J. Am. Chem. Soc. 1996, 118, 557-579
`
`557
`
`Stereocontrolled Total Synthesis of (+)-Paraherquamide B^
`Timothy D. Cushing,‡ Juan F. Sanz-Cervera,† and Robert M. Williams*
`Contribution from the Department of Chemistry, Colorado State UniVersity,
`Fort Collins, Colorado 80523
`ReceiVed August 7, 1995X
`
`Abstract: The convergent stereocontrolled, asymmetric total synthesis of (+)-paraherquamide B is described. Key
`features of this synthesis include (1) an improved procedure to effect reduction of unprotected oxindoles to indoles;
`(2) a complex application of the Somei/Kametani coupling reaction; (3) a high-yielding and entirely stereocontrolled
`intramolecular SN2¢ cyclization reaction that constructs the core bicyclo[2.2.2] ring system; (4) a mild Pd(II)-mediated
`cyclization reaction that constructs a complex tetrahydrocarbazole; and (5) the chemoselective reduction of a highly
`hindered tertiary lactam in the presence of an unhindered secondary lactam, utilizing precoordination of the more
`reactive secondary lactam to triethylaluminum.
`
`Introduction
`
`The paraherquamides are complex, heptacyclic, toxic mold
`metabolites with potent anthelmintic activity isolated from
`various Penicillium sp. The parent and most potent derivative,
`paraherquamide A (1), was first isolated from Penicillium
`parherquei in 1980 by Yamazaki.1 The simplest member,
`paraherquamide B (2), plus five other structurally related
`paraherquamides C-G (3-9) were isolated from Penicillium
`charlesii (fellutanum) (ATCC 20841) in 1990 at Merck & Co.2,3
`and concomitantly at SmithKline Beecham.4 More recently
`three additional related compounds were discovered by the same
`group at SmithKline.5 Interest in the paraherquamides has come
`from the finding that this class of alkaloids displays potent
`anthelmintic and antinematodal properties.6,7
`There are essentially three classes of broad-spectrum anthel-
`mintics currently in use:
`the benzimidazoles, the levamisoles/
`morantels, and the avermectins/milbemycins. Unfortunately, the
`first two groups have lost much of their utility due to the recent
`appearance of drug resistance built up by the helminths.7a,8 More
`^ Dedicated to Professor Ei-ichi Negishi on the occasion of his 60th
`birthday.
`† On leave from the Department of Organic Chemistry of the University
`of Valencia, Spain.
`‡ Present address: Tularik Inc., 270 East Grand Ave., South San
`Francisco, CA 94080.
`X Abstract published in AdVance ACS Abstracts, December 1, 1995.
`(1) (a) Yamazaki, M.; Fujimoto, H.; Okuyama, E.; Ohta, Y. Proc. Jpn.
`Assoc. Mycotoxicol. 1980, 10, 27. (b) Yamazaki, M.; Okuyama, E.;
`Kobayashi, M.; Inoue, H. Tetrahedron Lett. 1981, 22, 135.
`(2) (a) Blizzard, T. A.; Marino, G.; Mrozik, H.; Fisher, M. H.; Hoogsteen,
`K.; Springer, J. P. J. Org. Chem. 1989, 54, 2657. (b) Blizzard, T. A.; Mrozik,
`H.; Fisher, M. H.; Schaeffer, J. M. J. Org. Chem. 1990, 55, 2256. (c)
`Blizzard, T. A.; Margiatto, G.; Mrozik, H.; Schaeffer, J. M.; Fisher, M. H.
`Tetrahedron Lett. 1991, 32, 2437. (d) Blizzard, T. A.; Margiatto, G.; Mrozik,
`H.; Schaeffer, J. M.; Fisher, M. H. Tetrahedron Lett. 1991, 32, 2441. (e)
`Blizzard, T. A.; Rosegay, A.; Mrozik, H.; Fisher, M. H. J. Labelled Compd.
`Radiopharm. 1990, 28, 461.
`(3) (a) Ondeyka, J. D.; Goegelman, R. T.; Schaeffer, J. M.; Kelemen, L.
`J. Antibiot. 1990, 43, 1375. (b) Liesch, J. M.; Wichman, C. F. J. Antibiot.
`1990, 43, 1380.
`(4) Blanchflower, S. E.; Banks, R. M.; Everett, J. R.; Manger, B. R.;
`Reading, C. J. Antibiot. 1991, 44, 492.
`(5) Blanchflower, S. E.; Banks, R. M.; Everett, J. R.; Reading, C. J.
`Antibiot. 1993, 46, 1355.
`(6) Ostlind, D. A.; Mickle, W. G.; Ewanciw, D. V.; Andriuli, F. J.;
`Campbell, W. C.; Hernandez, S.; Mochaeles, S.; Munguira, E. Res. Vet.
