Anti-tick effect and cholinesterase inhibition caused by Prosopis juliflora alkaloids: in vitro and in silico studies

Braz J Vet Parasitol 2020; 29(2): e019819 | https://doi.org/10.1590/S1984-29612020036 This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Anti-tick effect and cholinesterase inhibition caused by Prosopis juliflora alkaloids: in vitro and in silico studies


Introduction
Parasitic diseases impair livestock health and can cause high mortality rate in cattle herds if parasitism rates are high. Rhipicephalus microplus is a constant threat to cattle because of the direct and indirect damage that it causes to animal health, thereby leading to diminished reproductive efficiency and milk and meat production (Hue et al., 2015). It is a hematophagous parasite that transmits diseases by acting as a vector for pathogens such as Babesia bovis and Anaplasma marginale (Adenubi et al., 2016). Use of synthetic acaricides is the most common strategy for tick control; however, increasing resistance to these acaricides has encouraged a search for new bioactive molecules from plants, as treatment alternatives (Rosado-Aguilar et al., 2017).
Prosopis juliflora, popularly known as "algaroba" and "algarobeira", is a shrub belonging to the Fabaceae family that is native to arid and semi-arid regions. This species was introduced into northeastern Brazil more than 50 years ago and is used as a food source for humans and animals because of its high production of pods and the high palatability and nutritional value of the pods (Pegado et al., 2006). Several types of biological activity have been reported for this plant, such as anthelmintic (Lima et al., 2017), insecticidal (Dhivya et al., 2018) and antibacterial (Odhiambo et al., 2015). These effects have been attributed to the alkaloids that are present in this species. The alkaloids of greatest pharmacological importance are juliprosopine and juliprosine .
The in vitro anticholinesterase activity of juliprosopine upon electric eels (Electrophorus electricus) acetylcholinesterase enzyme (AChE) was described by Choudhary et al. (2005). AChE is an essential enzyme in the nervous system of ticks and is the main target for organophosphate and carbamate pesticides (Zhou & Xia, 2009).
Because of the scarcity of information on the acaricide activity of P. juliflora, the aim of the present study was to evaluate the in vitro activity of the methanolic extract (ME) and alkaloid-rich fraction (AF) of this plant on the reproductive parameters of Rhipicephalus microplus, and to correlate this effect with inhibition of the AChE. Furthermore, in silico assays were performed to characterize the 3D structure of this tick's AChE1 and to predict the possible interaction mode of the major alkaloids of the AF at the active site of the AChE1.

Plant material
Pods from P. juliflora were collected in the municipality of Senhor do Bonfim, state of Bahia, Brazil, in September 2013. A voucher specimen was deposited at the Botanic Laboratory of Dr. Antônio Nonato Marques, Empresa Baiana de Desenvolvimento Agrícola S.A. (EBDA), Salvador, Bahia (number 5465).

Obtainment of the methanol extract and alkaloid-rich fraction
Air-dried and powdered pods (22.6 kg) from P. juliflora were macerated with 9 liters (L) of n-hexane for two days. After filtration, the solvent was evaporated under reduced pressure and the remaining plant material was subsequently extracted with methanol (MeOH) (9 L), using the same procedure. The yield of methanol extract (ME) was 0.19%. The ME was concentrated and dried under reduced pressure using a vacuum rotary evaporator Switzerland), which was used to furnish the alkaloid-rich fraction (AF). This was obtained by means of acidic/basic modified extraction as described by Ott-Longoni et al. (1980). The yield of AF was 0.043%.

LC-MS/MS analysis on the alkaloid-rich fraction
Liquid chromatography-tandem mass spectrometry (LC-MS/MS) was performed, coupled with an Esquire 3000 plus ion-trap mass spectrometer (Bruker Daltonics, Bremen, Germany) that was equipped with a CBM-20A controller, LC-20AD pump, SIL-20AC auto-sampler and SPD-20A detector, with an HPLC system connected to this. Separations were performed in a Phenomenex Luna C-18 column (250 × 4.6 mm), with 5-µm particle size. The elution gradient was run using a binary solvent system consisting of water containing 0.05% phosphoric acid (solvent A) and methanol (solvent B) at a constant flow rate of 1.0 mL/min. The gradient was as follows: 0 min, 0% B; 25 min, 100% B; 35 min, 100% B; 36 min, 20% B; 45 min, 20% B. The injection volume was 20 μL. Data were acquired using a UV detector at 280 nm and 360 nm. The capillary temperature was maintained at 300 °C and the electrospray capillary voltage at 4.5 kV. The LC/MS was performed in positive ionization mode and with the full scan (m/z 100-1,500).

In vitro studies
Rhipicephalus microplus samples R. microplus from POA strain (Porto Alegre strain) that is sensitive to acaricides currently on the market, and which were free of pathogens such as Babesia spp. and Anaplasma spp., was obtained from the Laboratório de Imunologia Aplicada à Sanidade Animal, Centro de Biotecnologia, Universidade Federal do Rio Grande do Sul (UFRGS). This colony has been maintained experimentally through feeding on Hereford calves (Bos taurus taurus) since 1976. Calves were purchased from an area that is naturally tick-free (Santa Vitória do Palmar, RS, Brazil; 33°32'2" S, 53°20'59" W) and were maintained in an isolated stall at the Federal Rural University of Rio Grande do Sul (UFRGS), Brazil, to avoid infestation by other ticks. These calves were handled in accordance with the institutional guidelines, as approved by the local ethics committee for animal use (Ethics Committee for Animal Experimentation of the Universidade Federal do Rio Grande do Sul).
Engorged females were carefully collected from the cattle, placed in Petri dishes with enough aeration and transported to the Laboratory of Toxicology, School of Veterinary Medicine and Animal Sciences, Federal University of Bahia, Salvador, BA, Brazil. These parasites were selected according to their integrity and motility. Following this, they were washed in distilled water, dried on absorbent paper and separated into two groups. In the first group, engorged females were used to evaluate the acaricidal activity up to 24 h after collection. In the second group, the females were incubated for two weeks (26 ± 2 °C and relative humidity > 80%), to yield larvae that were used to assess the anticholinesterase activity.

Adult Immersion Test (AIT)
The adult immersion test was conducted as described by Drummond et al. (1973). The ticks were weighed and separated into homogeneous groups with 10 females each, according to their weight (1.5 to 2 g). The engorged females were immersed for 5 min in 5 mL of the following treatments: ME (91, 127.6, 178.6, 250 and 350 mg/mL) and AF (16.9, 33.2, 65.1, 127.6 and 250 mg/mL) from P. juliflora; negative control (ethanol 70%); and positive control (Diazinon, 2.5 mg/mL), diluted as recommended by the manufacturer (Agener União Saúde Animal  ). Following the immersion, these ticks were dried on absorbent paper, placed in Petri dishes and incubated for 15 days (26 ± 2 °C and relative humidity > 80%), in order to evaluate oviposition. After this period, the eggs were weighed, transferred to glass tubes and incubated under the same conditions as described above. After 21 days, the larval hatching percentage was estimated visually using a stereomicroscope. Four repetitions were used for each of the treatments.

Larval Immersion Test (LIT)
The LIT was used to evaluate the effect of the AF of P. juliflora, most active fraction in AIT, against R. microplus larvae (Silva et al., 2009), at concentrations of 4.2 to 65.1 mg/mL. Approximately 100 larvae of 14 to 21 days of age were used, obtained through oviposition from untreated engorged female ticks and collected from naturally infested cattle (Catu, state of Bahia, Brazil). The larvae were put into 5-mL syringes, which were cut next to the needle. The syringe was closed using a fine-weft fabric fixed with an orthodontic rubber band, and the larvae were immersed in the treatments for five minutes and maintained at 27 ± 1 °C and 80 ± 5% relative humidity. An orifice of approximately 0.1 mm in diameter was made in the middle of the syringe. This procedure was repeated for each concentration of the AF, and for the positive control (Fipronil, 10mg/mL -diluted as recommended by the manufacturer) and negative control (70% ethanol). Larval mortality was recorded after 24 h. Only larvae that had the ability to walk were considered alive. All treatments were set up as three replicates for each concentration tested. Living and dead larvae were counted, and the percentage mortality was calculated as: ( ) % / 100 number of dead larvae total number of larvae x = mortality (1) In vitro anticholinesterase activity of larvae from R. microplus The anticholinesterase activity of the ME and AF was determined spectrophotometrically, in accordance with the methodology described by Ellman et al. (1961), as modified by Wright & Ahrens (1988). Samples of R. microplus larvae (100 mg) were macerated in deionized water (3 mL) and were centrifuged at 1,000 × g for 5 min. Then, the supernatants were used as the enzyme source. In a microtube, 50 µL of ME (final concentration: 0.26, 0.64, 1.6, 4 and 10 mg/mL) and AF (final concentration: 0.001, 0.005, 0.025, 0.125 and 0.625 mg/mL) of P. juliflora, the negative control (70% ethanol) and positive control (Eserine/Sigma-aldrich  ; final concentration: 0.014 mg/mL) were preincubated with AChE solution (200 µL) during 20 min. at 4 °C. For determination of AChE activity, phosphate buffer solution with pH 8.0 (0.1 M, 2,8 mL) and enzyme solution (200 μL) were homogenized and incubated at 35 °C for 2 min. Then, 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB 10 mM, 200 μL) and acetylthiocholine iodide (7.5 mM, 25 μL) were added to initiate the reaction. The absorbance was measured on the spectrophotometer (UV-Vis SP-220 Spectrophotometer, Biospectro, Japan) at 412 nm (0 and 6 min). Each sample was assayed with nine replicates. AChE percentage inhibition was calculated by using the equation: In silico studies Comparative modeling of AChE1 from R. microplus (RmAChE1) The amino acid sequence of the cholinergic domain (residues 51 to 528) of RmAChE1 (access code: A0A0F2P2D6), which is available on the UniProt server (UniProt, 2019), was aligned with the primary sequences from the crystallographic structure of the human AChE (HsAChE; PDB ID: 4EY7) and the Torpedo californica AChE (PDB ID: 2WFZ), using the Clustal Omega server (EMBL-EBI, 2019a). Both the human and the Torpedo californica AChE are available from the Protein Data Bank (RCSB PDB, 2019). From the alignment, 100 models of the RmAChE1 were built using the Modeler 9.18 software (Šali & Blundell, 1993). These were refined through four simulated annealing cycles (flags: autosched.slow; 500 interactions; refine.slow) and were classified according to the scoring function QMEAN6 (SIB, 2019) and the SSAgree values (Benkert et al., 2009). Next, the compatibility between the atomic model (3D) of the 10 best models, according to QMEAN score, and its own amino acid sequence (1D) was calculated using the Verify3D server (http://services.mbi.ucla.edu/Verify_3D/, Lüthy et al., 1992). An overall folding quality scale was used for each model. The stereochemistry quality was estimated using the Ramachandran plot (Ramachandran et al., 1963), which was calculated on the PDBsum server (EMBL-EBI, 2019b; Laskowski et al., 1997). The model with the best folding and stereochemistry parameters was used for the molecular docking assays.

Preparation of the RmAChE1 receptor structures and alkaloids
The 3D structure model of the RmAChE1 was prepared for docking assays using the Biopolymer module, which is available on the Sybyl  -X 2.1.1 platform (TRIPOS Associates, 2013). Firstly, hydrogen atoms were added and optimized to prioritize H-bond interactions. Histidine, glutamate and aspartate residues were manually checked for orientation, protonation and tautomeric states. The protonation state of residues was determined using Propka 3.1 Rostkowski et al., 2011), with pH = 8.0. Following this, AMBER Force Field 99 charges (Wang et al., 2000) were assigned to the protein. The same protocol was performed for the human isoform (HsAChE; 4EY7).
Major alkaloid structures from the AF (Figures 1 and 2), as well as donepezil, were sketched on the Marvin  Sketch 16.8.29 software (Marvin Sketch 16.8.29, 2017, ChemAxon, 2019 and were later converted into the 3D format using the Concord module (standard parameters), which is available on Sybyl  -X 2.1.1 (TRIPOS Associates, 2013). Afterwards, atomic Gasteiger-Huckel charges (Gasteiger & Marsili, 1980;Hou et al., 2013) were added and the structures were energy-minimized using the conjugate gradient method until the convergence criteria (0.001 kcal/mol or maximum number of interactions = 50,000) were reached at the water dielectric constant (ε = 80.4), in the Tripos force-field (Clark et al., 1989).

Molecular docking studies on piperidine alkaloids
Piperidine alkaloid (ligand) docking was carried out in the GOLD suite v. 5.5 (CCDC, Cambridge, UK). The search space was defined unrestrictedly as a radius of 18 Å, centered on the oxygen (gamma) of the catalytic residue Serine 222. The protein residues remained rigid throughout the calculation, while the ligands were kept flexible (additional N-pyramidal and ring-corner movements were allowed).
Docking simulations were performed using the Lamarckian genetic algorithm (Morris et al., 1998), which is a hybrid of a genetic algorithm and a local search algorithm that is available through GOLD 5.5 (Greenblatt et al., 2007;Jones et al., 1995). The docking parameters were set to default, except for the following: trials of 100 LGA runs for each ligand, initial population size of 250 individuals, random starting position and conformation and 2.5 x 10 6 generations at a selective pressure of 1:1, undergoing mutation (95%), crossover (95%) and migration (10%) in 10 islands and 5 niches. Each docking simulation produced 100 different docked conformations that were later ranked using the Piecewise Linear Potential function (ChemPLP), which was implemented in the GOLD suite (Greenblatt et al., 2007;Jones et al., 1995). Next, the complexes of RmAChE-alkaloids were visually analyzed using the Pymol 1.3 software (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC). This procedure was also performed for the donepezil structure, in relation to HsAChE.

Statistical analysis
The results obtained from the biological assays were expressed as the mean ± standard deviation (SD). The normality of data was assessed using the Shapiro-Wilk test. The acaricidal evaluation and anticholinesterase activity also underwent univariate analysis of variance (ANOVA) followed by post-hoc Tukey's honest significant test (Tukey´s HSD test, 5%), using the GraphPad Prism statistical software (version 5.0).

Chemical analysis
The LC-MS/MS chromatogram of the AF in the positive mode is shown in Figure 1. Figure Table 1 showed the effect of the ME and AF of P. juliflora on the reproductive parameters of adult females of R. microplus. The ME and AF induced significant reductions (p < 0.05) in the oviposition and hatching rates of the ticks, in a concentration-dependent manner. The efficacy of the AF (65.1 to 250 mg/mL) and ME (350 mg/mL) did not differ from that of the positive control (Diazinon; 2.5 mg/mL) (p > 0.05). The EC 50 values were 31.6 and 121 mg/mL for the AF and ME, respectively. Table 2 presents the effect of the AF against the R. microplus larvae. The AF had larvicidal activity against R. microplus, with LC 50 value of 13.8 mg/mL. The AF, at the highest concentration tested (65.1 mg/mL), showed a percentage of larvae mortality of 90%.

In vitro anticholinesterase activity against larvae of R. microplus
At the highest concentration (0.625 mg/mL), the AF inhibited AChE activity in 95% and did not differ from the positive control (eserine, 0.014 mg/mL) (p > 0.05). The IC 50 of the AF was 0.041 mg/mL. For the ME (10 mg/mL), the maximum percentage enzyme inhibition was 52% (Table 3). Table 1. Percentage oviposition, hatching and efficacy (mean ± standard deviation) among adult females of Rhipicephalus microplus, exposed to the immersion test with the methanolic extract (ME) and alkaloid-rich fraction (AF) of Prosopis juliflora. Different letters in columns indicate statistically significant difference (one-way ANOVA followed by the Tukey's HSD test, p < 0.05).

