studied the specificity of cytotoxic T lymphocytes 
(CTLs) derived from mice infected with Listeria 
monocytogenes. He isolated one clone (CD8^ with 
a/? receptors) that reacted with infected cells of sev- 
eral H 2 haplotypes. This clone detects a peptide, 
Fr38, isolated by high-pressure liquid chromatogra- 
phy as fraction 38, which is produced by the bacte- 
ria rather than the infected cells and has a blocked 
amino terminus. The clone reacts with Fr38-pulsed 
target cells from many strains but not from Mta~ 
strains, including the B10.CAS2 fibroblast line, and 
not with |82m-deficient cell lines. Because the clone 
recognizes Fr38-pulsed B10.CAS2 cells transfected 
with a cosmid that carries H-2M3'' as its only class I 
gene, it was concluded that H-2M3 presents the Lis- 
teria peptide. 
The Listeria-specific T cell clone failed to recog- 
nize Fr38 on cells from a number of wild-derived 
strains that can present the Mta peptide to bulk- 
cultured CTLs. These results suggest that the M3 
molecules from the wild and laboratory mice differ 
by a few amino acids, enough to abolish recognition 
by a single receptor species but not by a wider re- 
pertoire of receptors. These new alleles, which are 
being characterized by sequencing, show that 
H-2M3 is more polymorphic than was originally sus- 
pected. In this respect too, the gene is a neoclassical 
class I gene. Like classical class I antigens, H-2M3 is 
expressed from before day 8 of embryonic life, and 
its expression is inducible with interferon-7. 
H-2M2 and M3 are the most distal class I genes in 
the mouse MHC. Dr. Elsy Jones has isolated one yeast 
artificial chromosome (YAC) with M2 and two with 
M3, all in the range of 250 kb. Surprisingly, these 
clones contain no other MHC class I genes (which 
otherwise tend to be clustered less than 20 kb apart 
in the central D, Q, and T regions of the mouse 
MHC). By pulsed-field electrophoresis. Dr. Jones 
has also succeeded in linking the M1-M7-M8 and 
the M4-M5-M6 class I gene clusters from the proxi- 
mal part of the M region on an 870-kb Notl and a 
245-kb Sfll fragment. The spatial isolation of the 
expressed M2 and M3 genes may contribute to their 
conservation, protecting them from loss by unequal 
recombination between tandemly aligned, homolo- 
gous genes. 
RMA-S Cells and Qa-1 
RMA-S mutant cells produce empty, unstable class 
I molecules lacking peptides, and they cannot pre- 
sent antigens derived from endogenously synthe- 
sized proteins; surface display of stable, complete 
class I molecules can be restored by addition of ap- 
propriate synthetic peptides. Evan Hermel, a gradu- 
ate student in the laboratory, showed that the RMA-S 
mutation also affects display of the medial class I 
antigens Mta and Qa-1. In a collaboration with Dr. 
John Monaco (Medical College of Virginia, Rich- 
mond) , expression of both medial and major class I 
antigens was restored when the cells were trans- 
fected with a functional Tap- 2 gene, an MHC gene 
that encodes a member of the ABC superfamily of 
transporter proteins. 
In a collaboration with Drs. Carla Aldrich and 
James Forman (University of Texas Southwestern 
Medical School at Dallas), RMA-S cells were used to 
dissect the peptide specificity of Qa-1 ''-specific 
CTL clones. Qa-1 was the second medial class I anti- 
gen to be described; the Qa-l'' antigen of RMA-S 
cells is encoded by the T23 gene. Up to half the 
clones generated from a secondary in vitro response 
lyse RMA-S cells at least as well as RMA. These clones 
may recognize empty Qa-1 molecules or, more 
likely, Qa-1 molecules with a peptide, such as a sig- 
nal peptide, that does not require a functional TAP 
transporter to enter the endoplasmic reticulum. The 
rest of the clones require a functional Tap- 2 gene 
and hence recognize peptides that are not otherwise 
transported into the endoplasmic reticulum. 
Among such clones, some fail to recognize target 
cells that are homozygous for the H-2D region of the 
k haplotype; Qdm, a gene in this region, affects the 
ability of target cells to be recognized by anti-Qa-1 
CTL. The simplest explanation for this observation 
is that Qa- 1 can present a peptide derived from any 
H-2D molecule except H-2D'', and this peptide re- 
quires a functional TAP transporter too. 
A further distinction can be made among Tap-2- 
dependent, (Q^/m-independent clones. Some recog- 
nize RMA-S target cells that have been treated with 
oligomycin, an inhibitor of mitochondrial ATPase, 
which also increases lysis by Mta-specific CTLs, or 
with ionomycin, a calcium ionophore. The increase 
in intracellular calcium caused by the drugs may 
stimulate proteolysis, and the Tap-2 defect may be 
overcome by the increased concentration of pep- 
tides. One clone kills RMA-S treated with oligomy- 
cin but not ionomycin. 
Interspecies Divergence of B2m 
Mice are the only mammals in which allelic forms 
of /32m have been documented. Because of its inti- 
mate interaction with the peptide-binding a\ and 
al domains of MHC class I molecules, allelic differ- 
ences of i82m can subtly affect the antigenic struc- 
ture of MHC class I molecules and their ability to 
bind particular peptides. Evan Hermel has se- 
quenced the second exon of the B2m gene from 
several species of wild mice. This exon encodes 92 
of the 99 amino acids in mature /32m. Whereas four 
IMMUNOLOGY 325 
