![Molecular Encapsulation of Histamine H2-Receptor Antagonists by Cucurbit[7]Uril: an Experimental and Computational Study](http://data.docslib.org/img/a48e4084c10a8194323d01e40e23f370-1.webp)
molecules
Article Molecular Encapsulation of Histamine H2-Receptor Antagonists by Cucurbit[7]Uril: An Experimental and Computational Study
Hang Yin, Runmiao Wang, Jianbo Wan, Ying Zheng, Defang Ouyang and Ruibing Wang * State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Taipa, Macau 999078, China; [email protected] (H.Y.); [email protected] (R.W.); [email protected] (J.W.) [email protected] (Y.Z.); [email protected] (D.O.) * Correspondence: [email protected]; Tel.: +853-8822-4689
Academic Editors: Jianbo Xiao, Pinarosa Avato and Baodong Zheng Received: 17 July 2016; Accepted: 1 September 2016; Published: 6 September 2016
Abstract: The histamine H2-receptor antagonists cimetidine, famotidine and nizatidine are individually encapsulated by macrocyclic cucurbit[7]uril (CB[7]), with binding affinities of 6.57 (±0.19) × 103 M−1, 1.30 (±0.27) × 104 M−1 and 1.05 (±0.33) × 105 M−1, respectively. These 1:1 host-guest inclusion complexes have been experimentally examined by 1H-NMR, UV-visible spectroscopic titrations (including Job plots), electrospray ionization mass spectrometry (ESI-MS), and isothermal titration calorimetry (ITC), as well as theoretically by molecular dynamics (MD) computation. This study may provide important insights on the supramolecular formulation of H2-receptor antagonist drugs for potentially enhanced stability and controlled release based on different binding strengths of these host-guest complexes.
Keywords: cucurbit[7]uril; H2-receptor antagonists; molecular modelling; complexation; host-guest interaction
1. Introduction Among various macrocyclic molecules, an emerging family of molecular capsules known as cucurbit[n]urils (CB[n]s, n = 5–8, 10, 14) have recently attracted increasing attention in the field of pharmaceutical sciences and biomedical research [1,2]. CB[n]s consist of n glycoluril units that are connected by n pairs of methylene groups, with a lipophilic cavity in the middle that is accessible by a variety of guest molecules via two polar carbonyl-laced portals [3,4]. Due to its superior water-solubility, CB[7] (Figure1) is a particularly attractive capsule to host a variety of guest molecules of biomedical and medical interest, both in vitro and in vivo [1,2]. On the one hand, CB[7]’s safety profile and biocompatibility has been well studied with several in vitro, in vivo and ex vivo models [5–7]. For instance, several in vitro studies on cell cultures have shown that CB[7] exhibits very low toxicity at up to 1 mM concentrations. The effects observed for an intravenous single dose i.v. injection with a mouse model demonstrated that CB[7] has a very low acute toxicity at a dose of 250 mg/kg, based on a body weight change of less than 10% within 5 days of the injections [5]. The tissue specific toxicity including neuro-, myo- and cardiotoxicity of CB[7] has been examined with the use of ex vivo electrophysiological models. The study reported that 1 mM of CB[7] did not exhibit statistically measurable neurotoxicity, although myotoxic and cardiotoxic activities were observed in the presence of CB[7] concentrations of 0.3 mM [6]. Very recently, we have studied the developmental and organ-specific toxicity profiles of CB[7] with live zebrafish models and concluded that CB[7] is relatively safe and biocompatible at functional levels, which is consistent with previous in vitro, ex vivo and in vivo results [7].
Molecules 2016, 21, 1178; doi:10.3390/molecules21091178 www.mdpi.com/journal/molecules Molecules 2016, 21, 1178 2 of 11
On the other hand, even before these preliminary investigations of CB[7]’s safety profile had beenMolecules performed, 2016, 21, 1178 over a decade ago the Collins [8] and Kim [9] research groups independently2 of 11 pioneered the use of CB[7] for encapsulation of platinum complex-based anti-cancer agents, and they haveOn the demonstrated other hand, even that thebefore toxicity these ofpreliminary these agents investigations was reduced of CB[7] upon′s molecularsafety profile encapsulation had been performed, over a decade ago the Collins [8] and Kim [9] research groups independently pioneered by CB[7], presumably due to steric protection provided by the molecular capsule [8,9]. Following these the use of CB[7] for encapsulation of platinum complex-based anti-cancer agents, and they have works, CB[7] has drawn increasing attention as a host for the encapsulation of drug molecules demonstrated that the toxicity of these agents was reduced upon molecular encapsulation by CB[7], in recent years. Along this line, our research group has investigated CB[7]’s encapsulations of presumably due to steric protection provided by the molecular capsule [8,9]. Following these works, a variety of biomedically important molecules during the past years, including imidazolium- and CB[7] has drawn increasing attention as a host for the encapsulation of drug molecules in recent years. thiazolium-basedAlong this line, our model research drugs group [10 has,11 ],investigated the photosensitizer CB[7]’s encapsulations norharmane of [a12 variety], the of anti-peptic biomedically ulcer drugimportant ranitidine molecules [13], during the anticoagulant the past years, coumarin including and imidazolium- the associated and thiazolium-based model drug coumarin-6 model drugs [14 ], + vitamin[10,11], B the12 and photosensitizer its coenzyme norharmane [15], vitamin [12], Bthe1 [ 16anti],-peptic vitamin ulcer B6 [drug17], theranitidine neurotoxin [13], the MPTP/MPP anticoagulant[18 ], anestheticcoumarin agentsand the suchassociated as tricaine model [drug19], benzocainecoumarin-6 [14], and vitamin its metabolite B12 and itspara coenzyme-aminobenzoic [15], vitamin acid B [120 ], as[16], well vitamin as the B anti-cancer6 [17], the neurotoxin drug camptothecin MPTP/MPP+ [[18],21] andanesthetic the anti-tuberculosis agents such as tricaine drug [19], clofazimine benzocaine [22 ]. It hasand beenits metabolite generally para demonstrated-aminobenzoic by acid us [20], and as other well researchersas the anti-cancer that drugs drug camptothecin encapsulated [21] within and the CB[7] mayanti-tuberculosis lead to one or drug more clofazimine of several [22]. benefits It has been including generally improved demonstrated solubility, by us and chemical other researchers stability, and therapeuticthat drugs efficacy encapsulated as well within as reduced CB[7] side-effectsmay lead to [one14,18 or, 19more,21– of24 several]. For instance, benefits including a molecular improved capsule of coumarin-6@CB[7]solubility, chemical was stability, shown and to havetherapeutic significantly efficacy increased as well as bio-uptake reduced side-effects both in vitro [14,18,19,21–24].and in vivo, in comparisonFor instance, with a molecular the free coumarin-6 capsule of [coumarin-6@CB[7]14]. Of biomedical was relevance, shown weto have significantly demonstrated increased that CB[7] encapsulationbio-uptake both of a in neurotoxin vitro and MPTPin vivo,in in vivo comparisonmay lessen with the the neurotoxicity free coumarin-6 of the [14]. guest Of moleculebiomedical [18 ]. Additionally,relevance, we we have observed demonstrated that encapsulation that CB[7] encapsulation of anti-cancer of a drugneurot camptothecinoxin