|
|
||||||||

Cells Directly Recognize the Peptide Backbone in a Conformationally Dependent Manner1
Committee on Immunology and Ben May Institute for Cancer Research, University of Chicago, Chicago, IL 60637
| Abstract |
|---|
|
|
|---|

cells, detailed information concerning the structural nature
of these antigenic epitopes is lacking. In addition, the recent
descriptions of human TCR
cells recognizing mycobacterium-derived
low m.w. lipid molecules confirms that the spectrum and nature of
biologic structures that are capable of being recognized by TCR
cells are unclear. We have previously described a murine TCR
cell
clone, TgI4.4, that is reactive to herpes simplex virus (HSV)-1
glycoprotein I (gI). Unlike TCR
ß-mediated, MHC-restricted Ag
recognition but similar to Ig Ag recognition, TgI4.4 recognizes
purified gI directly, in the absence of Ag processing or presentation.
Since gI is a complex glycoprotein, the nature of the antigenic epitope
was investigated. First, gI recognition by TgI4.4 is conformationally
dependent, as revealed by denaturation and proteolytic experiments.
Secondly, the epitope recognized by TgI4.4 was mapped to the amino
terminus by using insertion mutants of gI. Lastly, TgI4.4 recognizes
the gI protein directly since completely deglycosylated forms of gI are
efficiently recognized. Therefore, TCR
cells are capable of
recognizing a variety of molecular structures, including proteins. The
ability of TgI4.4 to recognize a nonglycosylated form of gI suggests
that HSV-1 recognition by TCR
cells in vivo is not limited by
cell-specific glycosylation patterns or glycosylation-dependent
conformational influences. | Introduction |
|---|
|
|
|---|

cells prove to be a critical immunoregulatory population in both
bacterial and viral pathogenesis (1, 2), including herpes simplex virus
(HSV)4-1 infections (3).
However, in most cases the Ags recognized by the activated TCR
cells isolated from draining lymph nodes or pathogenic lesions remain
unknown (2, 4, 5). Despite the paucity of Ag-specific clones of
TCR
cells, the TCR
CDR3 regions sequenced from lymphoid
tissues and gut are very diverse, suggesting a potential for broad Ag
reactivity (6, 7). An abundance of recent evidence suggests that
TCR
cells recognize unprocessed Ags directly (5, 8). For
instance, as documented for HSV-1 glycoprotein I-reactive cells (3, 9)
as well as two MHC alloreactive TCR
cell clones (10), none of the
known factors involved in MHC class I or II Ag processing affect
TCR
Ag recognition. Furthermore, purified "whole" Ags are
able to stimulate TCR
cells directly (9, 11). This mode of Ag
recognition suggests Ig type recognition properties. Two structural
lines of evidence support these conclusions. First, Ag receptor CDR3
structures, analyzed from sequence databases, show that TCR
chains are structurally more similar to Ig heavy and light chains than
to TCR
ß-chains (12). Second, recent crystallographic data acquired
from a human TCRV
-chain reveal distinct configurations within both
the framework and CDR regions that implicate structures similar to both
the TCR and Ig molecules (13).
Structurally diverse types of glycoproteins ranging from heat shock
proteins (14, 15, 16, 17, 18); MHC class II (19), MHC class Ia (4), MHC class Ib
(20, 21); CD1c (22); bacterial superantigens (23), CD48 (24, 25); and
HSV-1 glycoprotein I can be recognized by TCR
cells (3, 26). In
fact, a subset of human TCR
cells have been described that are
reactive to mycobacterium-derived, phosphorylated low m.w. protease
resistant (LMP) lipid compounds (27, 28, 29). Interestingly, the presence
of the phosphate moiety is essential for recognition of the lipid
compound (30). Therefore, given the array of biologic structures
capable of being recognized by TCR
cells-lipids and
glycoproteins, it is essential to examine the nature of these molecular
interactions. We have described the recognition of HSV-1 glycoprotein I
by murine TCR
cells (3, 9). gI recognition by a representative
clone, TgI4.4, is direct and is independent of Ag processing or
presentation. In this study we further explore the nature of this
direct interaction and show that recognition of gI is highly dependent
on conformation, maps to the amino terminus, and is independent of
glycosylation. This structural analysis of TCR
cell Ag
recognition will have important implications toward the interpretation
of molecular structures and models of TCR
-mediated Ag recognition
derived from x-ray crystallographic data.
