Evaluation of the aggregation process in a mixture of propofol and benzocaine

I. León ab, A. Lesarri c and J. A. Fernández *a
aDepartment of Physical Chemistry, Faculty of Science and Technology, University of the Basque Country, Barrio Sarriena s/n, 48940 Leioa, Spain. E-mail: josea.fernandez@ehu.es
bGrupo de Espectrocopía Molecular (GEM), Edificio Quifima, Laboratorios de Espectroscopia y Bioespectroscopia, Unidad Asociada CSIC, Parque Científico UVa, Universidad de Valladolid, 47011, Valladolid, Spain
cDepartment of Physical Chemistry and Inorganic Chemistry, Faculty of Science, University of Valladolid, Valladolid E-47005, Spain

Received 11th July 2018 , Accepted 10th August 2018

First published on 10th August 2018


We report on a mass-resolved IR spectrosopic study on propofol–benzocaine aggregates. This is a complex system due to the several conformational isomers that both monomers may adopt and to the combination of functional groups they present, which allow the molecules to interact in many possible ways. However, our results demonstrate that a single conformation is favored for each stoichiometry. In the heterodimer, propofol acts as a proton donor to the ester group of benzocaine, while the whole cluster is stabilized by dispersive forces. These dispersive forces account for an important part of the system's stabilization energy as the calculations suggest. Propofol does not show any affinity for the amino group of benzocaine, even when a second molecule of propofol is introduced. These results demonstrate the difficulty in anticipating the aggregation preferences of even small organic molecules.


Introduction

While covalent bonds involve sharing electrons between interacting atoms, non-covalent interactions (NCIs) arise from the interactions between the electronic clouds of nearby molecules.1 The former produce very stable bonds that create new molecules or modify their chemical nature. The latter, on the other hand, are the core of flexible and dynamic structures, as the strength of the NCIs is significantly smaller than that of the real covalent bonds.1 They allow the molecules to create a dynamical environment in a large number of inorganic and organic processes, but especially in living organisms, where they are involved in most of the processes. The reduced interaction energy of the NCIs makes them a valuable tool to create stable-enough structures to carry out biological functions, while at the same time, they are easily dismantled when required. There are a large number of examples in the literature that highlight their importance: from the structure of water and ice, to the three-dimensional arrangement of macromolecular structures, such as DNA or proteins.2–4

Many experimental and computational research groups are contributing to characterizing NCIs.5–24 Our group is also engaged in such a task, using a combined experimental and computational approach. With the aid of supersonic expansions, we create the required conditions of temperature and absence of interfering interactions to form molecular aggregates that are free to adopt the most energetically efficient configurations. Using spectroscopic probes, we extract physical observables from such systems that are later compared with high-level quantum chemical calculations, allowing us to determine the structure of the aggregates and to evaluate the relative importance and strength of the NCI at play. In this way, we solved in the past several systems, such as eugenol and guaiacol dimers25 and their aggregates with water,26 solvation of propofol,27–29 the complex nucleation process of aniline10,30 or even the interaction preferences of cytosine.31

Following such studies, we present here an interesting system to illustrate not only the importance of non-covalent interactions, but also how small modifications of a simple molecule may lead to very different aggregation tendencies. Certainly, the final shape of a molecular aggregate is the resultant of all the forces at play; subtle changes in that balance may lead to completely different superstructures.32 Probably the most illustrative example of such a phenomenon is the conformational change that the ion channels of the cell membrane experience when a ligand attaches to the correct site (the active center): the interactions of an agonist with a limited portion of the superstructure suffices to modify its conformation (state) from closed to open.33–35

Here we explore the interactions between benzocaine (ethyl 4-aminobenzoate) and propofol (2,6-diisopropylphenol), Scheme 1. Despite the fact that both molecules are used as anesthetics, their combination may not be of pharmaceutical relevance, as their joint administration is not common. However, from a spectroscopic point of view, this is a very interesting system: both molecules have several interaction places and a combination of functional groups that suggest the existence of numerous interactions and conformational isomers. On one hand, propofol has a hydroxyl group, which is slightly acidic and therefore, prone to form hydrogen bonds as a proton-donor.


image file: c8cp04386h-s1.tif
Scheme 1 Propofol (2,6-diisopropylphenol) and benzozaine (ethyl 4-aminobenzoate).

