Mestrenova 12

LC analysis was performed in a Dionex Ultimate 3000 liquid chromatograph with an autosampler (WPS-3000TSL), a pump (LPG-3400SD), a column thermostat (TCC-3000SD), and a diode detector (DAD-3000) (Thermo-Fisher Scientific, Sunnyvale, CA, USA). The system was connected with Chromeleon 7 software.
Spectrophotometric measurements were conducted using a Multiskan GO Spectrophotometer (Thermo-Fisher Scientific). The %Inhibition was calculated from the formula
$$\% \mathrm{inhibition}=\left(\frac{{\mathrm{A}}_{\mathrm{Control}}-{ \mathrm{A}}_{\mathrm{Sample}}}{{\mathrm{A}}_{\mathrm{Control}}} \right)\times 100$$
MS spectra in the negative mode were recorded on an ESI-qTOF Compact mass spectrometer (Bruker, Bremen, Germany). The mass spectrometer was re-calibrated for every run [31 (link)]. ¹H-NMR, ¹³C-NMR, HSQC, HMBC, and COSY experiments were recorded on Avance 300 MHz spectrometer (Bruker) in DMSO-d6 and calibrated using the residual solvent peak. The data were processed with MestReNova 12 software (Mestrelab Research, Santiago de Compostela, A Coruña, Spain).

One-dimensional- and two-dimensional-NMR experiments were performed on Bruker Avance 300 MHz and 500 MHz spectrometers (Bruker BioSpin, Rheinstetten, Germany) using the residual solvent peaks as internal standards. NMR tests for glycosides were conducted in DMSO-d6, ellagitannins in acetone-d6 with D₂O (1+1, v/v), and acids and esters in CD₃OD. NMR data were analyzed by MestReNova 12 software (Mestrelab Research, Santiago de Compostela, Spain). Direct MS spectra (in negative mode) were recorded in water-methanol (1+1, v/v) on the ESI-qTOF Compact mass spectrometer (Bruker Daltonics, Bremen, Germany). LC-MS-derived spectra were recorded in an appropriate eluent (water-methanol acidified with formic acid) in negative mode. MS data were managed by Data Analysis 4.2 software (Bruker Daltonics). UV-VIS spectra of isolated compounds were measured in water-methanol (1+1, v/v; 0.01–0.03 mM) on a Cecil CE 3021 spectrometer (Cecil Instruments, Cambridge, UK).

For the feed sample 1 mL was dried at 50 °C for 48 h and dissolved in 0.6 mL of deuterium oxide (Sigma-Aldrich Co., St. Louis, MO, USA) or 0.6 mL of a mixture of D₂O and D₆-DMSO (ratio of 3:5). The same sample preparation was used for the retentates; however, the amount of sample dried was 200 µL instead of 1 mL. The instrument used for the heteronuclear single quantum coherence spectroscopy 2D-NMR was a Bruker Avance III HD 500 MHz spectrometer (Bruker BioSpin GmbH, Karlsruhe, Germany). A 5 mm broadband (BBO) probe was used together with a Z-gradient coil. The data was acquired with the pulse program “hsqcetgpsisp.2” with the following settings: 136 scans, 1.5 s relaxation delay, 10.3 µs pulse length, 11 ppm spectral width, 1538 FID size and frequency discrimination in F1 e/a. The data was illustrated, processed, and evaluated using MestReNova 12 (Mestrelab Research S.L., Santiago de Compostela, Spain). Processing methods used were base-line correction and phase correction. Semi-quantative calculations and assignments were done according to methods used in the literature [4 (link),20 (link),21 (link),22 (link),23 (link)].

3HB and 4HB copolymers were extracted using chloroform solvent in a Soxhlet extractor (Soxtec 2050, Foss, Denmark) from lyophilized cells harvested from fermentation broth. Subsequently, the extracted PHA was dissolved in chloroform and then precipitated with ten-fold volume of ethanol. After centrifugation at 12,000 rpm for 10 min, the resulted PHA was oven dried at 65 °C for 12 h prior to the subsequent studies. All fractionated polymers were analyzed by ¹³C NMR (Nuclear Magnetic Resonance) (Oxford-600, UK) for identifying block copolymer and/or random copolymer of 3HB and 4HB^{49 (link)}. MestReNova12 (Mestrelab Research, Spain) was used for spectra analysis. Bernoullian statistics method^{50 (link)} was employed to calculate the D-value of block copolymer and random copolymer (Supplementary Figs. 17, 19, and 21).

¹H-NMR map phase and baseline were adjusted using MestReNova 12 (Mestrelab Research, Santiago de Compostella, Spain). On the basis of the proton signal with the 5.233 ppm chemical shift of a-glucose, we corrected the chemical shift, phase, and baseline. The integral of 0.03 ppm equal-width splits was carried out in the map of 9.00–0.5 ppm. The water peak at 4.5–5.0 ppm was cut off. After the normalization of residual integral data, multivariate statistical analysis was carried out using SIMCA-p 14.0 (Umetrics, Sweden). Principle component analysis (PCA) was used to reflect inherent differences and similarities among samples and show the original classification status of data. Furthermore, partial least-squares discriminant analysis (PLS-DA) and orthogonal partial least-squares discriminant analysis (OPLS-DA) were carried out. The effectiveness of PLS-DA was verified by permutation tests (200 times). With the VIP value of OPLS-DA and the corresponding s-plot, different metabolites between groups were identified. Finally, information on the metabolites was imported to the pathway analysis module of MetaboAnalyst 4.0 (https://www.metaboanalyst.ca). In this module, pathway enrichment analysis was combined with pathway topology analysis for the screening of key metabolic pathways.