Review

The importance of synthetic chemistry in the pharmaceutical industry

See allHide authors and affiliations

Science  18 Jan 2019:
Vol. 363, Issue 6424, eaat0805
DOI: 10.1126/science.aat0805

Synthetic innovation in drug development

Chemical synthesis plays a key role in pharmaceutical research and development. Campos et al. review some of the advantages that have come from recent innovations in synthetic methods. In particular, they highlight small-molecule catalysts stimulated by visible light, enzymes engineered for versatility beyond their intrinsic function, and bio-orthogonal reactions to selectively modify proteins for conjugation. High-throughput techniques are also poised to accelerate methods optimization from small-scale discovery to large-scale production, and complementary machine-learning approaches are just coming into focus.

Science, this issue p. eaat0805

Structured Abstract

BACKGROUND

Over the past century, innovations in synthetic chemistry have greatly enabled the discovery and development of important life-changing medicines, improving the health of patients worldwide. In recent years, many pharmaceutical companies have chosen to reduce their R&D investment in chemistry, viewing synthetic chemistry more as a mature technology and less as a driver of innovation in drug discovery. Contrary to this opinion, we believe that excellence and innovation in synthetic chemistry continue to be critical to success in all phases of drug discovery and development. Moreover, recent developments in new synthetic methods, biocatalysis, chemoinformatics, and reaction miniaturization have the power to accelerate the pace and improve the quality of products in pharmaceutical research. Indeed, the application of new synthetic methods is rapidly expanding the realm of accessible chemical matter for modulating a broader array of biological targets, and there is a growing recognition that innovations in synthetic chemistry are changing the practice of drug discovery. We identify some of the most enabling recent advances in synthetic chemistry as well as opportunities that we believe are poised to transform the practice of drug discovery and development in the coming years.

ADVANCES

Over the past century, innovations in synthetic methods have changed the way scientists think about designing and building molecules, enabling access to more expansive chemical space and to molecules possessing the essential biological activity needed in future investigational drugs. In order for the pharmaceutical industry to continue to produce breakthrough therapies that address global health needs, there remains a critical need for invention of synthetic transformations that can continue to drive new drug discovery. Toward this end, investment in research directed toward synthetic methods innovation, furthering the nexus of synthetic chemistry and biomolecules, and developing new technologies to accelerate methods discovery is essential. One powerful example of an emerging, transformative synthetic method is the recent discovery of photoredox catalysis, which allows one to harness the energy of visible light to accomplish synthetic transformations on drug-like molecules that were previously unachievable. Furthermore, recent breakthroughs in molecular biology, bioinformatics, and protein engineering are driving rapid identification of biocatalysts that possess desirable stability, unique activity, and exquisite selectivity needed to accelerate drug discovery. Recent developments in the merging fields of synthetic and biosynthetic chemistry have sought to harness these molecules in three distinct ways: as biocatalysts for novel and selective transformations, as conjugates through innovative bio-orthogonal chemistry, and in the development of improved therapeutic modalities. The development of high-throughput experimentation and analytical tools for chemistry has made it possible to execute more than 1500 simultaneous experiments at microgram scale in 1 day, enabling the rapid identification of suitable reaction conditions to explore chemical space and accelerate drug discovery. Finally, advances in computational chemistry and machine learning in the past decade are delivering real impact in areas such as new catalyst design, reaction prediction, and even new reaction discovery.

OUTLOOK

These advances position synthetic chemistry to continue to have an impact on the discovery and development of the next generation of medicines. Key unsolved problems in synthetic chemistry with potential implications for drug discovery include selective saturation and functionalization of heteroaromatics; concise synthesis of highly functionalized, constrained bicyclic amines; and C-H functionalization for the synthesis of α,α,α-trisubstituted amines. Other areas, such as site-selective modification of biomolecules and synthesis of noncanonical nucleosides, are emerging as opportunities of high potential impact. The concept of molecular editing, whereby one could selectively insert, delete, or exchange atoms in highly elaborated molecules, is an area of emerging interest. Continued investment in synthetic chemistry and chemical technologies through partnerships between the pharmaceutical industry and leading academic groups holds great promise to advance the field closer to a state where exploration of chemical space is unconstrained by synthetic complexity and only limited by the imagination of the chemist, enabling the discovery of the optimal chemical matter to treat disease faster than ever before.

