Experimental FTIR and theoretical studies of gallic acid–acetonitrile clusters

Spectrochimica Acta Part A 86 (2012) 93–100

Contents lists available at SciVerse ScienceDirect

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa

Experimental FTIR and theoretical studies of gallic acid–acetonitrile clusters Namon Hirun a , Supaporn Dokmaisrijan b , Vimon Tantishaiyakul a,∗ a Drug Delivery System Excellence Center and Department of Pharmaceutical Chemistry, Faculty of Pharmaceutical Sciences, Prince of Songkla University, Hat-Yai, Songkla 90112, Thailand b School of Science, Walailak University, Tha Sala, Nakhon Si Thammarat 80161, Thailand

a r t i c l e

i n f o

Article history: Received 4 July 2011 Received in revised form 4 October 2011 Accepted 6 October 2011 Keywords: Gallic acid Acetonitrile FTIR DFT

a b s t r a c t Gallic acid (3,4,5-trihydroxybenzoic acid, GA) has many possible conformers depending on the orientations of its three OH and COOH groups. The biological activity of polyphenolic compounds has been demonstrated to depend on their conformational characteristics. Therefore, experimental FTIR and theoretical studies of the GA–solvent clusters were performed to investigate the possible most favored conformation of GA. Acetonitrile (ACN) was selected as the solvent since its spectrum did not interfere with the OH stretching bands of GA. Also of importance was that these OH groups, in addition to the carboxyl group, of the GA are the most likely groups to interact with receptors. The solution of GA in the ACN solution was measured and the complex OH bands were deconvoluted to four component bands. These component bands corresponded to the three OH bands on the benzene ring and a broad band which is a combination band of mainly the OH of the COOH group and the inter- and intramolecular H-bonds from the OH groups on the ring. The conformations, relative stabilities and vibrational analysis of the GA monomers and the GA–ACN clusters were investigated using the B3LYP/6-311++G(2d,2p) method. Conformational analysis of the GA monomer yielded four most possible conformers, GA-I, GA-II, GA-III and GA-IV. These conformers were subsequently used for the study of the GA:ACN clusters at the 1:1, 1:2 and 1:4 mole ratios. The IR spectra of the most stable structures of these clusters were simulated and the vibrational wavenumbers of the OH and C O groups were compared with those from the experiment. The FTIR component bands were comparable to the computed OH bands of the GA-I–(ACN)2 , GA-IV–(ACN)2 and GA-I–(ACN)4 clusters. Furthermore, the C O stretching bands and the bands in the regions of 1800–1000 cm−1 obtained by computing and the experiment were similar for these clusters. Thus, GA-I and GA-IV are the most preferable conformations of GA in ACN and perhaps in the polar environment around the receptor sites of GA. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Gallic acid (3,4,5-trihydroxybenzoic acid, GA) is one of the more interesting pharmaceutical components of plants with antioxidant, antimutagenic, anticarcinogenic, antihyperglycemic activities and a cardioprotective effect [1–3]. The biological activity of polyphenolic compounds, including GA has been shown to be highly dependent on their conformational characteristics [4]. Therefore, determination of their conformations is important to achieve a better understanding of their biochemical functions and in particular that of GA. At present four crystal structure confirmations for the GA monohydrate have been reported [5–7] and one anhydrous crystal conformation of GA has been prepared and investigated by our

∗ Corresponding author. Tel.: +66 74288864; fax: +66 74428239. E-mail addresses: [email protected] (N. Hirun), [email protected] (S. Dokmaisrijan), [email protected] (V. Tantishaiyakul). 1386-1425/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.saa.2011.10.009

group [8]. Due to the GA molecule containing three hydroxyl (OH) groups and a carboxyl (COOH) group, it can form both intra and intermolecular hydrogen bonds (H-bonds). Each GA polymorph has a different arrangement and conformation due to the distinct intra- and intermolecular H-bonds of GA and that of GA with water molecules for the monohydrate forms. Two main different orientations of the hydrogen atoms of the hydroxyl groups have been reported, as indicated by their crystal structures or their conformational analysis by quantum mechanical studies. According to Okabe et al. [6] and Cappelli et al. [9], all hydrogen atoms of the hydroxyl groups around the benzene ring are oriented in the same direction, and form two intramolecular H-bonds between the pair of hydroxyl groups at positions 3 and 4, and at positions 4 and 5 (Fig. 1). This orientation was also observed for the crystal structure of the anhydrous GA [8]. In contrast, Jiang et al. [5], Mohammed-Ziegler and Billes [10] and Billes et al. [7] described the reverse orientation of one hydrogen atom of the hydroxyl groups compared to the others. Consequently, in their structures, there is only one intramolecular H-bond available for

94

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100

Fig. 1. Molecular structure of GA with their atom numbering, together with a depiction of five specified torsions.

these conformations of GA. Furthermore, different orientations of the carboxyl group relative to the direction of the hydroxyl groups have also been observed in these reports. The conformations of various monomeric forms [9,10] and a dimeric form [7] of GA were previously investigated using molecular dynamic (MD) simulation [9], quantum mechanics using the B3P86/6-311G(d,p) calculation [10], and the Becke three parameter Lee–Yang–Parr functional (B3LYP) method with the use of the 6-31G* [7] or 6-311++G** [9] basis set. It should be noted that the minimum energy conformations reported in these studies are different due to the diverse group orientations of the starting conformations of GA. In some studies, the calculated infrared (IR) spectra of the minimum energy conformations were compared to the experimental spectra measured from the samples prepared in a KBr matrix. The calculated and the experimental wavenumbers showed some dissimilarities. Cappelli et al. [9] measured the ultraviolet (UV) spectra of GA in water and in acetonitrile (ACN), the spectra in both solvents were similar. The calculated and the experimental UV spectra were compared. Nevertheless, the comparison of the vibrational (IR) spectrum of GA in ACN was not investigated. In general, an IR spectrum will enable the showing of a large number of bands compared to the UV spectra and provides much more detailed information in terms of the intra- and intermolecular interactions of the specific groups. The FTIR spectrometer is a single-beam instrument, a background spectrum (air or solvent) should be first obtained before producing the sample spectrum. The ratio of the sample spectrum against the background will provide a result that looks like a double-beam spectrum of the sample. For the IR experiment, the solvents such as water and alcohol can strongly absorb IR. The spectra of these background solvents cannot be efficiently ratioed or subtracted out from the sample solutions especially at the region of the OH stretching bands. Usually, the background solvent should be relatively transparent in the spectral region of interest. ACN was selected as the solvent in this study since ACN does not show OH stretching bands. The OH bands on the benzene ring or the OH band of the carboxyl bands of GA will not be interfered with by the solvent. The incomplete ratio out of the background solvent may not affect the analysis of these groups which are most likely the significant groups, besides the carbonyl (C O) group, that interact with H-bond donors and acceptors with the receptors. ACN which is a polar solvent can interact intermolecularly with GA. This will possibly mimic the situation when GA interacts with its receptors in their relatively polar environment. For the theoretical study, the minimum energy conformers of the GA molecules with different orientations of the hydroxyl and the carboxyl groups in the implicit ACN solvent were searched, and subsequently used to form clusters with the ACN molecules. In this study, the isolated GA monomer included with the continuum solvent were not used for the IR prediction, since the continuum solvent has a less significant effect on the vibrational spectra. Thus, the lowest energy structures of the GA–(ACN)n clusters were used to predict the IR spectra of GA. In addition, the explicit solvent was

used in the models, because it has a significant effect on the GA vibrational spectra as shown by previous investigations [7,11]. The extent of ACN that interacts with GA has been previously investigated based on MD simulations [9]. This study reported only the radial distribution functions (RDFs) and corresponding coordination number (N(R)), which indicated the number of the ACN molecules at the distances from the GA atoms relative to the bulk density. However, the RDFs and their N(R) from this study cannot give information about the positions and orientations of ACN in the first solvation shell of the GA molecule. According to the RDFs of H(GA)· · ·N(ACN), the 1st maximum and minimum peaks for both the H10 and H14 atoms of GA (atom position on the structure is shown in Fig. 1) appear at 1.95 and ˚ respectively, but those for both H16 and H18 appear at 2.15 2.75 A, ˚ respectively. The N(R) value of the 1st solvation shell and 3.35 A, of each H atom of the OH group is equal to 1.05. This indicated that one ACN molecule was close to each of the H10 and H14 atoms but distant from each of the H16 and H18 atoms of GA. This study indicated that there are about 14 ACN molecules found in the first solvation shell of GA but the ACN residence times around the GA oxygen atoms were very low. In addition, the intramolecular H-bonds of GA were held and hindered the intermolecular H-bond formed between GA and ACN. It should be noted that a large amount of solvent in a model can overestimate the H-bond effects on the molecular moiety as previously reported [9]. In this study, GA–(ACN)n clusters where n is a small number of 1, 2 and 4 were employed. The comparison of the experimental and theoretical IR spectra especially at the OH and C O regions was investigated. 2. Materials and methods 2.1. Experimental GA was obtained from Fluka Chemie GmbH (Buchs, Switzerland). FTIR spectra were measured using a PerkinElmer Spectrum One FTIR spectrometer (Perkin-Elmer, Waltham, MA, USA). The preparation of the anhydrous form has been reported elsewhere [8]. For preparation of the KBr pellet, GA was ground with KBr in an agate mortar to a fine powder and then pressed into a disc under high pressure. The spectra were recorded from 4400 to 450 cm−1 at a 4 cm−1 resolution. For the solution of GA (107 mM) in ACN, the samples were injected into the FTIR liquid cell using the CaF2 windows. The pure solvent was recorded as a background. The spectrum obtained was the ratio of the single beam sample spectrum and the single beam background spectrum. Deconvolution of the IR spectra of GA in ACN at the OH regions was performed using GRAMS/AI (7.01) software (Thermo Galactic, Salem, NH, USA) by fitting the spectra with a Gaussian function. 2.2. Computational details All theoretical computations in this study were carried out using the Gaussian 03W program [12]. Initially, conformational analysis of the GA monomer was performed at the B3LYP/6-311++G(2d,2p) calculation [13–15]. All plausible conformers of GA were generated from the rotations of five specified torsions shown in Fig. 1. Each of these torsions was rotated approximately 0 and 180◦ . The possible conformers of GA with relative energies of less than 6 kcal mol−1 above the global minima were kept and reoptimized in the implicit ACN solvent using the same method. The SCRF-PCM model was applied to include the solvent effects on the structural stability. This model was tested to be reliable as previously described [9]. The four lowest-lying energy conformers of GA in the implicit ACN solvent (Fig. 2) were obtained and used for subsequent studies of the GA–(ACN)n clusters.

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100

95

Fig. 2. The four most stable conformers of GA in the implicit ACN solvent obtained from optimization at the B3LYP/6-311++G(2d,2p)/SCRF level of theory.

The structures of GA–(ACN)n clusters, where n = 1, 2 and 4, were generated and fully optimized in the implicit ACN solvent at the B3LYP/6-311++G(2d,2p) calculation. For the GA–(ACN)1 clusters, an ACN molecule was assigned to form a H-bond with the H10 of GA since in the closest contact of each OH group of GA, an ACN molecule can form a H-bond with H10 that is stronger than the other H atoms of GA [9]. The structures of the GA–(ACN)2 clusters were built by adding two ACN molecules to form a H-bond with GA at H10 and either H14 or H18 (Fig. 1) which is more accessible to the solvent than H16 . The structures of the GA–(ACN)4 clusters were generated by introducing four ACN molecules to form H-bonds with all the OH groups of GA. The optimized structures of the GA–(ACN)n clusters in each mole ratio were subsequently used for vibrational analysis. The vibrational calculations of these clusters with and without the inclusion of the continuum solvent were performed at the B3LYP/ 6-311++G(2d,2p) calculation in order to investigate the effects of the continuum solvent on the vibrational spectra. The normal modes of vibration were graphically displayed using the GaussView program. The theoretical vibrational wavenumbers of the GA–ACN clusters were scaled as previously described [7,10] to fit to the experimental wavenumbers, the OH and the C O groups were scaled by 0.99 and 0.999, respectively.

conformers to the complex experimental bands was also performed using the same approach [18]. The analysis of the three OH bands, observed at 3630, 3536 and 3350 cm−1 for the GA solution was thus investigated by this curve fitting technique. Deconvolution of these peaks gives four component bands at 3647, 3607, 3540 and 3353 cm−1 (Fig. 4). The three highest wavenumbers at 3647, 3607 and 3540 cm−1 may be assigned to the OH stretching bands of the hydroxyl groups on the benzene ring which are normally higher than the OH band from the acidic carboxyl group (COOH) [19]. The lowest wavenumber at 3353 cm−1 can be assigned to the OH of the carboxyl bands which are able to form intermolecular H-bonds with other OH groups on the rings and the nitrogen (–O–H· · ·N C–CH3 ) of ACN. As previously reported, the curve fitting can be used to facilitate the assignment of the complex OH broad bands, however, some IR component bands still cannot be resolved due to the contribution of the intramolecular or intermolecular H-bonds with other OH bands [16]. The spectra of GA in ACN shows a strong C O band at 1716 cm−1 which is much higher than that observed in the GA monohydrate (1702 cm−1 ) and anhydrous forms (1666 cm−1 ). ACN cannot form an H-bond with the C O group but this C O group in the solid state can be intermolecularly bonded to other H-bond donors. As a consequence, the vibrational wavenumber of this group shifts to the higher wavenumber in the ACN solution.

3. Results and discussion 3.1. Measurement and deconvolution of FTIR spectra The FTIR spectra of the anhydrous GA, GA monohydrate, GA in ACN solution (using ACN as a background) and ACN (using air as a background) are shown in Fig. 3. The experimental IR spectrum shows some strong bands from the ACN when using air as a background. These solvent peaks can still be observed in the GA sample solution when using the solvent as a background. The spectrum of GA in ACN over the range of 1366–1490 cm−1 was somewhat interfered with by the ACN peaks, due to the incomplete ratio out of the solvent background. Bands below 1000 cm−1 were not shown due to the cuts off of the CaF2 window. ACN also gives strong absorption bands at 2296 and 2255 cm−1 but they do not interfere with the GA bands (Fig. 3). The anhydrous form in the crystal state shows a sharp OH band at 3496 cm−1 , this peak disappears in the monohydrate form. The C O stretching band in the anhydrous form is observed at 1668 cm−1 , while that of the monohydrate is observed at 1702 cm−1 . The anhydrous crystal structure of GA has a similar spectrum to the dry form as previously reported [7]. The spectrum of GA in the ACN solution is different from those of the GA monohydrate or the anhydrous GA measured in the KBr discs (Fig. 3). The interpretation of the complicated OH absorption bands involving intramolecular and intermolecular H-bonds has previously been reported using the curve fitting method [16,17]. In addition, the assignment of the calculated bands of different

ACN

GA in ACN solution

GA monohydrate

anhydrous GA

4000

3500

3000

2500

2000

1500

1000

cm-1 Fig. 3. FTIR spectra of anhydrous GA, GA monohydrate, GA in the ACN solution measured using ACN as a background, and the ACN spectrum measured using air as a background.

96

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100

.20

Absorbance

.15

Peak1 Peak2 Peak3 Peak4 Exp GA-ACN Peak Fitting

.10

.05

0.00 3800

3600

3400

3200

3000

2800

cm-1 Fig. 4. Curve fitting for the OH stretching regions of GA in the ACN solution assigned to the three OH groups (Peak 1 to Peak 3) on the ring and one OH of the acid group (Peak 4).

3.2. Structural and energetic properties of the GA–(ACN)n clusters All local minimum energy structures of the GA–(ACN)n clusters, where n = 1, 2 and 4, and their relative total energies (E) in the kcal mol−1 unit are shown in Fig. 5. The E values are calculated from the summation of the relative structural energies and relative zero-point vibrational energies (ZPE). 3.2.1. The 1:1 cluster All optimized structures of the GA–(ACN)1 clusters (Fig. 5A) show a similar orientation of the ACN bound to GA. As expected, each ACN forms a strong intermolecular H-bond with the OH of the carboxyl group (H(GA)· · ·N C–CH3 ). The H(GA)· · ·N C–CH3 and ˚ O(GA)· · ·N C–CH3 distances should be equal to 1.86 and 2.83 A, respectively. The O–H· · ·N angles are in the range of 165–168◦ . The E values for each 1:1 cluster are almost equal. It seems that the conformations of GA in each of the 1:1 cluster do not affect their relative structural stabilities. 3.2.2. The 1:2 cluster Optimization of all possible structures of the GA–(ACN)2 clusters resulted in five local minimum energy structures (Fig. 5B). The orientations of the ACN molecule in each cluster are similar. The H-bond formations between ACN and the H10 (OH of carboxyl group) and the H14 (C3 -OH) or H18 (C5 -OH) of GA are nearly linear. ˚ but it is about All H(GA)· · ·N C–CH3 distances are about 2.8 A, 2.75 A˚ for GA-IV. For GA-IIa–(ACN)2 , the carboxyl O–H· · ·N and the hydroxyl O–H· · ·N angles are about 177 and 174◦ , respectively. Those for the other clusters are about 168 and 170◦ , respectively. According to the E values for these 1:2 clusters, GA-IIb–(ACN)2 , and GA-I–(ACN)2 , clusters are the lowest-lying energy structures. The E values for the GA-IV–(ACN)2 and GA-IIIb–(ACN)2 clusters are about 0.3 kcal mol−1 above the global minimum. That for the GA-IIa–(ACN)2 cluster has about a 1 kcal mol−1 higher energy than the GA-IIb–(ACN)2 cluster. These results demonstrate that the ACN molecules may have an influence on the relative structural energies and stabilities of these 1:2 clusters. 3.2.3. The 1:4 cluster Optimization of all the possible structures of the GA–(ACN)4 cluster yielded two local minimum energy structures for the GA-II–(ACN)4 and GA-I–(ACN)4 (Fig. 5C). The cluster of the GAIII–(ACN)4 is not stable on the potential energy surface (PES). The cluster of the GA-IV–(ACN)4 is a transition state (TS) structure. However, this conformation may be located at a saddle point on its PES. Transformation of a GA-IV–(ACN)4 cluster to a GAI–(ACN)4 cluster in the ACN environment hardly occurs at room

temperature because the torsional energy barrier of the –COOH· · ·ACN was estimated to be about 7 kcal mol−1 (Fig. 6). ˚ and angles (◦ ) are also shown in Fig. 5C. Their H-bond lengths (A) All 1:4 clusters have four ACN molecules forming H-bonds with all OH groups on the ring and the OH of the carboxyl groups of the GA molecule. The intermolecular H-bonds labeled with 1 and 2 in Fig. 5C are relatively strong bonds while those labeled with 3 and 4 are comparatively weak bonds. According to the E values of these three structures, the GA-I–(ACN)4 and GA-II–(ACN)4 clusters may be considered as the lowest-lying energy structures. Both have comparable E values, the GA-I–(ACN)4 cluster is only about 0.2 kcal mol−1 less stable than the GA-II–(ACN)4 cluster. According to the analysis of the structural and energetic properties of GA–(ACN)n cluster, the number and position of the ACN molecules around GA as well as the intermolecular interactions have a significant effect on the relative stabilities of the GA–ACN clusters. This is probably due to the different interactions between the ACN and GA with different mole ratios of GA:ACN. 3.3. The vibrational spectra of the GA–(ACN)n clusters The calculated stretching wavenumbers for the OH and C O groups of all GA clusters are shown in Fig. 7. All vibrational wavenumbers of these conformations are real so these conformations are at their minima on the PES. The vibrational spectra of all GA–(ACN)n clusters with the inclusion of the continuum solvent in the calculations present a very low (C O) of about 1700 cm−1 , compared to that from the experiment (1716 cm−1 ). Hence, the structural and energetic properties as well as the vibrational spectra of the GA–(ACN)n clusters with the inclusion of the continuum solvent are not presented here. In addition, the vibrational spectra of the GA–(ACN)n clusters without the inclusion of the continuum solvent were used to compare with the experimental value. 3.3.1. The 1:1 cluster The OH and C O stretching wavenumbers of the GA-I, GAII, GA-III and GA-IV clusters are shown in Fig. 7A. The C O stretching wavenumber for these clusters appear in the range of 1715–1718 cm−1 and this corresponds to the one observed in the experimental spectra of the GA solutions. For all conformation, the wavenumbers of the H-bonded OH groups on the ring are higher than those of the free OH group i.e., for GA-III, the (OH) band at C4 (3636 cm−1 ) is higher than the (OH) band at C3 (3554 cm−1 ). These are not in agreement with IR vibrational theory where the stretching vibration of a free OH group should appear at a higher wavenumber than that of a H-bonded OH group. Thus, GA with only one ACN molecule could not be used to represent the solvated structure of GA in ACN solution. The vibrational spectra of these GA–(ACN)1 clusters have been neglected in this study. 3.3.2. The 1:2 cluster The C O and OH stretching wavenumbers of the four lowestlying energy structures of the 1:2 cluster are shown in Fig. 7B. The C O stretching vibration is located at 1717 cm−1 for GA-IIb, GAI and GA-IIIB but that for GA-IV appears at 1719 cm−1 . These are close to those observed in the experimental spectra of the GA in ACN solution (1716 cm−1 ). The prediction for the OH stretching wavenumber on the ring from the GA-IIb and GA-IIIb clusters is irrelevant. Both conformations show lower vibrational wavenumber for the free OH groups at C5 compared to the H-bonded OH groups at C4 . Therefore, the calculated spectra of GA-IIb and GA-IIIb clusters have been neglected. Nevertheless, those values calculated from the GA-I–(ACN)2 and GA-IV–(ACN)2 clusters are equitable. The OH stretching wavenumber for the OH of the carboxyl group of GA-IV (3366 cm−1 )

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100

97

Fig. 5. The minimum energy structures and relative total energies of (A) the GA–(ACN)1 clusters, (B) GA–(ACN)2 clusters and (C) GA–(ACN)4 clusters together with their ˚ and angles (◦ ). H-bond lengths (A)

occurs at a lower wavenumber than that of GA-I (3442 cm−1 ). This may be due to the fact that the intermolecular H-bond between ACN and the OH of the carboxyl group of GA-IV is stronger than that of GA-I. The OH stretching wavenumber at C4 for the GA-I and GA-IV clusters appear at 3623 and 3632 cm−1 , respectively. These wavenumbers are lower than that at C5 (3641 cm−1 ) and C3 (3634 cm−1 ) for the GA-I and GA-IV clusters, respectively. According to the orientation of the OH groups at C4 , two intramolecular H-bonds can be formed as opposed to only one intramolecular H-bond formed by the OH group at the above mentioned C5 and C3 positions. The almost linear alignment of the H-bond formation between the ACN and OH at C3 and C5 of GA-I and GA-IV, respectively, together with the intramolecular H-bond with the nearby OH groups causes the wavenumbers of these OH groups

(3485 and 3493 cm−1 for the GA-I and GA-IV cluster, respectively) to appear at the lowest wavenumbers compared to those of the other two OH groups on the ring. 3.3.3. The 1:4 cluster The wavenumbers of the C O and OH stretching of the two lowest-lying energy structures of the GA-I–(ACN)4 and GA-II–(ACN)4 clusters are shown in Fig. 7C. The GA-I–(ACN)4 and GA-II–(ACN)4 clusters show (C O) bands at 1718 and 1719 cm−1 , respectively. The stretching wavenumber for the OH of the carboxyl group appear at lower wavenumber than for the hydroxyl bands. Comparison between the calculated OH stretching wavenumber for the GA-I–(ACN)2 , GA-IV–(ACN)2 , GA-I–(ACN)4 and GA-II–(ACN)4 clusters and the derived FTIR component bands

98

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100

Fig. 6. The torsional energy barrier of COOH· · ·ACN for the GA-IV–(ACN)4 cluster to the GA-I–(ACN)4 cluster.

from the GA in ACN are shown in Table 1. For the GA-II–(ACN)4 cluster, the calculated vibrational wavenumber of the OH groups are certainly different from the FTIR component OH bands. The calculated wavenumber of the OH groups for the other clusters are

comparable to each other and are close to the values obtained from the experiments. Nevertheless, the least comparable wavenumber between the calculated and experimental values for all structures are for the OH of the carboxyl bands. As mentioned above, this broad peak may not include only the OH of the carboxyl group, but may be combined with the intramolecular and intermolecular H-bonds from the other OH groups. Thus, the exact position for the OH of the carboxyl group may be difficult to obtain. The experimental FTIR spectra for the GA in the ACN solution and those for the calculated spectra of the GA-I–(ACN)2 and GA-IV–(ACN)2 and GA-I–(ACN)4 and GA-II–(ACN)4 clusters are shown in Fig. 8. The spectrum of the GA-II–(ACN)4 cluster is different from the GA spectrum obtained from the experiment. In addition the spectra of other clusters are comparable to each other and to the experimental values. The dissimilarity between the calculated and experimental wavenumbers of the GA-II–(ACN)4 cluster indicates that this cluster may not be the preferable structure in the real situation although, of the 1:4 clusters, its structure is the most stable. According to these results, GA can interact with ACN and the most favorable conformations may exist as GA-I and GA-IV. The

Fig. 7. Vibrational wavenumbers of the OH and C O groups of GA in the GA–ACN clusters.

Table 1 The (OH) and (C O) bands (in cm−1 ) of the FTIR component for GA in ACN solution and as calculated for the GA-I–(ACN)2 , GA-IV–(ACN)2 , GA-I–(ACN)4 and GA-II–(ACN)4 clusters. Deconvoluted experimental IR bands

3647 3607 3540 3353 1716

GA-I–(ACN)2

GA-IV–(ACN)2

GA-I–(ACN)4

GA-II–(ACN)4

Calculated

Assignment

Calculated

Assignment

Calculated

Assignment

Calculated

Assignment

3641 3623 3485 3422 1717

OH at C5 OH at C4 OH at C3 Carboxylic OH C O

3634 3632 3493 3366 1719

OH at C3 OH at C4 OH at C5 Carboxylic OH C O

3645 3620 3508 3444 1718

OH at C5 OH at C4 OH at C3 Carboxylic OH C O

3553 3582 3564 3406 1719

OH at C5 OH at C4 OH at C3 Carboxylic OH C O

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100

99

Fig. 8. The experimental FTIR spectra of GA in ACN solution (ACN background) and ACN (air background) as well as the calculated IR spectra of the GA-I–(ACN)2 , GA-IV–(ACN)2 , GA-I–(ACN)4 and GA-II–(ACN)4 clusters, (a) the whole range spectra and (b) the expanded spectra over the range of 1800–1000 cm−1 .

conformations of GA-I and GA-IV are similar to the conformations of the GA monohydrate reported by Okabe et al. [6] and the anhydrous GA described by Hirun et al. [8], respectively. These two conformations have the same orientation of the OH groups but possess different directions of the C O group as they relate to the OH groups on the ring.