| Materials and Methods |
|---|
|
|
|---|
CHO gIIg-expressing cells were grown and used to produce large
quantities of soluble gIIg, as previously described (9). LgI-1 cells
are LMtk- cells obtained from American Type Culture
Collection (ATCC, Manassas, VA) and stably transfected with the gI DNA,
as previously described (26). The TCR
clone TgI4.4
(105 cells) was passaged once a week with irradiated (20
Gy) splenic feeder cells (5 x 106 cells), mitomycin
C-treated (40 µg/ml for 30 min) LgI-1 cells (6 x
105 cells), and 30 U/ml rhIL-2, as previously described
(26). Ldld CHO mutant cells were obtained from ATCC
with permission from Dr. Monty Krieger (Massachusetts Institute of
Technology, Boston, MA) (31). All cells were grown in DMEM medium
supplemented with 10% FCS, 2 mM [l]-glutamine, 2 mM nonessential
amino acids, 100 U/ml penicillin, 100 µg/ml streptomycin, and 5
x 10-5 M 2-ME. The Ldld cells were cultured in
Hams F12 medium instead of the DMEM and contained the same
supplements. Tunicamycin was dissolved in DMSO and incubated with the
CHO gIIg cells at the indicated concentrations following addition of
fresh medium. The serum-free medium supplement, ITS, used for the
Ldld glycosylation experiments, consisted of insulin 0.625
mg/ml, transferrin 0.625 mg/ml, and selenium 0.625 mg/ml and was
obtained from Collaborative Research (Bedford, MA). Depending on the
glycosylation conditions, FCS (3%), UDP-galactose (20 µM), and
UDP-galactosamine (200 µM) were used to supplement the F12-ITS
media.
TgI4.4 bioassay
Resting TgI4.4 cells were stimulated with gIIg immobilized on
plastic, gIIg adsorbed to plastic immobilized goat anti-hIgG1
Fc-specific Abs, or gI-expressing cells for 48 h. The supernatant
was harvested and analyzed for secreted IFN-
by ELISA. This
consisted of a sandwich protocol of sequential steps of
anti-IFN-
- (H22 clone) coated wells, incubation of cell
supernatants, incubation of a secondary goat anti-IFN-
antisera,
and, finally, detection with alkaline phosphatase-coupled donkey
anti-goat Abs.
Abs and reagents
Goat anti-hIgG1 Fc Ab, affinity purified, (Cappel, Malvern,
PA) was used to coat tissue culture wells for adsorption of soluble
gIIg. Alkaline phosphatase-coupled goat anti-hIgG1 Fc Ab, affinity
purified (Cappel) was used for gIIg detection in Western blots. Abs for
the IFN-
ELISA have been previously described and were obtained from
Dr. Schrieber (Washington University, St. Louis, MO). The tunicamycin,
UDP-galactose, UDP-galactosamine, and trypsin-coupled agarose beads
were obtained from Sigma (St. Louis, MO). The protease endo-lys-c and
the N-glycanase were obtained from Boehringer Mannheim
(Indianapolis, IN). The G418 sulfate used for negative selection of
cells following DNA transfections was obtained from Life Technologies
(Rockville, MD).
Plasmids
The gIIg expression plasmid used for transfection of Ldld cells has been previously described (9). Linker scanning mutants of HSV-1 gI were constructed by S. Basu, et al., and consisted of the insertion of 10- to 12-mer linkers into various positions of the gI DNA, based on restriction enzyme sites (32). The mutants were sequenced to confirm the restoration of the correct reading frame.
