Interaction of 5-S-cysteinyl-dopamine with graphene oxide: an experimental and theoretical study for the detection of a Parkinson's disease biomarker

Isidro Badillo-Ramírez *a, Bruno Landeros-Rivera *b, Emmanuel de la O-Cuevas c, Rubicelia Vargas b, Jorge Garza b and José M. Saniger a
aInstituto de Ciencias Aplicadas y Tecnología, Universidad Nacional Autónoma de México, Circuito Exterior S/N, Ciudad Universitaria, 04510, Ciudad de México, Mexico. E-mail: isidrobadillor@gmail.com
bDivisión de Ciencias Básicas e Ingeniería, Departamento de Química, Universidad Autónoma Metropolitana-Iztapalapa, San Rafael Atlixco 186, Col. Vicentina, Iztapalapa, C. P. 09340, Ciudad de México, Mexico. E-mail: brunolanderos@hotmail.com
cUnidad Académica de Física de la Universidad Autónoma de Zacatecas, 98068 Zacatecas, Mexico

Received 20th July 2019 , Accepted 9th September 2019

First published on 9th September 2019


5-S-Cysteinyl-dopamine (CysDA) is a metabolite from dopamine oxidation found in patients with Parkinson's disease (PD), and has been suggested to be a biomarker for PD diagnosis. Therefore, the development of methodologies for its detection and identification is a matter of continuous interest for clinical neurology. Graphene oxide (GO) is an efficient 2D material known for its molecular adsorption capabilities, which also improves the spectroscopic signal and quenches the fluorescence of molecules adsorbed on it. For this reason, in the present work we examine the interaction of CysDA with GO. First, we demonstrate that Raman spectroscopy is a viable alternative analytical approach for the identification of CysDA, particulary due to the capacity of GO to quench the molecular fluorescence of the analyte. In order to enhance the interaction of CysDA with GO and to improve the detection of CysDA by vibrational spectroscopy, we studied the physicochemical processes involved in the adsorption of CysDA on GO through methods of Fourier transform infrared spectroscopy and quantum chemistry. These combined strategies revealed a wide variety of molecular interactions involved in the adsorption process. The stabilization of CysDA with GO is achieved primarily by hydrogen bonds and polarization effects, with a minor role of dispersion forces. In addition, we found that solvent molecules strengthens the adsorption of CysDA over GO. This work is the first reported strategy, as an alternative approach to the standard chromatographic methods, in the detection of CysDA, based on GO enhanced vibrational spectroscopy. The combination of Raman spectroscopy with graphenic substrates paves new ground for the future development of selective analytical platforms, which may ultimately enhance the detection of CysDA as a biomarker for PD.


Introduction

The 5-S-cysteinyl-dopamine (CysDA) molecule, which is shown in Fig. 1a, is a catechol-thioether metabolite formed from the nucleophilic addition of the L-cysteine amino acid to the dopamine–o-quinone intermediate, the first product in the spontaneous oxidation of dopamine (DA). Substantial experimental evidence has shown the toxicity of CysDA to dopaminergic neurons in the substantia nigra pars compacta (SNpc), which is the main region damaged in the physiopathology of Parkinson's disease (PD).1–3 In a recent report, we described the neurotoxic role of CysDA in cell neurodegeneration, which induces the pathogenesis and progress of PD.4 The CysDA metabolite has been detected in human brain tissues and cerebrospinal fluid, with particularly high concentrations in biological samples of PD patients.5–7 Therefore, the quantification of CysDA in biological fluids could be used as a biomarker to diagnose PD, as it has been proposed previously by some authors.4,8,9 Detection and quantification of CysDA in fluids of PD patients has been carried out by conventional electrochemical and chromatographic analytical techniques, such as high-performance liquid chromatography (HPLC).10,11 However, such analytical methods are expensive and require extensive time and complex preparation steps for analysis. In response, researchers are calling for the development of novel and complementary strategies to detect molecular biomarkers with similar HPLC sensitivity.
image file: c9nj03781k-f1.tif
Fig. 1 Lewis structures of 5-S-cysteinyl-dopamine (a) and graphene oxide (b); molecular models of CysDA (c) and GO (d), used in the theoretical calculations in this work.

Raman spectroscopy presents a viable alternative approach for analytical detection because it provides a quick-response, high sensitivity, label-free and susceptibility to molecular identification through the spectral fingerprint.12,13 However, in order to take advantage of the benefits of Raman spectroscopy, special attention has to be consider to increase the intrinsic low sensitivity of the Raman scattering and the interference of the fluorescence signal of biomolecules. For this purpose, the molecule under study is frequently adsorbed onto ad hoc nanostructured substrates, in order to enhance its Raman scattering cross-section and to quench the eventual fluorescence induced by the laser activation source of the Raman system. The scattering cross-section enhancement results in the increase of the Raman spectroscopy detection limit, even at the femtomolar level, greatly expanding the type and number of molecules that can be detected by this technique.14

The Surface Enhanced Raman Scattering (SERS) approach is widely used to increase the Raman scattering signal and to quench the fluorescence of the adsorbed molecule. SERS is based on the adsorption of biomolecules onto noble metal nanoparticle arrays.15–17 However, more recently, the adsorption of molecules onto graphenic substrates, in order to produce the Graphene Enhanced Raman Scattering (GERS) phenomena, is becoming an interesting alternative to SERS.18,19 Despite the fact that the scattering enhancement factor in GERS is lower than in SERS, the GERS approach offers important operational advantages. For example GERS produces lower spectral distortion and lower instability of the adsorbed molecules, which results in a more reliable detection.20

Graphene derived materials such as pristine graphene (PG), graphene oxide (GO) and reduced graphene oxide (rGO), have attracted attention to develop sensitive substrates. GERS supports have been designed for the detection of various molecules relevant for medical, biological, environmental and food research.21–25 GERS strategies have also proved successful for the detection of folic acid, a relevant biomarker of cancer. This has been accomplished through the use of a hybrid structure of graphene oxide with poly-diallyldimethyl ammonium chloride (GO-PDD) and silver nanoparticles (AgNP), leading to produce a linear response between 9 and 180 nM.26 By combining drop-coating deposition Raman (DCDR) spectroscopy and graphene-enhanced Raman spectroscopy (GERS), the detection of human recombinant interleukin-6 was achieved in an order lower than 1 pg mL−1.27 GO-nobel metal structures nanocomposites have also been designed for the detection of HIV-DNA in the order of fM concentration.28

