Annotation of the Staphylococcus aureus Metabolome Using Liquid

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN

Article

Annotation of the Staphylococcus aureus metabolome using liquid chromatography coupled to high-resolution mass spectrometry and application to the study of methicillin resistance Sandrine Aros-Calt, Bruno H. Muller, Samia Boudah, Celine Ducruix, Gaspard Gervasi, Christophe Junot, and François Fenaille J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.5b00697 • Publication Date (Web): 13 Oct 2015 Downloaded from http://pubs.acs.org on October 22, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Annotation of the Staphylococcus aureus metabolome using liquid chromatography coupled to high-resolution mass spectrometry and application to the study of methicillinresistance Sandrine Aros-Calt1,2, Bruno H. Muller2,Samia Boudah1,3,Céline Ducruix2, Gaspard Gervasi2, Christophe Junot1, François Fenaille1*

1

CEA, iBiTec-S, Service de Pharmacologie et d’Immunoanalyse, Laboratoire d’Etude du

Métabolisme des Médicaments, MetaboHUB-Paris, 91191 Gif-sur-Yvette cedex, France. 2

bioMérieux S.A., Innovation Unit, chemin de l'Orme; 69280, Marcy l'Etoile, France

3

GlaxoSmithKline - Centre de recherche F.Hyafil. 25 Avenue du Québec, 91140 Villebon-sur-

Yvette, France *Corresponding author E-mail: [email protected]. Phone: 33-1-69-08-79-54. Fax: 33-1-69-08-59-07.

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 45

ABSTRACT Staphylococcus aureuscan cause a variety of severe disease patterns, and can readily acquire antibiotic resistance.However, the mechanisms by which this commensal becomes a pathogen or develops antibiotic resistance are still poorly understood. Here, we asked whether metabolomics can be used to distinguish bacterial strains with different antibiotic susceptibilities. Thus, an efficient and robust method was first thoroughly implemented to measure the intracellular metabolites of S. aureus in an unbiased and reproducible manner.We also placed special emphasis on metabolome coverage and annotation, and usedboth hydrophilic interaction liquid chromatography and pentafluorophenyl-propyl columns coupled to high resolution mass spectrometryin conjunction withour spectral database developed inhouse to identifywith high confidence as many meaningfulS. aureus metabolites as possible. Overall, we were able to characterize up to 210 metabolites in S. aureus, which represents a substantial ~50% improvement over previouslypublished data.We then preliminarily compared the metabolic profiles of 10 clinically relevant methicillin-resistant and susceptible strains harvested at different time points during the exponential growth phase (without any antibiotic exposure). Interestingly, the resulting data revealed a distinct behavior of “slowgrowing”resistant strains, which show modified levels ofseveral precursors of peptidoglycan and capsular polysaccharidebiosynthesis.

Keywords Staphylococcus aureus, methicillin resistance, sample preparation, metabolomics, liquid chromatography,high-resolution mass spectrometry, metabolite identification.


Page 3 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION Staphylococcus aureus (S. aureus) is a Gram-positivecoccus that can be found as a commensal bacterium in the nose and on the skin ofabout one third of the global human population.1-3S. aureus is also a facultative pathogenic bacterium which can cause a wide range of infectious diseases, ranging from soft skin infections to severe systemic failures, such as sepsis and infective endocarditis.4 One other characteristic of S. aureus is its ability over the years to develop resistance to different classes of antimicrobial agents. This particular feature combined with drug companies’ reluctance to develop novel antibiotics,has led to serious public health concerns.5 Infections caused by antibiotic-resistant strains of S. aureuseven achieved epidemic proportions worldwidein several instances during the 1980s and 1990s.6One of the most pathogenicstrains is methicillin-resistant S. aureus (MRSA),which appeared shortlyafter the introduction of methicillin in the early 1960s, as a direct consequence ofits overuse.7The prevalence rate of MRSAvaries greatly from one country to another, with values ranging,for example,from below 1.5% in Iceland to above 50% in America and Asia in 2009.8MRSA infections are difficult to treat, and represent the leading cause of infection by a single infectious agent in the USA, with a mortality rate of 20%.9Recently, these bacteriahave become widespreadin many countries in both health care and community settings.2Some MRSA strains havealso developed multidrug resistance,and are resistant to almost all antibiotic families and even to glycopeptidessuch as vancomycin.10,11Altogether these data demonstrate the extraordinary capacity of S. aureus to adapt rapidly to a changing environment, which also strongly dictates its pathogenicity. Bacterial virulence and antibiotic resistance are often considered separately since they are not strictly linked to each other, but may also be regarded as a whole withclear associations and interactionsbetween them.12



