Department Of Gastroenterology, Ghent University, Ghent .

Supporting information for publicationJulie Vanden Bussche†*, Massimo Marzorati‡, Debby Laukens , and Lynn Vanhaecke†† Laboratory of Chemical Analysis, Ghent University, Merelbeke, Belgium.‡ Laboratory of Microbial Ecology and Technology, Ghent University, Ghent, Belgium Department of Gastroenterology, Ghent University, Ghent, Belgium.Validated High Resolution Mass Spectrometry-Based Approach For Metabolomic FingerprintingOf The Human Gut PhenotypeABSTRACT: Fecal samples are an obvious choice for metabolomic approaches, since they can beobtained non-invasively and allow studying interactions between the gut microbiota and the host. Theuse of ultrahigh performance liquid chromatography hyphenated to Orbitrap high-resolution massspectrometry (UHPLC-Orbitrap HRMS) in this field is unique. Hence, this study relied on OrbitrapHRMS to develop and validate a metabolic fingerprinting workflow for human feces and in vitrodigestive fluids. After chemometric sample extraction optimization, an aqueous dilution appearednecessary to comply to the dynamic range of the MS. The method was proven ‘fit-for-purpose’ througha validation procedure that monitored endogenous metabolites in quality control samples, whichdisplayed in both matrices an excellent linearity (R2 0.990), recoveries ranging from 93-105%, andprecision with CVs 15%. Finally, feces from 10 healthy individuals and 13 patients diagnosed withinflammatory bowel disease were subjected to metabolomic fingerprinting. 9553 ions were detected, aswell as differentiating profiles between Crohn’s disease or ulcerative colitis by means of (O)PLS-DAmodels. Additionally, samples from the dynamic gastrointestinal tract simulator (SHIME platform)were analyzed resulting in 6446 and 5010 ions for the proximal and distal colonic samples, respectively.Supplementing SHIME feed with antibiotics resulted in a significant shift (P 0.05) of 27.7 % of themetabolites from the proximal dataset and 34.3% for the distal one. As a result, the presentedfingerprinting approach provided predictive modeling of the gastrointestinal metabolome in vivo and invitro, offering a window to reveal disease related biomarkers and potential insight into the mechanismsbehind pathologies.TABLE OF CONTENT:- Figure S1: UHPLC-HRMS derived XIC of the selected reference standards detected in a QCpool of human feces (A) and of digestive fluids (B).- Figure S2: Principal component analysis score plot for the positive ionization mode on a datasetof healthy and IBD diagnosed individuals. (dark blue: healthy individuals, red: healthyvegetarians, yellow: CD active, turquoise: CD remission, purple: CU active, orange: CUremission)- Figure S3: Score plot of a partial least square analysis for the negative ionization mode on adataset of healthy (dark blue) and IBD diagnosed individuals (orange: CD and turquoise: UC).- Figure S4: Principal component analysis score plot for the positive ionization mode derived fromthe distal (left) and proximal (right) colonic vessel of a SHIME experiment, with no addition ofantibiotics (dark blue dot), a mild (light blue dot) and shock dosage (green dot).- Table S1: Overview of the known unknown metabolites used for the profiling metabolomicsmethod.- Table S2: Overview of the individual validation results for fecal samples as well as digestivefluids.- Table S3: The putative identification of the metabolites of the positive ionization database of theIBD cohort.S1Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationAlanineDedecanoicacidAcetic lPantothenicacidValine-d8Figure S1A: UHPLC-HRMS derived XIC of the selected reference standards detected in a QC sample ofhuman fecal samples.S2Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationAlanineAcetic acidPipecolic UDCAPantothenicacidValine-d8Figure S1B: UHPLC-HRMS derived XIC of the selected reference standards detected in a QC sample ofdigestive fluids.S3Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationFigure S2: Principal component analysis score plot for the positive ionization mode on a dataset of healthyand IBD diagnosed individuals. (dark blue: healthy individuals, red: healthy vegetarians, yellow: CD active,turquoise: CD remission, purple: CU active, orange: CU remissionFigure S3: Score plot of a partial least square analysis for the negative ionization mode on a dataset ofhealthy (dark blue) and IBD diagnosed