| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Microbiology, Immunology, and Molecular Genetics, University of California School of Medicine, Los Angeles, CA 90095;
Department of Molecular Biology, Skaggs Institute for Chemical Biology, La Jolla, CA 92037;
Division of Rheumatology, Immunology, and Allergy, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 02115;
§
La Jolla Institute for Allergy and Immunology, San Diego, CA 92121;
¶
Department of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, NY 10461; and
||
Molecular Biology Institute and Division of Dermatology, University of California School of Medicine, Los Angeles, CA 90095
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CD1 and MHC molecules share both amino acid homology and structural
similarity. In addition, both families of molecules possess Ag-binding
grooves composed of two long anti-parallel 
-helices placed over
a 
-sheet platform (2, 3, 4). Furthermore, like MHC I
molecules, CD1 forms heterodimers with
2-microgolobulin. However, the existence of
significant structural differences between the CD1 and the MHC I
Ag-binding grooves results in the selection of different classes of
ligands by these proteins. Unlike the MHC I peptide-binding groove,
which is lined with six pockets designated A-F, the putative CD1
Ag-binding groove, based on the crystal structure of mouse CD1d1, is
formed by two large pockets referred to as the A' and F' pockets
(2). Secondly, the two CD1 -helices are both closer
together and further from the -sheet platform making the CD1 groove
narrower and deeper than that of MHC I. Finally, unlike the MHC I and
II grooves which contain conserved polar amino acid residues that
interact with their respective antigenic peptides through hydrogen
bonding, the CD1 groove is almost entirely made up of nonpolar residues
that would make similar hydrogen bonding interactions with Ags
impossible. Such differences collectively make CD1 more appropriately
suited to present hydrophobic Ags.
Consistent with this structural model, studies of the human CD1b protein have revealed that this isoform of the CD1 family presents at least four lipid-containing Ags. The Ags described all possess the structural motif of two hydrophobic alkyl chains connected to a polar structure, which can be composed of carbohydrate moieties. These Ags include the mycolic acids (MA)6 (5), the phosphatidylinositol mannans (including lipoarabinomannan (LAM)) (6), the glycosylated MA derivatives (such as glucose monomycolate (GMM)) (7), and the gangliosides (8).
In this study, we investigated the role of the A' and F' pockets of human CD1b in the presentation of lipid Ags to T cells. Previous site-directed mutagenesis studies of lipid-interacting proteins mapped amino acids and domains critical to protein structure and function (9, 10, 11). With guidance from the crystal structure of mouse CD1d1 and the high degree of homology between mouse CD1 and human CD1 proteins, we performed alanine-scanning mutagenesis of the CD1b Ag-binding groove to identify amino acid residues critical to the presentation of lipid Ags. We compared the presentation of natural Mycobacterium phlei GMM and synthetic GMM (sGMM) because these two lipid Ags differ in the length of their hydrocarbon chains. Specifically, GMM possesses two hydrocarbon chains where the longer of the two can vary in length from C54-C62 and the shorter chain varies from C22-C24 (12). The sGMM molecule contains two hydrocarbon chains of C18 and C14 (7). We found residues from both the A' and F' pockets to be critical in the presentation of GMM and sGMM to the CD1b-restricted GMM-specific T cell line, LDN5. More importantly, the presentation of the two Ags was dependent on different subsets of residues located in distinct regions of the putative Ag-binding groove. These data indicate that both the A' and F' pockets of human CD1b are required for optimal presentation of lipid Ags to T cells and provide a mechanism by which lipid Ags of differing hydrocarbon tail length are accommodated and presented to T cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The cervical carcinoma HeLa cell line (American Type Culture Collection, Manassas, VA) was used as the APC type throughout these studies. Unless otherwise stated, these cells were cultured in DMEM/10% FCS (Life Technologies, Rockville, MD). The CD1b-transfected HeLa cell line HeLa-CD1b was maintained in the same media as the parent HeLa line with the addition of 500 µg/ml of G418 (Life Technologies). T cell assays were performed using the CD1b-restricted, GMM-specific T cell line, LDN5, which was maintained in RPMI 1640/8% FCS/2% human serum/1 nM IL-2 (Life Technologies) and stimulated with CD1+ APCs and M. phlei extract on a biweekly basis as described previously (7). The glycolipid Ags GMM and sGMM were purified as described previously (7).
