Cynthia A. Challener is a contributing editor to Pharmaceutical Technology.

Cynthia A. Challener

The incorporation of fluorine into APIs continues to be a key approach to improving important properties of candidate drug substances. While there are numerous methods for introducing trifluoromethyl groups, effective routes for selectively incorporating a single fluorine atom at a chiral center remain limited. Expensive, specialized fluorinating reagents and/or precious metal catalysts are often required to achieve the latter transformations. Researchers at GlaxoSmithKline (GSK) reported a simpler, cost-effective method for the preparation of fluorolactams via copper-catalyzed selective, direct fluorination using fluorine (F2) gas. This article summarizes the GSK research efforts (1).

	Although incorporation of fluorine atoms into APIs is known to increase their lipophilicity and impart other beneficial pharmacokinetic properties, there remain few commercial drugs with single fluorine atoms introduced at chiral quaternary carbons due to the difficulty in achieving this chemistry at scale. Numerous methods have been introduced in recent years for the enantioselective synthesis of these types of compounds, but most suffer from high cost.

	Specialized reagents based on Selectfluor (chloromethyl-4-fluoro-1,4-diazoniabicyclo[2.2.2]octane bis(tetrafluoroborate), transition-metal catalyzed reactions, and routes mediated by organocatalysts have all been reported. Unfortunately, all of these methods rely on reagents and/or ligands that often require multiple steps to synthesize and are typically too expensive to allow their use at commercial scale. In addition, most of these processes result in the production of large volumes of waste, which further reduces their practicality for large-scale manufacturing.

	As part of its environmental strategy, GSK aims to make its operations carbon neutral by 2050 (2). The company uses methods based on both qualitative and quantitative criteria to evaluate the atom economy and level of waste generation for proposed synthetic strategies. The obtained process mass intensity (PMI) values for different routes are compared to identify the most effective synthetic strategy.

	Scaling up a chiral fluorolactam building block

	Recently GSK researchers needed to scale up a process for the synthesis of chiral fluorolactam building blocks (enantiopure 2-fluoro-1,3-amidoesters) for the preparation of a series of spleen tyrosine kinase inhibitors as pre-clinical candidates. Three possible existing methods were considered: enantioselective fluorination of malonate esters using N-fluorobenzenesulfonimide (NFSI) and a Zn(OAc)

/DBFOX-Ph (4,6-dibenzofurandiyl-2,2’-bis(4-phenyloxazoline)) catalyst followed by formation of the amide, chiral alkylation of fluoromalonate derivatives followed by formation of the amide, or palladium-catalyzed fluorination using NFSI.

	Initially, the palladium-catalyzed route was selected for production of 100 g of the first chiral fluorolactam. The researchers found, however, that preparation of the catalyst involved a complex, multistep synthesis with low selectivity at large scale, leading to the need for extensive purification. A green metric analysis of the route identified issues with the enantioselective fluorination step due to the fact that NFSI is prepared using fluorine gas and the loss of material resulting from the low selectivity during catalyst preparation.

	As a second attempt, the GSK researchers considered the use of a fluoromalonate derivative. In this approach, a fluoromalonate ester would be converted to the desired fluorolactams via simple chemical transformations, which would then be subjected to enzyme-catalyzed resolution.

	The desired fluoromalonate was prepared via direct fluorination of the corresponding dimethyl malonate ester using fluorine gas. The reaction was catalyzed by copper nitrate and performed in acetonitrile, affording the desired product in high yield, even at the 50-g scale. Notably, the mass intensity was quite low.

	This fluoromalonate was then converted to the racemic fluorolactam derivative in three steps: Michael addition of acrylonitrile to form the nitrile, hydrogenation of the nitrile over palladium, and base-catalyzed ring closure. Hydrolase-catalyzed resolution of the nitrile, salt, and fluorolactam was then attempted.

	Ultimately after screening numerous enzymes, the GSK researchers found an effective method for the hydrolase-promoted amidation of the reduced salt to the fluorolactam in phosphate buffer with high enantioselectivity. The inexpensive enzyme CAL-B 10 000, a recombinant Candida Antartica Lipase B that is commercially available in multi-kilogram and tonne quantities, was selected. The enzyme was shown to be quantitatively recoverable and reusable three times without loss of activity.

	The researchers then further optimized each step in the conversion of the fluoromalonate ester to the reduced salt. The resulting three-step synthesis of the fluorolactam was found to have a PMI of 201, which was more than four times lower than that of the initial route based on palladium-catalyzed fluorination using NFSI (925).

