MIDA boronate allylation – synthesis of ibuprofen

MIDA boronates are among the most useful reagents for the Suzuki–Miyaura reaction. This chemistry typically generates new bonds between two aromatic rings, thereby restricting access to important areas of chemical space. Here we demonstrate the coupling of MIDA boronates to allylic electrophiles, including a new synthesis of the well-known COX inhibitor ibuprofen.


Introduction
N-Methyliminodiacetic acid (MIDA) boronates are one of the most useful organoboranes for the Suzuki-Miyaura crosscoupling strategy, as pioneered by Burke et al. (Fig. 1). 1 In addition to helping to suppress undesired side reactions such as protodeboronation, these BMIDA compounds have shown promise in enabling carbon-carbon bond formation through an iterative "building block" approach. This chemistry takes advantage to two key properties of MIDA boronates to enable iterative couplings: the orthogonal reactivity of these boronates under aqueous vs. anhydrous conditions, and crucially a "catchand-release" purication based on the chromatographic properties of the BMIDA functionality. By deprotecting MIDA boronates under aqueous conditions, followed by Pd-catalyzed coupling, a wide array of targets have been prepared using this methodology including natural products, pharmaceuticals, biological probes, and materials components. 2 Related work by Watson et al. has used a controlled speciation approach for iterative biaryl coupling of boronic acid pinacol esters with MIDA boronates. 3 While iterative strategies are widely used for generating biopolymers like polypeptides and polynucleic acids, the diversity of chemical bonds found in small molecules has hindered the development of an approach to preparing these compounds that is both modular and general. Iterative synthesis strategies that can address the challenging structural complexity posed by small molecules thus have the potential to transform the way synthetic chemistry is conducted. 4 Use of MIDA boronates in Pd-catalyzed coupling has enabled the synthesis of a wide array of targets. 2,4 A hallmark of these Pdcatalyzed C-C bond formations is that they are used almost universally to couple two carbons that are sp 2 hybridised, most commonly 2 aromatic rings. The relative ease with which at aromatic rings can be coupled has contributed to the proliferation of these structures in MedChem programmes, however it is now deemed partially responsible for the current crisis in compound attrition rate. 5 This drawback is also important because increased sp 3 -C content correlates to clinical success. 6 Very few examples of coupling MIDA boronates to sp 3 hybridised carbons are to be found in the literature, 7 and we chose to further explore this chemistry (Fig. 1b), including in the context of drug synthesis. Our approach builds upon the well precedented allylation of boronic acids, 8 while enabling the synthesis of new targets that could not be accessed using standard C sp 2-C sp 2 MIDA boronate coupling chemistry (Fig. 2).

Results and discussion
Our initial efforts to allylate aryl BMIDA compounds were quickly rewarded, as 2-naphthyl MIDA boronate could be efficiently coupled with allyl bromide using standard hydrolysis conditions (Fig. 3). Pleasingly, the reaction could be extended to aryl halide containing substrates functionalized at each position of a phenyl ring. Clearly these functional handles allow for further couplings. Furthermore, electron withdrawing nitro and cyano substituents were well tolerated, and a vinyl MIDA boronate could also be coupled. Use of an unprotected phenol resulted in a decreased yield of the desired product of C-C bond formation, due in large part to competing O-allylation, even when K 3 PO 4 was replaced with the milder base NaHCO 3 . 2-MIDA boronate substituted furan and N-Boc-pyrrole could also be allylated in moderate yield. Attempted allylation of other heterocycles including 3-pyridyl MIDA boronate was less successful, due in part to competing side reactions such as Nallylation. Interestingly, we were able to couple neo-pentyl BMIDA to cinnamyl chloride resulting in a new sp 3 -sp 3 C-C bond (9l). This process did not prove general, with other sp 3 -sp 3 candidates such as methyl BMIDA and cyclobutyl BMIDA having limited success (data not shown). Finally, we were able to couple a BMIDA-modied phenylalanine to cinnamyl chloride in excellent yield, demonstrating this methodology's potential use in peptide modication.
Next, we chose to examine the substrate tolerance on the allyl halide component by coupling 2-naphthyl MIDA boronate 7a with a range of commercial allyl halides (Fig. 4). 2-Methyl allyl bromide reacted successfully in comparable yield to the unfunctionalized system above. Coupling of cinnamyl chloride also proceeded well, and with excellent regiocontrol. BMIDA allylation with 1-methyallyl chloride and prenyl chloride resulted in decreased regioselectivity, though still good isolated yields.
In order to study the mechanism of the BMIDA allylation, we undertook a deuterium labelling study using dideuterioallyl bromide (Fig. 5). Coupling with 2-naphthyl MIDA boronate proceeded well, providing deuterated products 13 and 14 in 83% combined yield, but more importantly in a 1 : 1 ratio. This outcome is consistent with the intermediacy of p-allyl species   12, formed by oxidative addition to the allyl halide, and transmetallation with the boronic acid generated in situ.
Finally, we chose to exemplify the utility of the BMIDA allylation methodology by featuring it in the synthesis of the wellknown NSAID ibuprofen (Fig. 6). Our initial approach called for an enolate arylation reaction of a propionate ester with pbromo aryl BMIDA 4. 10 Several conditions proved unsuccessful, likely due to complications arising from the MIDA functional group. Recourse to a Negishi arylation 11 of a-bromo-t-butyl propionate however proved successful using QPhos as a ligand for Pd. 12 Pleasingly, BMIDA allylation of 16 proceeded well, providing the coupled product in 73% isolated yield. Completion of the synthesis of ibuprofen was realized by alkene hydrogenation and deprotection of the carboxylic acid. Attempts to use the same Pd source to effect multiple transformations in this synthesis met with limited success.

Conclusions
We have expanded the scope of MIDA boronate coupling to generate new bonds to sp 3 -hybridized carbons directly. The coupling is tolerant of a range of functional groups, including aryl halides, can be used with different allyl halide electrophiles, and is competent for sp 3 -sp 3 C-C bond formation. This transformation can serve as a useful addition to the BMIDAbased approach to generating small molecules, as shown by a new synthesis of the widely used drug ibuprofen.

Experimental section
Reactions involving air-sensitive agents and dry solvents were performed in glassware that had been dried in an oven (150 C) or ame-dried prior to use. These reactions were carried out with the exclusion of air using an argon atmosphere. NMR spectra were recorded on a Bruker DPX-400 spectrometer ( 1 H NMR at 400 MHz and 13 C NMR at 100 MHz) or a Bruker DPX-500 spectrometer ( 1 H NMR at 500 MHz and 13 C NMR at 125 MHz).
Chemical shis are reported in ppm. 1 H NMR spectra were recorded with CDCl 3 or d 6 -DMSO as the solvent using residual proton-containing solvent as internal standard, and for 13 C NMR spectra the chemical shis are reported relative to the central resonance of CDCl 3 or d 6 -DMSO. Signals in NMR spectra are described as singlet (s), doublet (d), triplet (t), multiplet (m), etc. or combination of these, which refers to the spin-spin coupling pattern observed. Spin-spin coupling constants reported are uncorrected. Two-dimensional (COSY, HSQC, HMBC, NOESY) NMR spectroscopy was used where appropriate to assist the assignment of signals in the 1 H and 13 C NMR spectra. IR spectra were obtained employing a Shimadzu FTIR-8400 instrument with a Golden Gate™ attachment that uses a type IIa diamond as a single reection element so that the IR spectrum of the compound (solid or liquid) could be detected directly (thin layer). High resolution mass spectra were recorded under FAB, ESI, EI and CI conditions by the analytical services at the University of Glasgow. Flash column chromatography was performed using forced ow of the indicated solvent system on EMD Geduran Silica Gel 60 as solid support and HPLC graded solvents as eluant. Reactions were monitored by thin layer chromatography (TLC) on Merck silica gel 60 covered aluminium sheets. TLC plates were developed under UV-light and/or with an acidic ethanolic anisaldehyde solution or a KMnO 4 -solution. All reagents were purchased from commercial suppliers and used without further purication unless otherwise stated.