Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Jan 12.
Published in final edited form as: Structure. 2011 Jan 12;19(1):80–89. doi: 10.1016/j.str.2010.10.007

Molecular insights into γδ T-cell costimulation by an anti-JAML antibody

Petra Verdino 1, Deborah A Witherden 2, M Sharon Ferguson 1, Adam L Corper 1, André Schiefner 1, Wendy L Havran 2, Ian A Wilson 1,3,*
PMCID: PMC3039130  NIHMSID: NIHMS253667  PMID: 21220118

Summary

γδ T cells bridge innate and adaptive immunity and function in immunosurveillance, immunoregulation, tumor cell recognition, and as first line of defense against microbial infection. Costimulation of epithelial γδ T-cell activation by the JAML receptor can be induced by interaction with its endogenous ligand CAR or by binding of the stimulatory antibody HL4E10. We, therefore, determined the crystal structure of the JAML-HL4E10 Fab complex at 2.95 Å resolution. HL4E10 binds the membrane-proximal domain of JAML through hydrophobic interactions that account for nanomolar affinity and long half life, contrasting with the fast kinetics and micromolar affinity of the hydrophilic CAR interaction with the membrane-distal JAML domain. Thus, despite different binding sites and mechanisms, JAML interaction with these two disparate ligands leads to the same functional outcome, namely JAML triggering and induction of cell signaling. Several characteristics of the HL4E10 antibody might then be harnessed in therapeutic applications, such as promoting healing of acute or chronic wounds.

Keywords: Crystal structure, antibody-antigen complex, binding characteristics, stimulatory antibody, γδ T cell, costimulation, cell signaling

Introduction

Dendritic epidermal γδ T cells (DETC) residing in the skin are prototypical intra-epithelial lymphocytes (Jameson et al., 2004) with a unique dendritic morphology (Boismenu and Havran, 1998). As the only resident T cell population in the epidermis, DETC are key players in tissue homeostasis; including tumor surveillance and wound repair (Jameson et al., 2002; Jameson et al., 2003; Jameson et al., 2004; Sharp et al., 2005). DETC use an invariant T-cell receptor (TCR) to bind unknown self-antigens expressed by neighboring cells during biological insults including infection, trauma or malignancy (Jameson et al., 2003; Jameson et al., 2004). Recent evidence suggests that for activation and for production of rapid and robust effector functions, epithelial γδ T cells, like αβ T cells, require accessory molecules to enhance the TCR-mediated signals (Witherden et al., 2010).

In αβ T cells, accessory molecules are utilized for modulating antigen responses and for the differentiation of T cells into phenotypically distinct effector cells. Costimulatory signals are important for clonal expansion and protective immune response, while inhibitory signals maintain T-cell self-tolerance and prevent autoimmunity (Croft, 2003; Kroczek et al., 2004; Salomon and Bluestone, 2001; Sharpe and Freeman, 2002; Watts, 2005). Thus, manipulation of immune modulatory interactions holds considerable promise in the clinic (Abken et al., 2002; Chambers et al., 2001; Martin-Orozco and Dong, 2006; Snanoudj et al., 2006; Stuart and Racke, 2002; Vincenti and Luggen, 2007; Weaver et al., 2008; Zang and Allison, 2007).

The importance of accessory molecules for activation of γδ T-cell responses has only recently been established (Whang et al., 2009; Witherden et al., 2010). Consistent with their distinct function, DETC do not express the αβ T-cell coreceptors CD4 and CD8, or the costimulatory molecules CD28 and ICOS, which are essential for αβ T-cell function (Boismenu and Havran, 1998; Haas et al., 1993; Jameson et al., 2004; Shires et al., 2001). Instead, the Junctional-Adhesion Molecule-Like protein, JAML (Moog-Lutz et al., 2003), was identified as the first costimulatory receptor specific to epithelial γδ T cells (Witherden et al., 2010). The interaction of JAML with Coxsackie and Adenovirus receptor, CAR (Bergelson et al., 1997; Guo et al., 2009; Luissint et al., 2008; Verdino et al., 2010a; Zen et al., 2005) on keratinocytes or with the stimulatory HL4E10 IgG antibody (Ab) induces potent costimulation, cytokine and growth factor production, activation of MAP kinase pathways. This stimulation ultimately leads to increased DETC survival and proliferation, and is important for modulating γδ T-cell responses during epithelial challenges, such as wound repair (Witherden et al., 2010).

Interestingly, HL4E10 elicits the same γδ T-cell responses as the natural ligand CAR, but does not compete with CAR for JAML binding (Witherden et al., 2010). Furthermore, HL4E10 can restore JAML-mediated costimulation of skin γδ T cells following blockage of CAR and, thus, reestablish proper wound healing (Witherden et al., 2010). Given the promising results of manipulating immune modulatory interactions of αβ T cells, humanized versions of HL4E10 might find application in the treatment of γδ T cell-associated diseases including chronic non-healing wounds. To gain a deeper understanding of the mechanisms of Ab-induced γδ T-cell costimulation, we investigated the molecular, functional, and structural characteristics of the HL4E10-JAML complex and compared those to the endogenous CAR-JAML interaction.

The JAML-HL4E10 complex crystal structure reveals that the antibody HL4E10 binds the membrane-proximal domain of JAML through hydrophobic interactions that account for nanomolar affinity and long half life in contrast to the extremely hydrophilic interaction of the natural ligand CAR with the membrane-distal domain of JAML which displays fast kinetics and only micromolar affinity. However, HL4E10 and CAR both induce JAML triggering, cell signaling through PI3K, and γδ T cell proliferation. Our findings support ligand-induced JAML dimerization/clustering as the mechanism for receptor triggering that, in the case of the natural ligand, is initiated by increased CAR expression. Furthermore, several characteristics of the HL4E10 antibody appear favorable for therapeutic applications, including its ability to elicit γδ T-cell responses indiscernible from those of the endogenous ligand CAR without interfering with the normal JAML-CAR interaction, as well as the high affinity, high specificity and long half-life of the antibody-receptor interaction.

Results

Crystal structure determination of the JAML-HL4E10 Fab complex

The entire glycosylated mouse JAML ectodomain (residues 1–260) was produced in SC2 cells essentially as described (Verdino et al., 2010a). The HL4E10 Fab was obtained by pepsin digestion and reduction of hamster anti-JAML IgG (λ) HL4E10 produced in a hybridoma cell line (Verdino et al., 2010b). The JAML-HL4E10 Fab complex was prepared for structural studies by incubating JAML with excess HL4E10 Fab and subsequent purification of the complex by size exclusion chromatography. Crystals were obtained in tetragonal space group P43212 and a complete data set to 2.95 Å resolution was collected at the Advanced Photon Source beamline 23ID-D (Argonne National Laboratory, Chicago, IL). The structure was determined by molecular replacement using unliganded structures of JAML (PDB entry 3MJ6) and the HL4E10 Fab (3MJ8) as templates and refined to Rcryst = 22.5% and Rfree = 28.7%. The final model consists of JAML residues Val8-Asp236 (residues 1–7, which correspond to the A-strand of the D1 Ig domain, and residues 237–260, which compose the C-terminal stalk that links the Ig domains to the cell membrane, as well as the C-terminal His6-tag, were not visible in the electron density maps), N-acetylglucosamine residues linked to Asn59 and to Asn69, and, due to crystal packing, a very well-defined, N-linked carbohydrate moiety (two N-acetylglucosamines, one fucose, three mannoses) linked to Asn105. The HL4E10 Fab light chain consisted of TyrL2-SerL212, and the heavy chain of GlnH1-GlyH228 (Table 1).

