Transition from Planar to Nonplanar Hydrogen Bond Networks in the

Apr 5, 2013 - formed by using supersonic expansions, which maintain the ... from cyclic to noncyclic hydrogen bond networks earlier than in similar sy...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/JPCA

Transition from Planar to Nonplanar Hydrogen Bond Networks in the Solvation of Aromatic Dimers: Propofol2‑(H2O)2−4 Iker León,† Judith Millán,‡ Emilio J. Cocinero,† Alberto Lesarri,§ and José A. Fernández†,* †

Department of Physical Chemistry, Faculty of Science and Technology, University of the Basque Country (UPV/EHU), B. Sarriena s/n, Leioa 48940, Spain ‡ Department of Chemistry, Faculty of Science, Agroalimentary Studies and Informatics, University of La Rioja, Madre de Dios, 51, Logroño 26006, Spain § Department of Physical Chemistry and Inorganic Chemistry, Faculty of Science, University of Valladolid, E-47011 Valladolid, Spain S Supporting Information *

ABSTRACT: Propofol (2,6-diisopropylphenol) is probably the most widely used intravenous general anesthetic. In this work, the interaction of propofol dimer with 2−4 water molecules was analyzed. The molecular aggregates were formed by using supersonic expansions, which maintain the molecules confined in a cold, collision-free environment. The clusters were then examined by using a number of massresolved laser-based spectroscopic techniques, including 2-color REMPI (resonance enhanced multiphoton ionization), UV/UV hole burning, and IR/UV double resonance. Two isomers were found for each stoichiometry, whose final structures were determined by comparison between the experimental data and those from density-functional-theory calculations (M06−2X/6-31+G(d)). The analysis of the observed structures allows the conclusion that the water molecules always form hydrogen bond networks, whose contribution to the cluster’s total binding energy increases with the number of water molecules. In the cluster with four water molecules, the two propofol molecules lose most of their contact points. In addition, the steric hindrance produces a change from cyclic to noncyclic hydrogen bond networks earlier than in similar systems.



INTRODUCTION

A general problem with the IVA and of many other bioactive substances is their poor solubility in water and therefore they cannot be administered as simple salts but as emulsions.3 For example, propofol (2,6-diisopropylphenol, Scheme 1), the most widely used IVA, is usually administered as an emulsion in 10% soybean oil, 2.25% glycerol, and 1.2% purified egg phosphatide. Such a mixture may favor bacterial growth, and therefore in some formulations EDTA or sodium metabilsulfite are added to inhibit bacterial growth.4 Thus, solubilization of IVA, and in particular of propofol, is a major issue, which requires complex formulations involving a number of chemicals, each one able to produce undesirable side effects, increasing also the injection volumes. The low solubility also affects the bioavailability of the active substance. Furthermore, the rate-limiting step in the absorption process for poorly water-soluble drugs is the dissolution rate of such drugs in the gastro-intestinal fluids.5 Several solutions have been developed to deal with the problem,6 including micronization,5 nanosuspension,7 eutectic mixtures,8 solid dispersions,9 complexation,10,11 microemulsions,12,13 hydrotropy,14,15 etc. The existence of such a large number of alternatives indicates that no universal procedure has been developed so far. In the quest for such a universal

Introduction of anesthesia by Crawford Long around the 1840s constituted a great advance in medicine.1 Although the first anesthetics were gaseous agents (such as ether, nitrous oxide, chloroform, or cyclopropane) the need for an intravenous anesthetic (IVA) was soon clear due to the problems that the volatile anesthetics present, such as flammability or the risk of hypoxia in the operating room. However, it was not until the 1930s that the first IVA, thiopental (Scheme 1), was introduced. Since then, many substances that work the same as IVA have been developed.1,2 Scheme 1. Propofol (2,6-Diisopropylphenol, Left) and Thiopental (Right)

Received: February 7, 2013 Revised: April 2, 2013 Published: April 5, 2013 © 2013 American Chemical Society

