| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of
*
Pathology and Immunology and
Chemistry, Washington University School of Medicine, St. Louis, MO 63110
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mAb 40F (8), which recognizes the class II I-Ak molecule, was used in flow cytometry, cell binding studies, and immunoaffinity chromatography. The mAb anti-HEL Ab, F10.6.6 (9), and a rabbit anti HEL antisera were used in flow cytometry and cell binding experiments. F10.6.6 is a conformation-dependent mAb that recognizes an epitope around the Arg68 residue (10). The AW3 mAb, generated in this laboratory (6), was used in cell binding studies for quantitation of the HEL48–63 peptide-class II I-Ak complex. The mAbs VAL-3 and 151.48-61 were used in the peptide immunoaffinity capture technique. VAL-3 recognizes the HEL fragment 31–47, specifically its FESNF sequence (7). The 151.48-61 Ab is specific for the HEL48–62 peptide. This Ab was generated by repeated s.c. injections of female CBA mice with 200 µg of the 48–61 peptide coupled to BSA by bromoacetyl succinimide (Sigma, St. Louis, MO). mAbs were purified from ascites by using protein A-Sepharose (Sigma). The ascites was generated by the injection of hybridoma cells into pristane (Sigma)-treated SCID mice. Anti-HEL rabbit polysera was purified by DEAE chromatography.
Cell lines and flow cytometric analysis
The murine B cell lymphoma line M12.C3.F6 expressing class II I-Ak (10) was transfected with a gene encoding a membrane form of HEL, as previously described (3). Two lines were examined containing I-Ak: M12-Ak mHEL-20 (referred to as mHEL-20) and M12-Ak mHEL-242 (referred to as mHEL-242; see below). The fusion protein was constructed by joining in-frame the entire HEL gene to a segment of the MHC class I Ld gene. To generate mutated HEL genes, site-directed mutagenesis was performed using the same plasmid. The mHEL(K52; K55) line was derived by transfecting M12.C3F6 cells with the mutated HEL cDNA consisting of a substitution of aspartic acid, at position 52 (the main anchor position), and isoleucine, at position 55, for lysine residues. Three other mutant cell lines consisted of a single amino acid substitution at position 46, 47, or 48 of the HEL cDNA for a proline residue. These were the same lines used in Ref. 3 . All cell lines were cultured in DMEM supplemented with 5% FCS.
The transfected cell lines were sorted for high class II I-Ak and membrane HEL expression. Cells were stained with 40F conjugated to FITC and biotinylated F10.6.6 for 40 min at 4°C. After two washes with chilled DMEM supplemented with 5% FCS, the cells were resuspended in 100 µl of R-PE-streptavidin (Molecular Probes, Eugene, OR) and incubated at 4°C for 40 min. The cells were subsequently washed twice and resuspended in medium. At least 20,000 cells were collected using the Becton Dickinson cell sorter (Mountain View, CA). Unstable cell lines were, in turn, subcloned into a 96-well flat-bottom plate at a frequency of one cell per well. Single colonies were screened for high surface expression using the FACScalibur (Becton Dickinson). Binding studies on B cell lymphoma lines were performed as previously described (6). Using the conformation-dependent Ab F10.6.6, we found that the mHEL-242 and mHEL(K52; K55) had the same amount of reactive material, calculated to be about 6,000 molecules/cell. The mHEL-20 contained about 4-fold more (21,000 sites/cell). The polyvalent rabbit anti-HEL recognized about 10,000 molecules/cell in mHEL-242 and mHEL(K52; K55) and about twice that amount in mHEL-20.
T cell assays
B cell lymphoma cells expressing wild-type or mutated forms of membrane HEL were titrated at 2-fold dilutions starting from 105 cells/well. The T cell hybridomas were added at a constant cell number of 105/well and incubated for 20 h. Subsequently, each well was assayed for levels of IL-2 using the IL-2-dependent cell line CTLL. In a total volume of 200 µl, 100 µl of supernatant was transferred to 10,000 CTLL cells and cultured for 18 h. In the last 8 h the cells were pulsed with 0.4 µCi/well of [3H]thymidine. The T cell hybridomas used were 3A9, A6A2, NH18, and NH115, which recognize peptides 48–62, 31–47, 18–33, and 115–129, respectively.