`Sci. 1990, 48, 260.
`(7) (a) Shoop, W. L.; Egerton, J. R.; Eary, C. H.; Suhayda, D. J. Parasitol.
`1990, 76 (2) 186. (b) Shoop, W. L.; Michael, B. F.; Haines, H. W.; Eary,
`C. H. Vet. Parasitol. 1992, 43, 259. (c) Schaeffer, J. M.; Blizzard, T. A.;
`Ondeyka, J.; Goegelman, R.; Sinclair, P. J.; Mrozik, H. Biochem. Pharmacol.
`1992, 43, 679. (d) Shoop, W. L.; Haines, H. W.; Eary, C. H.; Michael, B.
`F. Am. J. Vet. Res. 1992, 53, 2032.
`
`recently drug resistance to the avermectins has been observed
`in various parasites.9 The paraherquamides represent an entirely
`new structural class of antiparasitic agents, which promise to
`play a significant role in the near future. The relatively low
`culture yields of paraherquamide obtained for biological study
`have slowed the development and potential commercialization
`of these agents (Figure 1).
`As part of our ongoing efforts to elucidate the biosynthesis
`of the core bicyclo[2.2.2] ring system of the related alkaloids
`the brevianamides,10 we have applied methodology originally
`developed for the stereocontrolled total synthesis of (-)-
`brevianamide B11 to complete the first stereocontrolled total
`synthesis of (+)-paraherquamide B (12);12 the results of this
`study are described in full herein.
`The paraherquamides are structurally very similar to brevi-
`anamides A and B (17 and 16)13 and marcfortines A-C (13-
`15)14 with respect to the common core bicyclo[2.2.2] ring system
`that is derived from the cycloaddition of an isoprene unit across
`the amino acid R-carbons. This close structural similarity might
`imply a related biogenesis, and the structural features of these
`substances shall be described briefly from this standpoint. The
`paraherquamides and brevianamides A and B (17 and 16) appear
`to be derived from the condensation of tryptophan and proline,
`while the marcfortines are formed from the condensation of
`tryptophan and pipecolic acid. The origin of the methyl group
`in the pyrrolidine ring of paraherquamides A and C-E and
`VM55595-7 could in principle come from the methylation of
`proline, but it seems more likely that this amino acid residue is
`derived from isoleucine. The very low fermentation yield of
`paraherquamide B may be a manifestation of poor incorporation
`of cyclo-L-trp-L-pro into the subsequent biosynthetic machinery
`
`(8) Coles, G. C. Pestic. Sci. 1977, 8, 536.
`(9) (a) Van Wyk, J. A.; Malan, F. S. Vet. Rec. 1988, 123, 226. (b)
`Echevarria, F. A. M.; Trindade, N. P. Vet. Rec. 1989, 124, 147.
`(10) (a) Sanz-Cervera, J. F.; Glinka, T.; Williams, R. M. J. Am. Chem.
`Soc. 1993, 115, 347. (b) Sanz-Cervera, J. F.; Glinka, T.; Williams, R. M.
`Tetrahedron 1993, 49, 8471.
`(11) (a) Williams, R. M.; Glinka, T.; Kwast, E. J. Am. Chem. Soc. 1988,
`110, 5927. (b) Williams, R. M.; Glinka, T.; Kwast, E.; Coffman, H.; Stille,
`J. K. J. Am. Chem. Soc. 1990, 112, 808. (c) Williams, R. M.; Glinka, T.;
`Kwast, E. Tetrahedron Lett. 1989, 30, 5575.
`(12) A preliminary account of this work has appeared: Cushing, T. D.;
`Sanz-Cervera, J. F.; Williams, R. M. J. Am. Chem. Soc. 1993, 115, 9323.
`(13) (a) Birch, A. J.; Wright, J. J. J. Chem. Soc., Chem. Commun. 1969,
`644. (b) Birch, A. J.; Wright, J. J.; Tetrahedron 1970, 26, 2339. (c) Birch,
`A. J.; Russell, R. A. Tetrahedron 1972, 28, 2999. (d) Baldas J.; Birch, A.
`J.; Russell, R. A. J. Chem. Soc., Perkin Trans. 1 1974, 50.
`(14) (a) Polonsky, J.; Merrien, M. A.; Prange, T; Pascard, C. J. Chem
`Soc., Chem Commun. 1980, 601. (b) Prange, T.; Billion, M.-A.; Vuilhorgne,
`M.; Pascard, C.; Polonsky, J. Tetrahedron Lett. 1981, 22, 1977.
`
`0002-7863/96/1518-0557$12.00/0 © 1996 American Chemical Society
`
`
` P. 1
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`558 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
`
`Cushing et al.
`
`Figure 1.
