1
Late-Stage Functionalization by Transition Metal-Catalyzed C-H Functionalization

Erwan Brunard and Joanna Wencel-Delord

JMU Würzburg, Institute of Organic Chemistry, Faculty of Chemistry and Pharmacy, Am Hubland, 97074, Würzburg, Germany

1.1 Introduction

The 21st century is clearly marked by important environmental issues, global warming, expanding worldwide population, depletion of natural resources, and inflating energy needs. These issues clearly impact all domains of everyday life. However, chemistry is clearly the field that can provide real solutions to these major and urgent societal problems. Accordingly, the chemical industry has initiated the sustainable revolution, and extensive efforts are devoted to produce all types of chemicals in more sustainable manner, while decreasing the environmental footprint. However, the development of modern industrial process requires initial fundamental discoveries and the development of new synthetic tools, allowing the synthesis of desired organic products in more resource-, energy-, and step-efficient manner. Accordingly, the fundamental research in organic chemistry has been undergoing, since the last two decades, a real "sustainability revolution" [1, 2]. Various approaches have been proposed to allow more sustainable synthesis, among which visible-light-mediated photocatalysis, electrochemistry, continuous-flow reactions, and "benign solvent design" are selected examples. These strategies reached now the maturity to be considered valuable and general tools for organic synthesis. In parallel, C-H activation strategy profoundly changed the mindset of a synthetic chemist [3]. Indeed, a possibility of using non-prefunctionalized substrates in the transition-metal-catalyzed process presents a formidable advantage over the standard cross-coupling reactions. Over the past years, organic chemists have learned how to use latent and strong C-H bonds as functional motifs and directly convert them into desired functional groups. Over the years, the field of C-H activation expanded tremendously, allowing major breakthroughs and providing key improvements, such as efficient site-selectivity [4], use of abundant, cheap, and nontoxic metals [5], or the design of highly reactive catalytic systems able to promote the key C-H cleavage step under mild reaction conditions. Accordingly, C-H activation field, which was 15 years ago considered a scientific curiosity, now reached a totally new level of industrial applications in both production [6] and R&D industrial sectors [3c]. The C-H activation field holds, in particular, an exceptional potential for speeding up research in pharmaceutical industry [7] by unlocking the door toward rapid diversification of lead drug candidates via C-H late-stage functionalization. Indeed, if a direct C-H functionalization of a complex molecule can be done in one step, the diversity of lead's congeners can be rapidly accessed by avoiding traditional de novo, multistep synthesis of each product.

While the first, general example of the LSF to directly install diversity of motifs on celecoxib was reported as early as in 2011 by the group of Yu [8], over the last year, late-stage functionalization via C-H activation has advanced tremendously [9]. Indeed, over the years, academic groups have devoted considerable effort to go beyond the development of new C-H activation reactions that are compatible with benchmark substrates and to showcase the synthetic usefulness of these reactions by demonstrating their compatibility with complex molecules, such as drugs and natural products.

In parallel, industrial chemists tend to implement LSF mindset, often combined with high-throughput experimentation (HTE), to expand a chemical space. In this chapter, key achievements in the LSF via C-H activation are present. First, the LSF of aromatic compounds will be discussed, followed by the aliphatic ones. Each section is organized following the nature of the C-H transformation, including direct alkylation, arylation, alkenylation, and C-X bond formation (C-O, C-N, C-B, halogenation). Attention will also be focused on the selectivity issues and the mechanistic reasons controlling the site-selectivity. It is worth noting that while remarkable progress has been achieved in LSF of peptides [10], this topic has been covered extensively and therefore is considered out of the scope for this chapter.