Peptide binding reactions
Peptides were synthesized by F-moc chemistry (model 432A;
Applied Biosystems, Foster City, CA) using a Synergy 432A peptide
synthesizer (Applied Biosystems) and were purified by conventional
reverse phase HPLC (Waters 600E; Millipore, Bedford, MA). The purity
and sequence specificity of each peptide were analyzed using
matrix-assisted laser desorption ionization mass spectrometry, and the
peptides were found to be 
95% pure.
Peptide binding reactions were performed with detergent-solubilized I-Ak molecules, purified from the T2-Ak cell line (provided by Dr. P. Cress-well, Yale University School of Medicine, New Haven, CT). Purified I-Ak protein (12.5 pmol) was incubated with 0.25 pmol of the radiolabeled reference peptide (125I-YE-DYGILQINSR), which binds to I-Ak with high affinity (5). The reference peptide was iodinated by the chloramine-T method at a sp. act. of 0.5 mCi/1.5 nmol of peptide. For each binding reaction, known amounts of cold test peptide were added to the mixture and incubated for 72 h at room temperature. The peptide-Ak complex was separated from the free peptides by centrifuging the material through a Bio-Spin P6 gel filtration column (Bio-Rad, Hercules, CA). The excluded material was counted with a gamma counter (Wallac, Turku, Finland). A measure of binding strength is the ability of the test peptide to inhibit 50% binding of the radiolabeled reference peptide.
Isolation of I-Ak-associated HEL peptides
Naturally processed peptides of HEL were isolated by immunoaffinity chromatography as previously described (7) but with minor modifications that improved the isolation strategy. B lymphoma cells (2 x 109) expressing wild-type or mutated forms of membrane HEL were solubilized with 20 mM each of MEGA 8 and MEGA 9 detergents, in the presence of protease inhibitors (500 mM PMSF, 10 mM iodoacetamide, and 20 µg/ml leupeptin). The class II I-Ak molecules were recovered from the cell lysate with mAb 40F coupled to cyanogen bromide-activated Sepharose (Sigma). Sepharose was loaded into a 14-cm long Polyprep disposable chromatography column (Bio-Rad) and then washed with 10 column volumes (20 ml) of wash buffer (10 mM MEGA 8 and MEGA 9 detergent in PBS) and 50 column volumes of PBS (100 ml) to remove the detergents. The peptide-I-Ak complexes were then eluted with 5 ml of 0.1% trifluoroacetic acid (TFA; pH 1.9) and neutralized with 1 M Tris at pH 7.4.
Three evaluations were performed with the eluted complexes. First, the amount of peptide in the extract was quantitated by a recently developed ELISA (8). Biotinylated peptides were bound to an ELISA plate using streptavidin, and the Ab VAL-3 was used to detect it. First, a standard inhibition binding reaction was performed, incubating the Ab with free peptide before addition to the plate. The peptide extract was likewise incubated with the Ab. The amount of peptide in the extract was estimated using a calibration plot established with reference standards. Details of the procedure were analyzed in a recent study.6
In a second assay the peptide mixture was examined by mass spectrometry for HEL31–47 peptides. The extract was passed through a Centricon-10 concentrator (Amicon, Beverly, MA), and then the free peptide extract was incubated for 12 h at 4°C with 500 µl of a 50% (v/v) slurry of the anti-HEL31–47 Ab VAL-3 (10 mg of Ab was coupled to 0.5 g of cyanogen bromide-activated Sepharose). (The amounts of Ab to cell extract and HEL31–47-reactive material were standardized in our previous report (7).) Following the incubation and removal of Sepharose, the peptide extract was incubated for an additional 12 h with 500 µl of a 50% slurry of anti-HEL48–62 Ab 151.48-61 (15 mg of this Ab was coupled to 0.25 g of cyanogen bromide-activated Sepharose). From a 50% (v/v) slurry, 500 µl of Sepharose was added to the peptide extract. After each incubation period, the HEL peptides were recovered in an identical manner. The Sepharose was loaded into a 9-cm Polyprep disposable column (Bio-Rad) and washed with 120 column volumes of 0.01 M Tris and 40 column volumes of distilled water. The HEL peptides were then eluted from each column with 0.1% TFA and captured in presiliconized microfuge tubes. The samples were dried and examined immediately or were stored at -80°C.
