A new stereocontrolled total synthesis of the mast cell inhibitory alkaloid, (+)-monanchorin, via the wittig reaction of a stabilized ylide with a cyclic guanidine hemiaminal

Article abstract of DOI:10.1021/ol500616v

An asymmetric total synthesis of the mast cell inhibitor (+)-monanchorin is reported in which a Sharpless AD on 11 and a cyclic sulfate ring opening with an azide feature as key steps. After further manipulation, a novel guanidine-controlled ester reduction provided the guanidine-hemiaminal 25 which underwent Wittig olefination to give 27. Hydrogenation and a second guanidine-controlled reduction of the ester in 28, to obtain aldehyde 29, then set up a trifluoroacetic acid mediated cyclization to give (+)-monanchorin TFA salt.

Full text of DOI:10.1021/ol500616v

Letter
pubs.acs.org/OrgLett
A New Stereocontrolled Total Synthesis of the Mast Cell Inhibitory
Alkaloid, (+)-Monanchorin, via the Wittig Reaction of a Stabilized
Ylide with a Cyclic Guanidine Hemiaminal
Karl J. Hale* and Liping Wang
The School of Chemistry and Chemical Engineering and the Centre for Cancer Research and Cell Biology (CCRCB), Queen’s
University Belfast (QUB), Stranmillis Road, Belfast BT95 AJ, Northern Ireland, United Kingdom
S
* Supporting Information
ABSTRACT: An asymmetric total synthesis of the mast cell inhibitor (+)-monanchorin is reported in which a Sharpless AD on
11 and a cyclic sulfate ring opening with an azide feature as key steps. After further manipulation, a novel guanidine-controlled
ester reduction provided the guanidine-hemiaminal 25 which underwent Wittig olefination to give 27. Hydrogenation and a
second guanidine-controlled reduction of the ester in 28, to obtain aldehyde 29, then set up a trifluoroacetic acid mediated
cyclization to give (+)-monanchorin TFA salt.
It was Westphal3 who first recognized that mast cells
ast cells are leukocytes of bone marrow hematopoietic
congregate in a large number at the juncture of developing solid
Mstem cell origin that contribute significantly to the
tumors and normal healthy tissue, way back in 1891, and
although mature mast cells do indeed populate many normal
tissues, they generally do so only at a fairly low level, except for
the lung and airways, the digestive tract, around neurons, in
skin, or in wounds, where their presence in high abundance
helps combat unwanted pathogens and parasites.
It appears that tumors and tumor-associated mast cells exert
very profound effects on T-regulatory cell function that can
result in host immunosuppression and further promote
malignancy.1,4 Indeed, the interactions between mast cells
and regulatory T-cells appear to determine whether mast cells
combat pathogens exclusively (which is their normal function)
or whether they promote excessive tumor inflammation, tumor
immunity, and aggressive tumor growth.
Of possible mechanistic relevance to the latter finding is the
observation that tumor cells5 and mast cells6 both have the
ability to transfer mRNA and microRNA, peptides, lipids, and
nucleosides into other nearby cells via small nanosized vesicles
known as exosomes.5,6 The latter are released by a variety of
cell types, to allow communication and signaling to occur
between cells. In cases where mRNA is actually transferred and
eventually transfected, exosomes do this by fusing to and
eventually merging with target cells. Recipient cells are thus
capable of being genetically reprogrammed by exosomes!6
It is now fairly well established that tumor-secreted exosomes
can upregulate the expression of cytokines such as phospho-
establishment, growth, and metastatic spread of human
tumors.1 In this regard, the cytoplasm of mature mast cells is
packed full of large secretory granules that can release a variety
of powerful pro-angiogenic mediators when activated. These
include fibroblast growth factor (FGF), vascular endothelial
growth factor (VEGF), heparin, and angiopoietin-1 (Ang-1).
Mast cells are also known to secrete tumor-promoting
cytokines such as tumor necrosis factor-α (TNF-α), various
interleukins (ILs), and granulocyte-colony stimulating factor
(GM-CSF). They further release a range of tissue degrading
and tissue remodeling proteases that include tryptase (mMCP-
6), chymase (mMCP-4), the metalloexopeptidase carboxypep-
tidase A3 (CPA3), collagenases, cathepsins, and various matrix
metalloproteases that include MMP9.