However, it is somehow shielded by two hydrophobic isopropyl groups that limit its proton-donor functionality. On the other hand, benzocaine has a (slightly basic) amino group that can act both as a proton-donor or as a proton-acceptor. Furthermore, the oxygens of the ester group of benzocaine are also available to establish hydrogen bonds. Thus, one would anticipate that the main interaction between these molecules would be the formation of moderately strong hydrogen bonds. However, both molecules are also aromatic, and therefore stacking interactions will also be at play, modulating the final shape of the aggregate.

To complicate the picture, propofol may adopt five different conformations, differing in the relative orientation of its isopropyl groups, with two of those isomers being almost isoenergetic.29 Benzocaine also presents two conformers: while trans-benzocaine has the ethyl group lying in the plane of the molecule, the ethyl group forms an angle with the plane of the molecule in gauche-benzocaine.36

As it will be demonstrated, despite the multiple conformations that the aggregates may adopt, a single family of structures is favored, whose shape is the resultant of all the interactions at play.

Methods

Experimental

The experimental set-up employed in this work has been described in detail previously37 and therefore we will highlight here the most relevant aspects only. It consists of a time-of-flight mass spectrometer (Jordan Inc.), three sets of Nd/YAG/dye/doubling unit lasers (one Quantel Brilliant B/Fine Adjustment Pulsare and two Quantel Brilliant B/Lambda Physic Scanmates), and an Nd/YAG/differential frequency mixing unit (Innolas/Lioptec PulsarePro), together with electronics for synchronization and signal acquisition and storage. Separate mixtures of propofol (97%, Sigma-Aldrich) and benzocaine (99%, Sigma-Aldrich) were placed in a sample compartment in the gas line that feeds a pulsed valve (Jordan Inc.) and gently warmed up to 90 °C to achieve enough vapor concentration. The mixture of propofol and benzocaine was expanded into the ionization chamber of the mass spectrometer using a He backing pressure of 1 bar. The supersonic beam was skimmed before entering the ionization region of the mass spectrometer, where the 2-color R2PI (resonance enhanced two-photon ionization) and ion-dip infrared (IDIR) experiments were carried out. The laser energy of the UV lasers was maintained around 50–100 μJ per pulse, in order to prevent multi-photon processes.

Two-color R2PI experiments were done spatially and temporally overlapping the lasers and the fluence was reduced until a signal was obtained only with both lasers; i.e., blocking of one of the laser beams resulted in a complete loss of signal. In the IDIR experiments, the pump UV/IR laser was fired 100 ns prior to the probe UV laser and at half of its repetition rate, so real time active subtraction could be performed. Care was taken to achieve a perfect overlap between the population of molecules probed by both UV and IR lasers, to achieve the maximum depopulation possible without saturation.

Computational

The calculations were carried out in two stages. In the first one, molecular mechanics (MM) was used to explore the conformational landscape for the interaction of the molecules. The procedure used may be found in ref. 37. In the first step, Merk's Molecular Force Field, as described in Macromodel software (http://www.schrodinger.com), was used to map the conformational landscape. The structures were generated through a combination of the “Large scales Low Mode” (which uses frequency modes to create new structures) and Monte Carlo Minimization procedures. This method rendered a large number of species, some of them with minimal structural differences and therefore they were clustered into families using our own clustering code based on XCluster.38 This is the same procedure used in previous works on larger systems. However, an added difficulty was found in benzocaine–propofol compared with previous studies. The combination of multiple conformational isomers of the monomers with several interaction points resulted in a significant increase in the number of different structures found for each aggregate. Thus, in propofol⋯benzocaine, 1[thin space (1/6-em)]000[thin space (1/6-em)]000 trajectories (minimizations) produced nearly 1000 structures, with some of them appearing >1000 times and most of the rest >100. For propofol2⋯benzocaine, the same number of runs produced more than 50[thin space (1/6-em)]000 structures, with barely any repetition. In conclusion, the present system offers a tremendous variability compared with other systems of similar or even larger size. Furthermore, a homemade clustering algorithm had to be used, because Schrödinger's hierarchical clustering routine has a soft limit of ∼15[thin space (1/6-em)]000 input items. Even when eliminating redundant conformers using a RMSD setting in the conformational search it resulted in a number higher than this limit.

