| Differential NMR Metabolomics | |||
 |
|||
|
Drug activity and specificity is determined by distinct clustering patterns in PCA of NMR metabolomics data
Cellular activity of a drug or protein can be followed by monitoring changes in the metabolome using NMR NMR can be used to capture an image of the state of the cell based on the relative concentration of metabolites, small molecular-weight compounds present in the cell. The concentration of these metabolites are directly or indirectly related to the activity of associated enzymes or proteins. Consider the tricarboxylic acid (TCA) cycle:
The concentration of citrate and isocitrate are inherently dependent on the activity of aconitase, this is also true for the other metabolites in the TCA cycle. By simply lysing cells and transferring the soluble components to an NMR tube, a 1D 1H NMR spectrum will capture the relative concentration of the various metabolites in the cell based on the intensity of the associated NMR peaks.  Inactivating aconitase affects the concentration of metabolites in the TCA cycle The activity of a drug or protein can then be monitored by comparing the differential impact of the metabolome by using wild-type and mutant cells and treating these cells with/without the drug. The mutant cell line has the protein of interest (drug target) knocked-out, inactivated or diminished in activity. Simply, 10 separate cell cultures (25-50 mL) of the wild-type cells, mutant cells, wild-type cells in the presence of the drug and the mutant cells in the present of the drug are prepared. The result is a total of 40 1D 1H NMR spectra of the metabolome for each of the individual cell cultures. The major changes in the NMR spectra are identified by applying principal component analysis (PCA). PCA is a well established statistical technique that determines the directions of largest variations in the NMR data set. The PCA data is typically presented as a two-dimensional plot (scores plot) where the coordinate axis corresponds to the principal components representing the directions of the two largest variations in the NMR data set (PC1, PC2). PCA results in a reduction of the multivariable NMR spectra into a simpler coordinate system of PCs, where each NMR spectra is reduced to a single point in the PC coordinate axis. Similar NMR spectra will cluster together in PC coordinate space and variations along any of the PC axis will highlight experimental differences in the spectra.
There are four general clustering
patterns that will occur depending on the cellular activity and specificity
of the drug.
Consider the fungus,
Aspergillus nidulans,
where a mutant has eliminated the
activity of
urate oxidase (uaZ) in the
purine degradation pathway and
8-azaxanthine (AZA) is a known inhibitor of urate oxidase. The mutant
spores would be expected to have a higher concentration of metabolites
upstream of uaZ (Urate, Xanthine) relative to wild-type spores.
Thus, the NMR spectra of wild-type and mutant spores would form separate and distinct clusters in a PCA 2D scores plot. If AZA is inactive, then the wild-type spores with/without AZA would still cluster together and the mutant spores with/without AZA would form a separate cluster (see theoretical PCA clustering diagramed above). If AZA is selectively active against uaZ, then the wild-type spores with the addition of AZA will now cluster with mutant spores because both spores would have uaZ inactive and incur the same change in the metabolome. This is exactly what is observed.
Alternate clustering patterns would occur in the PCA if AZA is not selective (inhibits multiple proteins) or selectively inhibits a protein other than uaZ. In affect, our differential NMR metabolomics methodology provides a straight-forward method to determine the in vivo activity and selectivity of potential drug candidates. This is extremely valuable information prior to proceeding with expensive clinical trials, since most clinical fail because of lack of efficacy or toxic side-effects that may result from a lack of drug specificity.
M. R. Sadykov, M. E.
Olson, S.
Halouska, Y. Zhu, P. D. Fey,
R. Powers, and G. A. Somerville
(2008) "Tricarboxylic
acid cycle dependent regulation of
Staphylococcus epidermidis polysaccharide intercellular adhesin synthesis."
Journal of
Bacteriology/i>, 130(23):7621-7632.
From: Journal of Proteome Research (2007), 6(12):4608-4614. Metabolic Pathways Associated with DCS Activity Illustration of the mutant and wild-type Mycobacterium smegmatis cell lines and the affect of DCS NMR spectra illustrating major changes in metabolite concentrations PCA Results Metabolic pathways analyzed by NMR metabolomics PCA comparison of wild-type and mutant A. nidulans Theoretical PCA plot fro drug activity and specificity Affect of AZA on A. nidulans PCA Results NMR spectra illustrating major changes in metabolite concentrations Structures of relevant metabolites Additional NMR spectra illustrating changes in metabolite concentrations Comparison of NMR spectra and scores plot contribution Purine and pyrimidine biosynthetic super pathway
|
|
| |
