News Article | May 10, 2017
An analysis of available routes to olefin-containing sterol acetate 2a points to retrosynthetic deficiencies that still exist (Fig. 1a). Seven conventional steps are required to convert steroid derivative 1 to 2a (ref. 3), only one of which forms a strategic C–C bond. The entire strategy is built around the Wittig transform4, which requires redox-adjustment of the free carboxylate to the aldehyde, necessitating protecting-group manipulations. By this route only steroid 2a is accessible, because related methyl- and ethyl-containing sterol acetates (2b, 2c) would require more complex designs. The inefficiency of this sequence is a well known problem that has yet to be solved despite the attraction of a hypothetical method that would directly convert acylated 1 to 2a. Although olefins have the most richly developed reactivity and highest abundance (via the petrochemical industry), the diversity of alkyl carboxylic acid building blocks available is unmatched. If the aforementioned versatile olefin-based cross-coupling partners could be used in decarboxylative cross-coupling5, 6, 7, 8, 9, 10, novel synthetic pathways could be accessed. For example (Fig. 1b), one could envisage the total synthesis of diol-containing natural products such as 3 and 4 as arising from tartaric acid, perhaps the most inexpensive chiral building block available. This new disconnection could only be conceived in a decarboxylative fashion as the corresponding tartrate halides do not exist, and if they did, would most certainly not be stable. Here we present the invention of a general, scalable, chemoselective method for decarboxylative alkenylation that exhibits broad scope across a range of both olefin (from mono-substituted to fully substituted) and carboxylic acid (primary, secondary and tertiary) coupling partners (>60 examples, Fig. 1c). Decarboxylative alkenylation dramatically simplifies retrosynthetic analysis11. To demonstrate this, total syntheses of sixteen natural products across ten different natural product families spanning a range of steroids, polyketides, vitamins, terpenes, fragrances and prostaglandins are reported and directly compared to prior syntheses (see Methods for more information). The optimization of decarboxylative alkenylation is briefly summarized in Fig. 1d. In general, the reaction proceeded smoothly when using the tetrachloro-N-hydroxyphthalimide (TCNHPI, commercially available) esters, an inexpensive Ni(ii) source, and the abundant ligand 2,2′-bipyridine (bipy, L1). Using the piperidine-derived redox-active ester (RAE) 6, cyclohexenylation could be achieved at room temperature with 10 mol% of Ni(acac) •xH O, 10 mol% bipy, and 2.0 equivalents (equiv.) of alkenylzinc reagent 7 to furnish olefin 8 in 75% isolated yield. As has been demonstrated in previous work5, 6, 7, the RAE could also be generated in situ (entry 1 in the table of Fig. 1d) without any purification or even solvent removal. TCNHPI appears to be the optimal RAE (entries 2 and 3), and Ni(acac) •xH O proved superior to the more air- and moisture-sensitive NiCl •glyme (entry 4). In general, most common solvents were tolerated in this reaction (see Supplementary Information for details). Alternative ligands were also screened, and L1 was chosen as it is the least expensive (entries 5 and 6). Finally, as will be shown in several cases, the Fe-based catalytic system developed previously for RAE cross-coupling7, 12, 13 could be used as well (entry 7). The scope of this new decarboxylative alkenylation reaction is striking because all possible classes of olefin coupling partners could be employed with exquisite control of olefin geometry. The scope of alkenylzinc reagents that can be used in this coupling are exemplified in Fig. 2a, b. Simple mono-, di-, tri-, and tetra-substituted olefins are easily accessible (9–12, 16, 17). From a strategic perspective, the cycloalkenyl products (8, 13, 14) would be challenging to make in a more direct way from either the same starting material or even from a piperidone. The selection of which method to use when constructing an olefin (for example, Wittig or metathesis) is usually linked to the underlying mechanism of the process to enable production of the desired