Content uploaded by Mark R Wills
Author content
All content in this area was uploaded by Mark R Wills
Content may be subject to copyright.
The Journal of Immunology
Intracellular Sequestration of the NKG2D Ligand ULBP3 by
Human Cytomegalovirus
Neil J. Bennett,* Omodele Ashiru,
†
Fiona J. E. Morgan,* Yin Pang,* Georgina Okecha,*
Rob A. Eagle,
‡
John Trowsdale,
‡
J. G. Patrick Sissons,* and Mark R. Wills*
Human CMV (HCMV) encodes multiple genes that control NK cell activation and cytotoxicity. Some of these HCMV-encoded gene
products modulate NK cell activity as ligands expressed at the cell surface that engage inhibitory NK cell receptors, whereas others
prevent the infected cell from upregulating ligands that bind to activating NK cell receptors. A major activating NKR is the
homodimeric NKG2D receptor, which has eight distinct natural ligands in humans. It was shown that HCMV is able to prevent
the surface expression of five of these ligands (MIC A/B and ULBP1, 2, and 6). In this article, we show that the HCMV gene product
UL142 can prevent cell surface expression of ULBP3 during infection. We further show that UL142 interacts with ULBP3 and
mediates its intracellular retention in a compartment that colocalizes with markers of the cis-Golgi complex. In doing so, UL142
prevents ULBP3 trafficking to the surface and protects transfected cells from NK-mediated cytotoxicity. This is the first description
of a viral gene able to mediate downregulation of ULBP3. The Journal of Immunology, 2010, 185: 1093–1102.
Cytomegaloviruses use a range of mechanisms to modulate
cellular Ag presentation (1). They reduce levels of MHC
class I on the surface of infected cells to render these cells
less susceptible to lysis by MHC class I-restricted CD8
+
CTLs.
However, reduced surface MHC class I expression also leads to
reduced inhibitory signaling by MHC-specific receptors on NK
cells, which could increase susceptibility of infected cells to NK
cell-mediated cytotoxicity (2). Because downregulation of Ag pre-
sentation to evade T cell responses is critical for CMV reactivation
(3), CMVs have evolved mechanisms that inhibit NK cell function.
NK cells can be activated through a number of pathways: by
binding Ab through FcRs (4); through the action of cytokines, such
as IL-2 (5); and by direct interaction with target cells. These direct
interactions rely on a diverse range of receptors and ligands present
on NK and target cells (6, 7). NK cells express activating and
inhibitory cell-surface receptors. Activating receptors often recog-
nize cell-surface ligands induced by the stress of infection (8) or
transformation (9) and, in some cases, microbial proteins expressed
at the cell surface (10, 11), whereas many inhibitory receptors bind
MHC class I molecules as their ligands (12). Human CMV
(HCMV), also known as human herpes virus 5, encodes multiple
mechanisms to modulate NK cell responses by provision of inhib-
itory signals and suppression of activating signals.
Two mechanisms were described by which HCMV induces
inhibitory receptor signaling. In the first, the virus uses the host
HLA-E pathway to inhibit NK cells through the CD94/NKG2
heterodimeric inhibitory receptor (13, 14). This pathway is usually
dependent on binding peptides, derived from the classical MHC
class I proteins, which stabilize HLA-E and lead to its cell-surface
expression (14). Thus, reduced levels of HLA-A, -B, and
-C during HCMV infection cause reduced surface expression of
HLA-E and, consequently, less inhibitory signaling to NK cells.
HCMV is able to promote HLA-E expression at the cell surface
via the production of a viral protein UL40, which contains a mo-
nomeric peptide that binds HLA-E, thus allowing its cell-surface
expression (15, 16). The second inhibitory mechanism is the ex-
pression of a viral homolog of cellular MHC class I UL18 (17).
UL18 is trafficked to the cell surface with the aid of an unidenti-
fied HCMV protein (18), where it binds the inhibitory NK cell
receptor LIR-1. UL18 binds LILRB1 (LIR-1) with a much higher
affinity than MHC class I (19, 20), and this binding inhibits acti-
vation of LILRB1
+
NK cells but not LILRB1
2
NK cells (21).
NK cell activation can also be mediated by engagement of
activating receptors with their ligands on the surface of infected or
transformed cells. Of particular interest are ligands expressed in
response to viral infection. HCMV encodes at least five genes that
prevent the activation of NK cell receptor signaling. The pp65
tegument protein (UL83) is able to dissociate the CD3zsignaling
chain from NKp30 (22), whereas UL141 protein retains CD155
intracellularly. This reduces cell-surface levels of CD155 to pre-
vent binding of the activating NK cell receptors CD226 and CD96
(23). The remaining viral proteins interfere with NKG2D-
mediated NK cell activation. NKG2D is a major activating recep-
tor on NK cells, expressed on all human NK cells ex vivo (24).
Humans express at least eight NKG2D ligands (NKG2DLs):
MICA/B, ULBP1–3, RAET1E (ULBP4), RAET1G (ULBP5),
and RAET1L (ULBP6). The ligands can be expressed indepen-
dently of each other but are generally poorly expressed on normal,
uninfected tissues (25). Because expression of these ligands is
induced by infection (26), HCMV must prevent their upregulation
to minimize their effects. The viral UL16 protein binds ULBP1, 2,
and 6, as well as MICB (but not the related ULBP3 and MICA)
(27, 28); thus, it mediates their intracellular retention in the endo-
plasmic reticulum (ER) or Golgi to inhibit surface upregulation
(29, 30). MICB expression is also controlled by an HCMV-
*Department of Medicine, School of Clinical Medicine,
†
Department of Pathology,
and
‡
Cambridge Institute for Medical Research, University of Cambridge, Cam-
bridge, United Kingdom
Received for publication March 11, 2010. Accepted for publication April 30, 2010.
This work was supported by Wellcome Trust Grant 079591/Z/06/Z.
Address correspondence and reprint requests to Dr. Mark R. Wills at the current
address: University of Cambridge, Addenbrookes Hospital, Hills Road, Cambridge
CB2 0QQ, U.K. E-mail address: mrw1004@cam.ac.uk
Abbreviations used in this paper: CPE, cytopathic effect; ER, endoplasmic reticulum;
HCMV, human CMV; HFF, human foreskin fibroblast; MCMV, mouse CMV; MOI,
multiplicity of infection; NKG2DL, NKG2D ligand; RAd, recombinant adenovirus.
Copyright Ó2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00
www.jimmunol.org/cgi/doi/10.4049/jimmunol.1000789
encoded microRNA (UL112-1) that reduces MICB translation
(31). Viral microRNAs that target MICB were also identified in
EBV, HSV, and Kaposi’s sarcoma-associated herpesvirus (32).
Downregulation of surface expression of the closely related MICA
was also demonstrated in HCMV-infected cells mediated by
UL142 protein (33) and an unidentified gene (34).
HCMV expresses UL142 from the U
L
-b9region, which is
missing in highly passaged laboratory-adapted strains of the virus,
such as AD169 (35). UL142 is a member of the UL18 gene family
(36) and is predicted to have structural similarities to MHC class I,
with a1 and a2 domains but no a3 domain. In addition, we pre-
viously showed that UL142 localizes to the ER and cis-Golgi (37,
38) and that transduction of UL142 is sufficient to protect cells
from NK cell-mediated lysis, whereas reducing UL142 translation
with RNA interference during HCMV infection increases NK cell-
mediated cytotoxicity (37). UL142 can also downregulate the
products of full-length alleles of MICA by intracellular retention
in the Golgi complex (33, 38). Although these effects of UL142
have been proposed as the mechanism by which it exerts protec-
tion against NK cells, surface levels of truncated MICA protein
(encoded by the p008 allele) were not downregulated. Interest-
ingly, this allele is expressed by ∼66% of people in North America
(39) and is the most common MICA allele in most populations
studied, which led to the suggestion that UL142 may be driving
the selection of certain MICA alleles in the population (40).
Prevention of the cell-surface expression of NKG2DL is also ob-
served in mouse CMV (MCMV; or MuHV-1)-infected cells. Like
humans, mice have multiple NKG2DLs, including mH60, MULT-
1, and molecules of the RAE-1 family; MCMV was shown to pre-
vent cell-surface expression of all of these ligands using multiple
viral gene products. The m145 gene product prevents surface ex-
pression of MULT-1: mutant MCMVs lacking this gene are atten-
uated in vivo (41). All known RAE-1 isoforms are downregulated
by the m152 gene product (gp40), and the in vivo deficit observed
following infection with an m152-deleted virus can be minimized
by blocking NKG2D ligation (42). MCMV also downregulates
surface levels of mH60 with the m155 gene (43) in a protea-
some-dependent manner (44). Interestingly, the virus also shows
a certain amount of redundancy: specifically, the m138 gene prod-
uct, originally identified as a viral FcR (45), is able to downregu-
late expression of mH60, MULT-1 (46), and some isoforms of the
RAE genes (47).