Third, we examined a number of autologous peptides. Their isolation involved the purification of class II I-Ak molecules from 2 x 109 B lymphoma cells as described above. The 40F-Sepharose was loaded into the 14-cm Polyprep disposable chromatography column and washed with 40 ml of wash buffer, 60 ml of PBS, 60 ml of 0.01 M Tris (pH 7.4), and 10 ml of distilled water. The peptide extract was eluted with 0.1% TFA and then passed through a Centricon-10 concentrator. The final peptide extract was dried and stored until mass spectrometric analysis.
Analysis by mass spectrometry
Analysis of the class II I-Ak peptides was performed with a Finnigan liquid chromatography quadropole ion trap mass spectrometer (Finnigan, San Jose, CA). Briefly, the peptides were reconstituted in 40 µl of 2% acetonitrile/0.6% acetic acid solution, and 15 µl of this sample was injected into the capillary reverse phase HPLC (a Zorbax C18 0.3-mm x 25-cm column; Micro Tech, Sunnyvale, CA) connected on-line to the electrospray mass spectrometer. The total column effluent, which was maintained at a 4.5 µl/min flow rate, was admitted to the mass spectrometer. Data acquisition was started 10 min from the time of sample injection. The peptides were identified by mass (MS mode) in the scan range of m/z 600-1300 in the profile mode in which every three microscans were averaged to give one scan. In the tandem (MS/MS) mode, sequence information was derived by collision-induced dissociation. Sequence analysis was performed using an automated protein database sequencing program (SEQUEST) on an ICIS workstation (Finnigan), which compares the experimental mass of each product ion to the calculated ion mass.
An estimate of the amount recovered was made by having calibration curves generated with a set of sequence-specific synthetic peptides identical with the naturally processed HEL peptides identified previously (7). Each standard peptide was weighed using a microbalance (ATI CAHN) with a mass accuracy of ±2 µg. The peptides were weighed in the range of 200–400 µg and dissolved in 1 ml of a 50% acetonitrile solution. Injecting known amounts of standard peptide (500, 100, 50, 20, and 10 fmol) into the reverse phase HPLC columns on-line to the electrospray mass spectrometer allowed establishment of a total ion chromatogram. Calibration curves were constructed by calculating the area under the total ion chromatogram (y-axis) for each amount of peptide injected. For the native peptides for which there was no sequence-specific calibration curves, their recovery was quantified using as standards the synthetic peptides that closely resembled the native ones.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We previously reported that the immunodominant HEL48–63 epitope binds to I-Ak at a 20-fold higher abundance than the minor HEL31–47 determinant, and that this difference correlated with the stronger binding affinity for the class II I-Ak (7). To examine the interrelationship between these two determinants, we mutagenized the HEL cDNA to abrogate the selection of the 48–63 epitope (DGSTDYGILQINSRWW). This was performed by substituting the aspartic acid at position 52 and the isoleucine at position 55 with lysine residues. Asp52 is the main anchor residue of the 48–63 family, whereas Ile55 contributes little to the binding, but substitution of it with a lysine hinders binding. The 48–61 mutant peptide DGSTKYGKLQINSR failed to bind to detergent-solubilized class II I-Ak molecules. For example, while the 48–61 peptide (DGSTDYGILQINSR) inhibited the standard binding reaction with an IC50 of 73.9 nM, neither of the two mutant peptides (44–61 K52, K55 and 48–61 K52, K55) had any effect, even up to concentrations of 90 µM. (As explained before, 48–61, 48–62, and 48–63, the three major species, bind with the same affinity to I-Ak. For synthetic peptides we used 48–61 because of its better solubility.)