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 45

Although we are constantly improving our knowledge and understanding of the lifestyle and pathogenicity of S. aureus, little is known about its basic physiology and the mechanisms by which this commensal becomes a pathogen or develops antibiotic resistance.13,14With the application of several techniques from the omics toolbox, some progress has recently been made in this field. Indeed, several studies involving transcriptomics, proteomics, or metabolomics have provided relevant data regarding the behavior of S. aureusgrown under different conditions (growth-limiting conditions, low temperature, high salt concentration, or antibiotic exposure) and highlighted clear associations between virulence and metabolism.1524

Among those tools, metabolomics provides the most direct assessment of cellular phenotype

and is the technique most suited to quantitative and dynamic monitoring of variations in bacterial metabolism in response to particular environmental conditions. Liebeke et al demonstrated the potential of combiningproteomics- and metabolomics-based approaches to investigate the adaptation of S. aureus to glucose starvation, which resulted in more than 500 proteins and 94 metabolites monitored.Although changes in the proteome and metabolome were globally correlated, the metabolic profile displayed a much wider dynamic range. They especially noted a pronounced increase in concentrations of intracellular amino acids during the transition from exponential growth to glucose starvation, along with alterations in levels of enzymes of the tricarboxylic acid (TCA) cycleand gluconeogenesis.25 In another study, microarray analyses and global metabolite profiling were used to study the response of clinical MRSA strains to oxacillin exposure and demonstrated the potential of the cell to adapt and redirect its metabolism with dramaticallyincreased TCA cycle activity under these particular conditions.24By using 1H nuclear magnetic resonance (NMR), Ammons and coworkers investigated metabolic differences between MRSA and methicillin-susceptible S. aureus (MSSA) strains grown as biofilm or planktonic cultures, in relation to virulence traits. Under these conditions, the quantitative monitoring of 40 intracellular metabolites sufficed to


Page 5 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


distinguish MRSA and MSSA strains, but also biofilm and planktonic phenotypes based essentially on amino acid metabolism and TCA cycle activity.26 As shown by these different examples, the TCA cycle is often affectedby environmental changes or under growth limiting/stress conditions, while numerous studies have also demonstrated

its

role

in

regulatingor

affecting

staphylococcal

virulence

and/or

virulencedeterminant biosynthesis.16 For example, TCA activity was shown to be required for the biosynthesis of S. aureus capsular polysaccharide, one of its numerous virulence factors.27 Thus, metabolomics can be very helpful in gaining deeper insights into basal cell physiology and in deciphering the mechanisms underlying pathogenicity but can also prove highly relevant for distinguishing bacteria with different antibiotic susceptibilities.For instance, Weisenberg et al provided the first direct evidence that liquid chromatography coupled to mass spectrometry (LC/MS) metabolite profiling can readily and efficiently discriminate MRSA and MSSA strains grown in vitro under conventional conditions (i.e. without any antibiotic exposure).28Although highly promising, the latter study unfortunately did not identifyany metabolites responsible for the discrimination, which prevents better understanding of basal physiology underlying antibiotic susceptibility. In recent years, special emphasis has also been placed onthe study of the mode of action of antibiotics on S. aureus by metabolomics profiling. Antti and coworkers compared the metabolomes of clinical isolates of MRSA and MSSA strains grown in vitro in the presence of cloxacillin and vancomycin, and observed the ineffectiveness of the former antibiotic in modifying the metabolome of MRSA.29 Under these conditions, they were able to detect by gas chromatography coupled to mass spectrometry (GC/MS) up to 256 peaks that might correspond to metabolites, but only ~40% of those were putatively identified by comparison with databases.29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 45