The extent to which HCMV and MCMV use multiple redundant
mechanisms to prevent the cell-surface expression of NKG2DLs,
combined with the fact that all murine NKG2DLs are recognized
by at least one MCMV protein, suggests that HCMV is also able
to prevent cell-surface expression of all human NKG2DLs.
rUL16 protein was shown to have no binding to ULBP3 (27) or
RAET1E/ULBP4 and only very weak or marginal binding to
RAET1G/ULBP5 (28, 48); therefore, no HCMV genes that are able
to prevent the surface expression of ULBP3 or RAET1E/G have
been described. Because UL16 and MCMV m138 (45, 46) down-
regulate multiple NKG2DLs, we sought to determine whether
HCMV UL142 had a similar property.
In this article, we show that clinical strains of HCMVare able to
prevent surface expression of ULBP3. We demonstrate that UL142
interferes with cell-surface expression of ULBP3 and that expres-
sion of UL142 alone, by transduction or transfection, is able to re-
duce cell-surface expression of ULBP3 and cause its intracellular
retention in a compartment that colocalizes with markers of the cis-
Golgi. Immunoprecipitation of UL142 from ULBP3-expressing
cells coprecipitates ULBP3, and reduction of cell-surface expres-
sion of ULBP3 in UL142-expressing cells is sufficient to protect
cells from NK cell-mediated cytotoxicity.
Materials and Methods
Cells
Human foreskin fibroblasts (HFFs) were maintained in Eagle’s MEM
supplemented with 10% FCS, 2 mM L-glutamine, 10
5
IU penicillin/l, and
100 mg streptomycin/l (Invitrogen, Paisley, U.K.). U373 cells (an astrocy-
toma cell line) were maintained in the same medium, whereas 293T and
HeLa-M cells were maintained in DMEM supplemented as above.
Primary NK cell lines were established from PBMCs. PBMCs were
prepared from fresh heparinized venous blood samples by Ficoll-Hypaque
(Lymphoprep-Nyegaard, Oslo, Norway) density-gradient centrifugation.
NK cells were enriched by negative depletion using MACS columns (Mil-
tenyi Biotec, Auburn, CA). Briefly, 2 310
7
PBMCs were stained with
FITC-conjugated mAbs to CD3, CD14, and CD19 (BD Biosciences, San
Jose, CA), followed by anti-FITC MACS beads (Miltenyi Biotec), and
separated on an LS column. The pre- and postcolumn populations were
monitored by flow cytometry. The flow-through contained the negatively
selected, NK-enriched population. Polyclonal NK cell lines were estab-
lished by plating 2000 cells/well in 50 ml RPMI-20 (RPMI 1640 plus 10%
human AB serum and 10% FCS; Life Technologies) in 96-well U-bottom
microtiter plates (Costar, Corning, Corning, NY). Each well was stimu-
lated with 50 ml a mixture of 50 U/ml recombinant human IL-2 (National
Institute of Biological Standards and Control, Potters Bar, U.K.), 25,000
irradiated allogenic EBV-transformed B cell lines, and 25,000 irradiated
autologous PBMCs. Plates were incubated at 37˚C with 5% CO
2
and fed
with 50 ml RPMI-20 plus 50 U/ml recombinant human IL-2 on days 5 and
10. Polyclonal lines were established after 2 wk in culture and were main-
tained by feeding with 50 ml RPMI-20 plus 50 U/ml rIL-2 every 5 d.
Vector constructs
3xFLAG-tagged ULBP3 (denoted FLAG-ULBP3) was a gift from Richard
Apps and Ashley Moffett (Department of Pathology, University of Cam-
bridge) (49), and 3xFLAG-tagged ULBP2 (FLAG-ULBP2) was constructed
as described previously (50). N-terminal GFP-tagged UL142 was con-
structed as described previously (38). All plasmids were prepared using
a Nucleobond Maxi Xtra method (Machery Nagel, Du
¨ren, Germany) and
resuspended in Tris-EDTA or water. N-terminal GFP-tagged ULBP2 and
ULBP3 were constructed in a pcDNA3-based expression vector described
previously (51).
Viruses
HCMV strains AD169 (ATCC VR-538) and TB40/e (gift of Christian
Sinzger, University of Tu
¨bingen, Tu
¨bingen, Germany) were grown in HFFs.
Briefly, confluent 150-cm
2
flasks of HFFs were washed once with PBS and
infected with virus at a multiplicity of infection (MOI) of 0.1. One hour later,
fresh tissue culture media was added, and the infected cells were incubated
at 37˚C in 5% CO
2
. Four days later, when .90% of the cells were showing
signs of cytopathic effect (CPE), the tissue culture supernatant was
harvested and replaced with fresh media. This was repeated daily until
the cell monolayer was destroyed. The virus supernatants from each
harvest were aliquoted and frozen at 270˚C. An aliquot from each time
point was used to infect a 25-cm
2
tissue culture flask of HFFs, and CPE was
assessed 24 h later. Viral stocks demonstrating .90% CPE were retained,
and viral titer was determined by plaque assay.
Recombinant adenoviruses (RAds) expressing UL142 and GFP or
a control RAd expressing GFP alone (37) were gifts of Peter Tomesec and
Gavin Wilkinson (University of Cardiff, Heath Park, Cardiff, U.K.). U373
cells were infected at an MOI of 50, and HeLa-M cells were infected at an
MOI of 20; both cell types were incubated for 24–96 h.
Flow cytometry
Cells were harvested by trypsinization or in cell-dissociation media (Sigma-
Aldrich, St. Louis, MO), and ∼10
6
cells were pelleted in FACS tubes
(Falcon). Cells were washed in PBS, resuspended, and stained using
rPE-conjugated anti-MHC class I (BD Biosciences, San Jose, CA) or
anti-ULBP2 (R&D Systems, Minneapolis, MN), anti-ULBP3 (R&D Sys-
tems), or anti-MICA (Santa Cruz Biotechnology, Santa Cruz, CA) and an
Alexa Fluor 647-conjugated anti-mouse secondary layer, where required
(Invitrogen). Cells were washed in PBS, fixed in 2% paraformaldehyde,
and analyzed on a FACSCalibur instrument (BD Biosciences). Plots were
generated using WinMDI2.9 (Scripps Research Institute, La Jolla, CA).
NK cell cytotoxicity assays
Targets consisted of HeLa-M cells in six-well plates that were untrans-
fected, cotransfected with FLAG-ULBP3 and FLAG-UL142, or cotrans-
fected with FLAG-ULBP3 and pcDNA3. Transfections were carried out
1094 HCMV INFECTION PREVENTS SURFACE EXPRESSION OF ULBP3
using Fugene6 (Roche, Basel, Switzerland). Twenty-four hours posttrans-
fection, cells were washed once in PBS and harvested by trypsinization
(Sigma-Aldrich). K562 cells were used as a positive control for NK cell
cytotoxicity. Surface levels of ULBP3 were assessed by flow cytometry.
Target cells were loaded with [
51
Cr] (sodium chromate; Amersham Bio-
sciences, Piscataway, NJ and/or PerkinElmer, Wellesley, MA), incubated at
37˚C for 45 min, and then washed three times in RPMI-10. Primary human
polyclonal NK cells and
51
Cr-labeled target cells were mixed in 96-well
U-bottom plates in quintuplicate at different E:T ratios and incubated for
5 h at 37˚C, 5% CO
2
. Supernatants were harvested with a Biomek-1000
robot (Beckman Coulter, High Wycombe, U.K.), radioactive counts were
determined using a gcounter (COBRA, Packard Instrument, Meriden,
CT), and the percentage specific lysis was calculated.
Immunofluorescence microscopy
HeLa-M cells were grown on glass coverslips (VWR International, Leices-
tershire, U.K.). Cells were cotransfected with 3xFLAG vector, 3xFLAG-
UL142 and GFP-ULBP2, or GFP-ULBP3 using Fugene6. Cells were
incubated for 16 h, fixed in 4% paraformaldehyde for 30 min, and permea-
bilized in 0.1% Triton X-100. Cells were stained with rabbit anti-FLAG
(Sigma-Aldrich) and a mouse anti-gM130, anti-p230, or the IgG1 isotype
control Ab (all from BD Biosciences). Then, cells were stained with Alexa
Fluor 647-conjugated anti-mouse IgG andAlexa Fluor 568-conjugated anti-
rabbit IgG (both from Invitrogen). All coverslips were mounted in antifade
gold (Invitrogen) and visualized using a confocal microscope (Leica TCS
SP, Leica Microsystems, Deerfield, IL); images were analyzed using ImageJ
software (National Institutes of Health, Bethesda, MD).