It is now well established that the chemokines CCL2 and
CCL5, tumor-secreted stem cell factor (SCF, c-Kit ligand), and
SCF expressed on the outer cell surface of human tumors can
all lead to the chemotactic migration and local recruitment of
immature mast cells to tumors,1 and once there within the
tumor microenvironment, immature mast cells can eventually
differentiate, mature, and remain in a constant state of
activation through the presence of host activators such as the
C5a complement protein.2 Activated mast cells can thus
provide a rich and fertile secretory source of the above pro-
angiogenic and tissue remodelling factors which, when in the
vicinity of tumors, ultimately promote tumor vascularization,
tumor growth, and eventual patient demise by accelerating
metastatic spread.
Received: February 26, 2014
Published: April 7, 2014
© 2014 American Chemical Society
2154
dx.doi.org/10.1021/ol500616v | Org. Lett. 2014, 16, 2154−2157

Organic Letters
Letter
STAT3, phospho-SMAD2/3, IL-10,4 and TGF-β14 in regu-
latory T-cells (nT-reg/iT-reg), to help diminish the immune
response of cancer sufferers. Tumor-derived exosomes can
likewise upregulate granzyme B, perforin, and membrane-
associated death ligands such as FasL and TRAIL to activate
apoptosis in CD8+-T cells.4,7 It is possible that this
immunosuppressive reprogramming of T-cells by tumor-
derived exosomes7 involves the direct transfer of genetic
information via mRNA transfection, but it could also involve
the donation of key reprogramming proteins from the
exosome-budding tumors. It is thus conceivable that tumor-
derived exosomes could reprogram mast cells to promote host
immunosuppression, subvert T-regulatory cell function, and
even themselves behave rather like malignant tumor cells.
In light of this, the selective, yet controlled inhibition of
tumor-associated mast cell function (including exosome export)
could go a significant way toward restoring the normal healthy
immune response of some cancer patients, and possibly elicit a
cancer cure, if allied with simultaneous antitumor drug therapy
and surgery. However, such a hypothesis can only be tested
properly, if selectively cytotoxic small molecule inhibitors of
mast cell function can be identified, as opposed to simple mast
cell stabilizers (which will only prevent mast cell degranulation,
rather than actually stopping the mast cell exosome budding).
It was with these thoughts in mind that we became interested
in the recent report by McKee and co-workers8a that the
structurally novel bicyclic guanidine alkaloid, (+)-monanchorin,
showed cytotoxicity (IC50 = 11.7 μg/mL) in the NCI high-
throughput murine IC2 mast cell differential cytotoxicity
assay.8b The latter screens molecules for their ability to
selectively inhibit mast cells of the wild type, as well as those
carrying the D814Y mutation (which is c-Kit activating and
equivalent to that found in human D816Y mast cells).
Notwithstanding this assay providing information about
selective cytotoxicity, the original isolation team never revealed
whether (+)-monanchorin was indeed selectively cytotoxic
toward IC2 mast cells. Despite this, the very unusual structure
of (+)-monanchorin still makes it a most alluring small
molecule lead for further medicinal chemistry exploration and
exploitation, and it was with the latter in mind that we
ultimately decided to develop a new and highly practical
asymmetric total synthesis of (+)-monanchorin to assist those
efforts.
So far, only two total syntheses of (+)-monanchorin have
been achieved. The first, by Snider and Yu,9 unambiguously
defined the bicyclic ring structure of the natural product and
confirmed the absolute configuration as that depicted in
Scheme 1. The Snider pathway utilized a Shi catalytic
asymmetric epoxidation10 to create the key chiral epoxide
needed. This epoxide was then ring opened with azide ion to
obtain an inseparable 1.25:1 mixture of two regioisomers which,
after further synthetic manipulation, were ultimately separated
at the very final stages of the synthesis, the major isomer being
converted through to (+)-monanchorin by a novel acid-
mediated bicyclization process. The other route, developed by
Sutherland,11 exploited a highly stereoselective Pd-catalyzed
Overman trichloroacetimidate rearrangement12 to set the
nitrogen stereocenter present in the target (with 12:1
stereoselectivity), and although this did lead to a very elegant
and efficient total synthesis, it did require the use of noxious
and highly toxic trichloroacetonitrile in one of the steps. It also
employed (R)-glycidol as its chiral starting material, which
presently is sold at £18 a gram by the Aldrich Chemical Co. In
Scheme 1. Our Retrosynthetic Analysis of (+)-Monanchorin
light of these issues, we decided to embark on the development
of a new, much more convenient and cheap, asymmetric total
synthesis of (+)-monanchorin based upon the retrosynthetic
approach outlined in Scheme 1.