	The optimized alternative route based on direct fluorination using fluorine gas was found to be highly competitive from a PMI perspective compared to the approach based on methods previously reported in the literature. A large percentage of the PMI value for the optimized synthesis is due to the final resolution step, which leads to loss of 50% of the product (as is the case in all resolutions).

	Despite this loss, the cost of the fluorination-amidase process was also found to be significantly lower than that for syntheses based on specialized (and expensive) electrophilic fluorinating agents prepared from fluorine gas and palladium catalysis. The researchers attributed this result to the facile recycling of the enzyme catalyst and solvents and the use of an inexpensive copper catalyst for the fluorination step. They also noted that the low cost and straightforward synthesis make this strategy suitable for scale up.

. 18, 1313-1318 (2016).
	2. GSK, Green Chemistry, www.gsk.com/en-gb/behind-the-science/how-we-do-business/green-chemistry/ accessed Aug. 12, 2016.


	When referring to this article, please cite as C. Challener, “The Search for a Greener Fluorination Reaction," 

Articles

Researchers at GlaxoSmithKline report a greener and lower-cost route to chiral fluorolactams that is suitable for scale up.

The Search for a Greener Fluorination Reaction

The majority of small-molecule drugs in development today contain chiral centers, and according to Global Industry Analysts (GIA) nearly 95% of drugs will be chiral by 2020 (1). GIA estimates that the global chiral technology market, which is dominated by pharmaceutical applications (followed by agrochemicals and flavors and fragrances), will reach $5.1 billion by 2017.

	Because FDA requires manufacturers to investigate the properties (physicochemical, pharmacokinetic, etc.) of all enantiomers/diastereomers of chiral drugs to determine their individual safety and efficacy, the development of efficient chiral synthesis technologies remains a primary target for many academic and industrial researchers. Several transformations with significant potential for use in API manufacture have been published in the past year. A few of them are highlighted in the following.


	Researchers in the Doyle lab at Princeton University in New Jersey have developed a direct cross-coupling reaction to produce 1,2-dihydropyridines, versatile building blocks that are highly useful for API synthesis (2). The reaction involves the enantioselective Ni-catalyzed cross coupling of arylzinc reagents with pyridinium ions formed 

	The use of pyridine is of particular interest, according to Abigail Doyle, an associate professor of chemistry at Princeton and corresponding author of the article. “A highlight of the method is being able to use pyridine as a substrate because it’s inexpensive and abundant and has rarely been used in transition metal and asymmetric catalysis,” she said (3). Pyridine typically binds to nickel catalysts, inhibiting further catalytic activity. This issue was avoided by adding a slight excess of iso-butylchloroformate, an activating agent that reacts with pyridine to form an intermediate species that does not bind to the catalyst.

	The generated, enantioenriched 2-aryl-1,2-dihydropyridine products can be elaborated to numerous piperidine derivatives with little or no loss in enantiomeric excess. Nine different elaborations commonly used by medicinal chemists for drug development were performed to demonstrate their utility (3).

Radical strategy for enantioselective carbon-carbon bond formation


	Most enantioselective reactions proceed via polar reactivity, with few radical reactions providing high selectivities. Radical mechanisms are attractive for forming quaternary carbon centers, however, because radical intermediates are not significantly affected by steric factors, as is the case with cationic and anionic intermediates. Researchers at the Institute of Chemical Research of Catalonia (ICIQ) in Spain have recently reported a method that harnesses the potential of radicals for enantioselective synthesis of quaternary carbon stereocenters (4). Paolo Melchiorre and his group at ICIQ used a combination of photoredox and asymmetric organocatalysis to mediate enantioselective radical conjugate additions to β,β-disubstituted cyclic enones.

	A chiral organocatalyst containing a redox-active carbazole moiety was designed to generate iminium ions that are then stereoselectively trapped via an electron-relay mechanism by photochemically generated carbon-centered radicals. The reaction proceeds under mild reaction conditions with high enantioselectivity and without the need for transition-metal catalysts, making it a sustainable process (5). “Combining the potential of metal-free enantioselective organocatalysis and visible light photocatalysis can significantly advance the development of a more responsible and sustainable stereoselective chemistry while strengthening the chemistry toolbox to better face the challenges of modern organic chemistry,” noted Melchiorre. He added that “This catalyst may provide further opportunities to stereoselectively generate quaternary stereocenters that cannot be forged by using traditional polar reactivity.”

