Metabolic Response to Antibiotic Resistance
By: Carol A. Rouzer, VICB Communications
Published: February 1, 2013
The acquisition of antibiotic resistance in Nocardiopsis leads to major metabolic changes, including the production of novel secondary metabolites.
Bacteria can acquire antibiotic resistance by a number of mechanisms, including transfer of resistance elements from other bacteria and mutation of genes coding for proteins that serve as antibiotic targets. For some antibiotics, the target proteins play a role in transcription and translation. Examples are rifampicin, which binds to the bacterial RNA polymerase β subunit (rpoB), and streptomycin, which targets the 30S ribosomal subunit (Figure 1). Bacteria that are resistant to these antibiotics through mutations of the respective genes exhibit marked changes in transcription and translation that lead to alterations in global metabolism. This frequently includes increased production of secondary metabolites, compounds that are not directly related to the growth, development, or reproduction of the organism. Secondary metabolites are a major source of natural products, so antibiotic resistant bacteria may serve as a source of new molecules possessing valuable biological activities. To explore these possibilities and to better understand the impact of antibiotic resistance on bacterial primary and secondary metabolism, Vanderbilt Institute of Chemical Biology members John McLean and Brian Bachmann conducted an exhaustive investigation of the metabolomes of antibiotic resistant bacteria [D.K. Derewacz, et al. (2013) Proc. Natl. Acad. Sci. U.S.A., published online January 22, DOI:10.1073/pnas.1218524110].
Figure 1. Molecular basis for antibiotic resistance. (Left) Structure of rifampicin bound to the RNA polymerase β subunit showing the enzyme surface in green and the location of amino acids that are frequently mutated in resistant bacterial strains in red. (Right) Complex of streptomycin and its target, the 30S ribosomal subunit. Images reproduced from Wikimedia Commons under the Creative Commons Attribution-ShareAlike 3.0 Unported License.
The investigators focused their attention on Nocardiopsis sp. FU40 ΔApoS8. This nonpathogenic soil bacterium, which has been sequenced to high coverage, possesses nineteen identified secondary metabolite gene clusters. However, its only known secondary metabolite family is the macrolide polyketide apoptolidins. To simplify interpretation of their results, the investigators inactivated the apoptolidin pathway by replacement of the terminal enzyme (ApoS8) with an apramycin resistance cassette. Growth of these bacteria on medium containing rifampicin or streptomycin generated resistant strains. For their studies, the Bachmann and McLean labs selected five rifampicin-resistant and six streptomycin-resistant strains. All five of the rifampicin-resistant strains carried a G→A transition mutation in the rpoB gene. Three out of six streptomycin-resistant strains carried a mutation in rpsL, which codes for the S12 protein, a part of the 30S ribosomal subunit. The mutations conferring streptomycin resistance in the remaining three strains is a subject for future work.
The investigators cultured all resistant strains and the progenitor bacteria to the stationary phase on antibiotic-free medium. They then harvested the culture medium and cells for metabolomic analysis by ultra-high performance liquid chromatography (UPLC)-ion mobility mass spectrometry (IM-MS). IM-MS separates ions on the basis of their mobility through a drift tube filled with a neutral gas, usually helium. This separation, based primarily on charge-to-surface area ratio, adds an additional layer of species differentiation and identification to that afforded by the mass spectrometer, which is based on mass-to-charge ratio. Application of UPLC-IM-MS to the Nocardiopsis extracts yielded information on over 1,000 species per sample.
To analyze their extremely rich data sets, the McLean and Bachmann teams applied multivariate statistical analysis methods, among them principal component analysis (PCA) and orthogonal projection to latent structures-discriminant analysis (OPLS-DA). The goal was to develop a mathematical description of the metabolic features that most distinguished the antibiotic-resistant strains from each other and from the progenitor strain. PCA plots of each strain based on the two most distinguishing components indicated that most antibiotic-resistant strains were significantly different from the parent strain and that rifampicin-resistant bacteria tended to be distinguished from streptomycin-resistant strains. However, different strains resistant to each antibiotic showed significant variability, likely due to secondary mutations that help the bacteria adapt to the alterations in their transcription/translation machinery. Furthermore, there was substantial overlap in the characteristics of the rifampicin- and streptomycin-resistant strains (Figure 2).
