Bio-Guided Isolation of Antimalarial Metabolites from the Coculture of Two Red Sea Sponge-Derived Actinokineospora and Rhodococcus spp.

Coculture is a productive technique to trigger microbes’ biosynthetic capacity by mimicking the natural habitats’ features principally by competition for food and space and interspecies cross-talks. Mixed cultivation of two Red Sea-derived actinobacteria, Actinokineospora spheciospongiae strain EG49 and Rhodococcus sp. UR59, resulted in the induction of several non-traced metabolites in their axenic cultures, which were detected using LC–HRMS metabolomics analysis. Antimalarial guided isolation of the cocultured fermentation led to the isolation of the angucyclines actinosporins E (1), H (2), G (3), tetragulol (5) and the anthraquinone capillasterquinone B (6), which were not reported under axenic conditions. Interestingly, actinosporins were previously induced when the axenic culture of the Actinokineospora spheciospongiae strain EG49 was treated with signalling molecule N-acetyl-d-glucosamine (GluNAc); this finding confirmed the effectiveness of coculture in the discovery of microbial metabolites yet to be discovered in the axenic fermentation with the potential that could be comparable to adding chemical signalling molecules in the fermentation flask. The isolated angucycline and anthraquinone compounds exhibited in vitro antimalarial activity and good biding affinity against lysyl-tRNA synthetase (PfKRS1), highlighting their potential developability as new antimalarial structural motif.


Introduction
Exploring microbial forms of communication and utilising them in the production of secondary metabolites is of benefit in the process of natural products drug discovery [1]. Thus far, microbial secondary metabolites remain the major source for antimicrobial agents [2][3][4]. However, gene sequencing of many microbial genome showed that several species, mainly filamentous bacteria and fungi, apply a considerable part of their genes

Identification of Red Sea Sponge-Associated Actinobacteria
Two Red sea sponge-associated actinobacteria were isolated and taxonomically identified. Actinokineospora spheciospongiae strain EG49 was previously characterised [18,19]. The other actinobacterial strain was taxonomically identified as Rhodococcus sp. UR59, according to its morphology and its 16S rRNA genome sequence and phylogenetic analyses ( Figure 1). , x FOR PEER REVIEW 3 of 14 coculture extract led to the isolation and characterisation of a few active metabolites against P. falciparum. A potential antimalarial target is proposed on the basis of molecular docking experiments against a number of reported targets.

Identification of Red Sea Sponge-Associated Actinobacteria
Two Red sea sponge-associated actinobacteria were isolated and taxonomically identified. Actinokineospora spheciospongiae strain EG49 was previously characterised [18,19]. The other actinobacterial strain was taxonomically identified as Rhodococcus sp. UR59, according to its morphology and its 16S rRNA genome sequence and phylogenetic analyses ( Figure 1).

