Synthesis of (-)-Conduramine A1, (-)-Conduramine A2 and (-)-Conduramine E2 in Six Steps from Cyclohexa-1,4-diene

Article abstract of DOI:10.1021/acs.orglett.9b02914

A method to enable the synthesis of conduramines and their N-substituted derivatives (enantiopure or racemic form) in six steps (five steps for N-substituted derivatives) from cyclohexa-1,4-diene is reported. Key features of this reaction sequence include a preparation of benzene oxide that is amenable to multigram scale, and its efficient ring-opening upon treatment with a primary amine. Epoxidation of the resultant amino alcohols (40% aq HBF₄ then m-CPBA) is accompanied by hydrolytic ring-opening in situ to give the corresponding N-substituted conduramine derivatives directly. These may undergo subsequent N-deprotection to give the parent conduramines, as demonstrated by the preparation of enantiopure (-)-conduramine A1, (-)-conduramine A2, and (-)-conduramine E2 (the latter two for the first time). The selectivity of the epoxidation reaction is proposed to be the result of competitive ammonium-directed and hydroxyl-directed epoxidation processes, followed by either direct (S_N2-type) or conjugate (S_N2′-type) ring-openings of the intermediate epoxides.

Full text of DOI:10.1021/acs.orglett.9b02914

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Synthesis of (−)-Conduramine A1, (−)-Conduramine A2 and
(−)-Conduramine E2 in Six Steps from Cyclohexa-1,4-diene
Solange Da Silva Pinto, Stephen G. Davies,* Ai M. Fletcher, Paul M. Roberts, and James E. Thomson
Department of Chemistry, Chemistry Research Laboratory, University of Oxford, Mansfield Road, Oxford OX1 3TA, United
Kingdom
S
* Supporting Information
ABSTRACT: A method to enable the synthesis of conduramines and
their N-substituted derivatives (enantiopure or racemic form) in six
steps (five steps for N-substituted derivatives) from cyclohexa-1,4-
diene is reported. Key features of this reaction sequence include a
preparation of benzene oxide that is amenable to multigram scale, and
its efficient ring-opening upon treatment with a primary amine.
Epoxidation of the resultant amino alcohols (40% aq HBF4 then m-
CPBA) is accompanied by hydrolytic ring-opening in situ to give the
corresponding N-substituted conduramine derivatives directly. These may undergo subsequent N-deprotection to give the
parent conduramines, as demonstrated by the preparation of enantiopure (−)-conduramine A1, (−)-conduramine A2, and
(−)-conduramine E2 (the latter two for the first time). The selectivity of the epoxidation reaction is proposed to be the result of
competitive ammonium-directed and hydroxyl-directed epoxidation processes, followed by either direct (SN2-type) or
conjugate (SN2′-type) ring-openings of the intermediate epoxides.
onduramines are derivatives of conduritols (cyclohexa-5-
ene-1,2,3,4-tetraols) 1 in which one of the hydroxyl
C
functionalities has been replaced by an amino functionality.1
They have attracted interest due to their often potent
biological activity as selective glycosidase inhibitors, a class of
enzymes which are involved with the progression of a number
of diseases.1 The residue of conduramine F1, for example, is
found as a substructural unit within acarbose (glucobay) 2, a
drug used for the treatment of type 2 diabetes mellitus.2 This
property of the conduramines has resulted in a significant
amount of investigation being lavished on them. Interestingly,
it has been found that derivatives of conduramines often
possess enhanced biological activity over that of the parents;
for example, (−)-conduramine B1 3 itself has been shown to
lack any inhibitory activity against a range of glycosidases,
while the corresponding N-benzyl derivative 4 was found to
display selective activity.3 To further the interest in these
compounds, a number of natural products containing the
conduramine core are also known, including the Amaryllida-
ceae alkaloids narciclasine 54 and lycoricidine 6,5 both of which
contain the residue of conduramine A1 as a substructural unit
and possess desirable biological activity (Figure 1).
