Synthesis of natural and modified trapoxins, useful reagents for exploring histone deacetylase function

Article abstract of DOI:10.1021/ja9615841

Trapoxin, a cyclotetrapeptide isolated from the fungus Helicoma ambiens, profoundly affects mammalian cell growth and morphology. In this paper, we describe syntheses of trapoxin, [³H]trapoxin, and K-trap, a trapoxin-based affinity reagent. The

Full text of DOI:10.1021/ja9615841

10412
J. Am. Chem. Soc. 1996, 118, 10412-10422
Synthesis of Natural and Modified Trapoxins, Useful Reagents
for Exploring Histone Deacetylase Function
Jack Taunton, Jon L. Collins, and Stuart L. Schreiber*
Contribution from the Howard Hughes Medical Institute, Department of Chemistry and
Chemical Biology, HarVard UniVersity, 12 Oxford Street Cambridge, Massachusetts 02138
ReceiVed May 13, 1996X
Abstract: Trapoxin, a cyclotetrapeptide isolated from the fungus Helicoma ambiens, profoundly affects mammalian
cell growth and morphology. In this paper, we describe syntheses of trapoxin, [3H]trapoxin, and K-trap, a trapoxin-
based affinity reagent. These reagents have allowed us to achieve the first molecular characterization of histone
deacetylase, trapoxin’s cellular target.
Cancer results from a sequence of genetic changes that subvert
the normal mechanisms of control over cell growth and
morphology. The tools of molecular genetics have proven
invaluable in the identification of genes whose mutation or loss
results in an increased frequency of cellular transformation and
cancer.1 However, the biochemical basis of cell cycle and
morphological deregulation has not been fully illuminated.
During the past ten years, several natural products have been
isolated on the basis of their ability to revert the abnormal
morphology of oncogenically transformed cell lines back to the
“flat” morphology of the nontransformed, progenitor cell lines.
This structurally diverse group of natural products includes
FR901228,2 depudecin,3 radicicol,4 and trapoxin B (1).5 Trapoxin
belongs to a small family of hydrophobic cyclotetrapeptides,
all of which share three structural features: (1) a proline or
pipecolinic acid residue; (2) at least one amino acid whose
R-stereocenter is of the (R) configuration; and (3) a nonprotein-
ogenic amino acid, (2S,9S)-2-amino-8-oxo-9,10-epoxydecanoic
acid (Aoe). Other members of this family include chlamydocin
(2),6 HC-toxin (3),7 WF-3161 (4),8 and Cyl-2 (5),9 each
projecting a variable array of hydrophobic side chains from a
12 atom cyclic backbone (Figure 1). Several groups have shown
that reduction or hydrolysis of the epoxide results in a complete
loss of biological activity, suggesting that trapoxin may ir-
reversibly inactivate its receptor through covalent bond forma-
tion.5,6,10 This notion has been challenged, however, with the
recent isolation of a fungal metabolite that is structurally
identical to chlamydocin, except that the (9S)-9,10-epoxide has
been replaced by a (9R)-hydroxy group.11 Although the activity
of this compound toward mammalian cells has not been
Figure 1. The Aoe cyclotetrapeptide family of natural products.
reported, it is active at 0.3 µM in a plant growth assay and thus
raises interesting questions concerning the requirement of the
epoxyketone element for biological activity.
Subsequent to the publication of trapoxin’s structure and
detransformation activity, Yoshida and co-workers demonstrated
that trapoxin potently and irreversibly inhibits a relatively
uncharacterized enzymatic activity, histone deacetylase, which
is present in crude cell lysates.12 In addition, they showed that
trapoxin blocks histone deacetylation in vivo at the same
concentrations required for detransformation activity and in vitro
histone deacetylase inhibition (<10 nM). These important
observations placed trapoxin within another class of natural
products, the histone deacetylase inhibitors,13 which prior to the
discovery of trapoxin, had only two members, butyric acid (6)
and trichostatin A (7) (Figure 2). The cellular effects of
trapoxin, trichostatin, and butyrate are strikingly congruent.
These include, depending on the cell type, detransformation,
cell cycle arrest, and differentiation, but whereas butyrate is only
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) Hunter, T. Cell 1991, 64, 249-70.
(2) Ueda, H.; Nakajima, H.; Hori, Y.; Fujita, T.; Nishimura, M.; Goto,
T.; Okuhara, M. J. Antibiot. 1994, 47, 301-10.
(3) Matsumoto, M. et al. J. Antibiot. 1992, 45, 879-85.
(4) Kwon, H. J.; Yoshida, M.; Fukui, Y.; Horinouchi, S.; Beppu, T.
Cancer Res. 1992, 52, 6926-30.
(5) Itazaki, H. et al. J. Antibiot. 1990, 43, 1524-32.
(6) Closse, A.; Huguenin, R. HelV. Chim. Acta. 1974, 57, 533-45.
(7) Pringle, R. B. Plant Physiol. 1970, 46, 45-9.
(8) Umehara, K.; Nakahara, K.; Kiyoto, S.; Iwami, M.; Okamoto, M.;
Tanaka, H.; Kohsaka, M.; Aoki, H.; Imanaka, H. J. Antibiot. 1983, 36,
478-83.
(9) Hirota, A.; Suzuki, A.; Suzuki, H.; Tamura, S. Agric. Biol. Chem.
1973, 37, 643-7.
(10) Ciuffetti, L. M.; Pope, M. R.; Dunkle, L. D.; Daly, J. M.; Knoche,
H. W. Biochemistry 1983, 22, 3507-10.
(12) Kijima, M.; Yoshida, M.; Sugita, K.; Horinouchi, S.; Beppu, T. J.
Biol. Chem. 1993, 268, 22429-35.
(13) Yoshida, M.; Horinouchi, S.; Beppu, T. Bioessays 1995, 17, 423-
30.
(11) Gupta, S.; Peiser, G.; Nakajima, T.; Hwang, Y.-S. Tetrahedron Lett.
1994, 35, 6009-12.
S0002-7863(96)01584-3 CCC: $12.00 © 1996 American Chemical Society

Synthesis of Natural and Modified Trapoxins
J. Am. Chem. Soc., Vol. 118, No. 43, 1996 10413
Scheme 1
influenced by the chlamydocin (2) synthesis published by
Schmidt and co-workers.16 Thus, we planned a cyclization in
which bond formation occurs between an N-terminal Aoe
surrogate and C-terminal (R)-proline. It was in our approach
to the Aoe residue that we diverged from previous investigators,
as we needed an Aoe surrogate that would allow facile
incorporation of a radioisotope.
Figure 2. Butyric acid (6) and trichostatin (7), natural products that
inhibit histone deacetylase and promote morphological reversion of
transformed cells.
Following the method of McDougal and co-workers,21 we
were able to quantitatively monosilylate the C2 symmetric (+)-
2,3-(O)-isopropylidene-L-threitol. The exposed primary alcohol
of 8 was then oxidized under Swern conditions and homologated
with the triphenylphosphorane prepared from benzyl 3-bro-
mopropyl ether (Scheme 1). This reaction sequence provided
olefin 9 in 74% yield. Simultaneous olefin hydrogenation and
reductive debenzylation, followed by conversion of the alcohol
to the bromide, furnished 10 in 96% overall yield. The Grignard
reagent derived from 10 was transmetalated with CuBr/dimethyl
sulfide complex and alkylated with Cbz serine â-lactone 11
under conditions similar to those developed by Vederas and
co-workers.22 This reaction afforded, albeit in modest yield,
the fully functionalized Aoe surrogate 12. Elaboration of the
sensitive epoxyketone functionality was postponed until after
the cyclization, at which time a radioisotope could be conve-
niently introduced.
active at millimolar concentrations, both trichostatin and trapoxin
function in the nanomolar concentration range. The fact that
three disparate histone deacetylase inhibitors display parallel
biological activities provides evidence that inactivation of this
enzyme underlies most if not all of their cellular effects.
Although the reversible acetylation of specific lysine residues
near the N-terminus of the core histones has been studied for
over 30 years, more is known about the cellular consequences
of these post-translational modifications than about the enzymes
that catalyze the reactions.14 As the nucleosome, composed of
DNA wrapped around a histone octamer, is the fundamental
structural unit of the eukaryotic genome, changes in histone
acetylation are expected to play a role in the regulation of gene
expression. Much correlative evidence exists to support this
idea, but a molecular explanation of how histone acetylation
affects transcription and particularly how patterns of acetylation
are localized to discrete genomic domains, is lacking.
In order to explore the connections between histone acetyl-
ation, cell morphology, and cell growth, we have sought to
identify and characterize trapoxin’s molecular receptor. Our
approach to this problem has been a chemical one, and therefore
has relied upon a variety of synthetic trapoxin-based reagents.
In this paper, we report the total syntheses of trapoxin B, [3H]-
trapoxin B, and an affinity reagent that we call K-trap, in which
one of trapoxin’s phenylalanine residues is replaced by a
protected lysine residue. We have used the synthetic reagents
described herein to purify trapoxin’s biological target to
homogeneity. The peptide sequence derived from the purified
protein has allowed us to clone the human gene encoding the
elusive histone deacetylase (HD1).15
The tripeptide 13 was coupled to acid 12 under the influence
of EDC/HOBT (Scheme 2). Sequential deprotection of both
termini proceeded to give the amino acid 14 in 76% yield for
the three steps. After an exhaustive search, we found conditions
that successfully promoted closure of the 12-membered ring
tetrapeptide. Cyclization occurred in dilute DMF solution (0.2
mM) in the presence of DMAP (8 equiv) and BOP (6 equiv)
and provided the pure cyclotetrapeptide alcohol 15 in 51%
overall yield, following removal of the TIPS protecting group
with TBAF. Use of DMF was essential, as all other solvents
and solvent mixtures failed to effect this difficult cyclization.
Our attempts to emulate Schmidt and co-workers’ pentafluo-
rophenyl ester method16 were completely unsuccessful. In their
total synthesis of chlamydocin, this method was used in a similar
cyclization-the major, and apparently critical difference being
the presence of an aminoisobutyric acid (Aib) residue adjacent
to the Aoe residue in chlamydocin (2), as contrasted with the
phenylalanine found at this position in trapoxin.
Results and Discussion
Total Synthesis of Trapoxin B and [3H]Trapoxin B.
Numerous syntheses of Aoe-containing cyclotetrapeptides have
appeared,16-19 including an excellent synthesis of chlamydocin
by Baldwin and co-workers.20 In each of these syntheses, the
correct choice of a linear tetrapeptide precursor proved crucial
to the success of the cyclization. Our overall strategy was
Conversion of 15 to trapoxin (1) and [3H]trapoxin (1*) was
accomplished by the reaction sequence shown in Scheme 3.
After formation of the primary tosylate (78% yield), the
acetonide was removed with aqueous HCl. Epoxide formation
with DBU/MeOH occurred rapidly at 0 °C to give 16 (62%
yield, two steps). Based on the efficacy of the Moffatt oxidation
in the final step of Schmidt’s chlamydocin synthesis,16 we
treated epoxy alcohol 16 with the reagent combination DMSO,
DCC, and catalytic dichloroacetic acid. These exceptionally
mild conditions furnished totally synthetic trapoxin B (1) in 80%
(14) Loidl, P. Chromosoma 1994, 103, 441-9.
(15) Taunton, J.; Hassig, C. A.; Schreiber, S. L. Science 1996, 272, 408-
411.
(16) Schmidt, U.; Lieberknecht, A.; Griesser, H.; Bartkowiak, F. Angew.
Chem., Int. Ed. Engl. 1984, 23, 318-320.
(17) Jacquier, R.; Lazaro, R.; Raniriseheno, H.; Viallefont, P. Tetrahedron
Lett. 1986, 27, 4735-6.
yield, its spectroscopic characteristics (1H and 13C NMR, [R]20
,
D
(18) Kawai, M.; Gardner, J. H.; Rich, D. H. Tetrahedron Lett. 1986, 27,
1877-80.
(19) Schmidt, U.; Beutler, U.; Lieberknecht, A. Angew. Chem., Int. Ed.