Table 1.

JAML-HL4E10 Fab complex data collection and refinement statistics

Data collection
Space group P43212
Cell dimensions
    a, b, c (Å) 125.0, 125.0, 107.8
Wavelength (Å) 0.9793
Resolution (Å) 30.00-2.95 (3.06-2.95)*
Rmerge (%) 12.2 (61.7)
<I/σI> 6.8 (1.8)
Completeness (%) 97.7 (99.0)
Unique reflections 18,365 (1,816)
Redundancy 3.1 (2.9)
Refinement
Resolution (Å) 30.00-2.95
No. reflections work/test 17,200/927
Rcryst/Rfree (%) 22.5/28.7
No. atoms
    JAML 1,841
    Carbohydrates 99
    HL4E10 light chain 1,575
    HL4E10 heavy chain 1,579
B-values (Å2)
    JAML 72
    Carbohydrates 75
    HL4E10 light chain 74
    HL4E10 heavy chain 73
R.m.s. deviations
    Bond lengths (Å) 0.006
    Bond angles (°) 1.1
Ramachandran plot (%)
    Favored/allowed/outliers 94.2/5.5/0.3#
Open in a new tab
*

Highest resolution shell is given in parenthesis

#

TyrH97 and ProH149 are the only residues in the disallowed region, but are both located in loops with well-defined electron density.

Molecular structure of the JAML-HL4E10 Fab complex

Consistent with biochemical binding studies (Witherden et al., 2010), the JAML-HL4E10 complex crystal structure revealed that HL4E10 Fab binds the membrane-proximal, D2 Ig domain of JAML (Fig. 1). All six Fab CDRs (L1–3, H1–3) interact with the C-terminal JAML D2 domain and bury 1600 Å2 of the molecular surface area (820 Å2 on HL4E10 and 780 Å2 on JAML). The area and the shape complementarity {Sc=0.66; (Lawrence and Colman, 1993)} of the interface are comparable to those of typical antibody/antigen complexes (Lo Conte et al., 1999). Approximately 50% of the buried HL4E10 surface area (420 Å2 out of 820 Å2) is contributed by aromatic residues (34% tyrosine, 10% tryptophan, 7% phenylalanine) (Fig. 2,3). In addition, eight, scattered HL4E10 framework residues contact the JAML D1 BC-loop and bury 230 Å2 surface areas on both, HL4E10 and JAML (Fig. 1).

Figure 1. The stimulatory HL4E10 antibody binds the C-terminal, membrane proximal JAML D2 Ig domain.

Figure 1

Open in a new tab

Ribbon representation of the JAML-HL4E10 Fab complex. JAML and its two variable Ig domains are shown in light blue for the membrane-distal, N-terminal, D1 domain (residues 8–121) and in salmon for the membrane-proximal, C-terminal, D2 domain (residues 122–236). The JAML D1 A-strand (residues 1–7) and the C-terminal stalk region (residues 234–260, salmon dashes), which tethers the JAML Ig domains to the cell membrane, were disordered. Carbohydrate moieties attached to the three N-linked glycosylation sites (Asn59, Asn69, and Asn105) are shown in stick representation. The Fab fragment of the HL4E10 IgG is shown in gray for the light chain (variable domain VL and constant domain CL, residues 2–212) and dark gray for the heavy chain (variable domain VH and constant domain CH1, residues 1–228). All six Fab CDRs contact JAML (CDR L1, yellow; L2, cyan, L3, red; H1, blue; H2, pink; H3, green). The interaction of HL4E10 is focused on the C-terminal JAML D2 domain, in particular on the C’C”-loop, the D-strand, the DE-loop, and the BC-loop. CDR H3 is inserted between the JAML A’GFCC’C” and BED sheets and locks the JAML C’C”-loop and the C”-strand, which has unraveled from the β-sheet of the Ig-fold, into a conformation strikingly distinct from that found in the crystal structure of unliganded JAML

Figure 2. The JAML-HL4E10 complex is characterized by extensive hydrophobic interactions.

Figure 2

Open in a new tab

Molecular surface of JAML with the HL4E10 epitope colored in yellow (CDR L1 contacts), cyan (CDR L2), red (CDR L3), blue (CDR H1), magenta (CDR H2), and green (CDR H3). In the detail shown on the right, the electrostatic potential was mapped onto the molecular surface of JAML and contoured at ±60 kT/eV (blue/red). HL4E10 residues contacting JAML are shown as sticks above the JAML surface. The JAML-HL4E10 interface is noticeably hydrophobic and dominated by van der Waals’ interactions between aromatic HL4E10 (TyrL32, TyrL49, TrpL91, TrpL96, TyrH32, PheH96, TyrH97, TyrH99) and hydrophobic JAML residues (Val148, Met173, Phe178, Tyr169, Ile194, Leu190). Surprisingly, 36% of all HL4E10 contacts with JAML are contributed by only one residue, TyrH97, which is deeply inserted into a hydrophobic pocket in the JAML D2 domain (black arrow).

Figure 3. Close-up stereo representation of the JAML-HL4E10 interaction.

Figure 3

Open in a new tab

Insertion of HL4E10 CDR L1 and H3 into the cleft between JAML C’C”- and DE-loops. The Fab CDR loops are color-coded (CDR L1, yellow; L2, cyan, L3, red; H1, blue; H2, pink; H3, green) and JAML is shown in gray. The 2Fo-Fc electron density around the interacting residues is contoured at 1σ (light blue mesh, in panel A). Note the tyrosines that are engaged in hydrophobic stacking and hydrogen bonds via their polar hydroxyl groups. For example, TyrH97 of HL4E10 CDR H3 is packed against Tyr169 and Ile194 and the main chain of Asp170 and Ser171, and hydrogen bonds to the backbone carbonyl of Val148 and the amide of Lys150. Similarly, JAML Tyr169 stacks against HL4E10 TyrH97 and TyrL32 and hydrogens bonds with ThrH98 and GluL50.

The HL4E10 CDRs contact the BC-, C’C”-, and DE-loops of the JAML D2 Ig domain. The key feature of this interaction is the insertion of CDR H3 between two β-sheets, GFCC’C” and BED, of the JAML Ig-fold. While PheH96, TyrH97, TyrH99, and TyrL32 of the Fab are inserted into a cleft in the D2 domain of JAML, TrpL91 and TrpL96 of CDR L3 contribute additional hydrophobic contacts with the JAML C’C”-loop (Figs. 2,3). A total of 170 interactions are formed between JAML and HL4E10: 154 van der Waals contacts, 15 H-bonds, and a salt bridge between AspH101 (CDR H3) and JAML Lys146. TyrH97 in CDR H3 appears to be the key contributor to the JAML-HL4E10 interaction. Its side chain is deeply inserted into a pocket between the JAML C’C” and DE loops (Fig. 2) and contributes 62 van der Waals contacts (36% of all HL4E10 contacts with JAML), as well as two H-bonds from its hydroxyl group to the main chain of two deeply buried JAML residues, Val148 and Lys150 (Fig. 3B).