Internalization assays
HeLa-M cells were transfected as for immunofluorescence microscopy.
Fourteen hours after transfection, the cells were incubated at 4˚C, stained
with mouse anti-CD71 (eBioscience, Hatfield, U.K.), and mouse anti-
ULBP3 or anti-ULBP2 (both from R&D Systems) was added for 30 min
in ice-cold PBS. Cultures were washed five times with PBS and incubated
at 37˚C in prewarmed DMEM-10 for 2 h. Cells were fixed in 4% para-
formaldehyde for 30 min and stained using anti-FLAG to detect FLAG-
UL142, as well as an Alexa Fluor 647-conjugated anti-mouse secondary
Ab to detect internalized mouse Abs.
Immunoprecipitation
293T cells were cotransfected with GFP-tagged UL142 and 3xFLAG-tagged
ULBP2 or GFP-tagged UL142 and 3xFLAG-tagged ULBP3. At 16 h post-
transfection, GFP
+
transfected cells were isolated by cell sorting. Samples
for input controls were removed and analyzed by Western blot using anti-
ULBP2 and ULPB3, Abs as described below; cells were then lysed in
buffer containing 50 mM Tris-Cl [pH 8], 200 mM NaCl, 200 mM Na
3
VO
4
,
10 mM NaF, 0.5% Nonidet P-40, and 50 mM PMSF with 40 mg/ml BSA,
and cell membranes and nuclei were removed by centrifugation. Lysates
were precleared with an isotype-control Ab (R&D Systems) and protein
A-Sepharose and were immunoprecipitated using anti-GFP (Roche) or iso-
type control (R&D Systems). Beads were washed, and proteins were eluted
by boiling in Laemmli buffer and analyzed by Western blot.
Western blot analysis
Protein samples were separated by SDS-PAGE and transferred to HiBondC
nitrocellulose membranes. Membraneswere blocked with5% nonfat milk pro-
tein in 0.1% Tween-20 in PBS, and proteins were detected with anti-ULBP3
(R&D SystemsAF1517; 1:50) or anti-ULBP2(R&D Systems AF1298; 1:50),
followed by HRP-conjugated species-specific secondary Abs (Dako). HRP
signal was detected using an ECLPlus (Amersham Biosciences) chemilumi-
nescent HRP substrate and x-ray film (Kodak, Rochester, NY).
Results
Clinical strains of HCMV do not upregulate surface expression
of the NKG2DL ULBP3
Infection of fibroblast cell lines with HCMV strain AD169 was
shown to induce MICA, MICB, and ULBP1–3 mRNA expression
(26). At the protein level, MICB, ULBP1, and ULBP2 were not
expressed at the cell surface, but ULBP3’s cell-surface expression
was markedly increased by AD169 infection (52). The highly pas-
saged HCMV strain AD169 has a number of genetic differences
compared with low-passage clinical isolates of HCMV, including
an ∼13–15-kb deletion, termed the U
L
-b9region (35, 53). To
investigate the effects of different strains of HCMV on cell-
surface expression of ULBP3, HFF cells were infected with
AD169 or the low-passage clinical strain TB40e. Uninfected and
infected cells (24, 48, 72, and 96 h postinfection) were stained with
anti-ULBP3 and anti-MHC class I Abs and analyzed by flow
cytometry. Reduced MHC class I surface expression was used as
a marker of HCMV infection, and these infected cells were gated
and analyzed for ULBP3 expression; cells were also stained with
isotype-matched control Ab, which was negative in all cases. In
agreement with a previous analysis (52), uninfected HFFs did not
express ULBP3 on their cell surface; however, by 72 h postinfec-
tion with HCMV strain AD169, increased ULBP3 cell-surface ex-
pression was detected. In contrast, cells infected with HCMV strain
TB40e did not express ULBP3 on the cell surface at any time point
postinfection (Fig. 1). The results demonstrated that clinical strains
of HCMV are able to prevent ULBP3 cell-surface expression. Be-
cause the major difference between AD169 and TB40e virus strains
is the deletion of U
L
-b9, these results further suggested that a gene
in this genomic region is able to mediate this function.
UL142 decreases cell-surface expression of ULBP3
The U
L
-b9region encodes at least two genes (UL141 and UL142)
that modulate the NK cell response (23, 37). UL141 downregulates
levels of CD155 from the surface of infected cells, thus preventing
binding of the CD96 and CD226 activating receptors on NK cells
(23). Although it was demonstrated that UL142 is able to decrease
cell-surface expression of full-length alleles of MICA (33, 38), and
because HCMV UL16 is able to interact and prevent cell-surface
expression of multiple NKG2DL, we hypothesized that UL142
might also prevent ULBP3 cell-surface expression.
To determine whether UL142 in isolation was able to prevent
expression of ULBP3 at the cell surface, we used RAds expressing
UL142 and GFP or a control adenovirus expressing GFP alone.
U373 and HeLa-M cells were transduced with the RAds. U373
cells are homozygous for the full-length MICA allele *001,
whereas HeLa-M cells are homozygous for the truncated allele
*008. At 96 h posttransduction, cell-surface levels of MICA,
ULBP2, and ULBP3 were determined by staining with Ab and an-
alyzed by flow cytometry; cells were also stained with isotype-
matched control Ab, which was negative in all cases.
HeLa-M and U373 cells expressing UL142 exhibited decreased
cell-surface expression of ULBP3 compared with cells transduced
with the control adenovirus (Fig. 2). As previously reported (38),
expression of UL142 in U373 cells led to decreased cell-surface
expression of the full-length MICA allele *001 (Fig. 2A). How-
ever, UL142 expression had no effect on cell-surface expression of
the truncated allele MICA*008 in HeLa-M cells (Fig. 2B), as
expected. The cell-surface expression of the NKG2DL ULBP2
was not affected by UL142 expression in either cell line, providing
an important negative control.
UL142 can downregulate surface expression of ULBP3 during
HCMV infection
Although UL142 in isolation is able to downregulate surface levels
of ULBP3, we noted that TB40e was able to control ULBP3 levels at
early times (24–48 h) during infection, before UL142 expression
began at 72 h postinfection (37). This suggests that HCMVencodes
at least one additional gene in the U
L
-b9region that downregulates
surface levels of ULBP3. This meant that the use of a UL142
knockout virus to test the importance of UL142 during infection
was not feasible, because any phenotype might be masked by
the presence of the unidentified second gene. Therefore, we ex-
pressed UL142 in cells infected with the AD169 strain of virus
to attempt to rescue the AD169 phenotype to that of TB40e
The Journal of Immunology 1095
infection. HFFs were transduced with RAd expressing UL142 and
GFP (or a control adenovirus expressing GFP alone) and were
infected with AD169 24 h later. Surface levels of ULBP3 were
determined by flow cytometry; cells were also stained with iso-
type-matched control Ab. As expected, AD169 infection caused
upregulation in the levels of surface ULBP3 compared with un-
infected control cells. Transduction of AD169-infected cells with
the control adenovirus showed that the RAd had no intrinsic effect
on surface levels of ULBP3. Transduction of AD169-infected cells
with adenovirus expressing UL142 prevented surface expression of
ULBP3 (Fig. 3).
Expression of UL142 causes intracellular retention of ULBP3
in the cis-Golgi complex
UL16 mediates decreased surface expression of MICB and
ULBP1, 2, and 6 by intracellular retention in the ER or cis-Golgi
(27, 28). More recently, we showed that UL142 similarly retains
full-length alleles of MICA in the cis-Golgi (38). To determine
whether UL142 prevented cell-surface expression of ULBP3 via
a similar mechanism, we investigated the cellular localization of
ULBP3 in the presence of UL142. Because no Abs recognizing
endogenous/untagged UL142 were available, we used an N-
terminal FLAG-tagged UL142 construct that we previously
showed was able to mediate intracellular retention of full-
length alleles of MICA (38). HeLa-M cells were cotransfected
with FLAG-UL142 or the empty FLAG vector control and
a GFP-tagged ULBP3 or GFP-tagged ULBP2 construct. The
cells were stained with Abs to the FLAG tag (to visualize
UL142) and Golgi markers and then analyzed by immunofluo-
rescence microscopy.