`
`or may be the result of inefficient demethylation of the
`isoleucine-derived amino acid precursor.
`The oxidation state of the amino acid-derived dioxopiperazine
`moiety remains unchanged in the case of the brevianamides,
`but for the paraherquamides and the marcfortines the tertiary
`amide residue is enzymatically reduced to a monooxopiperazine,
`a process that is known.15 The tryptophan-derived indolyl side
`chain of the paraherquamides and marcfortines is oxidized to
`spiro-oxindoles while the indolyl side chain of the brevianamides
`oxidize to spiro-indoxyls. The paraherquamides, marcfortines,
`and brevianamides all incorporate one isoprene unit that forms
`the bridging bicyclo[2.2.2] ring structure. The paraherquamides
`and marcfortines differ from the brevianamides in that a second
`isoprene unit coupled with an oxidized form of tryptophan gives
`the dioxepin (or pyran) moiety. This is one of the most
`interesting and unique features of these compounds. The gem-
`dimethyl dioxepin ring found in paraherquamides A-E (1-5)
`and marcfortines A and B (13 and 14) is a unique ring system
`among natural products. A similar structural feature was
`discovered in the antifungal natural product strobilurin G (18),16
`but this dioxepin moiety lacks the double bond found in the
`other metabolites (Figure 1).
`As outlined in Scheme 1, a convergent synthesis of the
`enantiomer of paraherquamide B (12)17 was envisioned to
`contain four key carbon-carbon bond-forming reactions. The
`
`(15) Bond, R. F.; Boeyens, J. C. A.; Holzapfel, C. V.; Steyn, P. S. J.
`Chem. Soc., Perkin Trans. 1 1979, 1751.
`(16) Fredenhagen, A.; Hug, P.; Peter, H. H. J. Antibiot. 1990, 48, 661.
`(17) The enantiomer of the natural product was selected as the target
`due to the large relative cost difference between (S)- and (R)-proline.
`
`first task would involve the construction of a suitably R-alkylated
`proline derivative.11 The second important coupling would be
`the Somei/Kametani-type alkylation18 of a suitably protected
`(20)
`gramine derivative
`and the
`requisite piperazine-
`dione (19). The third and perhaps most crucial C-C bond-
`forming reaction, providing the core bicyclo[2.2.2] ring system,
`was a stereofacially controlled intramolecular SN2¢ cyclization
`reaction that sets the stereochemistry at C-20 (paraherquamide
`numbering) and concomitantly installs the isopropenyl group
`that will be utilized in the fourth C-C bond-forming reaction.
`This isopropenyl group, in turn, would be conscripted for an
`olefin-cation cyclization to provide the heptacyclic tetrahy-
`drocarbazole. Standard procedures to effect this transformation
`involve strong protic acids,11,19 and there was reason for concern
`about the reactivity of the more highly oxygenated indole (22)
`as a practical synthetic precursor to 23. The penultimate step,
`a regio- and stereofacially controlled oxidative spirocyclization
`reaction, must be accomplished to construct the desired spiro-
`oxindole. A number of these transformations were explored
`during the course of the investigations on the synthesis of (-)-
`brevianamide B,11 including a simple oxindole model study,11c
`which set a firm foundation for addressing some of the
`
`(18) (a) Somei, M.; Karasawa, Y.; Kaneko, C. Heterocycles 1981, 16,
`941. (b) Kametani, T.; Kanaya, N.; Ihara, M. J. Am. Chem. Soc. 1980, 102,
`3974.
`(19) (a) Stoermer, D.; Heathcock, C. H. J. Org. Chem. 1993, 58, 564.
`(b) Guller, R.; Dobler, M.; Borschberg, H.-J. HelV. Chim. Acta 1991, 74,
`1636. (c) Darbre, T.; Nussbaumer, C.; Borschberg, H.-J. HelV. Chim. Acta
`1984, 67, 1040. (d) Delpech, B.; Khuong-Huu, Q. J. Org. Chem. 1978, 43,
`4898.
`
`
` P. 2
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`Stereocontrolled Total Synthesis of (+)-Paraherquamide B
`
`J. Am. Chem. Soc., Vol. 118, No. 3, 1996 559
`
`Scheme 1
`
`Scheme 2
`
`stereochemical and regiochemical issues that would be faced
`in attacking the paraherquamide ring system.
`
`Results and Discussion
`Construction of the Dioxepinooxindole Ring System. The
`prenylated catechol ring system of the paraherquamides is an
`unusual oxidative cyclization product that previously has not
`been observed to occur in metabolites of mixed biogenetic
`origin. Although the parent 2H-1,5-benzodioxepin has been
`synthesized previously,20 to the best of our knowledge there
`has been no reported synthesis of the corresponding isoprenyl
`dioxepin ring system of paraherquamide. The synthesis of this
`ring system was explored in a simple model study employing
`prenylated catechol 24 (Scheme 2).21
`It was speculated that
`the requisite 7-endo-tet cyclization reaction would be facilitated
`by a stabilized tertiary carbocation provided by the prenyl
`substituent.
