Total Synthesis of Millingtonine

Article abstract of DOI:10.1002/anie.201602869

Millingtonine is a glycosidic alkaloid that exists as a pair of pseudo-enantiomeric diastereomers. Consideration of the likely biosynthetic origins of this unusual natural product has resulted in the development of a seven-step total synthesis. Results fr

Full text of DOI:10.1002/anie.201602869

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201602869
German Edition: DOI: 10.1002/ange.201602869
Total Synthesis
Total Synthesis of Millingtonine
Patrick D. Brown and Andrew L. Lawrence*
Abstract: Millingtonine is a glycosidic alkaloid that exists as
a pair of pseudo-enantiomeric diastereomers. Consideration of
the likely biosynthetic origins of this unusual natural product
has resulted in the development of a seven-step total synthesis.
Results from this synthetic work provide evidence in support of
a proposed network of biosynthetic pathways that can account
for the formation of several phenylethanoid natural products.
pathway towards millingtonine (1) has been proposed and,
as commented upon by the isolation team, “the mechanism of
insertion of this (C4N) unit between the two C6C2 units is
unknown”.[1]
There has been one previous total synthesis of milling-
tonine (1), reported in 2012 by the research groups of Ley,
Kirschning, and Baxendale.[2] The execution of a total syn-
thesis of this alkaloid is an impressive achievement, with the
intermediate structures en route to millingtonine (1) reported
to be “exceedingly prone” to rearrangement reactions. In
total, the Ley–Kirschning–Baxendale synthesis required
twelve linear steps from commercially available materials
(sixteen steps in total) and produced milligram quantities of
material. We were hopeful that if we could gain new insight
into how nature synthesizes millingtonine (1) we might be
able to develop a new, more step-economical, synthetic
strategy.
Previous biomimetic studies on other phenylethanoid
natural products provided some important clues as to the
potential origins of millingtonine (1).[3] We considered that
a phenylethanoid glycoside 2, which contains an ornithine-
derived N-linked putrescine unit, might represent a reason-
able biosynthetic precursor. Our biosynthetic proposal, which
is shown in Scheme 2, involves a network of pathways that can
account for the formation of several structurally distinct
natural products isolated from Bignoniaceae plants: corno-
side (3),[4] rengyolone (4),[5] incarviditone (5),[6] incarvillea-
tone (6),[7] incargranine B (7),[8] and millingtonine (1).[1]
In our proposal, diamine 2 can undergo an oxidative
dearomatization to form imine 8 (Scheme 2; pathway 1),
which following hydrolysis would give the known para-quinol
natural product cornoside (3). It has been shown that cleavage
of the glycosidic bond in cornoside (3) results in concomitant
oxa-Michael cyclisation to give the racemic natural product
rengyolone (4).[9] We previously investigated a biomimetic
domino-Michael dimerization of rengyolone (4) to access
incarviditone (5),[3a] a racemic homochiral dimer. From our
synthetic studies we were able to reassign the relative
stereochemistry of incarviditone (5) and isolate an unex-
pected racemic heterochiral dimer, which was subsequently
reported as a natural product named incarvilleatone (6).[7] An
alternative biosynthetic pathway from diamine 2 (Scheme 2;
pathway 2) involves oxidative deamination to give amino-
aldehyde 9, which would be expected to undergo intra-
molecular condensation to give enamine 10. We predicted
enamine 10 would then undergo rapid dimerization with its
corresponding iminium ion 11, which formed the basis of our
previous structural reassignment and biomimetic synthesis of
incargarnine B (7).[3c]
I
n 1996 Yamasaki and co-workers isolated the alkaloid
millingtonine (1) from Millingtonia hortensis, an ornamental
Bignonia plant more commonly known as the indian cork
tree.[1] Millingtonine (1) was isolated as a mixture of two
diastereomeric alkaloids, which contain a molecular frame-
work not previously known to exist in the natural world.
Conceptually, but not biosynthetically (see below), milling-
tonine (1) can be considered to consist of a racemic aglycone
core that is “resolved” into two pseudo-enantiomeric diaste-
reomers by the attachment of a pair of b-d-glucopyranosyl
units. Biosynthetically, millingtonine (1) is likely constructed
from two shikimate-derived C6C2 units, linked together by an
ornithine-derived C4N unit (Scheme 1). No biosynthetic
Scheme 1. Structure and retro-biosynthetic analysis of millingtonine
(1). Glc=d-glucopyranosyl.
[*] P. D. Brown, Dr. A. L. Lawrence
EaStCHEM School of Chemistry
University of Edinburgh, Joseph Black Building
David Brewster Road, Edinburgh, EH9 3FJ (UK)
E-mail: a.lawrence@ed.ac.uk
With these two divergent biosynthetic pathways in mind
we recognized the possibility that the two pathways could re-
converge to give millingtonine (1) (Scheme 2; pathway 3).
