*2.2. Sample Preparation*

The *B. obovata* fruit were hand-picked fresh at a maturity stage that was slightly under-ripe to optimise the phytochemicals and for ease of transportation. They were picked from trees near Darwin, Northern Territory, Australia, in 2015. The whole fruit were frozen, transported to Brisbane and remained frozen at −20 ◦C. A total of 554 g of the whole fruit with seeds intact (approximately 700 green plums) were lyophilised in a Christ Gamma 1-16 LSC Freeze Drying Unit (John Morris Scientific, Osterode, Germany), then flesh and seed were separated by blending the flesh to a powder in a kitchen food processor. The seeds were milled into a powder using a hammer mill (Lab Mill, Christy and Norris Ltd., Chelmsford, England). The composite flesh and seed powders were stored separately at room temperature, protected from light and in air-tight plastic containers.

#### *2.3. Accelerated Solvent Extractions*

Extractions were carried out using an Accelerated Solvent Extractor (ASE) (ASE 350, Dionex Corporation, Sunnyvale, CA, USA) with a method slightly modified from that in Navarro et al. [16]. Briefly, extractions were performed on 1 g samples of flesh powder or seed powder in triplicate using 100% methanol, 95% ethanol or distilled water as solvents, at 60 ◦C, 60 ◦C and 80 ◦C respectively with eight cycles. The extractions were filtered then evaporated until dry in a miVac evaporator (Genevac Ltd., Ipswich, England). The resulting extract powder was stored in air tight plastic containers at −20 ◦C.

#### *2.4. Extraction of Unbound Polyphenols*

The unbound polyphenol extraction method was slightly modified from Kammerer et al. [17]. Samples of 200 mg of flesh and seed powder were extracted in triplicate with methanol:water: hydrochloric acid (80:19:1) three times on a reciprocating shaker (RP1812, Paton Scientific, Stepney, SA, Australia) at 200 rpm for 2 h at ambient temperature, then centrifuged and the supernatant was removed. The supernatant and residue were stored separately at −20 ◦C.

#### *2.5. Extraction of Bound Polyphenols*

The extraction of the bound polyphenols was modified from Adom and Liu [18]. The residue from the unbound polyphenol extraction was hydrolysed for 1 h at 200 rpm in 2 M NaOH on a reciprocating shaker at ambient temperature. It was acidified to pH 2.0 with concentrated HCl and extracted five times with ethyl acetate. The ethyl acetate layers were dried under nitrogen at 40 ◦C (Dry Block Heater, Ratek Instruments, Boronia, VIC, Australia) before being reconstituted in 50% methanol and stored at −20 ◦C.

#### *2.6. In Vitro Antioxidant Capacity*

ASE extraction aliquots were made up in their extraction solvent to 1 mg/mL with dilutions, and tests were done in triplicate from the triplicate extracts of each solvent (*n* = 9). Tests were done in 96-well plates and read on a Tecan Microplate Reader (Tecan Infinite M200, Tecan Trading AG, Mannedorf, Switzerland) with Magellan Software (version 6.4, Tecan Trading AG, Mannedorf, Switzerland).

The total phenolics content (TPC) measured the reducing capacity of the flesh and seeds and was modified from Singleton and Rossi [19] and Ahmed et al. [20] with gallic acid standards on all extractions. Results are presented as g gallic acid equivalents (GAE)/kg dry weight (DW).

Radical scavenging activity was measured using a 2,2-diphenyl-1-picrylhydrazyl (DPPH) method modified from Yu and Moore [21], with Trolox standards at a concentration range of 5–35 μM/L. Briefly, equal amounts of control, standard or samples and 0.15 mM DPPH were mixed, incubated, and the absorbance read at 517 nm. Results were calculated to μM Trolox equivalents (TE)/g DW.

