 |
||||||||||||||||||||
Volume 82, Number 24 pp. 47-62 |
||||||||||||||||||||
Traditional methods thrive despite numerous hurdles, including tough luck, slow commercialization of catalytic processes |
||||||||||||||||||||
| |
||||||||||||||||||||
A survey by Frost & Sullivan estimates that in 2002, of the $7 billion in revenues worldwide from chiral products, 55% was generated by traditional technologies (chiral pool and separation), 35% by chemocatalysis, and 10% by biocatalysis. The survey projects that by 2005, worldwide revenues of $9.5 billion would be realized not much differently: 49% by chiral pool and separation, 36% by chemocatalysis, and 15% by biocatalysis. Demand for enantiopure chiral compounds continues to rise, primarily for use in pharmaceuticals but also in three other sectors: flavor and aroma chemicals, agricultural chemicals, and specialty materials. Demand from the drug industry is fueled by regulations governing chiral active pharmaceutical ingredients (APIs) and the recognition that enantiomers of a chiral compound could have dramatically different biological activities. Whereas chiral APIs previously were usually formulated as racemates, the preference now is for single enantiomers. Furthermore, the switch from a racemic to a single-enantiomer API is key to managing the life cycle, as well as improving the efficacy, of racemic drugs (C&EN, May 5, 2003, page 56). Such switches also contribute to the demand for optically pure compounds.
Meanwhile, in journals surveyed by Chemical Abstracts Service, the number of papers per year that are related to chiral technologies has tripled from just over 1,300 in 1994 to more than 4,400 in 2003, for a total of more than 24,000 chiral-technology-related papers published in the past 10 years. An overwhelming majority (72%) are about stereoselective or asymmetric syntheses. Yet when Hans-Ulrich Blaser, chief technology officer of Solvias, a Basel, Switzerland-based chemical company serving the life sciences industry, surveyed the literature three years ago for catalytic enantioselective processes, he found only 16 practiced on a commercial scale [Appl. Catal. A: Gen., 221, 119 (2001)]. Blaser's survey did not include commercial biocatalytic processes. More chemocatalytic processes may exist that he is not aware of, he says, because most companies do not publicize processes used in actual production. But clearly, the gap between R&D output and commercial application is wide. Research in academia is focused on molecule building at the bench scale, comments Enrico Polastro, vice president and senior industry specialist at Arthur D. Little Benelux, in Brussels. "What is easy to achieve in a flask might be extremely challenging in a reactor, given the heat- and mass-transfer considerations." Many factors contribute to the slow development of commercial-scale catalytic asymmetric processes. But the bottom line is cost. At the end of the day, the customer does not care what technology produces the required material. And to win business, suppliers must offer optically pure products at the most competitive price. At the economic level, some observers believe that a major damper to the practice of catalytic asymmetric chemistry is that most catalysts are not in the public domain. Many technologies are proprietary, exclusively available only to particular companies. When people consider using them, their first thought is not how wonderful the chemistry is but how much it will cost and how the price will be negotiated, observes Mukund S. Chorghade, president of the consulting company Chorghade Enterprises, Natick, Mass., and chief scientific officer of D&O Pharmachem, an intermediates and fine chemicals supplier based in Paramus, N.J. Complex negotiations turn off customers and prompt efforts to develop alternative methods, he adds. The situation has sparked a technological race that has produced a plethora of proprietary catalysts. Many of these catalysts are now available in research quantities from Strem Chemicals, Newburyport, Mass.: for example, ClMeOBIPHEP (Bayer Chemicals); CatAXium and CatASium families (Degussa AG); MonoPhos (DSM Pharmaceutical Products); Josiphos, Walphos, Mandyphos, and Rhophos (Solvias); and Synphos (Synkem). Companies develop their own catalysts to get around other companies' patents, says David Ager, competence manager for homogeneous catalysis at DSM. Very often, new catalysts are not any better than those already existing, but they give their inventors freedom to practice chemistries that previously were off-limits, he adds. In the view of Ronald Brandt, interim chief executive officer of the chiral catalyst developer Chiral Quest, Monmouth Junction, N.J., the task is easier for customers, because they set the price at which their molecule must be produced, whereas technology providers must work within the pricing already existing in the market. What's clear is that companies like Chiral Quest can't survive by only selling catalysts, unless those are very highly priced. For this reason, Brandt says, Chiral Quest, like other technology providers, offers multiple pricing options: Customers can buy a high-priced catalyst, pay royalties, or arrange anything in between. At the strategic level, customers have traditionally