Reactions of Aminopentadienal Derivatives with 5,6-Dihydropyridinium Salts as an Approach to Manzamine Alkaloids Based upon Biogenetic Considerations

Article abstract of DOI:10.1021/jo990604p

Recent results from our laboratory have prompted us to introduce a modification of the original Baldwin proposal for the biosynthesis of manzamine alkaloids. In this new model, aminopentadienal derivatives of general structures 2 or 3 can not only cyclize to give pyridinium salts such as cyclostellettamines but also react with 5,6-dihydropyridinium salts 1 leading to species such as 6 or 7, which would be key intermediates in the syntheses of manzamine A and halicyclamine B. On the basis of the chemistry depicted in Scheme 2, model reactions of dienes 9, 10, 12, and 13 with salts 1 have been investigated. Diene 9 gave an interesting rearrangement product 15a, while dienes 10 and 12 gave halicyclamine- and manzamine-type adducts 19 and 23, respectively. In addition, glutaconaldehyde derivative 13 gave adduct 32 possessing some characteristic features of sarain A.

Full text of DOI:10.1021/jo990604p

J . Org. Chem. 1999, 64, 7381-7387
7381
Rea ction s of Am in op en ta d ien a l Der iva tives w ith
5,6-Dih yd r op yr id in iu m Sa lts a s a n Ap p r oa ch to Ma n za m in e
Alk a loid s Ba sed u p on Biogen etic Con sid er a tion s
Karine J akubowicz, Kamel Ben Abdeljelil, Matthias Herdemann, Marie-The´re`se Martin,
Alice Gateau-Olesker, Ali Al Mourabit, Christian Marazano,* and Bhupesh C. Das
Institut de Chimie des Substances Naturelles, CNRS, 91198 Gif-sur-Yvette Cedex, France
Received April 8, 1999
Recent results from our laboratory have prompted us to introduce a modification of the original
Baldwin proposal for the biosynthesis of manzamine alkaloids. In this new model, aminopentadienal
derivatives of general structures 2 or 3 can not only cyclize to give pyridinium salts such as
cyclostellettamines but also react with 5,6-dihydropyridinium salts 1 leading to species such as 6
or 7, which would be key intermediates in the syntheses of manzamine A and halicyclamine B. On
the basis of the chemistry depicted in Scheme 2, model reactions of dienes 9, 10, 12, and 13 with
salts 1 have been investigated. Diene 9 gave an interesting rearrangement product 15a , while
dienes 10 and 12 gave halicyclamine- and manzamine-type adducts 19 and 23, respectively. In
addition, glutaconaldehyde derivative 13 gave adduct 32 possessing some characteristic features
of sarain A.
In tr od u ction
Manzamine A (Figure 1), isolated1 from a sponge
(Haliclona sp.) collected off Manzamo, Okinawa, has been
the subject of much synthetic effort during the past few
years.2 Interest in this field not only arose from the
challenging structure and biological activity of manz-
amine A itself but also from the recent discovery of a
large number of related natural alkaloids isolated from
sponges in the same order (Haplosclerida),3 some repre-
sentative members of which are depicted in Figure 1.
That these apparently different structures are bioge-
netically closely related was postulated by Baldwin and
Whitehead,4,5 who suggested that manzamine A and
keramaphidin B can arise from the intramolecular cy-
clization of a macrocyclic dihydropyridine intermediate.
Central to this hypothesis is the chemistry of dihydro-
pyridinium salts 1, which could result from the conden-
sation of an aldehyde and a primary amine with acrolein
(Scheme 1).
(1) Sakai, R.; Higa, T.; J efford, C. W.; Bernardinelli, G. J . Am. Chem.
Soc. 1986, 108, 6404-6405.
(2) For a recent review on synthetic approaches to manzamines,
see: Magnier, E.; Langlois, Y. Tetrahedron 1998, 54, 6201-6258. Total
synthesis of manzamine A: (a) Winkler, J . D.; Axten, J . M. J . Am.
Chem. Soc. 1998, 120, 6425-6426. (b) Martin, S. F.; Humphrey, J .
M.; Ali, A.; Hillier, M. C. J . Am. Chem. Soc. 1999, 121, 866-867.
(3) For comprehensive reviews, see: (a) Matzanke, N.; Gregg, R. J .;
Weinreb, S. M. Org. Prep. Proc. Int. 1998, 30, 3-51. (b) Andersen, R.
F igu r e 1. Representative members of manzamine alkaloids.
J .; Van Soest, R. W. M.; Kong, F. Alkaloids: Chemical and Biological
Perspectives; Pelletier, S. W., Ed.; Pergamon Press: Elsevier Science
1996; Vol. 10, pp 301-355. (c) Tsuda, M.; Kobayashi, J . Heterocycles
1997, 46, 765-794. (d) Crews, P.; Cheng, X.-C.; Adamczeski, M.;
Rodriguez, J .; J aspar, M.; Schmitz, F. J .; Traeger, S. C.; Pordesimo,
E. O. Tetrahedron 1994, 50, 13567-13574.
Recent results from our laboratory prompted us to
introduce a modification of the original Baldwin pro-
posal.6 In our new model, condensation of malonaldehyde
with an aldehyde and a primary amine would produce
key aminopentadienal derivatives (two possible regio-
isomers 2 and 3) that are very likely to cyclize in an acidic
(4) Baldwin, J . E.; Whitehead, R. C. Tetrahedron Lett. 1992, 33,
2059-2062.
(5) For model reactions leading to keramaphidin B and halicycla-
mine A skeletons, see: (a) Baldwin, J . E.; Bischoff, L.; Claridge, T. D.
W.; Heupel, F. A.; Spring, D. R.; Whitehead, R. Tetrahedron 1997, 53,
2271-2290 and references therein. (b) Gil, L.; Baucherel, X.; Martin,
M.-T.; Marazano, C.; Das, B. C. Tetrahedron Lett. 1995, 36, 6231-
6234.
