Protein Crystallography in the Study of Infectious Diseases 
subunits are very different in size, and the degree 
of amino acid sequence similarity is, if anything, 
less than what one would expect for two random 
sequences. Nonetheless, the structural similarity 
implies some distant evolutionary relationship 
between these families. 
Pseudomonas aeruginosa Pilin 
Pseudomonas aeruginosa is an opportunistic 
pathogen that infects burn victims and immuno- 
compromised patients. It is also one of the major 
pathogens infecting the lungs of cystic fibrosis 
patients. Filaments on its surface, termed pili, at- 
tach to epithelial cell surfaces, promoting colon- 
ization. Pili are formed from a helical array of 
identical pilin subunits. We have crystallized pi- 
lin from this organism (in collaboration with Wil- 
liam Paranchych, University of Alberta) and are 
attempting to improve the quality of the crystals 
to allow us to determine its structure. The three- 
dimensional structure would help us understand 
the details of cell-surface binding, which could 
be used to devise strategies to interfere with 
colonization. 
Computer-aided Drug Design 
Most of the drugs in use today were discovered 
by trial and error, and many of these have un- 
known mechanisms of action. Those with known 
mechanisms usually act by binding specifically to 
a drug receptor, often a protein molecule. Since 
crystallography allows us to examine the struc- 
tures of receptors and the details of their interac- 
tions with drugs, it should help in improving ex- 
isting drugs or even in inventing new ones. 
Progress has been made through crystallography 
in the former objective, but designing new drugs 
from scratch is still extremely difficult. We are 
trying to address this problem with the aid of 
computers. 
To design a new drug, one must first choose an 
appropriate receptor or target. In the design of an 
antibacterial drug, the key factor is selective tox- 
icity-, the drug should poison the pathogen but 
not the patient. One might choose, as a drug tar- 
get, an essential enzyme in a biochemical path- 
way unique to the bacterium, because that would 
minimize the chances of side reactions with host 
enzymes. In our work on drug design, we are as- 
suming that a good potential drug target has been 
chosen and its crystal structure determined. The 
problem, then, is how to exploit the structural 
information. There are probably billions of com- 
pounds that might be used as drugs; determining 
which of these might bind to, and interfere with, 
the target protein is far from trivial. 
We have chosen a "divide and conquer" ap- 
proach to drug design to reduce the magnitude of 
this problem. The vast set of possible compounds 
is made up of various combinations of a much 
smaller set of molecular fragments. We propose 
to design drugs by using a computer to dock 
members of a library of fragments to the region of 
the desired binding site, then to combine docked 
fragments to form chemically sensible mole- 
cules. We have tested the feasibility of fragment 
docking, and the results are sufficiently promis- 
ing that we will go on to test the feasibility of the 
next step. Fragment docking requires the calcula- 
tion of binding energies, which as yet can only be 
approximate and requires a great deal of com- 
puter time. But we believe that this method can 
provide a useful tool for drug design in the not- 
too-distant future, in view of the acceleration in 
the speed of computers and the improvements 
that are continually being made in the under- 
standing of molecular interactions. 
Our work is also supported by the Medical Re- 
search Council of Canada and the Alberta Heri- 
tage Foundation for Medical Research. 
A schematic illustration of the crystal structure 
of the B (binding) subunit of verotoxin-1, 
viewed from the top of the pentamer. Arrows 
indicate strands of j3-sheet, and cylinders indi- 
cate ahelices. 
From Stein, P.E., Boodhoo, A., Tyrrell, GJ., 
Brunton, f.L., and Read, R.J. 1992. Nature 
355:748-750. Copyright © 1992 Macmillan 
Magazines Limited. 
522 
