Total synthesis of the serine-threonine phosphatase inhibitor microcystin-LA

Article abstract of DOI:10.1021/ja961683e

Reversible protein phosphorylation, which is mediated by kinases and phosphatases, is a major control element of the cell. There is a diverse group of toxic natural products that inhibit certain phosphatases, thereby disrupting normal biochemical pathways

Full text of DOI:10.1021/ja961683e

J. Am. Chem. Soc. 1996, 118, 11759-11770
11759
Total Synthesis of the Serine-Threonine Phosphatase Inhibitor
Microcystin-LA
John M. Humphrey, James B. Aggen, and A. Richard Chamberlin*
Contribution from the Department of Chemistry, UniVersity of California,
IrVine, California 92697
ReceiVed May 20, 1996X
Abstract: Reversible protein phosphorylation, which is mediated by kinases and phosphatases, is a major control
element of the cell. There is a diverse group of toxic natural products that inhibit certain phosphatases, thereby
disrupting normal biochemical pathways. These toxins can be useful for dissecting the individual biochemical pathways
associated with each of these enzymes. This Article describes the first total synthesis of one such toxin, the cyclic
heptapeptide microcystin-LA. The synthesis features a convergent route that is amenable to analog preparation in
the search for selective phosphatase inhibitors. A new route to the unusual amino acid Adda is described, which
incorporates an efficient diastereoselective aspartate alkylation and diene synthesis via a Suzuki coupling reaction.
This work also features an efficient preparation of an N-methylalanine containing peptide via a Horner-Emmons
condensation and several difficult amino acid coupling reactions that relied heavily on Carpino’s remarkable HATU
reagent.
Introduction
There is a diverse group of structurally interesting microbial
defense toxins that act by inhibiting certain phosphatases,
thereby disrupting the normal biochemical pathways. Examples
include okadaic acid,10 tautomycin,11 calyculin,12 and the
microcystins,13 all of which are potent competitive inhibitors
of two major classes of phosphatases, protein phosphatase 1
(PP1), and protein phosphatase 2A (PP2A).14 Several endog-
enous phosphoproteins, e.g., phospho-DARP 32, are substrates
of PP2A but are selective inhibitors of PP1.3,4 In contrast, the
toxins inhibit both PP1 and PP2A, generally with little selectiv-
ity. The exceptions are cantharidin,15 thyrsiferyl 23-acetate,16
and okadaic acid,9 which are PP2A selective. As a result of
this selectivity, these toxins can be useful for dissecting the
individual biochemical pathways associated with each of these
enzymes, thereby providing a clearer picture of phosphatase
control in some systems.9 Further information could be gained
through the similar use of a PP1-selective inhibitor, but
compounds suitable for studies of this type are not yet known.
The key to discovering such a PP1 selective inhibitor may lie
in the chemical modification of existing nonselective inhibitors,
and as the first step toward this goal we report the first total
synthesis of a microcystin.17,18
There is considerable interest in the elucidation of signaling
pathways1 in the central nervous system, both between neurons,
via neurotransmitters, and within them. The elements respon-
sible for transmitting signals from membrane bound receptors
to other destinations within the neuron comprise a complex
system that is only partially understood. One important aspect
of this system is the reversible phosphorylation of proteins,
which is mediated by kinases (phosphorylation catalysts)2 and
phosphatases (dephosphorylation catalysts)3 that are indirectly
linked, through modulation by second messengers such as cAMP
or Ca2+, to the activation of dopamine and N-methyl-D-aspartate
(NMDA) receptors.4 Reversible protein phosphorylation is a
major control element of neurons and other cells, moderating
such diverse functions as neurotransmission, muscle contraction,
glycogen synthesis, T-cell activation, and cell proliferation.5-9
Excessive protein phosphorylation, whether through the activa-
tion of kinases or through the inhibition of phosphatases, can
lead to uncontrolled cellular proliferation, suggesting an active
role for the phosphatases in tumor suppression.5
X Abstract published in AdVance ACS Abstracts, November 15, 1996.
(1) For reviews on signaling pathways, see: (a) Rosen, M. K.; Schreiber,
S. L. Angew. Chem., Int. Ed. Engl. 1992, 31, 384. (b) Schreiber, S. L.; Liu,
J.; Albers, M. W.; Karmacharya, R.; Koh, E.; Martin, P. K.; Rosen, M. K.;
Standaert, R. F.; Wandless, T. J. Transplant. Proc. 1991, 23, 2839. (c)
Berridge, M. Bioessays 1995, 17, 491. (d) Magee, A. I. Cellular Signal.
1995, 165. (e) Langdon, S. P.; Smyth, J. F. Cancer Treat. ReV. 1995, 21,
65. (f) Edwards, D. R. TIPS 1994, 15, 239. (g) Elmoatassim, C.; Dornand,
J.; Mani, J. C. Biochim. Biophys. Acta 1992, 1134, 31. (h) Sun, H.; Tonks,
N. K. TIBS 1994, 19, 480.
(9) Fischer, E. Angew. Chem., Int. Ed. Engl. 1993, 32, 1130.
(10) For a recent review, see: Cohen, P.; Holmes, C. F. B.; Tsukitani,
Y. TIBS 1990, 15, 98.
(11) Ubukata, M.; Cheng, X. C.; Isono, K. J. Chem. Soc., Chem.
Commun. 1990, 244.
(12) Evans, D. A.; Gage, J. R.; Leighton, J. L. J. Am. Chem. Soc. 1992,
114, 9434, and references cited therein.
(13) (a) Botes, D.; Tuinman, A.; Wessels, P.; Viljoen, C.; Kruger, H.;
Williams, D. H.; Santikarn, S.; Smith, R.; Hammond, S. J. Chem. Soc.,
Perkin Trans. 1 1984, 2311. (b) Painuly, P.; Perez, R.; Fukai, T.; Shimizu,
Y. Tetrahedron Lett. 1988, 29, 11. (c) Watanabe, M. F.; Oishi, S.; Harada,
K.; Matsuura, K.; Kawai, H.; Suzuki, M. Toxicon 1988, 26, 1017.
(14) Takai, A.; Sasaki, K.; Nagai, H.; Mieskes, G.; Isobe, M.; Isons, K.;
Yasumoto, T. Biochem. J. 1995, 306, 657.
(15) Li, Y. M.; Casida, J. E. Proc. Natl. Acad. Sci. U.S.A. 1992, 89,
11867.
(16) Matsuzawa, S.; Suzuki, T.; Suzuki, M.; Matsuda, A. FEBS Lett.
1994, 356, 272.
(2) For reviews on kinases, see: (a) Serota-Taylor, S.; Radzio-Andzelm,
E. Structure 1994, 2, 345. (b) Murray, A. W. Chem. Biol. 1994, 1, 191.
(3) For a review on protein phosphatases, see: Cohen, P. Annu. ReV.
Biochem. 1989, 58, 453.
(4) (a) Greengard, P.; Browning, M. D. AdV. Second Messenger
Phosphoprotein Res. 1988, 21, 133. (b)Hemmings, H. C. J.; Nairn, A. C.;
Elliott, J. I.; Greengard, P. J. Biol. Chem. 1990, 265, 20369.
(5) For a review on the biological roles of protein phosphatases, see:
Cohen, P.; Cohen, P. T. W. J. Biol. Chem. 1989, 264, 21435.
(6) Cohen, P. The Structure and Regulation of Protein Phosphatases;
Raven Press: New York, 1990; pp 230-235.
(17) The synthesis of an Mdha containing segment of microcystin has
been reported: Zetterstro¨m, M.; Trogen, L.; Hammarstro¨m, L.-G.; Juhlin,
L.; Nilsson, B.; Damberg, C., Bartai, T.; Langel, U¨ . Acta Chem. Scand.
1995, 49, 696.
(7) For a recent review, see: Shenolikar, S.; Nairn, A. C. AdV. Second
Messenger Phosphoprotein Res. 1991, 23, 1.
(8) Krebs, E. G. Angew. Chem., Int. Ed. Engl. 1993, 32, 1122.
(18) Carmichael, W. W. Toxicon 1988, 26, 971.
S0002-7863(96)01683-6 CCC: $12.00 © 1996 American Chemical Society

11760 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Humphrey et al.
Scheme 1
The microcystins are produced by microcystis,18 a genus of
cyanobacteria (blue-green algae) that populates freshwater in
many parts of the world and is responsible for the poisoning
deaths of animals that consume or inhabit heavily contaminated
waters.13 The microcystins comprise a series of more than 4019
hepatotoxic and tumor-promoting cyclic heptapeptides with the
general structure cyclo-(D-Ala-X-(D)-erythro-â-methyl-iso-
Asp-Y-Adda-(D)-iso-Glu-N-methyldehyhdro-Ala-), where
the residues are numbered sequentially from (D)-Ala (1) to
N-methyldehydroalanine (7), and the letters X and Y represent
variable positions that are occupied by the only “natural” (L)
amino acids in the molecule. For example, with microcystin-
LA (1),20 leucine (L) fills the “X” site and alanine (A) occupies
the “Y” site. These toxins exhibit a number of synthetically
challenging structural features, including the unusual amino acid
Adda (cf. 4), iso-(D)-glutamic and iso-(D)-â-methyl aspartic acid
residues and N-methyl dehydroalanine (Mdha). Many of these
features are conserved from one toxin to the next, but the
variable regions X and Y can tolerate a fair amount of structural
diversity without complete loss of toxicity.21 The cyclic
pentapeptides nodularin22 and motuporin23 share some of these
features, and the latter has been synthesized by Schreiber and
Valentekovich.24 Recent X-ray diffraction studies of micro-
cystin-bound PP1,25 along with previous kinetic studies with
synthetically modified microcystin and nodularin derivatives,
have indicated that the structural features Adda, iso-Glu, and
Mdha play important roles in phosphatase recognition.26
Results and Discussion
Synthetic Plan. Our approach to the synthesis of microcys-
tin-LA begins with a division of the most challenging of these
structural components into the three fragments shown in Scheme
1: (i) the Mdha-containing fragment 2, (ii) the Adda fragment
4, and (iii) the iso-erythro-â-methyl aspartate fragment 5.
According to this strategy, the difficult N-methyl acylation
leading to fragment 2 would be completed early in the synthesis,
the fragments 2 and 5 would be prepared and joined in a 4 +
2 segment coupling to form the hexapeptide, and the Adda
component would be introduced near the end of the synthesis,
just prior to macrocyclization. Epimerization at the acid
terminal of fragment 2 during the 4 + 2 segment coupling was
to be minimized by employing the new segment coupling
techniques of Carpino,27a and epimerization of the â-amido Adda
derivative 4 during its incorporation was not expected to pose
a problem.24 Orthogonal protection of the advanced synthetic
intermediates would be accomplished with the Boc group and
with methyl and trichloroethyl esters, as shown in Scheme 1.
Our decision to close the macrocycle at Adda-Ala (position
A, Scheme 1) rather than at Adda-Glu (position B, Scheme 1)
was based on conformational preferences and stereoelectronics,
including the location of the Mdha residue within the cyclization
precursor. Specifically, tertiary amides exhibit little or no
preference for the trans conformations assumed by secondary
amides and can even prefer the cis forms if supported by other
favorable interactions such as intramolecular H-bonds.28 Cis
amides can promote cyclization in peptides by orienting the
reactive functionalities in closer proximity, thereby reducing the
entropy penalty that is usually incurred upon cyclization.
Furthermore, full advantage of this type of turn induction can
be achieved when the turn inducing structure lies midway along
the cyclization precursor.29ab
(19) Namikoshi, M. N.; Rinehart, K. L.; Sakai, R.; Stotts, R. R.; Dahlem,
A. M.; Beasley, V. R.; Carmichael, W. W.; Evans, W. R. J. Org. Chem.
1992, 57, 866.
(20) (a) Botes, D. P.; Kruger, H.; Viljoen, C. C. Toxicon 1982, 20, 945.
(b) Botes, D. P.; Tuinman, A. A.; Wessels, P. L.; Viljoen, C. C.; Kruger,
H.; Williams, D. H.; Santikarn, S.; Smith, R. J.; Hammond, S. J. J. Chem.
Soc., Perkin Trans. 1 1984, 2311.
(21) (a) Stoner, R. D.; Adams, W. H.; Slatkin, D. N.; Siegelman, H. W.
Toxicon 1989, 825. (b) Nishiwaki-Matsushima, R.; Fujiki, H.; Harada, K.;
Taylor, C.; Quinn, R. J. Bioorg. Med. Chem. Lett. 1992, 2, 673.
(22) (a) Botes, P. B.; Wessels, P. L.; Kruger, H.; Runnegar, M. T. C.;
Santikarn, S.; Smith, R. J.; Barna, J. C. J.; Williams, D. H. J. Chem. Soc.,
Perkin Trans. I 1985, 2747. (b) Rinehart, K. L.; Harada, K.; Namikoshi,
M.; Chen, C.; Harvis, C.; Munro, M. H. G.; Blunt, J.; Mulligan, P.; Beasley,
V.; Dahlem, A.; Carmichael, W. J. Am. Chem. Soc. 1988, 110, 8557.
(23) de Silva, E. D.; Williams, D. E.; Anderson, R. J.; Klix, H.; Holmes,
C. F. B.; Allen, T. M. Tetrahedron Lett. 1992, 33, 1561.
(24) Valentekovich, R. J.; Schreiber, S. L. J. Am. Chem. Soc. 1995, 117,
9069.
The conformation of the cyclization precursor is only one of
two major issues that demand consideration in the cyclization
step. The second is that of steric hindrance in the coupling
(26) Stotts, R. R.; Namikoshi, M.; Haschek, W. M.; Rinehart, K. L.;
Carmichael, W. W.; Dahlem, A. M.; Beasley, V. R. Toxicon 1993, 31, 783-
789.
(27) (a) Carpino, L. A.; El-Faham, A. J. Org. Chem. 1994, 59, 695. (b)
Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397. (c) Carpino, L. A.;
El-Faham, A.; Minor, C. A.; Albericio, F. J. Chem. Soc., Chem. Commun.
1994, 201. (d) Carpino, L. A.; El-Faham, A.; Albericio, F. Tetrahedron
Lett. 1994, 35, 2279. (e) Carpino, L. A.; El-Faham, A. J. Org. Chem. 1995,
60, 3561.
(25) Goldberg, J.; Huang, H.; Kwon, Y. G.; Greengard, P.; Nairn, A.
C.; Kuriyan, J. Nature 1995, 376, 745.
(28) Rich, D. H.; Mathiaparanam, P. Tetrahedron Lett. 1974, 46, 4037.

