PROKARYOTIC CHROMOSOME ORGANIZATION 
David R. Hillyard, M.D., Assistant Investigator 
I. Bacterial Genetics: Histone-like Proteins of 
Bacteria. 
The organization of bacterial DNA into a highly 
condensed yet functional form is dependent on its 
interaction with a family of ubiquitous scaffolding 
proteins. Included among them is a class of small 
and usually basic proteins whose direct and in- 
direct resemblance to eukaryotic histones has led 
to their being called histone-like proteins. Cur- 
rently almost nothing is known about the role 
of these proteins in vivo. The major activity of 
this laboratory is to identify and manipulate the 
genes for this class of proteins in Salmonella 
typhimurium. 
The most abundant of the histone-like proteins, 
HU, is extraordinarily conserved among bacteria. 
HU is able to bind and condense DNA, leading to 
the formation of nucleosome-like beaded struc- 
tures. In vitro HU facilitates site-specific recombina- 
tion needed for the inversion reaction of Salmo- 
nella flagellar phase variation, and it is essential for 
the transposition of bacteriophage Mu. The genes 
for the two Salmonella HU subunits, hupK and 
hupB, have been identified and inactivated. Sur- 
prisingly, the huph, hupB double mutants are 
viable, indicating that HU is not an essential pro- 
tein. However, gross alterations in the rates of in 
vivo Mu transposition, F plasmid stability, and 
flagellar phase variation are seen in cells that lack 
HU. When hupA and hupB mutant cells are tested, 
subunit-specific phenotypes are revealed. HU also 
seems to be required to maintain the normal distri- 
bution of supercoiled plasmid topoisomers. In ad- 
dition, HU mutant strains overexpress a 17 kDa 
protein that has properties linking it to the family 
of histone-like proteins. The hypothesis that 
changes in the distribution of this protein or other 
type II DNA-binding proteins compensate for the 
loss of HU is being explored. An insertional muta- 
tion in a histone-like gene of Salmonella has re- 
cently been generated that results in an extreme 
growth-defective phenotype only in huph, hupB 
double mutant backgrounds. In addition to HU, 
null mutations have now been generated for the 
Salmonella hns and himK genes, which also en- 
code proteins of this class. The long-range goal is to 
use genetics to reveal interactive phenotypes 
among the extended family of bacterial histone-like 
proteins. 
II. Conotoxin Genes. 
Among venomous animals, cone snails have de- 
veloped the capacity to envenomate an unusually 
broad diversity of natural prey successfully and in 
doing so have evolved an impressive array of toxins. 
Each species of Conus makes dozens of specific 
small toxic peptide ligands, which in many cases 
have high affinity to components of the nervous 
system. Little is known about the organization or 
evolution of toxin genes in any eukaryotic system. 
It is reasonable to think that the demands for toxin 
adaptation imposed on venomous animals may 
have led to novel genetic solutions. Work is contin- 
uing on the characterization of toxin genes from 
the moUusk-hunting snail, Conus textile. The full- 
length cDNA clones for the 27-amino acid King- 
Kong peptide have been obtained by screening 
cDNA libraries prepared from venom duct RNA. The 
sequence of the cDNA clone predicts an 80-amino 
acid propeptide with a single Lys-Arg processing 
site adjacent to the amino-terminal toxin amino 
acid. Using DNA probes specific for the amino-ter- 
minal portion of the propeptide, Dr. Hillyard's 
group has identified two additional Conus textile 
toxins. The structures of the three predicted pro- 
peptide molecules have been compared. In addi- 
tion to a highly conserved amino-terminal region 
(which is presumably processed from the final 
toxin), the King-Kong peptides contain a perfectly 
conserved arrangement of six cysteines within the 
processed toxin segments. However, between cyste- 
ine residues there is absolutely no amino acid con- 
servation. These cloning results and the fact that 
only a narrow range of cysteine arrangements oc- 
curs among conotoxin sequences suggest an un- 
usual pattern of diversification for these molecules. 
In addition, these results suggest how conotoxin 
molecules specifically fold into one specific disul- 
fide pattern (reduced and reoxidized peptide toxins 
assume many inactive configurations in addition to 
the active one). It is possible that information in- 
trinsic to the conserved amino-terminal portion of 
the propeptide is needed to ensure specific folding. 
Dr. Hillyard and his co-workers have recently 
constructed cDNA libraries from additional species 
of Conus, to expand the database to toxin families 
with dissimilar structural features and patterns of 
amino acid modification. In addition, work has 
begun on both the genetics and biochemistry of a 
Continued 
223 
