T CELL RECOGNITION AND SELECTION 
Mark M. Davis, Ph.D., Associate 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 di- 
rectly and others seem specifically designated to 
mobilize other cells, particularly B cells. Both of 
these activities can be mediated through the same 
recognition apparatus, the a-.^ T cell receptor 
(TCR) heterodimer, in close association with the 
CD3 polypeptides. A central feature of T cell recog- 
nition by the a: (3 receptor is that antigens are often 
(if not always) "seen" as peptide fragments com- 
plexed with either class I or class II histocompati- 
bility molecules. One goal is to try to understand at 
a biochemical and structural level exactly what 
TCRs "see," with what affinities, and whether there 
are any differences from the way antibodies com- 
bine with antigens. Another goal is to use transge- 
nic mice to analyze the selection of specific recep- 
tor chains and combinations in the thymus. This 
analysis should be informative with respect to both 
positive and negative models of thymic selection of 
the TCR repertoire and questions of affinity for an- 
tigen plus major histocompatibility complex (MHC) 
and MHC molecules alone. A third goal concerns 
the nature of the recently described 7:8 TCR found 
on a minority of mature T cells and on the surface 
of some T cell precursors. Transgenic systems in- 
volving the implanting of specific 78 genes in mice 
are providing some clues to the role of this antigen 
receptor in the immune system and their relation- 
ship to the a:P-bearing T cells. 
The T and B lymphocytes are also important as 
model systems for cellular differentiation and lin- 
eage relationships at a molecular level. Toward this 
end the isolation and characterization of possible 
regulatory molecules is useful. Although the studies 
described here deal largely with one such molecule 
(the product of the XLR gene family), the technol- 
ogy now exists to isolate other cell-type-specific, 
nuclear protein genes fairly rapidly. 
I. A Model for T Cell Recognition. 
Almost all of the sequence diversity in the TCR 
8-chain is concentrated in the V-J junctional region, 
which led Dr. Davis and his colleagues to examine 
more closely the issue of TCR diversity versus im- 
munoglobulins. Both TCR heterodimers (a: (3 or 
7:8) have significantly (10^-10^ x) greater diversity 
in their junctional residues than do immunoglobu- 
lin H:L combinations and quite a bit less (10^- 
10' X) diversity in their V:V combinations. This dif- 
ference may be explained by the fact that immuno- 
globulins are binding to single (and often quite 
large) antigenic entities, whereas TCRs are recog- 
nizing small fragments of antigens embedded in 
(larger) MHC molecules. Thus it seems likely that 
the junctional residues of TCR V regions are pri- 
marily interacting with antigens, whereas the re- 
mainder of the V region heterodimer is interacting 
with MHC determinants. Modeling of TCR V re- 
gions based on antibody structures and the recent 
solution of an MHC class I structure tend to sup- 
port this. Tests of this hypothesis involving hyper- 
variable region transfer mutagenesis and ablation 
experiments are under way. 
II. Expression of TCRs and MHC Molecules in a 
Solubilizable Form. 
To study T cell recognition in a cell-free system. 
Dr. Davis and his colleagues have developed a 
method for expressing TCR heterodimers in a form 
easily solubilizable. This involves replacing the 
normal transmembrane and cytoplasmic portions 
of TCR polypeptides with signal sequences for 
lipid-linked surface expression, derived from ei- 
ther decay-accelerating factor (DAF) or human 
alkaline phosphatase (HPAP). In at least one case, 
the resulting chimeric proteins form normal- 
appearing disulfide-linked a: (3 heterodimers. 
and Vp epitopes present on the original TCR 
are preserved in the lipid-linked form, as are 
their approximate spatial relationships. Molecules 
expressed in this way are easily cleavable with the 
enzyme phosphatidylinositol-specific phospho- 
lipase C and isolated from immunoaffinity columns. 
Recently, Dr. Davis used the HPAP signal sequences 
to express the mouse class II MHC molecule, I-E*', 
the MHC ligand for one TCR (2B4) that has been 
expressed in this fashion. Because the 2B4 T cell 
is specific for a fragment of pigeon cytochrome 
c complexed with I-E'', the key elements of this 
event may soon be able to be reconstituted in 
vitro. Quantities are also sufficient to produce 
enough material for structural studies. Thus Dr. 
Davis and his colleagues hope to move the analyses 
of T cell recognition from a cellular readout, of ne- 
cessity indirect, to direct biochemical and molecu- 
lar assays. 
Continued 
381 
