Diels–Alder Reactions During the Biosynthesis of Sorbicillinoids

Abstract The sorbicillinoids are a class of biologically active and structurally diverse fungal polyketides arising from sorbicillin. Through co‐expression of sorA, sorB, sorC, and sorD from Trichoderma reesei QM6a, the biosynthetic pathway to epoxysorbicillinol and dimeric sorbicillinoids, which resemble Diels–Alder‐like and Michael‐addition‐like products, was reconstituted in Aspergillus oryzae NSAR1. Expression and feeding experiments demonstrated the crucial requirement of the flavin‐dependent monooxygenase SorD for the formation of dimeric sorbicillinoids, hybrid sorbicillinoids, and epoxysorbicillinol in vivo. In contrast to prior reports, SorD catalyses neither the oxidation of 2′,3′‐dihydrosorbicillin to sorbicillin nor the oxidation of sorbicillinol to oxosorbicillinol. This is the first report that both the intermolecular Diels–Alder and Michael dimerization reactions, as well as the epoxidation of sorbicillinol are catalysed in vivo by SorD.


In Vitro Experimental Design and Results 21
In Vivo Experimental Design and Results 24 applied for purification was 50 mg/mL. T. reesei: A spore solution of the wild type strain or transformants grown on PDB-agar plates or PDB-agar plates supplemented with hygromycin B (100 µg/mL), respectively, were inoculated into 100 mL ME-medium in a 500 mL flask and incubated for seven days at 28 °C with 180 rpm shaking. Cultures were homogenized using a hand blender and cells were separated by filtration. The liquid supernatant was acidified with 2 M HCl to pH 3-4 and extracted twice with ethyl acetate (2 x 100 mL). Combined organic layers were dried over MgSO4 and solvent was removed under reduced pressure. The organic residue was dissolved in methanol to a concentration of 10 mg/mL, filtered over glass wool and analysed by LCMS. For preparative LCMS cultures were grown in a scale between 1-2 L and the concentration of the organic extract applied for purification was 50 mg/mL

Construction of A. oryzae Expression Vectors
T. reesei total RNA was extracted using TRIzol reagent (Thermo Fischer Scientific) and subsequently converted to cDNA using the High-Capacity RNA-to-DNA™ kit (Thermo Fischer Scientific) according to the manufacturer's instructions. The intron free cDNA was used for amplification of sorbicillinoids biosynthetic genes by PCR (See Table S2 for primers used in this study).
The tailoring genes sorC, sorD, sdr and p450 were cloned into any of the respective multigene expression vectors pTYGSarg/ade/met [1] via yeast homologous recombination under control of any of the constitutive promoters peno, padh or pgdpA (Table S3). The procedure was as followed: S. cerevisiae was streaked out on YPAD agar and incubated at 30 °C for 3-5 days. A single colony was transferred into 10 mL YPAD medium and incubated over night at 30 °C with 200 rpm shaking. This starter culture was added to 40 mL YPAD medium in a 250 mL flask and incubated for another 4.5 h at 30 °C with 200 rpm shaking. The culture was harvested by centrifugation at 3.000 g for 5 min. The pellet was washed with 25 mL ddH2O and centrifugation was repeated. The pellet was resuspended in 1 ddH2O, transferred into a 1.5 mL reaction tube and centrifuged at 18.000 g for 15 s. The pellet was resuspended in 400 µL ddH2O solution and aliquots of 50 µL were transferred into a separate 1.5 mL reaction tube. For each sample one aliquot was centrifuged at 18.000 g for 15 s and the pellet was dissolved in the transformation mixture consisting of 240 µL PEG solution (50 % (w/v) polyethylene glycol 3350),36 µL LiOAc (1 M), 50 µL denaturated salmon testis DNA (2 mg/mL in TE buffer), 34 µL DNA master mix containing the linearized vector and desired inserts obtained by PCR in equimolar concentration. The PCR fragments contain each 30 bp overlap at both 5' and 3' with the cut sites of the vector fragments to facilitate homologous recombination. Cells were first incubated for 30 min at 30 °C, then for 40 min at 42 °C. Cells were pelleted by centrifugation at 18.000 g for 15 s and supernatant was removed.
The pellet was resuspended in 500 µL ddH2O and 250 µL was spread on selective SM-Ura plates, which were incubated for four days at 30 °C. Constructed plasmid DNA was extracted from yeast cells using a Zymoprep ™ Yeast Plasmid Miniprep II kit (Zymo research, Orange, California, USA) and transformed into E. coli ccdB Survival cells by standard heat shock method for amplification. In same manner both PKS sorA and sorB were assembled in the entry vector pEYA which was subsequently used to transfer the target genes into any of the target multigene expression vectors pTYGSarg/ade/met by Gateway™ cloning.

