Novel South African Rare Actinomycete Kribbella speibonae Strain SK5: A Prolific Producer of Hydroxamate Siderophores including New Dehydroxylated Congeners

In this paper, we report on the chemistry of the rare South African Actinomycete Kribbella speibonae strain SK5, a prolific producer of hydroxamate siderophores and their congeners. Two new analogues, dehydroxylated desferrioxamines, speibonoxamine 1 and desoxy-desferrioxamine D1 2, have been isolated, together with four known hydroxamates, desferrioxamine D1 3, desferrioxamine B 4, desoxy-nocardamine 5 and nocardamine 6, and a diketopiperazine (DKP) 7. The structures of 1–7 were characterized by the analysis of HRESIMS and 1D and 2D NMR data, as well as by comparison with the relevant literature. Three new dehydroxy desferrioxamine derivatives 8–10 were tentatively identified in the molecular network of K. speibonae strain SK5 extracts, and structures were proposed based on their MS/MS fragmentation patterns. A plausible spb biosynthetic pathway was proposed. To the best of our knowledge, this is the first report of the isolation of desferrioxamines from the actinobacterial genus Kribbella.


Introduction
Minerals are essential for the growth, development, and propagation of living organisms. Iron, a crucial element, is involved in various cellular processes including oxygen metabolism, electron transfer, DNA and RNA biosynthesis, and as a catalyst in enzymatic processes where it serves as

