<
Characteristics of HSV an... Expression and properties... Interactions
of gC with h... Structural features of HS... Interactions of gD with
i... Potential significance of... Acknowledgements References

The most common manifestations of HSV infection are mucocutaneous
lesions, commonly called cold sores or fever blisters if they occur on
or near the lips. Such lesions can occur anywhere that the virus is
inoculated, however. Lesions on the fingers (herpes whitlow) used to
be an occupational hazard for dentists before the widespread use of
surgical gloves. Herpes genitalis refers to the lesions on genitalia
in sexually transmitted forms of disease. In addition, the cornea of
the eye can be infected to cause keratitis. These lesions usually
resolve without scarring and would not be a serious concern except
that they can recur frequently. The virus enters sensory and autonomic
neurons whose axons extend to the locale of the lesions and the virus
sets up latent infections from which it can periodically be
reactivated. Reactivated virus is transported back to the body surface
to cause recurrent lesions. Occasionally the virus can also be
transported to the central nervous system to cause encephalitis.
Fortunately, this occurs rarely but, unfortunately, the factors that
predispose to encephalitis are not known. Age must be one such factor
because HSV infections in newborn infants can result in severe
disseminated disease including neurological involvement. Two of the
most important cellular targets in HSV disease are epithelial cells of
skin and mucosa and neurons. Lymphocytes and other leucocytes can also
be infected, a phenomenon of unknown significance with respect to the
effectiveness of immune responses. Usually, spread of infection is by
cell-to-cell contact, not via a hematagenous or lymphatic route. Some
consideration will be given here to the requirements for viral entry
into leucocytes, epithelial cells and neurons.

There are two serotypes of HSV, HSV-1 and HSV-2. Lesions caused by
HSV-1 strains cannot be distinguished from those caused by HSV-2 but
there are distinct genetic and biological differences between members
of the two serotypes. For example, although both HSV-1 and HSV-2 can
infect either oral or genital sites, HSV-1 is more likely to
reactivate frequently from oral sites and HSV-2 is more likely to
reactivate from genital sites (Lafferty et al., 1987).

Characteristics of HSV and the herpesvirus family Go to:  Choose Top
of page HSV and disease Characteristics of HSV an... << Expression and
properties... Interactions of gC with h... Structural features of
HS... Interactions of gD with i... Potential significance of...
Acknowledgements References

The family is large and diverse but all members share certain features
in common. Each has a large double-stranded DNA genome encoding about
100-200 genes. In the virion, this genome is packed within an
icosahedral capsid displaying 162 capsomers or morphological elements.
The capsid is in turn coated with a layer of proteins called the
tegument, all of which are enclosed within a membrane composed of
lipids and more than a dozen viral proteins and glycoproteins. The
virion is designed to protect the viral genome from adverse conditions
in the extracellular environment and to permit cell invasion so that
the viral genome can be released to the cell nucleus for expression of
its genes. The outer membrane or envelope of the virion is its
infectivity organelle. Several of the glycoproteins in the envelope
are essential for viral entry into cells and others may influence this
process (Fig. 1).

The initial contact of HSV with a cell is believed to be binding of
the virion to glycosaminoglycan (GAG) chains of cell surface
proteoglycans. Heparan sulphate, one of several kinds of GAG, is
preferred and is considered to be a binding receptor, as opposed to an
entry receptor. Two of the virion glycoproteins, designated gB and gC,
are capable of binding to heparan sulphate and either appears to be
able to mediate the binding of virions to cell surface heparan
sulphate. Although this binding significantly enhances the efficiency
of HSV infection, it is not absolutely essential, at least not for the
infection of cultured cells. If gC is absent from the virion, specific
infectivity may be reduced as much as 10-fold as a result of reduced
efficiency of virus binding to cells (Herold et al., 1991), but
whether absence of gC has this effect is dependent on HSV serotype,
perhaps on HSV strain and on cell type (Cheshenko and Herold, 2002).
If both gB and gC are absent from the virion, binding to cells is
severely reduced (Herold et al., 1994). Infectivity is abolished also,
but this is in part because gB has an essential role in viral entry,
as outlined below. Deletion of the heparan sulphate-binding domain
from gB does not abrogate its essential role in entry (Laquerre et
al., 1998). If cells are devoid of GAGs, susceptibility to infection
is significantly reduced, but not abolished, unless entry receptors
for HSV are absent (Shieh et al., 1992; Gruenheid et al., 1993;
Banfield et al., 1995).

