4.1 General experimental procedure

LC-HR-MS analysis was performed on Thermo Vanquish UHPLC connected to a Proshell 120 EC-C18 column (2.1 × 100 mm, 1.9 μm) linked to an Orbitrap IQ-X Tribrid mass spectrometer. NMR spectra were recorded on a Bruker 600 MHz spectrometer equipped with a TCI cryoprobe at 25 °C. The 1H and 13C NMR chemical shifts were referenced to the DMSO-d6 solvent peaks at δH 2.50 and δC 39.52. All HPLC and LC–MS experiments were performed with a MeCN-H2O gradient solvent system.

4.2 Screening of plant extracts against Actinobacteria and antimicrobial activity assay

Hibiscus and other plant extracts were prepared by soaking 10 g of crushed dried plant parts in approximately 100 ml methanol with stirring for 1 h, then filtered and dried using a rotary evaporator. Stock solution of 900 and 225 mg/mL in DMSO were made for addition to Actinobacterial cultures. A total of 24 isolates were grown in 1.5 ml YEME liquid (3 g/L yeast extract, 3 g/L malt extract, 10 g/L glucose, 5 g/L peptone) in 12 well plates in three replicates. After 24 h, 30 µL of stocks of 900 and 225 mg/mL solutions of plant extracts were added to the culture to give final concentrations of 18 and 4.5 mg/mL, respectively. The cultures were incubated at 30 °C with shaking at 150 rpm for 7 days. For LC–MS profiling and antimicrobial activity assays,  1 mL of each culture was centrifuged for 5 min at 14,000 rpm. Supernatants were directly analysed by LC–MS (2 µL per injection) and disc diffusion assay (20 µL per each disc) against B. subtilis, E. coli, S. aureus, C. albicans and C. neoformans. Nutrient agar (NA) was used for all strains except C. neoformans, for which YPD (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose) was used.

4.3 HPLC fractionation/purification of thiolutin and aureothricin

To purify induced metabolites, strain MBN2-2 was grown in 250 ml liquid YEME with and without 4.5 mg/ml hibiscus flowers extract for 7 days at 30 °C with shaking at 150 rpm. Cultures were then centrifuged, and the supernatants were filtered (Whatman™ no. 1). The filtrate was passed through a C18 based cartridge (Thermo Hypersep C18 10 G) using flash chromatography. Initially it was washed with two volumes of deionized water, then the extract was eluted with two volumes of methanol. The methanol was evaporated by use of a Genevac, then the dried extract was resuspended in 1 ml methanol, loaded onto dental cotton and air-dried in a fume hood. The dental cotton containing the extract was loaded into a cartridge (10 × 30 mm) connected to a semi-preparative reverse-phase C18 Betasil column (21.2 mm × 150 mm). An Agilent 1260 Infinity II preparative HPLC was used for fractionation of the extract using the following method: initially, constant 5% acetonitrile for 5 min, followed by a linear gradient from 5 to 95% acetonitrile for 55 min, then isocratic at 100% acetonitrile for the next 5 min at a flow rate of 9 min/ml. The fraction collector was set to collect fractions at 60-s intervals throughout a 60-min runtime. Fractions were assayed against B. subtilis and active fractions analysed by LC–MS and NMR.

Thiolutin (1): 1H NMR [DMSO-d6]: 9.98 (1H, s), 7.35 (1H, s), 3.25 (3H, s), 2.02 (3H, s). 13C NMR [DMSO-d6]: 168.9, 166.2, 136.0, 132.4, 114.8, 111.0, 27.5, 22.4. HRESIMS [M + H]+m/z 229.0100 (calcd for C8H9N2O2S2, 229.0100).

Aureothricin (2): 1H NMR [DMSO-d6]: 9.90 (1H, s), 7.33 (1H, s), 3.25 (3H, s), 2.35 (2H, q, J = 7.5 Hz), 1.01 (3H, t, J = 7.6 Hz). HRESIMS [M + H]+m/z 243.0257 (calcd for C9H11N2O2S2, 243.0256).