The complexity of the system forced us to select a large number of the species found using MM for full optimization using DFT. At least two representative structures of each family, together with selected structures that appeared only once in a 30 kJ mol−1 stability window were submitted to full optimization at M06-2X/6-31+G(d) using the Gaussian 09 suite of programs.39 The structures were tested as true minima by a normal mode analysis. All the energy values presented in this work are ZPE corrected. Analysis of the nature of the strength of the interactions was carried out using NCIplot.40,41

Results

Electronic excited state

As stated above, the electronic excited state of propofol⋯benzocaine was recorded using the 2c-R2PI technique. Fig. 1 shows a comparison between the electronic excited state of propofoln⋯benzocaine (n = 1, 2) and those of the monomers. Propofol presents complex spectroscopy, which is the result of the contribution of four different conformational isomers to the spectrum.29 Likewise, contributions from the two conformers of benzocaine are clearly visible in its REMPI spectrum.36,42 A combination of both monomers results in a broad electronic spectrum of propofol⋯benzocaine that presents only a few resolved transitions starting at 34[thin space (1/6-em)]002 cm−1. Despite the fact that the nature of the spectrum does not enable us to clearly identify the electronic origin of the S1 ← S0 transition, the peaks around 34[thin space (1/6-em)]002 cm−1 highlight that there is a red-shift with respect to the bare benzocaine molecule. Addition of a second propofol molecule results in the disappearance of the discrete bands, leaving only the broad absorption band.
image file: c8cp04386h-f1.tif
Fig. 1 2-Color R2PI spectra of propofol, benzocaine, phenol⋯benzocaine, propofol⋯benzocaine and propofol2⋯benzocaine from lower to upper trace, respectively.

IR/UV spectroscopy

Although the electronic spectra of the aggregates are not very informative, it is still possible to extract relevant structural information using the technique of IR/UV double resonance. Tuning the probe laser at 34[thin space (1/6-em)]002 cm−1 we were able to record the mass-selected IR spectrum of the propofol⋯benzocaine heterodimer shown in Fig. 2. Despite the weakness of the R2PI signal and the complexity of obtaining an IR spectrum of such a large cluster, an IR spectrum with an excellent S/N ratio was obtained. Three peaks are clearly identified: the peaks at 3443 and 3537 cm−1 are due to the free NH symmetric and anti-symmetric stretching modes respectively, as the comparison with the bare molecule of benzocaine shows.43,44 The third peak located at 3478 cm−1 is due to the hydroxyl group of propofol. In the most stable conformer of the monomer, this band appears at ∼3651 cm−1. The red shifted position of this band in the spectrum of propofol⋯benzocaine indicates the existence of an interaction of moderate strength in the aggregate. Certainly, comparison with the IR/UV spectrum of phenol⋯benzocaine shows a significantly shorter red shift. As determined in previous work,44 the OH group of phenol establishes a strong hydrogen bond with the amino group of benzocaine. So either propofol interacts with a different site or its O–H⋯NH2 interaction presents a reduced strength due, very likely, to steric effects.
image file: c8cp04386h-f2.tif
Fig. 2 IR/UV double resonance spectra of trans-benzocaine, the most stable isomer of propofol, phenol⋯benzocaine, propofol⋯benzocaine and propofol2⋯benzocaine.

The addition of a second propofol molecule to form the propofol2⋯benzocaine heterotrimer results in a completely different IR spectrum, in which only two strong bands, presumably due to the two OH moieties, are visible. The one at 3561 cm−1 is slightly broader than expected, but the band at 3390 cm−1 is significantly broad, indicating a large anharmonicity, probably due to the formation of a strong hydrogen bond. The bands due to the NH2 stretches are not visible, perhaps due to a reduced intensity compared with the OHs or because they are convolved with the broad OH stretching bands. At this point, comparison with the computed structures is required to understand the experimental spectra.

Structure of propofol⋯benzocaine

The several conformers that the monomers may adopt anticipated the existence of numerous isomers for propofol⋯benzocaine. In line with this, 44 conformational isomers were found for propofol⋯benzocaine using DFT, in a 30 kJ mol−1 window, highlighting the conformational complexity of the system. The six most representative structures are shown in Fig. 3 and their predicted IR spectra are compared with the experimental spectrum in Fig. 4. The rest of the calculated structures together with their predicted IR spectra can be found in the ESI.
image file: c8cp04386h-f3.tif
Fig. 3 Most representative structures of propofol⋯benzocaine calculated at the M06-2X/6-31+G(d) level, with their relative stability in kJ mol−1 in brackets. All the computed structures may be found in the ESI.

image file: c8cp04386h-f4.tif
Fig. 4 A comparison between the propofol⋯benzocaine IR/UV spectrum and the predicted IR transitions of some selected structures. A correction factor of 0.955 was used to account for anharmonicity.

