Tracefinder 5

5 µl of each sample or palmitic acid standards in 90% acetonitrile was injected into a Vanquish LC system (Thermo Scientific, UK) using a flow rate of 0.25 mL min⁻¹. The analytical column was Aquity UPLC CSH C18 column (1.7 um particle size, 100 mm × 2.1 mm, Waters, Manchester, UK) held at 50 °C. Starting mobile phase composition was 40% solvent B (0.1% formic acid in 95% acetonitrile/5% IPA) in A (0.1% formic acid and 80% acetonitrile in water) increasing to 54% B over 5 minutes. For column wash and equilibration the %B was increased 70–99%B over 3 minutes, held at 95% B for 2 minutes then returned to 40% B for 5 mins. Column eluant was directed in to an Orbitrap Exploris 240 mass spectrometer (ThermoFisher Scientific, UK) and ionised using electrospray ionisation in negative polarity at 2500 V. Mass measurement used full scan mode with resolution of 120,000, a m/z range of 50–500. The maximum injection time was set automatically by the software. Samples and standards were analysed in triplicate and Tracefinder 5.1 (ThermoFisher Scientific, UK) was used to construct the calibration curve and determine peak areas of palmitic acid signals in the samples. Calibration standard concentrations analysed were 1 ng, 10 ng, 100 ng, 1 µg and 10 µg. Final data were normalised per cm of hyphae per g of starting material.

5 µl of each sample or the glucose and fructose standard mixture in 90% acetonitrile was injected into a Vanquish LC system (ThermoFisher Scientific, UK) using a flow rate of 0.25 mL min⁻¹. The analytical column was Accucore-150-Amide-HILIC (2.6 µm particle size, 150 mm × 2.1 mm, ThermoFisher Scientific, UK) held at 35 °C. Starting mobile phase composition was 90% solvent B (0.1% ammonium acetate in acetonitrile) in A (0.1% ammonium acetate and water) decreasing to 60% B over 5 minutes and held constant for 4 minutes before being re-equilibrated to 90% B after 2 minutes. Column eluant was eluted into an Orbitrap Exploris 240 mass spectrometer (ThermoFisher Scientific, UK) and ionised using electrospray ionisation in negative polarity at 2500 V. Mass measurement used full scan mode resolution of 120,000, a m/z range of 5–500. The maximum injection time was set automatically by the software. Samples and standards were analysed in triplicate and Tracefinder 5.1 (ThermoFisher Scientific, UK) was used to construct the calibration curve and determine peak areas of hexose signals in the samples. Calibration standard concentrations analysed were 1 ng, 10 ng, 100 ng, 1 µg and 10 µg. Final data were normalised per cm of hyphae per g of starting material.

Areas for the 20 compounds previously selected for the model were extracted in Thermo Fisher Scientific Trace Finder 5.1 software from the.RAW files. A compound fragmentation database was obtained from Compound Discoverer from previous runs. The quan master method was employed to allow summing up of all events in cases of metabolic features eluting across multiple peaks such as cytosine and cytosine-containing compounds where source fragmentation was observed, or butrylcarnitine where multiple elution peaks were observed. The ICIS detection algorithm was employed for all metabolic features; however, other detection algorithm settings (e.g., peak detection strategy, peak threshold type) and retention times settings (i.e., detection type and RT window) were tuned individually to each metabolic feature allowing for controlled selection of the peak area in every sample. Exported areas were further processed in R for normalization and statistical analysis.

Acquired LC–MS data were processed using the Thermo Scientific TraceFinder 5.1 software. Targeted metabolites were identified by matching accurate mass and retention times to the University of Iowa Metabolomics Core Facility’s in-house library of confirmed standards (Table S2). During data analysis in TraceFinder, mass tolerance was set to 2 millimass units (0.002 Daltons). After peak area integration by TraceFinder, NOREVA software was applied for signal drift correction on a metabolite-to-metabolite basis using the pooled QC sample that had been analyzed throughout the instrument run as the normalizing reference [17 (link)]. We further accounted for technical variation in sample quantity and autosampler injection volume by normalizing each metabolite signal intensity to the sum of all metabolite signal intensities within a sample to generate a ratiometric metabolite fingerprint. NOREVA and ratiometrically normalized individual QC samples were superimposed on a PCA plot, providing evidence of high run quality and accurate data normalization (Figure S1). Internal standards were used to test for differences in extraction efficiency amongst samples, which were not observed. Batch effects were not present in this study since all samples were run in one batch on the LC–MS. The processing blank showed no evidence of carryover or contamination.