Antibiotic resistance among infectious bacteria is an increasing problem worldwide, resulting in large part from the overuse of antibiotics. Writing in Nature Biotechnology, López-Igual et al.1 demonstrate a nifty way to selectively poison antibiotic-resistant Vibrio cholerae bacteria — the species that causes cholera — from the inside. The authors’ aim is to offer a highly targeted alternative to standard broad-brush antibiotics.
Our present scattergun overuse of antibiotics has caused several problems, one being the emergence of antibiotic-resistant bacteria. Another is that typical broad-spectrum antibiotics affect not only the target disease-causing (pathogenic) bacteria but also our normal beneficial bacteria, which protect us against infection and might influence many other aspects in humans, including weight, mood and allergies (see go.nature.com/31x0csa). The specificity of the approach proposed by López-Igual and colleagues could avoid both of these issues.
The authors’ method builds on the ability of bacteria to transfer certain pieces of genetic material (‘mobilizable’ DNA) on cell-to-cell contact, in a process known as conjugation. López-Igual et al. take advantage of this phenomenon to transfer a set of genes that encode a toxin (a protein called CcdB) and its antidote (a protein known as CcdA) from donor bacteria into their neighbours. The system is designed so that the toxin will be made only in V. cholerae and the antidote will be made only in the V. cholerae bacteria that are antibiotic-sensitive, so that just antibiotic-resistant V. cholerae will be killed (Fig. 1).
López-Igual et al. used several clever tricks to ensure that toxicity occurred only in the target cells. First, they engineered the toxin-encoding genes to be under the control of a Vibrio-specific protein, the transcription factor ToxR (which is essential for V. cholerae to cause disease). This means that, if the toxin-encoding genes were ever transferred by conjugation to other bacterial species, no toxin could be made. Second, to ensure that any antibiotic-sensitive V. cholerae recipients do not die, the authors included the gene that encodes the antidote on the mobilizable gene set. This antidote is made in all antibiotic-sensitive V. cholerae but not in antibiotic-resistant V. cholerae, because in the latter bacteria, antidote production is turned off by a repressor protein, SetR, which is encoded by, and also regulates, the antibiotic-resistance genetic element SXT in these bacteria.
However, simply having the antidote present might not be enough to prevent the killing of bacteria that are not intended to be targeted. The CcdB toxin is highly potent and acts rapidly, killing bacteria by causing extensive damage to their genetic material — it acts by inhibiting an enzyme called gyrase, locking it to DNA, which results in DNA breaks. The authors therefore built in a delay mechanism — generating a ticking time-bomb that becomes deadly only after some time.
To do this, they inserted a genetic module that encodes a special protein known as an intein into the toxin gene. Expression of the modified toxin gene produces a non-functional toxin, from which the intein protein excises itself over time through a process called splicing, thereby generating the functional toxin. The effect is to delay the deadly action of the toxin, allowing time for bacteria that receive it to respond. The time it takes for the toxin to mature allows a bacterium that is antibiotic-sensitive to produce enough antidote to survive. If the bacterium is antibiotic-resistant, however, no antidote is produced and, after the toxin has matured, the cell will die.
López-Igual et al. went on to show that their approach is not limited to a single type of toxin protein, but that others — namely, HigB2 (which targets a type of enzyme called an mRNase), RelE4 (which inhibits protein synthesis) and ParE2 (another gyrase inhibitor) — are also functional after intein-mediated splicing. Finally, they tested their method in three natural V. cholerae habitats: water, zebrafish larvae and crustacean larvae. They found that their approach could eradicate antibiotic-resistant V. cholerae in all three habitats. The specific regulators they used work only for V. cholerae, but the system could easily be adapted to target different bacteria.
Other studies2,3 have used bacterium-infecting viruses as well as conjugation methods to deliver DNA- or RNA-digesting enzymes called nucleases to target drug-resistant bacteria. How does López-Igual and colleagues’ approach compare with these? One advantage of their system is that fewer bacteria that can resist the internal threat evolve, with around one such ‘escape mutant’ per 107 bacteria that receive the toxin, which is a level of escape mutants that is least a hundred times lower than is found with a virus-based approach2,3.
Nevertheless, this escape rate is not low enough to prevent the development of resistance. One factor to consider is the typical population sizes of the bacteria being targeted4. If the population is larger than the size predicted to generate an escape mutant, then resistant bacteria will already be present. People who have cholera produce around 108 V. cholerae bacteria per gram of faecal material, and water reservoirs associated with an outbreak of the disease would probably host an even higher number of bacteria5. In such large populations, it is probable that thousands of mutant cells would be resistant to the toxin, and would be unaffected by the killing system described here. Further research should attempt to work out the mechanisms behind this resistance, and to find ways to optimize the system.
A problem with conjugation approaches in general is that they are inefficient, with only a few cells out of a hundred actually receiving the genes. Two things will probably be necessary to generate a functional therapy: first, the use of several toxins and/or delivery systems, to try to limit the number of escape mutants; and second, improvements in the efficiency of gene transfer. Although López-Igual and colleagues’ system will probably not immediately solve the problem of antibiotic-resistant cholera infections, it might make an important contribution to the arsenal of alternative treatments for critically ill people.
One final question is perhaps more profound. The severe watery diarrhoea that is characteristic of cholera causes an estimated 100,000 deaths every year6. With this grim statistic in mind, why target just the antibiotic-resistant bacteria when you could try to kill them all instead? A possible benefit could be a lower selection pressure on V. cholerae to become resistant to this new killing method. However, one would need to know in advance that the infection is indeed antibiotic-resistant, or it might be necessary also to use antibiotics in parallel to treat the infection. Perhaps the system described by López-Igual et al. should be viewed as an intriguing proof-of-concept of how selective antibiotic alternatives could be used in the future: it could easily be modified to target all V. cholerae, whether or not they are antibiotic-resistant.
Nature 570, 449-450 (2019)