HL4E10 binds a conformationally flexible region of JAML

No significant conformational changes occur in HL4E10 upon complex formation [RMSD of 0.7 Å for superimposition of the backbone atoms of the variable domains of complexed HL4E10 and unliganded HL4E10 {PDB 3MJ8; (Verdino et al., 2010b)}]. In contrast, the JAML region which constitutes the HL4E10 binding site adopts a strikingly different conformation in the JAML-HL4E10 complex as compared to unliganded JAML [PDB 3MJ6; (Verdino et al., 2010a)], whereas the rest of JAML is essentially unchanged (backbone atom RMSD of 1.1 Å for entire JAML; 1.1 Å for D1, 1.0 Å for D2). In the JAML-HL4E10 complex, JAML Asn172-Leu181 (C’C”-loop) and the C”-strand exhibit an average RMSD of 4.0 Å and the Phe178 side chain undergoes an 11 Å movement (Fig. 4A), mostly due to interaction with CDRs H3 and L3. Additionally, the N-terminal portion of the JAML C”-strand (Gly176-Phe178) peels away from the β-sheet and extends the C’C”-loop (Fig. 1,4). Disruption of the β-sheet structure accompanies movement of the C”-strand towards the HL4E10-binding site, and disrupts the normal β-sheet H-bonds observed between the C’- and C”-strands. Instead of four standard, β-sheet H-bonds in unliganded JAML, the conformational change and a register shift results in Ser168 in the C’-strand H bonding with Gln179 (rather than Phe178) in the C”-strand in the JAML-HL4E10 complex (Fig. 4B).

Figure 4. HL4E10 binds to a conformationally flexible region of JAML.

Figure 4

Open in a new tab

Wall-eyed stereo representation of the different conformations of the HL4E10 epitope of unliganded JAML (3MJ6, gray) and HL4E10-bound JAML (3MJ9, red). Residues are labeled in one letter code for unliganded JAML. (A) View of the HL4E10 epitope of JAML from the Ab perspective. The JAML C’C”-loop undergoes up to 7 Å backbone shifts and up to 11 Å shifts for side chains. (B) The JAML C”-strand has lost its β-strand character and moved ~3 Å towards the C’C”-loop. Instead of four standard H-bonds between the C’- and the C”-strand as in unliganded JAML (gray dashes), only two non-conventional H-bonds are formed between those strands in the JAML-HL4E10 complex (red dashes).

In the crystal structure of unliganded JAML [PDB 3MJ6; (Verdino et al., 2010a)], the region which constitutes the binding site for HL4E10 exhibits elevated average B-values (46 Å2) as compared to the rest of JAML (36 Å2), indicative of increased conformational flexibility. Thus, the C”- and D-strands of the JAML D2 domain appear to be sampling a variety of conformations in the unbound form and, upon HL4E10 binding, are locked into a distinct conformation, which contrasts to some extent with the general perception of Ig domain β-sandwich scaffolds as fairly rigid molecular structures.

HL4E10 and CAR interact with distinct binding kinetics and affinities with different JAML domains, but induce comparable PI3K recruitment and DETC proliferation

To further explore similarities and differences in the interaction of HL4E10 and CAR with the membrane-proximal and membrane-distal JAML domains, respectively (Fig. 5), binding kinetics and affinities of the respective JAML-ligand complexes were investigated by SPR (Fig. 6, S3). HL4E10 Fab binds surface-immobilized JAML with a fast on-rate (Kon ~1.3×105 M−1s−1) and a slow off-rate (Koff ~0.001 s−1) resulting in a half life of the interaction in the order of ~10 min. In contrast, the JAML-CAR interaction is characterized by a fast on-rate (Kon ~1×105 M−1s−1) and a biphasic dissociation where the majority of CAR dissociates rapidly (Koff ~0.5 s−1) with an estimated half life of ~1 s and the rest dissociates more slowly. A similar biphasic dissociation pattern has previously been observed for the CD2–CD48 interaction and been attributed to small amounts of multimeric aggregates in the sample (van der Merwe et al., 1993). The binding affinities of the two JAML complexes differ by about three orders of magnitude, with a KD of ~8 nM for JAML-HL4E10 and ~3–6 µM for JAML-CAR.

Figure 5. Comparison of JAML-binding to the stimulatory HL4E10 antibody or its endogenous ligand CAR.

Figure 5

Open in a new tab

JAML bound by HL4E10 is superimposed with JAML complexed with CAR (3MJ7) (N-terminal D1 Ig domain, light green, C-terminal D2 Ig domain, gold, disordered CAR D2 loops, dots). HL4E10 binds the membrane-proximal JAML D2 Ig domain (transparent salmon surface) through extensive hydrophobic interactions mediated by all six CDR loops. In contrast, CAR binds the A’GFCC’C” β-sheet of the JAML membrane-distal D1 domain (transparent light blue surface) via an overwhelmingly hydrophilic interface.

Figure 6. HL4E10 and CAR interaction with JAML as observed by SPR reveal distinct binding affinities and kinetics.

Figure 6

Open in a new tab

SPR binding data of HL4E10 Fab and CAR to immobilized JAML. Injection start and end points are indicated with red arrows. (A) HL4E10 Fab (2–500 nM) binding to JAML. Data (colored lines) are overlaid with a global fitting using a standard 1:1 Langmuir binding isotherm (black lines). The interaction is characterized by a fast on-rate and slow off-rate, apparent by the long half-life in the order of ~10 min. (B) CAR (0.3–20 µM) binding to JAML. Data are overlaid with a global fitting using a complex binding model accounting for heterogeneity of the ligand. The JAML-CAR interaction exhibits rapid on-and off-rates and a half-life of ~1 s. See also Figure S3.

HL4E10 and CAR induce potent costimulation of DETC through cytokine and growth factor production and activation of MAP kinase pathways (Witherden et al., 2010) and binding of phosphoinositide-3-kinase (PI3K) to the JAML intracellular domain (ICD) is a key component in JAML signaling (Verdino et al., 2010b). Consistent with their comparable fast on-rates, HL4E10 and CAR both induce rapid PI3K association within one minute after JAML ligation that is sustained for up to 30 mins (Fig. 7A). It has been shown that, in solution, only ligation by dimeric/multivalent ligands (such as a bivalent CAR-Fc fusion construct or intact HL4E10 IgG), but not monomeric ligands, leads to PI3K recruitment to JAML (Verdino et al., 2010b). However, when immobilized to a surface, monomeric HL4E10 Fab and CAR-His induce similar levels of DETC proliferation as dimeric JAML ligands (Fig. 7B).

Figure 7. PI3K recruitment and induction of DETC proliferation by HL4E10 binding to JAML.

Figure 7

Open in a new tab

(A) Anti-PI3K Western blots of JAML immunoprecipitated from epithelial γδ T cell lysates at various time points after stimulation with HL4E10 IgG. After 1 min, PI3K association to JAML is significantly elevated over the basal level. (B) Costimulation of γδ T-cell activation through HL4E10 IgG or Fab-binding to JAML. Proliferation of DETC to immobilized anti-CD3 IgG either alone (shaded bars) or in combination with anti-JAML HL4E10 IgG (solid bars), control IgG 1F4 (open bars), anti-JAML HL4E10 Fab fragment (striped bars) or control Fab (dotted bars) was assessed by 3H-thymidine incorporation. HL4E10 IgG and its Fab fragment both induce comparable activation of DETC when surface immobilized.

