Journal of Chemical Ecology, Vol. 26, No. 1, 2000
0098-0331/ 00/ 0100-0073$18.00/ 0 ? 2000 Plenum Publishing Corporation
73
ARTHROPOD?FROG CONNECTION:
DECAHYDROQUINOLINE AND PYRROLIZIDINE
ALKALOIDS COMMON TO MICROSYMPATRIC
MYRMICINE ANTS AND DENDROBATID FROGS
JOHN W. DALY,1,* H. MARTIN GARRAFFO,1 POONAM JAIN,1
THOMAS F. SPANDE,1 ROY R. SNELLING,2 C ?ESAR JARAMILLO,3
and A. STANLEY RAND3
1Laboratory of Bioorganic Chemistry
National Institute of Diabetes and Digestive and Kidney Diseases
National Institutes of Health
Bethesda, Maryland 20892-0820
2Los Angeles County Museum of Natural History
900 Exposition Boulevard,
Los Angeles, California 90007
3Smithsonian Tropical Research Institute
P.O. Box 2072,
Panama?, Panama
(Received January 5, 1999; accepted August 18, 1999)
Abstract?Neotropical poison frogs (Dendrobatidae) contain a wide variety
of lipophilic alkaloids, apparently accumulated unchanged into skin glands
from dietary sources. Panamanian poison frogs (Dendrobates auratus) raised
in a large, screened, outdoor cage and provided for six months with leaf-litter
from the frog?s natural habitat, accumulated a variety of alkaloids into the
skin. These included two isomers of the ant pyrrolizidine 251K; two isomers
of the 3,5-disubstituted indolizidine 195B; an alkaloid known to occur in
myrmicine ants; another such indolizidine, 211E; two pyrrolidines, 197B
and 223N, the former known to occur in myrmicine ants; two tricyclics,
193C and 219I, the former known to occur as precoccinelline in coccinellid
beetles; and three spiropyrrolizidines, 222, 236, and 252A, representatives of
an alkaloid class known to occur in millipedes. The alkaloids 211E, 197B,
and 223N appear likely to derive in part from ants that entered the screened
cage. In addition, the frog skin extracts contained trace amounts of four
alkaloids, 205D, 207H, 219H, and 231H, of unknown structures and source.
Wild-caught frogs from the leaf-litter site contained nearly 40 alkaloids,
*To whom correspondence should be addressed.
DALY ET AL.74
including most of the above alkaloids. Pumiliotoxins and histrionicotoxins
were major alkaloids in wild-caught frogs, but were absent in captive-raised
frogs. Ants microsympatric with the poison frog at the leaf-litter site and at
an island site nearby in the Bay of Panama? were examined for alkaloids. The
decahydroquinoline (? )-cis-195A and two isomers of the pyrrolizidine 251K
were found to be shared by microsympatric myrmicine ants and poison frogs.
The proportions of the two isomers of 251K were the same in ant and frog.
Key Words?Alkaloids, coccinellines, decahydroquinolines, indolizidines,
pyrrolidines, pyrrolizidines, dendrobatid frogs, myrmicine ants, coccinellid
beetles, millipedes.
INTRODUCTION
The nature of the dietary sources of the diverse lipophilic alkaloids detected in
extracts of skin of neotropical dendrobatid (Dendrobates, Epipedobates, Miny-
obates, and Phyllobates) frogs remains a major research challenge for chemi-
cal ecologists (Daly, 1998; Daly et al., 1997a). Such lipophilic alkaloids, when
detected in skin extracts from mantelline (Mantella) frogs of Madagascar,
bufonid (Melanophryniscus) toads of southeastern South America, and myo-
batrachid (Pseudophryne) frogs of Australia, undoubtedly also have a dietary
origin. Dendrobatid (Dendrobates, Epipedobates, and Phyllobates) and man-
telline (Mantella) frogs raised in captivity on fruit flies and crickets contain
no detectable alkaloids in skin extracts, but have the ability to efficiently and
selectively accumulate into skin dietary alkaloids provided to them (Daly et al.,
1980, 1992, 1994a,b, 1997b). A total of over 500 alkaloids that encompass over
20 structural classes have been detected in frog/ toad skin extracts. Because of
this large number, the individual alkaloids have been designated with boldface
numbers, representing the nominal mass, and, when necessary, with code letters
to distinguish those of the same nominal mass (for a current survey, see Daly et
al., 1999). Structures of each alkaloid class and certain individual alkaloids of
the present study are depicted in Figure 1.