Evolution of synthesis as a driver of innovation in drug discovery.

Past, present, and future advances in synthetic chemistry are poised to transform the practice of drug discovery and development.

Abstract

Innovations in synthetic chemistry have enabled the discovery of many breakthrough therapies that have improved human health over the past century. In the face of increasing challenges in the pharmaceutical sector, continued innovation in chemistry is required to drive the discovery of the next wave of medicines. Novel synthetic methods not only unlock access to previously unattainable chemical matter, but also inspire new concepts as to how we design and build chemical matter. We identify some of the most important recent advances in synthetic chemistry as well as opportunities at the interface with partner disciplines that are poised to transform the practice of drug discovery and development.

Over the past century, innovations in synthetic chemistry have greatly enabled the discovery and development of important life-changing medicines, improving the health of patients worldwide. In recent years, many pharmaceutical companies have chosen to reduce their R&D investment in chemistry, viewing synthetic chemistry more as a mature technology and less as a driver of innovation in drug discovery (13). Contrary to this opinion, we believe that excellence and innovation in synthetic chemistry continues to be critical to success in all phases of drug discovery and development. Moreover, recent developments in new synthetic methods, biocatalysis, chemoinformatics, and reaction miniaturization have the power to accelerate the pace and improve the quality of products in pharmaceutical research. The application of new synthetic methods is rapidly expanding the realm of accessible chemical matter for modulating a broader array of biological targets, and there is a growing recognition that innovations in synthetic chemistry are changing the practice of drug discovery (4, 5). Here, we identify some of the most enabling recent advances in synthetic chemistry as well as opportunities that we believe are poised to transform the practice of drug discovery and development in the coming years.

The pharmaceutical sector is currently facing multiple challenges: an increasing focus on complex diseases with unknown causal biology, a rapidly changing and highly competitive landscape, and substantial pricing pressures from patients and payers. In this challenging environment, drug discovery scientists must select biological targets of relevance to human disease and find safe and effective therapeutic molecules that appropriately modulate those targets. The current toolbox of synthetic methods and common chemical starting materials provides access to chemical space (6) that can be efficiently explored and mined to identify a suitable ligand and subsequently pursue studies of that preliminary lead compound toward its potential development as a successful drug. Brown and Boström have noted that a historical overreliance on just a few robust synthetic transformations (amide bond formation, sp2-sp2 C-C cross-coupling, and SNAr reactions) has biased the output of many drug discovery efforts, leading to narrow sampling of chemical space (7). In other cases, the lack of any reasonable method of synthesis has, at minimum, hampered thorough evaluation of chemical space or, at worst, prevented it completely.

Conversely, the discovery of breakthrough synthetic methods can truly transform the process of drug discovery. Innovation in synthetic chemistry provides opportunity to gain more rapid access to biologically active, complex molecular structures in a cost-effective manner that can change the practice of medicine. An outstanding example of the transformative power of synthetic chemistry in drug discovery is the application of carbenoid N-H insertion chemistry to the synthesis of β-lactam antibiotics (8). In the 1950s, the synthesis of antibiotics such as penicillin represented a formidable challenge to medicinal chemists, and broad exploration of structure-activity relationships (SAR) within this class of compounds was hindered by a lack of good methods of synthesis for these chemically sensitive structures. Indeed, the first chemical synthesis of penicillin took nearly a decade of dedicated effort to achieve (9) despite an intensive effort across multiple laboratories. This lack of synthetic accessibility prevented thorough evaluation of structurally related antibiotics that might have a broader spectrum of activity and an improved resistance profile. The application of intramolecular N-H carbenoid insertion chemistry (Fig. 1) to these structures provided a disruptive solution to the preparation of these fused β-lactams. This synthetic method was applied to the preparation of numerous natural and synthetic anti-infectives, including thienamycin (10), which subsequently led to the discovery and industrial manufacture of the antibiotic imipenem. In this example, synthesis enabled design, opening access to previously unattainable molecules of high therapeutic value.

