Electrochemical Characterization of Central Action Tricyclic Drugs by Voltammetric Techniques and Density Functional Theory Calculations

pharmaceuticals

Article Electrochemical Characterization of Central Action Tricyclic Drugs by Voltammetric Techniques and Density Functional Theory Calculations

Edson Silvio Batista Rodrigues 1, Isaac Yves Lopes de Macêdo 1, Larissa Lesley da Silva Lima 1, Douglas Vieira Thomaz 1 , Carlos Eduardo Peixoto da Cunha 1, Mayk Teles de Oliveira 1 , Nara Ballaminut 1, Morgana Fernandes Alecrim 1, Murilo Ferreira de Carvalho 1, Bruna Guimarães Isecke 1, Karla Carneiro de Siqueira Leite 1, Fabio Bahls Machado 1, Freddy Fernandes Guimarães 2, Ricardo Menegatti 1 , Vernon Somerset 3 and Eric de Souza Gil 1,*

1 Faculdade de Farmácia, Universidade Federal de Goiás, Goiânia GO 74690-970, Brazil 2 Bloco IQ-1, Campus Samambaia, Avenida Esperança s/n, Instituto de Química, Universidade Federal de Goiás, Goiânia CEP 74690-900, Brazil 3 Department of Chemistry, Faculty of Applied Sciences, Cape Peninsula University of Technology, Bellville 7535, South Africa * Correspondence: [email protected]; Tel.: +55-62-3209-6502

Received: 31 May 2019; Accepted: 17 July 2019; Published: 1 August 2019

Abstract: This work details the study of the redox behavior of the drugs cyclobenzaprine (CBP), amitriptyline (AMP) and nortriptyline (NOR) through voltammetric methods and computational chemistry. Results obtained in this study show that the amine moiety of each compound is more likely to undergo oxidation at 1a at E 0.69, 0.79, 0.93 V (vs. Ag/AgCl/KClsat) for CBP, AMP and NOR, p1a ≈ respectively. Moreover, CBP presented a second peak, 2a at E 0.98 V (vs. Ag/AgCl/KClsat) at pH p2a ≈ 7.0. Furthermore, the electronic structure calculation results corroborate the electrochemical assays regarding the HOMO energies of the lowest energy conformers of each molecule. The mechanism for each anodic process is proposed according to electroanalytical and computational chemistry ﬁndings, which show evidence that the methods herein employed may be a valuable alternative to study the redox behavior of structurally similar drugs.

Keywords: electrochemistry; oxidation mechanism; computational chemistry; electronic structure; cyclobenzaprine

1. Introduction Cyclobenzaprine (CBP), amitriptyline (AMP) and nortriptyline (NOR) are central-acting tricyclic drugs whose applicability in therapy makes them widely prescribed medicines. Although they share similar chemical structures (Figure1), these drugs display distinct biological activities, with CBP being a muscle relaxant used against acute musculoskeletal pain while AMP and NOR are antidepressants. Regarding CBP, the thermodynamics and kinetics of its redox processes are as yet unknown, so their elucidation is important to better understand the physicochemical features of this drug [1–3].

Pharmaceuticals 2019, 12, 116; doi:10.3390/ph12030116 www.mdpi.com/journal/pharmaceuticals Pharmaceuticals 2019, 12, 116 2 of 10 Pharmaceuticals 2019, 12, x FOR PEER REVIEW 2 of 10

FigureFigure 1. Chemical 1. Chemical structure structure of cyclobenzaprine of cyclobenzaprine (CBP), (CBP), amitriptyline amitriptyline (AMP) (AMP) and and nortriptyline nortriptyline (NOR). (NOR). Concerning the study of drug redox features, electrochemical methods such as voltammetry are highlyConcerning regarded, the since study both of kinetics drug redox and thermodynamicsfeatures, electrochemical are readily methods assayed such using as voltammetry minute solvent are volumes.highly regarded, These toolssince provideboth kinetics information and thermodynamics about the redox are behavior readily assayed of the compounds, using minute as solvent well as correlatedvolumes. These chemical tools phenomena provide information [4,5]. Moreover, about electrochemistrythe redox behavior grants of the information compounds, regarding as well the as oxidativecorrelated degradation chemical phenomena of these compounds [4,5]. Moreover, [6–8]. electrochemistry grants information regarding the oxidativeIn addition degradation to redox of featuresthese compounds acquired by [6–8]. electroanalytical findings, computational chemistry tools suchIn as addition density functional to redox features theory (DFT) acquired calculations by electroanalytical may assist in findings, the inference computational of physical-chemical chemistry properties.tools such as This density approach functional studies theory the electronic (DFT) calculations structure of moleculesmay assist based in the on inference a function of of physical- electron density.chemical The properties. optimization This approach of the spatial studies conformation the electronic of each structure molecule of molecules has been based performed, on a function and the molecularof electron orbitals density. energy The valueoptimization difference of between the spatial each conformation isomer have beenof each calculated molecule [4, 8 has,9]. Such been informationperformed, and is applicable the molecular to the orbitals prediction energy of possible value difference sites that are between susceptible each isomerto redox have reactions been andcalculated these data[4,8,9]. may Such be information correlated to is electrochemical applicable to the findings prediction in order of possible to propose sites that possible are susceptible oxidation pathwaysto redox reactions [4,8–11]. and these data may be correlated to electrochemical findings in order to propose possibleIn view oxidation of the pathways importance [4,8–11]. of the elucidation of their redox features, this work is aimed at the characterizationIn view of the importance of CBP, AMP of andthe elucidation NOR through of their electrochemical redox features, and this computational work is aimed chemistry at the methodscharacterization [8–12]. of Moreover, CBP, AMP the and correlation NOR through between electrochemical both approaches and was computational determined in chemistry order to elucidatemethods [8–12]. and propose Moreover, a possible the correlation electrochemical between oxidation both approaches pathway of thewas studied determined drugs. in order to elucidate and propose a possible electrochemical oxidation pathway of the studied drugs. 2. Results and Discussions