Western analysis
gIIg was enriched from 1 ml of cell supernatant using protein A-Sepharose beads (Pharmacia, Uppsala, Sweden). Beads were washed with PBS, and the bound gIIg was eluted using Laemmli sample buffer and boiling. Samples were resolved using 10% SDS-PAGE in a minigel apparatus and then transferred to nitrocellulose membrane (Schleicher & Schuell, Keene, NH) using a minigel transfer apparatus (Bio-Rad, Hercules, CA). The membrane was blocked with 3% BSA dissolved in TBST buffer (10 mM TRIS, 0.15 M NaCl, and 0.05% Tween 20) and then probed with alkaline phosphatase-coupled goat anti-human IgG Fc-specific Ab. Excess Ab was washed off, and the membrane was developed using BCIP and NBT substrates (Promega, Madison, WI).
gIIg treatments
gIIg in PBS was treated with trypsin-coupled Sepharose beads (Sigma) or with purified endo-lys-c (Boehringer Mannheim) at 37°C overnight. Trypsin beads were pelleted, and the supernatant, containing the proteolytic fragments, was used to coat plastic tissue culture wells for a TgI4.4 bioassay. The endo-lys-c sample was directly coated to plastic tissue culture wells, up to 100 µg/ml, for a TgI4.4 bioassay. Digestion was confirmed by SDS-PAGE and Coomassie blue staining. Reduction of gIIg was accomplished using various concentrations of 2-ME for 1 h at 25°C. Samples were dialyzed against PBS and then immobilized on plastic tissue culture wells for a TgI4.4 bioassay.
DNA transfections
Ldld CHO mutant cells were transfected using a kit from Specialty Media (Lavallette, NJ) that is based on a calcium-phosphate precipitation method. Cells were cotransfected with 10 µg gIIg expression plasmid and 1 µg of a neomycin resistance plasmid, pSV2neo. G418 sulfate was added at 24 h posttransfection at a concentration of 1 mg/ml. Colonies were isolated using cloning cylinders (Specialty Media) and screened for expression by the ability to stimulate TgI4.4. LMtk- cells were transiently transfected using a DEAE-dextran method. Cells in log-phase growth in a 100-mm plate were washed extensively and incubated with DMEM medium containing HEPES (10 mM), chloroquine (50 µM), DEAE-dextran (0.25 mg/ml), and 10 µg DNA in a final volume of 5 ml. After 34 h, the medium was aspirated and the cells were pulsed with DMEM medium containing 10% (v/v) DMSO for 80 s. Following the DMSO pulse, the cells were incubated in complete media and used in a TgI4.4 bioassay 40 h post transfection.
| Results |
|---|
|
|
|---|
A soluble fusion protein, gIIg, constructed with the extracellular
region of gI and the Fc portion of human IgG1 was secreted from CHO
cells and purified over an immunoaffinity column of protein
A-Sepharose. Purified gIIg was able to stimulate the TCR
clone
TgI4.4 to proliferate and to secrete IFN-
when immobilized directly
on plastic tissue culture wells or indirectly by adsorption to
immobilized anti-human IgG Fc-specific Abs, as previously shown
(9). The tertiary structure of gI is critical for TgI4.4 recognition
since thermal denaturation of gIIg abrogated recognition (data not
shown). Furthermore, mild reduction of intermolecular disulfide bonds,
accorded by treatment of gIIg with the reducing agent
ß-mercaptoethanol also abrogated recognition by TgI4.4, (Fig. 1
A). These observations
suggest that perturbations in the conformation of gIIg results in the
disordering of the antigenic epitope. Second, protease digestion of
gIIg was pursued to identify fragments capable of stimulating TgI4.4.
