Total synthesis and evaluation of vinblastine analogues containing systematic deep-seated modifications in the vindoline subunit ring system: Core redesign

Article abstract of DOI:10.1021/jm3014376

The total synthesis of a systematic series of vinblastine analogues that contain deep-seated structural modifications to the core ring system of the lower vindoline subunit is described. Complementary to the vindoline 6,5 DE ring system, compounds with 5,5, 6,6, and the reversed 5,6 membered DE ring systems were prepared. Both the natural cis and unnatural trans 6,6-membered ring systems proved accessible, with the latter representing a surprisingly effective class for analogue design. Following Fe(III)-promoted coupling with catharanthine and in situ oxidation to provide the corresponding vinblastine analogues, their evaluation provided unanticipated insights into how the structure of the vindoline subunit contributes to activity. Two potent analogues (81 and 44) possessing two different unprecedented modifications to the vindoline subunit core architecture were discovered that matched the potency of the comparison natural products and both lack the 6,7-double bond whose removal in vinblastine leads to a 100-fold drop in activity.

Full text of DOI:10.1021/jm3014376

Article
pubs.acs.org/jmc
Total Synthesis and Evaluation of Vinblastine Analogues Containing
Systematic Deep-Seated Modifications in the Vindoline Subunit Ring
System: Core Redesign
Kristin D. Schleicher, Yoshikazu Sasaki, Annie Tam, Daisuke Kato, Katharine K. Duncan,
and Dale L. Boger*
Department of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research Institute, 10550 North Torrey Pines
Road, La Jolla, California 92037, United States
S
* Supporting Information
ABSTRACT: The total synthesis of a systematic series of
vinblastine analogues that contain deep-seated structural
modifications to the core ring system of the lower vindoline
subunit is described. Complementary to the vindoline 6,5 DE
ring system, compounds with 5,5, 6,6, and the reversed 5,6
membered DE ring systems were prepared. Both the natural cis
and unnatural trans 6,6-membered ring systems proved
accessible, with the latter representing a surprisingly effective
class for analogue design. Following Fe(III)-promoted
coupling with catharanthine and in situ oxidation to provide
the corresponding vinblastine analogues, their evaluation
provided unanticipated insights into how the structure of the
vindoline subunit contributes to activity. Two potent analogues (81 and 44) possessing two different unprecedented
modifications to the vindoline subunit core architecture were discovered that matched the potency of the comparison natural
products and both lack the 6,7-double bond whose removal in vinblastine leads to a 100-fold drop in activity.
introduction of Vinca alkaloids into the clinic and the seminal
identification of their mechanism of action, tubulin remains
among the more successful therapeutic targets for cancer
chemotherapy.7 Previously, we reported the development of a
concise total synthesis of (−)- and ent-(+)-vindoline8−10
enlisting a tandem intramolecular [4 + 2]/[3 + 2] cycloaddition
cascade of 1,3,4-oxadiazoles,11 the extension of this method-
ology to the preparation of a series of related natural
products,12 and the subsequent asymmetric total synthesis of
vindoline and vindorosine (2).13 Central to this work was the
use of a biomimetic Fe(III)-promoted coupling14 of vindoline
with catharanthine (3) and the additional development of a
subsequent in situ Fe(III)-mediated alkene oxidation6i,j for
C20′-alcohol introduction that allows for their single-step
incorporation into total syntheses of vinblastine, related natural
products including vincristine, and key analogues in routes as
short as 8−12 steps.6j
INTRODUCTION
■
Vindoline (1),1,2 a major alkaloid of Catharanthus roseus,
comprises the more complex lower subunit of vinblastine
(4)2−5 and serves as both a biosynthetic3,4 and synthetic6
precursor to this natural product (Figure 1). Vinblastine and
vincristine (5), originally isolated in trace quantities from
Catharanthus roseus (L.) G. Don, inhibit assembly of tubulin
into microtubules, arresting mitosis. Half a century after the
Prior to these efforts, the majority of vinblastine analogues
have been prepared by modification of accessible peripheral
sites on the natural products, with the disclosure of a limited
number of analogues that contain deep-seated structural
changes.15 As a result, we expanded the scope of the Fe(III)-
promoted biomimetic coupling reaction to prepare vinblastine
Received: October 3, 2012
Published: December 19, 2012
Figure 1. Natural product structures.
© 2012 American Chemical Society
483
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
analogues containing key modifications to the upper velban-
amine subunit with replacements of the C16′ methyl ester,16
incorporation of a systematic series of C10′ and C12′
substituents,17 and introduction of previously inaccessible
C20′ functionality.18 In addition to serving to probe and
redefine the mechanistic details of the Fe(III)-promoted
coupling reaction,14b the studies have identified unappreciated
structural features of the natural products contributing to their
target binding and functional activity16 and have provided
altered structures that match or surpass17,18 the potency of the
clinical drugs.
With respect to modifications to the lower vindoline subunit,
including systematic changes to the C5 ethyl substituent,19 one
of the most remarkable observations confirmed in our initial
studies was the impact of removing the C6−C7 double bond.6j
Δ6,7-Dihydrovinblastine was found to be nearly 100-fold less
active than vinblastine, indicating that this region of the natural
product has a pronounced impact on its biological properties.
Because the X-ray crystal structure of vinblastine bound to
tubulin20 does not reveal an obvious stabilizing interaction with
the olefin that might account for the difference, we have begun
to explore its origin. In conjunction with our studies on the
asymmetric total synthesis of vindoline, analogues 6−9
containing a heteroatom β to the basic amine, including several
compounds in which the olefin-containing six-membered ring
was replaced with a saturated five-membered ring, were
prepared and coupled with catharanthine to afford the
corresponding vinblastine analogues (Figure 2).13b These C7-
core ring system. Complementary to the 6,5 DE ring system of
vindoline, compounds with 5,5, 6,6, and the reversed 5,6
membered DE ring systems were targeted to systematically
address the impact of ring size and conformation on activity.
Preparation of these analogues, not accessible from natural
product sources, further demonstrates the versatility of the
intramolecular [4 + 2]/[3 + 2] cycloaddition cascade. The
cumulative examination of the vinblastine analogues bearing the
key vindoline ring system modifications not only provided
analogues containing two different unprecedented deep-seated
changes to the core of the lower vindoline subunit whose
activity matched that of vinblastine, but it also revealed that the
impact of the vinblastine C6−C7 double bond is not the result
of a pronounced π-stabilizing interaction with tubulin and that
it does not appear to be related to its impact on the N9 pKa, its
lone pair orientation, or even the precise spatial location of N9.
In addition to discovering that alternative lower subunit ring
systems without the key double bond may be equally effective
even though the double bond removal in vinblastine results in a
100-fold loss in activity, the results of the studies suggest that
such deep-seated core redesign of biologically active natural
products should be more frequently explored than is presently
considered.
RESULTS AND DISCUSSION
■
Synthesis of the Vindoline Analogue Containing a 5,5
DE Ring System and Incorporation into the Correspond-
ing Vinblastine Analogue. Because 6 and 7 contained an
extraneous C7 hydroxymethyl substituent that made the
interpretation of their properties unclear, we targeted the C7
unsubstituted system 19. Relative to vinblastine, this entails
removal of the 6,7 double bond and ring contraction of the six-
membered D ring to provide the vinblastine analogue
containing a modified vindoline core 5,5 DE ring system.
A unique synthesis of the vindoline analogue 18 containing
this alternative 5,5 DE ring system was developed (Scheme 1).
