Uncovering biosynthetic relationships between antifungal nonadrides and octadrides

Maleidrides are a class of bioactive secondary metabolites unique to filamentous fungi, which contain one or more maleic anhydrides fused to a 7-, 8- or 9- membered carbocycle (named heptadrides, octadrides and nonadrides respectively). Herein structural and biosynthetic studies on the antifungal octadride, zopfiellin, and nonadrides scytalidin, deoxyscytalidin and castaneiolide are described. A combination of genome sequencing, bioinformatic analyses, gene disruptions, biotransformations, isotopic feeding studies, NMR and X-ray crystallography revealed that they share a common biosynthetic pathway, diverging only after the nonadride deoxyscytalidin. 5-Hydroxylation of deoxyscytalidin occurs prior to ring contraction in the zopfiellin pathway of Diffractella curvata. In Scytalidium album, 6-hydroxylation – confirmed as being catalysed by the α-ketoglutarate dependent oxidoreductase ScyL2 – converts deoxyscytalidin to scytalidin, in the final step in the scytalidin pathway. Feeding scytalidin to a zopfiellin PKS knockout strain led to the production of the nonadride castaneiolide and two novel ring-open maleidrides.


NMR
Instruments used; Varian 400-MR (400MHz), Varian VNMRS500 (500MHz), Bruker 500 Cryo (500MHz) or Varian VNMRS600 Cryo (600MHz). Chemical shifts (δ) quoted in parts per million (ppm) and coupling constants (J) in Hertz (Hz), rounded to 0.5 Hz intervals. Two-dimensional NMR techniques (HSQC, COSY, HMBC) were used routinely for the assignment of structures. Use of NOESY and TOCSY techniques is indicated as appropriate, and the identified correlations tabulated (Tables S5-10) Residual solvent peaks were used as the internal reference for proton and carbon chemical shifts.

Preparative LCMS
Compounds were purified using a Waters time-directed autopurification system compromising Waters 2767 autosampler, Waters 2545 pump system, Phenomenex Kinetex column (

Strains
Diffractella curvata CBS 591.74 was obtained from the CBS collection; Scytalidium album strains UAMH 3611 and UAMH 3620 from the University of Alberta Mold Herbarium and culture collection; Saccharomyces cerevisiae (Stratagene) strain YPH499 was used for plasmid assembly by yeast homologous recombination. Escherichia coli strain TOP10 (Invitrogen) was used as a host for all plasmids.

Growth and Fermentation Conditions
D. curvata was maintained on PDA at 25 °C. Agar plugs were used to inoculate 100 mL PDB in 500 mL non-baffled Erlenmeyer flasks and grown at 25 °C with shaking at 200 rpm. After growing for 1 week, the seed culture was homogenised and used to inoculate flasks containing fresh PDB. For metabolite production the cultures were grown for 8 days. The homogenised D. curvata seed culture was also used to inoculated static rice cultures (50 g rice and 50 ml water autoclaved in 500 mL Erlenmeyer flasks) which were grown for 21 -28 days at 25 °C.
S. album was maintained on MEA at 25 °C. Agar plugs were used to inoculate 100 mL GN in 500 mL non-baffled Erlenmeyer flasks and grown at 25 °C with shaking at 200 rpm. After growing for 1 week, the seed culture was homogenised and used to inoculate flasks containing MEB. For metabolite production the cultures were grown for 12 days. The homogenised S. album seed culture was also used to inoculate static rice-or oat-based cultures (50 g rice/oats and 50 ml of water autoclaved in 500 mL Erlenmeyer flasks) which were grown for 21 -28 days at 25 °C.

Fungal Nucleic Acid Preparation
Fungi were grown in an appropriate liquid medium, freeze dried and then ground under liquid nitrogen. Genomic DNA was prepared using the GenElute Plant Genomic DNA miniprep kit (Sigma). RNA was prepared using the RNeasy Plant Mini Kit (Qiagen) and DNA contamination was removed using the Qiagen RNase-free DNase set.

