Synthetic studies concerning the crinine alkaloid haemultine

Article abstract of DOI:10.1071/CH12473

The racemic form, (±)-1, of the structure originally assigned to the crinine alkaloid haemultine has been prepared for the first time. A key step involved the conversion of compound (±)-4 into the isomeric cis-C3a-arylhexahydroindole (±)-3 using a Pd0-cat

Full text of DOI:10.1071/CH12473

CSIRO PUBLISHING
Aust. J. Chem. 2013, 66, 30–39
Full Paper
http://dx.doi.org/10.1071/CH12473
Synthetic Studies Concerning the Crinine Alkaloid
Haemultine
A
A
A
A
Nadia (Yuqian) Gao, Xinghua Ma, Laurent Petit, Brett D. Schwartz,
A B
,
A
A
Martin G. Banwell,
Anthony C. Willis, Ian A. Cade,
A
and A. David Rae
A
Research School of Chemistry, Institute of Advanced Studies, The Australian National
University, Canberra, ACT 0200, Australia.
B
Corresponding author. Email: mgb@rsc.anu.edu.au
The racemic form, (ꢀ)-1, of the structure originally assigned to the crinine alkaloid haemultine has been prepared for the
first time. A key step involved the conversion of compound (ꢀ)-4 into the isomeric cis-C3a-arylhexahydroindole (ꢀ)-3
using a Pd0-catalysed intramolecular Alder-ene reaction. The amino-alcohol (ꢀ)-2 derived from the latter compound
reacted with paraformaldehyde in the presence of trifluoroacetic acid to give, via a Pictet–Spengler reaction, the target
(ꢀ)-1. The diastereoisomeric Mosher esters 15 and 16 obtained by coupling the racemate (ꢀ)-1 with the R-form, 14, of the
Mosher acid could be separated chromatographically and then reductively cleaved to give the enantiomerically pure
compounds (þ)-1 and (ꢁ)-1, respectively. The physical and spectroscopic data derived from the former enantiomer are
consistent with the proposition that the title natural product is, in fact, a mixture of (þ)-1 and its D2,3-double bond isomer.
Manuscript received: 15 October 2012.
Manuscript accepted: 9 November 2012.
Published online: 14 January 2013.
1
However, neither H nor 13C NMR spectroscopic data were
reported for their material. Rather, this group simply stated that
the compound was obtained as colourless needles with a melting
range of 170–1728C and that ‘its spectral data are similar with
those reported’ by Fales and Wildman.[4]
Introduction
The colourful bush lily Haemanthus multiflorus Martyn is
encountered in various parts of Africa and certain extracts of it
have been used in traditional medicine in this region for cen-
turies.[1] In 1958 Boit and Do¨pke reported[2] the isolation of the
natural product haemultine (mp 174–1758C, [a]D þ147) from
those extracts of the plant considered responsible for the
observed pharmacological effects.[3] On the basis of various
analytical and chemical correlation studies they assigned
structure 1 (Fig. 1, no absolute stereochemistry implied) to
haemultine and thereby categorised it as a member of the crinine
family of alkaloids, albeit the only one lacking an oxygenated
D-ring and, therefore, not fitting well with established
biogenetic theories (Fig. 1).[4]
Independent work carried out around the same time[5] sug-
gested that the chemical correlation studies employed by Boit
and Do¨pke might have resulted in various undetected isomer-
isation processes and thus raised the possibility that the structure
of the title alkaloid had been incorrectly assigned. Accordingly,
in 1961 Fales and Wildman[4] reinvestigated matters and con-
cluded that the semi-synthetic sample of haemultine that Boit
and Do¨pke had obtained during the course of their correlation
studies was, in fact, a mixture of compound 1 (mp 194–194.58C,
[a]D þ10) and its D2,3-double bond isomer (mp 201–2028C, [a]D
þ199). They also examined the extracts of two different samples
of Haemanthus multiflorus and could not detect haemultine in
either of them. This raised questions about the existence of the
title compound as a natural product. In the late 1980s a group
lead by Itokawa reported[6] the isolation of haemultine from
samples of Haemanthus multiflorus Martyn collected in Egypt.
The rather unusual oxygenation pattern associated with
haemultine[4] together with the lack of certainty about its
existence as a natural product and the ambiguities surrounding
its structure[4,7] suggests that carrying out synthetic studies in the
area would be a worthwhile endeavour. Rather surprisingly, no
such work appears to have been undertaken.[8,9] Accordingly,
the main objective of the research described here was to develop
unambiguous total syntheses of both enantiomeric forms of
the structure, 1, originally assigned to haemultine and to do so
using methodology recently developed within our group.[10,11]
A broader objective of such studies was to undertake various
biological evaluations of these and other synthetically derived
crinine-type systems for the purposes of establishing a compre-
hensive structure–activity relationship profile within the class.
The retrosynthetic analysis employed in developing synthe-
ses of targets (þ)-1 and (ꢁ)-1 is shown in Fig. 2. A key feature of
the approach was to involve resolution of the racemic form of
compound 1, something considered achievable by either frac-
tional crystallisation of the diastereoisomeric salts formed from
reaction with chiral acids, or through chromatographic separa-
tion of the derived Mosher esters. The racemic material was to
be prepared by subjecting the cis-C3a-arylhexahydroindole
(ꢀ)-2 to a Pictet–Spengler reaction so as to simultaneously
establish the C6-methylene and B-ring of target 1. Compound
(ꢀ)-2 would, in turn, be prepared using a new method for
Journal compilation Ó CSIRO 2013
www.publish.csiro.au/journals/ajc

Synthetic Studies on the Alkaloid Haemultine
31
2
HO
3
AgOCN
t-BuOH
D
H
1,4-dioxane
1,4-dioxan
79 %
O
O
Br
Br
C
A
Br Br
NH
NCO
(Ϯ)
N
Boc
6
5
-
6
(Ϯ)-7
B
1
(HO)2B
Fig. 1. Structure originally assigned to haemultine (1).
O
O
(i) TFA, CH2Cl2
ϩ
(ii) NsCl, DMAP
87 %
Br
NH
OH
Pd0
Na2CO3
77 %
Ns
H
O
9
(Ϯ)-8
N
O
OH
H
(ϩ)-1
Resolution
O
O
O
O
O
(i) NaH, DMF
ϩ
N
(ii) 1-bromo-2-butyne
86 %
NH
Ns
N
O
OH
Ns
(Ϯ)
-1
H
O
O
(Ϯ)
-
10
(Ϯ)-4
N
Pictet–Spengler
reaction
Scheme 1. DMAP, 4-(N,N-dimethylamino)pyridine; TFA, trifluoroacetic
(Ϫ)-1
acid.
Oxidative cleavage/
deprotection/
reduction
OH
H
was achieved using the reaction sequence shown in Scheme 1
and as originally established by Banwell and coworkers.[10]
Thus, the cyclopropane 5, which is readily available in multi-
gram quantities through addition of dibromocarbene to cyclo-
pentene,[13] was treated with silver cyanate thus inducing
electrocyclic ring-opening of the three-membered ring to give
an allylic cation that was trapped, in situ, to form the allylic
isocyanate (ꢀ)-6.[14] This last species was itself trapped by
added t-butanol to give the carbamate (ꢀ)-7[14] in 79 % yield.
