Peptide inhibition of human cytomegalovirus infection.

RESEARCH Open Access
Peptide inhibition of human cytomegalovirus
infection
Lilia I Melnik, Robert F Garry, Cindy A Morris*
Abstract
Background: Human cytomegalovirus (HCMV) is the most prevalent congenital viral infection in the United States
and Europe causing significant morbidity and mortality to both mother and child. HCMV is also an opportunistic
pathogen in immunocompromised individuals, including human immunodeficiency virus (HIV)- infected patients
with AIDS, and solid organ and allogeneic stem cell transplantation recipients. Current treatments for HCMV-
associated diseases are insufficient due to the emergence of drug-induced resistance and cytotoxicity, necessitating
novel approaches to limit HCMV infection. The aim of this study was to develop therapeutic peptides targeting
glycoprotein B (gB), a major glycoprotein of HCMV that is highly conserved across the Herpesviridae family, that
specifically inhibit fusion of the viral envelope with the host cell membrane preventing HCMV entry and infection.
Results: Using the Wimley-White Interfacial Hydrophobicity Scale (WWIHS), several regions within gB were
identified that display a high potential to interact with lipid bilayers of cell membranes and hydrophobic surfaces
within proteins. The ability of synthetic peptides analogous to WWIHS-positive sequences of HCMV gB to inhibit
viral infectivity was evaluated. Human foreskin fibroblasts (HFF) were infected with the Towne-GFP strain of HCMV
(0.5 MOI), preincubated with peptides at a range of concentrations (78 nm to 100 μM), and GFP-positive cells were
visualized 48 hours post-infection by fluorescence microscopy and analyzed quantitatively by flow cytometry.
Peptides that inhibited HCMV infection demonstrated different inhibitory concentration curves indicating that each
peptide possesses distinct biophysical properties. Peptide 174-200 showed 80% inhibition of viral infection at a
concentration of 100 μM, and 51% and 62% inhibition at concentrations of 5 μM and 2.5 μM, respectively. Peptide
233-263 inhibited infection by 97% and 92% at concentrations of 100 μM and 50 μM, respectively, and 60% at a
concentration of 2.5 μM. While peptides 264-291 and 297-315, individually failed to inhibit viral infection, when
combined, they showed 67% inhibition of HCMV infection at a concentration of 0.125 μM each.
Conclusions: Peptides designed to target putative fusogenic domains of gB provide a basis for the development
of novel therapeutics that prevent HCMV infection.
Introduction
Human cytomegalovirus (HCMV) is a ubiquitous oppor-
tunistic pathogen that belongs to the Betaherpesviridae.
The virulence of this pathogen is directly linked to the
immune status of its host. Primary HCMV infection is
generally asymptomatic in immunocompetent individuals,
although it causes a mononucleosis-like syndrome in
some. After primary HCMV infection, the virus establishes
lifelong latency and periodically reactivates with notable
pathological consequences. In contrast, HCMV infection
in immunocompromised patients such as AIDS patients
and solid organ and allogeneic stem cell transplantation
recipients causes serious disease [1]. Primary infection of
women during or right before pregnancy with HCMV is
the most common cause of congenital viral infection lead-
ing to significant morbidity and mortality. Congenital
HCMV infection is also associated with spontaneous abor-
tion, premature delivery, intrauterine growth restriction
(IUGR), and pre-eclampsia. The risk of primary infection
in a seronegative mother is 1 to 4%, which carries a 30 to
40% risk of congenital infection [2,3]. The majority of con-
genitally infected babies are asymptomatic at birth; how-
ever, 10 to 17% subsequently develop hearing defects or
neurodevelopmental sequelae [4]. Although the most
* Correspondence: cmorris2@tulane.edu
Graduate Program in Biomedical Sciences and Department of Microbiology
and Immunology, Tulane University, 1430 Tulane Avenue, New Orleans, LA,
70112 USA
Melnik et al. Virology Journal 2011, 8:76
http://www.virologyj.com/content/8/1/76
© 2011 Melnik et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

serious clinical sequelae are seen in cases where a mother
acquires a primary infection during pregnancy, down-
stream side effects are also seen in cases where latent
HCMV is reactivated [5] and where a mother is reinfected
with a different strain of the virus [6].