regarded as unacceptable reliance on technology over which they have no control or that is available from only one source. This attitude may be changing as technology providers demonstrate flexibility. Multiple sourcing can be arranged by sublicensing either to the customer or to a third party, says Michel Spagnol, vice president for strategic and technical marketing at Rhodia Pharma Solutions, Cranbury, N.J. The need for multiple sourcing also can be satisfied by production at multiple sites of a sole supplier, he adds. At the practical level from the technology developer's point of view, commercialization of catalytic asymmetric methods is tricky. For one thing, catalysis is still viewed as a high-risk step in fine chemicals production. "People in production units usually prefer noncatalytic reactions," Blaser says. "They have more experience with stoichiometric reactions, which are usually easier to control and more robust." The bigger problem is the still highly empirical nature of catalyst selection. When presented with a new molecule, no one can tell what catalyst is right simply from the target structure. That's because the understanding of reactions and catalytic mechanisms is far from complete. Analogies work, however, and scientists with a lot of experience in asymmetric catalysis are fairly successful in narrowing the field. "Catalyst selection is not just science, but also art. You need intuition, some luck, some feeling," Blaser says. The availability of substrates is also key. Sometimes synthesizing the appropriate substrate is more problematic than running the catalytic reaction itself, Ager says. Likewise critical is the availability of catalyst. A good catalyst is useless if one can't get it in commercial quantities. Many things can go wrong in process development, Polastro adds. Even very low levels of impurities can poison the system. The economics--of temperature control, mass transfer, agitation, solvent and catalyst recovery, among others--may not be favorable. Or the process may be so fragile that minor variations in operating parameters lead to significant changes in yield or quality. Time is another factor. Especially for products in development, the time to develop a catalytic asymmetric process may not meet the need to get materials out rapidly for testing. The window of opportunity is wider for established or mature products, for which developing a catalytic asymmetric step in the synthetic route could vastly improve production economics. FINALLY, good luck counts. In many cases, commercially viable catalytic chiral processes do not make it to prime time because the projects they support don't survive. When a product is not approved, a candidate does poorly in trials, or a promising lead is dropped, chiral chemistries that are ready to go are mothballed overnight. Ironically, it is usually under these circumstances that many successful process development stories are made public. For example, Avecia, based in Manchester, England, developed a commercial process to enantioselectively reduce m-nitroacetophenone to (S)-1-(3-nitrophenyl)ethanol. Everything was in place, ready to go to a 1,000-L scale, when the project died, according to John Blacker, technical manager for process technology. The asymmetric synthesis was developed to achieve better economics for a synthetic route in which the chiral alcohol is a key intermediate. With Avecia's proprietary transfer-hydrogenation catalyst CATHy, the conversion could be achieved in greater than 95% yield and greater than 95% enantiomeric excess with less than 0.1 mol % of catalyst; 25 kg of substrate could be converted in four hours in a 100-kg reactor. Because the chiral alcohol is an early intermediate, 95% enantiomeric excess is sufficient, Blacker says. "It's not worth spending a long time trying to get a perfect reaction at this point." Another recent example of a potentially commercially feasible but aborted catalytic asymmetric chemical process comes from a project Solvias carried out for the agrochemical company Syngenta, based in Basel. Syngenta required a route to a chiral mandelamide with fungicidal activity. After evaluating three enantioselective methods, Solvias prepared kilogram quantities of the material through reduction of methyl (4-chlorophenyl)glyoxalate with a ruthenium-(R)-MeOBIPHEP catalyst. The product, methyl (S)-4-chloromandelate, could be prepared in up to 94% enantiomeric excess. The free acid could be recrystallized to more than 99% enantiomeric excess. Reaction with the appropriate amine yields the required mandelamide. The process is far from optimized, but the efficiency achieved gives strong reason to believe that it would be feasible technically and economically, Blaser says. Syngenta lost interest in the compound class, however, and Solvias did no further work. But it is not uncommon that compounds that have been dropped are picked up again, he adds. If and when this project comes up again, the catalytic asymmetric chemistry might finally be developed to its full potential. According to Polastro, the method of choice at present for industrial-scale application still is traditional resolution. Others disagree, on the basis that the cost of resolution rapidly escalates with