The high specificity of the JAML-CAR interaction is evolutionary conserved

Sequence conservation analysis reveals that JAML and CAR homologs are found in mammals and other vertebrates including fowl (Fig. S1,S2). The evolutionary more ancient JAMLs from opossum, platypus, and chicken contain only the N-terminal, CAR-binding D1 Ig domain (Fig. S1), while JAML of placental mammals contains an additional D2 Ig domain. Several features, including ~44% sequence identity with D1, suggest that D2 arose from domain duplication. For CAR, the overall domain structure and entire sequence is highly conserved amongst species (Fig. S2) (Carson, 2001; Coyne and Bergelson, 2005; Freimuth et al., 2008; Howitt et al., 2003). Notably, the charged residues that provide high specificity to the JAML-CAR interaction are conserved amongst all JAML and CAR homologs suggesting that these charged residues are crucial for the physiological function of this receptor-ligand pair.

Discussion

The stimulatory Ab HL4E10 and the endogenous natural ligand CAR use entirely different receptor-binding mechanisms to induce JAML-mediated costimulation of epithelial γδ T-cell activation. The Ab and CAR elicited γδ T-cell responses, including activation of kinase cascades, cytokine production, and cell proliferation, are comparable (Witherden et al., 2010), yet the interactions of HL4E10 and CAR with JAML are completely distinct. First, HL4E10 binds the membrane-proximal JAML D2 domain, while CAR binds the membrane-distal JAML D1 domain. Second, while the total buried surface areas and shape complementarities are comparable (JAML-HL4E10: 1605 Å2, Sc=0.66; JAML-CAR: 1460 Å2, Sc=0.64), the composition of the JAML-HL4E10 and JAML-CAR interfaces are remarkably different: the former is dominated by hydrophobic interactions, while the latter is exceptionally hydrophilic due to numerous charged residues that engage in an interdigitating salt bridge network (Verdino et al., 2010a). Third, the different receptor-ligand interactions are reflected in distinct binding kinetics and three orders of magnitude difference in affinity.

Based on their fundamentally different modes of interaction with JAML, the central question arises as to how both ligands are able to induce essentially identical γδ T-cell responses through JAML engagement? Moreover, how do these findings impact our understanding of JAML receptor function and, thus, the molecular mechanism of epithelial γδ T-cell costimulation? The fact that HL4E10 and CAR bind different JAML Ig domains indicates that binding to the physiological receptor binding site on the D1 domain is not the only mode of binding and mechanism that can elicit receptor triggering. While the C”- and D-strands of the JAML D2 domain adopt different conformations in unliganded JAML and the JAML-HL4E10 complex, these differences appear to be an reflection of the inherent conformational flexibility of this particular region in JAML rather than a specific signaling mechanism, especially as it is difficult to envision how such changes could relay a signal along the flexible C-terminal stalk to the ICD. Furthermore, no significant structural rearrangements were observed upon CAR-binding to JAML (Verdino et al., 2010a), also suggesting a mechanism for JAML receptor triggering that is not caused by conformational changes.

Receptor crosslinking/clustering is a common mechanism of signal transduction in cell surface receptors including receptor tyrosine kinases or cytokine receptors (reviewed in (Lemmon and Schlessinger, 1998; Schlessinger, 2000)). For JAML, it has been shown that binding of bivalent, but not monovalent, ligands induces recruitment of phosphoinositide-3-kinase (PI3K) to the JAML ICD and subsequent cell signaling (Verdino et al., 2010a). Furthermore, the crystal structure of the JAML-CAR ectodomain complex revealed dimerization of the JAML-CAR complex that is mediated by interaction of two adjacent CAR molecules (Verdino et al., 2010a). The JAML-HL4E10 Fab structure does not contain any higher order oligomeric assemblies, consistent with the inability of monovalent HL4E10 Fab to induce PI3K recruitment in solution. However, if the Fab fragment is immobilized to a surface, it acts like a “quasi-multivalent” ligand and induces potent DETC proliferation comparable to that of intact bivalent HL4E10 IgG. These findings thus strongly support ligand-induced, JAML dimerization/clustering as a receptor triggering mechanism.

Expression profiles and the characteristics of the JAML-CAR interaction also suggest that JAML acts as a sensor on the DETC surface that is activated by clustering induced by increased CAR expression. First, JAML is constitutively expressed at low levels on DETC and is upregulated upon cellular activation (Witherden et al., 2010). Likewise, surface expression of CAR has been shown to be modified by inflammation, by non-inflammatory responses to tissue damage, by influence of neighboring cells (Carson, 2001; Freimuth et al., 2008), and is increased on stressed keratinocytes (Witherden et al., 2010). Second, the hydrophilic nature of the CAR epitope on JAML likely aids in JAML stability and solubilization on the cell surface in the absence of ligand. Third, the high specificity and fast kinetics of the JAML-CAR ectodomain interaction ensure a sensitive, rapid cellular JAML response to ligand binding that is also reflected by the fast association and dissociation of PI3K on the JAML ICD (Verdino et al., 2010a).

These findings all support the current view of skin γδ T cells as first responders to cellular stress that exhibit effector functions rapidly after biological or physical insult (Hayday, 2009). While the semi-activated state of skin γδ T cells enables rapid responses without significant delays (for example, due to clonal expansion), the powerful consequences of their full activation, such as inflammation and cytolysis, require tight control mechanisms. Costimulatory receptor-ligand pairs, such as JAML and CAR, which determine the final outcome of the primary TCR-ligand interaction, constitute an ideal check-point. The constitutive expression of JAML ensures high sensitivity and fast responses, while the pronounced specificity provided by the charge-complementary interface reduces the likelihood of false triggering. The striking conservation of the complementary charged residues in the JAML-CAR interface, even in the evolutionary more distant single Ig domain JAML receptors in non-placental mammals and fowl, also supports this hypothesis.

In αβ T cells, manipulation of immune-modulatory receptor signals as means to treat autoimmune diseases, cancers, and allograft rejection has become a promising new clinical strategy (Abken et al., 2002; Chambers et al., 2001; Martin-Orozco and Dong, 2006; Snanoudj et al., 2006; Stuart and Racke, 2002; Vincenti and Luggen, 2007; Weaver et al., 2008; Zang and Allison, 2007). The ability of the HL4E10 IgG to restore JAML-mediated costimulation of DETC in the absence of CAR ligation and to reestablish proper wound healing has indicated potential for the use of HL4E10 derivatives to modulate skin γδ T-cell responses in the clinic (Witherden et al., 2010). Our data reveal that HL4E10 might, indeed, be a good candidate for such applications. Significantly, HL4E10 induces JAML-mediated γδ T-cell responses that are indiscernible from those elicited by the natural ligand CAR (Witherden et al., 2010). However, as the epitope for HL4E10 is on a different domain from the CAR binding site, no competition is found with the endogenous JAML-CAR interaction. Thus, HL4E10 is less likely to interfere with CAR-related signaling in keratinocytes in vivo. As seen for CAR, HL4E10 binding to JAML is characterized by a fast on-rate; however, the extensive hydrophobic interface result in a ~500 times longer half-life for the JAML-HL4E10 complex. PI3K recruitment at the JAML ICD is tightly linked to the extracellular JAML-CAR interaction (Verdino et al., 2010a), where PI3K associates with JAML within 1 min of CAR binding and dissociates within 1 min after the JAML-CAR interaction has been disrupted. This rapid response seems to be facilitated by the fast on- and off-rates of CAR for JAML. Likewise, for the HL4E10-JAML interaction, the fast on-rate observed by SPR also coincides with rapid PI3K recruitment and it is likely that the long half-life of the complex could lead to sustained PI3K signaling. Furthermore, the rigid HL4E10 combining site provides optimal positioning of the key residues for JAML interaction, including TyrH97, which deeply protrudes into a pocket on the JAML D2 surface, and JAML Tyr169 which inserts into the HL4E10 combining site. These features all increase specificity and eliminate HL4E10 targeting of unrelated receptors. Indeed, the high specificity of HL4E10 for JAML has been demonstrated by lack of cross-reactivity with other Junctional Adhesion molecules including JAM-A, B, and C (Witherden et al., 2010).

