CluMSID: an R package for similarity-based clustering of tandem mass spectra to aid feature annotation in metabolomics.

SUMMARY: Compound identification is one of the most eminent challenges in the untargeted analysis of complex mixtures of small molecules by mass spectrometry. Similarity of tandem mass spectra can provide valuable information on putative structural similarities between known and unknown analytes and hence aids feature identification in the bioanalytical sciences. We have developed CluMSID (Clustering of MS2 spectra for metabolite identification), an R package that enables researchers to make use of tandem mass spectra and neutral loss pattern similarities as a part of their metabolite annotation workflow. CluMSID offers functions for all analysis steps from import of raw data to data mining by unsupervised multivariate methods along with respective (interactive) visualizations. A detailed tutorial with example data is provided as supplementary information. AVAILABILITY AND IMPLEMENTATION: CluMSID is available as R package from https://github.com/tdepke/CluMSID/and from https://bioconductor.org/packages/CluMSID/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The Alkylquinolone Repertoire of Pseudomonas aeruginosa is Linked to Structural Flexibility of the FabH-like 2-Heptyl-3-hydroxy-4(1H)-quinolone (PQS) Biosynthesis Enzyme PqsBC.

Pseudomonas aeruginosa is a bacterial pathogen that causes life-threatening infections in immunocompromised patients. It produces a large armory of saturated and mono-unsaturated 2-alkyl-4(1H)-quinolones (AQs) and AQ N-oxides (AQNOs) that serve as signaling molecules to control the production of virulence factors and that are involved in membrane vesicle formation and iron chelation; furthermore, they also have, for example, antibiotic properties. It has been shown that the ß-ketoacyl-acyl-carrier protein synthase III (FabH)-like heterodimeric enzyme PqsBC catalyzes the last step in the biosynthesis of the most abundant AQ congener, 2-heptyl-4(1H)-quinolone (HHQ), by condensing octanoyl-coenzyme A (CoA) with 2-aminobenzoylacetate (2-ABA), but the basis for the large number of other AQs/AQNOs produced by P. aeruginosa is not known. Here, we demonstrate that PqsBC uses different medium-chain acyl-CoAs to produce various saturated AQs/AQNOs and that it also biosynthesizes mono-unsaturated congeners. Further, we determined the structures of PqsBC in four different crystal forms at 1.5 to 2.7 Šresolution. Together with a previous report, the data reveal that PqsBC adopts open, intermediate, and closed conformations that alter the shape of the acyl-binding cavity and explain the promiscuity of PqsBC. The different conformations also allow us to propose a model for structural transitions that accompany the catalytic cycle of PqsBC that might have broader implications for other FabH-enzymes, for which such structural transitions have been postulated but have never been observed.

The Peptide Chain Release Factor Methyltransferase PrmC Influences the Pseudomonas aeruginosa PA14 Endo- and Exometabolome.

Pseudomonas aeruginosa is one of the most important nosocomial pathogens and understanding its virulence is the key to effective control of P. aeruginosa infections. The regulatory network governing virulence factor production in P. aeruginosa is exceptionally complex. Previous studies have shown that the peptide chain release factor methyltransferase PrmC plays an important role in bacterial pathogenicity. Yet, the underlying molecular mechanism is incompletely understood. In this study, we used untargeted liquid and gas chromatography coupled to mass spectrometry to characterise the metabolome of a prmC defective P. aeruginosa PA14 strain in comparison with the corresponding strain complemented with prmC in trans. The comprehensive metabolomics data provided new insight into the influence of prmC on virulence and metabolism. prmC deficiency had broad effects on the endo- and exometabolome of P. aeruginosa PA14, with a marked decrease of the levels of aromatic compounds accompanied by reduced precursor supply from the shikimate pathway. Furthermore, a pronounced decrease of phenazine production was observed as well as lower abundance of alkylquinolones. Unexpectedly, the metabolomics data showed no prmC-dependent effect on rhamnolipid production and an increase in pyochelin levels. A putative virulence biomarker identified in a previous study was significantly less abundant in the prmC deficient strain.

Untargeted LC-MS Metabolomics Differentiates Between Virulent and Avirulent Clinical Strains of Pseudomonas aeruginosa.

Pseudomonas aeruginosa is a facultative pathogen that can cause, inter alia, acute or chronic pneumonia in predisposed individuals. The gram-negative bacterium displays considerable genomic and phenotypic diversity that is also shaped by small molecule secondary metabolites. The discrimination of virulence phenotypes is highly relevant to the diagnosis and prognosis of P. aeruginosa infections. In order to discover small molecule metabolites that distinguish different virulence phenotypes of P. aeruginosa, 35 clinical strains were cultivated under standard conditions, characterized in terms of virulence and biofilm phenotype, and their metabolomes were investigated by untargeted liquid chromatography-mass spectrometry. The data was both mined for individual candidate markers as well as used to construct statistical models to infer the virulence phenotype from metabolomics data. We found that clinical strains that differed in their virulence and biofilm phenotype also had pronounced divergence in their metabolomes, as underlined by 332 features that were significantly differentially abundant with fold changes greater than 1.5 in both directions. Important virulence-associated secondary metabolites like rhamnolipids, alkyl quinolones or phenazines were found to be strongly upregulated in virulent strains. In contrast, we observed little change in primary metabolism. A hitherto novel cationic metabolite with a sum formula of C12H15N2 could be identified as a candidate biomarker. A random forest model was able to classify strains according to their virulence and biofilm phenotype with an area under the Receiver Operation Characteristics curve of 0.84. These findings demonstrate that untargeted metabolomics is a valuable tool to characterize P. aeruginosa virulence, and to explore interrelations between clinically important phenotypic traits and the bacterial metabolome.

Clustering of MS2 spectra using unsupervised methods to aid the identification of secondary metabolites from Pseudomonas aeruginosa.

Pseudomonas aeruginosa is an important opportunistic pathogen that produces a large arsenal of small molecule virulence factors and quorum sensing signal molecules. The annotation of these secondary metabolites in untargeted, mass spectrometry-based metabolomics is difficult, as many of them cannot be found in common metabolite databases, and as manual annotation is tedious. We therefore developed an algorithm named CluMSID that uses cosine similarities of product ion spectra and neutral loss patterns in combination with unsupervised clustering methods such as multidimensional scaling, density based clustering and hierarchical clustering to group structurally similar compounds and hence facilitate their annotation. The use of this tool allowed us to find clusters for several classes of primary and secondary metabolites, and helped identifying spectral similarities that would have gone unnoticed in standard untargeted metabolomics data analysis workflows. CluMSID enabled the annotation of 27 previously undescribed members of the canonical classes of alkyl quinolone quorum sensing signal molecules and provided evidence for the postulation of a new putative alkyl quinolone class. The CluMSID script written in R is open source and can be used by anyone in the metabolomics and natural product research community.