The interactions of HSV gB or gC with cell surface heparan sulphate
are also not sufficient for viral entry. Following binding of the
virus to the cell surface, cell entry requires that viral gD engage
any one of several entry receptors. Binding of gD to one of these
receptors triggers fusion of the viral envelope with a cell membrane,
and thus entry of the viral nucleocapsid and tegument into the cell
cytoplasm. This envelope-membrane fusion requires the action of other
viral envelope glycoproteins (gB and a heterodimer of gH-gL) in
addition to gD and the gD receptor (reviewed by Spear, 1993; Spear et
al., 2000).

The entry receptors discovered to date fall into three classes
(reviewed by Spear et al., 2000). They include HVEM (herpesvirus entry
mediator), a member of the TNF receptor family; nectin-1 and nectin-2,
members of the immunoglobulin superfamily; and specific sites in
heparan sulphate generated by certain 3-O-sulfotransferases. Mice can
be infected by HSV, and mouse and human forms of these receptors are
nearly indistinguishable in their HSV entry activity. HSV-1 and HSV-2
differ somewhat in receptor preferences. Whereas both HVEM and
nectin-1 are excellent entry receptors for both serotypes, nectin-2 is
virtually inactive for HSV-1 entry but does have weak entry activity
for HSV-2. The converse is true for 3-O-sulphate-modified heparan
sulphate.

Glycoproteins gB, gH and gL are structurally conserved among all
herpesviruses and probably have conserved essential roles in viral
entry. Glycoproteins gC and gD, on the other hand, are conserved among
most of the neurotropic alphaherpesviruses but have no recognizable
structural homologues in members of the other two branches of the
herpesvirus family.

Expression and properties of the HSV binding and entry receptors Go
to:  Choose Top of page HSV and disease Characteristics of HSV an...
Expression and properties... << Interactions of gC with h...
Structural features of HS... Interactions of gD with i... Potential
significance of... Acknowledgements References

Heparan sulphate is thought to be ubiquitously expressed, at least on
cells that stay put in tissues, as opposed to circulating cells of the
immune system. This does not necessarily mean that binding sites in
heparan sulphate for gB, gC or gD have the same distribution as
heparan sulphate. Heparan sulphate chains are synthesized as repeating
disaccharide units of N-acetyl-glucosamine and glucuronic acid and
then modified, in some regions of the chain but not others, by a
sequence of enzymatic reactions including de-acetylation of the
glucosamine, sulphation of the amino group, epimerization of the
glucuronic acid to iduronic acid, and O-sulphations at the 2-OH
position in the iduronic acid and the 6-OH and 3-OH positions in the
amino sugar (reviewed by Lindahl et al., 1998). These reactions
generate regions in heparan sulphate (approximately 6-12 residues)
that differ with respect to positions of epimerized uronic acid and
sulphate groups and that can bind proteins with great specificity. The
site in heparan sulphate to which gD can bind is generated by the
action of enzymes that can yield an octasaccharide of the following
structure: UA-GlcNS-IdoUA2S-GlcNAc-UA2S-GlcNS-IdoUA2S-GlcNH23S6S (Liu
et al., 2002). The demonstration that gD could bind to this sequence
in heparan sulphate built on findings that an expression plasmid
encoding mouse 3-O-sulfotransferase-3B could convert resistant Chinese
hamster ovary (CHO) cells to susceptibility to HSV-1 entry (Shukla et
al., 1999). Some isoforms of 3-O-sulfotransferases can generate
gD-binding sites (provided the other relevant enzymatic activities are
expressed) whereas others generate antithrombin-binding sites (Liu et
al., 1999; Shukla et al., 1999; Xia et al., 2002). The structures in
heparan sulphate to which gB and gC bind have not yet been determined.
However, it is clear from competition studies that gB and gC bind to
different structures and that HSV-1 and HSV-2 forms of gC also bind to
different structures (Gerber et al., 1995; Trybala et al., 2000).
Determination of the distribution of binding sites for gB, gC and gD
in heparan sulphate will require use of specific viral probes for
these sites. It cannot be done just by determining where various
enzymes are expressed, because it is not yet clear what influences the
overall coordination of their activities to generate specific sites.