scale-up. However, recent advances in separations are making large-scale resolutions more cost-effective, Chorghade says. He notes that in some cases, simulated-moving-bed technology (SMB) is cost-effective on a commercial scale. SMB is a form of multicolumn continuous chromatography. Six to eight columns are run in series. Feed enters specific columns at particular times, material moves out from one column into the next, and fractions are collected in separate evaporators. The system operates in continuous mode with solvent recycling. The equipment is commercially available from companies such as Novasep. Depending on the size of columns, separations can range in scale from grams to tons. The technology is used to make the APIs for two approved single-enantiomer drugs, escitalopram and sertraline. SMB is a winner at Aerojet Fine Chemicals, according to Aslam A. Malik, vice president for technology and business development. Whereas the past few years have been rough for custom chemical companies, "our sales have doubled over the past three years," he tells C&EN. The success is due to the value Aerojet can offer because of SMB, he says. In one exclusive-synthesis example that he mentions without specifics because of confidentiality agreements, use of SMB reduced six steps of chemistry required in a process to two. "That cut cost way down," he says. The biggest advantage of SMB is rapid development. Recalling recent work, Malik says a process developed on a 50-mm-diameter-column unit to everyone's satisfaction in October 2003 was operating in a 200-mm-diameter-column unit by December. "In a few months, we went from a few kilograms to 250 kg. But as far as the engineers were concerned, we could have gone all the way to metric-ton quantities. Because SMB is a physical separation, once it works, it works at any scale," he says. The biggest disadvantage is the high cost of acquiring the technology. Malik says the bill can reach $15 million by the time a facility is up and running. Strong justifications must be made for that kind of spending. According to Malik, one of the most convincing arguments was that SMB simultaneously involves engineering and chemistry expertise, which Aerojet developed through the company's long history in the defense business. Miconazole is an antifungal agent used to treat skin diseases. For topical applications, a single-enantiomer formulation may not be needed. However, miconazole may be considered for oral treatment of other diseases, including tuberculosis. Should a new oral-route use be found, the drug will likely have to be formulated as a single enantiomer. Furthermore, structurally related antifungal agents are taken orally, and some of those are produced as single enantiomers. Chromatographic experiments were conducted to determine the appropriate chiral stationary phase and mobile phase for separating the enantiomers of the racemic intermediate, Cox explains. The data then were used in computer simulations to optimize operating conditions for separation at a scale of 15 metric tons per year. The process was run on a small-scale SMB unit to confirm results. The process was valuated on the basis of the costs of raw material and other material and personnel inputs, as well as costs related to the outsourcing of an SMB operation. The valuation also embodied four key assumptions: Yield will be 96%. Product will be 99% enantiopure. The undesired enantiomer cannot be reracemized. And the separation will directly yield material of the required enantiopurity. According to the calculations, the additional cost to make one enantiomer of miconazole at 99% enantiomeric excess would be $121 per kg. This is the extra cost over the procedure that yields racemic product. The amount includes the cost of SMB operation ($95 per kg) and of the raw material that is discarded at the end ($26 per kg). To put that extra cost in context, Cox estimates that the chemical step to convert the intermediate to miconazole costs around $200 per kg. In this case, it will be more expensive to run the fairly straightforward chemical reaction than to separate the enantiomers.
What happens to the unwanted enantiomer also depends on the economics. Reracemizing and feeding the racemate back into the process is ideal but not always practical. In the miconazole case, the raw material costs $32 per kg. It is unlikely that reracemizing would be less costly in this example, Cox explains. People should not forget that the goal of chiral technologies--enantiopure product--also may be achieved with chemistry that already exists, notes David R. Dodds, founder of Dodds & Associates LLC, Manlius, N.Y., a consulting service for biotechnology and chemical companies. Process chemists seek the most robust, most productive, and least expensive synthetic route and aim to find it as fast as possible. Any reaction that can help reach this goal is useful. It is the overall process cost that will dictate which reactions will be used. And that cost covers not only reagents but also waste streams, utilities, equipment use, unit operations, and downstream requirements. Thus, it may be more commercially attractive to replace an elegant but expensive single reaction with several more mundane ones that have a lower total cost, he says. Such