Energy-minimized model of human CD1b
The homology model for human CD1b was constructed from the crystal structure of mouse CD1d1 (pdb code 1CD1) with which it shares 56% sequence identity. The amino acid sequence of mCD1d1 was replaced with that of human CD1b using the homology module of the program InsightII (Molecular Simulations, San Diego, CA). Hydrogen atoms were added, and amino acid functional groups were ionized corresponding to pH 7.0. Atomic partial charges and potentials for amino acids were taken from consistent valence force field as implemented in InsightII software (13). All energy minimization was performed using Discover program version 3.1 installed with InsightII interface as implemented on a Silicon Graphics workstation (Silicon Graphics, Mountain View, CA). Energy minimization was performed as a two-step procedure. The structure was optimized by 200 cycles of steepest descent followed by conjugate gradient minimization until the convergence was reached. All structural images were generated by InsightII software and processed using Adobe Photoshop (Adobe Systems, San Jose, CA).
Mutagenesis strategy
A 
1.4-kb XbaI fragment containing the entire CD1b
cDNA was subcloned from the eukaryotic expression construct
pSR-NEO-CD1b (14) into pBluescript
KS+ (Stratagene, San Diego, CA) for mutagenesis.
Mutagenesis was conducted per the manufacturer’s suggestions using the
QuickChange site-directed mutagenesis kit (Stratagene). Briefly, two
complementary 27-mer oligonucleotides containing an alanine
substitution in the middle codon (i.e., from the native sequence to
GCA, GCC, GCG, or GCT), were annealed onto 50 ng of the pBS-CD1b
template for 1 min at 55°C after denaturation and at 95°C for 1
min. The oligonucleotides were then extended at 68°C for 10 min using
the high fidelity, thermostable polymerase PFU (Stratagene). This
process was repeated for 18 cycles and the reaction product was
digested with Dpn I at 37°C for 30 min to remove template DNA. The
reaction product was transformed into DH5 competent cells. Plasmid
DNA obtained from resultant colonies (Qiagen, Valencia, CA) was
sequenced using the ABI Prism (PE Biosystems, Foster City, CA)
sequencing technology to verify mutagenesis. Once identified, the 5'
645-bp EcoRI fragment of each mutant construct containing
the mutation was completely sequenced with the exception of five
mutants that are outside (i.e., 3') of the EcoRI site (e.g.,
L161A, L162A, T165A, C166A, and Y169A). These fragments were religated
onto an EcoRI digested pBluescript construct encoding the
wild-type 3' portion of the CD1b gene. The five mutants located outside
of the internal EcoRI site were digested with
ClaI and BglII to generate a 726-bp fragment,
which was then religated onto a BglII/ClaI vector
containing the remaining 3' portion of the wild-type CD1b cDNA and
sequenced. Restriction analyses as well as directional PCR were used to
confirm all clones. This process eliminated 800 bp of sequencing per
clone. Once completed, the XbaI fragment containing each
complete mutant CD1b cDNA was religated onto the parental
XbaI digested pSR-NEO vector. Ligation orientation was
confirmed by restriction analysis as well as directional PCR.
Transfection of HeLa cells
Before transfection, 6 x 105 cells were plated onto 60-mm plates (Corning Glass, Corning, NY) and allowed to attach overnight in DMEM/10% FCS. In preparation for transfection, each plate was washed with 3 ml DMEM. Unless otherwise stated, all mutants and wild-type transfections were performed by incubating 5 µg of plasmid DNA (Qiagen) with 30 µl of the cationic lipid Superfect (Qiagen) for 5 min in 150 µl total volume of incomplete DMEM (without antibiotics). The DNA/lipid conjugates were then resuspended in 1 ml DMEM/10% FCS and plated onto the washed plates after the wash media was removed. Forty-eight hours after transfection, the transfectants were washed and detached in 1xPBS/1 mM EDTA. Transfectants were then washed, counted, and analyzed for surface expression and Ag presentation capability.