Construction of T. reesei Knockout Vectors
All vectors were built using yeast homologous recombination as described above. For the construction of pEY-TR-sorA the gDNA of T. reesei QM6a was used as a template for amplifying the right and left fragments for sorA , including about 1.5 kb for each fragment. The hygR cassette was amplified from the plasmid pTH-GS-eGFP. The primers (Table S2) were designed to include 30 bps overhangs to be recombined in yeast with compatible overhangs in a cut pEYA plasmid. For the construction of pEY-TR-sdr, about 1 kb were amplified from the left and right ends of the sdr gene from the gDNA of T. reesei QM6a, together with the hygR cassette fragment to be recombined by yeast homologous recombination. Similarly, about 1 kb length were amplified from the P450 gene together with a hygR cassette to build the KO vector pEYA-TR-p450.

Transformation of A. oryzae NSAR1
Spore suspension collected from a fresh A. oryzae NSAR1 DPY plate (approximately 5 days) was used to inoculate 50 mL (250 mL flask) of GN liquid culture and incubated for 16 h (28 °C, 110 rpm). Cells were collected by filtration over sterile miracloth, washed with 0.8M NaCl and suspended in 10 mL of filter-sterilised A. oryzae NSAR1 protoplasting solution (10 mg/mL lysing enzyme from Trichoderma harzianum, Sigma-Aldrich, 0.8M NaCl, 10 mM CaCl2). The suspension was incubated for 4 h at ambient temperature with gentle shaking. Protoplasts were released by pipetting, collected by centrifugation (3000 × g, 5 min) and directly suspended in the required amount of fungal transformation solution I (10mM CaCl2, 0.8M NaCl and 50mM Tris-HCl at pH 7.5). Vector DNA (≥1 μg in 10 μL of ddH2O) was mixed with 100 μL protoplasts and incubated on ice for 5 min. One millilitre of fungal transformation solution II (10mM CaCl2, 0.8M NaCl and 50 mM Tris-HCl at pH 7.5, 60% (w/v) PEG3350) was added and the mixture was incubated at ambient temperature for 20 min. Five millilitres of molten selective soft agar (CZD/S, CZD/S1 or CZD/S1 w/o methionine) was added and the mixture was poured over selective agar plates (CZD/S, CZD/S1 or CZD/S1 w/o methionine). Plates were incubated at 28 °C until colonies appeared, which were transferred to secondary plates of the respective selective agar. Vigorously growing colonies were transferred onto a third plate selective plate. For strains constructed in this study see Table S4.