DETC activation by both HL4E10 and CAR binding critically depends on suboptimal concentrations of CD3 indicating that TCR signaling is a prerequisite for JAML-mediated costimulation (Witherden et al., 2010). HL4E10 thus contrasts with Abs against the major αβ T-cell costimulatory receptor CD28, where conventional stimulatory Abs recognize an epitope close to the binding site for the endogenous ligands, CD80 (B7-1) and CD86 (B7-2), while TCR independent superagonists bind to a different epitope located elsewhere (Davis and van der Merwe, 2006; Evans et al., 2005; Luhder et al., 2003). The dependency of HL4E10 activity on primary TCR-CD3 signals without interfering with the natural JAML-CAR interaction might help avoid the adverse effects elicited by anti-CD28 superagonistic Abs in clinical trials (Suntharalingam et al., 2006; Waibler et al., 2008).

Summary

In summary, we demonstrate how JAML-mediated, epithelial γδ T-cell activation can be induced by two distinct ligands, monoclonal Ab HL4E10 and CAR, that use fundamentally different receptor-binding mechanisms. Recently, multiple different Abs against immune-modulatory receptors have proven successful in tumor immunotherapy, treatment of autoimmune diseases, or prevention of allograft rejection. Our studies reveal HL4E10 characteristics that are favorable for potential use of humanized versions of this antibody in therapy. Understanding the structure-function relationships of JAML, CAR, and HL4E10 and their role in immune modulation might thus aid in the development of novel therapeutics for treatment of γδ T-cell related diseases, such as chronic non-healing wounds, colitis, and cancer.

Experimental Procedures

Protein Expression and Purification

The ectodomain of mouse JAML comprising residues 1–260 of the mature protein including an engineered C-terminal His6-tag and a stop codon was expressed in SC2 cells, isolated from the supernatant by Ni-NTA affinity chromatography, and purified to homogeneity by size exclusion chromatography as previously described (Verdino et al., 2010a).

The monoclonal hamster anti-JAML IgG, HL4E10, was isolated from hybridoma cell supernatants by Protein A chromatography and digested for 3 hours with 4% pepsin in 1 M Na-acetate pH 5.5 in the presence of 20 mM cysteine. Fc and undigested IgG were removed by binding to a Protein A column. The Fab was further purified on Protein G and size exclusion columns. The cDNA sequence of HL4E10 IgG was determined (Verdino et al., 2010b).

Crystallization, Data Collection, Structure Determination, and Refinement

The JAML-HL4E10 complex was prepared by incubating JAML with 1.5-fold stoichiometric excess HL4E10 Fab for 2 h at 4°C in 20mM Tris-HCl pH8.0, 150mM NaCl with subsequent purification by size exclusion chromatography in the same buffer (Superdex 75 10/30, Amersham Pharmacia Biotech). Rod-shaped crystals (~0.5 × 0.02 × 0.02mm) were obtained at 22°C by mixing 1µl 6.5 mg/ml JAML-HL4E10 (in 20mM Tris-HCl pH8.0, 150mM NaCl) with 1µl 1.2–1.4 M Na-malonate, pH 6.0 by sitting drop vapor diffusion against 0.5ml reservoir solution. JAML-HL4E10 complex data were collected from a single crystal to 2.95 Å resolution at the Advanced Photon Source beamline 23ID-D (Argonne National Laboratory, Chicago, IL), and integrated and scaled with HKL2000 (Otwinowski and Minor, 1997).

The unliganded structures of JAML (3MJ6) (Verdino et al., 2010a) and the HL4E10 Fab (3MJ8) (Verdino et al., 2010b) were used as templates for molecular replacement. Solutions for JAML and the individual variable and constant domains of the Fab were found with PHASER (McCoy et al., 2005) in tetragonal space group P43212. The MR model was subjected to rigid body refinement and restrained all-atom refinement with simulated annealing with CNS (Brünger et al., 1998), and further improved by alternating cycles of model building with COOT (Emsley et al., 2010) and refinement with CNS and REFMAC5 (Winn et al., 2001).

Structure Validation and Analysis

The quality of the structure was verified with WHATCHECK (Hooft et al., 1996) and MOLPROBITY (overall clash score: 17.99, 95th percentile; Molprobity score: 2.43, 95th percentile) (Lovell et al., 2003). Root mean square displacements (r.m.s.d.) were determined with LSQMAN (Kleywegt, 1996). Hydrogen bonds were evaluated using HBPLUS (McDonald and Thornton, 1994), and van der Waals contacts were assigned with CONTACTSYM (Sheriff et al., 1987). Buried molecular surface areas were determined with MS (Connolly, 1993) with a 1.7 Å probe radius and standard van der Waals radii (Gelin and Karplus, 1979). Shape correlation (Sc) values were calculated with SC (Lawrence and Colman, 1993). Graphics were prepared with PYMOL (DeLano, 2002).

Surface Plasmon Resonance (SPR) Experiments

The pMT/BIP/V5-His A JAML expression vector (Verdino et al., 2010a) was used as a template to insert the BirA-tag sequence GGIFEAMKMELRD in the JAML stalk region between Leu258 and Asn259 with mPIPE cloning (Klock et al., 2008). The construct was transfected into SC2 cells, and BirA-tagged JAML was expressed and purified analogously as described for wt JAML (Verdino et al., 2010a). 2 mgs BirA-tagged JAML and a BirA-tagged control protein (peptide loaded MHC class II) in 50 mM Bicine pH 8.3 were biotinylated with 1/30 w/w BirA biotin protein ligase for 20 h at room temperature according the manufacturer’s protocol (Avidity LLC). Proteins were purified by size exclusion chromatography (Superdex 75 HR10/30) immediately prior to performing the SPR experiments. SPR measurements were performed on a BIAcore™ 2000 (GE Healthcare Inc.) at 25°C using 20 mM Tris, 150 mM NaCl pH 8.0 as running buffer. JAML and the control protein were captured in two different flow cells on a BIAcore™ SA sensor chip. Binding studies were performed for concentration series of CAR (0.146–300 µM) and HL4E10 Fab (0.122–2000 nM) at a flow rate of 30 µl/min with duplicates and in randomized order using BSA as control for nonspecific binding. For the JAML-CAR experiment, dissociation in buffer was used as regeneration method while, for the JAML-HL4E10 Fab experiment, regeneration was achieved through a 20 µl injection of 10 mM glycine pH 3.0.