The use of GO as substrate for GERS platforms has some advantages over pristine graphene, such as the ability of being dispersed in aqueous media or biological fluids. The presence of oxygenated functional groups over GO sheets, provide to it a wide possibility to form non-covalent interactions with the adsorbed biomolecule.29,30 In addition, a deliberate selection of the oxygenated species in GO should give rise to a modulation of the molecule–substrate interaction, where the Raman spectral amplification or the quenching effect could be maximized.31

In this work, the CysDA conjugate was adsorbed on a solid GO substrate, in order to quench the previously observed fluorescence signal of the molecule, while minimizing its chemical destabilization and spectral distortion. This strategy has allowed to obtain, for the first time, a clearly assignable Raman spectrum of the CysDA conjugate. In addition, with the purpose of optimizing the interaction of CysDA with GO, in a way to improve the detection of CysDA by Raman spectroscopy, the physicochemical processes involved in the adsorption of CysDA on GO was studied by Fourier transform infrared spectroscopy and quantum chemistry methods. The aim of the theoretical analysis is to study the role of the oxygen functional groups of GO over the CysDA adsorption. The forces involved in the interaction between CysDA and GO were analyzed through quantum chemistry methods, leading to determine the preferential sites of interaction. For this purpose, we used the quantum theory of atoms in molecules32 (QTAIM) and the non-covalent interaction index33 (NCI), which are ideal methods for this type of analysis.34,35 We employed density functional theory for the electronic structure computations due to the high number of atoms involved in the CysDA/GO complex, which forbid the use of correlation methods based on the wave-function. The use of QTAIM, NCI and DFT, along with other theoretical and experimental techniques, has been successfully applied to examine the interactions between a variety of biomolecules and some nanomaterials such as pristine, defected and oxidized graphene, as well as boron-nitride nanosheets.36–45 The findings presented in this work constitute a step towards the development of Raman based selective analytical platforms for the sensitive detection of CysDA as a PD biomarker.

Experimental

Materials

Dopamine hydrochloride, ceric ammonium nitrate, L-cysteine, Dowex 50Wx2, Dowex 50Wx8 and sulphuric acid were purchased from Sigma Aldrich (St. Louis, USA). Graphene oxide solution was obtained from Graphenea (San Sebastian, Spain).

Synthesis of CysDA

CysDA was synthesized with slightly modifications from the first experimental reports.46,47 In short, 20 millimoles of ceric ammonium nitrate (CAN) in 10 mL of H2SO4, 2 M were first added to 10 millimoles of dopamine hydrochloride, previously dissolved in 5 mL of H2SO4, 2 M. The mixture reaction was rapidly poured into a vigorously stirred solution of 20 millimoles of L-cysteine, dissolved in 5 mL of H2SO4, 2 M. A blue-green solution was obtained after the mixture, which was immediately passed through a column (2 × 40 cm) packed with Dowex 50Wx8 (H+ form), equilibrated with deionized water. The column was washed with water and then eluted with HCl, 0.5 M to eliminate unreacted dopamine, which was followed spectrophotometrically at λmax 280 nm. Then, elutions with HCl, 3 M followed by 4 M, gave a mixture of various cysteinyl-dopamine isomers: 2-S-cysteinyl-dopamine, 5-S-cysteinyl-dopamine and 2,5-S,S-dicysteinyl-dopamine. Eluted fractions from HCl 3 M and 4 M were combined and reduced to a small volume, and then passed through a column (2 × 40 cm) packed with Dowex 50Wx2, equilibrated with HCl 2 M. Further elutions were carried out with HCl, 2 M. Fractions were checked and followed spectrophotometrically to collect only those samples with the λmax at 254 nm and 292 nm, which is indicative that only 5-S-cysteinyl-dopamine was present. Finally, acidic solution was freeze-dried to obtain CysDA, as a pale brown residue. One part of the sample was stored at −20 °C to further spectroscopic characterization and the other part was dissolved in HCl, 10 mM for the GERS assay. The purity of the product was checked by reverse HPLC with spectrophotometric detection, showing only band absorptions at 254 nm and 292 nm, revealing no presence of unreacted dopamine or isomers (Fig. S1 in ESI). In addition, the structural identity of the molecule was confirmed by electrospray ion mass spectrometry, giving a 273 m/z molecular ion mass.48

GO characterization and preparation of GERS substrates

Commercial GO solution (4 mg mL−1) was first characterized by means of XPS, UV-vis, FTIR and Raman spectroscopies (Fig. S2–S4 in ESI). Spectral profiles are in good agreement with literature. From the aqueous commercial GO, a stock solution of 100 ppm was prepared in deionized water and sonicated for one hour, then used for the substrate preparation. A solid support of aluminum mirror was used for the deposition of GO. A drop of 10 μL of GO dispersion (100 ppm) was dropped over the metallic support and dried in a desiccator at room temperature. Finally, 3 μL of CysDA, at concentration of 0.1 mM, in acidic water, was dropped over the prepared GO/Al support and then dried under the same conditions. After the deposition of CysDA over the dried GO, the substrate was completely covered by the CysDA solution, without the formation of patterns similar to the typical “coffee ring”, which usually tends to concentrate the analyte in some specific regions. Therefore, we assumed that there was no excess of CysDA on the GO surface and the washing of the substrate was not a necessary requirement for Raman analysis.

Raman spectroscopy and GERS assays

Raman spectra measurements were recorded with a WITec alpha300 R (WITec GmbH, Ulm, Germany), using a 672 lines per mm grating and 532 nm laser light excitation, originated from a Nd:YVO4 green laser. The incident laser beam with a power of 3.56 mW was focused by a 100× objective, Zeiss, with 0.9 NA. First, in an attempt to obtain the normal Raman spectrum of CysDA, the freeze-dried compound was mounted over the aluminum mirror and several punctual spectra were recorded. However, sample hydration was very fast and an intense fluorescence signal covered the full Raman spectrum. In order to avoid that limitation, the prepared substrate of the CysDA/GO/Al system was analyzed. Five punctual Raman spectra were collected randomly in the middle zone of the dropped CysDA/GO complex. Spectra were acquired with 4 s integration time and 20 accumulations. Finally, spectra were cosmic ray removed, base line corrected and averaged for further band assignment.

Fourier transform infrared spectroscopy assay

Fourier Transform Infrared (FTIR) analysis were made in a Nicolet iS50R Thermo Scientific spectrometer, equipped with an attenuated total reflectance (ATR) diamond crystal. A small amount of the freeze-dried CysDA was spread over the diamond crystal and covered to avoid hydration. Spectra acquisition were collected with 32 scans, 4 cm−1 of spectral resolution, in the range of 500–3600 cm−1. GO characterization by FTIR was performed using a dried powder of GO, mounted over the diamond crystal and measured under the same aforementioned conditions. The study of the physicochemical interactions between GO and CysDA by means of FTIR spectroscopy was performed by mixing the dried powder of GO with the aqueous acidic CysDA to form a semi solid paste. The paste was spread over the diamond crystal and four IR spectra were collected under the same conditions. The working spectrum for each sample was the average of 4 recorded spectra.