Engl. 1989, 28, 333-4.
(21) McDougal, P. G.; Rico, J. G.; Oh, Y. I.; Condon, B. D. J. Org.
Chem. 1986, 51, 3388-90.
(22) Arnold, L. D.; Drover, J. C. G.; Vederas, J. C. J. Am. Chem. Soc.
1987, 109, 4649-59.
(20) Baldwin, J. E.; Adlington, R. M.; Godfrey, C. R. A.; Patel, V. K.
Tetrahedron 1993, 49, 7837-56.

10414 J. Am. Chem. Soc., Vol. 118, No. 43, 1996
Taunton et al.
Scheme 2
Scheme 3
HRMS) and biological activities indistinguishable from those
of the naturally derived substance.
treated with 20 nM [3H]trapoxin in the presence or absence of
400 nM unlabeled trapoxin for 30 min at 4 °C. A slurry of
activated charcoal was then added to adsorb unbound trapoxin,
leaving the specifically bound tracer in the supernatant.25 Figure
3 shows the results of these experiments in which the existence
of a trapoxin binding activity is clearly demonstrated. Simul-
taneous treatment of the extracts with 400 nM unlabeled trapoxin
reduced the amount of [3H]trapoxin recovered in the bound
fraction to background levels and thus confirmed that the
interaction detected by the assay is both specific and saturable.
In each of these experiments, the residual tracer left in the
supernatant most likely derives from nonspecific binding
between [3H]trapoxin and irrelevant, abundant proteins.
Given that our primary motivation for pursuing the total
synthesis of trapoxin lay in determining the molecular basis for
its detransformation activity, it was essential that our route allow
for radioisotope incorporation at a late stage. A major advantage
of our synthesis as compared to previous syntheses of related
Aoe-containing natural products rests in the intermediacy of
acetonide alcohol 15. The tritiated variant 15* was obtained
by oxidizing 15 with the Dess-Martin periodinane23 and
reducing the aldehyde with [3H]NaBH4 (Scheme 3). Compound
15* was converted to [3H]trapoxin (1*) by using a reaction
sequence similar to that used in the synthesis of trapoxin, which
served as an excellent model system for the preparation of the
tritiated compound. The identity of [3H]trapoxin was established
by the following criteria. (1) Starting with the NaBH4 reduction,
we performed a mock synthesis of [3H]trapoxin using the same
reaction conditions and concentrations as in the radioactive
synthesis. The material so obtained was spectroscopically
identical to naturally derived trapoxin B. (2) The TLC mobility
of synthetic [3H]trapoxin (>90% pure by TLC) was identical
to that of the natural material. (3) [3H]trapoxin was equipotent
to natural and synthetic trapoxin in its ability to morphologically
detransform V-sis NIH3T3 fibroblasts.5
Because trichostatin A (7) displays a spectrum of biological
activities nearly identical to that of trapoxin and is known to
inhibit cellular histone deacetylase activity, we asked whether
it could also antagonize [3H]trapoxin in our charcoal assay. At
a concentration of 1 µM, trichostatin fully competed with 20
nM [3H]trapoxin (Figure 3). The combined data imply that the
binding activity we detect is relevant to the cellular effects of
both trapoxin and trichostatin and further suggest that the
binding activity and the histone deacetylase activity are one and
the same.
Initial Characterization of a Trapoxin Binding Activity.
We used [3H]trapoxin to test for a binding activity in crude
nuclear extracts prepared from bovine thymus.24 Extracts were
(24) We focused on the nuclear proteins, as the level of trapoxin binding
activity in this fraction was twice that of the cytoplasmic fraction. Taunton,
J.; Schreiber, S. L., unpublished results.
(25) Receptor-Ligand Interactions: a Practical Approach; Hulme, E.
C., Ed.; Oxford University Press: New York, 1992.
(23) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-87.

Synthesis of Natural and Modified Trapoxins
J. Am. Chem. Soc., Vol. 118, No. 43, 1996 10415
Scheme 4
activity. Our main synthetic concern involved the choice of a
protecting group for the lysine ꢀ-amine. As this protecting group
would be the last to go, it would have to withstand acidic and
basic conditions and be removed with reagents that would leave
the epoxyketone function intact. Although the allyloxycarbonyl
(Alloc) group fulfilled these criteria, it was incompatible with
the Cbz group used to protect the Aoe surrogate 12 in our
trapoxin synthesis. For this reason, and because we desired a
more efficient synthesis of the Aoe fragment, we devised a
second route that also was substantially different from previous
syntheses of this amino acid. The route relies upon the Evans
chiral auxiliary to set the R-amino stereocenter and makes use
of the face selective electrophilic azidation of a chiral potassium
enolate, a reaction developed in the Evans’ laboratory.26
Starting with the same threitol derivative that was used in
our trapoxin B synthesis, the mono-TBS ether 17 was prepared
in quantitative yield (Scheme 4). After Swern oxidation of the
primary alcohol, the Aoe carbon backbone was completed by
coupling the unstable aldehyde with the ylide derived from
bromohexanoic acid and triphenylphosphine. Attachment of the
Evans (S)-oxazolidinone 19 to the resultant acid 18 furnished
imide 20, which was then converted into the tosylate 21 (88%
overall yield). As expected, the electrophilic azidation of 21
was highly diastereoselective, affording (S)-azido imide 22 in
73% isolated yield. Imide hydrolysis followed by simultaneous
hydrogenation of the azide and olefin gave the amino acid,
which was protected as the Fmoc carbamate 23. Although this
sequence is one step longer than our previous route, it easily
allows for the preparation of gram quantities of 23. An
additional improvement over our previous synthesis follows
from the early installation of the tosyl group, which reduces by
two the number of post-cyclization steps.
Figure 3. A trapoxin binding activity in nuclear extracts as revealed
by a charcoal adsorption assay.
Synthesis of K-Trap and the K-Trap Affinity Matrix. [3H]-
trapoxin provided us with a means of detecting a binding entity
in a complex mixture of macromolecules. Given that trapoxin’s
inhibition of histone deacetylase activity is irreversible at low
nanomolar concentrations, it seemed plausible that we could
specifically label the protein and follow it through several stages
of a conventional (and most likely laborious) purification.
However, we felt that an affinity-based purification would
provide the most expeditious path to revealing the molecular
identity of histone deacetylase. Hence we sought a versatile
trapoxin-like molecule [cf. K-trap (27)] that would lend itself
to the attachment of an affinity tag or a solid support.
The decision to replace the phenylalanine directly adjacent
to the Aoe residue with a protected lysine was based solely on
the observation that a nonconservative change at this position,
as seen in chlamydocin (2), does not lead to a decrease in
potency. We found that in the fibroblast detransformation assay,
chlamydocin was slightly more active than trapoxin B (our
unpublished results). We could therefore be reasonably certain
that the benzyl side chain was not required for detransformation
Standard peptide synthesis methods were used to assemble
tripeptide 24, which was then converted to the tetrapeptide 25
(Scheme 5). Concomitant removal of both the methyl ester and
the Fmoc protecting groups was followed by BOP/DMAP
mediated cyclization. Unexpectedly, the high dilution conditions
used in the trapoxin B synthesis (Scheme 2) produced the
cyclodimer as the major product. This problem was solved by
syringe pump addition of the tetrapeptide to a solution of BOP
(5.5 mM) and DMAP (7.4 mM) in dry DMF. After purification
by reverse phase HPLC, cyclotetrapeptide 26 was obtained in
58% overall yield. Completion of the K-trap synthesis pro-
(26) Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J. Am.
Chem. Soc. 1990, 112, 4011-30.

10416 J. Am. Chem. Soc., Vol. 118, No. 43, 1996
Taunton et al.
Scheme 5
Scheme 6
ceeded by analogy to the earlier trapoxin route (Scheme 6),
except that water soluble EDC was used instead of DCC in the
final Moffatt oxidation. Using this sequence of reactions, tens
of milligrams of pure K-trap (27) were prepared. K-Trap and
1
trapoxin exhibited virtually superimposable H and 13C NMR
spectral features corresponding to their shared structural ele-
ments, and substitution of Alloc-lysine for phenylalanine had
little effect on the ability of K-trap to inhibit histone deacetylase
activity in vitro. However, K-trap was approximately ten times
less potent than trapoxin in an in vivo cell growth assay
(inhibition of [3H]thymidine incorporation in MG-63 human
osteosarcoma cells; our unpublished results). This loss of
potency in vivo may reflect K-trap’s decreased cell permeability
and/or increased susceptibility to reductive inactivation in vivo.
K-Trap was deblocked with 5,5-dimethyl-1,3-cyclohexanedi-
one (dimedone) and catalytic palladium(0).27 The crude amine
was purified by reverse phase HPLC, lyophilized and coupled
immediately to Affi-Gel 10 (Biorad), a cross-linked agarose
matrix displaying N-hydroxysuccinimide ester functional groups.
In order to demonstrate that coupling had occurred without loss
of biological activity, we performed the following simple
experiment. Nuclear extracts were incubated with either the
K-trap matrix (28) or an ethanolamine-capped control matrix
for 12 h at 4 °C. The matrix beads were removed by
centrifugation, and the supernatants were assayed for both [3H]-
trapoxin binding and histone deactylase activity.15 Both activi-
ties were completely depleted by treatment with the K-trap
matrix and yet were not affected by the control matrix (Figure
4).
Figure 4. Specific depletion of trapoxin binding and histone deacetyl-
ase activities by the K-trap affinity matrix (28).
for purifying histone deacetylase. Nevertheless, our first
attempts to affinity purify the trapoxin receptor were frustrated
by high levels of nonspecifically associated proteins. These
initial experiments were carried out with a slightly modified
form of the K-trap affinity matrix in which a cleavable disulfide
linker had been appended to the solid support. A cleavable
linker was employed because we had assumed that trapoxin
irreversibly alkylated its target protein. However, concerns
about an irreversible covalent bond soon evaporated with the
synthesis of yet another trapoxin variant, I-trap (30) (Scheme
7).
Detours and Dead-Ends: Synthesis of I-Trap. After
completing the K-trap synthesis, we had all of the requisite tools
We designed I-trap based on our conviction that a covalently
labeled trapoxin target would be detected by SDS-PAGE/
autoradiography if we used an isotopically labeled compound
(27) Kunz, H.; Unverzagt, C. Angew. Chem., Int. Ed. Engl. 1984, 23,
436.

Synthesis of Natural and Modified Trapoxins
J. Am. Chem. Soc., Vol. 118, No. 43, 1996 10417
Scheme 7
with higher specific activity than [3H]trapoxin. Previous
attempts to visualize a [3H]trapoxin covalent adduct had failed,
and because autoradiographic detection of 125I is several orders
of magnitude more sensitive than tritium, a trapoxin variant
bearing two 125I atoms was synthesized. I-Trap (30) was
synthesized by analogy to K-trap (27), except that in this case,
we replaced the phenylalanine adjacent to proline in trapoxin
(1) with tyrosine, protecting the oxidatively sensitive phenol as
its allyl ether 29 (Scheme 7). Deprotection of 29 was followed
by rapid and quantitative iodination of the phenol with NaI/
chloramine-T.28 I-Trap behaved similarly to K-trap in its ability
to cause cell cycle arrest and to inhibit histone deacetylase
activity in vitro (our unpublished results).
The reaction conditions used to prepare I-trap were modified
with Na125I to give a single compound by TLC that comigrated
with the unlabeled compound (specific activity ∼5000 Ci/
mmol). A crude extract of nuclear proteins was treated with 1
nM 125I-trap in the presence or absence of 1 µM cold trapoxin.