Since carbamate-protected amines were considered likely to
prevent appropriate orbital overlap of the reacting groups in the
foreshadowed IMAE reaction,[10] compound (ꢀ)-7 was depro-
tected using trifluoroacetic acid (TFA) in dichloromethane. The
resulting primary amine was then treated with nosyl chloride in
the presence of 4-(N,N-dimethylamino)pyridine and triethyl-
amine thereby generating the sulfonamide (ꢀ)-8[10] in 87 %
yield. Suzuki–Miyaura cross-coupling[15] of compound (ꢀ)-8
with the commercially available aryl boronic acid 9 gave the
arylated cyclohexene (ꢀ)-10 (77 %) that was treated with
sodium hydride followed by 1-bromo-2-butyne in DMF to
provide substrate (ꢀ)-4[10] (86 %) required for the pivotal IMAE
reaction. All the spectroscopic data obtained on the reaction
products shown in Scheme 1 were in accord with the assigned
structures and matched those reported previously.[10]
O
O
O
HN
N
O
Ns
(Ϯ)
-
3
(Ϯ)-2
Pd0-catalysed
IMAE reaction
Ns ϭ Nosyl
Established
chemistry
O
O
N
Ns
Br Br
(Ϯ)
-
4
5
Fig. 2. The retrosynthetic analysis employed in developing syntheses of
targets (þ)-1 and (ꢁ)-1. IMAE, intramolecular Alder-ene.
generating cis-C3a-arylhexahydroindoles developed by
Banwell and coworkers,[10] specifically through the Pd-catalysed
intramolecular Alder-ene (IMAE) reaction of the previously
reported,[10] N-protected and propargylated 1-amino-2-aryl-2-
cyclohexene (ꢀ)-4 followed by selective oxidative cleavage of
the exocyclic double bond within product (ꢀ)-3. Compound
(ꢀ)-4 itself was to be generated from the ring-fused gem-
dibromocyclopropane 5 using established procedures as
detailed in the following section.[12]
Completion of the Synthesis of Racemic Haemultine [(ꢀ)-1]:
Implementation of the IMAE and Pictet–Spengler Reactions
With compound (ꢀ)-4 to hand the pivotal IMAE reaction was
examined. In the event, and in keeping with earlier observa-
tions,[10] upon subjecting this compound to reaction with
Pd(OAc)2 and the strongly s-donating and bidentate ligand
N,N0-bis(benzylidene)ethylenediamine (BBEDA) in refluxing
benzene,[16] the required and previously reported[10] cis-
C3a-arylhexahydroindole (ꢀ)-3 was obtained as a white, crys-
talline solid in 94 % yield (Scheme 2).
Results and Discussion
Synthesis of the Substrate Required for the IMAE Reaction
The synthesis of the N-protected and propargylated 1-amino-2-
aryl-2-cyclohexene (ꢀ)-4 required for the pivotal IMAE process

32
N. Gao et al.
Pd(OAc)2
BBEDA, C6H6
O
O
O
O
N
N
80ЊC, 7 h
94 %
C6
C7
Ns
Ns
C5
C4
(Ϯ)
-4
(Ϯ)-3
C19
C8
C18
(i) O3, CH2Cl2, Ϫ78ЊC, 90 s
(ii) DMS, CH2Cl2, Ϫ78ЊC, 0.25h
C9
C11
O10
or
N1
C17
O16
(i) K2OsO4⋅2H O, NMO, citric acid
2
C3
C12
O22
O21
O31
38 %
(brsm)
CH3CN, 18ЊC, 30 h
C2
S20
(ii) Phl(OAc)2, CH2Cl2, 18ЊC, 2 h
C13
O
O30
C23
C28
C15
O
N29
O
O
O14
O
O
N
C24
HN
Ns
(Ϯ)-11
PhSH, CsCO3
CH3CN
(Ϯ)
-12 (61 %)
C25
ϩ
C27
18ЊC, 3 h
0
O
C26
O
O
N
Fig. 3. ORTEP diagram derived from the single-crystal X-ray analysis of
the ketone (ꢀ)-11. Thermal ellipsoids are drawn at the 30 % probability
level. H atoms are shown as spheres of arbitrary radius.
(Ϯ)
-13 (19 %)
NaBH4, CH3OH
18ЊC, 1
h
92 %
HO
HO
accord with the assigned structure but final confirmation of this
was secured by single-crystal X-ray analysis. The derived
ORTEP diagram is shown in Fig. 3 while further details are
provided in the Experimental section.
H
O
H
O
Paraformaldehyde
N
HN
O
TFA, DCE
60ЊC, 18h
95 %
O
(Ϯ)-1
While various possible orderings of the steps required to
effect the conversion (ꢀ)-11 - (ꢀ)-1 can be envisaged, several
exploratory studies led us to conclude that the one shown in
Scheme 2 would be the most effective. Accordingly, the ketone
(ꢀ)-11 was subjected to reaction with thiophenol in the presence
of caesium carbonate, a reagent combination defined by
Fukuyama et al.[18] for the cleavage of nosyl groups under mild
conditions. As a consequence the deprotected and somewhat
unstable amino-ketone (ꢀ)-12 (61 %) was obtained together
with its chromatographically separable dehydro-derivative
(ꢀ)-13 (19 %). In an alternate approach, treatment of compound
(ꢀ)-11 with magnesium in methanol gave azaenone (ꢀ)-13
exclusively in 81–87 % yield. Exposure of compounds (ꢀ)-12
and (ꢀ)-13, either separately or as a mixture, to sodium
borohydride in methanol at 188C afforded the diastereoisome-
rically pure and b-configured alcohol (ꢀ)-2 in 96 % yield. The
stereochemical outcome of this reduction process was presumed
to be as illustrated on the basis that hydride would be delivered
preferentially to the exo-face of the C-3 carbonyl group within
the substrate. Confirmation of this was established by a single-
crystal X-ray analysis of the final product (ꢀ)-1 (see below).
Successive treatment of the amino-alcohol (ꢀ)-2 with paraform-
aldehyde followed by TFA resulted in a Pictet–Spengler
reaction. As a consequence the racemic modification of the
structure, 1, assigned to haemultine, was obtained as a crystal-
line solid and in near quantitative yield.
(Ϯ)-2
Scheme 2. BBEDA, N,N0-bis(benzylidene)ethylenediamine; DMS,
dimethyl sulphide; NMO, N-methylmorpholine N-oxide; TFA, trifluoroa-
cetic acid; DCE, 1,2-dichloroethane.
The selective oxidative cleavage of the exocyclic double
bond within compound (ꢀ)-3 was a necessary step associated
with obtaining the substrate, (ꢀ)-2, required for the Pictet–
Spengler reaction. Accordingly, various methods for carrying
out this conversion were explored on the basis that some
selectivity should be observed in this type of transformation
because the trisubstituted and exocyclic double bond should be
the more nucleophilic of the two within substrate (ꢀ)-3. In
the event, the desired selectivity was rather difficult to achieve.