Synthesis of Microcystin-LA
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11761
Scheme 2
residues, which can lead to some of the more frequently
observed side reactions during cyclization, such as dimerization
(or oligomerization) or epimerization of the C-terminal residue.
In general, cyclization attempts involving hindered amines (i.e.,
N-methyl residues, Val or Ile residues) are ill-advised,29b but
cyclization attempts between two (S)-amino acid residues can
also retard the cyclization reaction.29b,c In fact, for penta- and
hexapeptides consisting entirely of L residues, the cyclization
reaction often does not proceed until the acid terminal residue
epimerizes to the D configuration.29d Based on these observa-
tions, the Adda-Ala bond (position A) was chosen as the most
attractive site for ring closure since it involves a union between
(S) and (R) residues, and because the amine component is
relatively unencumbered with a sterically nondemanding R-vinyl
group.
hydroxyl group of an N-methyl threonine residue.24 A similar
strategy involves the elimination of sulfonium salts31a or
sulfoxides31b derived from cysteine or related residues. In a
different approach, Rich et al. demonstrated the propensity of
dehydro derivatives to N-alkylate selectively over saturated
residues, but good yields can be difficult to obtain, and some
substrates do not react.28,31c Our reluctance to rely on selective
N-alkylation and our desire for a flexible route that would allow
the incorporation of a variety of dehydro R groups for future
analog preparation prompted us to explore other possibilities
that would present a more direct route to this functionality.
Based on these considerations, modification of Schmidt’s
method31d,e seemed best suited for our purposes.
Our synthesis of Segment 2 (Scheme 2) thus begins with the
acid catalyzed condensation of methyl glyoxylate hemiacetal
732 and N-methyl benzyl carbamate 833 in ether, which gives a
ca. 65:35 equilibrium mixture of 8 and 9, respectively. We
made no attempt to drive the equilibrium, because the starting
materials are readily available in large quantity and the
purification is simple: water soluble 7 is easily removed during
aqueous workup, and N-methyl benzyl carbamate is recycled
after silica gel chromatography to give the desired condensation
product 9 in 27-37% yield (ca. 90% based on recovered
carbamate). Although 9 rapidly equilibrates with its precursors
7 and 8 in the presence of strong acid, pure samples are stable
for over 1 year at -15 °C. Conversion of 9 into the
phosphonylsarcosine derivative 10a34 was achieved in 75% yield
by activation with methanesulfonyl chloride followed by in situ
treatment with sodium iodide and trimethyl phosphite. After
saponification, the resultant acid 10b was appended to D-Ala-
L-Leu-OtBu35 (DCC/HOBt), to supply the tripeptide 11a in high
yield as a 1:1 mixture of diastereomers.
Preparation of Fragment 2. Having established this general
strategy, our synthesis of microcystin-LA begins with the largest
subunit, the Mdha containing fragment 2. Of the many available
routes to dehydro amino acids and peptides,30,31 relatively few
have been applied to the preparation of N-alkylated versions of
this functionality. In a recent example, Schreiber and Valentek-
ovich introduced the dehydro functionality of motuporin in the
final step by the hydroxide-catalyzed elimination of the free
(29) (a) Wipf, P. Chem. ReV. 1995, 95, 2115. (b) Cavelier-Frontin, F.;
Pe´pe, G.; Verducci, J.; Siri, D.; Jacquier, R. J. Am. Chem. Soc. 1992, 14,
8885. (c) Wenger, R. M. HelV. Chim. Acta 1984, 67, 502. (d) Ehrlich, A.;
Rothemund, S.; Brudel, M.; Beyermann, M.; Carpino, L. A.; Bienert, M.
Tetrahedron Lett. 1993, 34, 4781.
(30) For reviews on dehydro amino acid and peptide synthesis, see: (a)
Schmidt, U.; Lieberknecht, A.; Wild, J. Synthesis 1988, 159. (b) Stammer,
C. H. In Chemistry and Biology of Amino Acids, Peptides and Proteins;
John Wright and Sons: London, 1982; Vol. 6, pp 33-74. (c) R,â-
Dehydroamino acids and peptides: Noda, K.; Shimohigashi, Y.; Izumiya,
N. In The Peptides; Gross, E., Meienhofer, J., Eds.; Academic Press: New
York, 1983; Vol. 5, pp 285-339.
(31) (a) Noda, K.; Gazis, D.; Gross, E. Int. J. Peptide Protein Res. 1982,
19, 413. (b) Rich, D. H.; Tam, J. P. J. Org. Chem. 1977, 42, 3815. (c)
Rich, D. H.; Tam, J.; Mathiaparanam, P.; Grant, J. A. Synthesis 1975, 402.
(d) Schmidt, U.; Lieberknecht, A.; Schanbacher, U.; Beuttler, T.; Wild, J.
Angew. Chem., Int. Ed. Engl. 1982, 21, 770. (e) Schmidt, U.; Lieberknecht,
A.; Wild, J. Synthesis 1984, 53.
(32) (a) Bernstein, Z.; Ben-Ishai, D. Tetrahedron 1977, 33, 881. (b) Kelly,
T. R.; Schmidt, T. E.; Haggerty, J. G. Synthesis 1972, 544.
(33) Garcia, J.; Gonzalez, J.; Segura, R.; Vilarrasa, J. Tetrahedron 1984,
40, 3121.
(34) The corresponding diethylphosphonate derivative was prepared
previously by a different route: Shankar, R.; Scott, A. I. Tetrahedron Lett.
1993, 34, 231.

11762 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Humphrey et al.
The next coupling reaction required the free amine 11b and
Boc-D-Glu-OMe (6). We initially used Fmoc protection for this
purpose, anticipating later deprotection with an amine base prior
to macrocyclization. However, attempts to couple the secondary
amine 11b with Fmoc-Glu-OMe (HATU/Hu¨nig’s base; HATU
) O-(7-azabenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexaflu-
orophosphate, vide infra) were thwarted by competing Fmoc
deprotection. Concurrent model studies also implicated base
catalyzed imide formation at the aspartate residue as a potential
side reaction during Fmoc deprotection later in the synthesis.36ac
Since Cbz deprotection conditions were likewise incompatible
with microcystin functionality, we instead chose the Boc
protecting group. Our initial concerns over the stability of the
dehydroalanine residue to the acidic Boc deprotection conditions
were allayed by model studies demonstrating that the Mdha
residue survived prolonged exposure to trifluoroacetic acid.
In order to more conveniently prepare the known selectively
protected D-glutamate 637 required to proceed through the
sequence, we developed a one-pot synthesis from D-glutamic
acid. The procedure (Scheme 2) combines quantitative thionyl
chloride-induced esterification38 with selective γ-ester hydrolysis36b
and N-tert-butyloxycarbonylation and gives multigram quantities
of the protected amino acid 6 as its dicyclohexylamine salt in
63% overall yield.39 Regeneration of the free acid was
accomplished by acidification with aqueous NaHSO4 prior to
coupling with the N-methylamine 11b.
Although the reactivity of secondary amines such as 11b is
comparable with primary amines in SN2 reactions, such is not
the case with peptide coupling reactions, where low coupling
rates often lead to undesired side reactions and extensive
racemization. In these difficult coupling cases, the acid
component of a coupling pair needs more powerful carboxyl
activation, but this also raises concerns over racemization.
Furthermore, the traditional HOBt (1-hydroxybenzotriazole,
Scheme 2) based reagents (DCC/HOBt, BOP, etc.) and other
methods such as N-carboxyanhydride activation tend to give
suboptimal or inconsistent results.40 That the unsuccessful cases
involving HOBt often give benzotriazole active esters as major
products41 underscores the low reactivity of the benzotriazole
ester toward hindered secondary amines.42,43,27b-e
Scheme 3
for these difficult couplings. With HATU, acid activation
proceeds through the active ester derived from hydroxyazaben-
zotriazole (HOAt, Scheme 2), which is proposed to assist
coupling via intramolecular general base catalysis that results
in good reactivity toward N-methyl amines.27b-e Thus, the
HATU mediated coupling of the secondary amine 11b with (D)-
Boc-Glu-OMe (6) in DMF provided the phosphonate diaster-
eomers 12 in good yield. Without purification beyond an
aqueous workup, the diastereomers were treated with aqueous
formaldehyde and potassium carbonate, which afforded the
desired dehydropeptide 13, now a single isomer, in 75% yield
over three steps from 11a. Conversion into the acid 2 was
accomplished in high yield by TFA/thioanisole induced depro-
tection of the Boc and tert-butyl ester groups, followed by in
situ reprotection of the amine group with di-tert-butyldicarbon-
ate.
Preparation of Fragment 5. Having secured segment 2,
we next turned to the preparation of the D-erythro-â-methylas-
partate containing dipeptide 5. This effort begins with the
benzyl/phenylfluorenyl protected aspartate 14 (Scheme 3),
obtained in three steps and 78% overall yield from D-aspartic
acid as described.44a We have previously shown that 14 can
be alkylated (LHMDS/MeI) regio- and stereoselectively to give
the crystalline, syn-â-methyl diastereomer 15 exclusively and
in near quantitative yield.44 This method provides >30 g of
the requisite aspartate derivative in a high yielding route that
requires no chromatography. Conversion of 15 into the â-acid
16 was accomplished with excellent regioselectivity by treatment
with lithium hydroxide at 60 °C, although this reaction was also
accompanied by 30% epimerization at carbon 3 to give 17
(Scheme 3). Attempts to suppress this side reaction by
conducting the reaction at low temperature (0 °C to room
temperature) failed due to the inherent low reactivity of
compound 15,45 but the epimers could be easily separated by
selective crystallization of 16, which gave the desired acid in
70% yield after chromatography of the mother liquors (see
These difficulties with hindered amino acid couplings have
been addressed by several groups, and a number of specialized
reagents are now available that facilitate the union of sterically
hindered N-methyl amino acid derivatives.41b,43 The HATU
reagent developed by Carpino has proven particularly effective
(35) The dipeptide D-Ala-Leu-OtBu was prepared by a standard DCC/
HOBt protocol as described by Bodanski: Bodanski, M.; Bodanski, A. The
Practice of Petide Synthesis; Springer-Verlag: New York, NY, 1984.
(36) (a) Bodanski, M. Principles of Peptide Synthesis, 2nd ed.;
Springer-Verlag: Berlin Heidelberg, 1993; p 170. (b) Principles of Peptide
Synthesis, 2nd ed.; Springer-Verlag: Berlin Heidelberg, 1993; p 128. (c)
Bodansky, M.; Martinez, J. In The Peptides; Gross, E; Meienhofer, J. Eds.;
Academic Press: New York, 1983; Vol. 5.
(37) Schro¨der, E.; Klieger, E. Justus Liebigs Ann. Chem. 1964, 673, 196.
(38) Gmeiner, P.; Feldman, P. L.; Chu-Moyer, M. Y.; Rapoport, H. J.
Org. Chem. 1990, 55, 3068.
(39) The Fmoc and Cbz derivatives and the corresponding aspartic acid
derivatives are available in similar yields by this method. Experimental
details will be described elsewhere.
(40) (a) Cheung, S. T.; Benoiton, N. L. Can. J. Chem. 1977, 55, 911.
(b) Tung, R. D.; Rich, D. H. J. Am. Chem. Soc. 1985, 107, 4342. (c) Van
der Auwera, C.; Anteunis, M. J. Int. J. Peptide Protein Res. 1987, 29, 574.
(d) Jouin, P.; Poncet, J.; Dufour, M.-N.; Pantaloni, A.; Castro, B. J. Org.
Chem. 1989, 54, 617.
(41) (a) Anteunis, M. J. O.; Sharma, N. K. Bull. Soc. Chim. Belg. 1988,
97, 281. (b) Coste, J.; Frerot, E.; Jouin, P.; Castro, B. Tetrahedron Lett.
1991, 32, 1967.
(42) For examples of efficient HOBt-induced acylations of secondary
amines, see: (a) Chu, K. S.; Negrete, G. R.; Konopelski, J. P. J. Org .
Chem. 1991, 56, 5196. (b) White, J. D.; Amedio, J. C. J. Org. Chem. 1989,
54, 736. (c) Grieco, P. A.; Perez-Medrano, A. Tetrahedron Lett. 1988, 29,
4225. (d) reference 24.
(43) Diago-Meseguer, J.; Palomo-Col, A. L.; Fernandez-Lizarbe, J. R.;
Zugaza-Bilbao, A. Synthesis 1980, 547. (b) Van der Auwera, C.; Anteunis,
M. J. Bull. Soc. Chim. Belg. 1986, 95, 203. (c) Coste, J.; Dufour, M.-N.;
Pantaloni, A.; Castro, B. Tetrahedron Lett. 1990, 31, 669.
(44) (a) Humphrey, J. M.; Hart, J. A.; Bridges, R. J.; Chamberlin, A. R.
J. Org. Chem. 1994, 59, 2467. (b) Humphrey, J. M.; Hart, J. A.; Chamberlin,
A. R. Bioorg. Med. Chem. Lett. 1995, 5, 1315.