The results show that in the absence of UL142, ULBP2 and
ULBP3 were expressed at the cell surface in all transfected cells
(Fig. 4A). Unsurprisingly, some of these cells also showed limited
colocalization between GFP-ULBP2/3 and markers of the cis-
(GM130) and trans- (p230) Golgi complex (data not shown), pre-
sumably as a result of normal ULBP2/3 trafficking via the Golgi
complex to the cell surface.
FIGURE 2. Decreased cell-surface expression of ULBP3 in U373 or
HeLa-M cells expressing UL142. U373 or HeLa-M cells expressing
UL142 have decreased cell-surface expression of ULBP3. A, U373 cells
transduced with RAd UL142 (RAd expressing UL142 and GFP) or RAd
(control adenovirus expressing GFP alone) were stained for surface expres-
sion of ULBP3, ULBP2, an isotype control and MICA, or an isotype control.
U373 expresses the full-length MICA allele *001, which is downregulated
by UL142. As expected, ULBP2 was unaffected by UL142 expression;
however, ULBP3 surface expression was decreased. B, The experiment
was repeated using HeLa-M cells, which express the truncated MICA allele
*008. As expected, ULBP2 and MICA*008 were unaffected by UL142
expression; however, surface expression of ULBP3 was downregulated.
FIGURE 1. ULBP3 expression after HCMV infection. HFF cells in-
fected with the HCMV strain AD169 expressed increasing levels of
ULBP3 at the cell surface with increasing time postinfection, but ULBP3
was not expressed on the surface of cells infected with the HCMV strain
TB40e. HFF cells were infected with HCMV strain AD169 or the low-
passage strain TB40e for 24, 48, 72, or 96 h. The cells were stained with
anti-ULBP3 and anti-MHC class I Abs and analyzed by flow cytometry.
Infected cells were identified by gating for reduced surface expression
of MHC class I and then ULBP3 expression and isotype controls were
analyzed.
1096 HCMV INFECTION PREVENTS SURFACE EXPRESSION OF ULBP3
Cells that coexpressed FLAG-UL142 and GFP-ULBP3 lacked
any cell-surface expression of ULBP3. Instead, GFP-ULBP3
was detected in an intracellular compartment that exhibited exten-
sive colocalization with the marker of the cis-Golgi (GM130) (Fig.
4B). When expressed in isolation, most FLAG-UL142 proteins
predominantly colocalized with markers of the ER (38), whereas
in FLAG-UL142/GFP-ULBP3–cotransfected cells, FLAG-UL142
and GFP-ULBP3 were detected in the cis-Golgi (Fig. 4B). We also
cotransfected FLAG-UL142 with GFP-ULBP2 as a control; in this
case (Fig. 4B), ULBP2 was expressed at the cell surface in all
cotransfected cells. These results suggested that UL142 interfered
with the normal trafficking of ULBP3 to the cell surface by in-
tracellular retention in the Golgi complex.
Immunoprecipitation of UL142 coprecipitates ULBP3
UL16 directly interacts with ULBP1/2/6 and MICB to prevent the
surface expression of these NKG2DLs (27–30). Consequently, it
was possible that UL142 mediates intracellular retention of ULBP3
by a direct interaction. This hypothesis is supported by the fact that
GFP-ULBP3 and FLAG-UL142 are localized to the cis-Golgi in
cotransfected cells. To address this, immunoprecipitations were
carried out on lysates of 293T cells cotransfected with GFP-
UL142 and FLAG-ULBP3 or GFP-UL142 and FLAG-ULBP2.
At 16 h posttransfection, transfected cells were separated into
GFP
+
and GFP
2
populations using flow cytometric cell sorting.
Some of the cells from each of these populations were used for
Western blot as input controls for ULBP2 and ULBP3; the results
showed that both populations had about the same amount of these
proteins. Lysates were made from the rest of each population, and
immunoprecipitations were carried out using an anti-GFP Ab (to
precipitate GFP-UL142) or an isotype-control Ab and protein A-
Sepharose. Eluted proteins were denatured, run on an SDS-PAGE
gel, and analyzed by Western blotting using anti-ULBP3 or ULBP2
Abs. FLAG-ULBP3, but not FLAG-ULBP2, coprecipitated with
GFP-UL142 (Fig. 5), thus demonstrating that UL142 associates
with ULBP3 but not with ULBP2. However, this does not preclude
that the interaction between UL142 and ULBP3 is an indirect one,
requiring the involvement of another protein.
UL142 prevents export of ULBP3 to the cell surface
Our results showed that UL142 downmodulates expression of
ULBP3 at the cell surface, leading to intracellular retention of the
NKG2DL in the cis-Golgi (Fig. 4). Two mechanisms are possible:
UL142 prevents trafficking of nascent ULBP3 to the cell surface
or UL142 mediates internalization of cell surface ULBP3. To de-
termine the mechanism by which UL142 downmodulates surface
expression of ULBP3, Ab-internalization assays were carried out
to determine whether ULBP3 traffics to the cell surface in the
presence of UL142. At 14 h posttransfection, HeLa-M cells trans-
fected with FLAG-ULBP2 or FLAG-ULBP3 (with or without
GFP-UL142) were incubated with mouse anti-ULBP3 or anti-
ULBP2 Ab for 30 min at 4˚C. Cells were incubated at 37˚C for
an additional 2 h (for protein internalization to occur) and then
fixed, permeabilized, and stained with rabbit anti-FLAG Ab. This
was followed by incubation with anti-mouse Ab (to visualize in-
ternalized ULBP2/3) and anti-rabbit Ab (to visualize noninternal-
ized ULBP2/3).
Cells transfected with FLAG-ULBP3 alone showed clear inter-
nalization of surface ULBP3 (Fig. 6A). However, in cells cotrans-
fected with GFP-UL142 and FLAG-ULBP3, FLAG-ULBP3 was
absent from the cell surface, and no Ab internalization was ob-
served (Fig. 6A). To show that UL142 did not mediate a nonspecific
block on retrograde trafficking, internalization of ULBP2 and the
transferrin receptor CD71 was assessed. Cells expressing FLAG-
ULBP2, in the presence or absence of GFP-UL142, exhibited in-
ternalization of FLAG-ULBP2 (Fig. 6B). In addition, internaliza-
tion of CD71 was observed in cells cotransfected with GFP-UL142
and FLAG-ULBP3 (Fig. 6C), thus demonstrating that UL142 did
not mediate a global block on retrograde transport. The results
suggest that de novo synthesized ULBP3 was prevented from leav-
ing the Golgi and, thus, was unable to traffic to the cell surface.
UL142-mediated decrease in surface ULBP3 protects target
cells from NK cell-mediated lysis
NK cell cytotoxicity can be induced by signaling via the NKG2D
receptor. The UL142-mediated loss of ULBP3 on the surface of a tar-
get cell should reduce the ability of NK cells to lyse the target cell. To
test this, we carried out NK cell cytotoxicity assays using
HeLa-M cells. These were used as target cells because they do
not express the NKG2DLs ULBP1/2/3 (data not shown) and only
express low levels of the short allele of MICA, with which
UL142 is unable to interact (38); as such, we could transfect the
cells with ULBP3, to induce higher levels of cytotoxicity, and
cotransfected cells with ULBP3 and UL142, which should reduce
cytotoxicity to background levels. HeLa-M cells were cotrans-
fected with FLAG-ULBP3 and pcDNA3 alone or FLAG-UL142
as target cells against primary polyclonal human NK cells. As
predicted, transfection of HeLa-M cells with FLAG-ULBP3 led
to an increase in NK cell-mediated lysis of these target cells
FIGURE 3. UL142 expression in HCMV strain AD169-infected cells
substantially reduced surface ULBP3 expression. HFF cells were trans-
duced with adenovirus expressing UL142 or a control adenovirus. After
24 h, cells were infected with HCMV AD169 and incubated for an addi-
tional 24 h. Then cells were stained with anti-ULBP3 and anti-MHC class I
Abs and analyzed by flow cytometry. Infected cells were identified by
gating on cells with reduced surface expression of MHC class I and then
surface expression of ULBP-3 was analyzed.
The Journal of Immunology 1097
compared with untransfected HeLa-M cells. However, cotransfec-
tion of HeLa-M cells with FLAG-ULBP3 and FLAG-UL142 re-
duced the level of target cell lysis to that of untransfected
HeLa-M cells (Fig. 7A).
