The JI
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     
 

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stefanski, H. E.
Right arrow Articles by Hogquist, K. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stefanski, H. E.
Right arrow Articles by Hogquist, K. A.
The Journal of Immunology, 2000, 164: 3519-3526.
Copyright © 2000 by The American Association of Immunologists

Positive Selection Is Limited by Available Peptide-Dependent MHC Conformations1

Heather E. Stefanski, Stephen C. Jameson and Kristin A. Hogquist2

Department of Laboratory Medicine and Pathology and Center for Immunology, University of Minnesota, Minneapolis, MN 55455


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Recent data suggest that the diversity of self peptides presented in the thymus during development contributes to positive selection of a diverse T cell repertoire. We sought to determine whether a previously defined "hole in the immunological repertoire" could be explained by the absence of an appropriate selecting self peptide. The repertoire defect in question is the inability of bm8 mice to make an H-2K-restricted response to OVA. Like other OVA-specific, H-2K-restricted receptors, OT-I-transgenic T cells are not positively selected in bm8 mice. Using criteria we had previously established for identifying positive selection ligands, we found peptides that could restore positive selection of OT-I thymocytes in bm8 mice. Thus, the T cell repertoire can be limited by a requirement for specific self peptides during development. Data with MHC-specific Abs suggested that peptides might be able to force MHC residues to adopt different conformations in Kb vs Kbm8. This shows that peptides can potentially contribute to ligand diversity both directly (via variability in the solvent-exposed side chains) and indirectly (through their effect on the MHC conformation). Our data support a model where self peptide diversity allows selection of T cells specific for a broad range of MHC conformations.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Positive selection is the process whereby thymocytes differentiate into mature T cells after interacting with self peptide/MHC complexes on stromal cells. The requirement for such an interaction selects for cells with functional TCRs and results in MHC restriction of the repertoire (1). The precise role that self peptides play in this process has been difficult to fully understand. Although it is clear that peptides are required for positive selection, conflicting experimental evidence exists on whether or not the specificity of such peptides is important (2).

Peptide "add back" experiments directly showed that the nature of the MHC-bound peptide could influence cell fate (3, 4). In many situations, peptides that were closely related to the antigenic peptide for that receptor supported positive selection, whereas unrelated peptides did not (5, 6, 7, 8, 9). Although there are several examples in which unrelated peptides induced positive selection, including naturally occurring self peptides (10, 11, 12, 13, 14), even in these cases certain peptides supported selection whereas others did not. These experiments strongly suggest that peptide specificity is inherent in the positive selection process and imply that a diverse presentation of self peptides in the thymus is important for generating a diverse T cell repertoire.

Nonetheless, it seems unlikely that positive selection should be highly peptide specific, because that might severely limit repertoire size, and a high degree of peptide specificity during positive selection itself serves no apparent immunological purpose. Indeed, other data suggested that positive selection is not greatly impacted by the specificity of the self peptides presented. Here, the experimental approaches were designed to limit the available peptide/MHC repertoire to one face and to examine the impact on repertoire size (15, 16, 17, 18). It was shown that in such "single" peptide/MHC mice, a large and diverse repertoire was selected. This argues that peptide diversity, although present in the normal situation, is not important in T cell development. However, it has recently been shown that at least two of these approaches did not truly limit the peptide presentation to a single face (19, 20). Additionally, the small percentage of MHC molecules presenting diverse self peptides was shown to be important in selecting T cells. Indeed, it may prove to be impossible to experimentally limit the presented peptide diversity to purely one single complex in mice, and this leaves the issue of peptide specificity in selection unresolved.

In this work, we approached the issue another way by asking whether the absence of a particular self peptide could ever limit the repertoire. This approach builds on the observation that the mutant Kb molecule, Kbm8, does not select the same T cell repertoire that Kb does. Normally, the fact that two distinct MHC molecules select unique repertoires is not surprising; indeed it is a natural outcome of positive selection. However, in the case of Kb and Kbm8, it is surprising because the two alleles are highly similar in amino acid sequence and to date have not been distinguished serologically (21). There are four amino acid differences between Kb and Kbm8, and three of them lie in the bottom of the peptide binding groove (22). Of these, only two impact recognition, and they do so by influencing which peptides are presented, apparently without significantly affecting the structure of the {alpha}1 and {alpha}2 helices, which the TCR contacts (23, 24, 25, 26). bm8 mice are unable to mount an H-2K-restricted immune response to OVA. This has been shown to be due to absence of positive selection of OVA-specific T cells (27). In fact, an OVA-specific, Kb-restricted TCR transgene (OT-I) is also not selected in bm8 mice (28). This suggests that of the highly diverse pool of self peptides presented by Kbm8, none is sufficient to support positive selection of OVA-specific H-2K-restricted receptors such as OT-I. However, it remains possible that despite the similarity between Kbm8 and Kb, Kbm8 is unable to assume a structure that could support positive selection of such receptors and that the lack of an appropriate self peptide has no role in this positive selection defect. To test this, we searched for a peptide that could "restore" the positive selection of OVA-specific receptors in bm8 mice, thereby proving that the peptide specificity of positive selection could be so strict as to result in the loss of receptors from the repertoire.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Mice

OT-I is a C57BL/6 TCR transgenic (tg)3 strain, expressing a receptor specific for OVA/Kb (5). CD8tg is a C57BL/6 strain overexpressing CD8ß and CD8{alpha} under the control of the CD2 promoter (29). ß2M0 is a 129 x C57BL/6 strain with a targeted disruption of ß2-microglobulin (ß2m) on both chromosomes (30). B6.C-H-2bm8 (bm8) mice were generous gifts from Dr. J. Nikolic-Zugic (Sloan-Kettering Institute, New York, NY) and Dr. N. Gascoigne (Scripps Institute, La Jolla, CA).