HCMV has a double-stranded DNA genome of 235 kb
encoding approximately 165 genes [7]. It has a very
broad cellular tropism resulting in potential infection of
nearly every organ system. The ability of HCMV to
enter a wide range of cell types involves a complex
interaction between several viral envelope glycoproteins
and host cell surface receptors, although the entry of
herpesviruses into host cells is still poorly understood.
The HCMV virion envelope contains at least 20 virus-
encoded glycoproteins that are involved in cell attach-
ment and penetration [8]. Of these, glycoprotein B (gB)
is the most abundant glycoprotein [9] and is highly con-
served among the Herpesviridae [10]. Glycoprotein B
plays a critical role in the HCMV entry process. Initially,
gB along with gM/gN, is involved in tethering of virions
to heparan sulfate proteoglycans (HSPG) on the surface
of host cells. The short interaction of HCMV with
HSPG is followed by more stable interactions with one
or more viral cellular receptors, namely epidermal
growth factor receptor (EGFR) [11], platelet-derived
growth factor receptor (PDGFR) [12], and toll-like
receptor TLR-2 [13]. Glycoprotein B also interacts with
integrin avb3, a coreceptor that enhances HCMV entry
[14]. Integrins are known to synergise with EGFR as
well as with other receptors to activate signal transduc-
tion pathways [15-17]. To complete the entry process,
both viral and cellular membranes fuse, allowing the
release of virion-associated tegument and capsid pro-
teins into the cytoplasm. This final step of viral entry
into host cells requires gB and the gH/gL complex
[18-21].
Antibodies to HCMV gB have been shown not only to
block penetration of virions into cells, but also to limit
cell-to-cell infection, implying that gB plays a role in vir-
ion penetration into cells, cell-to-cell transmission, as
well as fusion of infected cells [20,22]. Recently, Isaacson
and coworkers used genetic complementation to con-
firm that gB is required for the fusion of viral and cellu-
lar membranes, virus entry, and cell-to-cell spread of
HCMV [23]. The importance of gB for viral infection
suggests that this viral envelope protein may be a
rational target for novel drug design.
HCMV infection is highly prevalent in the population
due to the ability of the virus to efficiently transmit
between hosts that harbour and periodically shed the
virus. HCMV is transmitted through direct exposure to
infected bodily secretions, including saliva, urine and
breast milk. Following infection, HCMV enters the
bloodstream and spreads to various organs including
kidney, liver, spleen, heart, brain, retina, esophagus,
inner ear, lungs, colon, and salivary glands [24]. The
ability of HCMV to infect a wide variety of cell types is
not due to the presence of high plasma levels of extra-
cellular virus, but is primarily due to cell-to-cell trans-
mission between mononuclear phagocytes (possibly
macrophages or dendritic cell precursors) and unin-
fected tissues [25].
The lack of a successful HCMV vaccine as well as the
toxicity and drug-induced resistance associated with cur-
rent therapeutics for HCMV indicate that this virus con-
tinues to pose a significant public health problem. Current
treatments for HCMV disease target viral replication and
can fail due to the emergence of drug-resistant virus var-
iants and induction of adverse effects. Hence, a new
approach in drug design against HCMV is required
[26,27]. Since HCMV and other herpesviruses establish a
lifelong latency in humans, antiviral therapy that inhibits
viral entry may serve as an alternative to the already exist-
ing and inadequate therapeutic agents. Here, we report the
design, development and characterization of peptides that
specifically inhibit viral infection and/or entry as a novel
approach to prevent HCMV infection.
Results
HCMV gB is a likely class III viral fusion protein
Structural studies place herpes simplex virus type 1 gB-1
[28] and Epstein-Barr virus gB into class III viral fusion
proteins (VFP) [29], which also includes VSV G [30],
members of the GP64 superfamily (baculovirus and tho-
gotovirus) [31] and tentatively bornavirus G [32].