Domain functions within UL142 could not be separated
Previous work using UL142-CD8 chimeric proteins demonstrated
that UL142 protein is localized to the ER and cis-Golgi by the
UL142 transmembrane domain (38). Using these same chimeras
in cotransfection experiments, it was also demonstrated that the
UL142 luminal domain is involved in intracellular retention of
full-length MICA alleles (38). We showed that ULBP3 coprecipi-
tates with UL142 (Fig. 5). This suggested that an association be-
tween UL142 and ULBP3 (directly or via a third-party protein)
leads to intracellular retention of the NKG2DL. Consequently, we
used the chimeric CD8-UL142 constructs (Fig. 7B) to determine
which domains of UL142 were involved in the retention of ULBP3
in the cis-Golgi. HeLa-M cells cotransfected with GFP-ULBP3 and
a Myc-tagged CD8-UL142 chimera were analyzed by immunoflu-
orescence microscopy (Fig. 7B,7C). None of the CD8-UL142 chi-
meras were able to mediate intracellular retention of GFP-ULBP3.
This showed that the cis-Golgi retention ability of UL142 cannot be
ascribed to any single domain of the UL142 protein.
Discussion
The HCMV gene UL16 is able to mediate the intracellular retention
of four different NKG2DLs (ULBP1/2/6 and MICB) (27, 28, 30,
52), whereas UL142 is known to mediate the intracellular retention
of a fifth NKG2DL (full-length alleles of MICA) (33, 37, 38).
Therefore, it seemed surprising that no mechanism for the control
of ULBP3 expression had been described, especially because
MCMV is able to interfere with the cell-surface expression of all
of the murine NKG2DLs. Thus, it seemed inherently plausible that
ULBP3 would be controlled in HCMV-infected cells and also pos-
sible that UL142 could downmodulate expression of other
NKG2DL(s), as is the case with UL16. In this study, we showed
that HCMV-infected cells can downregulate surface levels of
ULBP3, that only wild-type virus has this property, and that the
laboratory-adapted strain AD169, missing the U
L
-b9region, is
unable to control ULBP3 surface expression. We also
demonstrated that ULBP3 downregulation is mediated by UL142,
a gene that is located in this U
L
-b9region. UL142 is unable to
downmodulate mature ULBP3 from the cell surface; instead, it
promotes retention of nascent ULBP3 protein in the cis-Golgi,
thereby preventing trafficking of the NKG2DL to the cell surface.
UL142 also specifically coprecipitates with ULBP3, consistent
with there being a protein–protein interaction. The downregulation
of cell-surface ULBP3 induced by UL142 expression was sufficient
to protect cells from NK cell-mediated cytotoxicity.
Downregulation of MHC class I (“missing self”) and the in-
duction of ligands of activating NKRs (“stressed self”) are fea-
tures of many viral infections, including HCMV. These strategies
should render HCMV-infected cells susceptible to NK cell-
mediated cytotoxicity. However, to counter this, HCMV encodes
at least eight functions that modulate NK cell responses (54).
Although some of these mechanisms operate by providing inhib-
itory signals to NK cells to compensate for the lack of MHC class
I on the cell surface of infected cells, the majority prevent NK cell
FIGURE 4. Intracellular accumulation of ULBP3
in cells coexpressing ULBP3 and UL142. A, Cells
transfected with GFP-ULBP2 or GFP-ULPB3 showed
cell surface expression of both proteins. Original mag-
nification 3132 (left), 3130 (right). B, Cells co-
transfected with GFP-ULBP3 and FLAG-UL142 did
not express ULBP3 at the cell surface; however, cells
cotransfected with ULBP2-GFP and UL142 still ex-
pressed surface ULBP2. Original magnification 3126.
FIGURE 5. ULBP3 coprecipitated with UL142 from cotransfected cells.
293T cells were cotransfected with either GFP-UL142 and FLAG-ULBP3
or GFP-UL142 and FLAG-ULBP2. 16 h post transfection, cells were
separated into GFP positive and negative populations by cell sorting.
Lysates were made from the two separate populations. Immunoprecipita-
tions were carried out using either an anti-GFP Ab (to precipitate GFP-
UL142) or an isotype control Ab, and Protein A Sepharose. Eluted proteins
were denatured and analyzed by western blot using anti-ULBP3 or ULBP2
Abs. Input protein, some of the GFP positive and negative cell populations
were lysed and directly analyzed by western blot using both the ULBP2
and ULBP3 Abs.
1098 HCMV INFECTION PREVENTS SURFACE EXPRESSION OF ULBP3
activation by interfering with cellular ligands for activating recep-
tors on the NK cell. One of the major activating receptors on NK
cells is the homodimeric NKG2D. Human NKG2D has at least
eight ligands (ULBP1-6 and MICA/B), and HCMV infection
results in upregulated transcription of all of these ligands, with
the exception of ULBP4 (26, 28). Ligation of NKG2D by its
ligands provides a strong activating signal via the associated
DAP10 adaptor that can override the inhibitory signals mediated
by engagement of inhibitory NK cell receptors (55). Conse-
quently, survival of HCMV-infected cells would require viral
modulation of the expression of all induced NKG2DLs. HCMV
is particularly adept at minimizing NK cell activation through the
NKG2D pathway, with at least four viral genes (UL16, UL142,
UL112-1, and an unidentified gene) (27, 28, 30, 31, 33, 34, 37, 38,
52) combining to prevent expression of MICA/B and ULBP1/
2/6 at the cell surface. With so many mechanisms already identi-
fied, it seems possible that HCMV is particularly susceptible to
NKG2D-mediated responses, a hypothesis supported by evidence
from mouse models in which MCMV is the only virus known to
downregulate all ligands of murine NKG2D (41–47).
The NKG2DL MICA and MICB proteins are highly related in
their a1/2 domains (sharing 83% amino acid identity) and are
more distantly related to ULBP1–6, yet UL16 binds to MICB (but
FIGURE 6. UL142 retains ULBP3 in the cis-Golgi by preventing traffick-
ing to the cell surface. A, HeLa-M cells were transfected with FLAG-ULBP3
alone or in conjunction with GFP-UL142. Fourteen hours posttransfection,
transfected cells were incubated with mouse anti-ULBP3 Ab for 30 min at
4˚C and an additional incubation at 37˚C for 2 h (by which time intracellular
accumulations of ULBP3 could be observed). Cells were fixed, permeabi-
lized, and stained with anti-FLAG Ab to visualize noninternalized ULBP3.
This was followed by staining with an anti-mouse Ab to visualize cell-
surface–internalized Ab and, thus, ULBP3. Cells transfected with FLAG-
ULBP3 alone showed internalization of surface ULBP3; however, no inter-
nalization was observed in cells cotransfected with GFP-UL142 and FLAG-
ULBP3. Original magnification 3113 (top), 376 (bottom). B, As a control,
HeLa-M cells were transfected with FLAG-ULBP2 alone or in conjunction
with GFP-UL142. Cells that expressed ULBP2 in the presence or absence of
UL142 showed internalization of ULBP2. Original magnification 3113.
C, Cells cotransfected with GFP-UL142 and FLAG-ULBP3 and stained
with anti-transferrin receptor CD71 also showed CD71 internalization. Orig-
inal magnification 376.
FIGURE 7. UL142-mediated downregulation of ULBP3 surface expres-
sion protects against NK cell-mediated lysis. A, NK cytotoxicity assays
were carried out using HeLa-M cells transfected with ULBP3 alone or in
conjunction with UL142 as target cells against a primary polyclonal hu-
man NK cells. Lysis of HeLa-M cells cotransfected with ULBP3 and
UL142 was comparable to that of untransfected HeLa-M cells. All
domains of the UL142 protein were required for intracellular retention
of ULBP3. B, Three chimeras consisting of domains of UL142 and CD8
were constructed. Chimera EXT is a CD8 chimera consisting of the UL142
luminal domain. Chimera TM consists of the UL142 transmembrane do-
main, whereas chimera TAIL consists of the UL142 tail domain. C,
HeLa-M cells were cotransfected with GFP-ULBP3 and a CD8-UL142
chimera and then analyzed by immunofluorescence microscopy. None of
the CD8-UL142 chimeras was able to mediate intracellular retention of
ULBP3. Original magnification 388 (top), 363 (middle), 396 (bottom).
The Journal of Immunology 1099
not MICA) and to ULBP1/2/6 (but not ULBP3/4). A recent
structural analysis identified a small number of amino acids on
UL16 that mimics an area on NKG2D that interacts with its
ligands. Critically, NKG2DLs that bind UL16 have a glutamine or
glutamate at position 169. Ligands, such as MICA and ULBP3,
which have an arginine at this position, are not bound by UL16
(56). Thus, it was suggested that UL16 expression provides evo-
lutionary pressure by which viral immune evasion genes, such as
UL16, can drive diversification of the ligands for NKG2D to the
benefit of the host. Equally, as our data support, the virus is driven
to evolve other viral genes (UL142) to prevent the expression of
NKG2DLs that have escaped control by UL16.