Fig. 1 Synthetic method innovations enable discovery of important anti-infectives, imipenem and vaniprevir.

The development of targeted medicines for the treatment of chronic hepatitis C infection, a global health challenge (11), illustrates another key advance that innovative synthetic chemistry has contributed to drug discovery in recent years. The design and synthesis of hepatitis C virus (HCV) NS3/4a protease inhibitors represents a formidable challenge for medicinal chemists because the active site of this protease has a shallow, open binding site, and the enzyme possesses both genotypic and mutational diversity. Early studies of peptide-based inhibitors and subsequent molecular modeling suggested that construction of large, macrocyclic enzyme inhibitors could provide favorable ligand-protein binding and potent inhibition of this essential viral protease (12). The relatively flat and featureless protein surface requires a large ligand to gain sufficient binding affinity, while constrained macrocyclic ligands minimize the entropic cost of inhibitor binding. The application of ring-closing metathesis chemistry (13) has been transformative in the synthesis of many HCV NS3/4a protease inhibitors of varying ring sizes and complexity, including six approved drugs: simeprevir (14), paritaprevir (15), vaniprevir (16), grazoprevir (17), voxilaprevir (18), and glecaprevir (19). Ring-closing metathesis chemistry enabled the discovery of these and related macrocycles, allowing rapid assembly of complex bioactive molecules and broad exploration of SAR to address a range of properties.

In the two examples described above, the discovery of new synthetic pathways changed the way scientists thought about designing and building molecules, which broadened the accessible chemical space and thereby furnished molecules possessing the biological activity required in future drug candidates. The ability of the pharmaceutical industry to discover molecules to treat unmet medical needs and deliver them to patients efficiently in the face of an increasingly challenging regulatory landscape is dependent on continued invention of transformative, synthetic methodologies. Toward this end, investment in research directed toward synthetic methods innovation, furthering the nexus of synthetic chemistry and biomolecules, and developing new technologies to accelerate methods discovery is absolutely essential. Pertinent examples in these three areas are reviewed below.

Synthetic methods innovation

Over the past 20 years, several scientists have been recognized with the Nobel Prize for the invention of synthetic methodologies that have changed the way chemists design and build molecules. Each of these privileged methods—asymmetric hydrogenation, asymmetric epoxidation, olefin metathesis, and Pd-catalyzed cross-couplings—have broadly influenced the entire field of synthetic chemistry, but they have also enabled new directions in medicinal chemistry research. Of particular interest are new synthetic methods that enable medicinal chemists to control reactivity in complex, drug-like molecules, access non-obvious vectors for SAR development, and rapidly access new chemical space or unique bond formations. Recently, there have been several reported methods in these categories that have been rapidly adopted by medicinal chemists as a result of their practicality and broad utility.

Owing to the diverse biological activity of nitrogen-containing compounds, the discovery of Pd-catalyzed and Cu-catalyzed cross-coupling reactions of amines and aryl halides to form C-N bonds resulted in the rapid implementation of these synthetic methods in the pharmaceutical industry (20). The methodology addressed an unsolved problem to quickly and predictably access aromatic and heteroaromatic amines from simple precursors, and as a result it was rapidly adopted by medicinal chemists. Further development of these methodologies by process chemistry groups for scale-up has resulted in optimized ligands and precatalysts, as well as generally reliable protocols that have further advanced the application of this methodology in discovery programs. Consequently, aromatic C-N bonds are common features in pharmaceutical compounds (21), highlighting the tremendous impact that controlled construction of C-N bonds in aromatic compounds has had on medicinal chemistry programs. The next frontier is development of reliable methods to accomplish Csp3-N couplings (22).

As the development of transition metal–catalyzed processes has advanced, application of cutting-edge methods to the predictable activation of C-H bonds for functionalization of complex lead structures can enable novel vector elaborations, changing the way analogs are prepared (23). In particular, late-stage selective fluorination and trifluoromethylation of C-H bonds in an efficient, high-yielding, and predictable fashion permits the modification of lead compounds to give analogs that potentially possess greater target affinity and metabolic stability without resorting to de novo synthesis. Methodological advances have enabled preparation of fluorinated analogs of lead structures under either nucleophilic or electrophilic conditions (24). One promising recent example shows that electrophilic aromatic fluorination can occur under mild conditions with a palladium catalyst and an electrophilic fluorine source such as N-fluorobenzenesulfonimide (NFSI) (25). In addition, trifluoromethylation of a structurally diverse array of drug discovery candidates using zinc sulfinates, in the presence of iron(III) acetylacetonate, generated analogs with improved metabolic properties (26). Visible-light photoredox catalysis has been also been applied to the practical, direct trifluoromethylation of heteroarenes (27).

