Expanding antibiotic chemical space around the nidulin pharmacophore

Mahmud T. Morshed a, Daniel Vuong b, Andrew Crombie b, Alastair E. Lacey b, Peter Karuso a, Ernest Lacey ab and Andrew M. Piggott *a
aDepartment of Molecular Sciences, Macquarie University, NSW 2109, Australia. E-mail: andrew.piggott@mq.edu.au
bMicrobial Screening Technologies Pty. Ltd, Smithfield, NSW 2164, Australia

Received 5th March 2018 , Accepted 4th April 2018

First published on 4th April 2018


Reinvestigating antibiotic scaffolds that were identified during the Golden Age of antibiotic discovery, but have long since been “forgotten”, has proven to be an effective strategy for delivering next-generation antibiotics capable of combatting multidrug-resistant superbugs. In this study, we have revisited the trichloro-substituted depsidone, nidulin, as a selective and unexploited antibiotic lead produced by the fungus Aspergillus unguis. Manipulation of halide ion concentration proved to be a powerful tool for modulating secondary metabolite production and triggering quiescent pathways in A. unguis. Supplementation of the culture media with chloride resulted in a shift in co-metabolite profile to dichlorounguinols and nornidulin at the expense of the non-chlorinated parent, unguinol. Surprisingly, only marginal enhancement of nidulin was observed, suggesting O-methylation may be rate-limiting. Similarly, supplementation of the media with bromide led to the production of the corresponding bromo-analogues, but also resulted in a novel family of depsides, the unguidepsides. Unexpectedly, depletion of chloride from the media halted the biosynthesis of the non-chlorinated parent compound, unguinol, and redirected biosynthesis to a novel family of ring-opened analogues, the unguinolic acids. Supplementation of the media with a range of unnatural salicylic acids failed to yield the corresponding nidulin analogues, suggesting the compounds may be biosynthesised by a single polyketide synthase. In total, 12 new and 11 previously reported nidulin analogues were isolated, characterised and assayed for in vitro activity against a panel of bacteria, fungi and mammalian cells, providing a comprehensive structure–activity profile for the nidulin scaffold.


Introduction

Antibiotic-resistant microorganisms have become increasingly prevalent in the community in recent years and are now one of the most significant threats to human health.1 Despite overwhelming acknowledgement of the seriousness of this global health crisis,2 the development of next-generation antibiotics that are capable of combatting these superbugs has proven to be a significant challenge, particularly for mycobacteria.3–5 Many drug companies have downsized or eliminated their antibiotic programs in favour of more profitable diseases and consequently the current antibiotic drug discovery pipeline has slowed to a trickle. Clearly, there is an urgent need to consider alternative approaches to deliver next-generation antibiotics.

Nature has long been a valuable source of new antibiotics,6,7 with the majority of FDA-approved antibiotics either derived from, or inspired by, natural products.8,9 During the Golden Age of antibiotic discovery, from 1950–1970, more than half of the currently used antibiotic classes were discovered. Having exhausted much of the molecular low-hanging fruit, researchers have explored a variety of different approaches to gain access to additional antibiotic chemical space, including isolating microbes from unusual or extreme environments,10,11 identifying and prioritising novel organisms through chemotaxonomy and genomic analyses,12,13 unlocking silent microbial secondary metabolites through media optimisation, co-cultivation, epigenetic modification and heterologous expression,14–20 and developing techniques to cultivate previously uncultivable microbes.21–23 While these techniques have been very successful in identifying and accessing large numbers of novel secondary metabolites with antibiotic activities, very few of these molecules have progressed into clinical trials and none has been approved for therapeutic use in humans or animals.

An alternative strategy that has been successfully employed to deliver next-generation antibiotics involves revisiting molecules that were discovered during the Golden Age, but were either overlooked or not further developed due to the abundance of other promising candidates available at the time. These “forgotten” antibiotics are an invaluable reservoir of untapped chemical diversity, and given the current discovery crisis, the time is right to revisit these molecules in search of scaffolds with novel modes of action. Pleuromutilin (discovered 1950s, FDA-approved in 2007 as the synthetic analogue retapamulin) and daptomycin (discovered 1980s, FDA-approved in 2003) are two notable examples of recently approved antibiotics that had been relegated to the scientific literature for decades before being reborn as new drugs through simple synthetic modification and reformulation, respectively.

Nidulin and nornidulin (Fig. 1) are trichlorinated depsidone antibiotics produced by the fungus Aspergillus unguis. The compounds were first identified in 1945,24 isolated in 1946,25 elucidated partially in 1953[thin space (1/6-em)]26,27 and fully in 1963.28 Nidulin and nornidulin were initially reported to have marked inhibitory activity against Mycobacterium tuberculosis and M. ranae. More recently, the compounds were also shown to be active against methicillin-resistant Staphylococcus aureus (MRSA), with MIC values of 4 and 2 μg mL−1 respectively.29 Despite these potent antibacterial activities and desirable hydrophobic properties for sustained bioavailability, nidulin has not prospered as an antibiotic lead. In this study, we have revisited the nidulin scaffold and explored how facile nutrient changes can trigger A. unguis to produce a dynamic flux of bioactive metabolites with shifting selectivities and potencies. From this complex co-metabolite profile, we have expanded chemical space around the nidulin pharmacophore, enabling us to develop a detailed structure–activity relationship.


image file: c8ob00545a-f1.tif
Fig. 1 Previously reported depsidones related to nidulin.

Results and discussion

Optimisation of culture media

It is well known that even small changes in culture media and cultivation conditions can have a profound influence on the in vitro secondary metabolite profiles of fungi.30–32 Therefore, we initially cultivated A. unguis on a range of agars, liquid media and grains to explore the accessible (baseline) biosynthetic potential of this fungus. The organism produced consistent secondary metabolite profiles, which were typically dominated by unguinol, folipastatin and agonodepside B. High productivity was observed on rich and readily accessible carbon and nitrogen sources, such as yeast extract sucrose (YES) agar and liquid media, with lower levels of production on poorer media such as malt extract, glycerol casein, oatmeal and Czapek-Dox. There appeared to be no significant difference in the metabolite profile when A. unguis was cultivated on agar or in liquid media. On YES agar and in liquid media at 24 °C, increasing levels of metabolite production were observed after 7 days, continuing to 14 days (Fig. 2A). While A. unguis grew luxuriantly on grains (rice, wheat and barley), the organism failed to produce detectable levels of unguinol-related metabolites, with only low levels of the cyclic heptapeptide, unguisin A,33 observed.
image file: c8ob00545a-f2.tif
Fig. 2 HPLC traces of crude extracts of A. unguis cultivated for 14 days on (A) YES liquid media; (B) YES liquid media supplemented with 0.5% NaCl; (C) YES liquid media supplemented with 0.5% KBr.

A preparative-scale cultivation of A. unguis was performed on YES agar plates (200 × 15 g) at 24 °C for 14 days. The major metabolites were isolated by organic solvent extraction, solvent partition and reversed phase HPLC to yield the non-chlorinated metabolites unguinol,34 folipastatin,35 agonodepside B,36 2,7-dichlorounguinol37 (aspergillusidone C) and nidulin as the major products. Interestingly, the production levels of an unreported co-metabolite, 7-carboxyfolipastatin (1), were sufficient on YES agar to enable its purification and structure elucidation.

HR-ESI(+)-MS analysis of 7-carboxyfolipastatin (1) revealed a protonated molecule ([M + H]+m/z 425.1593) indicative of a molecular formula C24H24O7. The 1H and 13C NMR data for 1 in DMSO-d6 (Table 1) were very similar to those for the previously reported compound folipastatin,35 with the only significant differences being the absence of the H-7 aromatic proton at δH 6.41 and the presence of an additional exchangeable proton at δH 13.29 (br s) and a carbonyl carbon at δC 169.7. These differences, along with an increase in molecular mass of 44 u, suggested 1 was a novel 7-carboxy analogue of folipastatin. Detailed examination of the 2D NMR data for 1 (Table S1) confirmed the structure of 7-carboxyfolipastatin as shown in Fig. 3.