Fig. 1
B shows that trypsin-digested gIIg failed to stimulate
TgI4.4 when immobilized on plastic tissue culture wells. Since there
are 17 putative trypsin cleavage sites in gI, it is possible that none
of the fragments maintain the correct conformation. However, even
digestion of gIIg with endo-lys-C, a protease that generates only four
fragments of gI, failed to stimulate TgI4.4 when these peptides were
immobilized (Fig. 1
B). Thus, the conformation of the
tertiary structure of gI plays a crucial role in forming the epitope
recognized by TgI4.4.
|
The identification of the antigenic epitope of gI recognized by
TgI4.4 was pursued to gain a greater understanding of the molecular
interactions between the TCR and Ag. gI is expressed by HSV-1 and
localizes both on the cell surface of infected cells and on the viral
envelope in heterodimeric form with another glycoprotein, gE (33, 34).
TgI4.4 is capable of recognizing virus-infected cells as well as
purified gIIg, suggesting that the recognized epitope is independent
of, and not affected by, heterodimer formation. Interestingly,
heterodimeric association between gI and gE confers high
avidity-monomeric IgG Fc binding activity, whereas gE alone exhibits
low avidity-aggregated IgG Fc binding activity (33, 34). Since gE, but
not gI, contains sequence homology to mammalian Fc receptors (35), the
role of gI in converting the low avidity IgG Fc binding activity of gE
to high avidity binding is unknown. To elucidate the role of gI in Fc
binding, Basu. et al. generated linker scanning mutants of gI,
characterized these mutants, and identified regions of gI required to
confer high avidity Fc binding activity by the heterodimer (32). These
mutants were used to dissect the epitope specificity of TgI4.4. As seen
in Table I
, wt gI, transiently expressed
in L cells, was efficiently recognized by TgI4.4 as determined by the
induction of IFN-
secretion. Mutation of the core region of gI
prevented TgI4.4 recognition; however, these mutations appear to render
the gI molecule sensitive to misfolding since both Ab recognition and
Fc binding activity were abolished (32). Two closely spaced mutants, at
positions 43 and 63, also eliminated TgI4.4 recognition despite these
mutants ability to bind monomeric IgG Fc. Therefore, it is likely
that the amino terminus of gI, surrounding amino acids 4070, contains
the epitope recognized by TgI4.4.
|
gI is a complex glycoprotein containing three potential
N-linked carbohydrate sites and a putative domain of
O-linked carbohydrates. The role of these carbohydrates in
TgI4.4-mediated gI recognition was investigated. To assess the role of
N-linked carbohydrates, the glycosylation inhibitor
tunicamycin, which prevents the transfer of carbohydrates from the
dolichol core in the endoplasmic reticulum to the asparagine acceptor
in the protein, was used (36). CHO gIIg cells treated with tunicamycin
inhibited all N-linked glycosylations of gIIg as analyzed by
increased mobility on SDS-PAGE and subsequent Western blotting (Fig. 2
B). The absence of
N-linked glycosylation was confirmed by comparing the
mobility of both tunicamycin and N-glycanase-treated gIIg
(Fig. 2
B). As shown in Fig. 2
A, despite efficient
deglycosylation, TgI4.4 was still able to recognize
non-N-linked carbohydrate-modified gIIg when adsorbed to
anti-human IgG1-coated tissue culture wells.