Genome Sequencing, Transcriptomics and Bioinformatics
Genome sequencing and assembly was conducted by the Cambridge University DNA Sequencing Facility. D. curvata was sequenced using Nextera shotgun sequencing on a MiSeq platform and an assembly was generated using Newbler v2.9 1 to give a genome of ca 44.79 Mb, contained on 279 scaffolds with an N 50 of 319 Kb. S. album was sequenced using a combination of Nextera mate-pair and shotgun sequencing using a MiSeq platform and assembled using Newbler v2.9 1 to give a genome of 47.6 Mb over 19 scaffolds with a N 50 of ca 8. 29 Mb. An initial screen of the genome data for D. curvata and S. album, searching for homologues to the B. fulva byssochlamic acid gene cluster, revealed putative maleidride gene clusters within both genomes. In the D. curvata genome this gene cluster is located on scaffold 92, position 41603 -78875 nt. The S. album maleidride biosynthetic gene cluster (BGC) is located on scaffold 3, position 3125884 -3167631 nt.
RNA sequencing was also conducted by the Cambridge University DNA Sequencing Facility. Paired-end Illumina sequencing was conducted after TruSeq RNAseq library preparation. The sequencing reads were mapped to the genome assembly using Tophat. 2 Figure S1: RNAseq data generated from four different culture conditions, mapped to the zopfiellin biosynthetic gene cluster. Clear coregulation of the cluster can be seen, within a relatively un-transcribed region of the genome. See Figure S2 for RPKM data.    Figure S3: An ACT comparison of the zopfiellin and scytalidin BCGs identified likely homologous genes and highlighted the similarity between the two biosynthetic gene clusters. Histone H2A* -* ScyR9 belongs to the highly conserved H2A.F/Z family of histone H2A variants, which have been shown to play a role in transcriptional regulation and genomic stability. 23 As this protein is very highly conserved, it a reasonable to conclude that scyR9 plays no direct role in the biosynthesis of scytalidin, though it may indirectly impact production via a role in chromatin remodelling and transcriptional activation.     Figure S9: A Percent Identity Matrix for various α-ketoglutarate dependant dioxygenases, demonstrating that ScyL2 shares significant homology with RbtG and RbtB of the rubratoxin pathway, whereas ZopL9 is more unique, sharing most homology, albeit relatively low, with DES, the desaturase from the gibberellin pathway. Produced using Clustal Omega.
KGFQ Figure S10: Alignment of ScyL2 and ZopL9 with various characterised alpha-ketoglutarate dependant dioxygenases. Although low overall homology is observed, all sequences contain the conserved residues thought to bind iron and α-ketoglutarate. The H-X-D/E-X n -H motif is thought to be involved in binding Fe 2+ and an R-X-S motif that has been shown to bind 2-oxoglutarate. 26

Gene Disruption Procedures
Genes were knocked out using the bipartite method. 29 This approach involves splitting the selectable marker into two overlapping fragments, each of which are fused to regions homologous to the region to be targeted. (with a ~500 bp overlap). The splitting of the selectable marker means that homologous recombination is required to reconstruct the cassette and allow selection, which in turn improves the frequency of integration via homologous recombination, and thus gene disruption. A cassette containing the HygR gene for hygromycin resistance (accession number CAA83647) with the gpdA promoter and trpC terminator of A. nidulans, was used as the resistance marker in the transformations of both S. album and D. curvata. Knockout cassettes were constructed using homologous recombination in yeast using the plasmid pE-YA. These were then used as templates to produce the two bipartite fragments. The general approach is outlined in Figure S12. Knock-out cassettes were constructed for zopPKS, scyPKS and scyL2. The primer pair HygR1-F/HygR4-R was used to amplify the hygromycin cassette which formed the centre of the knock-out construct. The left and right flanking regions were amplified from either D. curvata or S. album genomic DNA using the following primer pairs (See Table S4  To amplify the required fragments for the transformation, the forward primer for the left flanking region (ZopPKS-LHF, ScyPKS-LHF or ScyL2-LHF) was combined with the primer HygR2-R, and HygR3-F was combined with the reverse primer for the right flanking region (ZopPKS-LHR, ScyPKS-LHR or ScyL2-LHR).

Genetic Characterisation of Knockout Strains
Agar plugs of transformants were used to inoculate 100 mL PDB (D. curvata) or MEB (S. album). Transformants were cultured at 25 °C with 200 rpm shaking for 3-5 days. The mycelia were freeze dried and then gDNA was extracted. PCR was used to test the integration of the knock-out constructs using primers designed within the resistance cassette and outside the homologous flanking regions ( Figure S13). To check for genetic purity of the transformants, primer LH-check was combined with the 'purity' primer for the specific gene; e.g. ZopPKS-LHcheck / ZopPKS-purity (See Table S4 for primer sequences). Transformants that were shown to be correct by PCR were then subjected to chemical analysis.

LH-check
HygR2-R RH-check HygR3-F Figure S13: PCR used to confirm correct integration for gene disruption Primer Table   Table S4: Primers used in this study. Compound primers, with tails (in bold) were designed to allow for yeast recombination. All primers were synthesized by Sigma.

S. cerevisiae transformation
Yeast homologous recombination was used to build knock-out constructs and was carried out as previously described. 30 An overnight culture of S. cerevisiae strain YPH499 (10 mL YPAD: 1% (w/v)yeast extract, 2% (w/v) bactotryptone, 2% (w/v) glucose, 0.04% (w/v) adenine sulphate, at 28 °C shaking at 200 rpm) was added to 40 mL of YPAD and incubated for 5 hours (28 °C shaking at 200 rpm). Cells were harvested by centrifugation; 3000 x g for 5 minutes, the supernatant discarded, and cells washed with 10 mL H2O. The pellet was resuspended in 1 mL 0.1 M LiOAc and transferred to a 1.5 mL microcentrifuge tube. Cells were pelleted at 14,000 x g and resuspended in 400 L 0.

D. curvata transformation
A seed culture of D. curvata grown in PDB was homogenised and used to inoculate a 500 mL non-baffled flask containing 100 ml of PDB. After 2 days the culture was transferred into two 50 ml centrifuge tubes and centrifuged at 6500 x g for 10 min. The supernatant was discarded and then the hyphae were washed with water followed by 0.8 M NaCl. Filter sterilized protoplasting solution (20 mL of 20 mg/ mL Trichoderma lysing enzyme, 5mg/ mL driselase in 0.8 M NaCl) was added to the mycelia and gently mixed for 3 h. To release the protoplasts from the hyphae the mixture was pipetted gently and then filtered through a sterile miracloth. The protoplasts were centrifuged at 3000 x g for 5 min and then washed with solution 1 (0.8 M NaCl, 10 mM CaCl 2 , 50 mM Tris-HCl pH 7.5). The supernatant was discarded and then the protoplasts were resuspended in solution 1 (300 µl). For each transformation 100 µl was transferred into a 10 ml tube. 10 µl of DNA (bipartite fragments for gene disruption) was added to each tube, gently mixed and incubated on ice for 10 min. 1 ml of solution 2 (0.8 M NaCl, 10mM CaCl 2 , 50 mM Tris-HCl pH 7.5, 1M PEG 3000) was added to each tube, mixed gently, and incubated at RT for 20 min. Molten PDB (50 °C) containing 0.8 % agar and 1 M sorbitol was added, gently mixed and then poured onto PDA plates which contained 1 M sorbitol. The plates were incubated at 25 °C for 4 days before an overlay containing 0.8 % PDA and hygromycin B (50 µg/ mL) was added. Colonies were then transferred to fresh PDA plates containing hygromycin B. Transformants were sub-cultured onto hygromycin plates a further two times.

S. album transformation
A seed culture of S. album grown in GN was homogenised and used to inoculate a 500 mL non-baffled flask containing 100 ml of GN. After 1 day shaking the culture was transferred into two 50 ml centrifuge tubes and centrifuged at 6500 x g for 10 min. The supernatant was discarded and then the hyphae were washed with water followed by 0.8 M NaCl. Filter sterilized protoplasting solution (20 mL of 20 mg/ mL Trichoderma lysing enzyme, 5mg/ mL driselase in 0.8 M NaCl) was added to the mycelia and gently mixed for 2 h. To release the protoplasts from the hyphae the mixture was pipetted gently and then filtered through a sterile miracloth. The protoplasts were centrifuged at 3000 x g for 5 min and then washed with solution 1 (0.8 M NaCl, 10 mM CaCl 2 , 50 mM Tris-HCl pH 7.5). The supernatant was discarded and then the protoplasts were resuspended in solution 1 (300 µl). For each transformation 100 µl was transferred into a 10 ml tube. 10 µl of DNA (bipartite fragments for gene disruption) was added to each tube, gently mixed and incubated on ice for 10 min. 1 ml of solution 2 (0.8 M NaCl, 10mM CaCl 2 , 50 mM Tris-HCl pH 7.5, 1M PEG 3000) was added to each tube and incubated at RT for 20 min. Molten MEA containing 0.8 % agar and 1 M sorbitol was added, gently mixed and then poured onto MEA plates which contained 1 M sorbitol. The plates were incubated at 25 °C for 2 days before an overlay containing 0.8 % MEA and hygromycin B (50 µg/ mL) was added. Colonies were then transferred to fresh MEA plates containing hygromycin B. Transformants were sub-cultured onto hygromycin plates a further two times.