a situation is likely to arise when an asymmetric step requires an expensive chiral catalyst or chiral auxiliary. The power of conventional chemistry is reflected by routes developed by companies such as Zambon, based in Milan, Italy. An example provided by Livius Cotarca, R&D manager for fine chemicals, is the separation of (R)-flurbiprofen from flurbiprofen, a racemic nonsteroidal anti-inflammatory agent. (R)-Flurbiprofen is being studied for the treatment of Alzheimer's disease and various cancers. Zambon's patented process has been scaled up to kilogram production. It requires no catalysts. The resolving agent is (R,R)-thiomicamine, an advanced intermediate in Zambon's route to thiamphenicol. Another example of the clever use of stoichiometric chemistry is the practical synthesis of l-ribose by HanChem, a company based in Daejeon, Korea. Interest in l-sugars is high for therapeutic and cosmetic applications, says Myung Joon Seo, the company's vice president for custom manufacturing. Demand for l-ribose is several metric tons per year, and prices range from $700 to $1,000 per kg, he adds. HanChem prepares l-ribose through a piperidine-induced, one-pot inversion of 2,3,5,6-di-O-isopropylidene-d-mannono-1,4-lactone [Tetrahedron Lett., 44, 3051 (2003)]. The reaction has been run at a scale of 5 kg. Seo hopes that the process will be competitive with the best so far, an enzymatic route from d-glucose. It would be competitive if HanChem can find a low-cost supplier of d-mannose. The price of d-mannose has increased since HanChem began the development work, he says. Still, the consensus is that catalytic asymmetric routes are the most desirable. When C&EN asked several companies about their best commercial or commercially viable chiral chemistry, most offered catalytic asymmetric chemistries. For the intermediates division of BASF, Ludwigshafen, Germany, the best story so far is (S)-methoxyisopropylamine, according to Henning Althoefer, manager for new business development. The compound is an intermediate for the single-enantiomer active ingredient of the herbicide Outlook, a chiral-switch product. Frontier, another BASF herbicide, contains the racemic active ingredient. A DEDICATED PLANT in BASF's facilities in Geismar, La., produces (S)-methoxyisopropylamine at a scale of several thousand metric tons per year, Althoefer says. Production is based on enzymatic acylation of a racemic amine by a proprietary ester. Only one enantiomer is acylated to an amide, which can be readily separated from the unreacted amine. The same principle is used to make BASF's ChiPros chiral amines in Ludwigshafen. Usually, Althoefer points out, the unwanted enantiomers are reracemized and fed back into the process. Similarly, Dowpharma uses biocatalytic resolution using a lipase to prepare enantiomerically pure ß-amino acids. Its lipase technology was developed in the late 1990s at Chirotech Technology Ltd., Cambridge, England, which is now a subsidiary of Dow Chemical. Karen E. Holt, technology leader for biocatalysis at Dowpharma, says that in previous resolutions of racemic ß-amino acids the amine is protected--that is, the substrate is derivatized both at the carboxyl group (as an ester) and the amino group (as an amide). She and others showed that protecting the amino group is unnecessary and that the resolution can be achieved in fewer steps than had been tried before [Tetrahedron Lett., 41, 2679 (2000)]. This route to 99% enantiopure ß-phenylalanines has been performed at metric-ton scale at Dowpharma, she says. Ian C. Lennon, technology leader for chemocatalysis at Dowpharma, points out that Chirotech chemists examined producing ß-amino acids by catalytic asymmetric hydrogenation of carbon-carbon double bonds in unsaturated substrates. But given the state of that technology in the late 1990s, they concluded that such a route would be more complex, with more stages and steps, and would produce more waste even though the key step is asymmetric. The complexity is due partly to the mixture of E and Z isomers produced in the synthesis of the required substrates. Most catalysts effectively hydrogenate the E isomer in high enantiomeric excess, and not the Z isomer. But the Z isomer is formed in greater amounts because it is more thermodynamically stable. In this scenario, the overall yield of enantioselective hydrogenation would be inferior compared with biocatalytic resolution. Chiral Quest recently has offered one solution to this problem in the form of the rhodium-TangPhos catalyst. According to Xumu Zhang, a chemistry professor at Pennsylvania State University and the company's founder and chief technology officer, this catalyst is indifferent to the geometric-isomer form of the substrate. In reactions run at mild enough conditions, rhodium-TangPhos hydrogenates E and Z isomers in 9899% enantiomeric excess. Because of the high electron-donating ability of TangPhos, turnovers of up to 10,000 can be achieved, he claims. The company is now scaling up production of the catalyst to kilogram quantities in anticipation of demand. ß-Amino acids are so important--and the lipase route to them so general--that many companies are trying to commercialize the chemistry. ß-Amino acids are intermediates for various drugs