Theoretical methods

In order to study the biomolecule/substrate interactions, as well as the effect of residual solvent molecules, 5-S-cysteinyl-dopamine/graphene oxide (CysDA/GO) complex was analyzed and classified in four groups: (1) [CysDA/GO]·10H2O (decahydrate), (2) [CysDA+/GO]·10H2O (decahydrate), (3) CysDA/GO, and (4) CysDA+/GO. The potential energy surfaces of 1 and 2 were explored with the simulated annealing method by the ASCEC code,49 computing the objective function with the PM7 method as implemented in MOPAC 2016.50,51 An open conformer of the CysDA molecule was used, avoiding the contact between the amino group of the dopamine and the carboxyl group of the cysteine (Fig. 1c). In view of the fact that the terminal carboxylic groups of the GO sheets have a strong influence on its physical and chemical properties,38,52 a finite model was chosen for this system (Fig. 1d). Emphasis was made on the effect of the oxygenated functional groups; three epoxide, five hydroxyl and six carboxyl units were randomly added to a 16-carbon ring polycyclic aromatic hydrocarbon, where only two aromatic rings were conserved. It has been shown that this type of representations of graphene derivatives simulate correctly their main features related to its absorption capabilities.39,41,44,53 The three lowest conformations of 1 and 2 were optimized at the B3LYP-D3/DZVP level of theory with the NWChem software.54 Once optimized, all water molecules were removed, and the systems were re-optimized with the same procedure, obtaining in this way minimum conformations for 3 and 4. We finally obtained twelve stable structures, three for each group. Additionally, interaction energies, along with their decomposition in electrostatic, polarization, dispersion and (Pauli) repulsion terms, of the water-free complexes were calculated using the CE-B3LYP/6-31G(d,p) approach with the Crystal Explorer package.55 Finally, a QTAIM and NCI analysis were performed with a grid-based algorithm accelerated by GPUs.56 Molecular graphs and NCI isosurfaces were drawn with VMD.57

Results and discussion

Raman spectra of CysDA adsorbed on GO substrate

The normal Raman spectrum of CysDA (see Fig. S5, ESI), was recorded by deposition of freeze-dried CysDA onto the aluminum support. The spectrum is mainly dominated by a broad band, from the intrinsic fluorescence of the molecule, which hides most of the expected Raman bands but revealing only one band around 1015 cm−1. In order to overcome this limitation, the prepared CysDA/GO/Al system was analyzed by Raman spectroscopy. The corresponding Raman spectrum of the CysDA/GO complex is presented in Fig. 2. The spectrum exhibits a well-defined and resolved set of vibrational bands, without the interference of fluorescence. This demonstrates the quenching efficiency of the fluorescence of CysDA, when deposited onto the GO substrate. It must be note here that, at the best of our knowledge, this is the first time that a Raman spectrum of the CysDA conjugate is shown in literature. The band assignment of the full Raman signals were done based on the expected vibrational modes in literature of the CysDA functional groups.58–60 Raman spectrum of CysDA/GO/Al system exhibits a broad but defined set of bands in the 3400–3300 cm−1 region, typically from the symmetric and asymmetric N–H stretching vibrational modes. Defined bands at 2966 cm−1 and 2895 cm−1 are originated from the contribution of aliphatic C–H and CH2 stretching modes. The strong and well defined band at 1605 cm−1 is originated from the C[double bond, length as m-dash]C aromatic ring stretching, with contribution of the protonated NH3+ asymmetric deformation. Region from 1500 cm−1 to 1200 cm−1 shows a broad and complex band, which is integrated by contribution in the deformation modes of NH3+, C–H and COOH, as well as the stretching modes of C–OH and C–N. The strongest and most conspicuous bands are in the 1150–1000 cm−1 interval, which is dominated by the characteristic 1096 cm−1 band, originated from the C[double bond, length as m-dash]C aromatic ring breathing mode of the DA moiety when it is attached to a sulfur atom.61 Bands in the 600–300 cm−1 interval are contributions from the O–C–O, C–N–N, C–C–S and aromatic ring deformations, as well as the characteristic C–S stretching vibrational mode. Detailed band assignment is presented in Table 1.
image file: c9nj03781k-f2.tif
Fig. 2 Raman spectrum of CysDA adsorbed onto the GO/Al support with wavenumber identification bands.
Table 1 Tentative band assignment of the wavenumber and vibrational Raman modes of the experimental CysDA/GO complex
Wavenumber (cm−1) Tentative band assignment
Abbreviations: str. (stretching), def. (deformation), bend. (bending), twist. (twisting), tors. (torsion), vibr. (vibration), sym. (symmetric) and asym. (asymmetric).
255 Ring def./C–S–C bend.
331 C–C–S bend.
377 Skeletal vibr./O–C–O in-plane bend/CH def.
456 C–O in-plane def./ring out-of-plane
571 CH in-plane ring def.
639 C–S str./CO tors./aliphatic chain C–C def.
724 CH out-of-plane/ring. def.
897 O–H out-of-plane def. vibr./carboxylic acids
997 C–N str./NH3+ twist/ring. def.
1096 C[double bond, length as m-dash]C–S ring breathing/C–O–C asym str.
1122 C–N str./NH twist./CH twist.
1295 CH twist./CH aromatic def./ring breathing
1336 OH scissoring/CH twisting/ring def.
1382 COOH bend/C–OH str.
1411 C–H def. due to sulfur/CH2 def./COH bend.
1480 NH3+ asym. def./CH scissoring
1602 C[double bond, length as m-dash]C str. ring def./OH scissoring
2895 C–H2 sym str.
2966 C–H2 asym str.
3303 N–H2 sym str./NH3+ sym str.
3345 N–H3+ sym str.
3375 N–H2 asym str.


Moreover, the quality of the Raman response of CysDA adsorbed over GO is reflected in the low contributions of the characteristic Raman bands of GO, D and G, at ∼1345 cm−1 and ∼1595 cm−1, respectively.31 Summarizing, a good quality Raman spectrum of CysDA is obtained when its acidic solution is deposited onto the GO/Al substrate, indicating that GO quenches the fluorescence of CysDA. This finding opens the possibility for further development of GO substrates, assisted Raman spectroscopy, for the detection and identification of this metabolite.

