Structure of alkaloid 275A, a novel 1-azabicyclo[5.3.0]decane from a Dendrobatid frog, Dendrobates lehmanni: Synthesis of the tetrahydrodiastereomers

Article abstract of DOI:10.1021/np0005098

The principal alkaloid 275A in skins of the Colombian poison frog Dendrobates lehmanni has been identified as the pyrrolo[1,2-α]azepane (1), the first occurrence in nature of this "izidine" system. Tetrahydro-1 proved identical to one of the four synthetic diastereomers, 2a-2d, thereby establishing that 1 has the 5Z,10E relative stereochemistry. Alkaloid 1 is often accompanied by other congeners, in particular a 5Z,10Z diastereomer 15, a dihydro analogue 16, and a ketone 17. Such izidines in frogs may arise from dietary ants, as do other classes of izidines.

Full text of DOI:10.1021/np0005098

J . Nat. Prod. 2001, 64, 421-427
421
Str u ctu r e of Alk a loid 275A, a Novel 1-Aza bicyclo[5.3.0]d eca n e fr om a
Den d r oba tid F r og, Den d r oba tes leh m a n n i: Syn th esis of th e
Tetr a h yd r od ia ster eom er s
H. Martin Garraffo,*,† Poonam J ain,† Thomas F. Spande,† J ohn W. Daly,† Tappey. H. J ones,‡
Lance J . Smith,‡ and Victor E. Zottig‡
Laboratory of Bioorganic Chemistry, NIDDK, NIH, Bethesda, Maryland 20892-0820, and Department of Chemistry,
Virginia Military Institute, Lexington, Virginia 24450
Received October 27, 2000
The principal alkaloid 275A in skins of the Colombian poison frog Dendrobates lehmanni has been
identified as the pyrrolo[1,2-a]azepane (1), the first occurrence in nature of this “izidine” system.
Tetrahydro-1 proved identical to one of the four synthetic diastereomers, 2a -2d , thereby establishing
that 1 has the 5Z,10E relative stereochemistry. Alkaloid 1 is often accompanied by other congeners, in
particular a 5Z,10Z diastereomer 15, a dihydro analogue 16, and a ketone 17. Such izidines in frogs may
arise from dietary ants, as do other classes of izidines.
Sch em e 1. Frog Skin Alkaloid 275A (1) and Four Synthetic
Tetrahydro Diastereomers (2a -2d )a
Dendrobatid frogs of Central and South America have
been a rich source of a wide variety of biologically active
alkaloids, comprising over 20 structural classes; code
names for the over 500 alkaloids found in frog skins are
provided in a recent review.1 In 1976, a C19 alkaloid of
molecular weight 275, which was code-named 275A (1),
was found to be a major alkaloid in skin extracts of a red
and black-banded dendrobatid frog, Dendrobates lehmanni
Myers and Daly, 1976, collected in montane forest near
Cali, Colombia.2 The frog, known previously only from
street markets in Cali, had been considered another
populational variant of the extremely variable species
Dendrobates histrionicus Berthold, 1845.3 The frog was
described as a new species, D. lehmanni,2 primarily on the
basis of the lack of histrionicotoxins, a class of alkaloids
present in all examined populations of D. histrionicus.4
Analysis of mating calls now supports the validity of D.
lehmanni as a separate species and was also the basis for
separation of the D. histrionicus complex into D. histrioni-
cus and D. sylvaticus Funkhauser, 1956.5
A gross structure (1 with stereochemistry undefined) was
presented for alkaloid 275A, based upon unpublished
results in a recent review.1 Alkaloid 1 was later one of a
group of mono- and bicyclic alkaloids studied with GC-CI
(NH3)-MS/MS and collision-activated dissociation.6 The
relative stereochemistry of 1 has now been established as
that of 2c by comparison of tetrahydro-1 with the four
synthetic diastereomers 2a -2d (Scheme 1).
Alkaloid 1 has been found as a major alkaloid only in D.
lehmanni, even in populations separated by over 100 km
(unpublished results). It has been detected by GC-MS, but
only rarely, as a trace alkaloid in populations of D.
speciosus Schmidt, 1857; D. pumilio Schmidt, 1858; D.
granuliferus Taylor, 1958; and D. auratus Girard, 1855
from Panama and Costa Rica (ref 4 and unpublished
results). We presume that 1 originates in frogs from an
alkaloid-containing small arthropod, on the basis of ongoing
studies of dendrobatid frogs, where other izidines and
decahydroquinolines have been shown to come from dietary
ants (ref 7 and references therein). The occurrence of 1 as
a
Diastereomer 2c results from hydrogenation of 1.
a significant alkaloid only in populations of D. lehmanni
poses interesting questions as to the distribution of the
putative alkaloid-containing arthropod. Montane forest
(850-1200 m) is the habitat for D. lehmanni, while most
populations of D. histrionicus occur at lower elevations
(<600 m). Thus, the putative arthropod containing 1 may
also be a montane species. Conversely, the arthropod(s)
containing the histrionicotoxins, also mainly C19 alkaloids,
which are in all populations of D. histrionicus, may not be
abundant at high elevations where D. lehmanni occurs but
only at lower elevations where D. histrionicus occurs.
However, a population of D. histrionicus from Altos de
Buey, Colombia, at 800 m, contained histrionicotoxins but
not 1.8 Feeding experiments indicated that D. lehmanni
readily accumulated histrionicotoxin into skin when fed
alkaloid-dusted fruit flies (unpublished results). Of the
other dendrobatid frogs found to contain 1, albeit in trace
amounts, only the Panamanian D. speciosus was from
montane forest (1250-1410 m).9
Resu lts a n d Discu ssion
Alkaloid 1 possessed a molecular formula of C19H33
N
(HREIMS) and had a nine-carbon side-chain containing a
terminal acetylene as shown by EIMS and GC-FTIR.
Figure 1 shows clearly the terminal acetylene with a νCtC
at 2119 cm-1, the νtCH as a split absorption at 3328 cm-1
,
and very weak Bohlmann bands (ca. 2800 cm-1).
* To whom correspondence should be addressed. Tel: (301)-496-1095.
Fax: (301)-402-0008. E-mail: garraffo@helix.nih.gov.
† NIH.
The acetylenic side-chain underwent facile R-cleavage on
EIMS. A weaker methyl loss was also detected, presumably
‡ Virginia Military Institute.
10.1021/np0005098 CCC: $20.00
© 2001 American Chemical Society and American Society of Pharmacognosy
Published on Web 03/24/2001

422 J ournal of Natural Products, 2001, Vol. 64, No. 4
Garraffo et al.
F igu r e 1. GC-FTIR spectrum of the alkaloid 275A (1) from the colombian frog Dendrobates lehmanni.
also from R-cleavage. Since the only unsaturation resided
in the acetylene moiety as proved by the uptake of four
hydrogens, 1 must have a bicyclic ring system. These
observations and the detection of no labile hydrogen
exchanges on CIMS with ND3, as well as very weak
Bohlmann bands, led us to propose initially a 6Z,10E10 4,6-
disubstituted quinolizidine structure for this alkaloid.11
After this proposal, a (6Z,10E)-4-methyl-6-n-propyl quino-
lizidine, 3, was detected in a Madagascan frog and also a
Brazilian ant;7 it had very weak Bohlmann bands, similar
to those observed with 1.
and the synthetic quinolizidines 4a -4d . However, analysis
1
of H NMR spectra obtained for 1 confirmed our earlier
conclusion of only one methyl group (d, J ) 6.8), assignable
to an R-methyl group (δ 1.45). We then began to study the
application to alkaloids of a seldom used MS technique
based upon collision-activated dissociation of the proto-
nated molecular ion, created by CI with NH3 reagent gas.
We found this tandem mass spectral technique particularly
useful when applied to the study of “izidine” alkaloids
where the separate rings and their attached substituents
are detected as major fragments.6 When applied to quino-
lizidine 3, GC-CI (NH3)-MS/MS produced the even mass
fragments of m/z 98 and 126. When this technique was
applied to 1, however, we found not the ions observed (m/z
98, 210) with the synthetic 4,6-disubstituted quinolizidine
isomers 4a -4d , i.e., protonated six-membered piperideines
with a methyl or nonyl substituent attached, but rather a
pyrroline carrying a nonynyl side-chain (m/z 192) and a
seven-membered dehydroazepane ring having a methyl
substitutent (m/z 112) (Scheme 2, cf. ref 6). This tandem
mass spectrometry technique revealed not only the nature
of each fused ring but also its substituent.
Using the same 2-methylpiperidine-6-n-butanal acetal
previously used in the synthesis of 37 and an n-nonyl
Grignard reagent, a mixture of four diastereomeric 4-meth-
yl-6-n-nonyl quinolizidines (4a -4d ) was synthesized. How-
ever, despite very close similarities of tetrahydro-1 and one
diastereomer, 4d , in MS and IR properties and GC behav-
ior, their retention times differed on co-injection. EIMS/
MS data also confirmed significant differences between
tetrahydro-1 and the two major diastereomers of 4 (see
Supporting Information).
Thus, 1 was proved not a quinolizidine but rather an
isomeric “izidine” with fused five- and seven-membered
rings commonly referred to as a 1-azabicyclo[5.3.0]decane,
although a pyrrolo[1.2-a]azepine nomenclature is also
accepted.12 We propose that such “5-7-izidines” be named
“lehmizidines” to emphasize their unique occurrence, as yet
in nature only in skin extracts of dendrobatid frogs, chiefly
from the Colombian dendrobatid frog D. lehmanni, and to
honor the late F. Carlos Lehmann Valencia for whom that
frog was named.2 CI (NH3)-MS/MS of tetrahydro-1 gener-
ated ions at m/z 112 and 196, proving the unsaturation to
be associated with the side-chain attached to the five-
membered pyrroline ring. For quick comparison purposes,
a mixture of the four diastereomers (2a -2d ) of tetrahy-
dro-1 was prepared by reductive amination of the triketone
5 (Scheme 3, unpublished work of T. H. J ones and J . S. T.
Gorman; see Supporting Information for the characteriza-
tion of 5).
We briefly considered branching in the nonyl chain as a
possible reason for the discrepancy between tetrahydro-1

Structure of Alkaloid 275A
J ournal of Natural Products, 2001, Vol. 64, No. 4 423
Sch em e 2. EIMS and CI-MS/MS Fragmentation Pathways Proposed for Alkaloid 275A (1) and Tetrahydro-1a
a
Fragment ion masses from tetrahydro-1 are in parentheses.
Sch em e 3. Lehmizidines from Reductive Amination of
Triketone 5
Sch em e 4. Lehmizidines from 2,5-Disubstituted Pyrrolidines
Gas chromatography of this mixture gave a separation
showing the four expected diastereomers and indicated that
the third-eluting component (2c) was identical in its GC
retention time, EIMS and CI (NH3)-MS/MS, and GC-FTIR
spectrum with tetrahydro-1. The EIMS fragmentations of
these diastereomers are virtually identical with one slight
difference noted; 2a and 2d show a slightly greater
R-cleavage of the C-5 methyl than do the other two isomers
(see Supporting Information).
We then set about designing efficient syntheses to
determine the relative stereochemistry of 2c. Two routes
were selected, one commencing from 2,5-disubstituted
pyrrolidines and the other from 2,7-disubstituted azepanes,
and, taken together, we were confident such syntheses
would prove the stereochemistry of 2c. A similar approach
was recently employed by us in the determination of the
relative stereochemistry of 3.7 First we prepared from 6
the cis and trans pyrrolidine ketals 7a and 7b (Scheme 4),
which were then cyclized by reductive amination after
deprotection. The major products from the cis/ trans mix-
ture were 2b and 2d . When nearly pure cis 7a was treated
in a similar manner, the major product was 2b, proving
that the pyrrolidine ring in 2b was cis and thus trans in
diastereomer 2d .
Since pure trans pyrrolidines are difficult to obtain by
reductive methods13 and since this route did not provide
useful amounts of 2c, we next devised a complementary
route (Scheme 5) through the appropriate azepanes based
on cyanoamine methodology. The requisite cyanoazepane
10 would require the cyclization of a protected seven-carbon
aldehyde with an amine at C-6. The precursor for this
compound, ketoacetal 8, was obtained in one step from
commercial starting materials using a Ni0-mediated con-
jugate addition.14 The aminoacetal 9 was produced nearly
quantitatively by reductive amination of 8 with benzy-
lamine in the presence of titanium isopropoxide.15
of cyanide ion and treatment with n-nonylmagnesium
bromide gave 2a , 2b, 2c, and 2d in a 45:13:11:31 ratio.
