Computer software helps discover new antibiotics

Researchers have discovered two new promising antibiotics effective against dangerous bacterial infections by sifting through the human microbiome with the help of a computer software.
Using computational methods to identify which genes in a microbe's genome ought to produce antibiotic compounds and then synthesising those compounds themselves, researchers at Rockefeller University in the US were able to discover these antibiotics without having to culture a single bacterium.
Researchers began by trawling publicly available databases for the genomes of bacteria that reside in the human body.
They then used specialised computer software to scan hundreds of those genomes for clusters of genes that were likely to produce molecules known as non-ribosomal peptides that form the basis of many antibiotics.

They also used the software to predict the chemical structures of the molecules that the gene clusters ought to produce. The software initially identified 57 potentially useful gene clusters, which researchers winnowed down to 30.
Researchers then used a method called solid-phase peptide synthesis to manufacture 25 different chemical compounds.

By testing those compounds against human pathogens, they successfully identified two closely related antibiotics, which they dubbed humimycin A and humimycin B.
Both are found in a family of bacteria called Rhodococcus - microbes that had never yielded anything resembling the humimycins when cultured using traditional laboratory techniques.
The humimycins proved especially effective against Staphylococcus and Streptococcus bacteria, which can cause dangerous infections in humans and tend to grow resistant to various antibiotics.

Further experiments suggested that the humimycins work by inhibiting an enzyme that bacteria use to build their cell walls - and once that cell-wall building pathway is interrupted, the bacteria die.
A similar mode of action is employed by beta-lactams, a broad class of commonly prescribed antibiotics whose effect often wanes as bacteria develop ways to resist them.
The scientists found that one of the humimycins could be used to re-sensitise bacteria to beta-lactams that they had previously outsmarted.
In one experiment, they exposed beta-lactam resistant Staphylococcus microbes to humimycin A in combination with a beta-lactam antibiotic, and the bugs once again succumbed.

Remarkably, that held true even when humimycin A had little effect by itself - a result that Brady attributes to the fact that both compounds work by interrupting different steps in the same biological pathway.
The study appears in the journal Nature Chemical Biology.