M1 and M3 muscarinic receptors may play a role in the neurotoxicity of anhydroecgonine methyl ester, a cocaine pyrolysis product

The smoke of crack cocaine contains cocaine and its pyrolysis product, anhydroecgonine methyl ester (AEME). AEME possesses greater neurotoxic potential than cocaine and an additive effect when they are combined. Since atropine prevented AEME-induced neurotoxicity, it has been suggested that its toxic effects may involve the muscarinic cholinergic receptors (mAChRs). Our aim is to understand the interaction between AEME and mAChRs and how it can lead to neuronal death. Using a rat primary hippocampal cell culture, AEME was shown to cause a concentration-dependent increase on both total [3H]inositol phosphate and intracellular calcium, and to induce DNA fragmentation after 24 hours of exposure, in line with the activation of caspase-3 previously shown. Additionally, we assessed AEME activity at rat mAChR subtypes 1–5 heterologously expressed in Chinese Hamster Ovary cells. l-[N-methyl-3H]scopolamine competition binding showed a preference of AEME for the M2 subtype; calcium mobilization tests revealed partial agonist effects at M1 and M3 and antagonist activity at the remaining subtypes. The selective M1 and M3 antagonists and the phospholipase C inhibitor, were able to prevent AEME-induced neurotoxicity, suggesting that the toxicity is due to the partial agonist effect at M1 and M3 mAChRs, leading to DNA fragmentation and neuronal death by apoptosis.

South America (1.8 and 1.2% annual prevalence rates, respectively), Oceania (1.5%), and Western and Central Europe (1%) 1 . It blocks the uptake of serotonin, norepinephrine and dopamine in presynaptic nerve terminals, as well as voltage-specific sodium channels, responsible for the local anesthetic effect 2 . Cocaine may cause intense vasoconstriction, endothelial cell dysfunction, oxidative stress and platelet aggregation [3][4][5] , which are responsible for the main systemic adverse effects of its abuse, including stroke, myocardial infarction, arterial dissection, vascular thrombosis, rhabdomyolisis and renal complications 6,7 .
There are two distinct chemical forms of cocaine: hydrochloride ('street' cocaine, 'coke'), a water-soluble powder which can be taken orally, intranasally or intravenously; and 'freebase' or 'crack' cocaine, which is cocaine without the hydrochloride moiety 8,9 . Crack cocaine is the smoked form of cocaine and has greater addictive potential than other routes of cocaine administration 10 . As crack cocaine has a low melting point (96-98 °C), the heating process quickly volatizes the cocaine, which is rapidly absorbed by the lungs and reaches the brain faster than any other route. Along with cocaine, anhydroecgonine methyl ester (AEME), a cocaine pyrolysis product, is also absorbed by the lungs 11 . Up to 80% of cocaine can be converted to AEME, depending on the temperature, the purity of the crack cocaine and the smoking devices 12 .
Little is known about AEME effects. To the best of our knowledge, the data available about this substance thus far includes only studies in the peripheral system, i.e., reductions in blood pressure and heart rate in rabbits with an increase in the respiratory rate 13 ; a negative inotropic effect in vitro possibly mediated by muscarinic cholinergic receptors (mAChRs), since it was reversed by atropine, a nonspecific muscarinic receptor antagonist 14 ; cardiovascular effects, e.g., hypotension and tachycardia in sheep, which are also antagonized by intravenous administration of atropine, again consistent with a muscarinic cholinergic effect 15 . However, the mechanisms of AEME in the central nervous system are poorly investigated and not well understood.
Our group first described the neurotoxicity of AEME and also the involvement of mAChRs in AEME-induced neuronal death in rat primary hippocampal cell cultures 16 . AEME seems to be a neurotoxic agent with greater neurotoxic potential than cocaine, showing an additive effect when combined. Caspase-3 activity in the hippocampal neurons was increased after 6 hours of exposure and seems to be one of the main mechanisms of AEME-induced neurotoxicity. Also, atropine prevented AEME-induced neurotoxicity, reinforcing that mAChRs are involved in AEME's effects. In addition, binding experiments with hippocampi membrane preparations have confirmed the affinity of AEME for muscarinic cholinergic receptors 16 .