In the course of studies to clarify the mechanism of a key ring
expansion reaction used in our asymmetric synthesis of
vindoline,13 we observed that warming a solution of the
primary alcohol 10 (vs tosylate 11) predictably did not lead to
ring expansion but did provide the unanticipated products 13
and 1421 derived from oxidative cleavage of the hydroxymethyl
substituent. Although the mechanistic details of this unusual
transformation are not yet fully defined, the conversion was
improved by conducting the reaction in acetonitrile (50 °C, 24
h, and 70 °C, 48 h) versus water−tetrahydrofuran (THF),
providing 13 (51%) in surprisingly good conversion.
Consistent with studies that indicated the reaction was
relatively insensitive to the presence of O2, that it was less
effective if O2 was deliberately introduced, and that it was more
effective with its careful exclusion, a plausible mechanism for
the generation of 13 entails reversible iminium ion generation
followed by elimination of formaldehyde to generate a
stabilized singlet carbene that undergoes a subsequent insertion
with the proximal C3 alcohol for reclosure of the N,O-acetal
(eq 1). Without further optimization, extending this reaction to
15, bearing the C16-methoxy substituent, provided 16 in
comparable conversion (48%, CH3CN, 50 °C, 72 h). Reductive
cleavage of the cyclic N,O-acetal (NaCNBH3, 10% i-PrOH−
THF, 3 equiv of HCl, 0 °C) afforded the key vindoline
analogue 18 (84%). Without optimization, 18 was coupled with
catharanthine (3) using the Fe(III)-promoted coupling and
Figure 2. Compounds used to probe changes to the C6−C7 region of
the vindoline subunit.
substituted six-membered or five-membered analogues only
matched or were a further 10-fold less potent than Δ6,7-
dihydrovinblastine, being 100- to 1000-fold less active than
vinblastine itself and indicating that the olefin impact may not
be due simply to its influence on the N9 pKa. In light of this
unappreciated major role for the C6−C7 region in establishing
functional activity and its unknown origin, we elected to probe
the steric and conformational features of the vindoline ring
system that might be responsible for these effects.
Herein, we report the synthesis of a series of analogues
bearing deep-seated structural modifications to the vindoline
484
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
matched human colon cancer cell line (HCT116/VM46) that
is resistant to vinblastine by virtue of overexpression of the drug
exporter Pgp. The vinblastine analogue 19 containing the ring
contracted D ring in the vindoline core proved to be 10-fold
more potent than 6 and 7, indicating that the C7
hydroxymethyl groups of the latter compounds were indeed
negatively impacting their properties (Figure 3).22 However,
Scheme 1
Figure 3. Activity of 5,5 DE ring system analogues.
unlike derivatives that follow, 19 remained 10- to 100-fold less
potent than vinblastine, displaying activity that was essentially
equivalent to 6,7-dihydrovinblastine (21) lacking the C6−C7
double bond. Similarly, although the anhydrovinblastine series
is 10-fold less potent, identical trends were observed. The
anhydrovinblastine analogue 20 was 10-fold less active than
anhydrovinblastine (22) itself, and 20 proved to be equipotent
to 6,7-dihydroanhydrovinblastine (23). In essence, the D ring
contraction in going from 21 to 19 or 23 to 20 had no impact
on the biological properties, but it also did not regain the
activity lost with removal of the C6−C7 double bond.
Synthesis of Vindoline Analogues Containing a 6,6 DE
Ring System and Incorporation into the Corresponding
Vinblastine Analogues. The vinblastine analogues contain-
ing modified vindoline subunits bearing a ring expanded 6,6 DE
ring system while not yet being the most potent variation,
proved to be among the most interesting and revealing. The
syntheses of the 6,6 DE ring system vindoline analogues closely
parallel our synthesis of vindoline itself, with the key difference
being the use of a cascade cycloaddition substrate with a
lengthened tether to the indole dipolarophile. Wittig olefination
of aldehyde 2410 provided 25 smoothly in 88% yield (Scheme
2). Hydrogenation of the olefin required use of Raney nickel to
subsequent oxidation to afford the vinblastine analogue 19, as
well as the anhydrovinblastine 20.
As in related studies,6j,16−19 the compounds were examined
for inhibition of cell growth against a mouse leukemia cell line
(L1210), a human colon cancer cell line (HCT116), and a
485
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
a variety of conditions (o-Cl2C6H4 or triisopropylbenzene, 0.5−
2 mM, 180−230 °C, 24−72 h). Both provided lactone products
derived from interception of the initial [4 + 2] cycloadditions
that stall and fail to progress through the [3 + 2] cycloaddition.
In contrast, both 35 and (E)-36, containing less substituted
initiating dienophiles, participated in the cycloaddition cascade
effectively (Scheme 3). Especially notable was the reaction of
Scheme 2
Scheme 3
(E)-36, lacking only the ethyl group of (E)-34, which provided
a single diastereomer 39 of the cascade cycloadduct in superb
conversion (68−87%, 2−0.1 mM) in relatively short reaction
times (6 h, 230 °C). This reaction was found to proceed
effectively even at the lower temperature of 180 °C in
triisopropylbenzene (TIPB, 50%, 2 mM), albeit requiring
longer reaction times (12 h), indicating that trisubstitution of
the initiating dienophile slows and, in the case of (E)-34,
ultimately precludes the subsequent [3 + 2] cycloaddition.
To initially assess the impact of a vindoline 6,6 DE ring
system, we elected to carry the cycloadduct 37 forward.
Although it was formed in lower conversions than 39, it serves
as a precursor to a series of vindoline analogues lacking only the
C4 acetoxy group, a substituent whose impact on the properties
of vinblastine are well established (10-fold reduction in
potency).6j This decision was fortunate in that it led to the
discovery of an unprecedented stereochemical impact on the
properties of the resulting vinblastine analogues. The cyclo-
adduct 37 could be resolved by chiral phase HPLC using a
semipreparative Daicel Chiralcel OD column either at this stage
or following conversion with Lawesson’s reagent23 to the
thiolactam 40 (Scheme 4), the absolute stereochemistry of
which was unambiguously established by X-ray crystallog-
raphy.24 Treatment of (+)-40, possessing the required absolute
stereochemistry, with Me3OBF4 and subsequent reduction of
the resulting S-methyl iminium ion (NaBH4)25 provided
exclusively 42 in superb yield (86%), bearing the unnatural
generate the saturated tether, as use of palladium on carbon
provided a mixture of the desired product 26 and overreduction
of the indole. Hydrolysis of the methyl ester cleanly provided
the corresponding carboxylic acid, which was converted to the
carboxamide 27 through reaction with 1,1′-carbonyldiimidazole
(CDI) followed by addition of NH4OH. Reduction of 27 by
LiAlH4 provided amine 28. After coupling with CDI to give 29,
reaction with methyl oxalylhydrazide and subsequent dehy-
drative ring closure (TsCl, Et3N) provided the 1,3,4-oxadiazole
30. Coupling of 30 with a series of carboxylic acids tethered to
candidate initiating dienophiles, including (E)-31 or (Z)-31,
generated the cycloaddition precursors 34−36.
In prior studies, significant differences in the rate and facility
of the cycloaddition cascade were observed in which either the
initiating [4 + 2] or subsequent [3 + 2] cycloaddition was
found to be rate limiting, depending on the substitution and
stereochemistry of the initiating dienophile.10,11 The (Z)-enol
ether dienophiles permit the direct introduction of the naturally
occurring C4 stereochemistry, while (E)-enol ether dienophiles
provide the C4 isomer requiring subsequent inversion of
configuration at this stereocenter. Typically, the reaction rate
for the (Z)-isomer is slower, being limited by a rate
determining [3 + 2] cycloaddition, and the reaction often
requires additional dilution to maximize conversion to the
desired cycloadduct. Unfortunately, neither (E)-34 nor (Z)-34
afforded the cascade cycloaddition product upon heating under
486
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
Scheme 4
Figure 4. Activity of saturated 6,6 DE ring system analogues.
this cis isomer, 10-fold more potent than 45, equipotent to the
natural product 46, and merely 10-fold less potent than
vinblastine in spite of large structural differences in the
vindoline portion of the molecule. This includes not only the
incorporation of the unnatural trans fused and ring expanded
6,6 DE ring system but also the removal of both the C6−C7
double bond and the C4 acetoxy group.
stereochemistry at C9 and a trans ring structure. Mistakenly
expecting that only the natural cis ring stereochemistry would
be active when incorporated into a vinblastine analogue,
considerable effort was spent searching for conditions that
would effect reductive oxido bridge cleavage to produce the
isomer 41. In these efforts, we found that 41 could be obtained
by a stepwise Raney nickel desulfurization of the thioamide
followed by acid-catalyzed oxido bridge opening and in situ
reduction of the resulting iminium ion with H2/PtO2.