image file: c8ob00545a-f3.tif
Fig. 3 Novel depsidones (1–5), depsides (6–9) and diphenyl ethers (10–12) isolated from Aspergillus unguis.
Table 1 1H (600 MHz) and 13C (150 MHz) NMR data in DMSO-d6 for depsidones 1–5
Pos. 7-Carboxyfolipastatin (1) 4,7-Dichlorounguinol (2) 7-Bromounguinol (3) 2-Chloro-7-bromounguinol (4) 7-Bromofolipastatin (5)
δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C
a Assignments are interchangeable within each column. b Not observed. c Assignment was supported by 1H–13C HMBC NMR data.
1 148.1 142.7 144.9 141.3 148.5
2 6.55, s 111.4 6.74, q (0.7) 115.6 6.56, dd (2.4, 0.8) 115.8 113.0 6.56, s 111.9
3 159.9c 157.8 161.9a 157.6 159.9
4 113.2 108.7 6.31, d (2.4) 104.6 6.58, s 105.1 113.2
4a 162.0 158.8 161.8a 160.5 161.2
5a 139.5 141.46a 141.0 141.0 142.5
6 136.8 134.2 135.9 135.9 135.8
7 114.6 116.6 108.0 108.2 108.5
8 160.3 149.0 149.5 149.7 149.6
9 114.0 117.1 116.9 117.0 116.8
9a 143.8 141.48a 142.2 142.1 142.3
11 163.9 161.5 162.8 161.7 162.9
11a 110.5c 112.2 111.2 119.2 109.9
1′ 133.7 129.4 130.9 130.7 131.8
2′ 4.93, qq (6.7, 1.5) 119.6 5.32, qq (6.8, 1.5) 126.8 5.25, qq (6.8, 1.4) 125.8 5.25, qq (6.8, 1.1) 126.1 5.33, qq (6.8, 1.5) 126.6
3′ 1.65, dq (6.7, 1.0) 13.9 1.74, dq (6.8, 1.1) 13.9 1.84, dq (6.8, 1.0) 13.4 1.85, dq (6.8, 1.1) 13.5 1.76, dq (6.8, 1.1) 13.7
4′ 1.81, dq (1.5, 1.0) 18.8 1.86, dq (1.5, 1.1) 17.4 1.86, dq (1.4, 1.0) 17.1 1.86, s 17.1 1.83, dq (1.5, 1.1) 17.3
1′′ 135.9 135.6
2′′ 5.34, qq (6.7, 1.6) 123.5 5.35, qq (6.8, 1.4) 123.9
3′′ 1.64, dq (6.7, 1.0) 14.1 1.65, dq (6.8, 1.1) 14.0
4′′ 1.78, dq (1.6, 1.0) 17.5 1.78, dq (1.4, 1.1) 17.4
1-Me 2.34, d (0.7) 20.8 2.33, s 20.7 2.38, s 18.2
4-Me 2.06, s 8.5 2.05, s 8.5
9-Me 2.03, s 9.0 2.18, s 10.3 2.17, s 10.6 2.17, s 10.5 2.18, s 10.4
3-OH 10.53, s 11.38, s 10.64, s 11.52, s 10.53, s
8-OH 9.43, s 9.22, s 9.28, s 9.28, s
7-CO2H 17.75, br s 169.7


Supplementation of culture media with chloride

Supplementation of culture media with metal halides has been previously employed by several groups to generate unnatural halogenated analogues of existing microbial metabolites.38–40 While moderate concentrations of chloride and bromide are generally well tolerated, the addition of fluoride or iodide generally inhibits microbial growth.41 In this study, we examined the effects of halide supplementation (NaCl and KBr) on the growth and secondary metabolite profile of A. unguis.

Cultivation of A. unguis in YES liquid medium supplemented with 0.5% NaCl dramatically altered the co-metabolite profile (Fig. 2B), with a 20-fold reduction in the non-chlorinated metabolites unguinol, folipastatin and agonodepside B, and comparable increases in the levels of the previously reported metabolites nornidulin and aspergillusidone C, as well as a novel metabolite, 4,7-dichlorounguinol (2). Similar co-metabolite patterns and abundances were observed using 0.25% and 1.0% NaCl. However, exposure to higher levels of NaCl resulted in significant salinotoxicity, characterised by mycelial whitening, failure to thrive and a complete loss of metabolite production. Interestingly, production levels of the monochloro depsidones after supplementation with NaCl were considerably lower than were observed for the dichloro metabolites. The accumulation of dichloro analogues suggests that incorporation of the first and second chloro groups in the biosynthesis of nidulin is rapid and limited by available chloride ions, while the incorporation of the third chloro group is slower and may proceed via an alternative chlorination pathway. As only a single monochloro metabolite, 2-chlorounguinol, was detectable by LCMS analysis, it is hypothesised that the pattern of chlorination of unguinol is first at C-2, then at C-7 and finally at C-4 to deliver nornidulin. The accumulation of nornidulin and low production levels of nidulin may also suggest that O-methylation of nornidulin is rate-limiting in the biosynthesis of nidulin. The absence of 2 in the non-supplemented medium suggests it may be a shunt metabolite accumulated under excessively saline conditions, although the presence of an alternative chlorination pathway cannot be discounted.

HR-ESI(+)-MS analysis of 4,7-dichlorounguinol (2) revealed a protonated molecule ([M + H]+m/z 395.0443/397.0411/399.0375 with 10[thin space (1/6-em)]:[thin space (1/6-em)]6[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C19H16Cl2O5. The 1H and 13C NMR data for 2 in DMSO-d6 (Table 1) were very similar to those for the previously reported compound unguinol,34 with the only significant difference being the absence of the resonances associated with aromatic protons H-4 and H-7. This difference, along with an increase in molecular mass of 70/72/74 u and a characteristic dichloro isotopic pattern in the mass spectrum, suggested 2 was the 4,7-dichloro analogue of unguinol. Detailed examination of the 2D NMR data for 2 (Table S2) confirmed the structure of 4,7-dichlorounguinol as shown in Fig. 3.

Supplementation of culture media with bromide

Supplementation of the A. unguis culture medium with 0.5% and 1.0% KBr also led to a major shift in co-metabolite profile (Fig. 2C), with a significant reduction in the abundance of unguinol and a corresponding increase in production of the previously reported38 metabolites 2-bromounguinol, 2,7-dibromounguinol and 4,7-dibromounguinol (aspergillusidones D–F respectively), as well as a novel monobrominated analogue, 7-bromounguinol (3), and a novel hybrid bromo/chloro analogue, 2-chloro-7-bromounguinol (4). Cultivation with KBr also resulted in similar reductions in the abundance of folipastatin and agonodepside B, with corresponding increases in the production of two novel brominated analogues, 7-bromofolipastatin (5) and 5-bromoagonodepside B (6).

HR-ESI(+)-MS analysis of 7-bromounguinol (3) revealed a protonated molecule ([M + H]+m/z 405.0331/407.0311 with 1[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C19H17BrO5. The 1H and 13C NMR data for 3 in DMSO-d6 (Table 1) were very similar to those for the previously reported compound unguinol,34 with the only significant difference being the absence of the resonance associated with aromatic proton H-7. This difference, along with an increase in molecular mass of 78/80 u and a characteristic monobromo isotopic pattern in the mass spectrum, suggested 3 was the 7-bromo analogue of unguinol. Detailed examination of the 2D NMR data for 3 (Table S3) confirmed the structure of 7-bromounguinol as shown in Fig. 3.

HR-ESI(+)-MS analysis of 2-chloro-7-bromounguinol (4) revealed a protonated molecule ([M + H]+m/z 438.9940/440.9916/442.9898 with 3[thin space (1/6-em)]:[thin space (1/6-em)]3[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C19H16BrClO5. The 1H and 13C NMR data for 4 in DMSO-d6 (Table 1) were very similar to those for the previously reported compound unguinol,34 with the only significant difference being the absence of the resonances associated with aromatic protons H-2 and H-7. This difference, along with an increase in molecular mass of 112/114/116 u and a characteristic monobromo/monochloro isotopic pattern in the mass spectrum, suggested 4 was either the 2-bromo-7-chloro or 2-chloro-7-bromo analogue of unguinol. A significant difference in the chemical shift for C-7 between 4 (δC 108.0) and 2 (7-Cl: δC 114.7) and a similar C-7 chemical shift between 4 and 3 (7-Br: δC 108.2) suggested C-7 must be brominated and hence C-2 must be chlorinated. Detailed examination of the 2D NMR data for 4 (Table S4) confirmed the structure of 2-chloro-7-bromounguinol as shown in Fig. 3.

HR-ESI(+)-MS analysis of 7-bromofolipastatin (5) revealed a protonated molecule ([M + H]+m/z 459.0802/461.0781 with 1[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C23H23BrO5. The 1H and 13C NMR data for 5 in DMSO-d6 (Table 1) were very similar to those for the previously reported compound folipastatin,35 with the only significant difference being the absence of the H-7 aromatic proton at δH 6.41. This difference, along with an increase in molecular mass of 78/80 u and a characteristic monobromo isotopic pattern in the mass spectrum, suggested 5 was the 7-bromo analogue of folipastatin. Detailed examination of the 2D NMR data for 5 (Table S5) confirmed the structure of 7-bromofolipastatin as shown in Fig. 3.

HR-ESI(+)-MS analysis of 5-bromoagonodepside B (6) revealed a protonated molecule ([M + H]+m/z 505.0852/507.0838 with 1[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C24H25BrO7. The 1H and 13C NMR data for 6 in DMSO-d6 (Table 2) were very similar to those for the previously reported compound agonodepside B,36 with the only significant difference being the absence of the H-5 aromatic proton. This difference, along with an increase in molecular mass of 78/80 u and a characteristic monobromo isotopic pattern in the mass spectrum, suggested 6 was the 5-bromo analogue of agonodepside B. Detailed examination of the 2D NMR data for 6 (Table S6) confirmed the structure of 5-bromoagonodepside B as shown in Fig. 3.

