concanavalin A sequence have failed to produce an 
IgG response that recognizes concanavalin A. Other 
turn sequences have been selected from known 
structures, v^^ith the additional criteria that the turn 
is known to be immunogenic. The important find- 
ing from this first design of a hybrid immunogen is 
that the conformation of the concanavalin A (3-turn 
was successfully transferred to a host protein in a 
native conformation. 
III. Genetic Analysis of a (i-Turn. 
A genetic system has been developed to explore 
the amino acid sequence requirements for the for- 
mation of a (3-turn, distant from the active site in 
staphylococcal nuclease (residues 27-31). An 
M13mpl8 construct has been prepared that allows 
an oligodeoxynucleotide mixture to be cloned into 
the nuclease gene, which has a random distribution 
of bases over codons 27-31. This oligonucleotide 
mixture was cloned into the vector to produce a li- 
brary of nuclease mutants, each with a unique se- 
quence at the P-turn site. Clones that displayed en- 
2ymatic activity on assay plates (~3-8%) were 
selected for DNA sequencing. Analysis of 130 ^-turn 
sequences from active nucleases indicates strong 
preferences for and against particular amino acids 
at each of the five positions. A further mutagenesis 
experiment, varying only the central two P-turn res- 
idues (28-29), resulted in a high proportion of en- 
zymatically active nuclease variants. A denaturant- 
dependent activity assay has been developed to 
estimate the impact of these substitutions on the 
stability of the nuclease variants. A series of "intra- 
turn" complementation mutants indicate structural 
interactions between sites within the P-turn. Rules 
for (3-turn formation may be derived from this ex- 
perimental approach. A determination of the influ- 
ence of these P-turn substitutions on the thermody- 
namic stability of pure protein variants is in 
progress. 
IV Role of Amino Acid Sequence in Determining 
P-Turn Type. 
Staphylococcal nuclease occurs in at least two 
folded conformations that are in slow exchange on 
the nuclear magnetic resonance (NMR) time scale. 
The interconversion of these folded conformers 
and the thermal unfolding kinetics have been inves- 
tigated by magnetization-transfer NMR experi- 
ments. This conformational heterogeneity is due to 
cis-trans isomerization about the peptide bond pre- 
ceding proline 117, which is predominantly in the 
cis configuration in the crystal structure. The pro- 
line isomerism hypothesis is also supported by vari- 
ant proteins prepared by site-directed mutagenesis, 
where proline 117 has been changed to a glycine 
(P117G) or a threonine (P117T). NMR spectra of 
these mutants do not indicate the conformational 
heterogeneity present in the wild-type protein. A 
number of additional amino acid substitutions in 
this P-turn region display altered cis-trans equilib- 
rium constants. Thus the free-energy contribution 
of amino acid residues to the stabilization of type 
VI P-turns can be evaluated using these methods. 
The P117G and P117T nuclease variants have 
been crystallized both with and without Ca^^ and a 
diphosphonucleotide inhibitor. The structures have 
been solved and refined at high resolution. The 
P117G mutation causes residues 115-118 to adopt 
a type I' P-turn structure, while the P117T muta- 
tion has produced a type I turn geometry. 
Studies of how amino acid substitutions in this 
region affect the equilibrium and kinetics of the 
isomerization, and the resulting crystal structures, 
will lead to a better understanding of the forces 
that stabilize protein molecules. 
V De Novo Enzyme Design. 
Dr. Fox and his colleagues are working toward 
the total design of a folded protein structure that 
should hydrolyze amino acid esters based on an 
a + p backbone fold. The approach has been to se- 
lect a protein fold and active-site geometry and 
then define an amino acid sequence that will favor 
that fold. Cycles of design, peptide synthesis, and 
detailed physical and structural analysis, followed 
by redesign of protein substructures and the final 
fold, should provide insight into the factors in- 
volved in stabilizing folded protein structures and 
in catalysis. A stable amphipathic a-helix has been 
designed as an independent element of the final 
structure. Studies are in progress to define the P- 
sheet component of the protein, using two 
stranded P-hairpin model systems. 
Dr. Fox is also Associate Professor of Molecular 
Biophysics and Biochemistry at Yale University. 
Continued 
578 