Satisfactory conversions were observed only at increased
hydrogen pressures (250 psi) and upon prolonged reaction
times (3.5 d), providing up to 28% of 41, with competitive
formation of its isomer 42 (42%). Without optimization, both
41 and the more readily accessible 42 were coupled with
catharanthine (3) to assemble the respective vinblastine
analogues 43 and 44 with an expanded E ring in the vindoline
core.
Because the removal of the C4 acetoxy group has been
established to contribute 10-fold to the potency of vinblastine
(4),6j both 43 and 44 were anticipated to be less active than
vinblastine. A more direct comparison of their biological
activity can be made to 4-desacetoxy-6,7-dihydrovinblastine
(45)6j and the additional natural product 4-desacetoxyvinblas-
tine (46).3h,26 Relative to 45, the change to the cis 6,6 DE
system in 43 resulted in a further 10-fold reduction in potency
(Figure 4). Remarkably, the vinblastine analogue 44 incorpo-
rating the synthetically more accessible unnatural trans fused
6,6 DE ring system was found to be 100-fold more active than
In light of these observations, we elected to establish the
impact of introducing the C6−C7 olefin into the analogues 43
and 44. Accordingly, α-hydroxylation of 37 and subsequent
acetylation of the secondary alcohol 47 afforded 48 (Scheme
5).27 Treatment of 48 with Lawesson’s reagent afforded
thiolactam 49 in acceptable yield, albeit with significant
amounts of recovered starting material. Efforts to drive the
reaction to completion using more reagent, higher reaction
temperatures, or longer reaction times simply led to full
consumption of 48 and the generation of additional byproducts
without improvements in the conversions to 49. Desulfuriza-
tion of 49 with Raney nickel and reductive cleavage of the
oxido bridge with NaBH4 provided a separable mixture of the
diastereomers 50 and 51. Removal of the acetyl groups in 50
and 51 was conducted by treatment with NaOMe to produce
the corresponding secondary alcohols 52 and 53 in high yields.
Subjection of 52 to Mitsunobu conditions for elimination
(PPh3, DEAD)28 yielded the desired vindoline analogue 54,
whereas such treatment of 53 resulted in formation of the
isomeric 7,8-anhydro product 55 instead of the desired 6,7-
anhydro analogue. Alternative elimination reaction conditions
(CCl4, PPh3) used by Kuehne29 were also unsuccessful for the
introduction of the required 6,7-olefin.
In light of the additional unsatisfactory, albeit unoptimized,
elements of this route, we examined alternative stages and
methods for the introduction of the C6−C7 double bond.
Instead of performing a late-stage elimination, we explored
487
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
Scheme 5
Scheme 6
what proved to be a much more effective route where the olefin
was inserted prior to removal of the lactam carbonyl oxygen
and opening of the oxido bridge. Treatment of optically active
37 with (PhSe)2 and LDA yielded the α-selenide product,
which upon oxidation with H2O2 gave the α,β-unsaturated
lactam 56 in exceptionally high yield (99%, Scheme 6).
Thionation of 56, carried out with Lawesson’s reagent,
provided 57 in excellent yield (80%). Reaction of 57 with
Me3OBF4 followed by NaBH4 reduction of the resulting S-
methyl iminium ion under neutral conditions, afforded a
diastereomeric mixture of the oxido bridge cleaved products 54
and 58 in good yield, again favoring formation of the more
interesting unnatural trans fused isomer 58 (3:1) upon
reduction of the intermediate iminium ion derived from
oxido bridge cleavage. Without optimization, both 54 and 58
were coupled with catharanthine to provide the respective 4-
desacetoxyvinblastine analogues 59 and 62, along with the
corresponding anhydrovinblastine and leurosidine analogues.
Like the observations made with vinblastine itself, the
introduction of the C6−C7 double bond in 59 (59 vs 43)
provided a >10-fold increase in activity (Figure 5), although it
remained less active than 4-desacetoxyvinblastine (46). In
contrast, introduction of the olefin into the already potent
unnatural trans fused series with 62 (62 vs 44) resulted in no
further gain in activity, and the compound remained equipotent
to the comparison natural product 4-desacetoxyvinblastine
(46). In essence, the combined studies comparing 43, 44, 59,
and 62 show that introduction of the saturated trans 6,6 DE
ring system on 44 is functionally equivalent to the natural
unsaturated cis 6,5 DE ring system found in vinblastine.
Figure 5. Activity of the unsaturated 6,6 DE ring system analogues.
488
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
Synthesis of Vindoline Analogues Containing the
Reversed 5,6 DE Ring System and Incorporation into
the Corresponding Vinblastine Analogues. Among the
most interesting and certainly the most surprising of the deep-
seated structural modifications were those that contained the
vindoline reversed 5,6 DE ring system. Here, not only is the
vindoline C6−C7 double bond removed with the D ring
contraction to a 5-membered ring, but the E ring is expanded
from a five-membered to a six-membered ring. To access the
targeted vindoline analogue 80 with this reversed 5,6 DE ring
system, a shorter two-carbon tether to the initiating dienophile
is required and was prepared starting from the commercially
available β-alanine benzyl ester 65 (Scheme 7). Teoc protection
Scheme 8
Scheme 7
and (Z)-34 and the more successful cascade cycloaddition of
the substrate containing a tether substituted dienophile
employed in the vindoline asymmetric total synthesis, which
also incorporated a shortened dipolarophile tether.13 While
these differences may be attributed in part to reduced entropic
costs encountered when initiating the reaction with the shorter
and/or substituted dienophile tether, it is also clear that the
further conformational restriction in the dipolarophile tether in
74 (74 vs 34) contributes to this success. Consistent with
expectations, (E)-74 underwent a more facile reaction (TIPB,
180 °C, 24 h, 41%, or xylene, 150 °C, 24 h, 43%), requiring
lower reaction temperatures and accommodating higher
reaction concentrations (2 mM) for the cycloaddition cascade
to provide 76.
Following resolution of the enantiomers of 75 on a Chiralcel
OD column (30% i-PrOH−hexanes, 2 cm × 25 cm, α = 1.4),
benzyl group removal (Pd(OH)2, 45 psi H2, MeOH, 16 h)
provided the secondary alcohol 77 (Scheme 9). Acetylation of
77 to give 78, followed by treatment with Lawesson’s reagent,
provided thiolactam 79, the absolute stereochemistry of which
was established by X-ray crystallography.32 Reductive desulfur-
ization and a final reductive oxido bridge cleavage cleanly
afforded the desired vindoline analogue 80. Without
optimization, single-step Fe(III)-promoted coupling of 80
with catharanthine (3) and in situ oxidation produced the
vinblastine analogue 81 along with the epimeric C20′
leurosidine analogue 82 and the anhydrovinblastine analogue
83.
Remarkably, this vinblastine analogue 81 proved to be
equipotent with vinblastine (Figure 6). Not only does 81 lack
the critical C6−C7 double bond of vinblastine, but it
incorporates both a ring contracted five-membered D ring
and a ring-expanded six-membered E ring. Individually, these
changes result in 100-fold (21 vs 4), 10- to 100-fold (19 vs 4),
and nearly 10-fold (59 vs 46) losses in activity but reinstate full
activity when combined into a single analogue containing the
deep-seated core modifications. Most notably, the C6−C7
double bond is no longer required within 81.
Because of the greater facility of the cycloaddition with (E)-
74, a route to 80 proceeding through 76 and requiring
inversion of the C4 stereocenter was also examined in studies
that also provided the opportunity to examine the C4
diastereomer of vinblastine and its related analogues. The
enantiomers of 76 were separated on a Chiralcel OD column
(30% i-PrOH−hexanes, 2 cm × 25 cm, α = 1.4). The benzyl
group of 76 was removed by hydrogenation (Pd(OH)2, 1 atm
H2, MeOH, 16 h) to afford 84, which proceeded under milder
conditions than required of its C4 isomer 75 (Scheme 10). The
of 65, hydrogenolysis of the benzyl ether, and formation of the
Weinreb amide30 67 set the stage for reaction with EtMgBr (3
equiv, 3 equiv of CeCl3, THF, 0 °C, 94%) to afford the ethyl
ketone 68. Wittig olefination with Ph3PCHOBn31 provided a
separable 1:1.2 mixture of the (Z)- and (E)-enol ethers 69,
which were independently converted to the cycloaddition
precursors 74. Teoc deprotection (Bu4NF) of (Z)- and (E)-69
and treatment of the liberated amines with carbonyldiimidazole
afforded (Z)- and (E)-70. Addition of methyl oxalylhydrazide
afforded 71, which underwent cyclization to the corresponding
oxadiazoles 72 in the presence of TsCl and Et3N. Coupling of
(Z)-72 and (E)-72 with 3-(N-methyl-6-methoxyindol-3-yl)-
propanoic acid (73), obtained from methyl ester 26 (2N
NaOH, MeOH, 90%), provided the cycloaddition substrates
(Z)-74 and (E)-74, respectively.
Cyclization of (Z)-74 proceeded in TIPB (230 °C, 24 h, 0.2
mM) to give the cycloadduct 75 possessing the natural C4
stereochemistry, albeit in modest yield and requiring dilution of
the reaction mixture to 0.2 mM to afford 75 in conversions up
to 24% (Scheme 8). Nevertheless, this reaction serves as an
interesting contrast to both the unsuccessful behavior of (E)-34
489
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
Scheme 9
Scheme 10
Scheme 11
Figure 6. Activity of the reversed 5,6 DE ring system analogue.
C4 alcohol in 84 was oxidized with Dess−Martin periodinane
to afford ketone 85. Selective reaction with Lawesson’s reagent
afforded the thiolactam 86, for which the absolute stereo-
chemistry was established by X-ray crystallography.33 Reductive
desulfurization with Raney nickel was followed by diaster-
eoselective reductive cleavage of the oxido bridge (PtO2, 45 psi
H2, 1:1 MeOH−EtOAc). This afforded oxido bridge cleavage
product as an inconsequential mixture of C4 ketone and
secondary alcohol 87, which was further treated with
LiAlH(Ot-Bu)3 (0 °C, THF) to obtain clean 87, with exclusive
hydrogen delivery to the ketone from the α-face. Without
optimization, acetylation of 87 also afforded the target
vindoline analogue 80.
isomer of the key vindoline analogue 80. Without optimization,
coupling of 90 with catharanthine (3) and in situ oxidation
afforded the corresponding vinblastine analogue 91.
For comparison purposes, the C4 epimer of vinblastine itself
was prepared and used to determine the effect of this single
point change in the natural product. This was accomplished by
treatment of (−)-vindoline with NaOMe in MeOH to afford 4-
desacetylvindoline (94, Scheme 12).10a Although SO3−pyridine
oxidation of 94 has been reported,8f oxidation with IBX (80 °C,
EtOAc) afforded the ketone 95 in higher yield in our hands.
Reduction of 89 with NaBH(OAc)3 in 1,2-dichloroethane
allowed for selective hydride delivery to the ketone from the β-
face to give 96, which was subsequently acetylated (48 h) to
give 4-epi-vindoline 97. Coupling with catharanthine (3) and in
situ oxidation proceeded to give 4-epi-vinblastine (98) and 4-
Additionally, the secondary alcohol 84 was acetylated to give
88 (Scheme 11). Treatment with Lawesson’s reagent afforded
thiolactam 89. Reductive removal of the thiolactam and a final
reductive oxido bridge cleavage cleanly afforded 90, the C4
490
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
remaining analogues represent, it is reasonable to expect that
they also accurately reflect qualitative and quantitative trends in
target binding effects, especially those of 81 and 44 that match
the activity of the corresponding natural products 4 and 46.6j
To confirm this assumption, tubulin binding affinities were
assessed using a well-established tubulin binding assay34
measuring the competitive displacement of 3H-vinblastine
from porcine tubulin. These studies, which are much more
sensitive and quantitatively much more accurate than those
using tubulin polymerization assays, established that 59 binds
tubulin with a lower affinity than vinblastine and confirmed that
81 binds tubulin with an affinity matching that of vinblastine
itself (Figure 8). These trends observed in the tubulin binding
Scheme 12
epi-leurosidine (99) in the typical 2:1 β:α diastereoselectivity,
along with 4-epi-anhydrovinblastine (100).
The C4 epimer of vinblastine (98) proved to be 10-fold less
potent than the natural product, being essentially equipotent
with 4-desacetoxyvinblastine, lacking the C4 acetoxy sub-
stituent (Figure 7). The analogue 91, containing both the
3
Figure 8. Competition between H-vinblastine and analogues for the
porcine tubulin binding site.
assays follow those observed in the functional cell-based assays,
further indicating that conclusions drawn from the latter assays
accurately reflect tubulin binding effects.
Examination of the X-ray structure of vinblastine bound to
tubulin20 does not reveal a stabilizing π-interaction of the C6−
C7 double bond with the protein, nor does it suggest that
removal of the double bond with Δ6,7-dihydrovinblastine (21)
introduces destabilizing steric interactions when modeled onto
the crystal structure (Figure 9). With the caveat that the crystal
structure is of a modest resolution (4.1 Å), there appears to be
substantial and adequate space in the largely hydrophobic
pocket that surrounds the DE ring system such that not only
would 21 not be expected to suffer destabilizing steric
interactions, but that the entire series of analogues may be
accommodated. The fact that compounds so close in structure
(4 vs 21) exhibit such disparate binding and functional activity
while those so spatially or conformationally dissimilar (81 vs 4
and 46 vs 44) are essentially equally active suggests that the
losses in binding affinity are not likely the result of sterically
destabilized binding. In fact, it is the analogues that are most
distinct from rather than similar to vinblastine that are most
active.
What does stand out is the potential impact that the
conformational or structural changes may have on the location
and orientation of the nitrogen (N9) lone pair as well as their
impact on the N9 basicity. However, here again the expected
impacts of the changes do not correlate with activity. Unlike
vinblastine (4), dihydrovinblastine (21), and the even less
active 43/59 series, where the N9 lone pair is oriented on the
β-face, that of 19 and the active 44/62 series is oriented to the
α-face. Similarly, even the spatial location of the N9 atom in the
Figure 7. Activity of C4 epimeric analogues.
vindoline reversed 5,6 ring system as well as the C4 epimeric
center, also displayed reduced cytotoxic activity, being
approximately 100-fold less potent than vinblastine and 81
and 10-fold less potent than epi-4-vinblastine (98).
Origin of the Impact of the Core Modifications. The
50−100-fold reduction in the cytotoxic activity of Δ6,7-
dihydrovinblastine (21) relative to vinblastine that we observe
in our cell-based assays correlates well with its 30−50-fold
reduction in affinity found in quantitative tubulin binding
assays.34 Thus, the diminished potency of 21 in our cell-based
assays is the direct result of its reduced target binding affinity
and not the result of other variables that conceivably could
affect the activity (e.g., cell permeability, metabolism). Given
the rather minor perturbations to the overall structure that the
491
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
When combined, the results of the evaluation of the modified
vinblastines indicate that the role of the C6−C7 double bond is
not derived from the π-unsaturation itself or its impact on the
N9 pKa or the nitrogen lone pair orientation. While many
tailoring single-point peripheral changes, such as installation of
the C4 acetoxy group, seem to be additive in their effects on the
biological properties vinblastine, this need not be the case for
structures containing more deep-seated changes to the core
structure of the natural product. Herein, such deep-seated
structural changes to the size and configuration of the vindoline
core ring system of vinblastine, accessible only by total
synthesis, provided exceptionally and even unexpectedly potent
vinblastine analogues.
CONCLUSIONS
■
Full details of the synthesis of a systematic series of vinblastine
analogues bearing deep-seated structural modifications to the
vindoline core ring system are disclosed. Not accessible from
natural product sources, analogues possessing 5,5, 6,6, and
reversed 5,6 membered DE ring systems were assembled using
a tandem intramolecular [4 + 2]/[3 + 2] cycloaddition cascade
reaction of 1,3,4-oxadiazoles as the key step. In the course of
converting the cycloadducts to the corresponding vindoline
analogues, an oxido bridge reduction of the cycloadducts
leading to 6,6 DE ring systems provided access to both the cis
and trans ring systems, resulting in the first study of vindoline
analogues with this structural feature and affording a
surprisingly effective new stereochemical class for exploration.
The vindoline analogues were found to be suitable substrates
for a biomimetic Fe(III)-promoted coupling and subsequent in
situ oxidation reaction, enabling access to vinblastine analogues
containing key ring system modifications.
Although single functional group removals from the
vinblastine lower subunit typically result in pronounced and
additive losses in activity,6j two unprecedented deep-seated
structural changes to the core structure were found that
maintain potent activity. Striking among these is 81, containing
the vindoline reversed 5,6 DE ring system, which proved
equipotent to vinblastine. Not only does 81 lack the critical
C6−C7 double bond of vinblastine whose removal results in a
100-fold loss in activity, but it also incorporates a ring
contracted five-membered D ring and a ring expanded six-
membered E ring, representing changes that individually result
in 10- to 100-fold reductions in activity. The behavior of the
isomeric analogues with the expanded 6,6 DE ring systems was
just as remarkable. Unlike the cis isomer 43 with the natural C9
stereochemistry that displayed diminished activity, the analogue
44 that not only lacks the C6−C7 double bond but also
contains the ring expanded six-membered E ring and the
unnatural (trans) stereochemistry was 100-fold more active,
matching the activity of natural 4-desacetoxyvinblastine (46)
and substantially exceeding that of the comparable 4-
desacetoxy-6,7-dihydrovinblastine (45). Cumulatively, the
studies demonstrate that the 100-fold impact of the vinblastine
C6−C7 double bond is not the result of a unique stabilizing π-
interaction with tubulin or its impact on the N9 pKa, its lone
pair orientation, or even the precise spatial location of N9. Just
as significantly, unlike modifications of peripheral substituents
or atom exchanges (tailoring effects), core structure redesign in
natural products is rarely explored, and the results detailed
herein with an important oncology drug indicate this may offer
far more opportunities than anticipated.35
Figure 9. X-ray of vinblastine20 ((a), PDB ID 1Z2B) and modeled
binding of dihydrovinblastine (21) (b), 44 (c), and 81 (d) with
tubulin viewed from the solvent accessible site into the binding pocket.
The catharanthine-derived velbanamine subunit is buried deep in the
binding pocket behind the vindoline subunit, which extends down the
pocket to the solvent interface. Asp β179, as found in the vinblastine-
tubulin X-ray structure, is highlighted (white, with the side chain
oxygen in red) on the surface representation and labeled in (a).
active 44/62 series is farthest from that found in vinblastine (4)
and, unlike all others, N9 in this series is not close enough to
participate in H-bonding to the C3 OH irrespective of the lone
pair orientation. Similarly, the projected basicity of N9 imposed
by the structural and conformational changes do not seem to
correlate with the observed changes in activity. Although 19,
like dihydrovinblastine (21), is expected to be more basic than
vinblastine itself, 81 and 21 would be expected to be very
similar yet display the most distinct activity. Similarly, 44 would
be expected to be more basic than 43, but it is the much more
active of the two isomers. Thus, the N9 basicity, the lone pair
orientation, and even the spatial location of the nitrogen do not
seems to correlate with activity.
Finally, the only charged, polar residues near or in the tubulin
binding pocket at the interface of the two tubulin heterodimers
are Lys α336, Lys α352, Lys β176, and Asp β179. Lys α336 and
Lys β176 are located too far from the N9 site to expect either to
play a role and are shielded from participating by intervening
atoms. While the full side chain of Lys α352 is not visible in the
X-ray electron density, it appears to be oriented away from
vinblastine. By contrast, Asp β179 is located ca. 4.3 Å away
from N9 in vinblastine in the cocrystal structure, moving from a
position close to a nucleotide binding site in the absence of
vinblastine to the vicinity of the drug when bound,20 and may
contribute a stabilizing electrostatic interaction (Figure 9). It is
conceivable that interactions with this mobile side chain residue
influence the relative binding of the analogues in this series. If
so, it is doing so in a way that is not dependent on the spatial
site or orientation of N9 and its lone pair, likely with the
adoption by Asp β179 itself of varied locations that are not
easily anticipated or modeled. These questions and whether a
basic N9 is even required are presently being examined.
492
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
A. E.; Watson, C. D.; Chapple, C. C. S.; Vukovic, J.; Misawa, M.
Extraction of 3′,4′-anhydrovinblastine from Catharanthus roseus.
Phytochemistry 1988, 27, 1713−1717.
ASSOCIATED CONTENT
* Supporting Information
Full experimental details and compound characterizations are
provided. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
́
(4) Reviews: (a) Blasko, G.; Cordell, G. A. Isolation, structure
elucidation, and biosynthesis of the bisindole alkaloids of Cathar-
anthus. In The Alkaloids; Brossi, A., Suffness, M., Eds.; Academic: San
́
Diego, 1990; Vol. 37, pp 1−76. (b) Sottomayor, M.; Barcelo, A. R.
AUTHOR INFORMATION
Corresponding Author
*Phone: (858) 784-7522. Fax: (858) 784-7550. E-mail:
boger@scripps.edu.
Notes
The Vinca alkaloids: From biosynthesis and accumulation in plant
cells, to uptake, activity and metabolism in animal cells. In Studies in
Natural Products Chemistry; Atta-ur-Rahman, Ed.; Elsevier: Amster-
dam, 2006; Vol. 33, pp 813−857.
■
(5) Reviews: (a) Pearce, H. L. Medicinal chemistry of bisindole
alkaloids from Catharanthus. In The Alkaloids; Brossi, A., Suffness, M.,
Eds.; Academic: San Diego, 1990; Vol. 37, pp 145−204. (b) Neuss, N.;
Neuss, M. N. Therapeutic use of bisindole alkaloids from
Catharanthus. In The Alkaloids; Brossi, A., Suffness, M., Eds.;
Academic: San Diego, 1990; Vol. 37, pp 229−240.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully acknowledge the financial support of the National
Institutes of Health (CA115526, CA042056) and the Skaggs
Institute of Chemical Biology. We wish to thank Dr. Raj K.