HVEM (also known as HveA, ATAR, TR2, TNFRSF-14) is expressed in a
variety of tissues and cell types, including T and B lymphocytes,
other leucocytes, epithelial cells and fibroblasts, but probably not
in neurons (Montgomery et al., 1996; Hsu et al., 1997; Kwon et al.,
1997; Marsters et al., 1997). HVEM is the principal receptor for HSV
entry into activated human T lymphocytes, but not for a number of
other human cell types (Montgomery et al., 1996). The viral ligand for
HVEM is gD (Whitbeck et al., 1997) whereas the natural cellular
ligands for HVEM are LIGHT and lymphotoxin- (Mauri et al., 1998).
LIGHT is also a ligand for the lymphotoxin- receptor (LTR). Studies of
the normal physiological roles of HVEM have focused on immune
responses and on LIGHT as the functional ligand (lymphotoxin- binding
is of lower affinity and of unknown significance). These studies
(reviewed by Croft, 2003; Granger and Rickert, 2003) have been aided
by the fact that patterns of expression for HVEM and LTR are largely
non-overlapping. For both mice and humans, HVEM is expressed in a
variety of tissues and cell types, as mentioned above, whereas LTR
expression is more restricted. With regard to immune responses, it is
relevant that HVEM, but not LTR, is abundantly expressed on NK-T cells
and naive CD8+ T cells and is expressed at variable levels on some
CD4+ T cells and dendritic cells. On the other hand, LTR is expressed
on the stromal cells of lymphoid organs and is critical for lymphoid
organ development. LIGHT expression is tightly regulated and is
detected on activated lymphocytes, NK cells and immature dendritic
cells. In vitro studies have shown that binding of LIGHT to HVEM can
provide a second signal for T cell activation. These signals are
transmitted via cytoplasmic TNF receptor-associated factors (TRAFs)
and lead to activation of NFB or JNK/AP-1. Several in vivo studies,
including some with LIGHT gene K/O mice and LIGHT gene transgenic
mice, have demonstrated that LIGHT-HVEM interactions contribute to
CTL-mediated immune responses, allograft rejection and
graft-versus-host disease.

Nectin-1 and nectin-2 are also expressed in a variety of tissues and
cell types, including epithelial cells, fibroblasts and neurons. The
cDNAs for nectin-1 and nectin-2 were originally described in the
literature as encoding poliovirus receptor-related proteins 1 and 2
(Prr1 and Prr2) (Eberléet al., 1995; Lopez et al., 1995). Because they
were not poliovirus receptors, they were renamed HveC and HveB
respectively (Geraghty et al., 1998; Warner et al., 1998). They were
then renamed nectin-1 and nectin-2, because Y. Takai and colleagues
had discovered their roles in cell adhesion. By now, at least four
nectins and other nectin-like molecules have been identified (reviewed
by Takai and Nakanishi, 2003). These cell surface glycoproteins are
members of the immunoglobulin superfamily and are closely related to
each other and to the poliovirus receptor. The nectins are cell
adhesion molecules that can co-localize with cadherins in adherens
junctions and can also engage in cell adhesions independent of
cadherins. The nectins dimerize in the plane of the membrane and
nectin homodimers engage in trans interactions with the same member or
other members of the nectin family on adjacent cells, or with HSV gD.
These trans interactions probably signal various intracellular events.
A variety of extracellular and intracellular signaling molecules, such
as scatter factor/hepatocyte growth factor, Ras, Cdc42 and Rac small G
proteins, regulate the formation and disruption of cell junctional
complexes in dynamic fashion. Nectin-1, -2 and -3 are broadly
expressed in a variety of tissues and cells. Nectin-4 appears to be
localized to the placenta in humans. Only nectin-1 and nectin-2 have
been shown to mediate HSV entry, through interactions with the viral
ligand, gD (Cocchi et al., 1998; Geraghty et al., 1998; Krummenacher
et al., 1998; Warner et al., 1998; Lopez et al., 2000).