Treatments
Concentration ( Table 3. Percentage inhibition (mean ± standard deviation) of in vitro AChE activity of larvae of R. microplus exposed to the methanolic extract (ME) and alkaloid-rich fraction (AF) of P. juliflora. Positive control: eserine (0.014 mg/mL); negative control: ethanol 70%. Different letters in columns indicate statistically significant difference (one-way ANOVA followed by the Tukey's HSD test, p < 0.05).  (Šali & Blundell, 1993), from high-resolution crystallography templates: human AChE (HsAChE, PDB ID: 4EY7) and Torpedo californica (TcAChE, PDB ID: 2WFZ). Alignment between the template primary sequence and the target (RmAChE1) showed moderate sequential identity (> 42%) ( Figure A.1 -Supplementary Material 1 ), which enabled proper use of the Modeler software (Webb & Sali, 2016;Šali & Blundell, 1993). After generation of RmAChE1 models, the quality evaluation was performed. The RmAChE1 best-ranked model showed that 98.8% of the residues were in the "allowed region" ( Figure 3A) and that 98.7% of the residues presented correct folding ( Figure 3B). Hence, it showed stereochemical characteristics, ( Figure B.1 -Supplementary Material) and folding that were compatible with the experimental model proteins (PDB ID templates: 4EY7 and 2WFZ). This therefore makes it possible to use this model in future docking studies.

Molecular docking assays on piperidine alkaloids
Docking studies usually require previous steps to evaluate the parameters that are to be used in the search (Cole et al., 2005;Jain, 2008). Briefly, a ligand co-crystallized with a protein is redocked with its receptor. Thus, the deviation values (root mean square deviation, RMSD) of the best ligand posed in relation to co-crystallized coordinates ought to be less than 2 Å (depending on the ligand size). Based on this hypothesis, donepezil was redocked to the receptor HsAChE (PDB ID: 4EY7) ( Figure C.1 -Supplementary Material) and the best-ranked pose showed an interaction profile similar to that observed for the co-crystallized structure (RMSD = 1.97 Å). The parameters used for the docking are enough for generation of reliable binding positions with HsAChE and these parameters were applied for RmAChE1 docking studies. Following this, docking of three alkaloids (juliprosinine, juliprosopine and prosopine) was performed at RmAChE1 (Figure 4).