Nevertheless, to optimize the performance of the graphenic platform it is important to get additional information about the interaction of CysDA with GO. This information cannot be deduced solely from the data of the obtained experimental Raman spectrum. For this reason, we combined an experimental FTIR spectroscopic analysis with a detailed theoretical simulation of the CysDA/GO complex.

It should be noted that the chemical interactions between the CysDA and the GO are not the only parameters to be considered. In the present study, the surrounding aqueous and acidic media should also play an important role in the entire adsorption picture in the CysDA/GO system. The full environmental conditions are frequently omitted in literature, even when a detailed study by theoretical simulation of the system is performed.

Chemical adsorption of CysDA on GO by FTIR spectroscopy and quantum chemical models

Due to the fact that FTIR spectral bands are not compromised by fluorescence interferences of the molecule under study, as is the case for Raman spectroscopy, it was possible to obtain the separated infrared spectra of the pure CysDA and the GO and compare them with the spectrum of the CysDA conjugate adsorbed on GO. In this way the changes of the corresponding band array could be correlate with the molecule/substrate interactions.

Nevertheless, the interpretation of the changes observed in those IR spectra is not straightforward. GO is not a stoichiometric compound, their sheets have a wide distribution size and they have not a homogeneous distribution of the oxygenated functional groups. The CysDA/GO system is neither a regular molecular cluster nor a really periodic system. Therefore, the probability that the CysDA was identically adsorbed onto GO, in all the cases and with the same intermolecular interactions is small. Therefore, CysDA adsorption on GO sheets should occurs following different adsorption modes and patterns. Hence the importance of the theoretical model, in order to provide a deeper insight about the possible kind of intermolecular interactions involved in the adsorption phenomena of the CysDA/GO system.

Fig. 3a shows the standard infrared spectrum of the lyophilized CysDA. Again, from the best of our knowledge, this is the first time that an experimental IR spectrum of CysDA is presented in literature. The assignment of the spectral bands was based on the information of classical table charts of IR vibrational modes.58,62 The stretching vibrational modes of NH3+, OH, aromatic CH and aliphatic CH/CH2 appearing at 3480 cm−1, 3438 cm−1, 3214 cm−1, and 3054 cm−1, respectively. The characteristic C[double bond, length as m-dash]O stretching vibrational mode of the carboxylic acid is shown at 1731 cm−1. The double and intense bands at 1627 cm−1 and 1612 cm−1 are due to the C[double bond, length as m-dash]C stretching mode of the benzene ring when the sulfur atom is adjacent to the catechol OH group. Bands at 1488 cm−1 and 1348 cm−1 are due to the O–H deformation and O–H bending modes of diol groups in the DA moiety, respectively. Bands at 1211 cm−1 and 1133 cm−1 are associated with the N–H in-plane bending and C–N stretching, respectively.63 The fully detailed band assignment is presented in Table 2. These assignments are in agreement with the calculated vibrational frequencies of CysDA molecule, first presented by Bagheri, which was obtained by means of molecular simulation in gas phase.64


image file: c9nj03781k-f3.tif
Fig. 3 FITR spectra of (a) lyophilized CysDA, (b) GO and (c) CysDA/GO complex.
Table 2 Tentative band assignment of the main FTIR spectral bands for CysDA, GO and the CysDA/GO complex
Wavenumber (cm−1) Tentative band assignment
CysDA CysDA/GO GO
Abbreviations: str. (stretching), def. (deformation), bend. (bending), twist. (twisting), arom. (aromatic).
938 Aliphatic. C–C–N str./NH3+ def.
980 C–O str.
1041 C–O str. of alkoxy
1049 CH2 twist.
1069 C–O str.
1166 O–C(O)–C str. of epoxy
1144 CH arom. def.
1133 NH3+ def./C–H arom. def.
1188 C–O str.
1197 C–O–H out-of-plane bend.
1211 C–N str./CH2 def.
1222 C–O str. of epoxy/C[double bond, length as m-dash]O def.
1266 1266 CH2 def.
1287 C–O str.
1348 O–H def./CH2 def.
1388 O–H in-plane bend.
1411 1418 C–O str./O–H def./–CH2CO– def.
1434 NH3+ in-plane bend/C–H def.
1488 1509 O–H def.
1612 C[double bond, length as m-dash]C str. of phenolic ring
1625 1620 C[double bond, length as m-dash]C str. of arom.
1627 C[double bond, length as m-dash]C str. of phenolic ring
1666 C[double bond, length as m-dash]C str. arom./C[double bond, length as m-dash]O str.
1731 1730 1729 C[double bond, length as m-dash]O str.
3050 2925 CH2 str. of aliphatic
3182 O–H str.
3216 3264 NH3+ str.
3440 3360 3357 O–H str.
3481 N–H2 str.


Fig. 3b shows the experimental IR spectrum of GO. The strong and broad bands between 3000–3600 cm−1 are due to stretching of O–H groups in GO, attached both at the edges or at the basal plane, forming a complex array of hydrogen bonds among neighboring GO sheets and with the remaining water molecules. Intense band at 1729 cm−1 is due to the oxidized C[double bond, length as m-dash]O carboxyl groups at the edges, while the broad band at 1620 cm−1 is related to the C[double bond, length as m-dash]C stretching mode of the remaining aromatic rings of non-oxidized graphitic domains. The broad band at 1388 cm−1 corresponds to OH deformation groups located at the edges of GO sheets, and bands around 1287 cm−1 are assigned to C–O stretching modes of surface internal epoxy groups, while bands between 1150 cm−1 and 1250 cm−1 are associated with random epoxy groups close to the edges. Finally, strong bands in the 950–1050 cm−1 interval correspond to the C–O stretching modes of alkoxy groups.65–67 The GO model described in the Experimental section (Fig. 1d) reasonably includes all those aforementioned functional groups, thus the model can be considered a close representative of the GO sample used in this work.

As mentioned above, the analysis of the interactions exhibited in the CysDA/GO structures found by our simulation is a good starting point for understanding the CysDA/GO FTIR experiment. The two most stable DFT optimized geometries of the complexes of groups 1, 2, 3 and 4 are depicted in Fig. 4. From the analysis of these structures, the following general trends can be deduced: the general equilibrium geometries of the CysDA/GO/H2O system [groups 1 and 2] shows that water molecules tend to gather together in cages in the opposite side of GO, and only one or two of them interact with GO and CysDA simultaneously. This behavior has been found before in other hydrated complexes.68,69 However, when water molecules where removed [groups 3 and 4], the CysDA and GO molecules change their previous structures to get closer, indicating a considerable effect of the presence of solvent molecules in the corresponding conformation of the CysDA/GO complexes, particularly for those functional groups involved in the intermolecular hydrogen bonds formation. It is important to point out that a spontaneous proton transfer from a hydroxyl group of GO to the amine group of the dopamine moiety was found in the process of geometry optimization of the most stable structure of 1, indicating that carboxylic acids of GO and amine group in dopamine act as an acid–base conjugated par in this system. This result gives further support to the proposal that under the acidic conditions of the experimental study, the CysDA should be found mostly in its protonated conformation, which also corroborates the previous results described in the experimental Raman and in the FTIR spectra.


image file: c9nj03781k-f4.tif
Fig. 4 Lower energy configurations of 1, 2, 3 and 4. The most stable structures are in the right column. Energy differences (ΔE) between each configuration are reported in kcal mol−1.

Molecular graphs and NCI isosurfaces of the most stable structures of 1, 2, 3 and 4 are depicted in Fig. 5 and 6, respectively. Although the number and sort of interactions is variable in the twelve studied complexes, Fig. 5 and 6 are representative of the general tendencies. Several type of intermolecular contacts, related to the occurrence of bond critical points, were found and classified in two primary groups: hydrogen bonds (HB) and non-hydrogen bonds (NHB). The former can be subdivided in strong (O–H⋯O, O–H⋯N and N–H⋯O) and weak (C–H⋯O, C–H⋯C, O–H⋯C, O–H⋯S and N–H⋯C) hydrogen bonds.


image file: c9nj03781k-f5.tif
Fig. 5 NCI isosurfaces of the lowest energy conformers of 1, 2, 3 and 4 drawn at 0.5 a.u.

image file: c9nj03781k-f6.tif
Fig. 6 Molecular graphs of the lowest energy structures of 1, 2, 3 and 4. Critical and ring critical points are depicted in yellow and green colors, respectively. Bond paths are shown in orange.

Geometrical, topological and energetic criteria were used to classify the non-covalent interactions involved in all structures. In the NHB group, a variety of interactions, (whose origin and subclassification is still a matter of debate,70–79) were recognized (C⋯C, C⋯O, C⋯N, C⋯S, O⋯O, O⋯N, O⋯S and H⋯H). Owing to scarce aromatic character of the GO model used in this work, the C⋯C contacts cannot be considered as π⋯π interactions. Except for one, the O⋯S contact found in the group 2 structure, which is identified as a chalcogen bond, all the other NHB interactions possess a dispersive behavior, as can be noticed in the green extended irregular surfaces on Fig. 5. Therefore, the adsorption process is compromised between strong specific intermolecular interactions like HB, which usually have an electrostatic character, and NHB contacts related to dispersive effects.

This feasibility of forming HB in the CysDA/GO complexes causes significant modifications in the adsorption phenomenon with respect to processes encountered in biomolecule/graphene structures, where the adsorbates tend to be oriented parallel to the graphene layer and the interaction is predominantly dispersive.36–41 In contrast, in the 1, 2, 3 and 4 structures the CysDA molecule can be either parallel to the GO (distorted) plane or bent and attached to the edge of GO to form HB with the outer hydroxyl or carboxyl groups. With the purpose of quantifying the effect of the different non-covalent interactions in the CysDA/GO systems, the total interaction energies of the three most stable structures of 3 and 4 (named 3a–3c and 4a–4c, respectively), as well as the electrostatic, polarization, dispersion and repulsion components, were computed (see Table 3).

Table 3 Total interaction energies and their decomposition in electrostatic (Elec.), polarization (Pol.), dispersion (Disp.) and repulsion (Rep.) contributions of the three most stable complexes of 3 and 4. All values are in kcal mol−1
Complex Elec. Pol. Disp. Rep. Total
3a −102.2 −21.9 −37.7 47.6 −114.3
3b −44.4 −9.3 −24.5 32.9 −45.3
3c −22.6 −6.3 −31.5 24.4 −36.0
4a −66.5 −17.9 −36.3 52.8 −67.9
4b −65.1 −15.8 −23.1 38.3 −65.7
4c −56.0 −21.5 −30.6 44.4 −63.7


It is observed that the electrostatic contribution prevails over other energetic components whenever the CysDA bends to form HB. This term is particularly large for 3a due to the proton transfer. On the other hand, in the cases where the CysDA is extended over the GO surface, as in 3c, dispersion is the dominant effect. Polarization is more relevant for the protonated and the proton-transfer complexes, increasing the total interaction energies of these systems, in comparison with those of the neutral complexes. Although the repulsion follows the same trend that the total interaction energies, it does not seem to have a clear relation with the HB formation. The interaction energies for the protonated complexes (4a–c) are less sensitive to the geometrical structure.

From the preceding analysis it is clear that in most of the cases the predominant component of the interaction in the CysDA/GO complex is caused substantially by HB, which is in agreement with the experimental data.42 This fact is expected to influence the IR spectrum profile of the CysDA/GO complex with respect to the isolated moieties (Fig. 3c). Indeed, the comparison of the corresponding IR spectrum shows that in the CysDA/GO system the maximum of the broad band of hydroxyl groups is red-shifted from 3351 cm−1 in GO and 3438 cm−1 in the CysDA to 3256 cm−1 and 3369 cm−1, respectively. These changes are followed by the shifting of the OH bending vibrational band from 1348 cm−1 in CysDA to 1418 cm−1 in the CysDA/GO. In addition, the shifted and lower intensity of the band at 1488 cm−1 in CysDA to 1511 cm−1 in the complex, supports the strong interaction between OH diols of the DA moiety with superficial GO functional groups.80 Moreover, remarkable band shifting is observed in the 1000–1200 cm−1 region, coming from the contribution from superficial C–O stretching from alkoxy and epoxy groups in GO, as well as NH3+ deformation in CysDA. Those variations are consequence of the multiple hydrogen bonding interactions and also from the influence of dispersive interactions over the aforementioned superficial functional groups in GO, such as C⋯O, C⋯N, O⋯O, O⋯N, and O⋯S.

On the other hand, experimental characteristic bands of the C[double bond, length as m-dash]C benzene ring stretching at 1612 cm−1 and 1627 cm−1 in CysDA, suffer a strong shift to 1666 cm−1 in the complex, which is accompanied by a minor shift from 1620 cm−1 to 1625 cm−1 in GO. Based on the results of the theoretical calculations, these changes can be associated with the weak O–H⋯C hydrogen bond formed between the hydroxylated groups of GO with the aromatic ring of CysDA (O–H⋯π interactions), which are expected to contribute to electrostatic and polarization interactions, along with C⋯C contacts (some of them related to π⋯π interactions), which are governed by dispersion.

Finally, an additional way to examine the influence of each type of interaction is by analyzing the values of the electron density at the bond critical points (ρBCP), as well as the application of the Espinosa–Molins–Lecomte (EML) approximation to estimate the relative strength of each hydrogen bond.81 As can be seen in Table S1 (ESI), the ρBCP is in most of the cases one order of magnitude larger for the HB, indicating that those prevail over NHB in the CysDA/GO complex stabilization, whether or not the CysDA is protonated. For the interaction energies calculated via the EML equation, a decreasing monotonic relation with ρBCP is observed (see Fig. 7). It can also be noted that some O–H⋯S contacts can be considered strong HB according to the energetic criteria, reaching up to −4.8 kcal mol−1. Besides, some HB, particularly N–H⋯O contacts, can reach values as large as −25.7 kcal mol−1, and can be separated into a very strong HB group. This result could explain the above-mentioned strong shifting in the 1300–1100 cm−1 IR region of the CysDA/GO complex, influenced by N–H in-plane bending and C–N stretching modes. The aforementioned outcome is also independent of the protonation state of CysDA. From Fig. 7 it is also clear that in the hydrated complexes the HB interaction, O–H⋯O, is enhanced.


image file: c9nj03781k-f7.tif
Fig. 7 Interaction energy (calculated via the EML equation) vs. ρBCP of the hydrated systems (top) and non-hydrated (bottom) systems.

Lastly, the classification of HB using the geometric criteria, which consists on analyzing the H⋯Y and X⋯Y distances (dH⋯Y and dX⋯Y,82 respectively, where X is the atom to which H is covalently bonded and Y is the atom forming the HB), as well as the X–H⋯Y angles (θ), reveals that the amount of strong HB contacts is slightly lower than that of the weak HB for the dehydrated complexes, but is larger for the former in the presence of water molecules. This result evidences an additional stabilization of solvent molecules in the CysDA/GO complexes (Fig. S6, ESI). Those results, in agreement with the energetic and the FTIR analysis, allow us to conclude that hydrogen bonding between the CysDA and GO, either in the surface or on its edge, is the most important interaction contributing to the adsorption process in the CysDA/GO system.

In agreement with the aforementioned findings, there are some previous experimental studies of the interaction between GO and dopamine (DA) in aqueous media, concluding that DA strongly adsorbs on GO mainly by non-covalent interactions. Even the interaction between DA and GO was found stronger in comparison to that of rhodamine6 green (R6G) over GO, analyzed in the same study.83 From this previous report, and taking into account that the CysDA molecule contains a DA moiety, a similar adsorption mechanism should be expected for CysDA over GO. The present work gives further elucidation of the adsorption process, which provides information for further design of a novel biosensor development, based on the detection of the CysDA metabolite.

Conclusions

In this work we demonstrate that using graphene oxide (GO) as a substrate is a good strategy for the detection and identification of CysDA by Raman spectroscopy. The interaction of CysDA with GO leads to the quenching of the fluorescence of CysDA. This strategy lead to obtain, for the first time, the Raman spectrum of CysDA with well-resolved bands that can be fully assigned. The assignment of these bands is consistent with the molecular structure of CysDA, showing that the interaction with GO does not result in a noticeable perturbation of its normal vibration modes. The FTIR experimental spectra of CysDA, GO and CysDA/GO, combined with quantum chemistry methods provide a way to understand the non-covalent interactions that are responsible for the adsorption of CysDA over GO. A wide variety of intermolecular interactions were found in the CysDA/GO system, classified as follow: hydrogen bonds (O–H⋯O, O–H⋯N, N–H⋯O, C–H⋯O, C–H⋯C, O, –H⋯C, O–H⋯S and N–H⋯C) and non-hydrogen bonds (C⋯C, C⋯O, C⋯N, C⋯S, O⋯O, O⋯N, O⋯S and H⋯H). Hydrogen bonds stabilize CysDA over GO by electrostatic and polarization effects, while the non-hydrogen bonds have a dispersive origin. The two types of interactions allow us to explain the main band shifting observed in the FTIR spectrum of the CysDA/GO complex, especially for the OH bending vibrational bands and the C[double bond, length as m-dash]C aromatic stretching modes of the CysDA ring. Furthermore, solvent molecules enhance the adsorption of CysDA over GO. The strategy used in this work could be applied for a further development of substrates based on GO-assisted Raman spectroscopy for the detection of CysDA, as a biomarker for PD. Moreover, considering the findings in this work, we propose that the functionalization of the GO substrate with amine groups will further enhance the CysDA adsorption. Our group will continue to test this hypothesis, in the near future, with additional experimental and theoretical analyses.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Authors thank to the financial support to UNAM trough the IN-111216 and IN-102917 PAPIIT-UNAM research projects. Financial support was also provided by Consejo Nacional de Ciencia y Tecnología (CONACYT) through the project CONACYT 2014-Fronteras de la Ciencia-2016. Isidro Badillo-Ramírez is a doctoral student from the Programa de Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México (UNAM) and was supported by CONACYT (grant number 599497). EOC thanks to the support of CONACYT (grant number 619615). IBR, JMS and EOC, thank to the Laboratorio Universitario de Caracterización Espectroscópica, LUCE-ICAT-UNAM for the spectroscopic characterization of the samples and to Dr Selene R. Islas for the technical support. BLR, RV and JG thank to the facilities provided by the Laboratorio de Supercómputo y Visualización en Paralelo at the Universidad Autónoma Metropolitana-Iztapalapa. BLR thanks to UAM for a posdoctoral fellowship.

Notes and references

  1. F. Zhang and G. Dryhurst, J. Med. Chem., 1994, 37, 1084 CrossRef CAS PubMed.
  2. C. Aureli, T. Cassano, A. Masci, A. Francioso, S. Martire, A. Cocciolo, S. Chichiarelli, A. Romano, S. Gaetani, P. Mancini, M. Fontana, M. dÉrme and L. Mosca, J. Neurosci. Res., 2014, 92, 347 CrossRef CAS PubMed.
  3. J. P. E. Spencer, P. Jenner, S. E. Daniel, A. J. Lees, D. C. Marsden and B. Halliwell, J. Neurochem., 1998, 71, 2112 CrossRef CAS PubMed.
  4. I. Badillo-Ramírez, J. M. Saniger and S. Rivas-Arancibia, Neurochem. Int., 2019, 129, 104514 CrossRef PubMed.
  5. E. Rosengren, E. Linder-Eliasson and A. Carlsson, J. Neural Transm., 1985, 63, 247 CrossRef CAS PubMed.
  6. B. Fornstedt, A. Brun, E. Rosengren and A. Carlsson, J. Neural Transm., 1989, 104, 931 Search PubMed.
  7. S. Ito, T. Kato and K. Fujita, Chromatographia, 1986, 21, 645 CrossRef CAS.
  8. F. C. Cheng, J. S. Kuo, L. G. Chia and G. Dryhurst, J. Neural Transm., 1996, 103, 433 CrossRef CAS PubMed.
  9. A. Carlsson and B. Fornstedt, Neurology, 1991, 41, 50 CrossRef CAS PubMed.
  10. D. Vauzour, J. T. Pinto, A. J. Cooper and J. P. E. Spencer, Biochem. J., 2014, 463, 41 CrossRef CAS PubMed.
  11. D. S. Goldstein, C. Holmes, P. Sullivan, Y. Jinsmaa, I. J. Kopin and Y. Sharabi, Parkinsonism Relat. Disord., 2016, 31, 79 CrossRef PubMed.
  12. D. I. Ellis, D. P. Cowcher, L. Ashton, S. O’Hagan and R. Goodacre, Analyst, 2013, 138, 3871 RSC.
  13. C. Krafft and J. Popp, Anal. Bioanal. Chem., 2015, 407, 699 CrossRef CAS PubMed.
  14. H. J. Butler, S. W. Fogarty, J. G. Kerns, P. L. Martin-Hirsch, N. J. Fullwood and F. L. Martin, Analyst, 2015, 140, 3090 RSC.
  15. F. Reil, U. Hohenester, J. R. Krenn and A. Leitner, Nano Lett., 2008, 8, 4128 CrossRef CAS PubMed.
  16. T. J. Moore, A. S. Moody, T. D. Payne, G. M. Sarabia, A. R. Daniel and B. Sharma, Biosensors, 2018, 8, 1 CrossRef PubMed.
  17. J. A. Dieringer, A. D. McFarland, N. C. Shah, D. A. Stuart, A. V. Whitney, C. R. Yonzon, M. A. Young, X. Y. Zhang and R. P. Van Duyne, Faraday Discuss., 2006, 132, 9 RSC.
  18. X. Ling and J. Zhang, Small, 2010, 6, 2020 CrossRef CAS PubMed.
  19. S. Huang, X. Ling, L. Liang, Y. Song, W. Fang, J. Zhang, J. Kong, V. Meunier and M. S. Dresselhaus, Nano Lett., 2015, 15, 2892 CrossRef CAS PubMed.
  20. P. G. Vianna, D. Grasseschi, G. K. B. Costa, I. C. S. Carvalho, S. H. Domingues, J. Fontana and C. J. S. De Matos, ACS Photonics, 2016, 3, 1027 CrossRef CAS.
  21. N. Zhang, L. Tong and J. Zhang, Chem. Mater., 2016, 28, 6426 CrossRef CAS.
  22. Y. Li, L. L. Qu, D. W. Li, Q. X. Song, F. Fathi and Y. T. Long, Biosens. Bioelectron., 2013, 43, 94 CrossRef CAS PubMed.
  23. Y. Zhang, S. Liu, L. Wang, X. Qin, J. Tian, W. Lu, G. Chang and X. Sun, RSC Adv., 2012, 2, 538 RSC.
  24. Y. Xie, Y. Li, L. Niu, H. Wang, H. Qian and W. Yao, Talanta, 2012, 100, 32 CrossRef CAS PubMed.
  25. J. Lee, J. Kim, S. Kim and D. H. Min, Adv. Drug Delivery Rev., 2016, 105, 275 CrossRef CAS PubMed.
  26. W. Ren, Y. Fang and E. Wang, ACS Nano, 2011, 5, 6425 CrossRef CAS PubMed.
  27. E. de la O-Cuevas, I. Badillo-Ramírez, S. R. Islas, C. Araujo-Andrade and J. M. Saniger, RSC Adv., 2019, 9, 12269 RSC.
  28. Z. Fan, R. Kanchanapally and P. C. Ray, J. Phys. Chem. Lett., 2013, 4, 3813 CrossRef CAS.
  29. B. Gupta, N. Kumar, K. Panda, V. Kanan and S. Joshi, Sci. Rep., 2017, 7, 1 CrossRef CAS PubMed.
  30. J. Dai, G. Wang, L. Ma and C. Wu, J. Mater. Sci., 2015, 50, 3895 CrossRef CAS.
  31. H. Yang, H. Hu, Z. Ni, C. K. Poh, C. Cong, J. Lin and T. Yu, Carbon, 2013, 62, 422 CrossRef CAS.
  32. R. F. W. Bader, Atoms in Molecules: A Quantum Theory, Clarendon Press, Oxford, 1990 Search PubMed.
  33. E. R. Johnson, S. Keinan, P. Mori-Sanchez, J. Contreras-Garcia, A. J. Cohen and W. Yang, J. Am. Chem. Soc., 2010, 132, 6498 CrossRef CAS PubMed.
  34. B. G. Oliveira, Struct. Chem., 2014, 25, 745 CrossRef CAS.
  35. R. Chaudret, B. de Courcy, J. Contreras-García, E. Gloaguen, A. Zehnacker-Rentien, M. Mons and J.-P. Piquemal, Phys. Chem. Chem. Phys., 2014, 16, 9876 RSC.
  36. H. T. Larijani, M. D. Ganji and M. Jahanshahi, RSC Adv., 2015, 5, 92843 RSC.
  37. D. L. Carballeira, N. Ramos-Berdullas, I. Pérez-Juste, J. L. C. Fajín, M. N. D. Cordeiro and M. Mandado, Phys. Chem. Chem. Phys., 2016, 18, 15312 RSC.
  38. S. J. Rodríguez, L. Makinistian and E. A. Albanesi, J. Phys.: Conf. Ser., 2016, 705, 012012 CrossRef.
  39. P. Singla, M. Riyaz, S. Singhal and N. Goel, Phys. Chem. Chem. Phys., 2016, 18, 5597 RSC.
  40. H. T. Larijani, M. Jahanshahi, M. D. Ganji and M. H. Kiani, Phys. Chem. Chem. Phys., 2017, 19, 1896 RSC.
  41. R. Zhiani, Appl. Surf. Sci., 2017, 409, 35 CrossRef CAS.
  42. S. Pandit and M. De, J. Phys. Chem. C, 2017, 121, 600 CrossRef CAS.
  43. D. Zhao, L. Li and J. Zhou, Appl. Surf. Sci., 2018, 428, 825 CrossRef CAS.
  44. J. Ebrahimi, M. G. Ahangari and M. Jahanshahi, Appl. Surf. Sci., 2018, 440, 778 CrossRef CAS.
  45. S. J. Rodríguez and E. A. Albanesi, Phys. Chem. Chem. Phys., 2019, 21, 597–606 RSC.
  46. L. Mosca, I. Tempera, E. Lendaro, L. Di Francesco and M. D’Erme, J. Neurosci. Res., 2008, 86, 954 CrossRef CAS PubMed.
  47. F. Chioccara and E. Novellino, Synth. Commun., 1986, 16, 967 CrossRef CAS.
  48. F. Zhang and G. Dryhurst, J. Med. Chem., 1994, 37, 1084 CrossRef CAS PubMed.
  49. J. J. P. Stewart, J. Mol. Model., 2013, 19, 1 CrossRef CAS PubMed.
  50. J. J. P. Stewart, J. Comput.-Aided Mol. Des., 1990, 4, 1 CrossRef CAS PubMed.
  51. MOPAC, Stewart Computational Chemistry, Colorado Springs, CO, USA, https://OpenMOPAC.net Search PubMed.
  52. B. Konkena and S. Vasudevan, J. Phys. Chem. Lett., 2012, 3, 867 CrossRef CAS PubMed.
  53. Y. Ai, Y. Liu, W. Lan, J. Jin, J. Xing, Y. Zou, C. Zhao and X. Wang, Chem. Eng. J., 2018, 343, 460 CrossRef CAS.
  54. M. Valiev, E. J. Bylaska, N. Govind, K. Kowalski, T. P. Straatsma, H. J. J. van Dam, D. Wang, J. Nieplocha, E. Apra, T. L. Windus and W. A. de Jong, Comput. Phys. Commun., 2010, 181, 1477 CrossRef CAS.
  55. M. J. Turner, J. J. McKinnon, S. K. Wolff, D. J. Grimwood, P. R. Spackman, D. Jayatilaka and M. A. Spackman, CrystalExplorer17, University of Western Australia, 2017, http://hirshfeldsurface.net Search PubMed.
  56. R. Hernández-Esparza, S. Mejía-Chica, A. D. Zapata-Escobar, A. Guevara-García, A. Martínez-Melchor, J. M. Hernández-Pérez, R. Vargas and J. Garza, J. Comput. Chem., 2014, 35, 2272 CrossRef PubMed.
  57. W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics, 1996, 14, 33 CrossRef CAS PubMed.
  58. D. Lin-Vien, The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, New York, 1991, p. 4225 Search PubMed.
  59. G. Varsanyi, Assignments for Vibrational Spectra of Seven Hundred Benzene Derivatives, Academia Kiado, Budapest, 1974 Search PubMed.
  60. S. S. Yi, K. Kim and M. S. Kim, J. Raman Spectrosc., 1993, 24, 213 CrossRef CAS.
  61. Y. J. Kwon, S. B. Lee, K. Kim and M. S. Kim, J. Mol. Struct., 1994, 318, 25 CrossRef CAS.
  62. L. J. Bellamy, The Infrared Spectra of Complex Molecules, Springer, Netherlands, 1975 Search PubMed.
  63. A. Lagutschenkov, J. Langer, G. Berden, J. Oomens and O. Dopfer, Phys. Chem. Chem. Phys., 2011, 13, 2815 RSC.
  64. A. Bagheri Gh, Bull. Chem. Soc. Ethiop., 2010, 24, 227 CrossRef.
  65. G. X. Wang, B. Wang, J. Park, J. Yang, X. P. Shen and J. Yao, Carbon, 2009, 47, 68 CrossRef CAS.
  66. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806 CrossRef CAS PubMed.
  67. J. I. Paredes, S. Villar-Rodil, A. Martínez-Alonso and J. M. D. Tascón, Langmuir, 2008, 24, 10560 CrossRef CAS PubMed.
  68. J.-R. Salazar-Cano, A. Guevara-García, R. Vargas, A. Restrepo and J. Garza, Phys. Chem. Chem. Phys., 2016, 18, 23508 RSC.
  69. E. García-Hernández, R. Flores-Moreno, Á. Vázquez-Mayagoitia, R. Vargas and J. Garza, New J. Chem., 2017, 41, 965 RSC.
  70. A. Sirohiwal, V. R. Hathwar, D. Dey, R. Regunathan and D. Chopra, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2017, 73, 140–152 CrossRef CAS PubMed.
  71. A. M. Koohi, Z. Mahdavifar and S. Noorizadeh, ChemistrySelect, 2017, 2, 2713 CrossRef CAS.
  72. D. J. Pascoe, K. B. Ling and S. L. Cockroft, J. Am. Chem. Soc., 2017, 139, 15160 CrossRef CAS PubMed.
  73. K. T. Mahmudov, M. N. Kopylovich, M. F. C. G. da Silva and A. J. Pombeiro, Dalton Trans., 2017, 46, 10121 RSC.
  74. S. Sarkar, M. S. Pavan and T. G. Row, Phys. Chem. Chem. Phys., 2015, 17, 2330 RSC.
  75. A. Bauzá, T. J. Mooibroek and A. Frontera, Angew. Chem., Int. Ed., 2013, 52, 12317 CrossRef PubMed.
  76. A. Bauzá, T. J. Mooibroek and A. Frontera, Chem. Rec., 2016, 16, 473 CrossRef PubMed.
  77. A. J. Edwards, C. F. Mackenzie, P. R. Spackman, D. Jayatilaka and M. A. Spackman, Faraday Discuss., 2017, 203, 93 RSC.
  78. G. Cavallo, P. Metrangolo, T. Pilati, G. Resnati and G. Terraneo, Cryst. Growth Des., 2014, 14, 2697 CrossRef CAS.
  79. J. S. Murray, P. Lane and P. Politzer, J. Mol. Model., 2009, 15, 723 CrossRef CAS PubMed.
  80. J. L. Chen, X. P. Yan, K. Meng and S. F. Wang, Anal. Chem., 2011, 83, 8787 CrossRef CAS PubMed.
  81. E. Espinosa, E. Molins and C. Lecomte, Chem. Phys. Lett., 1998, 285, 170 CrossRef CAS.
  82. G. R. Desiraju and S. Steiner, The Weak Hydrogen Bond in Structural Chemistry and Biology, International Union of Crystallography Monographs on Crystallography, 2001 Search PubMed.
  83. H. Ren, D. D. Kulkarni, R. Kodiyath, W. Xu, I. Choi and V. V. Tsukruk, ACS Appl. Mater. Interfaces, 2014, 6, 2459 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available: Experimental methods and results; molecular structures and topological data. See DOI: 10.1039/c9nj03781k

This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019