Synthesis of Microcystin-LA
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11763
Scheme 5
Figure 1.
Scheme 4
by formation of the diene σ bond in a Suzuki cross coupling
reaction.48 This protocol is mild, chemoselective, and has
worked remarkably well in highly functionalized substrates.
With this plan in mind, our synthesis of Adda originates with
the preparation of the backbone fragment, vinyl iodide 22
(Scheme 4), which was obtained from the â-methyl-D-aspartate
derivative 15 as described below.
Experimental section). The stereochemistry of 16 was con-
firmed by re-esterification (K2CO3/MeI/DMF) and comparison
with 15.
Although 16 can be coupled with amino esters in high yield
with HATU, the conditions required for the subsequent removal
of the benzyl/phenylfluorenyl N-protecting groups (H2/Pd-C/
TFA) were not compatible with preservation of the trichloroethyl
ester group we had chosen for the acid terminal of fragment 5.
So instead of coupling 16 and alanine trichloroethyl ester
directly, we first replaced the N-benzyl-N-phenylfluorenyl
groups with the Boc group (H2/Pd-C/TFA, then Boc2O) which
could later be removed with TFA without affecting the
trichloroethyl ester. The HATU/DIEA mediated union of the
Boc derivative 18 and alanine trichloroethyl ester46 gave better
yields of the dipeptide 19 (88%) than DCC/DMAP (71%),
although neither method showed any epimerization according
to 1H NMR analysis. The Boc dipeptide 19 was subsequently
N-deprotected with TFA to give 5 immediately prior to the 4
+ 2 segment coupling with fragment 2 (vida infra).
Synthesis of Fragment 4. With the tetrapeptide 2 and the
dipeptide 5 in hand, we next set about synthesizing the protected
Adda derivative 4. Several syntheses of Adda have appeared
in recent years,24,47 but most of these routes suffered from a
lack of selectivity during diene formation via the Wittig
condensation approach. We developed a new route that would
not only diverge from the methylated aspartate derivative 15,
which we had already used for fragment 5, but that would also
enable the mild and selective introduction of the Adda side chain
Whereas the route to 5 (Scheme 4) required regioselective
hydrolysis of the distal methyl ester of 15, Adda called for the
complementary selective hydrolysis of the R-ester. One excel-
lent method for effecting this type of deprotection is the copper-
(II) assisted hydrolysis of the diester ammonium salt.38 This
transformation (Scheme 4) requires N-deprotection of the
benzylphenylfluorenyl aspartate 15 (H2/Pd-C/TFA), which
produces the hydrochloride salt 23 after an appropriate workup
(see Experimental Section). Treatment of this salt with cupric
carbonate in water/ethanol at 50-60 °C for several hours,
followed by reductive workup with H2S, gave the requisite acid
24 with high selectivity and in nearly quantitative yield. Boc
protection was followed by active ester formation (DCC/N-
hydroxysuccinimide),49 and in situ reduction with sodium
borohydride (73%).50 Formation of the lactone 29, which occurs
spontaneously under basic conditions, is avoided by a final acid
quench to give the amino alcohol 27. The rate of lactonization
is reduced by the 2,3-cis carbamate and methyl substituents of
29 (Scheme 4),51 but chromatographed samples of 27 still slowly
lactonized at room temperature. When stored at -15 °C,
however, the alcohol solidified, and was stable for months when
kept in this manner. Swern oxidation52 of 27 gave the
epimerization-prone aldehyde 28 in excellent yield, which was
used immediately in the next step without purification. Trans-
formation of 28 into the requisite iodoalkene 22 was thus carried
out with CrCl2/CHI3 (53% yield) according to Takai’s proce-
dure53 to give exclusively the trans alkene isomer.
(45) The distal ester of this compound is surprisingly unreactive,
presumably due to steric hinderance from the N-phenylfluorenyl and
R-methyl groups. The compound does not react with hydrazine in refluxing
methanol. The failure of ester hydrazinolysis with hindered substrates has
been noted before: Davies, J. S.; Mohammed, A. K. J. Chem. Soc., Perkin
Trans.1 1981, 2982.
(48) (a) Miyaura, N.; Yamada, K.; Suginome, H.; Suzuki, A. J. Am.
Chem. Soc. 1985, 107, 972. (b) Miyaura, N.; Yamada, Y.; Suzuki, A.
Tetrahedron Lett. 1979, 20, 3437. (c) Miyaura, N.; Suginome, H.; Suzuki,
A. Tetrahedron Lett. 1981, 22, 127.
(46) (a) Alanine trichloroethyl ester was prepared through a higher
yielding modification of the literature procedure. See Carson, J. F. Synthesis
1979, 24. (b) Marinier, B.; Kim, Y. C.; Navarre, J.-M. Can. J. Chem. 1973,
51, 208.
(47) (a) Beatty, M. F.; Jennings-White, C.; Avery, M. A. J. Chem. Soc.,
Perkin Trans. 1 1992, 1637. (b) Namikoshi, M.; Rinehart, K. L.; Dahlem,
A. M.; Beasley, V. R.; Charmichael, W. W. Tetrahedron Lett. 1989, 30,
4349. (c) Chakraborty, T. K.; Joshi, S. P. Tetrahedron Lett. 1990, 31, 2043.
(49) Anderson, G. W.; Zimmerman, J. E.; Calahan, F. M. J. Am. Chem.
Soc. 1964, 86, 1839.
(50) Nikawa, J.; Shiba, T. Chem. Lett. 1979, 981.
(51) Similar substrates without 2,3-syn substituents appear to lactonize
spontaneously. See Baldwin, J. E.; North, M.; Flinn, A. Tetrahedron Lett.
1987, 28, 3167, and references therein.
(52) (a) Mancuso, A. J.; Swern, D. Synthesis 1981, 165. (b) Tidwel, T.
T. Synthesis 1990, 857.

11764 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Humphrey et al.
The synthesis of the second Suzuki coupling partner, the
boronic acid 20, begins with the known Evans aldol product
30 (Scheme 5).47a,54 Transformation into the Weinreb amide
31 (87%)55 was followed by methyl ether formation (NaH/MeI,
95%) and DIBAH reduction (79%)12 to the aldehyde 33. This
aldehyde was readily converted into the methyl alkyne 34b (via
the dibromoalkene 34a, not shown) in 85% yield via the Corey-
Fuchs protocol,56 but the subsequent transformation of 34b into
the requisite boronic acid 20 proved troublesome. Hydrobo-
rations of internal alkynes are known to proceed with moderate
to poor regioselectivity, but branching R to the alkyne (i.e.,
position 4 of 34b) reportedly raises this regioselectivity to
acceptable levels.57 Recent modifications, such as the employ-
ment of pinacolborane with58 or without59 catalysts such as Cp2-
ZrHCl, reportedly give significant improvements in hydrobo-
ration regioselectivities and efficiencies for a variety of terminal
and internal alkynes, but attempts to hydroborate alkyne 34b
according to these methods resulted only in its slow conversion
into complex product mixtures. When 34b was subjected to
the Suzuki-Miyaura protocol for hydroboration and dealkylative
boronic ester formation,60 the alkyne failed to react. Conversely,
catecholborane57 gave fairly efficient hydroboration when carried
out in a nonchelating solvent such as benzene but was essentially
nonselective (1.5:1 ratio of 20:21). Of the reagents we
investigated, dibromoborane dimethyl sulfide61 gave the best
results for the transformation of 34b to 20, but when the reaction
progressed past 50% conversion, products from methyl ether
cleavage and elimination began to appear. These side reactions
were suppressed in the presence of Ag2CO3 but not completely.
Nonetheless, Br2BH‚SMe2 hydroboration of 34b followed by
hydrolysis of the resultant dihaloborane gave 56% (74% based
on recovered 34b) of the boronic acids 20 and 21, in a 4:1 ratio
respectively, which were readily separated by chromatography.
The undesired isomer 21 was stable for weeks at -15 °C, but
the desired isomer 20 was unstable and was immediately
subjected to the next step.
Scheme 6
purification. Saponification of the methyl ester (LiOH/THF/
water) then gave the requisite Boc-ADDA-OH 4 in 88% yield
after silica gel chromatography.
Preparation of Fragment 3 and Macrocyclization. Having
prepared the components 2, 4, and 5, we were ready to assemble
the cyclization precursor 3 and began this process with the 4 +
2 segment coupling of fragments 2 and 5 (Scheme 6). Segment
couplings can be troublesome, because racemization generally
occurs faster during the activation of N-acyl R-amino acids than
it does during activation of the urethane-protected amino acids
used in conventional stepwise coupling protocols.27a This
problem has often led to the rejection of segment coupling
strategies in favor of stepwise coupling methods, at the
unfortunate sacrifice of convergence. However, through the
development of new reagents and improved reaction conditions,
low-racemization segment couplings are now possible in many
cases. For example, Carpino et al. have shown that acid
activation by HATU in the presence of collidine (rather than
the more commonly used Hu¨nig’s base) gives little epimerization
even in difficult cases.27a Using this method, the segment
acylation of 5 with 2 proceeded well, supplying the hexapeptide
38 in 80% yield with no observable epimerization. Deprotection
of the Boc group was accomplished with TFA immediately prior
to coupling with the Adda fragment 4.
For the final coupling reaction, which gives the cyclization
precursor 3, we relied again on the remarkable efficiency of
HATU.27c,e The heptapeptide 3 was thus obtained in satisfactory
yield and with no epimerization from the HATU/collidine
mediated coupling of the hexapeptide 39 and the Adda fragment
4 (Scheme 7). Deprotection of the trichloroethyl ester of 3
proceeded well (Zn/HOAc,46 84%, Scheme 7), and the resultant
acid 40 was activated for macrocyclization by treatment with
pentafluorophenol and DCC, followed by N-Boc deprotection
(TFA) of the resultant active ester 41.64a The original literature
protocols for peptide cyclization via the pentafluorophenyl ester
call for dropwise addition of the amine TFA salt to a mixture
of saturated aqueous NaHCO3 or Na2CO3 and chloroform,64 but
we were concerned about the potential for carbonation of the
amine group under these conditions. We therefore attempted
this reaction under both the literature conditions and by
substituting pH 9.5 phosphate buffer for carbonate. This change
resulted in a 10-20% improvement in yield, suggesting that
bicarbonate buffer may indeed not be optimal for this type of
cyclization. Macrolactamization was induced by the dropwise
addition of a 1 mM solution of the pentafluorophenyl ester TFA
salt to a two phase mixture of chloroform and pH 9.5 phosphate
buffer, which gave the cyclized diester in 56% yield from the
active ester 41 after reversed phase HPLC purification. Slow
addition was not necessary for the cyclization, since addition
over 10 min gave the same result as did syringe pump addition
over several hours. Our synthesis concludes with the lithium
hydroxide induced saponification of the methyl esters (LiOH/
With the Suzuki coupling partners 20 and 22 in hand, we
attempted the palladium catalyzed Suzuki cross-coupling reac-
tion.48 Our initial attempts with 1 M KOH as the base for this
coupling resulted in the formation of small amounts of the
product 35, but the bulk of the vinyliodide was converted into
a single isomer of the undesired diene 37, as verified by HRMS
1
and H NMR (no attempt was made to assign the stereochem-
istry of 37). We suspect that this diene forms via the pathway
shown in Figure 1, whereby the palladium catalyzed alkene
isomerization of 22 to 36 is followed by E2′ elimination to give
37.62 This side reaction was avoided by employing thallium
ethoxide63a in the presence of water,63b and under these new
conditions the palladium catalyzed Suzuki cross-coupling of
vinyl iodide 22 and the boronic acid 20 provided the fully
protected ADDA fragment 35 in 82% yield after HPLC
(53) Takai, K.; Nitta, K.; Ulimoto, K. J. J. Am. Chem. Soc. 1986, 108,
7408.
(54) (a) Gage, J. R.; Evans, D. A. Organic Synt. 1989, 68, 83. (b) Evans,
D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981, 103, 2127.
(55) Evans, D. A.; Kaldor, S. W.; Jones, T. K.; Clardy, J.; Stout, T. J. J.
Am. Chem. Soc. 1990, 112, 7001-7031.
(56) Corey, E. J.; Fuchs, P. L. Tetrahedron Lett. 1972, 3769.
(57) For an early review on hydroborations, see: Brown, H. C.;
Campbell, J. B. Aldrich. Acta 1981, 14, 3.
(58) Srebnik, M.; Pereira, S. Organometallics 1995, 14, 3127-3128.
(59) Knochel, P.; Tucker, C. E.; Davidson, J. J. Org. Chem. 1992, 57,
3482-3485.
(60) Suzuki, A.; Miyaura, T. N.; Kamabuchi, A. Synth. Commun. 1993,
23, 2851-2859.
(61) Brown, H. C.; Campbell, J. B. J. Org. Chem. 1980, 45, 389-395.
(62) Heck, R. F. Palladium Reagents in Organic Synthesis; Academic
Press: London, 1985; pp 19-20.
(63) (a) Uenishi, J.; Beau, J.-M.; Armstrong, R. W.; Kishi, Y. J. Am.
Chem. Soc. 1987, 109, 4756. (b) Smith, G. B.; Dezeny, G. C.; Hughes, D.
L.; King, A. O.; Verhoeven, T. R. J. Org. Chem. 1994, 59, 8151.
(64) (a) Schmidt, U.; Utz, R.; Lieberknecht, A.; Griesser, H.; Potzolli,
B.; Bahr, J.; Wagner, K.; Fischer, P. Synthesis 1987, 236. (b) Schmidt, U.;
Lieberknecht, A.; Griesser, H.; Talbiersky, J. J. Org. Chem. 1982, 47, 3261.