Data were analyzed with the BIAevaluation software (BIAcore/GE Healthcare Inc.) and with Scrubber-2 (Biologic Software Inc.). Binding kinetics and affinity for the JAML-HL4E10 reaction were determined by employing a global, mono-exponential, curve fit of the standard 1:1 Langmuir binding isotherm to the data. The JAML-CAR reaction exhibited rapid kinetics and some extent of complex binding behavior. Thus, we analyzed the SPR data in several different ways. We determined KD from equilibrium binding data of the JAML-CAR reaction, employed a mono-exponential decay fit to the fast portion of the dissociation phase to yield Koff, and then derived Kon from the equation KD = Koff/Kon. We also applied simultaneous Kon/Koff global fitting procedures to the kinetics data. The standard, mono-exponential, 1:1 Langmuir binding isotherm did not yield a satisfying fit to the data, but a complex model (heterogeneous ligand) resulted in a much better interpretation. However, equilibrium analysis and kinetic analysis of JAML-CAR SPR data both gave similar binding parameter values and affinity constants on the same order of magnitude as those determined previously by analytical ultracentrifugation (Verdino et al., 2010b).

PI3K-Binding Assays

7–17 DETC were cultured in complete RPMI 1640 (Invitrogen) + 10% heat-inactivated FCS + 20 U/ml IL-2 at 37°C and 5% CO2. 5×106 cells were seeded into Petri dishes and incubated over night. On the next day, the cells were serum- and IL-2 starved for 4 h at 37°C prior to incubation with 0.5 ml of 10 µg/ml CAR-Fc fusion protein (Witherden et al., 2010) or HL4E10 IgG (Verdino et al., 2010b) in DPBS. After 5 min incubation at 37°C, the cells were washed three times with ice-cold PBS and lysed in lysis buffer [10 mM Tris pH 7.6, 150 mM NaCl, 1 mM EDTA, 0.5% NP-40, 1 mM Na3VO4, 1 mM Na2MoO4, 5 mM MgCl2, 1× Complete Protease Inhibitor Tablet (Roche)]. Full-length JAML was immunoprecipitated from lysate volumes corresponding to 200 µg total protein using 25 µl Protein A beads and 5 µg purified anti-JAML HL4E10 IgG for at least 3 hours at 4°C under agitation. Beads were washed 4 times with 0.5 ml lysis buffer, and 25 µl 2× reducing sample loading buffer was added. After gently vortexing and boiling for 5 min at 100°C, the supernatants were collected by centrifugation and proteins separated by SDS-PAGE. Proteins were blotted onto a nitrocellulose membrane, blocked with 10% goat serum, 3% BSA for 2 hours at room temperature, and probed over night at 4°C with rabbit anti-PI3K p85 mAb 19H8 (Cell Signaling Technology). HRP-conjugated goat anti-rabbit IgG (Southern Biotechnology) was used as secondary antibody and the signal was developed in Supersignal West Pico ECL (Pierce).

Cell Proliferation Assays

Purified mAb IgGs and Fabs were diluted in ELISA coating buffer (50 mM Tris, 150 mM NaCl; pH 8.0 at room temperature) and immobilized to individual wells of 96-well, flat-bottom, microtiter ELISA plates in a final volume of 100 µl. The plates were incubated at 4°C overnight. Before adding cells, the plates were washed twice with ELISA coating buffer and blocked for 15 min with 100 µl complete DMEM media (Invitrogen) supplemented with 10% heat-inactivated FCS. 7–17 DETC were cultured at 5×104 cells per well. Cells were pulsed with 1 µCi 3H-thymidine at 24 hr after initiation of culture and harvested 14 h later. Cells were harvested onto glass fiber filters (Cambridge Technology), and 2 ml scintillation fluid was added to each sample. Counts were read on a Beckman LS3801 scintillation counter (Beckman Coulter). All data points were performed in triplicate and are presented as mean ± SD.

Sequence Analysis

Protein sequence homologs of mouse JAML (Genbank entries: mouse, NM_001005421; rat, XM_236198; pig, AK238962; cattle, NM_001080250; dog, XM_848338; horse, XM_001502755; human, NM_153206; chimpanzee, XM_508787; rhesus monkey, XM_001093298; platypus, XM_001519842; gray short-tailed opossum, XM_001380707; red jungle fowl, XM_417915) and mouse CAR (Genbank entries: mouse, NM_009988; rat, NM_053570; chimpanzee, XM_531394; dog, XM_535566; human, AF124598; rabbit, EF034116; orangutan, NM_001134075; rhesus monkey, XM_001107115; horse, XM_001497763; pig, XM_001924554; gray short-tailed opossum, XM_001369028; platypus, XM_001519711; red jungle fowl, XM_416681; western clawed frosh, NM_001011084; African clawed frosh, NM_001112833; zebrafish, BC164203) were retrieved with tBLASTn (Altschul et al., 1990) and aligned and manually edited with BIOEDIT (Hall, 1999).

Highlights

  • Antibody HL4E10 emulates the natural ligand CAR in activity, but not in its interaction with JAML

  • HL4E10 antibody binds a conformationally flexible region in the membrane-proximal domain of JAML

  • The hydrophobic antibody-receptor interface confers nanomolar affinity and long half-life to the HL4E10-JAML complex

  • HL4E10-induced clustering of JAML recruits PI3K and induces γδ T-cell proliferation

  • HL4E10 antibody has favorable characteristics for therapeutic applications

Supplementary Material

01

Acknowledgements

We thank the staffs of the Advanced Light Source BL 8.2.1, 8.2.2, and 4.2.2, the Advanced Photon Source BL 23ID-D, and the Stanford Synchrotron Radiation Lightsource BL 11-1, for support with data collection and crystal screening; E. Landais for providing control protein for the SPR study; J. G. Luz, M. A. Adams, R. L. Stanfield, Jonathan R. Hart, and D. A. Shore for helpful discussions. This study was supported by NIH grants AI042266 (I.A.W), CA58896 (I.A.W.), AI064811 (W.L.H.), the Skaggs Institute, and the Erwin-Schrödinger Fellowship J2313 of the Austrian Science Fund (P.V.). This is publication 20050-MB from The Scripps Research Institute.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The authors declare no conflict of interest. The coordinates and structure factors of the JAML-HL4E10 complex have been deposited in the Protein Data Bank with accession number 3MJ9.