-symmetrical, atropoisomeric functionalized biaryls


	Organocatalysis was also employed by scientists at Rice University in Texas to achieve the direct arylation of aromatic C(sp

)-H bonds in phenols and naphthols for the preparation of 1,1Ê¹-linked functionalized biaryls (6). Not only have the non-C

-symmetrical, atropoisomeric products not been previously synthesized, they are formed with complete regioselectivity, and the reaction is readily scalable and does not require an external oxidant. Results of density-functional theory calculations suggest that the reactions proceed via an asynchronous [3,3]-sigmatropic rearrangement of an 

	The biaryls are synthesized from readily available compounds, including quinone monoacetals and phenols/ naphthols. “Using these building blocks, we made 41 different chiral biaryl compounds in a relatively short time,” said corresponding author László Kürti (7). He added that the functionalized chiral biaryls are really versatile compounds. “You can use them outright as organocatalysts or complex them with transition metals to make new transition-metal catalysts. So the possibilities are unlimited. Moreover, these compounds can be used as building blocks en route to natural products with biaryl substructures in them.”


	Researchers at the University of Sydney in Australia have focused on the development of switchable phosphine ligands for transition metals whose chirality is modulated through an external trigger signal. These molecular motors make it possible to change the enantioselectivity in a catalytic process without the need to synthesize new catalysts. In the report, a photoswitchable chiral bisphosphine was prepared based on a unidirectional light-driven molecular motor that can be used to invert the stereoselectivity of a palladium-catalyzed desymmetrization reaction (8). As a result, each enantiomer can be formed using just one ligand.

	The bisphosphine ligand undergoes photochemical and thermal isomerization during unidirectional rotation around the motor central double bond, allowing for control of both the helicity of the ligand and the distance between the two phosphines. Notably, light-induced isomerizations in the ligand are endothermic, while heat-induced isomerizations are exothermic (9). The researchers report obtaining a nearly racemic mixture to very high enantiomeric ratios for each isomer (93:7 and 6:94) for the Pd-catalyzed desymmetrization reaction (8).

Understanding how chiral molecules behave


	Scientists at Newcastle University in England and the Swiss Federal Laboratories for Materials Science and Technology are learning about mechanisms of chirality through investigation of biomineralization, a process by which inorganic materials are formed with chiral structures following contact with proteins and polysaccharides, which act as modifiers during nucleation and crystal growth (10). The chirality of the biomolecules is transferred to the mineral structure; understanding how this process occurs can provide insights into the design of organic chiral transformations.

	In this report, Werner Hofer and his colleagues used scanning tunneling microscopy, photoelectron diffraction, and density functional theory to investigate how the chiral compound ‘buckybowl’ hemibuckminsterfullerene arranges copper surface atoms. “By understanding this process, we can now force materials to behave in a certain way, using biological plans to create the shapes and structures that we want. This has huge potential in the fields of materials design and drug synthesis,” Hofer said (11). In particular, the knowledge gained could facilitate the development of more efficient and effective enantioselective heterogeneous catalysts.
	Now that the proof of concept has been demonstrated, the challenge, according to the researchers, is to identify the right biomolecules that create the specific metal surface morphologies that are suitable for specific catalytic reactions (11).


	1. Global Industry Analysts, “Global Chiral Technology Market to Reach US$5.1 Billion by 2017, According to New Report by Global Industry Analysts, Inc.,” Press Release, April 4, 2012.
	2. J.P. Lutz, et al., 

. (2016). Advance article. DOI: 10.1039/C6SC00702C.
	3. Princeton University, “New Reaction Turns Feedstock Chemical Into Versatile, ChiraBuilding Block,” Press Release (Princeton, NJ, March 23, 2016).
	4. J.J. Murphy, et al., 

, 532 (7598), 218-222 (2016). DOI:10.1038/nature17438.
	5. Institute of Chemical Research of Catalonia, “Radical Route For The Synthesis Of Chiral Molecules,” Press Release (Tarragona, Spain, April 14, 2016).
	6. H. Gao, et al., Angewandte Chemie International Edition, 55 (2), 566-571 (2016).
	7. Rice University, “Chemical Design Made Easier,” Press Release (Houston, TX, Nov. 24, 2015).
	8. D. Zhao et al., 

., 6, 6652 (2015).
	9. The University of Sydney, “Molecular Motors For Chiral Catalysis,” Press Release (Sydney, Australia, October 2015).
	10. W. Xiao et al., 

, 8, 326-330 (2016).
	11. Newcastle University, “Nature’s Mirror-the Code for Chirality,” Press Release, (Newcastle upon Tyne, England, Feb. 8, 2016).

Recent research of efficient chiral synthesis technologies reveals potential uses in API manufacturing.