Figure 1. Structures of conduritol 1, acarbose (glucobay) 2,
(−)-conduramine B1 3, N-benzyl conduramine B1 4, narciclasine 5,
and lycoricidine 6.
The only source of these biologically interesting amino
polyols is by means of laboratory synthesis, as none of the
conduramine family themselves are naturally occurring.1
Unsurprisingly, therefore, several routes to these compounds
and their derivatives have been reported, although these are
often rather lengthy and give rise to only one or two
conduramine products.1,6 Given our previous experience
concerning the synthesis of dihydroconduramines,7−9 we
proposed that epoxidation (treatment with H+ then m-
CPBA) of allylic amino alcohols 7, derived from the ring-
opening of benzene oxide, would give the corresponding ring-
opened products 8 and 9, i.e., conduramines (R = H) or their
N-protected derivatives (R ≠ H) directly. This approach to
these compounds would be attractive in that it may allow the
Received: August 16, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b02914
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
preparation of several diastereoisomers of the targets for
biological profiling in very short order, and in any case would
provide further insight into the relative abilities of the
ammonium functionality versus the hydroxyl functionality to
direct the epoxidation reaction.10 Herein, we report our
preliminary findings within this area, which culminates in the
preparation of the racemic N-benzyl derivatives of condur-
amines A1, A2, and E2, and the enantiopure conduramines
(−)-A-1, (−)-A2, and (−)-E2 themselves, in six steps or fewer
from cyclohexa-1,4,-diene in each case (Figure 2).
as trans, due to the diagnostically large value of the 1H NMR 3J
coupling constant (3J = 12.1 Hz) between the protons attached
to the two stereogenic centers (Scheme 2).
Scheme 2. Preparation of Allylic Amino Alcohol 14
Treatment of 14 with 40% aq HBF4 and then m-CPBA gave
a 21:36:31:12 mixture of four compounds, subsequently
identified as N-benzyl conduramine A1 (18), N-benzyl
conduramine A2 (19), N-benzyl conduramine E2 (20), and
N-benzyl conduramine F2 (21), respectively. Purification by
preparatory t.l.c. gave 18 in 8% yield, 19 in 23% yield, 20 in
18% yield, and an impure sample of 21 in ∼5% yield, as single
diastereoisomers (>95:5 dr) in each case (Scheme 3). The
gross structures of 18−21 were assigned by analyses of the
typical range of 1D and 2D NMR spectra, while their relative
configurations were subsequently established following treat-
ment of each of 18−21 with H2 in the presence of
Pd(OH)2/C, which effected tandem hydrogenation of the
olefin functionality and hydrogenolytic removal of the N-
benzyl group to give the corresponding dihydroconduramines
A1 (22), A2 (23),8 E2 (24),9 and F2 (25);9,13 the samples of
Figure 2. Proposed synthesis of conduramines and their derivatives 8
and 9 from allylic amino alcohols 7.
Our first goal was the development of a synthesis of benzene
oxide amenable to scale-up, which was achieved by
modification of the method first reported by Gu
̈
nther.11
Treatment of cyclohexa-1,4-diene 10 with AcOOH in CH2Cl2
gave cyclohexa-1,4-diene monoepoxide 11 in 66% yield.
Treatment of 11 with Br2 in a mixture of CH2Cl2 and
CHCl3 delivered dibromide 12 in 91% yield (60% overall yield
from 10). When these two steps were telescoped, diluting the
initial reaction mixture with CHCl3 and adding Br2 to the same
reaction flask (i.e., obviating the isolation and purification of
11), dibromide 12 was isolated in 88% overall yield from 10,
on a multigram scale. Final treatment of 12 with DBU in Et2O
gave benzene oxide 13 in 54% yield; this was generated,
isolated, and immediately used, as required (Scheme 1).