4. Conclusion Based on the MD simulation and RDFs analysis, the number of ACN molecules located close to each OH group of the GA molecule was suggested to be 1.05 [9]. In this study, the appropriate number of ACN molecules that strongly interact with a GA molecule is probably 2 or 4. The 1:1 clusters do not represent the solvated structure of GA in the ACN. The GA-I–(ACN)2 and GA-IV–(ACN)2 clusters which have the two most accessible OH groups of GA interacting with ACN provide comparable vibrational wavenumbers of the OH groups with those from the experiment. The GA-I–(ACN)4 cluster for which all the OH groups on the GA molecule can interact with ACN also present good agreements of their OH stretching wavenumbers with the IR component bands of the OH groups. The order from the highest to the lowest wavenumbers of the OH bands for the GA-I–(ACN)2 and GA-I–(ACN)4 clusters correspond to the OH at C5 > OH at C4 > OH at C3 . Meanwhile, the sequence for the GA-IV–(ACN)2 cluster is OH at C3 > OH at C4 > OH at C5 . All OH groups on the ring are oriented in the same direction for both the GA-I and GA-IV conformations, the conformation of the OH group at C3 of GA-I are actually the same as that at the C5 of GA-IV. The numbering of these conformations is diverse due to the different orientation of their carboxyl groups. In conclusion, these calculations with the comparisons with the experimental values can be successfully used to predict the preferred GA conformation in ACN. GA prefers to exist as the GA-I and GA-IV conformations in ACN and possibly in the polar environment of its receptors. The interaction between these conformations and the receptors may subsequently elicit its biological activities.

Acknowledgments This work was supported by the Thailand Research Fund through the Royal Golden Jubilee Ph.D. Program through Grant No. PHD/0259/2549, National Research Council of Thailand through the Prince of Songkla University through Grant No. PHA520036S, the Higher Education Research Promotion and National Research University Project of Thailand, Office of the Higher Education Commission and PSU Grid. References [1] Y. Lu, F. Jiang, H. Jiang, K. Wu, X. Zheng, Y. Cai, M. Katakowski, M. Chopp, S.S.T. To, Eur. J. Pharmacol. 641 (2010) 102–107. [2] V.R. Punithavathi, P.S.M. Prince, R. Kumar, J. Selvakumari, Eur. J. Pharmacol. 650 (2011) 465–471. [3] D.H. Priscilla, P.S.M. Prince, Chem. Biol. Interact. 179 (2009) 118–124. [4] C.A. Gomes, T. Girao da Cruz, J.L. Andrade, N. Milhazes, F. Borges, M.P.M. Marques, J. Med. Chem. 46 (2003) 5395–5401. [5] R. Jiang, D. Ming, P.P.H. But, T.C.W. Mak, Acta Crystallogr. C 56 (2000) 594–595. [6] N. Okabe, H. Kyoyama, M. Suzuki, Acta Crystallogr. E 57 (2001) o764–o766. [7] F. Billes, I. Mohammed-Ziegler, P. Bombicz, Vib. Spectrosc. 43 (2007) 193–202. [8] N. Hirun, S. Saithong, C. Pakawatchai, V. Tantishaiyakul, Acta Crystallogr. E E67 (2011) o787. [9] C. Cappelli, B. Mennucci, S. Monti, J. Phys. Chem. A 109 (2005) 1933–1943. [10] I. Mohammed-Ziegler, F. Billes, J Mol. Struct. (THEOCHEM) 618 (2002) 259–265. [11] E. Tajkhorshid, K.J. Jalkanen, S. Suhai, J. Phys. Chem. B 102 (1998) 5899–5913. [12] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, J.A. Montgomery Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, S. Li, J.E. Knox, H.P. Hratchian, J.B. Cross, V. Bakken, D. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Gaussian Inc., Wallingford, CT, 2004. [13] S. Dokmaisrijan, V.S. Lee, P. Nimmanpipug, J. Mol. Struct. (THEOCHEM) 953 (2010) 28–38.

100 [14] [15] [16] [17]

N. Hirun et al. / Spectrochimica Acta Part A 86 (2012) 93–100 A.K. Rai, C. Song, Z. Lin, Spectrochim. Acta A 73 (2009) 865–870. A.K. Rai, X. Xu, Z. Lin, D.K. Rai, Vib. Spectrosc. 56 (2011) 74–81. T. Kondo, Cellulose 4 (1997) 281–292. M. Bonet, C. Quijada, F. Cases, Can. J. Anal. Sci. Spectrosc. 49 (2004) 234–239.

[18] E.D. Raczynska, K. Duczmal, M. Darowska, Vib. Spectrosc. 39 (2005) 37–45. [19] R.M. Silverstein, F.X. Webster, D.J. Kiemle, Spectrometric Identification of Organic Compounds, seventh ed., John Wiley & Sons, New Jersey, 2005.