Unbound I-trap was removed with charcoal as described above
with [3H]trapoxin, and proteins in the supernatant were treated
with 2% sodium dodecyl sulfate (SDS) buffer and separated
by polyacrylamide gel electrophoresis (PAGE; 12.5% gel). The
resulting phosphorimage (Figure 5) shows a labeled band (lane
1) that is efficiently competed by coincubating with unlabeled
trapoxin (lane 2). It is unlikely that this band corresponds to a
labeled protein, since the labeled species comigrates exactly with
free 125I-trap and the dye front (lane 3). This experiment
suggested, but did not prove, that trapoxin or a modified form
of trapoxin, dissociates from its target protein under denaturing
conditions, despite its ability to irreversibly inhibit histone
deacetylase activity. Proof that this is indeed the case came
from the successful elution of the purified (albeit denatured and
therefore catalytically inactive) histone deacetylase from the
K-trap affinity matrix (28), this time in the absence of a
cleavable linker.15
We speculate that the active site nucleophile that reacts with
trapoxin is also a good leaving group, for example, histidine,
aspartate, or glutamate. In this model, nucleophilic attack occurs
either R or â to the ketone, leading to a ring-opened intermediate
that is stabilized by numerous active site interactions. Upon
protein denaturation, these interactions disappear, causing the
nucleophile/leaving group to be ejected via â-elimination or
epoxide ring closure pathways. The model is reminiscent of
enzyme inactivation by the chloromethylketone serine protease
inhibitors, in which chloride is displaced by an active site
histidine.29 However, because the chloride leaving group of
these inhibitors is lost to the aqueous milieu, no kinetically
viable pathway exists for covalent bond reversal. Relevant to
this issue is a chloromethylketone variant of chlamydocin (4)
that was synthesized by Shute and co-workers and found to be
nearly as active as chlamydocin in a cell proliferation assay
(IC50 ∼ 10 nM).30 This reagent may enable the unequivocal
identification of the active site residue that is alkylated by
trapoxin and chlamydocin.
Purification and Cloning of the Trapoxin Target HD1, a
Nuclear Histone Deacetylase. Both [3H]trapoxin and the
K-trap affinity matrix played important roles in our purification
of histone deacetylase.15 First, a crude nuclear extract from
bovine thymus was fractionated by anion exchange chroma-
tography, and the fractions were screened for [3H]trapoxin
binding activity. Analysis of the same fractions for histone
deacetylase activity showed that the two activities precisely
coeluted. These fractions were far from pure, however, and
still contained hundreds of contaminating proteins. After affinity
chromatography with the K-trap matrix, we obtained the pure
(denatured) protein by SDS-PAGE. Peptide microsequencing
revealed it to be homologous to a putative yeast protein, Rpd3p.
Although Rpd3p exhibits no sequence homology to any protein
of known function, the RPD3 gene was shown to be required
for the proper transcriptional regulation of a subset of inducible
yeast genes.31
The peptide sequence information enabled us to obtain a
polymerase chain reaction (PCR) product from a human
complementary DNA (cDNA) library and, eventually, a full
length cDNA clone encoding the histone deacetylase, which
we call HD1. The predicted amino acid sequence is 60%
identical to yeast Rpd3p and contains no obvious structural
motifs. We have since shown that recombinant HD1 is a histone
deacetylase and that its activity is inhibited by trapoxin and
trichostatin. Currently, we are investigating its catalytic mech-
anism and substrate specificity as well its role in transcriptional
regulation, cell growth, and cellular transformation.
Histone deacetylase resisted molecular characterization for
over 30 years after Allfrey and co-workers first demonstrated
its existence in crude nuclear extracts.32 Through the chemical
synthesis of a natural inhibitor and variants based on its
structure, we were able to obtain the purified protein in two
chromatographic steps. Our approach illustrates the power of
synthetic organic chemistry as applied to a challenging cell
biology problem.
Experimental Section
General Experimental Methods. All reactions were performed in
flame or oven-dried glassware under a positive pressure of nitrogen or
argon. Air and moisture sensitive compounds were introduced via
syringe or cannula through a rubber septum. Tetrahydrofuran was
distilled from potassium/benzophenone ketyl. Diethyl ether was
distilled from sodium/benzophenone ketyl. Dichloromethane, toluene,
diisopropylethylamine, triethylamine, acetonitrile, and hexamethyl-
(30) Shute, R. E.; Dunlap, B.; Rich, D. H. J. Med. Chem. 1987, 30, 71-
8.
(28) Bolton, A. E.; Hunter, W. M. Biochem. J. 1973, 133, 529-539.
(29) Navia, M. A.; McKeever, B. M.; Springer, J. P.; Lin, T. Y.; Williams,
H. R.; Fluder, E. M.; Dorn, C. P.; Hoogsteen, K. Proc. Natl. Acad. Sci.
U.S.A. 1989, 86, 7-11.
(31) Vidal, M.; Gaber, R. F. Mol. Cell. Biol. 1991, 11, 6317-27.
(32) Allfrey, V.; Faulkner, R. M.; Mirsky, A. E. Proc. Natl. Acad. Sci.
U.S.A. 1964, 51, 786.
(33) Coulson, D. R. Inorg. Syn. 1972, 13, 121.

10418 J. Am. Chem. Soc., Vol. 118, No. 43, 1996
Taunton et al.
) 8.5 Hz, 1H), 4.51 (s, 2H), 3.89 (dd, J ) 11.0, 3.4 Hz, 1H), 3.80-
3.69 (m, 2H), 3.51 (t, J ) 6.8 Hz, 2H), 2.63-2.37 (m, 2H), 1.44 (s,
3H), 1.42 (s, 3H), 1.20-1.00 (m, 21H); 13C NMR (100 MHz, CDCl3)
δ 138.3, 131.8, 128.3, 127.5, 127.5, 108.8, 81.8, 73.0, 72.8, 69.6, 61.9,
28.4, 27.3, 26.9, 17.9, 11.8; HRMS (FAB, NaI) Calcd for C26H44O4Si
+ Na 471.2906, found 471.2912. Anal. Calcd for C26H44O4Si: C,
69.60; H, 9.88. Found: C, 69.50; H, 9.96.
(5S,6S)-1-Bromo-7-[(triisopropylsilyl)oxy]-5,6-(isopropylidene-
dioxy)heptane (10). A solution of olefin 9 (21.0 g, 0.047 mol) in 250
mL of EtOAc was treated with 10% Pd/C (8.4 g) and sparged with N2
for 10 min. Hydrogen gas was then bubbled through the heterogeneous
solution for 10 min and stirring was continued for 6 h under a blanket
of H2. The reaction mixture was filtered through Celite and concen-
trated to give 16.7 g (99% yield) of the saturated alcohol as a clear oil:
Figure 5. Reversibility of 125I-trap binding under protein denaturing
conditions. See text and Experimental Section for details.
disilazane were distilled from calcium hydride. Methanol was distilled
from magnesium methoxide. Dimethylformamide was distilled from
calcium hydride under reduced pressure and stored over 4-Å molecular
sieves. Analytical and preparative thin layer chromatography was
performed with E. Merck silica gel 60F glass plates, and flash
chromatography made use of E. Merck silica gel 60 (230-400 mesh).
Infrared spectra were recorded using a Nicolet 5PC FT-IR spectrometer.
1H NMR spectra were obtained on either a Bruker AM-500 (500 MHz)
or AM-400 (400 MHz). Combustion analyses were performed by
Atlantic Microlabs, Inc. (Norcross, GA).
Rf 0.16 (2:1 hexanes/ether); [R]20 -8.3° (c 0.047, CH2Cl2); IR (thin
D
film) 3434 (br), 2944, 2869, 1462, 1379, 1370, 1246, 1217, 1092, 1069,
1
882 cm-1; H NMR (400 MHz, CDCl3) δ 3.94 (dt, J ) 7.8, 3.6 Hz,
1H), 3.86 (dd, J ) 10.0, 3.9 Hz, 1H), 3.78-3.66 (m, 2H), 3.64 (t, J )
6.3 Hz, 2H), 1.77-1.42 (m, 7H), 1.39 (s, 3H), 1.36 (s, 3H), 1.16-
0.98 (m, 21H); 13C NMR (100 MHz, CDCl3) δ 108.3, 80.9, 79.2, 64.2,
62.7, 33.1, 32.65, 27.4, 26.9, 22.3, 17.9, 11.8; HRMS (FAB, NaI) Calcd
for C19H40O4Si + Na 383.2594, found 383.2568. Anal. Calcd for
C19H40O4Si: C, 63.28; H, 11.18. Found: C, 63.09; H, 11.22.
To a solution of the alcohol (4.0 g, 11.1 mmol, 1 equiv) and
triphenylphosphine (5.8 g, 22.2 mmol, 2 equiv) in 140 mL of THF at
room temperature was added CBr4 (7.4 g, 22.2 mmol, 2 equiv). After
stirring for 30 min, the reaction was diluted with hexanes and filtered
through silica gel (33% ether/hexanes wash). Purification by flash
chromatography (0-20% ether/hexanes) provided 4.6 g (97% yield)
of bromide 10 as a clear oil: Rf 0.81 (2:1 hexanes/ether); [R]20D -7.2°
(c 0.042, CH2Cl2); IR (thin film) 2944, 2867, 1462, 1379, 1370, 1242,
1094, 1069, 1013, 997, 884 cm-1; 1H NMR (400 MHz, CDCl3) δ 3.94
(dt, J ) 7.5, 3.2 Hz, 1H), 3.88 (dd, J ) 9.9, 3.8 Hz, 1H), 3.78-3.65
(m, 2H), 3.40 (t, J ) 6.8 Hz, 2H), 1.95-1.85 (m, 2H), 1.77-1.50 (m,
4H), 1.39 (s, 3H), 1.36 (s, 3H), 1.17-0.98 (m, 21H); 13C NMR (100
MHz, CDCl3) δ 108.4, 80.8, 79.1, 64.2, 33.5, 32.8, 32.6, 27.4, 26.9,
24.8, 17.9, 11.8; HRMS (FAB, NaI) Calcd for C19H39BrO3Si + H
423.1930, found 423.1926. Anal. Calcd for C19H40BrO3Si: C, 53.89;
H, 9.28. Found: C, 53.96; H, 9.26.
(2S,3S)-4-[(Triisopropylsilyl)oxy]-2,3-(isopropylidenedioxy)bu-
tanol (8). To a solution of NaH (60% in oil, 2.56 g, 0.064 mol, 1
equiv) in 125 mL of THF at 0 °C was added a solution of (+)-2,3-
(O)-isopropylidene-L-threitol (Lancaster, 10.4 g, 0.064 mol, 1 equiv)
in 125 mL of THF. After warming to room temperature and stirring
for 45 min, triisopropylsilyl chloride (13.7 mL, 0.064 mol, 1 equiv)
was added over 15 min. After stirring for 3 h, the reacion was poured
into saturated NaHCO3 and extracted with diethyl ether. The organic
fraction was dried over MgSO4, filtered, and concentrated in vacuo.
Purification by flash chromatography (0-33% ether/hexanes) gave 20.4
g (100% yield) of silyl ether 8 as a clear oil: Rf 0.40 (4:1 hexanes/
EtOAc); [R]20D +9.5° (c 0.017, CH2Cl2); IR (thin film) 3500 (br), 2944,
1
2869, 1464, 1381, 1369, 1248, 1217, 1084, 1071, 997, 884 cm-1; H
NMR (400 MHz, CDCl3) δ 4.02 (dt, J ) 7.8, 4.6 Hz, 1H), 3.96 (dd,
J ) 9.8, 3.9 Hz, 1H), 3.93-3.86 (m, 1H), 3.83-3.67 (m, 3H), 2.47-
2.39 (m, 1H), 1.40 (s, 3H), 1.39 (s, 3H), 1.17-1.02 (m, 21H); 13C
NMR (100 MHz, CDCl3) δ 109.0, 80.4, 78.1, 64.2, 62.80, 27.0, 26.9,
17.8, 11.8; HRMS (FAB, NaI) Calcd for C16H34O4Si + Na 341.2124,
found 341.2119. Anal. Calcd for C16H34O4Si: C, 60.33; H, 10.76.
Found: C, 60.20; H, 10.72.