Results and Discussion
The K. speibonae strain SK5 was isolated from a topsoil sample collected from Stellenbosch in the Western Cape Province of South Africa [17]. The strain was grown in an International Streptomyces Project medium 2 (ISP2) broth and an Amberlite XAD 16N resin was added after 14 days of incubation at 30 • C with constant shaking. Organic solvents, methanol (MeOH), ethyl acetate (EtOAc), and dichloromethane (CH 2 Cl 2 ), were used sequentially to extract the organics from the combined resin and culture broth. Then, the MeOH, EtOAc, and CH 2 Cl 2 extracts were subjected to a separate high-pressure liquid chromatography-diode array detection, high-resolution electrospray mass spectrometry (HPLC-DAD/HRESIMS) analyses for chemical profiling ( Figure S1). Further MS/MS and Global Natural Product Social (GNPS) molecular network analyses of the extracts revealed the presence of several siderophores and DKPs, some of which have not been reported previously ( Figure S2). The MeOH, EtOAc, and CH 2 Cl 2 extracts were combined and fractionated using a series of purification steps, including a modified Kupchan method [18], solid phase extraction (SPE), and HPLC to yield two new siderophores, speibonoxamine 1 and desoxy-desferrioxamine D 1 2, and four known hydroxamates, 3-6 and a DKP 7 ( Figure 1).
Molecules 2020, 25, x FOR PEER REVIEW 3 of 12 The K. speibonae strain SK5 was isolated from a topsoil sample collected from Stellenbosch in the Western Cape Province of South Africa [17]. The strain was grown in an International Streptomyces Project medium 2 (ISP2) broth and an Amberlite XAD 16N resin was added after 14 days of incubation at 30 °C with constant shaking. Organic solvents, methanol (MeOH), ethyl acetate (EtOAc), and dichloromethane (CH2Cl2), were used sequentially to extract the organics from the combined resin and culture broth. Then, the MeOH, EtOAc, and CH2Cl2 extracts were subjected to a separate high-pressure liquid chromatography-diode array detection, high-resolution electrospray mass spectrometry (HPLC-DAD/HRESIMS) analyses for chemical profiling ( Figure S1). Further MS/MS and Global Natural Product Social (GNPS) molecular network analyses of the extracts revealed the presence of several siderophores and DKPs, some of which have not been reported previously ( Figure S2). The MeOH, EtOAc, and CH2Cl2 extracts were combined and fractionated using a series of purification steps, including a modified Kupchan method [18], solid phase extraction (SPE), and HPLC to yield two new siderophores, speibonoxamine 1 and desoxy-desferrioxamine D1 2, and four known hydroxamates, 3-6 and a DKP 7 ( Figure 1).
The molecular formula of compound 1, isolated as a colourless amorphous solid, was deduced as C 27  The 1 H-NMR of 1 in DMSO-d 6 showed only six signals ( Figure S4), including four methylenes (δ H 3.00, 2.27, 1.36, 1.22), one methyl (δ H 1.77), and one NH (δ H 7.77) with integrals of 12, 8, 12, 6, 6, and 6, respectively. The number of carbon atoms observed in the HRESIMS was five times higher than the number of signals observed in the 13 C-NMR spectrum (δ C 171. 7, 169.4, 38.9, 31.4, 29.3, 24.3). These results suggested that compound 1 had a symmetrical structure and/or repeating motifs. Analysis of the 1 H-1 H COSY spectrum together with integrals of the signals in the 1 H-NMR spectrum revealed two main spin systems, one of which consisted of three repeating motifs H-3 through H-7, H3 through H-7 , and H-3" through H-7", and the other comprised two repeating motifs H-9 through H-10 and H-9 through H-10 . Careful examination of the 1D NMR data and 2D NMR data suggested that the symmetrical structure of 1 consists of two succinyls, two acyl cadaverine (AC) units, and one cadaverine moiety, characteristic of a hydroxamate siderophore. The heteronuclear multiple bond correlations (HMBC) from H-7 (δ H 3.00) to C-8 (δ C 171.7) and H-3" (δ H 3.00) to C-11 (δ C 171.7) established the connectivity of the AC subunits to the succinyl groups. Furthermore, the cross peaks from H-3 (δ H 3.00) to C-11 (δ C 171.7) and H-7 (δ H 3.00) to C-8 (δ C 171.7) established the attachment of the cadaverine moiety to the rest of the molecule.
The final structural analysis of 1 was confirmed by comparison with the spectroscopic data reported for desferrioxamine D 1 (3) in the literature [21], which differs from 1 in the presence of N-hydroxy groups in the structure. Therefore, compound 1 represents a new non-hydroxylated desferrioxamine, which was named speibonoxamine after the producing organism, Kribbella speibonae strain SK5. This is the first report of a non-hydroxylated desferrioxamine. Compound 1 may not be efficient in sequestering iron from the environment because of the lack of the hydroxamate moiety.
The molecular formula of compound 2, also isolated as a colourless amorphous solid, was established as C 27 Figure S9).
The mass of compound 2 was 32 mass units greater than 1 indicative of the presence of two additional oxygen atoms in the structure. Analysis of the 1 H-NMR data of 2 ( Figure S10) revealed chemical shifts similar to those of 1, except for signals at δ H 1.50 (H-6 and H-6 ) and 3.45 (H-7 and H-7 ) as the most distinguishable change. The observed proton (δ H 3.45) and carbon (δ C 47.6) chemical shifts in 2 were more downfield than in 1 (δ H 3.00, δ C 38.9), indicating that the methylenes C-7 and C-7 were deshielded by the adjacent N-hydroxy groups in 2. Likewise, the methylene signals at (δ H 1.50, δ C 26.5) were downfield in 2 compared to that in 1 (δ H 1.36, δ C 29.3) because of the attachment of C-6 and C-6 to the deshielded C-7 and C-7 in 2, respectively. Compound 2 was linked to desoxynocardamine (5) and desferrioxamine D 1 (3) in the GNPS molecular network with a mass difference of 2 and 16 Da, respectively ( Figure S32) indicating that 2 is an acyclic analogue of 5 or dehydroxylated analogue of 3. The structure of 2 was determined by analyses of the MS/MS fragmentation data together with a comprehensive interpretation of the 1D and 2D NMR data of 2 and comparison with literature data. Compound 2 is a new dehydroxy analogue of compound 3 and hence has been assigned the name desoxy-desferrioxamine D 1 [21].

Molecular Network Analysis
Dereplication of the metabolites produced by the K. speibonae strain SK5 was pivotal in the detection and isolation of the new compounds. Spectrometric data of the metabolites in the crude extract were searched against the natural product database, AntiBase (2017), with subsequent analysis of the data on the GNPS molecular networking platform [23], which grouped the compounds into clusters based on the similarity in MS/MS fragmentation patterns. Several known siderophores and DKPs were identified in the molecular network, including desferrioxamine H 11, ferrioxamine B 12, ferrioxamine E 13, ferrioxamine D 1 14, arthrobactin 15 and the DKP, hexahydro-3-(phenylmethyl)-pyrrolo[1,2-a]pyrazine-1,4-dione 16 ( Figure S32) [13,19]. Furthermore, three new dehydroxylated siderophores 8-10 were putatively identified in the molecular network, in addition to the new compounds, speibonoxamine 1 and desoxy-desferrioxamine D 1 2, isolated and reported in this study.
Compound 8 was linked to compound 2 when default parameters were used to generate the GNPS molecular network ( Figure S33 Table  S1). Inspection of the fragment ions of compounds 1, 2, and 8 was very useful in assigning the position of the hydroxyl group, especially fragment ion 411.2596 (C 20 H 35 O 5 N 4 ), which was present in 1 and 8, but absent in 2 ( Figures S34-S36). Although the new compound 8 could not be isolated due to a paucity of material, its structure was proposed based on the MS/MS fragmentation data analysis and was assigned the name didesoxy-desferrioxamine D 1 (Figure 2).
Molecules 2020, 25, x FOR PEER REVIEW 5 of 12 in 1 and 8, but absent in 2 ( Figures S34-S36). Although the new compound 8 could not be isolated due to a paucity of material, its structure was proposed based on the MS/MS fragmentation data analysis and was assigned the name didesoxy-desferrioxamine D1 ( Figure 2). The results suggest that the K. speibonae strain SK5 is a prolific producer not only of hydroxamate siderophores, but also dehydroxylated and non-hydroxylated desferrioxamines. Although the presence of siderophores has been previously identified in the genomes of Kribbella species [17,24], to our knowledge this is the first report of the isolation of hydroxamates from this genus.

Proposed Biosynthetic Pathway
Biosynthesis of the desferrioxamine class of hydroxamate siderophores has been elucidated in Streptomyces and is mediated by nonribosomal peptide synthetase-independent siderophore (NIS) synthetases [9,10,25,26]. Every NIS biosynthetic pathway identified to date contains at least one synthetase with a high sequence similarity to IucA/IucC and such synthetases have thus become the characteristic feature of these pathways [10]. In silico analysis of the annotated genome of the K. speibonae strain SK5 (GenBank accession number: SJJY00000000) identified the biosynthetic gene cluster (BGC) encoding the IucA/IucC synthetase that is likely responsible for producing 1-6 ( Figure  3A, Table 1). The results suggest that the K. speibonae strain SK5 is a prolific producer not only of hydroxamate siderophores, but also dehydroxylated and non-hydroxylated desferrioxamines. Although the presence of siderophores has been previously identified in the genomes of Kribbella species [17,24], to our knowledge this is the first report of the isolation of hydroxamates from this genus.

Proposed Biosynthetic Pathway
Biosynthesis of the desferrioxamine class of hydroxamate siderophores has been elucidated in Streptomyces and is mediated by nonribosomal peptide synthetase-independent siderophore (NIS) synthetases [9,10,25,26]. Every NIS biosynthetic pathway identified to date contains at least one synthetase with a high sequence similarity to IucA/IucC and such synthetases have thus become the characteristic feature of these pathways [10]. In silico analysis of the annotated genome of the K. speibonae strain SK5 (GenBank accession number: SJJY00000000) identified the biosynthetic gene cluster (BGC) encoding the IucA/IucC synthetase that is likely responsible for producing 1-6 ( Figure 3A, Table 1). The desferrioxamine (des) BGC in Streptomyces coelicolor strain M145 was proposed to comprise four genes, desABCD [25]. The spbA and spbB genes in the K. speibonae strain SK5 encode enzymes having a high sequence similarity to pyridoxal 5′-phosphate (PLP)-dependent decarboxylase and lysine-6-monooxygenase in Streptomyces strains, respectively [27]. These two enzymes are proposed to catalyze the decarboxylation of lysine 17 to form cadaverine 18 followed by hydroxylation to generate N-hydroxy-cadaverine 19 ( Figure 3B). The acyltransferase, SpbC, is likely to catalyze the Nacylation of 19 [9,25], which then undergoes ATP-dependent oligomerization and macrocylization catalyzed by SpbD and SpbE to produce the N-hydroxy-containing hydroxamate desferrioxamines 2-6 [28].
The isolation of speibonoxamine 1 from K. speibonae strain SK5 suggests that the biosynthetic enzyme, N-acyl-CoA transferase SpbC, may display substrate promiscuity, binding not only to Nhydroxy-cadaverine 19 but also cadaverine 18, the key difference between the biosynthesis of desferrioxamines and speibonoxamine 1. It is proposed that the common decarboxylation intermediate, cadaverine 18, undergoes spontaneous SpbC-catalyzed acylation, followed by SpbDand SpbE-catalyzed oligomerization of two molecules of succinate, two molecules of N-acetylcadaverine, and one molecule of cadaverine to generate 1.   The desferrioxamine (des) BGC in Streptomyces coelicolor strain M145 was proposed to comprise four genes, desABCD [25]. The spbA and spbB genes in the K. speibonae strain SK5 encode enzymes having a high sequence similarity to pyridoxal 5 -phosphate (PLP)-dependent decarboxylase and lysine-6-monooxygenase in Streptomyces strains, respectively [27]. These two enzymes are proposed to catalyze the decarboxylation of lysine 17 to form cadaverine 18 followed by hydroxylation to generate N-hydroxy-cadaverine 19 ( Figure 3B). The acyltransferase, SpbC, is likely to catalyze the N-acylation of 19 [9,25], which then undergoes ATP-dependent oligomerization and macrocylization catalyzed by SpbD and SpbE to produce the N-hydroxy-containing hydroxamate desferrioxamines 2-6 [28].
The isolation of speibonoxamine 1 from K. speibonae strain SK5 suggests that the biosynthetic enzyme, N-acyl-CoA transferase SpbC, may display substrate promiscuity, binding not only to N-hydroxy-cadaverine 19 but also cadaverine 18, the key difference between the biosynthesis of desferrioxamines and speibonoxamine 1. It is proposed that the common decarboxylation intermediate, cadaverine 18, undergoes spontaneous SpbC-catalyzed acylation, followed by SpbD-and SpbE-catalyzed oligomerization of two molecules of succinate, two molecules of N-acetyl-cadaverine, and one molecule of cadaverine to generate 1.