individuals (orange: CD and turquoise:UC).S4Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationFigure S4: Principal component analysis score plot for the positive ionization mode derived from the distal (left) andproximal (right) colonic vessel of a SHIME experiment, with no addition of antibiotics (dark blue dot), a mild (lightblue dot) and shock dosage (green dot).S5Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationTable S1: Compiled database for gastro-intestinal relevant metabolites (underlined: purchased and not detectedmetabolites).N Compound NameElementalCompositionN Compound nositol23-(4-hydroxphenyl)propionic acidC9H10O362Isobutyric acidC4H8O233-phenylpropionic acidC9H10O263Isovaleric acidC5H10O243-hydroxybenzoic acidC7H6O364L-alanineC3H7NO253-hydroxybutyric acidC4H7O365L-asparaginic neC9H11NO273-MethylhexanedioicL-glutamic 83,4-dihydroxyphenylacetic nylpropionic acidC9H10O469L-leucineC6H13NO2104-aminobutyric etic acidC8H8O3L-ornithineC5H12N2O2H134-methylvaleric acidC6H12O273L-prolineC5H9NO2145-aminovaleric acidC5H11NO274L-pyroglutamic 2O275L-serineC3H7NO316Adipinic acidC6H10O476L-threonineC4H9NO317Azelaic acidC9H16O477L-tyrosineC9H11NO318Acetic acidC2H4O278L-tryptophan19Benzoic acidC7H6O279L-valine20Beta-alanineC3H7NO280Lactic acid21Butyric acidC4H8O281Litocholic acidC24H40O322Cholic acidC24H40O582Linoleic feic acidC9H8O484Maleic acidC4H4O425Chenodeoxycholic 6MaltoseC12H22O1127Citrazinic inateC4H6O90MNU aldehyde31Cyclohexanocarboxylic C3H7NO2S92NDEA MA uctoseC6H12O696NPIP (piperidine)C5H10N2O37D-valine-d8C5[2]H8H3O2NPYR (N-Nitrosopyrrolidine)C4H8N2ON9738D- quinic acidC7H12O698OctanalC8H16O39Dodecanedioic acidC12H22O499Octanoic acidC8H16O240DopamineC8H12O2Cl100Oleic lC6H6O102Palmitic acidC16H32O243Phenylacetic acidC8H8O2103Panthothenic 8OC7H8OC5H10OS6Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publication45GalactoseC6H12O6105Pipecolinic acid46Glucaric acid (Saccharic acid)C6H9KO8106Propionic acidC3H6O247Glucitol (Sorbitol)C6H14O6107PutrescineC4H12N248Gluconic acid sodium 9SarcosineC3H7NO250Glucuronic acidC6H10O7110Sebacic acidC10H18O451Glutaric permineC10H26N453GlycocholateC26H43NO6113Stearic acidC18H36O254Glycolic acidC2H4O3114Suberic 7H14O116Thiazolidine-4-carboxylic acid57Hexanoic acidC6H12O2117Tauroursodeoxycholic acid58HexanalC6H12O118Ursodeoxycholic acid59IndolC8H7N119Valeric acid60Indole-3-acetic 5NO6SC24H40O4C5H10O2C5H4N4O2S7Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationTable S2. The analytical performance criteria for detection of 9 known unknown metabolites and an ISTD (valine-d8) in human fecal samples and SHIME suspensionsbased on Quality Control samples.Ionization modeRetention timeLinearity(R2)Instrumentalprecision (CV %)Alanine 0.97 0.0030.99410.0AcOH 1.72 0.0110.9927.3Pipecolic acid 1.54 0.0630.9942.3Spermidine 0.75 0.0120.992Inositol-1.01 0.0100.992Dodecanedioic acid-9.86 0.015Hexanoic acid 9.56 0.016UDCA-Panthotenic acidValine-d8Repeatability(CV %)Inter-day assayprecision (CV%)Recovery(mean SD, %)3.84.1104.0 8.012.013.099.2 1.93.24.3100.9 2.87.99.811.1101.6 3.44.95.58.3104.8 8.50.9951.34.15.099.8 0.40.99111.32.87.397.2 7.111.51 0.0370.99812.77.111.8101.5 2.2 5.22 0.0310.9982.25.87.7102.4 5.1 1.17 0.0040.9967.74.03.9102.4 6.6Alanine 0.95 0.0650.9931.01.65.9101.4 2.9AcOH 1.66 0.0080.9980.40.20.7103.7 4.5Pipecolic acid 1.47 0.0110.9971.50.62.397.8 4.7Spermidine 0.77 0.0250.9956.77.612.993.0 12.7Inositol-1.03 0.0160.9991.41.65.7101.1 6.9Azelaic acid 9.08 0.0150.9921.81.42.3102.0 3.1Acetylglucosamine 1.07 0.0250.9912.02.87.1101.0 6.6UDCA-11.46 0.0340.9930.80.15.896.0 7.7Panthotenic acid 5.20 0.0260.9995.01.63.899.9 0.3Valine-d8 1.31 0.0260.9962.71.88.597.9 4.5Human fecesSHIME fluidsS8Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Supporting information for publicationTable S3: The putative identification of the metabolites of the positive ionization database of the IBD cohort.This table is a separtaly loaded excel file.S9Ghent University - Faculty of Veterinary MedicineResearch Group Veterinary Public Health and Zoonoses – Laboratory of Chemical AnalysisSalisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60

Ghent University - Faculty of Veterinary Medicine Research Group Veterinary Public Health and Zoonoses – Laboratory of Chemical Analysis Salisburylaan, 133 B-9820 Merelbeke Belgium T 32-9-264.74.60 S1 . Julie Vanden Bussche†*, Massimo Marzorati‡, Debby Laukens , and Lynn Vanhaecke†