(2S,3S)-1-[(Triisopropylsilyl)oxy]-7-(benzyloxy)-2,3-(isopropyl-
idenedioxy)-4(Z)-heptene (9). A solution of DMSO (11.6 mL, 0.163
mol, 2.6 equiv) in 37 mL of CH2Cl2 was added via cannula to a cold
(-78 °C) solution of oxalyl chloride (7.1 mL, 81.6 mmol, 1.3 equiv)
in 203 mL of CH2Cl2. After stirring for 5 min, a solution of alcohol
8 (20 g, 62.8 mmol, 1 equiv) in 80 mL of CH2Cl2 was added dropwise.
After stirring for 15 min, Et3N (43.7 mL, 0.315 mol, 5.0 equiv) was
added, and the reaction was warmed to room temperature and stirred
for 1 h. The reaction mixture was diluted with 100 mL of water, and
the organic fraction was separated and concentrated in vacuo. The
residue was diluted with 200 mL of 50% hexanes/ether and washed
with water (3 × 100 mL) and brine (50 mL). The combined aqueous
fractions were extracted with ether (3 × 100 mL). The combined
organic fractions were dried over anhydrous Na2SO4, filtered, and
concentrated in vacuo. The crude aldehyde was placed under high
vacuum for 30 min and carried on directly to the next step.
(2S,8S,9S)-N-(Benzyloxycarbonyl)-2-amino-10-[(triisopropylsilyl)-
oxy]-8,9-(isopropylidenedioxy)decanoic Acid (12). To magnesium
turnings (14 mg, 3.26 mmol, 72 equiv based on 11) was added 1,2-
dibromoethane (1 µL) followed by a solution of bromide 10 (115 mg,
0.272 mmol, 6.0 equiv based on 11) in 0.210 mL of ether over 5 min.
After stirring for 2 h 40 min, the Grignard reagent was added to a cold
(-23 °C) solution of â-lactone 1122 (10 mg, 0.045 mmol, 1 equiv) and
CuBr/dimethyl sulfide complex (1.8 mg, 0.0085 mmol, 0.2 equiv) in
0.4 mL of THF and 20 µL of dimethyl sulfide. The reaction was stirred
for 2 h 20 min at -23 °C and diluted with degassed ether (1 mL)
followed by 1 N acetic acid (1 mL). After warming to 0 °C and stirring
for 15 min, the solution was treated with glacial acetic acid (100 µL)
and extracted with ether. The combined organic fractions were dried
over MgSO4, filtered, and concentrated in vacuo. Purification by flash
chromatography (5-10% MeOH/CHCl3) gave 10 mg (40% yield) of
acid 12 as a clear oil: Rf 0.24 (6% MeOH/CH2Cl2); [R]20 -15.3° (c
D
0.039, CH2Cl2); IR (thin film) 3321 (br), 2942, 2867, 1721, 1588, 1532,
1
1456, 1379, 1368, 1246, 1217, 1094, 1067, 884 cm-1; H NMR (400
MHz, DMSO-d6) δ 7.41-7.23 (m, 5H), 7.10 (br s, 1H), 5.01 and 4.98
(AB quartet, J ) 12.6 Hz, 2H), 3.91-3.78 (m, 2H), 3.78-3.67 (m,
2H), 3.58 (dt, J ) 7.8, 4.4 Hz, 1H), 1.72-1.14 (m, 10H), 1.28 (s, 3H),
1.26 (s, 3H), 1.13-0.95 (m, 21H); 13C NMR (100 MHz, DMSO-d6)
(partial) δ 155.8, 137.2, 128.3, 127.6, 107.6, 80.84, 77.5, 65.2, 63.5,
54.7, 32.8, 31.8, 28.9, 27.3, 26.8, 25.7, 25.3, 17.8, 11.3; HRMS (FAB,
NaI) Calcd for C30H51NO7Si + Na 588.3332, found 588.3340.
L-Phenylalanyl-L-phenylalanyl-D-proline Methyl Ester (13). An
ice cold solution of L-phenylalanyl-D-proline methyl ester (721 mg,
1.66 mmol, 1 equiv) in 11.1 mL of CH2Cl2 was treated with
N-methylmorpholine (NMM; 201 µL, 1.83 mmol, 1.1 equiv) followed
by a solution of N-(tert-butoxycarbonyl)-L-phenylalanine (440 mg, 1.66
mmol, 1 equiv) in 13.8 mL of CH2Cl2. Solid 1-hydroxybenzotriazole
(HOBT) (247 mg, 1.83 mmol, 1.1 equiv) was added, and stirring was
continued for 15 min, whereupon 1-(3-dimethylaminopropyl)-3-ethyl-
A heterogeneous suspension of the phosphonium bromide derived
from triphenylphosphine and benzyl 3-bromopropyl ether (46.3 g, 94.2
mmol, 1.5 equiv based on 8) in 345 mL of THF was cooled to -78 °C
and treated with n-BuLi (1.6 M in hexane, 58.9 mL, 94.2 mmol, 1.5
equiv). After stirring for 1 h at -23 °C, the reaction was cooled to
-78 °C and treated with a solution of the crude aldehyde in 266 mL
of THF. After 30 min at -78 °C and 45 min at 0 °C, the reaction
mixture was diluted with hexanes and filtered through silica gel (66%
hexanes/ether wash). The filtrate was concentrated in vacuo and
purified by flash chromatography (100% hexanes to 100% ether
gradient) to give 21 g of 9 (74% yield) as a clear oil: Rf 0.61 (5:1
hexanes/ether); [R]20 +0.7° (c 0.066, CH2Cl2); IR (thin film) 2944,
D
2867, 1464, 1454, 1379, 1368, 1244, 1148, 1103, 1082, 1016, 883,
1
862 cm-1; H NMR (400 MHz, CDCl3) δ 7.40-7.24 (m, 5H), 5.75
(dt, J ) 10.5, 7.4 Hz, 1H), 5.52 (tt, J ) 10.5, 1.4 Hz, 1H), 4.83 (t, J

Synthesis of Natural and Modified Trapoxins
J. Am. Chem. Soc., Vol. 118, No. 43, 1996 10419
carbodiimide (EDC) (351 mg, 1.83 mmol, 1.1 equiv) was added in
one portion. The reaction was stirred at 0 °C for 2 h and at room
temperature for 19 h. The reaction mixture was extracted with water,
1 N HCl, saturated NaHCO3, and again with water. The organic
fraction was dried over anhydrous MgSO4, filtered, and concentrated
in vacuo. Purification by flash chromatography (50-67% EtOAc/
hexanes) provided 700 mg (81% yield) of the Boc tripeptide, which
was quantitatively deblocked by stirring for 1 h in 50% TFA/CH2Cl2.
Azeotropic removal of TFA with heptane provided the TFA salt of
tripeptide 13 as a foam. The tripeptide was characterized as its Boc
carbamate: Rf 0.26 (50% EtOAc/hexanes); [R]20D +33.9° (c 0.019, CH2-
Cl2); IR (thin film) 3287 (br), 2978, 1750, 1713, 1691, 1640, 1523,
tography (2:1 EtOAc/hexanes) to give 39 mg of impure cyclotetrapep-
tide silyl ether, which was used immediately in the next reaction.
To a solution of the silyl ether (39 mg, 0.048 mmol, 1 equiv) in 4.8
mL of THF at room temperature was added TBAF (1.0 M in THF, 97
µL, 0.097 mmol, 2.0 equiv). The reaction was stirred for 1 h and
concentrated (<25 °C) in vacuo to ca. 1 mL. Purification by flash
chromatography (2:1 EtOAc/hexanes followed by 9:1 EtOAc/MeOH)
gave impure product. Further purification by preparative TLC (2 ×
0.5 mm plates; 5% MeOH/EtOAc) gave 20 mg (51% yield over two
steps) of alcohol 15 as a white foam: Rf 0.56 (10% MeOH/EtOAc);
[R]20D -99.5° (c 0.004, CH2Cl2); IR (thin film) 3289 (br), 2984, 2934,
2863, 1684, 1667, 1525, 1499, 1455, 1379, 1369, 1242, 1107, 1051
cm-1; 1H NMR (400 MHz, CDCl3) δ 7.54 (d, J ) 10.1 Hz, 1H), 7.33-
7.00 (m, 10H), 6.71 (br d, J ) 5.9 Hz, 1H), 6.57 (br d, J ) 5.3 Hz,
1H), 5.15 (dt, J ) 9.8, 6.3 Hz, 1H), 4.67 (d, J ) 6.4 Hz, 1H), 4.15 (q,
J ) 8.2 Hz, 1H), 3.91-3.55 (m, 8H), 3.35-3.15 (m, 3H), 3.01 (dd, J
) 13.0, 5.6 Hz, 1H), 2.28 (br d, J ) 9.9 Hz, 1H), 2.23-2.05 (m, 2H),
1.90-1.15 (unresolved, 11H), 1.41 (s, 3H), 1.40 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 175.2, 173.8, 172,81, 171.6, 137.0, 136.9, 129.1,
128.9, 128.7, 128.6, 126.9, 126.8, 108.6, 81.4, 76.9, 63.0, 62.1, 57.9,
54.1, 53.6, 46.9, 35.8, 35.1, 32.8, 29.2, 29.0, 27.4, 27.1, 25.8, 25.3,
24.9, 24.8; HRMS (FAB, NaI) Calcd for C36H48N4O7 + Na 671.3421,
found 671.3425.
1
1497, 1451, 1391, 1173 cm-1; H NMR (400 MHz, CDCl3) (major
rotamer) δ 7.30-7.10 (m, 10H), 6.80 (d, J ) 8.0 Hz, 1H), 4.97 (d, J
) 7.4 Hz, 1H), 4.88 (m, 1H), 4.38 (m, 1H), 4.27 (dd, J ) 7.8, 4.0 Hz,
1H), 3.65 (s, 3H), 3.44 (m, 1H), 3.15-2.92 (m, 3H), 2.87 (dd, J )
12.8, 9.6 Hz, 1H), 2.60 (m, 1H), 1.95-1.72 (m, 3H), 1.55-1.42 (m,
1H); 13C NMR (100 MHz, CDCl3) δ 172.0, 170.4, 169.4, 155.1, 136.5,
136.1, 129.4, 129.4, 128.5, 128.5, 127.0, 126.8, 58.7, 52.3, 52.1, 46.7,
39.6, 28.9, 28.2, 24.3; HRMS (FAB, NaI) Calcd for C29H37N3O6 + Na
546.2580, found 546.2603.
