Convulsant activity and neurochemical alterations induced by a fraction obtained from fruit (Oxalidaceae: Geraniales)

www.elsevier.com/locate/neuint

Neurochemistry International 46 (2005) 523–531

Convulsant activity and neurochemical alterations induced

by a fraction obtained from fruit Averrhoa

carambola (Oxalidaceae: Geraniales)

Ruither O.G. Carolino a, Rene O. Beleboni a, Andrea B. Pizzo b,Flavio Del Vecchio c, Norberto Garcia-Cairasco c,*, Miguel Moyses-Neto d,

Wagner F. dos Santos b, Joaquim Coutinho-Netto a

a Departamento de Bioquımica e Imunologia, Faculdade de Medicina de Ribeirao Preto, SP, Brazilb Departamento de Biologia, Faculdade de Filosofia, Ciencias e Letras de Ribeirao Preto,

Universidade de Sao Paulo, Ribeirao Preto, SP, Brazilc Departamento de Fisiologia, Faculdade de Medicina de Ribeirao Preto, Universidade de Sao Paulo,

Avenida Bandeirantes, 3900, Bairro Monte Alegre, 14049-900 Ribeirao Preto, SP, Brazild Divisao de Nefrologia, Departamento de Medicina Interna, Hospital das Clınicas da Faculdade de

Medicina de Ribeirao Preto, Universidade de Sao Paulo, Ribeirao Preto, SP, Brazil

Received 4 August 2004; received in revised form 8 February 2005; accepted 9 February 2005

Available onli

Abstract

We obtained a neurotoxic fraction (AcTx) from star fruit (Averrhoa carambola) and studied its effects on GABAergic and glutamatergic

transmission systems. AcTx had no effect on GABA/glutamate uptake or release, or on glutamate binding. However, it specifically inhibited

GABA binding in a concentration-dependent manner (IC50 = 0.89 mM). Video-electroencephalogram recordings demonstrated that following

cortical administration of AcTx, animals showed behavioral changes, including tonic-clonic seizures, evolving into status epilepticus,

accompanied by cortical epileptiform activity. Chemical characterization of AcTx showed that this compound is a nonproteic molecule with a

molecular weight less than 500, differing from oxalic acid. This neurotoxic fraction of star fruit may be considered a new tool for

neurochemical and neuroethological research.

# 2005 Elsevier Ltd. All rights reserved.

Keywords: Averrhoa carambola; Convulsant; Video-electroencephalogram; GABA/glutamate uptake, release and binding

1. Introduction

Star fruit or carambola (Averrhoa carambola, Oxalidaceae

family), originally from Asia, has become acclimatized in

many tropical countries, including Brazil. Star fruit prepara-

tions are used in traditional medicine in China and Malaysia to

treat headache, vomiting, coughing and restlessness. However,

Muir and Lam (1980) reported that injecting star fruit extract

into the peritoneal cavity of mice caused seizures and death.

Later, eight uremic patients were shown to develop intractable

hiccups following star fruit ingestion (Martin et al., 1993).

* Corresponding author. Tel.: +55 16 602 3330; fax: +55 16 633 0017.

E-mail address: [email protected] (N. Garcia-Cairasco).

0197-0186/$ – see front matter # 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.neuint.2005.02.002

More recently, Moyses-Neto et al. (1998) reported six patients

on a dialysis program, that developed various neurological

symptoms, and one patient died. Many observations of

intoxication following star fruits or juice ingestion by patients

with renal failure have been subsequently reported. Clinical

manifestations of such cases included persistent, intractable

hiccups, vomiting, variable degrees of disturbed conscious-

ness (mental confusion, psychomotor agitation), decreased

muscle strength, limb numbness, paresis, insomnia, paresthe-

sia and seizures (Chang et al., 2000; Lo et al., 2001; Chan et al.,

2002; Wu et al., 2002; Yap et al., 2002; Moyses-Neto et al.,

2003; Tse et al., 2003). Although the chemical nature of the

star fruit neurotoxin remains obscure, oxalic acid has been

proposed as a putative candidate (Chen et al., 2001).

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531524

g-Aminobutyric acid (GABA) and glutamate are,

respectively, the most predominant inhibitory and excitatory

neurotransmitters in the central nervous system, and an

imbalance between these systems may lead to hyperexcit-

ability, producing seizures (Fonnum, 1984; Bradford, 1995;

Engelborghs et al., 2000).

Certain plant toxins may alter synthesis, synaptic release

or postsynaptic action of GABA, leading to convulsions

(Chebib and Johnston, 2000). Ricinine, a neutral alkaloid

from Ricinus communis, evokes seizures probably by acting

on the benzodiazepine site of the GABAA receptor.