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under http://dx.doi.org/10.
1002/anie.201602869.
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
Scheme 2. Proposed network of biosynthetic pathways towards a family of plant-derived phenylethanoid natural products, including millingtonine
(1).
Thus, a Michael reaction between enamine 10 and cornoside
(3) would give an intermediate iminium ion 12, which would
rapidly ring-close through an oxa-Mannich reaction to give
millingtonine (1). Therefore, in our proposed biogenesis, the
two diastereomers of millingtonine (1) are the result of a lack
of stereochemical influence exerted by the sugars of enamine
10 and cornoside (3) in a crossed-dimerization, rather than
a late-stage glucosidation of a racemic aglycone.
In our previous work (Scheme 3),[3c] the incargranine B
framework (13) was accessed through a biomimetic domino
Mannich/SEAr (electrophilic aromatic substitution) reaction
sequence, initiated by acidic-deprotection of acetal-protected
amino-aldehyde 14.[10] It was envisaged that having para-
quinol 15[11] present under these acidic deprotection con-
ditions would allow for a biomimetic crossed-dimerization to
occur to give the millingtonine framework. However, after
extensive experimentation, screening various acidic reaction
conditions, no crossed-dimerizations could be achieved. Only
formation of the incargranine B and dia-incargranine B
aglycone structures (13 and 16) was observed. It was
concluded that the domino Mannich/SEAr reaction sequen-
ces, en route to the incargranine B frameworks, are too fast
for crossed-dimerization to compete. We considered, how-
ever, that a window of opportunity for crossed-dimerization
might exist if an N-aryl enamine (akin to 17) could be
accessed whilst avoiding formation of the highly electrophilic
N-aryl iminium species (akin to 18).
Many strategies were investigated before we settled upon
an alkene isomerization approach, using transition-metal
hydride catalysis.[12] Thus, commercially available 4-amino-
phenethyl alcohol 19 was protected as an acid-labile tert-
butoxycarbamate before a b-selective glucosidation was
Scheme 3. Previously reported biomimetic synthesis of the incargrani-
ne B aglycone 13.[3c] TBS=tert-butyldimethylsilyl.
achieved, using the pivaloyl-protected trichloroacetimidate
20 (Scheme 4).[13] Under the acidic glucosidation reaction
conditions, deprotection of aniline 21 occurred in situ to give
the pivaloyl-protected glucoside 22 in 53–55% yield over the
two steps. Condensation of aniline 22 with (Z)-1,4-dichlor-
obut-2-ene then gave N-aryl-2,5-dihydropyrrole 23 in 50–
52% yield.[14] It was found that the commercially available
compound RhHCO(PPh3)4 was a competent catalyst for the
2
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!

Angewandte
Communications
Chemie
Scheme 4. Biomimetic total synthesis of millingtonine (1). Boc=tert-butyloxycarbonyl, TMSOTf=trimethylsilyl trifluoromethanesulfonate, Glc=d-
glucopyranosyl, Piv=pivaloyl, TFA=trifluoroacetic acid.
isomerization of skipped enamine 23 into the conjugated
enamine 24.[15] All efforts to purify and isolate enamine 24,
however, were met with failure. Therefore, the search for
suitable reaction conditions for cross-dimerization of enam-
ine 24 with para-quinol 25, which was prepared in four steps
from tyrosol,[16] were pursued using freshly prepared solutions
of enamine 24. Pyrrolidine was investigated, in the hope of
establishing an iminium-organocatalytic cycle,[17] and gratify-
ingly led to crossed-dimerization for the first time, albeit in
very low yield. Control experiments, however, revealed that
the beneficial effect of adding pyrrolidine was down to the
unexpected formation of an aminal intermediate 26.[15]
Furthermore, inclusion of pyrrolidine at the beginning of
the rhodium-hydride-catalyzed isomerization reaction was
found to result in quantitative conversion of alkene 23 to
aminal 26.[18] Pivaloyl-protected cornoside 25 was then added
to this freshly prepared solution resulting in the formation of
a 1:1 diastereomeric mixture of crossed-dimers 27 which was
isolated in 50% yield. Formation of the cis,syn,cis-isomers 27
was an unexpected and presumably kinetically controlled
outcome, as we anticipated the desired cis,anti,cis ring system
would be thermodynamically more favorable. Our kinetic
versus thermodynamic reasoning was shown to be sound, as
exposure of the cis,syn,cis-isomers 27 to acidic conditions
resulted in isomerization to the desired cis,anti,cis-isomers 28
in 56–73% yield (Scheme 4). Presumably this isomerization
occurs via an acid-catalyzed retro-oxa-Mannich/iminium-
epimerization/oxa-Mannich reaction sequence. Finally,
removal of the pivaloyl groups under basic conditions gave
over 50 mg of millingtonine (1) in 67% yield. Thus we had
successfully synthesized millingtonine (1) in a longest linear
sequence of seven steps from commercially available materi-
als (ten steps in total), which compares favorably to the
previous state-of-the-art.[2]
The successful crossed-dimerization of aminal 26 with
para-quinol 25 (Scheme 4) stands in stark contrast to the
exclusive homodimerization observed when using acetal 14
(Scheme 3). Our failure to cross-dimerize acetal 14 was
attributed to the formation of a highly reactive N-aryl
iminium species 18, which leads to very fast and essentially
irreversible domino Mannich/SEAr reaction sequences
(Scheme 3). In our successful crossed-dimerization, however,
we still isolate a 14% yield of homodimers (29 and 30;
Scheme 4), indicating that an N-aryl iminium ion must still be
formed under these reaction conditions. We must then ask
how crossed-dimerization can compete. One possible explan-
À
ation is that it is more favorable for the endocyclic C N bond
À
of aminal 26 to cleave, in preference to the exocyclic C N
bond, to give the acyclic iminium ion 31 (Scheme 5a).