Adoption of photoredox catalysis in the pharmaceutical industry has been rapid, owing to the practicality of the process, the tolerance to functional groups in drug-like candidates, and the activation of nonconventional bonds in drug-like molecules (28). Application of photoredox catalysis to the Minisci reaction was reported, enabling the facile and selective introduction of small alkyl groups into a variety of biologically active heterocycles such as camptothecin (29). Photoredox catalysis has also been used for the direct and selective fluorination of leucine methyl ester to afford γ-fluoroleucine methyl ester with a decatungstate photocatalyst and NFSI (Fig. 2). Numerous processes have been reported to access γ-fluoroleucine methyl ester, a critical fragment of the late-stage drug candidate odanacatib; however, this method enables the most direct and efficient method to access this key building block in the fewest operations from a commodity feedstock (30). More recently, photoredox catalysis was used to generate diazomethyl radicals, equivalents of carbyne species, which induced site-selective aromatic functionalization in a diverse array of drug-like molecules (31). This represents the latest of a series of very diverse, practical, and potentially impactful uses of photoredox techniques to assemble libraries of drug-like scaffolds for screening.

Fig. 2 Synthetic methods with potential to enable drug discovery.

Although the preceding examples highlight the power of photoredox catalysis to accomplish previously unimaginable reactivity under very mild conditions (32, 33), even more remarkable transformations are being reported via synergistic catalysis, where both the photocatalyst and a co-catalyst are responsible for distinct steps in a mechanistic pathway that is only accessible with both catalysts present. For example, the combination of single-electron transfer–based decarboxylation with nickel-activated electrophiles has provided a general method for the cross-coupling of sp2-sp3 and sp3-sp3 bonds. This method establishes a new way of thinking about the carboxylic acid functional group as a masked cross-coupling precursor, expanding the synthetic opportunities for a functional group that is ubiquitous in chemical feedstocks (34). Furthermore, leveraging synergistic catalysis with photoredox has resulted in the discovery of milder conditions for C-O (35) and C-N cross-couplings, allowing application of these methods to more pharmaceutically relevant substrates (36). The concise synthesis of the antiplatelet drug tirofiban (37) is an excellent example of how the pharmaceutical industry can readily use this methodology to facilitate drug discovery and development. As research continues to surge in this field, additional breakthroughs are anticipated, and these will likely change how molecules are designed and built.

Intersection of synthetic chemistry with biomolecules

Biopolymers including proteins, nucleic acids, and glycans have evolved to achieve exquisite selectivity and function in a highly complex environment. These properties are of great interest to the pharmaceutical industry not only from a target perspective, but also from a therapeutic perspective. The success of monoclonal antibodies, peptides, and RNA-based therapies attests to the power that nature’s platforms offer to our industry and patients. Recent advances in merging the fields of synthetic and biosynthetic chemistry have sought to harness these molecules and to expand useful manipulation of biomolecules in three distinct ways: as catalysts for novel and selective transformations, as conjugates through innovative bio-orthogonal chemistry, and in the development of novel and improved therapeutic modalities.