Synthesis of Microcystin-LA
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11765
Scheme 7
THF/H2O, 0 °C), which gives a 50% yield of microcystin-LA
(1).65,66 The synthetic microcystin-LA co-eluted with authentic
material by reversed phase analytical HPLC (80:20 MeOH:0.2%
aqueous TFA), and the 1H NMR spectra of the samples in D2O
were identical.67,68
part to Carpino’s remarkable HATU reagent in several difficult
couplings of sterically hindered or racemization-prone amino
acid residues.
Experimental Section
Conclusion
General Methods. 1H and 13C NMR spectra were obtained on an
Omega 500 (500 MHz) or a General Electric GN-500 (500 MHz)
spectrometer. For spectra measured in organic solvents, data are
In summary, the first total synthesis of a microcystin described
herein features a convergent route that is amenable to analog
preparation in the search for selective phosphatase inhibitors.
It also features an efficient aspartate alkylation method that
provides large quantities of the alkylated aspartate 15, a key
intermediate en route to both the erythro-â-methyl aspartate and
Adda amino acid units. To our knowledge, this work marks
the first application of Schmidt’s phosphonoglycine method to
the preparation of N-alkyl dehydroamino acid containing
peptides and demonstrates the compatibility of the N-methyl-
dehydro residue with Boc and tert-butyl ester protection
schemes. Finally, the success of this route is due in no small
1
reported in ppm from internal tetramethylsilane for H NMR and in
1
ppm from the solvent for 13C NMR. For H spectra taken in D2O,
data are reported in ppm relative to HDO (3.80 ppm). Data are reported
as follows: chemical shift, multiplicity (app ) apparent, par obsc )
partially obscured, ovrlp ) overlapping, s ) singlet, d ) doublet, t )
triplet, q ) quartet, m ) multiplet, br ) broad, abq ) ab quartet),
coupling constant, and integration. Infrared (IR) spectra were taken
with a Perkin-Elmer Model 1600 series FTIR spectrophotometer.
Optical rotations were obtained with a JASCO DIP-360 digital
polarimeter. Melting points (mp) were obtained from a Laboratory
Devices Mel-Temp melting-point apparatus and are reported uncor-
rected. Elemental analyses were performed by Desert Analytics,
Tucson, AZ. Thin-layer chromatography (TLC) was performed on 0.25
mm Merck precoated silica gel plates (60 F-254), and silica gel
chromatography was performed using ICN 200-400 mesh silica gel.
Normal phase preparative HPLC was conducted with a 25 × 100 mm
Waters prep Nova-Pak HR silica cartridge (WAT038511), and reversed
phase preparative HPLC was conducted with a 25 × 100 mm Waters
prep Nova-Pak HR C18 cartridge (WAT038510). Reversed phase
analytical HPLC was conducted with a 4.6 × 250 mm Rainin
Microsorb-MV C18 column (86-200-C5). Solution pH measurements
were made with Whatman type CF pH 0-14 indicator paper strips.
Inert atmosphere operations were conducted under nitrogen passed
through a Drierite drying tube in oven or flame-dried glassware.
Anhydrous tetrahydrofuran (THF) was distilled first from calcium
hydride and then from potassium; anhydrous ether was distilled from
potassium. Triethylamine and methylene chloride were purified by
distillation from calcium hydride. Diisopropylethylamine (DIEA) and
DMF were dried over sieves. LHMDS (1.0 M in THF) was purchased
from Aldrich Chemical Company. HATU was purchased from
PerSeptive Biosystems Inc. D-Boc-Glu(OH)-OMe was prepared by the
one-pot procedure described in the text, and full experimental details
will be published elsewhere. Alanine trichloroethyl ester was prepared
through a higher yielding modification of the literature protocol,46a as
described below. N-Methyl benzylcarbamate was prepared from
benzylchloroformate and aqueous methylamine as suggested in the
literature.33 Methylglyoxylate hemiacetal was prepared according to
the literature procedure.32 All other reagents were used as purchased
from Aldrich, Sigma-Aldrich, or Acros unless otherwise stated.
(65) The remainder of the material is isomeric with the desired product
(HRMS and 1H NMR). Assuming it to result from base catalyzed
epimerization competing with saponification, we attempted a number of
alternative saponification methods but with no success. Cesium carbonate
(ref 66a) gave levels of isomerization comparable to those observed with
lithium hydroxide, and barium hydroxide gave more isomerization. Lithium
hydroperoxide (ref 66b-d) generated only small amounts of the desired
product along with multiple side products. The level of isomerization during
this step is consistent with levels observed by Schreiber and Valentekovich
for the analogous saponification step in their synthesis of motuporin: Dr.
Robert J. Valentekovich, The University of California at Irvine, personal
communication, 1996.
(66) (a) Kaestle, K. L.; Anwer, M. K.; Audhya, T. K.; Goldstein, G.
Tetrahedron Lett. 1991, 32, 327. (b) Misra, H. K.; Virzi, F.; Hnatowich,
D. J.; Wright, G. Tetrahedron Lett. 1989, 30, 1885. (c) Lucas, H.; Basten,
J. E. M.; van Dither, T. G.; Meuleman, D.; van Aelst, S. F.; van Boeckel,
A. A. Tetrahedron 1990, 46, 8207. (d) Dussault, P.; Lee, I. Q. J. Org. Chem.
1992, 57, 1952.
(67) The final product 1 was only slightly soluble in D2O, which
complicated characterization by 1H NMR analysis due to the large HDO
solvent peak. The water supression NMR method developed by Shaka
eliminated this problem. See: Hwang, T. L.; Shaka, A. J. J. Magn. Reson.
Series A 1995, 112, 275.
(68) Whereas the proton NMR of microcystin-LA in D2O exhibited a
single set of sharp resonances, but such was not the case with most of the
peptide intermediates. The 1H spectra of these compounds exhibited rotamer-
induced peak broadening and doubling at room temperature in CDCl3 or
DMSO-d6. Coalescence of the rotamers generally occurred at 70-90 °C in
DMSO-d6, which facilitated peak assignment and characterization. The
cyclized diester 42 exhibited severe peak broadening and doubling in
DMSO-d6 at room temperature, but, when heated to 70-90 °C, imide
formation occurred as evidenced by the absence of a methyl ester resonance
in the 1H NMR spectrum and by the loss of one unit of methanol according
to HRMS. The spectrum of 42 in CDCl3 exhibited the presence of one
major rotamer and <10% of a minor one, which began to coalesce at 45
°C.
N-Carbobenzyloxy-r-hydroxysarcosine Methyl Ester (9). To a
stirred solution of 35.3 g (214 mmol) of N-methylbenzylcarbamate (8)33

11766 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Humphrey et al.
and 25.7 g (214 mmol) of methylglyoxylate hemiacetal 732 in 214 mL
of anhydrous ether was added 5.0 g (21 mmol) of camphor sulfonic
acid at room temperature. After 30 h, the solution was shaken with
200 mL of 50% saturated aqueous NaHCO3. The water layer was
extracted with an additional 200 mL of ether, and the combined ether
portions were washed with brine, dried with MgSO4, filtered, and
concentrated to give 40 g of a clear oil containing mainly 9 and 8 (in
Rf 0.24 and 0.38 (EtOAc); IR: 3289, 2957, 1734, 1697, 1667, 1247,
1150, 1035; FAB MS m/e calcd for C26H43N3O9P+ (M + H+):
572.2737, found 572.2743. The high Rf diastereomer could be
selectively crystallized in low yield from ether/hexanes at -15°C over
1-2 weeks: mp 54-63 °C (dec); 1H NMR (500 MHz, DMSO-d6, 110
°C) δ 0.85 (d, J ) 6.1, 3H), 0.90 (d, J ) 6.4, 3H), 1.26 (d, J ) 7.1,
3H), 1.41 (s, 9H), 1.53 (app t, J ) 7.1, 2H), 1.64 (app dq, J ) 7.1, 7.1,
1H), 3.02 (s, 3H), 3.68, 3.69 (two ovrlp d, J ) 10.6 and 10.7, 6H),
4.17 (br app q, J ) 7.2, 1H), 4.38 (br app dq, J ) 7.4, 1H), 5.14 (s,
2H), 5.32 (d, J ) 23.9, 1H), 7.32 (m, 1H), 7.36 (app d, J ) 4.2, 4H),
7.83 (br d, J ) 7.7, 1H), 7.99 (br d, J ) 6.8, 1H). 1H NMR data for
the other isomer was obtained from a proton spectrum of the
diastereomeric mixture, and are as follows: 1H NMR (500 MHz,
DMSO-d6, 110 °C) δ 0.86 (par obsc d, J ) 5.9, 3H), 0.91 (par obsc d,
J ) 5.9, 3H), 1.26 (d, J ) 6.9, 3H), 1.41 (s, 9H), 1.53 (app t, J ) 7.1,
2H), 1.64 (app dq, J ) 7.1, 7.1, 1H), 3.00 (s, 3H), 3.69 (d, J ) 10.8,
6H), 4.17 (app q, J ) 7.2, 1H), 4.38 (br dq, J ) 7.4, 1H), 5.14 (s, 2H),
5.30 (par obsc d, J ) 23.9, 1H), 7.32 (m, 1H), 7.36 (app d, J ) 4.2,
4H), 7.86 (par obsc br m, 2H).
Boc-D-iso-Glu(OMe)-Me∆Ala-(D)-Ala-(L)-Leu-OtBu (13). 11a
(5.0 g, 8.75 mmol) was hydrogenated in 16 mL of methanol over 500
mg of 10% palladium on carbon for 2.5 h at 1 atm, then filtered, and
concentrated. The residue was reconcentrated twice from 5 mL of
toluene to remove residual methanol, and the remaining oil was
dissolved in 15 mL of DMF. Boc-Glu-OMe (6) (3.43 g, 13.1 mmol)
was then added, followed by 3.39 g (26.3 mmol) of DIEA, and, after
cooling to 0 °C, 4.99 g (13.1 mmol) of HATU. After stirring for 1 h
at 0 °C and 17 h at room temperature, the resultant orange solution
was partitioned between 200 mL of 50% saturated citric acid and 200
mL of ether. The organic portion was washed with 150 mL of 50%
saturated NaHCO3, and the aqueous portions were back extracted twice
with 150 mL each of ether. The combined organic portions were then
washed with brine, dried with MgSO4, filtered, and concentrated to
give an off-white foam consisting of the two diastereomers of 12, which
could be resolved by TLC, but were inseparable by silica gel
chromatography. Rf 0.26 and 0.36 (EtOAc); IR (thin film) 3326, 2962,
1728, 1712, 1697, 1666 cm-1; FAB MS calcd for M + Na+: 703.3295;
found: 703.3291. The NMR spectra for this 1:1 mixture of diaster-
eomers was further complicated by the presence of rotamers and spin
coupling to phosphorus. For this reason, and to obtain a higher overall
yield for the sequence, further purification and 1H NMR characterization
were delayed until after the Horner-Emmons reaction, which gives a
single stereoisomer: to the crude mixture of diastereomers was added
40 mL of THF, 20 mL of water, 4.8 g (35 mmol) of K2CO3, and 2.1
mL (ca. 10 equivalents) of 37% aqueous formaldehyde. The mixture
was stirred vigorously for 8 h at room temperature, then poured into
100 mL of water, and extracted twice with 100 mL of ether. The
organic extracts were washed with 100 mL of 50% saturated citric acid,
and the aqueous wash was back extracted with 100 mL of ether.
Combined organic portions were washed with brine, dried over MgSO4,
filtered, and concentrated to give 4.95 g of an off-white foam.
Chromatography (EtOAc) gave 3.80 g (75%) of 13 as a foam that
exibited no distinct melting point. Rf 0.38 (EtOAc); IR (thin film) 3320,
2971, 1722, 1658 cm-1; 1H NMR (500 MHz, DMSO-d6, 90 °C) δ 0.85
(d, J ) 6.8, 3H), 0.89 (d, J ) 6.5, 3H), 1.29 (d, J ) 7.6, 3H), 1.37 (s,
9H), 1.40 (s, 9H), 1.53 (app t, J ) 7.6, 2H), 1.63 (dq, J ) 6.8, 6.8,
1H), 1.82 (app dq, J ) 8.3, 8.3, 1H), 1.95 (app dq, J ) 7.2, 7.2, 1H),
2.32 (br app q, J ) 7.2, 1H), 3.01 (br s, 3H), 3.62 (s, 3H), 3.98 (br app
q, J ) 6.0, 1H), 4.17 (app q, J ) 7.2, 1H), 4.39 (dq, J ) 7.6, 7.6, 1H),
5.52 (br s, 1H), 5.96 (br s, 1H), 6.79 (br s, 1H), 7.74 (br s, 1H), 7.86
(br d, J ) 7.6, 1H); MS calcd for C28H49N4O9+ (M + H+): 585.3499;
found: 585.3489.
1
a 1:2 ratio, respectively) and a trace of 7 by H NMR analysis. The
material was chromatographed (3:7 EtOAc:hexanes) to give 14.9 g
(27%) of the title compound 9 and 24.5 g of recovered 8. Spectral
data for 9 are as follows: Rf 0.35 (1:1 EtOAc:hexanes); IR (thin film):
1
3408, 2954, 1750 cm-1; H NMR (500 MHz, DMSO-d6, 70 °C): δ
2.77 (s, 3H), 3.66 (s, 3H), 5.12 (s, 2H), 5.77 (br s, 1H); 6.61 (br s,
+
1H); 7.31-7.39 (m, 5H); MS m/e calcd for C12H16NO5 (M + H+):
254.1028. Found 254.1022.
N-Carbobenzyloxy-r-dimethylphosphonylsarcosine Methyl Ester
(10a).34 To 13.3 g (52.6 mmol) of 9 and 16.0 g (158 mmol) of
triethylamine in 100 mL of dry CH2Cl2 at -20 °C was added dropwise
6.63 g (57.9 mmol) of freshly distilled methanesulfonyl chloride. The
mixture was warmed to room temperature after 2 h and stirred for an
additional 15 h. The solvent was then removed under vacuum and
replaced with 50 mL of dry acetonitrile. Powdered potassium iodide
8.73 g (52.6 mmol) was added followed by 7.83 g (63.1 mmol) of
trimethylphosphite. After vigorous stirring for 22 h, the solvent was
removed under vacuum, and the residue was partitioned between EtOAc
and water. The aqueous layer was extracted with an additional portion
of EtOAc, and the combined organic layers were washed with brine,
dried over MgSO4, filtered, and concentrated to give an orange oil.
Chromatography (9:1 EtOAc:hexanes) gave 13.6 g (75%) of 10a as a
pale yellow oil: Rf 0.35 (EtOAc) IR (thin film) 2957, 1750, 1704 cm-1
;
1H NMR (500 MHz, CDCl3), two rotamers: δ 3.09 (s, 3H, minor
rotamer), 3.12 (s, 3H, major rotamer), 3.73, 3.75, 3.777, 3.781, 3.798,
3.803, 3.829, 3.833, 3.86 (9S, 9H from each rotamer,69 assigned to the
methyl esters of the phosphonate and the carboxylate groups), 5.11 (d,
J ) 12.3, 1H, minor rotamer), 5.15-5.22 (obsc d, 1H, minor rotamer),
5.19 (app dd, 2H, J ) 11.7, major rotamer), 5.32 (d, J ) 26.5, 1H,
minor rotamer), 5.56 (d, J ) 26.0, 1H, major rotamer), 7.32-7.38 (m,
5H from each rotamer); MS m/e calcd for C14H21NO7P (M + H+):
345.0972, found 345.0963.
N-Carbobenzyloxy-r-dimethylphosphonylsarcosine (10b).34 To
a solution of 10.9 g (31.6 mmol) of 10a in 30 mL of THF at room
temperature was added a solution of 1.52 g (36.3 mmol) of lithium
hydroxide in 20 mL of water. After stirring vigorously for 3 h, the
bulk of the THF was removed in vacuo and the remaining solution
was washed with one 25 mL portion of ether. The aqueous portion
was then acidified with 5 mL of 6 M HCl and extracted with three 20
mL portions of EtOAc. The combined organic layers were washed
with brine, dried with MgSO4, and concentrated to give 9.8 g (98%)
of a colorless oil that solidified upon standing and was pure according
1
to H NMR. Crystallization from EtOAc/hexanes gave 8.8 g (88%)
of 10b as a white solid: mp 94-96 °C; Rf 0.60 (80:20:15:1 CH2Cl2:
MeOH:HOAc:H20); IR (thin film) 3433-2578 br, 2956, 1739, 1700
1
cm-1; H NMR (500 MHz, CDCl3, major rotamer): δ 3.10 (s, 3H);
3.76 (d, J ) 11.2, 3H), 3.89 (d, J ) 11.2, 3H), 5.14 (d, J ) 12.4, 1H),
5.21 (d, J ) 12.4, 1H), 5.60 (d, J ) 26.7, 1H), 7.30-7.38 (m, 5H),
10.20 (br s, 1H); MS m/e calcd for C13H19NO7P+ (M + H+): 332.0894,
found 332.0991. analysis calcd for C13H18NO7P: C, 47.13; H, 5.49;
N, 4.23. Found: C, 46.92; H, 5.67; N, 4.40.
N-Carbobenzyloxy- r-dimethylphosphonylsarcosyl-(D)-alanyl-(L)-
leucine-tert-butyl Ester (11a). To 7.79 g (19.8 mmol) of N-Cbz-(D)-
Ala-(L)-Leu-OtBu in 40 mL of methanol was added 1.2 g of 10%
palladium on carbon. The mixture was hydrogenated at 1 atm for 2.5
h, then filtered through Celite, and concentrated. The residue was
reconcentrated from toluene and then from CH2Cl2/hexanes to remove
residual methanol. The residue was then dissolved in CH2Cl2, and 7.20
g (21.8 mmol) of 10b was added, followed by 2.94 g (21.8 mmol) of
HOBt and, at 0° C, 4.50 g (21.8 mmol) of DCC. The mixture was
stirred for 2 h at 0° C and for 12 h at room temperature. After filtering
through Celite and concentrating, the resultant oil was dissolved in
EtOAc and washed succcessively with saturated citric acid, saturated
NaHCO3, and brine. The solution was dried over MgSO4, filtered,
concentrated, and chromatographed (EtOAc) to give 10.3 g (91%) of
11a as a white foam consisting of a 1:1 mixture of two diastereomers:
Boc-D-iso-Glu(OMe)-Me∆Ala-(D)-Ala-(L)-Leu-OH (2). To 1.00
g (1.71 mmol) of 13 in 6 mL of TFA was added 212 mg (1.71 mmol)
of thioanisole. After 30 min, the solution was concentrated and then
reconcentrated once from CH2Cl2/hexanes and three times from 20 mL
portions of EtOAc. The remaining residue was diluted with 4 mL of
water, and 1.1 g (10.3 mmol) of Na2CO3 was added, followed by, after
cooling to 0 °C, 410 mg (1.88 mmol) of di-tert-butyldicarbonate in 4
mL of dioxane. The mixture was warmed to room temperature, stirred
for 12 h, then poured into a separatory funnel, diluted with 30 mL of
water, and washed twice with ether. The aqueous portion was acidified
to pH 2-3 with saturated citric acid and extracted twice with EtOAc.