Discussion
In the present study, the ME and AF of P. juliflora were shown to have acaricidal activity against R. microplus. The AF was four times more effective on adult females of R. microplus than was the ME, thus suggesting that the alkaloids present in this plant are the compounds responsible for this activity. In addition, the AF was more active on larvae (LC 50 =13.8 mg/mL) than on adult females (EC 50 =31.6 mg/mL) of R. microplus. These differences may be related to the greater mass of individual engorged adults and thinner cuticle of R. microplus larvae, which would make these stages more sensitive than the adult females (Cruz et al., 2016;Conceição et al., 2017). The LC-MS/MS analyses on the AF led to identification of three alkaloids (prosopine, juliprosinine and juliprosopine). These compounds were characterized through comparisons with data in the literature (Ott-Longoni et al., 1980;Ahmad et al., 1989;Singh & Swapnil, 2011;Singh & Verma, 2012;Santos et al., 2013). Other types of antiparasitic activity exhibited by this plant, such as insecticidal (Yadav et al., 2015;Dhivya et al., 2018) and anthelmintic (Lima et al., 2017), have also been described.
These findings are the first scientific reports on the anti-tick activity of the alkaloids of P. juliflora. The activity of alkaloid and non-alkaloid fractions of Leucas indica against Rhipicephalus (Boophilus) annulatus was studied previously, and only the alkaloid fraction (50 mg/mL) induced adult tick mortality (66.67%) and inhibition of fecundity (55.16%) (Divya et al., 2014).
From our evaluation of the anticholinesterase effect, the inhibition of AChE activity produced by the ME and AF of P. juliflora makes it possible to include them in the group of potent inhibitors (> 50% inhibition) of this enzyme, according to the classification of Vinutha et al. (2007). The AF was more active at a lower concentration (IC 50 = 0.041 mg/mL) than the ME (10 mg/mL; 52%), thus indicating that alkaloids are the bioactive compounds of P. juliflora responsible for the anticholinesterase activity. Previous studies reported that juliprosopine, isolated from P. juliflora, was active in vitro against the AChE of electric eels (Electrophorus electricus) (Choudhary et al., 2005).
Our work indicates that the alkaloids of P. juliflora have the same macromolecular target as do organophosphate and carbamate pesticides. According to Tan et al. (2011), inhibition of the AChE leads to increased levels of the neurotransmitter acetylcholine and to paralysis and death of the tick.
With the aim of understanding the interactions of these alkaloids (presented in the AF) with the AChE of R. microplus (RmAChE), in silico studies were performed on RmAChE1. The higher affinity with the substrate (acetylcholine; ACh) and higher conversion rates of RmAChE2 and RmAChE3 that were previously observed (Temeyer et al., 2010) suggest that RmAChE1 is very important for tick survival. However, the lack of data on the crystallographic structure of RmAChE limits both the designing of new acaricidal compounds and knowledge of the mechanism of action of these molecules (Williams et al., 2005;Lionta et al., 2014;Ferreira et al., 2018). Thus, comparative modeling (homology modeling) plays a significant role in reducing these limitations because it enables investigation, at the atomic level, of the molecular interactions of anticholinesterase compounds through molecular docking (Schmidt et al., 2014).
Since these in vitro assays with these compounds do not allow assessment of the molecular inhibition mechanism against RmAChE, docking studies were conducted. These were based on previous results relating to the non-competitive inhibition mechanism of juliprosopine against the AChE of E. electricus (Choudhary et al., 2005). Docking assays for other two alkaloids were conducted in relation to RmAChE1 because high Tanimoto coefficients (Bajusz et al., 2015;Rogers & Tanimoto, 1960) were achieved (juliprosinine T.C. > 95% and prosopine T.C. = 45%). This coefficient describes the percentage of structural similarity between two distinct compounds (Bajusz et al., 2015): the more similar their structures are, the higher the Tanimoto coefficient is.
The docking showed that the alkaloids seem to interact preferentially with the residues of the catalytic site, anionic subsite and PAS of RmAChE1 (Figure 4). Although the binding mode was distinct from that proposed before (Choudhary et al., 2005), the patterns of interactions were very similar (Figure 4). Our docking results corroborated what had been described for other known ligands of AChEs, such as tanshinone (Cheung et al., 2013), galantamine (Greenblatt et al., 2007), huperzine A and donepezil (Cheung et al., 2012). For instance, the residue tryptophan 286 (present in PAS of HsAChE) has an important π-π stacking interaction with donepezil (PDB ID: 4EY7) and tanshinone (PDB ID: 4M0E). However, at RmAChE1, this residue was replaced by threonine (Thr 301), thus leading to the loss of this interaction (Figure 4 and A.1), which reduced the affinity of these compounds against the enzyme (Swale et al., 2013). The docking results suggest that this lack of interaction was compensated by the π-cation interaction between the 2,3-dihydro-1H-indolizine or hexahydroindolizine from alkaloids and tyrosine 144 from RmAChE1 (Figure 4). π-cation and π-π interactions with tryptophan 86 (anionic site of the HsAChE) and hydrogen bonds with serine 203 also seem to be important for the mechanisms for inhibition caused by huperzine A (PDB ID: 4EY5), galantamine (PDB ID: 1DX6) and donepezil (PDB ID: 4EY7). These interactions were similar to our results between the 2-methyl-3-piperidinol and 2-(hydroxymethyl)-3-piperidinol groups and the tryptophan 103 and serine 222 residues of RmAChE1 (Figure 4).

Conclusion
The alkaloids of P. juliflora presented an in vitro acaricidal effect on the larvae and engorged females of R. microplus, and they inhibited acetylcholinesterase. An in silico assay on the main alkaloids obtained from the alkaloid-rich fraction (juliprosopine, juliprosinine and prosopine) suggested that these compounds preferentially interacted at the catalytic and PAS sites of RmAChE1. These interaction profiles are similar to those described for several known AChE inhibitors. To achieve a better description of this mechanism of action, additional in vitro studies on recombinant RmAChE1 and in silico molecular dynamic simulations are required.