Biochemical, pharmacological and behavioral studies

involving new convulsant compounds from plants and other

sources, may contribute towards a better understanding of

mechanisms involved in epilepsy, aiding in the development

of new therapeutic approaches for the treatment of this

disorder (Ferraz et al., 2000).

In the present work, we describe the isolation of the

convulsant fraction of the star fruit, and the investigation of

its actions on GABAergic and glutamatergic transmission

systems. In addition, we report its behavioral effects and

results of electrophysiological studies using video-electro-

encephalography (video-EEG). Preliminary data from these

studies have been presented elsewhere in abstract form

(Carolino et al., 2003; Garcia-Cairasco et al., 2002).

2. Experimental procedures

2.1. Material

CM-cellulose-52 was purchased from Whatman (UK)

and Matrex Cellufine A200 DEAE from Amicon Corpora-

tion (USA). Columns for HPLC were from Merck

(Germany) and Shimadzu Techno-Research, Inc. (Japan).

Highly purified ammonium bicarbonate (NH4HCO3) was

prepared by crystallizing this salt from a mixture of cold

ammonium hydroxide and dry ice at pH 7.8–8.0 (Sampaio

et al., 1983). Unlike some commercial brands of ammonium

bicarbonate, this product does not leave any residue

following lyophilization.

4-Amino-n-[2,3-3H]-butyric acid ([3H]-GABA) (89 Ci/

mmol), L-[U-14C]-glutamate (272 mCi/mmol) and L-

[G-3H]-glutamate (49 Ci/mmol) were obtained from Amer-

sham Biosciences (UK). Membrane filters (AAWP01300)

were purchased from Millipore (Brazil).

All other reagents were of analytical grade and purchased

from Merck (Germany) or Reagen (Brazil).

2.2. Fruit collection and preparation of the aqueous

extract

Star fruits were harvested from pesticide-free trees at the

Campus of the University of Sao Paulo, Ribeirao Preto, SP,

Brazil, washed, homogenized in distilled water 1:4 (w/v)

and centrifuged at 1465 � g for 5 min at 4 8C. The extract

obtained was used in the purification procedures described

below.

2.3. Isolation of the convulsant fraction from

the star fruit

The neurotoxic fraction (AcTx) was obtained by three

chromatographic steps, carried out at room temperature and

with absorbance being monitored at l = 280 nm. The

aqueous extract of star fruit was sequentially chromato-

graphed on anion exchange column (Matrex Cellufine A200

DEAE), then on cation exchange column (CM-cellulose-52)

and, finally, on reversed-phase high performance liquid

chromatography (HPLC). In each step of the fractionation

procedure the fractions obtained were assayed (as described

in Section 2.8) and the neurotoxic fraction was lyophilized

and submitted to the next step.

2.4. Neurotoxic fraction quantification

AcTx was quantified spectrophotometrically, using the

2,4,6-trinitrobenzene 1-sulfonic acid (TNBS) method

(Satake et al., 1960; Cayot and Tainturier, 1997). A standard

reference curve of glycine was used, and AcTx concentra-

tions were expressed in moles of glycine.

2.5. Amino acid analyses

AcTx (6.25 pmol of glycine) was hydrolyzed by HCl

(6N) in a sealed microcapillary tube at 110 8C for 24 h.

Amino acid analyses of the hydrolyzates were performed on

an automated analyzer equipped with a single (6 mm �220 mm, Beckman W-3 resin) column, using a ninhydrin

reagent system for colorimetric determination of amino

acids (Paula et al., 1998).

2.6. Analysis of molecular weight by ultrafiltration

AcTx (5.3 mM, final concentration) was dissolved in

phosphate buffered saline pH 7.4 and submitted to

ultrafiltration under N2 pressure (Amicon UM05 membrane,

Amicon Corp., USA). The presence of the neurotoxic

fraction on the filtrate was determined using assay for

neurotoxicity, as described in Section 2.8.

2.7. Animals

Wistar rats and male Swiss white mice were bred at the

University of Sao Paulo (Ribeirao Preto, SP, Brazil). Animals

were kept on a 12:12 h light:dark cycle, at room temperature,

and provided with food and water ad libitum. Animals were

decapitated without the use of anesthetics. All used procedures

followed the guidelines established in Guide for the Care and

Use of Laboratory Animals, approved by the Brazilian Society

for Neuroscience and Behavior. Every effort was made to

avoid unnecessary stress and pain to the experimental animals.

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531 525

2.8. Assay for neurotoxicity

Three microliters of different concentrations of each

chromatographic fraction were injected into the cisterna

magna of mice weighing between 15 and 20 g. Controls

received saline only. Chromatographic fractions capable of

inducing tonic-clonic convulsions or death were considered

neurotoxic.

2.9. Preparation of synaptosomes

Cerebral cortices from male Wistar rats (200–250 g) were

used to prepare synaptosomes as described by Gray and

Whittaker (1962). Synaptosomes were resuspended in

Krebs-phosphate buffer (NaCl 124 mM; KCl 5 mM;

KH2PO4 1.2 mM; CaCl2 0.75 mM; MgSO4 1.2 mM;

Na2HPO4 20 mM; glucose, 10 mM, pH 7.4) and centrifuged

for 20 min at 4 8C. Protein content was determined by the

Lowry method (Lowry et al., 1951), modified by Hartree

(1972).

2.10. Neurotransmitter uptake assays

Synaptosomes were resuspended in Krebs-phosphate

buffer and preincubated for 15 min at 25 8C (GABA assays)

or 37 8C (glutamate assays) in the absence or presence of

five increasing concentrations of AcTx (from 0.026 to

16.0 mM). Uptake assays were initiated by adding [3H]-

GABA (6.74 nM, final concentration) or L-[U-14C]-gluta-

mate (36 nM, final concentration), to synaptosome suspen-

sions (100 mg of protein ml�1), and incubated for 3 min at

the temperatures specified. All reactions were stopped by

centrifugation (3000 � g, 3 min at 4 8C). Supernatants were

discarded; pellets were washed twice with ice-cold distilled

water, homogenized in 10% trichloroacetic acid (TCA) and

centrifuged (3000 � g, 3 min at 4 8C). Aliquots of super-

natants were transferred to scintillation vials containing 5 ml

of the biodegradable scintillation cocktail ScintiVerse

(Fisher Scientific), and their radioactivity quantified in a

scintillation counter (Beckman, LS-6800) with a counting

efficiency of 85% for 14C, and of 35–40% for 3H. Results

were expressed as averages of uptake velocities with their

standard errors of the mean (S.E.M.), in fmol min�1 mg�1.

Nonspecific uptake was determined from samples incubated

in the presence of nipecotic acid 6 mM (GABA) or at 4 8C(glutamate). These values were subtracted from those of the

total uptake. Statistical analyses were performed using

Student’s t-test.

2.11. Neurotransmitter release assays

Synaptosomes (3 mg ml�1) were pre-loaded with 0.5 mM

[3H]-GABA or with 0.7 mM L-[U-14C]-glutamate in Krebs-

phosphate buffer, for 20 min at 25 or 37 8C, respectively.

Samples were centrifuged for 3 min at 7200 � g at 4 8C, and

pellets washed three times with the cold buffer. To assess

neurotransmitter release, the final pellet was resuspended in

Krebs-phosphate buffer and incubated in the absence or

presence of five increasing concentrations of AcTx (from

0.026 to 16.0 mM) for 3 min at 25 8C (GABA) or 37 8C(glutamate). Neurotransmitter release was also measured in

the presence of 50 mM KCl and 5 mM tetrodotoxin (TTX),

used to verify the functional properties of the synaptosomal

preparations. Reactions were stopped by centrifugation

(3000 � g, 3 min at 4 8C), and aliquots of supernatants and

pellets were separately transferred to scintillation vials

containing 5 ml of biodegradable scintillation cocktail;

radioactivity was quantified in a scintillation counter. The

amounts of GABA or glutamate released were calculated as

percentages of the total amounts of radiolabel present in the

synaptosomes at the beginning of the incubation period.

Statistical analyses were performed using Student’s t-test.

2.12. Preparation of synaptic membranes

Cerebral cortices from male Wistar rats (200–250 g) were

used to prepare synaptic membranes as described by Zukin

et al. (1974). Washed membranes were resuspended in

50 mM Tris–HCl buffer, pH 7.4 and their protein content

determined as described above.

2.13. Binding assays

Binding assays were initiated by adding [3H]-GABA

(10 nM, final concentration) or L-[G-3H]-glutamate

(25 nM, final concentration) to synaptic membranes

(1 mg of protein ml�1) in the absence or presence of five

increasing concentrations of AcTx (from 8.5 nM to

26.7 mM). Samples were incubated for 30 min at 25 8Cfor GABA, or 37 8C for glutamate binding assays.

Reactions were stopped by vacuum filtration through

Millipore membrane filters, followed by two washings

with cold buffer. The washed filters were transferred to

scintillation vials containing 5 ml of biodegradable

scintillation cocktail, and the radioactivity quantified by

a scintillation counter. Nonspecific binding was deter-

mined in the presence of nonlabeled GABA or glutamate

(each at 1 mM final concentration), respectively.

Specific binding was determined by subtraction of the

averages of nonspecific binding. Results were expressed as

averages � S.E.M. of percentages of controls. Statistical

analysis obtained in each experiment in the presence or

absence of AcTx, was performed using ANOVA one-way

test followed by Student–Newman–Keul’s post hoc test.

Dose-response curves were fitted by nonlinear regression

analysis using GraphPad Prism (Version 2.0, GraphPad

Software, San Diego, CA, USA).

2.14. Video-electroencephalographic recording

Wistar rats (n = 5; 200–250 g), were implanted with

chemitrodes in their cortices following the stereotaxic

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531526

coordinates: AP, 1.5 mm (bregma); L, 2.5 mm (midline); V,

1.5 mm (dura), according to Paxinos and Watson (1998).

Control electroencephalograms (EEG) (baseline) were

Fig. 1. Isolation of AcTx from the star fruit. The aqueous extract was

sequentially fractionated in DEAE (A) and CM-cellulose (B) columns, using

ammonium bicarbonate buffers (pH 7.8). The CM pool was applied to a

reversed-phase HPLC column (Shimadzu, PREP-ODS II, 20 mm � 250 mm,

5 mm) (C), previously equilibrated in0.05% trifluoroacetic acid (TFA) inMilli-

Q water and then eluted (phase B) with a linear gradient of water/acetonitrile/

TFA 35:65:0.05, at flow rate of 3 ml min�1. Eluates were monitored at

l = 280 nm. Dotted lines indicate mobile phase concentration. Denotes

active fractions.

recorded prior to (controls) and after vehicle (0.9% saline

solution) injection, for 30 min prior to AcTx application

(17 mM/1 ml, n = 2; 170 mM/1 ml, n = 3), during a total

recording time of 280 min. The electroencephalographic

recordings were made using a Cyberamp (Axon Instru-

ments; USA) and a Biopac (USA) MP100 amplifier and an

A/D converter, respectively. An All-in-wonder PRO video

card (ATI; USA) was used to synchronize the video signal

with the electroencephalographic recordings, which were

called video-EEG (Moraes et al., 2000). Only animals that

had the chemitrode position confirmed by histology were

used (Fig. 5C).

3. Results

3.1. Purification and amino acid analyses of AcTx

Fig. 1 shows chromatographic profiles of the three steps

of the fractionation of the star fruit aqueous extract. All

fractions obtained by each chromatographic step were

assayed for their ability to produce seizures or death in

mice. The profile of the star fruit extract applied on a

DEAE-cellulose column is illustrated in Fig. 1A. The

fractions presenting convulsant activity, were pooled,

lyophilized and rechromatographed on a CM-cellulose

column (Fig. 1B); the resulting active fractions were united

in a pool, which following reversed-phase HPLC column

produced the active fraction AcTx (Fig. 1C). Upon

biochemical characterization, this fraction failed to

demonstrate the presence of amino acids in its composition

(data not shown).

3.2. Absence of effects of AcTx on neurotransmitters

uptake and release

Fig. 2 shows that AcTx had no effect on GABA (hatched

bars) or glutamate (black bars) uptake. Each uptake assay

was accompanied in parallel by a lactic acid dehydrogenase

(LDH) activity measurement. No increase in LDH was

observed in the supernatants, indicating that the synapto-

somes maintained their integrity in the presence of the

concentrations of AcTx tested (data not shown).

Fig. 3 shows the results of the effect of five increasing

concentrations of AcTx on neurotransmitter release. AcTx

did not alter baseline release of GABA (hatched bars) or

glutamate (black bars). GABA and glutamate basal releases

were increased by 75 � 2% in the presence of 50 mM KCl,

and TTX did not alter the basal releases of these

neurotransmitters, as expected.

3.3. Effects of AcTx on receptor binding

To obtain a better insight into the mechanism of action of

AcTx, we studied the effects of this neurotoxic fraction on

GABA and glutamate binding (Fig. 4).

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531 527

Fig. 2. Effect of increasing concentrations of AcTx on [3H]-GABA

(hatched bars) and L-[14C]-glutamate (black bars) uptake in synaptosomes

from rat cerebral cortex. Student’s t-test showed no difference between the

uptake velocity of GABA or glutamate in the presence of any concentration

of AcTx used. Each bar is the average � S.E.M. of three independent

experiments, each done in triplicate.

Fig. 4. Effects of AcTx on specific [3H]-glutamate (A) and [3H]-GABA (B)

binding in synaptic membranes. No effects on [3H]-glutamate binding were

observed at any concentration of AcTx; in contrast, a dose-dependent

inhibition of [3H]-GABA binding was observed (IC50 = 0.89 mM). Incuba-

tions were performed with 10 nM [3H]-GABA or 25 nM [3H]-glutamate in

the absence and presence of AcTx (8.5 nM to 26.7 mM) in 50 mM Tris/HCl

buffer. Each point is the average � S.E.M. of percentage of binding in

controls. Data were obtained from three independent experiments, each

done in triplicate. Statistical analyses were made using one-way ANOVA

test (p < 0.05).