Library construction

An Escherichia coli library producing mixtures of octamer peptides of the sequence SXIXFXXL, where X is any of the 20 amino acids, was a gift of Marc Gavin and Michael Bevan (31). Insert DNA for the XXXNFEKL library was prepared from self-priming degenerate oligonucleotides (5'-CGT GGA TCC ATC GAG GGT AGG NNC/G NNC/G NNC/G AAC TTT GAA AAA CTG TAA TAA TGA CTG CAG TC-3'; N = G, A, T, or C) according to a standard protocol. The double-stranded product of the primer-extension reaction was cleaved by BamHI and PstI and ligated into pMAL-c (New England Biolabs, Beverly, MA) that had been linearized with the same two restriction enzymes. After transformation of competent DH5{alpha} E. coli, bacteria were cultured in SOC medium for 45–60 min and frozen at -80°C in 20% glycerol. Once adequate numbers of transformants were obtained (close to the theoretical complexity, see below), bacteria were thawed and seeded into eight 48-well plates (XXXNFEKL) or seven 48-well plates (SXIXFXXL) containing 0.75 ml Luria-Bertani (LB), 2 mg/ml glucose, 50 µg/ml ampicillin (LBGA) per well and the plates were incubated overnight at 37°C with 200 rpm agitation. To confirm the size of the libraries, aliquots of the thawed transformants were simultaneously plated on LB/ampicillin-agarose plates. Thus, it was determined that XXXNFEKL consisted of 4,800 clones (theoretical diversity, 8,000) and SXIXFXXL contained 189,000 (theoretical diversity, 160,000). Aliquots (0.25 ml) of the expanded 48-well cultures were pooled, made 20% glycerol, and stored as aliquots at -80°C.

To generate pure XXXNFEKL and SXIXFXXL peptides,)-peptide fusion proteins were prepared as described starting with aliquots of the pooled transformants. Bacteria were induced with isopropyl-ß-D-thiogalactopyranoside (Boehringer Mannheim, Indianapolis, IN) in 300-ml cultures, and fusion protein was purified on 12-ml bed volume amylose columns. After column elution with 10 mM maltose, fractions containing protein were pooled, quantitated by the BCA protein assay reagent (Pierce, Rockford, IL) and cleaved with the factor Xa (New England Biolabs) at room temperature overnight. Peptides were isolated by 10-kDa cutoff filtration (Centriprep 10, Amicon, Beverly, MA), made 0.1% trifluoroacetic acid, lyophilized, resuspended in 0.1% trifluoroacetic acid, and fractionated by reverse phase HPLC (Delta Pak C19-300A column, 3.9 mm x 30 cm; Waters 501, Millipore, Milford, MA). Fractions of 500 µl were collected, lyophilized, and resuspended in 200 µl distilled water.

Peptide preparation from libraries

In screening the libraries to isolate epitope expressing clones, it was not necessary to purify the maltose-binding protein-peptide fusion proteins by affinity chromatography. In the first round of screening, 96-well plates containing 1 ml LBGA were inoculated with 20 µl from each well of the 48-well frozen library stocks. They were incubated with agitation at 37°C to mid-log phase, isopropyl-ß-D-thiogalactopyranoside was added to 0.3 mM, and incubation was continued for another 3–4 h. The bacteria were pelleted and resuspended in 200 µl PBS, 1 mM EDTA, 0.5 mg/ml lysozyme, pH 8, and incubated at 4°C for 30 min followed by three freeze/thawings. The insoluble material was pelleted by centrifugation, and 50 µl of the supernatants were placed in wells of 96-well round-bottom plates containing 10 µl factor Xa (5 µg/96-well plate) diluted in PBS. Bacterial lysates were reacted with factor Xa for 4–5 h at 4°C. The peptides were then ready to be used in a 51Cr release assay without further modification. No increase in spontaneous release of labeled targets due to Factor Xa was noted. To find positive clones, a pool screening strategy was used. The bacterial cultures from wells that gave positive results in a CTL assay were diluted, and additional rounds of screening were done in the same manner until the cultures contained 1 clone/well. Plasmids from positive clones were prepared using a standard Miniprep procedure and were sequenced from the 3' end of the cloned oligonucleotide using the -40 primer.

Synthetic peptides and cell lines

Peptides were synthesized by Research Genetics (Huntsville, AL) and were used without further purification. Some peptides were kind gifts of Janko Nikolic-Zugic (Sloan-Kettering Institute, New York, NY). T2 (human TAP-1 and TAP-2 deficient) cells were transfected with Kb (a gift from Peter Cresswell, Yale University, New Haven, CT) or Kbm8 (a gift from Larry Pease, Mayo Foundation, Rochester, MN). Bwbm8 cells expressing Kbm8 were also a gift from Larry Pease.

CTL assays

The OVA/Kb-specific clone B3 or the OT-I-derived CTL clone CL.2 was used as responder. BWBM8 (expressing Kbm8) and EL4 (expressing Kb) cells were used as targets. CTL function was determined in a 51Cr release assay as described (5) using 3 x 104 CTL and 104 51Cr-labeled target cells in a 3-h assay.

MHC I stabilization assays

T2-Kb or T2-Kbm8 cells were incubated overnight at room temperature to force empty class I molecules to accumulate at the cell surface. Peptides were added at various concentrations and incubated at room temperature for 30 min and then at 37°C for 4 h. The cells were washed and stained for MHC I expression using the Kb-specific Abs Y3, 5F1, or 100.30, followed by goat anti-mouse Ig FITC.

Dulling assays

Peritoneal exudate cells (4) from bm8 mice, induced 5 days previously with 1 ml of thioglycolate, were plated at 1 x 105/well in a 96-well microtiter plate. After adherence for 1 h, monolayers were washed four times. Thymocytes (5 x 105) from OT-I, CD8tg, ß2M0 mice were isolated from 4- to 6-week-old mice and cocultured with the peritoneal exudate cells for 16–20 h at 37°C. Cells were stained for expression of CD4 (PE-RM4-5) and the non-Tg-encoded CD8.2 (biotin-2.43), followed by streptavidin-FITC, and analyzed as described previously (11).

