Epiquinamide: A Poison That Wasn't from a Frog That Was

Article abstract of DOI:10.1021/np8005452

In 2003, we reported the isolation, structure elucidation, and pharmacology of epiquinamide (1), a novel alkaloid isolated from an Ecuadoran poison frog, Epipedobates tricolor. Since then, several groups, including ours, have undertaken synthetic efforts to produce this compound, which appeared initially to be a novel, β2-selective nicotinic acetylcholine receptor agonist. Based on prior chiral GC analysis of synthetic and natural samples, the absolute structure of this alkaloid was established as (15,9aS)-1-acetamidoquinolizidine. We have synthesized the (1.R*,9aS*)-isomer (epiepiquinamide) using an iminium ion nitroaldol reaction as the key step. We have also synthesized ent-1 semisynthetically from (-)-lupinine. Synthetic epiquinamide is inactive at nicotinic receptors, in accord with recently published reports. We have determined that the activity initially reported is due to cross-contamination from co-occurring epibatidine in the isolated material.

Full text of DOI:10.1021/np8005452

J. Nat. Prod. 2009, 72, 243–247
243
Epiquinamide: A Poison That Wasn’t from a Frog That Was1
Richard W. Fitch,*,† Gordon D. Sturgeon,†,| Shaun R. Patel,‡,| Thomas F. Spande,§ H. Martin Garraffo,§ John W. Daly,§,# and
Richard H. Blaauw
Department of Chemistry, Indiana State UniVersity, 600 Chestnut Street, Science S35E, Terre Haute, Indiana 47809, Department of Chemistry,
Wabash College, P.O. Box 352, CrawfordsVille, Indiana 47933, Laboratory of Bioorganic Chemistry, National Institute of Diabetes and
DigestiVe and Kidney Diseases, National Institutes of Health, Department of Health and Human SerVices, Bethesda, Maryland 20892, and
Chiralix B.V., P.O. Box 31070, 6503 CB Nijmegen, The Netherlands
ReceiVed August 28, 2008
In 2003, we reported the isolation, structure elucidation, and pharmacology of epiquinamide (1), a novel alkaloid isolated
from an Ecuadoran poison frog, Epipedobates tricolor. Since then, several groups, including ours, have undertaken
synthetic efforts to produce this compound, which appeared initially to be a novel, ꢀ2-selective nicotinic acetylcholine
receptor agonist. Based on prior chiral GC analysis of synthetic and natural samples, the absolute structure of this
alkaloid was established as (1S,9aS)-1-acetamidoquinolizidine. We have synthesized the (1R*,9aS*)-isomer (epi-
epiquinamide) using an iminium ion nitroaldol reaction as the key step. We have also synthesized ent-1 semisynthetically
from (-)-lupinine. Synthetic epiquinamide is inactive at nicotinic receptors, in accord with recently published reports.
We have determined that the activity initially reported is due to cross-contamination from co-occurring epibatidine in
the isolated material.
Nicotinic acetylcholine receptors (nAChRs) are a class of ligand-
gated ion channels characterized by activation by the alkaloid
nicotine and are important targets for the development of
therapeutics.2-4 These receptors are involved in cellular signaling
in both the peripheral (PNS) and central (CNS) nervous systems.
In the PNS, nicotinic receptors mediate muscle contraction by
initiating action potentials at the muscle end-plate, as well as
coupling to mechanosensitive hair cells in the inner ear.5 In the
central nervous system, nAChRs mediate much of the fast synaptic
transmission in several areas of the brain. Activation of different
nAChR subtypes is considered to be responsible for the myriad
pharmacological effects of nicotine.2 Ganglionic-type receptors
mediate many of the toxic effects, while the central neuronal
subtypes, especially R4ꢀ2, are considered to be responsible for the
cognitive enhancing as well as the addictive properties of nicotine.4
Figure 1. Nicotinically active alkaloids from frogs.
The R7 homopentamer has been implicated in inflammation and
and the (1S,9aS)-(+)-enantiomer synthesized previously.13a The
schizophrenia, and it has been observed that a high percentage
(>80%) of schizophrenics are heavy smokers, with the behavior
absolute configuration of natural epiquinamide has been established
as (1S,9aS) by chiral gas chromatograpic analysis and comparison
with ent-1 and the natural product.13b Synthetic and biological
studies by two separate groups have demonstrated that the synthetic
enantiomers lack biological activity in several assays.16,17 Our group
has observed this as well with our synthetic isomers of epiquina-
mide. However, re-evaluation of the natural extract clearly indicated
activity in binding and functional fluorescence assays (see below).
Thus it is clear that epiquinamide itself lacks activity by itself, and
some other component is likely to be responsible for activity in
the purified compound. We have considered several possibilities
for this, and they are discussed herein.
possibly serving as a form of self-medication.2,3
Our interest in nAChRs stems from the ability of a number of
amphibian alkaloids6,7 to selectively modulate these receptors
(Figure 1). Epibatidine is a well-known highly potent nicotinic
agonist.8 Histrionicotoxin,9 gephyrotoxin,10 and several of the
pumiliotoxins11 are blockers of nicotinic receptors.
In 2003, we reported a novel alkaloid, epiquinamide (1), from
an Ecuadoran poison frog, Epipedobates tricolor,12 along with
a screening method based on functional fluorescence assays.12,13
Since then, several groups, including ours, have undertaken
synthetic efforts to produce this compound, which appeared
initially to be a novel ꢀ2-selective nicotinic acetylcholine
receptor agonist.13-18
Results and Discussion
We have synthesized the (1R,9aR)-enantiomer of epiquinamide
(ent-1) as well as the racemic (1R*,9aS*)-diastereomer (2), which
is described herein. We have compared them to the natural product
Chemistry. Our initial approach to epiquinamide was based on
methodology developed for the synthesis of lupine alkaloids used
by van Tammelen and Foltz.19 Construction of the quinolizidine
was envisioned as arising from alkylation and subsequent intramo-
lecular iminium ion aldol reaction of a cyclic imine with a dipolar
synthon capable of acting sequentially as an electrophile and
nucleophile. We see this approach as a general scheme amenable
to the core construction of any number of “izidine” azabicycles by
variation of the ring size of the imine, chain length of the dipolar
synthon, and substitution patterns thereof (Scheme 1). For the
* To whom correspondence should be addressed. Tel: (812) 237-2244.
Fax: (812) 237-2232. E-mail: rfitch@indstate.edu.
† Indiana State University.
‡ Wabash College.
§ National Institute of Diabetes and Digestive and Kidney Diseases, NIH.
Chiralix B.V.
| Undergraduate student.
# Deceased March 5, 2008.
10.1021/np8005452 CCC: $40.75
2009 American Chemical Society and American Society of Pharmacognosy
Published on Web 02/06/2009

244 Journal of Natural Products, 2009, Vol. 72, No. 2
Scheme 1. Alkylation-Iminium Ion Aldol Annulation
Fitch et al.
Scheme 4. Synthesis of ent-Epiquinamide
Scheme 2. Synthesis of epi-Epiquinamide
the axial configuration of the (1S*,9aS*)-trans-isomer was found
to be preferred, though only by half the energy difference (0.5 kcal/
mol). While significant, this energy difference is insufficient to
explain the overwhelming preference for the (1R*,9aS*)-cis-
diastereomer. However, anchimeric assistance by the quinolizidine
nitrogen further promotes the observed stereochemistry. Upon
formation of the nitronate, reprotonation of the carbon is postulated
to occur almost exclusively via suprafacial [1,3] proton transfer
from the nitrogen lone-pair to the aci-nitro carbon to give 5. This
essentially precludes formation of the desired (1S*,9aS*)-trans-
diastereomer. While this is not useful for the synthesis of epiquina-
mide, we envision the excellent degree of stereocontrol to be useful
for a number of “izidine” alkaloids (Scheme 1), and further studies
are underway in this area.
Scheme 3. Stereoselection in Formation of Nitroquinolizidine 5
present target, the most obvious sequence involves the reaction of
∆
We transformed nitroquinolizidine 5 to the acetamide by reduc-
tion over Raney nickel to afford diamine 6 (Scheme 2). Subsequent
acetylation of the crude amine afforded racemic amide 2 in 64%
yield over two steps (4% from dichlorobutane over five steps). The
NMR spectra were again consistent with the (1R*,9aS*)-cis-
diastereomer (epi-epiquinamide).17 In order to correlate with the
original isolation data that were obtained in acetone-d6 and CD3OD,
we obtained spectra in each solvent. The acetamide is soluble in
CDCl3 as well as the other two solvents and gave quite different
proton spectra in each (see Supporting Information), which we
ascribe to differences in conformational preference and/or associa-
tion (especially in CDCl3), perhaps as an acetamide dimer similar
to the behavior of carboxylic acids. In order to assess the activity
of each enantiomer of 2 independently, we made several attempts
to resolve diamine 6 using either D- or L-tartaric acid. This approach
has been used previously with excellent results with the similar
1,2-cyclohexanediamine.23 While we obtained nearly quantitative
yields of crystals using 0.5 equiv of the acid in aqueous methanol/
acetic acid, we observed only racemic 2 by chiral gas chromatog-
raphy on a cyclodextrin column after conversion to the free base
and acetylation.
1,2-piperideine (tetrahydropyridine, 3) with 1-iodo-4-nitrobutane
(4).
We prepared ∆1,2-piperideine (3, as the trimer)20 and 1-iodo-4-
nitrobutane (4)21 essentially according to the literature. Ni-
troiodobutane may be prepared either by direct Kornblum reaction
of 1,4-diiodobutane with silver nitrite or by a three-step sequence
from 1,4-dichlorobutane involving sequential Finkelstein monoio-
dination, Kornblum substitution of the iodide by nitrite, and a
second Finkelstein substitution of the remaining chlorine. While
the latter process is longer, the starting dichloride is only about
one tenth the cost of the diiodide and separation of the intermediates
provides a cleaner product (see Experimental Section). With both
synthons in hand, we set about the cyclization (Scheme 2), which
proceeded using either an excess of piperideine trimer or added
triethylamine as base to afford racemic 1-nitroquinolizidine 5 in
moderate yield as essentially a single diastereomer. While several
solvents were evaluated for this reaction, only alcoholic solvents
afforded reasonable yields, presumably because of enhanced
solubility of the charged intermediates involved in breakdown of
the tripiperideine and formation of the iminium salt.22
However, the relative configuration of the major diastereomer
of 5 was found to be epimeric to that of the natural product with
a diastereomeric ratio of ∼95:5 by GC-MS. The (1R*,9aS*)-cis-
stereochemistry was clear from the splitting pattern of the proton
adjacent to the nitro group, which displayed large (J ) 9.6 and
12.1 Hz) couplings indicating an axial orientation, whereas epi-
quinamide displays small couplings (largest J ) 7.6 Hz, NH
coupling, CH couplings <5 Hz) and typically appears as a
broadened singlet where the NH is exchanged, consistent with an
equatorial disposition. Attempts to epimerize the nitro group using
LDA in THF or boiling acetic acid failed to effect any significant
change. On examination of the system, the reason for this became
clear (Scheme 3). Assuming a chair-chair conformation of the
quinolizidine ring, the nitro group would normally be expected to
prefer an equatorial orientation. This was borne out by calculation
(MM2, Chem3D), which indicated a preference (0.9 kcal/mol) for
the equatorial configuration in the (1R*,9aS*)-cis-isomer. Surpris-
ingly, when the same calculation was carried out on the acetamide,
In order to access the correct diastereomer in enantiopure fashion,
we embarked on a semisynthesis from the commercially available
(1R,9aR)-trans-(-)-lupinine (Scheme 4). While the starting material
is rather expensive (∼$300/g), we needed very little material to
establish the stereochemistry of the natural product, as the semi-
synthesis was based on simple functional group manipulation of
the side chain.
The amide was prepared in straightforward fashion by oxidation
of the primary alcohol to the carboxylic acid (7) using the Jones
reagent in acetone.24 While the oxidation worked well by TLC,
workup was rather cumbersome owing to difficulties in separating
7 from the chromium salts. Ultimately, addition of silica gel and
quenching with ammonia and 2-propanol allowed most of the
chromium salts to be adsorbed and the amino acid separated by
filtration and purified by flash chromatography. Acid 7 was
subjected to a modified Curtius rearrangement using diphenylphos-
phoryl azide,25 hydrolyzed, and directly acetylated (Ac2O/pyridine)
to afford the amide (ent-1, 24% from lupinine). While this furnished

Epiquinamide
Journal of Natural Products, 2009, Vol. 72, No. 2 245
calcium in transfected HEK cells expressing R3 and ꢀ4 nicotinic
receptor subunits (Figure 2b).27 In all instances, synthetic epiquina-
mide diastereomers were found to be inactive, having no affinity
or functional activity up to 100 µM either as the free bases or as
methiodides. Curiously, the isosteric ammonium lupinoate and
nitroquinolizidine did show weak inhibition of binding at 100 µM
and minor functional inhibition of nicotinic receptor activation at
500 µM. This is too weak to be of great interest for nicotinic work,
but studies in other systems is ongoing given their similarity to
nipecotic acid, a well-known GABA agonist.28
Previous reports described racemic and enantiopure epiquina-
mide and its diastereomeric analogue to be inactive at nicotinic
receptors, in contrast to our initial observations regarding the
natural material.15,16 This prompted us to reexamine the natural
material, which retains activity, having approximately the same
potency as S-nicotine (Figure 2). We were initially puzzled by
this and examined the isolated material thoroughly for the
presence of contamination. The possibilities are (1) that epiquina-
mide coelutes with a very potent agonist; (2) that epiquinamide’s
activity is potentiated by a trace impurity; or (3) that the collected
fractions were contaminated by splashover from nearby wells
or a stray droplet from the fraction collector. A coeluting agonist
is possible, though unlikely. Potentiation of the reference nicotine
response in the assay was not observed, ruling out such a
coeluting substance. Interwell cross-contamination is the most
likely culprit here. The collection of the fractions was done in
96-well plates at 0.25 min intervals in six successive injections.
Epiquinamide was collected along one row and epibatidine
in the next successive row. Thus it is possible that droplet or
splashover contamination occurred.
Examination of isolated epiquinamide by gas chromatography-
mass spectrometry showed it to be >99% pure (Figure 3a). We
looked for epibatidine as a likely culprit, as it is present in significant
quantity in the extract and is an extraordinarily potent agonist. We
were unable to locate it by looking for the parent ions at m/z 208
and 210 (Cl isotopes). However, these ions are of very low
abundance in the electron impact mass spectrum of epibatidine.
By looking instead for the m/z 68 and 69 peaks, which are the
base peak and ∼95% abundance, respectively, we were able to
detect, by profiling these ions, a trace amount of epibatidine in the
isolated material (Figure 3b). Though sample amounts preclude
precise determination, by estimation, there is perhaps 0.1% of
epibatidine present in the sample. While tiny, this amount would
be consistent with the observed potency of the isolated material
from [3H]-epibatidine binding, which is approximately 1000-fold
less than that of epibatidine. Although it remains possible that we
have some other trace agonist, we believe epibatidine contamination
to be the cause of the observed activity. While it was not observed
during the original collection (each fraction was examined by GC-
MS), the concentration at which it would have been present would
have been nearly undetectable in the fractions. As further confirma-
tion, the original extract was rechromatographed and the bioactivity
reexamined (see Supporting Information). In the raw extract, weak
activity can be observed near the elution time of epiquinamide,
but is relatively insignificant and clearly not the robust activity
originally observed.
While these data are somewhat embarrassing to report, they
are not entirely unusual in the isolation of natural products.
Further, the results are a reminder that the presence of multiple
active substances in an extract (in this case compounded by small
amounts of sample available) can sometimes lead to erroneous
results from cross-contamination that are not always evident a
priori. This underscores the importance of obtaining synthetic
material whenever possible to corroborate both structure and
pharmacology. In this case, we were able to produce a correct
structure, but were misled by contamination in the pharmacology.
Often with more complex structures, the converse occurs. Thus,
Figure 2. Pharmacologic analysis of epiquinamide and intermedi-
ates: (a) affinity data; (b) functional data.
the target compound, the amide was very difficult to separate from
the symmetrical urea, formed as a result of reaction of hydrolyzed
amine with unhydrolyzed isocyanate during the workup process.
We then focused on the N-Boc derivative (8), which was prepared
by replacing toluene with t-BuOH as the solvent. While 8 proved
difficult to separate from aromatic byproducts, it could be smoothly
deprotected with trifluoroacetic acid and acetylated under Schotten-
Baumann conditions.13a This allowed for the neutral aromatic
byproducts to be separated from the acetamide by a simple
acid-base extraction, allowing isolation of ent-1 as a clean solid
without further chromatography. However, comparison of chiral
gas chromatographic retention times with that of the natural product
and co-injection clearly showed we did not have the natural
enantiomer.13b Similar comparison of a sample of (+)-epiquinamide
synthesized by Blaauw and co-workers13a did show identity and
established the absolute configuration of the natural product as
(1S,9aS), making his the first total synthesis of epiquinamide.
Noting the significant resemblance of epiquinamide to ace-
tylcholine, we prepared the methiodides of ent-1 and 2, along
with nitroquinolizidine 5 and lupinine, to produce closer
bioisosteres. Each was smoothly prepared by treatment of the
quinolizidine with methyl iodide in dry acetone, affording
crystalline methiodides (see Experimental Section).
Pharmacology. Natural epiquinamide, isolated as described in
our original report,12 and the synthetic isomers described herein
and earlier13 were evaluated for activity in both affinity and
functional assays. Affinity was assessed by [3H]-epibatidine binding
in rat cerebral cortical membranes, which primarily express the
R4ꢀ2 nicotinic receptor subtype (Figure 2a).26 Functional assays
were performed using fluorescence assays based on intracellular

246 Journal of Natural Products, 2009, Vol. 72, No. 2
Fitch et al.
Figure 3. GC-MS data for natural epiquinamide.
it is important that the pharmacognosist work in collaboration
with the synthetic/medicinal chemist to produce data that are
unequivocal. Indeed in this very extract, we have other com-
pounds, including phantasmidine, which are of significant
interest.5,6 However, because of similar concerns, we are com-
pleting the synthesis of phantasmidine prior to releasing its
structure and activity, which we hope to report soon.
In conclusion, we have synthesized three and examined the
pharmacology of all four stereoisomers of epiquinamide. The trans
relative configuration has been confirmed and the absolute config-
uration has been established as (1S,9aS). All four diastereomers
have been found to be inactive at nicotinic acetylcholine receptors,
and the misleading activity in the natural material is concluded to
be trace contamination by co-occurring epibatidine. This determi-
nation would not have been possible without synthetic material,
underscoring the importance of chemical synthesis in collaboration
with natural products isolation and pharmacology.
Acknowledgment. K. J. Lipscomb and R. J. Tucker are acknowl-
edged for performing exploratory experiments. A generous gift of HEK
cells expressing various subunit combinations for nicotinic receptors
by Drs. K. J. Kellar and Y. Xiao of Georgetown University Medical
School is gratefully acknowledged. This work was performed with funds
from the Department of Chemistry and College of Arts and Sciences
at Indiana State University (startup funding to R.W.F. and summer
stipend to G.D.S.), as well as a Sigma Xi Grant-in-aid (12030168 to
K.J.L.), the Lilly Endowment (Wabash College Indiana Internship
Initiative summer stipend to S.R.P. and Indiana State University
Promising Scholar Award to R.W.F.), and the National Science
Foundation (CHE0521075 for a 400 MHz NMR spectrometer at Indiana

Epiquinamide
Journal of Natural Products, 2009, Vol. 72, No. 2 247
Noonan, B. P.; Schargel, W.; Wheeler, W. C. Bull. Am. Mus. Nat.
Hist. 2006, 299, 1–262.
State University). Work performed at NIH was supported by the
intramural program of NIDDK.
(13) (a) Wijdeven, M. A.; Botman, P. N. M.; Wijtmans, R.; Schoemaker,
H. E.; Rutjes, F. P. J. T.; Blaauw, R. H. Org. Lett. 2005, 7, 4005–
4007. (b) Wijdeven, M.; Wijtmans, R.; Berg, R.; Noorduin, W.;
Schoemaker, H.; Sonke, T.; van Delft, F.; Blaauw, R.; Fitch, R.; Rutjes,
F. P. J. T.; Spande, T. F. Org. Lett. 2008, 10, 4001–4003.
(14) Tong, S. T.; Barker, D. Tetrahedron Lett. 2006, 47, 5017–5020.
(15) Huang, P. Q.; Guo, Z. Q.; Ruan, Y. P. Org. Lett. 2006, 8, 1435–
1438.
Supporting Information Available: Experimental procedures and
selected 1D and 2D NMR spectra for compounds and intermediates
are provided. This material is available free of charge via the Internet
at http://pubs.acs.org.
References and Notes
(16) Suyama, T. L.; Gerwick, W. H. Org. Lett. 2006, 8, 4541–4543.
(17) Kanakubo, A.; Gray, D.; Innocent, N.; Wonnacott, S.; Gallagher, T.
Bioorg. Med. Chem. Lett. 2006, 16, 4648–4651.
(18) Voituriez, A.; Ferreira, F.; Perez-Luna, A.; Chemla, F. Org. Lett. 2007,
9, 4705–4708.
(19) van Tamelen, E. E.; Foltz, R. L. J. Am. Chem. Soc. 1969, 91, 7372–
7377.
(20) DeKimpe, N.; Stevens, C. J. Org. Chem. 1993, 58, 2904–2906.
(21) Kurth, M. J.; Ahlberg-Randall, L. A.; Takenouchi, K. J. Org. Chem.
1996, 61, 8755–8761.
(22) Less polar solvents were observed to produce a precipitate without
appreciable further reaction. Though uncharacterized, we suspect it
to be the initial quaternary salt from reaction of intact tripiperidine
with the iodide. This is consistent with prior observation that protic
solvents promote equilibrium between trimer and monomer. See:
Schopf, C.; Komnak, A.; Braun, F.; Jacobi, E.; Borgmuth, M. L.;
Bullnehimer, M. Liebigs Ann. Chem. 1948, 559, 1–42. While slower,
we found the intermediate chloronitrobutane to be equally competent
in the annulation reaction.
(23) Gaslbøl, F.; Steenbøl, P.; Sørensen, B. S. Acta Chem. Scand. 1972,
26, 3605–3611.
(24) Schopf, C.; Thoma, O.; Schmidt, E.; Braun, W. Liebigs Ann. 1928,
465, 97–147.
(1) Presented at the 48th Annual Meeting of the American Society of
Pharmacognosy, Portland, ME, July 14-18, 2007; Abstract O4.
(2) Arneric, S. P.; Holladay, M.; Williams, M. Biochem. Pharmacol. 2007,
74, 1092–1101.
(3) Gundisch, D. Expert Opin. Ther. Pat. 2005, 15, 1221–1239.
(4) Tapper, A. R.; McKinney, S. L.; Nashmi, R.; Schwarz, J.; Desphande,
P.; Labarca, C.; Whiteaker, P.; Marks, M. J.; Collins, A. C.; Lester,
H. A. Science 2004, 306, 1029–1032.
(5) Elgoyhen, A. B.; Vetter, D. E.; Katz, E.; V. Rothlin*, C. V.;
Heinemann, S. F.; Boulter, J. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
3501–3506.
(6) Daly, J. W. J. Med. Chem. 2003, 46, 447–452.
(7) Daly, J. W.; Spande, T. F.; Garraffo, H. M. J. Nat. Prod. 2005, 68,
1556–1575.
(8) Spande, T. F.; Garraffo, H. M.; Edwards, M. W.; Yeh, H. J. C.; Pannell,
L.; Daly, J. W. J. Am. Chem. Soc. 1992, 114, 3475–3478.
(9) Daly, J. W.; Karle, I.; Myers, C. W.; Tokuyama, T.; Waters, J. A.;
Witkop, B. Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 1870–1875.
(10) Aronstam, R. S.; Daly, J. W.; Spande, T. F.; Narayanan, T. K.;
Albuquerque, E. X. Neurochem. Res. 1986, 11, 1227–1240.
(11) Daly, J. W.; Tokuyama, T.; Fujiwara, T.; Highet, R. J.; Karle, I. L.
J. Am. Chem. Soc. 1980, 102, 830–836.
(12) Fitch, R. W.; Garraffo, H. M.; Spande, T. F.; Yeh, H. J. C.; Daly,
J. W. J. Nat. Prod. 2003, 66, 1345–1350. The frogs from this collection
have since been reclassified as Epipedobates anthonyi on the basis of
geography: Graham, C. H.; Ron, S. R.; Santos, J. C.; Schneider, C. J.;
Moritz, C. EVolution 2004, 58, 1781–1793. This is an important
distinction given the dietary alkaloid source for these frogs, which is
presumably geographically dependent: Darst, C. R.; Menendez-
Guerrero, P. A.; Coloma, L. A.; Cannatella, D. C. Am. Nat. 2005,
165, 56–69. For a more comprehensive discussion and revision of the
systematics of dendrobatid frogs, see: Grant, T.; Frost, D. R.; Caldwell,
J. P.; Gagliardo, R.; Haddad, C. F. B.; Kok, P. J. R.; Means, B. D.;
(25) Ninomiya, K.; Shioiri, T.; Yamada, S. Tetrahedron 1974, 30, 2151–
2157.
(26) Houghtling, R. A.; Davila-Garcia, M. I.; Kellar, K. J. Mol. Pharmacol.
1995, 48, 280–287.
(27) Fitch, R. W.; Xiao, Y.; Kellar, K. J.; Daly, J. W. Proc. Nat. Acad.
Sci. U.S.A. 2003, 100, 4909–4914.
(28) Barrett-Jolley, R. Br. J. Pharmacol. 2001, 133, 673–678.
NP8005452