Transformation of T. reesei QM6a
Spore suspension collected from a fresh T. reesei PD plate was used to inoculate 50 mL (250 mL flask) of GN liquid culture and incubated for 16 h (28 °C, 110 rpm). Cells were collected by filtration over sterile miracloth, washed with washing solution (1.2 M sorbitol, 10 mM Tris-HCl pH 7.5) and suspended in 10 mL of filter-sterilised T. reesei protoplasting solution (10 mg/mL lysing enzyme from Trichoderma harzianum, 5 mg/mL driselase, 1.2 M sorbitol, 100 mM potassium phosphate pH 5.6). The suspension was incubated for 2 h at 28 °C with gentle shaking. Protoplasts were released by pipetting, collected by centrifugation (3000 × g, 5 min) and directly suspended in the required amount of resuspension solution (1 M sorbitol, 10 mM Tris-HCl pH 7.5). ). Vector DNA (≥1 μg, in 10 μL of ddH2O) was mixed with 200 μL protoplasts and 2 mL of transformation solution III (50mM CaCl2, 10 mM Tris-HCl at pH 7.5, 25% (w/v) PEG6000). Samples were incubated on ice for 20 min followed by incubation at ambient temperature for 5 min. 4 mL of resuspension solution were added and aliquots of 200 µL were mixed with 20 ml molten, 50 °C warm PD agar containing 50 µg/mL hygromycin B, spread over cultivation plates and incubated at 28 °C for 3-5 days until colonies were visible. Colonies were picked from the transformation plates and selected for three rounds on PD agar plates supplemented with 100 µg/ml hygromycin B to obtain pure colonies for further analysis. Positive transformants were grown on ME medium flasks at 28 °C for 7 days at 110 rpm for sorbicillinoid production and analysis. For strains constructed in this study see Table S4.

Cloning, Expression and Purification of SorC
For expression of sorC in E. coli BL21 (DE3) the expression plasmid pET-28/a-sorC (encoding for an N-terminal hexa-histidine tag) was built by restriction digest with NdeI and NotI, followed by ligation using T4 ligase. T. reesei cDNA was used as the DNA template ( Figure S13). Transformation of competent cells was performed based on a standard heat shock protocol. A pre-culture was grown overnight in LB-medium containing50 µg/ mL kanamycin at 37 °C with 200 rpm shaking. Each 1 mL of this seed culture was used to inoculate 100 mL 2TY-medium containing 50 µg/ mL kanamycin. Cells were grown at 37 °C and 200 rpm until an OD600 between 0.4-0.6 was reached. To induce protein expression Isopropyl-β-D-thiogalactopyranoside (IPTG) was added to a final concentration of 0.1 mM and cells were incubated for another 16 h at 16 °C and 200 rpm. Cells were harvested by centrifugation (3500 × g, 20 min) at 4 °C and resuspended in loading buffer (50 mM Tris-HCL pH 8.0, 150 mM NaCl, 10 mM imidazole, 10% glycerol (v/v)) and lysed by sonication. Cell debris was removed from the total lysate by centrifugation (10.000 × g, 40 min, 4 °C).
SorC containing a his6-tag (52.8 kDa) was purified by Ni 2+ affinity chromatography using a gravity column. 1 mL of Ni-NTA Agarose (Biozym) was washed with loading buffer and added to each 15 mL supernatant obtained in the previous step. The mixture was incubated for 1 h at 8 °C on an agitator with gentle shaking to facilitate protein binding. A 15 mL CHROMABOND® column (Macherey-Nagel) was activated with 20 % ethanol and equilibrated with loading buffer. Ni-NTA with bound protein was collected by gravity and washed twice with each 15 mL loading buffer, followed by another wash step with loading buffer (supplemented with 50 mM imidazole). Protein was eluted in 2 mL fractions with elution buffer (loading buffer + 500 mM imidazole). The buffer was exchanged to 50 mM phosphate buffer pH 8 by ultrafiltration with a molecular weight cut-off of 30 kDa and protein concentration was determined based on extinction coefficient and molecular weight using a spectrophotometer. Purified protein was assessed by SDS-PAGE ( Figure S14).