In this, we proposed to create the guanidine-hydroxy-
aldehyde 2 from the aldehyde 4 by treatment with anhydrous
trifluoroacetic acid.9,13 Compound 2 was then expected to
cyclize and dehydrate to form the seven-membered guanidi-
nium imine 1 which would subsequently undergo internal
nucleophilic addition to provide (+)-monanchorin. We believed
that aldehyde 4 would be derivable from the catalytic
hydrogenation of enal 5 which itself would be available from
the Wittig reaction of 6 with 7. Aldehyde 6 would potentially
derive from the α-azido ester 8 by azide reduction,
guanidinylation, N,O-isopropylidenation, and chelation-con-
trolled semireduction of the ester. A Sharpless AD14 and a
cyclic sulfate ring opening15 were planned for the synthetic
acquisition of 8.
We commenced the above route with the Wittig reaction
between n-hexanal and methyl triphenylphosphoranylidene
acetate in CH2Cl2 which afforded the (E)-alkene 11
predominantly in 85% yield (Scheme 2). This was then
subjected to Sharpless AD13 with AD-mix-α to give the syn-diol
10 cleanly in 75−84% yield and 92−93% ee. The
aforeomentioned ee estimate was based on NMR analysis of
the (R)-Mosher ester of azido alcohol 8, which itself was
derived from 10 following cyclic sulfate formation and azide
displacement under Sharpless’ now standard ring-opening
conditions, which furnished 8 exclusively. We progressed the
synthesis by catalytically hydrogenolyzing 8 to obtain the amine
13 and guanidinylating the latter with 14 in the presence of
stoichiometric silver(I) nitrate. Although both of these steps
went well, the subsequent N,O-isopropylidenation of 15 to
access 17 proved difficult to achieve, providing only the enol
ether 16 in a most disappointing 34% yield. The reaction of 15
with dimethoxypropane and p-TsOH in dry acetone was also
tested, but 17 was still not isolated.
2155
dx.doi.org/10.1021/ol500616v | Org. Lett. 2014, 16, 2154−2157

Organic Letters
Letter
Scheme 2. Our Initial Attempt at Preparing Aldehyde 4
Scheme 4. Attempted Execution of the Revised Plan
cleavable Boc groups remained intact and 20 was isolated in
good yield. Slow addition of DIBAL (2.45 equiv) to a −78 °C
solution of 20 in PhMe and stirring for a total of 2 h then
effected a highly efficient reduction of the ester to give the
cyclized guanidine-hemiaminal 19 in 93% yield after aqueous
workup. Unfortunately, the cyclic hemiaminal 19 failed to react
with ylide 7 under a range of conditions. We therefore
attempted the olefination of 19 with the more nucleophilic
methyl triphenylphosphoranylidene acetate, and to our delight,
this reaction proceeded successfully over 18 h at rt to give the
enoate 21 in 61% yield. The alkene in 21 was thereafter
reduced to the alkane, and another cyclic guanidine-intercepted
ester reduction was attempted with DIBAL at low temperature.
Again, this selective reduction worked out well, with aldehyde
18 arising. At this point in the route, we were confident that we
could complete the synthesis of (+)-monanchorin by a simple
CF3CO2H-induced O-desilylation and bicyclization (or vice
versa), but unfortunately, the TBS group did not cleave during
this attempted ring-closure process.
Given this disappointing result, we decided to change tack,
and return to the hydroxy-guanidine 15 (Scheme 3). We
Scheme 3. A Revised Retrosynthetic Plan
Given this situation, and the fact that Sutherland11 had
successfully cleaved a MOM-ether with dry CF3CO2H13 during
the final step of his (+)-monanchorin synthesis, we reasoned
that the use of a MOM-ether might produce a more satisfactory
outcome in our projected route. Accordingly, we returned once
more to the azido-alcohol 8 (Scheme 5), protected its hydroxyl
as a MOM-ether with MOMCl and Hunig’s base, and
hydrogenolyzed the azido group of 22 with 20% Pd(OH)2
on carbon in EtOAc. Amine 23 was then reacted with N,N-bis-
Boc-S-methylisothiourea 14 in the presence of AgNO3 and
Et3N at 0 °C. The result was the guanidine ester 24, which was
formed in 76% overall yield over three steps. DIBAL-reduction
now proceeded successfully, delivering the guanidine hemi-
aminal 25 in 79% yield. We next investigated its Wittig
olefination with the stabilized ylide 7. Not surpisingly, many
sets of reaction conditions failed to produce a satisfactory result,
but a partial success was eventually achieved when this reaction
was conducted in PhMe with 1.5 equiv of 7 at 70 °C in a sealed
tube over 18 h. However, the yield of enal 26 was typically poor
(31%) and there was a significant amount of accompanying
degradation; it also proved hard to completely purify 26 from
the triphenylphosphine oxide byproduct!