Cell wall biosynthesis represents a common target of antibiotics,30,31 but the metabolites involved in the corresponding pathways are often either not detected under common analytical conditions or are not included in publically available databases. A recent study investigated the impact on the S. aureus intracellular metabolome of 5 distinct antibiotics with different mechanisms of action(ciprofloxacin, erythromycin, fosfomycin, ampicillin, and vancomycin),

using

both

GC/MS and

LC/MS techniques

to

widen

metabolite

coverage.32Interestingly, the authors observed that the metabolome was systematically and drastically affected by the antibiotic treatment, with the most impacted pathways being the peptidoglycan pathway and purine and pyrimidine metabolism. By combining data from GC/MS and LC/MS, the authors obtained probably the most exhaustive S. aureus metabolome coverage reportedto date and identified 118 metabolites by use of pure analytical standards and 21 only by database comparison, which correspond to formal and putative annotations, respectively, according to the Metabolomic Standard Initiative criteria.33Most of the metabolites covering peptidoglycan synthesis were part of the putatively annotated compounds, due to lack of the corresponding commercially available molecules. In addition, 37unknown metabolites (~20% of the total metabolite pool) also presented significant (positive or negative) variations upon antibiotic exposure, some of them also potentially corresponding to cell wall precursors.32The latter studies demonstrate that considerable efforts are still required to annotate theS. aureus metabolome. S. aureus is predicted to be one of the most complex microorganisms as judged by the number of metabolites estimated from genome-scale models,despitequite variable estimations ranging from ~500 to ~1400 depending on the methodology used.34A quick comparison of the estimated and experimentally

observed

metabolite

numberssuggeststhat

a

fully

annotated

and

comprehensive metabolite library is still many years away.


Page 7 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


By comparing 6 distinct LC/MS conditions, we have recently demonstratedthat hydrophilic interaction liquid chromatography (HILIC) coupled to high-resolution mass spectrometry (HRMS) operating in the negative ion mode and a pentafluorophenyl-propyl (PFPP) column coupled to HRMS in the positive ion mode was the most efficient combined chromatographic method for profiling of the serum metabolome, resulting in ~260 formal or putative metabolite identifications.35 In the present paper, we have applied this strategy to the metabolite profiling of S. aureus, to achieve the highest number of metabolite annotations. Our initial compound database included 657 metabolites mainly of human origin, andwas further implemented with commercially available bacterial metabolites andother molecules described in the literature for further formal and putative identification, respectively.Before addressing biological questions, we first designed and implemented a robust and accurate sample preparation protocol to take an unbiased snapshot of S. aureus metabolism. This was achieved thanks to the quantitative monitoring of particularly stress-sensitive and/or chemically labile metabolites such as ATP. Using optimized solvent conditions to extract intracellular metabolites efficiently and reproducibly, we were able to characterize up to 210 distinct compounds. Lastly, a proof-of-concept study was realized with ten clinically relevant MRSA and MSSA strains that were grown in vitrowithout antibiotic and metabolically profiled at different time points. Under well-defined conditions, significant differences in their metabolic phenotypes were highlighted,implicatingthe biosynthesis of peptidoglycan and virulence determinants.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 45

MATERIALS AND METHODS Chemicals and reagents All authentic reference compounds were from Sigma-Aldrich (Saint Quentin Fallavier, France)and were of the highest level of purity. Carbonate ammonium and ammonium hydroxide were also from Sigma-Aldrich. Acetonitrile (ACN) was from SDS (Peypin, France), formic acid from Merck (Briare-le-Canal, France), while methanol and absolute ethanol were from VWR chemicals (Fontenay-sous-Bois, France). Deionized water was filtered through a Millipore Milli-Q water purification system.All buffers were prepared using ultrapure water (Milli-Q, Millipore). Mueller Hinton II broth (MHII, cation-adjusted)was from Becton Dickinson(Franklin Lakes, New Jersey, USA).15N-labeled AMP and ATPas well as

13

C,15N-labeled ATP were provided by Euriso-top (Saint-Aubin, France), while

15

N-ADP

was from Sigma-Aldrich.