4.4 Identification of elicitor compound from Hibiscus extract

Ten grams of crushed Hibiscus flowers was extracted with methanol for 1 h. Methanol was evaporated by rotary evaporator and the dried extract was subjected to initial fractionation on C18 cartridge (Thermo Hypersep C18 10 G) using flash chromatography, with 0, 20, 40, 60, 80 and 100 percent methanol. The first fraction (eluted with 0% MeOH) was found to induce the production of thiolutin and aureothricin, therefore, it was resuspended in 900 µL deionized water and fractionated by HPLC. The HPLC method was same as the one used for fractionation of strain MBN2-2. Fractions 8 and 11 were found to induce the production of both thiolutin and aureothricin and were analysed by LC–MS and NMR.

Hibiscus acid dimethyl ester (3): 1H NMR [DMSO-d6]: 6.80 (1H, s), 5.36 (1H, s), 3.76 (3H, s), 3.68 (3H, s), 3.21 (1H, d, J = 17.2 Hz), 2.68 (1H, d, J = 17.2 Hz). HRESIMS [M + H]+m/z 219.0499 (calcd for C8H11O7, 219.099).

Hydroxycitric acid 1,3-dimethyl ester (4): 1H NMR [DMSO-d6]: 4.13 (1H, s), 3.62 (3H, s), 3.55 (3H, s), 2.89 (1H, d, J = 15.5 Hz), 2.80 (1H, d, J = 15.5 Hz). HRESIMS [M + H]+m/z 237.0604 (calcd for C8H13O8, 237.0605).

4.5 Inducing production of thiolutin and aureothricin with hydroxycitric acid, citric acid, ammonium iron (III) citrate, iron (II) sulphate and iron (III) chloride

Strain MBN2-2 was grown in 1.5 ml YEME liquid for 24 h. Then from stock solutions of commercially available hydroxycitric acid, ammonium iron (III) citrate, citric acid, iron (II) sulphate, and iron (III) chloride were added to the MBN2-2 culture to give final concentrations of 1000, 750, 500, 250, 100, 50 and 25 µg/ml. Cultures were incubated for seven days at 30 °C with shaking. Each day a 60 µL of extract was taken for antimicrobial assay tests and LC–MS analysis.

4.6 Genomic DNA sequencing and assembly

Genomic DNA from the Streptomyces strain MBN2-2 was sequenced by the MicrobesNG DNA Sequencing Facility at the University of Birmingham using a combination of Illumina and Oxford Nanopore Long Reads. Illumina raw reads were trimmed using trim galore [32] and assembled with SPAdes [33] using Unicycler [34] as a SPAdes optimiser. Nanopore raw reads were assembled with Flye [35] and a hybrid assembly was performed with Unicycler [34] and Hybracter [36]. The genome contigs were polished by mapping the Illumina reads back to the assembled contigs with Mira 5.0 [37] and Pilon [38], and joined by manual inspection. The thiolutin/aureothricin biosynthetic gene cluster was identified by analysis of the genome sequence using antiSMASH [39].

4.7 Chrome Azurol S (CAS) assay

The CAS solution was made by the protocol outlined by Alexander and Zuberer [40]. Three separate solutions were prepared. Solution A consisted of 21.9 mg of hexadecyltrimethylammonium bromide (HDTMA) dissolved in 25 ml MiliQ water; Solution B contained 1.5 ml of 1 mM FeCl3.6H2O dissolved in 10 mM HCl, then mixed with 7.5 ml of 2 mM CAS; Solution C contained 9.76 g (2-(N-morpholino)ethanesulfonic acid) (MES) dissolved in 50 ml MiliQ water with pH adjusted to 5.6 with 1N KOH. Solution B was gradually added to solution A with continuous stirring, followed by addition of solution C. The resulting mixture was adjusted to a final volume of 100 ml with MilliQ water. Just before use, 87.3 mg of sulfosalicylic acid was added to the mixture. The stock solution of 5 mg/ml of the ethylenediaminetetraacetic acid (EDTA), citric acid, hydroxycitric acid, hibiscus acid dimethyl ester, and hydroxycitric acid 1,3-dimethyl ester were prepared in MilliQ water and diluted with CAS solution to achieve final concentrations of 25, 20, 15, 10, 5.0, 2.5, 1.0, 0.5, 0.25, 0.1, 0.05 and 0.01 mg/ml in a 100 µL total reaction volume. MilliQ water and EDTA were used as negative and positive controls, respectively.