Discovery and Reconstitution of the Cycloclavine Biosynthetic Pathway—Enzymatic Formation of a Cyclopropyl Group

The ergot alkaloids, a class of fungal-derived natural products with important biological activities, are derived from a common intermediate, chanoclavine-I, which is elaborated into a set of diverse structures. Herein we report the discovery of the biosynthetic pathway of cycloclavine, a complex ergot alkaloid containing a cyclopropyl moiety. We used a yeast-based expression platform along with in vitro biochemical experiments to identify the enzyme that catalyzes a rearrangement of the chanoclavine-I intermediate to form a cyclopropyl moiety. The resulting compound, cycloclavine, was produced in yeast at titers of >500 mg L−1, thus demonstrating the feasibility of the heterologous expression of these complex alkaloids.


SI-1.1 General materials and methods
Chanoclavine-I aldehyde 3, festuclavine 4 and agroclavine 5 were prepared or obtained as previously reported. 1 Acetonitrile was of LC-MS grade and was purchased from Sigma-Aldrich (St. Louis, MO, USA). All other chemicals and solvents were of analytical grade and used as purchased. All buffers and solutions were prepared in Milli-Q water.
Primers for cloning were synthesized by Integrated DNA Technologies (USA) and sequencing of the constructs was performed by Sourcebioscience (UK). Platinum Taq DNA polymerase (Life Technologies) or iProof High Fidelity DNA polymerase (BioRad) was used for PCR amplification of the genes. Restriction enzymes and T4 DNA ligase were purchased from New England Biolabs (USA). A Varian Cary 50 Bio Scanning Spectrometer was used to acquire UV-Vis spectra.

SI-1.2 DNA and protein sequence analysis
Computer-aided sequence analysis was done using Vector NTI 9.

SI-1.3 Construction and integration of gene expression cassettes for in vivo studies
Fungal genes used for in vivo expression were predicted from the genomic DNA cluster of For expression, all genes were cloned into expression vectors based on pRS313, pRS314, pRS315, and pRS316 4,5 . These vectors were provided with expression cassettes, comprising promoters and terminators amplified by PCR from yeast genomic DNA as previously described 5 . Five expression cassettes containing 1) a GPD1 promoter and a CYC1 terminator (G/C), 2) a PGK1 promoter and an ADH2 terminator (P/A), 3) a PDC1 promoter and an FBA1 terminator (P/F), 4) a TEF1 promoter and an ENO2 terminator (T/E), and 5) a TEF2 promoter and a PGI1 terminator (T/P) were constructed in the pRS vectors.
Integration vectors were constructed, based on these expression cassettes, as previously described for YORWΔ22 5,6 . Four additional constructs were made based on the integration sites YHRCΔ14, YMRWΔ15, YNRCΔ9, andYPRC τ3, described by Flagfeldt et al. 6 , and each construct was equipped with 3-5 expression cassettes.
The yeast strain used for cycloclavine 6 production was the Saccharomyces cerevisiae S288c, Mat α, NCYC3608, obtained from the National Collection of Yeast Cultures, Norwich, U.K. This basic strain has the KanMX antibiotic marker inserted into the Ura3 locus, and a non-functional HO locus 7 . This strain was subjected to 5 rounds of transformation, integrating a total of 21 gene expression cassettes (excluding markers) into the yeast genome to yield the cycloclavine producing strain (Table S1).

SI-1.4 Shake flask culture conditions
Engineered yeast strains were grown in standard SC broth with 2% glucose, minus relevant amino acids (ForMedium, Hunstanton, U.K.). Cultures for analysis were started from re-streaked, single colonies. These were grown overnight in standard SC broth, and then diluted to an optical density, at 600 nm, of 0.1 to start the main culture. Unless otherwise stated, main cultures were grown at 20°C with constant shaking at 150 rpm, 5 cm amplitude, for 72 hours in 250 mL shake flasks containing 25 mL medium.