MICA and ULBP3 possess MHC I-like a-1/2 extracellular
domains but with notable differences. ULBP3 is predicted to have
a cell-surface attachment via a GPI anchor, whereas MICA con-
tains a transmembrane domain (55). UL142 can downmodulate
cell-surface expression of the full-length MICA allele but not
the truncated allele *008. These two alleles of MICA have a very
high sequence identity, except for a premature stop codon that
truncates the transmembrane domain and tail of MICA*008. This
suggests that UL142 must, at least in part, recognize MICA in this
region, although the extracellular, transmembrane, and tail region
of UL142 were all required to prevent surface expression of full-
length MICA (38). A similar analysis of the UL142 domains that
might interact with ULBP3 (Fig. 7) again showed that none of the
UL142 domains in isolation were able to prevent the surface ex-
pression of ULBP3. UL142 resides in the ER as a property of its
transmembrane domain until it interacts with a newly synthesized
MICA or ULBP3 molecule. However, because intracellular reten-
tion of ULBP3 or MICA is reliant on the UL142 luminal domain,
it seems likely that there is an interaction between the UL142
luminal domain and the external domain of ULBP3 and MICA.
It remains unclear how UL142 is able to bind and retain the
disparate MICA and ULBP3 but not ULBP2, which is more
closely related to ULBP3 than MICA. There is little sequence
identity between the external domains of MICA and ULBP3;
however, the recent structural analysis of the interaction of
UL16 with various NKG2DLs highlighted that interaction can
depend on a small number of critical residues (56).
The data presented in this study, and the fact that ULBP3 is less
polymorphic than MICA (for which $70 alleles have been iden-
tified) (39, 57), suggest that downregulation of surface ULBP3
may represent the major mechanism by which UL142 protects
infected cells from NK cell cytotoxicity. Our results also suggest
that HCMV encodes at least one additional gene that down-
modulates surface expression of ULBP3 at early times post-
infection. Our previous work showed that UL142 is expressed
from 72 h postinfection (37). Interestingly, when ULBP3 ex-
pression is measured from 24–96 h postinfection with AD169 and
TB40e virus strains, it is expressed on the cell surface by 48 h
following infection by AD169 but not by TB40e (Fig. 1A). Be-
cause UL142 is not expressed at this time, we propose that at least
one more gene in the U
L
-b9region of HCMV is able to
downregulate surface levels of ULBP3. This would mean that
ULBP3, MICA, and MICB are each downregulated by at least
two genes in HCMV. This has parallels with HCMV-mediated
interference with the MHC class I pathway: the virus encodes
multiple genes (Us 2, 3, 6 and 11) that can interfere with normal
MHC class I processing and presentation, and these genes are
expressed at different points in the viral life cycle (58).
The presence of multiple viral modulators of some NKG2DLs
raises the question of whether there is a hierarchy of importance.
Perhaps those proteins modulated by two or more viral genes are
better activators of NKG2D; it was shown that mouse and human
NKG2DL have different binding affinities for NKG2D (59–61).
Alternatively, it is possible that multiple NKG2DLs exist to provide
the host with a range of mechanisms to distinguish different cell
types and tissues and infecting agents. Because HCMV infects
a range of cells and tissues, it is possible that the relative importance
of each NKG2DL may be more reliant on the location of its ex-
pression than on its ability to activate NK cells.
Upon infection, a variety of viruses was shown to induce
upregulation of cell-surface NKG2DL; subsequently, many of these
were shown to encode NKG2DL-associated evasion proteins (62–
66). Until now, no viral mechanism able to downregulate ULBP3
has been described. Human cells encode at least eight NKG2DLs;
HCMV has now been shown to encode genes that interfere with
six of these (MICA, MICB, ULBP1, ULBP2, ULBP3, and
ULBP6), leaving two additional ligands, ULBP4 and ULBP5. It
was shown that ULBP4 mRNA expression is not induced by
HCMV infection of fibroblasts and that UL16-Fc does not bind
ULBP4 (28); it is possible that this NKG2DL is not involved in
immunity to HCMV. Cell-surface–expressed ULBP5 is bound by
UL16-Fc, although much more weakly than the closely related
ligands ULBP2 and ULBP6 (28); however, it has not been for-
mally demonstrated that ULBP5 is prevented from being
expressed on the cell surface during HCMV infection.
The importance of HCMV modulation of multiple NKG2DLs
might not be solely related to viral evasion of NK cell responses,
because NKG2D receptor expression is not restricted to NK cells. In
humans, NKG2D is also expressed on all CD8
+
and gd T cells and
on some CD4
+
T cells, where it functions as a costimulatory mol-
ecule (67), and it may also induce TCR-independent killing (68).
More recently, it was shown that some dendritic cell subsets express
NKG2D and that signaling via the receptor can also induce target
cell killing (IFN-producing killer dendritic cells (69, 70). It is also
recognized that human dendritic cells subsets display cytotoxic
potential through a variety of stimuli and effector mechanisms
(69). Taking into account all of the evidence, the multiple mecha-
nisms that HCMV has evolved to prevent infected cells displaying
ligands for NKG2D, although clearly important for preventing NK
cell activation and cytotoxicity mediated through this receptor, also
seem likely to be important as a broader strategy to interfere with
innate and acquired immune responses to the virus.
Acknowledgments
We thank Professor John Sinclair, Dr. Emma Poole, and Dr. Michael Gill for
useful discussions.
Disclosures
The authors have no financial conflicts of interest.
References
1. Lin, A., H. Xu, and W. Yan. 2007. Modulation of HLA expression in human
cytomegalovirus immune evasion. Cell. Mol. Immunol. 4: 91–98.
2. Ljunggren, H. G., and K. Ka
¨rre. 1990. In search of the ‘missing self’: MHC
molecules and NK cell recognition. Immunol. Today 11: 237–244.
3. Pinto, A. K., M. W. Munks, U. H. Koszinowski, and A. B. Hill. 2006. Co-
ordinated function of murine cytomegalovirus genes completely inhibits CTL
lysis. J. Immunol. 177: 3225–3234.
4. Brier, A. M., L. Chess, and S. F. Schlossman. 1975. Human antibody-dependent
cellular cytotoxicity. Isolation and identification of a subpopulation of peripheral
blood lymphocytes which kill antibody-coated autologous target cells. J. Clin.
Invest. 56: 1580–1586.
5. Domzig, W., B. M. Stadler, and R. B. Herberman. 1983. Interleukin 2 dependence
of human natural killer (NK) cell activity. J. Immunol. 130: 1970–1973.
6. Moretta, L., C. Bottino, C. Cantoni, M. C. Mingari, and A. Moretta. 2001. Human
natural killer cell function and receptors. Curr. Opin. Pharmacol. 1: 387–391.
7. Biassoni, R. 2008. Natural killer cell receptors. Adv. Exp. Med. Biol. 640: 35–52.
8. Gleimer, M., and P. Parham. 2003. Stress management: MHC class I and class I-
like molecules as reporters of cellular stress. Immunity 19: 469–477.
1100 HCMV INFECTION PREVENTS SURFACE EXPRESSION OF ULBP3
9. Groh, V., R. Rhinehart, H. Secrist, S. Bauer, K. H. Grabstein, and T. Spies. 1999.
Broad tumor-associated expression and recognition by tumor-derived gamma
delta T cells of MICA and MICB. Proc. Natl. Acad. Sci. USA 96: 6879–6884.
10. Daniels, K. A., G. Devora, W. C. Lai, C. L. O’Donnell, M. Bennett, and
R. M. Welsh. 2001. Murine cytomegalovirus is regulated by a discrete subset of
natural killer cells reactive with monoclonal antibody to Ly49H. J. Exp. Med.
194: 29–44.
11. Lee, S. H., S. Girard, D. Macina, M. Busa
`, A. Zafer, A. Belouchi, P. Gros, and
S. M. Vidal. 2001. Susceptibility to mouse cytomegalovirus is associated with
deletion of an activating natural killer cell receptor of the C-type lectin super-
family. Nat. Genet. 28: 42–45.
12. Moretta, A., C. Bottino, M. Vitale, D. Pende, R. Biassoni, M. C. Mingari, and
L. Moretta. 1996. Receptors for HLA class-I molecules in human natural killer
cells. Annu. Rev. Immunol. 14: 619–648.
13. Posch, P. E., F. Borrego, A. G. Brooks, and J. E. Coligan. 1998. HLA-E is the
ligand for the natural killer cell CD94/NKG2 receptors. J. Biomed. Sci. 5: 321–
331.