The class II I-Ak M12C3F6 B lymphoma line was
transfected with the mutated HEL cDNA (K52; K55), resulting in the
stable expression of HEL as a membrane protein. We tested the ability
of these mHEL(K52; K55) mutant cells to stimulate HEL-specific T cell
hybridomas. By comparison to the wild-type mHEL line (mHEL-242),
mHEL(K52; K55) cells failed to trigger the 3A9 T cell hybridoma that
recognizes the HEL48–63 determinants (Fig. 1
). The T cell hybridomas recognizing
other epitopes of HEL secreted comparable levels of IL-2 when cultured
with mHEL or mHEL(K52; K55) B lymphoma cells. We obtained similar
results when these T cells hybridomas were cultured with fixed APC
(data not shown). Thus, the presentation of the other HEL determinants
(18–33, 31–47, and 115–129) was not changed in the absence of the
HEL48–62 epitope. Moreover, the extent of T cell
reactivity did not indicate a higher level of expression of these
epitopes.
|
Class II I-Ak molecules were purified from B
lymphoma cell lysates, and the self-peptides were released by addition
of 0.1% TFA. We quantitated the amounts of the 31–45 family of
peptides by a new approach that uses an ELISA inhibition assay. Fig. 2 indicates that there were no
differences in the amounts of the 31–47 family of peptides between the
line expressing the wild-type and that expressing the mutant HEL
molecule. A second experiment gave identical results with equal amounts
of peptides (not shown). (We are now studying the selection of the
18–33 family of HEL peptides. An initial estimate in the cell lines
gave identical contents (see Fig. 2).)
|
and Fig. 3). Seven of these peptides started at
residue 31 (Ala), six at residues 32 (Ala), and six at residue 33
(Lys). There was greater variation on the carboxyl terminus where the
peptides ended at residues 45, 46, 47, 48, 49, 51, and 52 (Table I
shows the major species identified). The peptide profile obtained from
the mutant mHEL(K52; K55) line was not significantly different. About
60% of peptides commenced at residue 31 in both mutant and wild-type
cells, whereas 70% of peptides ended at 46 or 47 at the carboxyl
terminus. Interestingly, in the mutant mHEL(K52; K55) line, the few
peptides ending at residue 51 were preferred over those few peptides
ending at 52.
|
|
Thus, the biochemical results are consistent with those presented in
Fig. 1
, where we did not find an elevated presentation of 31–47
peptides in the mutant mHEL line (K52; K55) in the absence of the
immunodominant 48–62 HEL peptide.
Identification of the 31–47 family of HEL peptides from proline mutant lines
The abundance of proline residues at the penultimate position of many naturally processed peptides has been well documented (11, 12, 13). Prolines stop exopeptidases involved in peptide trimming. Recent studies in our laboratory demonstrated that the presence of proline residues at the amino-terminal flanks of the 48–63 peptides resulted in the selection of longer peptides extended on the amino side, whereas the total amount of 48–63 peptides presented was the same (3). We are interested in these proline mutant lines because the major epitopes that were selected impinged on the carboxyl terminus of the selected minor determinant 31–47. The B lymphoma cells were transfected with the HEL cDNA consisting of single amino acid substitutions for a proline residue at positions 46, 47, and 48. Class II I-Ak molecules were purified from cell lysates of each B cell lymphoma line, and the HEL peptides were recovered by the anti-HEL31–47 peptide Ab VAL-3 and the anti-HEL48–61 peptide Ab 151.48-61.