Synthesis of Microcystin-LA
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11767
The combined extracts were washed with brine, dried with MgSO4,
filtered, and concentrated to an oil. Chromatography (1:10:90 HOAc:
MeOH:CH2Cl2) gave 817 mg (90%) of 2 as a white foam. A sample
precipitated as a white powder from ether/hexanes at -78 °C exhibited
no distinct melting point: Rf 0.40 (1:10:90 HOAc:MeOH:CH2Cl2); IR
(thin film) 3318, 2959, 1722, 1653 cm-1; 1H NMR (500 MHz, DMSO-
d6, 90 °C) δ 0.85 (d, J ) 6.4, 3H), 0.88 (d, J ) 6.8, 3H), 1.37 (s, 9H),
1.55 (t, J ) 6.8, 2H), 1.62 (app dq, J-13.5, 7.2, 1H), 1.80 (app dq, J )
14.7, 6.8, 1H), 1.95 (app dq, J ) 12.7, 6.0, 1H), 2.31 (m, 3H), 3.00 (s,
3H), 3.61 (s, 3H), 3.98 (br app q, J ) 5.2, 1H), 4.26 (app q, J ) 7.6,
1H), 4.40 (app dq, J ) 7.2, 7.2, 1H), 5.52 (br s, 1H), 5.98 (br s, 1H),
6.89 (br s, 1H), 7.79 (br s, 1H) 7.92 (br s, 1H); FAB MS m/e calcd for
temperature, diluted with 440 mL of fresh ether, and cooled in an ice
bath for 2 h. The resultant white precipitate was collected by filtration
and recrystallized from MeOH/Et2O to give 22.3 g (81%) of alanine
trichlorethyl ester p-toluene sulfonate as a white solid. The physical
properties of the product matched the literature values.46a The free
amine was generated as follows: the tosylate salt (998 mg 2.76 mmol)
was shaken with 30 mL of 50% saturated NaHCO3 and 30 mL of ether.
The aqueous portion was extracted with an additional 30 mL of ether,
and the combined organic extracts were dried with K2CO3, filtered,
and concentrated to give 531 mg (100%) of alanine trichloroethyl ester,
which was used directly in the next step.
Boc-D-erythro-â-Me-iso-Asp(OMe)-Ala-OTce (19). The acid 18
(360 mg 1.38 mmol) and 531 mg (2.76 mmol) of alanine trichloroethyl
ester were concentrated together from 5 mL of toluene, and the residue
was dissolved in 4 mL of DMF and cooled in an ice bath. To the
solution was added 535 mg (4.14 mmol) of DIEA, followed by 577
mg (1.52 mmol) of HATU. The solution was warmed to room
temperature after 1 h and stirred for an additional 8 h. The mixture
was worked up as for compound 13 to give 644 mg of an off-white
solid. Chromatography (1:1 EtOAc:hexanes, 36 g of SiO2) gave 563
mg (88%) of 19. Crystals were obtained from EtOAc/hexanes at -15
°C: mp 161-162 °C; Rf 0.44 (1:1 EtOAc:hexanes); IR (thin film) 3330,
+
C24H41N4O9 (M + H+): 529.2873. Found: 529.2895.
D-erythro-N-(9-Phenylfluoren-9-yl)-N-benzyl-â-methylaspartic Acid
r-Methyl Ester (16). To 4.19 g (8.28 mmol) of 15 in 150 mL of
THF was added 45 mL of MeOH, and the resultant solution was heated
to 60° C. Lithium hydroxide monohydrate (3.47 g, 82.8 mmol) in 90
mL of H2O was then added over 10 min, and the resultant cloudy
mixture was stirred at 60 °C for 24 h. After cooling to room
temperature, the mixture was washed twice with ether, and the combined
organic phases were back-extracted twice with water. The combined
aqueous phases were acidified with saturated citric acid and extracted
twice with EtOAc. The combined EtOAc phases were washed once
each with water and brine, dried over MgSO4, filtered, concentrated to
approximately 1/2 the original volume, and diluted with hexanes. The
desired epimer (1.64 g) was obtained as a crystalline solid, and the
mother liquors were chromatographed (65 g silica/g sample, 25:75
EtOAc:hexanes) to provide an additional 1.12 g (69% combined yield)
of the title compound 16: Rf 0.63 (1:1 EtOAc:hexanes); IR (thin film)
3055 br, 2933, 1700 cm-1; 1H NMR (500 MHz, CDCl3) δ 0.66 (d, J )
7 Hz, 3H), 2.43 (dq, J ) 7, 11.5 Hz, 1H), 3.00 (s, 3H), 3.57 (d, J )
12 Hz, 1H), 4.51 (d, J ) 13.5 Hz, 1H), 4.85 (d, J ) 13.5 Hz, 1H),
7.25-7.84 (m, 20H); HRMS m/e calcd for (M+H)+:492.2176. Found
492.2174. Anal. Calcd for C32H29NO4:C, 78.16; H, 5.96; N, 2.85.
Found: C, 78.14; H, 6.01; N, 2.87. The undesired epimer 17 was
obtained as 1.18 g (29%) of a white solid: Rf 0.33 (1:1 EtOAc:hexanes);
IR (thin film) 3055 br, 2933, 1716 cm-1; 1H NMR (300 MHz, CDCl3)
δ 1.10 (d, J ) 7.1 Hz, 3H). 2.18 (dq, J ) 11.3, 7.2 Hz, 1H). 2.88 (s,
3H), 3.41 (d, J ) 11.3 Hz, 1H), 4.05 (d, J ) 13.4 Hz, 1H), 4.35 (d, J
) 13.4 Hz, 1H), 7.18-7.75 (m, 20H); MS m/e calcd for (M+H)+:
492.2176, found 492.2168.
1
2979, 1750 cm-1; H NMR (500 MHz, CDCl3) δ 1.28 (d, J ) 7.6,
3H), 1.45 (s, 9H), 1.50 (d, J ) 7.2, 3H), 3.11 (m, 1H), 3.74 (s, 3H),
4.44 (dd, J ) 4.0, 9.9, 1H), 4.63 (d, J ) 11.9, 1H), 4.64 (par obsc dq,
J ) 7.2, 7.2, 1H), 4.96 (d, J ) 11.9, 1ÅH), 5.72 (d, J ) 8.7, 1H), 6.20
+
(d, J ) 6.4, 1H) HRMS m/e calcd for C16H26N2O7Cl3 (M + H+):
463.0805. Found: 463.0805.
D-erythro-â-Methyl Dimethylaspartate Hydrochloride (23). To
19.0 g (37.6 mmol) of 15 in 20 mL of CH2Cl2 and 40 mL of MeOH
was added 8.57 g (75.2 mmol) of TFA. The resultant solution was
hydrogenated over 3.8 g of 10% Pd-C at 52 psi for 8 h. The mixture
was then carefully filtered through Celite, and the filter pad was rinsed
with copious CH2Cl2 and MeOH. The CH2Cl2 was removed under
vacuum, and the resultant methanolic slurry was filtered through Celite
to remove the white solids. The filtrate was concentrated to an oil,
which was then partitioned between 20 mL of water and 40 mL of
ether. Three milliliters of concentrated HCl was added, and the mixture
was shaken. The aqueous portion was removed, cooled with ice, and
basified to pH 10.0 with cold concentrated ammonium hydroxide. The
mixture was then extracted twice with EtOAc, and the combined EtOAc
portions were washed with brine, dried with K2CO3, and concentrated
Boc-D-erythro-â-methyl Aspartic Acid r-Methyl Ester (18). A
Par bottle was charged with 2.71 g (5.51 mmol) of 16, 51 mL of EtOAc,
and 9 mL of CH2Cl2. After the solids had dissolved, 0.85 mL (1.10
mmol) of TFA was added, carefully followed by 405 mg of 10% Pd/
C. The mixture was hydrogenated at 50 psi for 14 hours, then filtered
through Celite, and concentrated. The resultant white crust was
partitioned between 50 mL of ether and 24 mL of water. The aqueous
layer was washed once with ether, and the combined organic portions
were extracted once with 20 mL of water. The combined aqueous
phases were then cooled to 0° C, and 2.92 g (27.1 mmol) of Na2CO3
was added, followed by 1.40 mL (6.06 mmol) of di-tert-butyldicar-
bonate in 33 mL of dioxane, and the resultant slurry was warmed to
room temperature, stirred vigorously for 25 h, and then partitioned
between ether and water. The aqueous layer was washed once with
ether, and the combined organic phases were back-extracted once with
water. The combined aqueous phases were then acidified with 1 M
NaHSO4 and extracted twice with EtOAc. The combined EtOAc phases
were washed twice with water and once with brine, dried over MgSO4,
filtered, and concentrated to a clear oil. Chromatography (30 g silica/g
sample, 1:49:50 HOAc:EtOAc:hexanes) gave 1.27 g (88%) of 18 as a
colorless oil: Rf 0.35 (1:10 MeOH:CH2Cl2); IR (thin film) 3355 br,
1
to give a colorless oil that was pure according to H NMR analysis.
The oil was immediately dissolved in dry ether, and a methanolic
solution of 37.6 mmol of HCl (prepared by carefully dissolving 2.7
mL (37.6 mmol) of acetyl chloride in 10 mL of MeOH at 0 °C) was
added to generate the HCl salt. The solvent was removed under vacuum
to give a white residue, which was then dissolved in CH2Cl2, diluted
with hexanes, and reconcentrated to give 7.25 g (91%) of 23 as a white
solid that was pure according to 1H NMR. Crystallization from MeOH/
EtOAc gave white needles: mp 124-126 °C; IR (thin film) 3413 br,
1
2956, 1745, 1234 cm -1; H NMR (500 MHz, D2O) δ 1.33 (d, J )
7.5, 3H), 3.49 (dq, J ) 4.1, 7.5, 1H), 3.76 (s, 3H), 3.85 (s, 3H), 4.51
(d, J ) 4.1, 1H); MS m/e calcd for (M - Cl-)+: 176.0919. Found
176.0917. Anal. Calcd for C7H14NO4Cl: C, 39.72; H, 6.68; N, 6.62.
Found: C, 39.72; H, 6.72; N, 6.50.
D-erythro-â-methylaspartic Acid â-Methyl Ester Hydrochloride
(24). To 2.00 g (9.45 mmol) of 23 in 32 mL of water and 8 mL of
EtOH was added 2.52 g (11.4 mmol) of basic cupric carbonate
(CuCO3‚Cu(OH)2), and the mixture was stirred vigorously at 70 °C
for 2.5 h. The mixture was then filtered through Celite, and the filter
pad was rinsed with copious water. The resultant blue solution was
diluted with more water to bring the final volume to 60 mL. Hydrogen
sulfide was then bubbled through the solution until an aliquot filtered
through Celite was colorless. After filtering the remaining black
mixture through Celite, the resultant clear, colorless solution was
concentrated under vacuum at 50 °C to constant mass, which provided
2977, 2611, 1711 cm-1 1H NMR (500 MHz, DMSO-d6) δ (major
;
rotamer) 1.04 (d, J ) 7.5 Hz, 3H), 1.37 (s, 9H), 2.81 (dq, J ) 6.5,7
hz, 1H), 3.60 (s, 3H), 4.37 (dd, J ) 6, 8.5 Hz, 1H), 6.95 (d, J ) 8.5
Hz, 1H); MS m/e calcd for (M+H)+: 262.1290, found 262.1289.
analysis calcd for C11H19NO6: C, 50.56; H, 7.34; N, 5.36. Found: C,
50.28; H, 7.19; N, 5.07.
Alanine Trichloroethylester, p-Toluene Sulfonate.46a To 6.78 g
(76 mmol) of L-alanine in 150 mL of toluene in a round bottom flask
equipped with a Dean-Stark trap was added 68 mL of trichloroethanol
and 21.6 g (114 mmol) of p-toluenesulfonic acid. The mixture was
refluxed with azeotropic water removal for 15 h, then cooled to room
1
1.83 g (98%) of 24 as a white solid that was pure according to H
NMR analysis: mp 270-280 °C (dec); IR (KBr) 3425-2367 br, 1719,
1
1619 cm-1; H NMR (500 MHz, D2O) δ 1.27 (d, J ) 7.3, 3H), 3.33
(dq, J ) 4.0, 11.1, 1H), 3.71 (s, 3H), 4.25 (d, J ) 3.74, 1H); MS m/e
calcd for C6H12NO4 (M - Cl-)+: 162.0763. Found 162.0768. Anal.