Figure 2. Plot of PCA results for the two most distinguishing components. Data are shown for five rifampicin-resistant (red) and six streptomycin-resistant (blue) strains plus the progenitor strain (black). Different symbols represent different strains, and identical symbols indicate replicates of the same strain. Note that the replicates are near to each other, indicating stability of the method. Distance along the horizontal axis indicates that the corresponding strains differ with regard to principal component 1, while distance along the vertical axis indicates differences with regard to principal component 2. Figure reproduced with permission from D.K. Derewacz, et al. [(2013) Proc. Natl. Acad. Sci. U.S.A., published online January 22, DOI:10.1073/pnas.1218524110, copyright 2013 Derewacz et al.].
Comparisons between pairs of individual strains indicated that the metabolic differences between them were on the scale of differences observed between bacteria of different phyla. When compared to the progenitor strain, resistant strains exhibited from 100 to over 300 unique features (average 280). In contrast, the progenitor strain exhibited only about 80 unique features when compared to most of the resistant strains. These results suggested a generalized derepression of transcription and translation, leading to an overall increase in metabolism among the antibiotic-resistant bacteria.
Loadings analysis of the PCA results indicated those spectral features that most strongly contributed to the differences between strains (Figure 3), and IM-MS data provided information needed to identify many of these metabolites. Of interest was the finding that many of the new metabolites in the resistant strains were of higher molecular mass (> 400 Da), and their mass spectra did not match those of primary metabolites. These findings suggested a substantial up-regulation of secondary metabolite formation. Detailed structural analysis by MS and NMR of the most abundant of these metabolites from one of the rifampicin-resistant strains revealed a new polyketide secondary metabolite based on a xanthene scaffold. The new metabolite, named mutaxanthene A, two minor variants, mutaxanthenes B and C, and four adducts comprised the majority of the most abundant distinguishing metabolites of this strain (Figure 4). The structure of the mutaxanthenes suggests that this class of secondary metabolites arises from the type II polyketide synthase pathway, and genomic data indicate that this pathway is present in Nocardiopsis sp. FU40 ΔApoS8. The investigators proposed a metabolic pathway for the biosynthesis of the mutaxanthenes based on known reactions of polyketide natural product synthesis.
Figure 3. Loadings analysis showing how each spectral feature in the metabolomic data contribute to the differences between strains. The inset shows the positions of rifampicin-resistant strain 4 (R4), streptomycin-resistant strain 5 (S5), and the progenitor strain (Wild type). Points in the loading diagram that are positioned along the vector indicating the position of each strain represent metabolites that are more strongly characteristic of that strain. For example, points 1 through 4 are characteristic of R4, 16, 17 and 18 are characteristic of S5, and 23 is characteristic of the progenitor strain. Metabolites designated 1 through 7 are all mutaxanthenes or adducts of mutaxanthenes. Figure reproduced with permission from D.K. Derewacz, et al. [(2013) Proc. Natl. Acad. Sci. U.S.A., published online January 22, DOI:10.1073/pnas.1218524110, copyright 2013 Derewacz et al.].
Figure 4. Structures of the mutaxanthenes.
Together the results indicate that acquisition of resistance to antibiotics that target the transcriptional or translational machinery leads to metabolome-scale changes. It is unclear whether these changes, which suggest a general derepression of biosynthesis, are an adaptive or a coincident trait. It is clear, however, that a full understanding of these changes is important to understanding the impact of antibiotic resistance in the clinic. Furthermore, the upregulation of secondary metabolite biosynthesis may be exploited in the future for the discovery of novel natural products such as the mutaxanthenes.