The dendrobatid frogs of the genus Dendrobates are known as ant spe-
cialists (Toft, 1980, 1995; Donnelly, 1991; Caldwell, 1996) and indeed rep-
resentatives of six of the more than 20 structural classes of lipophilic alka-
loids found in frog skin are now known to occur in extracts of myrmicine ants
(Jones et al., 1996, 1999; Spande et al., 1999; Daly et al., 1999 and references
therein). These classes are the 2,5-dialkylpyrrolidines, 2,6-dialkylpiperidines,
3,5-dialkylpyrrolizidines, 3,5-dialkylindolizidines, 4,6-dialkylquinolizidines, and
2,5-dialkyldecahydroquinolines. In addition to such ?ant alkaloids,? there are two
other classes of alkaloids found in frog skin that are known to occur in arthro-
pods, namely the tricyclic coccinelline class, known from coccinellid beetles (see
Daly et al., 1999 for references) and the spiropyrrolizidine class, known from a
DIETARY ANT-FROG ALKALOIDS 75
FIG. 1. Structures of various classes of alkaloids and certain individual alkaloids detected
in Panamanian poison frogs (Dendrobates auratus) as documented in Table 1. For further
structural details, including relative or absolute configuration of certain alkaloids and the
nature of the R substituents, see Daly et al. (1999).
millipede (Meinwald et al., 1975). Of the eight structural classes (a?h) of alka-
loids now known from frogs and arthropods, only a few individual alkaloids are
known from both frogs and arthropods as in the following tabulation (see Daly
et al., 1999):
Class a. Of the nearly twenty 2,5-disubstituted pyrrolidines reported from
myrmicine ants only five have been detected in frog skin extracts, while the re-
maining 11 from frog skin extracts have not been reported from ants.
Class b. Of the 20 or more 2,6-disubstituted piperidines found in myr-
micine ants, only three have been detected in frog skin extracts, while the remain-
ing 6 from frog skin extracts have not been reported from ants.
DALY ET AL.76
Class c. Of the five 3,5-disubstituted pyrrolizidines previously reported
from ants, three have been detected in frog skin extracts, while the remaining
18 from frog skin extracts have not been reported from ants.
Class d. Of the 10 or more 3,5-disubstituted indolizidines found in myr-
micine ants, only four also have been detected in frog skin extracts (see Jones
et al., 1996; Daly et al., 1999), while the remaining 16 from frog skin extracts
have not been reported in ants.
Class e. The one 4,6-disubstituted quinolizidine found in a myrmicine ant
(Jones et al., 1999) also has been detected in frog skin extracts; one other such
quinolizidine has been detected in frog skin.
Class f. Of the four 2,5-disubstituted decahydroquinolines recently reported
from myrmicine ants (Jones et al., 1999; Spande et al., 1999), two have been
detected in frog skin extracts, but nearly 40 other decahydroquinolines from frog
skin extracts have not been reported from myrmicine ants.
Class g. Of the five spiropyrrolizidines detected in frog skin extracts, only
nitropolyzonamine has been reported from millipedes. A related spiropyrrolidine,
polyzonimine, has been reported from both frog skin extracts and a millipede.
Class h. Only two of the tricyclic coccinelline class alkaloids reported in
beetles have been identified in frog skin extracts, but there are 15 or more other
tricyclic alkaloids in frog skin alkaloids that may prove to be of the coccinelline
class (see Daly et al., 1999).