11768 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Humphrey et al.
Calcd for C8H13NO4: C, 36.46; H, 6.13; N, 7.09. Found: C, 36.16;
H, 6.33; N, 6.97.
concentrated under vacuum to constant mass to give 248 mg (100%)
of 28 as a colorless oil: Rf
0.65 (40:60 EtOAc:hexanes); IR (thin film)
3377, 2967, 2833, 2722, 1722, cm-1; H NMR (500 MHz, CDCl3) δ
1.28 (d, J ) 7.5 Hz, 3H), 1.47 (s, 9H), 3.31 (dq, J ) 3.5, 7.5 Hz, 1H),
3.68 (s, 3H), 4.34 (dd, J ) 3.5, 9.5 Hz), 5.60 (d, J ) 9.0 Hz), 9.65 (s,
1H); MS m/e calcd for (M+H)+: 246.1342, found 246.1338. The
material was used immediately in the next step without further
purification.
1
Boc-D-erythro-â-methylaspartate â-Methyl Ester (25). To 5.29
g (26.8 mmol) of 24 in 53 mL of 1 M pH 12.5 phosphate buffer
(resultant pH ) 8.0) was added 53 mL of dioxane, and the resultant
mixture was cooled in an ice bath, and 1.5 g (26.7 mmol) of KOH was
added. When the base had dissolved, 7.01 g (32.2 mmol) of di-tert-
butyldicarbonate was added in one portion, and 10 min later the mixture
was allowed to warm to room temperature where stirring was continued
for 1.5 h (resultant pH ) 7.0). At this point, an additional 1.5 g (26.7
mmol) of KOH was added and stirring was continued for 8 h (final
pH ) 9.0). The mixture was washed with one portion of ether, and to
the remaining aqueous layer was added a solution of 6.6 mL of
concentrated HCl in 50 mL of 1 M pH 3.1 citrate buffer (final pH )
3.5). The mixture was then extracted twice with ether, and the
combined organic extracts were washed with brine, dried over MgSO4,
filtered, and concentrated to give a colorless oil. Crystallization from
EtOAc/hexanes at -20 °C gave 6.43 g of a white solid that was
contaminated by a trace of ninhydrin-positive baseline material ac-
cording to TLC analysis (4:1:95 MeOH:HOAc:CHCl3). Recrystalli-
zation (EtOAc/hexanes, -20 °C) gave 5.20 g of pure white crystals.
The mother liquors were combined and chromatographed to supply an
additional 1.79 g for a total of 6.99 g, 99% of 25: mp 72-73 °C; IR
trans-(3R,4S)-N-tert-Butoxycarbonyl-1-iodo-3-amino-4-car-
bomethoxy-1-pentene (22). The crude aldehyde 28 (248 mg 1.02
mmol) in 7.5 mL of dioxane was cannulated into a flask equipped with
a stirbar and charged with a slurry of 750 mg (6.09 mmol) of flame-
dried chromium(II) chloride in 1.3 mL of THF. To the resultant green
slurry was added 1.04 g (3.04 mmol) of iodoform, and the resultant
dark brown mixture was stirred vigorously in the dark for 19 h. The
mixture was then partitioned between ether and 50% saturated brine,
and the phases were separated. The aqueous phase was then saturated
with NaCl and extracted twice with ether. The combined organic
material was then washed twice with brine, dried over MgSO4, filtered,
and concentrated to give a yellow oil. The oil was chromatographed
(60 g silica/g sample, 1:9 EtOAc:hexanes) to give 197 mg (53%) of
22 as a colorless oil that solidified over time: Rf 0.63 (30:70 EtOAc:
hexanes); IR (thin film) 3422, 3044, 2978, 2300, 1711 cm-1; 1H NMR
(500 MHz, CDCl3) δ 1.21 (d, 7.5 Hz, 3H), 1.43 (s, 9H), 2.70 (m, 1H),
3.68 (s, 3H), 4.27 (m, 1H), 6.32 (dd, J ) 1, 14 Hz, 1H), 6.45 (dd, J )
6.5, 14 Hz, 1H); 13C NMR (125 MHz, CDCl3) δ 14.4, 28.3, 42.8, 51.9,
56.9, 78.4, 79.8, 144.0, 155.3, 174.8; MS m/e calcd for (M+H)+:
370.0516. Found 370.0516. Anal. Calcd for C12H20INO4: C, 39.03;
H, 5.47; N, 3.79. Found: C, 39.32; H, 5.47; N, 3.74.
1
(thin film) 3300-2600 br, 1724, cm-1; H NMR (500 MHz, CDCl3,
major rotamer of a 10:1 mixture) δ 1.27 (d, J ) 7.4 Hz, 3H), 1.44 (s,
9H), 3.27 (dq, J ) 3.7, 7.2, 1H), 3.70 (s, 3H), 4.53 (dd, J ) 3.7, 9.4,
1H), 5.51 (d, J ) 9.4, 1H); MS m/e calcd for (M+H)+: 262.1290.
Found 262.1298. Anal. Calcd for C11H19NO6 : C, 50.56; H, 7.34; N,
5.36. Found: C, 50.77; H, 7.42; N, 5.06.
(2S,3R)-N-tert-Butoxycarbonyl-2-methyl-3-amino-4-hydroxy Meth-
yl Butanoate (27). To 4.78 g (18.3 mmol) of 25 in 19 mL of dry
Glyme was added 2.21 g (19.2 mmol) of N-hydroxysuccinimide. After
cooling in an ice bath, 3.96 g (19.2 mmol) of DCC was added. The
mixture was stirred for 1 h at 0 °C and for 2.5 h at room temperature,
then diluted with 20 mL of dry ether, filtered through Celite, and
concentrated to give the active ester 26 as white solid. The active ester
was crystallized from ether/hexanes for characterization, but, since
crystallization or chromatography resulted in lower overall yields, it
was generally used crude in the next step. Spectral data for 26: IR
(thin film) 3353, 2980, 1740, 1710 cm-1; 1H NMR (500 MHz, CDCl3)
δ 1.36 (d, J ) 7.3 Hz, 1H), 1.46 (s, 9H), 2.83 (br s, 4H), 3.35 (dq, J
) 3.9, 7.3 Hz, 1H), 3.75 (s, 3H), 4.89 (dd, J ) 3.9, 10.0 Hz, 1H), 5.60
(d, 10.0H); MS m/e calcd for (M+H)+: 359.1448, found 359.1455.
Reduction: the crude active ester 26 was dissolved in 17 mL of dry
THF and cooled in an ice bath, and then 1.73 g (45.7 mmol) of NaBH4
was added. After 2 h, the mixture was warmed to room temperature
and stirred vigorously for an additional 15 h. The mixture was then
added dropwise to 400 mL of an ice cold, vigorously stirred solution
of 1 M pH 3.1 citric acid buffer. The resultant mixture was extracted
twice with ether, and the combined extracts were washed once with
water, once with brine, dried over MgSO4, filtered, and concentrated
to give an oil. Chromatography (1:1 EtOAc:hexanes) gave 3.30 g
(73%) of the alcohol 27 as a white solid: Rf 0.27 (1:1 EtOAc:hexanes);
IR (thin film) 3420 br, 2978, 1718 cm-1; 1H NMR (500 MHz, CDCl3)
δ 1.23 (d, J ) 7 Hz, 3H), 1.43 (s, 9H), 2.87 (dq, J ) 5.5, 6.5 Hz, 1H),
3.61 (dd, J ) 5.5, 11 Hz, 1H), 3.65 (dd, J ) 5.5, 11 Hz, 1H), 3.68 (s,
3H), 3.76 (overlap app dq, J ) 5.5, 8.5 Hz, 1H), 5.42 (d, J ) 8.5 Hz,
1H); 13C NMR (125 MHz, CDCl3) δ 14.9, 28.3, 39.7, 51.9, 54.8, 63.9,
79.7, 156.4, 175.9; MS m/e calcd for (M+H)+: 248.1498. Found
248.1500.
(2R,3S)-N-Methyl-N-methoxy-2-methyl-3-hydroxy-4-phenylbu-
tanamide (Weinreb Amide 31). To a slurry of 7.90 g (81.0 mmol)
of N,O-dimethyl hydroxylamine hydrochloride and 35 mL of THF in
a flame dried flask at 0 °C was added dropwise over 20 min 40.5 mL
(81.0 mmol) of 2 M trimethylaluminum in hexanes (caution: vigorous
effervescence), and the resultant solution was stirred for 30 min. A
solution of 10.0 g (29.2 mmol) of 30 in 20 mL of dry THF was then
added over 15 min, and, after stirring for 1 h, the solution was
cannulated into 350 mL of rapidly stirred, ice-cold 1 M HCl (caution:
vigorous effervescence). The mixture was then extracted twice with
400 mL of EtOAc, and the combined organic phases were washed once
each with water and brine, dried over MgSO4, filtered, and concentrated
under vacuum to give a clear oil. Chromatography (100 g silica/g
sample, 4:6 EtOAc:hexanes) gave 6.06 g (87%) of 31 as a clear oil:
Rf 0.55 (50:50 EtOAc:hexanes); IR (thin film) 3424, 3036, 2931, 1637
1
cm-1; H NMR (500 MHz, CDCl3) δ 1.23 (d, J ) 7.0 Hz, 3H), 2.71
(dd, J ) 7.0, 13.5 Hz, 1H), 2.85 (par obsc, 1H), 2.89 (dd, J ) 7.5,
13.5 Hz, 1H), 3.46 (br s, 3 H), 3.65 (br s, 1H), 4.11 (dt, J ) 3, 7.0 Hz,
1H), 7.21-7.30 (m, 5H); 13C NMR (125 MHz, CDCl3) δ 10.03, 31.75,
37.21, 40.01, 61.20, 72.86, 126.3, 128.4, 129.1, 138.3, 177.9; MS m/e
calcd for (M+H)+: 238.1443. Found 238.1447. Anal. Calcd for
C13H19NO3: C, 65.78; H, 8.09; N, 5.90. Found: C, 65.78; H, 8.04; N,
5.81.
(2R,3S)-N-Methyl-N-methoxy-2-methyl-3-methoxy-4-phenylbu-
tanamide (Weinreb Amide 32). To 3.81 g (16.1 mmol) of 31 in 56
mL of dry THF and 24 mL of dry DMF was added 10.0 mL (160
mmol) of iodomethane. The resultant solution was cooled to 0 °C,
and 1.93 g (48.1 mmol) of 60% sodium hydride in oil was added. After
stirring for 1.5 h, the solution was poored into 300 mL of ice cold 1 M
pH 7 phosphate buffer, and the mixture was extracted twice with 250
mL of EtOAc. The combined organic phases were washed once each
with water and brine, dried over MgSO4, filtered, and concentrated
under vacuum to give an oil. Chromatography (25 g silica/g sample,
25:75 EtOAc:hexanes) gave 3.84 g (95%) of 32 as a clear oil: Rf 0.45
(2R,3S)-N-tert-Butoxycarbonyl-2-amino-3-carbomethoxybuta-
nal (28). To 193 mg (1.52 mmol) of freshly distilled oxalyl chloride
in 2.5 mL of CH2Cl2 at -50° C was added 173 mg (2.03 mmol) of dry
DMSO. After 15 min, 250 mg (1.02 mmol) of 27 in 7 mL of CH2Cl2
was added via cannula over 10 min. The resultant cloudy mixture was
stirred for 1 h, at which point 416 mg (4.07 mmol) of dry triethylamine
in 1.5 mL of CH2Cl2 was added dropwise. The reaction was quenched
after 40 min by the addition of 0.5 mL of water and was then diluted
with 20 mL of hexanes. The mixture was washed twice with 5 mL of
ice cold 1 M NaHSO4, and the combined aqueous phases were back
extracted once with ether. The combined organic material was then
washed once with 10 mL of saturated NaHCO3, twice with 10 mL of
water, and once with 10 mL of brine, dried over MgSO4, filtered, and
(40:60 EtOAc:hexanes); IR (thin film) 3060, 2931, 2813, 1660 cm-1
;
1H NMR (500 MHz, CDCl3) δ 1.21 (d, J ) 7.0 Hz, 3H), 2.74 (dd, J
) 7.0, 14 Hz, 1H), 2.86 (dd, J ) 3.5, 14 Hz, 1H), 2.90 (obsc dq, 1H),
3.15 (s, 3H), 3.24 (s, 3H), 3.44 (br s, 3H), 3.61 (app dt, J ) 4, 7 Hz,
1H), 7.17-7.23 (m, 5H); 13C NMR (125 MHz, CDCl3) δ 13.43, 32.01,
38.48, 39.62, 58.84, 60.95, 83.75, 125.97, 128.04, 129.51, 138.77,
175.95; MS m/e calcd for (M+H)+: 252.1600. Found 252.1597. Anal.
Calcd for C14H21NO3: C, 66.89; H, 8.44; N, 5.57. Found: C, 66.79;
H, 8.35; N, 5.53.