Biocatalysis

Historically, the broad adoption of biocatalysis was held back by a limited availability of robust enzymes, a relatively small scope of reactions, and the long lead time required to optimize a biocatalyst through protein engineering (38). The invention of a recombinant engineered Merck/Codexis transaminase biocatalyst for the commercial manufacture of sitagliptin (Januvia) has inspired the broader application of biotransformations in the pharmaceutical industry (39). Tremendous advances have been made in molecular biology, bioinformatics, and protein engineering, enabling the development of biocatalysts with desired stability, activity, and exquisite selectivity. The impact of this area of research is exemplified by the 2018 Nobel Prize in Chemistry, recognizing Frances Arnold “for the directed evolution of enzymes.” As a result, biocatalysis has become more prevalent as a tool in drug discovery, as a valuable method for drug metabolite synthesis, and as a tool to enable rapid analog synthesis for SAR (40). For example, in 2013, the important discovery that cyclic guanosine monophosphate–adenosine monophosphate (2′,3′-cGAMP) is the endogenous agonist of STING, a protein involved in the activation of innate immune cells, triggered an intense interest in the synthesis of cyclic dinucleotide (CDN) analogs (41). Typically, the total synthesis of CDNs by purely chemical transformations requires long linear sequences and results in a time-consuming and low-yielding process. The optimization of STING agonists was greatly facilitated by the realization that the endogenous enzyme cGAS, responsible for the in vivo production of 2′,3′-cGAMP, could be engineered and harnessed for the biocatalytic production of non-natural CDNs (Fig. 3). The cyclization of various nucleotide triphosphate derivatives in a single biosynthetic step considerably reduced the cycle time and increased the yield of CDN synthesis, inspiring the design of novel agonists and the generation of SAR in this class (42). The continued investment in biocatalysis will lead to innovative solutions for unsolved problems in synthetic chemistry in both the discovery and development arenas. This will be driven by increased speed of protein engineering, access to enzymes with a variety of natural and even unnatural (43) catalytic activities, and the implementation of biocatalytic cascade catalysis to efficiently build complex chemical matter from simple starting materials (44).

Fig. 3 Biocatalytic synthesis of novel cyclic dinucleotides.

Bio-orthogonal chemistry

Achieving selective reactions with biopolymers such as proteins presents a host of unique challenges to the synthetic chemistry community; proteins have multiple reactive centers, charged residues, higher-order structure, and are usually handled in an aqueous environment. Nonetheless, the opportunity to create improved conjugates as therapies and imaging agents, or to induce covalent interactions to identify protein targets, represents important value to therapeutic drug discovery.

Methods for selective conjugation to biomolecules have undergone major synthetic evolution over the past 20 years. The discovery and development of a suite of click reactions has served as a powerful and broadly applied tool in protein bioconjugation (45). This highly bio-orthogonal and biocompatible reaction offers a powerful alternative to heterogeneous conjugation to surface lysines or engineered cysteines, and spurred the development of complementary expression technologies that could incorporate unnatural elements or recognition tags into biopolymers. This evolution in conjugation chemistry is best evidenced in the field of antibody-drug conjugates (ADCs): The first generation of ADCs were heterogeneous conjugates, whereas those of the second generation are now almost entirely homogeneous, with growing evidence that the site of conjugation is an important determinant of overall ADC performance (46).

The development of additional bio-orthogonal chemistries that can lead to selective reaction with biomolecules, particularly without the requirement for engineering a recognition element into the biomolecule, is an important new frontier for synthetic impact. Two recent examples of synthetic innovation suggest this toolset is expanding for proteins. In many cases, having the ability to conjugate at either the N or C terminus of a wild-type protein should avoid unintended disruption of its function or secondary structure. The development of selective N-terminal conjugation chemistry (47) and complementary application of decarboxylative alkylation chemistry to the C terminus of a protein substrate (48) offer new insights into achieving bio-orthogonal and highly site-selective conjugation with complex biomolecules (Fig. 4). These reactions take advantage of local differences in basicity and ionization potential respectively and, in doing so, leverage the complexity that biopolymers offer.

Fig. 4 Bio-orthogonal reactivity with proteins at N and C termini.

Synthetic innovation and therapeutic modalities

As these advances in synthetic, biorthogonal, and biosynthetic chemistry merge, so too do our capabilities to improve therapeutic modalities in the space between synthetic small molecules and expressed large monoclonal antibodies. Peptides, oligonucleotides, and bioconjugates have been advanced particularly for biological targets deemed “undruggable” by small-molecule and antibody platforms. Advances in these chemistries inspire new platforms and improve the breadth of biological targets that we can address. Two examples of innovation in therapeutic modalities through synthetic and biosynthetic chemistry are described below, although many others are being invented in academic and industrial settings.