Therefore, the problematic N-aryl iminium species is present
at lower concentration and can also be rapidly trapped by the
pyrrolidine. Thus enamine 32, which would also be expected
to be more nucleophilic than the N-aryl enamine 24, has
a sufficient window of opportunity to react with para-quinol
25 (Scheme 5a). It is interesting to consider whether this
might, therefore, be indicative of a similar mechanism
operating within the plant. Might pyrrolidine be mimicking
the coenzyme role of vitamin B6 (that is, pyridoxal phos-
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
Scheme 5. a) Mechanistic speculation for our successful crossed-dimerization. B) Plausible biosynthetic pathway towards dia-millingtonine (33)
involving vitamin B6 (i.e., pyridoxal phosphate) as a coenzyme. Enz=enzyme.
phate) in an interrupted oxidative-deamination reaction of
diamine 2 (Scheme 5b)?[19]
In summary, through synthesis alone we have been able to
probe the feasibility of our proposed biogenesis of milling-
tonine (1). We have been able to obtain evidence in support of
[1] T. Hase, K. Ohtani, R. Kasai, K. Yamasaki, C. Picheansoonthon,
a unified network of biosynthetic pathways that can account
Phytochemistry 1996, 41, 317 – 321.
for the formation of a family of phenylethanoid natural
[2] J. Wegner, S. V. Ley, A. Kirschning, A.-L. Hansen, J. M. Garcia,
products. Crucially, these new biosynthetic insights have
I. R. Baxendale, Org. Lett. 2012, 14, 696 – 699.
enabled the development of a highly step economical total
synthesis of millingtonine (1). Central to the development of
this successful synthesis was the use of transition-metal
hydride catalysis to access an otherwise unstable biomimetic
intermediate, an approach we envisage will prove applicable
to other biomimetic studies. The future isolation of the
cis,syn,cis isomer of millingtonine, namely dia-millingtonine
(33), from the natural environment would represent a new
case of “natural product anticipation”—an essentially unique
benefit of following biomimetic strategies in synthesis.[20]
[3] For previous biomimetic studies on related dimeric natural
products, see: a) P. D. Brown, A. C. Willis, M. S. Sherburn, A. L.
Lawrence, Org. Lett. 2012, 14, 4537 – 4539; b) K. Zhao, G.-J.
Cheng, H. Yang, H. Shang, X. Zhang, Y.-D. Wu, Y. Tang, Org.
Lett. 2012, 14, 4878 – 4881; c) P. D. Brown, A. C. Willis, M. S.
Sherburn, A. L. Lawrence, Angew. Chem. Int. Ed. 2013, 52,
13273 – 13275; Angew. Chem. 2013, 125, 13515 – 13517.
[4] J. S. Rosendal, A. Kjaer, N. B. Juhl, Acta Chem. Scand. 1973, 27,
367 – 369.
[5] K. Endo, H. Hikino, Can. J. Chem. 1984, 62, 2011 – 2014.
[6] Y. Q. Chen, Y. H. Shen, Y. Q. Su, L. Y. Kong, W. D. Zhang,
Chem. Biodiversity 2009, 6, 779 – 783.
[7] Y. P. Gao, Y. H. Shen, S. D. Zhang, J. M. Tian, H. W. Zeng, J. Ye,
H. L. Li, L. Shan, W. D. Zhang, Org. Lett. 2012, 14, 1954 – 1957.
[8] Y.-H. Shen, Y.-Q. Su, J.-M. Tian, S. Lin, H.-L. Li, J. Tang, W.-D.