Metabolite Extraction of D. curvata
For liquid cultures, the cultures were extracted after 8 days. To extract the cultures were acidified to pH 3 using 2 M HCl and the mycelia and broth were blended. The mycelia were removed by filtering and an equal volume of EtOAc was added to the broth. The aqueous layer was washed a further two times with EtOAc. The organic extracts were combined, dried with MgSO 4 and concentrated under vacuum. For D. curvata grown on rice, after 21 -28 days rice was soaked in EtOAc (500 mL), blended and acidified using HCl. Water was added and then the organic and aqueous layers separated. The aqueous layer was washed a further two times with EtOAc. The combined organic layers were dried with MgSO 4 and then concentrated under reduced pressure. The crude organic extract was dissolved in acetonitrile and defatted with hexane.

Metabolite Extraction of S. album
For liquid cultures, the culture was extracted after 12 days by acidifying and then blending. The broth was extracted using an equal volume of EtOAc. The combined organic extracts were dried with MgSO 4 and concentrated under reduced pressure. For S. album grown on rice or oats, after 21 -28 days, the rice or oats were soaked in EtOAc (500 mL), blended and acidified to pH 3 using HCl. Water was added and the organic and aqueous layers separated. The aqueous layer was washed a further two times with EtOAc. The combined organic layers were dried with MgSO 4 and then concentrated under reduced pressure. The crude organic extract was dissolved in acetonitrile and defatted by washing with hexane.           NMR data for castaneiolide 6    NMR data for zopfiellin 15      NMR data for scytalidin 16   NMR data for deoxyscytalidin 17      NMR data for 5-hydroxy-deoxyscytalidin 19       NMR data for 20 and 21

In vitro assays with ZopL9
Protein production The coding sequence for zopL9 was ordered as a codon optimised synthetic gene ( Figure SX), cloned into the expression vector pET151 (purchased from ThermoFisher scientific). The resulting plasmid; pET151-zopL9, was then transformed into Agilent BL21 gold cells according to the manufacturers protocol. A single colony was used to inoculate LB media (4 x 800 mL flasks) supplemented with 100 µg/ mL ampicillin. When the OD 600 reached 0.5 -0.7 protein expression was induced by the addition of IPTG (100 µL, 1 M stock). The cells were incubated overnight, then collected and resuspended in column buffer A (20 mM Tris pH 7.9, 10 mM imidazole, 500 mM NaCl, 10 % glycerol). Sonication was used to disrupt the cells before centrifugation at 15,000 rpm for 30 minutes. The supernatant was loaded onto a Nickel column (GE Healthcare HiTrap 5 mL) and eluted with column buffer A. The target protein was collected by eluting with column buffer B (20 mM Tris-HCl pH 7.9, 500 mM imidazole, 500 mM NaCl, 10 % glycerol). ZopL9 was purified using size exclusion chromatography and eluted into storage buffer (50 mM Tris-HCl pH 8, 20 % glycerol). Mass spectrometry and SDS page analysis confirmed the protein had the expected mass. Labelling studies S. album UAMH 3620 13 C labelling Sodium [1,2-13 C 2 ]-acetate (100 mg in 5 mL deionised H 2 O) was added to PDB (5 x 100 mL) on days 5 and 7. After 12 days the cultures were extracted to isolate labelled scytalidin and deoxyscytalidin. The labelling patterns for scytalidin and deoxyscytalidin were the same and were in agreement with the predicted biosynthetic pathway (Figure 3, main manuscript).