(2S,8S,9S)-2-Amino-10-[(triisopropylsilyl)oxy]-8,9-(isopropylidene-
dioxy)-decanoyl-L-phenylalanyl-L-phenylalanyl-D-proline (14). An
ice cold solution of tripeptide 13 (125 mg, 0.233 mmol, 1 equiv) in
1.6 mL of CH2Cl2 was treated with NMM (28 µL, 0.257 mmol, 1.1
equiv) followed by a solution of acid 12 (132 mg, 0.233 mmol, 1 equiv)
in 2 mL of CH2Cl2. Solid HOBT (35 mg, 0.257 mmol, 1.1 equiv) was
added, and stirring was continued for 15 min, whereupon EDC (49
mg, 0.257 mmol, 1.1 equiv) was added in one portion. After stirring
for 1 h at 0 °C and 20 h at room temperature, the reaction was applied
directly to a flash chromatography column. Elution with 60-67%
EtOAc/hexanes gave 183 mg (81% yield) of the N-(benzyloxycarbonyl)
tetrapeptide methyl ester as an amorphous solid: Rf 0.61 (2:1 EtOAc/
cyclo[(2S,8S,9S)-2-Amino-8-hydroxy-9,10-epoxy-decanoyl-L-phe-
nylalanyl-L-phenylalanyl-D-prolyl] (16). A solution of alcohol 15 (20
mg, 0.031 mmol, 1 equiv) in 0.62 mL of pyridine was treated with
DMAP (ca. 1 mg) followed by tosyl chloride (11.8 mg, 0.062 mmol,
2.0 equiv). After stirring for 4 h, the solvent was removed. The crude
product was dissolved in CH2Cl2 and applied to two preparative TLC
plates (0.5 mm). Elution with 1% MeOH/EtOAc gave 19 mg (78%
yield) of the tosylate as an amorphous solid: Rf 0.50 (2% MeOH/
EtOAc); [R]20 -75.3° (c 0.003, CH2Cl2); IR (thin film) 3287 (br),
D
2934, 1663, 1524, 1497, 1455, 1368, 1242, 1213, 1190, 1177, 1098,
hexanes); [R]20 +12.2° (c 0.024, CH2Cl2); IR (thin film) 3285 (br),
D
1
948 cm-1; H NMR (400 MHz, CDCl3) δ 7.80 (d, J ) 8.2 Hz, 2H),
2942, 2867, 1748, 1723, 1638, 1536, 1455, 1368, 1217, 1175, 1094,
7.54 (d, J ) 10.3 Hz, 1H), 7.35 (d, J ) 8.2 Hz, 2H), 7.32-7.07 (m,
10H), 6.50 (d, J ) 5.9 Hz, 1H), 5.15 (dt, J ) 9.8, 6.6 Hz, 1H), 4.68 (d,
J ) 6.7 Hz, 1H), 4.16 (q, J ) 8.8 Hz, 1H), 4.16-4.03 (m, 2H), 3.90-
3.61 (m, 6H), 3.37-3.17 (m, 3H), 3.02 (dd, J ) 13.9, 6.3 Hz, 1H),
2.45 (s, 3H), 2.28 (br d, J ) 8.6 Hz, 1H), 2.23-2.07 (m, 2H), 1.88-
1.15 (unresolved, 10H), 1.36 (s, 3H), 1.30 (s, 3H); 13C NMR (100 MHz,
CDCl3) δ 175.2, 173.8, 172.8, 171.6, 145.0, 137.0, 136.9, 129.9, 129.5,
129.1, 128.9, 128.7, 128.6, 128.1, 126.9, 126.8, 109.4, 78.2, 77.8, 69.2,
63.1, 57.9, 54.1, 53.6, 47.0, 35.8, 35.2, 32.9, 29.2, 29.0, 27.3, 26.7,
25.7, 25.3, 24.9, 24.8, 21.7; HRMS (FAB, NaI) Calcd for C43H54N4O9S
+ Na 825.3509, found 825.3497.
An ice cold solution of the acetonide tosylate (18 mg, 0.023 mmol)
in 1 mL of THF was treated with 1 mL of 5% aqueous HCl. Upon
warming to room temperature and stirring for 21 h, the THF was
removed in vacuo (<25 °C). The aqueous solution was diluted with
brine and extracted with CH2Cl2. The organic fraction was dried over
anhydrous MgSO4, filtered, and concentrated. The residue was purified
by preparative TLC (0.5 mm plate; 2% MeOH/EtOAc) to give 13 mg
(73% yield) of the diol tosylate as a clear oil.
1
882 cm-1; H NMR (400 MHz, CDCl3) (major rotamer) δ 7.40-6.98
(m, 15H), 6.85 (br s, 1H), 6.65 (d, J ) 7.4 Hz, 1H), 5.24 (d, J ) 8.2
Hz, 1H), 5.17-5.02 (m, 2H), 4.97-4.83 (m, 1H), 4.75-4.65 (m, 1H),
4.31 (dd, J ) 7.5, 4.1 Hz), 4.17-4.07 (m, 1H), 3.90 (dt, J ) 7.8, 3.7
Hz, 1H), 3.84 (dd, J ) 10.2, 4.3 Hz, 1H), 3.77-3.61 (m, 3H), 3.69 (s,
3H), 3.53-3.42 (m, 1H), 3.16-2.86 (m, 5H), 2.71-2.60 (m, 1H), 1.97-
1.69 (m, 4H), 1.68-1.17 (unresolved, 8H), 1.38 (s, 3H), 1.35 (s, 3H),
1.15-0.97 (m, 21H); 13C NMR (100 MHz, CDCl3) (major rotamer) δ
172.0, 171.4, 169.6, 169.3, 156.1, 136.2, 136.1, 136.0, 129.4, 128.5,
128.4, 128.1, 128.0, 127.0, 126.9, 108.3, 81.0, 79.0, 67.0, 64.2, 58.7,
55.01, 52.4, 52.2, 46.8, 39.5, 38.2, 33.4, 32.4, 29.4, 28.9, 27.4, 26.9,
26.0, 25.4, 24.3, 17.9, 11.8; HRMS (FAB, NaI) Calcd for C54H78N4O10-
Si + Na 993.5385, found 993.5388.
An ice cold solution of the Cbz tetrapeptide methyl ester (182 mg,
0.188 mmol, 1 equiv) in 3.2 mL of THF, 1.1 mL of MeOH, and 1.1
mL of water was treated with 1.0 N LiOH (563 µL, 0.563 mmol, 3.0
equiv). The reaction was stirred at 0 °C for 5 min and at room
temperature for 2 h. The reaction mixture was concentrated in vacuo
(room temperature), and the residue was dissolved in water and acidified
with 1.0 N HCl to pH 3. The product was extracted into EtOAc and
the combined organic layers were dried over MgSO4, filtered, and
concentrated to give an amorphous solid.
To this crude benzyl carbamate in 9.1 mL of THF was added 10%
Pd/C (200 mg). The mixture was stirred under an atmosphere of H2
for a total of 13.5 h. Filtration through Celite (MeOH wash) and
concentration under reduced pressure gave 144 mg (94% yield) of
deprotected tetrapeptide 14 as a foam.
cyclo[(2S,8S,9S)-2-Amino-10-hydroxy-8,9-(isopropylidenedioxy)-
decanoyl-L-phenylalanyl-L-phenylalanyl-D-prolyl] (15). A solution
of tetrapeptide 14 (50 mg, 0.061 mmol, 1 equiv) in 300 mL of
anhydrous DMF was treated with 4-(dimethylamino)pyridine (DMAP)
(60 mg, 0.486 mmol, 8.0 equiv) followed by benzotriazol-1-yloxytris-
(dimethylamino)phosphonium hexafluorophosphate (BOP) (161 mg,
0.364 mmol, 6.0 equiv). After stirring in the dark for 4 days, the solvent
was removed by vacuum distillation (<40 °C), and the residue was
dissolved in EtOAc and filtered through silica gel. The filtrate was
concentrated, and the residue was rapidly purified by flash chroma-
To the diol (6.2 mg, 8.1 µmol, 1 equiv) in 1.6 mL of MeOH at 0 °C
was added 1,8-diazabicylco[5.4.0]undec-7-ene (DBU) (1.0 M in THF,
41 µL, 41 µmol, 5.1 equiv). After stirring for 2 h, the reaction was
filtered through silica gel (10% MeOH/EtOAc wash), and the filtrate
was concentrated in vacuo. The crude product was purified by
preparative TLC (0.25 mm plate; 10% MeOH/EtOAc) to give 4.1 mg
(85% yield) of epoxy alcohol 16: Rf 0.50 (10% MeOH/EtOAc); [R]20
D
-78.9° (c 0.002, CH2Cl2); IR (thin film) 3283 (br), 3029, 2930, 2859,
1686, 1624, 1527, 1497, 1455, 1250, 1113 cm-1; 1H NMR (400 MHz,
CDCl3) δ 7.54 (d, J ) 10.3 Hz, 1H), 7.33-7.05 (m, 10H), 6.47 (d, J
) 5.4 Hz, 1H), 5.16 (dt, J ) 9.6, 6.5 Hz, 1H), 4.68 (d, J ) 5.9 Hz,
1H), 4.16 (q, J ) 8.2 Hz, 1H), 3.86 (dt, J ) 9.5, 4.6 Hz, 1H), 3.79-
3.62 (m, 3H), 3.58 (t, J ) 4.6 Hz, 1H), 3.43 (q, J ) 5.8 Hz, 1H),
3.35-3.17 (m, 3H), 3.06-2.92 (m, 2H), 2.83 (t, J ) 4.6 Hz, 1H), 2.72
(dd, J ) 4.6, 2.7 Hz, 1H), 2.35-2.10 (m, 2H), 2.00-1.15 (unresolved,
11H); 13C NMR (100 MHz, CDCl3) δ 175.2, 173.7, 172.7, 171.54,
137.0, 136.8, 129.0, 128.8, 128.6, 128.5, 126.9, 126.8, 71.4, 63.1, 57.8,
55.3, 54.0, 53.49, 46.9, 45.1, 35.7, 35.0, 34.1, 29.0, 28.9, 25.2, 25.0,

10420 J. Am. Chem. Soc., Vol. 118, No. 43, 1996
Taunton et al.
24.9, 24.8; HRMS (FAB, NaI) Calcd for C33H42N4O6 + Na 613.3002,
found 613.2994.
v) slurry (50 µL) of activated charcoal in 2.5% bovine serum albumin/
phosphate buffered saline was added, and the mixture was vortexed
briefly. After centrifugation at 14 000 rpm for 5 min, 0.35 mL of the
supernatant was combined with 3 mL of Aquasol scintillation fluid
(NEN) and counted with a Beckman LS 6500 scintillation counter. To
determine the inherent background of the assay, a control experiment
was performed in which 20 nM [3H]trapoxin was incubated with 0.5
mL buffer (25 mM tris pH 8, 10 mM NaCl, 10% glycerol). The data
in Figure 3 are the average values (with standard deviations) from three
independent assays.
Trapoxin B (1). A solution of epoxy alcohol 16 (2.5 mg, 4.25 µmol,
1 equiv) in 0.4 mL of benzene and 40 µL of DMSO was treated with
DCC (9 mg, 42.5 µmol, 10 equiv) followed by dichloroacetic acid (1.0
M benzene, 2.1 µL, 2.1 µmol, 0.5 equiv). After stirring for 2 h, more
dichloroacetic acid (1.0 M benzene, 2.1 µL, 2.1 µmol, 0.5 equiv) was
added, and stirring was continued for an additional 2 h. The reaction
was filtered through silica gel (10% MeOH/EtOAc wash), and the
filtrate was concentrated in vacuo. The residue was purified by
preparative TLC (0.5 mm plate). Elution with 5% MeOH/EtOAc gave
2 mg (80% yield) of trapoxin B as a white solid: Rf 0.53 (10% MeOH/
EtOAc); [R]20 -76.0° (c 0.50, MeOH) [lit.5 [R]20 -75.3° (c 0.517,
(2S,3S)-4-[(tert-Butyldimethylsilyl)oxy]-2,3-(isopropylidenedioxy)-
butanol (17). To an ice cold suspension of NaH (60% in oil, 1.22 g,
30.5 mmol, 1 equiv) in THF (60 mL) was added via cannula a solution
of (+)-2,3-(O)-isopropylidene-L-threitol (4.95 g, 30.5 mmol, 1 equiv)
in 50 mL of THF over 15 min. The suspension was allowed to stir for
45 min at room temperature, after which time tert-butyldimethylsilyl
chloride (4.60 g, 30.5 mmol, 1 equiv) in 10 mL of THF was added.
After stirring for 16 h, the reaction mixture was poured into 150 mL
of saturated NaHCO3 and extracted with ether (3 × 100 mL). The
extracts were dried over MgSO4, concentrated, and purified by flash
chromatography (3:1 hexanes/EtOAc) to give 8.43 g of compound 17
D
D
MeOH)]; IR (thin film) 3285 (br), 3061, 3028, 2928, 2856, 1684, 1525,
1
1499, 1455, 1343, 1306, 1262, 1242, 1113, 870, 742, 700 cm-1; H
NMR (500 MHz, CDCl3) δ 7.51 (d, J ) 10.3 Hz, 1H), 7.32-7.18 (m,
8H), 7.13 (d, J ) 6.8 Hz, 2H), 7.08 (d, J ) 10.3 Hz, 1H), 6.30 (d, J
) 5.6 Hz, 1H), 5.16 (dt, J ) 9.9, 6.3 Hz, 1H), 4.67 (d, J ) 6.2 Hz,
1H), 4.14 (q, J ) 7.9 Hz, 1H), 3.86 (dt, J ) 9.2, 4.7 Hz, 1H), 3.78-
3.64 (m, 2H), 3.42 (dd, J ) 4.7, 2.4 Hz, 1H), 3.34-3.18 (m, 3H),
3.06-2.97 (m, 2H), 2.86 (dd, J ) 5.9, 2.4 Hz, 1H), 2.41 (dt, J ) 17.5,
6.9 Hz, 1H), 2.35-2.13 (m, 3H), 1.87-1.68 (m, 3H), 1.66-1.15
(unresolved, 9H); 13C NMR (125 MHz, CDCl3) δ 207.5, 175.1, 173.7,
172.7, 171.5, 137.0, 136.9, 129.1, 128.9, 128.6, 128.5, 126.9, 126.8,
63.2, 57.9, 53.9, 53.5, 53.4, 46.9, 46.0, 36.2, 35.8, 35.1, 28.8, 28.6,
25.2, 24.9, 24.8, 22.8; HRMS (FAB) Calcd for C33H40N4O6 + H
589.3026, found 589.3020.