Bacterial strains Clinicalisolates ofmethicillin-resistant and susceptible Staphylococcus aureuswere from the bioMérieuxcollection of microorganisms. PCR assays were performed for the investigation of mecA and mecC genes in clinical isolates.Antibiotic susceptibility testing and minimal inhibitory concentration (MIC) of oxacillin were determined using an AST-P631 card mounted on a VITEK® 2 automated platformand Etest® strips (bioMerieux, Marcy l’Etoile, France), respectively. Detection of the protein PBP2a was performed using the “SLIDEX® MRSA Detection” test (bioMérieux) when specified. The results are summarized in Table S1.

Bacterial growth conditions S. aureus strains were first isolated from an overnight culture at 37°C on Columbia (COS) agar containing 5 % sheep blood (bioMérieux). All cultures described hereafter were


Page 9 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


performed in cation-adjusted MHII broth on a Minitron II rotary shaker at 220 rpm (Infors HT, Bottmingen-Basel, Switzerland) under aerobic conditions. Aerobic conditions (10% of total volume of Erlenmeyer) were obtained by filling 125 mL Erlenmeyer flasks with 12.5 mL of culture medium for the pre-culture, while the cultures were performed in 2 L baffled shake flasks containing 200 mL of culture medium. Thus, a few bacterial colonies from the agar plate were pre-cultured overnight at 37°C and an aliquot was withdrawn and diluted to an optical density at 600nm (OD600)of~0.1 in a fresh culture medium for the culture step. The earlyexponential phase correspondedfor all the strains investigated to anOD600of 1 as measured by a spectrophotometer (SpectronicGenesis, Milton Roy, Pont-Saint-Pierre, France) which was equivalent to 5x108 CFU/mL.Growth curves were determined for each strain studied and used to determine the OD at which the bacteria should be harvested to correspond to early, mid and late exponential phases.

Sampling of intracellular metabolites and metabolism quenching Fast filtration protocol. This protocol was adapted from published procedures.36,37Briefly, a 5 mL aliquot of cell culture broth was taken from the main culture and was rapidly filtered (in a few seconds)using polyethersulfone sterile membrane disc filters (Supor450, 0.45µm pore size, PALL, New York, USA) mounted on a Millipore filtration device (Darmstadt, Germany). The bacteria on the filter were quickly washed with 5 mL of 0.6% NaCl solutionat room temperature. The filter was rapidly transferred into a 50 mL Falcon tube containing 5 mL of cold 60% ethanol (w/v, ≤ -20°C), which also includes13C,15N-labeled ATP (9.6 µM final concentration in the LC/MS vial). The Falcon tube was subsequently quickly immersed in liquid nitrogen.The whole process was performed in less than 2 minutes. Centrifugation protocol.A 5 mL aliquot of cell culture broth was taken from the main culture and the bacteria were separated from the supernatant by centrifugation



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 45

(Herausmegafuge 40R Centrifuge, Thermo Fisher Scientific, Courtaboeuf, France) for 5 min at 10,000g at 4°C using 25mL Falcon tubes. The pellet was washed with 0.6% NaCl solution and centrifuged againfor 5 min at 10,000g at 4°C. After supernatant removal, 5 mL of cold 60% ethanol (w/v, ≤ -20°C) containing

13

C,15N-labeled ATP (see above) was used to

resuspend the pellet by vortexing. The tube was then immersed in liquid nitrogen. In this case, the whole process was completed within about 15 minutes. Protocol following on-filter culture.This protocol was adapted from the one previously described by Bennett et al.38FivemL of bacterial culture broth was withdrawn in the lag phase at an OD600of 0.1 using the filtration conditions described above. The filter containing the bacteria was placed face up on an agar plate loaded with cation-adjusted MHII. Cells were grown to the early exponential phase (OD600~1), and the metabolism was quenched within a few seconds by placing the filter in a 50 mL tube containing 5 mL of cold 60% ethanol as described above and immersing the tube in liquid nitrogen.