Fetal thymic organ culture (FTOC) and proliferation assays

FTOC was performed essentially as described (5) except OT-I bm8 animals were used and exogenous ß2m was not added. After 7 days, thymocytes were stained for expression of V{alpha}2 (the OT-ITg TCR {alpha}-chain), CD4, and CD8. Some thymocytes were tested for their capacity to respond to Ag in a 2-day proliferation assay determined by [3H]thymidine incorporation, as described previously (5).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Peptides that can be recognized by OVA-specific CTL in the context of Kbm8 are rare

Despite the fact that the endogenously processed OVA257–264 SIINFEKL (OVAp) binds Kbm8, it is unable to stimulate OVA-specific CTL in the context of Kbm8 (27, 32). Before searching for peptides that would restore positive selection of OVA-specific receptors, we sought to determine whether peptides could be targets for OVA-specific CTL in the context of Kbm8, and how frequent they are. To do this, we constructed a semirandom octamer peptide library predicted to bind both Kb and Kbm8. The library was based on the OVAp but allowed variation at the first three amino acids: XXXNFEKL. This design was chosen because the first three amino acids bind to the MHC in the vicinity of the Kbm8 mutations. Presumably, variability at these positions would allow for selection of an optimal Kbm8-binding peptide, whereas conservation at the carboxyl-terminal residues would enhance selection of OVAp mimetopes.

We used the pMal-C prokaryotic expression vector as described by Gavin et al. (31). Briefly, a self-priming degenerate oligonucleotide, encoding the XXXNFEKL sequence and three stop codons, was cloned into the pMal-C vector to encode a fusion protein in the following sequence: maltose-binding protein; factor Xa recognition site; and peptide. The degenerate peptides were then expressed and purified in batches from pooled E. coli clones. This material was fractionated into 101 fractions by HPLC and tested for the ability to target Kbm8 target cells for lysis by the OVA-specific CTL clones B3 or CL.2. Fig. 1Go shows a representative chromium release assay. The complexity of peptides tested in this assay was 4800. Fig. 1Go shows that ~40 HPLC fractions were able to target lysis when presented in the context of Kb. However, only two of these same fractions were able to target lysis when presented in the context of Kbm8. These data suggest the peptides that can target lysis of OVA-specific CTL when presented in context of Kbm8 are rare. Because individual bacteria express a unique peptide, it was possible to isolate clones based on CTL reactivity using a subpool screening strategy. We were able to determine the sequence of one of these peptides, SEINFEKL (E2). This peptide differs from OVAp only at position 2, with a glutamic acid instead of an isoleucine, and had been previously shown to target lysis in the context of Kbm8 (33). We did not determine the sequences of other peptides.



View larger version (34K):
[in this window]
[in a new window]
  		FIGURE 1. OVA-specific CTL detect more peptides from a semirandomized library when presented in the context of Kb than Kbm8. Peptides from the XXXNFEKL library were purified and fractionated by HPLC. The capacity of HPLC fractions to induce lysis by the OVA-specific CTL line B3 in context of Kb or Kbm8 was tested in a 3-h 51Cr release assay. Similar results were observed when the OVA-specific clone CL.2 was used. The peptide SEINFEKL was sequenced in the library, and this peptide resolves at fraction 50–51.

 
Due to the paucity of peptides that were able to restore CTL recognition when presented in context of Kbm8, we decided to test a more diverse library. We obtained the SXIXFXXL library (31), which has all of the dominant TCR contact residues mutated. Although this library has isoleucine fixed at position 3, we reasoned that this would not be a problem because a naturally occurring viral bm8-binding peptide has isoleucine in this position (33). We followed the same protocol as before, dividing the library into pools, inducing fusion protein production, purifying the peptides, separating them into fractions, and then screening HPLC fractions of these pools for CTL reactivity. Again we saw that Kb presented many more peptides to CTL compared with Kbm8, and we were unable to define any new peptides that restored recognition (data not shown).

These data, along with the data from the XXXNFEKL library, confirm that very few peptides can be presented in context of Kbm8 to OVA-specific CTL. Even though such peptides are rare, the fact that a peptide (in this case E2) can restore recognition when presented by Kbm8 suggests that peptides that could restore positive selection of OT-I T cells might also exist.

Defining candidate peptides for positive selection

The E2 peptide found using a library screening strategy is an agonist or a stimulatory peptide and therefore would most likely mediate negative selection when presented in the thymus, as has been shown to be the case (28). In OT-I and other systems, antigenic peptides that are mutated at the TCR contact residues to abolish agonistic activity were found in some cases to still influence T cell activation and positive selection (5, 7, 9). Therefore we varied the E2 peptide at the TCR contact residue positions 2 and 4, and we tested such peptides for the ability to mediate positive selection. On the basis of previous work, we expected that peptides likely to drive positive selection would have the following properties: 1) ability to bind the MHC allele; 2) not stimulatory for mature T cells (i.e., are negative in a CTL lysis assay); and 3) stimulatory for immature thymocytes. These variant peptides are shown in Table IGo.


View this table:
[in this window]
[in a new window]
  		Table I. Characteristics of various peptides presented by Kbm8

 
First, to identify peptides that bind Kbm8, we tested the variants in an MHC I stabilization assay. The half-maximal peptide concentrations are shown in Table IGo. Most of the peptides were able to bind to Kbm8, albeit with varying degrees of efficiency. A peptide from HSV was used as a positive control, binding Kbm8 with a half-maximal value of 15 nM. The S2E4 variant bound Kbm8 almost as well as HSV with a half-maximal value of 20 nM. Most of the other peptides had values between ~100 nM and 500 nM. D2 and K2 did not bind well to Kbm8.

Second, we measured agonist activity to mature T cells in a lysis assay using OVA-specific, Kb-restricted CTL as the effectors and Kbm8-positive cells as the targets, as described above for the library screening. Only three of the variant peptides were able to cause lysis of target cells: E2, D2, and K2 (Table IGo). E2 was the most stimulatory peptide with a half-maximal value of 10 nM. D2 and K2 were also able to target CTL recognition, although at lower efficiencies than E2. OVAp, which is the cognate ligand for OT-I when presented by Kb, was unable to stimulate OT-I CTL when presented in context of Kbm8. This is not due to the fact that OVAp cannot bind to Kbm8 (Table IGo and 33). None of the other variant peptides was able to target CTL recognition, even at high concentrations.