Metabolomics Analysis of the Coculture Extract of Actinokineospora spheciospongiae Strain EG49 and Rhodococcus sp. UR59 Using LC-HRMS
The analysis of the metabolomics data (Table 1) revealed 34 microbial secondary metabolites, of which 9 were detected from Actinokineospora spheciospongiae strain EG49 and the rest from Rhodococcus sp. UR59. Additionally, the analysis revealed the presence of diverse microbial chemical classes, namely, 10 angucyclines, 7 peptides, 3 macrolides, 3 anthraquinones, 2 polyenes, 2 polyethers, 2 phenolics, and 1 glycolipid. The predicted formula C16H18N2O4 was annotated as mitomycin-K [20,21], whereas C18H14O6 was dereplicated as fluostatin-B, an inhibitor of dipeptidyl peptidase III that was previously isolated from Streptomyces sp. TA-3391 [22]. Moreover, the predicted formulas C32H33O15 and C31H33O13 were dereplicated as actinosporin A and C, respectively, which were discovered from the culture of Actinokineospora spheciospongiae strain EG49 [23,24]. The formula C21H18O8 was dereplicated as daunomycinone, which was reported from Streptomyces co-
The formulas C 26 H 25 O 11 and C 25 H 24 O 8 were dereplicated as atramycin A and B, respectively. These isotetracenone metabolites were discovered from Streptomyces atratus BY90 [26]. Additionally, the suggested molecular formula C 18 H 12 O 5 was dereplicated as lagumycin B, which was previously isolated from Micromonospora sp. [27], while the formula C 16 H 12 O 5 was dereplicated as the isoflavonoid kakkatin that was reported from the soil-derived Streptomyces strain YIM GS3536. Moreover, it was discovered in another terrestrial Streptomyces sp. GW39/1530 [28,29]. Furthermore, the molecular formula C 9 H 9 NO 3 was dereplicated as erbstatin, a simple dehydrotyrosine derivative isolated from Streptomyces amnkusaensis [30,31]. Additionally, the molecular formula C 36 H 48 N 2 O 8 was dereplicated as ansatrienin A, previously detected in Streptomyces collinus [32]. Moreover, the formulas C 25 H 47 N 5 O 4 , C 26 H 49 N 5 O 4 , and C 28 H 53 N 5 O 4 were dereplicated as cyclic tetrapeptides rhodopeptin C1, C2, and B5, respectively, which were formerly reported in Rhodococcus sp. [33,34]. The formula C 32 H 48 N 6 O 9 was dereplicated as the peptide actinoramide B, which was detected in a marine bacterium highly corelated to the genus Streptomyces [35]. Likewise, the formula C 17 H 26 O 4 was dereplicated as cineromycin-B antibiotic that showed significant MRSA inhibition, which was isolated from the actinomycetales strain INA 2770 [36]. The formula C 19 H 27 N 5 O 7 was annotated as heterobactin B, a siderophore discovered from Rhodococcus erythropolis IGTS8 [37], while the formula C 26 H 39 NO 5 was dereplicated as piericidin-F, which was reported from Streptomyces sp. CHQ-64 [38]. Additionally, the formula C 27 H 39 NO 7 was annotated as migrastatin, which was reported as a tumour cell migration inhibitor and isolated from Streptomyces sp. MK929-43F1 [39]. Moreover, the formula C 24 H 46 N 6 O 8 was dereplicated as proferrioxamine-A1, a siderophore isolated from Streptomyces xinghaiensis NRRL B-24674T [40]. Furthermore, the formula C 23 H 38 O 5 was dereplicated as the 16-membered lactone protylonolide, which was identified as the metabolite of mycaminose idiotroph that has been obtained from Streptomyces fradiae KA-427 [41]. Moreover, the formula C 37 H 62 O 11 was dereplicated as the polyether 26-deoxylaidlomycin isolated from Streptoverticillium olivoreticuli IMET 43,782 [42], while the suggested formula C 35 H 58 O 10 was dereplicated as macrolide kaimonolide B, which was discovered in Streptomyces sp. no. 4155 and shown to significantly inhibit plant growth [43]. Furthermore, the formula C 25 H 44 O 7 was dereplicated as 8,15-dideoxylankanolide, which was reported in Streptomyces rochei 7434AN4 [44]. The molecular formula C 34 H 60 O 10 was identified as the polyether antibiotic ferensimycin-A, previously discovered in Streptomyces sp. no. 5057 [45]. Likewise, the formula C 26 H 46 N 6 O 5 was identified as the cytotoxic peptide lucentamycin C, which was reported from a marinederived actinomycete Nocardiopsis lucentensis CNR-712 [46]. Finally, the formula C 50 H 92 O 14 was dereplicated as glucolipsin-A, a glucokinase activator that has been isolated from Streptomyces puvpuvogenisclevoticus [47].
It is worth noting that the compounds listed in Table 1 were traced in the LC-HRESIMS analysis of the coculture extract. The producing strain for each compound was predicted on the basis of literature. However, mitomycin-K, 8,15-dideoxylankanolide, piericidin-F, migrastatin, kaimonolide B, rhodopeptin C1, rhodopeptin C2, and rhodopeptin B5 were also traced in the axenic culture of Rhodococcus sp. UR59. Additionally, actinosporins A and C, and UK-2B were also traced in the axenic culture of Actinokineospora spheciospongiae strain EG49. All other reported metabolites in Table 1 were not traced in the axenic cultures and were induced during the coculture fermentation.

Identification of the Isolated Compounds (1-8)
Chemical structures of the purified metabolites 1-8 from the coculture were assigned on the basis of comparing the LC-HRESIMS analysis, 1D and 2D NMR spectral data, and optical rotation measurements to the published literature ( Figure 2). Accordingly, compounds 1-3 have been previously isolated from Actinokineospora spheciospongiae strain EG49 and identified as the angucyclinone antibiotics actinosporin E, H, and G, respectively, through the activation of their cryptic gene cluster by N-acetylglucosamine [50]. Compound 4 was assigned as spoxazomicin C of the pyochelin family of antibiotics, which was previously isolated from the culture broth of the endophyte Streptosporangium oxazolinicum K07-0460T [51]. In contrast, compound 5 was previously identified as the angucyclinone antibiotic tetrangulol, which was previously isolated from Streptomyces rimosus [58] and recently from Amycolatopsis sp. HCa1 [59]. Compound 6 was previously discovered as capillasterquinone B, an anthraquinone that was isolated from the crinoid Capillaster multiradiatus [57]. Moreover, compound 7 was identified as L-tryptophanamide. We propose it as an artefact as it was not traced in the LC-MS analysis of either the axenic or the coculture extracts, and thus it was probably generated during the fractionation and purification process. Finally, compound 8 was isolated from Streptomyces sp. 517-02 [57] and identified as UK-2B, an antifungal antibiotic with similarity in structure to antimycin A [60].
Capillaster multiradiatus [57]. Moreover, compound 7 was identified as L-tryptophanamide. We propose it as an artefact as it was not traced in the LC-MS analysis of either the axenic or the coculture extracts, and thus it was probably generated during the fractionation and purification process. Finally, compound 8 was isolated from Streptomyces sp. 517-02 [57] and identified as UK-2B, an antifungal antibiotic with similarity in structure to antimycin A [60]. The same Actinokineospora spheciospongiae strain EG49 was subjected to N-acetyl-Dglucosamine (GluNAc)-mediated silent gene activation to produce new actinosporins E-H, the same actinosporins E (1), G (3), and H (2) discussed here under coculture and aglycone angucycline tetrangulol (5), which was not reported from the axenic culture treated with GluNAc [50]. The amino sugar GluNAc is a signalling molecule that can induce microbial secondary metabolism. It is present as a cell wall component in peptidoglycan or chitin in the bacterial or fungal cell wall, respectively [60,61]. Having observed a similar induction when Actinokineospora spheciospongiae strain EG49 was cocultured with Rhodococcus sp. UR59, we can assume that Rhodococcus sp. UR59 directed the biosynthesis of actinosporins in a similar way to GluNAc. This could be as exudation of GluNAc by one of the species into the coculture environment to trigger antibiotic production more likely from Rhodococcus sp. UR59 as defence molecules. The studies further support this and demonstrated that GluNAc is secreted by bacteria under malnourished conditions to signal antibiotic production against opposite competitors in the vicinity [62]. However, this requires further studies on the coculture medium to identify excreted GluNAc or compounds with similar signalling function.

Antimalarial Screening
Angucyclines are microbial secondary metabolites known as promising antimicrobial, anticancer, and antimalarial agents [63][64][65]. The core structure of angucyclines is characterised by a benz[α]anthracene ring, an angular tetracycline ring system [60]. The reported angucyclines can be categorised as aglycones such as saccharosporones A, B, and The same Actinokineospora spheciospongiae strain EG49 was subjected to N-acetyl-Dglucosamine (GluNAc)-mediated silent gene activation to produce new actinosporins E-H, the same actinosporins E (1), G (3), and H (2) discussed here under coculture and aglycone angucycline tetrangulol (5), which was not reported from the axenic culture treated with GluNAc [50]. The amino sugar GluNAc is a signalling molecule that can induce microbial secondary metabolism. It is present as a cell wall component in peptidoglycan or chitin in the bacterial or fungal cell wall, respectively [60,61]. Having observed a similar induction when Actinokineospora spheciospongiae strain EG49 was cocultured with Rhodococcus sp. UR59, we can assume that Rhodococcus sp. UR59 directed the biosynthesis of actinosporins in a similar way to GluNAc. This could be as exudation of GluNAc by one of the species into the coculture environment to trigger antibiotic production more likely from Rhodococcus sp. UR59 as defence molecules. The studies further support this and demonstrated that GluNAc is secreted by bacteria under malnourished conditions to signal antibiotic production against opposite competitors in the vicinity [62]. However, this requires further studies on the coculture medium to identify excreted GluNAc or compounds with similar signalling function.