Human and mouse mutations have provided some information about the
physiological roles of nectin-1 and nectin-2. A few kindreds of humans
with truncating mutations in the nectin-1 gene have been identified
(Suzuki et al., 2000). The phenotypes associated with loss of both
functional alleles include cleft lip/palate, hidrotic ectodermal
dysplasia, hair abnormalities, developmental defects of the hands and,
in some cases, mental retardation. Unfortunately, it has not been
possible to obtain sera from such individuals to test for evidence of
HSV infection. Fibroblasts and lymphoblastoid cells are available from
the Coriell Mutant Repository and exhibit the phenotypes expected with
respect to HSV entry. The fibroblasts expressed HVEM and nectin-2, but
not nectin-1, and were more resistant to HSV infection than wild-type
control cells. At high doses of virus, the cells could be infected
with appropriate virus strains, however, presumably via HVEM or
nectin-2 (F. Struyf and P.G. Spear, unpublished studies). It is
somewhat surprising that the phenotype of the nectin-1(-/-) genotype
is not more pronounced, especially with respect to neurological
defects. Nearly every neuron of the mouse peripheral and central
nervous systems expresses nectin-1 mRNA (Haarr et al., 2001) and Takai
and colleagues have defined junctions in the brain that involve
nectin-1/nectin3 trans interactions (Mizoguchi et al., 2002). The only
phenotype associated with knock-out of nectin-2 in the mouse is male
sterility (Bouchard et al., 2000). The cause is absence of appropriate
nectin-2/nectin-3 trans interactions between Sertoli cells and
spermatids (Mueller et al., 2003). The absence of more severe
phenotypes when nectin-1 or nectin-2 are not expressed, coupled with
co-expression of these proteins in many cell types, suggests that
there may be some redundancies of function in cells that normally
express both.

Interactions of gC with heparan sulphate and the C3b component of
complement Go to:  Choose Top of page HSV and disease Characteristics
of HSV an... Expression and properties... Interactions of gC with h...
<< Structural features of HS... Interactions of gD with i... Potential
significance of... Acknowledgements References

All alphaherpesviruses studied to date express a member of the gC
family that has both heparan sulphate-binding and C3b-binding
activities. As mentioned above, this glycoprotein is dispensable for
the replication of virus in cultured cells, despite its role in virus
binding to heparan sulphate. HSV strains obtained from patients almost
invariably express gC, however, and it is clear that gC can have a
role in viral virulence, as outlined below.

Protection from complement neutralization

The C3b-binding activity of gC is associated with protection of virus
from antibody-independent neutralization by complement components
(reviewed by Friedman, 2003). Both HSV-1 and HSV-2 retain full
infectivity in the presence of complement whereas mutants unable to
express gC are neutralized by complement, even in the absence of
anti-HSV antibodies (McNearney et al., 1987; Hidaka et al., 1991;
Gerber, et al., 1995; Friedman et al., 1996). Neutralization of
gC-negative virus requires complement activation and presence of C5
but does not require C6, C8 or factor D, indicating that the
alternative complement pathway and complement components beyond C5 are
not required (Friedman et al., 2000). Both HSV-1 gC and HSV-2 gC (gC-1
and gC-2) have been shown to bind to C3, C3b, iC3b, C3c, but not to
C3d (Tal-Singer et al., 1991; Kostavasili et al., 1997). HSV-1, but
not HSV-2, gC can block the binding of C5 and properdin to C3b (Fries
et al., 1986; Hung et al., 1994; Kostavasili et al., 1997). This
gC-1-specific blocking cannot provide a full explanation for the
ability of both serotypes to resist complement-mediated
neutralization. Some species specificity for interactions of
complement with alphaherpesvirus gCs has been reported (Huemer et al.,
1993). Also, it was reported that human and guinea pig sera, but not
mouse sera, could neutralize HSV-1 infectivity in the absence of
anti-HSV antibodies (Hidaka et al., 1991; Huemer et al., 1993).
However, compelling evidence was presented that HSV-1 mutants, unable
to express gC or altered in ability to bind to C3, were attenuated in
wild-type mice but exhibited near wild-type virulence in C3-deficient
mice (Lubinski et al., 1998; 1999).