SI-1.5 Fed-batch fermentation conditions
Production of cycloclavine 6 and festuclavine 4 in yeast appeared to be coupled to biomass production, and hence the fermentation process was a fed-batch that aimed to produce high biomass. Therefore, a conventional feeding regime was chosen and aeration and stirring regimes were aimed to avoid fermentative metabolism and minimize glucose accumulation as well as ethanol-and acetate formation.
The process was started as a batch using synthetic complete medium (SC), after which a feed was started that contained glucose, salts, vitamins, trace metals and amino acids. The The fermenter was equipped with two Rushton six-blade disc turbines. Air was used for sparging the fermenters. Temperature, pH, agitation, and aeration rate were controlled throughout the cultivation. The temperature was maintained at 20 °C. The pH was kept at 5.85 by automatic addition of 2M NH 4 OH during the starting batch phase and of 5M NH 4 OH during the feeding stage of the fermentation. The stirrer speed was set to 650 rpm and the aeration rate was kept to 1.0 vvm in order to prevent the Dissolved Oxygen (DO) dropping below 20%. The operating conditions used in the fermentation are summarized in Table S4.
For inoculation of the fermenter a 1-stage seed train was used. Seed cultures were prepared by inoculating 50 mL of medium in a 500-ml shake flask with 4 baffles (indents) to a starting OD 600 of ca. 0.4, using freshly grown colonies from an YPD plate. The seed medium consisted of SC-medium with 20 g/L glucose (Table S5). The shake flasks were placed on a shaking table with amplitude of 25 mm at 250 rpm at 20 ºC. The cells were grown into exponential phase until, after ~23 h., the OD 600 was ca. 3 (CDW ca. 0.9 g/L).
Prior to inoculation, an amount of the batch-medium in the fermenter equivalent to the amount of inoculum was removed and an aliquot of 10 mL of the seed culture was used for inoculation of the fermenter to a final volume of 0.32 L and a starting OD 600 of ca. 0.1 (CDW ca. 0.03 g/L). Fermentation was started as a batch of ca. 40 hours, with a starting volume of 0.32 L of SC-medium with 20 g/L glucose (Table S5).
After ca. 40 hours, a feed was started that contained a mixture of glucose, vitamins, tracemetals and salts, and was furthermore enriched with amino acids (Table S6). During said feeding phase, ammonium hydroxide (5M NH 4 OH) was used both as the nitrogen source and the base to control pH. Tables S7 and S8 list the compositions of the stock solutions of trace metals and vitamins, respectively.
The estimated total addition of feed was ca. 68 mL in 120 hours, according to a feed regime consisting of i) an exponential feeding phase of 30 hours, during which ca. 16 mL of feed were added, followed by ii) a second exponential feeding phase also of 30 hours during which ca. 15 mL of feed were added, and iii) a final constant feeding phase of 60 hours during which ca. 37 mL of feed were added.

SI-1.6 Cloning, overexpression, and purification of Aspergillus japonicus EasD, EasA, EasG and EasH for in vitro studies
The coding sequences (CDS) of EasD, EasA, EasG, and EasH were predicted from the genome sequence, using free on-line prediction tools, and synthesized with yeast codon optimization. The genes were cloned into plasmids pRS pRS313-316, which had been equipped with yeast promoters and terminators (see above). These vectors carry different auxotrophic markers allowing simultaneous expression in the yeast cell.
For heterologous protein expression, easA was amplified using the following primer pair: For heterologous protein expression, easH was amplified from a pRS vector (see above) and cloned into pEVE1914. This yeast expression vector provided an N-terminal FLAG-tag for protein purification, and allowed expression of the fusion protein from a Cup1 promoter.
Expression of EasH was carried out in synthetic complete medium (SC) without uracil.
Saccharomyces cerevisiae InvSci1 cells were grown to an OD 600 of 0.8 (2 days) prior to induction with copper(II) sulfate (300 µM) and grown for 16 hours at 20°C prior to harvesting.
Expression was carried out in LB media with kanamycin (50 µg/mL  Fig. S9).

SI-1.8 In vitro endpoint assays for EasA, EasG and EasH
In vitro assays used EasA, EasG and EasH purified as described above. Proteins were taken as frozen aliquots stored at -80 °C. Enzymes could be stored without detectable loss of activity for at least three months. A standard endpoint assay incubated EasA (0.5 μM,

SI-1.11 Mass spectrometry (in vitro assays)
Ultraperformance liquid chromatography was performed on a Waters Acquity UPLC system consisting of a binary pump, an online vacuum degasser, an autosampler, and a column compartment. Separation of the analytes was achieved on a Waters Acquity Ultra Performance BEH C18 column with 1.