Flow cytometry analysis of CD1b transfectants
After harvesting, two samples of 150,000 cells of each transfectant were incubated on ice in 100 µl of 1x PBS/2% FCS for 15 min before addition of Ab. Each mutant was then incubated with isotype control mouse IgG1 Ab (CalTag, Burlingame, CA) and anti-CD1b Ab BCD1b3 (15), washed, and incubated with goat anti-mouse F(ab')2-FITC-conjugated Ab (Caltag). Each sample was then washed, fixed, and analyzed using a FACScan apparatus (Becton Dickinson, Franklin Lakes, NJ) and WINmidi 2.6 software (shareware package, Joseph Troffer, The Scripps Institute, La Jolla, CA). Additional surface staining was performed using the anti-CD1b Ab MT-101 (Serotec, Oxford, U.K.) as the primary Ab. All transfections used at least three independent preparations of plasmid DNA and were conducted at least three independent times.
T cell assays and cytokine ELISA
Unless otherwise stated, T cell assays were performed using
3 x 104 HeLa transfectants and 2 x
104 LDN5 in 200 µl RPMI 1640/8% FCS/2% human
serum. After 18 h, 20 µl of supernatant were removed and
analyzed by ELISA for IFN-
as per the manufacturer’s suggestions
(Endogen, Woburn, MA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The CD1 proteins are a family of Ag-presentation molecules that
demonstrate modest amino acid sequence homology with MHC I
(1). In 1997, the crystal structure of mouse CD1d1
revealed close structural similarities between CD1 and MHC I (Fig. 1
) (2). Like MHC I, CD1
possesses an Ag-binding groove formed by two anti-parallel
-helices positioned over a platform of -sheets. However, unlike
MHC I, the identities of the CD1 residues critical for Ag presentation
have not been determined. Mouse CD1d1 and human CD1b demonstrate
significant amino acid homology, particularly in the residues that form
the Ag-binding groove (45% identity). An energy-minimized homology
model of human CD1b was generated based on the mouse CD1d1 structure,
which conserved both the A' and F' pockets (Fig. 1a).
Comparison of the human CD1b model with the known structure of mCD1d1
revealed close structural similarity, particularly within the region of
the -helices and -sheets that form the putative Ag-binding groove
(Fig. 1b). Based on this model, 36 residues in the putative
CD1b Ag-binding groove corresponding to the residues that line the A'
and F' pockets of the mouse CD1d1 Ag-binding groove were identified.
These residues were mutated by alanine substitution and analyzed to
identify residues that participate in CD1b Ag presentation to T cells
(Table I).
|
|
The cervical carcinoma-derived HeLa cell line was chosen to serve
as APCs because HeLa cells are devoid of CD1b expression, are readily
transfected with plasmid DNA, and can present Ag via CD1b once
transfected (S. A. Porcelli, unpublished observations). Transient CD1b
transfectants were comparable to stable CD1b transfectants in their
ability to present purified GMM or M. phlei soluble extract
(an alternative source of GMM) to the GMM-specific, CD1b-restricted T
cell line LDN5 (Fig. 2, a and
b). Thus, all subsequent assays were performed using the
transient transfection system as it facilitated the assessment of the
functional capacity of a large number of CD1b mutants.
|
). To evaluate the effects of CD1b
groove mutations on presentation of these Ags, Ag concentrations that
resulted in half-maximal activation of LDN5 by 3 µg/ml synthetic and
1 µg/ml M. phlei GMM were determined (data not shown).
|
Thirty-six CD1b mutants were individually transfected into HeLa
cells and analyzed for surface expression using the BCD1b3
mAb. The mutants displayed a wide spectrum of expression
profiles, a representative panel of which is depicted in Fig. 4. Of the mutants tested, 16 demonstrated
expression levels >85% of wild-type CD1b mean fluorescence intensity
(MFI) and were considered minimally affected by alanine substitution
(Table II). A total of 20 mutants
expressed CD1b at levels 
85% wild-type CD1b MFI. To investigate the
possibility that these decreases were a direct result of the loss of
the BCD1b3-binding epitope through mutagenesis, these mutants were
further analyzed using the anti-CD1b Ab MT-101. The BCD1b3 and
MT-101 Abs recognize independent epitopes in Ab competition experiments
(data not shown) (16, 17). All mutants, with the exception
of R140A, exhibited comparable levels of expression using either Ab
suggesting that the decreased surface expression patterns observed for
these mutants were not likely to be due to the loss of the BCD1b3
epitope. The R140A mutant exhibited significant decreases in cell
surface expression (43.6% wild-type CD1b MFI) using the BCD1b3 Ab but
was found to be expressed at greater than wild-type CD1b levels (127%
wild-type CD1b MFI) using the MT-101 Ab, indicating that the BCD1b3
epitope includes R140. In the case of the three mutated cysteine
residues (C131, C145, and C166), the observed loss of cell surface
expression is probably due to the abrogation of crucial disulfide
bonds between C131 and C145, as well as between C166 and the nongroove
residue C102. The mechanisms behind the observed decreased cell surface
expression for the remaining mutants remain unresolved.