1
23−25 gave H and 13C NMR spectra that matched those
previously reported8,9 (thus confirming their identities, and
hence the identities of 19−21), while diagnostic values of the
3
1H NMR J coupling constants observed for 22 allowed its
relative configuration (and hence that of 18) to be confidently
assigned. The observation of the four products 18−21 is
consistent with monoepoxidation of the diene functionality
within 14 being followed by a hydrolytic ring-opening reaction.
In order to gain some insight into the precise details of the
mechanism, the reaction was repeated using HBF4·OEt2 in
H218O (≥98% 18O) in place of 40% aq HBF4, which resulted in
formation of the same four compounds in the same ratio
(within experimental error), labeled with a single 18O atom
(92% incorporation of an 18O label in each case, as judged by
mass spectrometric analysis). The nature of the fragmentation
of these compounds rendered mass spectrometry unsuitable as
a tool to locate the position of the 18O label, and therefore
analysis of 16O/18O isotope-induced chemical shifts14 in the
13C NMR spectrum was employed to unambiguously locate
the 18O atom within each of these samples of 18−21: >95%
incorporation of the 18O label was observed at C(1) in all
cases, with all other positions showing negligible (<5%)
incorporation of the label (Scheme 3). On the basis of these
results, the following mechanistic hypothesis is proposed. As
one of the olefins bears an allylic hydroxyl functionality and the
other an allylic N-benzylamino functionality, both of which are
known to be able to direct the olefinic epoxidation reaction to
the proximal (syn) face in a six-membered-ring-system
(presumably by formation of a hydrogen bond in the transition
state),15 the rate of background (nondirected epoxidation) was
expected to be so low as to be a negligible contributor to the
selectivity observed in the epoxidation step. Thus, out of the
four possible regio- and diastereoisomeric products resulting
from monoepoxidation of 14, it is expected that only 16
(resulting from direction by the hydroxyl group) and 17
Scheme 1. Large Scale Preparation of Benzene Oxide 13
Benzene oxide 13 remains an under-utilized building block
in synthesis and its ring-opening, for example, has been
surprisingly little explored.12 We chose the ring-opening of 13
using benzylamine as a model system. Problems encountered
were lack of reactivity, or promotion of an undesired
rearrangement pathway giving phenol 15. However, after
optimization (variation of solvent, time, temperature, and the
presence of Lewis acids) it was found that treatment of 13 with
benzylamine in MeOH at 66 °C gave allylic amino alcohol 14
almost exclusively, and as a single diastereoisomer (>95:5 dr),
with only trace amounts (<5%) of phenol 15 being formed,
and upon purification 14 was isolated in 90% yield. The
relative configuration within 14 could be confidently assigned
B
DOI: 10.1021/acs.orglett.9b02914
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
showing no propensity for desymmetrization under these
conditions, as expected.18 These compounds were separated
chromatographically to give 26 in 19% yield and 27 in 21%
yield, along with a mixed fraction in 21% combined yield
(Scheme 4). The relative configuration within 27 was
unambiguously assigned by single crystal X-ray diffraction
analysis,19 with the absolute configuration following from the
known (R)-configuration of the α-stereocenter (Figure 3). The
dihedral angle between the two protons on the two stereogenic
centers in the solid state was 175°, and as before, the large
Scheme 3. Preparation of Racemic N-Benzyl Conduramines
A1 (18), A2 (19), E2 (20), and F2 (21)
value of the H NMR J coupling constant (3J = 12.9 Hz)
between these protons was consistent with an analogous
conformation being favored in solution and also indicative of
the trans relative configuration. On this basis, 26 was assigned
as the alternative trans diastereoisomer that would result from
ring-opening of the meso-epoxide 13 by the enantiopure
nucleophile; the large value of the 1H NMR 3J coupling
constant (3J = 11.5 Hz) between the protons on the two
stereogenic centers was supportive of this assignment.20
1
3
Scheme 4. Preparation of Allylic Amino Alcohols 26 and 27
a
The positions of the 18O labels when run with H218O (see text) are
PMP = para-methoxyphenyl.
shown in green; the proposed positions of the oxygen atoms derived
from m-CPBA are shown in either red (for 16 and derivatives) or blue
(for 17 and derivatives) for clarity. IUPAC numbering of 18−21 is
also shown.