stereochemistry of the newly formed olefin. Decarboxylative alkenylation divorces the C–C bond-forming event from such stereochemical concerns. As such, conventional techniques can be used to produce the precise E or Z geometry of an alkenyl-organometallic, which can then be used without any isomerization. For example, olefin 12 is produced as a 1:1.5 mixture of E/Z isomers because the starting commercial Grignard reagent from which it was derived exists as a mixture. Similarly, styrenyl derivative 15 could be procured with high geometrical purity (>20:1 E/Z) as controlled by the starting alkenylzinc species (derived from lithium–halogen exchange of the corresponding styrenyl iodide). The power of stereocontrolled alkyne carbometallation14 can be coupled with this method to produce geometrically pure alkenyl iodides. Alkenylzinc reagents derived therefrom (lithium–halogen exchange/transmetallation) afford tri-substituted olefin products such as 18 and 24 that would be otherwise challenging to make in a single step (in situ RAE) with complete stereopurity. One-pot alkyne hydrozirconation/transmetallation15 can also be used to access stereodefined E-olefins such as 20, 25 and 26. Despite the presence of low-valent Ni-species, butadiene-containing products such as 21 do not inhibit the reaction, and the E/Z ratio of the starting dienyl species is maintained. Experience from this laboratory has taught that cross-couplings at the D-ring of a steroid can be challenging16; therefore the formation of 19 from the easily obtained alkenylzinc species bodes well for future applications in such contexts. Pioneering work from the Knochel group17 has provided robust methods for generating vinylogous zinc reagents for use in cross-coupling chemistry. These species could also be used in the decarboxylative sense (Fig. 2b) to furnish a range of functionalized building blocks that may be otherwise challenging to access. Both cis and trans alkenylzinc reagents can be prepared with high geometric purity, leading smoothly to 27 and 28. This method provides a complementary strategy to olefin cross-metathesis18 to access such structures. It is also worth noting that the alkenylzinc reagent for butenolide (29) has never been prepared before. Cross-coupling using butenolides, a motif often found in natural products, as a nucleophile has only been accomplished through a Stille-coupling of the corresponding stannylated species16. The adducts with dimedone (30–37) represent an orthogonal pathway to Stork–Danheiser type adducts19 that, in some cases, would not easily be accessed (31, 32, 35). The magnesium bromide diethyl ether complex (MgBr •OEt ) was found to be an essential additive for reactions with alkenylzinc reagents derived from alkenyl iodides via lithium–halogen exchange, direct zinc insertion of electron-withdrawn α,β-unsaturated alkenyl iodides, and hydrozirconation/transmetallation of terminal alkynes. All alkenylzinc reagents were prepared as documented in Supplementary Information. Figure 2c outlines decarboxylative alkenylations using eighteen different primary carboxylic acids, only six of which were not commercially available (41, 52/53, 54, 56, 57 and 58/59) but very easily prepared. In contrast, substantially fewer of these electrophiles are available as a halide or an alcohol (from which a tosylate or halide could be made), whereas some substrates would be unstable if they were obtainable (43–46, 49 and 51). This points again to the undeniable convenience of a cross-coupling that employs readily available starting materials. Amino-acid-derived (38, 42), α-oxy (43–46), benzylic (48, 50, 51), fusidane-based (52, 53), and heterocycle-containing (55) acids could be successfully transformed into olefins of various types. Even peptides (58, 59) with unprotected residues are competent coupling partners under the reaction conditions. Efficient synthesis of naftifine (57), an antifungal pharmaceutical, was accomplished in high yield with excellent selectivity. Amino-acid-containing substrates showed no base-mediated erosion of enantiomeric excess (38, 42) or epimerization (58, 59), and benzylic acid 50 showed no loss of the aromatic bromide. Similarly, other known coupling partners in low-valent