`The first attempt at effecting this cyclization reaction
`
`(20) Guillaumet, G.; Coudert, G.; Loubinnoux, B. Angew. Chem., Int.
`Ed. Engl. 1983, 22, 64.
`
`(21) Williams, R. M.; Cushing, T. D. Tetrahedron Lett. 1990, 31, 6325.
`
`employed a phenylselenoetherification.22 Following a procedure
`of Clive,23 24 cyclized to 25 with either PhSeCl or N-
`phenylselenophthalimide (N-PSP),24 although in very low yield.
`The main byproducts came from the electrophilic addition across
`the double bond, electrophilic aromatic substitution of the phenyl
`ring by the phenyl selenide, and phenolic attack at the methylene
`producing the six-membered-ring product. The selenide 25 was
`treated with H2O2 and the resulting selenoxide thermally
`eliminated providing the unique dioxepin 26 in 49% yield.
`Due to the low yield of the phenylselenoetherification, an
`alternative procedure involving epoxidation followed by a Lewis
`acid-mediated ring closure was investigated.25 The prenylated
`catechol 24 was epoxidized with buffered m-CPBA to provide
`epoxide 27, which was treated with stannic chloride to give the
`dioxepin 28. A major side product in this reaction was a ketone,
`
`(22) (a) Nicolaou, K. C. Tetrahedron 1981, 37, 4097. (b) Nicolaou, K.
`C.; Magolda, R. L.; Sipio, W. J.; Barnette, W. E.; Lysenko, Z.; Joullie, M.
`M. J. Am. Chem. Soc. 1980, 102, 3784. (c) Clive, D. L. J. Tetrahedron
`1978, 34, 1049-1132.
`(23) (a) Clive, D. J. L.; Chiiattu, G.; Curtis, N. J.; Kiel, W. A.; Wong,
`C. K. J. Chem. Soc., Chem. Commun. 1977, 725. See also: (b) Liotta, D.;
`Zima, G. Tetrahedron Lett. 1978, 50, 4977. (c) Tiecco, M.; Testaferri, L.;
`Tingoli, M.; Bartoli, D.; Balducci, R. J. Org. Chem. 1990, 55, 429.
`(24) Nicolaou, K. C.; Claremon, D. A.; Barnette, W. E.; Seitz, S. P. J.
`Am. Chem. Soc. 1979, 101, 3704.
`(25) (a) Cookson, R. C.; Liverton, N. J. J. Chem. Soc., Perkin Trans. 1
`1985, 1589. (b) Kocienski, P.; Love, C.; Whitby, R.; Roberts, D. A.
`Tetrahedron Lett. 1988, 29, 2867. See also: (c) Nicolaou, K. C.; Prasad,
`C. V. C.; Somers, P. K.; Hwang, C.-K. J. Am. Chem. Soc. 1989, 111, 5335.
`
`
` P. 3
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`560 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
`
`Cushing et al.
`
`Scheme 3a
`
`a Reagents and conditions: (a) 4.0 equiv of NaOH, 1.0 equiv of 30% H2O2, 81-93%; (b) H2, Pd/C, AcOH, 92%; (c) 2.5 equiv of BBr3, CH2Cl2,
`-78 (cid:176)C, 99%; (d) 1.2 equiv of prenyl bromide, 1.1 equiv of K 2CO3, DMF, 0 (cid:176)C to room temperature, 52%; (e) m-CPBA, NaHCO3, CH2Cl2; (f) 1.2
`equiv of SnCl4, THF, 64%; (g) 1.6 equiv of NaBH4, 3.5 equiv of BF3(cid:226)OEt2, THF, 44-50%; (h) t-BuMe2SiCl, im, DMF, 40 (cid:176)C, 83%; (i) CH 2O,
`HNMe2, AcOH, H2O, 99%.
`
`resulting from a 1,2 hydride shift.26 A number of methods were
`explored to effect the dehydration of the secondary alcohol of
`dioxepin 28; the best result was realized with methyltriphenoxy-
`phosphonium iodide (MTPI) in HMPA to provide 26.27
`With a proven method accessible for the construction of the
`dioxepin ring system, attention was focused on constructing the
`requisite gramine derivative. Oxygenated indoles are notori-
`ously unstable and undergo facile autoxidation,28 photooxida-
`tion,29 dimerization, and polymerization.30
`In light of this
`problematic reactivity, our plan called for formation of the
`dioxepin ring system prior to indole (gramine) formation. The
`approach employed involved the formation of a suitably
`substituted oxindole (essentially a protected indole), followed
`by the construction of the dioxepin and final elaboration into
`the gramine derivative.