A significant decrease of GABA binding was observed in

the presence of AcTx, which at a concentration of 26.7 mM,

produced a maximal decrease of 65% relative to the control,

and showed an IC50 value of 0.89 mM (Fig. 4A). In contrast,

no significant effect on glutamate binding was observed

under our experimental conditions (Fig. 4B).

3.4. Video-electroencephalographic recording

Video-EEG recording was started after placing the

animals in the recording chamber. Control behavior and

EEGs were examined in the basal situation that is, in the

absence of vehicle or AcTx microinjections. Animals

explored the cage and, being awake, displayed typical

desynchronized, high frequency-low amplitude EEG activ-

ity. Subsequent microinjections of 1 ml of the 0.9% saline

Fig. 3. Effects of increasing concentrations of AcTx on [3H]-GABA

(hatched bars) or L-[14C]-glutamate (black bars) release. Synaptosomes

were preloaded with L-[14C]-glutamate (0.7 mM) or [3H]-GABA (0.5 mM)

for 20 min at 37 or 25 8C, respectively. Release was started by the addition

of AcTx (from 26 nM to 16 mM) and is expressed as % of neurotransmitter

released over control. Student’s t-test showed no difference between the

effect in controls and of each concentration of AcTx used. Each bar is the

average � S.E.M. of three independent experiments, each done in triplicate.

vehicle did not modify the EEG, nor induce any

behavioral alterations of the animals (Fig. 5A; two upper

recordings). After 17 mM AcTx injection, EEG activity

began to be noticed 1–3 min after injection, but behavior

expression changes were only seen after the 170 mM

injections. Animals submitted to cortical microinjection of

170 mM AcTx presented strong progression of epilepti-

form EEG activity from 10 to 240 min (Fig. 5), which

began between 1 and 1.7 min following injection (n = 3).

The continued phenomenon observed over such a long

period of time in both cases is called status epilepticus.

Since none of the animals died after the AcTx injection,

and although we usually ended seizures studies after

90 min of status epilepticus, in order to perform chronic

behavioral studies (spontaneous recurrent seizures) as well

as cellular studies, the behavior and EEG evolution of the

animals presenting the data of Fig. 5A and B, were

observed for 280 min or more. Behavioral and EEG

effects of a cortical AcTx (170 mM/1 ml) microinjection

are shown in Fig. 6.

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531528

Fig. 5. Evolution of epileptiform activity in rat 02 after cortical microinjection of AcTx (170 mM). (A) Control and saline recordings express typical

desynchronized EEG in the waking state. Ten minutes after AcTx injection, the EEG activity recording indicates subtle baseline alterations with poly-spikes that

thereafter, from 20 to 240 min, evolved into a sustained electrographic status epilepticus. Note that at 30 min EEG recording activity was of opposite polarity

when compared for example, to the EEG recording at 240 min. (B) Observe the very weak EEG epileptiform activity following cortical injection of 17 mM/1 ml

AcTx, in comparison with the strong EEG epileptiform activity shown after cortical injection of 170 mM/1 ml of the toxin. Notice also, as shown in (A), that in

the second half of the recording period (at around 140 min), there occurred a clear-cut inversion of EEG polarity.

4. Discussion

It has been described that star fruit ingestion may produce

intoxication in patients showing renal failure. Clinical

manifestations include variable degrees of mental confusion,

psychomotor agitation, insomnia, paresthesias and seizures

(Moyses-Neto et al., 2003). Neurotoxic activity had been

previously demonstrated by seizures elicited in rodents after

intracerebroventricular administration of the star fruit

extract (Moyses-Neto et al., 1998). Normally, neuronal

excitability is maintained by a balance between excitatory

and inhibitory neurotransmission. This balance is disturbed

during and following diverse neurological conditions

including epilepsy (Inglefield et al., 1995); it has been

hypothesized that the neurological effects observed in star

fruit intoxication could also be due to a putative imbalance

between the glutamatergic and GABAergic systems.

In this work, star fruit extract was fractionated by three

chromatographic procedures, leading to the isolation of a

convulsant fraction, referred to as AcTx. This compound is

highly homogeneous since AcTx HPLC re-injection resulted

in a single chromatographic peak. Moreover, AcTx is a

dialyzable compound that is freely filtered by a 500 cut-

off membrane, indicating a molecular weight lower than

500.

Like most plants in the Oxalidaceae family, star fruit

contains an abundant amount of oxalic acid. Because of the

similar neurological manifestation caused by injection of

many high oxalic acid-containing plant extracts (Sanz and

Reig, 1992), it is prudent to consider oxalic acid as a causal

agent of star fruit-associated toxicological encephalopathy

in dialysis patients (Chen et al., 2001).