Ni-chemistry20, such as aromatic C–O (48, 51) and C–Cl (45) bonds, and hydrolytically sensitive phenol acetate esters (54), remained untouched. Lewis-basic heteroatoms are tolerated (55, 57), and a formal synthesis of (β)-santalene could be accomplished in short order (addition of MeLi to ketone 56 and elimination leads to the natural product)21. Secondary (60–69) and tertiary bridgehead (70) RAEs can be easily alkenylated, as well, as shown in Fig. 2d. Of note here is that diastereoselective alkenylations can be predictably incorporated into synthesis plans, as demonstrated with substrates 64, 65, 68 and 69. Spirocyclic substrates 66 and 67 are of interest to a medicinal chemistry programme (Bristol-Myers Squibb); 67 highlights how ethylvinyl ether (zincated) can be used in this coupling as an alternative to Grignard/organolithium addition to an ester or Weinreb amide (see Supplementary Information for further details and comparison). To probe for the intermediacy of radical species in the decarboxylative alkenylation reaction, radical ring-opening experiments (Fig. 2e) were performed with 71 and 73, affording the corresponding ring-opened products 72 and 74, respectively. Of the 65 substrates depicted in Fig. 2, several were also performed with an in situ protocol (especially in cases where the RAE was not stable) or using an Fe-based system. Supplementary Information contains a detailed troubleshooting guide and a graphical user tutorial. The reaction has been field-tested at Bristol-Myers Squibb in many different contexts, and the robustness was demonstrated by conducting the reaction of 38 on the mole scale (63% yield, >600 g, >99% enantiomeric excess, carried out at Asymchem). As a testament to the robustness of this reaction, no substantial modifications to the general procedure were necessary when scaling this reaction up from the millimole to the mole scale. At present, there are not many obvious limitations for this method given that one can generally expect a serviceable yield across a range of substrates; there are relatively more limitations in the chemistry of preparing the alkenylzinc species. To illustrate a few of the vast number of ways to apply this new transformation, Fig. 3 summarizes the total synthesis of fifteen different natural products (full details of these sequences can be found in Supplementary Information). As mentioned above (Fig. 1a), steroidal substrate 2a exemplifies the inefficiency of previous approaches to olefin synthesis as a consequence of chemoselectivity issues surrounding the Wittig reaction and the incorrect oxidation state of the starting material (1). With decarboxylative alkenylation (Fig. 3a), the same starting material can be used and the desired product 2a can be accessed in two steps. Of note is that the current method can be used to make not only 2a but also related sterol acetates that would be otherwise inaccessible (in a direct fashion) such as 2b and 2c. The clerodane diterpene family consists of over 650 members that are broadly characterized as having a decalin framework appended to side chains with variegated substituents22. From a strategic perspective, it would be ideal if a single starting material could be used and divergently converted to multiple family members using a single reaction type. Decarboxylative alkenylation enables the synthesis of the same three natural isolates from 75, a material that is made in six steps from readily available (−)-5-methyl Wieland–Miescher ketone. Thus, in only nine steps from commercially available materials, a simple alkyne, an iodobutenolide and a bromofuran served as organozinc precursors that when coupled with 75, enabled access to (−)-kolavenol (76a), (−)- solidagolactone (76b), and (−)-annonene (76c), respectively. With decarboxylative alkenylation, one can access methyl trans-chrysanthemate (78, Fig. 2c) from commercially available caronic anhydride (77), which after one-pot methanolysis and radical cross-coupling delivers 78 in 31% yield with excellent diastereoselectivity (>20:1). Many applications of this methodology to polyketide synthesis can be envisaged where the decarboxylative approach allows for innovative uses of classic chiral building blocks in highly convergent ways. For