gC and HSV virulence

Although gC is dispensable for the replication of HSV in cultured
cells, it can have a role in HSV virulence in the intact animal.
Reports on the effects of gC deletions on HSV virulence are mixed. For
example, deletion of gC was found to have no effects on HSV-1
virulence in the mouse after corneal, intraperitoneal or intracerebral
inoculation (Minagawa et al., 1997). On the other hand, deletion of gC
abrogated HSV-1 replication in human skin implanted into SCID-hu mice
(Moffat et al., 1998), attenuated virulence in a rabbit seizure model
(Stroop and Schaefer, 1989) and attenuated virulence in mice
inoculated intradermally (zosteriform model), provided the mice
expressed the C3 component of complement (Lubinski et al., 1998;
1999). HSV-2 mutants have not been as extensively studied for
pathogenesis. An HSV-2 mutant with a deletion of 130 bp in the gC open
reading frame was reported to cause disease similar to that of
wild-type virus after intravaginal or intracerebral inoculation
(Johnson et al., 1986).

Structural features of HSV-1 and HSV-2 gC and mapping of binding
sites Go to:  Choose Top of page HSV and disease Characteristics of
HSV an... Expression and properties... Interactions of gC with h...
Structural features of HS... << Interactions of gD with i... Potential
significance of... Acknowledgements References

Members of the gC family are type I membrane glycoproteins with both
N-linked and O-linked glycans. gC-1 and gC-2 exhibit 65% identity in
amino acid sequence with most of the divergence in the N-terminal
region. This region in both gC-1 and gC-2 is mucin-like, with numerous
O-linked glycans, but relatively low conservation of amino acid
sequence. Part of this region is missing in gC-2. For gC-1, amino acid
substitutions in basic and hydrophobic residues between amino acids
129 and 160 and at position 247 (Fig. 2) significantly impaired
binding to heparan sulphate and also impaired binding of virus to
cells (Mardberg et al., 2001). Mutations that can reduce binding of
gC-2 to heparan sulphate have not been identified. Linker-insertion
mutations and amino acid substitutions in several regions of gC-1 and
gC-2 (Fig. 2) impaired binding to C3b (Seidel-Dugan et al., 1990; Hung
et al., 1992). It is evident from Fig. 2 that there is some overlap in
regions of gC-1 that are critical for binding to both heparan sulphate
and C3b. In fact, heparan sulphate can inhibit the binding of gC-1 to
C3b (Rux et al., 2002).

Deletion of amino acids 275-367 in gC-1 abrogated C3b binding whereas
deletion of amino acids 33-123 did not (Hung et al., 1992). However,
the latter deletion reduced somewhat the affinity of gC-1 for heparan
sulphate (Rux et al., 2002) and also prevented gC from blocking the
binding of C5 or properdin to C3b (Hung et al., 1994; Kostavasili et
al., 1997). Interestingly, HSV-1 mutants with deletion 275-367 in gC-1
were able to induce the secondary lesions occurring by zosteriform
spread in inoculated mice, but only in C3-deficient mice, not in
wild-type mice. The other deletion (33-123) had less effect and the
double deletion caused no more impairment of viral virulence in
wild-type mice than that observed with the 275-367 deletion alone.
Thus, the authors concluded that the C3b-binding activity of gC-1, not
its ability to block the binding of C5 or properdin to C3b, was most
important for protecting virus from complement-mediated inactivation
in vivo (Lubinski et al., 1999).

Although attention was first focused on gC for its ability to bind to
heparan sulphate and mediate the binding of virus to cells,
experimental evidence available to date suggests that the C3b-binding
activity of gC, conferring protection against complement inactivation,
may be more important for the contribution of gC to HSV virulence.
There are models of HSV disease in which gC has a critical role, as
described above, and in which the involvement of complement has not
yet been assessed. Clearly, more must be done to characterize these
models and also to investigate a possible relationship between heparan
sulphate binding and protection from complement.

Interactions of gD with its receptors Go to:  Choose Top of page HSV
and disease Characteristics of HSV an... Expression and properties...
Interactions of gC with h... Structural features of HS... Interactions
of gD with i... << Potential significance of... Acknowledgements
References

All alphaherpesviruses studied to date, except varicella-zoster virus
(the cause of chicken pox and shingles), express a member of the gD
family. It seems likely that members of the gD family encoded by
different herpesviruses have conserved the ability to bind to members
of the nectin family. Nectin-1 is highly conserved among mammalian
species (Shukla et al., 2000; Milne et al., 2001). HSV and bovine and
porcine herpesviruses can all use human, mouse and porcine forms of
nectin-1 for cell entry (Geraghty et al., 1998; Shukla et al., 2000;
Milne et al., 2001), in part explaining the broad host ranges of these
viruses, at least with respect to cultured cells. In the case of HSV,
the other known gD receptors are nectin-2, HVEM and specific sites in
heparan sulphate generated by 3-O-sulphotransferases.