In order to prove that AcTx differs from oxalic acid, both

compounds were submitted to HPLC chromatography run

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531 529

Fig. 6. Behavioral and EEG effects of a cortical AcTx (170 mM/1 ml) microinjection. (A) Digitalized behavior sequence (16 frames captured in a video-EEG

setup). Aligned frames allow the detection of subtle behavioral alterations such as forelimb (white rectangles) and head (white circles) and myoclonic activity.

(B) Observe the EEG window with hypersynchronous epileptiform activity coinciding with video frames ranging from the 1st to the 16th in (A). (C) Cellular

Nissl staining showing cortical localization of chemitrodes used for AcTx microinjection.

under the identical conditions. Retention times were 17 min

for AcTx and 4 min for oxalic acid (data not shown), ruling

out the possibility that the star fruit convulsant neurotoxin

under consideration could be oxalic acid. Preliminary

chemical characterization by amino acid analysis did not

detect any amino acid residue in AcTx, suggesting that it is a

nonproteic molecule.

Synaptosomes are a well-recognized model for studying

neurotransmitter nerve-terminal-related mechanisms since it

retains all machinery for the uptake, storage, release of

neurotransmitter, and ionic conductance, while being

sufficiently simple and homogeneous for meaningful

biochemical studies (Gray and Whittaker, 1962; Bicalho

et al., 2002; Wang and Sihra, 2003). In our experimental

conditions, the morphological integrity of synaptosomal

preparation was confirmed by electron microscopy and by

LDH assays (data not shown). In addition, results obtained

from TTX and KCl in GABA and glutamate release

experiments showed that synaptosomes were functional and

responsive for pharmacological treatments.

We demonstrated that AcTx had no effects on GABA or

glutamate uptake, ruling out the possibility of glutamate and

GABA uptake alterations as being responsible for the

convulsant action observed in star fruit intoxication.

Subsequent evaluation of AcTx effects on GABA and

glutamate release showed that as observed in the neuro-

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531530

transmitter uptake assays, AcTx did not alter these

processes, suggesting that enhanced glutamate release or

decreased GABA levels in the synaptic cleft respectively, are

not responsible for the convulsant effects observed in star

fruit intoxication.

To obtain insight into the neurochemical mechanism of

AcTx action by another approach, we studied its effects on

GABA and glutamate binding on synaptic membranes.

AcTx did not alter glutamate binding under our assay

conditions. In contrast, it inhibited GABA binding in a

concentration-dependent manner with an IC50 value of

0.89 mM. AcTx appears to be an inhibitory substrate for

GABA receptors whose antagonistic blockade could explain

its convulsant effects in uremic patients. Therefore, we

suggest that the convulsant effect of AcTx may be due to

postsynaptic blocking of GABA receptors.

Video-EEG recordings following cortical administration

of AcTx showed behavioral changes, including partial

limbic seizures (forelimb and head myoclonus), evolving to

a status epilepticus, accompanied by sustained cortical EEG

epileptiform activity, particularly after the 170 mM AcTx

injection. A preliminary characterization of the video-EEG

after crude star fruit juice had been given to animals bearing

induced acute renal failure (Garcia-Cairasco et al., 2002).

Star fruit juice also induced seizures when applied to cortical

areas, showing that convulsant activity is present in crude

star fruit extracts. The present data confirm the excitatory

profile of AcTx and star fruit extracts, whose mechanism,

as shown by the present data, seems to be related to

decreased inhibition, rather than to increased synaptic

excitation. Indeed, the fact that the video-EEG setup

detected a progressive and sustained EEG epileptiform

activity (increasing neuronal synchronization) induced by

AcTx is a characteristic of potent excitatory convulsants,

like pilocarpine and kainic acid (Leite et al., 2002).

In agreement with our data, a variety of convulsants from

plants that interact with GABA receptors has been reported,

including bicuculline, picrotoxinin, ricinine, virol A and

calycanthine (Curtis et al., 1970; Johnston, 1996; Ferraz

et al., 2000; Uwai et al., 2001; Chebib et al., 2003). The use

of these molecules is relevant for the improvement of

molecular knowledge about GABA binding sites and for the

elucidation of behavioral and biochemical mechanisms

involved in neurological disorders, especially those asso-

ciated with GABAergic dysfunction.

In summary, the present results provide a new insight into

the effects of a convulsant fraction from A. carambola that

acts primarily on GABA receptors. Although the chemical

nature of AcTx remains unknown, preliminary data indicate

that this compound is a nonpeptide molecule differing

from oxalic acid. To conclude, this neurotoxic fraction of

star fruit is presented as a tool possibly capable of providing

new insights into the molecular machinery of GABAergic

neurotransmission, which may contribute towards the

elucidation of mechanisms involved in neurological

disorders.