|
|
| Discussion |
|---|
|
|
|---|

cell-mediated
Ag recognition by defining the molecular requirements of recognition by
an Ag-specific TCR
cell clone. The HSV-1 glycoprotein I
(gI)-specific TCR
cell clone TgI4.4 recognizes gI in a direct
manner, independent of classical Ag processing or presentation (9). To
gain a greater understanding of the molecular interactions involved in
gI recognition by TgI4.4, conditions that alter the tertiary structure
of gI were used, and it was determined that the conformational status
of gI controls recognition. Using insertion mutants of gI, the epitope
recognized by TgI4.4 was located in the amino terminus, surrounding
amino acids 4070. The overall conformation and properties of gI were
not altered by these mutations since high avidity-monomeric IgG Fc
binding by the gIgE heterodimer was still intact. In fact, the closely
related gI glycoprotein derived from HSV-2 is not recognized by TgI4.4
(26), consistent with the high degree of divergence between these two
proteins in the N terminus (37). While it is clear that the amino
terminus contains the epitope, it is not clear whether the insertions
disrupted the primary structure of the epitope or whether the
insertions disrupted the local conformational status of the amino
terminus and thereby prevented recognition. Without an additional
measure of the conformation or function of this region of the amino
terminus, this cannot be conclusively determined. Reduction of gIIg
with ß-mercaptoethanol abrogated recognition by TgI4.4, suggesting
that the epitope is dependent on tertiary structure. Although the
epitope lacks cysteine residues, it is adjacent to a cluster of
cysteines that are conserved among HSV-1 glycoproteins (37), suggesting
that the disulfide bonded cysteines are somehow positioning the amino
terminus in an appropriate conformation/orientation. In addition, the
protease endo-lys-c used in this study did not cleave within the
identified epitope but instead separates the epitope from the rest of
the protein just before the cluster of cysteine residues. These data
suggest that the epitope of gI recognized by TgI4.4 is highly
conformationally dependent, independent of the ability of gI to
heterodimerize with gE, is solvent exposed, and, perhaps, is distal to
and protrudes away from cell membrane.
Given the recent descriptions of lipid reactive TCR
cells, it has
become important to define the nature of biologic structures recognized
by Ag-specific TCR
cells. The glycoprotein gI afforded a system
in which to determine whether or not TgI4.4 Ag specificity was directed
at the peptide, carbohydrate, or both. Interestingly, it was found that
gI recognition by TgI4.4 is directed at the peptide backbone and can
occur in the absence of any glycosylations. Since TgI4.4 appears to
recognize the fully deglycosylated gIIg less well, these results do not
rule out the possibility of subtle, qualitative differences in the
conformation of the epitope induced by carbohydrate modifications.
However, due to the variability in the amounts of gIIg produced in
response to the various treatments, it is just as possible that the
differences in stimulation of TgI4.4 by the various forms reflect
quantitative differences. In fact, the amino terminus, the location of
TgI4.4s epitope, is distal from predicted N-linked and
O-linked glycosylation sites (37). In parallel with the lack
of direct carbohydrate reactivity, it is interesting to point out that
the majority of CDR3 amino acids in the TCR
chain are hydrophobic in
nature (9), suggesting 1) a direct interaction between these two
regions and 2) epitope-driven selection for this TCR structure.
Finally, it was shown that TCR
cells play a critical role in
HSV-1 immunity and that gI-reactive TCR
cells may constitute a
large portion of the observed protective response (3). The findings
that gI recognition is both direct and independent of carbohydrates
suggests that TCR
cells may operate in a niche of recognition not
only of infected cells incapable of efficient Ag presentation but also
of many distinct cell types, independently of cell-specific
glycosylation patterns. Secondly, it is not clear whether gI
constitutes a major protective target of TCR
cell-mediated
protection in vivo. The use of an engineered gI-deficient HSV-1 to
address this question is not possible since this virus does not
propagate well in vivo and its virulence is attenuated (38, 39, 40). It
might now be possible to address this question by creating a
recombinant HSV-1 that contains mutations in the amino terminus that
disrupt TCR
cell gI recognition without altering the essential
functions of gI in HSV-1 virulence.