References

  1. Abken H, Hombach A, Heuser C, Kronfeld K, Seliger B. Tuning tumor-specific T-cell activation: a matter of co-stimulation? Trends Immunol. 2002;23:240–245. doi: 10.1016/s1471-4906(02)02180-4. [DOI] [PubMed] [Google Scholar]
  2. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J. Mol. Biol. 1990;215:403–410. doi: 10.1016/S0022-2836(05)80360-2. [DOI] [PubMed] [Google Scholar]
  3. Bergelson JM, Cunningham JA, Droguett G, Kurt-Jones EA, Krithivas A, Hong JS, Horwitz MS, Crowell RL, Finberg RW. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science. 1997;275:1320–1323. doi: 10.1126/science.275.5304.1320. [DOI] [PubMed] [Google Scholar]
  4. Boismenu R, Havran WL. γδ T cells in host defense and epithelial cell biology. Clin. Immunol. Immunopathol. 1998;86:121–133. doi: 10.1006/clin.1997.4468. [DOI] [PubMed] [Google Scholar]
  5. Brünger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, et al. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D. 1998;54:905–921. doi: 10.1107/s0907444998003254. [DOI] [PubMed] [Google Scholar]
  6. Carson SD. Receptor for the group B coxsackieviruses and adenoviruses: CAR. Rev. Med. Virol. 2001;11:219–226. doi: 10.1002/rmv.318. [DOI] [PubMed] [Google Scholar]
  7. Chambers CA, Kuhns MS, Egen JG, Allison JP. CTLA-4-mediated inhibition in regulation of T cell responses: mechanisms and manipulation in tumor immunotherapy. Annu. Rev. Immunol. 2001;19:565–594. doi: 10.1146/annurev.immunol.19.1.565. [DOI] [PubMed] [Google Scholar]
  8. Connolly ML. The molecular surface package. J. Mol. Graph. 1993;11:139–141. doi: 10.1016/0263-7855(93)87010-3. [DOI] [PubMed] [Google Scholar]
  9. Coyne CB, Bergelson JM. CAR: a virus receptor within the tight junction. Adv. Drug Deliv. Rev. 2005;57:869–882. doi: 10.1016/j.addr.2005.01.007. [DOI] [PubMed] [Google Scholar]
  10. Croft M. Co-stimulatory members of the TNFR family: keys to effective T-cell immunity? Nat. Rev. Immunol. 2003;3:609–620. doi: 10.1038/nri1148. [DOI] [PubMed] [Google Scholar]
  11. Davis SJ, van der Merwe PA. The kinetic-segregation model: TCR triggering and beyond. Nat. Immunol. 2006;7:803–809. doi: 10.1038/ni1369. [DOI] [PubMed] [Google Scholar]
  12. DeLano WL. The PyMOL Molecular Graphics System. San Carlos, CA, USA: DeLano Scientific; 2002. [Google Scholar]
  13. Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta Crystallogr D. 2010;66:486–501. doi: 10.1107/S0907444910007493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Evans EJ, Esnouf RM, Manso-Sancho R, Gilbert RJ, James JR, Yu C, Fennelly JA, Vowles C, Hanke T, Walse B, et al. Crystal structure of a soluble CD28-Fab complex. Nat. Immunol. 2005;6:271–279. doi: 10.1038/ni1170. [DOI] [PubMed] [Google Scholar]
  15. Freimuth P, Philipson L, Carson SD. The coxsackievirus and adenovirus receptor. Curr. Top. Microbiol. Immunol. 2008;323:67–87. doi: 10.1007/978-3-540-75546-3_4. [DOI] [PubMed] [Google Scholar]
  16. Gelin BR, Karplus M. Side-chain torsional potentials: effect of dipeptide, protein, and solvent environment. Biochemistry. 1979;18:1256–1268. doi: 10.1021/bi00574a022. [DOI] [PubMed] [Google Scholar]
  17. Guo YL, Bai R, Chen CX, Liu DQ, Liu Y, Zhang CY, Zen K. Role of Junctional Adhesion Molecule-Like protein in mediating monocyte transendothelial migration. Arterioscler. Thromb. Vasc. Biol. 2009;29:75–83. doi: 10.1161/ATVBAHA.108.177717. [DOI] [PubMed] [Google Scholar]
  18. Haas W, Pereira P, Tonegawa S. Gamma/Delta Cells. Annu. Rev. Immunol. 1993;11:637–685. doi: 10.1146/annurev.iy.11.040193.003225. [DOI] [PubMed] [Google Scholar]
  19. Hall TA. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Symp. Ser. 1999;41:95–98. [Google Scholar]
  20. Hayday AC. γδ T cells and the lymphoid stress-surveillance response. Immunity. 2009;31:184–196. doi: 10.1016/j.immuni.2009.08.006. [DOI] [PubMed] [Google Scholar]
  21. Hooft RW, Vriend G, Sander C, Abola EE. Errors in protein structures. Nature. 1996;381:272. doi: 10.1038/381272a0. [DOI] [PubMed] [Google Scholar]
  22. Howitt J, Anderson CW, Freimuth P. Adenovirus interaction with its cellular receptor CAR. Curr. Top. Microbiol. Immunol. 2003;272:331–364. doi: 10.1007/978-3-662-05597-7_11. [DOI] [PubMed] [Google Scholar]
  23. Jameson J, Ugarte K, Chen N, Yachi P, Fuchs E, Boismenu R, Havran WL. A role for skin γδ T cells in wound repair. Science. 2002;296:747–749. doi: 10.1126/science.1069639. [DOI] [PubMed] [Google Scholar]
  24. Jameson J, Witherden D, Havran WL. T-cell effector mechanisms: γδ and CD1d-restricted subsets. Curr. Opin. Immunol. 2003;15:349–353. doi: 10.1016/s0952-7915(03)00045-1. [DOI] [PubMed] [Google Scholar]
  25. Jameson JM, Sharp LL, Witherden DA, Havran WL. Regulation of skin cell homeostasis by γδ T cells. Front. Biosci. 2004;9:2640–2651. doi: 10.2741/1423. [DOI] [PubMed] [Google Scholar]
  26. Kleywegt GJ. Use of non-crystallographic symmetry in protein structure refinement. Acta Crystallogr. D. 1996;52:842–857. doi: 10.1107/S0907444995016477. [DOI] [PubMed] [Google Scholar]
  27. Klock HE, Koesema EJ, Knuth MW, Lesley SA. Combining the polymerase incomplete primer extension method for cloning and mutagenesis with microscreening to accelerate structural genomics efforts. Proteins. 2008;71:982–994. doi: 10.1002/prot.21786. [DOI] [PubMed] [Google Scholar]
  28. Kroczek RA, Mages HW, Hutloff A. Emerging paradigms of T-cell costimulation. Curr. Opin. Immunol. 2004;16:321–327. doi: 10.1016/j.coi.2004.03.002. [DOI] [PubMed] [Google Scholar]
  29. Lawrence MC, Colman PM. Shape complementarity at protein/protein interfaces. J. Mol. Biol. 1993;234:946–950. doi: 10.1006/jmbi.1993.1648. [DOI] [PubMed] [Google Scholar]
  30. Lemmon MA, Schlessinger J. Transmembrane signaling by receptor oligomerization. Methods Mol. Biol. 1998;84:49–71. doi: 10.1385/0-89603-488-7:49. [DOI] [PubMed] [Google Scholar]
  31. Lo Conte L, Chothia C, Janin J. The atomic structure of protein-protein recognition sites. J. Mol. Biol. 1999;285:2177–2198. doi: 10.1006/jmbi.1998.2439. [DOI] [PubMed] [Google Scholar]
  32. Lovell SC, Davis IW, Arendall WB, 3rd, de Bakker PI, Word JM, Prisant MG, Richardson JS, Richardson DC. Structure validation by Cα geometry: ϕ, ψ and Cβ deviation. Proteins. 2003;50:437–450. doi: 10.1002/prot.10286. [DOI] [PubMed] [Google Scholar]