Acknowledgements

We thank Vera L.A. Epifanio, Silvia H. Epifanio

(Biochemistry Department, FMRP-USP) for technical

assistance, Norberto Peporine Lopes and Leonardo Gobbo

Neto for critical reading of this manuscript. We also thank

FAPESP (numbers 03/00873-2, 00/08010-5 and 00/08101-

0), CAPES (scholarship to R.O.G. Carolino) and CNPq;

(numbers 473448/2003-3 and 500676/2003-8) for financial

support.

References

Bicalho, A.F., Guatimosim, C., Prado, M.A., Gomez, M.V., Romano-Silva,

M.A., 2002. Investigation of the modulation of glutamate release by

sodium channels using neurotoxins. Neuroscience 113, 115–123.

Bradford, H.F., 1995. Glutamate, GABA and epilepsy. Prog. Neurobiol. 47,

477–511.

Carolino, R.O.G., Beleboni, R.O., Moyses-Neto, M., Garcia-Cairasco, N.,

Coutinho Netto, J., 2003. Biological activity and purification of a

neurotoxin from Averrhoa carambola star fruit (Oxalidaceae). In:

Proceedings of the Seventh Symposium of the Brazilian Society on

Toxinology (SBTx), Pirenopolis, Brazil, 2002. J. Venom. Anim. Toxins

Incl. Trop. Dis. 9, 542.

Cayot, P., Tainturier, G., 1997. The quantification of protein amino groups

by the trinitrobenzenesulfonic acid method: a reexamination. Anal.

Biochem. 249, 184–200.

Chan, Y.L., Ng, H.K., Leung, C.B., Yeung, D.K.W., 2002. 31Phosphorous

and single voxel proton MR spectroscopy and diffusion-weighted

imaging in a case of star fruit poisoning. AJNR Am. J. Neuroradiol.

23, 1557–1560.

Chang, J.M., Hwang, S.J., Kuo, H.T., Tsai, J.C., Guh, J.Y., Chen, H.C., Tsai,

J.H., Lai, Y.H., 2000. Fatal outcome after ingestion of star fruit (Aver-

rhoa carambola) in uremic patients. Am. J. Kidney Dis. 35, 189–193.

Chebib, M., Duke, R.K., Duke, C.C., Connor, M., Mewett, K.N., Johnston,

G.A.R., 2003. Convulsant actions of calycanthine. Toxicol. Appl.

Pharmacol. 190, 58–64.

Chebib, M., Johnston, G.A., 2000. GABA-activated ligand gated ion

channels: medicinal chemistry and molecular biology. J. Med. Chem.

43, 1427–1447.

Chen, C.L., Fang, H.C., Chou, K.J., Wang, J.S., Chung, H.M., 2001. Acute

oxalate nephropathy after ingestion of star fruit. Am. J. Kidney Dis. 37,

418–422.

Curtis, D.R., Duggan, A.W., Felix, D., Johnston, G.A., 1970. GABA,

bicuculline and central inhibition. Nature 226, 1222–1224.

Engelborghs, S., D’Hooge, R., De Deyn, P.P., 2000. Pathophysiology of

epilepsy. Acta Neurol. Belg. 100, 201–213.

Ferraz, A.C., Pereira, L.F., Ribeirao, R.L., Wolfman, C., Medina, J.H.,

Scorza, F.A., Santos, N.F., Cavalheiro, E.A., Da Cunha, C., 2000.

Ricinine-elicited seizures: a novel chemical model of convulsive sei-

zures. Pharmacol. Biochem. Behav. 65, 577–583.

Fonnum, F., 1984. Glutamate: a neurotransmitter in mammalian brain. J.

Neurochem. 42, 1–11.

Garcia-Cairasco, N., Del Vecchio, F., Oliveira, J.A.C., Dantas, M., Moyses

Neto, M., 2002. Avideo-EEG study of the neurotoxic effects of star fruit

(Averrhoa carambola) juice ingestion in animals with renal failure. In:

ERA-EDTA Congress, Copenhagem.

Gray, E.G., Whittaker, V.P., 1962. The isolation of nerve ending from brain:

an electron-microscopic study of cell fragments derived by homogeni-

zation and centrifugation. J. Anat. 96, 79–87.

Hartree, E.F., 1972. Determination of protein: a modification of the Lowry

method that gives a linear photometric response. Anal. Biochem. 48,

422–427.

R.O.G. Carolino et al. / Neurochemistry International 46 (2005) 523–531 531

Inglefield, J.R., Perry, J.M., Schwartz, R.D., 1995. Postischemic inhibition

of GABA reuptake by tiagabine slows neuronal death in the gerbil

hippocampus. Hippocampus 5, 460–468.