MPTP in andvivo anti-tuberculosismay lessen the drugneurotoxicity clofazimine of the by CB[7]guest molecule has reduced [18]. theseAdditionally, drug’s inherentwe observed toxicities that encapsulation and maintained of anti-cancer their therapeutic drug efficacy,camptothecin as demonstrated and anti-tuberculosis by both in dr vitroug clofazimine and in vivo by CB[7] evidence has reduced [21,22]. these drug’s inherent toxicities andSimilarly, maintained the their encapsulation therapeutic efficacy, of an anti-pepticas demonstrated ulcer by drug both ranitidine,in vitro and whichin vivo isevidence also a histamine[21,22]. Similarly, the encapsulation of an anti-peptic ulcer drug ranitidine, which is also a histamine H2-receptor antagonist, by CB[7], protected the drug from thermal degradation, and this might be H2-receptor antagonist, by CB[7], protected the drug from thermal degradation, and this might be employed to extend the shelf-life of this drug and potentially enhance its therapeutic efficacy [13]. employed to extend the shelf-life of this drug and potentially enhance its therapeutic efficacy [13]. We We have recently extended our study to several other histamine H2-receptor antagonists (Figure1), have recently extended our study to several other histamine H2-receptor antagonists (Figure 1), namely namely cimetidine (CT), famotidine (FT) and nizatidine (NT), for their molecular encapsulation by cimetidine (CT), famotidine (FT) and nizatidine (NT), for their molecular encapsulation by CB[7] CB[7]experimentally experimentally and computationally, and computationally, and both and sets both of data sets have of data supported have supportedthe formation the of formation relatively of relativelystrong 1:1 strong host-guest 1:1 host-guest inclusion inclusion complexes complexes between between each of eachthese ofdr theseugs and drugs CB[7], and suggesting CB[7], suggesting that thatcomputational computational methods methods may may be beused used to topredict predict dr drug-carrierug-carrier interactions interactions and and preliminary preliminary drug drug formulationformulation screening. screening. More More importantly,importantly, different binding binding affi affinitiesnities between between each each of ofthese these drugs drugs with with CB[7]CB[7] may may find find application application for for potentiallypotentially controlled release release of of this this group group of of drug drug molecules. molecules.
FigureFigure 1. 1.Molecular Molecularstructures structures of cimetidine (CT), (CT), famotidine famotidine (FT), (FT), nizatid nizatidineine (NT) (NT) and and CB[7]. CB[7].
Molecules 2016, 21, 1178 3 of 11
Molecules 2016, 21, 1178 3 of 11 2. Results and Discussion 2. Results and Discussion 2.1. Experimental Study of Molecular Encapsulations of Histamine H2-Receptor Antagonists 2.1. Experimental Study of Molecular Encapsulations of Histamine H2-Receptor Antagonists 2.1.1. 1H-NMR Studies of the Encapsulation Sites 2.1.1. 1H-NMR Studies of the Encapsulation Sites The binding sites of these molecular encapsulation complexes were examined by 1H-NMR spectroscopy.The binding Generally, sites of a protonthese molecular resonance encapsulation will shift to complexes an upfield were resonance examined if it isby encapsulated 1H-NMR withinspectroscopy. the cavity Generally, of CB[n], a whereas proton resonance it will shift will to ashift downfield to an upfield resonance resonance if it is if located it is encapsulated outside of the cavitywithin but the close cavity to of the CB[ portaln], whereas of the CB[it willn]. shift Protons to a downfield that are well resonance outside if ofit is the located macrocyclic outside capsulesof the cavity but close to the portal of the CB[n]. Protons that are well outside of the macrocyclic capsules would not exhibit any resonance shifts. As illustrated in Figure2a, in D 2O solution, the aromatic protonwould H(1), not exhibit methyl any protons resonance H(6) shifts. and As methylene illustrated protons in Figure (H(2), 2a, in H(3)D2O solution, and H(4) the protons) aromatic of proton CT have shiftedH(1), methyl upfield protons in the presenceH(6) and methylene of CB[7] while protons only (H(2), the H(3) methyl and proton H(4) protons) H(5) has of shiftedCT have downfield, shifted upfield in the presence of CB[7] while only the methyl proton H(5) has shifted downfield, indicating indicating that the entire aromatic ring and methylene groups are located within the cavity of CB[7] that the entire aromatic ring and methylene groups are located within the cavity of CB[7] whereas whereas the methyl on nitrogen is sitting outside of the cavity but near the portal. When insufficient the methyl on nitrogen is sitting outside of the cavity but near the portal. When insufficient CB[7] CB[7] was present, appearance of only one set of NMR proton resonances of CT suggest that the was present, appearance of only one set of NMR proton resonances of CT suggest that the exchange exchange rate between the bound and free forms is fast with respect to the NMR time-scale. rate between the bound and free forms is fast with respect to the NMR time-scale. Similarly,Similarly, for for FT, FT, as as shown shown inin FigureFigure 22b,b, the aromatic proton proton H(1) H(1) and and methylene methylene protons protons (H(2), (H(2), H(3)H(3) and and H(4) H(4) protons) protons) ofof FTFT havehave shifted upfield upfield in in the the presence presence of of CB[7], CB[7], suggesting suggesting that that all allthese these groupsgroups of of FT FT are are encapsulated encapsulated withinwithin the cavity of of CB CB[7].[7]. In In contrast contrast with with the the case case of CT, of CT, two two separate separate setssets of of resonances resonances when when insufficientinsufficient CB[7] was adde addedd into into FT FT indicate indicate that that the the exchange exchange rate rate between between thethe bound bound and and free free forms formsof of FTFT isis slowslow on NMR NMR time-s time-scale.cale. For For NT, NT, as as shown shown by by Figure Figure 2c,2 thec, the aromatic aromatic protonproton H(1) H(1) and and methylene methylene protonsprotons (H(2), H(3), H(3), H( H(4)4) and and H(7)) H(7)) of of NT NT have have shifted shifted upfield upfield in the in the presencepresence of of CB[7] CB[7] whereas whereas thethe H(6)H(6) andand H(8) prot protonsons have have shifted shifted downfield, downfield, corresponding corresponding to the to the encapsulationencapsulation of of these these groupsgroups withinwithin the cavity cavity of of CB[7] CB[7] and and suggesting suggesting that that the the three three nitrogen nitrogen methyl methyl groupsgroups are are outsideoutside of of the the cavity. cavity. Because Because of the of activation the activation of theof H(5) the proton H(5) protonin D2O solution, in D2O the solution, proton the protonreadily readily underwent underwent exchange exchange with deuterium, with deuterium, so that there so that was there no visible was nosignal visible in the signal NMR in spectra. the NMR spectra.Like NMR Like spectra NMR spectraof FT, the of FT,two the separate two separate sets of resonances sets of resonances with insufficient with insufficient CB[7] exhibited CB[7] exhibited slow slowexchange exchange rates rates between between the bound the bound and free and forms free formsof NT ofon NT the onNMR the time-scale. NMR time-scale.