Chiral Advances Demonstrate Promise for API Synthesis

Pressures on drug manufacturers to reduce development timelines and costs while also providing safer and more effective medications are driving the pharmaceutical industry to seek more effective, alternative manufacturing solutions. In addition to continuous processes, both branded drug companies and contract development and manufacturing organizations are seeking greener routes to the production of complex APIs--routes that consume fewer resources and produce less waste. Replacement of more traditional reactions mediated by transition-metal catalysts with those facilitated by enzymes, or biocatalysts, is one option receiving growing attention. Jim Lalonde, senior vice-president of R&D with enzyme engineering company Codexis, shares his perspective on the benefits of biocatalysis for API synthesis.

: Why is biocatalysis considered to be greener than chemo (transition metal) catalysis?

: Biocatalysis is considered to be greener than transition metal catalysis in some key regards. Using the comparison of Merck’s second-generation process for sitagliptin (Rh/Josiphos asymmetric hydrogenation) with its third-generation biocatalytic transaminase process is illustrative, as it meets the green chemistry principles of safety, atom economy, waste prevention, and energy efficiency. Firstly, we eliminated the use of rhodium, which is considered toxic and had to be removed with an absorbent (which also generated a solid waste). Secondly, we eliminated the need for high-pressure, high-temperature hydrogenation and replaced it with a process that runs at ambient pressure and near ambient temperature. This change makes the biocatalytic process inherently safer, less energy-consumptive, and less expensive. The hydrogenation had to be performed in an expensive, custom high-pressure reaction vessel, while the biocatalytic transformation is conducted in standard reactors. The use of the transaminase catalyst also eliminated several synthetic steps, gave a higher overall yield, reduced the solvent demand, and resulted in a high purity product. The latter is often a feature of biocatalytic processes with engineered enzymes, as we can ‘evolve’ the enzyme until we have one that meets or exceeds product purity requirements. This high degree of customization is not typically possible with transition metal catalysts.

: What are the key changes that have led to growing interest in the use of biocatalysis in the pharmaceutical industry?

: The key advancement that has led to a large increase in the application of biocatalysis in the pharmaceutical industry has been the availability of enzyme engineering technologies. Previously, enzyme development took too long to allow the wide application of biocatalysts in commercial manufacturing, except for later-generation processes.

	Advances in genetics (genomic databases and next-gen sequencing), high-throughput biochemistry and robotics, data analysis, and computational modeling have enabled the rapid development of novel, customized biocatalysts that are optimized for API manufacture with improved efficiencies and reduced costs.

: When in the drug development cycle have biocatalysts most likely been used and why? Is this changing? If so, what is driving the change?

: Biocatalysts are used throughout the drug development process, from discovery through clinical Phases I, II, and III, to commercialization and even for generic drugs. The biggest change, however, is that biocatalysis is now becoming a first approach, rather than one of last resort. GlaxoSmithKline and Merck, through licensing our technology, now have the ability to rapidly demonstrate proof of concept with custom biocatalysts during the development of first scale-up routes to produce kilogram quantities. This milestone typically occurs in Phase I or II. With the in-house capability to then optimize an enzyme in months, they can meet their clinical development demands with enzyme-based routes. The second major change, which is just beginning to take hold, is that enzymes can enable the rapid synthesis of complex drug molecules that are not easily made using traditional organic chemistry. Consequently, many more candidates can be synthesized early in discovery programs.

: What are the most recent enzymes introduced to the market for API synthesis on a large scale? What might be expected in the next several years?

: Examples of new enzymes recently made available for API manufacture include an acylase specific for simvastatin manufacture that can regiospecifically acylate monocolin J, eliminating the need for protection and deprotection steps; imine reductases that reductively couple ketones with amines for the convergent synthesis of secondary and tertiary chiral amine drug targets; large libraries of transaminases and ketone reductases capable of enantiospecific synthesis of primary amines and secondary alcohols; and oxidases for the manufacture of asymmetric sulfoxides like esomeprazole and armodafinil.

	Of these transformations, currently there are only conventional catalysts for asymmetric oxidation, reduction of ketones, and asymmetric reductive amination. The enzymes are superior to conventional catalysts in that an enzyme with the targeted enantioselectivity can be developed for almost any substrate and is tolerant for virtually all functionality in the molecule.

	In development, there are enzymes for direct hydroxylation of unactivated C-H bonds, enzymes for direct synthesis of amides (no need for protection, activation, or deprotection steps), enzymes for regiospecific glycosylation of hydroxyl groups, and enzymes for the site-specific conjugation of proteins. There are no conventional transition metal catalysts for any of these transformations.

: What are the key advances in protein engineering and related technologies that have had the greatest impact on biocatalyst development in recent years? What might be expected in the next several years?
	

