STRUCTURAL MOLECULAR BIOLOGY 
Don C. Wiley, Ph.D., Investigator 
Dr. Wiley's laboratory has determined the struc- 
tures of membrane glycoproteins in order to study 
the mechanism of their biological activities. Cur- 
rently the structures of three polymorphic pro- 
teins—the influenza virus hemagglutinin (HA), the 
human class I histocompatibility antigens HLA-A2 
and HLA-A28, and the variable-surface glycoprotein 
from trypanosomes— are being studied by high-res- 
olution x-ray crystallography In each case, bio- 
chemical, physical, chemical, and recombinant DNA 
methods are being used to study the function of 
the proteins. 
I. Influenza Virus. 
The influenza hemagglutinin glycoprotein has at 
least three activities during viral infection. 1) It 
binds the virus to a cell surface by recognizing a si- 
alic acid receptor. 2) It accomplishes virus-cell entry 
by fusing the virus membrane to the cell's mem- 
brane after being triggered by the low pH of an en- 
docytic vesicle. 3) It undergoes structural changes 
from year to year that result in antigenic variation 
and the reoccurrence of influenza epidemics in pre- 
viously infected populations. 
A. Receptor binding. The structures of the HA from 
the 1968 Hong Kong virus and two single-amino 
acid substitution mutants were determined com- 
plexed with trisaccharide analogues of the cellular 
receptors of the virus. Those structures have now 
been refined, and a molecular model for the recog- 
nition of sialic acid by the virus has been published. 
The laboratory has begun synthesizing sialoside de- 
rivatives with substituents at various positions on 
the saccharide ring, both to permit further x-ray 
analyses to confirm the current model by locating 
specific positions on the ring and to explore which 
positions will accept substituents and still fit into 
the HA-binding site. A 4-0 acetylated sialoside has 
been complexed with crystalline HA, and x-ray data 
are being collected. 
A nuclear magnetic resonance (NMR) binding 
assay developed in collaboration with Dr. George 
Whitesides (Harvard University) has been used to 
determine the equilibrium dissociation constant be- 
tween hemagglutinin from wild-type and mutant vi- 
ruses with a(2,6)- and a(2,3)-linked sialosides. The 
dissociation constants are weak, ~2 mM. They are 
the same for monosaccharides and trisaccharides. 
supporting the crystal structure data, which indi- 
cate that the HA interacted strongly with only the 
sialic acid in a trisaccharide. 
A large NMR chemical shift observed for the sialic 
acid A^-acetyl methyl group is also consistent with 
the location of that methyl group over a tryptophan 
ring, as observed in the crystal structure. Such a 
ring amount shift would be expected from the crys- 
tal structure, arguing that the crystal structure of 
the complex is the same as the solution structure 
reported by the NMR signal. 
Efforts to design molecules to block receptor 
binding are beginning. 
B. Membrane fusion. A series of single-site muta- 
tions has been constructed on a cloned gene of the 
HA to explore the membrane fusion activity of the 
HA. Mutations that destabilize the trimer interac- 
tion in the globular HA^ chain raise the pH opti- 
mum for membrane fusion, arguing that the do- 
mains must rearrange during membrane fusion. 
Disulfide bonds introduced to lock the trimer inter- 
faces covalently with the HA^ domains destroy fu- 
sion activity while retaining receptor binding and 
antibody recognition. This indicates that these do- 
mains must be able to rearrange for fusion to occur. 
The experiments suggest that if the trimer interface 
could be stabilized by a drug molecule, membrane 
fusion, and therefore virus infectivity could be pre- 
vented. 
The structure of one of the single-site substitu- 
tions. Asp- 11 2 HA^ to Gly mutant HAs with an in- 
creased pH optimum for membrane fusion, has 
been determined by x-ray diffraction. It shows a 
loss of four intramolecular hydrogen bonds that 
stabilize the location of the amino-terminal "fusion 
peptide" of the HA, suggesting the origin of the mu- 
tant phenotype. 
II. Human Histocompatibility Antigens. 
The structure of HLA-A2, a class I histocompati- 
bility antigen, has been refined to 2.7 A resolution. 
The refined structure reveals details of the putative 
antigenic peptide found in a prominent cleft on the 
molecular surface. A series of well-defined pockets 
that appear to be recognition sites for amino acid 
side chains of processed antigens are now visible in 
the cleft. 
A second human histocompatibility antigen, HLA- 
Continued 
611 
