Biased T Cell Receptor Usage Directed
against Human Leukocyte Antigen DQ8-Restricted
Gliadin Peptides Is Associated with Celiac Disease
Sophie E. Broughton,1,8Jan Petersen,1,8Alex Theodossis,1Stephen W. Scally,1Khai Lee Loh,1Allan Thompson,2
Jeroen van Bergen,2Yvonne Kooy-Winkelaar,2Kate N. Henderson,1Travis Beddoe,1Jason A. Tye-Din,3
Stuart I. Mannering,4Anthony W. Purcell,1James McCluskey,5Robert P. Anderson,3,7,9Frits Koning,2,9Hugh H. Reid,1,9
and Jamie Rossjohn1,6,9,*
2Department of Immunohematology and Blood Transfusion, Leiden University Medical Center, 2333 ZA Leiden, The Netherlands
3The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, Victoria 3052, Australia
4St. Vincent’s Institute of Medical Research and Department of Medicine, University of Melbourne, St Vincent’s Hospital, Fitzroy,
Victoria 3065, Australia
5Department of Microbiology & Immunology, University of Melbourne, Parkville, Victoria 3010, Australia
6Institute of Infection and Immunity, Cardiff University School of Medicine, Heath Park, Cardiff CF14 4XN, UK
7Present address: ImmusanT, Inc., One Kendall Square, Suite B2004, Cambridge , MA 02139, USA
8These authors contributed equally to this work and are co-first authors
9These authors contributed equally to this work and are co-senior authors
Celiac disease is a human leukocyte antigen
(HLA)-DQ2- and/or DQ8-associated T cell-mediated
disorder that is induced by dietary gluten. Although
it is established how gluten peptides bind HLA-
DQ8and HLA-DQ2, it
peptide-HLA complexes are engaged by the T cell
receptor (TCR), a recognition event that triggers
disease pathology. We show that biased TCR
usage (TRBV9*01) underpins the recognition of
HLA-DQ8-a-I-gliadin. The structure of a prototypi-
cal TRBV9*01-TCR-HLA-DQ8-a-I-gliadin complex
shows that the TCR docks centrally above HLA-
DQ8-a-I-gliadin, in which all complementarity-deter-
mining region-b (CDRb) loops interact with the
gliadin peptide. Mutagenesis at the TRBV9*01-TCR-
HLA-DQ8-a-I-gliadin interface provides an energetic
basis for the Vb bias. Moreover, CDR3 diversity
accounts for TRBV9*01+TCRs exhibiting differing
reactivities toward the gliadin epitopes at various
deamidation states. Accordingly, biased TCR usage
is an important factor in the pathogenesis of DQ8-
mediated celiac disease.
is unclearhow such
The adaptive immune response is regulated by human leukocyte
antigen (HLA) molecules of the major histocompatibility complex
(MHC) via interactions with the clonally distributed antigen (Ag)-
specific ab T cell receptor (TCR) expressed on T cells. HLA class
II molecules, and in particular HLA-DQ8 that bears the b57 poly-
morphism, have long been associated with the development of
diseases, including type 1 diabetes and celiac disease (Sollid
et al., 1989; Spurkland et al., 1992; Todd et al., 1987). However,
there are few examples in which the mechanism of HLA-associ-
ated disease is well understood.
Celiac disease is a common, T cell-mediated, T cell inflamma-
tory disorder with autoimmune components, which is triggered
by the ingestion of dietary wheat gluten or related proteins
from rye and barley (Abadie et al., 2011; Jabri and Sollid,
2009). Celiac disease is predominantly limited to genetically pre-
disposed individuals, specifically those who express HLA-DQ2
(A1*0501-B1*0201) and/or HLA-DQ8 (A1*0301-B1*0302) (Sollid,
2002). In celiac disease patients, 90%–95% express HLA-DQ2
molecules and HLA-DQ2?patients usually express HLA-DQ8
(Karell et al., 2003). In HLA-DQ2-associated celiac disease,
intestinal T cell clones and lines are virtually all HLA-DQ2
restricted and recognize a range of peptides derived from each
of the subfractions of wheat gluten and homologous sequences
in barley hordeins and rye secalins (Anderson et al., 2000;
Arentz-Hansen et al., 2000, 2002; Sjo ¨stro ¨m et al., 1998; Vader
et al., 2003). Fine mapping of HLA-DQ2-restricted peptides
has revealed a preference for glutamate at anchor positions
P4 or P6, and sometimes P7, corresponding to glutamine resi-
dues susceptible to tissue transglutaminase (TG2) deamidation
(Arentz-Hansen et al., 2000, 2002; Qiao et al., 2005; Vader
et al., 2002b).
Despite HLA-DQ2 conveying a greater risk of celiac disease,
gluten-reactive CD4+T cells recognizing certain HLA-DQ8-
restricted determinants have been isolated from patients with
celiac disease who possess both HLA-DQ2 and HLA-DQ8 and
from patients who possess HLA-DQ8 but not HLA-DQ2 (Tollef-
sen et al., 2006; van de Wal et al., 1998a, 1998b, 1999). T cells
specific for the DQ8-a-I-gliadin determinant (EGSFQPSQE)
are commonly isolated from HLA-DQ2+HLA-DQ8+
donors and this gluten determinant is often immunodominant
Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc. 611
Tollefsen et al., 2006; van de Wal et al., 1998b). In contrast to
the more absolute deamidation dependence and relative pro-
tease resistance of the immunodominant gluten peptides in
HLA-DQ2-mediated disease, the immunodominant DQ8-a-I-
gliadin determinant derives from a protease-sensitive region of
a-gliadin, and a commonly recognized determinant from wheat
high molecular glutenin does not require deamidation (Shan
et al., 2002; Tye-Din and Anderson, 2008; van de Wal et al.,
A network of polar interactions enhance the binding of the
deamidated gliadin peptide DQ2-a-I-gliadin (PFPQPELPY) to
HLA-DQ2 (Kim et al., 2004) and DQ8-a-I-gliadin (EGSFQPSQE)
to HLA-DQ8 (referred to as DQ8-glia-a1) (Henderson et al.,
2007b; Sollid et al., 2012). Recently it has been suggested
that, because of the b57 Asp/Ala polymorphism in HLA-DQ8,
TCRs bearing a negatively charged residue within the CDR3b
loop are preferentially selected to interact with HLA-DQ8
complexed to native gluten peptides (Abadie et al., 2011;
Hovhannisyan et al., 2008). The subsequent binding of deami-
dated gluten peptides to HLA-DQ8 generates a stronger T cell
response and broadens the T cell repertoire to include TCRs
that do not contain a negatively charged residue within the
CDR3b loop. Presently, however, the nature of the ab T cell
repertoire directed against HLA-DQ8-restricted gliadin peptides
is unclear, and moreover it is unknown how these TCRs, which
are implicated in celiac disease pathology, engage the DQ8-
glia-a1 complex. We show that biased TCR usage underpins
DQ8-glia-a1 recognition and we provide a structural and ener-
getic basis for this key interaction in HLA-DQ8-mediated celiac
Biased TRBV9*01 Usage against DQ8-Glia-a1
To examine the T cell response in HLA-DQ8-mediated celiac
disease, we previously isolated and expanded three DQ8-
glia-a1-restricted T cell clones (SP3.4, SP4.6, LS1.2) from the
PBMCs of two celiac disease patients (patients ‘‘SP’’ and
‘‘LS’’) (Henderson et al., 2007b; Mannering et al., 2005), all of
which specifically bound the DQ8-glia-a1 tetramer (Figure S1
available online). In addition, three previously isolated DQ8-
glia-a1-specific T cell clones (S13, L3-12, S16) from small intes-
tinal biopsies of HLA-DQ2+HLA-DQ8+patients (patients ‘‘S’’ and
‘‘L’’) and one (T316) from a small intestinal biopsy of a patient
(patient ‘‘T’’) expressing the HLA-DQ8 transdimer (A*0501-
B*0302) were characterized (Kooy-Winkelaar et al., 2011).