The mAChRs belong to the G-protein coupled receptors and include five distinct subtypes, denoted as M 1 -M 5 mAChRs, which are widely distributed throughout the body. The odd-numbered mAChRs subtypes are coupled to G q /G 11 proteins and induce the hydrolysis of phosphoinositide lipids by phospholipases, while M 2 and M 4 mAChRs subtypes, coupled to G i /G 0 proteins, inhibit adenylyl cyclase activity. Thus, the activation of M 1 , M 3 and M 5 mAChRs subtypes produces inositol trisphosphate (IP 3 ) and intracellular calcium, while M 2 and M 4 mAChRs subtypes cause a downstream decrease in cAMP levels 17,18 . It is important to emphasize that all five mAChRs are expressed in the hippocampus 19,20 and they modulate hippocampal function through the inhibition of synaptic activity and/or the increase of neuronal excitability 21 .
To better understand the interaction between AEME and mAChRs, the present study investigated: 1) the effects of AEME on the production of total inositol phosphate and intracellular calcium release in rat primary hippocampal cell cultures, as well as DNA fragmentation as a result of caspase-3 activation, which was previously observed by our group 16 ; 2) the affinity and the action of AEME at mAChRs using Chinese hamster ovary (CHO) cells expressing all five individual rat mAChRs subtypes; 3) the effects of M 1 and M 3 selective antagonists on AEME-induced neurotoxicity, as well as the inhibition of phospholipase C (PLC). The comprehension of the mechanisms underlying AEME-induced neurotoxicity could contribute to explain why crack cocaine smoke is more devastating than other routes of cocaine administration.

Experiments with rat primary hippocampal cell culture. Effect of AEME on total [ 3 H]inositol
phosphates accumulation and on intracellular calcium release. AEME and the cholinergic agonist carbachol (control) (10 −8 to 10 −3 M) caused a concentration-dependent increase of total [ 3 H]inositol phosphate in the primary hippocampal cell culture. Maximum inositol phosphate accumulation was obtained with 10 −5 M (10 μ M) AEME and carbachol (Fig. 1A). The basal level of the total [ 3 H]inositol phosphate was 30626 ± 4250 dpm/10 6 cells.

Experiments with CHO cells. B max and K d determination for each mAChR subtype expressed in CHO
cells. The saturation binding of [ 3 H]NMS for each mAChR subtype expressed in CHO cells was specific and saturable and the specific binding fitted best a one-site model (see Supplementary Figure S1). An analysis of three experiments, each one performed in triplicate yielded dissociation constant (K d ) and binding capacity (B max ), summarized in Table 1.
Radioligand competition binding. [ 3 H]NMS displacement curves for AEME are shown in Fig. 3. The AEME K i values (μ M) were: 25.7 for rat M 1 , 19.5 for rat M 2 , 33.9 for rat M 3 , 24.5 for rat M 4 , and 29.5 for rat M 5 . The one-way ANOVA showed that the affinity for rat M 2 was greater than rat M 4 ; rat M 4 greater than rat M 3 ; and rat M 1 greater than M 3 (p < 0.05). Other than this observation, there was no difference among rat M 4 , rat M 1 and rat M 5 , and no difference between rat M 5 and rat M 3 . The averages of absolute values of controls are: 2274 ± 187, 888 ± 25, 3698 ± 303, 2831 ± 231, and 4059 ± 415 dpm/25 μ g of protein for rat M 1 , rat M 2 , rat M 3 , rat M 4 and rat M 5 , respectively.  AEME functional assays. Calcium mobilization assays (Figs 4 and 5) were performed to determine AEME's mechanism of action. AEME exhibited partial agonist-activity at the M 1 and M 3 mAChRs subtypes as observed by an increase in intracellular calcium fluorescence. The effective concentration that increased the response to 50% of maximum (EC 50 ) was greater than 100 μ M for both rat M 1 and rat M 3 , reaching 38.3% (rat M 1 ) and 27.2% (rat M 3 ) of the maximum acetylcholine response (Fig. 4) Table 2). In light of the ability of AEME to displace NMS, this suggested potential antagonist activity of AEME at rat M 5 . The averages of absolute values of controls are: 20307 ± 149, 18233 ± 86, and 23438 ± 205 arbitrary units/4 × 10 4 cells for rat M 1 , rat M 3 and rat M 5 , respectively; 21937 ± 178 and 20218 ± 165 arbitrary units/6 × 10 4 cells for rat M 2 and rat M 4 , respectively.