being developed, and demand ranges from several hundred kilograms to a few metric tons per compound, according to Karlheinz Drauz, Degussa's vice president for technology and R&D management. Like Dowpharma's, Degussa's route begins with racemic ß-amino acids prepared through one-pot synthesis. The amino acids are then esterified with a Degussa proprietary ester. A commercially available lipase selectively hydrolyzes only the (S)-esters, releasing (S)-ß-amino acids. The free acid precipitates directly from the reaction in greater than 99.5% enantiomeric excess and a chemical purity of 99%. Drauz says Degussa has conducted several multi-hundred-kilogram campaigns for various ß-amino acids, such as ß-phenylalanines, and is preparing for even larger campaigns for the near future. Befor adopting the biocatalytic route to ß-amino acids, Degussa
also tried catalytic enantioselective hydrogenation of "D-tert-Leucine is a missing link in the chiral map," Drauz says. Many asymmetric ligands, catalysts, and chiral auxiliaries are based on tert-leucine, but only those incorporating L-tert-leucine could be made easily. Now the complementary compounds based on D-tert-leucine can be prepared "on a roughly similar cost basis," he explains. "This is a great result for all asymmetric reactions based on tert-leucine." Furthermore, a demand exists for D-tert-leucine as a building block for new drug candidates. It also is needed in the synthesis of the enantiomers of drug candidates containing L-tert-leucine, which must also be evaluated during drug development. Biocatalysis also has been successfully applied to commercial production of single-enantiomer 3-hydroxybutyrates from prochiral ketones. One particular compound, ethyl (S)-4-chloro-3-hydroxybutyrate (ECHB), is an intermediate in the synthesis of cholesterol-lowering drugs such as Lipitor (atorvastatin) and Crestor (rosuvastatin). Sales of Lipitor alone reached $10.3 billion in 2003. Many companies are vying to supply the chiral side-chains in these compounds. At Daicel Chemical Industries, Tokyo, production capability for ECHB is more than 100 metric tons per year by whole-cell biocatalysis. Significant improvements to the process have been made since C&EN reported on it five years ago (C&EN, July 19, 1999, page 65). The key step is asymmetric hydrogenation of the carbonyl group of ethyl 4-chloroacetoacetate (ECAA) by a biocatalyst. That catalyst is a carbonyl reductase originally isolated from Kluyveromyces aestuarii and then expressed in Escherichia coli. ECAA is prepared chemically from diketene, a core raw material at Daicel. The hydrogen source is reduced nicotinamide adenine dinucleotide (NADH). Previously it was regenerated through a glucose dehydrogenase system that converts glucose to gluconic acid. Because of environmental concerns regarding the large amount of gluconic acid formed in the waste stream, Daicel switched to a formate dehydrogenase system, which converts formic acid to carbon dioxide, according to John R. Peterson, chief executive officer of Thesis Chemistry LLC, Mentor, Ohio, which is the marketing representative of Daicel in North America. A stable and active formate dehydrogenase was isolated from Mycobacterium vaccae, and its productivity was improved by site-specific mutation. The variant now in commercial use furnishes ECHB with greater than 99% enantiopurity at a rate of 49.9 g/L, much more productive than the wild-type enzyme (19.0 g/L) and even better than the optimized glucose dehydrogenase (45.6 g/L). Other targeted, site-specific mutations of the formate dehydrogenase have led to two other benefits: increased tolerance to ECAA and better performance in an organic solvent. According to Peterson, ECAA is toxic to whole cells bearing the wild-type enzyme, which is usually shut down by organic solvents as well. In this case, fewer cells are dying from ECAA, and the enzyme actually works better in organic solvents. Competition to come up with better ways to make ECHB is intense. Dowpharma teamed up with the San Diego-based biocatalysis company Diversa to develop a route that is likely based on nitrilases (C&EN, Feb. 18, 2002, page 86; Oct. 7, 2002, page 8). "The team is not in a position to comment at this time," a Dowpharma spokesperson says in response to C&EN's request for an update on the project's status. Chiral Quest has joined in with its ruthenium-C3-TunePhos catalyst for asymmetric hydrogenation of ECAA. Zhang estimates that as little as 1 kg of this catalyst can produce up to 9 metric tons of 98 to 99% enantiopure ECHB from ECAA. And according to Brandt, Chiral Quest is supplying kilogram amounts of the catalyst to a client for commercial production of ECHB. Wacker Specialties, Munich, Germany, is also targeting single-enantiomer 3-hydroxybutyrates as part of its mission to be "the number one chiral alcohol producer based on ketones," according to Hans Pommerening, director for organic fine chemicals. The company is a major producer of diketene, from which prochiral acetoacetates are readily made. To compare the economics of various options for making 3-hydroxybutyrates, Wacker analyzed two routes it operates to make (R)-3-hydroxybutyrate esters: biocatalytic reduction with an isolated enzyme and catalytic asymmetric