Similar treatment of a nearly pure sample of cis azepane
14a , obtained by the methodology used to provide the cis
pyrrolidine 7a , produced 2b and 2d , indicating the azepane
ring in these isomers to be cis and thus the azepane ring
in 2a and 2c to be trans.
These results allowed the assignment of cis pyrrolidine
and cis azepane rings to isomer 2b, i.e., a 5Z,10Z relative
stereochemistry. This is reflected in its slightly more
intense and broader Bohlmann bands relative to the other
diastereomers (see Figure 2, Supporting Information).
Diastereomer 2d then had a trans pyrrolidine ring and a
cis azepane ring or a 5E,10E stereochemistry. The stere-
ochemistry of the other two isomers, 2a and 2c, was
deduced by examining the material balances of 2a -2d
resulting from various mixtures of cis and trans pyrrolidine
ketals 7a /7b (see Scheme 4 and also Table 1 in Supporting
Information), in turn prepared by various reductants on
an intermediate pyrroline created from the N-chloro com-
pound with base (see Experimental Section).
The aminoacetal 9 was deprotected in the presence of
cyanide ion, and the resulting cyanoamine 10 was treated
with the Grignard reagent prepared from 2-(2-bromoethyl)-
1,3-dioxane to provide the N-benzylazepanes 13. Deben-
zylation gave a nearly 1:1 mixture of the isomeric dioxanyl
azepanes 14a /14b. Deprotection of 14a /14b in the presence

424 J ournal of Natural Products, 2001, Vol. 64, No. 4
Sch em e 5. Lehmizidines from 2,7-Disubstituted Azepanes
Garraffo et al.
diastereomer 15, a dihydro derivative 16, shown by EIMS
to have a molecular ion at m/z 277 with fragments at m/z
262 and 152 and by GC-FTIR to have a terminal olefin
and the same Bohmann band pattern as 1, a ketone
congener 289A (17: x ) 1-5), C19H31NO, and, tentatively,
a diene isomer of 1, 275G (18). A significant ion at m/z
165, seen with ketone 17 using the J EOL mass spectrom-
eter (see Experimental Section), indicates the carbonyl
group in the side-chain is not adjacent to the ring, and the
FTIR spectrum indicates it is not conjugated with the
terminal acetylene.
The major natural alkaloid 1 after hydrogenation cor-
responded on co-injection to the synthetic 2c, while the
minor isomer 15 after hydrogenation corresponded with 2b,
exhibiting identical retention times and EIMS and CI
(NH3)-MS/MS spectra (major fragments at m/z 112, 196 of
the same relative intensities; see Experimental Section)
and identical GC-FTIR spectra in the case of 2c and
tetrahydro-1 (insufficient material was available for an IR
spectrum on tetrahydro-15).
We had purified approximately 1 mg of 1 by HPLC and
obtained 1D and 2D proton NMR spectra. These data,
reported in Table 2 (Supporting Information), confirmed
the terminal acetylene and showed three downfield protons
(C-3, C-5, C-10) corresponding to CHN signals, one coupled
with the single methyl group, but were ambiguous regard-
ing the ring sizes and stereochemistry. In retrospect, we
saw connectivity in the five-membered ring in the COSY
spectrum of 1 but not in the seven-membered ring. Only a
single H-1- -H-2 cross-peak was detectible and was quite
weak and normally would have been ignored, were it not
that the structure became known in the interim. Initial
NMR work was attempted in CDCl3 but gave poor results.
Much better spectra were obtained in D2O, and from the
better separation of H-3, H-5, and H-10, we concluded that
a tertiary amine deuteriochloride had formed on exposure
to CDCl3. When the 11-methyl was decoupled, the signal
assigned to H-5 changed to a doublet (J ) 7.5 Hz),
indicating only one H-6- -H-5 coupling. We observed the
acetylenic hydrogen at δ 2.33 (J ) 2.7) coupled with the
H-18 methylene at δ 2.20. The latter signal on standing in
D2O for several days gradually changed from a triplet of
doublets (J ) 7.0, 2.7) to a triplet as the long-range coupling
to the acetylenic hydrogen (H-20) disappeared when that
hydrogen slowly exchanged with deuterium. Acetylenic
hydrogen exchange under neutral or slightly acidic condi-
tions is very unusual and so far as we know has only been
reported to occur via metal ion π-complexes.17 Our results
suggest an intramolecular general base deuteriochloride
catalysis.
The proportion of 2a in the cyclized mixtures rises
paralleling the increased percentage of cis pyrrolidine (7a )
in the starting mixture, proving that 2a has a cis pyrroli-
dine, while the azepane cyclizations proved 2a to have a
trans azepane ring. Thus 2a must have the 5E,10Z ster-
eochemistry, and finally by difference, 2c, the tetrahydro
derivative of the naturally occurring alkaloid, must have
a trans pyrrolidine and trans azepane ring, i.e., the 5Z,-
10E stereochemistry. This assignment is secured by the
observation that 2c increases with increases in the per-
centage of trans pyrrolidine, 7b in the starting 7a /7b
mixture (see Table 1, Supporting Information). Thus, the
first two isomers to elute from the GC column derive from
a cis pyrrolidine, and the pair to elute last arise from a
trans pyrrolidine. FTIR absorbance spectra of the four
synthetic diastereomers 2a -2d are presented in the Sup-
porting Information. A mixture of cis and trans 2-ethyl-7-
methyl azepanes (12) was also synthesized from 10, via
the N-benzyl intermediates 11 (Scheme 5), and GC-FTIR
spectra indicate these isomers can be distinguished by the
weak Bohlmann band shoulder evident in the cis isomer
but not the trans (see Figure 3, Supporting Information),
as is the case with piperidines and N-methylpyrrolidines.16
The elution order on our GC columns is the same as R,R′-
disubstituted-pyrrolidines and -piperidines with the cis
isomer eluting first.