(resulting from direction by the N-benzylammonium group)
are formed as intermediates in this reaction. Regioselective
ring-opening of 16 in an SN2-type fashion (with inversion of
configuration) gives 18, and an analogous process for 17 gives
19. Competitive SN′-type ring-opening of epoxide 17 would
result in the formation of 20 and 21. Although both SN1′- and
SN2′-types of hydrolytic ring-opening of 17 could give rise to
20 and 21 as a mixture of epimers,16 it is proposed that the
latter (i.e., SN2′-type process) is more likely in operation, as we
have observed no evidence of a predilection toward the
formation of a cation intermediate in any of the examples of
this reaction that we have previously studied.17 Importantly,
the lack of incorporation of the 18O label at C(4) within 20
indicated that SN′-type hydrolytic ring-opening of epoxide 16
was a negligible contributor to the production of 20 in this
reaction. Assuming this mechanistic rationale, the ratio of the
intermediate epoxides 16 and 17 in this reaction is therefore
represented by the ratio 18:(19+20+21), i.e., approximately
1:4, which is a measure of the relative directing abilities of the
two functionalities in this specific instance (Scheme 3). The
superior ability of the N-benzylamino substituent versus the
hydroxyl substituent to direct the epoxidation reaction (i.e.,
promote a faster reaction) is consistent with our previous
observations in a related system.10
Figure 3. X-ray crystal structure of 27 (selected H atoms have been
omitted for clarity).
Allylic amino alcohol 26 was arbitrarily selected for
elaboration to the corresponding conduramines (as 26 and
27 may be considered as pseudoenantiomeric for this
purpose). Treatment of 26 with 40% aq HBF4 and then m-
CPBA gave a 17:37:32:14 mixture of four compounds,
identified as N-α-methyl-p-methoxybenzyl conduramine A1
(30), N-α-methyl-p-methoxybenzyl conduramine A2 (31), N-
α-methyl-p-methoxybenzyl conduramine E2 (32), and N-α-
methyl-p-methoxybenzyl conduramine F2 (33), respectively.
Purification by preparatory t.l.c. gave 30 in 8% yield, 31 in 21%
yield, 32 in 15% yield, and an impure sample of 33 in ∼5%
yield, as single diastereoisomers (>95:5 dr) in each case. As
before, the gross structures of 30−33 were assigned by
analyses of the typical range of 1D and 2D NMR spectra and
the relative configurations within 31−33 were then assigned
With a route to racemic conduramine derivatives in hand, a
route to enantiopure materials was investigated. Ring-opening
of benzene oxide 13 upon treatment with enantiopure (R)-α-
methyl-p-methoxybenzylamine gave a mixture of two com-
pounds, 26 and 27, in an approximate 50:50 ratiothus
1
3
on the basis of the similarities between their H NMR J
coupling constants and those of the corresponding products
19−21 derived from amino alcohol 14.13 The absolute
C
DOI: 10.1021/acs.orglett.9b02914
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
configurations then followed from the known absolute
configurations of the stereocenters derived from the trans-
amino alcohol 26. The relative configuration within 30,
meanwhile, was subsequently assigned (vide infra). As the
ratio of 30−33 is the same (within experimental error) as the
ratio of the corresponding products 18−21 generated when 14
was subjected to the same reaction conditions, analogous
mechanisms to rationalize the formation of each of 30−33
from amino alcohol 26 may be surmised (Scheme 5).