example, tartaric acid is perhaps the cheapest enantiopure chemical that can be purchased (about US$1 per mole) and represents an ideal source of the 1,2-diol motif. The total syntheses of (−)-cladospolide B (3), (−)-iso-cladospolide B (83), and (+)-cladospolide C (4) illustrate how both enantiomers of tartaric acid can be used like simple ‘cassettes’, modularly incorporated to complete syntheses that are not only dramatically shorter than prior approaches but also more selective (Fig. 3d). A design based on radical cross-coupling of tartrate derived acids 79 and 84 sets the stage for a triply convergent approach wherein alkyl–alkyl cross-coupling (with alkylzinc reagent 80) precedes decarboxylative alkenylation (with either 81 or 85) to furnish 82 and 86 in only five steps with excellent control of olefin geometry. If the steps are counted from alkylzinc reagent 80, the sequence is six steps long, indicating that the 1,2-diol motif is no longer the bottleneck of the synthesis. This approach is the most direct and inexpensive known. Similarly, monomethyl succinate (87), a commodity chemical, can be coupled to a stereodefined alkenylzinc reagent to furnish 88 directly, which after deprotection and known macrolactonization, delivers (−)-phoracantholide J (89) in only three steps (eight steps including the synthesis of the alkenylzinc reagent). Prostaglandins are classic targets for total synthesis not only owing to their intriguing structures and exciting medicinal uses but also because they serve as a proving ground for the development of new methodologies (Fig. 3e)23. The commercially available Corey lactone (90) could be used in a four-step sequence wherein two steps are non-strategic (oxidation and one-pot hydrolysis/protection) and two install the key C–C bonds with the proper olefin geometry. Thus, sequential decarboxylative cross-coupling of E-(91) and Z-(92) alkenylzinc species to the requisite carboxylic acids provides a simple route not only to 93 but is also conceivably sufficiently flexible to access many new prostaglandin analogues in a combinatorial fashion. Aureonitol (95) is a tetrahydrofuran-containing natural product discovered in 1979 from Helichrysum aureonitens (Fig. 3f)24. A strategic decarboxylative dienylation of 94 delivers 95 in 32% yield with complete selectivity (>20:1 E/Z) as controlled by the chemistry used to fashion the diene nucleophile. Application of this transformation simplifies the synthesis because 94 can be made in seven simple steps from inexpensive (+)-xylose25, 26. Tocotrienols, members of the vitamin E family, are dietary supplements and have been reported to have an array of beneficial health effects (Fig. 3g)27. Current extraction methods from plant materials provide these compounds as a mixture, which is both difficult and costly to separate. Synthesis offers direct access to specific members of the tocotrienol family but contemporary efforts lack selectivity in olefin formation or require concessionary redox manipulations. Trimethylhydroquinone (96) can be employed to furnish 99 with high selectivity (>20:1 E/Z) and in only four steps overall since the farnesyl group can be directly coupled as a single fragment. Lyngbic acid (102), an inhibitor of quorum sensing in cyanobacteria, was prepared by Noyori asymmetric hydrogenation28 of the commercial β-keto ester 100 followed by decarboxylative cross-coupling delivers 102 with complete selectivity (>20:1 E/Z) in 51% isolated yield (about 98% enantiomeric excess). Olefins are ever-present functional groups that are found naturally and in every sector of chemical science. Their rich and robust chemistry make them integral to the planning, logic and reliable execution of multistep synthesis. This operationally simple method harnesses the reliable and programmable synthesis of olefin-containing zinc species and the unparalleled commercial availability and stability of alkyl carboxylic acids to access olefins in a powerful new way. Numerous applications can be anticipated of both this method and the strategy it enables in the contexts of chemoselective fragment coupling (convergent synthesis), homologation (as an alternative to Wittig olefination and related transforms), and stereospecific olefin installation.