Against what was anticipated, the most stable conformer, named structure 1, presents the propofol molecule sitting on top of the benzocaine molecule with a OH⋯O[double bond, length as m-dash]C hydrogen bond. This situation allows the two molecules to establish additional interactions between their respective aromatic rings and the aliphatic moieties of their partner. The extra stabilization given by these types of dispersive forces was discussed before45 and accounts for a significant portion of the interaction energy. In fact, the difference between structure 1 and the second most stable structure 2, is that, in the latter, benzocaine adopts a gauche configuration, reducing the interaction between the ethyl group of benzocaine and the aromatic ring of propofol. Such reduction results in a decrease of the binding energy of 2.5 kJ mol−1.

Interestingly, when propofol acts as a proton donor to the amino group of benzocaine in an OH⋯NH interaction such as in structure 6, the stability of the cluster is 4.6 kJ mol−1 smaller than that of the global minimum. Apparently, such a conformation has a similar stability to that of propofol lying on top of benzocaine, forming an O–H⋯π interaction (structure 7). Probably the two isopropyl groups that protect the OH moiety of propofol are responsible for the reduced interaction energy in structure 6. Finally, the interaction of propofol with benzocaine's oxygen on the ether group, as in structure 12, is the least stable interaction.

The comparison between the experimental IR spectrum and the predicted ones, Fig. 4, leads to an unequivocal assignment of the experimental spectrum to computed structure 1. Thus, the bands at 3443 and 3537 cm−1 correspond to the symmetric and antisymmetric stretches of the NH2 group, that does not take part in the intermolecular interactions, also explaining why they appear almost at the same position as in bare benzocaine's spectrum. The band at 3478 cm−1 is assigned to the stretching mode of propofol's OH moiety.

Structure of propofol2⋯benzocaine

While recording the spectrum of propofol2⋯benzocaine required fine optimization of all the experimental parameters, the calculations required to assign the IR/UV spectrum were equally challenging. The exploration of the conformational landscape using molecular mechanics points to an extraordinary complexity for a system of this size, which forced us to reduce the energetic window used to filter the structures produced by molecular mechanics, leading to a risk of missing important conformations during the stage of PES exploration. With the above-mentioned constraints, 60 structures were optimized using DFT: the 30 most stable structures found by MM and the 30 most different families according to the clustering parameters. Among them, 49 resulted in different structures and true minima, as it was demonstrated by the normal mode calculations. Fig. 5 shows some of the most representative structures. Although all the structures found below 10 kJ mol−1 from the global minimum present a different arrangement of the three molecules, they are all based on the same OH⋯OH⋯O[double bond, length as m-dash]C pattern. These structures are further stabilized by CH3⋯π interactions, highlighting once more the importance of these interactions. It is worthnoting that propofol does not show much interest in interacting with benzocaine's NH group, despite the fact that it is a good proton acceptor. Even the structures with cyclic hydrogen bond networks that are the most stable conformers in other systems, appear high in energy in propofol2⋯benzocaine (structure 16). The oxygen of the ethoxyl group of benzocaine does not seem to be very appealing to propofol, as the structures based on such interactions lie high in energy, according to our calculations.
image file: c8cp04386h-f5.tif
Fig. 5 Some selected structures of propofol2⋯benzocaine, calculated at the M06-2X/6-31+G(d) level, with their relative stability in kJ mol−1 in brackets. The complete set of computed structures may be found in the ESI.

Fig. 6 shows a comparison between the experimental IR spectrum and the predicted ones for the structures in Fig. 5. An excellent agreement between the experimental trace and the DFT prediction for structure 3 is observed. Certainly, the calculations correctly predict the appearance of only two bands in the spectrum, due to the stretches of the hydroxyl groups and the loss of intensity of the stretches of the amino group. Nevertheless, the broad band at 3390 cm−1 may hide the contribution from more than one species, as its asymmetric shape seems to indicate. Thus, the existence of an additional isomer, but with very similar structure, may not be completely ruled out.


image file: c8cp04386h-f6.tif
Fig. 6 Comparison between the propofol2⋯benzocaine IR/UV spectrum and the predicted IR transitions of some selected structures. A correction factor of 0.955 was used to account for anharmonicity.

Agreement with the spectrum predicted for structure 5 is also good. A close inspection of such an isomer shows that it is not very different from structure 3. Many other structures with slight modifications in the relative positions of the three molecules may exist, whose predicted spectra match the experimental observations. However, this fact does not alter the main conclusion, which is that the second propofol molecule prefers to attach to the first propofol, forming an OH⋯OH⋯O[double bond, length as m-dash]C hydrogen bond network, and that the rest of the interacting sites result in higher-energy conformers that were not detected in the experiment.

Discussion

Propofol has been demonstrated to be a complex system both from the computational and from the experimental point of view. However, assignment of the experimental spectra was straightforward, allowing us to propose a single conformational isomer to each stoichiometry. Although, in principle, propofol⋯benzocaine may be envisioned as an acid–base interaction, the isomers detected do not follow such rules and propofol prefers to interact with the ester side of benzocaine.

To explore the reasons behind such aggregation preferences, we compared the results from this work with those from previous work on phenol⋯benzocaine44 and toluene⋯benzocaine,43 and we analyzed the strength and nature of the non-covalent interactions using NCIplot (Fig. 7 and Table 1). Two isomers were found for toluene⋯trans-benzocaine, one with a combination of C–H⋯π and NH2⋯π interactions, and the second one attached exclusively by π⋯π interactions. The small difference in stability between them, ∼4 kJ mol−1 highlights the importance of the stacking interaction in these types of molecules. Phenol⋯benzocaine, on the other hand, favors formation of hydrogen bonds. The isomer detected in ref. 44 presents the expected O–H⋯NH2 leading interaction and at the same time, a secondary C–H⋯π interaction, accounting for 33 kJ mol−1 of binding energy. Surprisingly, our calculations point to a second isomer as the global minimum, but such a structure was not detected in previous work. In this second isomer, phenol is interacting with the C[double bond, length as m-dash]O group of benzocaine, forming a strong hydrogen bond which gives the complex ∼10 kJ mol−1 of extra stabilization.


image file: c8cp04386h-f7.tif
Fig. 7 Comparison between the analysis of the interactions using NCIplot for the assigned structure of propofol·benzocaine and those of phenol·benzocaine and toluene·benzocaine.
Table 1 Complexation energies of toluene⋯benzocaine, phenol⋯benzocaine and propofol⋯benzocaine calculated at the M06-2X/6-31+G(d) level
Aggregate Interaction D 0 (kJ mol−1)
a Structure detected in ref. 43. b Structure detected in ref. 44. c From this work.
Toluene⋯benzocaine NH2⋯π + CH3⋯πa 31.53
π⋯πa 27.21
Phenol⋯benzocaine OH⋯O[double bond, length as m-dash]C 42.96
OH⋯NH2b 33.03
Propofol⋯benzocaine OH⋯O[double bond, length as m-dash]Cc 61.04
OH⋯NH2 50.37


The stability of the isomer of propofol⋯benzocaine detected in the present work follows the same trend as phenol⋯benzocaine: interaction with the C[double bond, length as m-dash]O bond leads to a structure ∼10 kJ mol−1 more stable than the interaction with the NH2 moiety. The hydrogen bond is also weaker than in the case of phenol⋯benzocaine, as the position of the OH stretching vibration in the experimental spectrum shows (Fig. 2). However, propofol⋯benzocaine presents a significantly higher stability, due to the extensive interactions with the aromatic rings of both molecules, as highlighted by the NCI analysis (green surfaces in Fig. 7). Thus, the stability of the detected isomer, 61 kJ mol−1, is not far from the binding energy of phenol (∼43 kJ mol−1) plus toluene (∼27 kJ mol−1) to benzocaine. This is the resultant of a weaker hydrogen bond but stronger C–H⋯π interactions and it nicely illustrates how the introduction of isopropyl substituents can modulate the final structure of an aggregate and its affinity for other molecules.

