Single step syntheses of (1S)-aryl-tetrahydroisoquinolines by norcoclaurine synthases

The 1-aryl-tetrahydroisoquinoline (1-aryl-THIQ) moiety is found in many biologically active molecules. Single enantiomer chemical syntheses are challenging and although some biocatalytic routes have been reported, the substrate scope is limited to certain structural motifs. The enzyme norcoclaurine synthase (NCS), involved in plant alkaloid biosynthesis, has been shown to perform stereoselective Pictet–Spengler reactions between dopamine and several carbonyl substrates. Here, benzaldehydes are explored as substrates and found to be accepted by both wild-type and mutant constructs of NCS. In particular, the variant M97V gives a range of (1 S)-aryl-THIQs in high yields (48–99%) and e.e.s (79–95%). A co-crystallised structure of the M97V variant with an active site reaction intermediate analogue is also obtained with the ligand in a pre-cyclisation conformation, consistent with (1 S)-THIQs formation. Selected THIQs are then used with catechol O-methyltransferases with exceptional regioselectivity. This work demonstrates valuable biocatalytic approaches to a range of (1 S)-THIQs.


Supplementary Tables
Supplementary Table 1

Enzyme sequences
M97V with TEV cleavage site:

MHHHHHHSSGVDLGTENLYFQSMGIINQVSTVTKVIHHELEVAASADDIWTVYSWPGLAKHLPDLLG AFEKLEIIGDGGVGTILDVTFVPGEFPHEYKEKFILVDNEHRLKKVQMIEGGYLDLGVTYYMDTIHVVPT GKDSCVIKSSTEYHVKPEFVKIVEPLITTGPLAAMADAISKLVLEHKS
The TEV cleavage site is given in italics.

Protein crystallization and data processing
The positive difference density observed in the active site of the enzyme of the unrefined structure was significantly larger than what would be expected for a water molecule. Subsequent rounds of structure refinement led to a larger volume of difference density into which the ligand was built. It was only possible to model the ligand in one orientation to be consistent with the 'dopamine-first' mechanism.
The NCS mechanism is known to be highly dynamic, so it is likely that other minor, pseudoconformations of the ligand are also present.
To generate racemic standards of the 1-phenyl THIQs for chiral HPLC analysis, the phosphatemediated Pictet-Spengler reaction between dopamine and a variety of aldehyde derivatives (4a-k, Supplementary Figure 7) was investigated. 7 Reactions were performed using previously reported conditions. 7 Conversions were in the range (14 -86%), as shown in Figure S8. Significant precipitation was observed in the reaction samples and additional peaks were also observed during RP-HPLC analysis that did not correspond to the THIQ products isolated from enzymatic reactions. This suggests that side reactions (e.g. oxidations) occur due to the highly reactivity of the benzaldehydes. To minimise this, reactions were performed under an argon atmosphere and it is likely that some of the side-products adhered to the plastic Eppendorf tubes that the reactions were performed in.
For reactions between dopamine and linear or benzylic aldehydes, a ratio of the ortho:para isomers of product of 1:3 has been observed, so single regioisomer products were typically isolated by preparative-HPLC. For reactions between dopamine and a-methyl substituted aldehydes, only the para regioisomer was generated, likely due to steric hinderance. The conversions and regioisomer ratios of phosphate mediated reactions between dopamine and a variety of benzaldehyde analogues are given in Supplementary Figure 7. For the majority of substrates, a ratio of the two regioisomers, ortho:para of 8-10:1 was observed.

Chiral HPLC analyses
Chiral HPLC analysis (method indicated for each compound) was performed for the 1-aryl-THIQs generated by the NCS-mediated synthesis to determine the enantiomeric excess. Racemic standards (denoted by 'KPi') were prepared using a biomimetic phosphate mediated Pictet-Spengler reaction. 7 Retention times (tR) varied slightly depending on minor discrepancies in mobile phases and temperatures. Sample preparation for chiral HPLC injection: If chiral HPLC method 1 was used, samples were prepared using workup method 1 (Methods). Samples were prepared using workup method 2 (Methods), if chiral HPLC method 2 was used. An injection volume of 10 µL was used for all samples.

Enzymatic scale-ups and characterisation of resultant products
General method for enzymatic reactions: A solution of amine (final reaction concentration of 10 mM) and sodium ascorbate (final reaction concentration of 10 mM) was prepared in HEPES buffer (50 mM, pH 7.5 (except Figure 1c where pH 6 was used)). A solution of aldehyde (200 mM in MeCN) was prepared and the two solutions mixed in a 9:1 ratio. TfNCS (at approximately 10 mg mL -1 in 20 mM Tris, 50 mM NaCl, pH 7.5) was added and the reactions stirred at 37 o C. Control (no enzyme) reactions were performed using the same conditions but the TfNCS sample was substituted for enzyme buffer (20 mM Tris, 50 mM NaCl, pH 7.5).
Enzymatic reactions towards products for isolation and characterisation purposes were performed on a 10 mL scale using a 0.2 mg mL -1 final concentration of D29TfNCS-M97V (unless specified) for 3 h at 37 o C. Once the reaction was complete, workup method 3 was performed and the sample purified by preparative-HPLC (method indicated for each product). Methyltransferase reactions to give 7a and 7b were performed as described in the methods section and also prepared by workup method 3 then purified by preparative-HPLC (HPLC method indicated for each product). Isolated yields are given after purification by preparative-HPLC.

Preparation of calibration curves
Calibration curves were prepared by serial of a 2.5 mM stock of the product in 20% MeCN in water, which was isolated by preparative-HPLC purification (methods given for each compound isolated for characterisation purposes). Samples were analysed by achiral analytical HPLC method 1. Absorbance is given in mAU, measured at 280 nm.
room temperature. Solvents were removed under reduced pressure and the resulting residue azeotroped with methanol (3 x 10 mL). The product was partially purified by silica chromatography (100% dichloromethane) to give the product as a yellow oil (6.5 mg, 21%). The crude product was taken through to the next step without further purification.

N-Benzyl-2-(3,4-bis(benzyloxy)phenyl)ethan-1-amine
A solution of N- (3,4-bis(benzyloxy)phenethyl) benzamide (250 mg, 0.57 mmol) in THF (20 mL) was prepared under anhydrous conditions. Boron trifluoride diethyl etherate (36 µL, 0.29 mmol) was added and the reaction stirred under reflux for 10 min. Boron dimethyl sulfide complex (2 M in THF, 855 µL, 1.71 mmol) was added and the reaction mixture stirred for a further 3 h. The mixture was then cooled to 0 o C, 1 M HClaq added (10 mL) and stirred for 1 h. The reaction was then warmed to room temperature and stirred overnight. The pH was adjusted to 13 by addition of 2 M NaOH, the solution extracted with dichloromethane (3 x 20 mL). The organic phase was then dried with Na2SO4 and concentrated under reduced pressure to give the crude product as a grey oil which was taken through to the next step without further purification.