hydrogenation with ruthenium and a proprietary diphosphine ligand. On the basis of published literature, the company concluded that biocatalytic reduction by whole-cell fermentation is an inferior route, according to Pommerening. Wacker's isolated-enzyme biocatalytic route is based on an alcohol dehydrogenase from Lactobacillus brevis. The commercial application of this enzyme was developed by and is proprietary to Jülich Fine Chemicals (JFC). This company, based in Jülich, Germany, is one of Wacker's collaborators in developing enzymatic processes for fine chemicals. Wacker claims that the process yields 100% enantiopure product in 97% yield. It has been run at hundreds-of-kilograms scale. Wacker estimates that this route can produce commercial quantities of (R)-3-hydroxybutyrates at less than $100 per kg. By comparison, the chemocatalytic route using Wacker's proprietary ligand yields 98% enantiopure product in up to 95% yield. This route also has been run at hundreds-of-kilograms scale. Wacker estimates that the cost to produce (R)-3-hydroxybutyrates from this route will be 10 to 15% lower than by biocatalysis. Both routes use standard equipment. The biocatalytic route has an edge with regard to safety and health issues. It runs at ambient pressure and temperature, whereas the chemocatalytic route requires 100 °C temperatures and means for handling hydrogen, as well as a toxic solvent, methanol. On the other hand, the chemocatalytic route produces less than 100 g of organic waste per kg of product, compared with 2 L of waste per kg of product for the biocatalytic route. The biocatalytic route yields chiral alcohols of exceptionally high quality, says Thomas Maier, senior marketing manager at Wacker. For this reason, he says, Wacker is collaborating with JFC, as well as Prokaria, a Reykjavik company that discovers enzymes from Iceland's biodiversity pool, to develop processes for other specific chiral alcohols. At Bayer Chemicals, 3-hydroxyesters, including 3-hydroxybutyrates, are produced with enantiopurities of more than 99% in ton quantities per year by way of catalytic asymmetric hydrogenation of 3-ketoesters through Bayer's ruthenium-ClMeOBIPHEP catalyst, according to Ulrich Scholz, laboratory head for catalysis. At the request of a customer, Bayer has used the chemistry to prepare enantiopure (S,S)-pentan-2,4-diol from acetylacetone. Bayer prepared several kilograms of the enantiomer with greater than 99% enantiomeric excess and greater than 98% diastereomeric excess. Interestingly, the customer specializes in biocatalysis and has a biocatalytic route to the R,R enantiomer, Scholz says. "They were looking for a way to have both enantiomers in their portfolio." For Rhodia Pharma Solutions, the best chiral chemistry success so far is hydrolytic kinetic resolution (HKR) of racemic terminal epoxides, according to Spagnol. The technology was invented by Harvard University chemistry professor Eric N. Jacobsen and is licensed exclusively to Rhodia. The process produces tens of tons of single enantiomers of epichlorohydrin per year. An extremely versatile building block, epichlorohydrin provides access to a diverse range of intermediates of interest to the pharmaceutical industry. Rhodia also applies HKR to racemic propylene oxide, a commodity chemical, to prepare single-enantiomer propylene glycol. (R)-Propylene glycol is easily converted to (R)-propylene carbonate, an intermediate in the synthesis of the AIDS drug tenoforvir (Viread), from Gilead Sciences. This intermediate, produced in multiton quantities per year, is Rhodia's second largest product based on HKR technology, Spagnol says. Using a cobalt-salen catalyst, the Jacobsen hydrolytic kinetic resolution hydrolyzes only one enantiomer of a racemic epoxide to the corresponding diol. Because the diol and the unreacted epoxide differ greatly in their physical properties, they are readily separated. The process typically yields single-enantiomer epoxides in greater than 99% enantiomeric excess. That HKR consistently achieves excellent enantioselectivities was a major factor in its adoption by Rhodia, Spagnol says. With improvements on the original invention, including development of a second-generation catalyst, the process now uses less than 1 kg of catalyst to make 1 metric ton of product, Spagnol claims. The technology is so commanding that Rhodia is now the largest producer of single-enantiomer epichlorohydrins and (R)-propylene carbonate, he adds. Hard work and luck both contributed to Rhodia's success with HKR. According to Spagnol, from the time of licensing, it took at least four years of intense effort and investment to develop the technology to its current state. Meanwhile, demand for products derived from HKR has been increasing. With customers insisting on high-quality but low-cost supplies, the process had to be made extremely efficient. AN ALTERNATIVE ROUTE to single-enantiomer chiral epoxides is asymmetric epoxidation of olefins. At DSM, hopes are high for the reaction developed by chemistry professor Yian Shi at Colorado State University, Fort Collins. Called the Shi epoxidation, the reaction is catalyzed by a fructose-derived ketone and transforms trans alkenes that do not have to bear an allylic group to single-enantiomer