`The known pyruvic acid 29 (Scheme 3)31 (prepared in five
`steps from vanillin) was oxidatively decarboxylated32 to afford
`the phenylacetic acid 30, which was reductively cyclized to give
`the required oxindole 3133 in nearly quantitative yield.
`At this point, a method was needed to differentiate between
`the two phenolic substituents for the prenylation reaction. A
`number of attempted selective protecting group strategies were
`
`(26) For a related observation, see: Taylor, S. T.; Davisson, M. E.;
`Hissom, B. R., Jr.; Brown, S. J.; Pristach, H. A.; Schramm, S. B.; Harvey,
`S. M. J. Org. Chem. 1987, 52, 425.
`(27) Hutchins, R. O.; Hutchins, M. G.; Milewski, C. A. J. Org. Chem.
`1972, 37, 4191.
`(28) Houlihan, W. J.; Remers, W. A.; Brown, R. K. Indoles, Part one,
`The Chemistry of Heterocycles; John Wiley & Sons, Inc.: New York, 1972.
`(29) (a) Chan, A. C.; Hilliard, P. R., Jr. Tetrahedron Lett. 1989, 30, 6483.
`(b) d’Ischia, M.; Prota, G. Tetrahedron 1987, 43, 431.
`(30) This difficulty was observed in a short synthesis of the known
`6-acetoxy-7-methoxyindole (i). The unstable substance i was treated with
`TMSI, producing the dimer ii as the sole product.
`
`See: (a) Walker, G. N. J. Am. Chem. Soc. 1955, 77, 3844. (b) Burton, H.;
`Duffield, J. A.; Prail, P. F. G. J. Chem. Soc. 1950, 1062. (c) Beer, R. J. S.;
`Mcgrath, L.; Robertson, A.; Woodier, A. B. J. Chem. Soc. 1949, 2061. (d)
`Beer, R. J S.; Clarke, K.; Khorana, H. G.; Robertson, A. J. Chem. Soc.
`1948, 2223. (e) Chan, A. C.; Hilliard, P. R., Jr. Tetrahedron Lett. 1989, 30,
`6483. (f) d’Ischia, M.; Prota, G. Tetrahedron 1987, 43, 431. (g) Deibel, R.
`M. B.; Chedekel, M. R. J. Am. Chem. Soc. 1984, 106, 5884. (h) Heacock,
`R. A. Chem. ReV. 1959, 59, 181.
`(31) (a) Beer, R. J. S.; Clarke, K.; Davenport, H. F.; Robertson, A. J.
`Chem. Soc. 1951, 2029. (b) Bennington, F.; Morin, R. D.; Clark, L. C., Jr.
`J. Org. Chem. 1959, 24, 917.
`(32) Kosuge, T.; Ishida, H.; Inabe, A.; Nukaya, H. Chem. Pharm. Bull.
`1985, 33, 1414.
`
`explored, but nothing satisfactory was found; it was thus decided
`to forgo any protecting group for the 6-hydroxy position.
`Oxindole 31 was cleanly demethylated upon treatment with
`(clear) boron tribromide. The resulting oxindole 32 was
`subjected to the prenylation conditions, and the desired alkylated
`product 33 was obtained in 52% yield.34,35 Both of the methods
`discussed above for the formation of the seven-membered ring
`were examined. The phenylselenoetherification procedure failed
`on this substrate, and only products resulting from electrophilic
`aromatic substitution were formed.
`The alternative epoxidation/Lewis acid-mediated cyclization
`again proved to be successful on this substrate. The epoxidation
`reaction (m-CPBA) had to be buffered with NaHCO3, to prevent
`the formation of the six-membered-ring tertiary alcohol. In most
`cases, the reaction was worked up and taken on to the next
`step without purification (the labile epoxide tended to cyclize
`to the six-membered tertiary alcohol upon contact with silica
`gel). The incipient epoxide product was directly treated with
`SnCl4 in THF to provide the desired seven-membered-ring
`alcohol 34 (64% overall yield from 33).
`N-Alkylated oxindoles have been reported to be reduced to
`indoles by the use of DIBAL or LiAlH4;36 however, in the case
`of unsubstituted oxindoles, this reduction either fails or requires
`
`(33) This material has interesting chemical and physical characteristics.
`The solvent must be removed immediately after the hydrogenolysis to
`prevent the white product from turning to a black sludge. This oxindole 31
`would also change from a white color to a metallic gray simply by drying
`on the vacuum pump. These decomposition properties are no doubt due to
`the autoxidation of the indole tautomer form.