The peptide species isolated from these proline mutant lines commenced
at residue 31, 32, or 33 in similar proportion to that of the peptides
isolated from the wild-type mHEL line. Peptides isolated from the
Pro46 mHEL line ended predominantly at residues
46 and 47 (63%; Figs. 4 and 5). The peptides ending at residue 47
(41%) were preferred over those that ended at 46. Interestingly, a
significant number of peptides extended to residue 52 (17%). In the
Pro47 mHEL line, 92% of peptides end at residues
48 (Asp) or beyond, i.e., at 49 (Gly) or 52 (Asp). Peptides ending at
residues 46 and 47 were few (Fig. 4). Of the peptides extracted from
the Pro48 mHEL line, 73% ended at residue 46,
and none at 47. Thus, the effect of proline in extending the peptides
bound with Pro47 was not observed in the
Pro48 line.
|
|
Under these circumstances, was the selection of the immunodominant
determinant favored over that of the minor determinant? We confirmed
the results of our previous report, that the
Pro47 line expressed the same number of complexes
of the 48–63 family as the controls; binding with labeled AW Ab on
Pro47 resulted in 9.4 ng of
Ab/106 cells, in contrast to the standard m242
cells that exhibited 9.3. As shown in Fig. 2, the amount of the 31–46
family of peptides was the same in the Pro47
line. This result indicates that these two epitopes are selected
independently from separate HEL molecules (i.e., do not originate from
a single protein molecule).
Selection of autologous peptides from class II I-Ak
Because the dominant 48–62 peptide does not influence the
expression of the 31–47 peptide family, we examined whether HEL could
influence the expression of autologous epitopes. During the processing
of HEL, up to 20% of the class II molecules can be occupied by the
48–63 peptides among a spectrum of some several hundred self-peptides
processed and presented at steady state (6). Analysis of
class II I-Ak self-peptides in the
M12-Ak and mHEL-20 lines revealed an array of
self-peptides at different retention times (Fig. 6). None of the self-peptides was
presented to the same extent as the HEL48–62
peptide, and this was most obvious at 25 min. On the basis of
collision-induced dissociations (MS/MS mode), the complete sequences
and sources of prominent self-peptides were determined using the
SEQUEST database search. In the mHEL-20 line, the relative recovery of
the seven self-peptides analyzed was about the same as that observed in
the M12.C3.F6 line (Table II). In the
absence of HEL selection, we did not observe an increased presentation
of other self peptides; in particular, the Aßk
peptide with a binding strength similar to that of the
HEL48–62 peptide (5) was not
preferentially selected. The processing of HEL did not impede
presentation of the other self-peptides.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The scenario that we favor for HEL processing is that the protein, which in its native state is relatively resistant to proteolysis, needs first to be reduced (14, 15). After reduction and/or partial unfolding, the polypeptide is available to bind to I-Ak through the segments that contain a satisfactory binding core sequence, such as 52–61. This binding protects the peptide segment and the MHC molecule from proteolysis (16, 17, 18), but the portion of the polypeptide that extends beyond the binding site will be trimmed by amino and carboxyl peptidases. The results of our previous and present experiments, in which prolines were substituted on the amino-terminal flanks of 48–62 and caused longer peptides to be presented, support this interpretation. In agreement, some of the long natural peptides isolated from class II MHC molecules have been shown to contain prolines (11, 12). Our present findings indicate that this effect of proline also applies to the carboxyl terminus of the selected 31–47 segment. This and our previous biochemical analysis of the trimming of the proline-containing segments indicate that the trimming takes place up to the edge of the I-Ak molecule (19). When the distance between the proline residue and I-Ak increases, the effect of proline disappears, and the segment can again be trimmed to its original length (i.e., note the results with the Pro48 mutant). Prolyl dipeptidase and endopeptidase may be responsible for such trimming.
Despite the large predominance of 48–63 peptides compared with 31–49, binding of the former did not exclude the selection of the latter. There are two, not mutually exclusive, possibilities to explain these results. First, the lack of influence may be due to the low binding affinity of 31–47 for I-Ak. The lack of a 48–63 segment in HEL should not influence the ultimate concentration of HEL in vesicles, a parameter that should be critical for selection; for a low affinity binding segment such as 31–47, this amount of HEL may be limiting and too low to allow optimal binding affinity. Aside from this, other self peptides may have higher affinity for I-Ak and should preferentially bind. Second, the 31–47 family may be selected from a minor set of HEL molecules at a cellular site different from that where most of the 48–63 family is selected. Claims have been made that 31–47 is less dependent on the invariant chain or H-2DM than is 48–63 (20, 21). A segment with a low binding core may have a low chance of survival in a proteolysis-rich vesicle.
The presence of HEL peptides did not influence the presentation of the major autologous proteins, despite HEL providing a family of peptides that can occupy up to 20% of the I-Ak molecules. This result may be accounted for by either anatomical and/or affinity considerations as described above. Regardless, these results do tell us that it is difficult, if not impossible, to compete for presentation of autologous peptides even by a high affinity peptide, an issue that has been argued, without any chemical data, to support peptide therapy for the control of autoimmunity.
Finally, the issues of intramolecular competition of a protein for Ag processing has been raised as an explanation for the differences in T cell responses to peptides (22, 23). Without questioning such results, our findings and those of others (24) suggest that such phenomena need to be substantiated by chemical approaches independent of T cell readouts.
|
Footnotes |
|---|
2 R.G., C.V., and I.V. contributed equally to this work.
3 Current address: Walter and Eliza Hall Institute, Melbourne, Australia 3052.
4 Address correspondence and reprint requests to Dr. Michael Gross, Department of Chemistry, Washington University, 1 Brookings Drive, St. Louis, MO 63130; or Dr. Emil R. Unanue, Department of Pathology and Immunology, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110.
5 Abbreviations used in this paper: HEL, hen egg-white lysozyme; TFA, trifluoroacetic acid; mHEL, mutant HEL.
6 C. Velazquez and E. R. Unanue. 2000. Quantitation of lysozyme peptides bound to class II MHC molecules indicates large differences in content. Submitted for publication.
Received for publication April 17, 2000. Accepted for publication June 26, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  V. P. Yeung, J. Chang, J. Miller, C. Barnett, M. Stickler, and F. A. Harding Elimination of an Immunodominant CD4+ T Cell Epitope in Human IFN-{beta} Does Not Result in an In Vivo Response Directed at the Subdominant Epitope J. Immunol., June 1, 2004; 172(11): 6658 - 6665. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Suri, J. J. Walters, O. Kanagawa, M. L. Gross, and E. R. Unanue Specificity of peptide selection by antigen-presenting cells homozygous or heterozygous for expression of class II MHC molecules: The lack of competition PNAS, April 29, 2003; 100(9): 5330 - 5335. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. B. Lovitch, J. J. Walters, M. L. Gross, and E. R. Unanue APCs Present A{beta}k-Derived Peptides That Are Autoantigenic to Type B T Cells J. Immunol., April 15, 2003; 170(8): 4155 - 4160. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. H. Koonce, G. Wutz, E. J. Robertson, A. B. Vogt, H. Kropshofer, and E. K. Bikoff DM Loss in k Haplotype Mice Reveals Isotype-Specific Chaperone Requirements J. Immunol., April 1, 2003; 170(7): 3751 - 3761. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Velazquez, I. Vidavsky, K. van der Drift, M. L. Gross, and E. R. Unanue Chemical Identification of a Low Abundance Lysozyme Peptide Family Bound to I-Ak Histocompatibility Molecules J. Biol. Chem., November 1, 2002; 277(45): 42514 - 42522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Velazquez, R. DiPaolo, and E. R. Unanue Quantitation of Lysozyme Peptides Bound to Class II MHC Molecules Indicates Very Large Differences in Levels of Presentation J. Immunol., May 1, 2001; 166(9): 5488 - 5494. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
;
mHEL-242) or mutated form of HEL (•) were cultured with four
HEL-specific T cell hybridomas directed to different HEL epitopes.
mHEL(K52; K55) failed to stimulate only the 3A9 T cell hybridoma
directed to the 48–62 epitope. The response of the T cells directed to
the other epitopes was identical with that of the APC-expressing
unmutated HEL. Each box identifies and indicates the peptide
specificity of the T cells.