Finally, we used a double-positive (DP) dulling assay to detect peptides that stimulate immature thymocytes. Immature thymocytes have been shown to detect peptide ligands that are below the threshold of detection by any activation assay that utilizes mature T cells (34, 35). The dulling assay measures activation-induced coreceptor down-regulation on immature DP thymocytes. This assay has been successfully utilized to find positive selection ligands in the past (11). In a DP dulling assay, peritoneal macrophages are used as APC for thymocytes from preselection TCR Tg mice. In this case the OT-I, ß2-microglobulin-deficient thymocytes also express a CD8 transgene which further enhances their ability to detect weak ligands (11). In this assay, TCR stimulation leads to the decreased surface expression of CD4 and CD8 which is measured after 24 h of culture (11, 36). Although this assay detects both positive (weak) and negative (strong) selection ligands, those peptides that stimulate thymocytes but not CTL are good candidates for positive selection. Table IGo shows that a number of peptides could not be recognized as targets in the CTL assay but were detected in the dulling assay. Not surprisingly, the CTL agonists E2, D2, and K2 were able to induce down-regulation of CD4 and CD8, and because they function as agonists for mature CTL, they would be expected to mediate negative selection. The peptides OVAp, S2D4, S2E4, D4, and Q4 all scored positive in a dulling assay but negative in a lysis assay. OVAp is the cognate ligand for OT-I when presented in context of Kb and is not stimulatory to mature T cells in the context of Kbm8. However, in the dulling assay, OVAp is detected as a weak ligand with half-maximal activity at 100 nM. Although the thymocytes themselves (which are H-2Kb positive but deficient for ß2m) could potentially allow inefficient peptide presentation to each other in this assay, this was not the case for most of the peptides, because coreceptor down-regulation required the presence of APC (data not shown). The one exception to this was the OVAp peptide, which showed weak activity even in the absence of APC. Thus, we were unable to definitively determine whether the presentation of OVAp by Kbm8 would stimulate thymocytes. A number of peptide variants scored as null ligands (S2R4, E4, and R4), in that they bound Kbm8 but did not stimulate lysis or coreceptor down-regulation. In summary, these experiments defined the peptides S2D4, D4, Q4, and S2E4 as potential candidates for positive selection based on their ability to stimulate thymocytes while not being stimulatory to CTL.

A single peptide restores positive selection of functional CD8+ T cells

To determine whether these peptides could restore positive selection in OT-I bm8 mice, FTOC was performed. In this system, thymic lobes from OT-I bm8 mice at a gestational age of 16 days were excised, and sister lobes were cultured in the presence or absence of peptide for 7 days. At that time, the thymocytes were harvested and stained for TCR, CD4, and CD8. FTOC from Kb-positive OT-I animals display ~45% mature CD8+ T cells (data not shown and Ref. 5). In OT-I/bm8 cultures, however, the level of maturation was much lower (5% on average) as shown previously (Fig. 2Go and Ref. 28). Addition of a control Kbm8-binding peptide (HSV) had no effect on maturation in OT-I/bm8 cultures (Fig. 2Go). The candidate peptides D4 and S2D4 were able to induce maturation of CD8+ T cells (TCRhigh, CD8 single positive) from 5% to 31% and 27%, respectively (Fig. 2Go). D4 and S2D4 also increased the absolute number of TCRhighCD8+ T cells (Table IIGo). D4 caused an average 4.4-fold increase compared with no peptide, whereas the increase with S2D4 was 2.1-fold. Interestingly, the candidate peptides Q4 and S2E4 had no effect (Fig. 2Go, Table IIGo). It is possible, given the weak binding of Q4 to Kbm8, that it did not exhibit sufficient stability in serum containing media to influence development but was sufficient to cause bioactivity in the short term DP dulling assay. S2E4 was a very weak ligand in the dulling assay and perhaps was too weak to restore positive selection even at high concentrations in FTOC. Nonetheless, these data show that specific peptides can restore positive selection in OT-I bm8 mice.



View larger version (57K):
[in this window]
[in a new window]
  		FIGURE 2. D4 and S2D4 restore positive selection in OT-I bm8 fetal thymic organ culture. OT-I/bm8 lobes were cultured for 7 days and stained for TCR, CD4, and CD8 expression. For each culture condition, the histogram on the left represents TCR levels of total thymocytes, and the dot plot on the right represents the CD4 and CD8 expression after gating on TCRhigh cells. The percentage of TCRhigh CD8 single-positive cells among total live cells is shown in the upper right hand corner for each culture condition. The cells in the S2D4 condition were stained with CD8 FITC, whereas the rest of the conditions were stained with CD8 APC. The CD8 gate was based on CD8 single-positive cells from a control OT-I Kb lobe. These are representative plots of at least four lobes per condition.

 

View this table:
[in this window]
[in a new window]
  		Table II. S2D4 and D4 cause an increase in the number of TCRhighCD8+ T cells in FTOC

 
To determine whether the cells that were selected by S2D4 were functional, we tested them for their ability to proliferate in response to OVAp presented by Kb on tumor cells. None of the thymocytes from either culture proliferated when allowed to interact with APC alone. However, the thymocytes from the S2D4 cultures proliferated 5-fold more than those from control cultures in response to OVAp (Fig. 3Go). In addition, the cells selected in the presence of S2D4 expressed normal levels of CD8ß (data not shown) which has been shown to correlate with bioactivity in thymic cultured cells (37, 38). These data show that S2D4 drives the positive selection of functionally and phenotypically normal CD8+ T cells in OT-I/bm8 thymic cultures.



View larger version (22K):
[in this window]
[in a new window]
  		FIGURE 3. The CD8+ T cells selected by S2D4 are functional. Organ culture from OT-I bm8 was performed in the presence of 10 µM S2D4 or a control peptide HSV. Thymocytes were harvested and counted. A fraction of the cells were stained as in Fig. 2Go. The remaining cells from an equivalent number of lobes in each group were added to cultures with irradiated EL4 APCs with or without 100 nM OVA. [3H]Thymidine incorporation was determined after 48 h. The data shown are the mean (and SEM) of triplicate cultures from one experiment. Similar results were seen in two additional experiments. {square}, proliferation of thymocytes to APC alone; , APC pulsed with 100 nM OVA.