14. Braud, V. M., D. S. Allan, C. A. O’Callaghan, K. So
¨derstro
¨m, A. D’Andrea,
G. S. Ogg, S. Lazetic, N. T. Young, J. I. Bell, J. H. Phillips, et al. 1998.
HLA-E binds to natural killer cell receptors CD94/NKG2A, B and C. Nature
391: 795–799.
15. Tomasec, P., V. M. Braud, C. Rickards, M. B. Powell, B. P. McSharry, S. Gadola,
V. Cerundolo, L. K. Borysiewicz, A. J. McMichael, and G. W. Wilkinson. 2000.
Surface expression of HLA-E, an inhibitor of natural killer cells, enhanced by
human cytomegalovirus gpUL40. Science 287: 1031.
16. Ulbrecht, M., S. Martinozzi, M. Grzeschik, H. Hengel, J. W. Ellwart, M. Pla, and
E. H. Weiss. 2000. Cutting edge: the human cytomegalovirus UL40 gene product
contains a ligand for HLA-E and prevents NK cell-mediated lysis. J. Immunol.
164: 5019–5022.
17. Beck, S., and B. G. Barrell. 1988. Human cytomegalovirus encodes a glyco-
protein homologous to MHC class-I antigens. Nature 331: 269–272.
18. Maffei, M., F. Ghiotto, M. Occhino, M. Bono, A. De Santanna, L. Battini,
G. L. Gusella, F. Fais, S. Bruno, and E. Ciccone. 2008. Human cytomegalovirus
regulates surface expression of the viral protein UL18 by means of two motifs
present in the cytoplasmic tail. J. Immunol. 180: 969–979.
19. Cosman, D., N. Fanger, L. Borges, M. Kubin, W. Chin, L. Peterson, and
M. L. Hsu. 1997. A novel immunoglobulin superfamily receptor for cellular and
viral MHC class I molecules. Immunity 7: 273–282.
20. Chapman, T. L., A. P. Heikeman, and P. J. Bjorkman. 1999. The inhibitory re-
ceptor LIR-1 uses a common binding interaction to recognize class I MHC
molecules and the viral homolog UL18. Immunity 11: 603–613.
21. Prod’homme, V., C. Griffin, R. J. Aicheler, E. C. Wang, B. P. McSharry,
C. R. Rickards, R. J. Stanton,L. K. Borysiewicz, M. Lo
´pez-Botet,G. W. Wilkinson,
and P. Tomasec. 2007. The human cytomegalovirus MHC class I homolog UL18
inhibits LIR-1+ but activates LIR-1- NK cells. J. Immunol. 178: 4473–4481.
22. Arnon, T. I., H. Achdout, O. Levi, G. Markel, N. Saleh, G. Katz, R. Gazit,
T. Gonen-Gross, J. Hanna, E. Nahari, et al. 2005. Inhibition of the NKp30
activating receptor by pp65 of human cytomegalovirus. Nat. Immunol. 6: 515–
523.
23. Tomasec, P., E. C. Wang, A. J. Davison, B. Vojtesek, M. Armstrong, C. Griffin,
B. P. McSharry, R. J. Morris, S. Llewellyn-Lacey, C. Rickards, et al. 2005.
Downregulation of natural killer cell-activating ligand CD155 by human cyto-
megalovirus UL141. Nat. Immunol. 6: 181–188.
24. Bauer, S., V. Groh, J. Wu, A. Steinle, J. H. Phillips, L. L. Lanier, and T. Spies.
1999. Activation of NK cells and T cells by NKG2D, a receptor for stress-
inducible MICA. Science 285: 727–729.
25. Eagle, R. A., I. Jafferji, and A. D. Barrow. 2009. Beyond Stressed Self: Evidence
for NKG2D Ligand Expression on Healthy Cells. Curr. Immunol. Rev. 5: 22–34.
26. Eagle, R. A., J. A. Traherne, O. Ashiru, M. R. Wills, and J. Trowsdale. 2006.
Regulation of NKG2D ligand gene expression. Hum. Immunol. 67: 159–169.
27. Cosman, D., J. Mu
¨llberg, C. L. Sutherland, W. Chin, R. Armitage, W. Fanslow,
M. Kubin, and N. J. Chalupny. 2001. ULBPs, novel MHC class I-related mol-
ecules, bind to CMV glycoprotein UL16 and stimulate NK cytotoxicity through
the NKG2D receptor. Immunity 14: 123–133.
28. Eagle, R. A., J. A. Traherne, J. R. Hair, I. Jafferji, and J. Trowsdale. 2009.
ULBP6/RAET1L is an additional human NKG2D ligand. Eur. J. Immunol. 39:
3207–3216.
29. Ro
¨lle, A., M. Mousavi-Jazi, M. Eriksson, J. Odeberg, C. So
¨derberg-Naucle
´r,
D. Cosman, K. Ka
¨rre, and C. Cerboni. 2003. Effects of human cytomegalovirus
infection on ligands for the activating NKG2D receptor of NK cells: up-regulation
of UL16-binding protein (ULBP)1 and ULBP2 is counteracted by the viral UL16
protein. J. Immunol. 171: 902–908.
30. Dunn, C., N. J. Chalupny, C. L. Sutherland, S. Dosch, P. V. Sivakumar,
D. C. Johnson, and D. Cosman. 2003. Human cytomegalovirus glycoprotein
UL16 causes intracellular sequestration of NKG2D ligands, protecting against
natural killer cell cytotoxicity. J. Exp. Med. 197: 1427–1439.
31. Stern-Ginossar, N., C. Gur, M. Biton, E. Horwitz, M. Elboim, N. Stanietsky,
M. Mandelboim, and O. Mandelboim. 2008. Human microRNAs regulate stress-
induced immune responses mediated by the receptor NKG2D. Nat. Immunol. 9:
1065–1073.
32. Nachmani, D., N. Stern-Ginossar, R. Sarid, and O. Mandelboim. 2009. Diverse
herpesvirus microRNAs target the stress-induced immune ligand MICB to es-
cape recognition by natural killer cells. Cell Host Microbe 5: 376–385.
33. Chalupny, N. J., A. Rein-Weston, S. Dosch, and D. Cosman. 2006. Down-
regulation of the NKG2D ligand MICA by the human cytomegalovirus glyco-
protein UL142. Biochem. Biophys. Res. Commun. 346: 175–181.
34. Zou, Y., W. Bresnahan, R. T. Taylor, and P. Stastny. 2005. Effect of human
cytomegalovirus on expression of MHC class I-related chains A. J. Immunol.
174: 3098–3104.
35. Bradley, A. J., N. S. Lurain, P. Ghazal, U. Trivedi, C. Cunningham, K. Baluchova,
D. Gatherer, G. W. Wilkinson, D. J. Dargan, and A. J. Davison. 2009. High-
throughput sequence analysis of variants of human cytomegalovirus strains
Towne and AD169. J. Gen. Virol. 90: 2375–2380.
36. Dolan, A., C. Cunningham, R. D. Hector, A. F. Hassan-Walker, L. Lee,
C. Addison, D. J. Dargan, D. J. McGeoch, D. Gatherer, V. C. Emery, et al.
2004. Genetic content of wild-type human cytomegalovirus. J. Gen. Virol. 85:
1301–1312.
37. Wills, M. R., O. Ashiru, M. B. Reeves, G. Okecha, J. Trowsdale, P. Tomasec,
G. W. Wilkinson, J. Sinclair, and J. G. Sissons. 2005. Human cytomegalovirus
encodes an MHC class I-like molecule (UL142) that functions to inhibit NK cell
lysis. J. Immunol. 175: 7457–7465.
38. Ashiru, O., N. J. Bennett, L. H. Boyle, M. Thomas, J. Trowsdale, and
M. R. Wills. 2009. NKG2D ligand MICA is retained in the cis-Golgi apparatus
by human cytomegalovirus protein UL142. J. Virol. 83: 12345–12354.
39. Petersdorf, E. W., K. B. Shuler, G. M. Longton, T. Spies, and J. A. Hansen. 1999.
Population study of allelic diversity in the human MHC class I-related MIC-A
gene. Immunogenetics 49: 605–612.
40. Wilkinson, G. W., P. Tomasec, R. J. Stanton, M. Armstrong, V. Prod’homme,
R. Aicheler, B. P. McSharry, C. R. Rickards , D. Cochrane, S. Llewellyn-Lacey,
et al. 2008. Modulation of natural killer cells by human cytomegalovirus. J. Clin.
Viro l. 41: 206–212.