Under some of the best conditions identified thus far, and
involving exposure of substrate (ꢀ)-3 to ozone for 90 s at
ꢁ788C followed by rapid reductive work-up with dimethyl
sulfide, the target and crystalline ketone (ꢀ)-11 could be
obtained in 38 % yield based on recovered starting material
(brsm). Significant quantities of the ketodialdehyde arising from
cleavage of both double bonds with substrate (ꢀ)-3 seemed to be
formed as a result of the ozonolysis but attempts to convert this
very unstable material back into compound (ꢀ)-11 by exposing
it to the McMurry reagent[17] proved fruitless. The application of
conditions used by us on a related substrate,[9c] namely
K2OsO4ꢂ2H2O and N-methylmorpholine N-oxide (NMO) in
the presence of citric acid and then treating the resulting diols
with PhI(OAc)2, gave similar yields of ketone (ꢀ)-11. The
spectroscopic data derived from ketone (ꢀ)-11 were in complete
The 1H and 13C NMR spectroscopic data obtained on
compound (ꢀ)-1 are presented in Table 1 and these were in
complete accord with the assigned structure. In particular, the
former spectrum displayed two one-proton singlets arising from
the isolated protons of the aromatic ring, two mutually coupled

Synthetic Studies on the Alkaloid Haemultine
33
Table 1. 1H and 13C NMR spectroscopic data for ( )-haemultine [( )-1] and the mixture of L-(1)-tartrate salts of (1)- and (2)-haemultine [(1)-1
and (2)-1]
1H NMR [dH]
13C NMR [dC]
Synthetic (ꢀ)-1A
L-(þ)-Tartrate salts of (þ)-1 and (ꢁ)-1B
Synthetic (ꢀ)-1C
L-(þ)-Tartrate salts of (þ)-1 and (ꢁ)-1D
6.83, s, 1H
7.03, s, 1H
146.4
146.0
136.2
134.6
126.8
122.3
106.8
103.3
100.8
79.9
66.5
63.0
61.4
49.9
24.9
23.7
–
147.2 (C)
146.5 (C)
135.1 (C)
130.8 (C)
123.3 (C)
122.6 (C)
107.5 (C)
104.2 (C)
101.5 (CH2)
77.3 (CH)
66.9 (CH2)
61.2 (CH2)
58.4 (CH)
50.2 (C)
6.45, s, 1H
6.74, s, 1H
6.28, m, 1H
6.22, d, J 10.1, 1H
6.16, d, J 10.1, 1H
6.02, dd, J 10.1 and 5.7, 1H
5.88, m, 2H
5.97, s, 2H
4.30, d, J 16.7, 1H
4.50, d, J 16.0, 1H
3.90, m, 1H
4.03, d, J 16.0, 1H
3.70, d, J 16.7, 1H
3.91, m, 1H
3.30, m, 2H
3.67, dd, J 13.5 and 6.9, 1H
3.10, dd, J 12.4 and 4.6, 1H
3.45, d, J 12.8, 1H and 3.34, d, J 12.8, 1H
2.30–2.22, complex m, 2H
2.21–2.10, complex m, 2H
2.21–2.00, complex m, 2H
2.11–2.02, complex m, 2H
1.73, m, 1H
1.82, s, 1H
2.48, s, 2HE
–
–
–
–
–
–
–
–
–
24.0 (CH2)
21.6 (CH2)
174.2 (C)F
72.4 (CH)G
–
ARecorded in CDCl3 at 500 MHz.
BRecorded in (CD3)2SO at 400 MHz.
CRecorded in CDCl3 at 125 MHz.
DRecorded in (CD3)2SO at 100 MHz.
ESignal due to methine protons of tartrate residue.
FSignal due to carbonyl carbons of tartrate residue.
GSignal due to methine carbons of tartrate residue.
for a definite assignment of the relative configuration at
the hydroxy-bearing carbon. Accordingly, and given its crystal-
line form, the Pictet–Spengler product was subjected to single
crystal X-ray analysis. The derived ORTEP diagram is shown in
Fig. 4 while other details are presented in the Experimental
section. This analysis clearly demonstrates that the hydroxy
group attached to the C-ring sits over the D-ring rather than the
B-ring as seen in structure, 1, assigned to haemultine.
C15
O20
C14
O8
C16
C17
C7
C9
C4
C6
C1
C11
C18
C5
C19
C13
Resolution of (ꢀ)-1: Isolation and Characterisation
of the (1)- and (2)-Forms of Haemultine
O10
C12
N2
C3
With the racemic form, (ꢀ)-1, of the structure assigned to
haemultine in hand, methods for resolution into its constituent
enantiomers could be investigated. In a first attempt, compound
(ꢀ)-1 was reacted with stoichiometric amounts of L-(þ)-tartaric
acid in order to form the corresponding pair of diastereoisomeric
salts. The 1H and 13C NMR spectroscopic data obtained on the
mixture so formed (see Table 1) displayed a single set of signals
rather than the two that might have been expected given the
formation of diastereoisomers. Nevertheless, the successful
formation of the desired salts seems clear from, for example,
a comparison of the 13C NMR spectra of the free base and the
corresponding salts. Specifically, the differences in chemical
shifts due to the carbons surrounding the (protonated or unpro-
tonated) nitrogen support this proposition. However, despite
various attempts to do so, the salts could not be separated from
one another.
Fig. 4. ORTEP diagram derived from the single-crystal X-ray analysis of
(ꢀ)-haemultine [(ꢀ)-1]. Thermal ellipsoids are drawn at the 50 % probability
level. H atoms are shown as spheres of arbitrary radius.
(J 10.1) one-proton multiplets due to the non-equivalent olefinic
protons, and an AB spin-system (J 16.7) due to the geminally
related methylene protons of the C-ring. The {1H}13C NMR
spectrum showed the expected seventeen signals including eight
distinctly downfield of the resonances due to CDCl3 and arising
from the sp2-hybridised carbons of the aromatic ring or of the
double bond. The infrared spectrum showed no particularly
distinguishing features while the 70 eV electron impact (EI)
mass spectrum displayed a molecular ion at m/z 271 together
with a significant daughter ion at m/z 227 [corresponding to the
loss of H2C¼C(H)OH]. An accurate mass measurement on the
molecular ion established that it was of the expected composi-
tion, that is C16H17NO3. However, none of these data allowed
As a result of the difficulties detailed above, the reaction
sequence shown in Scheme 3 was used to provide compounds
(þ)-1 and (ꢁ)-1 in enantiomerically pure form. Thus, in the first
step of this simple sequence, the racemic modification of
compound 1 was reacted with the (R)-form, 14, of the Mosher

34
N. Gao et al.
OH
acid in the presence of 2,4,6-trichlorobenzoyl chloride and
triethylamine. After 1 h the reaction mixture was treated with
4-(N,N-dimethylamino)pyridine (DMAP) in toluene. The ensu-
ing mixture of esters 15 and 16 could be cleanly separated from
one another by flash chromatography and each was then
characterised by the usual methods.
H
O
O
HO2C
CF3
ϩ
N
Ph OMe
(Ϯ)
-
1
14
A comparison of the 1H and 13C NMR spectra derived from
the diastereoisomeric Mosher esters 15 and 16 is presented in
Table 2 and this reveals, generally speaking, only rather subtle
differences between the two sets of data. The most notable and
useful ones were encountered in the lower field regions of the
1H NMR spectra. In particular, the signals due to the olefinic
protons in compound 16 appear at lower field and are less well
differentiated than those due to their counterparts in isomer 15
(see bold entries in Table 2).
On the basis that the significant conformation of the Mosher
ester residue within each of compounds 15 and 16 is that in
which the C(H)–O–C(¼O)–C–CF3 ensemble is co-planar[19]
then, as illustrated in Fig. 5, it is the former diastereoisomer that
would experience greater deshielding (and differentiation) of
the olefinic protons. This is because these protons within
diastereoisomer 15 are most readily exposed to the anisotropic
magnetic shielding effect of the phenyl ring.
(i) 2,4,6-Cl3C6H2COCl, Et3N, THF
18ЊC, 1.5 h
(i) DMAP, toluene, 90–105ЊC, 5 h
O
CF3
O
OH
OMe
Ph
H
H
O
O
O
O
DIBAI-H
N
N
CH2CI2
18ЊC, 3 h
52%
(ϩ)-1
15 (39%)
ϩ
O
CF3
O
OMe
OH
Ph
H
H
In order to confirm the above-mentioned assignments of the
structures of compounds 15 and 16 each of them was subjected
to estercleavageusing diisobutylaluminium hydride(DIBAl-H),
thereby producing alcohols (þ)-1 (52 %) and (ꢁ)-1 (60 %),
respectively. The latter compounds were independently con-
verted into the corresponding 2,5-dibromobenzoates and a
single-crystal X-ray analysis was then undertaken on each, so
O
O
O
O
DIBAI-H
N
CH2CI2
18ЊC, 1.5h
60%
N
16 (44%)
(Ϫ)-1
Scheme 3. DMAP, 4-(N,N-dimethylamino)pyridine; DIBAl-H, diisobu-
tylaluminium hydride.
Table 2. 1H and 13C NMR spectroscopic data for Mosher esters 15 and 16
(Bold entries are signals due to the olefinic protons in compound 16 that appear at lower field and are less well differentiated than those due to their counterparts
(also in bold) in isomer 15)
1H NMR [dH]
Compound 16B
13C NMR [dC]
Compound 15A
Compound 15C
165.6
Compound 16C
165.6
7.48, m, 2H
7.55, m, 2H
7.42, m, 3H
7.40, m, 3H
146.8
146.8
6.96, s, 1H
6.93, s, 1H
146.6
146.5
6.57, s, 1H
6.48, s, 1H
134.6
134.8
6.08, d, J 5.8, 1H
6.17, d, J 10.2, 1H
132.3
132.1
5.91, s, 2H
5.94, m, 1H
132.2
132.0
5.78, m, 1H
5.91, s, 2H
129.5
129.5
5.15, m, 1H
5.08, m, 1H
128.3
128.3
4.38, d, J 15.7, 1H
4.33, d, J 16.0, 1H
127.3
127.4
3.75, d, J 15.7, 1H
3.70, d, J 16.0, 1H
126.3
126.7
–
3.93, dd, J 14.0 and 7.0, 1H
123.3 (CF3 quartet)
123.3 (CF3 quartet)
3.52, s, 3H
3.51, s, 3H
121.7
122.4
3.42, m, 2H
3.28, dd, J 14.0 and 3.5, 1H
106.6
106.6
3.11, t, J 9.4, 1H
3.10, m, 1H
104.1
104.0
2.00, m, 2H
2.05, m, 2H
101.0
100.9
1.65, m, 2H
1.65, m, 2H
84.1 (C–CF3 quartet)
84.5 (C–CF3 quartet)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
82.3
66.9
61.1
59.7
55.3
49.0
24.5
23.1
82.9
67.0
61.2
60.3
55.3
48.6
29.7
23.2
ARecorded in CDCl3 at 300 MHz.
BRecorded in CDCl3 at 500 MHz.
CRecorded in CDCl3 at 125 MHz.

Synthetic Studies on the Alkaloid Haemultine
35
for the purposes of resolving the racemic forms of such com-
pounds into their constituent enantiomers is also noteworthy.
Further applications of these protocols in natural products syn-
thesis will be reported in due course as will the outcomes of the
biological evaluations of compounds (þ)-1 and (ꢁ)-1.
Experimental
General Experimental Procedures
1H and 13C NMR spectra were recorded on a Varian Gemini
machine operating at 300 or 75MHz, respectively. Unless oth-
erwise specified, spectra were acquired at 208C in deutero-
chloroform (CDCl3) that had been stored over anhydrous sodium
carbonate. Chemical shifts are recorded as d values in parts per
million (ppm). Infrared spectra (nmax) were normally recorded on
a Perkin–Elmer 1800 Series FTIR Spectrometer and samples
were analysed as thin films on KBr plates (for liquids) or as a KBr
disk (for solids). Low-resolution electrospray ionisation (ESI)
mass spectra were recorded in positive-ion mode on a Micromass–
Waters LC-ZMD single quadrupole liquid chromatograph–
mass spectrometer while low- and high-resolution EI mass
spectra were recorded on a Fisons VG AUTOSPEC instrument.
Melting points were measured on a Stanford Research Systems
Optimelt-Automated Melting Point System and are uncorrected.
Optical rotations were measured at the sodium-D line (l 589 nm)
between 17 and 208C and at the concentrations (c, in g per
100 mL) indicated using spectroscopic grade chloroform
(CHCl3) as solvent. Analytical TLC was performed on aluminium-
backed 0.2 mm thick silica gel 60 F254 plates as supplied by
Merck. Eluted plates were visualised using a 254 nm UV lamp
and/or by treatment with a suitable dip followed by heating.
These dips included a mixture of vanillin/sulfuric acid/ethanol
(1g : 1 g : 18mL) or phosphomolybdic acid/ceric sulfate/sulfuric
acid (conc.)/water (37.5 g : 7.5 g : 37.5g : 720mL). The retarda-
tion factor (RF) values cited here have been rounded to the first
decimal point. Flash chromatographic separations were carried
out following protocols defined by Still et al.[20] with silica gel 60
(40–63 mm) as the stationary phase and using the AR- or HPLC-
grade solvents indicated. Starting materials and reagents were
generally available from the Sigma–Aldrich, Merck, TCI, Strem,
or Lancaster Chemical Companies and were either used as sup-
plied or, in the case of liquids, distilled when required. Drying
agents and other inorganic salts were purchased from the AJAX,
BDH, or Unilab Chemical Companies. THF, dichloromethane,
acetonitrile, and benzene were dried using a Glass Contour
solvent purification system that is based upon a technology
originally described by Grubbs et al.[21] Spectroscopic grade
solvents were used for all analyses. Where necessary, reactions
were performed under a nitrogen or argon atmosphere.
Fig. 5. Molecular model of compound 15 generated using Spartan 10
(MMFF94-derived geometries) with the ester residue and associated oxy-
methine proton (top centre) constrained in the required s-trans and coplanar
array and showing the consequent proximity of the olefinic hydrogens
(bottom centre) to the shielding zone of the phenyl group (bottom right).
firmly establishing their absolute configurations. Details are
presented in the Experimental section and the Supplementary
Material. Spectroscopic analyses of the alcohols themselves
revealed that the (þ)-form was obtained from precursor 15 while
its enantiomer was obtained from isomer 16. The circular
dichroism spectra recorded on each of the product enantiomers
were essentially mirror images of one another (see Supplemen-
tary Material) while the melting points and signs of the specific
rotations were consistent with those obtained by Fales and
Wildman[4] from their synthetically derived sample of the
former material. The discrepancy between the magnitudes of
the specific rotation of our sample of (þ)-1 [a]D þ75 (c ¼ 0.1,
CHCl3) and that reported for the material obtained by demethox-
ylation of haemanthamine[4] {[a]D þ10 (c ¼ 0.19, solvent not
specified)} may be attributed to the differences in solvent used
for obtaining the optical rotation data. Accordingly, we agree
with the suggestion[4] that the sample of haemultine obtained by
Boit and Do¨pke is likely to be a mixture of compound (þ)-1 and
its D2,3-double bond isomer. A similar interpretation could be
applied to the results of the study carried out by Itokawa and
coworkers.[6]
Specific Chemical Transformations
Conclusions
Compound (ꢀ)-11
Method A: A magnetically stirred solution of diene (ꢀ)-3[10]
(50 mg, 0.11 mmol) in dichloromethane/methanol (10 mL of a
1 : 1 v/v mixture) was cooled to ꢁ788C then sparged with ozone
generated using a 500 Model Fischer portable ozone generator.
After 90 s the stream of ozone was replaced with nitrogen and
the reaction mixture allowed to warm to 08C and then treated
with dimethyl sulfide (0.2 mL, 2.7 mmol). After a further 0.5 h
the reaction mixture was concentrated under reduced pressure
and the residue thus obtained subjected to flash chromatography
(silica, 35 : 65 v/v ethyl acetate/hexane elution) to afford two
fractions, A and B.
The studies reported here lend weight to the suggestion that the
purported natural product haemultine may in fact be, as sug-
gested by Fales and Wilman,[4] a mixture of compound (þ)-1 and
its D2,3-double bond isomer. However, the only way to determine
if this is truly so would be to reisolate the natural product (if this
can be done) and then subject it to analysis using modern spec-
troscopic techniques. Despite these residual ambiguities, the
present work demonstrates the utility of the IMAE/Pictet–
Spengler reaction sequence outlined in Fig. 2 as a means for
preparing crinine alkaloids, especially those bearing an oxygen
functionality in the C-ring. The capacity to use such functionality

36
N. Gao et al.
Concentration of fraction A (RF 0.5 in 1 : 1 v/v ethyl acetate/
The column used in the chromatographic separation men-
tioned immediately above was stripped with methanol and the
filtrate so obtained was concentrated under reduced pressure to
give a light-brown oil. This was subjected to flash chromatog-
raphy (silica, 0.3 : 99.7 v/v ammonia saturated methanol/
dichloromethane elution) to afford, after concentration of the
appropriate fractions (RF 0.8 in 1 : 1 v/v ethyl acetate/hexane),
the title compound (ꢀ)-12 (38 mg, 61 %) as a brown oil (Found:
[M þ H]þ, 258.1132. C15H15NO3 requires: [M þ H]þ,
258.1130]. dH (400 MHz) 6.74 (d, J 8.1, 1H), 6.62 (m, 2H),
6.10 (m, 1H), 5.90 (s, 2H), 5.46 (d, J 9.9, 1H), 3.67 (d, J 18.6,
1H), 3.60 (broad s, 1H), 3.42 (d, J 18.6, 1H), 2.36 (br s, 1H), 2.17
(m, 2H), 1.84 (m, 1H), 1.70 (m, 1H). dC (100 MHz) 218.1, 147.8,
146.6, 135.3, 131.1, 124.0, 121.4, 108.4, 108.2, 101.1, 63.7,
58.7, 55.1, 20.5, 19.7. nmax (KBr)/cmꢁ1 2921, 1742, 1503, 1488,
1440, 1243, 1038, 933, 811. m/z (ESI, þve) 270 (23 %,
[M þ Na]þ), 258 (100, [M þ H]þ), 240 (62).
hexane) afforded the title compound (ꢀ)-11 (16 mg, 38 % brsm)
as white crystals, mp 167–1698C (Found: Mþꢃ, 442.0837.
C21H18N2O7S requires: Mþꢃ, 442.0835). dH (300 MHz) 7.88
(d, J 8.0, 1H), 7.69 (m, 1H), 7.59 (m, 2H), 6.61–6.43 (complex
m, 3H), 6.12 (dt, J 10.0 and 4.0, 1H), 5.91 (m, 2H), 5.56 (d, J
10.0, 1H), 4.39 (t, J 6.4, 1H), 4.16 (ABq, J 16.0, 2H), 2.13 (m,
2H), 1.95 (m, 2H). dC (125 MHz) 208.3, 148.1, 147.9, 147.1,
133.8, 133.1, 131.7, 131.6, 131.3, 130.7, 124.9, 124.1, 120.8,
108.2, 107.8, 101.2, 65.2, 60.4, 53.2, 24.7, 21.5. nmax (KBr)/
cmꢁ1 2917, 1757, 1543, 1505, 1486, 1437, 1371, 1243, 1162,
1126, 1068, 1038, 912. m/z (EI, 70 eV) 442 (10 %, Mþꢃ), 255 (7),
201 (20), 200 (100), 141 (15).
Concentration of fraction B (RF 0.7 in 1 : 1 v/v ethyl acetate/
hexane) gave the starting diene (ꢀ)-3 (6 mg, 12 % recovery) that
was identical, in all respects, with an authentic sample.
Method B: A magnetically stirred solution of diene (ꢀ)-3
(321 mg, 0.71 mmol) in acetonitrile/water (15 mL of a 4 : 1 v/v
mixture) maintained at 188C was treated with K2OsO4ꢂ2H2O
(39 mg, 0.11 mmol), citric acid (1.4 g, 7.29 mmol), and NMO
(579 mg, 4.94 mmol). After 30 h the reaction mixture was
diluted with ethyl acetate (20 mL) followed by NH4Cl (50 mL
of a saturated aqueous solution), and the separated aqueous
phase was extracted with ethyl acetate (3 ꢄ 15 mL). The com-
bined organic phases were filtered through a short pad of TLC-
grade silica gel and the filtrate concentrated under reduced
pressure. The residue thus obtained was dissolved in dichlor-
omethane and then treated, at 188C and with magnetic stirring,
with PhI(OAc)2 (273 mg, 0.848 mmol). After 2 h the reaction
mixture was filtered and the filtrate concentrated under
reduced pressure to give a light-yellow oil that was subjected
to flash chromatography (silica, 3 : 5 v/v ethyl acetate/hexane
elution). Concentration of the relevant fractions (RF 0.5 in
4 : 2.5 : 5.5 v/v/v ethyl acetate/dichloromethane/hexane) gave
compound (ꢀ)-11 (120 mg, 39 %) as a white, crystalline solid
that was identical in all respects with the material produced by
Method A.
Method B: Magnesium (91 mg, 3.74 g atom) was added to a
magnetically stirred solution of compound (ꢀ)-11 (166 mg,
0.38 mmol) in methanol (10 mL) maintained at 188C. After 2 h
the reaction mixture was filtered through a short pad of TLC-
grade silica gel and the filtrate concentrated under reduced
pressure. The residue thus obtained was subjected to flash
chromatography (silica, 3 : 5 v/v ethyl acetate/hexane elution)
to give, after concentration of the appropriate fractions (RF 0.5 in
2 : 2.5 : 5.5 v/v/v ethyl acetate/dichloromethane/hexane), com-
pound (ꢀ)-13 (80 mg, 83 %) as a light-yellow oil that was
identical, in all respects, with the material obtained by Method
A as detailed above.
Compound (ꢀ)-2
Method A: A magnetically stirred solution of compounds
(ꢀ)-12 and (ꢀ)-13 (50 mg, ,0.19 mmol of a ,3 : 1 mixture
obtained as described immediately above) in THF/methanol
(5 mL of a 1 : 1 v/v mixture) maintained at 08C was treated with
NaBH4 (22 mg, 0.58 mmol). The resulting mixture was allowed
to warm to 188C over 1 h and then treated with NH4Cl (10 mL of
a saturated aqueous solution) and extracted with dichloro-
methane (3 ꢄ 10 mL). The combined organic phases were dried
(MgSO4), filtered, and concentrated under reduced pressure to
give a light-yellow oil. Subjection of this material to flash
chromatography (silica, 15 : 85 v/v ammonia saturated metha-
nol/dichloromethane elution) afforded, after concentration of
the relevant fractions (RF 0.5), the title compound (ꢀ)-2 (46 mg,
92 %) as a white, crystalline solid, mp 169–1718C (Found:
[M ꢁ Hꢃ]þ, 258.1135. C15H17NO3 requires: [M ꢁ Hꢃ]þ,
258.1130). dH (500 MHz) 6.89 (d, J 1.9, 1H), 6.82 (dd, J 8.2
and 1.9, 1H), 6.75 (d, J 8.2, 1H), 6.13 (m, 1H), 5.91 (s, 2H), 5.85
(d, J 10.4, 1H), 4.63 (t, J 7.1, 1H), 3.38 (m, 1H), 3.28 (m, 1H),
2.90 (m, 1H), 2.17 (m, 1H), 2.06 (m, 1H), 1.80 (br s, 2H) 1.70
(m, 1H), 1.60 (m, 1H). dC (125 MHz) 147.8, 146.0, 138.5, 130.5,
126.1, 120.1, 108.1, 107.6, 101.0, 79.5, 63.0, 52.7, 22.9, 20.1
(signal due to one carbon obscured or overlapping). nmax (KBr)/
cmꢁ1 3434, 3285, 3033, 2933, 2877, 1502, 1486, 1431, 1228,
1114, 1099, 1039, 937, 899, 873, 804. m/z (EI, 70 eV) 258 (1 %,
[M ꢁ Hꢃ]þ), 241 (5, [M ꢁ H2O]þꢃ), 201 (23), 200 (100).
Method B: NaBH4 (16 mg, 3.743 mmol) was added to a
magnetically stirred solution of compound (ꢀ)-13 (56 mg,
0.22 mmol) in methanol (5 mL) maintained at 188C. After 1 h
the reaction mixture was concentrated under reduced pressure
and the light-yellow oil thus obtained was subjected to flash
chromatography (silica, ammonia-saturated methanol elution)
Compounds (ꢀ)-12 and (ꢀ)-13
Method A: A magnetically stirred solution of ketone (ꢀ)-11
(108 mg, 0.24 mmol) in acetonitrile (5 mL) maintained under a
nitrogen atmosphere at 08C was treated with thiophenol (74 mL,
0.73 mmol) and caesium carbonate (258 mg, 0.79 mmol). The
ensuing mixture was allowed to warm to 188C, stirred at this
temperature for 1 h, and then treated with NH4Cl (10 mL of a
saturated aqueous solution) and extracted with CH2Cl2
(3 ꢄ 10 mL). The combined organic fractions were washed with
NH4Cl (1 ꢄ 10 mL of a saturated aqueous solution) and brine
(1 ꢄ 10 mL) before being dried (MgSO4), filtered, and concen-
trated under reduced pressure. The residue thus obtained was
subjected to flash chromatography (silica, 1 : 4 v/v ethyl acetate/
hexane elution) and concentration of the appropriate fractions
(RF 0.8 in 1 : 1 v/v ethyl acetate/hexane) afforded the title
compound (ꢀ)-13 (12 mg, 19 %) as a light-brown oil (Found:
M
þꢃ, 255.0892. C15H13NO3 requires: Mþꢃ, 255.0895). dH
(500 MHz) 8.04 (s, 1H), 6.79 (d, J 8.6, 1H), 6.61 (m, 2H),
6.11 (m, 1H), 5.95 (s, 2H), 5.39 (d, J 9.8, 1H), 4.46 (br s, 1H),
2.38 (m, 1H), 2.14 (m, 1H), 1.89 (m, 2H). dC (125 MHz) 201.8,
164.9, 148.0, 146.8, 133.9, 132.9, 123.7, 121.0, 108.5, 108.0,
101.2, 77.3, 56.3, 23.1, 19.7. nmax (KBr)/cmꢁ1 2917, 1735, 1504,
1489, 1439, 1246, 1234, 1210, 1039, 932. m/z (EI, 70 eV) 255
(60 %, Mþꢃ), 200 (100), 141 (30).

Synthetic Studies on the Alkaloid Haemultine
37
to give, after concentration of the appropriate fractions (RF 0.4 in
1 : 9 v/v ammonia-saturated methanol/methanol), compound
(ꢀ)-2 (55 mg, 96 %) as a light-yellow, crystalline solid that
was identical, in all respects, with the material obtained by
Method A detailed immediately above.
1743, 1503, 1482, 1265, 1239, 1170, 1039, 1030, 938, 737. m/z
(EI, 70 eV) 487 (100 %, Mþꢃ), 254 (45), 189 (43).
Concentration of fraction B (RF 0.5 in 7 : 2 : 1 v/v/v ethyl
acetate/hexane/methanol) afforded the title compound 15
(63 mg, 39 %) as a clear, colourless oil that solidified on
standing, [a]D þ32 (c 0.2, CHCl3) (Found: Mþꢃ, 487.1605.
C26H24F3NO5 requires: Mþꢃ, 487.1607). dH (300 MHz) see
Table 2. dC (125 MHz) see Table 2. nmax (KBr)/cmꢁ1 2948,
2915, 1745, 1482, 1270, 1240, 1169, 1121, 1038, 1026, 847,
733, 714. m/z (EI, 70 eV) 487 (100 %, Mþꢃ), 254 (57), 189 (52).
Compound (ꢀ)-1
A magnetically stirred solution of compound (ꢀ)-2 (54 mg,
0.21 mmol) in 1,2-dichloroethane (10 mL) was treated with
paraformaldehyde (32 mg, 1.06 mmol) and TFA (320 mL,
4.15 mmol). The resulting solution was heated at 608C for 18 h
and then cooled and concentrated under reduced pressure. The
ensuing residue was subjected to flash chromatography (silica,
3 : 17 v/v ammonia-saturated methanol/methanol) to afford,
after concentration of the relevant fractions (RF 0.4 in
methanol), the title compound (ꢀ)-1 (53 mg, 95 %) as a
Compound (1)-1
DIBAl-H (1.1 mL of a 1.0 M solution in dichloromethane,
1.1 mmol) was added dropwise to a magnetically stirred solution
of Mosher ester 15 (55 mg, 0.11 mmol) in dichloromethane
(8 mL) maintained at 188C under a nitrogen atmosphere. Stirring
was continued for a further 3.0 h and then the reaction mixture
was treated with methanol (2 mL) and additional dichloro-
methane (10 mL) before being poured into NH4Cl (25 mL of a
saturated aqueous solution). The separated aqueous phase was
extracted with dichloromethane (2 ꢄ 10 mL) and the combined
organic phases were then dried (MgSO4), filtered, and concen-
trated under reduced pressure to give a light-yellow oil. Subjec-
tion of this material to flash chromatography (silica, 3 : 17 v/v
ammonia-saturated methanol/methanol elution) gave, after con-
centration of the appropriate fractions (RF 0.4 in methanol), the
title compound (þ)-1[4] (16 mg, 52 %) as a white, crystalline
solid, mp 195–1968C (lit.[4] mp 194–194.5), [a]D þ75 (c 0.1,
CHCl3) {lit.[4] [a]D þ10 (c 0.19, solvent not specified)}. The
1H NMR, 13C NMR, IR, and mass spectroscopic data recorded
on this material were identical with those reported above for
racemic haemultine [(ꢀ)-1)].
white, crystalline solid, mp 202–2058C (Found: Mþꢃ
,
271.1208. C16H17NO3 requires: Mþꢃ
,
271.1208). dH
(500 MHz) see Table 1. dC (125 MHz) see Table 1. nmax
(KBr)/cmꢁ1 2907, 1502, 1481, 1323, 1307, 1238, 1064, 1038,
935, 850. m/z (EI, 70 eV) 271 (76 %, Mþꢃ), 270 (50), 228 (53),
227 (100), 213 (56).
L-(þ)-Tartrate Salts of (þ)- and (ꢁ)-Haemultine
A magnetically stirred solution of compound (ꢀ)-1 (10.0 mg,
0.04 mmol) in acetonitrile (2 mL) was treated with L-(þ)-tartaric
acid (5.5 mg, 0.04 mmol) and the resulting mixture heated at
508C until all the solid had dissolved and then treated with water
(,0.2 mL) and allowed to cool. The resulting solid was removed
by filtration to give a mixture of the title salts (15.5 mg, 100 %)
as clear, colourless crystals, mp 121–1258C (Found: [M þ H]þ,
272.1287; C16H17NO3ꢂC4H6O6 requires [M þ H]þ, 272.1287).
dH (400 MHz) see Table 1. dC (100 MHz) see Table 1. nmax
(KBr)/cmꢁ1 3422, 2914, 1662, 1489, 1354, 1301, 1254, 1115,
1082, 841, 688. m/z (ESI, þve) 272 (100 %, [M þ H]þ).
Compound (ꢁ)-1
Treatment of the Mosher ester 16 (54 mg, 0.11 mmol) with
DIBAl-H (550 mL of a 1.0 M solution in dichloromethane,
0.55 mmol) in essentially the same manner as described above
for the conversion 15 - (þ)-1 but using a reaction time of 1.5 h
afforded, after work-up and flash chromatography, compound
(2)-1 (18 mg, 60 %) as a white, crystalline solid, mp 195–
Compounds 15 and 16
2,4,6-Trichlorobenzoyl chloride (78 mL, 0.50 mmol) was
added to a magnetically stirred solution of (R)-(þ)-MTPA-OH
(14) (125 mg, 0.53 mmol) and triethylamine (230 mL,
1.65 mmol) in THF (4 mL) maintained at 188C under a nitrogen
atmosphere. After 1.5 h the reaction mixture was diluted with
toluene (5 mL) and the solution thus obtained added by a syringe
pump over 4 h to a magnetically stirred solution of compound
(ꢀ)-1 (90 mg, 0.33 mmol) and DMAP (405 mg, 3.32 mmol) in
toluene (10 mL) maintained at 908C under a nitrogen atmo-
sphere. The ensuing mixture was stirred at 1058C for 1 h and
then cooled and quenched with NaHCO3 (50 mL of a saturated
aqueous solution). The separated aqueous phase was extracted
with ethyl acetate (2 ꢄ 10 mL) and the combined organic phases
were washed with NH4Cl (50 mL of a saturated aqueous solu-
tion) before being dried (MgSO4), filtered, and concentrated
under reduced pressure. Subjection of the yellow oil thus
obtained to flash chromatography (silica, 1 : 20 : 20 -
1 : 10 : 10 v/v/v methanol/ethyl acetate/hexane gradient elution)
afforded two fractions, A and B.
1
1978C, [a]D ꢁ75 (c 0.25, CHCl3). The H NMR, 13C NMR,
IR, and mass spectroscopic data recorded on this material were
identical with those reported above for racemic haemultine
[(ꢀ)-1)].
2,5-Dibromobenzoate of Compound (þ)-1
A magnetically stirred solution of compound (þ)-1 (5.3 mg,
0.02 mmol) and 2,5-dibromobenzoic acid (11 mg, 0.04 mmol) in
dichloromethanewastreatedwithN,N0-dicyclohexylcarbodiimide
(DCC, 12 mg, 0.06 mmol) and DMAP (11 mg, 0.06 mmol).
After 16 h the reaction mixture was diluted with dichloro-
methane (10 mL) and the solution thus obtained treated with
NH4Cl (20 mL of a saturated aqueous solution). The separated
aqueous phase was extracted with dichloromethane (2 ꢄ 10 mL)
and the combined organic phases were then dried (MgSO4),
filtered, and concentrated under reduced pressure to give a light-
yellow oil. Subjection of this material to flash chromatography
(silica, 2 : 25 : 25 v/v/v methanol/ethyl acetate/hexane elution)
gave, after concentration of the appropriate fractions (RF 0.5 in
6 : 3 : 1 v/v/v ethyl acetate/hexane/methanol), the title ester
(7.9 mg, 77 %), as a white, crystalline solid, mp 162–1648C.
Concentration of fraction A (RF 0.6 in 7 : 2 : 1 v/v/v ethyl
acetate/hexane/methanol) afforded the title compound 16
(72 mg, 44 %) as a clear, colourless oil that solidified on
standing, [a]D ꢁ32 (c 0.1, CHCl3) (Found: Mþꢃ, 487.1606.
C26H24F3NO5 requires: Mþꢃ, 487.1607). dH (500 MHz) see
Table 2. dC (125 MHz) see Table 2. nmax (KBr)/cmꢁ1 2917,
(Found: Mþꢃ, 532.9664. C23H1979Br81BrNO4 requires Mþꢃ
,

38
N. Gao et al.
532.9660.) dH (600 MHz) 7.83 (d, J 3.0, 1H), 7.51 (d, J 8.4, 1H),
7.44 (dd, J 8.4 and 3.0, 1H), 6.96 (s, 1H), 6.52 (s, 1H), 6.20 (dm,
J 10.2, 1H), 6.03 (m, 1H), 5.93 (m, 2H), 5.23 (m, 1H), 4.39 (d, J
16.8, 1H), 3.82 (d, J 16.8, 1H), 3.61 (d, J 13.8, 1H), 3.51 (dd, J
13.8 and 7.2, 1H), 3.24 (m, 1H), 2.28 (m, 1H), 2.21 (m, 1H), 2.01
(m, 1H), 1.90 (m, 1H). dC (150 MHz) 163.7, 147.0, 146.7, 135.8,
135.6, 134.6, 134.1, 133.4, 131.8, 122.6, 121.0, 120.6, 106.7,
104.1, 101.0, 81.4, 67.0, 60.9, 59.5, 49.4, 29.7, 23.4, 22.7. nmax
(KBr)/cmꢁ1 2917, 2849, 1730, 1483, 1456, 1282, 1234, 1112,
1095, 1042, 1027, 1006, 940. m/z (EI, 70 eV) 535, 533 and 531
(13, 26, and 13 %, all Mþꢃ), 270 (100), 253 (47), 97 (61).
(ꢀ)-1 and (ꢀ)-11 were refined using the CRYSTALS program
package.[24] The crystallographic asymmetric unit for each of
the 2,5-dibromobenzoate derivatives of compounds (þ)-1 and
(ꢁ)-1 contain six distinct molecules that vary in terms of the
orientation of the benzoate residue with respect to the ABCD-
ring system of the alkaloid framework. These molecules are
arranged into pairs related by a non-crystallographic two-fold
screw axis and the pairs related to one another by approximate
z/3 translations. The data are twinned and the structures show
pseudo-symmetry requiring the use of the specialist structure
refinement program RAELS.[25] Atomic coordinates, bond
lengths and angles, and displacement parameters have been
deposited at the Cambridge Crystallographic Data Centre
(CCDC nos. 828020, 828019, 860930, and 860929 for com-
pounds (ꢀ)-1, (ꢀ)-11, and the 2,5-dibromobenzoate of com-
pounds (þ)-1 and (ꢁ)-1, respectively). These data can be
obtained free-of-charge from the Cambridge Crystallographic
Data Centre 12 Union Road, Cambridge CB2 1EZ, UK via
www.ccdc.cam.ac.uk/data_request/cif.
2,5-Dibromobenzoate of Compound (ꢁ)-1
Reaction of compound (ꢁ)-1 (8.4 mg, 0.03 mmol) with
2,5-dibrombenzoic acid (17 mg, 0.06 mmol) in the presence of
DCC (19 mg, 0.09 mmol) and DMAP (11 mg, 0.06 mmol) in the
same manner as described above for enantiomer (þ)-1 afforded,
after work-up and chromatographic purification, the title ester
(13 mg, 79 %) as a white, crystalline solid, mp 162–1648C. The
1H NMR, 13C NMR, IR, and mass spectrometric data recorded
on this material were identical with those reported above for its
enantiomer.
Supplementary Material
The X-ray crystal structures for the 2,5-dibromobenzoate
of compounds (þ)-1 and (ꢁ)-1, 1H and 13C NMR spectra
of compounds (ꢀ)-1, (ꢀ)-2, (ꢀ)-11 to (ꢀ)-13, the L-(þ)-
tartrate salts of (þ)- and (ꢁ)-haemultine, 15, 16, and the 2,5-
dibromobenzoate of compound (þ)-1, and circular dichroism
spectra derived from (þ)-1 and (ꢁ)-1 are available on the
Journal’s website.
Crystallographic Studies
Data for Compound (ꢀ)-1
C16H17NO3, M 271.32, T 200 K, monoclinic, space group
˚
P21/n, Z 4, a 7.5599(2), b 21.6876(5), c 8.0256(2) A, b 96.1285
3
(16)8; V 1308.33(6) A , Dx 1.377 g cmꢁ3, 2314 unique data
˚
(2ymax 508), R 0.035 [for 1835 reflections with I . 2.0s(I)];
Rw 0.087 (all data),
dichloromethane.
S 0.96. Crystals grown from
Acknowledgements
The authors thank the Australian Research Council and the Institute of
Advanced Studies for generous financial support.