Cloning, Expression and Purification of SorD
First, the synthetic vector pET100D/TOPO-sorD was purchased from Thermo Fischer encoding for a codon optimized ( Figure S15), intron-free sorD (65.0 kDa with N-terminal his6-tag) for expression in E. coli ( Figure S16A). Protein expression was performed as described above, varying the expression temperature between 12-37°C and the IPTG concentration between 0.1-1 mM. No overexpressed protein was observed ( Figure S16B). The expression vector plasmid pET-28/a-sorD (encoding for an N-terminal hexa-histidine tag) was constructed as described above, using NheI (5') and NotI (3'). Expression temperatures between 12-37°C and IPTG concentrations between 0.1-1 mM were tested, but all resulted in insoluble protein. Autoinduction using TB-medium was also not successful for both plasmids. For a new construct the first 27 amino acids (possible signal sequence) of sorD were removed, still resulting in insoluble protein. Usage of E. coli Codon Plus cells also met with failure.
For expression of sorD in Saccharomyces cerevisiae W303B the expression plasmid pESC-ura-sorD was built by restriction digest/ligation procedure using SalI (5') and KpnI (3'). Successful plasmid construction was confirmed by sequencing. Transformation of yeast was performed as described previously, with the 34 µL transformation mixture consisting only of the already assembled plasmid pESC-ura-sorD. Positive colonies were selected for two rounds on SM-ura agar and subsequently grown in glucose-ura medium (50 mL) for 2 days at 30 °C and 160 rpm. Cells were harvested by centrifugation, washed with ddH2O and subsequently dissolved in 1 mL ddH20. 200 µL were used to inoculate 50 mL galactose-ura medium to induce protein expression. Cells were incubated for 16 h at 30 °C and 160 rpm. After washing with ddH2O pelleted cells were dissolved in 1 mL PBS buffer (100 mM phosphate buffer pH 7.5, 8g/L NaCl) and lysed mechanically. The supernatant obtained after centrifugation was directly used for enzyme assays (assays were performed by adding 50 µL of the supernatant to the "standard" SorC assay) and was also analysed by SDS-PAGE. No putative target protein could be seen and the assays did not differ from any control.

Enzyme Assays with SorC
Enzymatic formation of 2a/2b: All analytical assays were performed on a 200 µL scale consisting of SorC (1 mg/mL) dissolved in 50 mM potassium phosphate buffer (pH 8), 10 mM NAD(P)H and 8 mM substrate 1a/1b (dissolved in acetone). Final concentration of acetone in the assay mixture was 5% (v/v), but the same results were obtained with 20 % (v/v). Assays were incubated at ambient temperature for one hour with gentle shaking. 200 mL MeCN was added and precipitated enzyme was separated by centrifugation. Supernatant was directly subjected to LCMS analysis.
Enzymatic formation 10a/10b and 15a: Assays were performed as described above, but modified slightly by adding an excess of NAD(P)H (20 mM) and raising the incubation time to 16-24 hours.
Enzymatic formation of bisorbicillinols 3a-c: All analytical assays were performed on a 200 µL scale consisting of SorC (1 mg/mL) dissolved in 50 mM potassium phosphate buffer (pH 8), 10 mM NAD(P)H and 8 mM substrate 1a/1b (dissolved in acetone). Final concentration of acetone in the assay mixture was 20 % (v/v). Assays were incubated at ambient temperature for one hour with gentle shaking. Assays were extracted with 400 µL CH2Cl2 or CHCl3 and the organic phase was removed by vacuum-centrifugation at 45 °C. Organic residue was dissolved in 100 µL MeOH and subjected to LCMS analysis. Formation of bisorbicillinols was also observed when the volume of acetone in the assays mixture was kept at 5% (v/v), indicating that the organic solvent inducing the dimerization is CH2Cl2/ CHCl3.
Enzymatic formation of spirosorbicillinols 6a/6b: Assays were performed as described for the formation of bisorbicillinols 3a-c, but adding an excess of scytolide 12 (20 mM) to the assay mixture.

Feeding Experiments
Cultures of the respective fungal transformants or the wt strain were inoculated into liquid cultures (50 mL DPY medium in 250 mL flask) and grown for two days at 28 °C with 110 rpm shaking. 2 mg of the respective compound (1a or 12) dissolved in DMSO were added and cultures were grown for another 16 h. Chemical extraction was performed as mentioned above.