| Acknowledgments |
|---|
| Footnotes |
|---|
2 Current address: Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305. ![]()
3 Address correspondence and reprint requests to Dr. Jeffrey A. Bluestone, Committee on Immunology and Ben May Institute for Cancer Research, MC1089, University of Chicago, 5841 South Maryland Avenue, Chicago, IL 60637. ![]()
4 Abbreviations used in this paper: HSV, herpes simplex virus; wt, wild type; GAL, galactose; GALNAC, galactosamine; gI, glycoprotein I; CHO, Chinese hamster ovary. ![]()
Received for publication March 19, 1998. Accepted for publication July 7, 1998.
| References |
|---|
|
|
|---|
cells: a specialized T-cell subset in the immune system. Annu. Rev. Cell Dev. Biol. 11:307.[Medline]
/
cells. Annu. Rev. Immunol. 11:637.[Medline]
/
cells protect mice from herpes simplex virus type 1-induced lethal encephalitis. J. Exp. Med. 185:1969.
/
on major histocompatibility complex antigen-specific CD3+, CD4-, CD8- T lymphocytes. J. Exp. Med. 168:1899.
/
T cells. Annu. Rev. Immunol. 14:511.[Medline]
-chain is diverse and distinct from that of fetal thymocytes. Nature 331:627.[Medline]
cells: mysterious cells of the immune system. Immunol. Res. 13:268.[Medline]

cell. J. Immunol. 152:5392.[Abstract]
T cells. Cell 76:29.[Medline]
T cell, G8. J. Exp. Med. 185:1223.
domain of a human
T-cell antigen receptor. Nature 391:502.[Medline]
by an antigen derived from Mycobacterium tuberculosis. Cell 57:667.[Medline]
T cells in the primary immune response to Mycobacterium tuberculosis. Science 244:713.
- T cells respond to mycobacterial heat-shock protein. Nature 340:309.[Medline]
9/V
2 T cells of a GroEL homolog on Daudi Burkitts lymphoma cells. Science 250:1269.
T cell receptor heterodimer. Science 245:746.
receptor-bearing T lymphocytes. Nature 330:262.[Medline]
T-cell receptors. Proc. Natl. Acad. Sci. USA 86:5928.
9 region. Nature 346:572.[Medline]
/
T cells, is encoded by an immunoglobulin superfamily gene (Blast-1) located in the CD1 region of human chromosome 1. J. Exp. Med. 173:1339.
T lymphocyte clone specific for herpes simplex virus glycoprotein I. J. Immunol. 148:983.[Abstract]
/
T cells are stimulated by protease-resistant ligands. Eur. J. Immunol. 20:1175.[Medline]
T cells. Proc. Natl. Acad. Sci. USA 91:8175.
T cells by nonpeptidic mycobacterial ligands. Science 264:267.
9V
2 T lymphocytes by mycobacterial low molecular weight ligand. Eur. J. Immunol. 24:1886.[Medline]
This article has been cited by other articles:
![]() |
A. A. Dandekar, K. O'Malley, and S. Perlman Important Roles for Gamma Interferon and NKG2D in {gamma}{delta} T-Cell-Induced Demyelination in T-Cell Receptor {beta}-Deficient Mice Infected with a Coronavirus J. Virol., August 1, 2005; 79(15): 9388 - 9396. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. A. Dandekar and S. Perlman Virus-Induced Demyelination in Nude Mice Is Mediated by {gamma}{delta} T Cells Am. J. Pathol., October 1, 2002; 161(4): 1255 - 1263. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Mertsching, A. L. Wurster, C. Katayama, J. Esko, F. Ramsdell, J. D. Marth, and S. M. Hedrick A mouse strain defective for {alpha}{beta} versus {gamma}{delta} T cell lineage commitment Int. Immunol., September 1, 2002; 14(9): 1039 - 1053. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Lundberg and G. A. Splitter gamma delta + T-Lymphocyte Cytotoxicity against Envelope-Expressing Target Cells Is Unique to the Alymphocytic State of Bovine Leukemia Virus Infection in the Natural Host J. Virol., September 15, 2000; 74(18): 8299 - 8306. [Abstract] [Full Text] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |