Distinct orientation of the alloreactive monoclonal CD8 T cell activation program by three different peptide/MHC complexes.

Abstract

We have characterized three different programs of activation for alloreactive CD8 T cells expressing the BM3.3 TCR, their elicitation depending on the characteristics of the stimulating peptide/MHC complex. The high-affinity interaction between the TCR and the K(b)-associated endogenous peptide pBM1 (INFDFNTI) induced a complete differentiation program into effector cells correlated with sustained ERK activation. The K(bm8) variant elicited a partial activation program with delayed T cell proliferation, poor CTL activity and undetectable ERK phosphorylation; this resulted from a low-avidity interaction of TCR BM3.3 with a newly identified endogenous peptide, pBM8 (SQYYYNSL). Interestingly, mismatched pBM1/K(bm8) complexes induced a split response in BM3.3 T cells, with total reconstitution of T cell proliferation but defective generation of CTL activity that was correlated with strong but shortened ERK phosphorylation. Crystal structures highlight the molecular basis for the higher stability of pBM8/K(bm8) compared to pBM1/K(bm8) complexes that exist in two conformers. This study illustrates the importance of the stability of both peptide/MHC and peptide/MHC-TCR interactions for induction of sustained signaling required to induce optimal CTL effector functions. Subtle allelic structural variations, amplified by peptide selection, may thus orient distinct outcomes of alloreactive TCR-based therapies.

Full text

Page 1

Distinct orientation of the alloreactive monoclonal
CD8 T cell activation program by three different peptide/
MHC complexes
Nathalie Auphan-Anezin1, Catherine Mazza1, Annick Guimezanes1,
Gregory A. Barrett-Wilt2, Felix Montero-Julian3, Alain Roussel4,
Donald F. Hunt2, Bernard Malissen1and Anne-Marie Schmitt-Verhulst1
1Centre d'Immunologie de Marseille-Luminy, CNRS-INSERM-Universite de la
M?diterran?e, Campus de Luminy, Marseille, France
2Department of Chemistry, University of Virginia, Charlottesville, USA
3Beckman Coulter Immunotech, Marseille, France
4AFM UMR6098 CNRS, Marseille, France
Wehavecharacterized threedifferentprograms of activation for alloreactive CD8 Tcells
expressing the BM3.3 TCR, their elicitation depending on the characteristics of the
stimulating peptide/MHC complex. The high-affinity interaction between the TCR and
the Kb-associated endogenous peptide pBM1 (INFDFNTI) induced a complete
differentiation program into effector cells correlated with sustained ERK activation.
The Kbm8variant elicited a partial activation program with delayed Tcell proliferation,
poor CTL activity and undetectable ERK phosphorylation; this resulted from a
low-avidity interaction of TCR BM3.3 with a newly identified endogenous peptide,
pBM8 (SQYYYNSL). Interestingly, mismatched pBM1/Kbm8complexes induced a split
response in BM3.3 Tcells, with total reconstitution of Tcell proliferation but defective
generation of CTL activity that was correlated with strong but shortened ERK
phosphorylation. Crystal structures highlight the molecular basis for the higher stability
of pBM8/Kbm8compared to pBM1/Kbm8complexes that exist in two conformers. This
study illustrates the importance of the stability of both peptide/MHC and peptide/
MHC-TCR interactions for induction of sustained signaling required to induce optimal
CTL effector functions. Subtle allelic structural variations, amplified by peptide
selection, may thus orient distinct outcomes of alloreactive TCR-based therapies.
Supporting information for this article is available at
http://www.wiley-vch.de/contents/jc_2040/2006/35895_s.pdf
Introduction
Alloreactive T cells, which are generally specific for
peptide/MHC (pMHC) complexes, as are self MHC--
restricted Tcells, have been evaluated for their potential
to provide high-affinity TCR reactive to antigens
associated with allogeneic MHC products [1, 2]. Thus,
monoclonal T cells can provide reagents endowed with
Correspondence: Nathalie Auphan-Anezin, Centre d'Immu-
nologie de Marseille-Luminy, Campus de Luminy, Case 906,
13288 Marseille, Cedex 09, France
Fax: +33-4-91-26-94-30
e-mail: auphan@ciml.univ-mrs.fr
Abbreviations: GzmB: granzyme B ? pMHC: peptide/MHC
(complexes) ? RFI: relative fluorescence intensity
Received 20/1/06
Revised 16/3/06
Accepted2/5/06
[DOI 10.1002/eji.200635895]
Key words:
CD8 effector T cell
? Crystal structure
? ERK ? Partial agonist
? TCR
Nathalie Auphan-Anezin et al.
Eur. J. Immunol. 2006. 36: 1856–1866
1856
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 2

effector function [3] and defined characteristics of
specificity and avidity [4, 5].
One of the hallmarks of recognition of pMHC by
abTCR, however, is that minor changes in the sequence
of peptide or MHC can profoundly alter the T cell
response, resulting in unpredictable effector functions.
Indeed, MHC alleles presenting discrete differences in
amino acids may influence T cell reactivity at two main
levels. First, identical peptides, such as those corre-
sponding to immunodominant viral epitopes, may
present subtle changes in conformation that alter the
capacity of the pMHC complex to interact with TCR
depending on the MHC allele with which they are
associated [6]. Second, MHC allelic variations may be
responsible for selecting different peptides for presenta-
tion during viral infection [7]. These rules concerning
theinteraction ofpeptideswithMHCallelic variantsalso
apply to the presentation of endogenous peptides. In
particular, the natural variant Kbm8, which arose bygene
conversion from the Kbgene, presents both Kb-shared
and uniquesets ofendogenously processed peptides [8].
In mice preconditioned by sublethal irradiation,
infusion of CD8 T cells expressing the Kb-specific
alloreactive TCR BM3.3 induced lethal graft-versus-host
disease (GVHD) in Kb-expressing hosts, whereas it
accelerated hematopoietic reconstitution in Kbm8hosts
as compared to syngeneic Kk-expressing hosts [9].
Differentialactivationevents observedinthetransferred
CD8 Tcells in vivo could be mimicked in vitro by BM3.3
Tcell stimulation with Kb- or Kbm8-expressing APC [10].
To determine whether such differential activation
programs are the result of quantitative differences in
the presentation of the same endogenous peptide
(pBM1) in the context of Kbm8or result from the
recognition of a fully distinct agonist, we analyzed Kbm8-
eluted peptides for reconstitution of BM3.3 TCR
reactivity on Tap-negative cells. We identified a new
peptide, designated pBM8 (SQYYYNSL), originating
from an endogenous RNA-binding protein that recon-
stituted TCRBM3.3 reactivityinthecontextof Kbm8.The
molecular bases for this allele-dependent peptide
selection were revealed by comparison of the crystal
structures of pBM8/Kbm8and pBM1/Kbcomplexes. This
study, therefore, identifies a novel endogenous pBM8/
Kbm8ligand for TCR BM3.3 that behaves as a natural
partial agonist, characterized by a lower avidity for and
increased rate of dissociation from the TCR BM3.3 as
compared tothe pBM1/Kbagonist ligand. These binding
characteristics are associated with the induction in naive
TCR BM3.3-expressing CD8 T cells of delayed ERK
phosphorylation as well as delayed proliferation and
partial effector program. Interestingly, exogenous addi-
tion of pBM1 to Kbm8APC reconstituted BM3.3 T cell
proliferation and strong ERK1/2 phosphorylation but
failed to induce sustained ERK1/2 activation and high
CTL activity, which we define as a split response.
Intriguingly, in crystals of the mismatched pBM1/Kbm8
complex, the pBM1 peptide afforded two distinct
structures. In one of them, pBM1 was aligned in a
manner similar to its position in Kb. In the other, it was
aligned in the position taken by pBM8 in Kbm8,
generating an epitope akin to that of the partial agonist
pBM8/Kbm8. Therefore, this study illustrates the con-
sequences of structural constraints imposed by allelic
pMHC complexes on the elicitation of functional
programs in monoclonal CD8 T cells.
Results and Discussion
Impact of polymorphic Kband Kbm8residues on
peptide conformation and stability of pMHC
complexes
Here we demonstrate that a peptide, pBM1 (INFDFNTI),
that is abundantly present in Kbeluates [11], could not
be detected in Kbm8eluates (Table 1). The screening was
performedwithBM3.3CTLeffectorsthatveryefficiently
recognize pBM1, even when added exogenously on Kb-
negative Kbm8target cells. Instead, a distinct peptide,
pBM8 (SQYYYNSL), could be identified in the Kbm8
eluate (Table 1) in spite of its 10-fold higher SC50value
(concentration of peptide inducing 50% of maximum
cytolysis) for lysis by CTL BM3.3 on Kbm8targets
(Table 2).
Except for the peptides arising from the Kb/Kbm8
molecules themselves, the peptide repertoire repre-
sented in the endoplasmic reticulum (ER) is expected to
Table 1. Identification of pBM8, a Kbm8-eluted peptide recognized by BM3.3 TCR CD8 effector cells
Fractions 40–41Fractions 73–75 Identified peptide
Kbm8eluatea)
17%2%pBM8: SQYYYNSL
Kbeluate 0% 35%pBM1: INFDFNTIb)
a)Kbm8-associated peptides were eluted from bm8 LPS blasts and fractionated by reverse-phase HPLC. Each HPLC fraction was
tested for epitope reconstitution on Kb-negative RMA.S-Kbm8targets in a CTL assay using BM3.3 TCR CD8 effector cells at an 8:1
ratio.
b)as reported in [11]
Eur. J. Immunol. 2006. 36: 1856–1866
Molecular immunology
1857
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 3

be the same in cells of B6 or bm8 origin. The fact that
pBM1 is not present (within the limits of our very
sensitive assays) at the surface of bm8 cells thus
indicates that it either failed to stabilize Kbm8molecules
in the ER or that pBM1/Kbm8complexes formed in the
ER dissociated before reaching the cell surface. Indeed,
both the affinity of the peptide-MHC association (10-
-fold higher for pBM8 ascomparedtopBM1withrespect
to Kbm8; Table 3) and the rate of pMHC dissociation
(Table 3) are in favor of pBM8/Kbm8as compared to
pBM1/Kbm8complexes. The affinity parameter may be
most relevant for peptides to successfully associate with
MHC molecules in the face of high concentrations of
competing peptides within the confinement of the ER.
The kinetics of peptide dissociation from the MHC may,
in turn, be critical during transfer of the pMHC from the
ERtothecellsurfacebecauseof thelowconcentrationof
free peptide in the Golgi apparatus [12, 13].
Three of the four amino acids that are polymorphic
between KbandKbm8areclusteredon thebsheetfloorof
the peptide-binding groove at positions 22, 23 and 24,
whereas the fourth polymorphic residue is located at
position 30 in a solvent-exposed loop remote from the
peptide-binding groove and from known sites of TCR
docking. In both Kband Kbm8, residue 23 points toward
the b2-microglobulin domain, whereas residues 22 and
24 project into the peptide-binding groove, contributing
to the architecture of the B pocket and establishing
contacts with the second amino acid (p2) of the bound
peptide. To determine how the structural differences
between Kband Kbm8impact the selection of antigenic
peptides and the stability of the resulting pMHC
complexes, we first solved the structure of pBM8/
Kbm8, the naturally occurring pMHC pair, to a resolution
of 1.9 ? (supplementary Table 1) and compared it tothe
structure of pBM1/Kb[14]. This analysis clearly points
to the high impact of the B pocket environment on these
properties of the pMHC complexes. Indeed, in the case
of the pBM1/Kbcomplex, a dense hydrogen-bonding
networkcenteredon Glu24 contributestothestabilityof
the complex, whereas in the pBM8/Kbm8complexes, the
potentially destabilizing role of the Glu24Ser substitu-
tion is compensated by the presence of a glutamine
residue at position 2 of pBM8 (Fig. 1A; see also
supplementary Table 2). Therefore, our results are also
consistent with the fact that position 24 of Kband Kbm8
determines the distinct repertoires of peptides accom-
modated by these two alleles and thereby plays an
important role in the alloreactivity that exists between
B6 and bm8 mice [15].
The pBM1/Kband pBM8/Kbm8natural complexes
went through various intracellular steps of proofreading
that led to their surface expression [12]. Accordingly,
each of the two peptides presents an optimal fit within
the B and C pockets of its corresponding MHC allele
(Fig. 1). As a result, the two pMHC complexes are
endowed with comparable stability (Table 3).
Distinct peptides selected by MHC allelic variants
elicit strong or partial CD8 T cell responses
The respective properties of strong and partial agonistof
pBM1/Kb
and pBM8/Kbm8
CD8 T cells [10] therefore fully rely on the nature and
the position of the peptide residues available for
interaction with the TCR. The structures of TCR
BM3.3 in complex with pBM1/Kband VSV8/Kb, a
previously characterized cross-reactive pMHC, showed
that residues p4, p6 and p7 of the antigenic peptides are
contacted by the TCR [14, 16]. In both TCR/pMHC
complexes, p6 constitutes a primary TCR contact and is
conserved between pBM1, VSV8 and pBM8, whereas p4
and p7 constitute secondary TCR contacts. A compar-
ison of the BM3.3/pBM1/Kb[14] and BM3.3/VSV8/Kb
[16] complexes shows that in the former, the position of
the amide group found at p6 is shifted by 1.9 ? toward
the TCR. As a result, the pBM1 Asnp6interacts more
efficiently with TCR BM3.3 and likely contributes to the
strong agonist property observed for this peptide as
complexesfor BM3.3
Table 2. Peptide concentrations required for half-maximum
cytolysis by BM3.3 effectors
MHCpeptideSC50(CTL)(nM)a)
Kb
pBM10.02?0.01
0.15?0.05
4.5?0.5
42.5?7.5
Kb
pBM8
Kbm8
pBM1
Kbm8
pBM8
a)SC50stands for the concentration of peptide inducing 50% of
maximum cytolysis
Table 3. Binding characteristics of pBM1 and pBM8 to Kband
Kbm8
MHCPeptideEC50(nM)a)
T1/2 (min)
Kb
none60
Kb
pBM125?5
>1000
>240b,c)
Kb
pBM875?5b)
52 Kbm8
none
Kbm8
pBM125?10
2.5?0.5
130?15b)
>240b,c)
Kbm8
pBM8
a)peptide concentration inducing 50% of maximum Kbor Kbm8
expression on TAP–/–cells
b)measured at a peptide concentration of 10–6M
c)at a peptide concentration of 10–8M, the half-life could be
measured only for the more stable pBM1/Kband pBM8/ Kbm8
complexes and was 90 min and 240 min, respectively
Nathalie Auphan-Anezin et al.
Eur. J. Immunol. 2006. 36: 1856–1866
1858
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 4

Figure 1. Views of the peptide-binding groove of Kband Kbm8bound to the pBM1 or pBM8 peptides. Comparison of the
hydrogen-bonding network found in the B pocket of the pBM8/Kbm8and pBM1/Kbcomplexes (A) orof thepBM1/Kbm8and pBM1/Kb
complexes (B). The Kband Kbm8MHC residues are shown in green and purple, respectively. Residues at positions 1, 2 and 3 of each
peptide are represented in pink for pBM8 bound to Kbm8, inyellow for pBM1 bound to Kband in cyan for pBM1 bound to Kbm8. MHC
main chains are in light grey, and oxygen and nitrogen atoms arein red and blue, respectively. Hydrogen bonds are represented as
dashed lines that are color coded as follows: black for pBM1/Kb, light pink for pBM8/Kbm8and light blue for pBM1/Kbm8. (C)
Superposition of the pBM1 peptide bound to Kband of pBM8 peptide bound to Kbm8. Peptides, MHC residues and hydrogen bonds
are depicted using the color code specified in (A) and (B). The MHC helices a1 and a2 are depicted in grey. Serine 99 on the b sheet
floor of the peptide-binding groove stabilizes each of the two peptides in a different manner. This results in a 1.45 ? shift between
the Ca atoms of residue p6 and in a different exposure of the side chain of this residue. (D, E) Two conformations are observed for
peptide residues 5 to 7 in the two pBM1/Kbm8complexes found per asymmetric unit. They are represented in cyan and are
superimposed in (D) and (E) to pBM1/Kb(yellow) and pBM8/Kbm8(pink). The pBM1/Kbm8Asnp6peptide residue adopts either a
“pBM1/Kb-likereferredtoaspBM1/Kbm8-OUT”conformation(D)ora“pBM8/Kbm8-likereferredtoaspBM1/Kbm8-IN”conformation(E).
The a1 and a2 MHC helices are depicted in grey. (F, G) 2Fo-Fc experimental electron density for the two pBM1 peptide
conformationscontouredat 1r. Thetwo peptides oftheasymmetricunit areplacedinanunbiasedextradensitycalculatedaftera
rigidbodyrefinementusingthepBM8/Kbm8modeldeprivedofantigenicpeptideandofwatermolecules.Thedifferentpanelswere
generated with Molscript [42] and Render [43] and PyMOL.
Eur. J. Immunol. 2006. 36: 1856–1866
Molecular immunology
1859
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 5

compared to VSV8, which poorly activates the BM3.3
Tcells [17]. One explanation for this different position-
ing of p6 is the presence of a hydrogen bond between
Ser99 of Kband Tyrp5of VSV8, which pulls the VSV8
backbone down into the bottom of the peptide-binding
groove in the VSV8/Kbcomplex [16]. Interestingly, a
similar interaction occurs between Tyrp5of pBM8 and
Kbm8, resulting in a 1.5 ? shift of the p6 Ca and a 2.5 ?
shift of the amide group toward the peptide-binding
groove as compared to that of the Asnp6found in pBM1/
Kb(Fig. 1C). As a consequence, the conformation
adopted by the Asnp6side chain of pBM8 is similar tothe
one observed for the Glnp6in BM3.3/VSV8/Kband
results in a 2-fold reduction of its solvent accessibility
when compared to that of pBM1 (80 ?2in pBM1/Kb
versus 40 ?2in pBM8/Kbm8). This poorer solvent
accessibility likely loosens the contacts of pBM8 with
TCRBM3.3andcontributestotheweakagonistproperty
of pBM8/Kbm8. It will be intriguing to understand how
TCR BM3.3 can adapt the positioning of its CDR3 to
interact with pBM8/Kbm8and compensate for the loss of
the numerous interactions involving residue p6.
pBM8/Kbm8complexes behave as partial agonists
for BM3.3 CD8 T cells
When assessing the functional consequences of this
qualitatively weak pBM8/Kbm8-TCR BM3.3 interaction,
we observed that increasing the concentration of pBM8
led to a slightly increased number of BM3.3 CD8 Tcells
recruited into proliferation, in line with an increase in
the probability of efficient interaction between the
TcellsandthespecificpMHCcomplexes(Fig.2,CFSEon
day 3). However, IL-2 production remained defective,
and CD25 expression was only slightly increased by
addition of pBM8. Therefore, even in the presence of a
high dose (10–6M) pBM8 peptide, Kbm8drove a partial
activation program in BM3.3 CD8 T cells. Importantly,
CTL activity of BM3.3 CD8 effector cells against
pBM8-loaded Tap–/–Kbm8targets was higher when
naive CD8 Tcells had been stimulated with bm8 APC in
the presence of exogenous IL-2 than when raised against
B6 APC (supplementary Fig. 1 and [18]). This was
correlated with a low level of TCR expression as a result
of receptor down-modulation that occurred in response
Figure 2. Effect of high doses of pBM8 and pBM1 peptides on primary responses of BM3.3 CD8 T cells induced by bm8 APC. (A)
CFSE-labeled BM3.3 CD8 cells were cultured for 2 or 3 days with either syngeneic CBA APC or B6 or bm8 APC in the absence or
presence of pBM8 and pBM1 peptides (10–6M). CFSE profiles of BM3.3 T cells are shown with the corresponding percentage of
non-divided cells. CD25 surface expression (day 2) as well as IL-2 (day 2) and GzmB (day 3) intracellular stainings are shown on
gated BM3.3 CD8 T cells. The percentage of CD8 cells positive for the corresponding marker as well as MFI (for CD25, in
parentheses)arereported.(B)CTLactivityofBM3.3CD8effectorcellsobtainedafter3 daysofculturewithsyngeneicCBAAPCorB6
or bm8 APC in the absence or presence of pBM8 or pBM1 peptides (10–6M) was tested on target cells expressing Kb(RMA) or not
(RDM4). Non-specific lysis on RDM4 targets never exceeded 2% (data not shown). Standard deviations were calculated for three
independent experiments.
Nathalie Auphan-Anezin et al.
Eur. J. Immunol. 2006. 36: 1856–1866
1860
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 6

to the agonist B6 but not in response to bm8 (data not
shown and [19]). Since lowered TCR expression might
increase reliance on CD8, we analyzed the CD8
dependency of BM3.3 CTL effectors sensitized by bm8
APC in the presence of IL-2. Results (supplementary
Fig. 2) showed that recognition of pBM8/Kbm8remained
fully inhibited by anti-CD8 mAb.
These data support the notion that the pBM8/Kbm8
complex behaves as a partial agonist, the recognition of
which is highly susceptible to changes in TCR BM3.3
expression and is highly dependent on the CD8
co-receptor, which may be involved in stabilization of
this weak interaction (Fig. 3).
Defective Ca2+signaling upon binding of pBM8/
Kbm8multimers to BM3.3 CD8 T cells
The pBM1/Kb-TCR interaction displays a relatively high
affinity (3 lM, [14]) and slow dissociation rate as
compared to the 3-fold higher KD (as measured by
multimer binding, Fig. 4B) and rapid dissociation of the
pBM8/Kbm8-TCR complexes (Fig. 4A). These kinetic
parameters are in line with the notion that T cell
activation generally correlates with the half-life of the
TCR-pMHC interaction, as originally postulated by the
kinetic proofreading model [20–22]. However, in the
faceofalimitingnumberofpMHCcomplexespresenton
the APC under physiological conditions, an excessive
half-life would hamper TCR serial engagement by
limiting available pMHC complexes [23].
Binding of fluorescently labeled pMHC multimers
can be combined with the use of the fluorescent
indicator of intracellular Ca2+, Indo-1, allowing a direct
evaluation of the consequences of TCR engagement on
signaling. Data in Fig. 4C show that even for a very low
TCR BM3.3 engagement by pBM1/Kbmultimers [peak
multimer binding relative fluorescence intensity (RFI)
=11 for 0.2 nM pBM1/Kb; Ca2+signal =64, see legend
to Fig. 4C] on naive BM3.3 CD8 T cells, a significant
calcium flux was induced, whereas no calcium signal
could be detected upon binding of pBM8/Kbm8multi-
mers, even for high TCR engagement (peak multimer
binding RFI =474 for 100 nM pBM8/Kbm8; Ca2+signal
=0). These binding studies characterize the pBM8/Kbm8
complex as a ligand of low avidity and high off-rate for
TCR BM3.3 (Fig. 4).
Figure 3.SensitivityoftheCTLactivityofBM3.3CD8effectorcellsto anti-CD8 mAb.51Cr-labeledRMA.Sand RMA.S-Kbm8cellswere
loadedwith different concentrations of the indicated peptide (pBM1, pBM8 or OVA) and incubatedwith BM3.3 CD8 effector cells at
an 8:1 ratio in the absence (open symbols) or presence (closed symbols) of anti-CD8 mAb. The percentage of specific lysis is
reported as a function of peptide concentration.
Eur. J. Immunol. 2006. 36: 1856–1866
Molecular immunology
1861
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 7

A split allo-response characterized by strong
proliferation and defective CTL activity is induced
by pBM1/Kbm8
The subtle manner in which the peptide-MHC interac-
tion can impact T cell activation programs is clearly
exemplified by the third pMHC ligand analyzed here,
resulting from the mismatchedassociation ofpBM1with
Kbm8. Indeed, addition of pBM1 to bm8 APC generated,
in a concentration-dependent fashion, a split respon-
sivenessinBM3.3CD8Tcells(Fig.2).Thisinvolvedtotal
reconstitution of proliferation and CD69 expression,
partial recovery of IL-2 secretion and CD25 up--
regulation, and defective CTL function, the latter being
correlated with deficient granzyme B (GzmB) expres-
sion (Fig. 2 and supplementary Fig. 3).
Of note, the opposite situation, in which pBM8 was
added to B6 APC, did not affect the BM3.3 CD8 T cell
activation profiles, presumably because sufficient en-
dogenous pBM1 peptide provided optimal stimulation
even in the presence of high concentrations of pBM8
(results not shown). This result suggests that pBM8/Kb
does not act as an antagonist for the BM3.3 TCR. A
similar conclusion was obtained for pBM8/Kbm8, since
APC from (B6 ? bm8) F1 mice were as efficient as B6
APC at inducing BM3.3 CD8 Tcell activation (results not
shown). Because of the dominant stimulatory capacity
of the endogenous pBM1/Kbcomplexes, we did not
further study the activation potential of pBM8/Kb
complexes.
To understand the structural basis of the weaker
stability displayed by the mismatched pBM1/Kbm8
complex (Table 3), we also solved its structure to
1.9 ? and compared it to that of the naturally occurring
pBM1/Kbcomplex. Supplementary Table 2 summarizes
the hydrogen-bonding network of the B pocket in the
pBM1/Kb, pBM1/Kbm8and pBM8/Kbm8complexes, and
Fig. 1B shows a superposition of the B pocket of the
pBM1/Kbm8and pBM1/Kbcomplexes. Although Asn70
of both MHC alleles and residues 1–3 of pBM1 lie in the
same conformation, a reorganization occurs inside the B
pocket, allowing pBM1 to accommodate the substitu-
tions found at positions 22 and 24 of Kbm8. As already
noted in the case of the pBM8/Kbm8complex, Tyr45
rotates by 60?and its hydroxyl group moves by 2 ? to
optimize its interaction with the shorter Ser24 residue.
Moreover, in pBM1/Kbm8a water molecule helps to fill
the void resulting from the Glu24Ser mutation and
maintain a hydrogen-bonding network between the
bottom of the peptide-binding groove and the a1 helix
(Fig. 1B). Therefore, the distinctive water-mediated
hydrogen-bonding network observed in the heterolo-
gous pBM1/Kbm8complex likely explains its reduced
affinity and shorter surface half-life as compared to the
homologous pBM1/Kb
and pBM8/Kbm8
(Table 3). A similar use of a compensatory water
molecule has already been observed in the structure of
an immunodominant HSV glycoprotein B peptide
complexed to Kbm8[6].
Importantly, our structural study of the pBM1/Kbm8
complex demonstrates how the modified chemical
environment of the B pocket, by influencing the
neighboring C pocket, can permit the repositioning of
the p5-p6 section of the bound peptide and results in the
occurrenceof twopeptideconformers.Thisdifferencein
conformation cannot be explained by any direct crystal
contact with the pBM1 peptide and is mainly limited to
residues p5 and p6 of the peptide C-terminus. The
largestshiftinconformationislocatedaround p6,where
the Ca deviates by 1.25 ? and the amide group by
2.78 ?. As a result, the two peptide conformers found in
the pBM1/Kbm8crystal adopt either a “pBM1/Kb-like”
(referred below as pBM1/Kbm8-OUT; Fig. 1D, F) or a
“pBM8/Kbm8-like” (referred below as pBM1/ Kbm8-IN;
Fig. 1E, G) conformation in which the amide group
found at p6 is positioned in a conformation similar to
that observed in the BM3.3/pBM1/Kband BM3.3/
VSV8/Kbcomplexes, respectively. It remains, however,
to be determined whether only one of the two peptide
conformations found in the pBM1/Kbm8crystal is
capable of productively interacting with the BM3.3
TCR. It is important to stress that the “pBM1/Kbm8-IN”
complexes
Figure 4. Binding of pBM8/Kbm8multimers fails to trigger
calcium signaling in naive BM3.3 CD8 T cells. (A, B) Binding of
pBM1/Kband pBM8/Kbm8multimers at different concentra-
tions on BM3.3 CD8 cellswas measured after 90 min at 4?C (see
Materials and methods). (A) The KD and half-life of pMHC
complexes are reported. (C) pBM1/Kbor pBM8/Kbm8multimers
were added to Indo-1-loaded BM3.3 CD8 cells. Staining with
multimers and calcium signaling were recorded simulta-
neously over a period of 15 min. Peak values for multimer
binding(RFI) arereportedin abscissaand for increase in Ca2+in
ordinate (for pBM1/Kbmultimers at 0.2 nM, 2.0 nM and 20 nM
andforpBM8/Kbm8multimersat100 nM,400 nMand1000 nM).
Nathalie Auphan-Anezin et al.
Eur. J. Immunol. 2006. 36: 1856–1866
1862
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 8

conformation found in the pBM1/Kbm8crystal corre-
sponds to a more stable conformation, since it displays
better thermal stability (the overall B factors for the
peptide are 30 ?2and 36 ?2in the “pBM1/Kbm8-IN” and
“pBM1/Kbm8-OUT” conformations, respectively). More-
over, the density map around residues p5 and p6 is of
much better quality in the case of the “pBM1/Kbm8-IN”
conformation (Fig. 1G). The presence of these two
conformers in the crystal may reflect the existence of
two pBM1/Kbm8conformations at the surface of the
APC, which could lower the probability of productive
engagement of TCR BM3.3 with the more antigenic
“pBM1/Kbm8-OUT” conformer in the face of the higher
stability of the “pBM1/Kbm8-IN” conformer. Thus,
increasing the peptide concentration may enhance the
probability of encounter of the “pBM1/Kbm8-OUT”
conformer, recognition of which by CTL BM3.3 would
be largely independent of CD8 as for pBM1/Kb.
Accordingly, at the CTL effector phase, the impact of
pBM1/Kbm8as compared to pBM1/Kbpresentation is a
200- to 300-fold displacement of the curve for CTL
activity versus peptide concentration in the absence or
presence of anti-CD8 mAb (Fig. 3). Similarly, pMHC
stabilization by substitution of the C-terminal residue of
a tumor-associated nonapeptide has recently been
described to reposition the peptide main chain and to
enhance its interaction with a TCR through its p4p5
residues, generating a ligand of higher affinity that is
less dependent on CD8 for Tcell binding [24]. The split
response may thus result from the low stability of the
pBM1/Kbm8complex, in particular in its more antigenic
“OUT” form. However, conformational changes may
further occur at the TCR-pMHC binding interface so that
the final peptide conformation in the TCR complex
might be distinct from both conformers [25, 26].
Partial and split responses in naive CD8 T cells
associated with ERK1/2 phosphorylation
ERK1/2 phosphorylation could not be detected in
responsetobm8APC(Fig.5).Nevertheless,intracellular
signals may accumulateto finally induce aweak ERK1/2
modification, in agreement with the notion that naive
T cell differentiation is triggered by a cumulative signal
which is reached at different time points for different
TCR ligands [27].
The mismatched assembled pBM1/Kbm8complex
revealed an important additional kinetic threshold that
impacts the differentiation program of naive CD8 Tcells.
Indeed, examination of the functions induced by a
strong short-lived (namely split) agonist such as pBM1/
Kbm8showed strong TCR-induced ERK1/2 phosphoryla-
tion but with a less sustained pattern than in response to
the full agonist (Fig. 5). This shortened ERK phosphor-
ylation was not the result of exogenous versus en-
dogenous peptide presentation, since exogenous pBM1
presented on Tap–/–B6 APC induced sustained ERK
modification (supplementary Fig. 4). Altogether, these
results point out that an increase in the initial signals
may be sufficient to drive entry into cell cycle. However,
asustainedactivationsignalisrequiredforacquisitionof
cytotoxic activity as well as for induction of IL-2
secretion (Fig. 2). These results further demonstrate
that CTL effector phase responses (Fig. 3; maximal lysis
for10–8MpBM1onKbm8targets)arefarlessdemanding
than the CTL differentiation process (Fig. 2; poor CTL
activity and GzmB induced in the presence of 10–6M
pBM1 on Kbm8APC), which appeared to be regulated by
multiple parameters. This is consistent with previous
data [28, 29] and with the notion that ligand
discrimination can be modulated by differentiation-
related changes in the stoichiometry of components of
the signaling network downstream of the TCR [30].
Indeed, CD8 T cell division appeared necessary for the
induction of CTL activity [10], but in the absence of
secondary signaling mediated by the high-affinity IL-2R,
CTL activity remained suboptimal [18]. The IL-2-
dependent complementation for induction of full CTL
Figure 5. Effect of exogenously added pBM8 and pBM1 peptides
on phosphorylation of ERK1/2 kinases upon activation of naive
BM3.3 CD8 T cells with bm8 APC. (A) BM3.3 CD8 T cells were
cultured with either CBA APC (grey histograms) or B6 or bm8
APC in the absence or presence of pBM8 and pBM1 peptides
(10–6M). As a positive control, we also included activation with
an anti-CD3 mAb. After 1 h, 3 h, 27 h and 42 h, intracellular
staining for phospho-ERK1/2 was performed, and histograms
(bold line) gated on CD8 T cells are shown. Numbers indicate
the percentage of phospho-ERK1/2 positive CD8 T cells.
Eur. J. Immunol. 2006. 36: 1856–1866
Molecular immunology
1863
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 9

lytic activity was mimicked by introduction of a
constitutive active STAT5 molecule, both instances
leading to stabilization of GzmB expression [18].
Production of IL-2 thus appears to remain a limiting
factor for the full development of CTL activity by pBM1/
Kbm8complexes (Fig. 2).
Previous observations using CD4 Tcell clones [31] or
T cell hybridomas [32] led to the general notion that
partial agonists induce transient ERK activation. En-
gagement of CD3 alone versus co-engagement of CD3
and CD4 was used as a model for partial versus full
agonist signaling in some of these studies [31], in line
with our observation that anti-CD3 mAb induced strong
but less sustained ERK phosphorylation than the full
agonist APC (Fig. 5). However, these models did not
allow discrimination of requirements for T cell prolif-
eration versus differentiation. Studies of naive CD8 OT-1
T cells responding to SIINFEKL or a variant partial
agonist peptide suggested that partial agonists fail to
induce significant early signaling but that the accumula-
tion of weak signals can be integrated to induce a
delayed response [27], a situation analogous to the
responsetopBM8/Kbm8.Thus,ourstudy furthershowsa
clear discrimination for CD8 Tcell responses between at
least twotypes of partial agonists associated with aweak
and delayed or a strong but transient ERK activation
(Fig. 5).
Concluding remarks
Altogether, our data are consistent with the notion that
transient versus sustained ERK1/2 phosphorylation may
control cell fate decisions towards proliferation or
differentiation, respectively. Similar conclusions have
emerged from studies on epidermal growth factor and
nerve growth factor signaling [33]. Determination of
whether a similar duality exists in the regulation of
TCR-mediated signaling [34] will require a refined
analysis of the cytoplasmic and nuclear pathways
controlling phosphorylation and dephosphorylation
[30] of ERK in T lymphocytes. We described three
pMHC structures that differentially orient the functional
program of an alloreactive CD8 T cell through fine
tuning of activation thresholds and showed how subtle
variations in the peptide-binding groove affect pMHC-
TCR interactions. The characterization of TCR/pMHC
models that elicit defined threshold-controlled re-
sponses will help in the definition of the structural
requirements and signaling pathways to be manipulated
for therapeutic goals.
Materials and methods
Mice
Mice (H-2k) transgenic for the alloreactive anti-KbTCR
(BM3.3) [19] were Rag1-deficient (Rag–/–) when indicated.
All animal experiments were in accordance with protocols
approved by the French and European Directives.
Cell purification and culture
CD8 T cells were purified from lymph nodes of TCR BM3.3
mice by negative selection as previously described [19].
CD8 T cells represented 90–98 % of the enriched population
and were incubated for 10 min at 37?C with 5 lM CFSE
(Molecular Probes, Oregon). APC were T cell-depleted and
irradiated splenocytes. When indicated, cultures were supple-
mented with pBM1 or pBM8 peptides (Schafer-N, DK).
Flow cytometric analyses
Antibodies used for immunofluorescence staining were:
FITC-anti-CD69, PCP-Cy5.5-anti-CD8a,
anti-CD25 and PE-anti-IL-2 (BD Pharmingen, CA). Phospho-
ERK staining was performed as described [35] with anti-
phospho-ERK1/2 (Cell Signaling Technology, MA) and FITC-
goat anti-rabbit (Caltag, Burlingame, CA). Intracellular
staining for GzmB (Caltag) was performed as described
[10]. For IL-2 intracellular staining, cells were harvested after
40 h of primary culture and restimulated 4 h with ionomycin
and PMA in the presence of brefeldin A. Calcium signaling was
measured as described [17]. After addition of an appropriate
dilution of PE-labeled pMHC multimers, both calcium flux and
PE fixation were quantified for 15 min on an LSR cytofluori-
meter (Becton Dickinson), and kinetic analyses were per-
formed using FlowJo software (TreeStar, CA). Peak values of
pMHC multimer binding and of the increase in Ca2+signal
above background are expressed as RFI.
allophycocyanin-
Cytotoxic assays
Cytotoxic activity of TCR BM3.3 CD8 cells was tested on
51Cr-labeled (sodium chromate, NEN, MA) RMA (H-2b) or
RDM4 (H-2k) targets during a 4 h incubation. When indicated,
purified anti-CD8 mAb (10 lg/mL) was added during the CTL
assay.
Characterization of Kbm8-eluted peptides
Kbm8
peptides were extracted and fractionated as previously
described [11]. Aliquots of the HPLC fractions were tested
for the presence of peptide(s) reconstituting the epitopes
recognized by BM3.3 effectors on the Kb-negative Kbm8-
expressing Tap-2–/–RMA.S cell line [36] or Tap-1–/–humanT2
line [37].
molecules were immunopurified, and endogenous
Nathalie Auphan-Anezin et al.
Eur. J. Immunol. 2006. 36: 1856–1866
1864
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 10

Peptide-dependent MHC stabilization assay
This assay is based on the detection of peptide-loaded MHC
molecules atthe surface
conformation-dependent anti-KbmAb 5F1 [38] that recog-
nizes peptide-loaded Kband Kbm8molecules equally well [39]
but fails to bind peptide-free MHC molecules. Following
incubation with peptide, Tap–/–cells were washed, and
residual binding of anti-KbmAb 5F1 was determined after
30 min incubation at 37?C [11, 36]. The peptide concentration
required to stabilize the half-maximum MHC class I surface
molecules is reported as EC50 (Table 3). To estimate the
half-life of pMHC complexes, binding experiments carried out
using 10–6M peptide were followed by various times of
incubation at 37?C.
of Tap–/–
cellsusingthe
Preparation of pMHC multimers
A tagged KbcDNA contained in a pET3a vector [14] was
substituted with the cDNA encoding Kbm8to produce Kbm8/
peptide binary complexes as described [14]. pBM1/Kband
pBM8/Kbm8multimers were prepared as described [17] by
Immunomics Operations (Beckman-Coulter Immunotech,
Marseille, France).
pMHC multimer binding and decay
CD8Tcells fromRag–/–TCR BM3.3micewereincubatedat4?C
with various concentrations of PE-coupled pBM1/Kbor pBM8/
Kbm8multimers, washed and formaldehyde-fixed. Concentra-
tions are expressed as monomeric peptide/Kbequivalents. KD
determination was done using Prism software (GrahPad
Software, CA). For decay analysis, CD8 Tcells were incubated
with 15 nM pBM1/Kbor pBM8/Kbm8multimers for 90 min at
4?C,washed threetimes andincubatedfor varioustimes at4?C
in the same buffer with 100 lg/mL anti-KbmAb 5F1 to avoid
re-fixation of pMHC to the TCR. At each time point, cells were
washed, formaldehyde-fixed and analyzed (FACScan, Becton
Dickinson).Dissociationcurvesareplottedasthenaturallogof
[(MFI at time t / MFI at time 0) ?100] versus time [40]. The
half-live of pMHC/TCR binding was calculated as t1/2 = ln2 /
slope.
Protein expression and purification
The Kbm8heavy chain and b2-microglobulin were expressed in
E. coli as inclusion bodies and refolded and purified with the
pBM1 or pBM8 peptides as described previously [16, 41]. The
pBM8/Kbm8and pBM1/Kbm8complexes where then concen-
trated to 6.5 mg/mL for crystallization assays in a buffer
containing 20 mM Tris pH 8.0 and 200 mM NaCl.
Crystallization and data collection/crystal structure
determination and refinement
For information on this, please see the Supplementary
Materials and Methods
Acknowledgements: This work was supported by
institutional grants from «Institut National de la Sant?
et de la Recherche M?dicale», «Centre National de la
Recherche Scientifique», «Association pour la Re-
cherche sur le Cancer» (to A.-M. S.-V.) and the European
Communities (project EPI-PEP-VAC QLK2-CT-00620 to B.
M ). We thank L. Leserman for criticism of the manu-
script.
References
1 Amrolia, P. J., Muccioli-Casadei, G., Yvon,E., Huls, H., Sili, U.,Wieder, E.
D.,Bollard,C.etal.,SelectivedepletionofdonoralloreactiveTcellswithout
loss of antiviral or antileukemic responses. Blood 2003. 102: 2292–2299.
2 Savage, P., Gao, L., Vento, K., Cowburn, P., Man, S., Steven, N., Ogg, G. et
al., Use of B cell-bound HLA-A2 class I monomers to generate high-avidity,
allo-restricted CTLs against the leukemia-associated protein Wilms tumor
antigen. Blood 2004. 103: 4613–4615.
3 Blattman, J. N. and Greenberg, P. D., Cancer immunotherapy: a treatment
for the masses. Science 2004. 305: 200–205.
4 Li, Y., Moysey, R., Molloy, P. E., Vuidepot, A. L., Mahon, T., Baston, E.,
Dunn, S. et al., Directed evolution of humanT-cell receptors with picomolar
affinities by phage display. Nat. Biotechnol. 2005. 23: 349–354.
5 Walter, S., Herrgen, L., Schoor, O., Jung, G., Wernet, D., Buhring, H. J.,
Rammensee, H. G. and Stevanovic, S., Cutting edge: predetermined
avidity of human CD8 T cells expanded on calibrated MHC/anti-CD28--
coated microspheres. J. Immunol. 2003. 171: 4974–4978.
6 Webb, A. I., Borg, N. A., Dunstone, M. A., Kjer-Nielsen, L., Beddoe, T.,
McCluskey, J., Carbone, F. R. et al., The structure of H-2 K(b) and K(bm8)
complexed to a herpes simplex virus determinant: evidence for a
conformational switch that governs T cell repertoire selection and viral
resistance. J. Immunol. 2004. 173: 402–409.
7 Macdonald, W. A., Purcell, A. W., Mifsud, N. A., Ely, L. K., Williams, D. S.,
Chang, L., Gorman, J. J. et al., A naturally selected dimorphism within the
HLA-B44 supertype alters class I structure, peptide repertoire, and T cell
recognition. J. Exp. Med. 2003. 198: 679–691.
8 van Bleek, G. M. and Nathenson, S. G., The structure of the antigen--
binding groove of major histocompatibility complex class I molecules
determinesspecific selection ofself-peptides.Proc.Natl.Acad.Sci. USA1991.
88: 11032–11036.
9 Auphan-Anezin, N. and Schmitt-Verhulst, A. M., Differential survival of
transferred CD8 Tcells and host reconstitution depending onTCR avidity for
host-expressed alloantigen. J. Immunol. 2001. 166: 7200–7207.
10 Auphan-Anezin, N., Verdeil, G. and Schmitt-Verhulst, A.-M., Distinct
thresholds for CD8 T cell activation lead to functional heterogeneity:
CD8 T cell priming can occur independently of cell division. J. Immunol.
2003. 170: 2442–2448.
11 Guimezanes,A.,Barrett-Wilt,G.A.,Gulden-Thompson, P.,Shabanowitz,
J., Engelhard, V. H., Hunt, D. F. and Schmitt-Verhulst, A. M.,
Identification of endogenous peptides recognized by in vivo or in vitro
generated alloreactive cytotoxic T lymphocytes: distinct characteristics
correlated with CD8 dependence. Eur. J. Immunol. 2001. 31: 421–432.
12 Garbi, N., Tanaka, S., van den Broek, M., Momburg,F. and Hammerling,
G. J., Accessory molecules in the assembly of major histocompatibility
complex class I/peptide complexes: howessential are they for CD8(+) T-cell
immune responses? Immunol. Rev. 2005. 207: 77–88.
13 Saito, Y., Peterson, P. A. and Matsumura, M., Quantitation of peptide
anchor residue contributions to class I major histocompatibility complex
molecule binding. J. Biol. Chem. 1993. 268: 21309–21317.
14 Reiser, J. B., Darnault, C., Guimezanes, A., Gregoire, C., Mosser, T.,
Schmitt-Verhulst, A. M., Fontecilla-Camps, J. C. et al., Crystal structure of
a Tcell receptor bound to an allogeneic MHC molecule. Nat. Immunol. 2000.
1: 291–297.
15 Hunt, H. D., Pullen, J. K., Dick, R. F., Bluestone, J. A. and Pease, L. R.,
Structural basis of Kbm8 alloreactivity. Amino acid substitutions on the
Eur. J. Immunol. 2006. 36: 1856–1866
Molecular immunology
1865
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 11

beta-pleated floor of the antigen recognition site. J. Immunol. 1990. 145:
1456–1462.
16 Reiser, J. B., Darnault, C., Gregoire, C., Mosser, T., Mazza, G., Kearney,
A., van der Merwe, P. A. et al., CDR3 loop flexibility contributes to the
degeneracy of TCR recognition. Nat. Immunol. 2003. 4: 241–247.
17 Guimezanes, A., Montero-Julian, F. and Schmitt-Verhulst, A. M.,
Structural and kinetic basis for low affinity cross-reactivity in T cell
allorecognition. Eur. J. Immunol. 2003. 33: 3060–3069.
18 Verdeil, G., Puthier, D., Nguyen, C., Schmitt-Verhulst, A.-M. and
Auphan-Anezin, N., STAT5- mediated signals sustain a TCR-initiated gene
expression program towards differentiation of CD8 T cell effectors. J.
Immunol. 2006. 176: 4834–4842.
19 Auphan, N., Ghosh, S., Flavell, R. A. and Schmitt-Verhulst, A.-M.,
Differential requirements for NF-jB and AP-1 transactivation in response to
minimal TCR engagement by a partial agonist in naive CD8 T cells. J.
Immunol. 1999. 163: 5219–5227.
20 Lyons, D. S., Lieberman, S. A., Hampl, J., Boniface, J. J., Chien, Y., Berg,
L. J. and Davis, M. M., A TCR binds to antagonist ligands with lower
affinities and faster dissociation rates than to agonists. Immunity 1996. 5:
53–61.
21 Rabinowitz, J. D., Beeson, C., Wulfing, C., Tate, K., Allen, P. M., Davis, M.
M. and McConnell, H. M., Altered Tcell receptor ligands trigger a subset of
early T cell signals. Immunity 1996. 5: 125–135.
22 Germain, R. N.and Stefanova,I., The dynamics of Tcell receptorsignaling:
complex orchestration and the key roles of tempo and cooperation. Annu.
Rev. Immunol. 1999. 17: 467–522.
23 Gonzalez, P. A., Carreno, L. J., Coombs, D., Mora, J. E., Palmieri, E.,
Goldstein, B., Nathenson, S. G. and Kalergis, A. M., Tcell receptor binding
kineticsrequiredforTcellactivation dependonthe densityofcognateligand
on the antigen-presenting cell. Proc. Natl. Acad. Sci. USA 2005. 102:
4824–4829.
24 Chen, J. L., Stewart-Jones, G., Bossi, G., Lissin, N. M., Wooldridge, L.,
Choi, E. M., Held, G. et al., Structural and kinetic basis for heightened
immunogenicity of T cell vaccines. J. Exp. Med. 2005. 201: 1243–1255.
25 Krogsgaard,M.,Prado, N.,Adams, E.J., He, X.L.,Chow,D. C.,Wilson, D.
B., Garcia, K. C. and Davis, M. M., Evidence that structural rearrangements
and/or flexibilityduringTCR bindingcancontributetoTcellactivation.Mol.
Cell 2003. 12: 1367–1378.
26 Lee, J. K., Stewart-Jones, G., Dong, T., Harlos, K., Di Gleria, K., Dorrell,
L., Douek, D. C. et al., T cell cross-reactivity and conformational changes
during TCR engagement. J. Exp. Med. 2004. 200: 1455–1466.
27 Rosette, C., Werlen, G., Daniels, M. A., Holman, P. O., Alam, S. M.,
Travers, P. J., Gascoigne, N. R. et al., The impact of duration versus extent
of TCR occupancy onTcell activation: a revision of the kinetic proofreading
model. Immunity 2001. 15: 59–70.
28 Faroudi, M., Utzny, C., Salio, M., Cerundolo, V., Guiraud, M., Muller, S.
andValitutti,S.,LyticversusstimulatorysynapseincytotoxicTlymphocyte/
target cell interaction: manifestation of a dual activation threshold. Proc.
Natl. Acad. Sci. USA 2003. 100: 14145–14150.
29 Purbhoo, M. A., Irvine, D. J., Huppa, J. B. and Davis, M. M., Tcell killing
does not require the formation of a stable mature immunological synapse.
Nat. Immunol. 2004. 5: 524–530.
30 Altan-Bonnet, G. and Germain, R. N., Modeling T cell antigen
discrimination based on feedback control of digital ERK responses. PLoS
Biol. 2005. 3: e356.
31 Chau, L. A., Bluestone, J. A. and Madrenas, J., Dissociation of intracellular
signaling pathways in response to partial agonist ligands of the T cell
receptor. J. Exp. Med. 1998. 187: 1699–1709.
32 Xi, H. and Kersh, G. J., Induction of the early growth response gene 1
promoter by TCR agonists and partial agonists: ligand potency is related to
sustained phosphorylation of extracellular signal-related kinase substrates.
J. Immunol. 2003. 170: 315–324.
33 Sasagawa, S., Ozaki, Y., Fujita, K. and Kuroda, S., Prediction and
validationof thedistinctdynamicsof transientandsustainedERK activation.
Nat. Cell Biol. 2005. 7: 365–373.
34 Tsukamoto, H., Irie, A. and Nishimura, Y., B-Raf contributes to sustained
extracellularsignal-regulatedkinaseactivationassociatedwithinterleukin-2
production stimulated through the Tcell receptor. J. Biol. Chem. 2004. 279:
48457–48465.
35 Chow, S., Patel, H. and Hedley, D. W., Measurement of MAP kinase
activation by flow cytometry using phospho-specific antibodies to MEK and
ERK: potential for pharmacodynamic monitoring of signal transduction
inhibitors. Cytometry 2001. 46: 72–78.
36 Dyall, R., Fremont, D. H., Jameson, S. C. and Nikolic-Zugic, J., T cell
receptor (TCR) recognition of MHC class I variants: intermolecular
second-site reversion provides evidence for peptide/MHC conformational
variation. J. Exp. Med. 1996. 184: 253–258.
37 Chattopadhyay, S., Theobald, M., Biggs, J. and Sherman, L. A.,
Conformational differences in major histocompatibility complex-peptide
complexes can result in alloreactivity. J. Exp. Med. 1994. 179: 213–219.
38 Bluestone, J. A.,Jameson, S., Miller, S. and Dick,R. 2nd, Peptide-induced
conformationalchangesinclassIheavychainsaltermajorhistocompatibility
complex recognition. J. Exp. Med. 1992. 176: 1757–1761.
39 Albert,F.,Buferne,M.,Boyer,C.andSchmitt-Verhulst,A.M.,Interactions
between MHC-encoded products and cloned T-cells. I. Fine specificity of
induction of proliferation and lysis. Immunogenetics 1982. 16: 533–549.
40 Savage, P. A., Boniface, J. J. and Davis, M. M., A kinetic basis for T cell
receptor repertoire selection during an immune response. Immunity 1999.
10: 485–492.
41 Zhang, W., Young, A. C., Imarai, M., Nathenson, S. G. and Sacchettini, J.
C., Crystal structure of the major histocompatibility complex class I H-2Kb
molecule containing a single viral peptide: implications for peptide binding
and T-cell receptor recognition. Proc. Natl. Acad. Sci. USA 1992. 89:
8403–8407.
42 Kraulis, P. J., MOLSCRIPT: a program to produce both detailed and
schematic plots of protein structures. J. Appl. Cryst. 1991. 24: 946–950.
43 Merritt, E. A. and Murphy, E. P., Raster3D Version 2.0 - A program for
photorealistic molecular graphics. Acta Cryst. 1994 D. 50: 869–873.
Nathalie Auphan-Anezin et al.
Eur. J. Immunol. 2006. 36: 1856–1866
1866
f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu