Melyrid beetles (Choresine): A putative source for the
batrachotoxin alkaloids found in poison-dart frogs
and toxic passerine birds
John P. Dumbacher*??, Avit Wako?, Scott R. Derrickson*, Allan Samuelson?, Thomas F. Spande
, and John W. Daly

*Smithsonian Conservation Research Center, Front Royal, VA 22630; ?Department of Birds and Mammals, California Academy of Sciences, San Francisco,
CA 94103; ?Herowana Village, Eastern Highlands Province, Papua New Guinea; ?Entomology, Bernice P. Bishop Museum, Honolulu, HI 96813; and

Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Department of
Health and Human Services, Bethesda, MD 20892
Contributed by John W. Daly, September 28, 2004
Batrachotoxins are neurotoxic steroidal alkaloids first isolated
from a Colombian poison-dart frog and later found in certain
passerine birds of New Guinea. Neither vertebrate group is
thought to produce the toxins de novo, but instead they likely
sequester them from dietary sources. Here we describe the pres-
ence of high levels of batrachotoxins in a little-studied group of
beetles, genus Choresine (family Melyridae). These small beetles
and their high toxin concentrations suggest that they might
provide a toxin source for the New Guinea birds. Stomach content
analyses of Pitohui birds revealed Choresine beetles in the diet, as
well as numerous other small beetles and arthropods. The family
Melyridae is cosmopolitan, and relatives in Colombian rain forests
of South America could be the source of the batrachotoxins found
in the highly toxic Phyllobates frogs of that region.
Pitohui  Ifrita  Phyllobates  APCI mass spectrometry  dietary arthropods
Batrachotoxins (BTXs) were discovered in the mid-1960s inskin extracts from a Colombian poison-dart frog (family
Dendrobatidae) (1, 2). The name for these unique steroidal
alkaloids was derived from the Greek ??batrachos,?? meaning
frog. During the following 25 years, BTXs were detected only in
frogs of the dendrobatid genus Phyllobates and not in scores of
other poison frogs (3). Only three species are toxic enough to be
used by native Americans for poisoning blow-dart tips: Phyllo-
bates terribilis, Phyllobates bicolor, and Phyllobates aurotaenia (4)
where BTXs are responsible for their toxicity. BTXs bind with
high affinity to voltage-gated sodium channels in nerve and
muscle membranes, locking them in an open state (5). Because
of their potency and specificity, batrachotoxins have been widely
used to study the function of sodium channels (6).
Poison-dart frogs that have been raised in captivity do not
contain detectable amounts of BTXs (7). This and other lines of
evidence indicate that dendrobatid poison frogs do not produce
batrachotoxins or other alkaloids de novo but that they likely
sequester such alkaloids from their diet (8, 9). Although other
frog-skin alkaloids have been identified in ants, millipedes, and
beetles (10?12), BTXs have not been found before in any insect
or plant source. Restrictions on field-work in Colombia have
prevented any search for the dietary origin of batrachotoxins in
the poison-dart frogs of that country. Thus, a putative dietary
source for the BTXs remained a mystery.
In 1992, the toxic principle from feathers and skin of bird
species of genus Pitohui, endemic to New Guinea, was isolated
and, remarkably, proved to be mainly a batrachotoxin, homo-
BTX (h-BTX) (13). BTXs were later found in New Guinea toxic
birds of the genus Ifrita (14). On a weight basis, the batracho-
toxins are among the most toxic natural substances known, being
250-fold more toxic than strychnine (4). It is believed that such
toxins provide some protection against the birds? natural ene-
mies, such as parasites (15, 16) and predators, including humans
(17, 18).
We now report analysis of extracts of the beetle genus
(Choresine: family Melyridae) suggested by New Guinean tra-
ditional village naturalists as a potential source of BTXs in
Pitohui and Ifrita. Villagers from Herowana (Crater Mountain
Wildlife Management Area, Eastern Highlands Province) iden-
tified beetles locally known as ??nanisani.?? Nanisani is also the
local name for the Blue-Capped Ifrita (Ifrita kowaldi), which
does carry BTXs. According to these local naturalists, the name
??nanisani?? refers specifically to the unusual tingling and numb-
ing sensation to the lips and face caused by contact with either
the beetles or the bird feathers.
Materials and Methods
Collected beetles were stored in glass vials in 100% ethanol for
transport at ambient temperatures to the National Institutes of
Health for analysis. The supernatant was withdrawn, and the
beetles were washed with 
1 ml of ethanol. This process was
repeated twice more, and the washes were combined with the
original supernatant. These washes were shown to extract vir-
tually all BTXs. Sonication (see footnote to Table 1) proved to
be superfluous in removing additional BTXs. The ethanol
extract was concentrated to 
200 
l under a nitrogen stream.
Single-beetle collections were concentrated to 10 
l. Beetles
were identified by A.S. and found to consist of four Choresine
species (see below). Voucher specimens were deposited in the
Bernice P. Bishop Museum.
A Hewlett?Packard 1100 liquid chromatograph with variable
wavelength detector set at 260 nm (the long-wavelength absorp-
tion maximum of the pyrrole moiety of BTXs), fitted with an
AQUA (Phenomenex, Torrance, CA) C-18 RP column (4.6
mm  25 cm; 5-
m particle size) and using a water?acetonitrile
gradient (90% water0.05% HOAc to 50% water0.05% HOAc
in 40min; 0.5 mlmin), was interfaced with a Finnigan LCQmass
spectrometer operating in the atmospheric pressure chemical
ionization mode. Interpretation of mass spectra and identifica-
tion of the BTX alkaloids are presented in greater detail
elsewhere (cf. ref. 14).
Results
Alkaloid profiles of the beetles surveyed are presented in Table
1. Characteristic or major mass spectral peaks, [MH], rep-
resented are the rearranged BTX-A acetate (mz 460), batra-
chotoxinin A (BTX-A) (mz 418), and the BTX-A crotonate
(mz 486), in an 
4:2:1 ratio in most collections. The BTX-A
crotonate usually emerged as two overlapping peaks of varying
proportions, corresponding to the trans and cis BTX-A O-
crotonate esters with the trans usually equivalent to or slightly
Abbreviations: BTX, batrachotoxin; h-BTX, homo-batrachotoxin; BTX-A, batrachotoxi-
nin A.
?To whom correspondence should be addressed. E-mail: jdumbacher@calacademy.org.
? 2004 by The National Academy of Sciences of the USA
www.pnas.orgcgidoi10.1073pnas.0407197101 PNAS  November 9, 2004  vol. 101  no. 45  15857?15860
CH
EM
IS
TR
Y
predominating over cis. Minor peaks were present for BTX (mz
539) and h-BTX (mz 553). Structures of these alkaloids are
shown in Fig. 1. In addition, several unknown BTX alkaloids
were detected.
Of the 391 beetles present in collections, 85% were Choresine
pulchra (Pic) (CP) and 14%wereChoresine semiopaca (Wittmer)
(CS). There were also four Choresine sp. A (CA) and one
Choresine rugiceps (Wittmer) (CR). Beetles were collected from
May to September 2001 (mainly August) and from May to
November 2002 (mainly November). All were collected in or
within 2 km of the village of Herowana (elevation, 
1,400 m).
Some collections were made within the village, and others were
made along the nearby Fio River. There is no discernable
difference in collections made in 2001 or 2002 in the proportions
of CS in the collections. A total of 40 separate collections were
analyzed. A summary of the alkaloid data for the following
collection types (I?VI) is presented in Table 1 (collections were
segregated as types by species present in the beetle mixtures,
although no evident pattern emerged in BTXs detected): type I,
29 separate collections of CP beetles; type II, 5 separate collec-
tions of CP
CS; type III, 3 separate collections of CP and CA
or CR; type IV, one 3-beetle collection of CS (which originated
within the village); type V, one single-beetle collection of CP;
type VI, one 17-beetle collection (not positively identified as to
species but likely all CP). One single-beetle collection of CP from
November 2002 of type I had no detectable BTXs, whereas the
single-beetle collection of CP (type V) from July 2001 had major
amounts of the crotonate ester of BTX-A. Alkaloids were
characterized by their protonated molecular ions [MH],
fragmentations, and liquid chromatography-retention times.