[3H]Trapoxin B (1*). A solution of alcohol 15 (2.3 mg, 3.6 µmol,
1 equiv) in 0.4 mL of CH2Cl2 was treated with the Dess-Martin
reagent23 (6 mg, 14.2 µmol, 3.9 equiv). After stirring for 1 h, the
reaction was filtered through silica gel (10% MeOH/EtOAc wash) and
concentrated. The residue was dissolved in CH2Cl2, and the filtration
was repeated to give the crude aldehyde.
An ice cold solution of the aldehyde in 0.5 mL of THF was treated
with [3H]NaBH4 (NEN; 500 µL of a 25 mCi/mL aqueous solution,
0.92 µmol, 13.6 Ci/mmol) and stirred for 1 h at 0 °C. The reaction
mixture was extracted with EtOAc (4×), and the combined organic
fractions were filtered through silica gel (10% MeOH/EtOAc wash).
Purification on a small pipet column (silica gel; 0-10% MeOH/EtOAc)
provided the [3H]alcohol 15*, which was used immediately in the next
reaction.
A solution of the 15* in 200 µL of pyridine containing DMAP (ca.
1 mg) was treated with p-toluenesulfonic anhydride (1.0 M in CH2Cl2,
71 µL, 71 µmol). After stirring for 2 h, the reaction was filtered through
silica gel (10% MeOH/EtOAc wash). The crude tosylate was purified
exactly as in the previous step and used immediately in the next reaction.
The [3H]acetonide tosylate was dissolved in 500 µL of THF, cooled
to 0 °C, and treated with 500 µL of 5% aqueous HCl. After stirring at
room temperature for 13 h, the reaction was cooled to 0 °C and
neutralized with solid NaHCO3. The product was extracted with EtOAc
(5×), and the combined organic layers were filtered through silica gel
(10% MeOH/EtOAc wash). After concentration in vacuo, the product
was used immediately in the next reaction.
(100% yield) as an oil: Rf 0.24 (1:4 EtOAc/hexanes); [R]20 +16.1°
D
(c 2.4, CHCl3); IR (thin film) 3480 (br), 2988, 2932, 2859, 1474, 1464,
1
1379, 1372, 1254, 1082, 837 cm-1; H NMR (400 MHz, CDCl3) δ
3.93 (m, 1H), 3.83-3.77 (m, 2H), 3.72 (dd, J ) 11.6, 4.5 Hz, 1H),
3.65-3.59 (m, 2H), 2.66 (br s, 1H), 1.35 (s, 3H), 1.33 (s, 3H), 0.83 (s,
9H), 0.02 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 109.0, 80.0, 78.0,
63.6, 62.6, 27.0, 26.8, 25.8, 18.2, -5.56, -5.59; HRMS (CI, NH3)
Calcd for C13H28O4Si + NH4 294.2101, found 294.2099.
(2S,3S)-10-[(tert-Butyldimethylsilyl)oxy]-8,9-(isopropylidenedioxy)-
6(Z)-decenoic Acid (18). To a cooled (-78 °C) solution of oxalyl
chloride (1.8 mL, 20.3 mmol, 1.3, equiv) in 50 mL of CH2Cl2 was
added dropwise DMSO (2.9 mL, 41 mmol, 2.6 equiv) in 9 mL of CH2-
Cl2. Silyl threitol 17 (4.3 g, 15.6 mmol, 1 equiv) in 20 mL of CH2Cl2
was then added dropwise. After stirring for 20 min, triethylamine (10.8
mL, 78 mmol, 5 equiv) was added, and the reaction was allowed to
warm to room temperature over a period of 1 h. The heterogeneous
mixture was extracted with water, concentrated, redissolved in 50%
hexanes/ether (200 mL), and extracted with water again (3 × 100 mL).
The combined aqueous fractions were then extracted with 50% hexanes/
ether, and the combined organic fractions were dried over Na2SO4 and
concentrated. The crude aldehyde was dried azeotropically from
benzene and used immediately.
In a separate flask, lithium hexamethyldisilazide was prepared by
adding n-BuLi (1.6 M in hexanes, 25.4 mL, 40.6 mmol, 2.6 equiv based
on 17) to hexamethyldisilazane (8.6 mL, 40.6 mL, 2.6 equiv) in 50
mL of THF at 0 °C for 15 min. This solution was then transferred to
a cooled (-20 °C) suspension of the phosphonium bromide derived
from 6-bromohexanoic acid and triphenylphosphine (9.28 g, 20.3 mmol,
1.3 equiv). After 30 min, the mixture was cooled to -78 °C, and the
aldehyde (see above) was added via cannula in 15 mL of THF. After
stirring for 30 min at -78 °C and 1 h at room temperature, the reaction
was quenched with water, and the organic solvents were removed in
vacuo. The mixture was acidified with saturated NaHSO4 and extracted
with EtOAc. The organic fraction was washed with brine, dried over
MgSO4, and concentrated. Purification by flash chromatography (2:1
hexanes/EtOAc, 0-0.5% AcOH) provided the acid 18 (3.6 g, 62%
yield) as an oil: Rf 0.22 (1:4 EtOAc/hexanes, 1% AcOH); [R]20D -0.76°
(c 3.4, CHCl3); IR (thin film) 3200 (br), 2932, 2859, 1740, 1711, 1252,
To an ice cold solution of the crude [3H]diol tosylate in 400 µL of
MeOH was added DBU (1.0 M in THF, 18 µL, 18 µmol). After stirring
at 0 °C for 2 h, the reaction mixture was filtered through silica gel
(10% MeOH/EtOAc wash), concentrated, and taken directly to the next
step.
A solution of the [3H]epoxy alcohol in 200 µL of benzene and 40
µL of DMSO was treated with DCC (8 mg, 35.5 µmol) followed by
dichloroacetic acid (1.0 M in benzene, 3.6 µL, 3.6 µmol). After stirring
for 2 h, the reaction mixture was filtered through silica gel (10% MeOH/
EtOAc wash). Purification on a small pipet column (silica gel; 5-10%
MeOH/EtOAc) gave [3H]trapoxin B (1*) (0.37 mg; specific activity
2.1 Ci/mmol). The compound was stored in DMSO stock solutions of
10 µM and 100 µM at -20 °C.
1
1146, 1080, 837, 777 cm-1; H NMR (400 MHz, CDCl3) δ 5.63 (m,
1H), 5.40 (m, 1H), 4.75 (m, 1H), 3.78 (dd, J ) 2.9, 8.1 Hz, 1H), 3.68-
3.60 (m, 2H), 2.33 (t, J ) 7.4 Hz, 2H), 2.3-2.0 (m, 2H), 1.5-1.4 (m,
2H), 1.41 (s, 3H), 1.40 (s, 3H), 0.88 (s, 9H), 0.06 (s, 3H), 0.04 (s, 3H);
13C NMR (100 MHz, CDCl3) δ 179.6, 135.2, 126.9, 108.8, 81.6, 72.7,
61.4, 33.8, 28.9, 27.3, 27.2, 26.8, 25.8, 24.2, 18.3, -5.4, -5.5; HRMS
(FAB, NaI) Calcd for C19H36O5Si + Na 395.2230, found 395.2236.
(4S)-4-Benzyl-3-[(8S,9S)-10-[(tert-butyldimethylsilyl)oxy]-8,9-(iso-
propylidenedioxy)-6(Z)-decenoyl]-2-oxazolidinone (20). To a cooled
(-78 °C) solution of acid 18 (139 mg, 0.374 mmol, 1 equiv) and
triethylamine (67 µL, 0.49 mmol, 1.3 equiv) was added pivaloyl chloride
(51 µL, 0.41 mmol, 1.1 equiv). After stirring for 15 min at -78 °C
and 45 min at 0 °C, the reaction was again cooled to -78 °C. To this
Charcoal Binding Assay with [3H]Trapoxin. A crude nuclear
extract was prepared from bovine thymus as described.15 The dialyzed
extract (0.5 mL in 25 mM tris pH 8, 10 mM NaCl, 10% glycerol) was
treated at 4 °C with 20 nM [3H]trapoxin (diluted from a 10 µM DMSO
stock) with and without 400 nM unlabeled trapoxin (or 1 µM
trichostatin, obtained from Wako Chemicals) for 30 min. A 10% (w/

Synthesis of Natural and Modified Trapoxins
J. Am. Chem. Soc., Vol. 118, No. 43, 1996 10421
solution was added via cannula the lithiated oxazolidinone 19 in 1.5
mL of THF, which had been prepared at -78 °C by treating (S)-(-)-
4-benzyl-2-oxazolidinone (119 mg, 0.67 mmol, 1.8 equiv based on 18)
with n-BuLi (2.3 M in hexane, 0.293 mL, 0.67 mmol, 1.8 equiv). After
15 min at -78 °C and 1.5 h at room temperature, the reaction was
quenched with 5% NaHSO4 (10 mL), and the solvents were removed
in vacuo. The residue was extracted with EtOAc (3×), and the
combined organic fractions were dried over Na2SO4 and concentrated.
Purification by flash chromatography (5:1 hexanes/EtOAc) afforded
the imide 20 (197 mg, 99% yield) as an oil: Rf 0.41 (1:4 EtOAc/
1H), 2.81 (dd, J ) 13.4, 9.4 Hz, 1H), 2.42 (s, 3H), 2.16 (m, 2H), 2.0-
1.5 (m, 4H), 1.35 (s, 3H), 1.32 (s, 3H); 13C NMR (100 MHz, CDCl3)
δ 170.7, 152.9, 145.0, 135.9, 134.8, 132.8, 129.8, 129.4, 129.0, 128.0,
127.4, 126.3, 109.8, 78.3, 73.3, 68.0, 66.6, 60.4, 55.3, 37.6, 30.7, 27.1,
27.0, 26.6, 25.9, 21.5; HRMS (FAB, NaI) Calcd for C30H36N4O8S +
Na 635.2151, found 635.2168.