Structure of gD and interface with HVEM

X-ray structures of a large portion of the HSV-1 gD ectodomain,
crystallized alone and in complex with a portion of HVEM (Carfi et
al., 2001), have revealed the following: (i) the core of the gD
ectodomain is an Ig-fold with unconventional disulphide bonding; (ii)
an N-terminal extension from the Ig-fold assumes a hairpin shape in
the complex with HVEM and has all the residues that make contact with
HVEM; (iii) this N-terminal extension is disordered in the crystals of
gD alone; (iv) a domain of gD downstream of the Ig-fold forms an
-helix which is sandwiched between the N-terminal hairpin (or
disordered region) and the Ig-fold. Residues in HVEM (Connolly et al.,
2002) and in the N-terminal hairpin of HSV-1 gD that are critical for
HVEM-gD interactions have been determined by mutational analyses, as
will be discussed more fully below for gD. Although the structure of
HSV-2 gD has not been determined, mutational analyses indicate that
HSV-1 and HSV-2 gD are very similar in their interactions with HVEM
and nectin-1.


Effects of gD mutations on physical and functional interactions with
entry receptors
The approach has been to introduce the desired mutations into the gD
open reading frame and then to test ability of the mutant gDs (i) to
bind to the various receptors and (ii) to participate with gB and
gH-gL in inducing membrane fusion. The former assay uses soluble forms
of gD, in some cases hybrids of the gD ectodomain fused to the Fc of
rabbit IgG, for quantification of their ability to bind to cells
transfected to express each of the relevant receptors (Geraghty et
al., 2000; 2001). For the latter assay, cells expressing gB, gH-gL,
various forms of wild-type or mutant gD (effector cells) are mixed
with target cell populations expressing each of the entry/fusion
receptors to quantify cell fusion activity (Pertel et al., 2001). It
is clear from these studies that the structural features of gD
critical for functional interactions with the various entry receptors
are distinct, as summarized below:

(i)  Deletion of amino acids 7-32, comprising all the HVEM contact
residues revealed in the X-ray structure (Fig. 3), eliminated physical
and functional interactions of gD-1 or gD-2 with all receptors (HVEM,
nectin-2 and 3-O-sulphated heparan sulphate) except nectin-1 (Yoon et
al., 2003). This applies to both the mouse and human forms of the
receptors. Binding of these mutant forms of gD-1 and gD-2 to mouse or
human nectin-1 and cell fusion activities with nectin-1 were
indistinguishable from those of wild-type gD-1 and gD-2, indicating
that the major contact site for nectin-1 in gD is downstream of amino
acid 32.

(ii)  Certain insertions or deletions in gD-1 eliminated binding and
activity with HVEM alone or with both HVEM and nectin-1, without
affecting expression of the gD-1 mutants on the cell surface (Jogger
et al., 2003). The former insertions are in the N-terminal region and
their effects are predictable from the X-ray structure of the gD-HVEM
complex. The latter insertions and deletion are in the major
alpha-helix mentioned above or in the E beta strand of the Ig fold
(Fig. 3). It seems likely that the latter mutations indirectly affect
the conformation of the HVEM contact sites in the N-terminal hairpin
and also affect the conformation of the nectin-1 contact sites, which
are undoubtedly different but have not yet been precisely defined.