Because gB is the most highly conserved envelope pro-
tein amongst the mammalian and avian herpesvirues
[25,33], gB of HCMV is likely to be a class III VFP and
shares structural features with gB of other members of
the Herpesviridae. Class III viral fusion proteins share
certain characteristics found in class I or class II viral
fusion proteins. The class III viral fusion proteins con-
tain an extended a-helix that trimerizes in the post-
fusion forms of the proteins [28,30,34], as has been
well-documented for the post-fusion forms of the class I
viral fusion proteins of orthomyxoviruses, retroviruses,
paramyxoviruses, arenaviruses, and coronaviruses [32].
Similarly, the class II viral fusion proteins of flaviviruses
and alphaviruses contain a fusion domain comprised
principally of b-sheets and “fusion loops.” Class III viral
fusion proteins also possess a fusion domain, as well as
several other features of class II viral fusion proteins,
suggesting that these two classes of proteins may share
a common progenitor [31].
The class III domain nomenclature used here can
apply to both class II and class III viral fusion proteins:
domain I (green), domain II (yellow), domain III (blue),
domain IV (stem domain, indigo) (Figure 1). This
Melnik et al. Virology Journal 2011, 8:76
http://www.virologyj.com/content/8/1/76
Page 2 of 11

unified nomenclature assigns domain II appellation to
the following: VSV G domain IV in the nomenclature of
Roche et al. [30], HSV-1 gB-1 and baculovirus gp64
domain I in the nomenclature of Heldwein et al. [28]
and Kadlec et al. [31] as the class III fusion domain,
which is structurally similar to class II viral fusion pro-
teins. In addition to minor adjustments in the ends of
domains, the current class III viral fusion protein num-
bering also combines two interacting domains into
domain III (I + II in Roche’s VSV G nomenclature, III +
IV in Heldwein’s HSV-1 gB-1 nomenclature and
Kadlec’s baculovirus nomenclature).
Identification of HCMV gB inhibitory peptides
The Wimley-White Interfacial Hydrophobicity Scale
(WWIHS) is an experimentally determined hydrophobi-
city scale that provides a quantitative description of a
protein partitioning and folding into membrane inter-
faces. WWIHS score-positive sequences may also inter-
act with hydrophobic surfaces within proteins, and are
often sequestered within pre-fusion forms of viral fusion
proteins. In addition to similarities in the overall struc-
ture of the post-fusion forms of class III VFP, there are
additional similarities in the distribution of WWIHS-
positive sequences (Figure 1, red). The similarities
include at least one extended “fusion loop” in the fusion
domain (domain II), and one or more WWIHS score-
positive sequences in domain III. With the exception of
the ACNPV GP64, each of these proteins contains
another WWIHS positive domain II sequence near the
“hinge” region adjacent to the domain. Herpesvirus gB
proteins have an additional WWIHS scale score-positive
sequence in domain I. In the case of class II and III
viral fusion proteins, the fusion loops in the fusion
domain often contain sequences with positive WWIHS
scores.
Previous studies have suggested that synthetic peptides
corresponding to or overlapping with sequences in viral
fusion proteins that have positive WWIHS scores can
sometimes serve as viral entry inhibitors [35-50]. For
example, Enfurvitide (Fuzeon) is a 36-amino acid pep-
tide that overlaps with a WWIHS score-positive
sequence in the transmembrane protein (TM) of HIV-1,
and prevents viral fusion and entry of the virus. To
identify regions of HCMV gB that have a high propen-
sity to interact with the lipid bilayer of cell membranes
and which potentially may serve as HCMV entry inhibi-
tors, we employed Membrane Protein eXplorer version
3.0 (http://blanco.biomol.uci.edu/mpex), a computer
program based on the WWIHS. Nine sequences with
significant positive WWIHS scores were identified
(Figure 2). As expected, several of these WWIHS
sequences corresponded to the predicted fusion domain
of HCMV gB, including the predicted fusion loops. One
of the WWIHS score-positive sequences spanned amino
acids 146 to 200 (peptide 146-200) that had a ΔG score
of 4.33 kcal/mol. This sequence was split into two smal-
ler peptides: 146-173 and 174-200. A second large
Figure 1 Wimley-White interfacial hydrophobicity scale score-positive sequences in class II and III viral fusion proteins. Sequences of a
representative class II viral fusion protein (Dengue virus E) and of class III viral fusion proteins with high potential to interface with lipid
membranes (red) were identified using Membrane Protein Explorer software (MpeX version 3.0). As discussed in the text, a class III domain
nomenclature is used here that can apply to both class II and III viral fusions proteins. The alternative domain number schemes used by Roche
et al. and Heldwein et al. are noted in parentheses. The Dengue virus (DENV E) stem domain sequence that was not included in the protein
used to determine the crystal structure has been added. The DENV stem has a positive WWIHS scale score, and corresponds to a previously
determined inhibitor of DENV and West Nile virus [35].