2.1.2. Results Electrochemical and Discussions Assays

2.1. ElectrochemicalIn order to electrochemically Assays characterize CBP, AMP and NOR, a pH study was conducted to investigate the best experimental conditions as well as the dependence of the oxidation process to In order to electrochemically characterize CBP, AMP and NOR, a pH study was conducted to protons. Thereafter, diﬀerential pulse voltammetry (DPV) was further used to evaluate the oxidation investigate the best experimental conditions as well as the dependence of the oxidation process to of each compound at a ﬁxed pH value, i.e., pH 7.0. Results are depicted in Figure2. protons. Thereafter, differential pulse voltammetry (DPV) was further used to evaluate the oxidation of each compound at a fixed pH value, i.e., pH 7.0. Results are depicted in Figure 2.

Pharmaceuticals 2019, 12, x 116 FOR PEER REVIEW 33 of of 10 Pharmaceuticals 2019, 12, x FOR PEER REVIEW 3 of 10

Figure 2. Differential pulse voltammetry results of 0.1 mM CBP (A), AMP (B) and NOR (C) at different pHFigure levels 2. DifferentialDi(pHfferential 3, 5, 7, 9, pulse 11) in voltammetry acetate buffer results solution ofof 0.1 (ABS) mM CBP and (phosphateA), AMP (B buffer) and NOR solution ( C) at(PBS). at different different pH levels (pH 3, 5, 7, 9, 11)11) inin acetateacetate bubufferffer solutionsolution (ABS)(ABS) andand phosphatephosphate bubufferffer solutionsolution (PBS).(PBS). The results indicate that pH 7.0 provides the best signal amplitude, and this condition apparentlyThe results results makes indicate indicateCBP prone that that pH to pH 7.0a second provides 7.0 provides oxidation the best the signalprocess best amplitude, signal (Figure amplitude, 2A). and thisThe conditionpH and study this apparently conditionfor AMP indicatedmakesapparently CBP that pronemakes pH 5.0 to CBP aprovides second prone oxidation theto besta second analytical process oxidation (Figure response process2A). (Figure The (Figure pH 2B), study and2A). for pH The AMP 9.0 pH is indicated the study best forfor that NORAMP pH (Figure5.0indicated provides 2C). that the pH best 5.0 analyticalprovides the response best analytical (Figure2 responseB), and pH (Figure 9.0 is 2B), thebest and forpH NOR9.0 is (Figurethe best2 forC). NOR (FigureConcerning 2C). CBP, CBP, AMP and NOR oxidation, the first first anodic peak, 1a, seemingly involves the equivalenceConcerning of of protons protons CBP, AMPand and electrons, electrons,and NOR hence hence oxidation, the the EEpa pa thevs.vs. pHfirst pH plot anodic plot rendered rendered peak, a1a, slope a seemingly slope of approximately of approximately involves the 59 −1 1 mV.pH59equivalence mV.pH, which− ,of which protons is associated is associatedand electrons, with with Nernstian hence Nernstian the processes Epa processesvs. pH (Figure plot (Figure rendered 2). 2Nonetheless,). Nonetheless,a slope of theapproximately theliterature literature has 59 reportedhasmV.pH reported−1 , thatwhich that proton-dependent is proton-dependent associated with electrochemical Nernstian electrochemical processes processes processes (Figure showcase showcase 2). Nonetheless, similar similar slope slope the valuesliterature values in in pHhas studiesreported [4,13,14]. [4 , that13,14 proton-dependent]. electrochemical processes showcase similar slope values in pH studiesThe [4,13,14]. tricyclic compounds showed showed similar similar profiles profiles for cyclic voltammetry (CV) (CV) at at different different scan rates Theand tricyclic seemed totocompounds havehave anan adsorptiveadsorptive showed similar oxidation oxidation profiles process, process, for evidenced cyclicevidenced voltammetry by by the the linearity linearity (CV) foundat found different in inIpa Ipa vs.scan vs. v 2 2 v(rrates (r= = 0.99)and 0.99) seemed (Figure (Figure 3to A)3A) have [ 4[4].]. an ThisThis adsorptive resultresult isis alsooxidationalso corroboratedcorroborated process, by byevidenced the literature by the [[15–17]15 linearity–17] since found compounds in Ipa vs. presentingv (r2 = 0.99) an (Figure abundance 3A) [4]. of This π-electronsπ-electrons result is tend also tocorroborated adsorb on carbon by the literaturesurfaces, such[15–17] as that since of compounds the carbon pastepresenting electrode an abundance (CPE) used of herein. π-electrons Furthermore, tend to adsorbthis behavior on carbon is quite surfaces, similar such to that as that of of other otherof the tricyclic tricyclic carbon compoundspaste electrode [13,14,16,18,19]. [13 ,(CPE)14,16, 18used,19]. herein. Furthermore, this behavior is quite similar to that of other tricyclic compoundsThe square [13,14,16,18,19]. wave voltammetry (SWV) was also similar for all tricyclic compounds and showed no reversibilityThe square inwave the voltammetry oxidation process process (SWV) (Figure (Figurewas also 3 3B).B). similar The The compositionfor composition all tricyclic of of compounds di differentialfferential and pulse pulse showed (DP) (DP) voltammogramsno reversibility inof of theeach oxidation compound process at at pH (Figure7.0, evidencing 3B). The their composition oxidative of processes, differential is depicted pulse (DP) in Figurevoltammograms 33C.C. of each compound at pH 7.0, evidencing their oxidative processes, is depicted in Figure 3C.

1 Figure 3. Cyclic voltammetry results results at at different different scan rates (12.5, 25, 50, 100, 250, 500 mV.s−−1)) of of 0.1 0.1 mM AMP in pH5 ACS (A); square wave voltammetry of 0.1 mM AMP in pH5 ACS (B); and differential mMFigure AMP 3. Cyclic in pH5 voltammetry ACS (A); squareresults at wave different voltammetry scan rates of (12.5, 0.1 mM25, 50, AMP 100, in250, pH5 500 ACSmV.s (−1B)); of and 0.1 pulse voltammetry of 0.1 mM CBP (—), AMP (- - -) and NOR (- – -) in pH 7 PBS (C). differentialmM AMP pulse in pH5 voltammetry ACS (A); squareof 0.1 mM wave CBP voltammetry (—), AMP (- of- -) 0.1and mM NOR AMP (- – -) in in pH5pH 7 ACSPBS ( (CB).); and differential pulse voltammetry of 0.1 mM CBP (—), AMP (- - -) and NOR (- – -) in pH 7 PBS (C). DPV findings findings allowed the determination of of the potential associated to to the oxidation of each compound. CBP, AMP and NOR showcased the first anodic peak, namely 1a at Ep1a 0.69, 0.79, compound.DPV findings CBP, AMP allowed and NOR the determination showcased the of first the anodic potential peak, associated namely 1a to at the Ep1a oxidation ≈ 0.69,≈ 0.79, of each0.93 compound. CBP, AMP and NOR showcased the first anodic peak, namely 1a at Ep1a ≈ 0.69, 0.79, 0.93

Pharmaceuticals 2019, 12, x FOR PEER REVIEW 4 of 10

Pharmaceuticals 2019,, 12,, x 116 FOR PEER REVIEW 4 4 of of 10 10 V (vs. Ag/AgCl/KClsat), respectively. CBP presented a second peak, 2a at Ep2a ≈ 0.98 V (vs. Ag/AgCl/KClsat) (Figure 3C). V (vs. Ag/AgCl/KClsat), respectively. CBP presented a second peak, 2a at Ep2a ≈ 0.98 V (vs. It can be observed that despite the common dibenzene–cycloheptane core among the tricyclic Ag/AgCl/KClsat)0.93 V (vs. Ag/AgCl (Figure/KClsat), 3C). respectively. CBP presented a second peak, 2a at Ep2a 0.98 V (vs. drugs, they undergo oxidation at different potentials and, thus, their structural differences≈ promote Ag/AgClIt can/KClsat) be observed (Figure that3C). despite the common dibenzene–cycloheptane core among the tricyclic changes in their kinetic behavior. drugs,It they can beundergo observed oxidation that despite at different the common potentials dibenzene–cycloheptane and, thus, their structural core differences among the promote tricyclic The SWV results evidenced that the redox processes associated with each anodic peak are changesdrugs, they in their undergo kinetic oxidation behavior. at different potentials and, thus, their structural differences promote irreversible, hence Ipa/Ipc is higher than one unit, and the CV findings suggested that the process is changesThe in SWV their results kinetic evidenced behavior. that the redox processes associated with each anodic peak are nonetheless adsorption-controlled [14,18,19]. irreversible,The SWV hence results Ipa/Ipc evidenced is higher than that theone redoxunit, and processes the CV associatedfindings suggested with each that anodic the process peak are is The next step involved the use of the voltammetric results and electronic structure calculations, nonethelessirreversible, adsorption-controlled hence Ipa/Ipc is higher [14,18,19]. than one unit, and the CV findings suggested that the process is in order to propose a possible oxidation pathway. nonethelessThe next adsorption-controlled step involved the use [ 14of, 18the,19 voltammetric]. results and electronic structure calculations, Therefore, these results were associated to molecular modeling in order to propose a possible in orderThe to next propose step involved a possible the oxidation use of the pathway. voltammetric results and electronic structure calculations, in oxidation pathway. orderTherefore, to propose these a possible results oxidation were associated pathway. to molecular modeling in order to propose a possible oxidationTherefore, pathway. these results were associated to molecular modeling in order to propose a possible 2.2. Molecular Electronic Structure Calculations oxidation pathway. 2.2. MolecularThe chemical Electronic structures Structure of CBP, Calculations AMP and NOR in their lowest energy conformers are presented in2.2. Figure Molecular 4. Electronic Structure Calculations The chemical structures of CBP, AMP and NOR in their lowest energy conformers are presented in FigureThe chemical4. structures of CBP, AMP and NOR in their lowest energy conformers are presented in Figure4.