Synthesis of 15a*
6 mg of 15a were dissolved in 2 mL MeOH and a slight excess of TMS-diazomethane (2 M in diethyl ether) were added dropwise while stirring. The sample was incubated for 2 h, the solvent was evaporated and the sample was directly analysed by NMR.

Biosynthetic Gene Cluster Analysis
Cluster identification: Draft genome of T. reesei QM6a was obtained from NCBI (WGS: AAIL02) and putative gene clusters were predicted using the secondary metabolites analysis tool fungiSMASH. [2] Among the 32 predicted gene clusters, the first shared 71 % homology with the sorbicillinoid biosynthetic gene cluster of P. chrysogenum. The respective cluster sequence was used as the query sequence for gene identification/ protein prediction with FGENESH. [3] Subsequently conserved domain analysis was applied using BLASTp. In same manner the sorbicillinoid biosynthetic gene cluster was identified in the draft genome of P. chrysogenum (WGS: JMSF01).
Selection of candidate genes: Bioinformatic analysis of the sorbicillinoid biosynthetic gene cluster in T. reesei QM6a revealed three candidate genes that, based on their predicted function, could be involved in sorbicillinoid biosynthesis: A second FMO (sorD), a short-chain dehydrogenase/ reductase (sdr) and a cytochrome P450 dependent monooxygenase (p450). Analysis of T. reesei cDNA showed that sorD and p450 were expressed under sorbicillinoid producing conditions, although the expression of p450 was weak.
Although the SorD proteins from T. reesei (XP_006961562) and P. chrysogenum (XP_002567557) possess the same conserved domains, both proteins share only 18.3 % identity and 30.2 % similarity based on protein alignment using the software geneious® (standard parameters).
ARTEMIS analysis: The alignment tool tBLASTx was used to create a comparison file between the sorbicillinoid BGC of T. reesei and P. chrysogenum. This file was used for identification of homologous proteins between the two cluster using ARTEMIS. [4] Results show that the following proteins are homologous: SorA, SorB, SorC, one MFS and one TF, but not SorD.

Confirming Expression of sorD in +sorABD Transformants
In order to confirm that sorD was truly expressed in +sorABD transformants their total RNA was extracted and converted to cDNA as described previously. Expression of sorD was confirmed by PCR ( Figure S17). The expression plasmid pTYGSade-sorA was used as a positive control and the freshly isolated RNA was used as a negative control to exclude contamination with genomic DNA.

Comparison of Sorbicillinoids Produced by +sorABCD Transformants Prior to Chemical Extraction
Transformants were grown for four days in liquid DPY-medium at 28°C with 110 rpm shaking. Cultures were homogenized using a hand blender and cells were separated by filtration. Crude supernatant was directly subjected to LCMS-analysis. LCMS-analysis revealed presence of dimeric sorbicillinoids already prior to chemical extraction ( Figure S10).

In Vitro Formation of 15a and 10ab
Sorbicillinol 2a was produced enzymatically by SorC using 1a as the substrate and extracted with ethyl acetate. Formation of 2a was confirmed by LCMS prior to incubation . 2a was incubated for 20 h in 50 mM phosphate (pH 8) buffer at ambient temperature either in presence or absence of NADPH. Upon longer incubation (16-24 h) epoxysorbicillinols 10 and the reduced sorbicillinol 15a are formed from sorbicillinol 2 during in vitro assay with SorC when an excess of NADPH is present ( Figure S12). Tables   Table S1: Media used during this study.

Figures Sorbicillinoid Compounds Identified by Mass and UV Absorption Profiles
UVλmax (MeOH): 216 nm, 250 nm, 361 nm UV-absorption, mass (identical) and fragmentation pattern (very similar) are matching spirosorbicillinol B 6b whose structure is fully characterized by NMR (see section NMR).