(6) Kaiser, A.; Billot, X.; Gateau-Olesker, A.; Marazano, C.; Das, B.
C. J . Am. Chem. Soc. 1998, 120, 0, 8026-8034.
10.1021/jo990604p CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/09/1999

7382 J . Org. Chem., Vol. 64, No. 20, 1999
J akubowicz et al.
Sch em e 1. Tw o P ossible Biogen etic P a th w a ys to
Six-m em ber ed Nitr ogen Heter ocycles Sta r tin g
fr om Acr olein 4 or Ma lon d ia ld eh yd e6
Sch em e 3. Syn th esis of Am in op en ta d ien a l
Der iva tives or Th eir Equ iva len tsa
Sch em e 2. P r op osed Biogen etic P a th w a ys to
Ma n za m in e A a n d Ha licyla m in e A
a
Key: (a) n-BuNH2 (2 equiv); (b) NaOH, H2O; (c) (i) NaH, THF;
(ii) HCO2H, Et2O; (d) n-BuNH2, CH2Cl2, Al2O3; (e) (i) n-BuLi, THF,
-20 °C; (ii) H+ (pH 3); (f) (i) Dowex 50WX8; (ii) NaOH.
the syntheses of manzamine A and halicyclamine B,
while reductive cyclization of 6 would give keramaphidin
B.
In this paper, we report details of a series of model
reactions intended to test the chemistry summarized in
Scheme 2.
Resu lts a n d Discu ssion
A. Syn th esis of Am in op en ta d ien a l Der iva tives or
Th eir Equ iva len ts. Since dihydropyridinium salts 1 are
readily available,8 the first problem to be addressed was
the regioselective synthesis of substituted aminopenta-
dienal derivatives corresponding to species 2 or 3.9 Ring
opening of Zincke salt 8 (Scheme 3) with 2 equiv of
n-butylamine gave the crystalline salt 9 in good yield. It
was anticipated that the hydrolysis of this intermediate
in the presence of sodium hydroxide would give a
separable mixture of the two aminopentadienal deriva-
tives 10 and 11. In fact, the immediate recording of the
1H NMR spectrum of the crude reaction mixture after
hydrolysis showed the initial formation of the two desired
isomers 10 and 11 in an 85:15 ratio. However, it was not
possible to isolate the minor isomer 11 since it readily
rearranged when chromatographed, or left in solution,
to give thermodynamically more stable isomer 10, which
was thus obtained as a single product in good yield. Since
the synthesis of 11 was not feasible by this approach,
we targeted the corresponding ester 12, the synthesis of
which was readily accomplished from commercial prod-
ucts via a Wittig-Horner reaction. We additionally
synthesized the sodium salt of 2-methylglutaconaldehyde
1310 also from commercially available precursors using
a Wittig reaction.
medium to give natural 3-alkylpyridinium salt deriva-
tives such as cyclostellettamines and pyridinium macro-
cyclic oligomers. These aminopentadienal derivatives can
not only produce pyridinium salts but also could alter-
natively condense with salts 1, this chemistry being at
the origin of our modified model6 for the construction of
manzamine A and halicyclamine B. This proposal is
summarized in Scheme 2. In this model, macrocycles such
as 4 and 5,7 both possessing dihydropyridinium and
aminopentadienal species, can react in two different ways
to give polycycles 6 and 7, respectively. These polycycles
could then be considered as advanced intermediates in
Nucleophilic additions of these new derivatives 9, 10,
12, and 13 to dihydropyridinium salts 1a were next
investigated.
(8) (a) Grierson, D. S.; Harris, M.; Husson, H.-P. J . Am. Chem. Soc.
1980, 102, 1064-1082. (b) For a review, see: Grierson, D. S. Org. React.
1990, 39, 85-295.
(9) For a review on the synthesis and reactions of glutaconaldehyde
and the corresponding amino derivatives (aminopentadienals), see:
Becher, I. Synthesis 1980, 589-612.
(10) For previous synthesis of glutaconaldehyde and derivatives, see
ref 8. For 2-alkyl derivatives, see also: Yanovskaya, L. A.; Kucherov,
V. F. Izv. Akad. Nauk. SSSR 1960, 2184; C. A.1961, 55, 14452. Free
glutacondialdehyde derivatives are very unstable intermediates (gluta-
conaldehyde is stable for only ∼30 min at ∼-70 °C).9
(7) Note that an aminopentadienal regioisomer of type 2 is required
for construction of the manzamine A skeleton, while both isomers of
type 2 and 3 can lead to the halicyclamine A skeleton (only the
regioisomer of type 3 is shown in Scheme 2).

Reactions of Aminopentadienal Derivatives
J . Org. Chem., Vol. 64, No. 20, 1999 7383
Sch em e 4a
Sch em e 5a
a
Key: (a) CH2Cl2, 64 h, 20 °C; (b) NaBH4, 2-propanol, 15% yield
for 15a .
B. Ad d ition of Dien e 9 to Sa lt 1a .11 The results of
the reaction of salt 1a 12 with diamino derivative 9 (free
base) are summarized in Scheme 4. The reaction was
monitored by GC-MS analysis of aliquots after reduction
with NaBH4. The major compounds detected were tet-
rahydropyridine 14 and a reduction product of 9 (struc-
ture not determined), which were accompanied by prod-
ucts seemingly corresponding to those arising from the
dimerization of salt 1a in the presence of a base.5 After
60 h, these products disappeared and a new adduct (15a )
was formed, which was recovered in 15% yield after
chromatography, along with major tetrahydropyridine 14
(37%). The structure of the adduct was established as
15a by an NMR study of the corresponding acetate
15b.13a A plausible mechanism for the formation of 15a
is shown in Scheme 4. Noteworthy is the fact that this
mechanism necessarily implies formation of the initial
adduct 16 resulting from the attack of one nitrogen of 9
on salt 1a . No product corresponding to the attack of a
nucleophilic carbon of 9, as required in our tentative
routes (Scheme 2), was detected. Albeit disappointing,
this result revealed some new interesting features of the
chemistry of reactants such as 1a and 9.
a
Key: (a) CH2Cl2, 12 h; (b) MeSO3H, H2O, MeOH, 50 °C;
(c) NaBH4, MeOH (19/20: 3/1 ratio, 23% yield from 1a ).
pyridinium salt 18, which was not isolated, but was
directly treated with an excess of NaBH4. GC-MS
analysis of the crude reaction mixture revealed the
formation of mainly two adducts 19 and 20 in a 3:1 ratio
along with traces of other minor diastereoisomers. These
two adducts were recovered in 23% overall yield, and
pure samples of each isomer were obtained by HPLC,
allowing structure resolution by NMR spectroscopy.13a
The structure 19 of the major isomer was confirmed
unambiguously by comparison with a related halicycla-
mine-type analogue resulting from previous model
reactions.5b This sequence thus represents an encourag-
ing result in favor of our proposed route to halicyclamine
A. The yields of 19 are consistently superior to those
obtained from our previous procedure,2b an added ad-
vantage being the differentiation of the two nitrogen
substituents. In addition, the minor adduct 20 can be
considered as a halicyclamine B14a model, albeit the
relative stereochemistry of the asymmetric center of the
tetrahydropyridine ring is inverted. We believe that this
epi structure is not of much significance since the
macrocyclic structure leading to the natural alkaloid can
C. Ha licycla m in e Mod el.11 Addition of aminopenta-
dienal 10 to salt 1a (Scheme 5) was more encouraging
since it proceeded smoothly in methylene chloride to form
a carbon-carbon bond, providing aminal 1713b that was
isolated in 55% yield. Treatment of this adduct with
methanesulfonic acid in water presumably gave the
(11) J akubowicz, K. Ph.D. Thesis in preparation.
(12) Pure salts 1a ,b were obtained from the corresponding methoxy
adduct according to our reported procedure: Gil, L.; Gateau-Olesker,
A.; Marazano, C.; Das, B. C. Tetrahedron Lett. 1995, 36, 707-710.
(13) (a) All complex structures were resolved by intensive NMR
spectroscopy studies including 1D and 2D NMR experiments (COSY
90, NOESY, HMQC, HMBC; see the Supporting Information). (b)
Structure 17 was confirmed by an X-ray analysis that will be reported
separately.
(14) (a) Harrisson, B.; Talapatra, S.; Lobkovsky, E.; Clardy, J .;
Crews, P. Tetrahedron Lett. 1996, 51, 9151-9154. (b) Charan, R. D.;
Garson, M. J .; Brereton, I. M.; Willis, A. C.; Hooper, J . N. A.
Tetrahedron 1996, 52, 9111-9120. (c) Clark, R. J .; Field, K. L.; Charan,
R. D.; Garson, M. J .; Brereton, I. M.; Willis, A. C. Tetrahedron 1998,
54, 8811-9126.

7384 J . Org. Chem., Vol. 64, No. 20, 1999
J akubowicz et al.
Sch em e 6. Ma n za m in e Mod ela
Sch em e 7
an equilibration process. Accordingly, the product 23 is
likely to be derived from the major primary adduct of the
cycloaddition reaction. This proposal is reinforced by the
fact that the intramolecular reduction of the very un-
stable iminium function19 probably takes place im-
mediately after the cycloaddition, thus preventing any
further equilibration. The inverted stereochemistry of the
exocyclic amino group in our experimental model (see 23)
when compared to the proposed intermediate 6 can be
explained by the same considerations. Thus, if the
intramolecular reduction process has “trapped” the ster-
eochemistry of the primary cycloadduct, further thermo-
dynamic equilibration of this center was not possible. The
intramolecular reduction might be avoided, for example,
by the temporary addition of a nucleophile to stabilize
the iminium function.19 If this is the case, it is conceivable
that the expected cycloadduct 25 (Scheme 7) could
rearrange to give the open intermediate 26, which would
then cyclize to give the thermodynamically more stable
iminium ion 27. Another isomerization mechanism would
be an elimination-addition process through diene 28.
Further work in this direction is now in progress.
a
Key: (a) CH2Cl2, 0-4 °C, 15 h; (b) Ac2O, pyridine (43% overall
for 23, 7% for 24).
significantly influence the stereochemistry of such reduc-
tion process. Pyridinium salts related to 18 are also likely
to be precursors of haliclonacyclamines A-C core skele-
tons.14b,c
D. Ma n za m in e Mod el.15 The cycloaddition reaction
of salt 1a with diene 12 (Scheme 6) proceeded smoothly
at 0-4 °C.16,17 The iminium adduct 21 was likely formed
first, but an unexpected intramolecular hydride transfer
occurred18 and the products of the reaction were esters
22 (resulting from the hydrolysis of the intermediate
imine), which were isolated as their acetyl derivatives
23 and 24. These two bicyclic adducts were isolated after
chromatography in 43% and 7% yield, respectively, and
their structure was elucidated by NMR spectroscopy.13a
No other adducts that could result from the reaction of
salt 1a with diene 12 were detected in significant
amounts as determined by GC-MS analysis of the crude
reaction mixture. The ease of this cycloaddition process
supports the model proposed in Scheme 2. In addition,
among the rather numerous Diels-Alder strategies2 for
the construction of manzamine A central AB rings, this
approach is certainly one of the most efficient since it
brings together all the important functionalities in a
single step.
E. An Ap p r oa ch to th e Sa r a in A Sk eleton .20
A
further series of results is depicted in Scheme 8. Methyl-
substituted glutaconaldehyde salt 13 adds to salt 1a in
a two-phase system (H2O/CH2Cl2), giving intermediate
29 that readily cyclized, affording a mixture of isomers
30 and 31 (approximately 5:95 ratio) in practically
quantitative yield. Filtration of this mixture through
silica gel using CH2Cl2 as eluent resulted in its complete
isomerization to give a 2:1 mixture of dialdehydes 32 and
33,13a which were isolated in 33% yield. This isomeriza-
tion also takes place on alumina with CH2Cl2 as eluent
but is slower. The relatively modest yield of the dialde-
hydes is probably due to competitive polymerization
processes. When the reaction of 1a with 13 was per-
formed in MeOH, formation of dialdehyde adducts 32 and
33 was also observed, but these products were isolated
in very low yield, again due to competing polymerization
processes. All data suggest that these relatively unstable
dialdehydes are in thermodynamic equilibrium and were
thus inseparable. Reduction of this mixture, followed by
acetylation of the resulting diols, gave two adducts that
were separated and characterized by NMR as 34 and 35
(1:2 ratio).13a The structures of the adducts 32 and 33
retain salient features of the sarain A core skeleton.
Adducts 32 and 33 thus possess the correct junction of
Separate treatment of the esters 23 and 24 with
t-BuOK resulted in each case in the recovery of a mixture
of these two esters in approximately a 1:1 ratio. This
result suggests that the stereochemistry of the ester
function in the major product 23 was not the result of
(15) Herdemann, M. Ph.D. Thesis in preparation.
(16) For the related cycloaddition of 1-methoxy-3-trimethylsilyloxy-
1,3-butadiene to aminonitrile derivatives of 1, see: Baldwin, J . E.;
Spring, D. R., Whitehead, R. C. Tetrahedron Lett. 1998, 39, 5417-
5420.
(17) Examples of push-pull dienes cycloadditions to classical di-
enophiles under high pressure: Branchadell, V.; Sodupe, M.; Ortuno,
R. M.; Oliva, A.; Gomez-Pardo, D.; Guinguant, A.; d’Angelo, J . J . Org.
Chem. 1991, 56, 4135-4141.
(18) For a discussion of analogous intramolecular reductions, see:
Cohen, T.; Onopchenko, A. J . Org. Chem. 1983, 48, 4531-4537. Note
that treatment of the reaction mixture with NaBH4 in MeOH prior to
acetylation gave the same results and is therefore unnecessary.
(19) Mitch, C. H. Tetrahedron Lett. 1988, 29, 6831-6834.
(20) Ben Abdeljelil, K. Ph.D. Thesis in preparation.

Reactions of Aminopentadienal Derivatives
J . Org. Chem., Vol. 64, No. 20, 1999 7385
Sch em e 8. An Ap p r oa ch to Sa r a in A Sk eleton a
Sch em e 9. Ma lon a ld eh yd e Biogen etic Scen a r io to
Sa r a in A
(AcOEt/heptane, gradient from 10:90 to 50:50) to afford pure
product 15a (325 mg, 15% yield) as a pale yellow oil. An
analytical sample of 15a was treated with pyridine and Ac2O
to give the corresponding acetate N-(6-ben zyl-2-bu tyl-8,10-
dim eth yl-2,6-diaza-tr icyclo[5.3.1.03,8u n dec-9-en -11-ylm eth -
1
yl)-N-bu tyla ceta m id e (15b): IR (film) 3024, 1649 cm-1; H
NMR (CDCl3, 400 MHz) δ 0.88 (t, J ) 7.2 Hz, 3 H), 0.97 (t,
J ) 7.4 Hz, 3 H), 1.24-1.38 (m, 7 H), 1.34 (s, 3 H), 1.55 (m, 2
H), 1.76 (d, J ) 11.5 Hz, 1 H), 1.82 (s, 3 H), 1.85 (m, 1 H), 2.04
(d, J ) 10.2 Hz, 1 H), 2.09 (s, 3 H), 2.18 (m, 1 H), 2.37 (m, 1
H), 2.44 (m, 1 H), 2.46 (m, 1 H), 2.66 (dt, J ) 3.5, 12.0 Hz, 1
H), 2.86 (dd, J ) 4.7, 13.7 Hz, 1 H), 3.07 (s, 1 H), 3.14 (dd,
J ) 9.6, 13.7 Hz, 1 H), 3.30 (m, 1 H), 3.36 (m, 1 H), 3.40 (d,
J ) 13.0 Hz, 1 H), 3.63 (d, J ) 13.0 Hz, 1 H), 5.57 (s, 1 H),
7.19-7.35 (m, 5 H); 13C NMR (CDCl3, 100 MHz) δ 14.0, 14.2,
20.3, 20.7, 21.5, 21.8, 22.3, 26.8, 30.7, 31.2, 35.6, 38.6, 42.4,
48.3, 48.9, 57.8, 59.8, 60.7, 61.2, 61.8, 126.7, 128.2 (2 C), 128.6,
129.7 (2 C), 139.9, 140.3, 170.5; MS (EI) m/z (rel intensity)
451 (M•+, 13), 360 (10), 164 (100), 91 (92); HRMS (CI) calcd
for C29H46N3O m/z 452.3641, found m/z 452.3631.
Ha licycla m in e Mod el. 3-(8-Ben zyl-2-bu tyl-9-m eth yl-
2,8-d ia za b icyclo[3.3.1]n on -3-en -4-yl)-2-m et h ylp r op en a l
(17). 5,6-Dihydropyridinium salt 1a (2.68 g, 7.5 mmol) was
slowly added to aminopentadienal 10 (1.03 g, 6.2 mmol) in dry
CH2Cl2 (30 mL). After the mixture was stirred for 12 h at room
temperature, the solvent was removed under reduced pressure.
The crude reaction mixture was chromatographed on a short
column of alumina using AcOEt/heptane (1:1) as eluent.
Aminal 17 (1.2 g, 55% yield) was obtained as a yellow oil. Pure
crystals of 17 were obtained from CH2Cl2/heptane: mp 135
°C; IR (CHCl3) 2964, 2871, 2818, 2705, 1649, 1563, 1197, 1144,
1084 cm-1; 1H NMR (300 MHz, CDCl3) δ 0.88 (t, J ) 14 Hz, 3
H), 1.26 (m, 2 H), 1.34 (m, 1 H), 1.36 (d, J ) 7 Hz, 3 H), 1.46
(dt, J ) 6, 7 Hz, 3 H), 1.85 (m, 1 H), 1.93 (s, 3 H), 2.15 (ddt,
J ) 4.5, 13, 13 Hz, 1 H), 2.33 (ddd, J ) 3.5, 12.5, 12.5 Hz, 3
H), 2.56 (dd, J ) 3.5, 11 Hz, 1 H), 2.91 (m, 1 H), 3.23 (m, 2 H),
3.56 (d, J ) 13.5 Hz, 1 H), 3.73 (d, J ) 13.5 Hz, 1 H), 3.85 (bs,
1 H), 6.55 (s, 1 H), 6.98 (s, 1 H), 7.21-7.37 (m, 5 H), 9.15 (s, 1
H); 13C NMR (75.47 MHz, CDCl3) δ 10.1, 13.8, 15.7, 19.9, 25.3,
32.1, 32.4, 33.7, 43.7, 55.7, 59.6, 74.5, 112.3, 124.1, 127.2, 128.4
(2C), 128.5 (2C), 138.8, 149.1, 153.8, 194.0. MS (EI) m/z (rel
intensity) 352 (M•+, 100), 91 (48). Anal. Calcd for C23H31N2O:
C, 77.59; H, 8.77; N, 7.86. Found: C, 77.72; H, 8.97; N, 7.83.
1′-Ben zyl-1-bu tyl-5,3′-d im eth yl-1,2,3,6,1′,2′,3′,4′,5′,6′-d ec-
a h yd r o[3,4′]bip er id in yl (19). Methanesulfonic acid (0.5 mL,
7.7 mmol) was added dropwise to a stirred suspension of
aminal 17 (1.2 g, 3.4 mmol) in a mixture of H2O (40 mL) and
MeOH (4 mL) at 50 °C. The reaction mixture was warmed to
50 °C and maintained for 3 h at this temperature. The heating
bath was then removed, and the crude mixture was slowly
added to a solution of NaBH4 (10 equiv) in MeOH at 0 °C and
then stirred for 16 h at room temperature. The reaction
mixture was quenched with a solution of 2 N aqueous HCl,
basified with 5% aqueous NaOH, and then extracted with CH2-
Cl2. The combined organic extracts were dried over MgSO4 and
concentrated in vacuo. The residue was chromatographed over
a
Key: (a) CH2Cl2, 20 °C, (30/31: 5/95 ratio, 95% yield); (b)
Al2O3, CH2Cl2 (32/33: 3/2 ratio, 33% yield); (c) (i) NaBH4, MeOH;
(ii) Ac2O, pyridine.
the two six-membered rings, the quaternary carbon
center, and carbons for the construction of the five-
membered nitrogen heterocycle.
Albeit incomplete, this model represents the first
experimental evidence supporting the well-accepted idea
that the biosyntheses of sarain A and manzamines are
closely related.3 In this context, one of the possible
biogenetic scenarios leading to sarain A from malonal-
dehyde chemistry (cf Scheme 1)6 is depicted in Scheme
9. Such chemistry would avoid the intermediacy of
dihydropyridinium and dihydropyridine species (our
previous proposal),12 whose chemistry is difficult to
control (vide infra).
Exp er im en ta l Section
Ad d ition of Dien e 9 to Sa lt 1a . To a solution of dihydro-
pyridinium salt 1a (1.92 g, 5.4 mmol, prepared from 1-benzyl-
4-methoxy-3-methyl-1,4,5,6-tetrahydropyridine and p-toluene-
sulfonic acid)12 in dry CH2Cl2 (20 mL) was added derivative 9
(free base, 1.2 g, 5.4 mmol) in dry CH2Cl2 (20 mL). The mixture
was stirred at room temperature for 64 h and then cooled to
-78 °C. 2-Propanol (30 mL) and NaBH4 (611 mg, 16 mmol)
were added. After the mixture was stirred for 4 h at room
temperature, an aqueous solution of 2 N HCl was added
dropwise under vigorous stirring. The resulting mixture was
basified with 2 N NaOH and the product extracted with CH2-
Cl2. Removal of solvent under reduced pressure left a crude
product, which was purified by chromatography over alumina

7386 J . Org. Chem., Vol. 64, No. 20, 1999
J akubowicz et al.
alumina using AcOEt/heptane (5:95) as eluent to give a
mixture of adducts 19 and 20 (261 mg, 23% yield) in a 3:1
ratio as calculated by GC analysis. Pure samples of each
adduct were obtained after HPLC on silica gel using a mixture
CH2Cl2/MeOH/Et3N (99:1:0.2) as eluent. Major adduct 19: IR
(film) 2956, 2932, 2874, 2801, 2755, 2357, 1660, 1455, 1377
(IE) m/z 384 (M•+), 326 (M•+ - NHCOCH3), 310 (C3H6O2), 186,
146 (C10H12N), 91 (C7H7).
An Ap p r oa ch to th e Sa r a in A Sk eleton . 3-(8-Ben zyl-
9-m eth yl-2-oxa -8-a za bicyclo[3.3.1]n on -3-en -4-yl)-2-m eth -
ylp r op en a l (31). 5,6-Dihydropyridinium salt 1a (1.2 g, 3.92
mmol), in CH2Cl2 (15 mL), was added dropwise to a solution
of salt 13 (432 mg, 3.24 mmol) in H2O (25 mL). After vigorous
stirring of the resulting biphasic mixture for 2 h at room
temperature, the organic phase was decanted and collected.
The remaining H2O phase was extracted with CH2Cl2. The
combined organic extracts were dried over MgSO4, and the
solvent was evaporated under reduced pressure to give an
unseparable mixture of adducts 30 and 31 (900 mg, 94% yield)
in a 5:95 ratio (determined by 1H NMR spectroscopy): MS (IC)
m/z 298 (MH+). Major adduct 27: 1H NMR (CDCl3, 300 MHz)
δ 1.25 (d, J ) 6.9 Hz, 3 H), 1.41 (m, 1 H), 1.92 (s, 3 H), 2.1 (m,
1 H), 2.19 (m, 1 H), 2.55 (dd, J ) 12.1, 4.1 Hz, 1 H), 2.68 (m,
1.H), 2.82 (m, 1H), 3.65 (d, J ) 13.7 Hz, 1 H), 3.92 (d, J )
10.3 Hz, 1 H), 4.75 (d, J ) 2.5 Hz, 1 H), 6.50 (s, 1 H), 7.30 (s,
1 H), 7.32 (m, 5 H), 9.29 (s, 1 H);13C NMR (CDCl3, 75 MHz) δ
10.3, 15.6, 23.4, 31.4, 32.1, 43.1, 58.9, 91.9, 117.4, 127.4, 128.5,
128.8, 130.8, 138.5, 150.5, 156.6, 195.2.
1
cm-1; H NMR (300 MHz, CDCl3) δ 0.85 (d, J ) 6 Hz, 3 H),
0.92 (t, J ) 7 Hz, 3 H), 1.04 (m, 1 H), 1.24-1.41 (m, 3 H),
1.42-1.54 (m, 3 H), 1.54-1.61 (m, 2 H), 1.64 (bs, 3 H), 1.85
(ddd, J ) 2.5, 11, 11 Hz, 1 H), 1.96 (dd, J ) 9, 9 Hz, 1 H), 2.37
(dt, J ) 2, 8 Hz, 2 H), 2.53 (bd, J ) 15 Hz, 1 H), 2.72 (m, 2 H),
2.80 (dd, J ) 1.5, 8 Hz, 1 H), 2.89 (m, 1 H), 3.00 (bd, J ) 15
Hz, 1 H), 3.46 (dd, J ) 12, 12 Hz, 2 H), 5.14 (bs, 1 H), 7.18-
7.40 (m, 5 H); 13C NMR (75 MHz, CDCl3) δ 14.2, 17.0, 20.9,
21.1, 26.3, 29.3, 32.8, 35.8, 46.2, 50.7, 54.7, 57.2, 58.6, 62.4,
63.4, 124.8, 126.9, 128.2 (2 C), 129.2 (2 C), 132.5, 138.6; MS
(EI) m/z (rel intensity) 340 (M+., 59), 249 (50), 186 (58), 91
(100); HRMS (CI) calcd for C23H37N2 (MH+) 341.2958, found
341.2956. Minor adduct 20: IR (film) 2956, 2932, 2874, 2801,
2756, 1652, 1455, 1377 cm-1; 1H NMR (400 MHz, C6D6) δ 0.80
(d, J ) 6.6 Hz, 3 H), 0.98 (t, J ) 7.4 Hz, 3 H), 1.01 (d, J ) 7.0
Hz, 3 H), 1.40 (m, 2 H), 1.47 (m, 1 H), 1.50 (m, 1 H), 1.56 (m,
2 H), 1.58 (m, 1 H), 1.80 (m, 1 H), 1.82 (m, 1 H), 1.87 (m, 1 H),
1.96 (dd, J ) 7.7, 10.8 Hz, 1 H), 2.40 (m, 2 H), 2.48 (m, 1 H),
2.70 (dd, J ) 5.6, 10.8 Hz, 1 H), 2.73 (bd, J ) 15.0 Hz, 1 H),
2.91 (bd, 2 H), 3.01 (bd, J ) 15.0 Hz, 1 H), 3.44 (s, 2 H), 5.40
(bs, 1 H), 7.19-7.46 (m, 5 H); 13C NMR (75 MHz, C6D6) δ 14.3,
17.5, 19.6, 21.0, 29.7, 31.2, 32.6, 34.2, 50.7, 54.3, 54.7, 58.5,
58.8, 62.3, 63.7, 126.7, 127.20-129.22 (5 C), 139.6, 140.0;
HRMS (CI) calcd for C23H37N2 (MH+) 341.2962, found 341.2959.
2-Ben zyl-8,9-d im et h yl-2-a za b icyclo[3.3.1]n on -6-en e-
6,8-d ica r ba ld eh yd es 32 a n d 33. Adduct 31 (60 mg) was
dissolved in a solution of CH2Cl2/MeOH (99:1, 8 mL). Silica
gel (1.2 g) was added, and the resulting heterogeneous mixture
was vigorously stirred during 4 h at room temperature. After
filtration and removal of solvent under reduced pressure, an
inseparable mixture of dialdehydes 32 and 33 was obtained
as a pale yellow oil (20 mg, 33% yield) in a 2:1 ratio: MS (IC)
1
m/z 298 (MH+). Major isomer 32: H NMR (CDCl3, 400 MHz)
Ma n za m in e Mod el. 8-Acetyla m in o-2-ben zyl-7,8a -d i-
m eth yl-1,2,3,4,4a ,5,8,8a -octa h yd r oisoqu in olin e-5-ca r box-
ylic Acid Eth yl Ester (23). To dihydropyridinum salt 1b [2.4
mmol, prepared from 1-benzyl-4-methoxy-3-methyl-1,4,5,6-
tetrahydropyridine (651 mg) and camphorsulfonic acid (697
mg)],12 dissolved in anhydrous CH2Cl2 (20 mL), was added at
0 °C diene 12 (661 mg, 1.25 mmol) in CH2Cl2 (20 mL). After
10 min at 0 °C, the solution was stirred for 15 h at 4 °C. The
reaction mixture was then poured into a mixture of saturated
NaHCO3 and H2O. The CH2Cl2 phase was collected, and the
H2O layer was extracted with CH2Cl2. The combined organic
fractions were dried over MgSO4. The solvent was removed
under reduced pressure. The residue was dissolved in CH2Cl2
(60 mL), and to this solution were added pyridine (1.21 mL)
and Ac2O (0.71 mL, 7.5 mmol). After the mixture was stirred
during 1.5 h at ambient temperature, crushed ice and satu-
rated NaHCO3 were added. The resulting mixture was ex-
tracted with CH2Cl2 and dried over MgSO4 and the solvent
removed under reduced pressure. The residue (1.49 g) was
chromatographed over silica gel (150 g) using a heptane/
AcOEt/CH3OH/Et3N mixture (gradient from 90:10:0:0.1 to 79:
20:1:0.1). The major adduct 23 was isolated as a pale yellow
oil (360 mg, 43% yield): IR (CHCl3) 3444, 3007, 2945, 2811,
1721, 1661 cm-1; 1H NMR (CD3OD, 400 MHz) δ 7.31 (m, 5 H),
5.49 (bs, 1 H), 4.42 (bs, 1 H), 4.16 (q, J ) 7.1 Hz, 2 H), 3.52 (d,
J ) 13.4 Hz, 1 H), 3.47 (d, J ) 13.4 Hz, 1 H), 3.26 (m, 1 H),
2.56 (m, 1H), 2.37 (m, 1 H), 2.34 (bd, J ) 11.8 Hz, 1 H), 2.17
(bd, J ) 11.8 Hz, 1 H), 2.08 (m, 1 H), 1.99 (m, 1 H), 1.94 (s, 3
H), 1.62 (bs, 3 H), 1.40 (m, 1 H), 1.25 (t, J ) 7.1 Hz, 3 H), 1.17
(s, 3 H); 13C NMR (CD3OD, 100 MHz) δ 14.5, 20.7, 22.6, 25.0,
26.9, 38.7, 39.6, 44.7, 50.0, 56.7, 56.8, 62.0, 64.3, 120.7, 128.0,
129.2, 129.9, 135.7, 139.7, 173.6, 175.7; MS (IE) m/z 384 (M+),
326 (M+ - NHCOCH3), 310 (C3H6O2), 186, 146 (C10H12N), 91
(C7H7); HRMS (CI) calcd. for C23H33N2O3 (MH+) 385.2490,
found 385.2466. Minor isomer 24 (60 mg, 7% yield): IR (CHCl3)
δ 1.00 (m, 1 H), 1.25 (s, 3 H), 1.39 (d, J ) 7.12 Hz, 3 H), 1.84
(m, 1 H), 2.2 (m, 1 H), 2.37 (dd, J ) 14, 4.8 Hz, 1 H), 2.60 (dt,
J ) 14.6, 3.2 Hz, 1 H), 2.82 (s, 1 H), 3.85 (s, 1 H), 3.85 (d, J )
14.1 Hz, 1 H), 3.95 (d, J ) 14.1 Hz, 1 H), 6.95 (s, 1 H), 7.27
(m, 5 H), 9.49 (s, 1 H), 9.65 (s, 1 H);13C NMR (CDCl3, 75 MHz)
δ 17.7, 18.6, 21.9, 30.9, 30.9, 41.0, 56.0, 62.0, 65.3, 127.1, 128.0,
128.4, 140.0, 144.5, 151.9, 192.5, 199.9. Minor isomer 33
(characteristic signals): 1H NMR (CDCl3, 400 MHz) δ 1.05 (m,
1 H), 1.34 (1.34 (d, J ) 6.8 Hz, 3H), 1.47 (s, 3 H), 1.67 (m, 1
H), 2.79 (d, J ) 3 Hz, 1 H), 3.10 (s, 1 H), 3.92 (d, J ) 2.1 Hz,
1 H), 4.02 (d, J ) 2.1 Hz, 1 H), 6.68 (d, J ) 1.3 Hz, 1 H), 9.44
(s, 1 H), 9.47 (s, 1 H); 13C NMR (CDCl3, 75 MHz) δ 17.1, 19.0,
21.0, 30.4, 31.0, 41.4, 55.7, 62.0, 65.3, 127.0, 128.0, 128.4, 141.0,
146.0, 151.3, 192.2, 199.9.
Acetic Acid 6-Acetoxym eth yl-2-ben zyl-8,9 d im eth yl-2-
a za bicyclo[3.3.1]n on -6-en -8-ylm eth yl Ester s 34 a n d 35.
The above mixture of dialdehydes 32 and 33 was reduced with
an excess of NaBH4 in MeOH at ambient temperature during
3 h. After usual extraction, the crude diol mixture was treated
with an excess of Ac2O in pyridine. After removal of pyridine
and excess Ac2O under reduced pressure, the crude mixture
was purified by chromatography over alumina using EtOAc/
heptane (3:1) as eluent. A fraction containing a mixture of 34
and 35 (1:2 ratio) and a fraction of practically pure 35 was
thus obtained, allowing detailed NMR analysis. Minor isomer
34: 1H NMR (CDCl3, 400 MHz) δ 1.06 (m, 1 H), 1.07 (s, 3 H),
1.35 (d, J ) 6.8 Hz, 3 H), 1.79 (m, 1 H), 2.07 (s, 3 H), 2.09 (s,
3 H), 2.13 (d, J ) 3.4 Hz, 1 H), 2.15 (m, 1 H) 2.39 (dd, J ) 12.4
Hz, J ) 4.9 Hz, 1 H), 2.60 (s, 1 H), 2.7 (m, 1 H), 3.75 (d, J )
14.3 Hz, 1 H), 3.96 (d, J ) 14.3 Hz, 1 H), 4.08 (d, J ) 10.4 Hz,
1 H), 4.13 (d, J ) 10.7 Hz, 1 H), 4.41 (d, J ) 12.9 Hz, 1 H),
4,53 (d, J ) 12.9 Hz, 1 H), 5.68 (s, 1 H); 13C NMR (CDCl3, 75
MHz) δ 17.9, 20.5, 21.2, 25.1, 31.8, 35.5, 41.1, 44.2, 61.8, 65.7,
66.7, 70.0 126.8, 128.1, 128.26, 128.3, 131.6, 142.0, 172.0;
HRMS (EI) calcd for C23H31NO4 385.2253, found 385.2253.
Major isomer 35: 1H NMR (CDCl3, 300 MHz) δ 1.05 (m, 1 H),
1.05 (s, 3 H), 1.25 (d, J ) 6.7 Hz, 3 H), 1.90 (m, 1 H), 2.05 (s,
3 H), 2.08 (s, 3 H), 2.45 (dd, J ) 12.2, 5 Hz, 1 H), 2.60 (s, 1 H),
2.85 (m, 1 H), 3.75 (d, J ) 11 Hz, 1 H), 3.85 (d, J ) 11 Hz, 1
H), 3.88 (d, J ) 5 Hz, 1 H), 3.92 (d, J ) 5 Hz, 1 H), 4.4 (d,
J ) 12.9 Hz, 1 H), 4.55 (d, J ) 12.9 Hz, 1 H), 5.50 (s, 1 H),
7.25 (m, 5 H), 7.30 (m, 1 H), 7.35 (m, 1 H); 13C NMR (CDCl3,
75 MHz) δ 17.6, 21.0, 21.5, 22.0, 32.4, 35.2, 41.9, 44.4, 61.7,
62.2, 66.6, 70.8, 126.7, 128.2, 131.9, 137.7, 170.9, 171.2.
3254, 3016, 2826, 1721, 1661 cm-1 1H NMR (CD3OD, 400
;
MHz) δ 7.24 (m, 5 H), 5.65 (bs, 1 H), 4.15 (q, J ) 7.1 Hz, 3 H),
3.60 (d, J ) 13.1 Hz, 1 H), 3.40 (m, 1 H), 3.30 (d, J ) 13.1 Hz,
1 H), 2.96 (bd, J ) 11.0 Hz, 1 H), 2.88 (d, J ) 11.8 Hz, 1 H),
1.95 (s, 3 H), 1.92 (d, J ) 11.8 Hz, 1 H), 1.84 (m, 1 H), 1.74 (m,
1 H), 1.69 (bs, 3 H), 1.50 (m, 1 H), 1.25 (t, J ) 7.1 Hz, 3 H),
1.14 (m, 1 H), 1.02 (s, 3 H); 13C NMR (CD3OD, 75 MHz) δ 14.6.
21.5, 23.4, 25.2, 28.1, 36.9, 42.6, 43.3, 53.8, 55.3, 61.8, 64.1,
65.2, 120.1, 128.3, 129.3, 130.5, 135.4, 138.5, 173.1, 175.3; MS

Reactions of Aminopentadienal Derivatives
J . Org. Chem., Vol. 64, No. 20, 1999 7387
Con clu sion
products depicted in Figure 1. Another interesting feature
of this chemistry is the generation of very different and
fairly complex polycyclic molecules such as 15a , 19, 23,
and 32 using in all cases the same dihydropyridinium
species 1 as well as simple and very closely related dienes
such as 9, 10, 12, and 13.
The above results gave some experimental support in
favor of our recently proposed6 modification of the Bald-
win and Whitehead hypothesis concerning the biogenesis
of manzamine alkaloids.4 This model cannot be ruled out,
but, as it stands, it suffers from competitive reactions,
in particular oxidoreductive reactions between dihydro-
pyridine intermediates, a process that is difficult to
control (or requires an enzymatic process).21 Aminopen-
tadienal derivatives are not prone to oxidation and thus
do not present the disavantages inherent to dihydropyr-
idine reactivity. The chemistry of aminopentadienal
derivatives offers further opportunities for undertaking
experiments toward “biomimetic synthesis”22,23 of natural
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures for preparation of dienes 9, 10, 12, and 13 and
procedures used for NMR determination of complex structures.
Copies of 1H and 13C NMR spectra of dienes 9, 10, 12, 13,
adducts 15b, 17, 19, 20, 23, 24, 31, 32-33 (mixture at
equilibrium), 34, and 35 with attribution of signals and
schemes showing observed NOEs for compounds 23 and 35.
This material is available free of charge via the Internet at
http://pubs.acs.org.
J O990604P
(21) This model has recently received additional support by a short
synthesis of keramaphidin B, albeit in low yield: Baldwin, J . E.;
Claridge, T. D. W.; Culshaw, A. J .; Heupel, F. A.; Lee, V.; Spring, D.
R.; Whitehead, R. C.; Boughtflower, R. J .; Mutton, I. M.; Upton, R. J .
Angew. Chem., Int. Ed. 1998, 37, 2661-2663.
(22) A pertinent definition has been formulated by C. H. Heath-
cock: “A particularly powerful strategy is “biomimetic synthesis”, in
which we try to guess how Nature might assemble a particular
molecule and then try to mimic this hypothetical route in the
laboratory”.
(23) For examples of particularly efficient biomimetic syntheses of
natural alkaloids, see, inter alia: Robinson, R. J . Chem. Soc. 1917,
762. Heathcock, C. H. Angew. Chem., Int. Ed. Engl. 1992, 31, 665-
681. Franc¸ois, D.; Lallemand, M.-C.; Selkti, M.; Tomas, A.; Kunesch,
N.; Husson, H.-P. Angew. Chem., Int. Ed. Engl. 1998, 37, 104-105.
For a review with more examples, see: Tietze, L.; Beifuss, U. Angew.
Chem., Int. Ed. Engl. 1993, 32, 131-312.