(iii)  The N-terminal region of gD-1 or gD-2 is required for
functional interactions with nectin-2, but not nectin-1, and the
actual amino acid sequence within the first 53 amino acids determines
whether nectin-2 can be recognized as a cell fusion or entry receptor.
Nectin-2 is a very poor entry/fusion receptor for gD-1 but more active
for gD-2. Hybrids of gD-1 and gD-2 containing the first 53 amino acids
from gD-2 (only 6 amino acid differences from the gD-1 sequence)
resembled gD-2 in cell fusion activity with nectin-2 (Zago and Spear,
2003). Single or double substitutions into the gD-1 sequence were not
as effective at conferring activity with nectin-2 as all 6
substitutions. Alternatively, certain single amino acid substitutions
in gD-1, at positions conserved between gD-1 and gD-2 (L25P, Q27P,
Q27R, L28A), can enable gD-1 to use nectin-2 as a receptor for entry
and cell fusion (Warner et al., 1998; Lopez et al., 2000; Connolly et
al., 2003; Yoon et al., 2003). The L25P, Q27P and Q27R mutations in
gD-2 have no effect on the ability of gD-2 to mediate
nectin-2-dependent cell fusion (Yoon et al., 2003). Various lines of
evidence indicate that nectin-2 interacts with the wild-type form of
gD-2 at much lower affinity than does nectin-1. It seems likely that
the major contact site on gD for nectin-2 is downstream of amino acid
32, as is true for nectin-1, but that there may be secondary contact
sites in the N-terminal region that are critical for interaction with
nectin-2 but not nectin-1. Alternatively, the N-terminus may affect
the conformation of other domains of gD-2 in a way that is critical
for interactions with nectin-2 but not with nectin-1.


Preliminary results indicate that multiple amino acid substitutions
downstream of amino acid 32 in gD-1 or gD-2 can abrogate functional
and physical interactions with nectin-1 and nectin-2 but not with HVEM
(C. Jogger, S. Manoj, D. Myscofsky and P.G. Spear, unpublished data).
By introducing various gD mutations into the viral genome, it should
be possible to obtain viral mutants capable of using only nectin-1 or
only HVEM as an entry receptor. These mutants will be invaluable for
assessing the role of each receptor in infection of certain target
cells in cultured cells and in experimental animals and for studies
designed to determine whether interactions of gD with each receptor,
either during viral entry or viral replication, influences the cell
response to virus infection.

Potential significance of alternative usage of receptors by HSV Go
to:  Choose Top of page HSV and disease Characteristics of HSV an...
Expression and properties... Interactions of gC with h... Structural
features of HS... Interactions of gD with i... Potential significance
of... << Acknowledgements References

The ability of HSV-1 and HSV-2 to use multiple receptors for entry
into cells could in part reflect simple redundancy but there may be
greater significance to the existence of multiple receptors. First,
these viruses may use different receptors to enter different cell
types, governed by the natural patterns of receptor expression. For
example, primary human T lymphocytes express HVEM, but not nectin-1,
and HVEM can be used for HSV entry into these cells (Montgomery et
al., 1996; Geraghty et al., 1998). Cells of neuronal origin express
nectin-1, but not HVEM (Montgomery et al., 1996; Geraghty et al.,
1998; Mauri et al., 1998), and it was reported that nectin-1 could
mediate HSV entry into rat neurons (Richart et al., 2003). Second,
different receptors could be used for entry into the same cell type
under different conditions. Epithelial cells express nectin-1,
nectin-2 and HVEM (Geraghty et al., 1998; Warner et al., 1998) but the
nectins can be localized to junctions and remain inaccessible to virus
until the junctions are disrupted (Yoon and Spear, 2002). Therefore,
entry of virus into cells of an intact epithelium may require either
damage to the epithelium or use of a receptor such as HVEM, instead of
nectin-1. On the other hand, cell-to-cell spread of HSV infection in
an intact epithelium could depend on use of one of the nectins because
exit of virus from polarized cells is directed to lateral surfaces
(Johnson et al., 2001) where nectins would be localized. In fact, it
has been shown that nectins can enhance the cell-to-cell spread of HSV
infection in cell monolayers, as assessed by plaque size (Sakisaka et
al., 2001). Third, single amino acid substitutions in HSV-1 or HSV-2
gD can alter receptor usage, as described above, suggesting that
natural polymorphisms could influence the cell types infected and
therefore pathogenesis. Fourth, interaction of gD with one of its
receptors (either in the same cell or on an adjacent cell) may
transduce signals or interfere with signal transduction via the
natural ligand. For example, cell junctions could be modified or
appropriated by virus, through interactions of gD with one of the
nectins (Krummenacher et al., 2003). Also, the normal responses of
leucocytes to foreign antigen or to inflammatory cytokines could be
affected by interactions of gD with HVEM. It has been reported that
cells expressing gD can inhibit T cell responses (La et al., 2002).
The viral mutants altered in gD may permit investigation of some of
these possibilities.