(2S,8S,9S)-N-(Fluorenylmethoxycarbonyl)-2-amino-10-[(p-tolu-
enesulfonyl)oxy]-8,9-(isopropylidenedioxy)decanoic Acid (23). To
azide 22 (1.06 g, 1.73 mmol, 1.0 equiv) in 3:1 THF/water (40 mL)
was added LiOH (0.124 g, 5.25 mmol, 3.0 equiv) at 0 °C. The reaction
was quenched after 1 h with 1 N HCl (5.3 mL). The mixture was
concentrated and extracted with EtOAc. After drying over MgSO4 and
concentrating in vacuo, the crude azido acid was dissolved in 9:1 THF/
water (90 mL), treated with 10% Pd/C, and hydrogenated at 1 atm H2
for 36 h. The black mixture was filtered through Celite and
concentrated to give the crude amino acid. This compound was
dissolved in 20 mL of 25% DMF/THF and treated at 0 °C with 2,6-
lutidine (0.3 mL, 2.59 mmol, 1.5 equiv) and fluorenylmethyl-N-
hydroxysuccinimidylcarbonate (0.87 g, 2.59 mmol, 1.5 equiv). After
1.5 h, water (20 mL) was added, and the volatile solvents removed in
vacuo. The mixture was acidified with saturated NaHSO4 to pH 2 and
extracted with ether (3 × 30 mL). The organic phases were washed
with water and brine and dried over MgSO4. Purification by flash
chromatography (4:5:1 hexanes/EtOAc/MeOH, 0-1% AcOH) gave
hexanes); [R]20 +31.1° (c 0.80, CHCl3); IR (thin film) 2930, 2857,
D
1786, 1701, 1499, 1385, 1250, 1213 cm-1; 1H NMR (400 MHz, CDCl3)
δ 7.33-7.23 (m, 3H), 7.19 (m, 2H), 5.64 (m, 1H), 5.40 (m, 1H), 4.74
(m, 1H), 4.64 (m, 1H), 4.19-4.12 (m, 2H), 3.76 (m, 1H), 3.66-3.61
(m, 2H), 3.27 (dd, J ) 10.0, 3.3 Hz, 1H), 2.97-2.86 (m, 2H), 2.73
(dd, J ) 3.7, 9.6 Hz, 1H), 2.24-2.20 (m, 1H), 2.14-2.08 (m, 1H),
1.72-1.65 (m, 2H), 1.49-1.4 (m, 2H), 1.40 (s, 3H), 1.39 (s, 3H), 0.88
(s, 9H), 0.05 (s, 3H), 0.04 (s, 3H); 13C NMR (100 MHz, CDCl3) δ
172.9, 153.3, 135.2, 135.1, 129.3, 128.8, 127.2, 126.9, 108.6, 81.7,
72.8, 66.0, 61.6, 55.0, 37.8, 35.2, 29.0, 27.5, 27.2, 26.8, 25.8, 23.7,
18.2, -5.4, -5.5; HRMS (FAB, NaI) Calcd for C29H45NO6Si + Na
554.29140, found 554.2907.
(4S)-4-Benzyl-3-[(8S,9S)-10-[(p-toluenesulfonyl)oxy]-8,9-(isopro-
pylidenedioxy)-6(Z)-decenoyl]-2-oxazolidinone (21). Hydrogen fluoride/
pyridine complex (0.82 g) was cooled to 0 °C and treated with pyridine
(2.3 mL) and THF (3.9 mL). This solution was then added to silyl
ether 20 (2.7 g, 5.08 mmol, 1.0 equiv), and the resulting mixture was
allowed to warm to room temperature. After 3 h, additional HF/
pyridine/THF solution (5 mL) was added. The reaction was quenched
30 min later by the addition of saturated NaHCO3 followed by extraction
with ether (2 × 50 mL). After drying with MgSO4, the combined
organic fractions were concentrated and redissolved in dry CH2Cl2 (10
mL). Diisopropylethylamine (1.0 mL, 5.8 mmol, 1.1 equiv) was added,
and the solution was cooled in ice. Tosyl chloride (1.1 g, 5.8 mmol,
1.1 equiv) and DMAP (0.124 g, 1.02 mmol, 0.2 equiv) were then added.
After 4 h, more DMAP (total ) 5.0 mmol, 1.0 equiv) and tosyl chloride
(total ) 7.5 mmol, 1.5 equiv) were added, and the reaction was stirred
for a total of 15 h. The mixture was diluted with EtOAc (100 mL),
washed with 5% NaHCO3, dried over MgSO4, and purified by flash
chromatography (1:3 EtOAc/hexanes) to yield 2.57 g of tosylate 21
(89% yield) as an oil: Rf 0.19 (1:3 EtOAc/hexanes); [R]20D +32.9° (c
3.8, EtOAc); IR (thin film) 2936, 1778, 1700, 1599, 1454, 1360, 1177,
Fmoc amino acid 23 (582 mg, 53% over three steps) as an oil: [R]20
D
-4.0° (c 1.4, CHCl3); IR (thin film) 3250, 2986, 2938, 1719, 1522,
1
1366, 1177, 980 cm-1; H NMR (400 MHz, CDCl3) δ 7.8-7.7 (m,
4H), 7.6-7.5 (m, 2H), 7.4-7.25 (m, 6H), 5.40 (d, J ) 8.2 Hz, 1H),
4.55-4.35 (m, 3H), 4.20 (m, 1H), 4.08 (dd, J ) 10.6, 4.0 Hz, 1H),
4.03 (dd, J ) 10.6, 4.5 Hz, 1H), 3.8-3.75 (m, 2H), 2.40 (s, 3H), 1.95-
1.83 (m, 1H), 1.8-1.65 (m, 1H), 1.55-1.45 (m, 2H), 1.5-1.2 (m, 6H),
1.33 (s, 3H), 1.28 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 177.1, 156.1,
145.0, 143.8, 141.2, 132.6, 129.9, 128.0, 127.7, 127.1, 125.1, 120.0,
109.4, 78.1, 77.7, 69.1, 67.1, 53.7, 47.1, 32.7, 32.1, 28.9, 27.2, 26.6,
25.5, 25.0, 21.6; HRMS (FAB, NaI) Calcd for C35H41NO9S + Na
674.2400, found 674.2419.
NE-(Allyloxycarbonyl)-L-lysyl-L-phenylalanyl-D-proline Methyl
Ester (24). The trifluoroacetate salt of L-phenylalanyl-D-proline methyl
ester (566 mg, 1.45 mmol, 1 equiv) was combined with NR-(tert-
butoxycarbonyl)-Nꢀ-(allyloxycarbonyl)-L-lysine (481 mg, 1.45 mmol,
1 equiv), NMM (0.176 mL, 1.6 mmol, 1.1 equiv), HOBT (196 mg,
1.6 mmol, 1.1 equiv), and EDC (306 mg, 1.6 mmol, 1.1 equiv) in CH2-
Cl2 (18 mL). After 18 h, the crude mixture was purified directly by
flash chromatography (67-80% EtOAc/hexanes) to give the tripeptide
(559 mg, 64% yield) as an oil. Boc deprotection was carried out in 16
mL of 1:1 CH2Cl2/trifluoroacetic acid for 1 h. Solvents were removed,
and the crude trifluoroacetate was concentrated several times from 50%
heptane/CH2Cl2. This compound was characterized as the free base
24: [R]20D +23.0° (c 3.9, CHCl3); IR (thin film) 3320 (br), 2948, 2876,
1
948, 820, 664 cm-1; H NMR (400 MHz, CDCl3) δ 7.73 (d, J ) 8.3
Hz, 2H), 7.3-7.1 (m, 7H), 5.64 (m, 1H), 5.29 (m, 1H), 4.7-4.5 (m,
2H), 4.15-4.05 (m, 3H), 3.98 (dd, J ) 10.8, 4.7 Hz, 1H), 3.73 (m,
1H), 3.22 (dd, J ) 13.4, 3.1 Hz, 1H), 3.0-2.7 (m, 2H), 2.70 (dd, J )
13.4, 9.6 Hz, 1H), 2.38 (s, 3H), 2.2-2.0 (m, 2H), 1.7-1.6 (m, 2H),
1.45-1.35 (m, 2H), 1.31 (s, 3H), 1.28 (s, 3H); 13C NMR (100 MHz,
CDCl3) δ 173.0, 153.4, 144.8, 136.5, 135.3, 133.0, 129.8, 129.3, 128.9,
128.0, 127.2, 125.8, 109.7, 78.4, 73.1, 67.9, 66.2, 55.1, 37.9, 35.2, 28.9,
27.4, 27.1, 26.6, 23.8, 21.5; HRMS (FAB, NaI) Calcd for C30H37NO8S
+ Na 594.2137, found 594.2148.
1
1744, 1721, 1644, 1530, 1447, 1246; H NMR (400 MHz, CD3OD,
mixture of rotamers) δ 7.32-7.17 (m, 5H), 5.91 (m, 1H), 5.28 (d, J )
17.2 Hz, 1H), 5.16 (d, J ) 9.2 Hz, 1H), 4.93 (m, 1H), 4.50 (m, 2H),
4.25 (m, 1H), 3.86 (m, 1H), 3.66 (s, 3H), 3.2-2.9 (m, 6H), 2.0-1.3
(m, 10 H); 13C NMR (100 MHz, CDCl3, mixture of rotamers) δ 172.8,
172.2, 171.0, 169.9, 156.3, 136.3, 133.0, 129.4, 129.2, 128.3, 126.9,
126.6, 117.3, 65.2, 59.3, 58.6, 52.5, 52.2, 51.8, 46.8, 46.5, 40.5, 39.4,
37.4, 30.9, 29.5, 28.8, 24.3, 22.5, 21.9; HRMS (FAB, NaI) Calcd for
C25H36N4O6 + Na 511.2533, found 511.2524.
(4S)-4-Benzyl-3-[(2S,8S,9S)-2-azido-10-[(p-toluenesulfonyl)oxy]-
8,9-(isopro-pylidenedioxy)-6(Z)-decenoyl]-2-oxazolidinone (22). To-
sylate 21 (2.57 g, 4.5 mmol, 1.0 equiv) was azeotropically dried with
toluene under high vacuum and dissolved in dry THF (110 mL). To
the cooled solution (-78 °C) was added potassium hexamethyldisilazide
(0.67 M in toluene, 10.1 mL, 6.8 mmol, 1.5 equiv). After 30 min, a
cooled solution (-78 °C) of 2,4,6-triisopropylbenzenesulfonylazide
(2.09 g, 6.8 mmol, 1.5 equiv) in 19 mL of THF was added via cannula.
Exactly 3 min later, AcOH (1.2 mL) was added, and the mixture
warmed to 35 °C with a water bath. After 1.5 h, the mixture was
concentrated and partitioned between EtOAc and 5% NaHCO3 (50 mL
each). The organic fraction was dried over Na2SO4, concentrated, and
purified by flash chromatography (25% EtOAc/hexanes) to provide the
azide 22 (2.0 g, 73% yield) as an oil: Rf 0.20 (1:2 EtOAc/hexanes);
(2S,8S,9S)-N-(Fluorenylmethoxycarbonyl)-2-amino-10-[(p-tolu-
enesulfonyl)oxy]-8,9-(isopropylidenedioxy)decanoyl-NE-(allyloxycar-
bonyl)-L-lysyl-L-phenylalanyl-D-proline Methyl Ester (25). The
trifluoroacetate salt of tripeptide 24 (155 mg, 0.252 mmol, 1.1 equiv)
was combined with Fmoc amino acid 23 (145 mg, 0.229 mmol, 1.0
equiv), NMM (0.028 mL, 0.25 mmol, 1.1 equiv), HOBT (34 mg, 0.25
mmol, 1.1 equiv), and EDC (48 mg, 0.25 mmol, 1.1 equiv) in CH2Cl2
(3.2 mL). After 18 h, the crude mixture was purified directly by flash
chromatography (0-2-5% MeOH/CH2Cl2) to give tetrapeptide 25 (206
[R]20 +67.0° (c 3.5, EtOAc); IR (thin film) 2986, 2108, 1782, 1703,
D
1368, 1177, 982 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.75 (d, J ) 8.4
Hz, 2H), 7.35-7.15 (m, 7H), 5.52-5.51 (m, 1H), 5.36 (m, 1H), 4.92
(dd, J ) 9.2, 4.2 Hz, 1H), 4.647 (m, 1H), 4.59 (m, 1H), 4.25 (m, 1H),
4.18 (dd, J ) 9.1, 2.9 Hz, 1H), 4.09 (dd, J ) 10.8, 4.0 Hz, 1H), 4.01
(dd, J ) 10.8, 4.6 Hz, 1H), 3.77 (m, 1H), 3.29 (dd, J ) 13.4, 3.2 Hz,
mg, 81% yield) as an oil: Rf 0.45 (EtOAc); [R]20 +3.1° (c 0.91,
D
CHCl3); IR (film) 3298 (br), 2938, 1721, 1638, 1532, 1451, 1177, 984
1
cm-1; H NMR (400 MHz, CDCl3, mixture of rotamers) δ 7.8-7.7
(m, 4H), 7.7-7.6 (m, 2H), 7.4-7.1 (m, 11H), 7.95-7.8 (br, 1H), 5.9-
5.75 (m, 2H), 5.45 (br s, 1H), 5.3-5.1 (m, 2H), 5.0-4.85 (br m, 1H),

10422 J. Am. Chem. Soc., Vol. 118, No. 43, 1996
Taunton et al.