The occurrence of the above eight structural classes in frogs and arthropods
includes data from the present paper and is summarized in Table 1. The present
report of trans-251K in ants brings to 22 the number of alkaloids known to
occur in both frogs and arthropods. Thus, the vast majority of over 500 alkaloids
detected in frog skin extracts have not yet been identified from a possible dietary
source. Furthermore, until the present report, none of the 22 alkaloids had been
identified in arthropods from a region where the frogs occur.
Arthropods obtained with Berlese funnels from leaf-litter have been re-
ported to provide a variety of alkaloids to captive-raised poison frogs (Dendro-
bates auratus) in Panama? (Daly et al., 1994b). No known myrmicine ant alka-
loids were detected, but one beetle alkaloid was detected. In the present study,
arthropods, contained in the fresh leaf-litter that was placed in large, screened
outdoor cages during raising of the poison frog Dendrobates auratus, provided
14 of the 40 alkaloids found in skin extracts of wild-caught frogs from the same
leaf-litter site. Ant alkaloids, a beetle alkaloid, and putative millipede alkaloids
were among those detected in the skin extracts of the captive-raised frogs. Two
ant alkaloids, both pyrrolidines, were detected only in the captive-raised frogs.
Pumiliotoxins, an allopumiliotoxin, and histrionicotoxins were major skin alka-
loids in the wild-caught frogs but were not detected in captive-raised frogs pro-
vided with leaf-litter as the sole source of dietary arthropods. We also report the
occurrence of a decahydroquinoline and a pair of diastereomeric pyrrolizidines
DIETARY ANT-FROG ALKALOIDS 77
TABLE 1. SUMMARY OF ALKALOIDS OF EIGHT STRUCTURAL CLASSES REPORTED FROM
ARTHROPODS AND/ OR FROG SKIN EXTRACTSa
Reported from
Arthropods Only Only
Structural class and frog skin arthropods frog skin
2,5-Disubstituted pyrrolidines 5 ?15b 11
2,6-Disubstituted piperidines 3 ?17b 16
3,5-Disubstituted pyrrolizidines 3 2b 18
3,5-Disubstituted indolizidines 4 ?6b 16
4,6-Disubstituted quinolizidines 1b 0 1
2,5-Disubstituted decahydroquinolines 2 2b ?40
Spiropyrrolizidines + spiropyrrolidines 2c 0 4
Tricyclic coccinelline class 2 >10d 15
aIncludes present report (see Daly et al., 1999).
bMyrmicine ants.
cMillipedes.
dCoccinellid beetles.
that occur in myrmicine ants and in skin extracts of microsympatric poison
frogs.
METHODS AND MATERIALS
Tadpoles of the poison frog Dendrobates auratus were collected near the
entrance to Quarry Heights on Cerro Anco?n (Ancon Hill), in the city of Panama?,
from late November 1993 until January 15, 1994. Fourteen transformed frogs
were obtained from January 26 to March 16, 1994. The frogs were raised for
five to seven months as follows at the base of Cerro Anco?n behind the Tupper
building of the Smithsonian Tropical Research Institute.
Four juvenile frogs were placed on January 26 in an outdoor cage (1.8 m
wide ? 2.8 m long ? 2.3 m high) screened with metallic mosquito netting to
impede access of insects from the outside. Three frogs survived until they were
killed for analysis of skin alkaloids on September 11, 1994 [weight 2.1?2.3 g,
snout?vent length (SVL) 2.8?3.0 cm]. The frogs? diet in this control cage was
provided by an excess of fruit flies (Drosophila) in the cage, breeding on bananas
and Drosophila medium. Leaf litter (frozen for two weeks to eliminate insects)
placed along the inside borders of the cage was sprayed daily with water to
maintain humidity. Although some small insects did appear in the cage during
the experiment, the main source of food was the excess of breeding fruit flies.
Six juvenile frogs were placed during the period of February 10 to March
16 in another outdoor cage, also screened with metallic mosquito netting. Four
frogs survived until sacrifice for analysis of skin alkaloids on September 11,
DALY ET AL.78
1994 (weight 1.4?1.8 g, SVL 2.7?2.9 cm). A fifth also survived but was very
emaciated (weight 1.1 g, SVL 2.6 cm) and was not included in the analysis.