Figure 4. Lowest energy conformers of CBP (A), AMP (B) and NOR (C). Figure 4. Lowest energy conformers of CBP (A), AMP (B) and NOR (C). Figure 4. Lowest energy conformers of CBP (A), AMP (B) and NOR (C). HOMO displays the spread of electronic density in specific regions of the molecule, depicting HOMO displays the spread of electronic density in speciﬁc regions of the molecule, depicting the the likely region of the compound’s first oxidation (Figure 5). likelyHOMO region displays of the compound’s the spread ﬁrst of electronic oxidation density (Figure 5in). specific regions of the molecule, depicting the likely region of the compound’s first oxidation (Figure 5).

Figure 5. HOMO distribution of CBP (A), AMP (B) and NOR (C), in the molecules’ respective lowest Figureenergy 5. conformers. HOMO distribution of CBP (A), AMP (B) and NOR (C), in the molecules’ respective lowest energy conformers. Figure 5. HOMO distribution of CBP (A), AMP (B) and NOR (C), in the molecules’ respective lowest Therefore, the order of E in anodic processes also follows this trend, with E correlating with energy conformers. pa p1a HOMO.Therefore, This process the order was of also Epa showcasedin anodic processes in a previous also follows outreach this by trend, our group with[ 4E],p1a as correlating well as in otherwith HOMO.reports inThis the process literature was [20 also]. showcased in a previous outreach by our group [4], as well as in other Therefore, the order of Epa in anodic processes also follows this trend, with Ep1a correlating with reports in the literature [20]. HOMO.In thisThis sense, process the was amine also moietyshowcased of each in a compoundprevious outreach is the regionby our ingroup which [4], the as well HOMO as in orbital other In this sense, the amine moiety of each compound is the region in which the HOMO orbital is reportsis distributed in the literature (Figure5 ).[20]. This suggests that this group is more likely to be the electro-active site distributed (Figure 5). This suggests that this group is more likely to be the electro-active site involved involvedIn this in sense, the ﬁrst the anodicamine process,moiety ofE p1aeach compound0.85 V (vs. is Ag the/AgCl region/KClsat). in which It isthe known HOMO that orbital amine is ≈ distributed (Figure 5). This suggests that this group is more likely to be the electro-active site involved

Pharmaceuticals 2019, 12, x FOR PEER REVIEW 5 of 10 Pharmaceuticals 2019, 12, 116 5 of 10 in the first anodic process, Ep1a ≈ 0.85 V (vs. Ag/AgCl/KClsat). It is known that amine moieties usually havemoieties high usually electro-oxidation have high electro-oxidation potentials (i.e., potentials greater than(i.e., greater 0.65 V), than what 0.65 corroborates V), what corroborates the results the depictedresults depicted herein. herein. The higher The higher potential potential of peak of peak 1a for 1a forNOR NOR may may be be attributed attributed to to its its electronic electronic configurationconﬁguration [ [17,20–27];17,20–27]; nonetheless, nonetheless, the the HOMO HOMO energy energy levels levels also suggested also suggested that this that region this requires region requiresa higher a overpotential higher overpotential to undergo to its undergo redox reaction, its redox whereas reaction, CBP whereas and AMP CBP have and higher AMPHOMO have higher levels HOMOand are levels thus more and are easily thus oxidized more easily (Figure oxidized6). (Figure 6).

FigureFigure 6. 6. OrbitalOrbital energy plotplot ofof CBP,CBP, NOR NOR and and AMP, AMP, depicting depicting the the LUMO LUMO (blue (blue line), line), HOMO HOMO (red (red line) line)and lowerand lower occupied occupied orbitals orbitals (black (black lines). lines). Orbital energies correlate with the oxidation of compounds according to an electrode’s Fermi level, Orbital energies correlate with the oxidation of compounds according to an electrode’s Fermi where higher HOMO values yield anodic processes at lower potentials and vice versa [4]. Furthermore, level, where higher HOMO values yield anodic processes at lower potentials and vice versa [4]. Mulliken partial charges display electron charge density in different regions of a molecule, which may Furthermore, Mulliken partial charges display electron charge density in different regions of a elucidate regions likely to undergo electro-oxidation in a similar manner to that of HOMO orbitals molecule, which may elucidate regions likely to undergo electro-oxidation in a similar manner to that (Table1). of HOMO orbitals (Table 1). The unsaturated C15 (Figure4) in CBP showed the most negative partial charge value (Table1). The unsaturated C15 (Figure 4) in CBP showed the most negative partial charge value (Table 1). Therefore, CBP C15 is prone to electro-oxidation, being a possible electroactive site responsible for CBP Therefore, CBP C15 is prone to electro-oxidation, being a possible electroactive site responsible for second electro-oxidation. CBP second electro-oxidation. In order to evaluate the number of electrons in the anodic process of AMP, CBP and NOR, the In order to evaluate the number of electrons in the anodic process of AMP, CBP and NOR, the Laviron equation [1x, 2x] was calculated as follows: Laviron equation [1x, 2x] was calculated as follows: 22.3RT.3RT E퐸pa == E0 + logvlogv ((11− α))nFnF − wherewhere EEpapa isis the the oxidation oxidation potential, potential, E’ E’ is isthe the equilibrium equilibrium potential, potential, R Ris isthe the universal universal gas gas constant, constant, T isT the is the temperature, temperature, α αis isthe the transference transference coefficient, coefficient, n nis is the the number number of of transferred transferred electrons, electrons, F F is is the the Faraday’sFaraday’s constant constant and and v v is is the the scan scan rate rate [28–30]. [28–30]. ConsideringConsidering α == 0.5 duedue toto the the irreversibility irreversibility of of the the process process [3x], [3x], the the values values of n of calculated n calculated for AMP, for AMP,CBP and CBP NOR and were NOR 2.3097, were 2.5616 2.3097, and 2.5616 2.0788, and respectively. 2.0788, respectively. Thus, the numberThus, the of number electrons of involved electrons in involvedthe oxidation in the of oxidation AMP, CBP of andAMP, NOR CBP could and NOR be inferred could be as inferred 2 and, therefore, as 2 and, theretherefore, was nothere significant was no significantdifference indifference the oxidation in thekinetics oxidation of thesekinetics drugs. of these drugs. TheseThese findings findings were were used used to to propose the electro-oxidation pathway depicted in Figure7 7..

Pharmaceuticals 2019, 12, 116 6 of 10

The proposed electro-oxidation mechanisms contemplate the equivalence of protons and electrons in the ﬁrst oxidation, i.e., peak 1a. The participation of two electrons (n = 2) in the ﬁrst oxidation processes, inferred from the Laviron equation, was taken into account in the proposal of the oxidation mechanism (Figure7). Moreover, the mechanism depicted herein is similar to other reports in the literature concerning the oxidation of aminated tricyclic compounds [25]. Regarding the second oxidation process undergone by CBP (i.e., peak 2a), literature reports have shown that the vinylic bridge at the start of the aliphatic chain C15 (Figure4) is the most probable site for electron loss. Some reports on forced degradation studies have shown products similar to those of our proposed mechanism [17,20,25,27].

Table 1. Mulliken partial charges of CBP, AMP and NOR, represented in percentages of negative and positive charge of each individual atom with hydrogens summed.

CBP AMP NOR Mulliken Mulliken Mulliken Atom Number Atom Number Atom Number Charge Charge Charge C1 0.042401 C1 0.112381 C1 0.119449 − − − C2 0.039541 C2 0.442957 C2 0.447457 C3 0.027008 C3 0.409320 C3 0.398639 − C4 0.010317 C4 0.221850 C4 0.235249 − − C5 0.010578 C5 0.158772 C5 0.143385 − − C6 0.016292 C6 0.316322 C6 0.302634 − − C7 0.027150 C7 0.996592 C7 0.994888 − C8 0.016568 C8 0.127010 C8 0.050367 − − − C9 0.050069 C9 0.080692 C9 0.001748 − C10 0.032538 C10 0.584003 C10 0.490114 − − − C11 0.015741 N11 0.173777 N11 0.121114 C12 0.010368 C12 0.020840 C12 0.005466 − C13 0.021583 C13 0.057774 C13 0.036809 C14 0.054836 C14 0.253322 C14 0.238812 − − − C15 0.066515 C15 0.088678 C15 0.129595 − − − C16 0.112695 C16 0.311043 C16 0.321218 C17 0.019144 C17 0.196833 C17 0.204345 − − − C18 0.142342 C18 0.174628 C18 0.128244 − − N19 0.416923 C19 0.415570 C19 0.424846 − − − C20 0.139806 C20 0.184025 C20 0.143199 C21 0.133751 C21 0.014030 Pharmaceuticals 2019, 12, 116 7 of 10 Pharmaceuticals 2019, 12, x FOR PEER REVIEW 7 of 10

Figure 7. Proposed CBP, AMP and NOR oxidation mechanism.

3. Materials and Methods 3. Materials and Methods 3.1. Reagents, Samples and Solutions 3.1. Reagents, Samples and Solutions All electrolyte salts, solvents and reagents were of analytical grade, and were used without All electrolyte salts, solvents and reagents were of analytical grade, and were used without further purification. Electrolyte solutions were prepared with double-distilled Milli-Q water further purification. Electrolyte solutions were prepared with double-distilled Milli-Q water (conductivity 0.1 µS cm 1; Millipore S. A., Molsheim, France). CBP, AMP and NOR standards − −1 (conductivity ≤ ≤0.1 µS cm ; Millipore S. A., Molsheim, France).1 CBP, AMP and NOR standards (United States Pharmacopea) were used to prepare 1 mmol L− stock solution immediately prior to (United States Pharmacopea) were used to prepare 1 mmol L−1 stock solution immediately prior to the experiments. the experiments. 3.2. Electrochemical Assays 3.2. Electrochemical Assays Voltammetric measurements were performed using a potentiostat/galvanostat PGSTAT® model ® 204 withVoltammetric a FRA32M measurements module (Metrohm were Autolab,performed City, using Country) a potentiostat/galvanostat integrated with NOVA PGSTAT 2.1® software. model ® All204 measurementswith a FRA32M were module performed (Metrohm in Autolab, a 1 mL one-compartmentCity, Country) integrated electrochemical with NOVA cell 2.1 coupled software. to a three-electrodeAll measurements system were consisting performed of ain carbon a 1 mL paste one-compartment electrode (CPE) electrochemical working electrode cell prepared coupled withto a three-electrode system consisting of a carbon paste electrode (CPE) working electrode prepared with 100 mg of graphite powder and 30 µL of mineral oil, a Pt wire counter electrode and a Ag/AgCl/KClsat reference100 mg of electrodegraphite powder (both purchased and 30 µL from of mineral Lab solutions, oil, a Pt Swireão Paulo, counter Brazil). electrode and a Ag/AgCl/KClsat referenceThe experimentalelectrode (both conditions purchased for from diff erentialLab solutions, pulse voltammetrySão Paulo, Brazil). (DPV) were: pulse amplitude The experimental conditions for differential pulse voltammetry1 (DPV) were: pulse amplitude of of 50 mV, pulse width of 0.5 s and scan rate of 10 mV s− . All voltammetric assays were performed 50 mV, pulse1 width of 0.5 s and scan rate of 10 mV s−1. All voltammetric assays were performed in 0.1 in 0.1 mol L− acetate buffer solution (ACS), pH 3.0, 5.0 and phosphate buffer solution (PBS), pH 7.0, −1 9.0mol and L acetate 11.0. Voltammetric buffer solution assays (ACS), were pH performed3.0, 5.0 and inphosphate buffers purged buffer solution with nitrogen (PBS), gaspH 7.0, in order 9.0 and to 11.0. Voltammetric assays were performed in buffers purged with nitrogen gas in order to avoid oxygen interference. The electrochemical area of the carbon paste electrode was 7.77 mm2, as

Pharmaceuticals 2019, 12, 116 8 of 10 avoid oxygen interference. The electrochemical area of the carbon paste electrode was 7.77 mm2, as estimated in a previous work of our group [31]. The CPE was cleaned and the surface was renewed after each assay. DP voltammograms were background-subtracted and baseline-corrected to provide better visualization. All experiments were conducted in triplicate and data were analyzed using Origin Pro 9® software (Northampton, MA, USA).

3.3. Molecular Electronic Structure Calculations The computational chemistry protocol used herein was adapted from previous work conducted by our group [4] and consisted of: semi-empiric (PM6, i.e. Parameterization Method 6) and density functional theory (DFT) calculations, which were used for the electronic structure investigation and optimization of CBP, AMP and NOR. A 6-311G++ basis set was employed in DFT. The B3LYP exchange correlation function was used in DFT calculations. The full geometry optimizations of the molecules were computed considering the two levels of theory applied in this study. The ﬁrst step in the determination of the most stable molecular geometries was performed by a conformational structural geometry search using the PM6 method, which is much faster than DFT calculation. The lowest energy conformers of CBP, AMP and NOR from the PM6 results were re-optimized by DFT calculations. For illustration purposes, the orbital distributions of HOMO for AMP and NOR, and those of HOMO/HOMO-1 for CBP were rendered. The distribution of Mulliken charges in each chemical structure was also investigated. The values of LUMO, HOMO, and lower energy occupied orbitals were calculated for all structures. All the calculations were performed using the Gaussian 09 computational chemistry software package [10]. All molecule ﬁgures were rendered in Jmol 14 [12].

4. Conclusions The voltammetric and electronic structure calculation findings corroborated the elucidation of CBP, AMP and NOR possible oxidation sites, redox reaction pathway and electrochemical profiling. The results obtained in this study show that the amine moiety of each compound is more likely to undergo oxidation at 1a at E 0.69, 0.79, 0.93 V (vs. Ag/AgCl/KClsat) for CBP, AMP and NOR, p1a ≈ respectively. Moreover, CBP presented a second peak, 2a at E 0.98 V, relative to the oxidation of p2a ≈ C15 due to the high negative charge observed in Mulliken partial distribution. Structural changes in compounds with the same core, such as the dibenzene–cycloheptane core, may yield compounds with different redox dynamics. Thus, electrochemical and computational chemistry studies are versatile approaches that can be used in the investigation of redox behavior.