 
Peptides imprint unique conformations on the {alpha} helices of Kb and Kbm8

Our data suggest that peptides exist that can mediate positive selection of OVA-specific receptors when presented by Kbm8. Thus, it is surprising that naturally occurring Kbm8 molecules, which present thousands of unique self peptides (24, 25), cannot mediate selection. Perhaps Kbm8 simply does not bind any of the self peptides that OT-I utilizes for selection. We have estimated that the number of such peptides is small and have determined the sequence of one of these (11). This peptide, CP{alpha}1, is naturally processed and presented from the ubiquitously expressed F-actin capping protein. We tested whether Kbm8 could bind CP{alpha}1 and found that it could (Table IIIGo). This is not surprising given the previous observation that the repertoire of peptides that Kb and Kbm8 present is overlapping (25, 39, 39). Indeed, we tested 16 naturally occurring self peptides presented by Kb and found that 5 of these peptides bind Kbm8 and Kb equivalently. Eight of the peptides displayed slightly lower binding to Kbm8 and 3 displayed much lower binding to Kbm8 (Table IIIGo).


View this table:
[in this window]
[in a new window]
  		Table III. Relative binding of K2 self peptides to K2 and Kbm8

 
This suggests that lack of peptide binding may not entirely explain the inability of bm8 to mediate selection of OVA-specific receptors. Rather, bm8 may bind the same peptides that mediate positive selection when presented by Kb, but such peptides either adopt a unique conformation in the Kbm8 groove or force a unique conformation upon the Kbm8 {alpha}1 and {alpha}2 helices, such that T cell recognition is altered. The fact that Kb and Kbm8 are serologically indistinguishable by a panel of Kb-specific Abs might seem to argue against the idea that the {alpha}1 and {alpha}2 helices assume unique conformations (21). However, these staining experiments were done using splenocytes or tumor cells that present many different peptides. It has been shown that a number of Kb-specific Abs are peptide sensitive (40). In other words, although they bind the {alpha}1 and {alpha}2 helices, this binding can be influenced by the nature of the groove-associated peptide. To determine whether differences in Ab binding to Kb vs Kbm8 could be detected, we tested different peptides in an MHC I stabilization assay and used a panel of 5 peptide-sensitive Kb-specific Abs. Two Abs, 5F1 and 100.30, were able to discriminate between Kb and Kbm8 (Fig. 4Go, top and middle). This discrimination was apparent only when certain peptides, such as E4 or PTP, were used to stabilize the MHC. The absence of staining on Kb/E4 or Kbm8/PTP was not due to the inability of those peptides to bind the MHC because the peptide-insensitive Ab Y3 recognized these complexes (Fig. 4Go, bottom) on both Kb and Kbm8. The data suggest that peptide/Kbm8 complex is able to adopt a conformation different from that of Kb when certain peptides are presented in the peptide binding groove.



View larger version (48K):
[in this window]
[in a new window]
  		FIGURE 4. Kbm8 and Kb assume distinct conformations when certain peptides are bound. T2-Kb or T2-Kbm8 cells were pulsed with 10 µM peptide for 4 h and then stained with the Kb-specific Abs 5F1, 100.30, or Y3. Y axis, mean fluorescence intensity (MFI) when the peptide is present minus the mean fluorescence intensity of the no peptide control. {square}, T2-Kb cells; {blacksquare}, T2-Kbm8 cells. A representative of three experiments is shown.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
bm8 mice do not positively select H-2K-restricted OVA-specific T cells (27, 28). We wanted to determine whether this defect was due to the absence of an appropriate selecting self peptide during thymic development or whether OVA-specific receptors are unable to interact with peptide/Kbm8 due to structural differences imposed by the mutations. Although the dominant OVA epitope, OVAp, can bind Kbm8, it is not stimulatory to Kb-restricted CTL when presented by Kbm8 (27, 32). This suggests that Kbm8 may be structurally unrecognizable despite the high level of similarity between the Kb and Kbm8 alleles. Yet, the definition of variants of the OVAp peptide that can be recognized in the context of Kbm8 (28, 33) proves that this is not so. OVA-specific Kb-restricted receptors can efficaciously interact with Kbm8 given the right peptide. Thus, the inability of Kbm8 to positively select such receptors is surprising. The notion that the defect is due to the absence an appropriate selecting peptide has been previously raised (27). In this work, we confirm this possibility by defining peptides that restore positive selection of H-2K-restricted OVA-specific T cells in bm8 mice. Such peptides were found by first defining a peptide that could be an agonist when presented by Kbm8 (E2) and then screening TCR contact residue variants of this peptide (Table IGo) using criteria previously shown to be successful in identifying positive selection ligands. Two of these, S2D4 and D4, were able to restore positive selection in bm8 organ cultures. This simple experiment proves that peptide diversity is critically important for positive selection in normal mice and that the absence of an appropriate self peptide can restrict the repertoire.

Why are none of the many self peptides normally presented by Kbm8 able to positively select OVAp-specific T cells? One possibility is that positive selection is highly peptide specific, and we have simply stumbled on an example in which receptors are lost to the repertoire because of this high level of stringency. However, there is much evidence that positive selection is, in fact, not highly peptide specific (1). Indeed, a lower level of specificity for positive selection (compared with negative selection or Ag recognition) is inherent in the affinity model of selection, for which there is mounting evidence (41). This model supposes that low affinity interactions, which are by definition more prevalent, mediate positive selection of thymocytes. Why then does Kbm8 have this positive selection defect? We believe that the data from the randomized peptide library screening shed some light on this issue. Using two different semirandom peptide libraries, we found that peptides that could stimulate OT-I CTL were rare when presented by Kbm8, compared with being abundant when presented by Kb. This suggests that Kbm8 does not easily adopt a conformation that is recognizable by OT-I- like receptors. One difficulty in drawing such a conclusion is that we did not screen a completely randomized octamer library. Thus, perhaps the amino acids phenylalanine at position 5 and leucine at position 8 (which were not varied in either library) restrict the ability of Kbm8 to present to OVAp-specific CTL. This seems unlikely given that these are deeply buried MHC anchor residues that were shown to be shared between Kb and Kbm8 peptide repertoires (24). Therefore, the relative paucity of peptides that can stimulate OT-I CTL by bm8 APC suggests that it is something about the structure of the Kbm8 molecule itself that is very different from Kb, and this difference may be more important for positive selection.

Because Kb and Kbm8 are identical in the {alpha}1 and {alpha}2 helices, it was assumed that the MHC residues that contact the TCR are in a similar conformation in both molecules. This is supported by the fact that the many Kb-specific Abs also bind Kbm8 (21). The amino acid differences between Kb and Kbm8 lie in the peptide binding groove, and indeed, it has been shown that the two molecules present distinct peptide repertoires (25). This has pointed to a role for the bound peptide in cases where Kb and Kbm8 function differently (27, 32, 39), but it is known that the MHC-bound peptide can influence the conformation of the {alpha}1 and {alpha}2 residues (40, 42, 43). Thus, we considered the possibility that OVAp-like peptides force the {alpha}1 and {alpha}2 helices to lie in different conformations in Kbm8 compared with Kb, thus altering the structure of the peptide-MHC complex. This was confirmed directly with the Kb-specific Abs 5F1 and 100.30. These Abs recognizes both Kb and Kbm8 on spleen cells or tumor cells, where a variety of peptides are displayed (32). However, when the MHC molecules were stabilized at the cell surface with single peptides, we found that in some cases the epitope was completely destroyed on Kb, but not on Kbm8, or vice versa. The loss of an Ab epitope could be due to either steric blockade by bulky side chains of the peptide (40) or a conformational change in the structure of the MHC residues (39). We favor the latter possibility for two reasons. First, the peptides E4 and Q4 give opposite staining patterns with 5F1. These two side chains are similar in size; thus, it is difficult to imagine that one would sterically block Ab recognition, while the other would not. It is possible, however, that the charge differences between the glutamic acid and glutamine could account for the differences in Ab recognition. However, it is surprising that this charge difference does not affect Ab recognition when the peptide is presented by Kbm8. Secondly, Kb and Kbm8 stain differently with Ab 100.30 when bound to the PTP peptide. In a previous analysis of 128 peptide variants, the pattern of staining on Kb with 100.30 did not suggest steric hindrance by peptide side chains (40). These data, then, suggest that the two alleles can adopt different "subconformations" in a peptide-dependent manner. This may explain the inability of bm8 to activate CTL, despite its ability to bind the cognate ligand (32, 39). We hypothesize that the defect in positive selection in bm8 mice lies in the relative inability of Kbm8 to adopt the appropriate subconformation which is critical for engaging OT-I like receptors. Because this subconformation is driven by the groove-associated peptide, the defect in selection can be restored by certain peptides, like S2D4 and D4.

In summary, self peptides contribute to the diversity of MHC-peptide "faces" in two ways. First, the solvent-exposed peptide side chains, which vary greatly among self peptides, contribute to diversity, and that this diversity can indeed influence selection has already been shown (6). Furthermore, the side chains buried in the MHC, which have a more limited diversity among self peptides (44), can also impose different conformations on the MHC {alpha} helices. Our experiments presented here suggest that this diversity of peptide-dependent MHC conformations can influence positive selection as well.


   Acknowledgments
 
We thank Janko Nikolic-Zugic for peptides and mice, Nick Gascoigne for mice, and Marc Gavin and Mike Bevan for bacterial libraries. We also thank members of the Hogquist and Jameson laboratories for helpful discussion and critical review of the manuscript.


   Footnotes
 
1 This work was supported by National Institutes of Health Grants AI39560 (K.A.H.) and AI38903 (S.C.J.) and a Searle Scholars Award (K.A.H.). Back

2 Address correspondence and reprint requests to Dr. Kristin A. Hogquist, Box 334 FUMC, 420 Delaware St. SE, Minneapolis, MN 55455. E-mail address: Back

3 Abbreviations used in this paper: tg, transgenic; FTOC, fetal thymic organ culture; DP, double positive; OVAp, OVA257–264 SIINFEKL; ß2m, ß2-microglobulin. Back

Received for publication October 13, 1999. Accepted for publication January 18, 2000.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Sebzda, E., S. Mariathasan, T. Ohteki, R. Jones, M. Bachmann, P. Ohashi. 1999. Selection of the T cell repertoire. Annu. Rev. Immunol. 17:829.[Medline]
  2. Bevan, M. J.. 1997. In thymic selection, peptide diversity gives and takes away. Immunity 7:175.[Medline]
  3. Hogquist, K. A., M. A. Gavin, M. J. Bevan. 1993. Positive selection of CD8+ T cells induced by major histocompatibility complex binding peptides in fetal thymic organ culture. J. Exp. Med. 177:1469.[Abstract/Free Full Text]
  4. Ashton-Rickardt, P. G., L. Van Kaer, T. N. M. Schumacher, H. L. Ploegh, S. Tonegawa. 1993. Peptide contributes to the specificity of positive selection of CD8+ T cells in the thymus. Cell 73:1041.[Medline]
  5. Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, F. R. Carbone. 1994. T cell receptor antagonist peptides induce positive selection. Cell 76:17.[Medline]
  6. Jameson, S. C., K. A. Hogquist, M. J. Bevan. 1994. Specificity and flexibility in thymic selection. Nature 369:750.[Medline]
  7. Ashton-Rickardt, P. G., A. Bandeira, J. R. Delaney, L. Van Kaer, H.-P. Pircher, R. M. Zinkernagel, S. Tonegawa. 1994. Evidence for a differential avidity model of T cell selection in the thymus. Cell 76:651.[Medline]
  8. Sebzda, E., V. A. Wallace, J. Mayer, R. S. M. Yeung, T. W. Mak, P. S. Ohashi. 1994. Positive and negative thymocyte selection induced by different concentrations of a single peptide. Science 263:1615.[Abstract/Free Full Text]
  9. Smyth, L., O. Williams, R. Huby, T. Norton, O. Acuto, S. Ley, D. Kioussis. 1999. Altered peptide ligands induce quantitatively but not qualitatively different intracellular signals in primary thymocytes. Proc. Natl. Acad. Sci. USA 95:8193.[Abstract/Free Full Text]
  10. Hu, Q., C. R. B. Walker, C. Girao, J. T. Opferman, J. Sun, J. Shabanowitz, D. F. Hunt, P. G. Ashton-Rickardt. 1997. Specific recognition of thymic self peptides induces the positive selection of cytotoxic T lymphocytes. Immunity 7:221.[Medline]
  11. Hogquist, K. A., A. J. Tomlinson, W. C. Kieper, M. A. McGargill, M. C. Hart, S. Naylor, S. C. Jameson. 1997. Identification of a naturally occurring ligand for thymic positive selection. Immunity 6:389.[Medline]
  12. Nakano, N., R. Rooke, C. Benoist, D. Mathis. 1997. Positive selection of T cells induced by viral delivery of neopeptides to the thymus. Science 275:678.[Abstract/Free Full Text]
  13. Ignatowicz, L., W. Rees, R. Pacholczyk, H. Ignatowicz, E. Kushnir, J. Kappler, P. Marrack. 1997. T cells can be activated by peptides that are unrelated in sequence to their selecting peptide. Immunity 7:179.[Medline]
  14. Pawlowski, T. J., M. D. Singleton, D. Y. Loh, R. Nerg, U. D. Staerz. 1996. Permissive recognition during positive selection. Eur. J. Immunol. 26:851.[Medline]
  15. Ignatowicz, L., J. Kappler, P. Marrack. 1996. The repertoire of T cells shaped by a single MHC/peptide ligand. Cell 84:521.[Medline]
  16. Miyazaki, T., P. Wolf, S. Tourne, C. Waltzinger, A. Dierich, N. Barois, C. Benoist, D. Mathis. 1996. Mice lacking H2-M complexes, enigmatic elements of the MHC class II peptide-loading pathway. Cell 84:531.[Medline]
  17. Marten, W. D., G. G. Hicks, S. K. Menduatta, H. I. Leva, H. E. Ruley, L. Van Kaer. 1996. H-2 M mutant mice are defective in the peptide loading of class II molecules, antigen presentation and T cell repertoire selection. Cell 84:543.[Medline]
  18. van Santen, H. M., A. Woolsey, P. G. Ashton Rickardt, L. Van Kaer, E. J. Baas, A. Berns, S. Tonegawa, H. L. Ploegh. 1995. Increase in positive selection of CD8+ T cells in TAP1-mutant mice by human ß2-microglobulin transgene. J. Exp. Med. 181:787.[Abstract/Free Full Text]
  19. Grubin, C. E., S. Kovats, P. deRoos, A. Y. Rudensky. 1997. Deficient positive selection of CD4 T cells in mice displaying altered repertoires of MHC class II-bound self peptides. Immunity 7:197.[Medline]
  20. Barton, G. M., A. R. Rudensky. 1999. Requirement for diverse, low-abundance peptides in positive selection of T cells. Science 283:67.[Abstract/Free Full Text]
  21. Hammerling, G. J., E. Rusch, N. Tada, S. Kimura, U. Hammerling. 1982. Localization of allodeterminants on H-2Kb antigens determined with monoclonal antibodies and H-2 mutant mice. Proc. Natl. Acad. Sci. USA 79:4737.[Abstract/Free Full Text]
  22. Pfaffenbach, G. M., H. Uehara, J. Geliebter, S. G. Nathenson, D. H. Schulze. 1991. Analysis of the H-2Kbm8 mutant: correlation of structure with function. Mol. Immunol. 28:697.[Medline]
  23. Hunt, H. D., J. K. Pullen, R. F. Dick, J. A. Bluestone, L. R. Pease. 1990. Structural basis for Kbm8 alloreactivity: amino acid substitutions on the ß-pleated floor of the antigen recognition site. J. Immunol. 145:1456.[Abstract]
  24. Van Bleek, G. M., S. G. Nathenson. 1991. The structure of the antigen-binding groove of major histocompatibility complex class I molecules determines specific selection of self peptides. Proc. Natl. Acad. Sci. USA 88:11032.[Abstract/Free Full Text]
  25. Rohren, E. M., L. R. Pease, H. L. Ploegh, T. N. M. Schumacher. 1993. Polymorphisms in pockets of major histocompatibility complex class I molecules influence peptide preference. J. Exp. Med. 177:1713.[Abstract/Free Full Text]
  26. Pullen, J. K., H. D. Hunt, L. R. Pease. 1991. Peptide interactions with the Kb antigen recognition site. J. Immunol. 146:2145.[Abstract]
  27. Nikolic Zugic, J., M. J. Bevan. 1990. Role of self peptides in positively selecting the T-cell repertoire. Nature 344:65.[Medline]
  28. Barnden, M. J., W. R. Heath, F. R. Carbone. 1997. Down-modulation of CD8 beta-chain in response to an altered peptide ligand enables developing thymocytes to escape negative selection. Cell. Immunol. 175:111.[Medline]
  29. Robey, E. A., B. J. Fowlkes, J. W. Gordon, D. Kioussis, H. von Boehmer, F. Ramsdell, R. Axel. 1991. Thymic selection in CD8 transgenic mice supports an instructive model for commitment to a CD4 or CD8 lineage. Cell 64:99.[Medline]