As the AEME antagonist effect appeared most robust at the rat M 5 mAChR subtype, Schild analyses were performed to determine antagonist affinity (pA2). Using this technique, we estimated the pA2 and Schild slopes, which were 4.87 ± 0.10 and 0.95 ± 0.07, respectively (Fig. 6). These results suggest that AEME is a competitive orthosteric antagonist at rat M 5 . The average of absolute values of control is 21255 ± 124 arbitrary units/4 × 10 4 cells.  Effect of M 1 and M 3 selective mAChR antagonists and the PLC inhibitor on AEME-induced neurotoxicity. MTT viability assay ( Fig. 7) was performed to determine the involvement of either M 1 or M 3 mAChRs, as well as the PLC inhibitor (U73122) on AEME-induced neurotoxicity. After 24 hours of exposure, 10 −3 M AEME presented a significant decrease in neuronal viability, presenting 67.2% of viable cells (F 9,220 = 20.87; p < 0.001). The incubation of AEME in the presence of the M 1 selective antagonist, pirenzepine, prevented its neurotoxicity (98.4% of viable cells; p < 0.001 compared with AEME), as well as with the M 3 selective antagonist, p-fluorohexahydro-sila-difenidol hydrochloride (p-F-HHSiD) (85.4% of viable cells; p < 0.05 compared with AEME) and iPLC (U73122) (102.6% of viable cells; p < 0.01 compared with AEME).

Discussion
This manuscript comprises the first description of the mechanisms of AEME at mAChRs.  16 , indicating that mAChRs may be a target of AEME. Although the pKi values were very similar, we observed a slight preference for the M 2 mAChR. In the hippocampus of male rats, immunoprecipitation studies indicated a predominance of the M 1 subtype (55%) and low expression of M 2 (17%), M 3 (10%) and M 4 (15%) 22 . Their order of abundance has been ranked as M 1 ≫ M 2 > M 3 = M 4 > M 5 19,20 . In hippocampus of female rats in proestrus, immunoprecipitation studies also confirmed that, although hippocampus expresses all mAChR subtypes, the population of M 1 receptors is predominant. In addition, the amount of M 2 in the hippocampus is higher than the  Considering that the expression of M 1 receptor is higher than the others mAChR subtypes in rat hippocampus 19,20,22 , in the present study we showed that this substance is a partial agonist at M 1 and M 3 mAChRs subtypes through the increase in the intracellular calcium, even in the presence of low concentrations of acetylcholine. Also, 100 μ M AEME slightly shifted the acetylcholine concentration-response curves to the right for the M 2 and M 4 mAChR subtypes, characterizing a weak antagonist effect. The same antagonist effect was also observed for M 5 mAChR; however, the effect of AEME seems to be relatively higher for mAChRs with lower EC 50 for acetylcholine. It is important to note that AEME increased, in a concentration-dependent manner, total [ 3 H]inositol phosphate, indicating that AEME may act as an agonist at the M 1 and/or M 3 mAChR subtypes.
The intracellular cascade involved in the activation of the M 1 mAChRs increases the intracellular calcium 23 , which in turn might be involved in the activation of the caspase signaling, leading to neuronal death. In fact, Shih et al. 24 showed that arecoline, a mAChR agonist, induced neuronal death by apoptosis at concentrations from 50-200 μ M. The generation of reactive oxygen species, the decrease in antioxidant defenses and the activation of caspase-3 were some of the mechanisms studied. Besides the chemical similarity among AEME and arecoline, with similar pK i values in binding studies on cloned M 1 -M 5 mAChRs 25 , it seems that these substances share the same neuronal death pathways, culminating in the activation of caspase-3 16,24 . This executioner enzyme can mediate the catabolic process, characterizing the end-stage of apoptosis 26 . The current manuscript corroborates with this finding, suggesting that one of the mechanisms involved in DNA fragmentation observed 24 hours of exposure to 10 −3 M AEME could be explained by the previous activation of caspase-3. Whether, the DNA fragmentation might be a consequence of this intracellular cascade activation remains to be explored. AEME-induced neurotoxicity could be triggered by an intracellular calcium increase, in a concentration-dependent manner, observed at concentrations starting at 10 −4 M (100 μ M). The endoplasmic reticulum contains calcium release channels which can be activated by IP 3 through IP 3 receptors 27 . Thus, the increase in free cytosolic calcium observed at AEME concentrations starting at 10 −7 M (0.1 μ M) and reaching a maximum at 10 −5 M (10 μ M) may be triggered by the increases in inositol phosphate accumulation, in particular IP 3 . This outcome may be explained by the partial agonist effect at M 1 and M 3 mAChRs as observed in the in vitro CHO cells data for AEME. It is important to emphasize that the AEME-induced neurotoxicity investigated by Garcia et al. 16 occurred after a long-term exposure, i.e., different concentrations of AEME for 12 and 24 hours.
Few studies have evaluated the effects of AEME at mAChRs [13][14][15] . Our group was the first one to correlate the neurotoxic effects of AEME with the activation of mAChRs, as the nonspecific mAChR antagonist atropine was able to prevent AEME-induced neurotoxicity 16 . Although M 4 and M 5 mAChRs subtypes are less expressed in the hippocampus 19,20 and AEME showed an antagonist effect when interacting with these subtypes, we believe that these effects are not likely to explain AEME-mediated neurotoxicity because of the preventive atropine effect 16 . However, we cannot discard their importance. M 2 mAChRs subtypes are found in this brain region in cholinergic synaptic terminals controlling acetylcholine and other neurotransmitter release, e.g., glutamate, which could promote excitotoxicity at high concentrations 28 . Also, some substances with an antagonist effect at the M 4 mAChR subtype, when injected in the dorsal hippocampus of rats, induce retrograde amnesia, disrupting memory consolidation 29 .
Several substances, e.g., the muscarinic toxins (MTs), a group of small proteins isolated from the venom of some snakes, have a high selectivity and affinity for individual mAChRs subtypes with competitive antagonist, allosteric modulator, and potential agonists effects 30 . MT2 toxin, for example, activates M 1 , M 3 and M 5 mAChRs, leading to a significant increase in intracellular calcium 23 . According to Bashkatova et al. 31 , the stimulation of M 1 mAChRs by the injection of M 1 agonist McN-A-343 increased nitric oxide and lipid peroxidation in the striatal tissues, the same effect observed for the psychostimulant amphetamine. Several cells, including neurons, present the constitutive form of nitric oxide synthase, which is a calcium/calmodulin-dependent enzyme rapidly activated in response to intracellular calcium increase, leading to nitric oxide production 32,33 . Depending on the pathophysiological conditions, it could be overproduced, resulting in cellular toxicity and death by oxidative stress and lipid peroxidation 34 . Interestingly, MT7, a potent non-competitive antagonist toxin at the M 1 mAChR subtype, with no antagonist activity at the M 3 or M 5 mAChRs subtypes 23 , was able to prevent amphetamine-induced nitric oxide generation and the lipid peroxidation process. The authors attributed that the activation of M 1 mAChRs might play a critical role in the neurotoxic process induced by amphetamine 31 . Our previous study showing the preventive effect of mAChRs antagonist atropine in the AEME-induced neurotoxicity 16 corroborate with these findings. Moreover, we demonstrated that M 1 selective antagonist (pirenzepine) and the M 3 selective antagonist (p-F-HHSiD), were able to prevent or reduce AEME-induced neurotoxicity. In addition, the PLC inhibitor U73122 was also able to prevent AEME toxicity indicating that these effects are mediated by PLC activation. Taking together, these results indicate that the neurotoxic effects of AEME may involve the activation of both M 1 and M 3 mAChRs.
To the best of our knowledge, this is the first study demonstrating AEME mechanism of action at the mAChRs. Its partial agonist effect at M 1 and M 3 mAChRs may be the cause of AEME neurotoxicity, once the selective antagonists were able to prevent it. The IP 3 accumulation and the increase in free cytosolic calcium, as well as oxidative stress, could result in mitochondrial dysfunction, affecting the electron transfer chain, leading to ATP depletion and neuronal death by caspase activation, which causes DNA fragmentation 35 . This study corroborates with our previous study, reinforcing the idea that AEME is more than a crack cocaine biomarker; it may play a crucial role in several CNS disorders, including cognitive deficits of crack cocaine users 36 , as they are exposed to a mixture of cocaine, AEME and others (e.g. solvents) that could enhance the risk of a neurotoxic effect.

Material and Methods
Anhydroecgonine methyl ester (AEME). Cocaine was gently donated by the Criminal Institute of São Paulo to the Laboratory of Toxicological Analyses (School of Pharmaceutical Sciences, University of São Paulo) for research purposes. Briefly, cocaine was purified (95%) and converted into its salt form, cocaine hydrochloride, by bubbling hydrochloric acid into mixture of purified cocaine dissolved in diethyl ether. Then, AEME was synthesized, using cocaine hydrochloride as start material and purified as previously described by our research group 16 . The AEME product (purity > 98%) was confirmed by proton nuclear magnetic resonance ( 1 H-NMR) and electrospray ionization-mass spectrometry (ESI-MS).
For the muscarinic CHO cells studies, AEME was purchased from Lipomed ® (purity > 98%). Experiments with rat primary hippocampal cell culture. Hippocampal cell culture and immunohistochemical characterization. Hippocampal neurons were dissociated from hippocampi of E18-E19 Wistar rat embryos, as described previously [37][38][39][40] . Pregnant rats were anesthetized with sodium pentobarbitone 55 mg/kg and the fetuses were rapidly decapitated to remove their hippocampi. The tissue was placed into a Petri dish containing 100 U/mL penicillin and 100 μ g/mL streptomycin (Gibco) in a cooled Neurobasal medium (Gibco). Hippocampi were washed with Hank's Balanced Salt Solution (HBSS) and submitted to a mechanical fragmentation using appropriate scissors. Hippocampi fragments were then transferred to a 0.25% trypsin in Earl's Balanced Salt Solution (EBSS) solution pH 7.2-7.4 and were incubated for 10 minutes at 37 °C. After the incubation period, cells were washed with an EBSS solution containing 277.5 U/mL DNAse (Sigma) and 10% fetal bovine serum (FBS) (Gibco) and centrifuged at 300 g (Eppendorf 5804R) for 2 minutes at 20 °C. Neurons were isolated by mechanical dissociation in an EBSS solution (with DNAse and fetal bovine serum) using Pasteur pipettes with different diameters sizes and centrifuged for 5 minutes (300 g). The tissue was then resuspended in Neurobasal medium (Gibco) supplemented with 0.5 mM L-glutamine, 25 μ M L-glutamic acid, 100 U/mL penicillin, 100 μ g/mL streptomycin and 2% B27 supplement (Gibco) to reduce glial cell proliferation 40,41 . The cells were seeded onto 0.01% poli-L-lysine-coated multiwell culture plate and maintained at 37 °C in a humidified atmosphere of 5% CO 2 , for 7-8 days, the time required for maturation of hippocampal neurons, forming a network of functional synaptic contacts 42 . On the second day, half of the old medium was replaced by the same volume of a fresh medium with the same composition. On the seventh day, cells were incubated with AEME in several concentrations for different time periods, depending on the experiment. Hippocampal neurons were plated on poli-L-lysine-coated 24-well culture plate at a density of 2 × 10 5 cells/cm 2 . The culture cells were immunohistochemically characterized with MAP2 (neuronal marker) and GFAP (astrocytic marker) showing a predominance of 92% of neurons and 8% of astrocytes 16,43 .

Effect of AEME on total [ 3 H]inositol phosphates accumulation.
Hippocampal cells (1 × 10 6 cells/well) were allowed to equilibrate for 10 minutes with a nutrient solution of the following composition (mM): NaCl 157.00; KCl 5.60; CaCl 2 0.27; MgCl 2 3.00; NaHCO 3 1.80; glucose 5.50 (pH 7.0-7.2) at 37 °C under constant shaking. Cells were incubated with 5 μ Ci of myo[ 3 H]inositol (specific activity 18.0 Ci/mmol) for 80 minutes, and lithium chloride (10 mM) for additional 30 minutes. Afterwards, the cells were incubated in the absence (basal level) and in the presence of carbachol (positive control) and AEME (10 −8 to 10 −3 M) for 10 minutes. Cells were removed from the plate using a 0.1 M NaOH solution and washed three times with nutrient solution, transferred to 2 mL of methanol:chloroform (2:1 v-v) at 4 °C and homogenized with a Ultra-Turrax T25 homogenizer at 9500 rpm. Chloroform (0.62 mL) and H 2 O (0.93 mL) were added to the homogenate, and the solution was centrifuged for 5 minutes at 1,000× g to separate aqueous and organic phases 44 . Total [ 3 H]inositol phosphate was measured as previously described by Ascoli et al. 45 with slight modification. The aqueous layer was neutralized with 0.1 M HCl and then mixed with 1 mL anion-exchange resin (Dowex AG-X8, formate form, 200-400 mesh), allowed to equilibrate for 30 minutes at room temperature, under agitation, and centrifugated at 1,000× g, for 5 minutes at 4 °C. The resin was then washed sequentially, with 10 mM myo-inositol (2 mL) and 5 mM sodium tetraborate/60 mM sodium formate (2 mL Effect of AEME on intracellular calcium release. Hippocampal cells were removed from the plates and previously incubated with Fluo4-AM dye (Invitrogen) dissolved in Neurobasal medium (Gibco ® ) for one hour 46 . The cells were centrifuged (300 g for 5 minutes at 20 °C), washed with phosphate buffer solution (PBS) without calcium, and plated in microscopy confocal (Zeiss LSM 510, Meta) microplates pretreated with poli-L-lysine. The fluorescence was measured for 90 seconds in the absence and/or in the presence of 10 −5 , 10 −4 and 10 −3 M AEME. After this period, 250 mM KCl was added in order to promote neuronal death and, thus, verify its viability. All measurements were performed with a PBS free-calcium solution. Digitalized images were analyzed through the software Origin 5.
Effects of AEME on DNA fragmentation. DNA fragmentation was assessed using propidium iodide following the method described by Lima et al. 47 . Hippocampal cells were incubated with 10 −4 or 10 −3 M AEME for 12 and 24 hours. Briefly, exposed cells were removed from the culture plate and centrifuged (500 g for 5 minutes) and lysed with a fragmentation buffer (100 mg/mL propidium iodide, 0.1% sodium citrate, and 0.1% Triton-X) for 2 hours at room temperature in the dark. DNA fragmentation was analyzed in a flow cytometry (FACSCalibur, Becton Dickinson, CA) using the FL2-A channel (λ excitation = 585 nm and λ emission = 642 nm) and 10 million cells were evaluated per experiment.

Experiments with CHO cells. CHO cell culture. Chinese hamster ovary (CHO) cells individually
and stably expressing individual rat mAChRs (rM 1 to rM 5 mAChRs) were purchased from the American Type Culture Collection (ATCC) and cultured according to ATCC recommendations. To generate stable rM 2 and rM 4 cell lines for use in calcium mobilization assays, rM 2 and rM 4 -expressing cells were stably transfected with a chimeric G protein (G qi5 ) using Lipofectamine 2000 (Invitrogen). rM 2 and rM 4 cells were grown in Ham's F-12 medium containing 10% heat-inactivated fetal bovine serum (FBS), 2 mM GlutaMAX (Gibco ® ), 20 mM HEPES, 500 μ g/mL G418 sulfate, 200 μ g/mL hygromycin B and an antibiotic-antimycotic solution (Gibco ® ). rM 1 , rM 3 and rM 5 mAChR-CHO cells were grown in the same medium without hygromycin B 48 . CHO cells were cultured in a 150 mm-diameter cell culture-treated dishes (Corning ® ) at 37 °C and 5% CO 2 until complete confluence.
Membrane preparation and equilibrium radioligand binding assays. Membranes were prepared from CHO cells stably expressing individual rat mAChRs (rM 1 to rM 5 ). Briefly, media was removed, cells were washed with PBS and harvested with PBS into a centrifuge bottle. Cells were spun at 2,000 g for 5 minutes at 4 °C. After harvesting, the supernatant was gently discarded and cold membrane preparation buffer [Hanks' balanced salt solution (HBSS; Invitrogen), 4.17 mM sodium bicarbonate, 20 mM HEPES and 0.5 mM EDTA] was added (10 mL/dish). Cells were then homogenized for 5-10 seconds with a rotary homogenizer (Polytron ® ) and centrifuged at 20,000g for 20 minutes at 4 °C. The homogenization/ centrifugation cycle was repeated twice. After the last centrifugation, the supernatant was gently discarded and the pellet was resuspended in cold membrane preparation buffer (up to 2 mL). All binding reactions were performed in a total volume of 1 mL containing 25 μ g of membrane protein. Non-specific binding was determined in the presence of 1 μ M atropine. l-[N-methyl-3 H]scopolamine ([ 3 H]NMS; GE Healthcare) saturation binding was performed to calculate both the index of the receptor density (B max ) and the NMS dissociation constant (NMS K d ) values for each subtype of mAChR. The competition binding reactions were carried out in 96-well deep-well plates with membrane protein in the presence of appropriate concentrations of AEME or vehicle, and 0.1 nM [ 3 H]NMS for 3 hours at room temperature.
The equilibrium binding was terminated by rapid filtration using a 96-well harvester (Brandel ® ). Filters were washed three times with ice-cold harvesting buffer and were dried overnight. Radioactivity was counted using a TopCount NXT (PerkinElmer ® ) and counts were normalized to the maximal specific binding in the presence of vehicle [48][49][50] . Saturation and competition binding data were analyzed using a weighted nonlinear least-squares interactive curve-fitting program GraphPad Prism (GraphPad Prism Software version 5.0). A mathematical model for one or two binding sites was applied. The equilibrium dissociation constant (K D ) and the binding capacity (B max ) were determined 51 . The inhibition constant (K i ) was determined from competition curves using the Cheng and Prusoff equation 52 .