Extraction of intracellular metabolites Following quenching, tubes containing bacteria on filters in the extraction solution were vortexed 10 times on ice to remove the cells from the filter. Then, a 1 mL aliquot of thecell suspension was transferred into 2 mL tubes containing 0.1 mm glass beads (Bertin Technologies, Montigny-le-Bretonneux, France). Cell disruption was performed by three cycles in a Precellys 24 homogenizer (Bertin Technologies) for 30seconds at 3800rpm and at ~4°C. The glass beads and cell debris were separated from the supernatant by centrifugation for 5 minutes at 4°C and 10,000g. A 400µL aliquot ofthe supernatant was withdrawn and further vacuum dried using a SpeedVac instrument (Thermo Fisher Scientific, Courtaboeuf, France) and stored at -80°C until analysis.


Page 11 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For ZIC-pHILIC analysis, samples were reconstituted in 32µL of a 10 mM ammonium carbonate buffer (pH 10.5, adjusted with ammonium hydroxide) containing ethylmalonic acid, 15

N-labeled aspartic acid,

13

C-labeled glucose and ampicillin (each at 3µg/mL, used as

standards for monitoring of LC/MS performances) and 15N-labeled AMP, ADP and ATP (at 2.5, 2.5, 5 µg/mL, i.e. 3.4, 5.8, and 9.8 µM, respectively) as internal standards. The samples were centrifuged 5 min at 10,000g and 4°C and the resulting supernatants were further diluted with 48 µL of acetonitrile. For PFPP analysis, dried samples were redissolved in 80µL of 5% acetonitrile containing 0.1% formic acid,15N-labeled aspartic acid,

13

C-labeled glucose and

ampicillin. Samples were vortexed for 20 seconds and centrifuged for 5 min at 10,000g and 4°C.

Preparation of quality control samples A quality control (QC) sample was prepared by pooling numerous metabolite extracts from different S. aureus strains (both MRSA and MSSA strains) and also containing the internal standards described above.This QC sample was injected regularly every 10 samples throughout the acquisition batch. Two, four, and eight times diluted QC samples were also prepared and used to distinguish real MS signals from background noise by monitoring signal decrease upon dilution.

LC/MS analysis of intracellular metabolites The analytical conditions were essentially as described previously.35 LC/MS analyses were performed with either a Nexerachromatographic system (Shimadzu, Champs sur Marne, France)for the sample preparation step and then with an Ultimate 3000 (Dionex, Thermo Fisher Scientific) for the metabolome annotation and the application to MRSA and MSSA strains.The LC system exchange was accompanied by a ~1-min shift in retention times. Both



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 45

chromatographic systems wereused coupled to an ExactiveOrbitrap mass spectrometer (Thermo Fisher Scientific) fitted with an electrospray (ESI) source and operating in the positive and negative ion modes for PFPP and HILIC analyses, respectively. The chromatographic separations were performed usingboth a Discovery HSF5 PFPP column(5µm, 2.1×150mm; Sigma-Aldrich) maintained at 30°C and

a Sequant ZIC-

pHILIC(5µm, 2.1×150mm)column at 15°C (Merck, Darmstadt, Germany) operated under gradient elution, as follows. For PFPP, mobile phases were 0.1% formic acid in water (A)and 0.1% formic acid in acetonitrile (B), at a flow rate of 250µL/min. The elution consisted of an isocratic step of 2min at 5% phase B, followed by a linear gradient from 5 to 100% of phase B in 18min. These proportions were kept constant for 4min before returning to 5% of phase B and letting the system equilibrate for 6min. Concerning HILIC conditions, mobile phase A was 10 mMammonium carbonate pH10.5 (adjusted with ammonium hydroxide) while mobile phase B was100% acetonitrile, and the flow rate was 200µL/min. Elution started with an isocratic step of 2min at 80% B, followed by a linear gradient from 80 to 40% of phase B from 2 to 12min. The chromatographic system was then rinsed for 5min at 0% B, and the run was ended with an equilibration step of 15min. Chromatographic conditions common to both modes include the temperature of the autosampler compartment (4°C), the injection volume (5 µL), the injection loop option (partial loop with needle overfill), and the presence of an on-line prefilter. The Exactivemass spectrometer was operated with capillary voltage at −3kV in the negative ionization mode and 5kV in the positive ionization and a capillary temperature set at 280°C. The sheath gas pressure and the auxiliary gas pressure (nitrogen) were set at 60 and 10 arbitrary units, respectively. The detection was performed from m/z 75 to 1000 in both ionization modes using a resolution set at 50,000 at m/z 200 (full width at half maximum).