epoxides with enantiomeric excesses usually greater than 95%. Pharmaceutical applications of the reaction have been licensed exclusively to DSM. According to DSM's Ager, DSM has used the original methodology, in which the oxidant is potassium peroxomonosulfate (Oxone), to make around 100 kg of a custom-synthesis product for a pharmaceutical customer. The substrate scope is broad, including trisubstituted olefins and olefins bearing a wide range of functional groups. But until recently, the Shi epoxidation was ineffective for cis olefins and terminal olefins. In 2002, the reaction was extended to these substrates through a catalyst that is a glucose-derived nitrogen analog of the fructose-derived ketone in the original invention. The Shi epoxidation complements the chemistry developed by chemistry Nobel Laureate K. Barry Sharpless. The Sharpless epoxidation, which has been licensed to PPG Fine Chemicals, is catalyzed by a titanium-tartrate complex and requires allylic alcohol substrates. "In many cases, newer technologies have proven to be more cost-effective than the Sharpless epoxidation," according to PPG's website. At Bayer Chemicals, meanwhile, Arne Gerlach, laboratory head for specialty chemicals, has spent part of the past three years developing the asymmetric epoxidation reaction called the Juliá-Colonna oxidation. Here the substrate is an enone; the product is a chiral epoxy ketone; the oxidant is hydrogen peroxide; and the chiral catalyst is poly-l-leucine, which can be recycled. Gerlach says the substrate scope of this reaction is limited to substituted enones. On the other hand, the reaction is atom economical and requires mild reaction conditions, an inexpensive oxidant, and a recyclable catalyst. These advantages were attractive enough to invest in its development, Gerlach says. According to Gerlach, the original reaction was not practical. Throughput was low, and workup was complicated. An industrially viable process now patented by Bayer Chemicals involves a simpler workup and a phase-transfer cocatalyst, which was not part of the original invention. A reliable synthesis of poly-l-leucine also has been established. In the Bayer process, the enone substrate is dissolved in an organic phase, such as toluene. Hydrogen peroxide and an inorganic base, to generate peroxide anion, are in an aqueous phase. Poly-l-leucine, which is not soluble in water or organic solvent, constitutes a third phase. The phase-transfer cocatalyst--tetrabutylammonium bromide--escorts the peroxide to the organic phase. This innovation cut the reaction time from more than 90 minutes to seven minutes in experiments with 75 mg of a chalcone substrate. In the organic phase, the peroxide delivers an oxygen atom to the substrate. It is believed that the substrate is bound to poly-l-leucine through the carbonyl of the enone so that only one face of the olefin double bond is available to accept it. The mixture is stirred at 700800 rpm to ensure efficient mixing of the three phases. After the reaction, the catalyst is filtered out, the liquid phases are separated, and the product is concentrated from the organic phase. Practiced on 100 g of the same chalcone substrate, Bayer's epoxidation protocol converts 75% of the starting material to an epoxy ketone with 95.5% enantiopurity. Recrystallization raises the enantiopurity to more than 99%. "We have now set the stage to produce multikilogram quantities with this technology," Gerlach says. Metolachlor has two elements of chirality: a chiral axis and a stereogenic center. The active ingredient in Dual therefore consists of four stereoisomers. In 1982, it was established that the biological activity resides only in the two S diastereomers. That knowledge spurred the search for a commercially viable chiral switch. The chiral switch is based on catalytic enantioselective hydrogenation of an imine. The reaction was developed in what was then the central research laboratories of Ciba-Geigy, in Basel, by chemists led by Blaser, Hans-Peter Jalett, Benoit Pugin, and Felix Spindler. One key to the success is the catalyst discovered as the work progressed--iridium complexed to a ferrocenyl diphosphine ligand now well known as Josiphos. The iridium-Josiphos catalyst delivers both satisfactory enantiomeric excess and an amazing turnover number of more than 1 million. That and the relatively mild reaction conditions--hydrogen pressure of 80 bar and a reaction temperature of 50 °C--yield a highly efficient, cost-competitive process. The story of (S)-metolachlor has been told, most recently in a chapter of the book "Asymmetric Catalysis on Industrial Scale: Challenges, Approaches, and Solutions," edited by Blaser and Elke Schmidt (Wiley, 2004). This account gives the impression that development proceeded as if according to a well-planned blueprint. "It's never the whole story in publications," says Blaser. "It's so much more complex." Earlier, in a much more nuanced report [Adv. Synth. Catal., 344, 17 (2002)], Blaser aptly called the 13-year (198194) effort an odyssey and compared the quest for the right catalyst with navigating a labyrinth. When the search began, almost nothing was known about enantioselective reduction of imines, catalytic asymmetric hydrogenations that could