(a)
Figure 2. Cont.
Molecules 2016, 21, 1178 4 of 11 Molecules 2016, 21, 1178 4 of 11
(b)
(c)
1 FigureFigure 2. 1H-NMR 2. H-NMR (400 (400 MHz) MHz) spectra spectra of (ofa) ( cimetidine;a) cimetidine; (b )(b famotidine) famotidine and and (c ()c nizatidine,) nizatidine, in in thethe absenceabsence and inand the in presence the presence of 0.6 of and0.6 and 1.8 equiv.1.8 equiv. of CB[7]of CB[7 (with] (with increasing increasing CB[7] CB[7] concentrations concentrationsfrom from bottombottom upwards) in D2O. upwards) in D2O. 2.1.2. Job Plot and ESI-MS Studies 2.1.2. Job Plot and ESI-MS Studies The binding stoichiometries of FT-CB[7] and NT-CB[7] were studied by the continuous variation Thetitration binding (Job plot) stoichiometries and monitored of FT-CB[7] by UV-visible and NT-CB[7] spectroscopy were in studied deionized by water the continuous (Figure S1). variation During titration (Job plot) and monitored by UV-visible spectroscopy in deionized water (Figure S1).
Molecules 2016, 21, 1178 5 of 11
Molecules 2016, 21, 1178 5 of 11 During the Job’s method titration, the total concentration of the guest and the host were the same in theeach Job sample.′s method If a Jobtitration, plot shows the total a maximum concentration at 0.5 (theof the ratio guest of the and concentration the host were of the the host same to thein each total sample.concentration If a Job of plot the host shows and a the maximum guest), it at suggests 0.5 (the that ratio the of binding the concentration stoichiometry of betweenthe host theto the host total and concentrationthe guest is 1:1. of the The host Job plotand the of FT@CB[7] guest), it suggests monitored that by the UV-visible binding stoichio spectroscopymetry between at 269 nm the (Figure host and3a) thereached guest ais maximum 1:1. The Job at plot the ratioof FT@CB[7] of 0.5 for monitore (CB[7])/([CB[7]]d by UV-visible + [FT]), spectroscopy suggesting thatat 269 the nm binding (Figure ratio 3a) reachedof FT and a maximum CB[7] is mainly at the ratio 1:1. of Similarly 0.5 for (CB[7])/([CB[7]] for NT, the Job + plot[FT]), of suggesting NT@CB[7] that (Figure the binding3b) also ratio reached of FT anda maximum CB[7] is mainly at 0.5, implying1:1. Similarly that for the NT, binding the Job ratio plot of of NT NT@CB[7] and CB[7] (Figure is mainly 3b) also 1:1 reached at this concentration a maximum atlevel 0.5, usedimplying in this that study. the binding Because ratio the absorbance of NT and CB[7 of CT] is monitored mainly 1:1 by at UV-visible this concentration spectroscopy level wasused not in thisinfluenced study. Because by the additionthe absorbance of CB[7], of CT the monitored binding stoichiometry by UV-visible of spectroscopy CT@CB[7] could was not not influenced be examined by theby thisaddition method. of CB[7], Additionally, the binding the stoichiometry binding stoichiometries of CT@CB[7] of allcould three not complexes be examined were by also this examined method. Additionally,by ESI-MS analysis, the binding which stoichiometries may provide direct of all evidence three complexes of the binding were ratio also of examined these supramolecular by ESI-MS analysis,complexes. which As expectedmay provide (Figure direct S2), evidence a doubly of charged the bindingm/z peakratio of (m/z these= 708.23, supramolecular calculated complexes. at 708.24) Asfor expected CT-CB[7] (Figure sample S2), indicated a doubly a 1:1charged binding m/z stoichiometrypeak (m/z = 708.23, for this calculated complex. at Similarly,708.24) for the CT-CB[7] doubly samplecharged indicatedm/z peaks a 1:1 for binding the FT@CB[7] stoichiometry complex for this (m/ complex.z = 750.70, Similarly, calculated the doubly at 750.70) charged and NT@CB[7] m/z peaks forcomplex the FT@CB[7] (m/z = 747.73, complex calculated (m/z = 750.70, at 747.74) calculated supported at 750.70) that of bothand NT@CB[7] these two complexescomplex (m existed/z = 747.73, as 1:1 calculatedhost-guest at pairs. 747.74) supported that of both these two complexes existed as 1:1 host-guest pairs.
(a)
(b)
FigureFigure 3. 3. JobJob plots plots of of FT@CB[7] FT@CB[7] complexes complexes ( (aa)) and and NT@CB[7] NT@CB[7] complexes complexes ( (bb).).