In the first case, it has long been appreciated that a critical element of the success of oligonucleotide-based therapies was the introduction of phosphorothioates into the oligo backbone, which afforded improved stability to biological matrices as well as improved membrane permeability to aid with cytosolic delivery. Although these and other improvements in stability and delivery have advanced the field and enabled novel therapeutics to enter the clinic, many oligo-based therapies require high doses to overcome barriers to delivery, and their use is limited by their toxicity. Further improvements in stability and potency of the oligonucleotide should contribute to a widening of the therapeutic index and dose lowering. Interestingly, the chemistry used to introduce stabilizing phosphorothioates leaves each center as a mixture of two P-stereoisomers. Therefore, most clinical phosphorothioate-containing oligos that have 20 base pairs are, in reality, a large mixture of stereoisomers (219), each with different potency and stability characteristics. The ability to control phosphorothioate chemistry through an oxazaphospholidine approach by Wada and colleagues (49) led to a practical and scalable platform (50) for stereopure antisense oligonucleotides that demonstrate preclinical superiority to the corresponding stereomixtures.

Within the peptide arena, there has been a growing recognition that cyclic peptides offer improved starting points for drug discovery programs relative to their linear counterparts, largely due to improvements in entropic cost for binding and proteolytic stability. Early display platforms developed to discover cyclic peptides relied on disulfide formation, and more recently on posttranslational introduction of bis-electrophiles that can cyclize peptides with two cysteine residues (51). Through combined application of a ribozyme biocatalyst to enable unnatural amino acid incorporation into peptides, and then bio-orthogonal chemistry for cysteine cyclization through that unnatural amino acid, the Suga lab has developed an improved mRNA display platform (52) that has demonstrated tremendous potential to identify peptide ligands for challenging targets. The merging of chemical synthesis and biosynthesis within a common platform inspires further exploration of cyclic peptide modality; the introduction of selection pressures and forced evolution into this platform begins to resemble aspects of natural product generation that has historically inspired both organic synthesis and drug discovery.

Technologies to accelerate innovation

High-throughput experimentation

Given the need to invent and rapidly deliver medicines to patients, the pharmaceutical industry must invest in capabilities with the potential to radically accelerate the discovery and industrialization of transformative synthetic methodologies. High-throughput screening in biology has been the foundation of hit discovery for decades, and in recent years, the pharmaceutical industry has strategically invested in the creation of high-throughput experimentation (HTE) tools for chemistry that enable scientists to test experimental hypotheses with hundreds of arrayed experiments (53). In the same time frame required for traditional single-reaction evaluation, the different parameters that determine reaction outcome, discrete variables (catalysts, reagents, solvents, additives), and continuous variables (temperatures, concentrations, stoichiometries) can be holistically explored in parallel (54). As a result, the synthetic chemist now has access to exponentially larger amounts of experimental data than ever before. One recent example of the use of end-to-end HTE in process development was the discovery and development of an organo-catalyzed, enantioselective, aza-Michael reaction for the commercial manufacture of the antiviral letermovir (Fig. 5) (55). In this work, a series of efficient synthetic pathways were envisioned by chemists and key transformations were evaluated in parallel using HTE. The emergence of an H-bonding catalysis mechanism was initially discovered with moderate enantioselectivity and low conversion using chiral phosphoric acids. Rapid evaluation of a large number of diverse scaffolds with H-bonding capability in this transformation resulted in the discovery of an efficient and highly selective bis-sulfonamide catalyst. Further HTE work enabled the mechanistic understanding of the transformation, leading to optimization of both the catalyst structure and definition of optimal processing conditions. In this study and in many others (56, 57), novel bond-forming reactions were conceived by scientists, discovered through HTE, and then rapidly industrialized for the commercial manufacture of late-stage drug candidates.

Fig. 5 High-throughput experimentation to accelerating reaction discovery.