: Current engineering advances are such that wild-type enzymes are no longer used, and virtually all biocatalysts are engineered. Computational advances, metagenomics, gene synthesis, and sequencing advances have had the largest impact. Computational advances include molecular modeling in which docking of a target substrate into enzyme models can be performed 

 with sufficient accuracy to dramatically accelerate development of suitable biocatalysts. Data mining and machine learning algorithms that can correlate the performance data for a biocatalyst with the structure of the enzyme also guide the design of more optimal enzymes.

	Genomic databases, which contain sequence information on a vast and growing number of organisms, provide access to diverse enzyme starting points for engineering programs. The cost to synthesize and sequence genes has dropped exponentially, so we now typically sequence thousands of variants per week per development program. This information is used in the data mining algorithms. The combined effect of all of these advances is such that an enzyme development program that took several man-years a decade ago now can happen in a few months.

	Current molecular models of enzymes are static and based on X-ray crystallography, which is only a snap shot of a dynamic structure. As computational models improve with increased computational speed, we can expect to develop predictive models for protein structure that do not require crystallographic structure determination and molecular dynamics that will better model the kinetics of catalysis. Using these predictive models, one should need to only make and test a small number of variants to come up with a commercial biocatalyst. We should also be able to expand the types of enzymes to ones that catalyze non-natural transformations that are desirable to organic chemists.

: What about whole cell production of pharmaceutical intermediates and APIs? What is the state of this technology?

: Whole cell production of pharmaceuticals has been established for decades for the manufacture of natural product drugs (e.g., antibiotics such as streptomycin and the immunosuppressant Tacrolimus). Semisynthetics, or drugs that are made from fermentation followed by chemical modification, often have enhanced efficacy, however. Recently, routes to natural and semisynthetic drugs that are rare in nature (e.g., artemisinin from the sweet wormwood plant and opiods from poppies) have been produced using synthetic biology; the complete biosynthetic pathway for these drugs has been installed in an easily fermentable yeast, vastly lowering the cost of production and simplifying the supply chain for these valuable drugs. In addition, we can now create enzymes that act on non-natural molecules and even ones that catalyze non-natural reactions, which will allow the direct fermentation of a much wider range of drug molecules.

: Is there anything else about biocatalysis as a green alternative that is worth noting?
	

: We continue to be surprised by the breadth of chemistries possible using biocatalysis. Pharma companies estimate that anywhere from 25-75% of their pipeline can benefit from biocatalytic methods. In the area of drug discovery, there is a trend toward molecules of higher complexity with higher specificity and thus both lower dosage requirements and fewer side effects. Enzymes that have been designed for selective transformations allow the rapid creation of libraries of these complex molecules.


	When referring to this article, please cite as C. Challener, “Going Green with Biocatalysis," 

Enzymatic catalysis offers pharma manufacturers a way to implement the Principles of Green Chemistry.

Going Green with Biocatalysis

	Residual undesired proteins-host-cell proteins (HCPs) and high-molecular-weight (HMW) and low-molecular-weight (LMW) species-have the potential to affect the safety and efficacy of biopharmaceuticals. As a result, the levels of these protein impurities in biologic drug substances and drug products must be controlled and are typically considered as critical quality attributes. The properties of the different contaminants vary significantly. In addition, the identification and quantification of some of these residual proteins, particularly in the presence of large concentrations of the product protein, can be challenging. Use of multiple orthogonal analytical methods and purification techniques is often required. Advances in mass spectrometry are, however, enabling more detailed analyses and thus improved monitoring of residual proteins during downstream processing.

	The most common protein-based impurities that downstream processes are developed to control include high- and low-molecular-weight species, charge-variants, and host-cell proteins.

	HMW species include dimers, trimers, tetramers, etc., formed of monomers that can be either covalently or non-covalently linked, according to John Moscariello, vice-president of process development for CMC Biologics. “Often, these large protein impurities consist of misfolded monomers in which surfaces of the monomer are exposed that typically would not be in the monomeric form. They are believed to primarily impact safety, but also can impact efficacy,” he says. He does note, however, that there are data suggesting that some aggregated species are well tolerated. Even so, regulatory agencies expect biopharmaceutical manufacturers to control HMW species through specifications for the drug substance, drug product, and end-of-shelf life.

	LMW species, which include clipped species and half molecules for compounds that are intended to be dimeric, such as monoclonal antibodies and bispecific antibodies, must also be controlled during downstream processing. These molecules can impact both safety due to the potential to cause immunogenicity responses and efficacy as the result of missing structural features, such as complementarity determining regions, which play a role in determining the antigen-binding specificities of antibodies and T-cell receptors, according to Moscariello.