Melnik et al. Virology Journal 2011, 8:76
http://www.virologyj.com/content/8/1/76
Page 3 of 11

segment within the fusion domain of HCMB gB had a
ΔG score of 3.39, and was split, consequently, into smal-
ler peptides: 233-263 and 264-291. An additional pep-
tide, 297-315 corresponding to another fusion domain
sequence, was also synthesized, along with an additional
4 peptides corresponding to other WWIHS-positive
domains of HCMV gB. To prevent dimer formation,
cysteines were replaced with alanines.
Nine synthetic peptides corresponding to sequences
with significant WWIHS scores were synthesized and
examined for their ability to inhibit HCMV infection of
HFF cells. Peptides that were most effective are pre-
sented here (Table 1). All synthetic peptides were tested
at the following concentrations: 100 μM, 50 μM, 25 μM,
10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.625 μM, 0.3125 μM,
0.156 μM, and 0.078 μM. HFF were seeded at a density
of 3.5×105 cells in each well of a 24-well plate 24 hours
prior to infection. HFF were washed with 1× DPBS and
mock- or virus-infected for 90 minutes at RT with the
Towne-GFP strain of HCMV (0.5 MOI) preincubated
with different concentrations of inhibitory peptides at
37°C for 90 minutes. After infection, virus was removed
and Dulbecco’s modified Eagle medium (DMEM) sup-
plemented with 10% fetal bovine serum (FBS), penicillin
G (100 U/mL), streptomycin (100 mg/mL), and Gluta-
MAX (2 mM) was added to each well and cells were
incubated at 37°C for 48 hours. GFP-positive cells were
visualized 48 hours post-infection by fluorescence
microscopy and then quantified using flow cytometry.
Peptide 174-200, for instance, demonstrated 80% inhi-
bition of viral infection at a concentration of 100 μM,
and 51% and 62% inhibition at concentrations of 5 μM
and 2.5 μM, respectively (Figure 3). Peptide 233-263
inhibited viral infection by 97% and 92% at concentra-
tions 100 μM and 50 μM, respectively, and by 60% at a
concentration of 2.5 μM (Figure 4). The scrambled pep-
tide (control), of peptide 233-263, was unable to inhibit
HCMV infection significantly (data not shown). While
peptide 264-291, alone, showed inhibition of 70.5%, at a
concentration of 5 μM (Figure 5), peptide 297-315
tested alone showed 40% inhibition at a concentration
of 50 μM (Figure 6). None of the remaining peptides
showed significant inhibition of HCMV infection at any
of the concentrations tested (data not shown). In addi-
tion to testing individual peptides in viral infectivity
assays, peptides were similarly tested in combination.
Interestingly, when peptides 174-200 and 233-263 were
tested together, no significant inhibitory effect was
shown (Figure 7). On the contrary, peptides 264-291
and 297-315 tested together displayed 67% inhibition at
a concentration of 0.125 μM each (Figure 8). Represen-
tative fluorescent and bright light images of HFF cells
infected with the Towne-GFP strain of HCMV (0.5
MOI) preincubated with or without peptide 233-263 at
a concentration of 100 μM were taken 48 hours post-
infection (Figure 9).
Discussion
The WWIHS is a computational approach, based upon
an experimentally determined algorithm to estimate the
propensity of an amino acid sequence to interact with
lipid membrane interfaces [49]. Using this method, we
identified several regions of HCMV gB with high inter-
facial hydrophobicity. Peptides that are analogous to
several of these regions inhibited HCMV infectivity at
low μM concentrations (Figure 3, 4, 5, 6, 7 and 8).
