The discovery of penicillin in the 1920s revolutionised the treatment of infection, with a host of life-threatening diseases becoming simple to cure. But the bacteria fought back, mutating to render themselves immune to the drugs. Resistance is rising, with methicillin-resistant Streptococcus aureus, or MRSA, making headlines, and new drugs are desperately needed to fight the new generations of bacteria.

Yet there has been no truly new class of antibiotics in 15 years, with medicinal chemists relying on modifying existing drugs to create new ones. But the chemistry can be difficult in the lab, never mind at a production scale. As bacteria are extremely efficient at creating antibiotic substances, perhaps they could be genetically modified to make new ones. And this is one of the focuses of Joe Spencer’s research.

‘We’re still carrying out classical biosynthetic studies aimed at working out how particular molecules were made,’ Joe explains. ‘But now, with the advent of molecular biology and the ability to sequence genes very rapidly, it’s become easier to find the genes encoding the enzymes that synthesise natural products. And having found the genes, we can tinker with them, and programme them to modify the antibiotic structure. The resistance mechanisms are known for all the most important classes of antibiotics, which means there is a fighting chance of getting around the problem.’

As techniques advance and knowledge of more pathways become clear, Joe believes it may become possible to bolt together genes from different bacteria. This approach could give rise to novel compounds as well as accessing natural products that are only available in minute amounts.

‘Marine polyketides such as discodermolide, which is a potent anticancer drug, are difficult to obtain because the bacteria in the sponge can’t be grown in the lab,’ Joe explains. ‘With our detailed knowledge of how polyketides are made, it may be possible to include genes from routes we already know, to construct polyketide synthase enzymes capable of making discodermolide.’ This would be very useful, as it would be good to have in a large quantity – but it’s not available by total synthesis as it’s too expensive, and it can’t be cultured easily.

Joe explains that once an interesting compound has been chosen, it’s a case of extracting the genomic DNA from the organism that makes it. Then the DNA is chopped up randomly, and put into a library, which should contain at least one copy of all the genes.

If any information about enzymes in the pathway is known, then it’s a case of fishing for the gene that makes it in the library –it’s easy to work out what enzyme a string of DNA in a gene is coded to make. Then, because all the antibiotic-making genes are clustered together, once you’ve pinpointed the right area of the gene, it’s much easier to sequence the other genes nearby to find out the rest of the enzymes involved.

‘Now, with gene sequencing becoming much faster, it’s possible to randomly sequence parts of the gene,’ he says. ‘As these may be similar to other genes in known antibiotic-making pathways, genes can be found in a hit-and-miss way rather than rationally.’

One of Joe’s numerous research collaborations is with David O’Hagan at St Andrews, looking at how fluorine is incorporated into molecules by bacteria. ‘While fluorine is common in chemicals – and around 20% of all medicines on the market contain fluorine – it’s very rare in nature,’ he explains. ‘The natural abundance of free fluoride ions is very low compared to chloride ions, and also fluoride is difficult to oxidise, which is the major process that enzymes use to incorporate chloride and bromide.’

However, in 2002 a fluorinating enzyme was discovered in the bacterium Streptomyces cattleya. ‘The bug is hard to handle synthetically,’ Joe explains. ‘So we found the genes that make the key enzyme in the pathway – fluorinase – and genetically modified E. coli bacteria, which are much easier to work with, so that they will produce the enzyme in large quantities. This enabled us to crystallise the protein and solve the fluorinase structure, and establish that it catalyses the SN2 attack of fluoride on S-adenosylmethionine.

‘This opens up the possibility of using bugs to make fluorinated molecules in a controlled way,’ Joe concludes, ‘and hopefully we will be able to create some much-needed novel antibiotics.’