`(34) The undesired regioisomer was obtained in less than 1% yield, and
`the bis-alkylated material was produced in only 8.3% yield. This selectivity
`is presumably a manifestation of the domination of inductive effects of the
`amide functionality directing the alkylation to the 7-position.
`(35) The structure of compound 33 was confirmed by simply tosylating
`33 and comparing the product (37) to the same substance prepared from
`31. The two independently synthesized products were identical in every
`way.
`
`(36) (a) Kishi, Y.; Nakatsuka, S.; Fukuyama, T.; Havel, M. J. Am. Chem.
`Soc. 1973, 95, 6494. (b) Robinson, B. Chem. ReV. 1969, 69, 785.
`
`
` P. 4
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`Stereocontrolled Total Synthesis of (+)-Paraherquamide B
`
`J. Am. Chem. Soc., Vol. 118, No. 3, 1996 561
`
`Scheme 4a
`
`a Reagents and conditions: (a) 36, 0.5 equiv of PBu3, MeCN, 51%; (b) DMAP, Et3N, BOC2O, CH2Cl2, 90%; (c) 5 equiv of LiCl, 1.5 equiv of
`H2O, HMPA, 100 (cid:176)C, 66%; (d) 3.0 equiv of n-Bu4NF, THF, 79%; (e) 1.9 equiv of LiCl, 4.0 equiv of collidine, 4.0 equiv of MsCl, DMF, 86%; (f)
`t-BuMe2SiOTf, 2,6-lutidine,CH2Cl2, 76%; (g) 10 equiv of NaH, benzene, 11%.
`
`Figure 2.
`
`In 1972 it was reported37 that
`more vigorous conditions.
`substituted and unsubstituted oxindoles could be reduced to the
`corresponding indole in high yields with borane in THF at 0
`(cid:176) C. Oxindole 34 was subjected to these conditions (1.0 M BH3/
`THF, Aldrich), but with no reaction. However, when oxindole
`34 was treated with 1.6 equiv of NaBH4 and 3.5 equiv of BF3(cid:226)-
`OEt2 in THF for 1 day (0 (cid:176)C to room temperature), the desired
`indole 35 was obtained in 43-50% yield. The indole 35 was
`treated with a warm solution of TBDMSCl and imidazole in
`DMF, to provide the required O-silylated indole, which was
`easily converted to the gramine 36 through the well-known
`Mannich procedure (Scheme 3).
`Construction of the Bicyclo[2.2.2] Ring System. To probe
`the stability of the dioxepin-indole in subsequent transforma-
`tions, a model study involving the previously synthesized
`racemic piperazinedione 3838 was investigated (Scheme 4).
`Indole 36 was condensed with the piperazinedione 38 following
`the Somei/Kametani conditions18 to give the desired syn product
`39 (a racemic mixture of two diastereomers) in 51% yield. The
`relative stereochemistry of this substance was evident by an
`examination of the 1H NMR spectrum. There is a large upfield
`shift of the proline ring protons of 39 ((cid:228) Ha, Hb, Hc; 0.03-
`0.19 (m), 0.43-0.52 (m), 0.62-0.72 (m) ppm).
`It is well-
`known that N-alkylated piperazinediones prefer to adopt a boat-
`like conformation due to the planar geometry of the amides and
`A-1,3 steric interactions of N-alkyl residues. This forces the
`
`substituents on the amino acid R-carbons to adopt either
`pseudoaxial or pseudoequatorial dispositions. In conformer B
`(Figure 2) the carbomethoxy group is sterically congested by
`the bulky isopentenyl group,
`favoring the alternate boat
`conformer (A), which positions the indole ring under the
`piperazinedione, positioning the two pyrrolidine protons Ha and
`Hb directly over the shielding cone of the aromatic indole ring
`system; the corresponding anti-isomer cannot adopt this type
`of conformation. Furthermore, a consideration of the mecha-
`nism of the Somei/Kametani reaction18 supports the expectation
`that the syn-isomer (39) should be the major product. The
`gramine derivative (36) reacts with tributylphosphine to form
`a bulky (tributylphosphino)indole intermediate that can only
`approach from the less congested face of the piperazinedione
`enolate, away from the isopentenyl moiety.
`A similar phenomenon was observed when 39 was subjected
`to the decarbomethoxylation procedure (LiCl, H2O, HMPA)
`directly. The two main products isolated were the syn-isomer
`40 and the anti-isomer 41, in a ratio of 3.3:1.0 (Figure 3). These
`stereochemical assignments were made by comparing the 1H
`NMR spectral data of 40 and 41. There was a pronounced
`upfield shift of three pyrrolidine ring protons in compound 41,
`a shift that is not observed for diastereomer 40.
`Piperazinedione 39 was first converted to the BOC-protected
`indole 42, which was subsequently subjected to a decar-
`bomethoxylation reaction supplying the syn-diastereomer 43 as
`
`(37) Sirowej, H.; Khan, S. A.; Plieninger, H. Synthesis 1972, 84.