	HCPs are generated during the production of recombinant proteins. The host cells used for fermentation/cell culture produce HCPs necessary for the cells to function (growth, gene transcription, protein synthesis, etc.) and during cell death. They must also be controlled during downstream processing because they have the potential to elicit an immune response in patients due to the use of non-human cell lines (typically Chinese hamster ovary or CHO cells) for production. Some HCPs can also affect drug product stability. “Regulatory agencies expect robust control of these species through drug substance release testing,” Moscariello observes.

	The properties of the different types of residual protein contaminants in biologic drugs vary significantly, leading to the need to use different analytical methods to detect and monitor each. The presence of HMW species, for instance, is frequently determined using size-exclusion chromatography (SEC) coupled with high-pressure liquid chromatography (HPLC) or ultra-high performance liquid chromatography (UHPLC), while LMW species are detected using a capillary electrophoresis-sodium dodecyl sulfate (CE-SDS) (reduced and non-reduced) technique. HCPs are typically measured using enzyme-linked immunosorbent assays (ELISA).

	“SEC-HPLC (or UHPLC) is a very robust assay and typically has very good resolution for HMW species, but it is often difficult to resolve LMW species,” Moscariello says. As a result, CE-SDS is used for the detection of LMW protein impurities. The SDS denatures the protein, and if needed a reducing agent can be added to break any disulfide bonds. As in SDS-PAGE (polyacrylamide gel electrophoresis), the LMWs are separated based on size. For monoclonal antibodies, when running a reduced CE-SDS, typically highly resolved intact heavy and light chains, as well as any clipped species or non-reducible HMW species (i.e., covalently linked compounds) can be observed, according to Moscariello. To detect half-antibodies, however, non-reduced CE-SDS is frequently used because they are difficult to detect when using a reducing agent.

	Although ELISA is the most common technique for monitoring HCPs, the selectivity of this method is dependent on the antibodies used in the ELISA kit, and consequently some HCPs can potentially be missed, according to Peter Levison, senior marketing director, downstream processing with Pall Life Sciences. HCP ELISAs employ a polyclonal mixture of anti-host cell proteins, but often do not provide complete coverage for all of the HCPs expressed during a specific process. This situation is particularly true for commercial kits. “While these kits allow for off-the-shelf detection of a wide range of HCPs for various cell lines, including CHO, they likely have significant gaps in HCPs for each specific process. As a result, ELISAs are known to show significant variability and it is difficult to get accurate values,” Moscariello comments. On the other hand, he notes that the development of product-specific HCP assays with high coverage is costly and time consuming.

	Most companies are exploring more advanced techniques for the detection of HCPs and other protein impurities that are designed to overcome the limitations of existing methodologies. In particular, Moscariello notes significant investment in process analytical technology (PAT) tools for online or at-line monitoring of HMWs and other contaminants. For instance, he points to the use by Amgen of multi-angle light scattering (MALS) to monitor protein aggregates at the effluent of a cation-exchange column.

	Several advances in detection techniques for HCPs are also being made. Use of isotope tags for relative and absolute quantification (iTRAQ) HPLC technology can enable better detection of host-cell proteins, and when coupled with comprehensive genome sequences, can enable better identification of HCPs using established mass spectrometry (MS) databases, according to Levison. Data obtained from techniques such as surface-enhanced laser desorption/ionization-time of flight (SELDI-TOF) MS in combination with ELISA results are also allowing semi-quantification of total HCPs.

	The first step in the purification of bioprocess fluids taken from bioreactors is clarification. In this step, the biologic-drug substance is separated from cell debris and turbidity is removed, leaving a clear harvested cell-culture fluid. This step is typically achieved via depth filtration or centrifugation and does also lead to the reduction of undesired protein impurities. Pall Life Sciences introduced the Cadence Acoustic Separator technology in April 2016-an alternative clarification method for cell-culture bioprocess fluids, including those from continuous perfusion processes-based on acoustic wave separation technology licensed in 2015 from FloDesign Sonics (FDS).

	Following clarification, orthogonal chromatographic steps during protein purification are generally used to further reduce HCP, HMW, and other contaminants to acceptable levels, according to Levinson. “The most powerful step to remove HCPs is affinity chromatography,” asserts Moscariello. There are issues with this method, however. While almost universally adopted for monoclonal antibodies, this method is often not applicable for non-fragment-crystallizable (Fc)-region-containing products. There are significant product-HCP interactions with antibodies. It is common to try to disrupt these interactions with harsh washes, particularly on affinity resins, which requires that they be salt-tolerant. Levison adds that the introduction of new mixed-mode modalities has provided additional selectivities that are providing improved HCP clearance.