Addition of the second propofol results in the formation of a cooperative hydrogen bond network, with the subsequent reinforcement of the pre-existing hydrogen bond, as highlighted by the longer red shift in the OH stretching. In the computed structures, the O–H⋯O[double bond, length as m-dash]C distance changes from 1.91 Å to 1.81 Å and the angle moves from 151° to 155° with the addition of the second propofol. It is worth noting that the three molecules prefer the formation of the above-described linear hydrogen bond network instead of a cyclic one, as observed in previous similar systems,27,30,46–49 which adds an extra hydrogen bond to the network. Such a structure is ∼10 kJ mol−1 higher in energy, at the calculation level used in this work. The reason behind this preference is not simple, but it has to be a more favorable combination of interactions. These results encourage us to continue the exploration of the non-covalent interactions in systems of increasing complexity to be able, one day, to predict the aggregation preferences of the organic molecules.

Conclusions

Propofol–benzocaine has been demonstrated to be a complex but interesting system. The several conformational isomers that the monomers may adopt suggested a complex conformational landscape. However, only a single isomer was detected for the heterodimer and another one for the heterotrimer. Conversely, the calculations showed a complex conformational panorama, with many stable structures populating a narrow stability window. This means that the barriers for isomerization are low, so they can be overcome during the cooling stage, and the system can easily reach the global minimum. It is very likely that the large and relaxed π⋯π and CH⋯π interactions play an important role in such a process.

Propofol exhibited a strong propensity towards interacting with the C[double bond, length as m-dash]O group of benzocaine and to form linear hydrogen bond networks, against what is the normal rule in this kind of system: formation of a cyclic hydrogen bond network. Somehow, the balance of interactions favors the former type of structures, although the exact reason is not easy to visualize. These results also highlight how difficult it is to a priori predict the behavior of organic molecules and the need for further data on systems of increasing complexity.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The research leading to these results has received funding from the Spanish MINECO and FEDER (CTQ2015-68148). I. L. thanks the Basque Government for a pre-doctoral fellowship. This work has been possible thanks to the support of the computing infrastructure of the i2BASQUE academic network and the SGI/IZO-SGIker network. We would like to thank the technical support from the personnel from the UPV/EHU laser facility. We also would like to thank R. Laplaza and J. Contreras for their help with NCIPlot.

