processing mutant T2 suggests a role for MHC- 
linked genes in class II antigen presentation. //m- 
miinol 148:2586-2590. 
Roche, P. A., Marks, M.S., and Cress well, P. 1991. 
Formation of a nine-subunit complex by HLA class 
II glycoproteins and the invariant chain. Nature 
354:392-394. 
Spies, T., Cerundolo, V., Colonna, M., Cresswell, 
P., Townsend, A., and DeMars, R. 1992. Presenta- 
tion of viral antigen by MHC class I molecules is 
dependent on a putative peptide transporter he- 
terodimer. Nature 355:644-646. 
Wei, M.L., and Cresswell, P. 1992. HLA-A2 mole- 
cules in an antigen-processing mutant cell con- 
tain signal sequence-derived peptides. Nature 
356:443-446. 
MOLECULAR APPROACHES TO LYMPHOCYTE RECOGNITION AND DIFFERENTIATION 
Mark M. Davis, Ph.D., Investigator 
T lymphocytes play a number of critical roles in 
the immune system. Some subsets of T cells are able 
to kill virally infected or transformed cells directly; 
others seem specifically designated to mobilize 
other cells, particularly B cells, in the course of an 
immune response. Both activities can be mediated 
through the same recognition apparatus, the a:/? T 
cell receptor (TCR) heterodimer, in close associa- 
tion with the CD3 polypeptides. A central feature of 
T cell recognition by this receptor is that antigens 
are often (if not always) "seen" as peptide frag- 
ments complexed with either class I or class II mole- 
cules of the major histocompatibility complex 
(MHC) . One goal of the laboratory is to understand 
at a biochemical and structural level how this is ac- 
complished topologically, with what affinity, and 
whether this phenomenon is distinguishable funda- 
mentally from antibody-antigen interactions. 
These pursuits have involved expressing both 
TCR and MHC class II heterodimers in a lipid-linked 
form so that they may be easily cleaved off cell sur- 
faces and used to reconstitute T cell recognition (to- 
gether with specific peptide antigens) . Having tens 
of milligrams of soluble TCR and MHC proteins has 
enabled Dr. Davis and his colleagues to embark on 
many projects. The development of a new strategy to 
delineate which parts of the TCR are contacting 
which residues on the peptide is also providing new 
information on the topology and "rules of engage- 
ment" concerning T cell recognition. 
Another goal is to use transgenic mice to analyze 
the selection of specific receptor chains and hetero- 
dimers in the thymus. This should be informative 
with respect to both the removal of self-reactive 
TCRs in the thymus (negative selection) and the en- 
hanced maturation and export of T cells having a 
good "fit" with one or more of the thymic MHC 
molecules (positive selection). These mice are also 
a valuable source of physiologically normal T cells 
at different stages of differentiation to use in defin- 
ing the requirements for activation and differentia- 
tion of these cells in vivo. 
Deriving a Topology for TCR-mediated 
Recognition of Peptide-MHC Complexes 
Dr. Davis and others have proposed a model 
(based on patterns of TCR diversity and MHC and 
immunoglobulin [Ig] structures) in which the V-J 
junctional regions of TCRs are largely or solely re- 
sponsible for peptide binding, whereas germline V 
region-encoded loops might contact the a helices 
of MHC molecules. Correlative data from many labo- 
ratories indicate an important role for the V-J junc- 
tional sequences in T cell specificity, with specific 
residues sometimes appearing in V„, V^, or both in 
response to a specific peptide-MHC combination. 
In one series of experiments. Dr. Davis and his 
colleagues have attempted without success to 
transfer peptide specificity from one TCR to another 
by swapping V(D)J junctional sequences and other 
CDR loops. As negative results in this type of experi- 
ment are difficult to interpret, the group recently 
developed a way of addressing the issue that has 
turned out to be informative. The method works by 
1 ) introducing changes in the peptide at positions 
that alter T cell reactivity but not MHC binding, 2) 
immunizing mice that are transgenic for either a- or 
jS-chain TCR that recognizes the original peptide (to 
hold half of the TCR "constant"), and then 3) ana- 
lyzing the endogenous TCR usage of responding T 
cells. In this fashion the laboratory has elicited re- 
ciprocal changes in TCR a- and /3-chain CDR3 se- 
quences in response to changed amino acids on the 
peptide. This shows that there is direct TCR-peptide 
contact, that it is largely the province of CDR3 resi- 
dues, and that both a and |8 chains can participate in 
peptide recognition. 
One surprising outcome of the above studies is 
322 