News Article | April 19, 2017
The new method, decarboxylative alkenylation, easily turns carboxylic acids, a relatively cheap, abundant and diverse class of compounds, into alkenes (also called olefins), another large family of compounds commonly used for pharmaceuticals and other applications. It thus should facilitate the discovery and development of a great variety of new drugs and other chemical products. "This method dramatically simplifies the syntheses of olefins; in fact, it has really changed the way I think about making molecules," said principal investigator Phil S. Baran, Darlene Shiley Professor of Chemistry at TSRI. The method, published in this week's edition of Nature, essentially supersedes reactions that have been in chemistry textbooks and in widespread industrial and academic use for decades. Chief among these is the Wittig reaction, discovered in 1954, for which Germany's Georg Wittig was awarded a share of the Nobel Prize for Chemistry in 1979. The Wittig reaction enables the making of many olefins from precursor compounds, and even though it tends to require a many-step process, chemists have continued to rely heavily on it up to the present. "Organic chemists have endured this burdensome 'analog' process for decades with little complaint," Baran said. "Now with this new method we're bringing olefination into the digital era." The new method originated with a breakthrough by Baran and his laboratory that was described a year ago in the Journal of the American Chemical Society. That transformation allowed the building of many complex drug and other molecules starting from carboxylic acids, using inexpensive metal catalysts. Carboxylic acids include numerous bulk-produced chemicals, as well as many abundant natural molecules, among them the amino-acids that cells use to build proteins. "That advance opened the door to a lot of other possibilities, and since then we've been addressing as many of them as we can, starting with the most important," Baran said. The new method enables chemists to turn carboxylic acids into olefins in relatively few steps, using nickel or iron catalysts. In one demonstration, Baran's team produced sterol acetate, a natural olefin with potential as a drug building block, from a standard precursor in two steps—whereas the traditional procedure using the Wittig reaction requires seven steps. To show the vast scope of the new method, Baran's team employed it to make nearly 70 diverse olefins with reactions that were greatly streamlined and simplified compared to traditional methods. "Essentially, with this approach we are taking the most diverse functional group and converting it into the most versatile one," Baran said. The new method allows chemists much better control over the geometry of the resulting molecules, and moreover simplifies the conceptual side of olefin synthesis. "It enables a primitive form of synthesis logic in which you can just start with a cheap compound that is structurally related and convert it with relative ease into an olefin," Baran noted. In one demonstration, the team synthesized the natural antibiotic cladospolide, which is otherwise hard to make, starting from tartaric acid, a bulk-produced carboxylic acid that structurally resembles cladospolide. The paper includes 15 other total or near-total syntheses from cheap starting compounds of natural products that were previously difficult or impractical to synthesize, including prostaglandins, aureonitol, and tocotrienols. The pharmaceutical company Bristol-Myers Squibb, which has a Research Agreement with TSRI, works with the Baran Laboratory under the Agreement and is already using the new method in at least one of its drug development programs. Baran noted that another collaborator, the contract drug-maker Asymchem, has also demonstrated the suitability of the new method for pharmaceutical manufacturing by scaling up a sample reaction to yields on the order of a kilogram. Explore further: New molecule-building method opens vast realm of chemistry for pharma and other industries
Gage J.R.,Asymchem Inc. |
Guo X.,Asymchem Life Science Tianjin Co. |
Tao J.,Asymchem Life Science Tianjin Co. |
Zheng C.,Asymchem Inc.
Organic Process Research and Development | Year: 2012
The design and use of a flow reactor for scaling up an exothermic nitration reaction in a safe manner on production scale is described. The flow reactor is made of a jacketed, stainless steel coil. Two charging methods, pump and nitrogen pressure were tested on a substituted pyridine substrate. The transfer from kilogram scale in the laboratory to 100-kg scale in the plant was successfully accomplished. © 2012 American Chemical Society.
Frantz M.-C.,University of Pittsburgh |
Pierce J.G.,University of Pittsburgh |
Pierce J.M.,University of Pittsburgh |
Kangying L.,Asymchem Inc. |
And 3 more authors.
Organic Letters | Year: 2011
Chemical equations presented. JP4-039 is a novel nitroxide conjugate capable of crossing lipid bilayer membranes and scavenging reactive oxygen species (ROS). An efficient and scalable one-pot hydrozirconation- transmetalation-imine addition methodology has been developed for its asymmetric preparation. Furthermore, this versatile methodology allows for the synthesis of cyclopropyl and fluorinated analogs of the parent lead structure. © 2011 American Chemical Society.