A Revised Modular Approach to (–)‐trans‐Δ8‐THC and Derivatives Through Late‐Stage Suzuki–Miyaura Cross‐Coupling Reactions

A revised modular approach to various synthetic (–)‐trans‐Δ8‐THC derivatives through late‐stage Suzuki–Miyaura cross‐coupling reactions is disclosed. Ten derivatives were synthesized allowing both sp2‐ and sp3‐hybridized cross‐coupling partners with minimal β‐hydride elimination. Importantly, we demonstrate that a para‐bromo‐substituted THC scaffold for Suzuki–Miyaura cross‐coupling reactions has been initially reported incorrectly in recent literature.

Scheme 1. A) Synthesis of (-)-trans-Δ 8 -THC using (-)-verbenol (2) and olivetol (1a) by Mechoulam et al.; [9] B) Synthesis of (-)-trans-Δ 9 -THC-Br using multistep synthesis by Carreira et al.; [10] C) Our revised modular synthesis of (-)-trans-Δ 8 -THC derivatives. ated a growing interest in the preparation of new (synthetic) cannabinoids (Scheme 1A). In particular, the introduction of unnatural substituents on the resorcinol building block was shown to improve selectivity of THC analogues for CB 1 or CB 2 . Despite various strategies that have been developed over the years, [1] the synthesis of THC derivatives remains a significant challenge. Therefore, a generally applicable modular approach allowing late-stage synthetic modification of cannabinoids would be very useful. As an example, an elegant method to synthesize challenging Δ 9 -THC derivatives via late-stage Suzuki-Miyaura cross-coupling reactions was recently reported by Carreira et al. (Scheme 1B). [10] Yet, the preparation of the Δ 9 -THC-Br precursor required a multistep sequence and did not provide access to the corresponding (-)-trans-Δ 8 -THC (Δ 8 -THC) derivatives. [6] Herein we report a revised one-step synthetic approach to Δ 8 -THC, Δ 8 -propyl-THC and halogenated Δ 8 -THC scaffolds, which have been used in SAR studies. [6] We also demonstrate that recent reports concerning the synthesis of para-substituted THC derivatives are incorrect, [11] and by studying the regioselectivity of various resorcinol derivatives with (-)-verbenol (2) we deliver proof of the correct assignment of the two possible regioisomers. Finally, both regioisomeric scaffolds were functionalized through late-stage Suzuki-Miyaura cross-coupling reactions with sp 2 -and sp 3 -hybridized organoboron reagents (Scheme 1C).