|
|
To determine the functional significance of decreased CD1b cell
surface expression on Ag presentation, the amount of input wild-type
CD1b plasmid DNA was titrated to modulate the level of CD1b expressed
on the transfectants. The minimum amount of conformationally active
CD1b expression necessary to produce an optimal LDN5 T cell response
was 63% of the MFI (achieved using 0.75 µg of plasmid DNA)
observed under the maximum wild-type DNA concentration of 5
µg/transfection (Fig. 5). Simultaneous
analyses of two mutants, W40A and L161A, demonstrated the
reproducibility of our cell surface expression and Ag presentation data
(data not shown). Decreasing the amount of input wild-type CD1b DNA
from 5.0 to 0.75 µg decreased the MFI of the transfectants from
148.9 to 95.2 MFI. More importantly, the percentage of cells
transfected decreased from 55% to 16% in these experiments
(data not shown). Analyses of the alanine-substitution mutants
demonstrated that although alanine substitution decreased the CD1b
expression level in many of the mutants assayed, it did not
significantly affect the number of cells transfected as observed in the
DNA titration experiments described in Fig. 5. Therefore, we concluded
that mutants expressing 
63% wild-type MFI possessed sufficient cell
surface CD1b expression to maximally stimulate T cells and assayed the
Ag presentation capabilities of only those mutants that exhibited CD1b
expression levels >63% wild-type CD1b MFI.
|
Of the 27 mutations meeting the experimentally determined surface
expression criteria (i.e., 63% wild-type MFI), six presented both
GMM or sGMM at >70% the Ag presentation capacity of wild-type CD1b as
determined by T cell release of IFN- (Fig. 6, a and
b). Thus, it is unlikely that these residues play a major
role in CD1b Ag presentation. Eight CD1b mutants presented both GMM or
sGMM at levels 70% of wild-type CD1b (Fig. 6, d and
e). None of these 14 residues demonstrated noteworthy Ag
presentation preferences between the two Ags as revealed by the ratios
of the relative presentations of GMM vs sGMM by these mutants (Fig. 6, c and f).
|
0.6667 (Fig. 7, a–c). The remaining nine mutants were preferentially
affected in the presentation of sGMM yielding GMM/sGMM relative
presentation ratios 1.5 (Fig. 7, d–f). T157 contributes
atoms to both the outer and inner surface of the CD1b groove in our
model and thus its mutation could potentially inhibit presentation
through the loss of either a TCR contact residue and/or a critical Ag
interaction residue. We find the second possibility more likely as the
T157A presentation of GMM was unaffected in our study while the
presentation of sGMM to the same TCR was significantly inhibited.
|
, a–c. The residues affecting presentation of both
GMM and sGMM are located in both the A' and F' pockets. It is
interesting to note that the residues that cause a specific inhibition
in the presentation of GMM are located deep inside the A' pocket while
the residues that specifically affect the presentation of sGMM are
located between the center of the groove and the F' pocket. Thus, our
data would indicate that distinct residues in the A' and F' pockets are
involved in the presentation of lipid Ags with different
hydrocarbon chain lengths.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The expression of MHC I on the cell surface requires stable interaction
with bound peptides within the Ag-binding groove (18, 19).