Figure S3 A chanoclavine-I 2 producing yeast was used for ectopic expression of combinations of ergot alkaloid genes
Expression of easD and easG from Aspergillus japonicus (i) led to production of chanoclavine-I aldehyde 3 ([M + H] + = 255), as would be expected from the activity of easD. However, the forward reaction was not complete, as seen by the concommittant accumulation of chanoclavine-I 2 ([M + H] + = 257). No new products were detected, and none of the two products seem to be a substrate for easG. Expression of easD in combination with easA (ii), both from Aspergillus japonicus, resulted in production of a new compound, with an [M + H] + of 257. This compound was not purified, but is likely to be dihydro-chanoclavine-I aldehyde 7, which was also seen in the in vitro assay ( Figure S21). Expression of easD (Aspergillus japonicus) in combination with easA and easG from either Aspergillus japonicus (iii) or from Aspergillus fumigatus (iv) resulted in the production of festuclavine 4 ([M + H] + = 241).

Figure S4
The ratio between cycloclavine 6 and festuclavine 4 production depends on the copy number of easH A yeast strain expressing the entire cycloclavine pathway, comprising a single copy of easH (eash x 1), produced cycloclavine 6 and festuclavine 4 approximately in the ratio 3:2. Addition of 1, 2, or 3 extra gene copies increased this ratio in a dose dependent manner. With 3 additional copies (easH x 4) the ratio was more than 7:1 of cycloclavine 6 to festuclavine 4.  EasH requires all of the co-factors (excluding L-ascorbic acid which is only preventing from iron oxidation) to produce cycloclavine 6. This fact supports hypothesis that EasH is a Fe(II) and α -ketoglutarate dependent hydroxylase.

Figure S11 Standard curve for cycloclavine 6.
Standard curve for cycloclavine 6 concentration and peak area response observed by LC-MS. The concentrations 1.25-0.039 µM represent the full range of cycloclavine 6 product that was analyzed by mass spectrometry.

Figure S12 Time course of EasH assays
Conditions are as described in the methods for in vitro endpoint assays unless otherwise specified. Cycloclavine 6 concentration increases with time.

Figure S13 Concentration of cycloclavine 6 increases with increasing EasH concentration and with time
Conditions are as described in the methods for in vitro endpoint assays unless otherwise specified.

Figure S14 Cycloclavine 6 formation dependence on concentration of chanoclavine-I aldehyde 3 and time
Concentration of cycloclavine 6 increases with increasing chanoclavine-I aldehyde 3 concentration and with time. Conditions are as described in the methods for in vitro endpoint assays unless otherwise specified.

Figure S15
The pH optimum of cycloclavine 6 formation by EasA, EasG and EasH (A. japonicus) was determined to be pH = 7 Assay conditions are as described in the methods for in vitro endpoint assays. Buffers used were: 100 mM K 2 HPO 4 buffers at pH: 5, 6, 7 and 8 were used. Figure S17 Cycloclavine 6 and festuclavine 4 production with different ratio of enzymes Cycloclavine 6 and festuclavine 4 production with different ratio of enzymes: EasA: EasG: EasH (A. japonicus). The highest concentration of cycloclavine 6 was observed with ten-fold excess of EasH. The highest concentration of festuclavine 4 was observed when Aj_EasA was in 10-fold excess.
Figure S18 UV spectra of EasH (a) EasH which contains bound NADP + (after first purification by FLAG-tag); (b) EasH subjected to a second purification step (by gel filtration column), which loses activity in within 48 hours, appears to have lost the NADP + cofactor during the second purification step. Conditions are as described in the methods for in vitro endpoint assays unless otherwise specified. Chanoclavine-I aldehyde 3, agroclavine 5 or festuclavine 4 are clearly not the substrates for cycloclavine 6 production by Aj_EasH (alone).      Table S6) 30 [ml/L] Trace metal stock solution (see Table S7) 15 [ml/L] *Yeast Nitrogen Base