Results and Discussion
Inspired by the seminal work of Mechoulam et al. we investigated whether the electrophilic aromatic substitution of commercially available olivetol (1a) with (-)-verbenol (2), directly followed by cyclization to afford Δ 8 -THC could also be effected with Brønsted acids (see: Experimental Section). Reaction under the influence of TfOH in CH 2 Cl 2 at 0°C provided the thermodynamic isomer Δ 8 -THC in 33 % isolated yield as the sole product. Unlike weaker Brønsted acids, TfOH was successfully used for both Friedel-Crafts alkylation and subsequent cyclization at room temperature. We also envisioned that this transformation could be used to create a Δ 8 -THC scaffold for late-stage derivatization through Pd-catalyzed cross-coupling reactions. Thus, initially using readily available phloroglucinol (1b), Δ 8 -THChydroxy analogue 3 was prepared using TfOH in 53 % yield (Scheme 2). Selective triflation with Tf 2 O at 0°C of the least hindered para-hydroxy substituent resulted in Δ8-THC-triflate 4 in 56 % yield. Scheme 2. Synthesis of Δ 8 -THC-triflate (4) using phloroglucinol (1b) and subsequent regioselective triflation.
Unfortunately, all attempts of triflate 4 to undergo sp 2 -sp 3 Suzuki-Miyaura coupling utilizing various ligands, solvents and different organoboron reagents failed to give the desired products (see: Supporting Information I). Presumably, oxidative addition onto the electron-rich aromatic system did not occur, since in most cases triflate 4 was recovered. [12] During the preparation of this manuscript, Studer et al. reported the sp 2 -sp 2 Suzuki-Miyaura cross coupling with triflate 4 to obtain arylsubstituted THC derivatives, [13] but were unable to prepare biologically more relevant sp 3 -substituted THC derivatives [14] through direct cross-coupling reactions.
The electrophilic aromatic substitution/cyclization protocol of 5 with (-)-verbenol (2) surprisingly provided different results than recently published by Studer et al. [11a] and Dethe et al. [11b] (Scheme 3). In our hands, a mixture of regioisomers 6 and 7 was obtained, with the ortho-substituted regioisomer 6 being the main product, meaning that electrophilic aromatic substitution of 5 did not only take place on the "activated" C2-position but also on the equivalent C4-and C6-positions. [16] The Dethe and Studer groups reported formation of the para-isomer 7 as the sole product, however, the structure was initially incorrectly assigned. Our characterizations are in line with the para-bromosubstituted Δ 9 -THC derivatives by Carreira et al., [10] describing similar NMR shifts and coupling constants. The discrepancy in the assignment of the regioisomers was clarified using a variety of NMR experiments (see: Supporting Information II). Careful analysis of the 1 H-NMR spectrum showed clear proof of the difference between regioisomers 6 and 7, indicated by a 0.7 Hz difference in 4 J 3′,5′ coupling constant between the two aromatic protons and their distinguishable chemical shifts ( Figure 1). This was further confirmed by HMBC NMR analysis showing a correlation between proton H-1 and C-2′.
Since the undesired regioisomer was formed predominantly, we studied the intrinsic regioselectivity of the electrophilic aromatic substitution hoping that by changing the halide of the resorcinol system the ratio could be positively influenced. Starting from 5-chloro-and 5-iodoresorcinol (8 and 9, respectively) four halide-substituted THC analogues 16/17 and 18/19 were prepared. Despite the difference in size of the halides, no clear trend in regioselectivity was observed, since in all cases orthosubstitution was preferred over para-substitution. This preference has also been observed in literature, [16,17] and is most likely due to the deactivating effect exerted by the halide on the aromatic ring. Selective para-substitution was only observed in case of the alkyl-substituted THC regioisomers 13a and 13b. This is underlined by Baek et al., [18] who already showed in 1992 that electrophilic aromatic substitution of alkyl resorcinols preferentially takes place at the C2-position. For the halide-substituted THC analogues the highest amount of parasubstitution and total yield were obtained starting from 5-bromoresorcinol (5, Table 1, entry 2). These bromo-substituted synthons for Suzuki-Miyaura cross-coupling reactions were used to derivatize the pharmacologically relevant C3′-and C5′-positions of Δ 8 -THC. [19]  To investigate the reactivity of bromides 6 and 7, various Pd-catalyzed cross-coupling reactions were evaluated. Classical Heck, Kumada, Stille, and Negishi reactions were investigated, but all led to degradation of the THC scaffold, were low yielding and/or hard to reproduce. The Suzuki-Miyaura cross-couplings of 6 and 7 were successful and provided six different Δ 8 -THC derivatives (Scheme 4). Use of Pd(dppf )Cl 2 as the catalyst in combination with Cs 2 CO 3 , MeOH and potassium trifluorobo-rates (BF 3 K salts) [10] worked best in our hands and afforded the products 10a-c and 11a-c in yields ranging from 17 up to 78 %. NMR data of the ortho-substituted derivatives 10a-c were in agreement with those obtained in earlier studies, [13] although they were previously reported to be para-substituted (see: Supporting Information III). Notably, 10b was formed as an inseparable mixture of atropisomers (R a , S a ), but could be analyzed using advanced NMR techniques (see: Supporting Information IV). Scheme 4. The Suzuki-Miyaura cross-coupling of isomers 6 and 7 to give Δ 8 -THC derivatives using sp 2 -hybridized organotrifluoroborate substrates.
To extend this method to a modular approach, we studied conditions that would allow the synthesis of more challenging substrates involving sp 2 -sp 3 cross-coupling. It was found that Pd(OAc) 2 combined with RuPhos and NaOH facilitated coupling with sp 3 -hybridized reagents with minimal -hydride elimination. [20] The BF 3 K salts, used as substrates for cross-coupling reactions, were prepared in a straightforward manner from the corresponding boronic acids under non-etching conditions. [21] Elaborating on the essential difference of regioisomers 6 and 7, we converted 7 into naturally occurring Δ 8 -THC (13a) and Δ 8propyl-THC (13b) by successful Suzuki-Miyaura cross-coupling (Scheme 5). The spectroscopic data of 13a and 13b were in agreement with previously conducted experiments (see: Experimental Section). The versatility of this new modular route towards Δ 8 -THC was extended to the preparation of THC derivatives 12a-b.

Conclusions
In conclusion, we developed a synthetically versatile experimental procedure to synthesize Δ 8 -THC and a range of deriva-Scheme 5. The Suzuki-Miyaura cross-coupling of isomers 6 and 7 to give Δ 8 -THC (derivatives) using sp 3 -hybridized organotrifluoroborate substrates.
tives. Six unique halide-substituted THC analogues were prepared through an electrophilic aromatic substitution/cyclization protocol of three different halide resorcinols with verbenol, which are scaffolds for Suzuki-Miyaura cross-coupling reactions. Regioselectivity of the Friedel-Crafts alkylations was evaluated and shown to be primarily ortho-directing, most likely due to electronic effects. The use of bromo-substituted Δ 8 -THC in recent literature was wrongly reported to provide para-substituted products and is rectified. Our revised modular approach proved to be suitable for sp 2 -and sp 3 -hybridized substrates and led to the synthesis of ten different pharmacologically relevant Δ 8 -THC derivatives. We envision that this modular procedure can be extended to Δ 9 -THC derivatives using double bond isomerization [22] or starting from para-mentha-2,8-dien-1-ol, which is currently being studied in our laboratories.

Experimental Section
Supporting Information (see footnote on the first page of this article): copies of 1D and 2D NMR spectra and extensive NMR studies are provided in Supporting Information.

General information:
NMR spectra were recorded on a Bruker Avance III 400 MHz or a Bruker 500 MHz spectrometer and the compounds were assigned using 1 H NMR, 13 C NMR, 11 B NMR, 19 F NMR, COSY, HSQCED and HMBC spectra. Chemical shifts were reported in parts per million (ppm.) relative to reference (CDCl 3 : 1 H: 7.26 ppm. and 13 C 77.16 ppm; CD 3 OD: 1 H: 3.31 ppm. and 13 C 49.00 ppm; (CD 3 ) 2 SO: 1 H: 2.50 ppm. and 13 C 39.52 ppm.) NMR data are presented in the following way: chemical shift, multiplicity (s = singlet, bs = broad singlet, d = doublet, t = triplet, dd = doublet of doublets, ddd = doublet of doublet of doublets, dtd = doublet of triplet of doublets h = heptet, m = multiplet and/or multiple resonances) and coupling constants J in Hz. Reactions were monitored using TLC F 254 (Merck KGaA) using UV absorption detection (254 nm) and by spraying them with cerium ammonium molybdate stain (Hannesian's stain) followed by charring at ca 300°C. Mass spectra were recorded on a JEOL AccuTOF CS JMS-T100CS (ESI) mass spectrometer. Melting points (m.p.) were determined using a Büchi Melting Point B-545. Automatic flash column chromatography was executed on a Biotage Isolera Spektra One using SNAP or Silicycle cartridges (Biotage, 30-100 μm, 60Å) 4-50 g. Reactions under protective atmosphere were performed under positive Ar./N 2 flow in flame-dried flasks. Atom-numbering of the THC compounds is derived from an earlier reported NMR assignment in literature. [19] 2. General procedures General procedure I for potassium trifluoroborate salt synthesis from boronic acid (22-25): [21] Boronic acid (1 equiv.) was dissolved in acetonitrile (0.1M), KF (4 equiv.) in water (1M) was added at r.t. and the reaction was left stirring for 5 min. 2,3-Dihydroxysuccinic acid (2.05 equiv.) dissolved in THF (0.3M) (heat was required) was added dropwise to the vigorously stirred biphasic mixture and a white precipitate formed immediately. The reaction was diluted with acetonitrile and filtered. The flask and filter were rinsed with acetonitrile and the filtrate was concentrated in vacuo. The residue was dried under high vacuum affording the trifluoroborate salt as pure product (22)(23)(24)(25).