3396

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

solution, a deep knowledge of the drug-water interaction is of utmost importance. In principle, propofol, as phenol, presents an OH group that may form hydrogen bonds with water, and therefore it should present some solubility in water. However, the steric hindrance imposed by the isopropyl groups makes such interaction less stable than in phenol,16−19 and consequently, the solubility of propofol in water is 2 orders of magnitude lower than that for phenol. In previous studies, the structures of phenol clusters containing up to nine water molecules were determined,16−20 showing that water gathers around the hydroxy moiety, trying always to form the same polyhedral structures found in ice. The same steric hindrance prevents propofol from forming strong hydrogen bonds in the homodimer, which in turn presents strong C−H···π interactions and an unexpectedly high aggregation energy.21 In this work, we increase the knowledge on the solvation of propofol, extending the study to higher order clusters, namely the propofol dimer with 2−4 water molecules. For that purpose, we use a methodology successfully tested in previous works: The clusters are formed by using a supersonic expansion of a mixture of propofol and water. Then, they are interrogated by using a number of mass-resolved experimental techniques that allow obtaining their electronic and IR spectra and determining the number of isomers in each stoichiometry. Such results are compared with the predictions by using density functional theory (DFT) and a suitable basis set. In this way, the structure of all the isomers found can be determined. Although the methodology is similar to that of previous works, application to such large systems requires very careful optimization of the experimental setup. The conditions to produce the clusters are achievable only through a delicate adjustment of all the experimental parameters. Furthermore, a significantly harder computational effort has been necessary to explore the conformational landscape and to carry out the optimizations.

Figure 1. Variation of the mass spectrum of an expansion of a propofol/water/He mixture in the vicinity of the propofol dimer at different valve-laser delays. At short delays, detection of propofol homomers is favored (1). As the delay increases, PPF2Wm clusters are detected (2). If the delay is increased further, the intensity of PPF2Wm clusters increases, but PPF1Wm+10 clusters are also detected (3). Depending on the delay, a clear separation between monomer−water and dimer−water clusters can be achieved (a−c).

Computations. A detailed description of the computation procedure may be found elsewhere.16,20 Briefly, a first exploration of the conformational landscape was performed by using molecular mechanics (MMFFs force field), as implemented in Macromodel (Schrodinger Inc.). A combination of Monte Carlo and “Large scale-Low modes” procedures was used to carry out a thorough conformational search. A minimum of 100 000 runs were employed for each stoichiometry, usually resulting in thousands of structures, most of them redundant. Once redundancies were eliminated, the remaining structures were grouped into families, attending to their structural similarity. At least one member of each family and all the lower energy structures were subjected to further optimization at the M06−2X/6-31+G(d) level, imposing no restrictions. The calculation level was chosen as a compromise between calculation accuracy and computation time. The M06−2X method in combination with Popple’s basis sets has been demonstrated to yield accurate results for similar systems.16−18,20 At the calculation level chosen, optimization of a single structure of PPF2W4 takes on average ∼48 h (wall time, 4 processors), while frequency calculations take around 36 h (wall time), so the use of larger basis sets would be impractical. All the optimizations were followed by a vibrational harmonic frequency calculation to check which structures are true minima and to obtain the zero-point energy (ZPE) corrections. Thus, all the relative energies presented in this work contain such a correction. The cluster’s binding energy was also estimated by subtracting the energy of the components



EXPERIMENTAL METHODS A detailed description of the experimental system may be found elsewhere and therefore only the most relevant details are presented here.21 Propofol (Sigma-Aldrich) and water vapor were seeded in He or Ar and expanded inside the chamber of a time-of-flight (TOF) mass spectrometer, using a pulsed valve (both Jordan Inc.). The relatively high vapor pressure of propofol makes any warming unnecessary. The jet formed is interrogated by using up to three lasers. All the REMPI (resonance enhanced multiphoton ionization) experiments were carried out with use of a two-color detection scheme. Due to an accidental coincidence, the peaks of propofol2(H2O)n (hereafter PPF2Wn) are very close in the mass spectrum to those of propofol 1 (H 2 O) n+10 (hereafter PPF1Wn+10). Figure 1 shows a mass spectrum of a propofol/ water/He expansion, recorded by tuning the laser to 37 383 cm−1, which ensures that its wavelength is not in resonance with any transition of the detected species. Depending on the laser-valve delay, it is possible to favor the detection of propofol homomers, PPF1Wm or PPF2Wn clusters. Careful optimization of the experimental variables was critical to record the spectra presented in this work. UV/UV and IR/UV double resonance experiments were carried out by using a 2-color detection scheme and real time active subtraction to increase the s/n (signal-to-noise) ratio. See ref 21 for the details. 3397

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

than 1500 cm−1. Surprisingly, the absorption almost disappears in the spectrum of PPF2W3, which looks cleaner. As discussed in previous works18 the origin of such broad absorption may not be the accumulation of low-frequency vibrations. A similar behavior was found for PPF1W4−6, concluding that it could be related to the floppiness of the cluster’s structures. However, as will be shown below, the structures of the isomers assigned in this work, although complex, do not seem to present the floppiness of PPF1W2. The s/n ratio decreases with the size of the cluster, partly due to a drop in intensity, but still a reasonably well-resolved spectrum was obtained for the PPF2W4 species. Regarding the frequency shifts, the spectra of PPF2W3 and PPF2W4 appear in the same region, slightly red-shifted with respect to the PPF2W2 spectrum. It is worth noting that it was possible to avoid fragmentation during the study, carefully adjusting the experimental parameters. We will now analyze the spectroscopy of each cluster independently. Propofol2(H2O)2. Despite the REMPI spectrum of this species being dominated by a broad absorption, it was possible to find contributions from at least two different isomers (Figure 3), with origin bands at 36 087 and 36 152 cm−1, respectively.

from the cluster’s total energy. The basis set superposition error was estimated by using the counterpoise procedure of Boys and Bernardi.22 The structures presented in this work are denoted as WnSm, where n refers to the number of water molecules in the cluster and m to the energetic order of the structure, starting with the lowest energy one.



RESULTS Figure 2 shows a comparison between the 2-color spectrum of propofol and PPF2W0−4. As demonstrated in previous

Figure 2. Two-color REMPI spectra of propofoln(H2O)m, n = 1,2; m = 0−4. The 1:0, 2:0, and 2:1 spectra are taken from refs 16 and 21.

Figure 3. UV/UV hole burning traces of PPF2W2, obtained probing the transitions at 36 087 and 36 152 cm−1 respectively. The REMPI (R2PI) spectrum in the 36 100−36 400 cm−1 region is also shown for comparison. The peaks denoted with asterisks indicate the transitions employed for recording the hole burning spectra for each isomer.

studies,16,23 the propofol spectrum contains contributions from four different isomers, which differ in the relative orientation of the isopropyl groups. The energy difference between those four conformers is less than 5 kJ/mol, although the isomerization barrier between the two most stable ones is small enough to allow depopulation of one of the species if Ar is used instead of He. Formation of the dimer results in a reduction in the number of species to only two, with very similar structures:21 both present a C−H···π interaction, while the OH moieties present either a dipole−dipole interaction (isomer 1) or a weak hydrogen bond (isomer 2). This weak hydrogen bond is reinforced in the three detected isomers of PPF2W1.21 It is also worth mentioning that the 000 transition of the propofol dimer was not located, probably due to geometry changes upon ionization.21 The spectrum of PPF2W2 presents a very small shift to the blue. This suggests that the bottom of the excited state potential energy surface (PES) may have not been reached, due to the conformational changes that these species usually experience upon excitation. The spectrum of PPF2W2 is dominated by a broad absorption, which extends for more

The broad absorption also appears in the hole burning trace of isomer 2, and therefore it is either due to this isomer or it shows up because it has a significant contribution to the intensity of the transition at 36 152 cm−1, which is the one probed to record the hole burning trace. To obtain structural information from PPF2W2, the IR/UV spectra of the two detected isomers were recorded, probing the transitions marked with asterisks in Figure 3. The results are presented in Figure 4, together with the predicted spectra for some representative calculated structures. Clearly, the spectra of both isomers are markedly different: while the spectrum of isomer 1 presents three strong transitions around ∼3439 cm−1, and at least two transitions in the free OH stretch region, the spectrum of isomer 2 presents a group of four OH stretches between 3258 and 3443 cm−1 and two very well resolved transitions in the free OH stretching region. The calculations for this system resulted in 18 fully optimized structures (Figure S1 of the Supporting Information), a 3398

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

Figure 4. IR/UV hole burning traces obtained for the two isomers found for PPF2W2, probing the transitions at 36 087 and 36 152 cm−1, respectively. The spectra predicted for some representative calculated structures are also shown for comparison. The relative stability of each structure (in kJ/mol) is also shown in brackets. A correction factor of 0.956 was applied to the computed spectra to account for anharmonicity. All the calculated spectra can be found in Figure S2 of the Supporting Information.

significantly lower number of conformers than for other smaller propofol clusters, due to the relatively higher stability of those configurations with cyclic hydrogen bond networks. Examining the calculated structures, it is clear that the two propofol molecules reach an optimum position when they can establish C−H···π weak hydrogen bonds, while the water molecules form a hydrogen bond network with the two OH moieties. Thus, not many possibilities are left for the system. The lowest energy structure (W2S1) presents the two water molecules interacting with each other and with the two propofol molecules. Such a hydrogen-bonded network has no correspondence with those found in pure water clusters. Nevertheless, the comparison with the experimental results clearly demonstrates that such an isomer is not detected. Regarding the comparison between experimental and simulated spectra, a discrepancy in the intensities is clearly appreciated. One must keep in mind that the calculated frequencies are obtained by using a normal-mode analysis, while the experimental spectrum is obtained by using a double resonance technique and the existence of hydrogen bonds surely includes anharmonic contributions. Most of the calculated structures present an eight-member ring hydrogen bond network, although some other conformations are also possible, like W2S9 or W2S12, in which no cyclic networks are found. The comparison with the experimental

traces clearly demonstrates that such structures cannot reproduce the observed spectra. Certainly, assignment of isomer 1 to the calculated structures is surprisingly straightforward for this stoichiometry, as only W2S7 is able to reproduce the presence of three peaks around 3439 cm−1. It is also clear from the spectrum that the structure of isomer 2 presents the two propofol moieties forming C−H···π hydrogen bonds, and that the water molecules are forming a cyclic hydrogen bond network. However, several very similar calculated structures are able to reproduce the observed spectral features. We therefore assign isomer 2 to an W2S2-type structure. The complete set of calculated spectra is collected in Figure S2 of the Supporting Information, while Table S1 in the Supporting Information collects the relative stability and calculated binding energies of the isomers found. The structure of the two assigned isomers in Cartesian coordinates can also be found at the end of the Supporting Information. Propofol2(H2O)3. Two isomers were found for this cluster, as demonstrated by the UV/UV hole burning experiment in Figure 5. However, it is not clear where the 000 transitions of the two isomers are located. Clearly, the strong transition at 36 033 is not the true origin of isomer 1, which presents a weak feature at 35 991 cm−1. A progression in a ∼ 30 cm−1 mode is also visible in the spectrum of this isomer. On the other hand, the first discrete transition of isomer 2 is located at 36 198 cm−1, 3399

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

of the REMPI spectrum do not appear in any of the two hole burning traces, but they are due to fragmentation from higher order clusters. Nevertheless, the existence of further isomers may not be completely ruled out. It was possible to record the IR/UV spectra of the two isomers by using the two transitions highlighted with asterisks in Figure 5, obtaining the traces shown in Figure 6, together with a molecular representation and the spectra of some representative calculated structures. The IR spectrum of isomer 1 shows nine transitions in the 3100−3800 cm−1 region, indicating that the OH stretches are nonequivalent, and therefore that each one is surrounded by a particular environment, thus discarding formation of simple, cyclic hydrogen bond networks. In addition, the strong shift to the red of some of the transitions, compared with those in the PPF2W2 spectrum, points to a reinforcement of some of the hydrogen bonds. Finally, only eight transitions were expected (2 × 3 water molecules plus two propofolic OH stretches), while nine features were detected. The extra feature may be due to an additional isomer with overlapping transitions or to a Fermi resonance. Additional IR/IR/UV/UV experiments would be needed to elucidate the origin of this extra peak,20,24,25 although the weak signal intensity seriously

Figure 5. UV/UV hole burning traces of PPF2W3, obtained probing the transitions at 36 033 and 36 235 cm−1, respectively. The REMPI (R2PI) spectrum in the 35 900−36 400 cm−1 region is also shown for comparison.

but the broad absorption in the hole burning trace seems to indicate that the true origin was not reached. Some transitions

Figure 6. IR/UV hole burning traces obtained for the two isomers found for PPF2W3, probing the transitions at 36 033 and 36 235 cm−1, respectively. The spectra predicted for some representative calculated structures are also shown for comparison. The relative stability of each structure is also shown in brackets. A correction factor of 0.956 was applied to the computed spectra to account for the anharmonicity. All the calculated spectra can be found in Figure S3 of the Supporting Information. 3400

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

structures with complex 3D shapes. Figure 8 also shows the spectra of some selected calculated structures. The complete collection of 42 calculated structures may be found in Figure S5 of the Supporting Information, while the predicted IR spectra are collected in Figure S6 of the Supporting Information. In most of the calculated structures, the two aromatic rings interact with the isopropyl groups of the other molecule, while the water molecules aggregate and interact with the hydrophilic side of the chromophores, forming a hydrogen bond network. Three kinds of networks are predominantly found: prism-like structures, although one of the edges is missing (W4S1 and W4S3 for example), book-like (for example W4S2, W4S4, and W4S6), and cycles (for example, W4S11, W4S13, and W4S16). Each of these structures has its own signature in the IR. Thus, while the isomers with book-like structures present spectra that reproduce the experimental traces, those with prism-like networks have the vibrational transitions grouped at higher frequencies in the IR spectrum and therefore they do not reproduce the observed spectra. The predicted spectra for the structures with cyclic networks do not reproduce well the recorded spectra either. On these considerations and following energetic criteria, the two experimental traces are assigned to the calculated structures W4S2 (isomer 1) and W4S5 (isomer 2). The structure of these two isomers in Cartesian coordinates may be found at the end of the Supporting Information, together with the rest of the assigned structures.

compromises the success of such complicated experiments. On the other hand, isomer 2 presents eight transitions in the 3250−3700 cm−1 region. Three of them are very close, pointing to the existence of three very equivalent OH bonds. Thousands of structures were found for this cluster by using molecular mechanics. However, after grouping the structures into families, only 24 optimized structures were obtained as shown in Figure S3 (Supporting Information). Following the trend observed in the most stable calculated structures of PPF2W3, the hydrophobic sides of the propofol molecules are close to each other establishing several C−H···π interactions, while the water molecules form more or less complicated hydrogen bond networks. In some of them, the two propofol molecules are relatively separated, pointing to an increase in the relative importance of the contribution of the hydrogen bonds to the cluster’s total stability. W3S1, the global minimum, is the structure whose predicted spectrum better reproduces the experimental trace obtained for isomer 1. In this structure, all the OH bonds are connected in a hydrogen bond network. We therefore assign isomer 1 to the structure W3S1. However, no single assignment was found for isomer 2: at least two structures, W3S2 and W3S3, reproduce reasonably well the experimental features in the IR spectrum. A close inspection of both structures shows that they are very similar, with the only difference on the position of the propofol moieties, which are slightly farther away in the latter. Therefore, we assign isomer 2 to the more stable W3S2 and very likely the barrier for W3S3 to isomerize into W3S2 is too low to allow trapping population in this third isomer. Propofol2(H2O)4. The spectrum of PPF2W4 presents a large abundance of well-resolved lines. Nevertheless, two isomers are detected, as demonstrated in the hole burning spectra presented in Figure 7. The 000 transition of isomer 1 appears



DISCUSSION The experimental detection of the hydration clusters of the propofol dimer makes possible a structural comparison with the geometries of the pure water clusters, phenol−water and propofol−water systems. This comparison is valuable considering the small number of intermolecular clusters of this size examined with mass-resolved excitation spectroscopy. Figure 9 collects the structures of the isomers assigned in this work, together with the structure of several other systems found in the literature and reoptimized at the same calculation level employed in this work (M06−2X/6-31+G(d)). The comparison with pure water clusters gives us an estimation of the relative contribution of the hydrogen bond networks to the overall cluster stabilization energy, while comparison with the clusters of phenol and propofol monomer allows estimating the influence of the isopropyl groups and of the hydrophobic interactions in the cluster’s final structure. As observed in the figure, the structure of the water tetramer is an 8-member ring. Such a structure is maintained with small changes in the clusters of phenol and propofol with three water molecules. The two detected isomers of PPF2W2 maintain the same hydrogen bond network, and at the same time the position of the two propofol molecules is not much different from that in the propofol dimer: they still are able to maintain the C−H···π interactions that give the cluster a relatively high stabilization energy, compared with similar systems. The water pentamer presents a single isomer. This structure is maintained in phenol·W4, as shown in Figure 9. Apparently the pentameric structure is particularly stable, and therefore water has a strong tendency to form it. However, such a structure was not found experimentally in PPF1W4, in which the water molecules prefer to form an eight-member ring instead, leaving the fourth water molecule outside, interacting both with the water and the aromatic rings. Such a structure is highly dynamic and was suggested as being responsible for the absence of discrete features in the REMPI spectrum.18

Figure 7. UV/UV hole burning traces of PPF2W4, obtained probing the transitions at 36 018 and 36 124 cm−1, respectively. The REMPI (R2PI) spectrum in the 36 000−36 400 cm−1 region is also shown for comparison.

at 36 018 cm−1, and its spectrum exhibits two progressions of vibronic bands with 10 cm−1 spacing. On the other hand, the 000 transition of isomer 2 appears at 36 046 cm−1 and it is separated by ∼1 cm−1 from a vibronic transition of isomer 1. Thus, to avoid interferences, the peak at 36 124 cm−1 was used to record the UV/UV hole burning of isomer 2. Figure 8 shows the IR/UV spectra of the two isomers of PPF2W4 recorded by using the two transitions highlighted with asterisks in Figure 7. Both spectra present a large abundance of lines, scattered along the 700 cm−1 scanned, pointing to 3401

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

Figure 8. IR/UV hole burning traces obtained for the two isomers found for PPF2W4, probing the transitions at 36 018 and 36 124 cm−1, respectively. The spectra predicted for some representative calculated structures are also shown for comparison. The relative stability of each structure is also shown in brackets. A correction factor of 0.956 was applied to the computed spectra to account for the anharmonicity. All the calculated spectra can be found in Figure S6 of the Supporting Information.

Figure 9. Comparison between the structures of the isomers of PPF2W2−4 and those of PPF1W3−5, phenolW3−5, and (water)4−6. The structures of the phenol clusters were taken from ref 26 and reoptimized at the M06−2X/6-31+G(d) level, while those of water clusters were taken from refs 27 and 28 and similarly reoptimized. 3402

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

exposing the hydroxyl groups to water. An extrapolation to a more concentrated solution would require data on larger clusters, containing several propofol units and an increasing number of water molecules, but we can extrapolate that the propofol units will form larger aggregates that may even have a micellar shape. This last condition would depend on whether they can arrange such that the aromatic rings are in the center establishing π−π or N−H···π interactions, leaving the hydroxyl groups in the outside of the micelle. Theoretical studies demonstrate that at least the benzene molecules can gather around an alkali metal ion, forming a spherical aggregate.29 Leaving aside the differences between the system of ref 29 and propofol hydrated clusters, such a study demonstrates that it is geometrically possible to form globular aggregates of aromatic molecules. Our group is currently working with higher propofol/water clusters to shed more light on this issue.

According to the calculations, structures in which the water molecules form a ten-member ring are possible, but they were not found experimentally. The results for PPF2W2 confirm such conclusions: the water molecules form an eight-member ring in the two detected isomers, while the other propofol molecule accommodates to establish O−H···O and C−H···π interactions. It is worth noting that, apparently, the system prefers to optimize the position of the hydroxy group to maximize its interaction with the water molecules. Consequently, the distance between the two aromatic rings increases from that found in lower order clusters. Thus, the hydrogen bonds seem to have a stronger influence in the cluster’s final structure than the hydrophobic interactions. The cluster of six water molecules can adopt at least four different shapes, which, attending to their relative stability, are the following: prism, cage, book, and hexagon.28 However, if the Gibbs free energy is taken into account instead of enthalpy, the hexagon becomes the most favorable one, followed by the book, cage, and prism.28 Nevertheless, no isomer with a hexagonal network was found for the clusters of phenol or propofol with water. It is also worth noting that in a recent work by Pérez et al. using microwave spectroscopy, water6 clusters with prism, cage, and book geometries were found, but not the isomer with hexagonal geometry.27 In the two isomers reported for phenolW5 the water molecules are forming bookand cage-type structures. On the other hand, two isomers were also found for PPF1W5, one of them with a book-like network and the other one with four water molecules forming an eightmember ring.18 The latter lies relatively high in energy and therefore the assignment should be taken with caution. However, if the assignment is correct, it may be giving information on the kinetics of the cluster formation process. Finally, the two isomers detected for PPF2W4 present book-like networks. It is worth mentioning that in isomer 2 there is almost no contact between the hydrophobic sites of the two propofol molecules, highlighting the importance of the hydrogen bond interactions over those that take place between the hydrophobic sides of the chromophores. Thus, the hydrophobic interactions are limited to those compatible with the frame imposed by the hydrogen bond network. The evolution in the structures determined in this work resembles what is observed in the transition from a concentrated to a diluted solution: when a solution is saturated, the solute molecules can cluster together, as in PPF2W2−3. Addition of more solvent can eventually result in the disaggregation of such solute clusters. Another important observation is that the transition from (planar) cyclic to (nonplanar) noncyclic hydrogen bond structures occurs with six water molecules in pure water clusters and with the fifth water molecule in phenol/water clusters. However, such transition takes place with four water molecules in the hydrated clusters of propofol and with three water molecules in propofol2 hydrated clusters. The preference for noncyclic structures in the case of propofol may be a consequence of the steric hindrance introduced by the isopropyl groups or may be due to entropic reasons. A comparison with additional systems should be necessary to confirm such hypotheses, but to the best of our knowledge, this is the first vibrational analysis of a molecular aggregate containing two aromatic molecules and up to four water molecules. Regarding the solubility of IVA, this study confirms the tendency of hydrophobic sides of propofol to aggregate,



CONCLUSIONS In this work the spectroscopy of PPF2W2−4 clusters has been analyzed by using several mass-resolved, laser-based spectroscopic techniques. By using 2-color REMPI the spectrum of the S1 electronic state of the species was obtained. All three clusters exhibit spectra composed of a broad absorption and discrete features, the broad absorption being more important for PPF2W2. Small shifts in the electronic origins of all three clusters were observed. With use of UV/UV hole burning spectroscopy it was demonstrated that each spectrum contains contributions from at least two isomers, whose IR/UV spectra were obtained by probing selected bands. The comparison between those IR spectra and the results from calculations at the M06−2X/6-31+G(d) level demonstrates that the water molecules form hydrogen bond networks that resemble those found on pure water clusters. The importance of the contribution of the hydrogen bond interaction to the total cluster’s binding energy increases with the number of water molecules. These changes are reflected in the distance between the hydrophobic sides of the two propofol molecules, also increasing with the number of water molecules. This effect is especially evident for PPF2W4 isomer 2, in which the importance of the C−H···π interactions is drastically reduced. Experiments on larger clusters are on the way to confirm the conclusions of this work.



ASSOCIATED CONTENT

S Supporting Information *

All calculated structures, together with their predicted IR spectra and energetic data, and Cartesian coordinates for the assigned structures. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Fax: +34946013500. Tel: +34946015387. E-mail: josea. [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We would like to thank Dr. Grazia M. Cereghetti for her advice on the solubility of drugs issue. The research leading to these results has received funding from Spanish Ministry of Science and innnovation MICINN/MINECO (Consolider-Ingenio 3403

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404

The Journal of Physical Chemistry A

Article

(19) León, I.; Cocinero, E. J.; Rijs, A. M.; Millán, J.; Alonso, E.; Lesarri, A.; Fernández, J. Formation of water polyhedrons in propofolwater clusters. Phys. Chem. Chem. Phys. 2013, 15, 568−575. (20) Leon, I.; Millan, J.; Cocinero, E.; Lesarri, A.; Castano, F.; Fernandez, J. A. Mimicking Anaesthetic-Receptor Interaction: a Combined Spectroscopic and Computational study of Propofol···Phenol. Phys. Chem. Chem. Phys. 2012, 14, 8956−8963. (21) León, I.; Millán, J.; Castaño, F.; Fernández, J. A. A Spectroscopic and Computational Study of Propofol Dimers and Their Hydrated Clusters. ChemPhysChem 2012, 13, 3819−3826. (22) Boys, S. F.; Bernardi, F. Calculation of Small Molecular Interactions by Differences of Separate Total Energies - Some Procedures with Reduced Errors. Mol. Phys. 1970, 19, 553−566. (23) Lesarri, A.; Shipman, S. T.; Neill, J. L.; Brown, G. G.; Suenram, R. D.; Kang, L.; Caminati, W.; Pate, B. H. Interplay of Phenol and Isopropyl Isomerism in Propofol from Broadband Chirped-Pulse Microwave Spectroscopy. J. Am. Chem. Soc. 2010, 132, 13417−13424. (24) Weiler, M.; Bartl, K.; Gerhards, M. J. Infrared/ultraviolet quadruple resonance spectroscopy to investigate structures of electronically excited states. Chem. Phys. 2012, 136, 114202−114206. (25) Shubert, V. A.; Zwier, T. S. IR/IR/UV Hole-Burning: Conformation Specific IR Spectra in the Face of UV Spectral Overlap. J. Phys. Chem. A 2007, 111, 13283−13286. (26) Luchow, A.; Spangenberg, D.; Janzen, C.; Jansen, A.; Gerhards, M.; Kleinermanns, K. Structure and energetics of phenol(H2O)(n), n ≤ 7: Quantum Monte Carlo calculations and double resonance experiments. Phys. Chem. Chem. Phys. 2001, 3, 2771−2780. (27) Pérez, C.; Muckle, M. T.; Zaleski, D. P.; Seifert, N. A.; Temelso, B.; Shields, G. C.; Kisiel, Z.; Pate, B. H. Structures of Cage, Prism, and Book Isomers of Water Hexamer from Broadband Rotational Spectroscopy. Science 2012, 336, 897−901. (28) Shields, R. M.; Temelso, B.; Archer, K. A.; Morrell, T. E.; Shields, G. C. Accurate Predictions of Water Cluster Formation, (H2O) (n = 2−10). J. Phys. Chem. A 2010, 114, 11725−11737. (29) Marques, J. M. C.; Llanio-Trujillo, J. L.; Albertí, M.; Aguilar, A.; Pirani, F. Alcali-ion Microsolvation with Benzene Molecules. J. Phys. Chem. A 2012, 116, 4947−4956.

2010/CSD2007-00013, CTQ2012-39132, and CTQ201122923) and from UPV/EHU (UFI 11/23). I.L. thanks the GV for pre- and postdoctoral fellowships and E.J.C. thanks the Spanish Ministry (MICINN) for a “Ramón y Cajal” Contract. Computational resources from the SGI/IZO-SGIker and from i2BASQUE academic network were used for this work. Technical and human support provided by the Laser Facility of the SGIKER (UPV/EHU, MICINN, GV/EJ, ESF) is also gratefully acknowledge. J.M. thanks the CLR for additional support (FOMENTA09/03).