	There is also a lot of work being done with precipitating and flocculating agents, according to Moscariello. He notes that caprylic acid has been shown to significantly remove HCPs from product streams without significant loss of product.

	Advances in detection technologies are having an impact on the capability of downstream processing to remove undesired protein impurities, according to Moscariello “The ability to identify specific HCPs using state-of-the-art MS techniques has allowed downstream development scientists to assess the charge, size, and hydrophobicity of these HCPs and exploit their differences from the product of interest and develop more robust processes,” Moscariello adds.

	“Improvements in analytics and genomic/proteomic analyses may well provide better awareness of the presence of contaminants and therefore facilitate their clearance,” Levinson concludes.

	When referring to this article, please cite it as C. Challener, "Protein Impurities Pose Challenges,” 

Multiple methods are required for detecting and removing protein impurities.

Protein Impurities Pose Challenges

Continuous flow manufacturing for small-molecule drug synthesis received a vote of confidence in April 2016 when FDA approved a change from batch to continuous processing for Janssen Products’ Prezista (darunavir), a drug for the treatment of HIV-1 infection. This decision is the first such approval for a change from batch to continuous manufacturing, and the second for a continuous process; Vertex received approval in July 2015 for the continuous production of its cystic fibrosis drug Orkambi (lumacaftor/ivacaftor).

	In a blog posted on April 12, 2016, Lawrence Yu, FDA’s deputy director of the Office of Pharmaceutical Quality in the Center for Drug Evaluation and Research, expounded on the benefits of continuous manufacturing: greater reliability, safety, efficiency, responsiveness/flexibility, and reduced costs (1).

	He stated, “Although it is not easy for drug manufacturers to transition from batch to continuous manufacturing, there are significant rewards. FDA encourages others in the pharmaceutical industry to consider similar efforts.”

	Pharmaceutical companies are certainly interested as well, as indicated by the numerous reports of API synthetic methods designed to be operated under continuous flow conditions. A summary of selected processes is presented in the following.


	Researchers in China report a continuous-flow process for the preparation of 

-nitroaniline to afford a diazonium chloride intermediate, which is then subjected to azo coupling with sodium thiomethoxide to give 1-(methylthio)-2-(3-nitrophenyl)diazene, followed by dediazonization to provide 

-nitrothioanisole in high yield (2). A flow process was developed to minimize accumulation of the energetic intermediate diazonium salt and enable large-scale production.

	The same scientists also developed a continuous flow process for the selective nitration of 1-methyl-4-(methylsulfonyl)benzene in 98% yield using 80 weight percent (wt%) sulfuric acid (3). The short residence time of 5 s led to a significant reduction in byproduct formation. Notably, the spent acid could be reused multiple times after concentration under a vacuum.

	Swedish researchers established a protocol for the transcarbamylation/transesterification of sulfonylcarbamates with alcohols via microwave heating under continuous-flow conditions (4). Using transparent borosilicate tube reactors, the scientists prepared two series of O-alkylsulfonylcarbamates. They also applied the methodology to the synthesis of three angiotensin II type 2 receptor ligands.

	Scientists at the University of Graz (Austria) and Noramco report the continuous flow production of hydrocodone from thebaine via selective olefin reduction to form 8,14-dihydrothebaine followed by hydrolysis (5). Selective transfer hydrogenation using hydrazine hydrate (N

 is not possible under batch conditions due to the high explosion potential of the reaction mixture. The reported continuous high-temperature/high-pressure method is efficient, safe, and scalable. The process involves four consecutive feeds of liquid N

O with separate residence times for completion of the reaction in 1 h.

	Researchers from The Netherlands developed an intensified, solvent-free continuous process for the conversion of alcohols to chlorides using hydrogen chloride gas (6). A separate dry zone was used to deliver the hydrogen chloride gas and prevent corrosion. Less than half the equivalents of HCl are needed due to the use of the gas instead of hydrochloric acid, and only water is generated as a byproduct. The reaction is typically complete within 20 min.

	Scientists from Eli Lilly and Company and Vanderbilt University reported the conversion of a batch-based stereoselective aza-Henry reaction between an arylnitromethane and Boc-protected aryl aldimine using a homogeneous Brønsted acid-base catalyst to an automated intermittent-flow process with product crystallization and slow reagent addition (7). The organocatalytic enantioselective reaction included recycle of the catalyst, resulting in increased process intensity maintaining a high catalyst concentration in the reactor. The researchers performed a continuous campaign with the automated system and obtained high throughput and high selectivity while minimizing catalyst usage.