Synthesis of Microcystin-LA
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11769
(2R,3S)-2-Methyl-3-methoxy-4-phenylbutanal (33). To 1.94 g
(7.72 mmol) of 32 in 53 mL of dry CH2Cl2 at -100 °C was added
over 10 min 10.3 mL (15.4 mmol) of 1.5 M diisobutylaluminum hydride
in toluene. After 1.5 h, 2 mL of dry methanol was added, and the
solution was warmed to 0 °C. To the resultant solution was added 50
mL of ice-cold 1 M tartaric acid, and the mixture was stirred vigorously
for 20 min at 0 °C. The phases were separated, and the aqueous phase
was extracted twice with CH2Cl2. The combined organic phases were
dried over MgSO4, filtered through Celite, and concentrated under
vacuum to a clear oil. Chromatography (100 g silica/g sample, 6:94
EtOAc:hexanes) gave 1.17 g (79%) of 33 as a clear oil: Rf 0.36 (10:
90 EtOAc:hexanes); IR (thin film) 3055, 2933, 2822, 2711, 1722 br,
obsc. mult., 1H), 3.12 (dd, J ) 4.5, 13.5 Hz, 1H), 3.37 (ddd, J ) 2,
4.5, 10 Hz, 1H), 3.42 (s, 3H), 6.21 (q, J ) 7 Hz, 1H), 7.20-7.33 (m,
5H). Chromatography fractions containing unreacted starting material
were purified via HPLC (0.5:99.5 EtOAc:hexanes) to give 69 mg (18%)
of recovered 34b.
Boc-Adda-OMe (35). To 184 mg (0.74 mmol) of 20 was added
222 mg (0.60 mmol) of 22, 5 mL of THF, and 43 mg (2.4 mmol) of
water. The colorless solution was cooled to 0 °C, and 49 mg (0.07
mmol) of Pd[PPh3]4 in 1.5 mL of dry THF was added via cannula.
After 5 min, 377 mg (1.51 mmol) of TlOEt was added to the orange
solution over a 10 min period. After 2 h, the light brown suspension
was partitioned between EtOAc and water, and the aqueous phase was
extracted twice with EtOAc. The combined organic phases were
washed twice with water and once with brine, dried over MgSO4,
filtered, and concentrated under vacuum to give a dark residue. The
residue was filtered through silica (10:90 EtOAc:hexanes) to give an
orange oil, which was then purified by HPLC (4:96 EtOAc:hexanes)
to give 219 mg (82%) of 35 as a yellow oil: Rf 0.39 (20:80 EtOAc:
hexanes); IR (thin film) 3424, 3025, 2931, 1713 cm-1; 1H NMR (500
MHz, CDCl3) δ 1.02 (d, J ) 6.5 Hz, 3H), 1.22 (d, J ) 7 Hz, 3H), 1.45
(s, 9H), 1.59 (d, J ) 1 Hz, 3H), 2.58 (ddq, J ) 6.5, 7, 9.5 Hz, 1H),
2.67 (dd, J ) 7.5, 14 Hz, 1H), 2.72-2.77 (m, 1H), 2.80 (dd, J ) 4.5,
13.5 Hz, 1H), 3.18 (ddd, J ) 4.5, 6.5, 7.5 Hz, 1H), 3.23 (s, 3H), 3.66
(s, 3H), 4.36 (br, 1H), 5.27-5.30 (m, 1H), 5.38 (d, J ) 9.5 Hz, 1H),
5.43 (dd, J ) 6.5, 15.5 Hz, 1H), 6.18 (d, J ) 15.5 Hz, 1H), 7.17-7.28
(m, 5H); 13C NMR (125 MHz, CDCl3) δ 12.6, 14.3, 16.1, 28.3, 36.6,
38.2, 44.0, 51.6, 54.4, 58.6, 79.2, 86.9, 125.1, 125.9, 128.1, 129.4, 132.4,
135.9, 136.1, 139.3, 155.5, 175.3; MS m/e calcd for (M+H)+:
446.2906. Found 446.2899.
1
1600 cm-1; H NMR (500 MHz, CDCl3) δ 1.18 (d, J ) 7 Hz, 3H),
2.39 (dq, J ) 3.5, 7 Hz, 1H), 2.72 (dd, J ) 7, 13.5 Hz, 1H), 3.00 (dd,
J ) 6.5, 13.5 Hz, 1H), 3.26 (br s, 3H), 3.90 (ddd, J ) 3.5, 6.5, 6.5 Hz,
1H), 7.21-7.32 (m, 5H), 9.69 (s, 1H); 13C NMR (125 MHz, CDCl3) δ
7.5, 37.4, 48.9, 58.0, 81.6, 126.5, 128.6, 129.2, 138.0, 204.3; MS m/e
calcd for (M+NH4)+: 210.1495. Found 210.1503.
Dibromoalkene 34a. To 7.59 g (22.9 mmol) of carbon tetrabromide
in 35 mL of dry CH2Cl2 at -20° C was added over 5 min 11.1 g (42.3
mmol) of triphenylphosphine in 12 mL of dry CH2Cl2. After 15 min
the resultant orange solution was cooled to -78 °C, and 2.20 g (11.4
mmol) of 33 in 11 mL of dry CH2Cl2 was added by cannula over 10
min. The solution faded to a light yellow after 10 min, at which point
it was concentrated under vacuum to a brown crust. Chromatography
(80 g silica/g sample, 1:99 EtOAc:hexanes) gave 3.80 g (95%) of 34a
as a colorless oil: Rf 0.39 (5:95 EtOAc:hexanes); IR (thin film) 3055,
3022, 2922, 2822, 1605 cm-1; 1H NMR (500 MHz, CDCl3) δ 1.07 (d,
J ) 7 Hz, 3H), 2.59 (ddq, J ) 5, 7, 9.5 Hz, 1H), 2.73 (dd, J ) 5.5, 14
Hz, 1H), 2.81 (dd, J ) 7, 14 Hz, 1H), 3.25 (s, 3H), 3.30 (ddd, J ) 5.5,
5.5, 7.5 Hz, 1H), 6.36 (d, J ) 9.5 Hz, 1H), 7.20-7.31 (m 5H); 13C
NMR (125 MHz, CDCl3) δ 13.7, 37.9, 41.6, 58.6, 85.0, 88.7, 126.2,
128.4, 129.3, 138.7, 141.2; MS m/e calcd for (M+NH4)+: 363.9912.
Found 363.9902.
Boc-Adda (4). To 80 mg (0.18 mmol) of 35 was added 4.7 mL of
THF and 0.91 mL (0.90 mmol) of 1 M LiOH. The resultant mixture
was stirred vigorously for 50 h at ambient temperature and then was
partitioned between ether and water. The aqueous phase was washed
with ether, and the combined ether phases were back-extracted once
with water. The combined aqueous phases were then acidified with 1
M NaHSO4 and extracted twice with EtOAc. The combined EtOAc
phases were washed twice with water and once with brine, dried over
MgSO4, filtered, and concentrated to an oil. The sample was filtered
through silica (20:79:1 EtOAc/hexanes/HOAc) to give 68 mg (88%)
of 4 as an oil: Rf 0.44 (25:74:01 EtOAc:hexanes:HOAc); IR (thin film)
(4R,5S)-4-Methyl-5-methoxy-6-phenyl-2-hexyne (34b). To 2.55
g (7.33 mmol) of 34a in 100 mL of dry THF at -78 °C was added
dropwise over 10 min 13.7 mL (25.6 mmol) of 1.87 M nBuLi in
hexanes, and the solution was stirred for 1 h. To the resultant purple
solution was added 4.6 mL (73.4 mmol) of iodomethane, and the
solution was then warmed to ambient temperature and stirred for 2 h.
The solution was then partitioned between ether and 5% NaSO3, and
the phases were separated. The organic phase was washed once each
with water and brine, dried over MgSO4, filtered, and concentrated
under vacuum to an oil. Purification via MPLC (1:99 EtOAc:hexanes)
gave 1.31 g (89%) of 34b as a colorless oil: Rf 0.24 (1:99 EtOAc:
hexanes); IR (thin film) 3066, 2922, 1944, 1600, cm-1; 1H NMR (500
MHz, CDCl3) δ 1.19 (d, J ) 7 Hz, 3H), 1.83 (d, J ) 2.5 Hz, 3H), 2.55
(ddq, J ) 2.5, 6.5, 7 Hz, 1H), 2.85 (dd, J ) 7.5, 14 Hz, 1H), 2.98 (dd,
J ) 4, 14 Hz, 1H), 3.24 (ddd, J ) 4, 6.5, 7 Hz, 1H), 3.27 (s, 3H),
7.20-7.30 (m, 5H); 13C NMR (125 MHz, CDCl3) δ 3.7, 17.4, 30.1,
37.9, 58.5, 77.5, 81.1, 85.9, 126.0, 128.1, 129.6, 139.2; MS m/e calcd
for (M+NH4)+: 220.1702. Found 220.1700. Anal. Calcd for
C14H18O: C, 83.11; H, 8.99. Found: C, 83.42; H, 9.07.
1
3055, 2966, 2866, 1727 br cm-1; H NMR (500 MHz, DMSO-d6) δ
0.94 (d, J ) 7 Hz, 3H), 0.96 (d, J ) 7 Hz, 3H), 1.35 (s, 9H), 1.51 (br
s, 3H), 2.52 (par obsc m, 2H), 2.63 (dd, J ) 7, 14 Hz, 1H), 2.72 (dd,
J ) 5, 14 Hz, 1H), 3.16 (s, 3H), 3.23 (ddd, J ) 5, 5.5, 7 Hz, 1H), 4.22
(app q, J ) 8 Hz, 1H), 5.37 (par obsc dd, J ) 7.5, 15.5 Hz, 1H), 5.39
(d, J ) 10.5 Hz, 1H), 6.08 (d, J ) 15.5 Hz, 1H), 6.91 (d, J ) 8.5 Hz,
1H), 7.15-7.27 (m, 5H); MS m/e calcd for (M+H)+: 432.2751. Found
432.2761.
Boc-D-iso-Glu(OMe)-Me∆Ala-(D)-Ala-(L)-Leu-(D)-erythro-â-Me-
iso-Asp(OMe)-(L)-Ala-OTce (38). Dipeptide 19 (314 mg, 0.68 mmol)
was mixed with 2 mL of TFA for 30 min and concentrated, and the
residue was partitioned between EtOAc and saturated NaHCO3. The
aqueous layer was extracted with two additional portions of EtOAc,
and the combined organic material was washed with brine, dried with
K2CO3, filtered, and concentrated. To the residue was added 431 mg
(0.82 mmol) of the tetrapeptide 2 and 4 mL of toluene. The mixture
was then concentrated, and the residue was dissolved in 2 mL of DMF.
To the solution was added 247 mg (2.04 mmol) of collidine and, after
cooling to 0 °C, 310 mg (0.82 mmol) of HATU. The solution was
stirred for 1 h at 0 °C, and for 19 h at room temperature. The mixture
was worked up as for compound 13, substituting 1:1 Ether:EtOAc for
the extractions, to give 729 mg of crude product. Chromatography
(5:95 isopropyl alcohol:CH2Cl2 to elute the nonpolar impurities, then
10:90 isopropyl alcohol:CH2Cl2 to elute the product) gave 476 mg
(80%) of 38: Rf 0.49 (1:9 isopropyl alcohol:EtOAc) IR (thin film) 3301,
2957, 1747, 1651 cm-1; 1H NMR (500 MHz, DMSO-d6, 90 °C) δ 0.83
(d, J ) 6.4, 3H), 0.87 (d, J ) 6.8, 3H), 1.06 (d, J ) 7.2, 3H), 1.30 (d,
J ) 7.2, 3H), 1.36 (obsc d, 3H), 1.37 (s, 9H), 4.69 (par obsc t, J ) 7.2,
2H), 1.60 (m, 1H), 1.82 (m, 1H), 1.96 (m, 1H), 2.32 (m, 2H), 2.96
(par obsc m, 1H), 3.00 (br s, 3H), 3.60 (s, 3H), 3.61 (s, 3H), 3.97 (br
app q, J ) 5.6, 1H), 4.28 (app q, J ) 7.2, 1H), 4.37 (par obsc m, 1H),
4.38 (par obsc m, 1H), 4.52 (dd, J ) 5.6, 8.4, 1H),4.80 (d, J ) 12.3,
1H), 4.92 (d, J ) 12.3, 1H), 5.52 (br s, 1H), 5.96 (br s, 1H), 6.80 (br
Boronic Acid 20. To 400 mg (1.98 mmol) of 34b in 1 mL of dry
CH2Cl2 was added 545 mg (1.98 mmol) of Ag2CO3, and the mixture
was cooled to 0 C. To the mixture was then added 6.0 mL (6.0 mmol)
of 1 M Br2BH‚SMe2 in CH2Cl2, and the resultant mixture was stirred
vigorously in the dark for 11 h at 0 °C. The mixture was then poured
into 40 mL of rapidly stirring ice cold 0.5 M pH 3.1 citrate buffer and
extracted twice with EtOAc. The combined organic phases were
washed once with water and twice with brine, dried over MgSO4,
filtered, and concentrated under vacuum to an oil. Rapid chromatog-
raphy (100 g silica/g oil, 25:75 EtOAc:hexanes) gave 220 mg (45%)
of 20 as a colorless oil, which required immediate use to avoid
significant decomposition: Rf 0.20 (30:70 EtOAc:hexanes); IR (thin
1
film) 3031, 2940, 2858, 1628, cm-1; H NMR (500 MHz, CDCl3) δ
1.07 (d, J ) 6.5 Hz, 3H), 1.70 (d, J ) 1.5 Hz, 3H), 2.73 (dd, J ) 8,
14 Hz, 1H), 2.78 (par obsc mult., 1H), 2.85 (dd, J ) 4.5, 14 Hz, 1H),
3.27 (obsc. mult., 1H), 3.27 (s, 3H), 6.64 (dd, J ) 1.5, 9.5 Hz, 1H),
7.19-7.29 (m, 5H). Also isolated during the chromatography was 54
mg (11%) of the undesired isomer 21 as a clear oil: Rf 0.32 (30:70
1
EtOAc:hexanes); H NMR (500 MHz, CDCl3) δ 1.08 (d, J ) 7 Hz,
3H), 1.22 (d, J ) 7 Hz, 3H), 2.66 (dd, J ) 10, 13.5 Hz, 1H), 2.68 (par.