These seven DQ8-glia-a1-restricted T cell clones were ana-
lyzed for their ab TCR gene usage, CDR3 sequences, and
dose response by means of DQ8-glia-a1 deamidated peptide
analogs (Table 1; Figure 1). Clone LS1.2 used TRAV17*01-
TRBV6-1*01 pairing (Table 1). The T cell clones S16 and T316
1*01, respectively. The two clones isolated from patient SP,
SP3.4 and SP4.6, shared the same TCR b chain, TRBV9*01,
with an alternative TCR a chain pairing of TRAV26-2*01 and
TRAV38-2-DV8*01, respectively (Table 1). Further, the T cell
clones S13 and L3-12 were TRAV26-2*01-TRBV9*01+(Table
1). Notably, the three clones (SP3.4, S13, and L3-12) character-
ized that exhibit identical TRAV26-2-TRBV9*01 usage were
derived from three celiac disease patients (two adults, one child)
who are not related to each other. We also characterized the
TRAV-TRBV usage of two other previously isolated T cell clones,
S12 and T15 (Kooy-Winkelaar et al., 2011). S12 is specific for
DQ8-glut-H1, whereas T15 cross-reacts with glia-g1 presented
by either HLA-DQ8 or HLA-DQ8.5 (Kooy-Winkelaar et al.,
2011; van de Wal et al., 1999). Although both the T15 and S12
T cell clones exhibited different TRAV gene usage, they both
used the TRBV9*01 gene (Table 1). Although no sequence
homology was observed between the CDR3 loops of the nine
DQ8-restricted clones, the presence of a non-germline-encoded
Arg residue in the CDR3a loop of the TRBV9*01+DQ8-glia-a1-
specific TCRs was notable (Table 1).
The public TRBV9*01 usage in six independent HLA-DQ8-
restricted T cell clones, specific for a1-gliadin (or other gliadin
determinants), suggested that biased TCR usage was an impor-
tant factor underpinning this response. To investigate this, we
tested the expression of TRBV9*01 on CD4+T cells present in
peripheral blood of HLA-DQ8+and HLA-DQ8?individuals and
Table 1. The Variable Gene Usage of the HLA-DQ8-Gliadin or -Glutenin-Restricted TCRs
SP3.4DQ8-glia-a1 26-2*0145*01YYC ILRDGRGGADGLT FG
9*011*012-7*01CAS SVAVSAGTYEQ YFG
DQ8-glia-a126-2*0154*01YYC ILRDSRAQKLV FG
9*011*012-7*01CAS SAGTSGEYEQ YFG
DQ8-glia-a126-2*0149*01YYC ILRDRSNQFY FG9*011*012-5*01CAS STTPGTGTETQ YFG
DQ8-glia-a1 38-2-DV8*0131*01 YFC AYRSARGARLM FG
9*011*01 2-7*01 CAS SVAVSAGTYEQ YFG
DQ8-glia-a113-1*0237*01/02 YFC AGGSSNTGKLI FG4-2*01 - 2-2*01CAS SQDIRNTGEL FFG
17*0144*01YFC ATDFPGTASKLT FG
6-1*011*011-3*01CAS SEALPGRSGNTI YFG
8-3*0136*01 YFC AVGETGANNLF FG 6-1*012*02 2-1*01CAS SEARRYNEQ FGP
20*026*01 YLC AVQASGGSYIPT FG9*011*01 2-3*01CAS SNRGLGTDTQ YFG
DQ8-glut-H129-DV5*0149*01YFC AASAYPGNQFY FG9*012*012-5*01CAS SVYDGRGETQ YFG
Junctionally encoded residues are underlined. DQ8 epitope nomenclature according to Sollid et al. (2012). See also Figure S1 and Tables S1 and S2.
aTCR variable gene usage as defined in IMGT-V-QUEST Database (Brochet et al., 2008).
bThe TRBV usage and CDR3b sequences of clones S12, S13, and S16 have been published previously (Hovhannisyan et al., 2008).
cDQ8/8.5-glia-a1 = T316 T cell clone recognizes a-I-gliadin presented by HLA-DQ8 or HLA-DQ8.5.
dDQ8/8.5-glia-g1 = T15 T cell clone recognizes glia-g1 presented by HLA-DQ8 or HLA-DQ8.5.
DQ8-Mediated Celiac Disease
612 Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc.
of four celiac disease patients and one non-celiac disease
patient. On average, 2% of peripheral blood CD4+T cells ex-
pressed TRBV9*01 (n = 9, range 0.8%–3.7%, no marked differ-
ence between HLA-DQ8+and HLA-DQ8?healthy controls)
(Table S1). In contrast, elevated amounts of TRBV9*01 expres-
sion was observed in the four gluten-reactive T cell lines isolated
from celiac disease patients (average 15.8%, range 6.8%–
32.5%), whereas no such increase was observed in the T cell
line from a non-celiac disease patient (0.9%) (Table S1).
To formally demonstrate that TRBV9*01 expression correlates
with gliadin reactivity in HLA-DQ8+celiac disease patients, we
isolated TRBV9*01+and TRBV9*01?T cells from a gluten-reac-
tiveTcelllinederived fromabiopsyofaceliacdisease patientby
per well or 100 cells per well, expanded, and subsequently
tested for reactivity against a deamidated gliadin preparation
(Glia-TG2) and the deamidated DQ8-glia-a1 peptide (Table
S2). Although the results clearly demonstrate that gliadin reac-
tivity is found in both the TRBV9*01?and TRBV9*01+fractions,
there was a strong selection for reactivity with the DQ8-glia-a1
peptide in the TRBV9*01+cell fraction, as indicated by the fact
that 11 out of 12 gliadin-reactive T cell clones and 9 out of 9
T cell lines were found to specifically react with the DQ8-glia-
a1peptide (Table S2).Incontrast, in theTRBV9*01?cell fraction,
3 out of 10 T cell clones and 5 out of 7 T cell lines responded to
gliadin but only one of the cell lines and none of the clones re-
sponded to the DQ8-glia-a1 peptide. Similar results were found
with a gliadin-reactive T cell line from a second celiac disease
patient (not shown). Thus, TRBV9*01 expression is elevated
in gluten-reactive T cell lines derived from biopsies of celiac
disease patients and correlates with reactivity toward the DQ8-
glia-a1 peptide. Accordingly, biased TRBV9*01 usage appears
to be a prominent feature of the HLA-DQ8-restricted response
to the a-I-gliadin determinant.
positions, three TRAV26-2*01+-TRBV9*01+T cell clones and
three TRAV26-2*01?-TRBV9*01?T cell clones were tested
against a concentration range of four variants of the a-I-gliadin
peptide with a Q and/or E at position p1 and p9 (Figure 1). We
observed striking differences in sensitivity of the T cell clones
with an at least 100-fold difference between clone LS1.2 and
L3-12 (the least and most sensitive, respectively). Both the
TRAV26-2*01-TRBV9*01-positive and -negative T cell clones
displayed unique reactivity patterns, but in all cases stronger
responses were observed against peptides in which either one
or both of the Q residues at P1 and P9 were replaced by an E.
tivity toward the deamidated peptides. Namely, whereas clone
SP3.4 preferred the P1Q-P9E peptide to the P1E-P9Q variant,
clone L3-12 displays the opposite pattern and responded more
strongly to the P1E-P9Q peptide. In contrast, the third TRAV26-
2*01-TRBV9*01+T cell clone, S13, did not display a strong pref-
erence for either the P1Q-P9E or P1E-P9Q variant (Figure 1).
Thus, the three TRAV26-2*01+-TRBV9*01+T cell clones dis-
played unique reactivity patterns and are differentially affected
Figure 1. Deamidation Dependence of DQ8-Glia-a1-Restricted TCRs
Response of the T cell clones SP3.4 (A), S13 (B), L3-12 (C), LS1.2 (D), S16 (E), and T316 (F) to a concentration range of four variants of the DQ8-glia-a1 peptide:
P1Q-P9Q, black line; P1E-P9Q, blue line; P1Q-P9E, red line; P1E-P9E, purple line. All measurements were performed in triplicate. Data are mean ± SEM.
DQ8-Mediated Celiac Disease
Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc. 613
by deamidation of the Q at p1 and p9. Because the TRAV-TRBV
usage is conserved between these three clones, it indicates that
CDR3 sequence variability underpins the differing degrees of
deamidation dependence of these T cell clones.
