Antibiotic discovery is always welcome. Here’s a new one from a well-searched area (Actinomycetes extracts), and the authors (university teams from McMaster, Indiana, and Montreal) did a lot of groundwork to make sure that they weren’t just going to rediscover known agents. That’s a serious problem with broad-based antibiotic screening, as you’d imagine, and there are several ways to try to get around it. One is to engineer the organisms being screened – you can try them with and without various resistance mechanisms to see where you stand. But in this case, the search concentrated on wild-type Actinomycetes that showed unusual biosynthetic gene clusters by sequencing. A phylogenetic tree was constructed for these (various non-ribosomal protein synthesis genes) and the search was biased towards the least-homologous ones and the ones that showed the least similarity in known self-resistance genes. That latter part should, in theory, send you more often to the natural products that have new modes of action, as opposed to minor variations on known themes. If you’ve ever wondered how bacteria manage not to kill themselves with their own antibiotics, that’s a big part of the answer: they have their own resistance mechanisms already in the can – often several of them, and their genes tend to be clustered with the biosynthesis genes for the natural product and regulated simultaneously.
This approach led to “corbomycin”, shown at right. It’s broadly similar to the known peptidoglycan antibiotics such as complestatin and chloropeptin I but with some extra ring closures, giving it quite a structure. These agents are known to be mostly active against gram-positive organisms, but they maintain this activity even against vancomycin-resistant strains, which makes their mechanism of action of interest. There have been some proposals in the past, but this work identifies that actual MOA as binding to peptidoglycan and thus inhibit a range of autolysin activity, which inhibits cell wall synthesis. Inhibiting any single autolysin enzyme would probably be a losing strategy, since there’s a lot of redundant activity in that area, so the bind-to-the-natural-substrate mechanism is a better way to go and should (as in the case of vancomycin) lead to slower development of resistance.
The paper demonstrates topical activity in a mouse model. It’s unknown what might happen when you try to dose corbomycin systemically, although you’d have to figure that some significant formulations work might be involved, since that’s not a very soluble-looking compound. But I’m glad to see it and also glad to see the mechanism of this whole class illuminated. Now what we need is something like this that messes with the gram-negatives, too!