41. Krmpotic, A., M. Hasan, A. Loewendorf, T. Saulig, A. Halenius, T. Lenac,
B. Polic, I. Bubic, A. Kriegeskorte, E. Pernjak-Pugel, et al. 2005. NK cell
activation through the NKG2D ligand MULT-1 is selectively prevented by the
glycoprotein encoded by mouse cytomegalovirus gene m145. J. Exp. Med. 201:
211–220.
42. Lodoen, M., K. Ogasawara, J. A. Hamerman, H. Arase, J. P. Houchins,
E. S. Mocarski, and L. L. Lanier. 2003. NKG2D-mediated natural killer cell
protection against cytomegalovirus is impaired by viral gp40 modulation of
retinoic acid early inducible 1 gene molecules. J. Exp. Med. 197: 1245–1253.
43. Hasan, M., A. Krmpotic, Z. Ruzsics, I. Bubic, T. Lenac, A. Halenius,
A. Loewendorf, M. Messerle, H. Hengel, S. Jonjic, and U. H. Koszinowski.
2005. Selective down-regulation of the NKG2D ligand H60 by mouse cytomeg-
alovirus m155 glycoprotein. J. Virol. 79: 2920–2930.
44. Lodoen, M. B., G. Abenes, S. Umamoto, J. P. Houchins, F. Liu, and L. L. Lanier.
2004. The cytomegalovirus m155 gene product subverts natural killer cell an-
tiviral protection by disruption of H60-NKG2D interactions. J. Exp. Med. 200:
1075–1081.
45. Tha
¨le, R., P. Lucin, K. Schneider, M. Eggers, and U. H. Koszinowski. 1994.
Identification and expression of a murine cytomegalovirus early gene coding for
an Fc receptor. J. Virol. 68: 7757–7765.
46. Lenac, T., M. Budt, J. Arapovic, M. Hasan, A. Zimmermann, H. Simic,
A. Krmpotic, M. Messerle, Z. Ruzsics, U. H. Koszinowski, et al. 2006. The her-
pesviral Fc receptor fcr-1 down-regulates the NKG2D ligands MULT-1 and H60.
J. Exp. Med. 203: 1843–1850.
47. Arapovic, J., T. Lenac, R. Antulov, B. Polic, Z. Ruzsics, L. N. Carayannopoulos,
U. H. Koszinowski, A. Krmpotic, and S. Jonjic. 2009. Differential susceptibility
of RAE-1 isoforms to mouse cytomegalovirus. J. Virol. 83: 8198–8207.
48. Wittenbrink, M., J. Spreu, and A. Steinle. 2009. Differential NKG2D binding to
highly related human NKG2D ligands ULBP2 and RAET1G is determined by
a single amino acid in the alpha2 domain. Eur. J. Immunol. 39: 1642–1651.
49. Apps, R., L. Gardner, J. Traherne, V. Male, and A. Moffett. 2008. Natural-killer
cell ligands at the maternal-fetal interface: UL-16 binding proteins, MHC
class-I chain related molecules, HLA-F and CD48. Hum. Reprod. 23: 2535–
2548.
50. Bacon, L., R. A. Eagle, M. Meyer, N. Easom, N. T. Young, and J. Trowsdale.
2004. Two human ULBP/RAET1 molecules with transmembrane regions are
ligands for NKG2D. J. Immunol. 173: 1078–1084.
51. Boyle, L. H., A. K. Gillingham, S. Munro, and J. Trowsdale. 2006. Selective
export of HLA-F by its cytoplasmic tail. J. Immunol. 176: 6464–6472.
52. Welte, S. A., C. Sinzger, S. Z. Lutz, H. Singh-Jasuja, K. L. Sampaio, U. Eknigk,
H. G. Rammensee, and A. Steinle. 2003. Selective intracellular retention of virally
induced NKG2D ligands by the human cytomegalovirus UL16 glycoprotein. Eur.
J. Immunol. 33: 194–203.
53. Cha, T. A., E. Tom, G. W. Kemble, G. M. Duke, E. S. Mocarski, and
R. R. Spaete. 1996. Human cytomegalovirus clinical isolates carry at least 19
genes not found in laboratory strains. J. Virol. 70: 78–83.
54. Lanier, L. L. 2008. Evolutionary struggles between NK cells and viruses. Nat.
Rev. Immunol. 8: 259–268.
55. Eagle, R. A., and J. Trowsdale. 2007. Promiscuity and the single receptor:
NKG2D. Nat. Rev. Immunol. 7: 737–744.
56. Mu
¨ller, S., G. Zocher, A. Steinle, and T. Stehle. 2010. Structure of the HCMV
UL16-MICB complex elucidates select binding of a viral immunoevasin to di-
verse NKG2D ligands. PLoS Pathog. 6: e1000723.
57. Zhang, Y., A. M. Lazaro, B. Lavingia, and P. Stastny. 2001. Typing for all known
MICA alleles by group-specific PCR and SSOP. Hum. Immunol. 62: 620–631.
58. Reddehase, M. J. 2002. Antigens and immunoevasins: opponents in cytomega-
lovirus immune surveillance. Nat. Rev. Immunol. 2: 831–844.
59. McFarland, B. J., and R. K. Strong. 2003. Thermodynamic analysis of de-
generate recognition by the NKG2D immunoreceptor: not induced fit but rigid
adaptation. Immunity 19: 803–812.
The Journal of Immunology 1101
60. O’Callaghan, C. A., A. Cerwenka, B. E. Willcox, L. L. Lanier, and
P. J. Bjorkman. 2001. Molecular competition for NKG2D: H60 and RAE1
compete unequally for NKG2D with dominance of H60. Immunity 15: 201–211.
61. Sutherland, C. L., N. J. Chalupny, K. Schooley, T. VandenBos, M. Kubin, and
D. Cosman. 2002. UL16-binding proteins, novel MHC class I-related proteins,
bind to NKG2D and activate multiple signaling pathways in primary NK cells. J.
Immunol. 168: 671–679.
62. Ward, J., M. Bonaparte, J. Sacks, J. Guterman, M. Fogli, D. Mavilio, and
E. Barker. 2007. HIV modulates the expression of ligands important in triggering
natural killer cell cytotoxic responses on infected primary T-cell blasts. Blood
110: 1207–1214.
63. Cerboni, C., F. Neri, N. Casartelli, A. Zingoni, D. Cosman, P. Rossi, A. Santoni,
and M. Doria. 2007. Human immunodeficiency virus 1 Nef protein down-
modulates the ligands of the activating receptor NKG2D and inhibits natural
killer cell-mediated cytotoxicity. J. Gen. Virol. 88: 242–250.
64. Wen, C., X. He, H. Ma, N. Hou, C. Wei, T. Song, Y. Zhang, L. Sun, Q. Ma, and
H. Zhong. 2008. Hepatitis C virus infection downregulates the ligands of the
activating receptor NKG2D. Cell. Mol. Immunol. 5: 475–478.
65. Schepis, D., M. D’Amato, M. Studahl, T. Bergstro
¨m, K. Ka
¨rre, and L. Berg.
2009. Herpes simplex virus infection downmodulates NKG2D ligand expression.
Scand. J. Immunol. 69: 429–436.
66. Thomas, M., J. M. Boname, S. Field, S. Nejentsev, M. Salio, V. Cerundolo,
M. Wills, and P. J. Lehner. 2008. Down-regulation of NKG2D and NKp80
ligands by Kaposi’s sarcoma-associated herpesvirus K5 protects against NK cell
cytotoxicity. Proc. Natl. Acad. Sci. USA 105: 1656–1661.
67. Groh, V., R. Rhinehart, J. Randolph-Habecker, M. S. Topp, S. R. Riddell, and
T. Spies. 2001. Costimulation of CD8alphabeta T cells by NKG2D via
engagement by MIC induced on virus-infected cells. Nat. Immunol. 2: 255–260.
68. Verneris, M. R., M. Karami, J. Baker, A. Jayaswal, and R. S. Negrin. 2004. Role
of NKG2D signaling in the cytotoxicity of activated and expanded CD8+ T cells.
Blood 103: 3065–3072.
69. Larmonier, N., J. Fraszczak, D. Lakomy, B. Bonnotte, and E. Katsanis. 2010.
Killer dendritic cells and their potential for cancer immunotherapy. Cancer
Immunol. Immunother. 59: 1–11.
70. Su
¨ss, G., and K. Shortman. 1996. A subclass of dendritic cells kills CD4 T cells
via Fas/Fas-ligand-induced apoptosis. J. Exp. Med. 183: 1789–1796.
1102 HCMV INFECTION PREVENTS SURFACE EXPRESSION OF ULBP3