1.2 Late-Stage Functionalization of Aromatic Compounds

1.2.1 LSF via Direct Alkylation

LSF of drug candidates is particularly appealing when allowing the direct introduction of small groups, impacting significantly the biological activity of a drug molecule while not altering excessively the overall structure. Accordingly, LSF methylation is particularly attractive, allowing to rapidly explore the "magic methyl effect" on a given molecule. An early example of late stage methylation was published in 2013 by Yu [11]. The catalytic system combining Pd(OAc)2 catalyst together with a monoprotected amino acid ligand and Li2CO3 was initially optimized to perform direct C-H methylation of aryl carboxylic acids using methyl tetrafluoroborate as the methylating agent (Figure 1.1a). The potential of this strategy could be further demonstrated by synthesizing a therapeutically relevant methylated derivative of BMS-98947-0550-1. The superiority of the carboxylate function as a directing group over the ester motif warrant fully site-selectivity of this reaction, furnishing the modified compound in synthetically useful 55% yield. Subsequently, Pilkarski [12] and Larrosa [13], designed, respectively, Rh and Ru-catalyzed direct methylation of phenyl pyridine and imine derivatives (Figure 1.1b, c). While RhCp* catalytic system was designed to promote direct methylation using methyl tetrafluoroborate reagent under mechanochemical, solvent-free conditions, the Ru-metallacyclic catalyst performed well in the presence of quaternary anilinium salt. Both protocols could be subsequently used for LSF of drugs, delivering, in high yields, methyl derivatives of etoricoxib and diazepam.

Figure 1.1 Late-stage functionalization via direct alkylation. a) Pd-catalyzed late-stage methylation of BMS-98974-0550-1 developed by Yu et al. [11] b) Rh-catalyzed methylation of etoricoxib developed by Pilarski et al. [12] c) Ru-catalyzed methylation of diazepam developed by Larrosa et al. [13] d) Co-catalyzed methylation of rucaparib developed by Johansson, Ackermann et al. [14] e) Mn-catalyzed methylation of diazepam developed by Ackermann et al. [15] f) Ru-catalyzed methylation of emapunil and zolpidem developed by Ackermann, Johansson et al. [16] g) Ru-catalyzed methylation of carprofen developed by Lan, Zhao et al. [17].

Regarding the great potential of LSF for R&D sector, several companies conduct their own research programs. Considering the "magic methyl effect," the researchers from AstraZeneca in collaboration with Ackermann focused on developing a general protocol for LS-methylation of complex molecules [14]. Use of HTE approach allowed them to rapidly screen hundreds of reaction conditions, aiming in a general protocol compatible with a large diversity of compounds incorporating distinctive functional groups that could play the role of DGs. This exceptional approach resulted in establishing a Co-based protocol, enhancing direct methylation of 26 structurally very different simple aromatics. Remarkably, the same protocol allowed direct methylation of 16 drug-like molecules, clearly demonstrating its unique efficiency for the LSF of biologically active compounds. The regioselectivity in this reaction is generally controlled by a directing effect of coordinating groups (amides, carbonyl, N-heterocycles), thus allowing a predictable and site-selective methylation (Figure 1.1d). Although in some cases rather low yields were reached, this approach enables a tremendous research acceleration compared to the traditional de novo synthesis, as exemplified on rucaparib. Co-catalyzed methylation furnished the desired methylated derivative in one step in 39% yield, whereas de novo synthesis would require a 12-step protocol.

While methylation reactions are highly appealing from drug-design perspective, LSFs via alternative alkylation reactions have also been described. The group of Ackermann explored Mn-catalyzed direct allylation reactions (Figure 1.1e) [15]. Although initially developed for C2-functionalization of indole scaffolds with remarkable applications for peptide modifications, this transformation performed equally well using diazepam as the starting material. The desired alkylated diazepam was thus isolated in 65% yield, and the excellent site-selectivity results from ortho-directing properties of the imine group.

A complementary, distal meta-selective direct alkylation of complex molecules was recently reported by Ackermann and Johansson (Figure 1.1f) [16]. The HTE approach allowed expedient evaluation of various reaction conditions, targeting a reaction protocol compatible with a diversity of aromatic substrates. After the extensive optimization, [Ru(O2CMes)2(p-cymene)] was determined as the optimal catalyst in the presence of a monophosphine ligand L1, allowing direct meta-coupling between aromatics bearing various DGs and bromoalkanes....