Other “izidines”, namely, 3,5-disubstituted pyrrolizidines
and indolizidines and 4,6-disubstituted quinolizidines,
found in frog skin appear to derive from dietary ants,7 and
thus it appears likely that lehmizidines will have an ant
origin. However, at this time no simple lehmizidines, such
as 1 from dendrobatid frogs, have been detected in ants
nor anywhere else in nature. The common occurrence of
the C19 trans pyrrolidine 19 in myrmicine ants indigenous
In most extracts of D. lehmanni, there were detected
along with major amounts of 1, minor amounts of a

Structure of Alkaloid 275A
J ournal of Natural Products, 2001, Vol. 64, No. 4 425
Sch em e 6. Possible Biosynthetic Conversion of Ant
Pyrrolidine 19 to a Lehmizidine System
proportions of 1:15:16:17 ≈ 10:1:1:2 based on GC-MS chro-
matograms. The extract also contained several pumiliotoxins
(alloPTX 253A, alloPTX 267A, PTX 307A, PTX 323A) and a
number of trace alkaloids. The alkaloid fraction from skins of
a yellow and black morph of D. lehmanni collected in the R´ıo
Anchicaya´ drainage in 1992 contained large amounts of 1 and
also minor amounts of 15 in a ratio of 9:1. A variety of other
alkaloids were present, but in relatively small amounts.
Alkaloid fractions from two such frogs were combined and
purified by HPLC using an HP 1090 with flow rate of 1 mL/
min and a gradient of H2O-CH3CN with an HP Asahipak
column (4.6 mm × 25 cm) to provide approximately 1 mg of 1.
Ca ta lytic Hyd r ogen a tion of 1. Tetrahydro-1 was obtained
from an estimated 50 µg of 1 in a 1 mL screw cap vial fitted
with a Mayo-type Teflon adaptor cap and containing 100 µL
of MeOH, a 7 mm Teflon stirring bar, and a trace of 5% Pd/C
(Degussa) catalyst. After reduction at room temperature for 2
h under 2 atm pressure of H2 (generated using an Alltech
electrolytic generator), the H2 atmosphere was replaced with
N2 and the solution was filtered through a 4 mm syringe filter
(Alltech, #2092). The filter was then rinsed with another 100
µL of methanol in portions and the combined filtrate concen-
trated to 50 µL for GC-FTIR and GC-EIMS.
2-Meth yl-2-(5,8-d ioxoh ep ta d ecyl)-1,3-d ioxola n e (6). A
mixture containing 2.52 g (13.9 mmol) of 1-dodecen-3-one,24
2.32 g of 2-methyl-2-(5-oxopentyl)-1,3-dioxolane25 (13.9 mmol),
0.8 mL of triethylamine, and 0.38 g of 3-benzyl-5-(2-hydroxy-
ethyl)-4-methylthiazolium chloride was stirred and heated
under argon overnight. The mixture was taken up in 50 mL
of diethyl ether and filtered through a short Florisil column.
The solvent was removed in vacuo, and the residue was
recrystallized from methanol to provide 2.78 g (56.5% yield)
of 6 as a waxy solid, mp 34-36 °C: FTIR 2934, 1723, 1457,
1216, 1125, 1065 cm-1; EIMS m/z 354 (M+, 0.5), 339 (2), 321
(1), 249 (2), 199 (3), 165 (1), 128 (11), 111 (2), 99 (3), 87 (100),
55 (11), 43 (38); HREIMS m/z 353.2693 (calcd for C21H37O4,
(M - H) +, 353.2692); 355.2857 (calcd for C21H39O4 (M + H)+,
355.2848).
to the western hemisphere18 suggests that 19 in ants might
share the same biosynthetic pathway as 1 or even be its
biosynthetic precursor (see Scheme 6). The bicyclic ana-
logues of monocyclic ant venom alkaloids are well known,
and indeed, pyrrolidine 19 occurs along with analogous
pyrrolizidines in other Monomorium species.19 A cis pyr-
rolidine isomer (present as a trace accompanying 19, cis:
trans ≈ 1:100) might lead to a dihydro-15. Acetylenic
groups have not been detected in ant pyrrolidines.
Although simple lehmizidines represent a new alkaloid
class, the Stemona alkaloids contain this ring system (cf.
ref 20 and references therein) as do the Cephalotaxus
alkaloids.21 Several recent syntheses of the simple lehmizi-
dine system have been reported.22 Earlier work was
undertaken chiefly in the context of an abnormal Clem-
mensen reduction product encountered by Clemo.23 We
chose to develop different stereoselective routes from
pyrrolidines and azepanes (Schemes 3 and 4) to allow the
unambiguous and efficient determination of diastereomer
structures.
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. Vapor phase FTIR
spectra were obtained using a Hewlett-Packard (HP) model
5965B detector interfaced with an HP 5890 gas chromatograph
fitted with either a J &W 15 m × 0.25 mm DB-1 column or a
30 m × 0.25 mm Restek Rtx-5 Amine column. An HP
ChemStation was used to process the data. Mass spectra were
obtained in the EI mode using a Shimadzu QP-5000 GC-MS
equipped with a 30 m × 0.32 mm Rtx-5 column. A Finnigan
GCQ mass spectrometer with the same column was used for
EI-MS/MS and CI (NH3)-MS/MS and a Finnigan Ion Trap
Model 7 for some earlier EIMS data on various extracts. High-
resolution mass spectrometry (HREIMS) used a J EOL SX102
instrument equipped with a 15 m × 0.20 mm HP-5 column
for GC-EIMS and GC-HREIMS. Direct probe work with the
same instrument was done in the FAB+ mode with xenon gas
and a poly(ethylene glycol) reference and glycerol matrix.
Sou r ce of Biologica l Ma ter ia ls. Specimens of D. leh-
manni were originally purchased from street vendors of Cali,
Colombia, and later collected in montane forest of the R´ıo
Anchicaya´ drainage nearby in the department of Cauca in the
1970s.2 Voucher specimens are in the American Museum of
Natural History (New York City). A subsequent collection in
1983 was from steep slopes in the same general area near R´ıo
Clara, since most of the higher forest had been destroyed
(unpublished data). Further specimens of unknown locales
have come through the pet trade or were provided to us after
confiscation by the U.S. Fish and Wildlife Service. We are
indebted to Volker Ennenbach for specimens from the R´ıo
Anchicaya´ drainage and from the R´ıo Daqua drainage a few
kilometers to the north and to Dr. J ack Frankel for specimens
from the R´ıo Ta´pare drainage near San J ose´ de Palmar, Choco,
Colombia, some 150 kilometers to the north. Such specimens
were obtained in the early 1990s.
cis- a n d tr a n s-2-Meth yl-2-[4-(5-n -n on yl-2-p yr r olid yl)-
bu tyl]-1,3-d ioxola n e (7a /7b). A solution containing 0.358 g
(1.01 mmol) of 6 in 8 mL of methanol was treated sequentially
with 20 mg of KOH, 0.27 g of NaCNBH3, and 0.092 g of NH4-
OAc. The mixture was stirred overnight, and 40 mg of NaBH4
was added. After 1 h, the solvent was removed in vacuo and
the residue partitioned between 10% NaOH and diethyl ether.
The ether layer was dried over anhydrous K2CO3, and the
solvent was removed in vacuo to provide 0.3 g (90%) of an oily
residue that contained two major components in a 51:49 ratio
having identical EIMS: EIMS m/z 339 (M+, 1), 324 (4), 294
(3), 213 (4), 212 (45), 197 (11), 196 (84), 182 (8), 168 (13), 152
(7), 126 (9), 99 (8), 87 (77), 82 (20), 69 (17), 68 (18), 55 (19), 43
(100); HREIMS m/z 340.3200 (calcd for C21H42 NO2 (M + H) +
340.3216).
,
cis-2-Meth yl-2-[4-(5-n -n on yl-2-p yr r olid yl)bu tyl]-1,3-d i-
oxola n e (7a ). A solution containing 0.36 g of a mixture of 7a
and 7b in 10 mL of anhydrous CH2Cl2 was treated with 0.2 g
of N-chlorosuccinimide and stirred for 2 h. The solvent was
removed in vacuo, and the residue was taken up in 10 mL of
ethanol, treated with 1.0 g of NaOH, and heated at reflux for
2 h. After cooling, the solvent was removed in vacuo and the
residue was partitioned between water and diethyl ether. GC-
MS analysis of the ether layer showed only two isomeric
1-pyrrolines. EIMS: m/z 336 ((M - H)+, 1), 322 (4), 294 (11),
292 (17), 250 (8), 222 (13), 209 (30), 180 (3), 166 (19), 124 (20),
110 (10), 97 (15), 96 (28), 87 (100), 83 (42), 82 (61), 55 (20), 43
(96), 41 (40), and m/z 336 ((M - H)+, 1), 322 (3), 294 (5), 292
(9), 250 (14), 238 (8), 194 (6), 183 (8), 182 (70), 180 (18), 87
(100), 83 (35), 55 (15), 43 (79), 41 (34). Hydrogenation of this
mixture over 5% Rh/Al2O3 in ethanol provided an 83:17 ratio
of 7a and 7b. Treatment of similarly prepared pyrroline
mixtures with a slight excess of DIBAL under argon at -15
°C provided 7a and 7b in an 89:11 ratio on one occasion and
a 93:7 ratio on another (Table 1, Supporting Information).
2-(5-Oxoh exyl)-1,3-d ioxola n e (8). A solution containing
25 mL of pyridine, 50 mL of THF, 7 g of Zn dust, 8 mL (98
Ch a r a cter iza tion of Alk a loid s. The profiles of alkaloids
in skin extracts of different specimens of D. lehmanni obtained
in different years and different locations in Western Colombia
were assessed by GC-MS analysis. The profile and amounts
of alkaloids differed considerably in these extracts, suggesting
variability in alkaloid prey items. The original extract from
the R´ıo Anchicaya´ drainage collected in J anuary 19734 was
characterized as follows: Alkaloid 275A (1) was a major
alkaloid and was accompanied by 15, 16, and 17 with the

426 J ournal of Natural Products, 2001, Vol. 64, No. 4
Garraffo et al.
mmol) of methyl vinyl ketone, and 5 g of nickel(II) chloride
was heated and stirred under argon until the mixture became
reddish-brown (0.5 h). A solution containing 9.05 g (50.0 mmol)
of 2-(2-bromoethyl)-1,3-dioxolane in 2 mL of pyridine was
added dropwise, and the mixture was refluxed overnight. After
cooling, it was diluted with 200 mL of diethyl ether and filtered
through Celite. The solvent was removed in vacuo, and the
residue was filtered through a short silica gel column with
diethyl ether to provide, after evaporation of the solvent, 8.35
g (97.0%) of 8. A sample had bp 73-77 °C (0.3 mmHg): EIMS
m/z 171 ((M - H) +, 1), 154 (1), 139 (1), 129 (1), 114 (2), 99 (2),
84 (4), 73 (100), 55 (4), 43 (44); HREIMS m/z 173.1191 (calcd
for C9H17O3 (M + H)+, 173.1178); 171.1023 (calcd for C9H15O3
(M - H)+, 171.1021).
(M+, 0.5), 226 (2), 212 (1), 198 (2), 184 (1), 170 (2), 152 (2), 138
(4), 126 (7), 113 (11), 112 (100), 96 (4), 87 (11), 84 (18), 70 (18),
56 (13), 41 (36); HREIMS m/z 228.1966 (calcd for C13H26NO2
(M + H)+, 228.1964).
cis-2-Meth yl-7-[3-(1,3-dioxan -2-yl)pr opyl]azepan e (14a).
A 0.1 g sample of cis and trans 14a /14b was treated sequen-
tially with N-chlorosuccinimide and NaOH as described for
the synthesis of 7a to provide a 1:1 mixture of the isomeric
1-dehydroazepane acetals: EIMS m/z 225 (M+, 2), 224 (4), 210
(2), 197 (2), 139 (20), 124 (60), 111 (62), 110 (80), 96 (46), 87
(90), 83 (17), 82 (20), 68 (30), 67 (33), 55 (40), 42 (50), 41 (100)
and m/z 225 (2), 224 (2), 149 (12), 148 (10), 138 (30), 125 (100),
124 (36), 110 (54), 97 (15), 87 (14), 84 (14), 82 (15), 80 (15), 44
(42), 41 (100). Hydrogenation of this mixture over 5% Rh/Al2O3
provided a 98:2 mixture of the azepanes 14a /14b.
Cycliza tion s to 3-n -Non yl-5-m eth ylleh m izid in es (2a -
2d ). A. F r om Dioxa n yl Azep a n es (14a /14b). A solution
containing 0.1 g of a 1:1 mixture of 14a /14b in 5 mL of THF,
5 mL of 10% HCl, and 0.5 mL of 70% HClO4 was stirred
overnight. The THF was evaporated, 20 mL of CH2Cl2 was
added, and the mixture was carefully brought to a methyl
orange end-point with KCN and stirred for 20 h. The mixture
was then made basic by the addition of a slight excess of KCN,
and the organic layer was separated and dried over anhydrous
K2CO3. The solvent was removed in vacuo, and the residue
was taken up in 10 mL of THF and treated with an excess of
n-nonylmagnesium bromide to provide after the usual workup
2a , 2b, 2c, and 2d in a 45.2:13.4:10.5:30.9 ratio. HREIMS: m/z
280.3013 (calcd for C19H38N (M + H)+, 280.3004); m/z 278.2859
(calcd for C19H36N (M - H)+, 278.2848). When a 98:2 mixture
of 14a /14b was treated under similar conditions, a 1.4:25:0.2:
74 mixture of 2a , 2b, 2c, and 2d was obtained. HREIMS: m/z,
280.3014 (calcd for C19H38N (M + H)+, 280.3004); 278.2859
(calcd for C19H36N (M - H)+, 278.2848).
Cycliza tion s to 3-n -Non yl-5-m eth ylleh m izid in es (2a -
2d ). B. F r om P yr r olid in e Keta ls (7a /7b). Solutions contain-
ing ca. 50 mg of 7a /7b in 5 mL of THF were treated with 1
mL of 10% HCl and 3 drops of 70% HClO4 and stirred
overnight. The mixture was made basic with 10% NaOH and
partitioned between diethyl ether and water. The ether layer
was evaporated, and the residue was taken up in 10 mL of
MeOH, then treated with a few drops of 10% NaOH and a
slight excess of NaCNBH3, and stirred for 6 h. The mixture
was carefully acidified with 10% HCl and made basic with 10%
NaOH, and after concentration in vacuo, the residue was
partitioned between diethyl ether and water. Gas chromato-
graphic analysis of the ether layers showed the presence of
3-n-nonyl-5-methyllehmizidines 2a -2d in the ratios shown in
Table 1 (Supporting Information).
2-(5-Ben zyla m in oh exyl)-1,3-d ioxola n e (9). The keto-
acetal 8 prepared above was treated with 5.46 mL (50.0 mmol)
of benzylamine and 18.4 mL (1.3 equiv) of titanium(IV)
isopropoxide for 2 h. After the addition of 50 mL of anhydrous
ethanol and 2.1 g of NaCNBH3, the mixture was stirred
overnight. The mixture was diluted with 250 mL of diethyl
ether and treated with 4 mL of 10% NaOH and 8 mL of water
followed by 20 g of anhydrous K2CO3. The mixture was filtered
through Celite, and after the solvent was removed in vacuo,
the residue was filtered through a short Florisil column with
ether. Subsequent evaporation of the solvent provided 7.64 g
(59.9%) of nearly pure 8. A sample purified by Kugelrohr
distillation had bp 130-140 °C (0.15 mmHg): EIMS m/z 262
((M - H)+, 1), 248 (2), 218 (1), 206 (1), 186 (1), 172 (1), 160
(1), 146 (1), 134 (73), 106 (7), 91 (100), 73 (25), 65 (9), 45 (16);
HREIMS m/z 264.1954 (calcd for
264.1964).
C
16H26NO2 (M + H)+,
N -B e n zy l-2-m e t h y l-7-[3-(1,3-d io x a n -2-y l)p r o p y l]-
a zep a n e (13). A solution containing 3.21 g (12.2 mmol) of 2-(5-
benzylaminohexyl)-1,3-dioxolane 9 in 5 mL of THF was cooled
in an ice bath and treated with 20 mL of 10% HCl and 2.5 mL
of 70% HClO4 and stirred at room temperature for 12 h. A
small aliquot was neutralized with 10% NaOH, and analysis
of the organic phase showed that the starting material was
completely hydrolyzed. The mixture was then diluted with 50
mL of H2O, and after the addition of 70 mL of CH2Cl2, the
two-phase mixture was adjusted to a methyl orange end-point
by the careful addition of solid KCN and stirred for 5 h. The
mixture was made slightly basic by the careful addition of
excess KCN, the organic phase was separated and dried over
anhydrous K2CO3, and the solvent was removed in vacuo. The
residue 10, was dissolved in 10 mL of anhydrous THF and
added under argon to a 1.8-fold excess of the Grignard reagent
prepared from 2-(2-bromoethyl)-1,3-dioxane.26 The mixture
was worked up by the addition of 150 mL of diethyl ether and
3 mL of 10% NaOH, followed by anhydrous K2CO3. The
mixture was filtered through Celite, and the solvent was
removed in vacuo to provide 4.1 g of residue, 80% of which
consisted of 13, present as a pair of isomers: FTIR (mixture
of both isomers) 3068, 3030, 2973, 2961, 2931, 2853, 2732,
1600, 1492, 1456, 1378, 1280, 1243, 1216, 1148, 1097, 1023,
1005, 928, 892, 854, 799 cm-1; EIMS m/z 317 (M+, 1), 302 (2),
274 (1), 260 (1), 240 (1), 226 (4), 216 (1), 203 (16), 202 (100),
186 (1), 174 (5), 160 (2), 146 (4), 127 (7), 106 (5), 91 (93), 69
Ch a r a cter iza tion of leh m izid in e 275A (1): EIMS m/z
275 (1), 260 (4), 152 (100), 124 (3), 110 (2), 96 (1), 82 (1), 67
(3); EIMS/MS (on m/z 260) 138 (100); (on m/z 152) 124 (58),
110 (67), 96 (27), 70 (100); (on m/z 138) 110 (24), 96 (17), 84
(100), 70 (30); CIMS (NH3) 276; CIMS (ND3): 277; CI (NH3)-
MS/MS m/z 192 (43), 112 (100), 109 (43), 95 (60), 81 (18); FTIR
(see Figure 1) 3328, ca. 2800, 2119, 1457, 1364, 1247 cm-1; 1H
NMR (see Table 2, Supporting Information); HREIMS m/z,
275.2621 (calcd for C19H33N, 275.2613); 274.2171 (calcd for
(9), 55 (11), 41 (27); HREIMS m/z 318.2427 (calcd for C20H32
-
C
19H32N, 274.2164); 260.2380 (calcd for C18H30N, 260.2378).
NO2 (M + H)+, 318.2433); 316.2273 (calcd for C20H30NO2 (M
- H)+, 316.2277).
Ch a r a cter iza tion of tetr a h yd r o-275A: FTIR (identical to
2c, see Figure 2, Supporting Information) 2932, 2865, ca. 2800,
1460, 1365, 1212 cm-1; EIMS m/z 279 (2), 278 (2), 264 (8), 250
(2), 236 (2), 222 (3), 153 (8), 152 (100), 138 (6), 124 (3), 110
(3), 96 (2), 82 (2); EIMS (direct probe) m/z 279 (3), 264 (5), 250
(5), 222 (3), 180 (3), 152 (100), 138 (5), 124 (5), 110 (6), 96 (5),
83 (12), 70 (22), 55 (13); CIMS (i-C4H10, direct probe) 289 (100),
294 (35), 268 (52); CI (NH3)-MS/MS m/z 196 (44), 112 (100),
109 (23), 95 (61), 81 (18).
Ch a r a cter iza tion of leh m izid in e 275A′ (15): FTIR 3328,
2865, ca. 2800 (broader than for 1 above, resembling that of
2b, see Figure 2, Supporting Information), ca. 2110, 1457, ca.
1360, ca. 1250-1150 cm-1; EIMS m/z 275 (6), 274 (9), 260 (17),
246 (4), 232 (3), 218 (3), 194 (3), 178 (3), 166 (2), 152 (100),
138 (3), 136 (3), 124 (5), 110 (4), 95 (3), 81 (4), 67 (6), 55 (6);
cis- a n d tr a n s-2-Meth yl-7-[3-(1,3-d ioxa n -2-yl)p r op yl]-
a zep a n e (14a /14b). A solution containing 4.1 g of 13, 4.0 g of
10% Pd/C, and 4.2 g of ammonium formate in 100 mL of MeOH
was heated at reflux for 5 h. Analysis of an aliquot showed
that none of the starting material remained. Occasionally
when repeating this step, there would be no reaction, then the
mixture would be filtered, new catalyst and ammonium
formate added, and the reflux resumed. When all the starting
material had reacted, the mixture was filtered through Celite,
the solvent was removed in vacuo, and the residue was
partitioned between diethyl ether and 10% NaOH. Evaporation
of the solvent provided 2.5 g (85%) of a mixture, 70% of which
was 14a /14b, which were analyzed together: FTIR (broad GC
peak) 2960, 2931, 2852, 1379, 1148, 1003 cm-1; EIMS m/z 227

Structure of Alkaloid 275A
J ournal of Natural Products, 2001, Vol. 64, No. 4 427
(9) Edwards, M. W.; Daly, J . W.; Myers, C. W. J . Nat. Prod. 1988, 51,
1188-1197.
H4-5, CI (NH3)-MS/MS m/z 196 (28), 112 (100), 109 (33), 95
(71), 81 (23); GC peak overlaps that of 16 below (15:16 ≈ 1.6:
1)
(10) We have adopted the stereochemical notation of Sonnet, which was
developed originally for indolizidines but has also been used in many
cases with other R,R′-disubstituted “izidines” to describe their relative
configurations. The hydrogen at the lowest numbered substituent (C-
3, in the case of 1) is used as a reference for the other substituent
(C-5) and the bridgehead C-10 hydrogen, and they are either on the
same face (Z) or opposite face (E). Cf.: Sonnet, P. E.; Netzel, D. A.;
Mendoza, R. J . Heterocycl. Chem. 1979, 16, 1041-1047. R* and S*
descriptors, while occasionally used, are more cumbersome and not
as intuitive.
Ch a r a cter iza tion of leh m izid in e 277A (16): FTIR 3085,
2864, ca. 2800, 1643, 991, 912 cm-1; EIMS m/z 277 (5), 276
(6), 262 (12), 248 (2), 234 (3), 220 (4), 194 (2), 178 (3), 162 (3),
152 (100), 124 (5), 110 (4), 96 (4), 82 (4), 67 (7), 55 (10).
Ch a r a cter iza tion of leh m izid in e 289A (17): FTIR 3328,
ca. 2800, ca. 2120, 1723, 1452, 1360, 1249, 1214, 1120 cm-1
;
HREIMS m/z 289.2399 (calcd for C19H31NO (M+), 289.2406);
288.2306 (calcd for C19H30NO (M+ - 1), 288.2327); 274.2177
(calcd for C18H28NO (M+ - CH3) 274.2171); 152.1436 (calcd
for C10H18N, (M+ - C9H13O), 152.1439); EIMS m/z 289 (1), 274
(6), 208 (2), 178 (2), 165 (4), 152 (100), 124 (4), 110 (3), 82 (3),
67 (3); EIMS (J EOL instrument) m/z 289 (10), 288 (10), 274
(50), 260 (6), 246 (6), 232 (6), 208 (10), 193 (6), 180 (10), 165
(26), 152 (100), 124 (16), 110 (10), 96 (10), 81 (10); CIMS (NH3)
290; CIMS (ND3) 291.
(11) Daly, J . W.; Garraffo, H. M.; Spande, T. F. In The Alkaloids; Cordell,
G. A., Ed.; Academic Press: San Diego, CA, 1993; Vol. 43, Chapter
3, pp 185-288.
(12) (Ring index, number 8037: octahydropyrrolo[1,2-a]azepine). For an
early review, see: Mosby, W. L. In Heterocyclic Compounds, Weiss-
berger, A., Ed.; Interscience: New York, 1961 (Heterocyclic Systems
with Bridgehead Nitrogen Atoms; Part 2, Chapter 10, pp 957-1000).
The following are early references to the synthesis of the parent
unsubstituted ring system: (a) Leonard, N. J .; Wildman, W. C. J .
Am. Chem. Soc. 1949, 71, 3100-3102. (b) Leonard, N. J .; Goode, W.
E. J . Am. Chem. Soc. 1950, 72, 5404-5407. (c) Leonard, N. J .; Felley,
D. L.; Nicolaides, E. D. J . Am. Chem. Soc. 1952, 74, 1700-1703. (d)
Cohen, L. A.; Witkop, B. J . Am. Chem. Soc. 1955, 77, 6595-6600. (e)
Djokic´, S.; Seiwerth, R. Croat Chem. Acta 1957, 29, 97-100; Chem.
Abstr. 52, 11818f. (f) Lukesˇ, R.; J anda, M. Collect. Czech Chem.
Commun. 1960, 25, 1612-1617; Chem. Abstr. 54, 18516c.
(13) Bacos, D.; Celerier, J . P.; Marx, E.; Saliou, C.; Lhommet, G.
Tetrahedron Lett. 1989, 30, 1081-1082.
Ack n ow led gm en t. We thank Lewis Pannell and Noel
Whittaker, LBC, NIDDK, NIH, for the HREIMS data and
Herman J . C. Yeh, LBC, NIDDK, NIH, for the 2D 1H NMR
data on 1. J effrey S. T. Gorman prepared and carried out the
triketone reaction that produced the first reference mixture
of lehmizidines 2a -2d . J ason M. Wilham, an NIH UGSP
summer student, obtained some of the GC-FTIR spectra. The
authors T.H.J . and L.J .S. gratefully acknowledge support from
the J effress Memorial Trust.
(14) Manchand, P. S.; Yiannikouros, G. P.; Belica, P. S.; Madan, P. J . Org.
Chem. 1995, 60, 6574-6581.
(15) Mattson, R. J .; Pham, K. M.; Leuck, D. J .; Cowen, K. A. J . Org. Chem.
1990, 55, 2552-2554.
(16) Garraffo, H. M.; Simon, L. D.; Daly, J . W.; Spande, T. F.; J ones, T.
H. Tetrahedron 1994, 50, 11329-11328.
(17) cf. Ginnebaugh, J . P.; Maki, J . W.; Lewandos, G. S. J . Organomet.
Chem. 1980, 190, 403.
(18) J ones, T. H.; Blum, M. S.; Howard, R. W.; McDaniel, C. R.; Fales, H.
M.; DuBois, M. D.; Torres, J . J . Chem. Ecol. 1982, 8, 285-300.
(19) J ones, T. H.; Stahly, S. M.; Don, A. W.; Blum, M. S. J . Chem. Ecol.
1988, 14, 2197-2211.
(20) Xu, R.-S.; Lu, Y.-J .; Chu, J .-H.; Iwashita, T.; Naoki, H.; Naya, Y.;
Nakanishi, K. Tetrahedron 1982, 38, 2667-2670.
Su p p or tin g In for m a tion Ava ila ble: The compositions of mix-
tures of 2a -2d from cyclizations of pyrrolidines 7a /7b (Table 1), the
1H NMR data on 1 (Table 2), the spectral characterization of 4-methyl-
6-n-nonylquinolizidines (4a -4d ) and the triketone 5, the synthesis of
cis- and trans-2-ethyl-7-methylazepanes (12) via the N-benzyl inter-
mediates (11), the FTIR spectra of 2a -2d (Figure 2) and cis and trans
12 (Figure 3), and the MS characterization of the four synthetic
diastereomers of tetrahydro-1 (2a -2d ) are available free of charge via
the Internet at http://pubs.acs.org.
(21) Huang, L.; Xue, Z. In The Alkaloids; Brossi, A., Ed.; Academic Press:
New York, 1984; Vol. 23, Chapter 3, pp 157-226.
Refer en ces a n d Notes
(22) (a) Dorta, R. I.; Francisco, C. G.; Suarez, E. Chem. Commun. 1989,
1168-1169. (b) Hammond, G. B.; Plevy, R. G. Org. Prep. Proced. Int.
1991, 23, 735-739. (c) Moeller, K. D.; Rothfus, S. L. Tetrahedron Lett.
1992, 33, 2913-2916. (d) Moeller, K. D.; Hanau, C. E. Tetrahedron
Lett. 1992, 33, 6041-6044. (e) Hammond, G. B.; Plevey, R. G, J .
Fluorine Chem. 1993, 63, 1-2. (f) Milligan, G. L.; Mossman, C. J .;
Aube´, J . J . Am. Chem. Soc. 1995, 117, 10449-10459. (g) Morimoto,
Y.; Iwahashi, M. Synlett. 1995, 12, 1221-2.
(23) Cf.: Prelog, V.; Seiwerth, R. Chem. Ber. 1939, 72, 1639-1642. Clemo,
G. R.; Ramage, G. R. J . Chem. Soc. 1931, 437-439, and also refs
12a-c.
(24) J ones, T. H.; Franko, J . B.; Blum, M. S.; Fales, H. M. Tetrahedron
Lett. 1980, 21, 789-792.
(1) Daly, J . W.; Garraffo, H. M.; Spande, T. F. In Alkaloids: Chemical
and Biological Perspectives; Pelletier, S. W., Ed.; Pergamon Press:
New York, 1999; Vol. 13, Chapter 1, pp 1-161.
(2) Myers, C. W.; Daly, J . W. Bull. Am. Mus. Nat. Hist. 1976, 157, 173-
262.
(3) Silverstone, P. A. Nat. Hist. Mus. Los Angeles Cty. Sci. Bull. 1975,
21, 1-55.
(4) Daly, J . W.; Myers, C. W.; Whittaker, N. Toxicon 1987, 25, 1023-
1095.
(5) Lo¨tters, S.; Glaw, F.; Ko¨hler, J .; Castro, F. Herpetozoa 1999, 12, 23-
38.
(6) Garraffo, H. M.; Spande, T. F.; J ones, T. H.; Daly, J . W. Rapid
Commun. Mass Spectrom. 1999, 13, 1-11.
(7) J ones, T. H.; Gorman, J . S. T.; Snelling, R. R.; Delabie, J . H. C., Blum,
M. S.; Garraffo, H. M.; J ain, P.; Daly, J . W.; Spande, T. F. J . Chem.
Ecol. 1999, 25, 1179-1193.
(25) Kim, S.; Kim, Y. G.; Kim, D.-I. Tetrahedron Lett. 1992, 33, 2565-
2566.
(26) Stowell, J . C. J . Org. Chem. 1976, 41, 560-561.
(8) Daly, J . W.; Spande, T. F.; Whittaker, N.; Highet, R. J .; Feigl, D.;
Nishimori, N.; Tokuyama, T.; Myers, C. W. J . Nat. Prod. 1986, 49,
265-280.
NP0005098