Affinity for the DQ8-Glia-a1 Epitope
and LS1.2 TCRs (data not shown). Native gel-shift analysis
demonstrated that the SP3.4 TCR ligated to DQ8-glia-a1, yet
did not bind to an irrelevant pMHC class II (pMHC-II) (data not
shown). As determined by surface plasmon resonance (SPR)
analysis, the steady-state affinity (KD) of the SP3.4, L3-12, and
S13 TCRs for DQ8-glia-a1 were 11.4 ± 0.6 mM, 7.0 ± 0.5 mM,
and 1.0 ± 0.1 mM, respectively (Figure 2). However the affinity for
the LS1.2 TCR interaction with DQ8-glia-a1 was weaker (50.4 ±
2.2 mM) (Figure 2). These data correlate with the much weaker
reactivity of the LS1.2 clone in the gluten-specific proliferation
assay compared to the other clones (Figure 1). We also deter-
mined the kinetic rate constants for the SP3.4 TCR-DQ8-glia-a1
interaction, which exhibited association (ka) and dissociation (kd)
Figure 2. Affinity Data for the DQ8-Glia-a1-
Binding analysis of TCRs SP3.4, L3-12, S13, and
LS1.2 to DQ8-glia-a1 via surface plasmon reso-
nance. Concentration series of each TCR was
passed over surface immobilized DQ8-glia-a1.
Right column: Measured response curves of
single dilution series for each TCR. Left column:
Curve fits for TCR-DQ8-glia-a1 KDdetermination
with single ligand binding model. Data from
multiple (n =) measurements was combined after
normalizing each equilibrium response curve
against the calculated response maximum. Data
are mean ± standard deviation.
rate constants of 2.53 3 104± 0.29 3
104M?1s?1and 0.224 ± 0.029 s?1,
respectively (Figure 2), values that fall
within the range for typical TCR-pMHC
interactions. However, the affinity values
of the SP3.4 TCR, L3-12 TCR, and S13
TCR for DQ8-glia-a1 were higher than
that typically observed for microbial and
non-self TCR-pMHC-II (?30 mM) (Cole
et al., 2007) and much higher than the
low affinity for autoreactive TCR-pMHC
complexes (?100–200 mM) (Bulek et al.,
2012; Deng and Mariuzza, 2007). Indeed,
the affinity of the SP3.4 TCR, L3-12 TCR,
and S13 TCR toward TCR-DQ8-glia-a1
TCR-pMHC-I complexes (?1–10 mM)
(Clements et al., 2006; Cole et al., 2007;
Godfrey et al., 2008).
Structural Overview of the SP3.4
To establish how the SP3.4 TCR interacts
of the SP3.4 TCR-DQ8-glia-a1 complex at 3.2 A˚resolution
(Figures 3A and 3B; Table 2). Further, we determined the struc-
ture of the SP3.4 TCRin the nonliganded state at3.2 A˚resolution
(Table 2). Together with the binary DQ8-glia-a1 complex previ-
ously determined (Henderson et al., 2007b), we were thus able
to establish the extent of plasticity in the SP3.4 TCR-DQ8-glia-
The SP3.4 TCR docks centrally above HLA-DQ8, at approxi-
mately 70?across the long axis of the Ag-binding cleft,
where the Va and Vb domains of SP3.4 TCR sits above the
b and a helix of HLA-DQ8, respectively (Figure 3C). The buried
surface area (BSA) at the SP3.4 TCR-DQ8-glia-a1 interface
is ?880 A˚2, a value at the lower end of the range for TCR-
pMHC-II complexes (Table S3; Rudolph et al., 2006). Thus, the
overall docking mode of the SP3.4 TCR-DQ8-glia-a1 complex
is comparable to other TCR-pMHC structures determined to
date (Burrows et al., 2010) and does not display the ‘‘nonstan-
dard’’ docking modes that have generally been associated
with autoimmune TCR-pMHC-II complexes (Deng and Mar-
iuzza, 2007), which may be a consequence of the SP3.4 TCR
DQ8-Mediated Celiac Disease
614 Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc.
recognizing a non-self Ag restricted to HLA-DQ8. Further, the
HLA-DQ8-restricted TCR does not reside over the b57 position
of HLA-DQ8 (Figures 3B and 4).
To engage the DQ8-glia-a1, the CDR loops of the SP3.4 TCR
show limited plasticity, with a movement in the CDR1a loop
(main chain rmsd, 1.2 A˚) and CDR3a loop (main chain rmsd for
observed atoms, 2.0 A˚), whereas all the CDRb loops do not
appreciably change conformation upon ligation (not shown).
Further, the Ag-binding clefts of the nonliganded DQ8-glia-a1
(Henderson et al., 2007b) and DQ8-glia-a1 from the ternary
complex structure overlay closely, with an rmsd of 0.35 A˚, and
there are only two side chains that move notably to accom-
modate the SP3.4 TCR. These are the side chains of Val67
and Arg70 from HLA-DQ8, which reorients to avoid steric
clashes with Tyr114b and, consequently, forms interactions
with CDR3a and the CDR3b loops of the SP3.4 TCR (Figure S3).
Thus, in comparison to the plasticity of the CDR loops seen
in many TCR-pMHC interactions, a relatively rigid interaction
underpins the SP3.4 TCR-DQ8-glia-a1 complexation, although
some changes were observed in the Ca domain that may relate
to signaling (Beddoe et al., 2009; Ishizuka et al., 2008; Garcia
et al., 1998; Tynan et al., 2007). The SP3.4 forms a number of
van der Waals (vdw) interactions and ten hydrogen bonds with
DQ8-glia-a1 (Table S3). The Va and Vb domains of the SP3.4
TCR contribute 43% and 57% to the BSA, respectively, at the
DQ8-glia-a1 interface with the Va and Vb chains contacting
both helices of HLA-DQ8. Within this interface, all six CDR loops
contribute to the interaction, but to varying degrees (Table S3).
The CDR1a and CDR2a loops of the SP3.4 TCR contribute
14% and 5% to the BSA, respectively, whereas the CDR1b
and CDR2b loops contribute 6% and 11%, respectively.
(A and B) Structural overview (A) and close up (B)
of the Ag-binding interface.
(C) The SP3.4 TCR-DQ8-glia-a1 footprint.
The b57 position is colored magenta. CDR1a,
CDR2a, and CDR3a are colored red, pink, and
cyan, respectively; CDR1b, CDR2b, and CDR3b
colored orange, purple, and blue, respectively.
The DQ8 a and b chains are colored light gray and
green, respectively. The a-I-gliadin peptide is in
yellow. The SP3.4 a and b chain framework
regions are colored dark gray and lime green,
respectively. N and C termini of SP3.4 and DQ8
are shown as blue and red spheres, respectively.
See also Figure S3 and Table S3.
Notably, the CDR3a and CDR3b loops
contribute 23% and 29% of the BSA,
respectively, and thus the CDR3 loops
dominate contacts, interacting with both
the a-I-gliadin peptide and HLA-DQ8.
The SP3.4 TCR exclusively interacts
with the HLA-DQ8 a chain via vdw con-
tacts, binding to a region of the a1 helix
The interactions are mediated principally
inwhichPhe58sitsunderneaththe CDR3a loop, with its aromatic
ring packing against Arg110a. Arg66b, a framework residue adja-
cent to the CDR2b loop, packs against Thr61 and Leu60, while
Tyr57b interacts with Ala64 and points toward His68, with the
The b1 helix of HLA-DQ8 sits underneath the CDR1a and
CDR2a loops, but also mediates contacts with the CDR3a and
CDR3b loops largely as a result of the long side chain of Arg70
that stretches across the Ag-binding cleft, and subsequently
forms vdw interactions and H bonds with Gly109a and
Thr113b of SP3.4 TCR. Thr36a and Tyr38a from the CDR1a
phobic focal point on the b1 helix, interacting with Ala73, Thr77,
and His81 of HLA-DQ8 (Figure 4B). Accordingly, the SP3.4 TCR
adopts a standard and central docking topology over HLA-DQ8.
Interactions with the a-I-Gliadin Peptide
Although the extent of contacts between the SP3.4 TCR and
HLA-DQ8 is relatively small, the interaction with the a-I-gliadin
peptide is, by comparison, more extensive, with the CDR3a
loop and the CDR1b, CDR2b, and CDR3b loops contacting posi-
tions P1-Glu, P2-Gly, P3-Ser, P5-Gln, and P8-Gln of the peptide
(Figure 4C). Of the ten H bonds present at the SP3.4 TCR-DQ8-
glia-a1 interface, eight are mediated by peptide residues. This
distribution of H bonds suggests that the largest energetic
contribution to SP3.4 TCR-DQ8-glia-a1 binding is provided by
polar interactions with the peptide. A key residue appears to
be Arg110aof the CDR3aloop with its side chain oriented along-
side the peptide and toward its N terminus. The guanidinium
moiety of Arg110a packs tightly against Phe58a of HLA-DQ8
while also forming four H bonds to the backbone and side chains
DQ8-Mediated Celiac Disease
Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc. 615
of the a-I-gliadin peptide (at P1-Glu and P3-Ser) (Figures 4C and
S4). In the unliganded structure, the CDR3a loop is disordered.
However, upon ligation, its conformation, in particular Arg110a,
is stabilized by an intrachain salt bridge to the adjacent
Asp108a and by a H bond between Asp108a and the backbone
The P5-Gln and P8-Gln are the prominent upward-facing resi-
dues of the gliadin peptide, with the CDR3b loop being posi-
tioned in between these two residues and P8-Gln having to
reorientate its side chain to accommodate the CDR3b loop.
P5-Gln contacts the CDR3a and CDR3b loops of the SP3.4
TCR, forming a H bond with the main chain of Arg110a and
vdw contacts with Ser 111b (Figure 4C). P8-Gln contacts the
CDR1b, CDR2b, and the CDR3b loops and forms H bonds with
the main chain of Val108b and side chain of Tyr57b, from the
CDR3b and CDR2b loops, respectively, as well as vdw contacts
with Ala109b and Leu37b from the CDR3b and CDR1b loops,
respectively. Accordingly, the P5-Gln-X-X-P8-Gln sequence
mediates many contacts with the SP3.4 TCR.
Peptide Library Scan
role in the interaction with the SP3.4 TCR, the specificity of
the SP3.4 T cell clone was analyzed for reactivity to a library
of 1,363 native and in silico deamidated gliadin-derived
18-mers comprising 8,114 10-mer candidate determinants, via
Table 2. Data Collection and Refinement Statistics
Resolution limits (A˚)
Cell dimensions (A˚)
a = 111.71,
b = 134.33,
c = 140.30
a = 124.56,
b = 124.56,
c = 61.10, g = 120?
Total number of
Number of unique
Data completeness (%)
91.0 (93.8) 87.2 (89.7)
8.4 (26.1) 7.2 (38.9)
<I/s(I)> 9.1(2.6)14.6 (2.2)
Multiplicity 4.6 (4.6)2.0 (2.0)
R factor (%)c
R free (%)d
Number of Atoms
Protein and peptide11,879 6,107
Average B Factor (A˚2)
Main chain atoms 43.885.3
Side chain atoms42.684.4
Rmsd bond lengths (A˚)
Rmsd bond angles (?)0.9751.049
See also Figure S2.
aValues in parentheses refer to the highest resolution bin.
bRpim= Sh[1/(N?1)]1/2SijIi(h)?<j(h) >jShSiIi(h) where I is the observed
intensity and <I> is the average intensity of multiple observations from
dCalculated as for the Rfactor(see footnote c) using 5% of reflections.
eOutput from the MolProbity Server (Chen et al., 2010).
Figure 4. Interactions at the Interface of SP3.4 TCR-DQ8-Glia-a1
(A) DQ8-a chain-SP3.4 contacts.
(B) DQ8-b chain-SP3.4 contacts.
(C) a-I-gliadin peptide-SP3.4 contacts.
The TCR and pHLA are displayed in cartoon format, with key residues shown
as sticks. The residue color scheme is as for Figure 3. vdw interactions are
highlighted by beige dashes and hydrogen bonds by blue dashes. See also
Figure S4 and Table S4.
DQ8-Mediated Celiac Disease
616 Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc.
an overnight interferon-g (IFN-g) ELISpot assay (Tye-Din et al.,
2010). The SP3.4 T cell clone recognized a wide variety of dea-
midated and native gliadin peptides, all containing highly homol-
ogous sequences to the core P1-P9 of the a-I-gliadin peptide
(Table S4; Figure 1). Indeed, peptides containing all permuta-
tions of P1-Gln or Glu and P9-Gln or Glu were among the
peptides eliciting the strongest response. These data support
the earlier T cell proliferation results indicating that the SP3.4
T cell clone recognizes both deamidated and native gliadin
peptides in the context of HLA-DQ8 (Henderson et al., 2007b).
Moreover, six of the top ten peptides possess the Gln-X-X-Gln
motif, thereby highlighting the role this motif plays in SP3.4
TCR recognition of DQ8-glia-a1.
Energetic Basis of the TRBV9*01 Bias
Next, we examined the energetic basis underpinning the biased
TRBV9*01 usage directed toward the DQ8-glia-a1 complex.
Solvent-exposed SP3.4 TCR residues from the TCRb chain,
whose side chains interacted with DQ8-glia-a1 were selected
for substitution to alanine. We also mutated Arg110a and
Asp108a from the CDR3a loop, because Arg110a represented
the most prominent DQ8-glia-a1 contact point on the Va chain,
and Asp108a appeared to stabilize the conformation of
Arg110a. In total, 13 SP3.4 TCR amino acid substitutions were
made and included CDR3a: Asp108aSer, Arg110aAla, Ar-
g110aLys; CDR1b: Leu37bAla; CDR2b: Tyr57bPhe, Tyr57bAla;
Vb framework: Arg66bAla; CDR3b: Val108bAla, Ala109bGly,
Val110bAla, Ser111bAla, Thr113bAla, Tyr114bAla. All of the
mutant SP3.4 TCRs expressed and refolded with similar yield
asto wild-type (WT)SP3.4 TCRand exhibited similar biophysical
properties to the WT SP3.4 TCR (not shown). The impact of the
mutants on the affinity of the SP3.4 TCR-DQ8-glia-a1 interaction
were assessed by SPR analysis. The effects of the SP3.4 TCR
mutants were grouped into three categories: no effect,
a moderate effect (KDbetween 35 and 57 mM), and a marked
effect (KD> 57 mM or KD< 2.3 mM) (Figures 5 and S5).
Mutation of Arg110a to Ala or Lys abrogated DQ8-glia-a1
binding, thereby underscoring the key role this residue plays
in contacting the a-I-gliadin determinant and HLA-DQ8. Further,
the Arg110aLys mutant highlighted the importance of its guan-
danium group in forming multiple H bonding contacts with
DQ8-glia-a1. Mutation of Asp108a also abrogated DQ8-glia-a1
binding, thereby highlighting the importance of the Asp108a-X-
Arg110a sequence and associated conformation in enabling
binding (Figures5A and 5B). Of the nine positions mutated within
the TCRb chain, only two residues (Leu37b and Tyr57b) were
critically important for the interaction; one residue (Ser111b)
moderately impacted on the affinity of the interaction, and muta-
tion of five residues (Val108b, Ala109b, Val110b, Thr113b, and
Tyr114b) had no impact on the interaction with DQ8-glia-a1
(Figures 5A and 5B). The moderate impact of the Ser111bAla
mutant could be attributed to the loss of a stabilizing contact
with the CDR3a loop as opposed to a direct reduction in the
number of contacts with the P5-Gln residue. Thus, despite the
CDR3b loop contributing 29% BSA to the SP3.4 TCR-DQ8-
glia-a1 interface, the amino acid side chains within the CDR3b
loop of the SP3.4 TCR are not a critical factor in mediating
binding with DQ8-glia-a1. This indicates that a diverse repertoire
of CDR3b sequences would potentially be permissive to DQ8-
The germline-encoded Leu37b and Tyr57b residues are posi-
tioned closely together, situated above HLA-DQ8 a chain, with
Leu37b interacting with the His68 side chain and Tyr57b packing
against the main chain of the a1 helix. Both of these residues
form vdw contacts with P8-Gln of the a1-gliadin determinant,
with Tyr57b also forming a H bond with P8-Gln, although the
effect of the Tyr57bPhe mutant shows that this interaction is
not energetically required for DQ8-glia-a1 ligation. Accordingly,
residues from within the CDR3a loop, Leu37b from the CDR1b
loop, and Tyr57b from the CDR2b loop are the ‘‘hot-spot’’ resi-
dues underpinning the SP3.4 TCR-DQ8-glia-a1 interaction
providing a basis for the TRBV9*01 bias (Figures 5A and 5B).
Although it is established that CD4+T cells play a pivotal role in
DQ8- and DQ2-mediated celiac disease, little was known
Figure 5. Energetic Basis of TRBV9*01 Bias
Energetic hot spots of the SP3.4-DQ8-glia-a1 interface.
and energetically important SP3.4 contact residues shown as sticks and
colored according to the changed KDof the mutant (red, KD> 57 mM; yellow,
KDbetween 35 and 57 mM; green, KD< 2.3 mM). The corresponding contact
residues of DQ8 are shown in gray (a chain), cyan (b chain), and pink (peptide).
Interactions between the highlighted residues are shown as dashes (blue, H
bonds; beige, vdw contacts).
(B) Comparison of the measured equilibrium dissociation constants of SP3.4
mutants. Differences in KDwere classed as moderate (KDbetween 35 and
57 mM, yellow bar) and marked (KD> 57 mM [red] or KD< 2.3 mM [green]).
Horizontal lines show 1, 3, and 5 times multiples of the WT SP3.4 KD.
Data are mean ± standard deviation. See also Figure S5.
DQ8-Mediated Celiac Disease
Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc. 617
regarding the nature of the responding T cell repertoire and how
the pathogenic TCRs interacted with the gliadin determinants
complexed to the disease-associated HLA class II molecules.
Previously it was shown that DQ8-native gliadin complexes
select TCRs that harbor a negatively charged residue within
gliadin determinants to HLA-DQ8 generates a stronger T cell
response and diversifies the T cell repertoire (Hovhannisyan
et al., 2008). We show that biased TCR usage underpins the
adaptive immune response to the deamidated DQ8-glia-a1
peptide complexed to HLA-DQ8. Biased TCR usage is often
linked to TCR recognition of atypical (featured or featureless)
pMHC landscapes in viral immunity (Turner et al., 2006).
However, biased Vb6.7 (TRBV7-2*01) usage has also recently
been described in the adaptive immune response to DQ2.5-
glia-a2, although the structural basis for this biased gene usage
remains unclear (Qiao et al., 2011). Our present findings extend
this to HLA-DQ8 and provide a structural basis for the observed
biased TCR usage in HLA-DQ8-associated celiac disease.
To address the basis of TRBV9*01 bias, we determined the
structure of a prototypical TRBV9*01 TCR, SP3.4, in complex
with DQ8-glia-a1, and undertook associated mutagenesis
and affinity studies. The TRAV26-2-TRBV9*01-HLA-DQ8-glia-
a1 structure did not exhibit an atypical binding mode and the
central docking topologyalsoprecludedthe SP3.4TCRinteract-
ing with theb57residue on HLA-DQ8,a polymorphic sitelocated
at the base of the C-terminal end of the Ag-binding cleft, which is
linked to increased risk of autoimmune disease in humans and
mice possessing the HLA-DQ8 and I-Ag7alleles, respectively.
tive TCR-I-Ag7HEL11-27complex (Yoshida etal., 2010),the SP3.4
TCR did not interact with residues of the P9 pocket, the P9-Glu,
or the b57 position.
Despite the TRBV bias, the non-germline-encoded CDR3
loops dominated the interactions with DQ8-glia-a1; however,
the mutagenesis data indicated that the CDR3b loop was
dispensable for the interaction, which is consistent with its
lack of sequence conservation among DQ8-glia-a1-restricted
TCRs. These observations resonate with CDR3b variability in a
recently described antiviral response that exhibited a biased
T cell repertoire (Day et al., 2011). Instead, the energetic hot
spot within the TRBV9*01 chain resided with two residues only,
Leu37b and Tyr57b, in the CDR1b and CDR2b loops, respec-
tively. Recently, the structure of a TRBV9*01+TCR (termed
TK3) complexed to a pMHC-I has been reported, permitting
a broader perspective on the role of this CDR2b loop in MHC-
restricted immunity (Gras et al., 2010). Namely, Tyr57b of the
TK3 TCR solely contacts the virally encoded peptide, whereas
the principal role of Tyr57b in SP3.4 TCR is to mediate contacts
with HLA-DQ8. This clearly highlights the variable role identical
germline-encoded regions can play in MHC immunity. Indeed,
it was surprising that the TRBV bias is essentially limited to two
residues forming essential specificity-governing contacts with
HLA-DQ8, suggesting that the TRBV9*01 usage may relate to
other factors, such as TRAV-TRAJ pairing, analogous to
a recently described observation (Stadinski et al., 2011; Turner
and Rossjohn, 2011).
Some of the DQ8-a1-glia-restricted TCRs displayed a lesser
dependence on P1 and/or P9 deamidation, suggesting that
higher-affinity TCRs provide additional stability to the TCR-
pMHC-II complex, which ordinarily would be provided by the
greater stability of the deamidated HLA-DQ8 complex. In
contrast, the LS TCR possessed much weaker affinity for
the DQ8-glia-a1 complex and was more sensitive to deamida-
tion of the a1-gliadin determinant, thereby suggesting that
TCR-pMHC-II affinity is linked to the absolute requirement for
deamidation dependence on the interaction. This resonates
with a recent observation whereby autoimmunity can be
triggered by high-affinity TCRs binding to weak self-pMHC
complexes (Yin et al., 2011). Interestingly, a differing extent of
deamidation dependence was observed in the three biased
TRAV26-2*01-TRBV9*01 TCRs, indicating that the differing
CDR3 usage can modulate the fine specificity of the DQ8-
a1-glia-mediated response, which is also consistent with
TCRs being able to recognize other gliadin
determinants. Moreover, the variable degree in which the T cell
clones depend on deamidation at the P1 and/or P9 position
may reflect the fact that deamidation of the two Q residues at
the P1 and P9 position is variable, indicating that all four variants
of the DQ8-glia-a1 peptide probably exist in vivo and elicit T cell
For a full understanding of celiac disease pathogenesis, it will
be necessary to unravel the interactions between the lamina
propria and epithelial compartments (DePaolo et al., 2011).
There is no doubt that the T cell response to gluten pep-
tides bound to HLA-DQ2 and/or DQ8 plays a crucial role.
Accordingly, we report a celiac disease-associated TCR-
pMHC-II ternary complex and thereby provide insight into the
structural requirements that govern TCR recognition of antigenic
HLA-DQ8-restricted gliadin peptides. The mode of recognition
by the SP3.4 TCR is likely to be shared by a number of
TRBV9*01+TCRs implicated in HLA-DQ8-mediated celiac
disease. Thus, it may be important to target such selected
TCRs in immunotherapy.
Gliadin (Sigma G3375) was incubated for 4 hr at 37?C in 10-fold excess with
chymotrypsin (Sigma C3142) in ammonium bicarbonate (pH 8) and finally
boiled for 15 min. Protein concentration was determined by the BCA method
(Pierce, USA). Deamidation with guinea pig liver TG2 (Sigma T5398) was as
described previously (Anderson et al., 2000, 2005).
Isolation of Gluten-Specific T Cell Lines and Clones from Intestinal
Polyclonal gluten-specific T cell lines were isolated from the small intestine of
celiac disease patients, as described (Vader et al., 2002b). T cell clones were
isolated from HLA-DQ2 or HLA-DQ8 heterozygous patients, except clones
T15 and T316, which were isolated from a rare patient that expresses nei-
ther HLA-DQ2 or HLA-DQ8 but does express the HLA-DQ8 transdimer
(HLA-DQ8.5) (A*0501-B*0302) (Kooy-Winkelaar et al., 2011). The study was
approved by the Medical Ethics Committees of the Free University Medical
Center and the Leiden University Medical Center. Written informed consent
was obtained from each subject before enrollment. TRBV9*01-positive cells
in peripheral blood and T cell lines and clones isolated from small intestinal
biopsies were stained with a FITC-labeled monoclonal antibody specific
for TRVB9 (clone BL37.2; Beckman Coulter). Cells were either analyzed for
expression of TRBV9*01 or sorted into TRBV9-positive and -negative cell frac-
tions by fluorescence activated cell sorting. See Supplemental Experimental
Procedures for more information.
DQ8-Mediated Celiac Disease
618 Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc.
T Cell Proliferation Assays
Proliferation assays were performed in triplicate in 150 ml IMDM supple-
mented with glutamine (Life Technologies, Invitrogen, Grand Island, NY) and
10% human serum in 96-well flat-bottom plates. In brief, APCs were loaded
with antigen for 2 hr, after which 15,000 T cells were added. We used
30,000 mitomycin C-treated (Sigma) EBV-LCL cells (cell line BSM, HLA-
DQ8+: DQA1*0301; DQB1*0302) as APCs. Unless otherwise indicated, TG2-
treated gliadin (Glia-TG2) was used at a final concentration of 450 mg/ml,
and synthetic peptides were used at the concentrations indicated. After
48 hr at 37?C, cultures were pulsed with 0.5 mCi [3H]thymidine and harvested
18 hr later.
Peptide Library Design
A comprehensive wheat gliadin 18-mer peptide library was designed to
encompassall unique10aminoacidsequencesin131entriesforTriticum aes-
GenBank (October 2006). The 1,363-member library was synthesized in
batches of 96 peptides as Pepsets (Pepscan, Netherlands). Peptides were
initially dissolved in 50% aqueous acetonitrile (v:v) to 12.5 mM (?25 mg/ml).
Rather than pretreat peptides with and without TG2, the peptide library was
designed to encompass all unique 10-mers with and without deamidation
based on in silico deamidation according to the motifs Gln-X1-Pro-X3and
Gln-X1-X2-Y3,where X1and X3are any amino acids except proline, Pro is
proline, X2is any amino acid, and Y3is any hydrophobic amino acid (Phe,
Tyr, Trp, Ile, Leu, Val, Met) (Fleckenstein et al., 2002; Vader et al., 2002a).
The peptide library was assessed by IFN-g ELISpot assay at a concentration
of 25 mg/ml.
Overnight IFN-g ELISpot assays (Mabtech; Sweden) with 96-well plates
(MSIP-S45-10; Millipore) were performed by a modification to the manufac-
turer’s instructions as previously described (Anderson et al., 2000, 2005).
See Supplemental Experimental Procedures.
TCR Isolation and Sequence Analysis
Total RNA was extracted from the LS1.2, SP3.4, and SP4.6 T cell clones via
TRIzol reagent (Invitrogen). cDNA clones encoding the TCR a and b chains
were produced from the total RNA with the 50RACE System for Rapid Ampli-
instructions. TotalRNAwasisolated from L3-12,S13,S16,T316,T15,and S12
T cell clones employing the RNeasy Mini Kit (QIAGEN) and cDNA synthesized
with superscript reverse transcriptase III (Invitrogen, USA) according to the
manufacturer’s instructions. See Supplemental Experimental Procedures.
Protein Expression and Purification
TCRs were expressed, refolded, and purified with an engineered disulfide
linkage in the constant domains essentially as previously described (Boulter
et al., 2003; Garboczi et al., 1996). The extracellular domains of the HLA-
DQ8 a and b chains were coexpressed as soluble protein in High Five insect
cells (Trichoplusia ni BTI-TN-5B1-4 cells; Invitrogen), via a baculovirus expres-
sion system as described previously (Henderson et al., 2007a). Purified DQ8-
glia-a1 was mixed in a 1:1 ratio with purified SP3.4 TCR, and the resultant
SP3.4-DQ8-glia-a1 complex was then purified by size exclusion chromatog-
raphy (Superdex 200; GE Healthcare).
DQ8-Glia-a1 Tetramer Production and Staining
incorporated on the C terminus of the b chain by means of the BirA enzyme
(Beckett et al., 1999; O’callaghan et al., 1999). Tetramer reagents were gener-
ated by the addition of NeutrAvidin R-phycoerythrin (PE) conjugate (Invitrogen)
a1 tetramer were washed and resuspended in culture medium (Iscove’s modi-
fied Dulbecco’s medium [IMDM] [Invitrogen] supplemented with 5% [v/v]
pooled human serum [PHS], 2 mM glutamine [Glutamax, Invitrogen], 5 3
10?5M 2-mercaptoethanol [Sigma-Aldrich], and 100 mM nonessential amino
acids [Invitrogen]) in a final volume of 50 ml. PE-DQ8-glia-a1 tetramer
(50 mg/ml final concentration) was added to the cells and staining performed
in the dark at 37?C for 2 hr. Cells were washed in PBS containing 0.1% (w/v)
BSA and stained for 20 min on ice, with fluorescein isothiocyanate (FITC)-
conjugated anti-human CD4 (BD Bioscience). Cells were washed again and
resuspended in PBS with 0.1% (w/v) BSA. Propidium iodide was added to
the cells immediately prior to flow cytometric analysis. Cells were analyzed
on a FACSAria instrument (BD Bioscience), and in each experiment, optimal
compensation and gain settings were determined with unstained and single-
The SP3.4 TCR and the SP3.4-DQ8-glia-a1 complex, in 10 mM Tris (pH 8),
150 mM NaCl, were concentrated to 10 mg/ml and 5 mg/ml, respectively. In
method at 20?C. Crystals of the unliganded TCR were obtained in 3.8 M
sodium formate, 0.1 M Tris (pH 7.5), 5% PEG 4k. The ternary complex was
crystallized in 0.2 M sodium acetate, 0.1 M MES (pH 6.5), 17% PEG 8k.
Data collection, processing, structure determination, refinement, and valida-
tion were conducted with standard crystallography software (see Supple-
mental Experimental Procedures; Table 2).
Surface Plasmon Resonance Measurement and Analysis
Equilibrium affinity constants of the TCR-DQ8-glia-a1 interactions were deter-
mined by surface plasmon resonance with a Biacore 3000 instrument. Bio-
tinylated DQ8-glia-a1 was immobilized at different concentrations on two
flow-cell surfaces of a Biacore CAPture sensor chip (600–800 RU and
1,500–1,700 RU, respectively) and biotinylated DR4-CLIP (600-800 RU) was
used as a negative control. After two 1 min injections of 1 mM biotin, equilib-
rium affinities were determined at 25?C in HBS buffer (10 mM HEPES-HCl [pH
7.4], 150 mM NaCl, and 0.005% surfactant P20 supplied by the manufacturer).
Decreasing concentrations of each TCR were passed over all four flow cells at
10 ml/min for 2 min. The maximum concentration for all dilution series was
192 mM, with the exception of the LS TCR, for which the set-up was altered
because of low protein yields (injection for 1 min at a flow rate of 5 ml/min
and a maximum concentration of 93 mM). The final response curves were ob-
tained by subtracting the signal of the untreated flow cell from the signals of
each sample cell, and equilibrium response curves of each dilution series
were normalized against Rmaxto account for the differential surface loading.
The equilibrium dissociation constant, KD, was subsequently obtained by
fitting the combined data with Sigmaplot12 (Systat software).
The coordinates and structure factors for the SP3.4-DQ8-glia-a1 complex and
unliganded SP3.4 structures have been deposited to Protein Data Bank Japan
(PDBj) with entry codes 4GG6 and 4GG8, respectively.
Supplemental Information includes Supplemental Experimental Procedures,
five figures, and four tables and can be found with this article online at
This work was supported by the National Health and Medical Research
Council, the Australian Research Council, an NHMRC Independent Research
Institutes Infrastructure Support Scheme grant #361646, Victorian State
Government Operational Infrastructure Support, the Celiac Disease Consor-
tium, an Innovative Cluster approved by The Netherlands Genomics Initiative,
and the Dutch government (grant BSIK03009). We thank the staff at the MX1
beamline of the Australian Synchrotron, Victoria, Australia for assistance
with data collection. We thank J. Dromey for assistance. T.B. is supported
by Pfizer Australian Research fellowship; J.A.T.-D. is supported by a NHMRC
Australian Research Training Fellowship; S.E.B. is supported by an NHMRC
PhD scholarship; A.W.P. is supported by an NHMRC SRF; and J.R. is sup-
ported by an Australia Fellowship from the NHMRC. J.A.T.-D. and R.P.A. are
coinventors of patents pertaining to the use of gluten peptides in therapeutics,
diagnostics, and nontoxic gluten. R.P.A. and J.A.T.-D. are shareholders in
DQ8-Mediated Celiac Disease
Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc. 619
Nexpep Pty Ltd and R.P.A. is a shareholder in ImmusanT Inc. R.P.A. is Chief
developing a peptide-based therapy and diagnostic suitable for celiac
Received: August 17, 2011
Accepted: July 10, 2012
Published online: October 11, 2012
Abadie, V., Sollid, L.M., Barreiro, L.B., and Jabri, B. (2011). Integration of
genetic and immunological insights into a model of celiac disease pathogen-
esis. Annu. Rev. Immunol. 29, 493–525.
Anderson, R.P., Degano, P., Godkin, A.J., Jewell, D.P., and Hill, A.V. (2000).
In vivo antigen challenge in celiac disease identifies a single transglutami-
nase-modified peptide as the dominant A-gliadin T-cell epitope. Nat. Med.
Anderson, R.P., van Heel, D.A., Tye-Din, J.A., Barnardo, M., Salio, M., Jewell,
D.P., and Hill, A.V. (2005). T cells in peripheral blood after gluten challenge in
coeliac disease. Gut 54, 1217–1223.
Arentz-Hansen, H., Ko ¨rner, R., Molberg, O., Quarsten, H., Vader, W., Kooy,
Y.M., Lundin, K.E., Koning, F., Roepstorff, P., Sollid, L.M., and McAdam,
S.N. (2000). The intestinal T cell response to alpha-gliadin in adult celiac
disease is focused on a single deamidated glutamine targeted by tissue trans-
glutaminase. J. Exp. Med. 191, 603–612.
Arentz-Hansen, H., McAdam, S.N., Molberg, O., Fleckenstein, B., Lundin,
K.E., Jørgensen, T.J., Jung, G., Roepstorff, P., and Sollid, L.M. (2002).
Celiac lesion T cells recognize epitopes that cluster in regions of gliadins
rich in proline residues. Gastroenterology 123, 803–809.
Beckett, D., Kovaleva, E., and Schatz, P.J. (1999). A minimal peptide substrate
in biotin holoenzyme synthetase-catalyzed biotinylation. Protein Sci. 8,
Nielsen, L., Pang, S.S., Dunstone, M.A., Liu, Y.C., et al. (2009). Antigen ligation
triggers a conformational change within the constant domain of the alphabeta
T cell receptor. Immunity 30, 777–788.
Boulter, J.M., Glick, M., Todorov, P.T., Baston, E., Sami, M., Rizkallah, P., and
Jakobsen, B.K. (2003). Stable, soluble T-cell receptor molecules for crystalli-
zation and therapeutics. Protein Eng. 16, 707–711.
customized and integrated system for IG and TR standardized V-J and V-D-J
sequence analysis. Nucleic Acids Res. 36 (Web Server issue), W503-8.
Bulek, A.M., Cole, D.K., Skowera, A., Dolton, G., Gras, S., Madura, F., Fuller,
A., Miles, J.J., Gostick, E., Price, D.A., et al. (2012). Structural basis for the
killing of human beta cells by CD8(+) T cells in type 1 diabetes. Nat.
Immunol. 13, 283–289.
Burrows, S.R., Chen, Z., Archbold, J.K., Tynan, F.E., Beddoe, T., Kjer-Nielsen,
L., Miles, J.J., Khanna, R., Moss, D.J., Liu, Y.C., et al. (2010). Hard wiring of
T cell receptor specificity for the major histocompatibility complex is under-
pinned by TCR adaptability. Proc. Natl. Acad. Sci. USA 107, 10608–10613.
Chen, V.B., Arendall, W.B., 3rd, Headd, J.J., Keedy, D.A., Immormino, R.M.,
Kapral, G.J., Murray, L.W., Richardson, J.S., and Richardson, D.C. (2010).
MolProbity: all-atom structure validation for macromolecular crystallography.
Acta Crystallogr. D Biol. Crystallogr. 66, 12–21.
Clements, C.S., Dunstone, M.A., Macdonald, W.A., McCluskey, J., and
Rossjohn, J. (2006). Specificity on a knife-edge: the alphabeta T cell receptor.
Curr. Opin. Struct. Biol. 16, 787–795.
Cole, D.K., Pumphrey, N.J., Boulter, J.M., Sami, M., Bell, J.I., Gostick, E.,
Price, D.A., Gao, G.F., Sewell, A.K., and Jakobsen, B.K. (2007). Human
TCR-binding affinity is governed by MHC class restriction. J. Immunol. 178,
Day, E.B., Guillonneau, C., Gras, S., La Gruta, N.L., Vignali, D.A.A., Doherty,
P.C., Purcell, A.W., Rossjohn, J., and Turner, S.J. (2011). Structural basis for
enabling T-cell receptor diversity within biased virus-specific CD8+ T-cell
responses. Proc. Natl. Acad. Sci. USA 108, 9536–9541.
Deng, L., and Mariuzza, R.A. (2007). Recognition of self-peptide-MHC
complexes by autoimmune T-cell receptors. Trends Biochem. Sci. 32,
DePaolo, R.W., Abadie, V., Tang, F., Fehlner-Peach, H., Hall, J.A., Wang, W.,
Marietta, E.V., Kasarda, D.D., Waldmann, T.A., Murray, J.A., et al. (2011).
Co-adjuvant effects of retinoic acid and IL-15 induce inflammatory immunity
to dietary antigens. Nature 471, 220–224.
Fleckenstein, B., Molberg, O., Qiao, S.W., Schmid, D.G., von der Mu ¨lbe, F.,
Elgstøen, K., Jung, G., and Sollid, L.M. (2002). Gliadin T cell epitope selection
by tissue transglutaminase in celiac disease. Role of enzyme specificity and
pH influence on the transamidation versus deamidation process. J. Biol.
Chem. 277, 34109–34116.
Garboczi, D.N., Ghosh, P., Utz, U., Fan, Q.R., Biddison, W.E., and Wiley, D.C.
(1996). Structure of the complex between human T-cell receptor, viral peptide
and HLA-A2. Nature 384, 134–141.
Garcia, K.C., Degano, M., Pease, L.R., Huang, M., Peterson, P.A., Teyton, L.,
and Wilson, I.A. (1998). Structural basis of plasticity in T cell receptor recogni-
tion of a self peptide-MHC antigen. Science 279, 1166–1172.
Godfrey, D.I., Rossjohn, J., and McCluskey, J. (2008). The fidelity, occasional
promiscuity, and versatility of T cell receptor recognition. Immunity 28,
Gras, S., Chen, Z., Miles, J.J., Liu, Y.C., Bell, M.J., Sullivan, L.C., Kjer-Nielsen,
L., Brennan, R.M., Burrows, J.M., Neller, M.A., et al. (2010). Allelic polymor-
phism in the T cell receptor and its impact on immune responses. J. Exp.
Med. 207, 1555–1567.
Henderson, K.N., Reid, H.H., Borg, N.A., Broughton, S.E., Huyton, T.,
Anderson, R.P., McCluskey, J., and Rossjohn, J. (2007a). The production
and crystallization of the human leukocyte antigen class II molecules
HLA-DQ2 and HLA-DQ8 complexed with deamidated gliadin peptides impli-
cated in coeliac disease. Acta Crystallogr. Sect. F Struct. Biol. Cryst.
Commun. 63, 1021–1025.
Henderson, K.N., Tye-Din, J.A., Reid, H.H., Chen, Z., Borg, N.A., Beissbarth,
T., Tatham, A., Mannering, S.I., Purcell, A.W., Dudek, N.L., et al. (2007b). A
structural and immunological basis for the role of human leukocyte antigen
DQ8 in celiac disease. Immunity 27, 23–34.
Ciszewski, C., Curran, S.A., Murray, J.A., David, C.S., et al. (2008). The role of
HLA-DQ8 beta57 polymorphism in the anti-gluten T-cell response in coeliac
disease. Nature 456, 534–538.
Ishizuka, J., Stewart-Jones, G.B., van der Merwe, A., Bell, J.I., McMichael,
A.J., and Jones, E.Y. (2008). The structural dynamics and energetics of an
immunodominant T cell receptor are programmed by its Vbeta domain.
Immunity 28, 171–182.
Jabri, B., and Sollid, L.M. (2009). Tissue-mediated control of immunopa-
thology in coeliac disease. Nat. Rev. Immunol. 9, 858–870.
Karell, K., Louka, A.S., Moodie, S.J., Ascher, H., Clot, F., Greco, L., Ciclitira,
P.J., Sollid, L.M., and Partanen, J.; European Genetics Cluster on Celiac
Disease. (2003). HLA types in celiac disease patients not carrying the
DQA1*05-DQB1*02 (DQ2) heterodimer: results from the European Genetics
Cluster on Celiac Disease. Hum. Immunol. 64, 469–477.
Kim, C.Y., Quarsten, H., Bergseng, E., Khosla, C., and Sollid, L.M. (2004).
Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in
celiac disease. Proc. Natl. Acad. Sci. USA 101, 4175–4179.
Kooy-Winkelaar, Y., van Lummel, M., Moustakas, A.K., Schweizer, J., Mearin,
M.L., Mulder, C.J., Roep, B.O., Drijfhout, J.W., Papadopoulos, G.K., van
Bergen, J., and Koning, F. (2011). Gluten-specific T cells cross-react between
HLA-DQ8 and the HLA-DQ2a/DQ8b transdimer. J. Immunol. 187, 5123–5129.
Mannering, S.I., Dromey, J.A., Morris, J.S., Thearle, D.J., Jensen, K.P., and
Harrison, L.C. (2005). An efficient method for cloning human autoantigen-
specific T cells. J. Immunol. Methods 298, 83–92.
O’callaghan, C.A., Byford, M.F., Wyer, J.R., Willcox, B.E., Jakobsen, B.K.,
McMichael,A.J., and Bell, J.I. (1999). BirAenzyme: production and application
DQ8-Mediated Celiac Disease
620 Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc.
in the study of membrane receptor-ligand interactions by site-specific
biotinylation. Anal. Biochem. 266, 9–15.
Qiao, S.W., Piper, J., Haraldsen, G., Oynebra ˚ten, I., Fleckenstein, B., Molberg,
O., Khosla, C., and Sollid, L.M. (2005). Tissue transglutaminase-mediated
formation and cleavage of histamine-gliadin complexes: biological effects
and implications for celiac disease. J. Immunol. 174, 1657–1663.
Qiao, S.-W., Ra ´ki, M., Gunnarsen, K.S., Løset, G.A., Lundin, K.E.A., Sandlie, I.,
and Sollid, L.M. (2011). Posttranslational modification of gluten shapes TCR
usage in celiac disease. J. Immunol. 187, 3064–3071.
Rudolph, M.G., Stanfield, R.L., and Wilson, I.A. (2006). How TCRs bind MHCs,
peptides, and coreceptors. Annu. Rev. Immunol. 24, 419–466.
Shan, L., Molberg, O., Parrot, I., Hausch, F., Filiz, F., Gray, G.M., Sollid, L.M.,
and Khosla, C. (2002). Structural basis for gluten intolerance in celiac sprue.
Science 297, 2275–2279.
Sjo ¨stro ¨m, H., Lundin, K.E., Molberg, O., Ko ¨rner, R., McAdam, S.N.,
Anthonsen, D., Quarsten, H., Nore ´n, O., Roepstorff, P., Thorsby, E., and
Sollid, L.M. (1998). Identification of a gliadin T-cell epitope in coeliac disease:
general importance of gliadin deamidation for intestinal T-cell recognition.
Scand. J. Immunol. 48, 111–115.
Sollid, L.M. (2002). Coeliac disease: dissecting a complex inflammatory
disorder. Nat. Rev. Immunol. 2, 647–655.
Sollid, L.M., Markussen, G., Ek, J., Gjerde, H., Vartdal, F., and Thorsby, E.
(1989). Evidence for a primary association of celiac disease to a particular
HLA-DQ alpha/beta heterodimer. J. Exp. Med. 169, 345–350.
Sollid, L.M., Qiao, S.-W., Anderson, R.P., Gianfrani, C., and Koning, F. (2012).
Nomenclature and listing of celiac disease relevant gluten T-cell epitopes
restricted by HLA-DQ molecules. Immunogenetics 64, 455–460.
Spurkland, A., Sollid, L.M., Polanco, I., Vartdal, F., and Thorsby, E. (1992).
HLA-DR and -DQ genotypes of celiac disease patients serologically typed to
be non-DR3 or non-DR5/7. Hum. Immunol. 35, 188–192.
Stadinski, B.D., Trenh, P., Smith, R.L., Bautista, B., Huseby, P.G., Li, G., Stern,
L.J., and Huseby, E.S. (2011). A role for differential variable gene pairing in
creating T cell receptors specific for unique major histocompatibility ligands.
Immunity 35, 694–704.
to susceptibility and resistance to insulin-dependent diabetes mellitus. Nature
Tollefsen, S., Arentz-Hansen, H., Fleckenstein, B., Molberg, O., Ra ´ki, M.,
Kwok, W.W., Jung, G., Lundin, K.E., and Sollid, L.M. (2006). HLA-DQ2
and -DQ8 signatures of gluten T cell epitopes in celiac disease. J. Clin.
Invest. 116, 2226–2236.
Turner, S.J., and Rossjohn, J. (2011). ab T cell receptors come out swinging.
Immunity 35, 660–662.
Turner, S.J., Doherty, P.C., McCluskey, J., and Rossjohn, J. (2006). Structural
determinants of T-cell receptor bias in immunity. Nat. Rev. Immunol. 6,
Tye-Din, J., and Anderson, R. (2008). Immunopathogenesis of celiac disease.
Curr. Gastroenterol. Rep. 10, 458–465.
Tye-Din, J.A., Stewart, J.A., Dromey, J.A., Beissbarth, T., van Heel, D.A.,
Tatham, A., Henderson, K., Mannering, S.I., Gianfrani, C., Jewell, D.P., et al.
(2010). Comprehensive, quantitative mapping of T cell epitopes in gluten in
celiac disease. Sci. Transl. Med. 2, 41ra51.
Tynan, F.E., Reid, H.H., Kjer-Nielsen, L., Miles, J.J., Wilce, M.C., Kostenko, L.,
Borg, N.A., Williamson, N.A., Beddoe, T., Purcell, A.W., et al. (2007). A T cell
receptor flattens a bulged antigenic peptide presented by a major histocom-
patibility complex class I molecule. Nat. Immunol. 8, 268–276.
Vader, L.W., de Ru, A., van der Wal, Y., Kooy, Y.M., Benckhuijsen, W., Mearin,
M.L., Drijfhout, J.W., van Veelen, P., and Koning, F. (2002a). Specificity of
tissue transglutaminase explains cereal toxicity in celiac disease. J. Exp.
Med. 195, 643–649.
Vader, W., Kooy, Y., Van Veelen, P., De Ru, A., Harris, D., Benckhuijsen, W.,
Pen ˜a, S., Mearin, L., Drijfhout, J.W., and Koning, F. (2002b). The gluten
response in children with celiac disease is directed toward multiple gliadin
and glutenin peptides. Gastroenterology 122, 1729–1737.
Vader, L.W., Stepniak, D.T., Bunnik, E.M., Kooy, Y.M., de Haan, W., Drijfhout,
J.W., Van Veelen, P.A., and Koning, F. (2003). Characterization of cereal
toxicity for celiac disease patients based on protein homology in grains.
Gastroenterology 125, 1105–1113.
van de Wal, Y., Kooy, Y., van Veelen, P., Pen ˜a, S., Mearin, L., Papadopoulos,
G., and Koning, F. (1998a). Selective deamidation by tissue transglutaminase
strongly enhances gliadin-specific T cell reactivity. J. Immunol. 161, 1585–
van de Wal, Y., Kooy, Y.M., van Veelen, P.A., Pen ˜a, S.A., Mearin, L.M.,
Molberg, O., Lundin, K.E., Sollid, L.M., Mutis, T., Benckhuijsen, W.E., et al.
(1998b). Small intestinal T cells of celiac disease patients recognize a natural
pepsin fragment of gliadin. Proc. Natl. Acad. Sci. USA 95, 10050–10054.
van de Wal, Y., Kooy, Y.M., van Veelen, P., Vader, W., August, S.A., Drijfhout,
J.W., Pen ˜a, S.A., and Koning, F. (1999). Glutenin is involved in the gluten-
driven mucosal T cell response. Eur. J. Immunol. 29, 3133–3139.
Yin, Y., Li, Y., Kerzic, M.C., Martin, R., and Mariuzza, R.A. (2011). Structure of
a TCR with high affinity for self-antigen reveals basis for escape from negative
selection. EMBO J. 30, 1137–1148.
Yoshida, K., Corper, A.L., Herro, R., Jabri, B., Wilson, I.A., and Teyton, L.
(2010). The diabetogenic mouse MHC class II molecule I-Ag7 is endowed
with a switch that modulates TCR affinity. J. Clin. Invest. 120, 1578–1590.
DQ8-Mediated Celiac Disease
Immunity 37, 611–621, October 19, 2012 ª2012 Elsevier Inc. 621