Table 2 1H (600 MHz) and 13C (150 MHz) NMR data for depsides 6–9 in DMSO-d6
Pos. 5-Bromoagonodepside B (6) Unguidepside A (7) 3-Bromounguidepside A (8) Decarboxyunguidepside A (9)
δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C
a Assignments are interchangeable.
1 113.1 108.3 106.9 107.9
2 154.5a 160.0 158.9 160.5
3 112.7 6.22, d (2.4) 100.4 96.1 6.21, d (2.4) 100.4
4 154.2a 161.1 159.1 161.1
5 103.2 6.21, d (2.4) 109.8 6.50, s 110.9 6.22, d (2.4) 110.0
6 142.5 140.3 140.3 140.5
7 166.0 166.9 167.8 167.7
1′ 112.0 112.1 112.8 6.76, d (1.7) 109.2
2′ 159.0 158.8 158.5 156.1
3′ 116.9 116.8 116.6 115.2
4′ 151.6 151.6 151.0 149.8
5′ 6.30, s 113.7 6.45, s 114.1 6.57, s 114.0 6.60, d (1.7) 109.6
6′ 145.5 145.4 145.4 141.5
1′′ 137.0 136.9 136.7 134.1
2′′ 5.30, qq (6.5, 1.4) 121.5 5.30, qq (6.7, 1.4) 121.5 5.32, qq (6.8, 1.4) 121.7 5.81, qq (6.8, 1.2) 121.5
3′′ 1.65, dq (6.5, 1.0) 13.7 1.65, dq (6.7, 1.1) 13.7 1.65, dq (6.8, 1.0) 13.7 1.73, dq (6.8, 1.0) 14.0
4′′ 1.84, dq (1.4, 1.0) 18.2 1.85, dq (1.4, 1.1) 18.2 1.85, dq (1.4, 1.0) 18.1 1.90, dq (1.2, 1.0) 15.0
1′′′ 136.0
2′′′ 5.31, qq (6.5, 1.4) 123.9
3′′′ 1.66, dq (6.5, 1.0) 13.4
4′′′ 1.87, dq (1.4, 1.0) 17.2
3-Me 2.12, s 10.0
6-Me 2.34, s 21.3 2.40, s 22.6 2.38, s 21.7
3′-Me 1.99, s 9.0 2.01, s 9.2 1.99, s 9.3 1.96, s 9.1
2-OH 9.69, s 10.34, s 11.16, s 10.48, s
4-OH 9.42, s 9.99, s 11.13, s 10.01, s
2′-OH 11.41, br s 11.39, br s 11.31, br s 9.56, s
1′-CO2H 13.62, br s 171.7 13.63, br s 171.8 13.64, br s 171.6


In addition to the brominated analogues described above, supplementation of the culture media with KBr also led to the production of two novel metabolites that were not observed on other culture media. The parent structure, unguidepside A (7), is the putative biosynthetic precursor of the previously reported depsidone aspergillusidone A (Fig. 1), while incorporation of bromide into 7 gave rise to 3-bromounguidepside A (8). Clearly, the effects of bromide addition are more complex than hitherto understood. While A. unguis produces good levels of mono and dibrominated metabolites, the profile does not directly mimic the chlorinated metabolites. Bromide appears to be a partial substrate, with the fungus unable to introduce the third bromo-substituent to give the fully brominated analogues of nornidulin and nidulin. Interestingly, bromide also has the ability to stimulate alternative biosynthetic pathways to widen the organism's biosynthetic repertoire.

HR-ESI(+)-MS analysis of unguidepside A (7) revealed a protonated molecule ([M + H]+m/z 373.1272) indicative of a molecular formula C20H20O7. The 1H and 13C NMR data for 7 in DMSO-d6 (Table 2) were very similar to those for the previously reported compound aspergillusidone A,37 with the only significant differences being the presence of additional aromatic proton at δH 6.45 (s) and an additional exchangeable proton at δH 10.34 (s). These differences, along with an increase in molecular mass of 2 u, suggested 7 was a ring-opened analogue of aspergillusidone A. Diagnostic HMBC correlations from the new aromatic proton to C-1′, C-3′, C-4′ and C-1′′ and a ROESY correlation from the new exchangeable proton to H-3 confirmed ring opening at the C-5 – oxygen bond. Detailed examination of the remaining 2D NMR data for 7 (Table S7) confirmed the structure of unguidepside A as shown in Fig. 3.

HR-ESI(−)-MS analysis of 3-bromounguidepside A (8) revealed a deprotonated molecule ([M − H]m/z 449.0240/451.0216 with 1[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C20H19BrO7. The 1H and 13C NMR data for 8 in DMSO-d6 (Table 2) were very similar to those observed for 7, with the only significant difference being the absence of the resonance associated with aromatic proton H-3. This difference, along with an increase in molecular mass of 78/80 u and a characteristic monobromo isotopic pattern in the mass spectrum, suggested 8 was the 3-bromo analogue of 7. Detailed examination of the 2D NMR data for 8 (Table S8) confirmed the structure of 3-bromounguidepside A as shown in Fig. 3.

Salt-leached media

While there are several literature accounts describing the supplementation of culture media with halides, to our knowledge there have been no accounts describing the opposite scenario of halide depletion. Our initial attempts to precipitate chloride from the culture media with 0.5% AgNO3 had negligible impact on the co-metabolite profile of A. unguis, while higher concentrations of AgNO3 proved toxic. It is possible that considerable pools of inaccessible chloride are bound within polysaccharides and proteins and these pools only become available as the fungus catabolises the complex media components.

An alternative strategy of halide depletion was developed based on leaching chloride from the culture media with steam. This was accomplished by cooking pearl barley in a pressure cooker and allowing steam to leach soluble components into the water condensate. The procedure, reminiscent of Soxlet extraction, affords the robust grains depleted of readily solubilised components. “Leached” barley offers considerable advantages for solid cultivation where the microbes have little mycelial penetration or digestive capacity to break-down the grains.

While grains prepared by absorption were satisfactory media for cultivating A. unguis, they had little or no capacity to support production of nidulin-related metabolites. Unexpectedly, cultivation on the leached barley gave high levels of unguinol-related metabolites. This “activation” in the absence of soluble factors had previously not been encountered in our in-house comparative studies of hundreds of fungi and suggested the presence of an unrecognised factor controlling fungal metabolite biosynthesis in the grains. Early in the cultivation (Day 7), the culture was dominated by the same repertoire of chlorinated metabolites that were observed using YES media, with only trace levels of the non-chlorinated metabolites unguinol, unguidepside A, folipastatin and agonodepside B. The chloro-analogues were less abundant than observed using YES media and their distributions were markedly different, with the levels of nornidulin and nidulin being proportionally higher than the dichlorounguinols. By Day 14, the levels of the chlorinated metabolites had plateaued, while non-chlorinated metabolites increased steadily until Day 21. At this time, the non-chlorinated metabolites represented over 80% of the extractable metabolites. Extraction and fractionation of the culture revealed that unguinol biosynthesis had been redirected along a new pathway to yield moderate levels of four novel metabolites, decarboxyunguidepside A (9), unguinolic acid (10), decarboxyunguinolic acid (11) and 5-chlorounguinolic acid (12). Trace levels of the corresponding decarboxy analogue of 12 were also detected by LCMS analysis, but were insufficient for isolation and formal characterisation.

HR-ESI(−)-MS analysis of decarboxyunguidepside A (9) revealed a deprotonated molecule ([M − H]m/z 327.1235) indicative of a molecular formula C19H20O5. The 1H and 13C NMR data for 9 in DMSO-d6 (Table 2) were very similar to those observed for 7, with the only significant differences being the absence of resonances associated with a carboxylic acid and the presence of an additional aromatic proton at δH 6.76, d (1.7 Hz). These differences, along with a decrease in molecular mass of 44 u, suggested 9 was the decarboxylated analogue of 7. Detailed examination of the 2D NMR data for 9 (Table S9) confirmed the structure of decarboxyunguidepside A as shown in Fig. 3.

HR-ESI(+)-MS analysis of unguinolic acid (10) revealed a protonated molecule ([M + H]+m/z 345.1324) indicative of a molecular formula C19H20O6. The 1H and 13C NMR data for 10 in DMSO-d6 (Table 3) were very similar to those for the previously reported compound unguinol,34 with the only significant differences being the presence of two additional exchangeable protons at δH 12.78 (br s) and δH 8.52 (br s), and the deshielding of the carbonyl carbon from δC 162.5 to 169.0. These differences, along with an increase in molecular mass of 18 u, suggested 10 was the novel hydrolysed (ring-opened) analogue of unguinol. Detailed examination of the 2D NMR data for 10 (Table S10) confirmed the structure of unguinolic acid as shown in Fig. 3.

Table 3 1H (500 MHz) and 13C (125 MHz) NMR data in DMSO-d6 for diphenyl ethers 10–12
Pos. Unguinolic acid (10) Decarboxyunguinolic acid (11) 5-Chlorounguinolic acid (12)
δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C δ H, mult (J in Hz) δ C
1 115.3 6.00, br m 106.9 116.8
2 157.6 159.5 154.2
3 5.76, d (2.3) 99.6 5.85, d (2.2) 99.4 6.02, s 99.6
4 159.4 157.8 154.3
5 6.22, d (2.3) 111.0 6.11, br m 108.9 113.0
6 138.3 138.7 134.2
1′ 148.6 148.3 148.0
2′ 110.8 110.5 110.6
3′ 153.3 152.3 152.9
4′ 6.20, s 106.0 6.15, s 105.7 6.20, s 105.7
5′ 136.4 135.8 135.9
6′ 132.1 131.7 131.2
1′′ 133.7 133.5 132.9
2′′ 5.48, qq (6.8, 1.3) 124.4 5.36, qq (6.8, 1.2) 123.2 5.44, qq (6.7, 1.4) 124.0
3′′ 1.56, dq (6.8, 1.0) 14.3 1.52, dq (6.8, 1.0) 13.7 1.55, dq (6.7, 1.0) 13.8
4′′ 1.75, dq (1.3, 1.0) 17.0 1.72, dq (1.2, 1.0) 16.6 1.74, dq (1.4, 1.0) 16.5
6-Me 2.22, s 20.2 2.10, s 21.2 2.27, s 17.3
2′-Me 1.95, s 9.3 1.96, s 9.0 1.96, s 8.9
4-OH 9.59, s 9.09, s 10.22, s
1′-OH 8.50, br s 8.38, s 8.41, br s
3′-OH 9.12, s 9.01, s 9.16, s
1-CO2H 12.76, br s 169.5 12.94, br s 168.1


HR-ESI(−)-MS analysis of decarboxyunguinolic acid (11) revealed a deprotonated molecule ([M − H]m/z 299.1286) indicative of a molecular formula C18H20O4. The 1H and 13C NMR data for 11 in DMSO-d6 (Table 3) were very similar to those observed for 10, with the only significant differences being the absence of resonances associated with a carboxylic acid and the presence of an additional aromatic proton at δH 6.00 (br m). These differences, along with a decrease in molecular mass of 44 u, suggested 11 was the decarboxylated analogue of 10. Detailed examination of the 2D NMR data for 11 (Table S11) confirmed the structure of decarboxyunguinolic acid as shown in Fig. 3.

HR-ESI(−)-MS analysis of 5-chlorounguinolic acid (12) revealed a deprotonated molecule ([M − H]m/z 377.0795/379.0765 with 3[thin space (1/6-em)]:[thin space (1/6-em)]1 relative abundance) indicative of a molecular formula C19H19ClO6. The 1H and 13C NMR data for 12 in DMSO-d6 (Table 3) were very similar to those observed for 10, with the only significant difference being the absence of one of the two aromatic protons. This difference, along with an increase in molecular mass of 34/36 u, suggested 12 was a monochlorinated analogue of 10. A diagnostic HMBC correlation from the remaining aromatic proton (δH 6.02, s) to C-1′′ and a ROESY correlation from the aromatic proton to 3′-OH suggested the chlorine atom must be on C-5. Detailed examination of the 2D NMR data for 12 (Table S12) confirmed the structure of 5-chlorounguinolic acid as shown in Fig. 3.

By leaching the grains, we reduced access to chloride ions, which was reflected in the reduced abundance of chloro-metabolites produced by the fungus. While the complete removal of chloride is likely not achievable in complex media, we were able to fundamentally alter the co-metabolite profile of A. unguis, leading to the production of four unprecedented metabolites. Using leached media, we identified two previously unrecognised aspects of fungal metabolite biosynthesis. Firstly, there must be a soluble factor that blocks unguinol-related biosynthesis in grains prepared by the absorption method, as leaching the grains restored unguinol biosynthesis. Secondly, in the presence of low chloride levels, A. unguis has the metabolic dexterity to produce a novel family of related metabolites. Thus, manipulating the concentration of chloride is a powerful tool for increasing metabolite diversity within a chemical class. As anticipated, addition of exogenous chloride (0.05%) to the leached media resulted in a 20-fold increase in the production of the mono- and di-chlorinated metabolites, with a 2-fold increase in nornidulin and nidulin production.

Precursor-directed biosynthesis with unnatural salicylic acids

In addition to supplementing the culture media with halides, we also investigated an alternative strategy of accessing novel nidulin-related analogues by feeding the fungus appropriately substituted unnatural salicylic acids. When YES liquid medium was supplemented with the putative biosynthetic precursor orsellinic acid (4-hydroxy-6-methylsalicylic acid), A. unguis grew well and produced the same co-metabolite profile as the non-supplemented medium (Fig. 4A), with no significant change in the levels of productivity (Fig. 4B). However, supplementation with 2-methylresorcinol as a simple analogue of the second putative biosynthetic precursor, aspergillusphenol A, resulted in the complete loss of production of all unguinol-related metabolites, and a corresponding increase in the production of two unrelated cyclic peptides, unguisins A and B33 (Fig. 4C). As 2-methylresorcinol exhibited no obvious mycotoxicity against A. unguis or other fungi, it was reasoned that the compound must be acting as an inhibitor of unguinol biosynthesis.
image file: c8ob00545a-f4.tif
Fig. 4 HPLC traces of crude extracts of A. unguis cultivated for 14 days on (A) YES liquid media; (B) YES liquid media supplemented with 1 μg mL−1 orsellinic acid; (C) YES liquid media supplemented with 1 μg mL−1 2-methylresorcinol.

Supplementation of the culture media of A. unguis with a small collection of unnatural salicylic acids, including 6-methyl-, 6-fluoro- and 6-methoxysalicylic acids, 3-methylorsellinic acid, orsellinic acid dimethyl ether and 2,4-dihydroxybenzoic acid, did not reveal any trace of the corresponding unnatural unguinol analogues. While orsellinic acid can be detected as a minor polar metabolite of A. unguis and the fungus can tolerate high levels of exogenous orsellinic acid, it is still uncertain whether the compound is a substrate for unguinol biosynthesis. Given the biosynthetic machinery of A. unguis appears to offer little latitude for incorporating unnatural subunits, it is possible that the biosynthesis of unguinol does not involve the coupling of two discrete aromatic monomers, but rather employs a single polyketide synthase (PKS) that both synthesises and joins two aromatic rings to yield a depside. This depside is oxidised by a cytochrome P450 to the corresponding depsidone, which in turn is hydrolysed by specific esterases to yield the diphenyl ether. A similar biosynthetic pathway has been proposed for depsides and depsidones from lichens.42 Nonetheless, it is convenient to conceptualise the metabolite cohort of A. unguis as arising from three biosynthetic subunits – orsellinic acid, aspergillusphenol A and aspergillusphenol A carboxylic acid (Fig. 5). These three subunits can be combined in six different ways, giving rise to eighteen theoretically possible variants. Twelve of these variants have been reported from Nature, while nine are present in the co-metabolite profile of A. unguis – depsides I–IV, depsidones I–IV, and diphenyl ethers I. The orsellinic acid “self-condensation” Type V variants are absent in A. unguis, but have been observed in fungal endophytes and symbionts, as exemplified by the depsidone corynesidone D, from the fungus Corynespora cassiicola,43 and the depside and diphenyl ethers lecanoric acid and notatic acid respectively, from lichens.44 The remaining possible combinations of building blocks are, as yet, unreported from Nature. Within this tight biosynthetic strategy, media variation or supplementation has provided access to over half of the naturally occurring scaffolds, with 15 depsidones, 4 depsides and 4 diphenyl ethers represented.


image file: c8ob00545a-f5.tif
Fig. 5 Proposed biosynthetic relationship between the five types (I–V) of depsides, depsidones and diphenyl ethers observed in Nature, which can be conceptualised as arising from three building blocks: orsellinic acid, aspergillusphenol A and aspergillusphenol A carboxylic acid.