  33. Luhder F, Huang Y, Dennehy KM, Guntermann C, Muller I, Winkler E, Kerkau T, Ikemizu S, Davis SJ, Hanke T, et al. Topological requirements and signaling properties of T cell-activating, anti-CD28 antibody superagonists. J. Exp. Med. 2003;197:955–966. doi: 10.1084/jem.20021024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Luissint AC, Lutz PG, Calderwood DA, Couraud PO, Bourdoulous S. JAM-L-mediated leukocyte adhesion to endothelial cells is regulated in cis by α4β1 integrin activation. J. Cell Biol. 2008;183:1159–1173. doi: 10.1083/jcb.200805061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Martin-Orozco N, Dong C. New battlefields for costimulation. J. Exp. Med. 2006;203:817–820. doi: 10.1084/jem.20060219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McCoy AJ, Grosse-Kunstleve RW, Storoni LC, Read RJ. Likelihood-enhanced fast translation functions. Acta Crystallogr. D. 2005;61:458–464. doi: 10.1107/S0907444905001617. [DOI] [PubMed] [Google Scholar]
  37. McDonald IK, Thornton JM. Satisfying hydrogen bonding potential in proteins. J. Mol. Biol. 1994;238:777–793. doi: 10.1006/jmbi.1994.1334. [DOI] [PubMed] [Google Scholar]
  38. Moog-Lutz C, Cave-Riant F, Guibal FC, Breau MA, Di Gioia Y, Couraud PO, Cayre YE, Bourdoulous S, Lutz PG. JAML, a novel protein with characteristics of a junctional adhesion molecule, is induced during differentiation of myeloid leukemia cells. Blood. 2003;102:3371–3378. doi: 10.1182/blood-2002-11-3462. [DOI] [PubMed] [Google Scholar]
  39. Otwinowski Z, Minor W. Processing of x-ray diffraction data collected in oscillation mode. Methods Enzymol. 1997;276:307–326. doi: 10.1016/S0076-6879(97)76066-X. [DOI] [PubMed] [Google Scholar]
  40. Salomon B, Bluestone JA. Complexities of CD28/B7: CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu. Rev. Immunol. 2001;19:225–252. doi: 10.1146/annurev.immunol.19.1.225. [DOI] [PubMed] [Google Scholar]
  41. Schlessinger J. Cell signaling by receptor tyrosine kinases. Cell. 2000;103:211–225. doi: 10.1016/s0092-8674(00)00114-8. [DOI] [PubMed] [Google Scholar]
  42. Sharp LL, Jameson JM, Cauvi G, Havran WL. Dendritic epidermal T cells regulate skin homeostasis through local production of insulin-like growth factor 1. Nat. Immunol. 2005;6:73–79. doi: 10.1038/ni1152. [DOI] [PubMed] [Google Scholar]
  43. Sharpe AH, Freeman GJ. The B7-CD28 superfamily. Nat. Rev. Immunol. 2002;2:116–126. doi: 10.1038/nri727. [DOI] [PubMed] [Google Scholar]
  44. Sheriff S, Hendrickson WA, Smith JL. Structure of myohemerythrin in the azidomet state at 1.7/1.3 Å resolution. J. Mol. Biol. 1987;197:273–296. doi: 10.1016/0022-2836(87)90124-0. [DOI] [PubMed] [Google Scholar]
  45. Shires J, Theodoridis E, Hayday AC. Biological insights into TCRγδ+ and TCRαβ+ intraepithelial lymphocytes provided by serial analysis of gene expression (SAGE) Immunity. 2001;15:419–434. doi: 10.1016/s1074-7613(01)00192-3. [DOI] [PubMed] [Google Scholar]
  46. Snanoudj R, de Preneuf H, Creput C, Arzouk N, Deroure B, Beaudreuil S, Durrbach A, Charpentier B. Costimulation blockade and its possible future use in clinical transplantation. Transpl. Int. 2006;19:693–704. doi: 10.1111/j.1432-2277.2006.00332.x. [DOI] [PubMed] [Google Scholar]
  47. Stuart RW, Racke MK. Targeting T cell costimulation in autoimmune disease. Expert Opin. Ther. Targets. 2002;6:275–289. doi: 10.1517/14728222.6.3.275. [DOI] [PubMed] [Google Scholar]
  48. Suntharalingam G, Perry MR, Ward S, Brett SJ, Castello-Cortes A, Brunner MD, Panoskaltsis N. Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N. Engl. J. Med. 2006;355:1018–1028. doi: 10.1056/NEJMoa063842. [DOI] [PubMed] [Google Scholar]
  49. van der Merwe PA, Brown MH, Davis SJ, Barclay AN. Affinity and kinetic analysis of the interaction of the cell adhesion molecules rat CD2 and CD48. EMBO J. 1993;12:4945–4954. doi: 10.1002/j.1460-2075.1993.tb06188.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Verdino P, Witherden DA, Havran WL, Wilson IA. The molecular interaction of CAR and JAML recruits the central cell signal transducer PI3K. Science. 2010a;329:1210–1214. doi: 10.1126/science.1187996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Verdino P, Witherden DA, Podshivalova K, Rieder SE, Havran W, Wilson IA. cDNA sequence and Fab crystal structure of HL4E10, a hamster IgG lambda light chain antibody stimulatory for γδ T cells. 2010b doi: 10.1371/journal.pone.0019828. Submitted. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Vincenti F, Luggen M. T cell costimulation: a rational target in the therapeutic armamentarium for autoimmune diseases and transplantation. Annu. Rev. Med. 2007;58:347–358. doi: 10.1146/annurev.med.58.080205.154004. [DOI] [PubMed] [Google Scholar]
  53. Waibler Z, Sender LY, Merten C, Hartig R, Kliche S, Gunzer M, Reichardt P, Kalinke U, Schraven B. Signaling signatures and functional properties of anti-human CD28 superagonistic antibodies. PLoS ONE. 2008;3:e1708. doi: 10.1371/journal.pone.0001708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Watts TH. TNF/TNFR family members in costimulation of T cell responses. Annu. Rev. Immunol. 2005;23:23–68. doi: 10.1146/annurev.immunol.23.021704.115839. [DOI] [PubMed] [Google Scholar]
  55. Weaver TA, Charafeddine AH, Kirk AD. Costimulation blockade: towards clinical application. Front. Biosci. 2008;13:2120–2139. doi: 10.2741/2829. [DOI] [PubMed] [Google Scholar]
  56. Whang MI, Guerra N, Raulet DH. Costimulation of dendritic epidermal γδ T cells by a new NKG2D ligand expressed specifically in the skin. J. Immunol. 2009;182:4557–4564. doi: 10.4049/jimmunol.0802439. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Winn MD, Isupov MN, Murshudov GN. Use of TLS parameters to model anisotropic displacements in macromolecular refinement. Acta Crystallogr. D. 2001;57:122–133. doi: 10.1107/s0907444900014736. [DOI] [PubMed] [Google Scholar]
  58. Witherden DA, Verdino P, Rieder SE, Garijo O, Mills RE, Teyton L, Fischer WH, Wilson IA, Havran WL. The junctional adhesion molecule JAML is a costimulatory receptor for epithelial γδ T cell activation. Science. 2010;329:1205–1210. doi: 10.1126/science.1192698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zang X, Allison JP. The B7 family and cancer therapy: costimulation and coinhibition. Clin. Cancer Res. 2007;13:5271–5279. doi: 10.1158/1078-0432.CCR-07-1030. [DOI] [PubMed] [Google Scholar]
  60. Zen K, Liu Y, McCall IC, Wu T, Lee W, Babbin BA, Nusrat A, Parkos CA. Neutrophil migration across tight junctions is mediated by adhesive interactions between epithelial coxsackie and adenovirus receptor and a junctional adhesion molecule-like protein on neutrophils. Mol. Biol. Cell. 2005;16:2694–2703. doi: 10.1091/mbc.E05-01-0036. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS253667-supplement-01.doc (9.5MB, doc)

RESOURCES