Data for Compound (ꢀ)-11
C21H18N2O7S, M 442.45, T 200 K, monoclinic, space group
˚
P21/n, Z 4, a 12.8018(2), b 11.8557(3), c 13.7369(3) A, b
References
112.4650(12)8; V 1926.69(7) A , Dx 1.525 g cmꢁ3, 4387 unique
3
˚
[1] S. C. Chhabra, B. L. A. Mahunnah, E. N. Mshiu, J. Ethnopharmacol.
1987, 21, 253. doi:10.1016/0378-8741(87)90103-6
[2] H.-G. Boit, W. Do¨pke, Chem. Ber. 1958, 91, 1965. doi:10.1002/CBER.
19580910928
data (2ymax 558), R 0.035 [for 3903 reflections with I . 2.0s(I)];
Rw 0.094 (all data), S 0.99. Crystals grown from dichloro-
methane/hexane.
[3] M. M. Iwu, Handbook of African Medicinal Plants, 1993 (CRC Press:
Boca Raton, FL) and references therein.
Data for the 2,5-Dibromobenzoate of Compound (þ)-1
C23H19Br2NO4, M 533.22, T 200 K, monoclinic, space group
[4] H. M. Fales, W. C. Wildman, J. Org. Chem. 1961, 26, 1617.
doi:10.1021/JO01064A076
[5] H. M. Fales, W. C. Wildman, J. Am. Chem. Soc. 1958, 80, 4395.
doi:10.1021/JA01549A073
˚
P21, Z 12, a 7.4967(3), b 36.8043(12), c 21.8414(7) A, b
3
90.5106(9)8; V 6026.0(4) A , Dx 1.763 g cmꢁ3, 20816 unique
˚
data (2ymax 508), R 0.059 [for 11928 reflections with I . 3.0s(I)];
Rw 0.068 [I .3.0s(I)], S 1.54. Crystals grown from chloroform/
ethanol/ethyl acetate.
[6] O. M. Abdallah, A. A. Ali, H. Itokawa, Phytochemistry 1989, 28, 3248.
doi:10.1016/0031-9422(89)80322-X
[7] Both enantiomeric forms of the crinine alkaloid framework are
encountered in nature. See, for example: C.-K. Chen, F.-H. Lin,
L.-H. Tseng, C.-L. Jiang, S.-S. Lee, J. Nat. Prod. 2011, 74, 411.
doi:10.1021/NP100819N
[8] (a) For representative examples of crinine alkaloid syntheses reported
over the preceding three decades or so see: L. E. Overman, S. Sugai,
Helv. Chim. Acta 1985, 68, 745. doi:10.1002/HLCA.19850680324
(b) W. H. Pearson, F. E. Lovering, Tetrahedron Lett. 1994, 35, 9173.
doi:10.1016/0040-4039(94)88457-9
Data for the 2,5-Dibromobenzoate of Compound (ꢁ)-1
C23H19Br2NO4, M 533.22, T 200 K, monoclinic, space group
˚
P21, Z 12, a 7.4853(3), b 36.7672(16), c 21.8086(3) A, b
3
90.4044(15)8; V 6001.9(4) A , Dx 1.770 g cmꢁ3, 26662 unique
˚
data (2ymax 55.28), R 0.070 [for 15459 reflections with I. 3.0s(I)];
Rw 0.081 [I .3.0s(I)], S 1.88. Crystals grown from chloroform/
methanol/ethyl acetate.
(c) C. Bru, C. Guillou, Tetrahedron 2006, 62, 9043. doi:10.1016/
J.TET.2006.07.005
(d) M. Bohno, K. Sugie, H. Imase, Y. B. Yusof, T. Oishi, N. Chida,
Tetrahedron 2007, 63, 6977. doi:10.1016/J.TET.2007.05.041
(e) N. T. Tam, J. Chang, E.-J. Jung, C.-G. Cho, J. Org. Chem. 2008, 73,
6258. doi:10.1021/JO8008353
(f) J.-D. Liu, S.-H. Wang, F.-M. Zhang, Y.-Q. Tu, Y.-Q. Zhang, Synlett
2009, 3040 and references therein.
Structure Determination
Images were measured on a Nonius Kappa CCD diffracto-
˚
meter (MoKa, graphite monochromator, l 0.71073 A) and data
extracted using the DENZO package.[22] Structure solution was
by direct methods (SIR92).[23] The structures of compounds

Synthetic Studies on the Alkaloid Haemultine
39
[9] (a) Various syntheses of crinine alkaloids have been reported by our
group see: M. G. Banwell, J. E. Harvey, K. A. Jolliffe, J. Chem. Soc.,
Perkin Trans. 1 2001, 2002. doi:10.1039/B106369N
(b) A. D. Findlay, M. G. Banwell, Org. Lett. 2009, 11, 3160.
doi:10.1021/OL901230W
[16] B. M. Trost, C. Pedregal, J. Am. Chem. Soc. 1992, 114, 7292.
doi:10.1021/JA00044A049
[17] A. R. Carroll, W. C. Taylor, Aust. J. Chem. 1990, 43, 1439 and
references cited therein. doi:10.1071/CH9901439
[18] T. Fukuyama, C.-K. Jow, M. Cheung, Tetrahedron Lett. 1995, 36,
(c) L. Petit, M. G. Banwell, A. C. Willis, Org. Lett. 2011, 13, 5800.
doi:10.1021/OL2023938
[10] A. L. Lehmann, A. C. Willis, M. G. Banwell, Aust. J. Chem. 2010, 63,
6373. doi:10.1016/0040-4039(95)01316-A
[19] T. R. Hoye, C. S. Jeffrey, F. Shao, Nat. Protoc. 2007, 2, 2451.
doi:10.1038/NPROT.2007.354
1665. doi:10.1071/CH10359
[20] W. C. Still, M. Kahn, A. Mitra, J. Org. Chem. 1978, 43, 2923.
doi:10.1021/JO00408A041
[11] For a summary of the various alkaloid syntheses that have been
undertaken within our group see: M. G. Banwell, N. (Y.) Gao, B. D.
Schwartz, L. V. White, Top. Curr. Chem. 2012, 309, 163. doi:10.1007/
128_2011_217
[12] (a) For discussions on the application of gem-dihalocyclopropanes
in natural product synthesis see: M. G. Banwell, D. A. S. Beck,
P. C. Stanislawski, M. O. Sydnes, R. M. Taylor, Curr. Org. Chem.
2005, 9, 1589. doi:10.2174/138527205774370469
(b) M. G. Banwell, A. L. Lehmann, R. S. Menon, A. C. Willis, Pure
Appl. Chem. 2011, 83, 411. doi:10.1351/PAC-CON-10-10-21
[13] J. Sonnenberg, S. Winstein, J. Org. Chem. 1962, 27, 748. doi:10.1021/
JO01050A012
[14] M. G. Banwell, C. J. Cowden, Aust. J. Chem. 1994, 47, 2235.
doi:10.1071/CH9942235
[21] A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen, F. J.
Timmers, Organometallics 1996, 15, 1518. doi:10.1021/OM9503712
[22] DENZO–SMN. Z. Otwinowski, W. Minor, Processing of X-ray diffrac-
tion data collected in oscillation mode, in Methods in Enzymology,
Volume 276: MacromolecularCrystallography, Part A (EdsC.W.Carter,
Jr, R. M. Sweet) 1997, pp. 307–326 (Academic Press: New York, NY).
[23] SIR92. A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi,
M. C. Burla, G. Polidori, M. Camalli, J. Appl. Cryst. 1994, 27, 435.
[24] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J. Watkin,
J. Appl. Cryst. 2003, 36, 1487. doi:10.1107/S0021889803021800
[25] A. D. Rae, RAELS06, 2006 (The Australian National University:
Canberra).
[15] N. Miyaura, T. Yanagi, A. Suzuki, Synth. Commun. 1981, 11, 513.
doi:10.1080/00397918108063618