provide guidance were few, and the number of ligands to choose from was limited. But even now, with more ligands, more screening capabilities, and more people with hands-on experience, finding the right catalyst is still the key, Blaser says. Optimizing and scaling up are routine by comparison. Lack of rapid analytical techniques was also a handicap at the early stages. Enantiomeric excesses were measured from optical rotations or by nuclear magnetic resonance spectrometry. Chromatographic methods were not developed until later. "Good analytics are crucial with any project," Blaser says. "If they are not reliable or are slower than the throughput of screening systems, they are useless." Initially, progress was slow as hypotheses about routes and catalysts failed. Nevertheless, the research continued because Blaser and his team were trusted by and had the strong backing of the company. At one point, management suspended the project as it considered whether another compound altogether should be developed to replace metolachlor. When finally it was decided to stick with the chiral switch, the project sped up with the full force of Ciba-Geigy's organizational expertise and institutional know-how. "You must have somebody who says, 'This problem is important, and we want to solve it,'" Blaser tells C&EN. "That was the situation at Ciba-Geigy's agrochemical division. It was their most important product, it eventually came off patent, and they needed a replacement." Having a champion and years of development time are rare in the current climate in which speed to market is everything. The window of opportunity is narrow, and success will be elusive without certain elements in place, Blaser says. These elements include a suitable library of ligands, auxiliaries, and metal precursors; access to rapid screening of catalytic systems; access to very good analytical equipment; and highly experienced specialists. The last is especially key, given that so many catalysts are being invented but their inventors do not always have experience with process development. Licensed catalysts are not worth much until somebody applies them to scale, he adds. The full potential of enantioselective catalysis is far from realized. As process chemists acquire more experience with catalytic systems, as researchers gain better understanding of enantioselective reactions and mechanisms, and as optically pure products progress through pipelines and eventually to markets, so will the promise of the catalytic chiral technology incrementally be fulfilled.
|
||||||||||||||||||||
| Chemical & Engineering
News ISSN 0009-2347 Copyright © 2004 |
THE
COST OF using patented chiral technology can be prohibitive,
Polastro points out. And the right price can be hard to figure out.
Assigning a value to a specific technology in the overall context of
a final product being developed is not an exact science. Suppliers
and customers will not discuss how they calculate these costs.
THE
CHEMISTRY
involves enantioselective transfer of hydrogen from formic
acid to the substrate. The aromatic nitro group is highly reducible
and would have been attacked by other catalysts, Blacker points out.
"This reaction gives a good example of the selectivity you can get
with a catalyst."
To
illustrate the economics of SMB, Geoffrey B. Cox, vice president and
general manager for separation solutions at 
,ß-unsaturated
eneamides, but it proved inferior. With Degussa's rhodium-MalPhos catalyst,
only 95% enantiopurity could be achieved. At least one recrystallization
would be needed to raise the enantiopurity to that required for pharmaceutical
applications. Furthermore, the route yields a derivatized ß-amino
acid. Releasing the amino acid requires another chemical step and additional
workup. "An enzyme that gives greater than 99.5% enantiomeric excess
in one step directly from the solution is hard to beat," Drauz says.
Degussa
is developing methods to reracemize the undesired enantiomers while
also seeking potential uses for them. It turns out that (R)-ß-amino
acid esters are excellent resolution agents. According to Drauz, in
late March, Degussa revealed that it has successfully used ethyl (R)-3-amino-3-phenylpropionate
in a process to resolve racemic tert-leucine for kilogram-scale
production of D-tert-leucine. This compound previously could
not be accessed through biocatalytic means, he says. Now a chemical
resolution is available.
THE
BIGGEST DIFFERENCE is that throughput for the chemocatalytic
route is three to four times higher than for biocatalysis. To achieve
the high enantiomeric excess of the biocatalytic route, the process
must run with dilute solutions. That means lower throughput rates,
Pommerening explains. Both routes are poised for operation at multiton
scale, he adds. The customer's requirement for purity and price will
dictate which route to take.
With
a capacity of more than 10,000 tons per year, the largest-scale enantioselective
catalysis at present produces the penultimate intermediate in the synthetic
route to (S)-metolachlor. This compound is the active ingredient of
Dual Gold, a broad-spectrum herbicide against grass weeds. The predecessor
of Dual Gold is Dual, with racemic metolachlor as the active ingredient.
Dual came to the market in 1976.