`
`(38) Williams, R. M.; Glinka, T. Tetrahedron Lett. 1986, 27, 3581.
`
`
` P. 5
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`562 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
`
`Cushing et al.
`
`Figure 3.
`
`Scheme 5a
`
`a Reagents and conditions: (a) 3.8 equiv of CAN (0.33 M), 2:1 CH3CN/H2O, 2 h, 79%; (b) (i) 2 equiv of NaBH4, EtOH; (ii) t-BuPh2SiCl, im,
`DMF, 75%; (c) (i) 1.0 equiv of n-BuLi, 1.1 equiv of MeOCOCl, -78 (cid:176)C; (ii) 2.2 equiv of LiN(SiMe 3)2, 1.1 equiv of MeOCOCl, THF, -100 (cid:176)C,
`93%; (d) 36, 0.5 equiv of PBu3, CH3CN, reflux, 73%; (e) LiCl, HMPA, 100 (cid:176)C ( syn/anti 3:1), 89%; (f) Me3OBF4, Na2CO3, CH2Cl2 (syn, 81%; anti,
`62-71%); (g) (i) BOC2O, DMAP, Et3N, CH2Cl2; (ii) n-Bu4NF, THF (syn, 90%; anti, 85%); (h) NCS, Me2S (syn, 74-81%; anti, 86%).
`
`the major product. Compound 43 (the minor, anti-diastereomer
`was not utilized) was desilylated to provide the diol 44, which
`was converted to the allylic chloride 45. Careful treatment of
`45 with t-BuMe2SiOTf, to prevent transesterification of the BOC
`groups,39 gave the desired product 46 in 76% yield. Allylic
`chloride 46 was subjected to 10 equiv of NaH in refluxing
`benzene, but the reaction proved extremely sluggish. After 5
`days, the desired product 47 was obtained in a poor 11% yield
`(19% based on recovered 46; accompanied by extensive
`decomposition). The syn-isomer 47 was the only stereoisomer
`formed in this reaction; the corresponding anti-isomer was not
`detected. While this reaction demonstrated the potential viability
`of the stereoselective intramolecular SN2¢ reaction, work on the
`racemic model system was halted, due to the low yield in this
`
`(39) Sakaitani, M.; Ohfune, Y. J. Am. Chem. Soc. 1990, 112, 1150.
`
`key transformation coupled with perceived difficulties associated
`with removing the N-p-methoxybenzyl group.
`Total Synthesis of (+)-Paraherquamide B. Starting from
`the known piperazinedione 48 (prepared in eight steps from (S)-
`proline),11 the enal 49 was obtained in 79% yield by treatment
`of 48 with a 0.33 M solution of ceric ammonium nitrate (Scheme
`5).40 The resulting product (49) was reduced with NaBH4 and
`protected with t-BuPh2SiCl in a two-step process to give the
`silyl ether 50 in 75% yield. Compound 50 was then subjected
`to a two-step, one-pot acylation providing the required substrate
`51 in 93% yield. The crude material was a mixture of epimers
`in a ratio of approximately 4:1 (syn:anti).
`Interestingly this
`mixture had a tendency to epimerize during column chroma-
`
`(40) Yoshimura, J.; Yamaura, M.; Suzuki, T.; Hashimoto, H. Chem. Lett.
`1983, 1001.
`
`
` P. 6
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`Stereocontrolled Total Synthesis of (+)-Paraherquamide B
`
`J. Am. Chem. Soc., Vol. 118, No. 3, 1996 563
`
`Figure 4.
`
`Scheme 6
`
`tography, resulting in an increase in the proportion of the syn-
`isomer. The two products were combined and condensed with
`the gramine 36 providing the indole 52 in 73% yield as a mixture
`of two diastereomers (epimeric at
`the secondary alcohol
`stereogenic center). Interestingly, the imidic carbamate group
`was also cleaved in the course of this reaction. Compound 52
`was subjected to the decarbomethoxylation procedure, affording
`a 3:1 mixture of 53 (syn) and 54 (anti) in 89% combined yield.
`The lactam 53 could be converted to the N-BOC-protected
`allylic chloride 55 in four steps and in good overall yield
`(Scheme 6), but numerous attempts to effect the SN2¢ reaction
`on this substrate failed. These reactions were capricious and
`were accompanied by the occasional appearance of the spiro-
`lactones 56 and 57, formed in low yield <5% (Figure 4).