Calcium mobilization assay. CHO cells expressing rM 1 , rM 3 and rM 5 were plated at 4 × 10 4 cells per well, whereas rM 2 and rM 4 were plated at 6 × 10 4 cells per well, in standard growth media (as described previously under "cell culture") in 96-well plates 24 hours before assay and were incubated overnight at 37 °C in 5% CO 2 . On the day of the assay, media was removed, cells were washed with calcium assay buffer [Hanks' Balanced Salt Solution (HBSS; Invitrogen), 20 mM HEPES, 4.17 mM sodium bicarbonate, 2.5 mM probenecid (Sigma), pH 7.4] and maintained in calcium assay buffer containing 2.3 μ M Fluo4-AM dye (Invitrogen). Cells were incubated for 45 minutes (37 °C, 5% CO 2 ) for dye loading. Fluo4-AM dye was removed, cells were washed and replaced with 40 μ L of calcium assay buffer. AEME concentration-response curves were determined in single-add experiments. For calcium fluorescence measurements of AEME potency ("double-add" calcium assay), 100 μ M AEME was added 20 seconds after the beginning of data collection and increasing concentrations of acetylcholine were added 100 seconds later via Flexstation II (Molecular Devices). Fluorescence measurement continued for a total of 200 seconds of acquisition time using an excitation wavelength of 488 nm and an emission wavelength Scientific RepoRts | 5:17555 | DOI: 10.1038/srep17555 of 525 nm. For Schild analyses, fixed concentrations of AEME (10 −6 to 10 −3.5 M) were added before the acetylcholine concentration-response curve in the same way described previously for the "double-add" calcium assay. All compounds were dissolved in dimethyl sulfoxide (DMSO) and then in calcium assay buffer to obtain a final DMSO concentration of 0.3%. All of the peaks of the calcium response were normalized to baseline and then as a percentage of the maximum acetylcholine response. These values were fit using GraphPad Prism version 5.0 to determine EC 50 values and Schild slope and pA2 in the Schild analyses 48,53,54 . Effect of M 1 and M 3 selective mAChR antagonists and the PLC inhibitor on AEME-induced neurotoxicity. Neuronal viability was evaluated in the hippocampal cell culture using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma) reduction assay, with some modification 16 . Briefly, after 24 hours of incubation with 10 −3 M AEME in the absence and in the presence of 10 nM pirenzepine hydrochloride (M 1 selective antagonist; Sigma), 10 nM p-F-HHSiD (M 3 selective antagonist; Sigma) and the specific inhibitor of different isoforms of PLC U73122 (10 nM) 55 (Sigma), all the medium was removed, and 100 μ L of MTT solution, containing 5 mg/mL MTT in PBS and neurobasal medium without phenol red (1:9, v/v) were added. After 3 hours of incubation with MTT at 37 °C in a humidified atmosphere of 5% CO 2 , the MTT solution was removed and 200 μ L of dimethyl sulfoxide was added to each well. After 30 minutes of shaking, the absorbance was measured at 570 nm in a multiwell plate reader (BioTek Synergy H1 Hybrid Reader). Potassium chloride (250 mM) was used as a positive control of neuronal death. Muscarinic antagonists, as well as the iPLC, were added 30 minutes prior to incubation with AEME. The antagonist concentration used was near the pK i 25,56 . A control containing 0.002% DMSO was used, since the iPLC stock solution was previously dissolved in this solvent. The assay was performed in quadruplicate, and the results were expressed as a percentage of the control value 16 . Statistical analyses. Data were analyzed by one-way analysis of variance (ANOVA) followed by Newman-Keuls multiple comparison post-hoc test. Acetylcholine concentration-response curve in the presence of 100 μ M AEME was analyzed by Student's t-test. P < 0.05 was considered statistically significant. All data were plotted and analyzed by GraphPad Prism software version 5.0. Data were reported as mean ± SEM.