Christa Kletter1

Sabine Glasl1
Armin Presser2
Ingrid Werner1
Gottfried Reznicek1
Samdan Narantuya1
Selim Cellek3
Ernst Haslinger2
Morphological, Chemical and Functional Analysis of
Johann Jurenitsch1 Catuaba Preparations

Abstract
Fourteen commercial samples of the popular Brazilian aphrodisiac
Catuaba specified as bark drugs of Anemopaegma, Erythroxylum
and Trichilia species were examined for identity and purity. Only
Original Paper
retention times, ESI-MS) given. The structure elucidation of the
two main alkaloids, catuabine D and its hydroxymethyl deriva-
tive, is presented. The 1H- and 13C-NMR assignments of these al-
kaloids are discussed with regard to literature data. Neither
aqueous nor methanolic extracts of the Trichilia catigua refer-

Downloaded by: Dot. Lib Information. Copyrighted material.

a minority of the examined Catuaba samples contained the crude ence material nor alkaloid-enriched fractions of commercial
drugs claimed on the labels. More than half of the products were samples showed any effect on the rabbit corpus cavernosum in
adulterated with different crude drugs. The majority of the sam- an in vitro test.
ples contained a bark originating from Trichilia catigua. The TLC
fingerprints confirmed the heterogeneity, in 50 % of the samples Key words
tropane alkaloids of various concentrations were detected. TLC Anemopaegma Erythroxylum Trichilia aphrodisiac micros-
and HPLC methods for separation and identification of the tropane copy tropane alkaloids
alkaloids were developed and their analytical data (RF values,

Introduction
The introduction of sildenafil for the treatment of erectile dys-
function triggered a boom of herbal medicines which are
claimed to improve sexual pleasure, libido and erectile function.
A large number of different products is available in drugstores
and on the internet promising improved sexual performance
without any side effects. A quality control of such products con-
cerning plant origin, homogeneity and chemical composition
does not exist. Nevertheless, such preparations originating from
natural sources often are suggested and believed to be safe. One
of these herbal medicines is Catuaba which was a popular reme-
dy in Brazil in the mid-nineteenth century. It is still used there as
993
an analgesic, central nervous system stimulant, tonic and aphro-
disiac, mainly by preparing hot infusions to drink [1]. Depending
on the consulted source, the Catuaba-furnishing plants vary and
belong to different genera and families: Anemopaegma (Bigno-
niaceae), Erythroxylum ( = Erythroxylon, Erythroxylaceae), Ilex
(Aquifoliaceae), Micropholis (Sapotaceae), Phyllanthus (Euphor-
biaceae), Secondatia (Apocynaceae), Tetragastris (Burseraceae),
Trichilia (Meliaceae) and species of the Myrtaceae [1], [2], [3].
The confusion about Catuaba dates back to the beginning of the
20th century. In 1904, J. A. da Silva determined a Catuaba-furnish-
ing plant as Erythroxylon catuaba. Until today the identification
of this species collected in Bahia, Northern Brazil, is doubtful, be-
cause no type specimen exists and da Silva's plant illustration

Dedicated to Professor Adolf Nahrstedt on the occasion of his retirement as Editor-in-Chief of Planta Medica
Affiliation
1
Institute of Pharmacognosy, University of Vienna, PharmaCenterVienna, Vienna, Austria
2
Institute of Pharmaceutical Chemistry and Technology, Karl Franzens University, Graz, Austria
3
Wolfson Institute for Biomedical Research, University College London, London, UK

Correspondence
Ass. Prof. Dr. Ch. Kletter Institute of Pharmacognosy PharmaCenterVienna University of Vienna
Althanstrae 14 1090 Vienna Austria Fax: +43-1-4277-9552 E-mail: Christa.Kletter@univie.ac.at

Received May 5, 2004 Accepted July 6, 2004

Bibliography
Planta Med 2004; 70: 9931000  Georg Thieme Verlag KG Stuttgart New York
DOI 10.1055/s-2004-832627
ISSN 0032-0943
 does not fit to an Erythroxylum species [1], [2], [4]. Several years
later, another da Silva specified Catuaba from the South of Brazil
as the rhizome of Anemopaegma mirandum (Chamisso) Alph. De
Cand. [5], [6]. As the latter da Silva was the editor of the first
Brazilian Pharmacopoeia, a monograph of Catuaba citing
Anemopaegma mirandum as plant source was included in the
first edition of 1929. In 1939, da Cunha, identified the plant
source of the northern Catuaba as Tetragastris catuaba [1]. An-
other expert claimed in 1966 that the plant illustration pres-
ented by da Silva in 1904 resembles a Trichilia species [2]. In
1998, Marques again raised the question of the origin of the
Northern Catuaba and specified the drug-furnishing plant as
Trichilia catigua Adr. Juss. [7]. Such heterogeneity leads un-
doubtedly to difficulties regarding the identification of the tra-
Original Paper
This paper presents an examination of fourteen commercial Ca-
tuaba samples including single crude drugs and mixtures which
are sold in Brazil and Europe. Morphological, anatomical and
chemical analyses using TLC and HPLC were performed to check
the identity and purity of the samples with respect to the given
declarations. The phytochemical analysis focused on the detec-
tion of tropane alkaloids, which are well-known compounds in
Erythroxylum species [11]. To evaluate an aphrodisiac action, the
effects of aqueous and methanolic extracts of Trichilia catigua
and tropane alkaloid-enriched extracts of two commercial Ca-
tuaba samples on the rabbit corpus cavernosum were investiga-
ted.

ded bark. Consequently, information on the medicinal effects of
the drug is unreliable if the plant source is not clearly defined.
Only few investigations exist on the medicinal properties of
Anemopaegma mirandum and Trichilia catigua. An early publi-
cation (1937) reports that plant extracts (leaves, stems and rhi-
zomes) of Anemopaegma mirandum induce hypotension, hypo-
Materials and Methods
Catuaba samples
Fourteen commercial Catuaba samples from various origins (Bra-
zil, Germany, The Netherlands, United Kingdom, Austria) were
acquired in health shops, via internet and via personal contacts

Downloaded by: Dot. Lib Information. Copyrighted material.

thermia, peripheral vasodilation and renal vasoconstriction in
dogs [8]. Hydro-alcoholic extracts of Trichilia catigua bark
were shown to evoke relaxation of the rabbit corpus caverno-
sum [9] and inhibit pain induced by acetic acid and formalin
[10].
to Brazil. They were available as chopped or powdered bark sam-
ples or plant extract and were claimed to originate from
Anemopaegma mirandum, A. glaucum, Erythroxylon catuaba or
Trichilia catigua (Table 1).

Table 1 Quality control of Catuaba samples by means of microscopy and TLC

Sample Claimed Drug Sample Microscopic Examination TLC Country

No Plant Source of Origin
Trichilia catigua Adulteration Alkaloids
994 bark

1 Anemopaegma mirandum chopped bark + additional bark ++ Brazil

no 11
2 chopped bark + Germany
3 capsules 500 mg, +?a additional bark, sand Germany
powdered bark
4 Erythroxylaceae chopped bark + two different Germany
additional barks
5 Anemopaegma glaucum capsules 400 mg, +?a additional bark The Netherlands
powdered bark
6 Trichilia catigua authentic drug, + Brazil
chopped bark
7 Anemopaegma glaucum chopped bark + additional bark United Kingdom
8 Trichilia catigua authentic drug, + Brazil
chopped bark
9 Erythroxylon catuaba Martius chopped bark + additional bark no 11 + Austria
10 powdered bark +?a starch, clusters of Brazil
transparent and
brown parenchyma
11 chopped bark +++ Brazil
12 Anemopaegma mirandum capsules 250 mg, +?a ? + Brazil
powdered bark
13 Anemopaegma mirandum capsules 390 mg, maize starch ++ Brazil
dried extract
14 Trichilia catigua Adr. Juss. capsules 500 mg, +?a ? ++ Brazil
powdered bark
15 Trichilia catigua chopped bark + additional bark no 11 ++ Brazil
16 Trichilia catigua chopped bark + additional bark Brazil

a
The identity of powdered Trichilia catigua bark could not be determined by microscopic means alone, because the characteristic tissue arrangement got destroyed during the powdering
processes.

Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000
Reference samples of Trichilia catigua Adr. Juss. bark were
provided by R. Wasicky, Brazil (no 6) and L. C. Marques, Universi-
ty of Maring, Brazil (no 8, collected in Caitit, Bahia, Brazil). As
no samples of stem bark of Erythroxylum vacciniifolium could be
acquired, either in Europe or Brazil, we used the twig bark of her-
barium specimens as reference material. The herbarium samples
of Erythroxylum vacciniifolium Mart. were supplied by the fol-
lowing institutions: Botanische Staatssammlung Mnchen, M-
0 066 895, Rodovia Guaratuba-Itabu, Paran, Brazil; Herbarium
Ulm, Sammlung Gottsberger, no 73 902 17.5A, Perube, Serra do
Mar, Brazil; Herbarium Ulm, no 69 495, Serra do Cabral, Brazil.
The compounds catuabine D and its hydroxymethyl derivative
had been isolated from sample 1 [12], [13] and were used for
the development of analytical systems (TLC, HPLC) to screen dif-
ferent samples for this type of alkaloids. All chemicals, reagents
and solvents used were of HPLC or analytical grade and were
supplied from Merck (Darmstadt, Germany), J. T. Baker (Deven-
ter, The Netherlands) or Riedel-de-Han (Seelze, Germany).
Microscopy
We based the anatomical examinations on literature data of
Anemopaegma mirandum rhizome [5], [6], Trichilia catigua bark
[7], and on our own investigations of authentic reference materi-
al of Trichilia catigua (bark samples no 6, 8) and Erythroxylum
vacciniifolium (twig bark of herbarium specimens). The anatomi-
cal details of Trichilia catigua bark as cited in the literature were
scrutinised for accuracy and completeness and the results includ-
ed in the microscopic description of the bark presented in this
paper. No anatomical characteristics of Anemopaegma glaucum
a substitute of A. mirandum according to Brazilian sources [3]
could be traced in the literature. Due to missing microscopic
literature data regarding Erythroxylum vacciniifolium bark we re-
corded the microscopic characters of E. vacciniifolium twig bark
for further comparison with the Catuaba samples.
Sample preparation
For TLC and HPLC fingerprints, 0.5 g powdered or chopped sam-
ple were moistened with 0.15 mL of concentrated ammonia and
mixed with 5 mL dichloromethane. After further addition of 0.1
mL of concentrated ammonia the mixture was extracted by shak-
ing gently for 1 h at room temperature. Filtration over Na2SO4
and evaporation yielded an extract which was re-dissolved in
100 mL dichloromethane: 30 mL thereof were used for TLC, 5 mL
for HPLC.
For the in vitro assays the alkaloid-enriched extracts of the sam-
ples 1 and 11 were prepared in the same way as for TLC and HPLC
fingerprints, starting with 1 g of the sample and using double
amounts of solvents and reagents as indicated above. The yields
for samples 1 and 11 were 0.018 g and 0.020 g, respectively. The
residue was suspended in physiological saline (0.9 % NaCl) and
used for the in vitro assays.
The aqueous extracts of the samples 1 and 8 were prepared by
infusing 10 g powdered drug with 100 mL boiling water and
drawing for 10 min, followed by filtration and lyophilisation.
The yields for samples 1 and 8 were 1.7 g and 1.4 g, respectively.
The extracts were dissolved in physiological saline and used for
the in vitro assays. The methanolic extracts were obtained by ex-
tracting 10 g drug with 100 mL of methanol for 20 min at 40 8C.
After filtration the methanol was removed under vacuum which
gave yields of 2.4 g and 2.0 g for samples 1 and 8, respectively.
These residues were dissolved in physiological saline and used
for the in vitro assays.
TLC
Analyses were performed on silica gel 60 F254 TLC plates
(20 20 cm, Merck, Darmstadt, Germany) running two mobile
phases: dichloromethane-acetone (97 + 3) showed the pattern
of the apolar compounds, whereas a mixture of toluene-ace-
tone-methanol-ammoniaconc (45 + 45 + 7 + 3) was developed for
the separation of the alkaloids. Detection was carried out under
UV 254 nm, 366 nm and by spraying with the alkaloid-selective
Original Paper
potassium iodoplatinate reagent (0.25 mL of 5 % hexachloroplati-
nic (IV) acid solution mixed with 2.25 mL of 10 % potassium io-
dide solution and diluted with 5 mL H2O; the reagent has to be
used immediately after preparation [14]).
HPLC and MS
Downloaded by: Dot. Lib Information. Copyrighted material.
Analyses were carried out on a Merck Hitachi liquid chromato-
graph (L-7100 pump; L-7450 diode array detector, monitoring
wavelength 275 nm; D-7000 interface, D-7000 HSM software,
Merck, Darmstadt, Germany) using a Supersorb RP SelectB col-
umn (4 mm, 125 3 mm, Merck, Darmstadt, Germany) as station-
ary phase. Mixtures of methanol and a phosphate buffer (0.05 M
NaH2PO41 H2O, adjusted with 85 % H3PO4 to pH 2.5) served as
mobile phase, best results were achieved by a two-step gradient
system starting at 10 % methanol up to 25 % methanol within 15
min (rate = 1 %/min) followed by a rate of 0.5 %/min up to 55 %
methanol within 60 min. The flow rate was 1.0 mL/min. Fractions
were collected, extracted with dichloromethane after addition of
ammonia, evaporated and measured off-line on a PE Sciex API 995
150EX mass spectrometer using the Ionspray source in the posi-
tive ESI mode.
NMR and HR-MS
NMR spectra were recorded on a Varian Unity Inova 400 NMR
spectrometer at 297 K using sample tubes of 5 mm diameter
(Kontes Glass Company, The Gerresheimer Group, Dsseldorf)
with tetramethylsilane (TMS) as internal standard. A dual probe
head with shielded z-gradients or broadband probe (400 MHz)
was used. HMBC experiments were optimised for a long-range
coupling constant of 8 and 4 Hz, respectively. Before NOE experi-
ments were performed, dissolved oxygen was removed by bub-
bling argon through the solution.
HR-MS was performed on a PE-SCIEX QStar QTOF mass spectro-
meter using the Ionspray source in the positive ESI mode, exact
mass calibration with quinine giving m/z = 325.1916 (100 %),
[MH]+.
Structure elucidation
As reported previously, two tropane alkaloids had been isolated
by us from sample 1. Their structures were elucidated as catua-
bine D (2) and its 7-exo-hydroxy-N-methyl derivative (1) [12],
[13]. Simultaneously, a Swiss working group identified the same
compounds and specified their NMR data in CDCl3 [15]. We used
acetone-d6 as solvent to identify possible OH groups. However,
our NMR experiments showed several differences compared to
Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000
 the reported NMR data, particularly in the range of the bridge-
heads C-1 and C-5. Proton and carbon resonances were clearly
assigned by combining the information of one- and two-dimen-
sional NMR-techniques (COSY, HSQC, HMBC, selective TOCSY-
and NOE-experiments). To determine the relative configuration
of 1 and 2 we performed selective 1D-NOE experiments: irradia-
tion at the resonance of H-6 in compound 1 gave NOEs at the sig-
nals of H-4eq, H-5, H-7endo and H-17, in addition positive NOEs of
H-1 and H-6 were observed after irradiation of H-7. In compound
2 H-6 showed NOE effects to H-4eq, H-5, H-7endo and H-17,
whereas from H-7endo NOEs at H-2eq, H-6, H-7exo and H-17 were
induced. These experiments allowed the assignment of the
endo and exo protons as well as the proper determination of
C-1 and C-5. The recorded data were in perfect accordance with
Original Paper
the calculated structures generated from energy minimum com-
putations using the MM2 force field (Fig. 1).
7-exo-Hydroxy-N-methylcatuabine D (1): UV (45 % MeOH):
lmax = 275 nm; 1H- and 13C-NMR data, see Table 2; HR-MS: m/z =
388.1882 (100 %), [MH]+ (calculated for C20H26N3O5 : 388.1872);
[a]20
D : + 6.5 (c 0.092, EtOH).
Catuabine D (2): UV (48 % MeOH): lmax = 275 nm; 1H- and 13C-
NMR data, see Table 2; HR-MS: m/z = 358.1743 (100 %), [MH]+
(calculated for C19H24N3O4 : 358.1767). [a]D20: 28.5 (c 0.047
EtOH).
Functional test on the rabbit Corpus cavernosum
Male New Zealand White rabbits (3.0 3.5 kg, Harlan, UK) were
sacrificed by an overdose of pentobarbitone (Euthesate, Willows
Francis Veterinary, UK) injected into the marginal vein of the ear.
The corpus cavernosum was excised and cut longitudinally to ob-
996 tain four identical strips (4 10 mm). Each preparation was
cleaned of adherent tissues and mounted horizontally between
Fig. 1 Energy minimum computation of 1 and 2 with selected NOE correlations.
Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000
two ring electrodes (4 mm diameter) in superfusion chambers
(37 8C) as described previously [16], [17]. The chambers were
perfused with modified Krebs' solution at a constant flow of 1.0
mL/min by means of peristaltic pumps (Miniplus 2, Gilson). One
end of the preparation was tied to a Grass FT 03C force-displace-
ment transducer connected to a Linearcorder WR 3101 (Graph-
tec, Tokyo, Japan) for registration of isometric changes in tension.
The preparations were given a pre-tension (0.6 g) and allowed to
equilibrate for 90 min. The preparations were stimulated electri-
cally (electrical field stimulation; EFS) for 5 s with trains of rec-
tangular pulses of 50 V, 0.3 ms pulse duration and at a range of
frequency of 0.5 25 Hz, delivered by Grass S88 stimulators.
The mechanical responses were also recorded on a computer by
a specialised data acquisition system (Axon Instruments, USA).
The composition of the modified Krebs' solution was (mM):
NaCl 136.9, KCl 2.7, CaCl2 1.8, MgSO4 0.6, NaHCO3 11.9, KH2PO4
0.5, glucose 11.5 and gassed with 5 % CO2 in O2 (pH 7.0 7.2). Aqu-
eous and methanolic extracts of the samples 1 and 8 as well as
alkaloid-enriched dichloromethane extracts of the samples 1
and 11 were tested for their direct and indirect relaxant activity
Downloaded by: Dot. Lib Information. Copyrighted material.
on the rabbit corpus cavernosum. The direct relaxant activity
was examined by elevating the tone of the rabbit corpus caverno-
sum with phenylephrine (3 mM; EC80) followed by application of
the samples. Direct relaxant activity is registered if the tone of
the tissue decreases as shown using an NO donor (sodium nitro-
prusside), prostacyclin or forskolin (reference compounds). For
the indirect relaxant activity the tissues were stimulated trans-
murally to release nitric oxide from their internal nerves to elicit
relaxation (EFS-induced relaxation). During the ensuing relaxa-
tion the samples were applied, indirect relaxant activity is regis-
tered by an increase either in magnitude or duration of the re-
laxation as caused by the reference compounds such as phospho-
diesterase type 5 inhibitors (i. e., sildenafil citrate) [16], [17]. The
samples were dissolved in physiological saline and were intro-
 1
Table 2 H- and 13C-NMR data of 1 and 2 (400 and 100 MHz, acetone-d6, d, J in Hz)

7-exo-Hydroxy-N-methylcatuabine D (1) Catuabine D (2)

Position dC dH a dC dH

1 68.0 3.09 3.11 (m) 60.7 3.31 3.35 (m)

2ax 30.2 2.21 2.29 (m) 35.0 2.13 2.19 (m)
2eq 1.70 (d,15) 1.69 (d br, 14.9)
3 67.1 5.21 (t br, 5.2) 67.1 5.17 (t, 5.1)
4ax 30.2 2.25 2.33 (m) 33.6 2.17 2.24 (m)
4eq 1.74 (d,15) 1.89 (d br, 14.9)
5 64.9 3.29 3.31 (m) 66.6 3.25 (s br)
6 78.6 5.75 (d, 6.1) 80.0 5.77 (dd, 7.5, 3.1)
7exo 76.6 3.51 (OH, d, 8.5) 37.1 2.20 2.27 (m)

Original Paper
7endo 4.72 (t br, 6.7) 2.68 (dd, 13.8, 7.5)
8 37.6 2.62 (CH3, s) 40.1 2.53 (CH3, s)
9 161.4 161.3
10 123.6 124.0
11 118.4 6.92 6.95 (m) 115.6 6.82 6.84 (m)
12 108.2 6.07 (dd, 3.9, 2.6) 110.4 6.19 6.22 (m)
13 130.2 6.94 6.96 (m) 124.1 7.03 7.06 (m)

Downloaded by: Dot. Lib Information. Copyrighted material.

14 36.7 3.94* (CH3, s) 10.91 (NH, s br)
15 160.6 160.8
16 123.5 123.4
17 118.5 7.16 (dd, 4.0, 1.8) 118.3 7.05 7.08 (m)
18 108.6 6.15 (dd, 4.0, 2.5) 108.5 6.12 (dd, 4.0, 2.5)
19 130.8 7.01 7.03 (m) 130.7 7.00 7.01 (m)
20 36.7 3.93* (CH3, s) 36.7 3.94 (CH3, s)

a
Assignments marked with an asterisk are interchangeable

duced by infusion into the perfusate by injection into the tubing regard to the given specifications (Table 1). Ten of the 14 pro-
at the entrance to the perfusion chamber at a rate of 100 mL/min ducts had labels indicating the plant sources of the respective 997
using a syringe pump (Harvard Apparatus, Model 22, USA). drug or at least the plant family, five samples had no plant speci-
fication at all. The declarations of four samples bark of
Anemopaegma mirandum (no 1, 12) and A. glaucum (no 5, 7)
Results and Discussion were questionable anyway. Both cited Anemopaegma species re-
present shrubs and no trees which only produce a thin bark un-
Microscopic characters of Trichilia catigua bark likely to be used for trading. Furthermore, Catuaba samples ori-
The fibrous bark disintegrates easily, has a greyish cork outside ginating from Anemopaegma mirandum do not consist of the
and is of distinct reddish-brown colour on the inner side. A hot bark but the rhizomes of the plant [5]. The specification of an-
water infusion turns red. In cross-section (Fig. 2) the outermost
tissue is a cork layer followed by a few rows of parenchyma
which may contain oxalate druses or crystals in varying num-
bers. Usually, there are groups of stone cells or a disrupted layer
of stone cells right below this parenchyma. The secondary cortex
is built up of parenchyma and groups of pitted fibres which are
arranged in tangential rows. The fibre bundles are accompanied
by cell rows that have many solitary crystals (Fig. 3). Occasional-
ly, oxalate druses or crystals appear in the parenchyma of the
secondary cortex. A brownish-yellow secretion is seen in many
parenchyma cells and in the intercellular spaces. The medullary
rays are built up of one row of parenchyma cells which are
turned into stone cells in between the fibre bundles. The bark
contains many rounded, solitary and compound starch grains
measuring 2 to 8 mm in diameter.

Examinations of the Catuaba samples

We examined the identity and purity of 14 commercial Catuaba Fig. 2 Cross section of Trichilia catigua bark, 100. A: layer of stone
products (chopped drugs, powdered drugs, one dried extract) in cells; B: fibre bundle.

Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000

Fig. 3 Fibre bundles and medullary rays of Trichilia catigua bark, long-
itudinal section, 400. A: pitted fibre; B: stone cell of the medullary
Original Paper
Chemical analyses
Dichloromethane extracts of the fourteen commercial samples
(see sample preparation and Table 1) were screened by TLC and
HPLC. Detection under UV 254 nm and 366 nm showed big differ-
ences of the fingerprints after TLC in both mobile phases. The sam-
ples (no 14, 15, 16) declared as Trichilia catigua did not correspond
to the references (no 6, 8) according to UV 366 nm, which might be
explained by adulterations that contributed additional substances
of unknown identity and, therefore, caused changes of the TLC fin-
gerprints. A TLC check on the sesquiterpenes of the calamene-type
and on flavalignans reported previously for Trichilia catigua [18],
[19] could not be realised, as neither reference substances nor ana-
lytical data about these rare compounds were available.

ray; C: row of single crystals.
other sample bark of Erythroxylon catuaba (no 9) was
equally questionable. Erythroxylum catuaba is considered to
be a doubtful species, as discussed earlier in this communica-
Seven out of fourteen commercial samples (no 1, 9, 11, 12, 13, 14
and 15) showed a positive reaction for alkaloids after TLC (Fig. 4).
The pattern differed with regard to quality and quantity. Among
these seven samples, only one sample (no 11) was characterised
by alkaloid zones with high intensity (RF 0.3 0.8). The others re-
vealed alkaloid spots of much lower intensity in the same RF

Downloaded by: Dot. Lib Information. Copyrighted material.

tion. range indicating that the samples were mixtures consisting of
sample 11 and one or more additional alkaloid-free barks. For
Microscopic analyses further characterisation of these alkaloids an HPLC system was
The microscopic analyses of the samples pointed at Trichilia established running a two-step gradient of methanol-water on
catigua as the main plant source, irrespective of the products' de- Supersorb RP SelectB material at pH 2.5. The addition of different
clarations. All but one (no 11) chopped crude drug samples con- acids such as formic acid, acetic acid, trifluoroacetic acid or phos-
tained a bark which exhibited the same macroscopic and micro- phoric acid was tested as well as the elution with various gradi-
scopic characters as authentic Trichilia catigua bark. No rhizome ent systems. However, best separations were achieved by run-
parts of Anemopaegma mirandum could be found in any Catuaba ning a phosphate buffer in a two-step gradient. During the HPLC
products, not even in those samples which should originate from runs UV spectra were recorded on-line, giving typical curves for
this plant. Two samples (no 7, 16) were adulterated with bark the tropane alkaloids with lmax at 275 nm and a slight shoulder
chops of an unknown plant species and one sample (no 4) even
998 contained pieces of at least two different bark drugs. These adul-
terations were recognised by different anatomical details such as
diverging arrangements of stone cells and/or fibres which were
sometimes accompanied by rows of crystal-containing cells.
Some fibres had conspicuous yellow cell walls. Another chopped
sample without species declaration (no 11) consisted of a bark
which was also found as additive in other three Catuaba samples
(no 1, 9, 15). The bark sample no 11 showed a positive alkaloid re-
action in the TLC screening (Fig. 4). This bark exhibited anatomi-
cal characters which corresponded greatly to those of the twig
bark of Erythroxylum vacciniifolium herbarium specimens. Due
to the small amount of twig bark samples we could not run a
TLC to confirm the microscopic findings. Unfortunately, no au-
thentic stem bark of Erythroxylum vacciniifolium was available
for the investigations. All powdered drug samples (no 3, 5, 10,
12, 14) regardless of their declarations exhibited characteristics
(such as the presence of fibres and stone cells) which were also
typical of authentic Trichilia catigua bark powder. However,
these samples could not be assigned microscopically beyond
doubt to Trichilia catigua, because the characteristic structural
pattern the specific arrangement of fibres and stone cell layers
got inevitably destroyed during the powdering processes.
Three of these drug powders (no 3, 5, 10) were adulterated with
cells supposedly originating from other bark and fruit drugs. One Fig. 4 TLC of dichloromethane extracts of the Catuaba samples under
UV 254 nm and after detection with the alkaloid-selective reagent po-
Catuaba product (no 13) specified as dried extract of
tassium iodoplatinate reagent. Stationary phase: silica gel 60 F254
Anemopaegma mirandum contained maize starch and showed a Merck; mobile phase: toluene-acetone-methanol-ammoniaconc. (45
positive alkaloid reaction. + 45 + 7 + 3); detection: UV 254 nm (above), potassium iodoplatinate
reagent (below).

Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000
Table 3 Alkaloids from sample 11 after HPLC and TLC (Fig. 5): RF values, retention times, quasi molecular ions and UV spectrum. * Not detected
on TLC

Peak tR [min] [M+H]+ m/z RF UV spectrum

a 4.5 265 > 281 0.36

b 5.8 281 > 265 0.06
c 7.5 265 0.39
d 9.1 265 n. d.*
e 26.7 404 0.11
1 28.1 388 0.49
2 29.5 358 0.47
f 31.0 374 n. d.*
3 36.0 430 0.63

Original Paper
4 38.9 372 0.57

Downloaded by: Dot. Lib Information. Copyrighted material.

at 240 nm (Table 3). The HPLC fractions resulting from sample 11
(Fig. 5) were collected, extracted with dichloromethane after addi-
tion of ammonia and used for TLC and MS analyses after evapora-
tion of the organic solvent. Thus the compounds were character-
ised by their retention times, RF values, UV and mass spectra (Ta-
ble 3), all of them showed UV spectra typical for tropane alkaloids.
Fractions 1 and 2 (Fig. 5) were identified as catuabine D (2) and its
hydroxymethyl derivative (1) by co-injection with the pure sub-
stances. According to MS fractions 3 and 4 correspond to the re-
cently published structures 3 and 4 [15] (Fig. 6). For the fractions
a f quasi-molecular ions of 265 amu (a, b, c and d), 281 amu (a
and b), 404 amu (e) and 374 amu (f) were detected (Table 3) which
correspond to different isomers according to [15] and [20] (Fig. 6). 999

Fig. 5 shows the HPLC analyses of the samples 1, 11 and 15, all
of them exhibit positive alkaloids spots on TLC. Even though

Fig. 6 Structures of the tropane alkaloids 1 4 [12], [13], [15] and

suggested isomers for a f according to [20].

we used the same drug amounts for extraction, the peaks in

Fig. 5 HPLC of dichloromethane extracts of samples 1, 11, 15 and TLC
of sample 11. HPLC: stationary phase: Supersorb RP SelectB 4 mm,
125 3 mm; mobile phase: methanol-water (0.05 M NaH2PO4, H3PO4
pH 2.5), 10 55 % in 15 min (rate 1 %/min), 25 55 % in 60 min (rate
0.5 %); flow rate: 1.0 mL/min; detection: UV-DAD, monitoring wave-
length 275 nm. TLC: stationary phase: silica gel 60 F254 Merck; mobile
phase: toluene-acetone-methanol-ammoniaconc (45 + 45 + 7 + 3); de-
tection: UV 254 nm. * Not detected on TLC.
HPLC differed in intensity. According to the retention times
and UV spectra the tropane alkaloids 1 (tR 28 29 min) and 2
(tR 29 30 min) seem to be present in all three samples, al-
though only in traces in samples 1 and 15. Sample 15 was char-
acterised by additional peaks (tR 23, 27, 31, 32 min) for which a
tropane ester structure could be excluded due to the differing
UV spectra. These differences in alkaloid concentration con-
firmed the assumption that several bark samples were mixed
with the tropane alkaloid-containing drug no 11. The referen-
ces, samples 6 and 8 (Trichilia catigua) were devoid of alka-
loids.
Functional tests
The in vitro tests on the rabbit corpus cavernosum were carried
out with extracts of authentic Trichilia catigua bark (no 8), ex-
tracts of Catuaba sample 1 and tropane alkaloid-enriched frac-
tions of Catuaba products (no 1, 11). The samples were applied

Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000
 in 1/1000, 1/100, 1/10 dilutions and in undiluted form. None of
the applications caused any change in the tone or in the magni-
tude or duration of the EFS-induced relaxations.
Conclusion
The combination of microscopic and phytochemical methods to
check the quality of the drug samples led to a quick characterisa-
tion and assessment of the products. Microscopy was an ideal
tool to detect Trichilia catigua bark as the main ingredient of at
least seven samples and allowed us to find various adulterations.
Due to the scarce information on the chemical compounds of
Trichilia catigua bark, a drug-specific phytochemical check of
Original Paper
the samples could not be realised. Furthermore, a comparison of
the TLC fingerprint with authentic reference material gave no
evidence of the drug's presence in the samples, because the adul-
terations and mixtures of barks led to an invalid fingerprint chro-
matogram. Other phytochemical investigations indicated the
presence of alkaloid esters of tropanol with pyrrolecarboxylic
acid in several Catuaba samples. The unique acid moiety of the
alkaloids esters pointed to the Catuaba-furnishing species
Erythroxylum vacciniifolium Mart. [15], [20], [21], [22]. Trichilia
catigua reportedly does not contain any alkaloids [7], neither
have alkaloids been traced so far in those Anemopeagma species
which are traded as Catuaba. The poor quality of the examined
products raises the question about the safety of such herbals.
Taking into account that, e. g., tropane alkaloids exert various ef-
fects and plants with unknown identity may contain toxic con-
stituents with unpredictable interactions, the ingestion of such
products bears heavy risks. Regarding the claimed aphrodisiac
effects of the Catuaba products, a check of aqueous and metha-
1000 nolic extracts of Trichilia catigua bark and the supposed
Erythroxylum vacciniifolium bark drug did not give evidence of
such properties, at least not in the applied in vitro test system
on the rabbit corpus cavernosum. Nevertheless, one hypothesis
to explain the popularity of Catuaba in Brazil for decades of years
might be an effect on the central erection/libido centres, evi-
dence for this is missing up to now.
Acknowledgements
We are most grateful to Prof. Dr. Richard Wasicky, Sao Paolo, Bra-
zil, who helped us unceasingly in acquiring reference material for
Trichilia catigua and to locate relevant Brazilian publications. We
also express our thanks to Prof. Dr. L. C. Marques, University of
Maring, Centre of Pharmacy and Pharmacology, Laboratory of
Pharmacognosy, Maring, Paran, Brazil, for donating authentic
bark samples of Trichilia catigua. We thank the University of Ulm,
Herbarium, and the Botanische Staatssammlung Mnchen for
providing samples of Erythroxylum vacciniifolium. We greatly ap-
preciate the support of Dr. Karl Stifter, Vienna, Austria, who raised
the Catuaba topic and initiated the contacts with Brazil and sup-
plied us with various Brazilian commercial Catuaba samples. We
also thank Frans v. d. Dungen, Helmut Dreyssig and Daglef Mor-
Kletter C et al. Morphological, Chemical and Planta Med 2004; 70: 993 1000
tenthaler who purchased Catuaba samples from various markets
in Brazil. Diploma theses were carried out by Anna Schneider, Sa-
bine Schlucker, Monika Kaniak and Wolfgang Stindl.
References
1
Daly DC. The genus Tetragastris and the forests of eastern Brazil. Stud-
ies in neotropical Burseraceae III. Kew Bull 1990; 45: 179 94
2
Ducke A. A catuaba na botnica sistemtica, cientfica e pseudo-cien-
tfica. Rev Bras Farm 1966; 47: 267 72
3
Corra PM. Diccionrio das plantas teis do Brasil e das exticas culti-
vadas. Ministerio da Agricultura. Rio de Janeiro 1931; 2: 150 2
4
Hamet R. Sur l'origine botanique des drogues dsignes au Brsil sous
le nom de Catuaba. Comptes Rendues Paris 1936; 203: 1178 9
5
Da Silva RAD. Catuaba (Anemopaegma mirandum (Chamisso) Alph. De
Candolle). Pharmacopeia dos Estados Unidos do Brasil 1926; Compan-
hia Editora Nacional, So Paulo, Brasil:
6
Da Silva RAD. Catuaba. Rev Flora Med, 1934: 211 24
7
Marques LC. Contribuico ao esclarecimento da identidade botnica
da droga vegetal Catuaba. Espec. de Capa,1998; Mar/Abr: 8 11
8
Hamet R. Sobre alguns effeitos physiologicos da droga brasileira con-
hecida pelo nome de folhas de Catuaba. Comptes Rendus des Sances
de la Socit de Biologie et de ses filiales 1937; 124: 904 7
Downloaded by: Dot. Lib Information. Copyrighted material.
9
Antunes E, Gordo WM, de Oliveira JF, Teixeira CE, Hyslop S, De Nucci G.
The relaxation of isolated rabbit corpus cavernosum by the herbal
medicine Catuama (and its constituents). Phytother Res 2001; 15:
416 21
10
Vaz ZR, Mata LV, Calixto JB. Analgesic effect of the herbal medicine Ca-
tuama in thermal and chemical models of nociception in mice. Phyto-
ther Res 1997; 11: 101 6
11
Evans WC. The comparative phytochemistry of the genus Erythroxylon.
J Ethnopharmacol 1981; 3: 265 77
12
Glasl S, Presser A, Werner I, Haslinger E, Jurenitsch J. Tropane alka-
loids from a Brazilian bark traded as `Catuaba'. Sci Pharm 2003; 71:
113 9
13
Glasl S, Presser A, Werner I, Haslinger E, Jurenitsch J. Erratum to Tro-
pane alkaloids from a Brazilian bark traded as `Catuaba'. Sci Pharm
2004; 72: 97
14
Merck E. Anfrbereagenzien fr Dnnschicht- und Papier-Chromato-
graphie. Darmstadt:1980: Reagens Nr. 177, p 58
15
Zanolari B, Guilet D, Marston A, Queiroz EF, Paulo MQ, Hostettmann K.
Tropane alkaloids from the bark of Erythroxylum vacciniifolium. J Nat
Prod 2003; 66: 497 502
16
Cellek S, Moncada S. Nitrergic control of peripheral sympathetic re-
sponses in the human corpus cavernosum: a comparison with other
species. Proc Natl Acad Sci USA 1997; 94: 8226 31
17
Kasakov L, Cellek S, Moncada S. Characterization of nitrergic neuro-
transmission during short- and long-term electrical stimulation of
the rabbit anococcygeus muscle. Br J Pharmacol 1995; 115: 1149
54
18
Pizzolatti MG, Venson AF, Smnia Jnior A, Smnia EFA, Braz-Filho R.
Two epimeric flavalignans from Trichilia catigua (Meliaceae) with an-
timicrobial activity. Z Naturforsch 2002; 57c: 483 8
19
Garcez WS, Garcez FR, Ramos L, Camargo MJ, Damasceno jr. GA. Ses-
quiterpenes from Trichilia catigua. Fitoter 1997; 68: 87 8
20
Zanolari B, Wolfeneder JL, Guilet D, Marston A, Queiroz EF, Paulo MQ,
Hostettmann K. On-line identification of tropane alkaloids from
Erythroxylum vacciniifolium by liquid chromatography-UV detection-
multiple mass spectrometry and liquid chromatography-nuclear
magnetic resonance spectrometry. J Chromatogr A 2003; 1020: 75
89
21
Graf E, Lude W. Alkaloide aus Erythroxylum vacciniifolium Martius, 1.
Mitt. Isolierung von Catuabin A, B und C. Arch Pharm (Weinh) 1977;
310: 1005 10
22
Graf E, Lude W. Alkaloide aus Erythroxylum vacciniifolium Martius, 2.
Mitt. Strukturaufklrung von Catuabin A, B und C. Arch Pharm
(Weinh) 1978; 311: 139 52