When tested in combination certain combinations of
peptides (peptides 264-291 and 297-315) displayed
increased inhibition of infectivity at concentrations of
125 nM (Figure 8). These results suggest that the
HCMV inhibitory peptides identified here may serve as
Figure 2 Determination of regions within gB that display a
high propensity to interact with the lipid surface of cell
membranes by using Wimley-White Interfacial Hydrophobicity
Scale (WWIHS). WWIHS identifies segments of proteins that prefer a
transbilayer helix conformation to an unfolded interfacial location.
We used the Interface Scale of the Membrane Protein explorer
(MpeX version 3.0) computer program to identify these particular
segments of HCMV gB. The Interface scale measures a residue’s free
energy of transfer within an unfolded polypeptide chain from water
to a phosphocholine bilayer. We identified nine segments of HCMV
gB that display high propensity to interact with the lipid surface of
cell membrane, and designed peptides, ranging from 19 to 31
amino acids in length, that are analogous to the identified regions
of gB.
Table 1 Amino acid sequences of HCMV gB peptides
Peptide Amino acid sequence Position
174-200 WEIHHINKFAQAYSSYSRVIGGTVFVA 174-200
233-263 WHSRGSTWLYRETANLNAMLTITTARSKYPY 233-263
264-291 HFFATSTGDVVYISPFYNGTNRNASYFG 264-291
297-315 FFIFPNYTIVSDFGRPNAA 297-315
HCMV gB peptides were synthesized based on the amino acid sequence
determined from GenBank accession no. DAA00160 (Human Herpesvirus 5
strain AD169).
Melnik et al. Virology Journal 2011, 8:76
http://www.virologyj.com/content/8/1/76
Page 4 of 11

Figure 3 Inhibition of HCMV infection by peptide 174-200. HFF were seeded at a density of 3.5×105 cells in each well of a 24-well plate 24
hours prior to infection. The inhibitory effect of peptide 174-200 was evaluated by infecting human foreskin fibroblasts (HFF) with the Towne-
GFP strain of HCMV (0.5 MOI) preincubated with different concentrations of inhibitory peptide 174-200 at 37°C for 90 minutes. GFP-positive cells
were visualized 48 hours post-infection by fluorescence microscopy and then quantified using flow cytometry. Significant reductions in the
number of GFP-positive cells compared to HCMV infected cells are denoted by a * (p < 0.05), ** (p < 0.01), and *** (p < 0.001, determined using
one-way ANOVA and Tukey’s post test). All structural figures of HSV-1 gB in the post-fusion configuration were generated using MacPyMOL [56]
and FreeHand (Macromedia). Different domains of gB are shown in the ribbon structures: yellow-fusion domain II, purple-stem, green-domain I,
and blue-domain III with extended a-helices, which are involved in trimerization. Peptides targeting different domains of HSV-1 gB that
correspond to HCMV gB domains are shown in black.
Figure 4 Inhibition of HCMV infection by peptide 233-263. HFF were seeded at a density of 3.5×105 cells in each well of a 24-well plate 24
hours prior to infection. The inhibitory effect of peptide 233-263 was evaluated by infecting human foreskin fibroblasts (HFF) with the Towne-
GFP strain of HCMV (0.5 MOI) preincubated with different concentrations of inhibitory peptide 233-263 at 37°C for 90 minutes. GFP-positive cells
were visualized 48 hours post-infection by fluorescence microscopy and then quantified using flow cytometry. Significant reductions in the
number of GFP-positive cells compared to HCMV infected cells are denoted by a * (p < 0.05), ** (p < 0.01), and *** (p < 0.001, determined using
one-way ANOVA and Tukey’s post test). All structural figures of HSV-1 gB in the post-fusion configuration were generated using MacPyMOL [56]
and FreeHand (Macromedia). Different domains of gB are shown in the ribbon structures: yellow-fusion domain II, purple-stem, green-domain I,
and blue-domain III with extended a-helices, which are involved in trimerization. Peptides targeting different domains of HSV-1 gB that
correspond to HCMV gB domains are shown in black.
Melnik et al. Virology Journal 2011, 8:76
http://www.virologyj.com/content/8/1/76
Page 5 of 11