Biological screening

The 12 new and 11 previously reported A. unguis metabolites were screened for in vitro activity against the Gram-positive bacteria Bacillus subtilis and Staphylococcus aureus, the Gram-negative bacterium Escherichia coli, the yeast Saccharomyces cerevisiae, and the mouse myeloma NS-1 cell line (Table 4). As the mode of action of nidulin and its analogues is unknown, it is prudent to map selectivity as a key aspect in the selection of appropriate scaffolds for development. Nidulin is a potent antibacterial with activity against Gram-positive bacteria exclusively. None of the analogues tested showed activity our indicative Gram-negative E. coli strain.
Table 4 In vitro biological activities of new compounds 1–12 and related previously reported compounds isolated from Aspergillus unguis
Compound Class Typea Minimum inhibitory concentration (μg mL−1)
Bs Sa Sc NS-1
a Refer to Fig. 5. b Bacillus subtilis (ATCC 6633). c Staphylococcus aureus (ATCC 25923). d Saccharomyces cerevisiae (ATCC 9763). e NS-1 (ATCC TIB-18) mouse myeloma cells. No significant activity was observed for any of the test compounds against Escherichia coli (ATCC 25922) or Candida albicans (ATCC 10231). f Decarboxylated.
7-Carboxyfolipastatin (1) Depsidone IV 20.6 41.1 >100 >100
4,7-Dichlorounguinol (2) Depsidone I 3.1 6.3 6.3 50.0
7-Bromounguinol (3) Depsidone I 3.1 6.3 6.3 50.0
2-Chloro-7-bromounguinol (4) Depsidone I 2.6 2.6 2.6 42.2
7-Bromofolipastatin (5) Depsidone III 1.6 3.1 >100 12.5
5-Bromoagonodepside B (6) Depside IV 8.3 66.4 33.2 66.4
Unguidepside A (7) Depside II 100.0 >100 >100 100.0
3-Bromounguidepside A (8) Depside II 25.5 >100 51.0 >100
Decarboxyunguidepside A (9) Depside I 3.1 12.5 100.0 25.0
Unguinolic acid (10) Diphenyl ether I >100 >100 >100 >100
Decarboxyunguinolic acid (11) Diphenyl ether If 25.0 100.0 >100 100.0
5-Chlorounguinolic acid (12) Diphenyl ether I 25.0 100.0 >100 25.0
Nidulin Depsidone I 0.8 6.3 >100 27.2
Nornidulin Depsidone I 1.6 6.3 >100 >100
Unguinol Depsidone I 3.1 12.5 >100 12.5
Folipastatin Depsidone III 0.8 1.6 >100 6.3
Emeguisin A Depsidone III 0.8 1.6 6.3 12.5
2-Chlorounguinol Depsidone I 12.5 25.0 25.0 25.0
Aspergillusidone C Depsidone I 3.1 6.3 6.3 25.0
Aspergillusidone D Depsidone I 2.9 2.9 2.9 46.7
Aspergillusidone E Depsidone I 3.1 6.3 6.3 25.0
Aspergillusidone F Depsidone I 2.9 5.8 2.9 46.6
Agonodepside B Depside IV 3.1 12.5 >100 50.0
Ampicillin Control 0.2 3.1
Clotrimazole Control 0.4
5-Fluorouracil Control 0.1


Nidulin displayed excellent selectivity for bacteria over yeast (SIY = 125; MIC S. cerevisiae/MIC B. subtilis) and mammalian cells (SIM > 125; MIC NS-1/MIC B. subtilis). The difference in potency between B. subtilis and S. aureus is noteworthy, but consistent with the ampicillin positive control. Nornidulin demonstrated comparable selectivity, albeit with slightly less potency. The non-chlorinated parent analogue, unguinol, was marginally less potent against bacteria and maintained selectivity over yeast (SIY > 125), but did not exhibit selectivity over mammalian cells (SIM = 4 and 1 for B. subtilis and S. aureus, respectively).

The role of chloro substituents in sustaining selectivity was reinforced by folipastatin, which gave comparable potency against both bacteria and maintained selectivity over yeast, but did not exhibit the high selectivity over mammalian cells demonstrated by nidulin (SIM = 6.8 and 3.9 for B. subtilis and S. aureus, respectively). The introduction of bromine into the unguinol scaffold in metabolites 3–5 and aspergillusidones D–F maintained activity against the Gram-positive bacteria and reasonable selectivity over mammalian cells, but was also associated with potent activity against yeast (SIY ∼1 to 2).

The presence of the carboxy group in depsides 7 and 8, diphenyl ethers 10 and 12 and depsidone 1 led to a significant reduction in antibacterial potency. However, this limitation was not universal, with depside 6 and agonodepside B maintaining moderate activity against B. subtilis. Despite the low potency of the unguidepsides, decarboxylation of 7 to give 9 improved antibiotic potency and restored selectivity over yeast and mammalian cells.

Access to ring-opened analogues by manipulation of chloride levels offered glimpses of the role of the benzoate and phenoxy bridges of the dioxepanone ring system. The diphenyl ethers offer little space for synthetic development as all analogues were either inactive or only weakly active. However, the depsides offer wider synthetic latitude. For example, decarboxyunguidepside A (9), although less potent than nidulin, offers comparable selectivity against both yeast and mammalian cells, suggesting a more targeted antibiotic scaffold. Ready access to more complex salicylates and substituted phenols makes 9 a useful lead for synthetic investigation. Of the seven scaffolds isolated from A. unguis, only depside I and depsidone I provide metabolites with acceptable levels of selectivity for further exploration. Indeed, the potent activity displayed by some substituents against yeast and mammalian cells suggests an active site common to eukaryotic and prokaryotic organisms that has undergone some evolutionary change to accommodate selectivity, but is nonetheless preserved. This site of action presents an ideal target for A. unguis by providing the maximum protection from attack and predation of resources for the minimum biosynthetic effort.

Conclusions

Microbes producing halogenated metabolites provide an opportunity to generate “unnatural” analogues through manipulation of halide ion concentration in the culture media. In this study, we employed a nidulin-producing strain of Aspergillus unguis as a model for halide-containing secondary metabolites. We found that supplementation of the culture media with chloride was not a simple additive process and that bromide did not simply act as a competitive substrate but also as an inhibitor of halide incorporation. Perhaps most notably, deprivation of halides by leaching salts from the culture media revealed hitherto unseen metabolic pathways in the fungus. Such manipulation of halide levels allowed us to expand chemical space around the nidulin pharmacophore with 12 novel analogues, thereby providing a fuller understanding of the structure–activity relationships associated with this potent and selective antibiotic scaffold.

While Nature may well offer fewer and fewer opportunities for antibiotic discovery, we are just beginning to understand how a microbe's metabolic repertoire and dexterity can be manipulated and exploited to our advantage. Over 70 years of bioassay-guided discovery has harvested the most active antibiotics, often rejecting the lesser finds as incidental. In this multidrug-resistant era, such “forgotten” antibiotics gain renewed value. From only 3 building blocks, A. unguis has the biosynthetic potential to generate a suite of chemical scaffolds, which can be modulated by subtle variations in media. These scaffolds exhibit differing potencies and selectivities against bacteria, fungi and higher eukaryotes and reflect the “yin and yang” of defence and attack the fungus requires to establish colonies on a diverse range of substrates and ecological niches. From these metabolites, we have identified two scaffolds – a depside and a depsidone – with sufficient selectivity for further development.

Experimental

General

1H NMR and 13C NMR spectra were recorded in 5 mm Pyrex tubes (Wilmad, USA) on either a Bruker Avance II DRX-600K 600 MHz or Bruker Avance III HD 500 MHz spectrometer. All spectra were obtained at 25 °C, processed using Bruker Topspin 3.5 software and referenced to residual solvent signals (DMSO-d6δH 2.49/δC 39.5 ppm). High resolution electrospray ionisation mass spectra (HRESIMS) were obtained on a Q Exactive Plus hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) by direct infusion. Electrospray ionisation mass spectra (ESIMS) were acquired on an Agilent 1260 UHPLC coupled to an Agilent 6130B single quadrupole mass detector.

Analytical HPLC was performed on a gradient Agilent 1260 Infinity quaternary HPLC system consisting of a G4212B diode array detector. The column was an Agilent Poroshell 120 EC-C18 (4.6 × 50 mm, 2.7 μm) eluted with a 1 mL min−1 gradient of 10–100% acetonitrile/water (0.01% TFA) over 8.33 min. Semi-preparative HPLC was performed on a gradient Agilent 1260 Infinity quaternary HPLC system coupled to a G4212B diode array detector. The column was an Agilent Zorbax SB-C18 (9.4 × 250 mm, 5 μm) eluted with a 4.18 mL min−1 gradient of 10–100% acetonitrile/water (0.01% TFA) over 40 min. Preparative HPLC was performed on a gradient Shimadzu HPLC system comprising two LC-8A preparative liquid pumps with static mixer, SPD-M10AVP diode array detector and SCL-10AVP system controller with standard Rheodyne injection port. The columns used in the purification of the metabolites were selected from either a Hypersil C18 spring column (50 × 150 mm, 5 μm; Grace Discovery), a Platinum-EPS C18 column (50 × 150 mm, 5 μm; Grace Discovery) or a Zorbax SB-C18 column (50 × 150 mm, 5 μm; Agilent) eluted isocratically with acetonitrile/water mixtures containing 0.1% TFA modifier at 60 mL min−1.

Optimisation of culture media

Optimisation of culture media for A. unguis was performed using a range of liquid, agar and grain-based media. The liquids and agars (glycerol casein, Czapek-Dox, malt extract, oatmeal, yeast extract sucrose) were prepared according to the recipes in Table S13. The grains (pearl barley, jasmine rice, basmati rice and cracked wheat) were prepared by hydration (∼30 mL water in 250 mL flask) during sterilisation at 121 °C for 40 min. All the media were inoculated with a spore suspension of A. unguis and incubated at 24 °C for 14 days. The liquid media were incubated both with shaking (100 rpm) and as static cultures. The cultures were sub-sampled (1 mL for liquid media and 1 g for agar and grain media) and extracted with MeOH (2 mL) for a minimum of 1 h on a wrist shaker, centrifuged (15[thin space (1/6-em)]700g for 5 min) and analysed by LCMS. The major metabolites were analysed by HPLC retention time, UV spectroscopy and mass spectrometry.