  30. Koller, B. H., P. Marrack, J. W. Kappler, O. Smithies. 1990. Normal development of mice deficient in ß2M, MHC class I proteins, and CD8+ T cells. Science 248:1227.[Abstract/Free Full Text]
  31. Gavin, M. A., B. Dere, III A. G. Grandea, K. A. Hogquist, M. J. Bevan. 1994. Major histocompatibility complex class I allele-specific peptide libraries: identification of peptides that mimic an H-Y T cell epitope. Eur. J. Immunol. 24:2124.[Medline]
  32. Nikolic-Zugic, J., F. R. Carbone. 1990. The effect of mutations in the MHC class I peptide binding groove on the cytotoxic T lymphocyte recognition of the Kb restricted ovalbumin determinant. Eur. J. Immunol. 20:2431.[Medline]
  33. Dyall, R., D. H. Fremont, S. C. Jameson, J. Nikolic-Zugic. 1996. T cell receptor (TCR) recognition of MHC class I variants: intermolecular second-site reversion of an MHC mutation by substituted peptides provides evidence for peptide/MHC conformational variation. J. Exp. Med. 184:253.[Abstract/Free Full Text]
  34. Davey, G. M., S. L. Schober, B. T. Endrizzi, A. K. Dutcher, S. C. Jameson, K. A. Hogquist. 1998. Pre-selection thymocytes are more sensitive to TCR stimulation than mature T cells. J. Exp. Med. 188:1867.[Abstract/Free Full Text]
  35. Lucas, B., I. Stefanova, K. Yasutomo, N. Dautigny, R. N. Germain. 1999. Divergent changes in the sensitivity of maturing T cells to structurally related ligands underlies formation of a useful T cell repertoire. Immunity 10:367.[Medline]
  36. McGargill, M. A., K. A. Hogquist. 1999. Antigen-induced co-receptor down-regulation on thymocytes is not a result of apoptosis. J. Immunol. 162:1237.[Abstract/Free Full Text]
  37. Chidgey, A., R. Boyd. 1997. Agonist peptide modulates T cell selection thresholds through qualitative and quantitative shifts in CD8-coreceptor expression. Int. Immunol. 9:1527.[Abstract/Free Full Text]
  38. Hogquist, K. A., J. L. Bonnevier. 1998. Development of peptide-selected CD8+ T cells in fetal thymic organ culture occurs via the conventional pathway. J. Immunol. 161:3396.
  39. Chattopadhyay, S., M. Theobald, J. Biggs, L. A. Sherman. 1994. Conformational differences in major histocompatibility complex-peptide complexes can result in alloreactivity. J. Exp. Med. 179:213.[Abstract/Free Full Text]
  40. Hogquist, K. A., III A. G. Grandea, M. J. Bevan. 1993. Peptide variants reveal how antibodies recognize major histocompatibility complex class I. Eur. J. Immunol. 23:3028.[Medline]
  41. Alam, S. M., P. J. Travers, J. L. Wung, W. Nasholds, S. Redpath, S. C. Jameson, N. R. J. Gascoigne. 1996. T-cell receptor affinity and thymocyte positive selection. Nature 381:616.[Medline]
  42. Bluestone, J. A., S. Jameson, S. Miller, R. Dick. 1992. Peptide-induced conformational changes in class I heavy chains alter major histocompatibility complex recognition. J. Exp. Med. 176:1757.[Abstract/Free Full Text]
  43. Fremont, D. H., M. Matsumura, E. A. Stura, P. A. Peterson, I. A. Wilson. 1992. Crystal structures of two viral peptides in complex with murine MHC class I H-2Kb. Science 257:919.[Abstract/Free Full Text]
  44. Rammensee, H. G., K. Falk, O. Rotzschke. 1993. Peptides naturally presented by MHC class I molecules. Annu. Rev. Immunol. 11:213.[Medline]
  45. Falk, K., O. Rotzschke, S. Stevanovic, G. Jung, H. G. Rammensee. 1991. Allele-specific motifs revealed by sequencing of self peptides eluted from MHC molecules. Nature 351:290.[Medline]
  46. Tallquist, M. D., T. J. Yun, L. R. Pease. 1996. A single T cell receptor recognizes structurally distinct MHC/peptide complexes with high specificity. J. Exp. Med. 184:1017.[Abstract/Free Full Text]
  47. Malarkannan, S., F. Gonzalez, V. Nguyen, G. Adair, N. Shastri. 1999. Alloreactive CD8+ T cells can recognize unusual, rare, and unique processed peptide/MHC complexes. J. Immunol. 157:4464.[Abstract]
  48. Malarkannan, S., M. Afkarian, N. Shastri. 1995. A rare cryptic translation product is presented by Kb major histocompatibility complex class I molecule to alloreactive T cells. J. Exp. Med. 182:1739.[Abstract/Free Full Text]
  49. Malarkannan, S., T. Serwold, V. Nguyen, L. A. Sherman, N. Shastri. 1996. The mouse mammary tumor virus env gene is the source of a CD8+ T-cell-stimulating peptide presented by a major histocompatibility complex class I molecule in a murine thymoma. Proc. Natl. Acad. Sci. USA 93:13991.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 J. Immunol.Home page
A. I. Webb, N. A. Borg, M. A. Dunstone, L. Kjer-Nielsen, T. Beddoe, J. McCluskey, F. R. Carbone, S. P. Bottomley, M.-I. Aguilar, A. W. Purcell, et al.
The Structure of H-2Kb and Kbm8 Complexed to a Herpes Simplex Virus Determinant: Evidence for a Conformational Switch That Governs T Cell Repertoire Selection and Viral Resistance
J. Immunol., July 1, 2004; 173(1): 402 - 409.
[Abstract] [Full Text] [PDF]

Home page
 J. Immunol.Home page
T. K. Starr, M. A. Daniels, M. M. Lucido, S. C. Jameson, and K. A. Hogquist
Thymocyte Sensitivity and Supramolecular Activation Cluster Formation Are Developmentally Regulated: A Partial Role for Sialylation
J. Immunol., November 1, 2003; 171(9): 4512 - 4520.
[Abstract] [Full Text] [PDF]

Home page
 J. Immunol.Home page
T. Dao, J. M. Blander, and D. B. Sant'Angelo
Recognition of a Specific Self-Peptide: Self-MHC Class II Complex Is Critical for Positive Selection of Thymocytes Expressing the D10 TCR
J. Immunol., January 1, 2003; 170(1): 48 - 54.
[Abstract] [Full Text] [PDF]

Home page
 J. Immunol.Home page
Y. Ito, S. Arai, N. S. C. van Oers, I. Aifantis, H. von Boehmer, and T. Miyazaki
Positive Selection by the Pre-TCR Yields Mature CD8+ T Cells
J. Immunol., November 1, 2002; 169(9): 4913 - 4919.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stefanski, H. E.
Right arrow Articles by Hogquist, K. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stefanski, H. E.
Right arrow Articles by Hogquist, K. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS