MACROMOLECULES OF THE CENTRAL DOGMA OF MOLECULAR BIOLOGY 
Thomas A. Steitz, Ph.D., Investigator 
The laboratory of Dr. Steitz has been using single- 
crystal x-ray crystallography combined with molecu- 
lar genetics to elucidate the structure and function 
of proteins and nucleic acids that are involved in 
DNA replication, recombination, transcription (and 
its regulation), and some aspects of protein synthe- 
sis. Major achievements of the past year include de- 
termination of the structures of 1) a Klenow frag- 
ment • DNA complex showing bound duplex DNA 
and 2) HIV reverse transcriptase complexed with 
the nonnucleotide inhibitor Nevirapine. 
Transcription and Its Regulation 
The catabolite gene activator protein (CAP) from 
Escherichia coli is a dimer of 22,500-molecular- 
weight subunits that activates transcription from 
certain E. co// operons in the presence of cAMP. The 
crystal structure of this sequence-specific DNA- 
binding protein shows that each subunit consists of 
a cAMP-binding domain and a smaller domain in- 
volved in binding the DNA. 
The crystal structures of CAP cocrystallized with 
either a 30- or 32-bp DNA fragment have been 
solved. The most striking result is that the DNA du- 
plex is kinked 43° at two positions in both solved 
crystals, but the bending at the ends of the DNA 
differs as a result of differences in crystal packing. 
Data at 2.8-A resolution from cocrystals of CAP com- 
plexed with a 46-bp fragment should yield a struc- 
ture in which crystal-packing effects on the end of 
DNA are different. The role of this CAP-induced DNA 
bending in transcription activation is being pursued 
by biochemical and structural studies of CAP and 
RNA polymerase subunits complexed with pro- 
moter DNA. 
Recombination 
The recA protein of E. coli plays a major and es- 
sential role in general recombination by catalyzing 
the formation of three- and four-strand DNA struc- 
tures that involve homologous pairing. The crystal 
structure of recA protein has been refined to 2.3-A 
resolution. The protein forms the same helical fila- 
ment in the crystal as seen at far lower resolution in 
electron micrographs of recA bound to DNA. 
Similarities of the ADP binding site to the OTP 
binding site of the ras p21 oncogene have led to 
hypotheses concerning the allosteric mechanism of 
coupling between ATP and DNA binding that may be 
general for many NTPases. Crystals of a recA-ATP 
analogue complex have been grown, and attempts 
to crystallize complexes with DNA are being 
pursued. 
The next enzyme in the recombination pathway that 
cleaves the HoUiday junction, ruvC, has been crystal- 
lized in a form suitable for structure determination. 
(These studies on recombination have been supported 
by the National Institutes of Health.) 
DNA Synthesis 
Important progress has been made during the past 
year in refining the structure of Klenow fragment 
complexed with 1 1 base pairs of duplex DNA con- 
taining a three-nucleotide 3' overhang that is ob- 
served to bind in the 3 -5' exonuclease active site. 
This editing complex shows duplex DNA bound at a 
right angle to the large cleft that contains the poly- 
merase active site. This position for the duplex prod- 
uct of DNA synthesis implies that the DNA must 
make a large bend to put the primer terminus into 
the polymerase active site. (This research was sup- 
ported by the American Cancer Society.) 
The most significant advance in the previous year 
has been the preliminary determination of the struc- 
ture of HIV reverse transcriptase (RT) complexed 
with a noncompetitive inhibitor, Nevirapine, dis- 
covered at Boehringer Ingelheim. Complexation 
with this inhibitor proved crucial to obtaining crys- 
tals that diffract to 3.0-A resolution. The 3.5-A reso- 
lution structure of the p66 /p5 1 heterodimer is strik- 
ingly asymmetric: while the polymerase domain of 
p66 has a large cleft analogous to that of Klenow 
fragment, p51 shows no such cleft. One of the four 
polymerase subdomains has the same structure and 
conserved catalytic residues as Klenow fragment, 
suggesting that the catalytic subdomains of polymer- 
ases evolved from a common ancestor. 
An A-form RNA-DNA hybrid can be model-built 
into the deep cleft that extends between the poly- 
merase and RNase H active sites. Nevirapine binds in 
a pocket adjacent to the DNA and at the base of a 
thumb-like protrusion. Mutations in RT that are re- 
sistant to Nevirapine alter protein side chains that 
contact the inhibitor. They presumably lower the 
affinity of the enzyme for this component by chang- 
ing the shape of the binding site. Mutations resistant 
to AZT (3'-azido-3'-deoxythymidine) and DDI (di- 
deoxyinosine) are in residues that contact the tem- 
plate strand, making the structural interpretation of 
their phenotype unclear. (The reverse transcriptase 
studies were supported in part by the National Insti- 
tutes of Health.) 
STRUCTURAL BIOLOGY 489 