The frogs? diet in this case was provided each week by fresh leaf litter gathered
on Cerro Anco?n from the area where Dendrobates auratus were common and
where the tadpoles were collected. Wingless fruit flies were used to supplement
the diet every two weeks. Insects from the area surrounding the cage undoubtedly
contributed to the diet to some extent.
Four juvenile frogs were introduced on January 31 into a large glass terrar-
ium (30 gallon) placed in an adjacent cage and sealed to restrict access of insects
from the outside. Two frogs survived until sacrifice for analysis of skin alkaloids
on September 11, 1994 (weight 1.6 and 1.8 g, SVL 2.5 and 2.5 cm). The frogs?
diet in this terrarium consisted entirely of wingless Drosophila provided every
second day.
Two adult Dendrobates auratus from the site on Cerro Anco?n were sacri-
ficed on September 29, 1994 (weight 1.7 and 2.7 g, SVL 2.9 and 3.2 cm) for
analysis of skin alkaloids.
From June 20 until August 25, 1996, fifty separate ant colonies were col-
lected near the entrance to Quarry Heights on Cerro Anco?n at the site where
Dendrobates auratus were common. Winged ants were obtained from only
15 of these colonies. On August 29, 1996, eleven separate ant colonies were
collected on Isla Taboga near a site where adult Dendrobates auratus previ-
ously had been collected (Daly et al., 1994b) for analysis of skin alkaloids.
No winged ants were collected on Isla Taboga. Ants were placed in a small
volume (?0.5 ml) of methanol in a NUNC vial for detection of alkaloids by
gas chromatographic?mass spectral analysis with a Finnigan model 800 ion trap
detector, interfaced with a Varian model 3400 gas chromatograph fitted with a
30 m ? 0.32 mm ID, RTX 5 (Restek) fused silica-bonded column programmed
from 1008 to 2808C at the rate of 108C/min.
Frog skins were extracted and alkaloid fractions were prepared and ana-
lyzed by gas chromatography?mass spectrometry and vapor-phase infrared spec-
troscopy (see Daly et al., 1994b and references therein). The gas chromatograms
depicted in Figure 2 were obtained with a 6-ft 1.5% OV-1 packed column (2 mm
ID) with a flame ionization detector. A sample of 2 ml of a methanolic alkaloid
fraction equivalent to 2 mg wet weight skin was injected at a column temper-
ature of 1508C. After the solvent maximum passed (0.3 min), the column was
programmed to 2808C at 108C/min.
Certain methanolic ant extracts and frog alkaloid fractions were analyzed
on a chiral Supelco Beta-DEX 120 fused silica capillary column (30 m ? 0.25
mm ID, 0.25-mm film thickness) at 1508C.
Voucher specimens of frogs and ants are deposited in the American Museum
of Natural History, New York, New York, and in the Los Angeles County
Museum of Natural History, Los Angeles, California, respectively.
DIETARY ANT-FROG ALKALOIDS 79
FIG. 2. Gas chromatographic profiles for alkaloids in skins of poison frogs (Dendrobates
auratus). (A) Terrarium raised frog [snout?vent length (SVL) 25 mm] with diet of wing-
less fruit flies. (B) Outdoor cage-raised frog (SVL 30 mm) on diet mainly of fruit flies.
(C) Outdoor cage-raised frog (SVL 26 mm) on diet from fresh leaf litter. (D) Outdoor
cage-raised frog (SVL 27 mm) on diet from fresh leaf litter. (E) Wild-caught frog (SVL
32 mm) from Cerro Anco?n. Peaks marked with x are nonalkaloidal artifacts.
RESULTS AND DISCUSSION
One of the frogs raised in the large outside glass terrarium with a diet of
only wingless fruit flies had no detectable alkaloids in the skin extract (see Figure
DALY ET AL.80
2A), as expected from previous studies (Daly et al., 1992, 1994a,b). However,
the other frog did have a trace amount of indolizidine 195B (data not shown),
an alkaloid presumably obtained from a myrmicine ant. It has been noted that
exclusion of small insects from the terrarium was not completely successful.
Frogs raised in the screened outdoor cage in which fruit flies were by far
the most abundant food source contained as a major alkaloid in skin extracts the
3-butyl-5-methyl indolizidine 195B, an alkaloid known to occur in myrmicine
ants as monomorine I. Traces of another isomer of 195B and another indolizidine
211E were present, along with trace amounts of two 2,5-disubstituted pyrro-
lidines, 197B, and 223N, all of which presumably are derived from small myr-
micine ants, which, along with other small insects, were able to penetrate the
control cage despite the mosquito screen. In addition, such ants may have been
targeted to some extent as prey items, since the frogs in the control cage were
provided with an abundance of fruit flies as food. The flame ionization gas chro-
matogram for one of three frogs is shown in Figure 2B. The chromatograms for
the other two frogs were similar. The occurrence of alkaloids in skin extracts
from these captive-raised frogs is presented in Table 2.
The frogs raised in the screened outdoor cage that was provided each week
with fresh leaf litter from Cerro Anco?n as a source of dietary arthropods had at
least 16 different alkaloids in skin extracts. The chromatograms for two frogs are
shown in Figures 2C and 2D. The chromatograms for the other two frogs were
similar to these, but with lesser amounts of alkaloids. The major alkaloid in all
four frogs was the spiropyrrolizidine oxime ether 236, an alkaloid presumably
obtained from a millipede. In lesser amounts there were the indolizidine 195B
and an isomer and two diastereomers of the pyrrolizidine 251K, presumably all
from myrmicine ants; the tricyclic 193C (precoccinelline), presumably from a
coccinellid beetle; another tricyclic 219I; and four alkaloids, 205D, 207F, 219H,
and 231H, of unknown structural class. Trace amounts of another indolizidine
211E, two pyrrolidines 197B and 223N, and two other spiropyrrolizidine oximes
222 and 252A were present (see Table 2).
One wild-caught frog from Cerro Anco?n had over 40 alkaloids in its skin
extract (Figure 2E). The other wild-caught frog had a similar profile, but at
about twofold higher levels (data not shown). Thirty six of the major, minor,
and trace alkaloids from these frogs are tabulated in Table 2. Other alkaloids
were present in such trace amounts that they could not be well characterized.
The major alkaloids were histrionicotoxins. These were not detected in the frogs
raised in outside cages provided with fresh leaf litter from Cerro Anco?n, nor were
the pumiliotoxins 307A and 323A, which were significant alkaloids from skin of
the wild-caught frogs. The spiropyrrolizidine 236, which was the major alkaloid
in captive-raised frogs, was a minor alkaloid in the wild-caught frogs. The two
isomers of pyrrolizidine 251K, the indolizidine 195B, the tricyclics 193C and
219I, and alkaloid 219H, all of which were minor alkaloids in the captive-raised
DIETARY ANT-FROG ALKALOIDS 81
TABLE 2. ALKALOIDS IN SKIN EXTRACTS OF PANAMANIAN POISON FROG
(Dendrobates auratus)
I II Wild-caught
Captive-raised Captive-raised in
in outdoor cages outdoor cages on III, IV,
Alkaloida on Drosophila leaf litter arthropods Ancon Hill Isla Tabogab
Histrionicotoxins
235A +
259A + +
261A +
283A +++ +
285A +++
285C +++
287A ++
Pumiliotoxins
251D + +
277B +
307A ++ +
307B + +
323A ++ ++
Allopumiliotoxins
253A +
267A ++
305Ac +
323B ++ +
339A +
2,5-Decahydroquinolines
cis-195A ++ ++
211A +
219A +++, +
243A + ++, +
269AB +
3,5-Pyrrolizidines
249I +
cis-251K ++ ++ ++
trans-251K + + +
265Jc +
3,5-Indolizidines
195B +++, + ++, + ++, +
211E + +
5,8-Indolizidines
203A +
235B ++ +
5,6,8-Indolizidines
195G +
223A +
DALY ET AL.82
TABLE 2. CONTINUED
I II Wild-caught
Captive-raised Captive-raised in
in outdoor cages outdoor cages on III, IV,
Alkaloida on Drosophila leaf litter arthropods Ancon Hill Isla Tabogab
233G +
249H + ++
263D +
277C +
Gephyrotoxins
287C +
Spiropyrrolizidines
222 + + +
236 +++ ++ +
252A + + +
Tricyclics
193C ++ + +
205B ++
207Jc +
219Ic ++ ++
2,5-Pyrrolidines
trans-197B + +
223N + +
Unclassified
205Dc + +
207Fc ++ +
217Cc +
219Hc ++ +
231Hc + +
243Ec +
271Cc +
aFor structures see Figure 1 and Daly et al. (1999). The numbers preceding the ring system indicate
the positions of substitution. The relative amounts (+++ c major, ++ c minor, + c trace) follow
the notation used in prior publications. For example, the major alkaloid 195B in column I (Figure
2B) was present in amounts of about 80 mg/ frog. The major alkaloid 236 of column II (Figure 2C)
was present in amounts of about 20 mg/ frog. The major alkaloids 283A, 285A, 285C of column
III (Figure 2E) were present in amounts of 150?200 mg/ frog. The major alkaloid 219A of column
IV was present in amounts of about 50 mg/ frog (see Daly et al., 1994b). The trace alkaloids were
present in amounts of 1 mg or less per frog, while the minor alkaloids were present in significantly
lower amounts than the major alkaloid(s) in each extract. Where two entries are given and separated
by commas, this indicates two isomers in the order of elution from the gas chromatograph.
bData are from Daly et al. (1994b). Some of the alkaloids have been reclassified in light of additional
data.
cAlkaloids of undefined or incompletely defined structures.
DIETARY ANT-FROG ALKALOIDS 83
frogs, were also minor alkaloids in wild-caught frogs. However, several of the
minor alkaloids of the wild-caught frogs, most notably the decahydroquinoline
195A, the tricyclic 205B, the allopumiliotoxin 323B, and the 5,8-disubstituted
indolizidine 235B were not detected in the captive-raised frogs (see Table 2).
Conversely, the two pyrrolidines, 197B and 223N, detected as trace alkaloids in
extracts from both groups of frogs raised in the outdoor cages, were not detected
in the wild-caught frogs.
There are significant differences between the present results in which poison
frogs (Dendrobates auratus) from Cerro Anco?n were raised in large, screened,
outdoor cages on arthropods provided each week on fresh leaf litter from Cerro
Anco?n and previous results in which frogs of the same species were raised on
arthropods driven into indoor terraria with heat lamps from Berleze funnels con-
taining leaf litter from Cerro Anco?n (Daly et al., 1994b). In both studies the puta-
tive millipede alkaloid 236 was the major alkaloid in skin extracts of captive-
raised frogs. The tricyclic beetle alkaloid 193C (precoccinelline) was a minor
alkaloid in both studies, as were the tricyclic 219I and alkaloid 219H. The ant
alkaloids 195B and 251K of the present study were not detected in the Berleze
funnel paradigm, suggesting that myrmicine ants are not readily obtained from
leaf litter with that paradigm and/ or that the myrmicine ants were able to enter
the outdoor cages despite the screening. The latter explanation in all likelihood
accounts for the presence of indolizidine 195B, which was the major skin alka-
loid in frogs raised in the screened outdoor cage on an abundance of fruit flies
(Figure 2B). The most inexplicable difference between the two studies is that
the 19-carbon histrionicotoxins, a 19-carbon gephyrotoxin, and a 19-carbon dec-
ahydroquinoline were detected as minor alkaloids in one frog and as trace alka-
loids in two other frogs raised in terraria with the Berleze funnel paradigm but
were not detected at all in captive-raised frogs of the present study. It might
be noted that the earlier results were with frogs raised in terraria on arthropods
from leaf litter obtained from June onwards until death of the frogs seven months
later in December, 1992 (Daly et al., 1994b). The present results were with frogs
raised in outdoor cages provided with fresh leaf litter from February through
August 1994. Thus, in addition to the difference of Berleze funnel extraction
versus outdoor cages, the arthropods of the earlier study came more from the
rainy season than those of the present study. The levels and range of alkaloids
found in the skin of dendrobatid frogs presumably reflect many factors, such as
the availability and suitability of alkaloid containing prey items, prey selection
by the frog species, and the selectivity and degree of expression of the alkaloid
sequestering system in each frog species.
In an effort to identify possible sources of the 16 alkaloids found in skin of
the captive-raised frogs, ants from a total of 61 terrestrial nests were analyzed,
both from the leaf litter area on Cerro Anco?n, where the poison frog Dendro-
bates auratus is common, and from an island, Isla Taboga, in the nearby Bay
DALY ET AL.84
of Panama?, where Dendrobates auratus is also common. The profiles of alka-
loids in wild-caught Dendrobates auratus from these two sites differ consid-
erably (see Daly et al., 1994b, and Table 2). From the 50 ant nests collected
on Cerro Anco?n, the alate queens, but not the wingless workers, from one nest
were found to contain the decahydroquinoline cis-195A, which was also a minor
alkaloid component in skin extracts from the wild-caught Dendrobates auratus
from Cerro Anco?n (Figure 2D). Both ant and frog contained the (? )-enantiomer
of cis-195A as shown by gas chromatographic analysis on a chiral capillary col-
umn that separates the two enantiomers. The ant was identified as a Solenopsis
(Diplorhoptrum) sp. Alate queens of the same species from another nest also con-
tained decahydroquinoline cis-195A. Ant extracts from the remaining 48 nests
did not contain alkaloids. The species were not identified. This and the following
example represent the first reported instances of an alkaloid present in skin of
a dendrobatid frog and in a microsympatric arthropod available to the frog as a
prey item. The other example is from Isla Taboga, where wingless ants, identified
as Megalomyrmex silvestri Wheeler, from two of 11 nests were found to contain
two diastereomers of the 3,5-disubstituted pyrrolizidine 251K in the same ratio
(cis-251K?trans-251K, 3 : 1) as in skin extracts from microsympatric Dendro-
bates auratus on Isla Taboga. 5E,9E-3-Butyl-5-hexylpyrrolizidine (trans-251K)
has been previously reported from a myrmicine ant, Megalomyrmex foreli, of
Costa Rica (Jones et al., 1996) at a site at which dendrobatid frogs do not occur.
Extracts from ants of the other nine nests on Isla Taboga contained no alkaloids,
nor were the ants identified.
In summary, fresh leaf litter, collected mainly during the dry season in
Pacific Panama?, apparently can provide dietary sources for some of the many
alkaloids detected in skin extracts of poison frogs (Dendrobates auratus) found
at the same site. The major classes of alkaloids provided by the leaf-litter arthro-
pods are pyrrolizidines and indolizidines known from myrmicine ants, a tricyclic
known from coccinellid beetles, and spiropyrrolizidines, putatively from mil-
lipedes. In addition, one species of myrmicine ants collected at the leaf-litter
site was found to contain the decahydroquinoline (? )-cis-195A, which was also
present as a minor alkaloid in skin of the microsympatric population of the poi-
son frog Dendrobates auratus, while another species of myrmicine ants from a
small island contained a cis- and a trans-pyrrolizidine 251K in the same 3 : 1
ratio that was present in skin of a microsympatric population of the poison frog
Dendrobates auratus.
Acknowledgments?The Instituto Nacional de Recursos Naturales Renovables of Panama is
gratefully acknowledged for permission to collect specimens necessary for these studies. The authors
thank Roberto Iba?n?ez for help in collecting the tadpoles and for suggestions during the course of
the experiments.
DIETARY ANT-FROG ALKALOIDS 85
REFERENCES
CALDWELL, J. P. 1996. The evolution of myrmecophagy and its correlates in poison frogs (family
Dendrobatidae). J. Zool. (London) 240:75?101.
DALY, J. W. 1998. Thirty years of discovering arthropod alkaloids in amphibian skin. J. Nat. Prod.
61:162?172.
DALY, J. W., MYERS, C. W., WARNICK, J. E., and ALBUQUERQUE, E. X. A. 1980. Levels of
batrachotoxin and lack of sensitivity of its action in poison-dart frogs (Phyllobates). Science
208:1383?1385.
DALY, J. W., SECUNDA, S. I., GARRAFFO, H. M., SPANDE, T. F., WISNIESKI, A., NISHIHARA, C., and
COVER, J. F., JR. 1992. Variability in alkaloid profiles in neotropical poison frogs (Dendrobati-
dae): Genetic versus environmental determinants. Toxicon 30:887?898.
DALY, J. W., SECUNDA, S. I., GARRAFFO, H. M., SPANDE, T. F., WISNIESKI, A., and COVER, J. F.,
JR. 1994a. An uptake system for dietary alkaloids in poison frogs (Dendrobatidae). Toxicon
32:657?663.
DALY, J. W., GARRAFFO, H. M., SPANDE, T. F., JARAMILLO, C., and RAND, A. S. 1994b. Dietary
source for skin alkaloids of poison frogs (Dendrobatidae)? J. Chem. Ecol. 20:943?955.
DALY, J. W., GARRAFFO, H. M., and MYERS, C. W. 1997a. The origin of frog skin alkaloids: An
enigma. Pharm. News 4:9?14.
DALY, J. W., GARRAFFO, H. M., HALL, G. S. E., and COVER, J. F., JR. 1997b. Absence of skin
alkaloids in captive-raised Madagascan mantelline frogs (Mantella) and sequestration of dietary
alkaloids. Toxicon 35:1131?1135.
DALY, J. W., GARRAFFO, H. M., and SPANDE, T. F. 1999. Alkaloids from amphibian skins, pp. 1?161,
in S. W. Pelletier (ed.). Alkaloids, Chemical and Biological Perspectives, Vol. 13. John Wiley
& Sons, New York.
DONNELLY, M. A. 1991. Feeding patterns of the strawberry poison frog, Dendrobates pumilio
(Anura: Dendrobatidae). Copeia 3:723?730.
JONES, T. H., TORRES, J. A., SPANDE, T. F., GARRAFFO, H. M., BLUM, M. S., and SNELLING, R. R.
1996. Chemistry of venom alkaloids in some Solenopsis (Diplorhoptrum) species from Puerto
Rico. J. Chem. Ecol. 22:1221?1236.
JONES, T. H., GORMAN, J. S. T., SNELLING, R. R., DELABIE, J. H. C., BLUM, M. S., GARRAFFO, H.
M., JAIN, P., DALY, J. W., and SPANDE, T. F. 1999. Further alkaloids common to ants and frogs:
Decahydroquinolines and a quinolizidine. J. Chem. Ecol. 25:1179?1193.
MEINWALD, J., SMOLANOFF, J., MCPHAIL, A. T., MILLER, R. W., EISNER, T., and HICKS, K. 1975.
Nitropolyzonamine: A spirocyclic nitro compound from the defensive glands of a milliped
(Polyzonium rosalbum). Tetrahedron Lett. 1975:2367?2370.
SPANDE, T. F., JAIN, P., GARRAFFO, H. M., PANNELL, L. K., YEH, H. J. C., DALY, J. W., FUKUMOTO,
S., IMAMURA, K., TOKUYAMA, T., TORRES, J. A., SNELLING, R. R., and JONES, T. H. 1999.
Structures of decahydroquinolines from dendrobatid poison frogs and a myrmicine ant: Use of
1H- and 13C-NMR in conformational analysis. J. Nat. Prod. 62:5?21.
TOFT, C. A. 1980. Feeding ecology of thirteen syntopic species of anurans in a seasonal tropical
environment. Oecologia 45:131?141.
TOFT, C. A. 1995. Evolution of diet specialization in poison-dart frogs (Dendrobatidae). Herpeto-
logica 51:202?216.