Dihydrofolate reductases in some folate-requiring bacteria with low trimethoprim susceptibility

Trade-offs with stability modulate innate and mutationally acquired drug resistance in bacterial dihydrofolate reductase enzymes

Structural stability is a major constraint on the evolution of protein sequences. However, under strong directional selection, mutations that confer novel phenotypes but compromise structural stability of proteins may be permissible. During the evolution of antibiotic resistance, mutations that confer drug resistance often have pleiotropic effects on the structure and function of antibiotic-target proteins, usually essential metabolic enzymes. In the present study, we show that trimethoprim (TMP)-resistant alleles of dihydrofolate reductase from Escherichia coli (EcDHFR) harboring the Trp30Gly, Trp30Arg or Trp30Cys mutations are significantly less stable than the wild-type, making them prone to aggregation and proteolysis. This destabilization is associated with a lower expression level, resulting in a fitness cost and negative epistasis with other TMP-resistant mutations in EcDHFR. Using structure-based mutational analysis, we show that perturbation of critical stabilizing hydrophobic interactions in wild-type EcDHFR enzyme explains the phenotypes of Trp30 mutants. Surprisingly, though crucial for the stability of EcDHFR, significant sequence variation is found at this site among bacterial dihydrofolate reductases (DHFRs). Mutational and computational analyses in EcDHFR and in DHFR enzymes from Staphylococcus aureus and Mycobacterium tuberculosis demonstrate that natural variation at this site and its interacting hydrophobic residues modulates TMP resistance in other bacterial DHFRs as well, and may explain the different susceptibilities of bacterial pathogens to TMP. Our study demonstrates that trade-offs between structural stability and function can influence innate drug resistance as well as the potential for mutationally acquired drug resistance of an enzyme.

Use of folic acid casei medium reveals trimethoprim susceptibility of Lactobacillus species

AIM

Lactobacilli have been reported to have intrinsic resistance to trimethoprim. The susceptibility of lactobacilli to trimethoprim on different media was investigated in order to search for a phenotypic test method that could indicate the presence of acquired resistance genes.

METHODS AND RESULTS

Strains of Lactobacillus acidophilus, Lact. paracasei, Lact. rhamnosus and Lact. plantarum were susceptibility tested with E-tests on folic acid casei medium (FACM), MRS and defined medium 1. The effects of addition or removal of nucleosides and thymidine phosphorylase were investigated. E-tests on FACM yielded reproducible minimal inhibitory concentrations (MICs) for trimethoprim but addition of nucleosides was necessary for growth of Lact. acidophilus. MICs for the tested strains were 0.125-0.19, 0.25-3 and 0.064-0.19 microg ml(-1) for Lact. paracasei, Lact. rhamnosus and Lact. plantarum, respectively. With the addition of deoxyuridine and deoxyadenosine to FACM the MICs of Lact. acidophilus were 0.064-1 microg ml(-1).

CONCLUSIONS, SIGNIFICANCE AND IMPACT OF THE STUDY

Lactobacilli do not have intrinsic resistance to trimethoprim. The results show that trimethoprim susceptibility testing of the tested Lactobacillus species is possible and indicate that transferable resistance genes are absent in all the tested strains.

Antimicrobial resistance in Staphylococcus aureus

Recognized since 1883 as a common cause of infection, Staphylococcus aureus' preantimicrobial-era bacteremia mortality rate was 82%. The mortality of that era threatens to return as evidence of growing vancomycin resistance undermines the utility of vancomycin therapy. Successful treatment of S. aureus infections requires knowledge of its antimicrobial resistance capacity.

High-level resistance to trimethoprim in clinical isolates of Campylobacter jejuni by acquisition of foreign genes (dfr1 and dfr9) expressing drug-insensitive dihydrofolate reductases

The pathogenic bacterium Campylobacter jejuni has been regarded as endogenously resistant to trimethoprim. The genetic basis of this resistance was characterized in two collections of clinical isolates of C. jejuni obtained from two different parts of Sweden. The majority of these isolates were found to carry foreign dfr genes coding for resistant variants of the dihydrofolate reductase enzyme, the target of trimethoprim. The resistance genes, found on the chromosome, were dfr1 and dfr9. In about 10% of the strains, the dfr1 and dfr9 genes occurred simultaneously. About 10% of the examined isolates were found to be negative for these dfr genes and showed a markedly lower trimethoprim resistance level than the other isolates. The dfr9 and dfr1 genes were located in the context of remnants of a transposon and an integron, respectively. Two different surroundings for the dfr9 gene were characterized. One was identical to the right-hand end of the transposon Tn5393, and in the other, the dfr9 gene was flanked by only a few nucleotides of a Tn5393 sequence. The insertion of the dfr9 gene into the C. jejuni chromosome could have been mediated by Tn5393. The frequent occurrence of high-level trimethoprim resistance in clinical isolates of C. jejuni could be related to the heavy exposure of food animals to antibacterial drugs, which could lead to the acquisition of foreign resistance genes in naturally transformable strains of C. jejuni.

Low trimethoprim susceptibility of anaerobic bacteria due to insensitive dihydrofolate reductases

All the 28 Bacteroides fragilis strains investigated were susceptible to sulfamethoxazole (minimal inhibitory concentration < 16 mug/ml) and resistant to trimethoprim (TMP; minimal inhibitory concentration > 4 mug/ml). Synergism between sulfamethoxazole and TMP was present in all strains at a ratio of 1:1. The few clostridia investigated proved more resistant to both compounds. Dihydrofolate reductases from B. fragilis, C. perfringens, and some other anaerobic species were isolated. Inhibition profiles with six structurally different inhibitors revealed major differences in all enzymes. For 50% inhibition, the enzyme from B. fragilis and all clostridia required concentrations of TMP which were between several hundredfold and 1,000-fold higher than those required for the enzyme of Escherichia coli, whereas the enzyme from Propionibacterium acnes only needed a threefold higher concentration. In vitro activities of TMP were seen to correspond to the activity at the enzymatic level in B. fragilis and P. acnes, but correspond to a much lesser extent to the activity at the enzymatic level in clostridia, where a poor penetration is assumed to be involved. Dihydrofolate reductase inhibitors other than TMP were found to be as active as TMP both at the enzyme and in vitro. In B. fragilis, higher concentrations of exogenous thymidine were required for increasing the minimal inhibitory concentration of TMP than in E. coli and probably also in C. perfringens.