11770 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Humphrey et al.
s, 1H), 7.74 (br d, J ) 8.0, 1H), 7.88 (br m, 2H), 8.29 (br d, J ) 6.8,
1H); FAB MS calcd for C35H56Cl3N6O13 (M+H)+: 873.2970.
Found: 873.2969.
EtOAc. The solution was cooled in an ice bath, and 13 mg (0.064
mmol) of DCC was added. The mixture was stirred for 2 h at 0 °C
and for 22 h at room temperature, then filtered through Celite,
concentrated, and passed through a plug of dry silica gel (10:1 EtOAc:
isopropyl alcohol) to give 50 mg (83%) of 41: Rf 0.38 (9:1 EtOAc:
Boc-Adda-(D)-iso-Glu(OMe)-Me∆Ala-(D)-Ala-(L)-Leu-(D)-erythro-
â-Me-iso-Asp(OMe)-(L)-Ala-OTce (3). Compound 38 (94 mg 0.11
mmol) was mixed with 1 mL of TFA for 30 min. The solution was
then concentrated, and the resultant oil was partitioned between EtOAc
and saturated aqueous NaHCO3. The aqueous portion was extracted
with two additional portions of EtOAc, and the combined organic
extracts were then washed with brine, dried with K2CO3, filtered, and
concentrated. To the resultant residue was added 40 mg (0.090 mmol)
of Boc-Adda-OH as a solution in CH2Cl2. The solution was again
concentrated and then reconcentrated from 2 mL of toluene. To the
resultant residue was added 0.5 mL of DMF and 33 mg (0.27 mmol)
of collidine. After cooling the solution in an ice bath, 41 mg (0.11
mmol) of HATU was added. After stirring the solution for 1 h at 0 °C
and 14 h at room temperature, 1 mL of saturated NaHCO3 was added.
The mixture was stirred for 10 min and then was partitioned between
ether and water. The aqueous layer was extracted with two additional
portions of ether, and the combined organic layers were washed with
brine, dried with MgSO4, and concentrated. Chromatography (EtOAc
to elute nonpolar material, then 1:10 isopropyl alcohol:EtOAc to elute
the product) gave 74 mg (69%) of 3: Rf 0.39 (10:5 EtOAc:isopropyl
+
acetone); FAB MS calcd for C53H81N7O15 (loss of C6F5): 1056.
Found: 1056. The active ester 41 was used in the next step without
further purification: to 33 mg (0.027 mmol) of 41 was added 1 mL of
freshly distilled TFA. After 50 min, the mixture was concentrated and
then reconcentrated twice from CHCl3. The residue was then dissolved
in 27 mL of CHCl3 and added dropwise over 40 min to a vigorously
stirred mixture of 27 mL of CHCl3 and 27 mL of 0.5 M, pH 9.5
phosphate buffer. After stirring the mixture for an additional hour,
the layers were separated. The chloroform layer was passed through
a plug of cotton, and the aqueous portion was extracted twice with
EtOAc. The combined EtOAc layers were washed with brine and added
to the CHCl3 portion. The solution was then dried with MgSO4 and
concentrated. Reversed phase HPLC (52:48 CH3CN:10 mM HOAc)
provided 14 mg (56%) of 42. Rf 0.35 (2:8 isopropyl alcohol:CH2Cl2);
IR (thin film) 3296, 2927, 1741, 1649 cm-1 1H NMR (500 MHz,
;
CDCl3) δ 0.844 (d, J ) 6.4, 3H), 0.900 (d, J ) 6.4, 3H), 1.02 (d, J )
6.8, 3H), 1.12, 1.13 (two ovrlp d, J ) 6.8, 7.2, 6H), 1.29 (d, J ) 7.2,
3H), 1.37 (d, J ) 7.2, 3H), 1.56 (obsc m, 2H), 1.60 (s, 3H), 1.66 (m,
1H), 1.75 (m, 2H), 2.21 (m, 1H), 2.45 (m, 1H), 2.53-2.64 (m, 2H),
2.68 (dd, J ) 7.6, 13.9, 1H), 2.80 (obsc dd, 1H), 2.81 (obsc m, 1H),
3.20 (m, 1H), 3.24 (s, 3H), 3.31 (s, 3H), 3.69 (s, 3H), 3.94 (s, 3H),
4.18 (dq, J ) 6.0, 6.0, 1H), 4.38 (dq, J ) 8.0, 8.0, 1H), 4.56 (dd, J )
2.0, 8.4, 1H), 4.65 (app q, J ) 9.1, 1H), 4.72 (m, 1H), 4.86 (dq, J )
4.0, 1H), 5.28 (dd, J ) 8.8, 15.5, 1H), 5.30 (s, 1H), 5.41 (d, J ) 9.9,
1H), 5.99 (s, 1H), 6.25 (d, J ) 15.5, 1H), 6.73 (d, J ) 9.1, 1H), 6.81
(d, J ) 8.4, 1H), 6.89 (d, J ) 9.6, 1H), 7.19 (app t, J ) 7.6, 3H),
7.25-7.28 (obsc m, 2H), 7.90 (d, J ) 6.0, 1H), 8.28 (br d, J ) 8.4,
alcohol); IR (thin film) 3299, 2957, 1747 cm-1 1H NMR (500 MHz,
.
DMSO-d6, 70 °C) δ 0.82 (d, J ) 6.4, 3H), 0.86 (d, J ) 6.8, 3H), 0.96
(d, J ) 6.8, 3H),1.00 (d, J ) 7.2, 3H), 1.06 (d, J ) 7.2, 3H), 1.29 (d,
J ) 7.2, 3H), 1.36 (par obsc d, 3H), 1.37 (s, 9H), 1.50 (app t, J ) 6.8,
2H), 1.55 (d, J ) 0.8, 3H), 1.59 (m, 1H), 1.82 (m, 1H), 1.96 (m, 1H),
2.21-2.35 (m, 2H), 2.55 (m, 2H), 2.65 (dd, J ) 7.6, 14.0, 1H), 2.74
(dd, J ) 4.8, 14.0, 1H), 2.74 (dd, J ) 4.8, 14.0, 1H), 2.97 (par obsc m,
1H), 2.98 (br s, 3H), 3.17 (s, 3H), 3.24 (ovrlp dt, J ) 5.2, 7.2, 1H),
3.588 (s, 3H), 3.593 (s, 3H), 4.05 (br app q, J ) 6.0, 1H), 4.22 (br app
q, J ) 6.8, 1H), 4.27 (app q, J ) 7.6, 1H), 4.33-4.38 (m, 2H), 4.51
(dd, J ) 5.6, 8.8, 1H), 4.79 (d, J ) 11.9, 1H), 4.93 (d, J ) 11.9, 1H),
5.38 (d, J ) 9.5, 1H), 5.44 (dd, J ) 3.2, 15.9, 1H), 5.48 (br s, 1H),
5.98 (br s, 1H), 6.07 (d, J ) 15.9, 1H), 6.32 (br s, 1H), 7.15-7.19 (m,
3H), 7.24-7.27 (m, 2H), 7.79 (d, J ) 9.2, 1H), 7.92 (br ovrlp d, 2H),
8.01 (br d, J ) 7.2, 1H), 8.35 (d, J ) 7.2, 1H). FAB MS calcd for
C55H83N7O15Cl3+ (M+H)+: 1186. Found: 1186. FAB MS calcd for
the molecular envelope of C55H82N7O15Cl3Na+ (M + Na)+: 1208.5
(83.3%), 1209.5 (54.6%), 1210.5 (100.0%), 1211.5 (57.8%), 1212.5
(46.0%), 1213.5 (22.1%), 1214.5 (10.0%). Found: 1208.5 (83.2%),
1209.5 (57.0%), 1210.5 (100.0%), 1211.5 (57.8%), 1212.5 (44.0%),
1213.5 (20.0%), 1214.5 (6.4%).
+
1H) FAB MS calcd for C48H72N7O12 (M+H)+: 938.5238. Found:
938.5222.
Microcystin-LA (1). To 6.0 mg (0.0064 mmol) of 42 in 1 mL of
THF at 0 °C was added 1 mL of cold 0.1 M LiOH. After stirring
vigorously for 5 h at 0 °C, the mixture was acidified with 1 M NaHSO4
(ca. 4 drops) and extracted three times with EtOAc. The combined
organic portions were washed with brine, dried with MgSO4, filtered,
and concentrated. Preparative reversed phase HPLC (72:28 MeOH:
0.2% aqueous TFA) gave 3 mg (50%) of an unidentified isomer (FAB
MS calcd for C46H68N7O12+ (M+H)+ 910.4925. Found 910.4920) and
3 mg (50%) of microcystin-LA (1): IR (thin film) 3307, 2926, 1640,
Boc-Adda-D-iso-Glu(OMe)-Me∆Ala-(D)-Ala-(L)-Leu-(D)-erythro-
â-Me-iso-Asp(OMe)-(L)-Ala-OH (40). To 70 mg (0.0589 mmol) of
3 in 2 mL of HOAc was added 0.5 g of zinc powder. The mixture
was stirred vigorously for 4 h and then filtered through Celite. The
filter cake was rinsed with copious EtOAc, and the filtrate was
concentrated. Chromatography of the residue (10:1:0.1 CH2Cl2:
isopropyl alcohol:HOAc to elute the nonpolar material, then 10:2:0.1
to elute the product) gave 52 mg (84%) of 40: Rf 0.31 (10:2:0.1 CH2-
Cl2:isopropyl alcohol:HOAc); IR (thin film) 3301, 2958, 1743, 1650
cm-1; 1H NMR (500 MHz, DMSO-d6, 90°C) δ 0.82 (d, J ) 6.4, 3H),
0.86 (d, J ) 6.4, 3H), 0.97 (d, J ) 6.8, 3H), 1.01 (d, J ) 7.2, 3H),
1.04 (d, J ) 7.2, 3H), 1.25 (d J ) 7.2, 3H), 1.29 (d, J ) 7.2, 3H), 1.37
(s, 9H), 1.51 (app t J ) 7.6, 2H), 1.55 (s, 3H), 1.60 (par obsc app dq,
J ) 13.5, 6.8, 1H), 1.84 (app dq, J ) 14.3, 6.4, 1H), 1.96 (app dq, J
) 13.1, 6.4, 1H), 2.29 (m, 2H), 2.57 (m, 2H), 2.66 (dd, J ) 13.9, 7.2,
1H), 2.74 (dd, J ) 13.9, 4.8, 1H), 2.95 (par obsc m, 1H), 2.98 (br s,
3H), 3.18 (s, 3H), 3.24 (ovrlp t, J ) 5.2, 7.2, 1H), 3.59 (s, 3H), 3.60
(s, 3H), 4.06 (br app q, J ) 7.6, 1H) 4.14-4.28 (three ovrlp m, 3H),
4.37 (app dq, J ) 7.6, 1H), 4.48 (dd, J ) 5.6, 8.8, 1H), 5.38 (d, J )
9.6, 1H), 5.45 (par obsc dd, J ) 6.8, 15.9, 1H), 5.46 (par obsc br s,
1H), 5.98 (br s, 1H), 6.07 (d, J ) 15.5, 1H), 6.25 (br s, 1H), 7.18 (m,
3H), 7.25 (app t, J ) 7.6, 2H), 7.77 (br d, J ) 8.8, 1H), 7.82-7.86 (br
m, 2H), 7.92 (br d, J ) 6.8, 1H), 7.96 (br d, J ) 6.4, 1H); FAB MS
calcd for C53H81N7O15 (M+H)+: 1056.6. Found: 1056.6. FAB MS
calcd for the molecular envelope of C53H80NaN7O15 (M + Na)+: 1078.6
(100.0%), 1079.6 (63.4%), 1080.6 (22.7%), 1081.6 (5.9%). Found:
1078.6 (100.0%), 1079.6 (58.8%), 1080.6 (20.8%), 1081.6 (5.0%).
1504, 1257 cm-1 1H NMR (500 MHz, D2O; some resonances are
;
obscured by the solvent peak) δ 0.860 (d, J ) 6.7, 3H), 0.900 (d, J )
6.7, 3H), 1.02 (two overlp d, J ) 7.3, 6H), 1.05 (d, J ) 7.3, 3H), 1.28
(J ) 7.3, 3H), 1.39 (d, J ) 7.3, 3H), 1.67 (m, 2H), 1.70 (s, 3H), 1.90-
2.05 (m, 3H), 2.60 (m, 1H), 2.69-2.80 (three ovrlp m, 3H), 2.92 (dd,
J ) 5, 14, 1H), 3.01 (dd, J ) 7.9, 7.9, 1H), 3.16 (m, 1H), 3.27 (s, 3H),
3.39 (s, 3H), 3.47 (m, 1H), 3.96 (app t, J ) 7.9, 1H), 4.30 (m, 1H),
4.39 (m, 1H), 5.54 (s, 1H), 5.51-5.64 (par obsc m, 2H), 5.90 (s, 1H),
6.31 (d, J ) 17.1, 1H), 7.28 (app d, J ) 6.7, 3H), 7.36 (app t, J ) 6.7,
2H); FAB MS calcd for C46H68N7O12 (M+H)+ 910.4925. Found
910.4942. The 1H NMR spectrum of the synthetic sample was identical
to the 1H NMR spectrum of authentic microcystin-LA, and the synthetic
sample coeluted with authentic microcystin-LA by analytical reversed
phase HPLC (80:20 MeOH:0.2% aqueous TFA).
Acknowledgment. We are indebted to Professor Richard
E. Moore for an authentic sample of microcystin-LA. We are
grateful to Dr. Catherine J. Leblanc for early synthetic studies
of Adda and to Professor. A. J. Shaka and Gilbert Mackin for
assistance in NMR characterization of the final product. We
also thank Professor Peter Wipf, Dr. Michael F. Cristofaro, and
Dr. Robert J. Valentekovich for helpful discussions. Partial
support by the National Institutes of Health (NS-27600) is
gratefully acknowledged.
Microcystin-LA Dimethyl Ester (42). To 52 mg (0.049 mmol) of
40 was added 24 mg (0.13 mmol) of pentafluorophenol in 1 mL of
JA961683E