reasoned that if the hydroxyl group of 15 could be protected as
an OTBS ether, we might be able to effect a controlled semi-
reduction of the ester to the aldehyde, which could be
immediately intercepted by the pendant guanidine group to
give the guanidine hemiaminal 19. A Wittig reaction with 7
might then yield an enal whose olefin could be reduced
selectively. Acid-mediated deprotection could then possibly
complete the pathway to (+)-monanchorin in a manner
analogous to that shown in Scheme 1. Although this was a
very bold and direct synthetic plan, a search of the literature did
not reveal a single precedent for performing a Wittig olefination on
an uncharged cyclic guanidine-hemiaminal like 19 with a stabilized
phosphorus ylide,16 and a second issue that could potentially
arise, if this reaction did succeed, was in situ Michael cyclization
of the liberated guanidine anion17 onto the product enal.
Despite these significant risks, the great brevity and overall
simplicity of this route still attracted us toward it.
Alcohol 15 was thus protected with TBSOTf (1.1 equiv) and
2,6-lutidine (2.2 equiv) in dry CH2Cl2 (Scheme 4) at −78 °C
for 1 h and rt for a further 45 min. Fortunately, the potentially
2156
dx.doi.org/10.1021/ol500616v | Org. Lett. 2014, 16, 2154−2157

Organic Letters
Letter
Scheme 5. A New Total Synthesis of (+)-Monanchorin
ASSOCIATED CONTENT
* Supporting Information
■
S
NMR and IR spectra and HRMS and full experimental
procedures. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: k.j.hale@qub.ac.uk.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank QUB for a studentship (L.W.) and a starter grant
(K.J.H.).
■
REFERENCES
■
(1) Reviews: (a) Ribatti, D.; Crivellato, E. Biochim. Biophys. Acta
2012, 1822, 2. (b) Khazai, K.; Blatner, N. R.; Khan, M. W.; Gounari,
F.; Gounaris, E.; Dennis, K.; Bonertz, A.; Tsai, F.-N.; Strouch, M. J.;
Cheon, E.; Phillips, J. D.; Beckhove, P.; Bentrem, D. J. Cancer
Metastasis Rev. 2011, 30, 45. (c) Maltby, S.; Khazaie, K.; McNagny, K.
M. Biochim. Biophys. Acta 2009, 1769, 19.
(2) (a) Nitta, H.; Wada, Y.; Kawano, Y.; Murakami, Y.; Irie, A.;
Taniguchi, K.; Kikuchi, K.; Yamada, Y.; Suzuki, K.; Honda, J.; Wilson-
Morifuji, M.; Araki, N.; Eto, M.; Baba, H.; Imamura, T. Clin. Cancer
Res. 2013, 19, 2004. (b) For our work on C5a antagonism with
aurantimycin and the A83586C-citropeptin hybrid, see: Assem, E.-S.
K.; Peh, K. H.; Wan, B. Y.; Manaviazar, S.; Walters, M. A.; George, J.
H.; Hale, K. J. Inflamm. Res. 2008, 57 (Issue 1 Suppl.), 21.
(3) Westphal, E. In Ferbenanlytische Untersuchungen Zur Histologie
Und Klinik Des Blutes; Erlich, P., Ed.; Hirshwald: Berlin, 1891; p 17.
(4) Szajnik, M.; Czystowska, M.; Szczepanski, M. J.; Mandapathil, M.;
Whiteside, T. L. PLoS One, 2010, 5, Issue 7, e11469. DOI:10.1371/
journal.pone.0011469 (accessed Jan 13, 2014).
(5) (a) Hannafon, B. N.; Ding, W.-Q. Int. J. Mol. Sci. 2013, 14, 14240.