	Researchers from Eli Lilly have also achieved a continuous Grignard reaction using a continuous stirred tank reactor (CSTR) (8). The process was developed after mapping the operation space using knowledge about CSTR process parameters, the reaction kinetics, magnesium (Mg) sequestration efficiency, equipment characterization, and the impact of process disturbances on steady-state conditions. A reaction kinetics model was also developed, and computational fluid dynamics (CFD) models were applied to the equipment design. This approach enabled the development of a highly robust and safe commercial-scale process.

	Researchers at AztraZeneca, in combination with various European university groups, have focused on the development of continuous Buchwald-Hartwig reactions. In one report, the scientists used the highly active palladium N-heterocyclic carbene (NHC) precatalyst [Pd(IPr*)(cin)Cl] 4 (IPr* = 1,3-bis(2,6-bis(diphenylmethyl)-4-methylphenyl)imidazol-2-ylidene; cin = η3-cinnamyl), a universal approach for continuous workup and purification, and a methodology for catalyst recycling and reuse (9). Before implementing the flow reaction, the catalyst’s stability in water was investigated and an effective workup methodology was developed that allows separation of the Buchwald-Hartwig product into the aqueous stream as a salt while retaining the aryl bromide starting material in the organic stream with the catalyst.

	The flow reaction was performed in a microreactor at the gram scale using continuous workup conditions involving acidic extraction of the product and on-stream recycling of the catalyst in the organic phase (10). The custom-designed rector for running heterogeneous reactions in flow combines the flexibility of CSTRs with the smooth operation, low residence time distribution and excellent heat transfer capability of a conventional flow reactor, according to the researchers.

Low-temperature batch to continuous processes


	The conversion of low-temperature batch reactions to continuous processes is attractive because in many cases higher temperatures can be used to achieve equal or better results. For example, scientists from Eisai report the continuous production of eribulin mesylate via the diisobutylaluminum hydride reduction of an ester followed by n-butyllithium-mediated coupling of the resultant aldehyde to a sulfone fragment to afford a sulfone-diol as a key intermediate, all without the need for cryogenic conditions (11). The reduction was run at -50 °C instead of -70 °C, yet a better selectivity and yield were obtained, while the butyllithium-mediated coupling when run under flow conditions at 10 °C proceeded with a better conversion rate than the batch reaction when performed at -70 °C.

	Meanwhile, scientists at Bayer Technology Services report the running of a continuous process for the lithiation of fluoroaromatics in a microreactor demonstration unit (12). A higher temperature was possible in this case, too, leading to higher space-time yields compared to those observed for the batch reaction. Temperature control was found to be critical, though; optimization was necessary to identify the most appropriate reaction conditions, hardware, and reactant purification methods to achieve minimal salt precipitation and prevent runaway conditions.

	Investigation of continuous processing has not been limited to chemical syntheses. There has also been significant interest in developing continuous biocatalytic reactions, which are more complex and challenging to implement. German researchers reported the integration of microfluidic reactor technology with solid-liquid biocatalytic reactions (13). As a case study, they reported the biocatalytic conversion of styrene to (S)-styrene oxide using catalytic biofilms in a microreactor.

., Article ASAP, Web publication, March 24, 2016.
	3. Z. Yu, et al., 

. 20 (2), 199-203 (2016).
	4. I.Kumpina, et al., 

. 20 (2), 440-445 (2016).
	5. B. Pieber, D.P. Cox, and C.O. Kappe, 

. 20 (2), 376-385 (2016).
	6. S. Borukhova, T. Noël, and V. Hessel, 

Process Res. Dev. 20 (2), 568-573 (2016).
	7. S.V. Tsukanov, et al., 

 20 (2), 215-226 (2016).
	8. S.-W. Wong , et al., 

. 20 (2), 540-550 (2016).
	9. A.Chartoire, et al., 

. 20 (2), 551-557 (2016).
	10. S. Falss, et al., 

. 20 (2), 558-567 (2016).
	11. T. Fukuyama, et al., 

. 20 (2), 503-509 (2016).
	12. S. Laue, V. Haverkamp and L. Mleczko

. 20 (2), 480-486 (2016).
	13. R. Karande, A. Schmid, and K. Buehler, 

Research efforts around the world are demonstrating that the application of flow chemistry for small-molecule API synthesis.

Flow Chemistry Discoveries Advance Continuous Processing

Pharmaceutical Technology
	Vol. 40, No. 7
	Pages: 28–29

When referring to this article, please cite as C. Challener, "Expanding the Chiral Toolbox," 

Recent chiral advances demonstrate promise for API synthesis.