Page 13 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The software interface was Xcalibur (version 2.1, Thermo Fisher Scientific). The mass spectrometer was calibrated externally before each analysis in both ESI polarities using the manufacturer's predefined methods and the recommended calibration mixture provided by the manufacturer. LC/ESI-MS/MS experiments were performed using a Dionex Ultimate chromatographic system combined with a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific), under non-resonant collision-induced dissociation conditions using higher-energy C-trap dissociation (HCD), at normalized collision energies (NCE) ranging from 15 to 30%.

Quantification of adenylated compounds and energy charge determination AMP, ADP and ATP were quantified by the isotope dilution method using their 15N-labeled homologues, with a procedure similar to that described by Martano et al.39 Peak integration was performed with the Quanbrowser module of Xcalibur (version 2.1). From those quantitative measurements (expressed as moles per liter), we calculated the adenylated energy charge (AEC) by using molar concentration with the following formula: AEC =

ATP + 0.5 × ADP ATP + ADP + AMP

Data processing All raw data were manually inspected using the Qualbrowser module of Xcalibur version 2.1 (Thermo Fisher Scientific), while the Quanbrowser module was used for peak detection and integration ofinternal standards. The file converter module of Xcaliburwas used to convert the raw files to CDF files.Automatic peak detection and integration were performed using the matched filter algorithm withinthe XCMS software package.40Grouping of features was performed using the CAMERA software.41



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 45

In order to remove analytical drift induced by clogging of the ESI source observed in the course of analytical runs, chromatographic peak areas of each variable present in the XCMS peaklists were normalized using the LOESS algorithm.42Features generated from XCMS were filtered according to the following criteria: (i) the correlation between QC dilution factors and areas of chromatographic peaks (filtered variables should have coefficients of correlation above 0.5 in order to account for metabolites occurring at low concentrations and which are no longer detected in the most diluted samples), (ii) repeatability (the coefficient of variation obtained on chromatographic peak areas of QC samples should be below 30%) and (iii) ratio of chromatographic peak area of biological to blank samples above a value of 3.

Metabolite annotation and identification Feature annotation was performed considering a ±10 ppm mass tolerance and using our inhouse spectral database,35,43 as well as the publically available databases KEGG,44 HMDB,45 and METLIN.46Metabolite identification was further confirmed by additional MS/MS experiments. To be identified, ions had to match at least 2 orthogonal criteria (among accurate measured mass, isotopic pattern, MS/MS spectrum and retention time) to those of an authentic chemical standard analyzed under the same analytical conditions, as proposed by the Metabolomics Standards Initiative.33 In the absence of an available authentic chemical standard, metabolites of interest were putatively annotated, based on accurate measured mass and interpretation of the MS/MS spectra when available.

Statistical analysis

The data sets were imported into SIMCA-P (version 11.0, Umetrics, Umea, Sweden) to perform multivariate dataanalyses such as principal component analysis (PCA) and partial least-squares discriminant analysis (PLS-DA). All data were mean-centered and unit variance 14 ACS Paragon Plus Environment

Page 15 of 45

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(UV)-scaled in SIMCA-P. The PLS-DA models were validated by several permutation analyses (100 times). The discriminant metabolites were selected by combiningmultivariate variable importance in the projection (VIP) obtained from the PLS-DA model, and univariate p-values(non-parametric Mann-Whitney statistical test). The metabolites were considered as discriminant when VIP>1.5 and p-value

Annotation of the Staphylococcus aureus Metabolome Using Liquid

Recommend Documents