4.6-4.15 (br m), 4.1-3.9 (m, 2H), 3.8-3.45 (br m), 3.61 (br s, 3H),
3.2-2.9 (br m, 4H), 2.75-2.65 (m, 1H), 2.39 (s, 3H), 1.9-1.2 (br m),
1.29 (br s, 3H), 1.23 (s, 3H); 13C NMR (100 MHz, CDCl3, mixture of
rotamers) δ 172.0, 171.7, 170.5, 169.6, 156.4, 144.8, 143.8, 141.2,
136.0, 133.1, 132.8, 129.8, 129.3, 129.1, 128.4, 128.3, 127.9, 127.6,
127.0, 125.0, 119.8, 117.2, 109.2, 96.0, 78.0, 77.7, 69.1, 67.0, 65.2,
59.3, 58.6, 55.0, 53.3, 52.9, 52.4, 52.2, 52.1, 47.1, 46.7, 40.3, 39.3,
32.8, 32.3, 29.2, 29.0, 28.8, 27.2, 26.6, 25.5, 25.2, 24.2, 22.4, 22.1,
21.9, 21.5; MS (FAB, NaI) Calcd for C60H75N5O14S + Na 1144, found
1144.
white solid (22.5 mg, 66% yield over three steps): [R]20 -83.6° (c
D
0.45, CHCl3); IR (thin film) 3293 (br), 2934, 1715, 1682, 1661, 1526,
1
1454, 1244; H NMR (400 MHz, CDCl3) δ 7.3-7.15 (m, 5H), 7.06
(d, J ) 10.2 Hz, 1H), 6.60 (d, J ) 5.4 Hz, 1H), 5.88 (m, 1H), 5.26 (d,
J ) 17.1 Hz, 1H), 5.18 (d, J ) 10.3 Hz, 1H), 5.10 (m, 1H), 4.81 (br
s, 1H), 4.65 (d, J ) 6.9 Hz, 1H), 4.52 (m, 2H), 4.21 (m, 1H), 3.82 (m,
1H), 3.49 (m, 1H), 3.39 (m, 1H), 3.2-3.0 (m, 3H), 3.0-2.9 (m, 2H),
2.83 (dd, J ) 5.7, 1.3 Hz, 1H), 2.40 (m, 1H), 2.4-2.1 (m, 5H), 1.9-
1.7 (m, 2H), 1.7-1.4 (m, 6H), 1.4-1.2 (m, 8H); 13C NMR (100 MHz,
CDCl3) δ 207.5, 174.9, 174.2, 172.8, 171.6, 156.3, 136.8, 133.0, 129.0,
128.6, 126.8, 117.6, 77.2, 65.4, 61.4, 57.8, 54.0, 53.4, 46.9, 46.1, 40.5,
36.2, 35.7, 29.7, 29.2, 28.8, 28.6, 25.3, 24.9, 24.8, 23.6, 22.7; HRMS
(FAB, NaI) Calcd for C34H47N5O8 + Na 676.3322, found 676.3354.
K-Trap Affinity Matrix (28). K-Trap (27) (1 mg, 1.5 µmol, 1
equiv) was concentrated to dryness from a CH2Cl2 stock solution in a
0.5 mL conical vial. A solution of 5,5-dimethyl-1,3-cyclohexanedione
(0.06 M in THF, 0.2 mL, 12 µmol, 8 equiv) was added, and the mixture
was degassed by three cycles of freeze-pump-thaw. A freshly
prepared solution of tetrakis(triphenylphosphine)palladium33 (0.01 M
in degassed THF, 40 µL, 0.4 µmol, 0.27 equiv) was then added, and
the reaction vial was immediately placed in a 35 °C water bath. After
30 min, the mixture was concentrated to dryness, dissolved in 1 mL of
MeOH, and filtered through Celite, which was rinsed with 0.5 mL
MeOH and 0.5 mL of water. Purification by semipreparative HPLC
(2 × 1 mL injections; Beckman Ultrasphere ODS, 1 × 25 cm; 30%
MeOH/0.1% aqueous TFA for 5 min, followed by 30-100% MeOH
gradient over 25 min; 5 mL/min flow rate), afforded the amine as its
TFA salt. This material was dissolved in a mixture of DMSO (0.1
mL) and THF (2 mL) and added to 1 mL of packed Affi-Gel 10
(Biorad), which had been washed with 10 mL of THF in a 5 mL
polyethylene column (Biodrad). The slurry was treated with diisopro-
pylethylamine (0.01 mL) and shaken for 2 h at room temperature.
Ethanolamine (0.01 mL) was then added, and the slurry was shaken
for an additional 3 h. The K-trap affinity matrix was washed with 15
mL of isopropyl alcoholl and stored as a 50% isopropyl alcohol slurry
at -20 °C.
cyclo[(2S,8S,9S)-2-Amino-10-[(p-toluenesulfonyl)oxy]-8,9-(isopro-
pylidenedioxy)-decanoyl-NE-(allyloxycarbonyl)-L-lysyl-L-phenylala-
nyl-D-prolyl] (26). To a THF/water solution (3:1, 2.6 mL) of
tetrapeptide 25 (114 mg, 0.102 mmol, 1.0 equiv) was added 1 N LiOH
(0.347 mL, 0.347 mmol, 3.4 equiv). After 2.5 h, the reaction was
neutralized by the addition of 1 N HCl (0.347 mL), diluted with MeCN
(10 mL), and extracted with hexanes (3 × 5 mL). The MeCN fraction
was concentrated, azeotropically dried with benzene (3 × 2 mL), and
dissolved in dry DMF (2.5 mL). This solution was added by syringe
pump over 12 h to a DMF solution (110 mL) of BOP (270 mg, 0.61
mmol, 6 equiv) and DMAP (100 mg, 0.82 mmol, 8 equiv). After the
addition was complete, the reaction was stirred for 2 h and the DMF
was removed by vacuum distillation (bath temperature <40 °C). The
residue was partitioned between EtOAc (50 mL) and 0.5 M citrate (pH
5, 20 mL) and washed with 5% K2CO3 and brine. The organic phase
was dried over Na2SO4, concentrated, and dissolved in MeCN (2 mL).
Purification by preparative reverse phase HPLC (Rainin Dynamax 300-
A; 30-100% MeCN/0.1% aqueous TFA gradient over 30 min; 20 mL/
min flow rate) provided cyclotetrapeptide 26 (51 mg, 58% yield) as a
white solid: Rf 0.37 (5% MeOH/EtOAc); [R]20 -63.8° (c 0.65,
D
CHCl3); IR (thin film) 3297 (br), 2938, 1663, 1524, 1454, 1368, 1244,
1
1177 cm-1; H NMR (400 MHz, CDCl3) δ 7.77 (d, J ) 8.2 Hz, 2H),
7.33 (d, J ) 8.2 Hz, 2H), 7.3-7.15 (m, 5H), 7.06 (d, J ) 10.2 Hz,
1H), 6.60 (d, J ) 6.0 Hz, 1H), 5.87 (m, 1H), 5.26 (dd, J ) 17.2, 1.1
Hz, 1H), 5.17 (d, J ) 10.3 Hz, 1H), 5.08 (m, 1H), 4.82 (br m, 1H),
4.65 (d, J ) 7.5 Hz, 1H), 4.6-4.45 (m, 2H), 4.22 (m, 1H), 4.08 (dd,
J ) 10.8, 4.1 Hz, 1H), 4.03 (dd, J ) 10.8, 4.4 Hz, 1H), 3.9-3.7 (m,
1H), 3.50 (m, 1H), 3.22-3.1 (m, 4H), 2.94 (dd, J ) 13.5, 6.1 Hz, 1H),
2.43 (s, 3H), 2.35-2.25 (m, 1H), 2.25-2.05 (m, 4H), 1.8-1.2 (m, 15H),
1.33 (s, 3H), 1.27 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 174.9, 174.2,
172.8, 171.5, 156.2, 145.0, 136.8, 133.0, 132.7, 129.9, 128.6, 128.0,
126.8, 117.5, 109.3, 78.1, 77.8, 77.2, 69.1, 65.4, 61.3, 57.8, 54.1, 53.4,
46.8, 40.5, 35,7, 32.8, 29.2, 29.0, 28.9, 28.8, 27.2, 26.6, 25.6, 25.3,
24.8, 23.5, 21.6; HRMS (FAB, NaI) Calcd for C44H61N5O11S + Na
890.3986, found 890.4012.
Charcoal Binding with 125I-Trap Followed by SDS-PAGE.
Nuclear extract (0.2 mL) was incubated with 1 nM 125I-trap with and
without 1 µM unlabeled trapoxin for 30 min at 4 °C. Charcoal was
added and sedimented (see above), and 0.15 mL of the clear supernatant
was treated for 30 min at room temperature with 50 µL of nonreducing
sample buffer for sodium dodecylsulfate polyacrylamide gel electro-
phoresis (4× SDS-PAGE sample buffer: 8% SDS, 40% glycerol, 0.8%
bromophenol blue, 0.25 M tris pH 6.8). As a control, pure 125I-trap
(0.5 nM, 0.15 mL) was processed identically for electrophoresis.
Samples (40 µL) were electrophoresed (12.5% gel) until the dye front
had migrated through 80% of the gel. The gel was stained with
Coomasie brilliant blue, dried, and exposed to an imaging plate (Fuji).
Images were processed with the Fujix BAS1000 image analyzer using
MacBAS imaging software.
K-Trap (27). Cyclotetrapeptide 26 (45 mg, 0.052 mmol) was
dissolved in 1.0 mL of THF and treated with 1.0 mL of 6.2% aqueous
HCl for 16 h. The reaction mixture was diluted with saturated NaHCO3,
extracted with CH2Cl2 (5×), and dried over Na2SO4. The crude diol
was next dissolved in a mixture of DMSO (1 mL) and MeOH (5mL),
to which was added DBU (25 µL, 0.16 mmol, 3.2 equiv) at 0 °C. After
1 h, the volume was reduced to 2 mL, and 20 mL of EtOAc was added
and extracted sequentially with water, 0.5 M citrate (pH 5), half-
saturated NaHCO3, and brine. The aqueous fractions were separately
back extracted with CHCl3 (5 × 10 mL). The combined organic
fractions were dried over Na2SO4 and concentrated to give the crude
epoxy alcohol, which was further dried by concentrating from toluene
(2 × 0.5 mL).
Acknowledgment. This paper is dedicated to Professor
Yoshito Kishi for his leadership and vision in organic synthesis
and its surrounding disciplines. This research was supported
by the National Institute of General Medical Sciences. Mass
spectral data were obtained by Dr. Andrew Tyler and colleagues
at the Harvard University Mass Spectrometry Facility. We
acknowledge the NIH BRS Shared Instrumentation Grant
Program (1 S 10 RR01748-01A1) and NSF (CHE88-14019) for
providing NMR facilities. We thank Timothy Jamison and
Julian Simon for critically reading the manuscript. J.T. thanks
the National Science Foundation and Eli Lilly and Company
for funding. J.L.C. was funded by the American Cancer Society.
S.L.S. is an investigator at the Howard Hughes Medical Institute.
To the epoxy alcohol in CH2Cl2 (3 mL) was added sequentially
DMSO (0.25 mL), EDC (52 mg, 0.27 mmol, 8 equiv), and dichloro-
acetic acid (45 µL, 0.016 mmol, 0.5 equiv). After 8 h, the reaction
was quenched with half-saturated NaHCO3 (4 mL) and extracted with
CH2Cl2 (4 × 5 mL). The combined organic fractions were dried over
Na2SO4, concentrated, redissolved in MeOH (2 mL), and filtered
through Celite. Purification by preparative HPLC (Rainin Dynamax
300-A; 0.1% aqueous TFA for 5 min, followed by a 0-50% MeCN
gradient over 25 min; 18 mL/min flow rate) provided K-trap (27) as a
JA9615841