In the experiments presented here, we were able to significantly
inhibit expression of CD1b by alanine substitution of a subset of the
residues that line its Ag-binding groove (i.e., F10, V12, L66, F77,
G98, C131, C145, I154, and C166). The mechanism behind the observed
decreases in CD1b surface expression for the six noncysteine residues
(i.e., F10, V12, L66, F77, G98, and I154) remains unresolved. One
explanation is that these residues are required for interaction with a
natural ligand (20). Alternatively, mutagenesis of the
putative human CD1b groove residues may affect association of CD1b with
2-microglobulin, which is required for CD1b
traffic to the cell surface (21). The mCD1d1 protein
appears to use amino acids located within the
3 domain and residues 119–124 of the
2 domain to bind
2-microgloblin. Based on the current molecular
model of CD1b, we predict that CD1b associates with
2-microgloblin by a mechanism generally
similar to that of mCD1d1 although the analogous loop formed by
residues 119–124 in mCD1d1 is shortened in human CD1b (S.K. and
I.A.W., unpublished observations). Regardless, we find the possibility
that the mutants listed above affect
2-microgloblin association unlikely as these
mutations are located outside of the areas that are believed to be
involved in CD1/2-microgloblin interaction. We
cannot rule out additional potential mechanisms for decreased cell
surface expression of these mutants such as decreases in RNA/protein
stability, changes in CD1b structure, and/or changes in intracellular
traffic as a result of mutagenesis.
It has become evident that MHC I binds peptides through the formation of hydrogen bonds between the pockets of MHC I and the amino acid sidechains of bound peptides (18, 19, 22, 23, 24, 25, 26, 27, 28). Additional stabilizing interactions occur between the charged peptide N and C termini and the ends of the MHC I groove. Similar studies of MHC II have demonstrated that the MHC II sidechains and backbone form hydrogen bonds along the entire length of the peptide in addition to specific interactions between peptide amino acid sidechains and pockets within the MHC II Ag-binding groove (29, 30, 31, 32, 33). In the current study, we found the substitution of eight of the hydrophobic residues within the CD1b Ag-binding groove (i.e., W40, I74, F88, I96, L114, F123, L124, and L162) to decrease lipid Ag presentation function for both GMM and sGMM. As these residues are all hydrophobic, this implicates hydrophobic interactions to be critical for the presentation of lipid Ags by CD1b. Similar to previous mutagenesis studies of lipid-binding proteins, we observed that single amino acid substitutions significantly altered both protein structure and function (9, 10, 11).
In addition, as GMM and sGMM have considerably different hydrocarbon chain lengths for their diacyl chains, the aforementioned eight hydorphobic residues may be involved in the presentation of lipid Ags containing a variety of hydrocarbon chain lengths. Interestingly, six of these eight amino acid residues demonstrate evolutionary conservation between mouse CD1d1 and all four human CD1 isoforms (a-d), being either identical or substituted with an amino acid of similar hydrophobic properties. The conservation of these six amino acids combined with their role in presentation of both GMM and sGMM would suggest that these residues are critical to the presentation of a variety of lipid Ags by the CD1 proteins. However, we cannot rule out the possibility of the introduction of more general conformational changes in the CD1b protein that directly affect TCR interactions or inhibit Ag loading of CD1b through other mechanisms.
It is tempting to speculate about the conformation with which these lipid Ags interact with the A' and F' pockets of CD1b, yet our data do not differentiate between two different models. In one model of CD1b/GMM interaction, the larger of the two hydrocarbon chains of GMM (ranging in length from C54-C62) would interact with the larger A' pocket, thereby placing the shorter hydrocarbon chain (C22-C24) in the F' pocket. Consistent with this hypothesis, GMM, which possesses the longest hydrocarbon chain of the four present in GMM and sGMM (C18:C14), appears to specifically use the distal A' pocket for optimal presentation. However, based on the model of human CD1b, the CD1b A' pocket is unlikely to be large enough to accommodate such a large ligand in the absence of further modification. The alternative explanation is also plausible, that the smaller chain of GMM inserts into the A' pocket while the larger chain extends across the F' pocket and possibly protrudes outside of the binding groove of the F' pocket (34, 35). Even though the longer chain of sGMM is smaller than either chain of GMM, the orientation of GMM and sGMM would need to be identical to permit T cell recognition by the same clone. A more detailed analysis of additional synthetic GMM molecules containing more subtle differences in their alkyl chains and/or further crystallographic analysis of GMM Ags bound to CD1b should resolve the correct orientation of these lipid Ags within the CD1b Ag-binding groove.
Overall, our data indicate that both the A' and F' pockets of CD1b are required for the presentation of GMM to T cells. Consistent with this hypothesis, analyses of the Ags presented by CD1b indicate that T cell specificity is conferred by the biochemical structure of the polar region for LAM, GMM, and the ganglioside GM1 (6, 7, 8). In contrast, the lipid portions of the ligands LAM and GMM were found to be essential for Ag presentation and presumed to be involved in binding to CD1b. Surface plasmon resonance studies of CD1b-lipid interaction further support this hypothesis as they provide direct evidence for lipid-CD1 interaction as being dependent on the lipid portion of the Ags studied (36). The demonstration that both the A' and F' pockets are required for optimal presentation provides a mechanism by which human CD1b can present a variety of lipid Ags to T cells in the immune response to microbial pathogens.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Amgen, Amgen Center, Thousand Oaks, CA 91320.
3 Current address: Structural Biology Laboratory, Sincrotrone Trieste Elettra, 34012 Basovizza (TS), Italy.
4 Current address: Merck and Co., Inc., One Merck Drive, P.O. Box 100, WS3AB-40, Whitehouse Station, NJ 08889.
5 Address correspondence and reprint requests to Dr. Robert L. Modlin, 52-121 CHS, University of California School of Medicine, 10833 Le Conte Avenue, Los Angeles, CA 90095.
6 Abbreviations used in this paper: MA, mycolic acids; LAM, lipoarabinomannan; GMM, glucose monomycolate; sGMM, synthetic GMM; MFI, mean fluorescence intensity; hcD1b, human CD1b; mCD1d1, mouse CD1d1.
Received for publication August 14, 2000. Accepted for publication December 5, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
+ T cells. Nature 372:691.[Medline]
RIII (CD16). Clin. Immunol. Immunopathol. 62:16.[Medline]
This article has been cited by other articles:
|
|
 |
|
  Y. Fujita, T. Naka, M. R. McNeil, and I. Yano Intact molecular characterization of cord factor (trehalose 6,6'-dimycolate) from nine species of mycobacteria by MALDI-TOF mass spectrometry Microbiology, October 1, 2005; 151(10): 3403 - 3416. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Fujita, T. Naka, T. Doi, and I. Yano Direct molecular mass determination of trehalose monomycolate from 11 species of mycobacteria by MALDI-TOF mass spectrometry Microbiology, May 1, 2005; 151(5): 1443 - 1452. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Batuwangala, D. Shepherd, S. D. Gadola, K. J. C. Gibson, N. R. Zaccai, A. R. Fersht, G. S. Besra, V. Cerundolo, and E. Y. Jones The Crystal Structure of Human CD1b with a Bound Bacterial Glycolipid J. Immunol., February 15, 2004; 172(4): 2382 - 2388. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Im, K. O. A. Yu, P. A. Illarionov, K. P. LeClair, J. R. Storey, M. W. Kennedy, G. S. Besra, and S. A. Porcelli Direct Measurement of Antigen Binding Properties of CD1 Proteins Using Fluorescent Lipid Probes J. Biol. Chem., January 2, 2004; 279(1): 299 - 310. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Rauch, J. Gumperz, C. Robinson, M. Skold, C. Roy, D. C. Young, M. Lafleur, D. B. Moody, M. B. Brenner, C. E. Costello, et al. Structural Features of the Acyl Chain Determine Self-phospholipid Antigen Recognition by a CD1d-restricted Invariant NKT (iNKT) Cell J. Biol. Chem., November 28, 2003; 278(48): 47508 - 47515. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. P. Grant, E. M. Beckman, S. M. Behar, M. Degano, D. Frederique, G. S. Besra, I. A. Wilson, S. A. Porcelli, S. T. Furlong, and M. B. Brenner Fine Specificity of TCR Complementarity-Determining Region Residues and Lipid Antigen Hydrophilic Moieties in the Recognition of a CD1-Lipid Complex J. Immunol., April 15, 2002; 168(8): 3933 - 3940. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D. De Silva, J.-J. Park, N. Matsuki, A. K. Stanic, R. R. Brutkiewicz, M. E. Medof, and S. Joyce Lipid Protein Interactions: The Assembly of CD1d1 with Cellular Phospholipids Occurs in the Endoplasmic Reticulum J. Immunol., January 15, 2002; 168(2): 723 - 733. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