HTE tools have also begun to have an impact in drug discovery (58). As new catalytic methods emerge that redefine which bonds can be forged, the breadth of the resulting substrate scope is poorly understood, as most test substrates commonly demonstrated in the literature are simple and not representative of the complex functionality common in drug candidates. Pre-dosed, reaction-specific HTE screening kits, containing a lab’s most successful and general catalyst systems, are used in discovery chemistry labs to enable the rapid identification of reaction conditions that work for these complex substrates. Additionally, HTE has recently been leveraged to benchmark emerging methods against different catalytic procedures through the creation of arrays of complex, drug-like substrates known as informer libraries (59) or through addition of diverse molecular fragments that can disrupt catalysis (60, 61). The use of these diagnostic methods allows exploration of the relationship between reaction types and diverse complex substrate structures, thus enabling synthetic practitioners to make better decisions about which synthetic methods to prioritize in their problem-solving. Additionally, miniaturization of HTE to nanomole scale—for example, by automated nanomole-scale batch (62) and flow (63) approaches—now enables the execution of more than 1500 simultaneous experiments at microgram scale in 1 day for rapid identification of suitable reaction conditions to explore chemical space and accelerate drug discovery. This capability is augmented by advances in rapid high-throughput analytics, such as MISER (multiple injections in a single experimental run) and MALDI (matrix-assisted laser desorption/ionization) mass spectrometry techniques (64), which have enabled the analysis of as many as 1536 reactions in very short time frames. Finally, nanomole HTE can also expedite the preparation of diverse, complex arrays of molecules and, when coupled directly with biological testing, can radically alter how drug discovery is performed (65).

Computational methods

The use of computer-assisted methods to guide synthetic chemistry is emerging as an important component in the practice of drug discovery. Advances in computational chemistry and machine learning in the past decade are delivering real impact in areas such as new catalyst design (66) or showing considerable promise in others such as reaction prediction (67). The application of deep learning methods has the potential to uncover new chemical reactions, expanding the access to new pharmaceutical chemical matter. Granda et al. (68) have reported promising results toward this end. By combining automated synthesis with machine learning, they reported the discovery of four chemical transformations with differentiated novelty.

Recently, computer-guided design has been successfully applied to the preparation of catalysts that provide asymmetric control of a cycloisomerization reaction (69). Computational methods were used to evaluate the catalytic pathway of a previously unknown reaction, leading to the hypothesis that the electronics of the catalyst ligand influence both the rate and stereoselectivity of the transformation. Application of quantum methods such as density functional theory (DFT) provided optimal ligand designs with markedly enhanced rate and selectivity over the original ligand. A second example where the use of computational methods aided in the design of a superior catalyst is reported in the synthesis of a pronucleotide (ProTide, Fig. 6) (70). Achieving selective phosphoramidation of a nucleoside at the 5′ hydroxyl over the 3′ hydroxyl with stereocontrol at the phosphorus center is highly challenging. A combination of mechanistic studies using a variety of chiral catalysts and DFT calculations of a proposed transition state further informed by experimental observations led to the rational design of a dimeric phosphoramidation catalyst with an improved rate and excellent stereoselectivity.

Fig. 6 Application of computational modeling to new catalyst design.

Despite these successes, the process for rational computational design of a catalyst is arduous, requiring the modeling of multiple mechanistic pathways and refinement of numerous molecules and transition states. A program for automating much of this process has been reported (71), and the advancement of such methods as well as the continual increase in processing power will drive further use of these tools in the future.

The application of machine learning to synthetic problems has also generated considerable interest and excitement. One area of active research is the use of algorithms for synthetic route planning to a target molecule (72, 73). Segler et al. combined Monte Carlo tree search and three neural networks to identify potential synthetic routes (74). The success of the approach was qualitatively evaluated through a double-blind A/B test, where 45 chemistry students showed no preference between machine-suggested synthetic routes versus literature routes for representative target molecules. Machine learning has additionally been applied to forward reaction prediction (75). Neural networks were used to predict the major product of a reaction using an algorithm that assigns a probability and rank to potential products. Additionally, machine learning was used to successfully predict the performance of a single reaction, a Buchwald-Hartwig amination, against multiple variables: reactants, catalysts, bases, and additives (76). Application of machine learning holds considerable promise for synthetic optimization of targets far exceeding those described herein, toward predicting routes, main products, side products, and optimal conditions, among others. The continued advancement of these methods leverages the wealth of public information in the scientific and patent literature as well as within pharmaceutical institutions. The quality, breadth, depth, and density of the data within the domain of the predictions is critical for driving toward high-accuracy models. Inclusion of examples of both successful and unsuccessful transformations is also highly important. HTE is a highly attractive, complementary technology for augmenting existing datasets by generating model-suitable data, maximizing information content through careful design of experiments and capacity to deliver large volumes of data in a rapid and cost-effective manner.