Zhang, Helv. Chim. Acta 2010, 93, 2393 – 2396.
Acknowledgements
P.D.B. thanks the University of Edinburgh for the provision of
a studentship. The Royal Society is thanked for the award of
a Research Grant. A.L.L. thanks Prof. Mick Sherburn
(Australian National University) and Prof. Guy Lloyd-Jones
FRS (University of Edinburgh) for helpful discussions.
[9] K. Endo, K. Seya, H. Hikino, Tetrahedron 1989, 45, 3673 – 3682.
[10] There exists the possibility that this dimerization proceeds
through a concerted cycloaddition mechanism, with dimers 16
and 13 the result of endo and exo processes, respectively. For
further discussion, see: a) D. A. Powell, R. A. Batey, Org. Lett.
2002, 4, 2913 – 2916; b) V. K. Kouznetsov, Tetrahedron 2009, 65,
2721 – 2750.
Keywords: alkaloids · biomimetic synthesis · domino reactions ·
natural products · phenylethanoids
[11] The para-quinol 15 was synthesized in two steps, see Ref. [3a] for
details.
4
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!

Angewandte
Communications
Chemie
[12] a) W. A. Herrmann, M. Prinz, Applied Homogeneous Catalysis
Vitamin B6: Pyridoxal Phosphate, Part B, Coenzymes and
with Organometallic Compounds, 2nd ed., Wiley-VCH, Wein-
heim, 2002; b) T. J. Donohoe, T. J. C. OꢀRiordan, C. P. Rosa,
Angew. Chem. Int. Ed. 2009, 48, 1014 – 1017; Angew. Chem. 2009,
121, 1032 – 1035.
Cofactors, Vol. 1, Wiley Interscience, 1986.
[20] For recent examples of natural product anticipation from
biomimetic studies, see: a) D. P. OꢀMalley, K. Li, M. Maue,
A. L. Zografos, P. S. Baran, J. Am. Chem. Soc. 2007, 129, 4762 –
4775; b) P. Sharma, B. Lygo, W. Lewis, J. E. Moses, J. Am. Chem.
Soc. 2009, 131, 5966 – 5972; c) M. Gavagnin, E. Mollo, G.
Cimino, Rev. Bras. Farmacogn. 2015, 25, 588 – 591; d) A. L.
Lawrence, R. M. Adlington, J. E. Baldwin, V. Lee, J. A. Kershaw,
A. L. Thompson, Org. Lett. 2010, 12, 1676 – 1679; e) V. Sofiyev,
J.-P. Lumb, M. Volgraf, D. Trauner, Chem. Eur. J. 2012, 18, 4999 –
5005; f) S. Strych, G. Journot, R. P. Pemberton, S. C. Wang, D. J.
Tantillo, D. Trauner, Angew. Chem. Int. Ed. 2015, 54, 5079 –
5083; Angew. Chem. 2015, 127, 5168 – 5172; g) P. Ellerbrock, N.
Armanino, M. K. Ilg, R. Webster, D. Trauner, Nat. Chem. 2015,
7, 879 – 882.
[13] P. Zimmermann, R. Sommer, T. Bꢁr, R. R. Schmidt, J. Carbo-
hydr. Chem. 1988, 7, 435 – 452.
[14] J. M. Bobbitt, L. H. Amundsen, R. I. Steiner, J. Org. Chem. 1960,
25, 2230 – 2231.
[15] See the Supporting Information for details and related model
studies.
[16] See the Supporting Information for details.
[17] A. Erkkilꢁ, I. Majander, P. M. Pihko, Chem. Rev. 2007, 107,
5416 – 5470.
[18] Determined by comparison to an internal standard using
1H NMR spectroscopy, see the Supporting Information for
details.
[19] a) D. Dolphin, R. Poulson, O. Avramovic, Vitamin B6: Pyridoxal
Phosphate, Part A, Coenzymes and Cofactors, Vol. 1, Wiley
Interscience, 1986; b) D. Dolphin, R. Poulson, O. Avramovic,
Received: March 22, 2016
Published online: && &&, &&&&
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Total Synthesis
P. D. Brown,
A. L. Lawrence*
&&&& — &&&&
Total Synthesis of Millingtonine
Hidden symmetry: A seven-step total
synthesis of the glycosidic alkaloid mill-
ingtonine was developed by considering
its likely biosynthetic origins. Synthetic
results provide evidence in support of
a proposed network of biosynthetic
pathways that can account for the for-
mation of several phenylethanoid natural
products. Glc=d-glucopyranosyl, Piv=
pivaloyl.
6
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!