For the preparative-scale cultivation, YES agar (3.0 kg) was prepared according to recipe in Table S13 and sterilised by autoclaving at 121 °C for 30 min. A spore suspension of a 7-day-old Petri plate of A. unguis was used to inoculate 200 Petri plates (9 cm) each containing 15 g of sterile yeast extract sucrose medium. The flasks were incubated at 24 °C for 14 days. By this time, the culture has grown extensively and reached maximum metabolite productivity.

The pooled cultures were extracted with acetone (2 × 2.0 L) and the combined extracts were reduced in vacuo to give an aqueous residue. The residue was then partitioned against ethyl acetate (2 × 1 L) and dried in vacuo to give a crude extract (9 g). The crude extract was partitioned against hexane and 90% MeOH/H2O (2 × 400 mL). The methanolic (5.24 g) fraction was evaporated to dryness to give final crude extract. The methanolic extract was further fractionated by preparative HPLC into seven fractions. Fraction 3 (9.8–12.2 min) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18, isocratic 35% MeCN/H2O containing 0.01% TFA, 4.18 mL min−1) to yield 7-carboxyfolipastatin (1) (tR = 19.7 min; 3.0 mg). An isolation scheme is presented in Fig. S1.

Supplementation of media with chloride

YES liquid medium was prepared according to recipe in Table S13, and aliquots (50 mL) of this medium were transferred to each of 12 Erlenmeyer flasks (250 mL). A different quantity of NaCl or KBr (0.25 g, 0.50 g, 1.00 g, 1.25 g, 2.50 g or 10.0 g) was added to each flask to give a range of halide salt concentrations from 0.25% to 10.0%, and the media were sterilised by autoclaving at 121 °C for 30 min. A spore suspension of a 7-day-old Petri plate of A. unguis was used to inoculate the flasks. The flasks were incubated at 24 °C for 14 days on a rotary shaker at 100 rpm. The cultures were sub-sampled as described for the optimisation of culture media.

For preparative-scale chloride-supplemented cultivation, yeast extract sucrose liquid medium (3.5 L) was prepared according to recipe in Table S13, and NaCl (17.5 g) was added to give a final concentration of 0.5%. A spore suspension of a 7-day-old Petri plate of A. unguis was used to inoculate 70 × 250 mL Erlenmeyer flasks, each containing 50 mL of sterile yeast extract sucrose medium. The flasks were incubated at 24 °C for 14 days, by which time the culture has grown extensively and reached maximum metabolite productivity. The pooled liquids were partitioned against ethyl acetate (2 × 2 L) and the combined extracts were concentrated in vacuo to a crude residue (5.3 g). The enriched residue was fractionated by preparative HPLC (platinum C18, isocratic 50% stepwise to 80% MeCN/H2O containing 0.01% TFA, 60 mL min−1) to yield 7 fractions (Fr. 1–7). Fr. 3 (141 mg) was purified by preparative HPLC (Alltima C18, isocratic 70% MeCN/H2O containing 0.01% TFA, 20 mL min−1) to yield 4,7-dichlorounguinol (2) (tR 14.01 min, 18.7 mg) and 2,7-dichlorounguinol (tR 15.54 min, 74.1 mg). An isolation scheme is presented in Fig. S2.

Supplementation of media with bromide

For preparative-scale bromide-supplemented cultivation, yeast extract sucrose liquid medium (1.5 L) was prepared according to recipe in Table S13, and KBr (15 g) was added to give a final concentration of 1%. A spore suspension of a 7-day-old Petri plate of A. unguis was used to inoculate 30 × 250 mL Erlenmeyer flasks, each containing 50 mL of sterile yeast extract sucrose medium. The flasks were incubated at 24 °C for 14 days, by which time the culture has grown extensively and reached maximum metabolite productivity. The pooled cultures (3.6 L) were filtered, the mycelia were extracted with acetone (2 × 1 L) and the combined extracts were evaporated in vacuo to give an aqueous residue. The residue was then partitioned against ethyl acetate (2 × 2 L) and dried to give a crude extract (3.1 g).

The crude extract was further fractionated by preparative HPLC. The fractions were subsampled and analysed by C18 analytical HPLC. The fractions containing comparable metabolites were combined and evaporated to dryness, yielding 13 fractions (Fr. 1–13). A portion of Fr. 3 (109 of 123 mg) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18, isocratic 40% MeCN/H2O containing 0.01% TFA, 4.18 mL min−1) to yield 7-bromounguinol (3) (tR = 18.2 min; 7.0 mg). A portion of Fr. 4 (127 of 154 mg) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18, isocratic 35% MeCN/H2O containing 0.01% TFA, 4.18 mL min−1) to yield unguidepside A (7) (tR = 18.7 min; 10.7 mg). A portion of Fr. 6 (46 of 180 mg) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18, isocratic 20% MeCN/H2O containing 0.01% TFA, 4.18 mL min−1) to yield 3-bromounguidepside A (8) (tR = 20.8 min; 9.8 mg). Fr. 7 (113.0 of 192 mg) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18, gradient of 35–80% MeCN/H2O containing 0.01% TFA, 4.18 mL min−1) to yield 2-chloro-7-bromounguinol (4) (tR = 23.1 min; 8.1 mg). A portion of Fr. 11 (134 of 190 mg) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18, isocratic 20% MeCN/H2O containing 0.01% TFA, 4.18 mL min−1) to yield 7-bromofolipastatin (5) (tR = 24.3 min; 3.4 mg). Fr. 13 (57 mg) was purified by semi-preparative HPLC (Agilent Zorbax SB-C18) with isocratic 50% MeCN + 0.01% TFA (flow rate 4.18 mL min−1). After separation of this fraction, 5-bromoagonodepside B (6) (tR = 20.1 min; 23.4 mg) was isolated. An isolation scheme is presented in Fig. S3.

Silver nitrate treatment

YES liquid medium was prepared according to recipe in Table S13, and aliquots (50 mL) of this medium were transferred to each of 6 Erlenmeyer flasks (250 mL). A different quantity of AgNO3 (0.01 g, 0.05 g and 0.1 g) was added to each flask to give silver nitrate concentrations from 0.01% to 0.1%, and the media were sterilised by autoclaving at 121 °C for 30 min. A spore suspension of a 7-day-old Petri plate of A. unguis was used to inoculate the flasks. The flasks were incubated at 24 °C for 14 days on a rotary shaker at 100 rpm. The cultures were sub-sampled on Days 7, 14 and 21 as described for the optimisation of culture media.

Salt-leached media

Pearl barley (800 g) was washed exhaustively with distilled water (6 × 2 L), then poured into pressure cooker containing 3 L of boiling distilled water and cooked under pressure for 9 min. The leached pearl barley was then washed with distilled water (1.5 L) and drained. Two additional quantities of pearl barley (800 g) were processed in an identical manner. Sixty Erlenmeyer flasks (250 mL), each containing 80 g of leached pearl barley per flask, were autoclaved at 121 °C for 40 min. The flasks were then each inoculated with three agar squares (1 × 1 cm) of A. unguis and incubated at 24 °C for 21 days.

The pooled cultures (3.6 kg) were extracted with acetone (2 × 5 L) overnight on rotary platform (125 rpm) and the combined extracts were concentrated in vacuo to an aqueous residue (200 mL). The residue was partitioned against ethyl acetate (2 × 2.0 L) to provide a crude extract (55.5 g). The crude extract was redissolved in 10% H2O/MeOH (500 mL) then partitioned against hexane (2 × 1 L) to remove the lipids and provide an enriched extract (35.7 g).

The enriched extract was adsorbed onto silica gel, which was dry-loaded onto a silica gel column (140 g, 300 × 50 mm). The column was washed once with hexane (500 mL), then eluted with 50% hexane/CHCl3 (500 mL), 25% hexane/CHCl3 and CHCl3 (500 mL), followed by a stepwise gradient of 1, 2, 4, 8, 16, 32 and 100% MeOH/CHCl3 (500 mL each step), to yield 11 fractions (Fr. 1–11). Fr. 5 (723 mg) was purified by preparative HPLC (Hypersil C18, isocratic 60% MeCN/H2O containing 0.01% TFA, 60 mL min−1) to yield decarboxyunguidepside A (9) (tR 20.36 min, 24.3 mg). Fr. 6 (2.1 g) was purified by preparative HPLC (Hypersil C18, isocratic 60% MeCN/H2O containing 0.01% TFA, 60 mL min−1) to yield decarboxyunguinolic acid (11) (tR 11.40 min, 128.1 mg). Fr. 8 (501 mg) was purified by preparative HPLC (Hypersil C18, isocratic 50% MeCN/H2O containing 0.01% TFA, 60 mL min−1) to yield unguinolic acid (10) (tR 9.20 min, 99.2 mg) and 5-chlorounguinolic acid (12) (tR 11.40 min, 22.1 mg). An isolation scheme is presented in Fig. S4.

Precursor-directed biosynthesis with unnatural salicylic acids

YES liquid medium was prepared according to recipe in Table S13, and aliquots (50 mL) of this medium were transferred to each of eight Erlenmeyer flasks (250 mL). Stock solutions (0.5 mg mL−1 in DMSO) of orsellinic acid and seven other unnatural biosynthetic building blocks (6-methylsalicylic acid, 6-fluorosalicylic acid, 6-methoxysalicylic acid, 3-methylorsellinic acid, orsellinic acid dimethyl ether, 2,4-hydroxybenzoic acid and 2-methylresorcinol) were prepared and an aliquot (100 μL) of each solution was added to a different inoculated flask. The flasks were incubated at 24 °C for 14 days on a rotary shaker at 150 rpm. The cultures were sub-sampled on Days 7, 14 and 21 as described for the optimisation of culture media.

Characterisation of compounds

7-Carboxyfolipastatin (1). White amorphous solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 190 (3.84), 215 (4.63), 268 (4.12), 312 (3.83) nm; IR (ATR) νmax 2978, 2917, 2859, 1709, 1661, 1602, 1577, 1442, 1405, 1378, 1358, 1256, 1207, 1176, 1112 cm−1; NMR (600 MHz, DMSO-d6) see Table 1 and Table S1; HR-ESI(+)MS m/z 425.1593 [M + H]+ (calculated for C24H25O7+, 425.1595).
4,7-Dichlorounguinol (2). White solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 205 (5.06), 282 (4.25), 321 (4.28) nm; IR (ATR) νmax 3497, 3398, 2360, 1718, 1603, 1558, 1407, 1365, 1356, 1337, 1306, 1264, 1241 cm−1; NMR (600 MHz, DMSO-d6) see Table 1 and Table S2; HR-ESI(+)MS m/z 395.0443 [M + H]+ (calculated for C19H1735Cl2O5+, 395.0448).
7-Bromounguinol (3). Brown solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 201 (5.08), 219 (4.90), 262 (4.39), 323 (3.68) nm; IR (ATR) νmax 3432, 2976, 1711, 1617, 1572, 1406, 1353, 1318, 1258, 1212, 1184 cm−1; NMR (600 MHz, DMSO-d6) see Table 1 and Table S3; HR-ESI(+)MS m/z 405.0331 [M + H]+ (calculated for C19H1879BrO5+, 405.0332).
2-Chloro-7-bromounguinol (4). White amorphous solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 191 (5.09), 203 (5.07), 224 (4.88), 260 (4.32), 290 (4.09) nm; IR (ATR) νmax 2916, 1729, 1597, 1572, 1412, 1354, 1334, 1226, 1173, 1106 cm−1; NMR (600 MHz, DMSO-d6) see Table 1 and Table S4; HR-ESI(+)MS m/z 438.9940 [M + H]+ (calculated for C19H1779Br35ClO5+, 438.9942).
7-Bromofolipastatin (5). Brown amorphous solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 190 (4.81), 281 (3.83) nm; IR (ATR) νmax 2985, 1727, 1600, 1577, 1489, 1405, 1354, 1317 cm−1; NMR (600 MHz, DMSO-d6) see Table 1 and Table S5; HR-ESI(+)MS m/z 459.0802 [M + H]+ (calculated for C23H2479BrO5+, 459.0802).
5-Bromoagonodepside B (6). Brown amorphous solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 190 (4.93), 220 (4.88), 269 (4.44), 316 (4.22) nm; IR (ATR) νmax 2918, 1655, 1607, 1403, 1297, 1253, 1183 cm−1; NMR (600 MHz, DMSO-d6) see Table 2 and Table S6; HR-ESI(+)MS m/z 505.0852 [M + H]+ (calculated for C24H2679BrO7+, 505.0856).
Unguidepside A (7). Brown amorphous solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 191 (5.06), 215 (5.03), 263 (4.65), 308 (4.47) nm; IR (ATR) νmax 3362, 2979, 1643, 1577, 1486, 1450, 1385, 1305, 1249, 1174, 1099 cm−1; NMR (600 MHz, DMSO-d6) see Table 2 and Table S7; HR-ESI(+)MS m/z 373.1272 [M + H]+ (calculated for C20H21O7+, 373.1282).
3-Bromounguidepside A (8). Brown amorphous solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 190 (4.85), 217 (4.77), 271 (4.44), 310 (4.16) nm; IR (ATR) νmax 2914, 1650, 1606, 1573, 1408, 1295, 1246, 1209, 1181, 1106 cm−1; NMR (600 MHz, DMSO-d6) see Table 2 and Table S8; HR-ESI(−)MS m/z 449.0240 [M − H] (calculated for C20H1879BrO7, 449.0241).
Decarboxyunguidepside A (9). UV (MeOH) λmax (log[thin space (1/6-em)]ε) 216 (4.39), 265 (4.13), 304 (3.68) nm; IR (ATR) νmax 3393, 2970, 1714, 1654, 1622, 1583, 1449, 1410, 1311, 1253, 1200, 1173 cm−1; NMR (500 MHz, DMSO-d6) see Table 2 and Table S9; HR-ESI(−)MS m/z 327.1235 [M − H] (calculated for C19H19O5, 327.1238).
Unguinolic acid (10). Brown solid; UV (MeOH) λmax (log[thin space (1/6-em)]ε) 200 (4.92), 219 (4.82), 285 (4.04) nm; IR (ATR) νmax 3191, 2985, 2927, 2360, 2339, 1690, 1604, 1425, 1377, 1320, 1260, 1205, 1153, 1085, 1002 cm−1; NMR (500 MHz, DMSO-d6) see Table 3 and Table S10; HR-ESI(+)MS m/z 345.1324 [M + H]+ (calculated for C19H21O6+, 345.1333).
Decarboxyunguinolic acid (11). UV (MeOH) λmax (log[thin space (1/6-em)]ε) 205 (4.85), 281 (3.81) nm; IR (ATR) νmax 3361, 2916, 1593, 1494, 1422, 1376, 1314, 1210, 1152, 1126 cm−1; NMR (500 MHz, DMSO-d6) see Table 3 and Table S11; HR-ESI(−)MS m/z 299.1286 [M − H] (calculated for C18H19O4, 299.1289).
5-Chlorounguinolic acid (12). UV (MeOH) λmax (log[thin space (1/6-em)]ε) 207 (5.08), 286 (4.16) nm; IR (ATR) νmax 3308, 2985, 2113, 1697, 1590, 1424, 1378, 1338, 1211, 1159 cm−1; NMR (500 MHz, DMSO-d6) see Table 3 and Table S12; HR-ESI(−)-MS m/z 377.0795 [M − H] (calculated for C19H1835ClO6, 377.0797).

Biological screening

Test compounds were dissolved in DMSO to provide 10 mg mL−1 stock solutions (or between 1 and 10 mg mL−1 for compounds of limited quantities). An aliquot of each stock solution was transferred to the first lane of rows B to G in a 96-well microtitre plate and two-fold serially diluted across the 12 lanes of the plate to provide a 2048-fold concentration gradient. Bioassay medium was added to an aliquot of each test solution to provide a 100-fold dilution into the final bioassay, thus yielding a test range of 100 to 0.05 μg mL−1 in 1% DMSO. In each assay, row A contained no test compound (0% inhibition) and row H was left uninoculated (100% inhibition).

Cytotoxicity assay13

NS-1 (ATCC TIB-18) mouse myeloma cells purchased from ATCC were inoculated in 96-well microtitre plates (190 μL) at 50[thin space (1/6-em)]000 cells per mL in DMEM (Dulbecco's Modified Eagle Medium + 10% fetal bovine serum (FBS) + 1% penicillin/streptomycin (Life Technologies)) and incubated in 37 °C (5% CO2) incubator. At 48 h, resazurin (250 μg mL−1; 10 μL) was added to each well and the plates were incubated for a further 48 h. MIC end points were determined visually.

Antibacterial assay13

Bacillus subtilis (ATCC 6633) and Staphylococcus aureus (ATCC 25923) were used as indicative species for Gram-positive antibacterial activity and Escherichia coli (ATCC 25922) for Gram-negative antibacterial activity. A bacterial suspension (50 mL in 250 mL flask) was prepared in nutrient broth by cultivation for 24 h at 250 rpm, 28 °C. The suspension was diluted to an absorbance of 0.01 absorbance units per mL, and 10 μL aliquots were added to the wells of a 96-well microtitre plate, which contained the test compounds dispersed in nutrient agar (Amyl) with resazurin (12.5 μg mL−1). The plates were incubated at 28 °C for 48 h during which time the uninhibited control wells changed from a blue to light pink colour. MIC end points were determined visually.

Antifungal assay13

The yeasts Candida albicans (ATCC 10231) and Saccharomyces cerevisiae (ATCC 9763) were used as indicative species for antifungal activity. A yeast suspension (50 mL in 250 mL flask) was prepared in 1% malt extract broth by cultivation for 24 h at 250 rpm, 24 °C. The suspension was diluted to an absorbance of 0.005 and 0.03 absorbance units per mL for C. albicans and S. cerevisiae, respectively. Aliquots (20 μL and 30 μL) were applied to the wells of a 96-well microtitre plate, which contained the test compounds dispersed in malt extract agar containing bromocresol green (50 μg mL−1). The plates were incubated at 24 °C for 48 h during which time the uninhibited control wells changed from a blue to yellow colour. MIC end points were determined visually.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank Dr M. McKay (APAF, Macquarie University) for the acquisition of HRMS data. This research was funded, in part, by the Australian Research Council (FT130100142 to AMP) and Macquarie University (MQRTP scholarship to MTM).

Notes and references

  1. World Health Organization, Antimicrobial resistance: global report on surveillance, World Health Organization, Geneva, 2014 Search PubMed .
  2. US Centers for Disease Control and Prevention, Antibiotic resistance threats in the United States, Atlanta, 2013 Search PubMed .
  3. L. L. Silver, Future Microbiol., 2015, 10, 1711–1718 CrossRef CAS PubMed .
  4. L. L. Silver, Clin. Microbiol. Rev., 2011, 24, 71–109 CrossRef CAS PubMed .
  5. H. W. Boucher, G. H. Talbot, J. S. Bradley, J. E. Edwards, D. Gilbert, L. B. Rice, M. Scheld, B. Spellberg and J. Bartlett, Clin. Infect. Dis., 2009, 48, 1–12 CrossRef PubMed .
  6. A. M. Piggott and P. Karuso, Mar. Drugs, 2005, 3, 36–63 CrossRef CAS .
  7. A. M. Piggott and P. Karuso, Comb. Chem. High Throughput Screening, 2004, 7, 607–630 CrossRef CAS PubMed .
  8. D. J. Newman and G. M. Cragg, J. Nat. Prod., 2016, 79, 629–661 CrossRef CAS PubMed .
  9. D. J. Newman and G. M. Cragg, J. Nat. Prod., 2012, 75, 311–335 CrossRef CAS PubMed .
  10. R. K. Pettit, Mar. Biotechnol., 2011, 13, 1–11 CrossRef CAS PubMed .
  11. Z. E. Wilson and M. A. Brimble, Nat. Prod. Rep., 2009, 26, 44–71 RSC .
  12. J. I. Pitt, L. Lange, A. E. Lacey, D. Vuong, D. J. Midgley, P. Greenfield, M. I. Bradbury, E. Lacey, P. K. Busk, B. Pilgaard, Y.-H. Chooi and A. M. Piggott, PLoS One, 2017, 12, e0170254 Search PubMed .
  13. H. J. Lacey, D. Vuong, J. I. Pitt, E. Lacey and A. M. Piggott, Aust. J. Chem., 2016, 69, 152–160 CrossRef CAS .
  14. M. M. Zhang, F. T. Wong, Y. Wang, S. Luo, Y. H. Lim, E. Heng, W. L. Yeo, R. E. Cobb, B. Enghiad, E. L. Ang and H. Zhao, Nat. Chem. Biol., 2017, 13, 607–609 CrossRef CAS PubMed .
  15. B. K. Okada, M. R. Seyedsayamdost and M. R. Seyedsayamdost, FEMS Microbiol. Rev., 2017, 41, 19–33 CrossRef PubMed .
  16. J. S. Zarins-Tutt, T. T. Barberi, H. Gao, A. Mearns-Spragg, L. Zhang, D. J. Newman and R. J. M. Goss, Nat. Prod. Rep., 2016, 33, 54–72 RSC .
  17. T. Netzker, J. Fischer, J. Weber, D. J. Mattern, C. C. Konig, A. A. Brakhage, V. Valiante and V. Schroeckh, Front. Microbiol., 2015, 6, 299 Search PubMed .
  18. D. K. Derewacz, B. C. Covington, J. A. McLean and B. O. Bachmann, ACS Chem. Biol., 2015, 10, 1998–2006 CrossRef CAS PubMed .
  19. U. R. Abdelmohsen, T. Grkovic, S. Balasubramanian, M. S. Kamel, R. J. Quinn and U. Hentschel, Biotechnol. Adv., 2015, 33, 798–811 CrossRef CAS PubMed .
  20. A. A. Brakhage and V. Schroeckh, Fungal Genet. Biol., 2011, 48, 15–22 CrossRef CAS PubMed .
  21. C. Lok, Nature, 2015, 522, 270–273 CrossRef CAS PubMed .
  22. D. Nichols, N. Cahoon, E. Trakhtenberg, L. Pham, A. Mehta, A. Belanger, T. Kanigan, K. Lewis and S. Epstein, Appl. Environ. Microbiol., 2010, 76, 2445–2450 CrossRef CAS PubMed .
  23. L. L. Ling, T. Schneider, A. J. Peoples, A. L. Spoering, I. Engels, B. P. Conlon, A. Mueller, T. F. Schäberle, D. E. Hughes and S. Epstein, Nature, 2015, 517, 455–459 CrossRef CAS PubMed .
  24. J. M. Kurung, Science, 1945, 102, 11–12 CAS .
  25. G. H. Hogeboom and L. C. Craig, J. Biol. Chem., 1946, 162, 363–368 CAS .
  26. F. M. Dean, A. Robertson, J. C. Roberts and K. B. Raper, Nature, 1953, 172, 344–344 CrossRef CAS PubMed .
  27. F. Dean, J. C. Roberts and A. Robertson, J. Chem. Soc., 1954, 1432–1439 RSC .
  28. W. F. Beach and J. H. Richards, J. Org. Chem., 1963, 28, 2746–2751 CrossRef CAS .
  29. Y. Zhang, J. Mu, Y. Feng, L. Wen and J. Han, Nat. Prod. Res., 2014, 28, 503–506 CrossRef CAS PubMed .
  30. J. Silber, A. Kramer, A. Labes and D. Tasdemir, Mar. Drugs, 2016, 14, 137 CrossRef PubMed .
  31. K. M. VanderMolen, H. A. Raja, T. El-Elimat and N. H. Oberlies, AMB Express, 2013, 3, 1 CrossRef CAS PubMed .
  32. T. O. Larsen, J. Smedsgaard, K. F. Nielsen, M. E. Hansen and J. C. Frisvad, Nat. Prod. Rep., 2005, 22, 672–695 RSC .
  33. J. Malmstrom, J. Nat. Prod., 1999, 62, 787–789 CrossRef CAS PubMed .
  34. N. Kawahara, S. Nakajima, Y. Satoh, M. Yamazaki and K.-I. Kawai, Chem. Pharm. Bull., 1988, 36, 1970–1975 CrossRef CAS .
  35. K. Hamano, M. Kinoshita-Okami, A. Hemmi, A. Sato, M. Hisamoto, K. Matsuda, K. Yoda, H. Haruyama, T. Hosoya and K. Tanzawa, J. Antibiot., 1992, 45, 1195–1201 CrossRef CAS PubMed .
  36. S. Cao, A. S. Lee, Y. Huang, H. Flotow, S. Ng, M. S. Butler and A. D. Buss, J. Nat. Prod., 2002, 65, 1037–1038 CrossRef CAS .
  37. S. Sureram, S. Wiyakrutta, N. Ngamrojanavanich, C. Mahidol, S. Ruchirawat and P. Kittakoop, Planta Med., 2012, 78, 582–588 CrossRef CAS PubMed .
  38. S. Sureram, C. Kesornpun, C. Mahidol, S. Ruchirawat and P. Kittakoop, RSC Adv., 2013, 3, 1781–1788 RSC .
  39. J. E. Hochlowski, M. Jackson, R. R. Rasmussen, A. M. Buko, J. J. Clement, D. N. Whittern and J. B. McAlpine, J. Antibiot., 1997, 50, 201–205 CrossRef CAS PubMed .
  40. W. Keller-Schierlein, R. Muntwyler and H. Zaehner, Experientia, 1969, 25, 786 CrossRef CAS PubMed .
  41. M. Preisitsch, S. E. Heiden, M. Beerbaum, T. H. J. Niedermeyer, M. Schneefeld, J. Herrmann, J. Kumpfmueller, A. Thuermer, I. Neidhardt, C. Wiesner, R. Daniel, R. Mueller, F.-C. Bange, P. Schmieder, T. Schweder and S. Mundt, Mar. Drugs, 2016, 14, 21 CrossRef PubMed .
  42. D. Armaleo, X. Sun and C. Culberson, Mycologia, 2011, 103, 741–754 CrossRef CAS PubMed .
  43. F. B. C. Okoye, S. Lu, C. S. Nworu, C. O. Esimone, P. Proksch, A. Chadli and A. Debbab, Tetrahedron Lett., 2013, 54, 4210–4214 CrossRef CAS .
  44. J. A. Elix, A Catalogue of Standardized Chromatographic Data and Biosynthetic Relationships for Lichen Substances, Canberra, 3rd edn, 2014, published by the author Search PubMed .

Footnote

Electronic supplementary information (ESI) available: NMR spectra, UV-vis spectra and tabulated 2D NMR data. See DOI: 10.1039/c8ob00545a

This journal is © The Royal Society of Chemistry 2018