Johnston, G.A.R., 1996. GABAA receptor pharmacology. Pharmacol. Ther.

69, 173–198.

Leite, J.P., Garcia-Cairasco, N., Cavalheiro, E.A., 2002. New insights from

the use of pilocarpine and kainate models. Epilepsy Res. 50, 93–103.

Lo, K.Y., Tong, G.M.W., Wong, P.N., Mak, S.K., Wong, A.K.M., 2001.

Persistent hiccup in a continuous ambulatory peritoneal dialysis patient

following ingestion of star fruit. Hong Kong J. Nephrol. 3, 45–46.

Lowry, O.H., Rosenbrouch, N.J., Farr, A.L., Randall, R.J., 1951. Protein

measurement with the folin phenol reagent. J. Biol. Chem. 193, 265–275.

Martin, L.C., Caramori, J.S.T., Barreti, P., Soares, V.A., 1993. Intractable

hiccups induced by carambola (Averrhoa carambola) ingestion in

patients with end stage renal failure. J. Bras. Nefrol. 15, 92–94.

Moraes, M.F., Del Vecchio, F., Terra, V.C., Garcia-Cairasco, N., 2000. Time

evolution of acoustic ‘information’ processing in the mesencephalon of

Wistar rats. Neurosci. Lett. 284, 13–16.

Moyses-Neto, M., Robl, F., Coutinho-Netto, J., 1998. Intoxication by star

fruit (Averrhoa carambola) in six dialysis patients? Nephrol. Dial.

Transplant. 13, 570–572.

Moyses-Neto, M., Costa, J.A.C., Garcia-Cairasco, N., Coutinho-Netto, J.,

Nakagawa, B., Dantas, M., 2003. Intoxication by star fruit (Averrhoa

carambola) in 32 uraemic patients: treatment and outcome. Nephrol.

Dial. Transplant. 18, 120–125.

Muir, C.K., Lam, C.K., 1980. Depressant action of Averrhoa carambola.

Med. J. Malaysia XXXIV, 279–280.

Paula, C.A., Sousa, M.V., Salgado, M.C.O., Oliveira, E., 1998. Purification

and substrate specificity of an angiotensin converting elastase-2 from

the rat mesenteric arterial bed perfusate. Biochim. Biophys. Acta 1388,

227–238.

Paxinos, G., Watson, C.H., 1998. The Rat Brain in Stereotaxic Coordinates.

Academic Press, New York, p. 256.

Sampaio, S.V., Laure, C.J., Giglio, J.R., 1983. Isolation and characterization

of toxic proteins from the venom of the Brazilian scorpion Tityus

serrulatus. Toxicon 21, 265–277.

Sanz, P., Reig, R., 1992. Clinical and pathological findings in fatal plant

oxalosis. A review. Am. J. Foren. Med. Pathol. 13, 342–345.

Satake, K., Okuyama, T., Ohashi, M., Shinoda, T., 1960. The spectro-

photometric determination of amine, amino acid and peptide with 2,4,6-

trinitrobenzene 1-sulfonic acid. J. Biochem. 47, 654–660.

Tse, K.C., Yip, P.S., Lam, M.F., Choy, B.Y., Li, F.K., Lui, S.L., Lo, W.K.,

Chan, T.M., Lai, K.N., 2003. Star fruit intoxication in uraemic

patients: case series and review of the literature. Int. Med. J. 33,

314–316.

Uwai, K., Ohashi, K., Takaya, Y., Oshima, Y., Furukawa, K.I., Yamagata,

K., Omura, T., Okuyama, S., 2001. Virol A, a toxic trans-polyacetylenic

alcohol of Cicuta virosa, selectively inhibits the GABA-induced Cl�

current in acutely dissociated rat hippocampal CA1 neurons. Brain Res.

889, 174–180.

Wang, S.J., Sihra, T.S., 2003. Opposing facilitatory and inhibitory modula-

tion of glutamate release elicited by cAMP production in cerebrocortical

nerve terminals (synaptosomes). Neuropharmacology 44, 686–697.

Wu, C.C., Denq, J.C., Tsai, W.S., Lin, S.H., 2002. Star fruit-induced

neurotoxicity in two patients with chronic renal failure. J. Med. Sci.

22, 75–78.

Yap, H.J., Chen, Y.C., Fang, J.T., Huang, C.C., 2002. Star fruit: a neglected

but serious fruit intoxication in chronic renal failure. Dial. Transplant.

31, 564–597.

Zukin, S.R., Young, A.B., Snyder, S.H., 1974. Gamma-aminobutyric acid

binding to receptor sites in the rat central nervous system. Proc. Natl.

Acad. Sci. U.S.A. 71, 4802–4807.