(b) Taylor, D. D.; Gercel-Taylor, C. Semin. Immunopathol. 2011, 33,
441. (c) Huber, V.; Filipazzi, P.; Iero, M.; Fais, S.; Rivoltini, L. J.
Translat. Med. 2008, 6, 63.
(6) (a) Valadi, H.; Ekstrom, K.; Bossios, A.; Lee, J. J.; Lotvall, J. O.
Nat. Cell Biol. 2007, 9, 654. (b) Lee, Y.; El Andolousi, S.; Wood, M. J.
A. Hum. Mol. Genet. 2012, R1−R10.
(7) Peng, P.; Yan, Y.; Keng, S. Oncol. Rep. 2011, 25, 749.
(8) (a) Meragelman, K. M.; McKee, T. C.; McMahon, J. B. J. Nat.
Prod. 2004, 67, 1165. (b) Henrich, C. J.; Goncharova, E. I.; Wilson, J.
A.; Gardella, R. S.; Johnson, T. R.; McMahon, J. B.; Takada, K.;
Bokesch, H. R.; Gustafson, K. R. Chem. Biol. Drug Res. 2007, 69, 321.
(9) Yu, M.; Snider, B. B. Org. Lett. 2009, 11, 1031.
(10) Zhao, M.-X.; Shi, Y. J. Org. Chem. 2006, 71, 5377.
(11) Zaed, A. M.; Sutherland, A. Org. Biomol. Chem. 2010, 8, 4394.
(12) Overman, L. E. J. Am. Chem. Soc. 1976, 98, 2901.
(13) For the mechanism of “dry” CF3CO2H-mediated acetal
deacetalation, see: Li, W.; Li, J.; Wu, Y.; Fuller, N.; Markus, M. A. J.
Org. Chem. 2010, 75, 1077.
(14) Sharpless, K. B.; Amberg, W.; Bennani, Y. L.; Crispino, G. A.;
Hartung, J.; Jeong, K. S.; Kwong, H. L.; Morikawa, K.; Wang, Z. M. J.
Org. Chem. 1992, 57, 2768.
Fortunately, the room temperature Wittig reaction of methyl
triphenylphosphoranylidene acetate with the guanidine-hemi-
aminal 25 proceeded cleanly when conducted in CH2Cl2 at rt
for 18 h; a 64% yield of the desired enoate 27 was obtained. As
was the case with 21, enoate 27 had to be used immediately for
the subsequent catalytic hydrogenation step, without any undue
delay, otherwise self-destructive internal Michael addition and
polymerization gradually started to consume it (when in highly
concentrated form). The hydrogenation step gave the methyl
ester 28 in 91% yield.
A selective −75 °C semireduction of the methyl ester in 28
was now attempted with DIBAL (2.5 equiv) in order to obtain
the seven-membered hemiaminal, which underwent sponta-
neous ring opening upon aqueous workup to give the open-chain
aldehyde 29 (presumably to relieve ring strain). Unfortunately,
all of our different attempts to purify this compound by SiO2
flash chromatography led to a much more polar pair of
products being formed, which we suspected were the seven-
membered cyclic guanidine hemiaminals. Given the difficulties
associated with purifying this mixture, we eventually decided to
treat crude 29 directly with CF3CO2H in CH2Cl2 and, after 4 d
at reflux, the TFA-salt of (+)-monanchorin was isolated in 40%
overall yield, after SiO2 flash chromatography. Importantly, the
product had NMR spectra and spectral data that matched
closely with those published by Snider9 and Sutherland.11
In summary, we have developed a new and fully stereo-
controlled asymmetric synthesis of the mast cell inhibitory
alkaloid, (+)-monanchorin, that delivers this molecule as its
TFA salt in high ee and high overall yield (7.2%, 12 steps).
(15) Hale, K. J.; Manaviazar, S.; Delisser, V. M. Tetrahedron 1994, 50,
9181.
(16) For the opening of a positively charged guanidium hemiaminal
with a nonstabilized phosphorus ylide, see: Arnold, M. A.; Day, K. A.;
Duron, S. G.; Gin, D. Y. J. Am. Chem. Soc. 2006, 128, 13255.
(17) Evans, D.; Murphy, P. J. Chem. Commun. 2011, 47, 3225.
2157
dx.doi.org/10.1021/ol500616v | Org. Lett. 2014, 16, 2154−2157