Molecules 2016, 21, 1178 6 of 11
Molecules 2016, 21, 1178 6 of 11 2.1.3. Binding Affinity Studies by UV-Visible and NMR Spectroscopic Titration 2.1.3. Binding Affinity Studies by UV-Visible and NMR Spectroscopic Titration Binding affinity is always considered as a key parameter to evaluate non-covalent binding behaviorBinding in host-guest affinity is interactions.always considered With as gradual a key parameter addition to of evaluate increasing non-covalent amounts binding of CB[7] behavior from 0 to 6.0in equivalents host-guest interactions. to a solution With of0.05 gradual mM addition FT in deionized of increasing water, amounts the absorbance of CB[7] from at 283 0 tonm 6.0 equivalents monitored by UV-visibleto a solution spectroscopy, of 0.05 mM gradually FT in deionized decreased water, (Figure the 4absorbancea). A modest at 283 hypsochromic nm monitored shift by was UV-visible observed, presumablyspectroscopy, due gradually to the aromatic decreased ring (Figure of FT 4a). was A encapsulated modest hypsochromic in the cavity shift ofwas CB[7]. observed, The fittingpresumably curve of thedue absorbance to the aromatic at 283 ring nm of against FT was encapsulated the host concentration in the cavity showed of CB[7]. good The fitting agreement curve of with the absorbance a 1:1 binding at 283 nm against the host concentration showed good agreement with a 1:14 binding−1 stoichiometry stoichiometry model and provided a binding constant Ka = 1.30 (±0.27) × 10 M . The binding affinity model and provided a binding constant Ka = 1.30 (±0.27) × 104 M−1. The binding affinity of NT@CB[7] of NT@CB[7] was examined by the same method as that used for FT@CB[7]. With the gradual addition was examined by the same method as that used for FT@CB[7]. With the gradual addition of increasing of increasing amounts of CB[7] from 0 to 3.0 equivalents to a solution of 0.04 mM NT, the absorbance amounts of CB[7] from 0 to 3.0 equivalents to a solution of 0.04 mM NT, the absorbance at 315 nm at 315 nm increased with a modest hypsochromic shift (Figure4b). The fitting curve of the absorbance increased with a modest hypsochromic shift (Figure 4b). The fitting curve of the absorbance5 at− 1283 nm at 283 nm against the concentration provided a binding constant Ka = 1.05 (±0.33) × 10 M and also against the concentration provided a binding constant Ka = 1.05 (±0.33) × 105 M−1 and also gave evidence gaveto support evidence a 1:1 to supportbinding stoichiometry. a 1:1 binding stoichiometry. DueDue to to the the lack lack of of responsiveness responsiveness of the CT CT’s′s absorbance absorbance to to complexation complexation with with CB[7], CB[7], the the binding binding 1 1 behaviorsbehaviors of of CT@CB[7] CT@CB[7] was was instead instead monitoredmonitored by 1H-NMRH-NMR spectroscopy spectroscopy in in D D2O.2O. During During the the 1H-NMRH-NMR titration,titration, the the concentration concentration ofof CTCT waswas kept consta constantnt at at 1.0 1.0 mM mM in in the the presence presence of ofan an increasing increasing concentrationconcentration of of CB[7] CB[7] (from (from 00 toto 3.03.0 equivalents).equivalents). The The change change in in the the chemical chemical resonances resonances was was clearly clearly observedobserved (Figure (Figure S3), S3), and and the the H(5) H(5) protonproton of CT wa wass chosen chosen as as a aprobe probe for for the the binding binding behavior behavior because because thethe resonance resonance of of H(5) H(5) proton proton was was thethe mostmost trackable one one among among all all of of the the protons. protons. The The non-linear non-linear least least squaressquares fitting fitting curve curve (Figure (Figure4 c)4c) of of the the chemical chemical resonanceresonance of of H(5) H(5) against against the the concentration concentration of ofCB[7] CB[7] 3 3 −1 −1 yieldedyielded a bindinga binding affinity affinity of ofK Kaa == 6.57 (±0.19) (±0.19) × ×1010 M M. .
(a)
(b)
Figure 4. Cont.
Molecules 2016, 21, 1178 7 of 11 Molecules 2016, 21, 1178 7 of 11
(c)
Figure 4. Absorbance spectra along with the corresponding fitting curves of FT@CB[7] (a) and Figure 4. Absorbance spectra along with the corresponding fitting curves of FT@CB[7] (a) and NT@CB[7] (b) as monitored by UV-visible spectrometry. Fitting curve of H(5) proton NMR resonance NT@CB[7] (b) as monitored by UV-visible spectrometry. Fitting curve of H(5) proton NMR resonance of CT against the concentration of CB[7] in D2O at 25 ◦°C (c). of CT against the concentration of CB[7] in D2O at 25 C(c). 2.1.4. ITC Titration Studies 2.1.4. ITC Titration Studies Isothermal titration calorimetry (ITC) is a powerful technique to give not only binding stoichiometry (N)Isothermal and binding titration affinity calorimetry (Ka) but also (ITC)thermodynamic is a powerful parameters technique (∆H, ∆G to and give T∆S not) by monitoring only binding stoichiometrythe micro-level (N) andthermal binding change affinity of a given (Ka) system. but also An thermodynamic aqueous solution parameters of CB[7] for (ITC∆H ,was∆G loadedand T ∆S) by monitoringinto the titration the micro-level syringe and soluti thermalons changeof CT, FT, of or a NT given in deionized system. water An aqueous were loaded solution in the of titration CB[7] for ITC wascell, individually, loaded into and the titrationITC titration syringe analysis and of solutions these samples of CT, was FT, conducted or NT in at deionized 25 °C. The water binding were loadedparameters in the titration including cell, N (binding individually, stoichiometry), and ITC K titrationa (binding analysisaffinity), ∆ ofH these(enthalpy samples change), was ∆G conducted (Gibbs free◦ energy change) and T∆S (entropy change), was auto-analyzed by MicroCal PEAQ-ITC Analysis at 25 C. The binding parameters including N (binding stoichiometry), Ka (binding affinity), ∆H Software 1.1.0.1262 (for ITC titration isotherms, refer to Figure S4), and summarized in the Table 1. (enthalpy change), ∆G (Gibbs free energy change) and T∆S (entropy change), was auto-analyzed by
MicroCalTable PEAQ-ITC 1. Thermodynamic Analysis parameters Software of 1.1.0.1262 non-covalent (for interactio ITC titrationns of CT@CB[7], isotherms, FT@CB[7] refer and to NT@CB[7] Figure S4), and summarizedencapsulation in the Table complexes1. at 25 °C, derived from ITC, NMR and UV-vis titrations (all uncertainties are standard deviations). Table 1. Thermodynamic parameters of non-covalent interactions of CT@CB[7], FT@CB[7] and NT@CB[7]Complexes encapsulation Method complexesN at 25Ka (M◦C,−1) derivedΔH from(kCal/mol) ITC, NMRΔG (kCal/mol) and UV-vis TΔ titrationsS (kCal/mol) (all ITC 1.44 (± 0.38) × 104 −5.69 ± 0.16 uncertaintiesCT@CB[7] are standard deviations).0.99 −13.23 ± 4.01 −7.57 ± 0.01 NMR 6.57 (± 0.19) × 103 −5.21 ± 0.02 ITC 4.95 (± 0.36) × 104 −6.40 ± 0.04 FT@CB[7] 1.12 −1 −11.08 ± 0.38 −4.68 ± 0.01 Complexes MethodUV-vis N 1.30Ka (M (± 0.27)) × 104 ∆H (kCal/mol) ∆G (kCal/mol)−5.62 ± 0.12 T∆S (kCal/mol) ITCITC 1.441.34 (± 0.38) (± 0.21)× 10 ×4 104 −5.69−5.64± 0.16± 0.09 NT@CB[7]CT@CB[7] 0.990.95 −13.23−15.60± 4.01 ± 2.18 −7.57−9.96± 0.01 ± 0.01 NMRUV-vis 6.571.05 (± 0.19) (± 0.33)× 10 ×3 105 −5.21−6.85± 0.02± 0.19 ITC 4.95 (± 0.36) × 104 −6.40 ± 0.04 FT@CB[7] 1.12 −11.08 ± 0.38 −4.68 ± 0.01 ITC tests haveUV-vis further confirmed1.30 the (± 0.27) 1:1 ×binding104 stoichiometry of− these5.62 ± complexes,0.12 consistent with ITC 1.34 (± 0.21) × 104 −5.64 ± 0.09 Job plotsNT@CB[7] and ESI-MS results.0.95 For all of the three complexes,−15.60 ± enthalpy2.18 change was the− main9.96 ± contributor0.01 UV-vis 1.05 (± 0.33) × 105 −6.85 ± 0.19 to these non-covalent interactions, suggesting hydrogen bondings and ion-dipole interactions are the main driving forces. This is likely due to strong interactions between protonated amines of the drug moleculesITC tests haveand carbonyl further confirmedportals of the the host. 1:1 binding Of note, stoichiometry the binding affinities of these de complexes,rived from ITC consistent tests are with Job plotsdifferent and from, ESI-MS but results. generally For comparable all of the three with, complexes,the previously enthalpy discussed change results was from the UV-visible main contributor and to theseNMR non-covalent spectroscopic interactions,titrations. The suggesting most significant hydrogen difference bondings between and Ka ion-dipolevalues derived interactions from ITC are the mainand NMR/UV-Vis driving forces. methods This isis likelyfor the due NT@CB[7] to strong complexes, interactions where between the Ka from protonated UV-Vis aminestitration ofis the drugapproximately molecules and 7.8 carbonyl fold of the portals value from of the ITC host. titration. Of note, To the the best binding of our knowle affinitiesdge, derived the much from difference ITC tests are differentin Ka values from, is reasonable but generally as binding comparable constants with,may vary the quite previously significantly discussed if different results methods from are UV-visible used [25]. Additionally, the binding constants varying from 103 to 105 in aqueous solutions seem to be and NMR spectroscopic titrations. The most significant difference between K values derived from comparable with those recently observed on other drug-CB[7] complexes [14,16,20,22],a attesting the ITC and NMR/UV-Vis methods is for the NT@CB[7] complexes, where the K from UV-Vis titration potential of these complexations in biomedical sciences. a is approximatelyAccording 7.8 to foldthe formula: of the value ΔG = − fromRTlnK ITC (T = titration. 297.5 K), the To theΔG values best of of our the knowledge,complexes could the muchbe differencereadily in calculatedKa values from is reasonablethe experimental as binding results based constants on Ka may value vary derived quite from significantly the UV-visible if different and methods are used [25]. Additionally, the binding constants varying from 103 to 105 in aqueous solutions seem to be comparable with those recently observed on other drug-CB[7] complexes [14,16,20,22], attesting the potential of these complexations in biomedical sciences. Molecules 2016, 21, 1178 8 of 11
MoleculesAccording 2016, 21, 1178 to the formula: ∆G = −RTlnK (T = 297.5 K), the ∆G values of the complexes8 of 11 could be readily calculated from the experimental results based on Ka value derived from the CT − FT − NT − UV-visibleNMR spectroscopic and NMR titrations spectroscopic (ΔG = titrations 5.21 kCal/mol, (∆GCT Δ=G−5.21 = 5.62 kCal/mol, kCal/mol∆ GandFT =ΔG−5.62 = kCal/mol6.85 kCal/mol). and ∆TheseGNT =values−6.85 are kCal/mol generally). Theseconsistent values with are the generally Gibbs free consistent energy withchanges the Gibbsmeasured free by energy ITC changesmethods measured(Table 1). by ITC methods (Table1).
2.2.2.2. ComputationalComputational StudyStudy ofof Host-GuestHost-Guest ComplexesComplexes CT,CT, FTFT andand NTNT areare generallygenerally hydrophilichydrophilic moleculesmolecules withwith hydrophobichydrophobic moieties,moieties, whilewhile CB[7]CB[7] hashas aa hydrophobichydrophobic cavity.cavity. Therefore,Therefore, thethe waterwater moleculesmolecules havehave strongerstronger influenceinfluence onon thethe guestguest moleculesmolecules andand theirtheir bindingbinding withwith CB[7]CB[7] inin thesethese systems.systems. FigureFigure5 5 showed showed the the complexes’ complexes’ structures structures of of these these drug@CB[7]drug@CB[7] complexescomplexes byby molecularmolecular dynamicsdynamics (MD)(MD) simulationsimulation wherewhere waterwater moleculesmolecules cancan bebe builtbuilt in.in. TheThe aromaticaromatic ringring andand methylenemethylene groups of CT, FTFT andand NTNT areare locatedlocated withinwithin thethe cavitycavity ofof CB[7],CB[7], whichwhich areare consistentconsistent withwith thethe bindingbinding geometriesgeometries derivedderived fromfrom ourour experimentalexperimental resultsresults (Figure(Figure2 2),), as as ourour NMRNMR experimentsexperiments showedshowed thatthat thethe entireentire aromaticaromatic ringring andand methylenemethylene groupsgroups ofof CTCT werewere locatedlocated withinwithin thethe cavitycavity ofof CB[7]CB[7] whereaswhereas thethe methylmethyl onon nitrogennitrogen isis sittingsitting outsideoutside ofof thethe cavitycavity butbut nearnear thethe portal.portal. TheThe aromaticaromatic andand methylenemethylene groupsgroups ofof FTFT and NT were encapsulated within the cavity of CB. ThreeThree nitrogennitrogen methylmethyl groupsgroups ofof NTNT werewere outside outside of of the the cavity. cavity.
FigureFigure 5.5. SnapshotsSnapshots ofof CT@CB[7]CT@CB[7] ((leftleft),), FT@CB[7]FT@CB[7] ((middlemiddle)) andand NT@CB[7]NT@CB[7] ((rightright)) atat 3030 ns.ns.
TableTable2 2shows shows the the thermodynamic thermodynamic parameters parameters including including enthalpy enthalpy and and entropy entropy changes changes as as well well as as GibbsGibbs freefree energiesenergies ofof thethe complexescomplexes thatthat werewere calculatedcalculated byby thethe MM_GBSAMM_GBSA methodmethod [[26].26]. Interestingly, thethe calculatedcalculated datadata alsoalso suggestedsuggested thatthat enthalpyenthalpy changeschanges areare thethe mainmain contributorscontributors towardstowards thesethese non-covalentnon-covalent interactionsinteractions whereaswhereas entropyentropy changeschanges contributedcontributed negatively,negatively, consistentconsistent withwith ourour ITCITC resultsresults (Table(Table1 ).1). Entropic Entropic contribution contribution is is associated associated with with the the changes changes of of water water molecules molecules within within the the CB[CB[nn]] cavity, but this this might might be be compensated compensated by by the the ordered ordered structure structure of host-guest of host-guest self-assembly self-assembly and andassociated associated water water molecules molecules by the by portals the portals of the of host the molecules. host molecules. The observed The observed discrepancy discrepancy in values in valuesbetween between the calculated the calculated and the experimental and the experimental methods has methods been observed has been frequently observed frequentlybefore, as different before, asexperimental different experimental and modeling and approaches modeling may approaches have re maysult in have different result values. in different There values. are two There possible are tworeasons possible for such reasons differences: for such on differences: the one hand, on the the one systems hand, of the MD systems simulation of MD were simulation ideal systems were idealwith systemsonly one with ligand, only one one macrocycle ligand, one and macrocycle water molecule and waters, while molecules, the real while soluti theons real contained solutions much contained more muchsolute moreand solvent solute molecules; and solvent on molecules; the other hand, on the the other force hand,field may the be force necessary field may to be be further necessary optimized to be furtherfor CB[7] optimized systems, for which CB[7] deserves systems, further which in-depth deserves investigations. further in-depth investigations.
TableTable 2. 2.Binding Binding freefree energies energies (kCal/mol) (kCal/mol) ofof drug@CB[7]drug@CB[7] complexescomplexes byby MM_GBSAMM_GBSA method.method. kCal/mol CT@CB[7] FT@CB[7] NT@CB[7] kCal/mol CT@CB[7] FT@CB[7] NT@CB[7] ΔETOT −37.55 ± 5.24 −33.12 ± 3.93 −32.47 ± 3.36 ∆ETOT TΔSTOT −37.55−20.10± ±5.24 1.09 −20.86−33.12 ± 1.09± 3.93−19.68 ± 1.63−32.47 ± 3.36 T∆STOT ΔG −20.10−17.45± ±1.09 4.92 −12.25−20.86 ± 3.77± 1.09−12.79 ± 2.97−19.68 ± 1.63 ∆G −17.45 ± 4.92 −12.25 ± 3.77 −12.79 ± 2.97 3. Materials and Methods
3.1. Materials CB[7] was synthesized according to a literature method [27,28]. CT, FT and NT were purchased from TCI® (Shanghai, China) and used as received.
Molecules 2016, 21, 1178 9 of 11
3. Materials and Methods
3.1. Materials CB[7] was synthesized according to a literature method [27,28]. CT, FT and NT were purchased from TCI® (Shanghai, China) and used as received.
3.2. Instrumentation The 1H-NMR spectra were acquired using an Ultra Shield 400 PLUS NMR spectrometer (Bruker, Karlsruhe, Germany). The ESI-MS spectrometry analysis was conducted using a LTQ OrbiTrap XL instrument (Thermo, Waltham, MA, USA) equipped with an ESI/APcI multiprobe. The UV-visible spectroscopic analysis was performed using a DR6000 UV-visible spectrometer (HACH, Düsseldorf, Germany) with a 1.0 cm path length quartz cell. The isothermal titration calorimetry data was studied with a PEAQ-ITC instrument (Malvern, Northampton, MA, USA).
3.3. Complexes Preparation and Characterization Stock solutions of CT, FT and NT, each at 1 mM concentration, were prepared with Milli-Q water (Merck Millipore, Darmstadt, Germany). The solutions of FT and NT for the Job plot titration and binding affinity titration were prepared by diluting the stock solution to concentrations of 0.05 and 0.04 mM, respectively. The solutions of CB[7] for the Job plot titration were prepared by diluting 1 mM CB[7] to the same concentration as the solutions of FT and NT, respectively. The complexes solutions of FT and NT employed for the binding constant titrations were 0.05 mM FT in the presence of 6.0 equivalents of CB[7], and 0.04 mM NT in the presence of 3.0 equivalents of CB[7], respectively. Due to the poor optical properties of CT and the fact that CB[7] encapsulation doesn’t influence its absorbance properties, both the Job plot titration and binding constant titration of CT@CB[7] were determined by 1H-NMR analysis. The solutions of CT for the binding titration by 1H-NMR analysis were prepared by dissolving 1 mM CT in D2O in the absence and in the presence of various amounts of CB[7] (up to 3.0 equivalents). The ITC solutions of 0.05 mM CT, 0.05 mM FT and 0.1 mM NT were diluted from their 1 mM stock solutions, respectively. The ITC solution of CB[7] was prepared at 2 mM with Milli-Q water.