Chadha for the X-ray structures disclosed herein. We also thank
Katerina Otrubova for assistance with the tubulin binding
assays. D.K. and K.K.D. were Skaggs Fellows.
(6) (a) Mangeney, P.; Andriamialisoa, R. Z.; Langlois, N.; Langlois,
Y.; Potier, P. Preparation of vinblastine, vincristine, and leurosidine,
antitumor alkaloids from Catharanthus species (Apocynaceae). J. Am.
Chem. Soc. 1979, 101, 2243−2245. (b) Kutney, J. P.; Choi, L. S. L.;
Nakano, J.; Tsukamoto, H.; McHugh, M.; Boulet, C. A. A highly
efficient and commercially important synthesis of the antitumor
Catharanthus alkaloids vinblastine and leurosidine from catharanthine
and vindoline. Heterocycles 1988, 27, 1845−1853. (c) Kuehne, M. E.;
Matson, P. A.; Bornmann, W. G. Enantioselective syntheses of
vinblastine, leurosidine, vincovaline and 20′-epi-vincovaline. J. Org.
Chem. 1991, 56, 513−528. (d) Bornmann, W. G.; Kuehne, M. E. A
common intermediate providing syntheses of ψ-tabersonine, coro-
naridine, iboxyphylline, ibophyllidine, vinamidine, and vinblastine. J.
Org. Chem. 1992, 57, 1752−1760. (e) Kuehne, M. E.; Zebovitz, T. C.;
Bornmann, W. G.; Marko, I. Three routes to the critical C16′−C14′
parf relative stereochemistry of vinblastine. Syntheses of 20′-desethyl-
20′-deoxyvinblastine and 20′-desethyl-20′-deoxyvincovaline. J. Org.
Chem. 1987, 52, 4340−4349. (f) Magnus, P.; Mendoza, J. S.; Stamford,
A.; Ladlow, M.; Willis, P. Nonoxidative coupling methodology for the
synthesis of the antitumor bisindole alkaloid vinblastine and a lower-
half analog: solvent effect on the stereochemistry of the crucial C-15/
C-18′ bond. J. Am. Chem. Soc. 1992, 114, 10232−10245.
(g) Yokoshima, S.; Ueda, T.; Kobayashi, S.; Sato, A.; Kuboyama, T.;
Tokuyama, H.; Fukuyama, T. Stereocontrolled total synthesis of
(+)-vinblastine. J. Am. Chem. Soc. 2002, 124, 2137−2139.
(h) Kuboyama, T.; Yokoshima, S.; Tokuyama, H.; Fukuyama, T.
Stereocontrolled total synthesis of (+)-vincristine. Proc. Natl. Acad. Sci.
U. S. A. 2004, 101, 11966−11970. (i) Ishikawa, H.; Colby, D. A.;
Boger, D. L. Direct coupling of catharanthine and vindoline to provide
vinblastine: total synthesis of (+)- and ent-(−)-vinblastine. J. Am.
Chem. Soc. 2008, 130, 420−421. (j) Ishikawa, H.; Colby, D. A.; Seto,
S.; Va, P.; Tam, A.; Kakei, H.; Rayl, T. J.; Hwang, I.; Boger, D. L. Total
synthesis of vinblastine, vincristine, related natural products, and key
structural analogues. J. Am. Chem. Soc. 2009, 131, 4904−4916.
(7) Jordan, M. A.; Wilson, L. Microtubules as a target for anticancer
drugs. Nature Rev. Cancer 2004, 4, 253−265.
ABBREVIATIONS USED
■
CDI, 1,1′-carbonyldiimidazole; DEAD, diethyl azodicarbox-
ylate; DMAP, 4-dimethylaminopyridine; EDCI, 1-ethyl-3-(3-
dimethylaminopropyl)carbodiimide; HMDS, hexamethyldisila-
zide; IBX, 2-iodoxybenzoic acid; LDA, lithium diisopropyla-
mide; pyr, pyridine; THF, tetrahydrofuran; TsCl, p-toluene-
sulfonyl chloride; Et3N, triethylamine; TIPB, triisopropylben-
zene; Teoc, 2-trimethylsilylethyl carbonyl; PPh3,
triphenylphosphine
REFERENCES
■
(1) Gorman, M.; Neuss, N.; Biemann, K. Vinca alkaloids. X. The
structure of vindoline. J. Am. Chem. Soc. 1962, 84, 1058−1059.
(2) Moncrief, J. W.; Lipscomb, W. N. Structures of leurocristine
(vincristine) and vincaleukoblastine. X-ray analysis of leurocristine
methiodide. J. Am. Chem. Soc. 1965, 87, 4963−4964.
(3) (a) Noble, R. L.; Beer, C. T.; Cutts, J. H. Role of chance
observations in chemotherapy: Vinca rosea. Ann. N. Y. Acad. Sci. 1958,
76, 882−894. (b) Noble, R. L. Catharanthus roseus (Vinca rosea): The
importance and value of a chance observation. Lloydia 1964, 27, 280−
281. (c) Svoboda, G. H.; Neuss, N.; Gorman, M. Alkaloids of Vinca
rosea Linn. (Catharanthus roseus G. Don). V. Preparation and
characterization of alkaloids. J. Am. Pharm. Assoc., Sci. Ed. 1959, 48,
659−666. For related, naturally occurring Vinca alkaloids, see the
following. Leurosidine: (d) Svoboda, G. H. Alkaloids of Vinca rosea
(Catharanthus roseus). 1X: Extraction and characterization of
leurosidine and leurocristine. Lloydia 1961, 24, 173−178. Deoxyvin-
blastine: (e) De Bruyn, A.; Sleechecker, J.; De Jonghe, J. P.; Hannart, J.
Alkaloids from Catharanthus roseus. Proton NMR study of 20′S-
deoxyvinblastine and its N-2′ borane complex. Bull. Soc. Chim. Belg.
1983, 92, 485−494. (f) Neuss, N.; Gorman, M.; Cone, N. J.;
Huckstep, L. L. Vinca alkaloids. XXX. Chemistry of the deoxyvinblas-
tines (deoxy-vlb), leurosine (vlr), and pleurosine, dimeric alkaloids
from Vinca rosea Linn. (Catharanthus roseus G. Don). Tetrahedron Lett.
1968, 9, 783−787. N-Desmethylvinblastine: (g) Simonds, R.; De
Bruyn, A.; De Taeye, L.; Verzele, M.; De Pauw, C. N-Deformyl
vincristine, a bisalkaloid from Catharanthus roseus. Planta Med. 1984,
50, 274−276. Desacetoxyvinblastine: (h) Neuss, N.; Barnes, A. J.;
Huckstep, L. L. Vinca Alkaloids. XXXV. Desacetoxyvinblastine, a new
minor alkaloid from Vinca rosea (Catharanthus roseus). Experientia
1975, 31, 18−20. Desacetylvinblastine: (i) Svoboda, G. H.; Barnes, A.
J. Alkaloids of Vinca rosea Linn. (Catharanthus roseus G. Don) XXIV.
Vinaspine, vincathicine, rovidine, desacetyl VLB, and vinaphamine. J.
Pharm. Sci. 1964, 53, 1227−1231. Anhydrovinblastine: (j) Goodbody,
(8) Racemic syntheses: (a) Ando, M.; Buchi, G.; Ohnuma, T. Total
̈
synthesis of ( )-vindoline. J. Am. Chem. Soc. 1975, 97, 6880−6881.