REFERENCES

(1) Franks, N. P.; Lieb, W. R. Molecular and Cellular Mechanisms of General-Anesthesia. Nature 1994, 367, 607−614. (2) Molecular and Basic Mechanisms of Anesthesia; Urban, B. W.; Barann, M., Eds.; Pabst Science Publishers: Lengerich, Germany, 2002. (3) Savjani, K. T.; Gajjar, A. K.; Savjani, J. K. Drug Solubility: Importance and enhancement techniques. ISRN Pharmaceutics 2012, 2012, 1−10. (4) Brunton, L. L.; Lazo, J. S.; Parker, K. L. Goodman & Gilman’s The Pharmacological Basis of Therapeutics; McGraw Hill: New York, NY, 2006. (5) Mohanachandran, P. S.; Sindhumol, P. G.; Kiran, T. S. Enhancement of Solubility and Dissolution Rate: an Overview. Pharm. Globale 2010, 4, 1−10. (6) Pinnamaneni, S.; Das, N. G.; Das, S. K. Formulation approaches for orally administered poorly soluble drugs. Pharmazie 2002, 57, 291−300. (7) Patel, V. R.; Agrawal, Y. K. Nanosuspension: An approach to enhance solubility of drugs. J Adv. Pharm. Technol. Res. 2011, 2, 81− 87. (8) Liua, D.; Feib, X.; Wanga, S.; Jianga, T.; Sua, D. Increasing solubility and dissolution rate of drugs via eutectic mixtures: itraconazole−poloxamer188 system. Asian J. Pharm. Sci. 2006, 1, 213−221. (9) Linn, M.; Collnot, E. M.; Djuric, D.; Hempel, K.; Fabian, E.; Kolter, K.; Lehr, C. M. Soluplus (R) as an effective absorption enhancer of poorly soluble drugs in vitro and in vivo. Eur. J. Pharm. Sci. 2012, 45, 336−343. (10) Choi, J. M.; Kim, H.; Cho, E.; Choi, Y.; Lee, I. S.; Jung, S. Solubilization of haloperidol by acyclic succinoglycan oligosaccharides. Carbohydr. Polym. 2012, 89, 564−570. (11) Ma, D.; Hettiarachchi, G.; Nguyen, D.; Zhang, B.; Wittenberg, J. B.; Zavalij, P. Y.; Briken, V.; Isaacs, L. Acyclic cucurbit[n]uril molecular containers enhance the solubility and bioactivity of poorly soluble pharmaceuticals. Nat. Chem. 2012, 4, 503−510. (12) Kawakami, K.; Ebara, M.; Izawa, H.; Sanchez-Ballester, N. M.; Hill, J. P.; Ariga, K. Supramolecular Approaches for Drug Development. Curr. Med. Chem. 2012, 19, 2388−2398. (13) Hu, L. D.; Yang, J. X.; Liu, W.; Li, L. Preparation and evaluation of ibuprofen-loaded microemulsion for improvement of oral bioavailability. Drug Delivery 2011, 18, 90−95. (14) Kim, J. Y.; Kim, S.; Pinal, R.; Park, K. Hydrotropic polymer micelles as versatile vehicles for delivery of poorly water-soluble drugs. J. Controlled Release 2011, 152, 13−20. (15) Cui, Y.; Xing, C. Y.; Ran, Y. Q. Molecular Dynamics Simulations of Hydrotropic Solubilization and Self-Aggregation of Nicotinamide. J. Pharm. Sci. 2010, 99, 3048−3059. (16) Leon, I.; Cocinero, E.; Millan, J.; Jaeqx, S.; Rijs, A.; Lesarri, A.; Castano, F.; Fernandez, J. A. Exploring Microsolvation of the Anesthetics Propofol. Phys. Chem. Chem. Phys. 2012, 14, 4398. (17) Leon, I.; Cocinero, E. J.; Millan, J.; Rijs, A. M.; Usabiaga, I.; Lesarri, A.; Castano, F.; Fernandez, J. A. A Combined Spectroscopic and Theoretical Study of Propofol·(H2O)3. J. Chem. Phys. 2012, 137, 074303. (18) Leon, I.; Cocinero, E. J.; Lesarri, A.; Castaño, F.; Fernandez, J. A. A Spectroscopic Approach to the Solvation of Anesthetics in Jets: Propofol(H2O)n, n = 4−6. J. Phys. Chem. A 2012, 116, 8934−8941. 3404

dx.doi.org/10.1021/jp401386y | J. Phys. Chem. A 2013, 117, 3396−3404