New Approach to Antibiotics

A novel type of conjugate of carbohydrates and protein--known as a "glycodendriprotein"--may provide a new route to antibiotics.

Chemist Benjamin G. Davis of the University of Oxford, in England; microbiologist Marjorie M. (Kelly) Cowan of Miami University of Ohio, Oxford; and coworkers have shown that by linking branched carbohydrate-tipped structures known as glycodendrimers to a protein-degrading enzyme, they can inhibit the infectivity of the bacterium Actinomyces naeslundii [J. Am. Chem. Soc., 126, 4750 (2004)].

Davis and his colleagues are targeting bacteria earlier in the infection process than most antibiotics do. "We're trying to block infection before it even gets going," he says. "The glycodendrimer inhibits the binding, but then the enzyme that is attached to the glycodendrimer swings around, chews up the protein on the surface, and renders it unable to grab hold of the host it wants to infect."

The glycodendrimers consist of sugar molecules on the tips of an antenna-like scaffold formed by such dendrimeric core molecules as TREN [tris(2-aminoethyl)amine] or mesitylene. Testing glycodendrimers with different numbers of antennae, the researchers found that a pair of antennae, with a total of two sugar residues, was the optimal number for the antibacterial system. For an alternative carbohydrate-binding protein, they found that a pair of branched antennae displaying four sugars was best.

The glycodendrimers are attached to the protease subtilisin through a cysteine linkage. Because subtilisin contains no naturally occurring cysteines, the positions where cysteines are inserted determine the locations of the carbohydrate on the protein. The carbohydrate targets and binds to receptors on the surface of A. naeslundii. Then the subtilisin chews up the protein adhesin on the bacterial cell surface. The carbohydrates successfully direct an otherwise nonspecific protease to the target protein. "We deliberately chose an enzyme that was very broad so that all we had to do was retool the enzyme with the appropriate targeting ligand," Davis says.

Davis believes the approach of synthesizing glycoproteins is flexible enough to be applied to a variety of systems. "There are a vast array of carbohydrate-binding proteins out there," he says. "You can basically pick the carbohydrate you need and then retool the end of the dendrimer for that function."

Carolyn R. Bertozzi, professor of chemistry at the University of California, Berkeley, who also studies glycoconjugates, calls this work an interesting approach to blocking bacterial adhesion and targeting receptors for degradation.

"While bacterial proteins can become resistant to drugs via rapid mutation, the actual carbohydrate-binding residues in these bacterial adhesins are under pressure to remain conserved, as they are vital for the colonizing activity of the bacterial cell," Bertozzi says. "Thus, any approach that targets the carbohydrate-binding activity is less likely to suffer from the selection of resistant strains. Although its reduction to clinical practice is a ways off, as a concept, I think this is an interesting combination of antiadhesive therapy and receptor-mediated drug delivery that will prompt new ways of thinking about antibiotic design."