(b) Kutney, J. P.; Bunzli-Trepp, U.; Chan, K. K.; De Souza, J. P.;
Fujise, Y.; Honda, T.; Katsube, J.; Klein, F. K.; Leutweiler, A.;
Morehead, S.; Rohr, M.; Worth, B. R. Total synthesis of indole and
dihydroindole alkaloids. 14. A total synthesis of vindoline. J. Am. Chem.
Soc. 1978, 100, 4220−4224. (c) Ban, Y.; Sekine, Y.; Oishi, T. A new
synthetic route to ( )-vindoline. A synthesis of the Buchi’s tetracyclic
̈
intermediate. Tetrahedron Lett. 1978, 19, 151−154. (d) Takano, S.;
Shishido, K.; Sato, M.; Yuta, K.; Ogasawara, K. New synthetic routes to
synthons for the synthesis of functionalized aspidosperma alkaloids. J.
Chem. Soc., Chem. Commun. 1978, 943−944. Takano, S.; Shishido, K.;
Matsuzaka, J.; Sato, M.; Ogasawara, K. A new synthesis of the synthons
for the functionalized aspidosperma alkaloids via α-ketocarbonium ion
intermediate. Heterocycles 1979, 13, 307−320. (e) Danieli, B.; Lesma,
G.; Palmisano, G.; Riva, R. Rapid access to the highly oxygenated
493
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
aspidosperma alkaloids vindoline, vindorosine, and cathovaline. J.
Chem. Soc., Chem. Commun. 1984, 909−911. (f) Andriamialisoa, R. Z.;
Langlois, N.; Langlois, Y. A new efficient total synthesis of vindorosine
and vindoline. J. Org. Chem. 1985, 50, 961−967. (g) Danieli, B.;
Lesma, G.; Palmisano, G.; Riva, R. Aspidosperma alkaloids.
Conversion of tabersonine into vindoline. J. Chem. Soc., Perkin Trans.
1 1987, 155−161. (h) Zhuo, S.; Bommezijn, S.; Murphy, J. A. Formal
total synthesis of ( )-vindoline by tandem radical cyclization. Org.
Lett. 2002, 4, 443−445.
(9) Enantioselective syntheses: (a) Feldman, P. L.; Rapoport, H.
Synthesis of (−)-vindoline. J. Am. Chem. Soc. 1987, 109, 1603−1604.
(b) Kuehne, M. E.; Podhorez, D. E.; Mulamba, T.; Bornmann, W. G.
Biomimetic alkaloid syntheses. 15. Enantioselective syntheses with
epichlorohydrin: total syntheses of (+)-, (−)-, and ( )-vindoline and a
synthesis of (−)-vindorosine. J. Org. Chem. 1987, 52, 347−353.
(c) Cardwell, K.; Hewitt, B.; Ladlow, M.; Magnus, P. Methods for
indole alkaloid synthesis. A study of the compatibility of the indole-
2,3-quinodimethane strategy for the synthesis of 16-methoxy-
substituted aspidosperma-type alkaloids. Synthesis of (+)- and
(−)-16-methoxytabersonine. J. Am. Chem. Soc. 1988, 110, 2242−
2248. (d) Kobayashi, S.; Ueda, T.; Fukuyama, T. An efficient total
synthesis of (−)-vindoline. Synlett 2000, 883−886.
(10) (a) Choi, Y.; Ishikawa, H.; Velcicky, J.; Elliott, G. I.; Miller, M.
M.; Boger, D. L. Total synthesis of (−)- and ent-(+)-vindoline. Org.
Lett. 2005, 7, 4539−4542. (b) Ishikawa, H.; Elliott, G. I.; Velcicky, J.;
Choi, Y.; Boger, D. L. Total synthesis of (−)- and ent-(+)-vindoline
and related alkaloids. J. Am. Chem. Soc. 2006, 128, 10596−10612.
(11) (a) Wilkie, G. D.; Elliott, G. I.; Blagg, B. S. J.; Wolkenberg, S. E.;
Soenen, D. R.; Miller, M. M.; Pollack, S.; Boger, D. L. Intramolecular
Diels−Alder and tandem intramolecular Diels−Alder/1,3-dipolar
cycloaddition reactions of 1,3,4-oxadiazoles. J. Am. Chem. Soc. 2002,
124, 11292−11294. (b) Elliott, G. I.; Fuchs, J. R.; Blagg, B. S. J.;
Ishikawa, H.; Tao, H.; Yuan, Z.-Q.; Boger, D. L. Intramolecular Diels−
Alder/1,3-dipolar cycloaddition cascade of 1,3,4-oxadiazoles. J. Am.
Chem. Soc. 2006, 128, 10589−10595. For reviews of heterocyclic
azadiene cycloaddition reactions, see: (c) Boger, D. L. Diels−Alder
reactions of azadienes. Tetrahedron 1983, 39, 2869−2939. (d) Boger,
D. L. Diels−Alder reactions of heterocyclic azadienes. Scope and
applications. Chem. Rev. 1986, 86, 781−793. (e) Boger, D. L.;
Weinreb, S. M. Hetero Diels−Alder Methodology in Organic Synthesis;
Academic: San Diego, 1987.
and an expanded scope of the Fe(III)-mediated vinblastine coupling
reaction. J. Am. Chem. Soc. 2012, 134, 13240−13243.
(15) Reviews: (a) Kuehne, M. E.; Marko, I. Syntheses of vinblastine-
type alkaloids. In The Alkaloids; Brossi, A., Suffness, M., Eds.;
Academic: San Diego, 1990; Vol. 37, pp 77−132. (b) Potier, P.
Synthesis of the antitumor dimeric indole alkaloids from Catharanthus
species (vinblastine group). J. Nat. Prod. 1980, 43, 72−86. (c) Kutney,
J. P. Biosynthesis and synthesis of indole and bisindole alkaloids in
plant cell cultures: a personal overview. Nat. Prod. Rep. 1990, 7, 85−
103. (d) Kutney, J. P. Plant cell cultures and synthetic chemistrya
potentially powerful route to complex natural products. Synlett 1991,
11−19. (e) Kutney, J. P. Plant cell culture combined with chemistry: a
powerful route to complex natural products. Acc. Chem. Res. 1993, 26,
559−566. For recent studies, see: (f) Kuehne, M. E.; Bornmann, W.
G.; Marko, I.; Qin, Y.; Le Boulluec, K. L.; Frasier, D. A.; Xu, F.;
Malamba, T.; Ensinger, C. L.; Borman, L. S.; Huot, A. E.; Exon, C.;
Bizzarro, F. T.; Cheung, J. B.; Bane, S. L. Syntheses and biological
evaluation of vinblastine congeners. Org. Biomol. Chem. 2003, 1,
2120−2136. (g) Miyazaki, T.; Yokoshima, S.; Simizu, S.; Osada, H.;
Tokuyama, H.; Fukuyama, T. Synthesis of (+)-vinblastine and its
analogues. Org. Lett. 2007, 9, 4737−4740.
(16) Tam, A.; Gotoh, H.; Robertson, W. M.; Boger, D. L.
Catharanthine C16 substituent effects on the biomimetic coupling
with vindoline: preparation and evaluation of a key series of vinblastine
analogues. Bioorg. Med. Chem. Lett. 2010, 20, 6408−6410.
(17) Gotoh, H.; Duncan, K. K.; Robertson, W. M.; Boger, D. L. 10′-
Fluorovinblastine and 10′-fluorovincristine: synthesis of a key series of
modified Vinca alkaloids. ACS Med. Chem. Lett. 2011, 2, 948−952.
(18) Leggans, E. K.; Barker, T. J.; Duncan, K. K.; Boger, D. L.
Iron(III)/NaBH4-mediated additions to unactivated alkenes: Synthesis
of novel C20′-vinblastine analogues. Org. Lett. 2012, 14, 1428−1431.