3.4. MD Computation MD simulations were performed using the Amber14 and Amber Tools 14 software package [29]. All molecules were built using the Leap module with Amber GAFF force field and Antechamber module by AM1-BCC charge method. The Amber Tools 14 were used to build the starting structure of the drug and CB[7], as well as drug-CB[7]complexes with TIP3P water model of 10 Å (Table3). After energy minimization, 30-ns simulations were performed and the protocol was similar to those described in our previous publications [30–32]. The MM_GBSA method was used to calculate the enthalpy (∆H) and the Normal Mode Analysis was performed for the entropy (T∆S)[26]. The 100 MD snapshots from the last 1 ns of each system were used for binding free energy calculations.
Table 3. Simulation settings of the systems.
Parameter CT-CB[7] FT-CB[7] NT-CB[7] Water shell (Å) 10 10 10 Atom number of CB7 126 126 126 Atom number of drug 34 37 44 Atom number of Cl- 1 1 1 Number of water 1413 1736 1365 Total atom number 4400 5374 4265 Molecules 2016, 21, 1178 10 of 11
4. Conclusions In summary, we have investigated both experimentally and computationally the molecular interactions between a promising drug carrier cucurbit[7]uril and a group of H2-histamine antagonists, including cimetidine, famotidine and nizatidine. We have found that the non-covalent interactions between these drugs and the host molecules are mainly enthalpy driven, as hydrogen bonding and ion-dipole interactions play a dominant role in the complexation processes. In addition, molecular modeling techniques could help us investigate the binding behaviors between these drugs and a macrocyclic carrier and in this case confirmed that these complexations are mainly enthalpy driven. These results, taken together, support the use of CB[7] as a potential drug carrier with tunable binding strength that may ultimately benefit formulation and delivery of specifically selected drugs.
Supplementary Materials: Supplementary materials can be accessed at: http://www.mdpi.com/1420-3049/21/ 9/1178/s1. Acknowledgments: Macau Science and Technology Development Fund (FDCT/020/2015/A1), SKL-QRCM Operating Fund (Open Fund 003) and University of Macau Research Fund (SRG2014-00025-ICMS-QRCM) are gratefully acknowledged for providing financial support. Author Contributions: H.Y. and R.W. conceived and designed the experiments; H.Y. and R.M.W. performed the experiments; H.Y., D.O., J.W., Y.Z. and R.W. analyzed the data; J.W. and Y.Z. contributed reagents/materials/analysis tools; H.Y., D.O. and R.W. wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.
References
1. Gavvala, K.; Koninti, R.K.; Sengupta, A.; Hazra, P. Cucurbit[7]uril assisted ultraviolet to visible fluorescence switch of a heart medicine. Phys. Chem. Chem. Phys. 2014, 16, 2823–2826. [CrossRef][PubMed] 2. Tonga, G.Y.; Jeong, Y.; Duncan, B.; Mizuhara, T.; Mout, R.; Das, R.; Kim, S.T.; Yeh, Y.-C.; Yan, B.; Hou, S.; et al. Supramolecular regulation of bioorthogonal catalysis in cells using nanoparticle-embedded transition metal catalysts. Nat. Chem. 2015, 7, 597–603. [CrossRef][PubMed] 3. Macartney, D.H. Encapsulation of drug molecules by cucurbiturils: Effects on their chemical properties in aqueous solution. Isr. J. Chem. 2011, 51, 600–615. [CrossRef] 4. Ghosh, I.; Nau, W.M. The strategic use of supramolecular pKa shifts to enhance the bioavailability of drugs. Adv. Drug Deliv. Rev. 2012, 64, 764–783. [CrossRef][PubMed] 5. Uzunova, V.D.; Cullinane, C.; Brix, K.; Nau, W.M.; Day, A.I. Toxicity of cucurbit[7]uril and cucurbit[8]uril: An exploratory in vitro and in vivo study. Org. Biomol. Chem. 2010, 8, 2037–2042. [CrossRef][PubMed] 6. Oun, R.; Floriano, R.S.; Isaacs, L.; Rowan, E.G.; Wheate, N.J. The ex vivo neurotoxic, myotoxic and cardiotoxic activity of cucurbituril-based macrocyclic drug delivery vehicles. Toxicol. Res. 2014, 3, 447–455. [CrossRef][PubMed] 7. Chen, H.; Chan, J.Y.W.; Yang, X.; Wyman, I.W.; Bardelang, D.; Macartney, D.H.; Lee, S.M.Y.; Wang, R. Developmental and organ-specific toxicity of cucurbit[7]uril: In vivo study on zebrafish models. RSC Adv. 2015, 5, 30067–30074. [CrossRef] 8. Wheate, N.J.; Day, A.I.; Blanch, R.J.; Arnold, A.P.; Cullinane, C.; Grant Collins, J. Multi-nuclear platinum complexes encapsulated in cucurbit[n]uril as an approach to reduce toxicity in cancer treatment. Chem. Commun. 2004, 1424–1425. [CrossRef][PubMed] 9. Jin Jeon, Y.; Kim, S.-Y.; Ho Ko, Y.; Sakamoto, S.; Yamaguchi, K.; Kim, K. Novel molecular drug carrier: Encapsulation of oxaliplatin in cucurbit[7]uril and its effects on stability and reactivity of the drug. Org. Biomol. Chem. 2005, 3, 2122–2125. [CrossRef][PubMed] 10. Wang, R.; Yuan, L.; Macartney, D.H. Inhibition of C(2)-H/D exchange of a bis(imidazolium) dication upon complexation with cucubit[7]uril. Chem. Commun. 2006, 2908–2910. [CrossRef] 11. Li, S.; Miao, X.; Wyman, I.W.; Li, Y.; Zheng, Y.; Wang, Y.; Macartney, D.H.; Wang, R. High-affinity host-guest complex of cucurbit[7]uril with a bis(thiazolium) salt. RSC Adv. 2015, 5, 56110–56115. [CrossRef] 12. Wang, R.; Wyman, I.W.; Wang, S.; Macartney, D.H. Encapsulation of a β-carboline in cucurbit[7]uril. J. Incl. Phenom. Macrocycl. Chem. 2009, 64, 233–237. [CrossRef]