Author Contributions: E.S.B.R., L.L.d.S.L., C.E.P.d.C., M.F.A. and M.F.d.C. were responsible for the electrochemical analysis; I.Y.L.d.M., D.V.T. and F.F.G. were responsible for the computational chemistry analysis; N.B., M.T.d.O., K.C.d.S.L. and E.d.S.G. were responsible for the writing—original draft preparation; B.G.I., F.B.M., R.M. and V.S. were responsible for the writing—review and editing. Funding: This research was funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), grant number [409287/2018-0]. Acknowledgments: The authors wish to thank Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Fundação de Amparo à Pesquisa do Estado de Goiás (FAPEG) and Coordenadoria de Aperfeiçoamento de Pessoal (CAPES) for their financial support and scolarships. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Yang, Y.W.; Macdonald, J.B.; Nelson, S.A.; Sekulic, A. Treatment of vismodegib-associated muscle cramps with cyclobenzaprine: A retrospective review. J. Am. Acad. Dermatol. 2017, 77, 1170–1172. [CrossRef] [PubMed] Pharmaceuticals 2019, 12, 116 9 of 10

2. De Santana Santos, T.; Calazans, A.C.M.; Martins-Filho, P.R.S.; Da Silva, L.C.F.; De Oliveira, E.; Silva, E.D.; Gomes, A.C.A. Evaluation of the muscle relaxant cyclobenzaprine after third-molar extraction. J. Am. Dent. Assoc. 2011, 142, 1154–1162. [CrossRef][PubMed] 3. Zhang, T.T.; Xue, R.; Wang, X.; Zhao, S.W.; An, L.; Li, Y.F.; Zhang, Y.Z.; Li, S. Network-based drug repositioning: A novel strategy for discovering potential antidepressants and their mode of action. Eur. Neuropsychopharmacol. 2018, 28, 1137–1150. [CrossRef][PubMed] 4. De Macêdo, I.Y.L.; Garcia, L.F.; Menegatti, R.; Guimarães, F.F.; Lião, L.M.; de Carvalho, F.S.; Torres Pio dos Santos, W.; Verly, R.M.; Arotiba, O.A.; de Souza Gil, E. Electrochemical characterizations of darbufelone, a di-tert-butylphenol derivative, by voltammetric techniques and density functional theory calculations. Electrochim. Acta 2018, 268, 462–468. [CrossRef] 5. de Siqueira Leite, K.C.; Garcia, L.F.; Sanz, G.; Colmati, F.; de Souza, A.R.; da Costa Batista, D.; Menegatti, R.; de Souza Gil, E.; Luque, R. Electrochemical characterization of a novel nimesulide anti-inflammatory drug analog: LQFM-091. J. Electroanal. Chem. 2018, 818, 92–96. [CrossRef] 6. Thomaz, D.V.; de S. Leite, K.C.; Moreno, E.K.G.; Garcia, L.F.; Alecrim, M.F.; Macêdo, I.Y.L.; Caetano, M.P.; de Carvalho, M.F.; Machado, F.B.; de Gil, E.S. Electrochemical study of commercial black tea samples. Int. J. Electrochem. Sci. 2018, 13, 5433–5439. 7. Torres, S.; Brown, R.; Szucs, R.; Hawkins, J.M.; Zelesky, T.; Scrivens, G.; Pettman, A.; Taylor, M.R. The application of electrochemistry to pharmaceutical stability testing—Comparison with in silico prediction and chemical forced degradation approaches. J. Pharm. Biomed. Anal. 2015, 115, 487–501. [CrossRef] 8. Zhang, Y.H.; Xia, Z.N.; Qin, L.T.; Liu, S.S. Prediction of blood-brain partitioning: A model based on molecular electronegativity distance vector descriptors. J. Mol. Graph. Model. 2010, 29, 214–220. [CrossRef] 9. Verstraelen, T.; Bultinck, P. Can the electronegativity equalization method predict spectroscopic properties? Acta Part A Mol. Biomol. Spectrosc. 2015, 136, 76–80. [CrossRef] 10. Frisch, M.J.; Trucks, G.W.; Schlegel, H.B.; Scuseria, G.E.; Robb, M.A.; Cheeseman, J.R.; Scalmani, G.; Barone, V.; Petersson, G.A.; Nakatsuji, H.; et al. Gaussian 09, Revision, A.02; Gaussian, Inc.: Wallingford, CT, USA, 2016. 11. Hasanzadeh, N.; Nori-Shargh, D. Correlations between hardness, electronegativity, anomeric effect associated with electron delocalizations and electrostatic interactions in 1,4,5,8-tetraoxadecalin and its analogs containing S and Se atoms. Comput. Theor. Chem. 2015, 1051, 1–9. [CrossRef] 12. Jmol: An Open-Source Java Viewer for Chemical Structures in 3D. 2009. Available online: http://jmol. sourceforge.net/ (accessed on 22 January 2019). 13. Enache, T.A.; Oliveira-Brett, A.M. Phenol and para-substituted phenols electrochemical oxidation pathways. J. Electroanal. Chem. 2011, 655, 9–16. [CrossRef] 14. Ban, Z.; Kätelhön, E.; Compton, R.G. Voltammetry of porous layers: Staircase vs. analog voltammetry. J. Electroanal. Chem. 2016, 776, 25–33. [CrossRef] 15. Ferancová, A.; Korgová, E.; Mikó, R.; Labuda, J. Determination of tricyclic antidepressants using a carbon paste electrode modified with β-cyclodextrin. J. Electroanal. Chem. 2000, 492, 74–77. [CrossRef] 16. Ivandini, T.A.; Sarada, B.V.; Terashima, C.; Rao, T.N.; Tryk, D.A.; Ishiguro, H.; Kubota, Y.; Fujishima, A. Electrochemical detection of tricyclic antidepressant drugs by HPLC using highly boron-doped diamond electrodes. J. Electroanal. Chem. 2002, 521, 117–126. [CrossRef] 17. Wang, J.; Golden, T.; Ozsoz, M.; Lu, Z. Sensitive and selective voltammetric measurements of tricyclic antidepressants using lipid-coated electrodes. J. Electroanal. Chem. 1990, 298, 217–226. [CrossRef] 18. Uchida, Y.; Kätelhön, E.; Compton, R.G. Cyclic voltammetry with non-triangular waveforms: Electrochemically irreversible and quasi-reversible systems. J. Electroanal. Chem. 2018, 810, 135–144. [CrossRef] 19. Harada, R.; Hosomi, T.; Masai, H.; Terao, J. Reversible and stable redox behavior of a Pt (II) bis(dithiobenzoate)-type complex attributed to rotaxane-based stabilization. Tetrahedron Lett. 2018, 59, 2930–2933. [CrossRef] 20. De Toledo, R.A.; Santos, M.C.; Suffredini, H.B.; Homem-de-Mello, P.; Honorio, K.M.; Mazo, L.H. DFT and electrochemical studies on nortriptyline oxidation sites. J. Mol. Modeling 2009, 15, 945–952. [CrossRef] 21. Rodrigues, A.S.; Nunes, M.J.; Lopes, A.; Silva, J.N.; Ciríaco, L.; Pacheco, M.J. Electrodegradation of naphthalenic amines: Influence of the relative position of the substituent groups, anode material and electrolyte on the degradation products and kinetics. Chemosphere 2018, 205, 433–442. [CrossRef] Pharmaceuticals 2019, 12, 116 10 of 10

22. Shabbir, M.; Akhter, Z.; Ahmad, I.; Ahmed, S.; Bolte, M.; McKee, V. Synthesis and bioelectrochemical behavior of aromatic amines. Bioorg. Chem. 2017, 75, 224–234. [CrossRef] 23. Liu, Y.; Zhao, D.; Zhao, Z. Study of degradation mechanisms of cyclobenzaprine by LC-MS/MS. Anal. Methods 2014, 6, 2320–2330. [CrossRef] 24. Darwish, H.W.; Naguib, I.A.; Darwish, I.A. Five modified classical least squares based models for stability indicating analysis of cyclobenzaprine HCl with its major degradation products: A comparative study. Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 2018, 204, 598–602. [CrossRef][PubMed] 25. Harde, M.T.; Wankhede, S.B.; Chaudhari, P.D. A validated inherent stability indicating HPTLC method for estimation of cyclobenzaprine hydrochloride in tablets and use of MS–QTOF in characterization of its alkaline stress degradation product. Bull. Fac. Pharm. Cairo Univ. 2016, 54, 145–156. [CrossRef] 26. Wang, J.; Bonakdar, M.; Morgan, C. Voltammetric measurement of tricyclic antidepressants following interfacial accumulation at carbon electrodes. Anal. Methods 1986, 58, 1024–1028. 27. Suckow, R.F.; Cooper, T.B. Simultaneous determination of imipramine, desipramine, and their 2-hydroxy etabolites in plasma by ion-pair reversed-phase high-performance liquid chromatography with amperometric detection. J. Pharm. Sci. 1981, 70, 257–261. 28. Gil, E.S.; Kubota, L.T. Electrochemical properties of Doyle catalyst immobilized on carbon paste in the presence of DNA. Bioelectrochem. Bioenerg. 2000, 51, 145–149. [CrossRef] 29. Laviron, E. General expression of the linear potential sweep voltammogram in the case of diffusionless electrochemical systems. J. Electroanal. Chem. 1979, 101, 19–28. [CrossRef] 30. Munir, A.; Bozal-Palabiyik, B.; Khan, A.; Shah, A.; Uslu, B. A novel electrochemical method for the detection of oxymetazoline drug based on MWCNTs and TiO2 nanoparticles. J. Electroanal. Chem. 2019, 844, 58–65. [CrossRef] 31. Thomaz, D.V.; de Aguiar Filho, A.M.; de Macêdo, I.Y.L.; Rodrigues, E.S.B.; Gil, E.S. Predictive modelling to study the electrochemical behaviour of PdO, TiO2 and perovskite-type LaFeO3 modified carbon paste electrodes. Path Sci. 2018, 5, 4001–4007. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).