Many had not been detected in prior studies of passerine birds
of Papua New Guinea (13, 14).
Several previously uncharacterized BTX alkaloids are re-
ported in Table 1 that had not been seen either in frog or bird:
a molecular weight (MW) 524 alkaloid that lacks the N-methyl
from the homomorpholine ring of the BTX nucleus; a MW
435437 chloro analog of BTX-A, where the 20-hydroxyl group
is replaced by chlorine; a MW 499 material, likely a homocro-
tonate of BTX-A, fragmenting to mz 400; and two alkaloids of
MW 513, one giving an mz 400 fragment and the other giving
a fragment atmz 428 (perhaps an ethyl ketal of theO-crotonate
of BTX-A) and perhaps a pyrrole-hydroxylated BTX (MW 564)
and pyrrole-ethoxylated BTX (MW 582) or h-BTX (MW 596).
Additional characterization will be required. Some of these
alkaloids are likely to be artifacts.
It initially appeared that C. semiopaca (collection type IV)
might be a richer source of batrachotoxin than other Choresine
species. In a collection of type II where C. semiopaca was mixed
with C. pulchra (25 CS and 28 CP, August 2001), BTX was a
major alkaloid and h-BTX was also detected. Another mixed
collection of type II (32 CS and 9 CP, September 2001), however,
had no detectable BTX or h-BTX.
Discussion
Whereas the birds always had h-BTX  BTX, the situation is
reversed in the beetles, where h-BTX is scarcely detectable or
absent. In frogs of the genus Phyllobates, BTX levels exceed
those of h-BTX by 
2-fold. The level of BTX in one collection
of three C. semiopaca was found to be 1.8 
g per beetle or
roughly one-tenth the amount present in a complete skin of
either P. bicolor or P. aurotaenia but far less than that found in
a P. terribilis skin, which has ca. 1 mg. The amounts of BTX in
beetle collections were, however, extremely variable. Our atmo-
spheric pressure chemical ionization mass spectrometer was
sufficiently sensitive, even in the full scan mode (mz 50?900
atomic mass units), to detect BTXs in a single beetle.
Beetles are not known to synthesize steroidal skeletons, such
as that found in the BTXs, although they do use the mevalonate
pathway to form terpenes. Instead, steroids and triterpenoids in
beetles are presumed to originate from plant phytosterols (cho-
lesterol, stigmasterol, sitosterol, ergosterol, etc.), which are
modified by the beetle or possibly by a symbiont or collection of
Table 1. Summary of alkaloids detected in collections of Choresine beetles
Collection
type (no.
analyzed)
BTX-A
418
BTX-A
O-acetate
(rearranged)
460
BTX-A
O-crotonate
486 BTX 539 h-BTX 553 Other BTXs
I (29) ?  ? nd? nd? 514 (?) (2), 500 (), 525 (?), 583 (), and others
II (5)    nd? nd? 500 (?), 514 (), 565 (), 583 (), 597 ()
III (3)    565
IV (1)  nd    525 (), 514 (), 436438 ()
V (1)    nd nd
VI* (1)      500 (), 514 ()
The protonated molecular ion is shown in bold. nd, not detectable; , trace (detected with single ion profiling); , minor; , major. Contents ranged
widely in different collections listed in I, II, and IV. For example, in the 29 collections reported as type I, BTX-A and BTX-A O-crotonates ranged from minor to
major, BTX-A ranged from nd to minor, and h-BTX ranged from nd to trace.
*This collection in ethanol (17 beetles, not identified as to species) was sonicated in fresh ethanol for a 30-min period at room temperature after the initial extract
had been decanted. The original and sonicated extracts were combined for analysis. The sonication step added little to the BTXs extracted.
Fig. 1. Structures of the most common batrachotoxins found in Choresine
beetles.
15858  www.pnas.orgcgidoi10.1073pnas.0407197101 Dumbacher et al.
symbionts (19). The possible role of symbionts in the production
of beetle BTXs is being investigated.
The four species of Choresine beetles in our collections range
in body length from
5.4 to 6.2 mm and in body breadth (across
elytra) from 2.25 to 2.65 mm and average, in dried specimens,

5.7 mm in length and 2.3 mm in breadth. Each of these species
has a series of glandular vesicles that may protrude from the
lateral margins of abdominal segments 1?6 (Fig. 2), plus addi-
tional large vesicles from the prothorax (anterior and lateral
margins) and metathorax (lateral margins). The elytra are
shaped such that vesicles can evert even when the elytra are
closed. Such vesicles are characteristic of Melyrids of the sub-
family Malachiinae to which Choresine belongs. The systematics
of this group for the Pacific area, treated mainly by the late
Walter Wittmer, now remain a challenge for the relatively few
contemporary entomologists studying Melyrid beetle taxonomy.
To our knowledge, there have been no reports of secondary,
bioactive compounds from beetles of the family Melyridae.
Little is known about the natural history of these Choresine
beetles. Males and females are distinguishable only with a
dissecting microscope. Choresine beetles, although apparently
not brightly colored, do display a metallic blue-violaceous col-
ored elytra that contrasts strongly with the yellow and blackish
coloration of the pronotum (Fig. 2). Contrasting colors have
some aposematic value, and this contrast is probably more
conspicuous in sunlight. In Herowana, where the local people
captured and preserved hundreds, most beetles were caught
flying early in the morning or were found under leaves of garden
plants. It was uncommon for a collector to catch more than

10?15 beetles per day. No adult Choresine beetles were found
aggregated in groups. Local naturalists were unable to provide
any information about larvae, host plants, or other habits, other
than that adults were more commonly found in or around fresh
thatch bunches used for roofing material. Local men warned that
when working in fields or with thatch, one should take precau-
tions to prevent these beetles from alighting in an eye or in facial
perspiration because the beetles cause a burning sensation that
can be severe.
Collections in museums of the continental United States
offered few additional clues. These beetles are often among
insects collected when fogging canopies; however, it is rare to
collect more than a few Choresine beetles per tree. Other
collections found at the California Academy of Sciences con-
tained specimens obtained by netting individuals in grass of
Madang Province (D. Kavanaugh, personal communication).
BTXs have been identified from five Pitohui species as well as
I. kowaldi (13, 14). We examined the stomach contents of 11 I.
kowaldi, 17 Pitohui dichrous, 12 Pitohui ferrugineus, 8 Pitohui
kirhocephalus, and 1 Pitohui nigrescens, for a total of 49 individ-
ual birds. Stomach contents were classified by arthropod group
Table 2. Summary of bird stomach contents
Species
No. of
individuals
Choresine-sized contents
Plant matterArthropods Insects Choresine
All species 49 44 30 1 27
P. dichrous 17 15 10 1 14
I. kowaldi 11 10 8 0 4
Other Pitohui 21 19 12 0 12
Each column shows the number of individuals whose stomach contained the particular food type.
Fig. 2. A C. pulchra beetle with an insert showing vesicles.
Dumbacher et al. PNAS  November 9, 2004  vol. 101  no. 45  15859
CH
EM
IS
TR
Y
and insect order, where recognizable, and plant and seed ma-
terial. We verified that Pitohui and Ifrita do feed on small
Choresine-sized insects (present in up to 30% of all individuals)
(Table 2). Furthermore, one Pitohui did contain a Choresine in
its stomach that was identified asChoresine nigroviolacea (Cham-
pion). Although we have not surveyed the toxin content in C.
nigroviolacea, these studies demonstrate that toxic birds do feed
on a variety of insects of the same size and even from the same
genus, and we anticipate that they probably do feed onC. pulchra
and other Choresines in areas where those species are present.
The present results suggest that Choresine beetles are potentially
a direct source of batrachotoxins for toxic New Guinea birds. An
alternative hypothesis, that both the birds and the beetles may get
their toxins from some common third source, such as a plant,
seems unlikely, because these stomach contents and other
reports (18) indicate that Ifrita is primarily, probably exclusively,
insectivorous, although plant matter was also found in their
stomachs, most likely arising from accidental ingestion of
mosses, etc. while foraging. Only Pitohui species had seeds in
their stomachs.
Over the last several years, we have used multiple methods to
search for natural sources of BTXs, including radiotelemetry of
Pitohui in NewGuinea, examination of Pitohui and Ifrita stomach
contents, survey of hundreds of insects and plants that might
formulate Pitohui diets, and follow-up of leads provided by many
local naturalists.
The present discovery emphasizes the value of traditional
knowledge. Pitohui birds appear to be omnivorous and feed on
a wide variety of insects, spiders, millipedes, other arthropods,
fruit, and even smaller vertebrates. They do not appear to be diet
specialists. In tropical rainforests, therefore, the diet of an
omnivore, such as the Hooded Pitohui (P. dichrous), may contain
literally thousands of species. Thus, the observations and folk-
lore of traditional people can and did provide labor-saving leads
and an incredible and essential store of knowledge and insights
for scientists working in unfamiliar places.
We thank the people of Herowana Village for their hard work and
interest; Greg Greene for assistance in the field; A.Mack, D.Wright, and
R. Sinclair of the Wildlife Conservation Society for logistical assistance;
S. Miller, T. Erwin, and W. Steiner, Jr., of the Smithsonian Institution
for early entomological discussions and assistance with taxonomy; andD.
Kavanaugh of the California Academy of Sciences for additional infor-
mation on field observations and a critique of an earlier version of the
manuscript. J.P.D. was supported by National Science Foundation Grant
DEB 0108247.
1. Daly, J. W., Witkop, B., Bommer, P. & Biemann, K. (1965) J. Am. Chem. Soc.
87, 124?126.
2. Tokuyama, T., Daly, J. & Witkop, B. (1969) J. Am. Chem. Soc. 91, 3931?3938.
3. Daly, J. W. (1998) J. Nat. Prod. 61, 162?172.
4. Myers, C. W., Daly, J. W. & Malkin, B. (1978) Bull. Am. Mus. Nat. Hist. 161,
307?366.
5. Albuquerque, E. X., Daly, J. W. & Witkop, B. (1971) Science 172,
995?1002.
6. Strichartz, G., Rando, T. & Wang, G. K. (1987) Annu. Rev. Neurosci. 10,
237?267.
7. Daly, J. W., Myers, C. W., Warnick, J. E. & Albuquerque, E. X. (1980) Science
208, 1383?1385.
8. Daly, J. W., Garraffo, H. M., Spande, T. F., Jaramillo, C. & Rand, A. S. (1994)
J. Chem. Ecol. 20, 943?955.
9. Daly, J. W., Secunda, S. I., Garraffo, H. M., Spande, T. F., Wisnieski, A. &
Cover, J. F. J. (1994) Toxicon 32, 657?663.
10. 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. & Spande, T. F. (1999) J. Chem. Ecol. 25,
1179?1193.
11. Saporito, R. A., Donnelly, M. A., Hoffman, R. L., Garraffo, H.M. &Daly, J.W.
(2003) J. Chem. Ecol. 29, 2781?2786.
12. Saporito, R. A., Garraffo, H. M., Donnelly, M. A., Edwards, A. L., Longino,
J. T. & Daly, J. W. (2004) Proc. Natl. Acad. Sci. USA 101, 8045?8050.
13. Dumbacher, J. P., Beehler, B. M., Spande, T. F., Garraffo, H. M. & Daly, J. W.
(1992) Science 258, 799?801.
14. Dumbacher, J. P., Spande, T. & Daly, J. W. (2000) Proc. Natl. Acad. Sci. USA
97, 12970?12975.
15. Dumbacher, J. P. (1999) Auk 116, 957?963.
16. Mouritsen, K. N. & Madsen, J. (1994) Oikos 69, 357?358.
17. Kocher-Schmid, C. (1993) Muruk 6, 1?15.
18. Majnep, I. S. & Bulmer, R. (1977) Birds of My Kalam Country (Auckland Univ.
Press, Auckland).
19. Jungnickel, H. & Dettner, K. (1997) Mitt. Dtsch. Ges. Angew. Entomol. 11,
895?898.
15860  www.pnas.orgcgidoi10.1073pnas.0407197101 Dumbacher et al.