Non-structural carbohydrate analysis
We adapted methods from Mullin et al . (2021) for sugar and starch analyses. Briefly, ground samples were extracted with 1.6 mL ultra-pure water in a 2 mL glass tube. We placed tubes in a rack and enclosed the rack in steam above a tray of boiling water for 60 min and then closed the tube with its cap, and vortexed them for 30 sec. The resulting extract was separated into two 0.5 mL aliquots for soluble sugars and starch and stored at +4 °C.
For soluble sugars (glucose, fructose, sucrose), tubes containing a 0.5 mL aliquot were centrifuged at 18,213 rcf for 15 min and the resulting supernatant was transferred into new 2 mL tubes containing 1.0 ml of methanol and vortexed for 30 sec. Samples were incubated at room temperature for 90 min and centrifuged at 18,213 rcf for 10 min. Afterwards, 0.5 ml of extracts were transferred to 2 ml glass vials and stored at −40 °C. Sugars were analyzed using an Ultra High-Performance Liquid Chromatography (UHPLC, 1290 Infinity Agilent Tech., Santa Clara, CA, USA) fitted with an InfinityLab Poroshell 120 HILIC-Z column (2.1 x 100 mm 2.7µm, Agilent Tech.) and an Evaporative Light Scattering Detector (ELSD, 1290 ELSD II, Agilent Tech.). A gradient analysis was performed using a binary solvent system of ultra-pure water with 0.02% v/v ammonium hydroxide (channel A) and acetonitrile (HPLC-grade) with 0.02% v/v ammonium hydroxide (channel B) flowing at 0.2 ml min-1. We calculated quadratic standard curves from three dilutions prepared from analytical standards to quantify glucose (Chemical Purity: 99%), fructose (CP: 99%), and sucrose (CP: 99.5%). Soluble sugars were identified based on the retention time of the analytical standards. Concentrations of soluble sugars were reported as µg mg-1 of dry weight.
Starch extraction was conducted using the second aliquot (0.5 mL) of sample extract. We used a series of enzymatic digestions to convert starch into gluconate-6-phosphate as adapted from Lahr & Sala (2014) and Cale et al. (2019a). Extractions began with the steam bath procedure described above. Following the steam bath, samples were vortexed for 30 sec, then a 0.5 ml aliquot that included suspended solids was transferred to a new 2 mL tube containing 0.5 mL of α-amylase solution (0.75 mg enzyme mL−1; Sigma-Aldrich) to convert starch to maltose and similar polysaccharides. The tubes were immediately vortexed for 30 sec, incubated in a water bath at 50° C for 16 hm and then centrifuged (18,213 rcf) for 20 min. The resulting 0.5 mL of the supernatant was transferred into a new 2 mL tube containing with 0.5mL of amyloglucosidase solution 2.5 g enzyme (Sigma-Aldrich) in 50 mL sodium acetate buffer (0.1 M, pH 4.5). The tubes were incubated in a 50 °C bath for 16 h to convert maltose and similar polysaccharides to glucose and fructose and then centrifuged (18,213 rcf) for 15 min. We transferred 0.02 mL of the final glucose extracts to 2 mL glass vials with starch derived glucose, which were immediately quantified by spectrophotometric analysis using the Synergy Microplate Reader H1 (BioTek, Winooski, VT, USA) at an absorbance of 340 nm. Using methods reported in Mullin et al . (2021), we quantified the concentration of starch using two calibration curves. The first curve estimated glucose concentrations from the sample absorbance at 340 nm, and the second estimated starch concentration from glucose concentration.