Zopfiellin 13 C labelling
Sodium [1,2-13 C 2 ]-acetate (50 mg in 6 mL of deionised H 2 O) was added to PDB (3 × 100 mL) D. curvata cultures on days 4 and 5. After 14 days the culture was extracted and labelled zopfiellin 15 was purified. The samples were analysed by 13 C NMR and compared to the unlabelled zopfiellin 15 to determine the level and position of the incorporation. The carbon signals assigned to C-2 and C-6 were enhanced and appeared as singlets. Two carbonyl carbons, C-3' and C-7', were doublets (J CC 61 Hz) coupled to C-3 and C-7 respectively. One side of the molecule has four intact acetates (from C-8' to C-5''), suggesting that the original polyketide is a tetraketide. The other side of the molecule has three intact acetates (C-4'/C-4; C-1'''/C-2'''; C-3'''/C-4'''). C-5 was labelled, but not coupled, suggesting the loss of one carbon. Feeding Compounds to Diffractella curvata Strain zopPKS zopPKS was fed with scytalidin 16 or deoxyscytalidin 17 purified from S. album cultures. 5-hydroxydeoxyscytalidin 19, purified from zopPKS cultures fed with deoxyscytalidin, was also fed back to strain zopPKS.
To achieve these feeding studies, a seed culture of ΔzopPKS was homogenised and used to inoculate 500 mL non-baffled flasks containing 100 mL PDB. After 3 days, 5 mg per 100 mL flask, of compounds 16, 17 or 19 (solubilized in DMSO or MeOH at a concentration of 20 mg/ mL) were fed. The feeding was repeated on day 5. After 12-14 days the cultures were extracted with EtOAc and the products analysed using LCMS.
To isolate compounds 6, 20 and 21, a total of 150 mg of scytalidin was fed to 15 x 100 mL cultures of zopPKS.
The cultures were extracted with EtOAc and preparative LCMS using a 40 -95 % gradient yielded 10 mg of 21, 3.4 mg of 20 and 0.7 mg of 6.

Synthesis
All reagents were sourced from commercial suppliers and were used without further purification unless stated otherwise. All reactions using anhydrous solvents were performed using standard Schlenk syringe-septa techniques, with flame dried glassware under a positive pressure of nitrogen. Anhydrous DCM was dried by passing through a modified Grubbs system of alumina columns, manufactured by Anhydrous Engineering. Flash column chromatography was performed according to the procedures used by Still et al. 35

X-Ray Crystallography
X-ray diffraction experiments on 16 and 18 were carried out at 100(2) K on a Bruker APEX II diffractometer using Mo-K α radiation (λ = 0.71073 Å), while 17 was carried out at 100(2) K Bruker Microstar rotating anode diffractometer using Cu-K α (λ = 1.54178 Å). Intensities were integrated in SAINT 36 from several series of exposures measuring 0.5° in ω or φ and absorption corrections based on equivalent reflections were applied using SADABS. 37 Structures 16 and 17 were solved using ShelXS 38 , while 18 was solved using ShelXT 39 all of the structures were refined by full matrix least squares against F 2 in ShelXL 38, 40 using Olex2 41 . All of the non-hydrogen atoms were refined anisotropically. While all of the hydrogen atoms were located geometrically and refined using a riding model. In the case of 16 the molecule displayed disorder, the occupancies of the fragments was determined by refining them against a free variable with the sum of the two sites set to equal 1, the occupancies were then fixed at the refined values and restraints were used to maintain sensible geometries and thermal parameters. In addition Squeeze within Platon 42, 43 was used to remove disordered solvent from the lattice of 16 that could not be sensibly modelled. Crystal structure and refinement data are given in Table S5. Crystallographic data for compounds 16, 17

Assignment of zopfiellin stereochemistry
The substituent eclipsing the phenyl group of the MTPA unit experiences an aromatic shielding interaction resulting in an upfield chemical shift for protons affected. From comparing the chemical shifts, it is evident that the phenyl group in (S)-MTPA-zopfiellin shields the cyclooctadiene ring protons (6H 2 and 5-H) as apparent by their significant upfield shift relative to (R)-MTPA ( Figure S72). The opposite effect is observed for 2''-HH which shifts downfield in (S)-MTPA-zopfiellin. This indicates that the 2''-HH signal is more shielded in (R)-MTPAzopfiellin than it is in (S)-MTPA-zopfiellin.
These results reveal that the phenyl group in (S)-MTPA-zopfiellin interacts with the cyclooctadiene ring rather than the alkyl chain. Inputting this information into the syn-co-planar conformation for the (S)-MTPA derivative gives R 1 as the alkyl chain and R 2 as the cyclooctadiene ring. The absolute stereochemistry of the secondary alcohol has therefore been assigned as R.