Fitness gains hamper efforts to tackle drug resistance

It has long been assumed that resistance to antibiotics reduces the fitness of disease-causing bacteria, but experiments on Salmonella Typhi, the bacteria that causes Typhoid fever, are now challenging this view.

T he effectiveness of the antibiotics used to treat diseases caused by bacteria is increas ingly being limited by the spread of resist ance to these drugs. For a long time it was thought that most of the mutations that lead to antimicrobial resistance in bacteria led to a reduc tion in fitness, which meant that the resistant bacteria did not grow as fast as those that were susceptible to antimicrobials. These fitness costs are important because they determine how quickly resistance is likely to develop, or to disappear if antibiotic use is discontinued (Andersson and Hughes, 2010). However, resistant bacteria often develop secondary mutations that can compen sate for these fitness costs, and these mutations allow the bacteria to thrive, even in antimicrobial free environments (Lenski, 1998). When this occurs, resistance is likely to continue even if antibiotic use is stopped. This is why public health bodies around the world are so worried by the spread of antibiotic resistance.
Typhoid fever is caused by eating food or drinking water contaminated with Salmonella Typhi bacteria. According to the World Heath Organisation (WHO), the drug of choice for treat ing typhoid fever is ciprofloxacin, which belongs to the fluoroquinolone family of antibiotics (WHO, 2003). However, resistance to the fluoroquin olones is increasing all over the world (Parry and Beeching, 2009). Now, in eLife, Stephen Baker of the Oxford University Clinical Research Unit in Vietnam and coworkers in Vietnam and the UK have compared the fitness of various mutants of Salmonella Typhi that are resistant to fluoroquin olones with the fitness of a strain that is still suscep tible to these drugs (Baker et al., 2013).
The fluoroquinolones target enzymes that untangle knots in bacterial genomes. Resistance to these antibiotics in Salmonella Typhi involves changes in these target enzymes that reduce the ability of the drugs to bind to them. These changes occur by substitution mutations in two genes, gyrA and/or parC, that encode these enzymes. Advances in sitedirected mutagenesis allowed Baker and coworkers to construct eleven fluoroquinoloneresistant mutants and one sus ceptible control mutant with known nucleotide substitutions from a susceptible, weakened labo ratory strain of Salmonella Typhi. The resistant mutants were generated by substitutions at two positions in the gyrA gene and one in the parC Copyright Dutta. This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

ANTIMICROBIAL RESISTANCE
Fitness gains hamper efforts to tackle drug resistance It has long been assumed that resistance to antibiotics reduces the fitness of disease-causing bacteria, but experiments on Salmonella Typhi, the bacteria that causes Typhoid fever, are now challenging this view. Image DNA pyrosequencing (y-axis) and colony counting (x-axis) provide similar estimates for the frequency of antibiotic resistance mutations gene, which were introduced in combinations to create single, double or triple mutants. Seven of the eleven labengineered mutants have been found in clinical isolates, whilst the remaining four had not been observed in nature before.
To start each in vitro competition assay, a series of cultures containing equal numbers of bacteria from the susceptible parent strain and one of the 12 labengineered mutant strains were prepared. These cultures were then grown in the absence of any antibiotics for several generations. Every 24 hr, some of each culture mixture was removed and analysed ( Figure 1). More culture was also diluted into fresh antibioticfree growth media, and the entire process was repeated for a period of 15 days. Colony counting was used to calculate the number of bacterial generations during each period.
Whether each mutant had a selective advan tage or disadvantage was determined from the frequency of the engineered mutated genes at each time point, which was measured by specific pyrosequencing with the extracted DNA. Statistical models were used to calculate the selection co efficient and fitness difference of each mutant when compared to the parent strain. Since the fate of a particular mutation depends on its interaction with its environment, including all the other genes found in the genome (Cordell, 2002), the effects of interactions between genes, often known as 'epistasis', were also calculated for the double and triple mutants.
Interestingly, six resistant mutants (two single mutants and the four double mutants tested) could outcompete the susceptible Salmonella Typhi parent strain. This indicates that these mutations conferred increased fitness, even in the absence of the antimicrobial. This goes against the dogma that resistance mutations always have a fitness cost, and can explain the frequent occur rence of certain resistant mutants in the anti biotic free environment. The single mutant with highest selective advantage (called S83F) con tributed 1.3% more offspring per generation than the parent strain, and the most competitive dou ble mutant (S83FD87N, which has two changes within the gyrA gene) contributed 7.4% more offspring per generation. In contrast, the triple mutant (S83FD87GS80I) contributed 1% less offspring per generation than the parent strain: this indicates a negative interaction between these mutations, with the third mutation wiping out the fitness advantage of the double mutants.
Since Salmonella Typhi lives within host cells during an infection, and is restricted to humans, Figure 1. Fluoroquinolone-resistant mutants of Salmonella Typhi vs susceptible strains. Fluoroquinolone-resistant mutants (red) with either one, two or three mutations in the gyrA and parC genes were cultured with a fluoroquinolone-susceptible parent strain (white) in liquid media that did not contain any antibiotics. Every 24 hr, over a period of 15 days, these growth cultures were subcultured into fresh antibiotic-free media, and samples were taken for colony counting and DNA extraction. Pyrosequencing of extracted DNA was used to determine the frequency of each mutation and the parent strain at each time-point, and to calculate the fitness of each mutant compared to the parent strain. Increases in fitness allowed some single and double mutants (red arrow) outcompeted the parent strain, while a reduction in fitness meant that the triple mutant lost out to the parent strain (blue arrow).
the results obtained in this in vitro experimental system using a weakened, lab strain would not be the same with wild strain and ex vivo or in vivo models. However, as acknowledged by Baker et al., all models have their own inherent limita tions and currently there are no available systems (except humans and primates) that exactly mimic Salmonella Typhi infection cycle.
The process of evolution in any organism is determined by the order in which mutations occur (Kimura, 1985). Baker et al. have con cluded that single mutations in gyrA gene that reduce susceptibility to fluoroquinolones (espe cially the S83F mutation) provide the starting point for the evolution of more resistant double mutants, which maintain the resistance, even in the absence of antibiotics. This would contribute to the global spread of fluoroquinoloneresistant Salmonella Typhi, and would also undermine efforts to control the spread of drugresistant typhoid through the prudent use of antimicro bial drugs.

Shanta Dutta is in the Bacteriology Division, National
Institute of Cholera and Enteric Diseases, Kolkata, India shanta1232001@yahoo.co.in Competing interests: The author declares that no competing interests exist.