Chelating activity was measured using a ferrous ion chelating (FIC) assay modified from Decker and Welch [22], Kuda et al. [23] and Wang et al. [24], with ethylenediaminetetraacetic acid (EDTA) as a positive control. Briefly, 200 μL of solvent controls or samples and 10 μL of 1 mM ferrous chloride were mixed and the absorbance read at 562 nm to obtain blank results. 15 μL of 2.5 mM ferrozine was added, it was incubated (10 min, dark, ambient temperature), and the absorbance read at 562 nm. The blank was subtracted from the result reading and % chelating was calculated using the solvent control.

## *2.7. Antimicrobial Activity*

Antimicrobial activity was tested on each type of ASE extraction using a Kirby–Bauer disc diffusion assay modified from Dussault et al. [25]. Standard Plate Count Agar (APHA CM0463, Oxoid Ltd., Hampshire, England) was spread with either *Staphyloccocus aureus* strain 6571 (NCTC— National Collection of Type Cultures, Health Protection Agency Centre for Infection, London, UK) or *Escherichia coli* strain 9001 (NCTC) at a McFarland turbidity of absorbance 0.1 at 540 nm. Triplicates of 75 μL of concentrated extract in 20% ethanol were added to aseptically placed discs on the bacteria coated agar, with 20% ethanol as the control. The concentration of extract on each disc was equivalent to approximately 130 mg DW of flesh freeze dried powder and 60 mg DW of seed freeze dried powder. Plates were incubated (Sanyo Incubator, MIR-154, Sanyo Electric Co., Ltd., Osaka, Japan) at 37 ◦C for 24 h. The diameter of the zone of inhibition was measured in mm under a lighted magnifying glass using a 150 mm digital caliper (Craftright Engineering Works, Jiangsu, China).

#### *2.8. Scanning Electron Microscopy of Antimicrobial Activity*

The same *S. aureus* and *E. coli* strains were grown for 7 h in tryptone soy yeas<sup>t</sup> extract broth (TSYEB) at 37 ◦C. A total of 75 μL of concentrated green plum flesh water extract in 20% ethanol (from approximately 130 mg DW of flesh powder) was added to 1 mL bacteria and broth samples, and 75 μL of 20% ethanol was added to the controls, and incubated for 24 h at 37 ◦C. The samples and controls were washed three times in sterile phosphate buffered saline and fixed in 3% glutaraldehyde. They were adhered to poly-L-lysine-coated (1 mg/mL) coverslips and dehydrated in ethanol before being dried in a critical point dryer (Tousimis Research Corporation, Rockville, MD, USA). Coverslips

were attached to stubs with double-sided carbon tabs and coated with gold before samples were imaged using scanning electron microscopy (SEM) on a Jeol Neoscope JCM 5000 (Jeol Ltd., Tokyo, Japan) at an accelerating voltage of 10 kV.

#### *2.9. Phytochemical Quantification and Identification*

The phytochemicals in the extracts of the free and bound polyphenols as well as the ASE extracts were quantified by ultra pressure liquid chromatography—photo diode array detector (UPLC-PDA) following a modified method of Gasperotti et al. [26]. The compounds were separated on a Waters Acquity HSS T3 (100 × 2.1 mm; 1.8 μm) column at 40 ◦C using the following gradient: 5–20%B (3 min), isocratic (1.3 min), 20–45%B (4.7 min), and 45–100%B (2 min). Mobile phase A consisted of water and mobile phase B of acetonitrile, both containing 0.1% formic acid. The flow was set to 0.4 mL/min and 5 μL of the pre-filtered sample (0.2 μm, Pall, Cheltenhan, VIC, Australia) was injected into the system. The compounds were quantified using external calibrations of the individual phenolic compounds and the peak identity was confirmed by LC-MS.

The same method as described above was used on a ultra high pressure liquid chromatography— mass spectrometry (UHPLC-MS) QExactive (Thermo Fisher Scientific, Bremen, Germany) for peak identification. The LCMS was used in the negative mode, and were run against 39 different standards as a screening study. The seed unbound extraction had anthocyanin identified on a UHPLC-MS QExactive in the positive mode and quantified on a UHPLC with a PDA detector (Agilent, Wilmington, DE, USA).