Scheme 6. Preparation of Enantiopure Conduramines
(−)-A1 (34), (−)-A2 (35), and (−)-E2 (36)
Scheme 5. Preparation of Enantiopure N-α-Methyl-p-
methoxybenzyl Conduramines A1 (30), A2 (31), E2 (32),
and F2 (33)
Despite the modest diastereoselectivity of the epoxidation
reactions and multiple hydrolytic ring-opening reactions
resulting in the production of mixtures of isomeric products,
the low step-count results in overall yields which are
comparable to or better than those of other multistep
processes, as well as allowing rapid production of unknown
conduramines and derivatives for biological profilingthese
derivatives may display biological activity superior to that of
the parent compounds. In the case of an N-α-methyl-p-
methoxybenzyl protecting group, the corresponding enantio-
pure conduramines may be accessed upon deprotection.
Further investigations, including investigation of the diaster-
eoselectivity of the epoxidation reaction itself, are currently
ongoing in our laboratory and these results will be reported in
due course.
Finally, treatment of 30−32 with Et3SiH in the presence of
TFA effected removal of the α-methyl-p-methoxybenzyl
fragment to give (−)-conduramine A1 (34), (−)-conduramine
A2 (35), and (−)-conduramine E2 (36), in 82%, 79%, and
58% yield, respectively, and in >95:5 dr in each case. The
former, (−)-conduramine A1 (34), has been previously
described21 (as has its antipode),22 and thus the relative and
absolute configurations of 30 were secured. The latter two,
(−)-conduramine A2 (35) and (−)-conduramine E2 (36),
have not been previously described (Scheme 6).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.9b02914.
■
S
Experimental details, characterization data, and copies of
1H and 13C NMR spectra (PDF)
Accession Codes
In conclusion, the synthesis of a range of conduramines and
their N-protected derivatives has been achieved in very short
order from cyclohexa-1,4-diene. The key steps of the process
involve formation of benzene oxide and its ring-opening with a
primary amine. Epoxidation (HBF4 then m-CPBA) of the
resultant allylic amino alcohols is controlled by hydrogen-
bonding to either the allylic hydroxyl functionality or the in
situ formed allylic N-benzylammonium moiety, with the latter
being superior over the former in its ability to promote
epoxidation of the proximal olefin. Hydrolytic ring-opening of
the resultant epoxide intermediates occurs in situ, via either a
direct (SN2-type) process or a conjugate (SN2′-type) process,
giving a mixture of the corresponding amino triol products.
CCDC 1940566 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
via www.ccdc.cam.ac.uk/data_request/cif, or by emailing
data_request@ccdc.cam.ac.uk, or by contacting The Cam-
bridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: steve.davies@chem.ox.ac.uk.
ORCID
Stephen G. Davies: 0000-0003-3181-8748
D
DOI: 10.1021/acs.orglett.9b02914
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) For a review, see: Łysek, R.; Vogel, P. Tetrahedron 2006, 62,
2733.
(2) DiNicolantonio, J. J.; Bhutani, J.; O’Keefe, J. H. Open Heart
2015, 2, 327.
(3) Łysek, R.; Schutz, C.; Vogel, P. Bioorg. Med. Chem. Lett. 2005,
15, 3071.
(4) Ceriotti, G. Nature 1967, 213, 595.
(5) Okamoto, T.; Torii, Y.; Isogai, Y. Chem. Pharm. Bull. 1968, 16,
1860.
(6) For selected, more recent examples of syntheses of condur-
amines and their N-substituted derivatives, see: (a) Chang, Y.-K.; Lo,
H.-J.; Yan, T.-H. J. Chin. Chem. Soc. 2010, 57, 24. (b) Lu, P.-H.; Yang,
C.-S.; Devendar, B.; Liao, C.-C. Org. Lett. 2010, 12, 2642. (c) Kuno,
S.; Higaki, K.; Takahashi, A.; Nanba, E.; Ogawa, S. MedChemComm
2015, 6, 306. (d) Myeong, I.-S.; Kim, J.-S; Lee, Y.-T.; Kang, J.-C.;
Park, S.-H.; Jung, C.; Ham, W.-H. Tetrahedron: Asymmetry 2016, 27,
823. (e) Harit, V. K.; Ramesh, N. G. J. Org. Chem. 2016, 81, 11574.