For additional applications, see: (b) Barker, T. J.; Boger, D. L. Fe(III)/
NaBH4-mediated free radical hydrofluorination of unactivated alkenes.
J. Am. Chem. Soc. 2012, 134, 13588−13591. (c) Leggans, E. K.;
Duncan, K. K.;. Barker, T. J ; Schleicher, K. D.; Boger. D. L. A
remarkable series of vinblastine analogues displaying enhanced activity
and an unprecedented tubulin binding steric tolerance: C20′ urea
derivatives. J. Med. Chem. 2013, 56, DOI: 10.1021/jm3015684.
(19) Va, P.; Campbell, E. L.; Robertson, W. M.; Boger, D. L. Total
synthesis and evaluation of a key series of C5-substituted vinblastine
derivatives. J. Am. Chem. Soc. 2010, 132, 8489−8495.
(12) (a) Elliott, G. I.; Velcicky, J.; Ishikawa, H.; Li, Y.; Boger, D. L.
Total synthesis of (−)- and ent-(+)-vindorosine: tandem intra-
molecular Diels−Alder/1,3-dipolar cycloaddition of 1,3,4-oxadiazoles.
Angew. Chem., Int. Ed. 2006, 45, 620−622. (b) Yuan, Z. Q.; Ishikawa,
H.; Boger, D. L. Total synthesis of natural (−)- and ent-(+)-4-
desacetoxy-6,7-dihydrovinblastine and natural and ent-minovine:
oxadiazole tandem intramolecular Diels−Alder/1,3-dipolar cyclo-
addition reaction. Org. Lett. 2005, 7, 741−744. (c) Ishikawa, H.;
Boger, D. L. Total synthesis of (−)- and ent-(+)-4-desacetoxy-5-
desethylvindoline. Heterocycles 2007, 72, 95−102. (d) Campbell, E. L.;
Zuhl, A. M.; Liu, C. M.; Boger, D. L. Total synthesis of
(+)-fendleridine (aspidoalbidine) and (+)-1-acetylaspidoalbidine. J.
Am. Chem. Soc. 2010, 132, 3009−3012. (e) Lajiness, J. P.; Jiang, W.;
Boger, D. L. Divergent total syntheses of (−)-aspidospermine and
(+)-spegazzinine. Org. Lett. 2012, 14, 2078−2081. (f) Wolkenberg, S.
E.; Boger, D. L. Total synthesis of anhydrolycorinone using sequential
intramolecular Diels−Alder reactions of a 1,3,4-oxadiazole. J. Org.
Chem. 2002, 67, 7361−7364.
(13) (a) Kato, D.; Sasaki, Y.; Boger, D. L. Asymmetric total synthesis
of vindoline. J. Am. Chem. Soc. 2010, 132, 3685−3687. (b) Sasaki, Y.;
Kato, D.; Boger, D. L. Asymmetric total synthesis of vindorosine,
vindoline, and key vinblastine analogues. J. Am. Chem. Soc. 2010, 132,
13533−13544.
(14) (a) Vukovic, J.; Goodbody, A. E.; Kutney, J. P.; Misawa, M.
Production of 3′,4′-anhydrovinblastine: a unique chemical synthesis.
Tetrahedron 1988, 44, 325−331. (b) Gotoh, H.; Sears, J. E.;
Eschenmoser, A.; Boger, D. L. New insights into the mechanism
(20) Gigant, B.; Wang, C.; Ravelli, R. B. G.; Roussi, F.; Steinmetz, M.
O.; Curmi, P. A.; Sobel, A.; Knossow, M. Structural basis for the
regulation of tubulin by vinblastine. Nature 2005, 435, 519−522.
(21) Frequently, synthetic exploration was carried out initially with
the unnatural enantiomer series, but we have chosen only to depict the
natural enantiomer series for the sake of clarity. The structures of 13
(CCDC 827890) and 14 (CCDC 827891), as well as 11 (CCDC
827889), were unambiguously established with single-crystal X-ray
structures conducted on samples in the unnatural enantiomer series.
(22) Analogues were tested using L1210 murine leukemia, HCT116
colon cancer, and an established companion vinblastine resistant cell
line (HCT116/VM46). See: (a) Dumontet, C.; Sikic, B. I.
Mechanisms of action of and resistance to antitubulin agents:
microtubule dynamics, drug transport, and cell death. J. Clin. Oncol.
1999, 17, 1061−1070. (b) Kavallaris, M.; Verrills, N. M.; Hill, B. T.
Anticancer therapy with novel tubulin-interacting drugs. Drug Res.
Updates 2001, 4, 392−401.
(23) Yde, B.; Yousif, N. M.; Pedersen, U.; Thomsen, I.; Lawesson, S.
O. Studies on organophosphorus compounds XLVII. Preparation of
thiated synthons of amides, lactams, and imides by use of some new
P,S-containing reagents. Tetrahedron 1984, 40, 2047−2052.
(24) The structure and absolute stereochemistry of 40 were
established by X-ray (CCDC 863666) conducted on a sample in the
unnatural enantiomer series.
(25) Raucher, S.; Klein, P. A convenient method for the selective
reduction of amides to amines. Tetrahedron Lett. 1980, 21, 4061−
4064.
494
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495

Journal of Medicinal Chemistry
Article
(26) Barnett, C. J.; Cullinan, G. J.; Gerzon, K.; Hoying, R. C.; Jones,
W. E.; Nevlon, W. M.; Poore, G. A.; Robinson, R. L.; Sweeney, M. J.;
Todd, G. C.; Dyke, R. W.; Nelson, R. L. Structure−activity
relationships of dimeric Catharanthus alkaloids. 1. Desacetylvinblastine
amide (vindesine) sulfate. J. Med. Chem. 1978, 21, 88−96.
(27) The structure of racemic 48 was established by X-ray (CCDC
863669).
(28) (a) Mitsunobu, O. The use of diethyl azodicarboxylate and
triphenylphosphine in synthesis and transformation of natural
products. Synthesis 1981, 1. (b) Iimori, T.; Ohtsuka, Y.; Oishi, T.
Regioselective dehydration in cyclic system with triphenylphosphine-
azodicarboxylate. Tetrahedron Lett. 1991, 32, 1209−1212.
(29) Kuehne, M. E.; Okuniewicz, F. J.; Kirkemo, C. L.; Bohnert, J. C.
Studies in biomimetic alkaloid syntheses. 8. Total syntheses of the C-
14 epimeric hydroxyvincadifformines, tabersonine, a (hydroxymethyl)-
D-norvincadifformine and the C-20 epimeric pandolines. J. Org. Chem.
1982, 47, 1335−1343.
(30) Nahm, S.; Weinreb, S. M. N-methoxy-N-methylamides as
effective acylating agents. Tetrahedron Lett. 1981, 22, 3815−3818.
(31) Tam, T. F.; Fraser-Reid, B. 1,2-O-Isopropylidene furanoses as
chiral precursors for α-methylene-γ-butyrolactones. J. Chem. Soc.,
Chem. Commun. 1980, 556−558.
(32) The structure and absolute stereochemistry of 79 were
established by X-ray (CCDC 863668) conducted on material in the
unnatural enantiomer series.
(33) The structure and absolute stereochemistry of 86 were
established by X-ray (CCDC 863667) conducted on material in the
unnatural enantiomer series.
(34) Owellen, R. J.; Donigian, D. W.; Hartke, C. A.; Hains, F. O.
Correlation of biologic data with physicochemical properties among
the Vinca alkaloids and their congeners. Biochem. Pharmacol. 1977, 26,
1213−1219.
(35) All compounds tested for biological activity were >95% pure.
The purity was established by HPLC using two different reverse phase
gradient solvent systems.
495
dx.doi.org/10.1021/jm3014376 | J. Med. Chem. 2013, 56, 483−495