STRUCTURAL STUDIES OF GENE REGULATORY PROTEINS 
Carl O. Pabo, Ph.D., Associate Investigator 
Dr. Pabo's research has focused on the structure 
and design of proteins that regulate gene expres- 
sion. The laboratory is attempting to understand 
how proteins recognize specific sites on double- 
stranded DNA and how the bound proteins regu- 
late gene expression. This information will eventu- 
ally be used to design novel DNA-binding proteins 
for research, diagnosis, and therapy. During the 
past year, Dr. Pabo's research has used a high-reso- 
lution crystal structure of the X repressor-operator 
complex as the basis for continued genetic and 
structural analysis of repressor-operator interac- 
tions. Crystallographic studies of several other 
DNA-binding proteins are in progress, and the labo- 
ratory has continued attempts to crystallize key reg- 
ulatory proteins from the human immunodefi- 
ciency virus (HIV). The laboratory also has 
continued development of a database of protein- 
DNA interactions and development of programs for 
computer-aided protein design. Several tight-bind- 
ing variants of the \ repressor have been designed 
and are being prepared for experimental tests. 
I. Structural and Genetic Studies of Repressor- 
Operator Interactions. 
Prokaryotic repressors provide useful model sys- 
tems for the study of protein-DNA interactions, and 
Dr. Pabo's laboratory is studying the repressor from 
bacteriophage X. The laboratory recently solved the 
crystal structure of a complex that contains the 
DNA-binding domain of repressor and a 20 base 
pair synthetic operator site. This structure showed 
how the helix-turn-heltx motif is used for DNA rec- 
ognition and led to important new insights about 
protein-DNA interactions. Refinement and analysis 
of this model have continued during the past year. 
Crystallographic refinement has shown how the ex- 
tended amino-terminal arm fits into the major 
groove near the center of the operator site. Since 
this structural motif has not been seen in other re- 
pressor-operator complexes, site-directed mutagen- 
esis was used to define the critical contacts. These 
studies show that three lysine residues in the arm 
play a central role in site-specific recognition. 
Crystallographic studies of another prokaryotic 
repressor— the arc repressor from Salmonella bac- 
teriophage P22— are in progress. Genetic data sug- 
gest that this protein does not use the helix-turn- 
helix motif for recognition, and the first heavy-atom 
derivative has given an initial low-resolution map of 
the structure. The search for other heavy-atom de- 
rivatives is in progress, and the laboratory has ob- 
tained small cocrystals of the arc repressor-operator 
complex. 
II. Physical and Structural Studies of Eukaryotic 
DNA-binding Proteins. 
The homeodomain is a conserved DNA-binding 
domain that was discovered in a set of proteins that 
regulate Drosophila development and was later ob- 
served in many other eukaryotic regulatory pro- 
teins. Although the intact proteins often are much 
larger, the homeodomain itself contains about 60 
amino acids. Sequence comparisons and nuclear 
magnetic resonance (NMR) studies have indicated 
that the homeodomain contains a helix-turn-helix 
motif Since about 100 homeodomains have been 
sequenced and since these bind to closely related 
DNA sites, the homeodomains provide an interest- 
ing system for studying protein-DNA interactions. 
The laboratory is attempting to crystallize represen- 
tative homeodomains from yeast, Drosophila, and 
humans. Dr. Cynthia Wolberger has recently grown 
small cocrystals that contain the DNA-binding do- 
main of the a2 repressor (which regulates mating 
type expression in yeast) and a duplex DNA frag- 
ment that contains two a2 binding sites. In collabo- 
ration with Dr. Thomas Kornberg (University of Cal- 
ifornia at San Francisco), the laboratory also has 
grown microcrystals of the homeodomain from the 
engrailed gene oi Drosophila. 
Another structural motif— the zinc finger do- 
main—has also been discovered in eukaryotic DNA- 
binding proteins. These domains each contain 
about 30 amino acids and have conserved cysteine 
and histidine residues that bind to zinc. To under- 
stand this structural motif and its role in nucleic 
acid recognition, Dr. Pabo and his colleagues have 
been studying peptides that contain zinc finger do- 
mains. The first peptide to be synthesized and char- 
acterized was a single finger that corresponded to 
the second domain from transcription factor IIIA of 
Xenopus. This peptide folds in the presence of zinc, 
but it binds DNA nonspecifically. The laboratory has 
now synthesized tandemly repeated domains that 
have higher nonspecific binding constants, and ex- 
periments are in progress to determine whether 
these peptides also show site-specific binding. (The 
Continued 
591 