13. Wang, R.; Macartney, D.H. Cucurbit[7]uril host-guest complexes of the histamine H2-receptor antagonist ranitidine. Org. Biomol. Chem. 2008, 6, 1955–1960. [CrossRef][PubMed] Molecules 2016, 21, 1178 11 of 11
14. Miao, X.; Li, Y.; Wyman, I.; Lee, S.M.Y.; Macartney, D.H.; Zheng, Y.; Wang, R. Enhanced in vitro and in vivo uptake of a hydrophobic model drug coumarin-6 in the presence of cucurbit[7]uril. MedChemComm 2015, 6, 1370–1374. [CrossRef] 15. Wang, R.; MacGillivray, B.C.; Macartney, D.H. Stabilization of the base-off forms of vitamin B12 and coenzyme B12 by encapsulation of the α-axial 5,6-dimethylbenzimidazole ligand with cucurbit[7]uril. Dalton Trans. 2009, 3584–3589. [CrossRef][PubMed] 16. Li, S.; Yin, H.; Wyman, I.W.; Zhang, Q.; Macartney, D.H.; Wang, R. Encapsulation of vitamin B1 and its phosphate derivatives by cucurbit[7]uril: Tunability of the binding site and affinity by the presence of phosphate groups. J. Org. Chem. 2016, 81, 1300–1303. [CrossRef][PubMed] 17. Li, W.; Li, S.; Wyman, I.W.; Macartney, D.H.; Zhang, Q.; Zheng, Y.; Wang, R. Supramolecular encapsulation of vitamin B6 by macrocyclic nanocontainer cucurbit[7]uril. J. Nanomater. 2015, 2015.[CrossRef] 18. Li, S.; Chen, H.; Yang, X.; Bardelang, D.; Wyman, I.W.; Wan, J.; Lee, S.M.Y.; Wang, R. Supramolecular inhibition of neurodegeneration by a synthetic receptor. ACS Med. Chem. Lett. 2015, 6, 1174–1178. [CrossRef][PubMed] 19. Chen, H.; Chan, J.Y.W.; Li, S.; Liu, J.J.; Wyman, I.W.; Lee, S.M.Y.; Macartney, D.H.; Wang, R. In vivo reversal of general anesthesia by cucurbit[7]uril with zebrafish models. RSC Adv. 2015, 5, 63745–63752. [CrossRef] 20. Li, S.; Yin, H.; Martinz, G.; Wyman, I.W.; Bardelang, D.; Macartney, D.H.; Wang, R. Supramolecular encapsulation of benzocaine and its metabolite para-aminobenzoic acid by cucurbit[7]uril. New J. Chem. 2016, 40, 3484–3490. [CrossRef] 21. Yang, X.; Wang, Z.; Niu, Y.; Chen, X.; Lee, S.M.Y.; Wang, R. Influence of supramolecular encapsulation of camptothecin by cucurbit[7]uril: Reduced toxicity and preserved anti-cancer activity. MedChemComm 2016, 7, 1392–1397. [CrossRef] 22. Li, S.; Chan, J.Y.-W.; Li, Y.; Bardelang, D.; Zheng, J.; Yew, W.W.; Chan, D.P.-C.; Lee, S.M.Y.; Wang, R. Complexation of clofazimine by macrocyclic cucurbit[7]uril reduced its cardiotoxicity without affecting the antimycobacterial efficacy. Org. Biomol. Chem. 2016, 14, 7563–7569. [CrossRef][PubMed] 23. Chu, J.J.; Apps, M.G.; Wheate, N.J. Chemical factors affecting cucurbit[n]uril formulation into ocular dosage forms: Excipient binding, solubility, corneal permeability and antibiotic encapsulation. Supramol. Chem. 2014, 26, 648–656. [CrossRef] 24. Saleh, N.i.; Al-Handawi, M.B.; Bufaroosha, M.S.; Assaf, K.I.; Nau, W.M. Tuning protonation states of tripelennamine antihistamines by cucurbit[7]uril. J. Phys. Org. Chem. 2016, 29, 101–106. [CrossRef] 25. Barrow, S.J.; Kasera, S.; Rowland, M.J.; del Barrio, J.; Scherman, O.A. Cucurbituril-based molecular recognition. Chem. Rev. 2015, 115, 12320–12406. [CrossRef][PubMed] 26. Genheden, S.; Ryde, U. The mm/pbsa and mm/gbsa methods to estimate ligand-binding affinities. Expert Opin. Drug Discov. 2015, 10, 449–461. [CrossRef][PubMed] 27. Day, A.; Arnold, A.P.; Blanch, R.J.; Snushall, B. Controlling factors in the synthesis of cucurbituril and its homologues. J. Org. Chem. 2001, 66, 8094–8100. [CrossRef][PubMed] 28. Bardelang, D.; Udachin, K.A.; Leek, D.M.; Margeson, J.C.; Chan, G.; Ratcliffe, C.I.; Ripmeester, J.A. Cucurbit[n]urils (n = 5–8): A comprehensive solid state study. Cryst. Growth Des. 2011, 11, 5598–5614. [CrossRef] 29. Case, D.A.; Babin, V.; Berryman, J.T.; Betz, R.M.; Cai, Q.; Cerutti, D.S.; CheathamIII, T.E.; Darden, T.A.; Duke, R.E.; Gohlke, H.; et al. Amber 14; University of California: San Francisco, CA, USA, 2014. 30. Ouyang, D.; Zhang, H.; Herten, D.-P.; Parekh, H.S.; Smith, S.C. Structure, dynamics, and energetics of sirna-cationic vector complexation: A molecular dynamics study. J. Phys. Chem. B 2010, 114, 9220–9230. [CrossRef][PubMed] 31. Ouyang, D.; Zhang, H.; Parekh, H.S.; Smith, S.C. The effect of pH on pamam dendrimer-sirna complexation —Endosomal considerations as determined by molecular dynamics simulation. Biophys. Chem. 2011, 158, 126–133. [CrossRef][PubMed] 32. Wang, R.; Zhou, H.; Siu, S.W.I.; Gan, Y.; Wang, Y.; Ouyang, D. Comparison of three molecular simulation approaches for cyclodextrin-ibuprofen complexation. J. Nanomater. 2015, 2015.[CrossRef]
Sample Availability: Samples of the compounds are available from the authors.
© 2016 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/).