`It
`seems likely that the failure of 55 to cyclize in the desired
`fashion can be attributed to nonbonding interactions between
`the tert-butoxycarbonyl group and the pendant dioxepin in-
`dole.41,42
`These observations dictated that a suitable amide protecting
`group would have to be selected that was less electron
`withdrawing and less sterically demanding than both the tert-
`butoxycarbonyl and the p-methoxybenzyl groups. The loss of
`the lactam methoxycarbonyl group in the alkylation of 51 with
`the gramine 36 was presumably due to N f N acyl transfer to
`dimethylamine, a byproduct of the Somei/Kametani reaction.
`This appears to be a general reaction that was used to selectively
`deprotect the N-tert-butoxycarbonyl group of 58 without de-
`blocking the N-tert-BOC-protected indole. Thus, refluxing a
`
`(41) The formation of the two spiro compounds 56 and 57 is presumably
`due to the increased electrophilicity of the N-acylated amide. Apparently,
`trace moisture in the reaction mixture caused the production of hydroxide,
`which then hydrolyzed the reactive amide bond. The resulting carboxylic
`acid cyclized in an SN2¢ fashion, furnishing the spiro lactones.
`(42) (a) Giovannini, A.; Savoia, D.; Umani-Ronchi, A. J. Org. Chem.
`1989, 54, 228. (b) Flynn, D. L.; Zelle, R. E.; Grieco, P. A. J. Org. Chem.
`1983, 48, 2424.
`
`solution of 58 and Me2NH in CH3CN furnished compound 59
`in 92% yield43 (Scheme 7).
`The strategy planned for the reduction of the tertiary amide
`called for the protection of the secondary lactam as a lactim
`ether,44 and this group seemed suitable for use earlier in the
`synthesis and appeared compatible with the SN2¢ cyclization.
`Thus, syn-isomer 53 was treated with 20 equiv (optimum) of
`Na2CO3 and 5 equiv of Me3OBF4 in dichloromethane for 4 h,
`to yield 81% of compound 60. Even though the next two
`reactions could be carried out in a stepwise fashion, it proved
`most convenient to convert 60 directly to the protected diol 62
`in a one-pot, two-step sequence. Diol 62 was then subjected
`to the chlorination procedure successfully used in the conversion
`of diol 44 to the allylic chloride 45. Unfortunately, under these
`conditions, the reaction failed and the lactim ether was cleanly
`deblocked back to the lactam. This problem was finally solved
`by following the procedure of Corey,45 which called for the
`addition of compound 62 to a mixture of N-chlorosuccinimide
`and dimethyl sulfide, which yielded the chloride 64 in 81%
`yield.
`Allylic chloride 64 was reprotected with t-BuPh2SiOTf to
`provide 66 in 77-82% yield. The stage was now set to effect
`the SN2¢ reaction. Compound 66 was refluxed in benzene with
`20 equiv of sodium hydride, resulting in a very clean and high-
`yielding cyclization reaction furnishing the desired product 68
`in 93% yield (Scheme 8).
`(43) This result is noteworthy, especially in light of a report that tert-
`butoxycarbonyl-protected amides are cleaved to the tert-butoxycarbonyl-
`protected amines with DEAEA (2-(N,N-diethylamino)ethylamine) in CH3CN
`at room temperature; see: Grehn, L.; Gummarsonn, K.; Ragnarsson, U.
`Acta Chem. Scand. B 1987, 41, 18. However, the substrates examined in
`that report were all open-chain amides. Interestingly it is known that BOC-
`protected lactams can be cleaved by base but it is the amide bond that is
`broken as was observed on substrate 55. Recently it has been reported that
`Mg(OMe)2 will also cleave lactam carbamates including BOC-protected
`lactams; see: Wei, Z.-Y.; Knaus, E. E.; Tetrahedron Lett. 1994, 35, 847.
`(44) Williams, R. M.; Brunner, E. J.; Sabol, M. R. Synthesis 1988, 963.
`(45) Corey, E. J.; Kim, C. U.; Takeda, M. Tetrahedron Lett. 1972, 13,
`4339.
`
`
` P. 7
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`564 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
`
`Cushing et al.
`
`Scheme 7
`
`Scheme 8
`
`This last series of reactions was also carried out in parallel
`on the anti-isomer 54. Following the same sequence (five steps)
`we obtained the fully protected chloride 67 in good yield. The
`chloride 67 was then refluxed in benzene with the required
`amount of sodium hydride to yield the same product (68, 85%
`
`yield) as that obtained from 66. The yields of 68 from both
`routes were very high, and the undesired anti-diastereomer was
`not detected. The high degree of facial selectivity observed in
`the cyclizations to 68 and 47 is quite interesting and warrants
`It is generally accepted that SN2¢ reactions
`some comments.
`
`
` P. 8
`
`UT Ex. 2027
`SteadyMed v. United Therapeutics
`IPR2016-00006
`
`

`
`Stereocontrolled Total Synthesis of (+)-Paraherquamide B
`
`J. Am. Chem. Soc.