General Experimental Procedures
NMR spectra were obtained on a BRUKER Ascend 600 (Bruker, Billerica, MA, USA) Prodigy cryoprobe at 600 and 150 MHz for 1 H and 13 C nuclei, respectively. DMSO-d 6 (δ H 2.50, δ C 39.7), CD 3 OD (δ H 3.30, δ C 49.0), and CDCl 3 (δ H 7.25, δ C 77.00) were used for preparing samples for NMR experiments. High resolution mass spectrometric data were obtained using a Thermo Instruments MS system (LTQ XL/LTQ Orbitrap Discovery, Thermo Scientific, Bremen, Germany) coupled to a Thermo Instruments HPLC system (Accela PDA detector, Accela PDA autosampler, and Accela pump). The MS was run in a positive high-resolution mode (60,000) and MS/MS at a resolution of 7500 and a low-resolution negative mode. The following conditions were used: Capillary voltage 45 V, capillary temperature 260 • C, auxiliary gas flow rate 10-20 arbitrary units, sheath gas flow rate 40-50 arbitrary units, spray voltage 4.5 kV, mass range 100-2000 amu (maximum resolution 30,000). HPLC separations were carried out using a reverse phase C18 ACE 10 µM 10 × 250 mm column connected to an Agilent Technologies 1200 series HPLC system equipped with an Agilent Technologies 1200 series quad pump and Agilent Technologies 1200 series DAD. The amberlite XAD 16N resin was obtained from Sigma-Aldrich, Johannesburg, South Africa. All solvents used throughout were of a HPLC-grade and purchased from both Merck and Sigma-Aldrich (Johannesburg, South Africa).

Isolation and Characterization of the Strain
The K. speibonae strain SK5 was isolated from a topsoil sample collected from the town of Stellenbosch in the Western Cape Province of South Africa using a newly developed Kribbella-selective medium [17].

Fermentation
A liquid stock culture of the K. speibonae strain SK5 was inoculated into a 15 mL International Streptomyces Project medium 2 (ISP2) broth (yeast extract 4 g, malt extract 10 g, glucose 4 g, in 1 L H 2 O, pH 7.3 [29]) in a 250 mL Erlenmeyer flask and incubated for five days at 30 • C with shaking. Then, the entire culture was inoculated into a 50 mL ISP2 broth in a 500 mL Erlenmeyer flask and incubated at 30 • C with shaking for four days. This culture was subsequently split into three parts, which were used as the inocula for 3 × 100 mL ISP2 broths in 1000 mL Erlenmeyer flasks and incubated at 30 • C with shaking for four days. Then, each 100 mL culture was inoculated into a 1000 mL ISP2 broth in a 5000 mL Erlenmeyer flask and incubated at 30 • C with shaking. After 14 days of incubation, the Amberlite XAD 16N resin (50 g/L) was added under sterile conditions to each flask and further incubated for 6 h at 30 • C with shaking. Subsequently, the cultures were harvested and filtered under pressure using a piece of glass wool placed in a Buchner funnel. The filtrate was partitioned in a separating funnel with an equal volume of ethyl acetate. Then, the cell mass mixed with the Amberlite XAD 16N resin (containing the adsorbed organics) was extracted sequentially with MeOH (3×), then EtOAc (3×), and finally, CH 2 Cl 2 (3×). All the organic extracts were concentrated under a reduced pressure to give 3.43 g of the MeOH extract, 0.51 g of the EtOAc extract, and 0.19 g of the CH 2 Cl 2 extract. The extracts were subjected to HPLC-DAD/HRESIMS analyses.

HPLC-DAD/HRESIMS Analyses
Chemical profiling of the MeOH, EtOAc, and CH 2 Cl 2 extracts was performed by HPLC-DAD/ HRESIMS analyses. Each extract (10 µL, 0.1 mg/mL in MeOH) was injected and chromatographically separated on a C18 reverse-phase HPLC column (ACE 10 µM 10 × 250 mm) using a linear gradient from 95% solvent A (0.1% formic acid in water) to 100% solvent B (0.1% formic acid in acetonitrile) for 25 min, followed by 100% B for 5 min at a flow rate of 1.5 mL/min. The DAD of the HPLC was scanned from 200-400 nm. The MS was run in a positive high-resolution mode (60,000) and MS/MS at a resolution of 7500 and a low-resolution negative mode. The Xcalibur software was used to process the raw data. The exact mass and molecular formula of each peak was entered as a single query in the commercially available AntiBase (2017) Natural Compound Identifier (https://www.wiley.com/enus/AntiBase%3A+The+Natural+Compound+Identifier-p-9783527343591) to ascertain whether the data matched any compound in the database. The HPLC-DAD/HRESIMS profiles of the three extracts showed similar chemical profiles, hence they were combined to obtain 4.13 g.

Molecular Networking
Raw data obtained from the LC-MS/MS system were converted to a mzXML format using the ProteoWizard tool MSconvert (version 3.0.10051, Vanderbilt University, Nashville, TN, USA) [30]. All mzXML data were uploaded to the Global Natural Products Social (GNPS) molecular networking (MN) webserver3 (http://gnps.ucsd.edu) and analyzed using the MN workflow [23]. The data were filtered by removing all MS/MS fragment ions within +/− 17 Da of the precursor m/z. MS/MS spectra were window filtered by choosing only the top six fragment ions in the +/− 50 Da window throughout the spectrum. The precursor ion mass tolerance was set to 0.02 Da and a MS/MS fragment ion tolerance of 0.02 Da. Then, a network was created where edges were filtered to have a cosine score above 0.7 and more than three matched peaks. Further, edges between two nodes were kept in the network if and only if each of the nodes appeared in each other's respective top 10 most similar nodes. Finally, the maximum size of a molecular family was set to 100, and the lowest scoring edges were removed from molecular families until the molecular family size was below this threshold. The spectra in the network were then searched against GNPS' spectral libraries. The library spectra were filtered in the same manner as the input data. All matches kept between the network spectra and library spectra were required to have a score above 0.7 and at least three matched peaks. The output of the molecular network was visualized using the Cytoscape version 3.7.2 (https://cytoscape.org/) [31] and displayed using the settings "preferred layout" with "directed" style. The nodes (compounds) originating from the culture medium and solvent control (MeOH) were excluded from the original network to enable visualization of only the K. speibonae strain SK5 metabolites derived from the MeOH, EtOAC, and CH 2 Cl 2 extracts of the cultures.

Conclusions
The K. speibonae strain SK5 is a prolific producer of hydroxamate (desferrioxamine) siderophores and their dehydroxy analogues. Two new desferrioxamines, speibonoxamine and desoxy-desferrioxamine D 1 , four known hydroxamates, and a DKP were isolated from K. speibonae strain SK5. Furthermore, several new and known siderophores were identified in the K. speibonae strain SK5 extracts, and their plausible structures were determined by the MS/MS fragmentation analysis. This is the first report of siderophores isolated from the genus Kribbella. The proposed speibonoxamine pathway ( Figure 3B) suggests a biosynthetic machinery distinct from that reported for desferrioxamine biosynthesis.