Future directions

As we have discussed, breakthroughs in synthetic chemistry have proven to be the inspiration for the discovery and development of new medicines of important therapeutic value. Despite the many advances described above, the pace and breadth of molecule design is still constrained because of unsolved problems in synthetic chemistry. Many opportunities still remain to advance the field, such that synthetic chemistry will never constrain compound design or program pace, and should actually inspire access to uncharted chemical space in the pharmaceutical industry.

Recently, we conducted a summit with key opinion leaders to assess the state of field and to identify areas of research in synthetic methods that would have critical impact in the pharmaceutical industry. Key unsolved problems in synthetic chemistry included selective saturation and functionalization of heteroaromatics, concise synthesis of highly functionalized, constrained bicyclic amines, and C-H functionalization for the synthesis of α,α,α-trisubstituted amines. Other areas, such as selective functionalization of biomolecules and synthesis of noncanonical nucleosides, were identified as emerging areas of high potential impact. We envision that partnerships between the pharmaceutical industry and leading academic groups in the field hold great promise to spur the invention of disruptive synthetic chemistry to address these areas.

The most intriguing idea to emerge from the discussion was the concept of molecular editing, which would entail insertion, deletion, or exchange of atoms in highly functionalized compounds at will and in a highly specific fashion. Many innovations discussed above possess elements of this aspirational goal; however, a truly general method of this type would substantially change the pace of drug discovery and reduce constraints on compound design. Figure 7 prospectively illustrates how analogs of a complex lead scaffold might be accessed via site selective C-H functionalization, heteroaromatic reduction, ring expansion, or ring contraction. The power to modify this scaffold directly and specifically not only avoids a potentially lengthy synthesis of analogs, but also removes any limitation of molecular design imposed by synthetic hurdles. We anticipate that breakthroughs in the area of molecular editing will improve the pace and quality of molecule invention, enabling the introduction of new and important medicines at a faster rate.

Fig. 7 Molecular editing to enable drug discovery.

Outlook

Synthetic chemistry has historically been a powerful force in the discovery of new medicines and is now poised to have an even greater impact to accelerate the pace of drug discovery and expand the reach of synthetic chemistry beyond the traditional boundaries of small-molecule synthesis. New methods of synthesis can greatly expand the rate of molecule generation while also providing opportunities to routinely synthesize complex molecules in the course of drug discovery. Manipulation of biomolecules either as catalytic reagents (i.e., engineered enzymes) or as substrates for site-specific modulation is becoming more accessible and creating new opportunities for producing novel therapeutic entities. Academic research continues to be an important venue for producing novel reactivity, and rapid application of new methods has the potential to further drive molecule invention in drug discovery. New technologies such as HTE, automation, and new analytical methods are accelerating the discovery of new reaction methods. Further, integration of computational reaction modeling with the vast quantities of experimental data generated by nanoscale HTE has the potential to build more informative models that can predict successful reaction conditions or even discover new reactions. The field of predictive chemical synthesis remains nascent, but opportunities to build prognostic algorithms via machine-learning processes are likely to expand in the coming years. Continued investment in synthetic chemistry and chemical technologies has the promise to advance the field closer to a state where exploration of chemical space is unconstrained by synthetic complexity and is only limited by the imagination of the chemist. Advancements in synthetic chemistry are certain to remain highly relevant to the mission of inventing new medicines to improve the lives of patients worldwide.

References and Notes

Acknowledgments: We thank M. Kress, J. Hale, L.-C. Campeau, D. Schultz, S. Krska, D. DiRocco, and A. Walji for their critical review of the manuscript; C. T. Liu for preparation of Fig. 1; and D. MacMillan, R. Sarpong, M. Gaunt, F. Arnold, and G. Dong for their participation in the Disruptive Chemistry Summit at Merck. The authors declare no competing financial interests.

Stay Connected to Science

Navigate This Article