(f) Prasad, K. R. Tetrahedron 2018, 74, 6689. (g) Harit, V. K.;
Ramesh, N. G. Org. Biomol. Chem. 2019, 17, 5951.
́
(7) Brennan, M. B.; Csatayova, K.; Davies, S. G.; Fletcher, A. M.;
Green, W. D.; Lee, J. A.; Roberts, P. M.; Russell, A. J.; Thomson, J. E.
J. Org. Chem. 2015, 80, 6609.
(8) Da Silva Pinto, S.; Davies, S. G.; Fletcher, A. M.; Roberts, P. M.;
Thomson, J. E. Synthesis 2018, 50, 64.
(9) Da Silva Pinto, S.; Davies, S. G.; Fletcher, A. M.; Roberts, P. M.;
Thomson, J. E. J. Org. Chem. 2018, 83, 9939.
(10) Brennan, M. B.; Davies, S. G.; Fletcher, A. M.; Lee, J. A.;
Roberts, P. M.; Russell, A. J.; Thomson, J. E. Aust. J. Chem. 2015, 68,
610.
̈
(11) Vogel, E.; Gunther, H. Angew. Chem., Int. Ed. Engl. 1967, 6,
385.
(12) For example, see: (a) DeMarinis, R. M.; Berchtold, G. A. J. Am.
Chem. Soc. 1969, 91, 6525. (b) Jeffrey, A. M.; Yeh, H. J. C.; Jerina, D.
M.; De Marinis, R. M.; Foster, C. H.; Piccolo, D. E.; Berchtold, G. A.
J. Am. Chem. Soc. 1974, 96, 6929. (c) Rastetter, W. H.; Chancellor, T.;
Richard, T. J. J. Org. Chem. 1982, 47, 1509−1512. (d) Bertozzi, F.;
Crotti, P.; Del Moro, F.; Feringa, B. L.; Macchia, F.; Pineschi, M.
Chem. Commun. 2001, 2606.
(13) Further details are contained within the SI.
(14) For a review of isotope effects in nuclear shielding, see Hansen,
P. E. Prog. Nucl. Magn. Reson. Spectrosc. 1988, 20, 207.
(15) Davies, S. G.; Fletcher, A. M.; Thomson, J. E. Org. Biomol.
Chem. 2014, 12, 4544.
(16) Unlike SN2-type processes, SN2′-type processes are known to
be non-stereospecific; see: Smith, M. B. March’s Advanced Organic
Chemistry: Reactions, Mechanisms, and Structure, 7th ed.; John Wiley &
Sons Inc.: Hoboken, NJ, 2013.
(17) See refs 6−10, 15, and references contained therein.
(18) For the related ring-opening of racemic 3,4-epoxycyclohexene
with a chiral amine, see: Knapp, S.; Sebastian, M. J.; Ramanathan, H.;
Bharadwaj, P.; Potenza, J. A. Tetrahedron 1986, 42, 3405−3410.
(19) Crystallographic data (excluding structure factors) have been
deposited with the Cambridge Crystallographic Data Centre as
supplementary publication number CCDC 1940566.
(20) These data also add further support to the assignment of the
trans relative configuration of 14.
(21) Myeong, I.-S.; Kim, J.-S.; Lee, Y.-T.; Kang, J.-C.; Park, S.-H.;
Jung, C.; Ham, W.-H. Tetrahedron: Asymmetry 2016, 27, 823.
́
(22) Norsikian, S.; Soule, J.-F.; Cannillo, A.; Guillot, R.; Tran Huu
Dau, M.-E.; Beau, J.-M. Org. Lett. 2012, 14, 544.
E
DOI: 10.1021/acs.orglett.9b02914
Org. Lett. XXXX, XXX, XXX−XXX