Compensatory mutations

There is another factor related to the development of significant antibiotic resistance in bacteria that has not been discussed with reference to mycobacteria. When components of essential macromolecular components (such as ribosomes) are altered by mutation, as in antibiotic-resistant strains, the mutants are frequently growth defective, which can lead to reduction in the pathogenicity of the antibiotic-resistant strains as compared to the drug-sensitive wild-type. This may be more apparent in the case of an intracellular pathogen that is exposed to severe growth-limiting stress in the host. Recent research with Gram-negative pathogens has revealed the possibility of additional factors associated with the acquisition of successful antibiotic resistance genotypes; these factors warrant study in mycobacteria as well.

Several researchers (Lenski 1997, Levin et al 1997, Schrag & Perrot 1996) have studied the effects of the acquisition of an antibiotic resistance plasmid, or a mutation conferring antibiotic resistance, on the ability (fitness) of bacteria to survive in competition with wild-type organisms or to tolerate a stressful environment (pathogenesis). It was found that the change from sensitive to resistant had negative effects on fitness; in other words, the antibiotic-resistant mutants were enfeebled. However, during prolonged growth of antibiotic-resistant strains, variants arose spontaneously that were as fit as the wild-type strains without a change in the resistant phenotype. This restoration of fitness was due to compensatory mutations occurring intra- or extragenically with reference to the original resistance allele. Acquisition of such mutations led to resistant strains that could compete successfully with wild-type strains for growth in a hostile environment (Table 2). Even more striking are experiments by Bjorkman et al (1998) which have analysed the appearance of virulent streptomycin-resistant S. typhimurium in mouse infection studies. While the majority of infecting organisms did not survive, rapid selection of fast-growing (surviving) strains took place; these strains had acquired additional mutations (sometimes intracodon!) compensating the growth defect without altering the resistant phenotype. In some cases, two compensatory mutations were identified. These analyses were extended to mutants resistant to rifampicin (rpoB) and nalidixic acid (gyrA) in S. typhimurium, and similar multiple mutations were identified. The general conclusion is that all antibiotic-resistant microbial pathogens are probably multiple mutants: that is, the initial mutation to generate the resistance phenotype, plus the compensating mutation(s) to maintain virulence. There may be more to resistance than meets the eye!

It is reasonable to ask if the same situation applies to mycobacteria; do streptomycin-resistant and rifampicin-resistant strains of M. tuberculosis isolated during the course of infection possess both target modification and associated compensation? Is this likely to be true for all cases of antibiotic resistance in mycobacterial pathogens? In an analysis of rpoB mutations conferring rifamycin resistance (Cole 1994), it was noted that double mutations and deletions were quite common in M. tuberculosis (see Fig. 2.). Are these the result of compensation? Early studies with E. coli showed that there is antagonism between rpsL and rpoB mutations (Chakrabarti & Gorini 1975, 1977); yet coexisting mutations of this kind are common in multidrug-resistant M. tuberculosis. Are additional compensatory changes needed to re-establish the fitness of such strains? Is the primary resistance mutation growth limiting, and does the secondary mutation restore the normal growth characteristics of the organism? Or, on the contrary, since M. tuberculosis grows so sluggishly at the best of times, there may be no need for compensatory mutations in this genus (Cole 1994).

TABLE 2 Overcoming the 'cost' of resistance: restoring fitness






rps'L*, rpsD, rpsE,?






gyrA *,?

^represents second-site mutations; ? represents additional unidentified mutations.

^represents second-site mutations; ? represents additional unidentified mutations.

FIG. 2. The multiplicity of mutations leading to rifampicin resistance in the gene for the beta subunit of RNA polymerase (rpoB) in Escherichia coli (above sequence) and Mycobacterium tuberculosis (below sequence). For details see Cole (1994).

In the studies mentioned above it should be noted that compensatory mutations, when isolated from their cognate resistance mutation, often reduce survival fitness of the host. Comparable studies with mycobacteria need to be carried out. What biochemical mechanisms may be involved in determining fitness? Do compensatory mutations simply prevent mutant reversion or suppress the adverse effect(s) of the structural mutant? Do the compensatory mutations modulate interaction between pathogen and host? Finally, it will be of considerable interest to examine the genotype of streptomycin- or clarithromycin-resistant isolates that have altered rRNA genes (rrnA or rrnB) to see what type of compensatory change, if any, is associated with this type ofmutation.

The development of resistance to antibiotics by bacteria is inescapable. Unhappily, the resistance mechanisms and the means by which they are established in human mycobacterial infections are still not well understood. The presence of compensatory mutations may complicate the situation, since the fitness 'cost' of resistance is likely to be a significant factor in determining both survival and virulence. Combinations of drugs and fully maintained courses of therapy are the most obvious and successful approaches to the treatment of susceptible strains at the present time. For multidrug-resistant strains, the only available treatment appears to be the use of cocktails of different classes of antibiotics, which must be rational combinations based on good biochemical evidence of drug action. However, novel antibiotics are needed, and if what I have proposed is correct, the use of anti-mutators might be justified, although molecules with this property of lowering mutation rates have not been well studied and have never been employed in medicine.

When all is said and done, it seems that the development of an effective vaccine is the ultimate solution to the problem of tuberculosis.

I wish to thank D. Davies for invaluable assistance in preparing this manuscript, and the Canadian Bacterial Diseases Network and the National Science and Engineering Council of Canada for their support of my research leading to these cogitations.

0 0

Post a comment