Antimalarial Screening
Angucyclines are microbial secondary metabolites known as promising antimicrobial, anticancer, and antimalarial agents [63][64][65]. The core structure of angucyclines is characterised by a benz[α]anthracene ring, an angular tetracycline ring system [60]. The reported angucyclines can be categorised as aglycones such as saccharosporones A, B, and C [60], and glycosylated angucyclines such as pseudonocardones A−C [63] and urdamycinone E, urdamycinone G, and dehydroxyaquayamycin isolated from fungal and bacterial strains [62]. However, different antimalarial activity profiles between aglycones and glycosylated angucyclines have not been explained.
The potential antiparasitic effectiveness of the angucycline scaffold and the promising antimalarial effect exhibited by the total extract of the coculture of Actinokineospora spheciospongiae strain EG49 and Rhodococcus sp. UR59 (IC 50 value of 0.13 µg/mL, Table 2) when screened against Plasmodium falciparum have encouraged us to perform large-scale coculture fermentation. Large-scale fermentation followed by liquid-liquid fractionation and HPLC purification of the active sub-fraction led to the isolation of eight metabolites. The antimalarial screening of the isolated compounds indicated that the angucycline glycosides 1-3 and aglycone 5 and the anthraquinone 6 exhibited antimalarial effect with IC 50 values in the range of 9-13.5 µg/mL in comparison to the IC 50 value of the positive control chloroquine (0.022 µg/mL). The activity of the compounds 1-3, 5, and 6 was further studied by docking against a few known drug targets to suggest these compounds as potential leads to be developed for enhanced activity. It worth noting that the isolated molecules did not show the expected antimalarial activity, which could be attributed to either the synergistic effect of microbial metabolites in the coculture extract or the presence of minor molecules that were too scarce to be isolated even after large-scale fermentation.

Compound
Binding

General Experimental
Extract purification was conducted by preparative Agilent 1100 series HPLC equipped with gradient pump and DAD using a reversed-phase Sunfire (C18, 5 µm, 10 × 250 mm, serial no. 226130200125). All 1D and 2D NMR spectral data were acquired using a JEOL ECZ-R500 NMR spectrometer equipped with a Royal 5 mm combined broadband and inverse probe. Thermo LTQ Orbitrap coupled to an HPLC system was utilised to acquire HRESIMS data using capillary temperature of 260 °C, capillary voltage of 45 V, sheath gas flow rate of 40-50 arbitrary units, auxiliary gas flow rate of 10-20 arbitrary units, spray voltage of 4.5 kV, and mass range of 100-2000 amu (maximal resolution of 60,000). Optical rotations and UV spectra acquisition were acquired using a Perkin-Elmer 343 polarimeter and Perkin-Elmer Lambda2 UV-VIS spectrometer, respectively.

Actinomycetes Isolation
Callyspongia sp. was collected from Hurghada (Red Sea, Egypt) at a depth of 5 m and latitude 27°17′01.0′′ N and longitude 33°46′21.0′′ E. The sponge specimen was identified by Prof. El-Sayd Abed El-Aziz (Department of Invertebrates Lab., National Institute of

General Experimental
Extract purification was conducted by preparative Agilent 1100 series HPLC equipped with gradient pump and DAD using a reversed-phase Sunfire (C18, 5 µm, 10 × 250 mm, serial no. 226130200125). All 1D and 2D NMR spectral data were acquired using a JEOL ECZ-R500 NMR spectrometer equipped with a Royal 5 mm combined broadband and inverse probe. Thermo LTQ Orbitrap coupled to an HPLC system was utilised to acquire HRESIMS data using capillary temperature of 260 • C, capillary voltage of 45 V, sheath gas flow rate of 40-50 arbitrary units, auxiliary gas flow rate of 10-20 arbitrary units, spray voltage of 4.5 kV, and mass range of 100-2000 amu (maximal resolution of 60,000). Optical rotations and UV spectra acquisition were acquired using a Perkin-Elmer 343 polarimeter and Perkin-Elmer Lambda2 UV-VIS spectrometer, respectively.

Actinomycetes Isolation
Callyspongia sp. was collected from Hurghada (Red Sea, Egypt) at a depth of 5 m and latitude 27 • 17 01.0 N and longitude 33 • 46 21.0 E. The sponge specimen was identified by Prof. El-Sayd Abed El-Aziz (Department of Invertebrates Lab., National Institute of Oceanography and Fisheries, Egypt). The sponge was transported in a plastic bag in seawater to the laboratory and washed thoroughly with sterile seawater. The surface sterilised specimen was cut into pieces of ≈1 cm 3 , followed by vigorous homogenising with 10 volumes of sterile seawater in a pre-sterilised mortar. Serially diluted supernatant (10 −1 , 10 −2 , 10 −3 ) was subsequently plated on to the sterile agar plates. For the isolation of different actinomycetes, we used M1, ISP2, and marine agar (MA) media were used [18]. The isolation of slow-growing actinomycetes was performed by supplementing all media with filtered 25 µg/mL nalidixic acid, 25 µg/mL nystatin, and 100 µg/mL cycloheximide. The inoculated plates were stored in an incubator for 6-8 weeks at 30 • C. Subculturing of distinct colony morphotypes resulted in pure strains. Rhodococcus sp. UR59 was cultured on ISP2 medium and preserved in 20% glycerol at −80 • C. On the other hand, Actinokineospora spheciospongiae strain EG49 was previously recovered and identified from the Red Sea sponge Spheciospongia vagabunda [18].

Molecular Identification and Phylogenetic Analysis
With reference to Hentschel et al., we carried out 16S rRNA gene amplification, cloning, and sequencing using 27F and 1492RRNA as universal primers [18]. By using the Pintail programme, we identified chimeric sequences [67]. The sequence's genus level affiliation was validated using the Project Classifier of the Ribosomal Database. All the sequences were classified at the genus level by the RDP Classifier (g 16srrna, f allran) and confirmed with the SILVA Incremental Aligner (SINA) [68]. Using the SINA Web Aligner, an alignment was determined again (variability profile: bacteria). The Gap-only position with trimALL was eliminated (-noallgaps). The best fitting model was initially calculated for phylogenetic tree construction with the Model Generator. To produce the phylogenetic tree, we applied RAxML (-f a-m GTRGAMMA-x 12345-p 12345 -# 1000) and the estimated model with 1000 bootstrap resamples. With Interactive Tree of Life (ITOL) [69], visualisation was achieved. The BLAST with the accession number MW453143 was deposited at Genebank.

Co-Cultivation and Extract Preparation
Rhodococcus sp. UR59 and Actinokineospora spheciospongiae strain EG49 were cultivated on liquid media M1 and ISP2 as axenic and cocultures. A total of 20 mL of 3-day-old culture of Rhodococcus sp. was used for large scale fermentation. Rhodococcus sp. UR59 was transferred to 20 × 2 L Erlenmeyer flasks containing 1 L of ISP2 medium pre-inoculated with 20 mL of 4-day-old Actinokineospora spheciospongiae strain EG49 and left for 7 days at 25 • C and 180 rpm in a shaker incubator. After fermentation, the culture was filtered, and the supernatant was extracted twice with ethyl acetate (1.5 L each) followed by evaporation under vacuum to provide the ethyl acetate extract (850 mg).

Metabolic Profiling
For mass spectrometry analysis, the dry ethyl acetate extracts from different microbial and coculture samples were dissolved in MeOH at 1 mg/mL and subjected to metabolic analysis using LC-HRESIMS according to Abdelmohsen et al. [23]. An Acquity UPLC system coupled to a Synapt G2 HDMS qTOF hybrid mass spectrometer (Waters, Milford, CT, USA) was used to acquire the HRMS data using capillary temperature at 320 • C, spray voltage at 4.5 kV, and mass range of m/z 150-1500; both positive and negative ESI modes were applied. The MS was processed using MZmine 2.20 on the basis of the defined parameters [23]. The chromatogram builder and chromatogram deconvolution were detected and followed by mass ion peaks. The isotopes were differentiated by grouper isotopic peaks and the missing peaks were depicted using the gap-filling peak finder. Then, molecular formula prediction and peak identification were conducted from the processed positive and negative ionisation mode datasets. Finally, the peaks were dereplicated against the Dictionary of Natural Products (DNP) database.

Antimalarial Screening
The Malstat assay was used as mentioned earlier to assess the compounds' antimalarial effect [70,71]. The compounds were dissolved in DMSO (Sigma Aldrich, Taufkirchen, Germany) at concentrations ranging from 50 µg/mL to 0.4 µg/mL, and synchronised P. falciparum 3D7 ring stage cultures were placed in duplicate at a parasite level of 1% in 96-well plates (200 µL/well). Chloroquine (CQ; Sigma Aldrich, Taufkirchen, Germany) was used as a positive control. The P. falciparum 3D7 parasite was cultured with the compounds at 37 • C in 5% O 2 , 5% CO 2 , and 90% N 2 for 72 h. After this, 20 µL was transferred to 100 µL of the Malstat reagent (0.1% Triton X-100, 1 g of L-lactate, 0.33 g Tris, and 33 mg of APAD (3-acetylpyridine adenine dinucleotide; Taufkirchen, Germany)) dissolved in 100 mL of distilled water (pH 9.0) in a 96-well microtiter plate. The plasmodial lactate dehydrogenase (LDH) activity was then evaluated by adding to the Malstat reaction 20 µL of a 1:1 mixture of diaphorase (1 mg/mL) and nitro blue tetrazolium (NBT). The optical densities were estimated at 630 nM, and the IC 50 values were determined using the GraphPad Prism software version 5 from variable-slope sigmoidal dose-response curves (GraphPad Software Inc., La Jolla, CA, USA).

Molecular Docking
Docking analysis was carried out using the Discovery Studio 2.5 software (Accelrys Inc., San Diego, CA, USA). Completely automatic docking tool using "Dock ligands (CDOCKER)" procedure operating on Intel Core i32370 CPU @ 2.4 GHz 2.4 GHz, RAM Memory 2 GB under the Windows 10.0 system. Furthermore, these docked compounds were assembled using a software Chem 3D ultra 12.0 (Cambridge Soft Corporation, USA (2010)), and then sent to the Discovery Studio 2.5 software. From this, an automatic protein formulation procedure was conducted through the MMFF94 forcefield with the binding site sphere recognised by the software. The receptor was recorded as "input receptor molecule" in the CDOCKER protocol explorer. Establishing this, the test compounds were subjected to force fields to obtain the minimum energy structure. These poses were ranked and studied thoroughly, showing the best ligand-HDAC interactions from the calculations and 2D and 3D examinations [72,73].

Conclusions
Microbial coculture continues to prove its efficiency in triggering the production of cryptic microbial secondary metabolites. Mixed cultivation of two Red Sea-derived actinobacteria, namely, Actinokineospora spheciospongiae strain EG49 and Rhodococcus sp. UR59, resulted in the induction of several non-traced metabolites in their axenic cultures. Interestingly, actinosporins E-H were reported to be induced when the axenic culture of the Actinokineospora spheciospongiae strain EG49 was treated with the signalling molecule GluNAc. Such induction was comparable to that made by the Rhodococcus sp. UR59 in the coculture environment, providing the effectiveness of co-cultivation in the discovery of microbial metabolites yet to be discovered in the axenic fermentation with the potential that could be comparable to adding signalling molecules in the fermentation flask. Additionally, the induced actinosporins exhibited a promising antimalarial effect that is likely to be through the inhibition of P. falciparum lysyl-tRNA synthetase, which requires further investigation as an interesting structural motif for the development of new antimalarial therapeutics.