Notes and references

  1. P. Hobza and K. Muller-Dethlefs, Non-Covalent Interactions, RCS Publishing, Cambridge, 2010 Search PubMed.
  2. J. Lehn, Science, 2002, 295, 2400–2403,  DOI:10.1126/science.1071063.
  3. D. A. Dougherty, Science, 1996, 271, 163–168,  DOI:10.1126/science.271.5246.163.
  4. G. J. Bartlett, A. Choudhary, R. T. Raines and D. N. Woolfson, Nat. Chem. Biol., 2010, 6, 615 CrossRef CAS.
  5. C. A. Southern, D. H. Levy, J. A. Stearns, G. M. Florio, A. Longarte and T. S. Zwier, J. Phys. Chem. A, 2004, 108, 4599–4609,  DOI:10.1021/jp0496093.
  6. P. Hobza and J. Sponer, Chem. Rev., 1999, 99, 3247–3276,  DOI:10.1021/cr9800255.
  7. J. Rezac and P. Hobza, Chem. Rev., 2016, 116, 5038–5071,  DOI:10.1021/acs.chemrev.5b00526.
  8. R. A. Jockusch, R. T. Kroemer, F. O. Talbot and J. P. Simons, J. Phys. Chem. A, 2003, 107, 10725–10732,  DOI:10.1021/jp0351730.
  9. F. Dietrich, D. Bernhard, M. Fatima, C. Perez, M. Schnell and M. Gerhards, Angew. Chem., Int. Ed. Engl., 2018, 57, 9534–9537,  DOI:10.1002/anie.201801842.
  10. S. Habka, V. Brenner, M. Mons and E. Gloaguen, J. Phys. Chem. Lett., 2016, 7, 1192–1197,  DOI:10.1021/acs.jpclett.6b00454.
  11. A. Bhattacherjee and S. Wategaonkar, Phys. Chem. Chem. Phys., 2016, 18, 27745–27749,  10.1039/c6cp05469b.
  12. W. Caminati, L. Evangelisti, G. Feng, B. M. Giuliano, Q. Gou, S. Melandri and J. U. Grabow, Phys. Chem. Chem. Phys., 2016, 18, 17851–17855,  10.1039/c6cp01059h.
  13. J. A. Frey, C. Holzer, W. Klopper and S. Leutwyler, Chem. Rev., 2016, 116, 5614–5641,  DOI:10.1021/acs.chemrev.5b00652.
  14. A. Poblotzki, H. C. Gottschalk and M. A. Suhm, J. Phys. Chem. Lett., 2017, 8, 5656–5665,  DOI:10.1021/acs.jpclett.7b02337.
  15. R. A. Mata and M. A. Suhm, Angew. Chem., Int. Ed. Engl., 2017, 56, 11011–11018,  DOI:10.1002/anie.201611308.
  16. T. Forsting, H. C. Gottschalk, B. Hartwig, M. Mons and M. A. Suhm, Phys. Chem. Chem. Phys., 2017, 19, 10727–10737,  10.1039/c6cp07989j.
  17. K. K. Mishra, S. K. Singh, P. Ghosh, D. Ghosh and A. Das, Phys. Chem. Chem. Phys., 2017, 19, 24179–24187,  10.1039/c7cp05265k.
  18. Gas-Phase IR Spectroscopy and Structure of Biological Molecules, ed. A. M. Rijs and J. Oomens, Springer International Publishing, Heidelberg, New York, Dordrecht, London, 2015 Search PubMed.
  19. O. Mo, M. Yanez, I. Alkorta and J. Elguero, J. Chem. Theory Comput., 2012, 8, 2293–2300,  DOI:10.1021/ct300243b.
  20. Y. Hu, J. Guan and E. R. Bernstein, Mass Spectrom. Rev., 2013, 32, 484–501,  DOI:10.1002/mas.21387.
  21. M. Schmies, A. Patzer, M. Schuetz, M. Miyazaki, M. Fujii and O. Dopfer, Phys. Chem. Chem. Phys., 2014, 16, 7980–7995,  10.1039/c4cp00401a.
  22. K. Mizuse, N. Mikami and A. Fujii, Angew. Chem., Int. Ed. Engl., 2010, 49, 10119–10122,  DOI:10.1002/anie.201003662.
  23. A. M. Rijs, M. Kabelac, A. Abo-Riziq, P. Hobza and M. S. de Vries, ChemPhysChem, 2011, 12, 1816–1821 CrossRef CAS.
  24. A. Shahi and E. Arunan, Phys. Chem. Chem. Phys., 2014, 16, 22935–22952,  10.1039/c4cp02585g.
  25. A. Longarte, C. Redondo, J. A. Fernández and F. Castaño, J. Chem. Phys., 2005, 122 Search PubMed.
  26. A. Longarte, I. Unamuno, J. A. Fernández, F. Castaño and C. Redondo, J. Chem. Phys., 2004, 121, 209–219 CrossRef CAS.
  27. I. León, J. Millán, E. J. Cocinero, A. Lesarri and J. A. Fernández, Angew. Chem., Int. Ed., 2013, 52, 7772–7775 CrossRef.
  28. I. León, J. Millán, E. J. Cocinero, A. Lesarri and J. A. Fernández, Angew. Chem., Int. Ed., 2014, 53, 12480–12483,  DOI:10.1002/anie.201405652.
  29. I. León, E. Cocinero, J. Millán, S. Jaeqx, A. Rijs, A. Lesarri, F. Castaño and J. A. Fernández, Phys. Chem. Chem. Phys., 2012, 14, 4398 RSC.
  30. I. León, I. Usabiaga, P. F. Arnaiz, A. Lesarri and J. A. Fernández, Chem. – Eur. J., 2018, 24, 10291–10295,  DOI:10.1002/chem.201802015.
  31. J. González, I. Usabiaga, P. F. Arnaiz, I. León, R. Martínez, J. Millán and J. A. Fernández, Phys. Chem. Chem. Phys., 2017, 19, 8826–8834,  10.1039/C6CP08476A.
  32. K. Müller-Dethlefs and P. Hobza, Chem. Rev., 2000, 100, 143–168,  DOI:10.1021/cr9900331.
  33. G. M. Lipkind and H. A. Fozzard, Mol. Pharmacol., 2005, 68, 1611–1622 CAS.
  34. V. Yarov-Yarovoy, J. Brown, E. M. Sharp, J. J. Clare, T. Scheuer and W. A. Catterall, J. Biol. Chem., 2001, 276, 20–27,  DOI:10.1074/jbc.M006992200.
  35. V. Yarov-Yarovoy, J. C. McPhee, D. Idsvoog, C. Pate, T. Scheuer and W. A. Catterall, J. Biol. Chem., 2002, 277, 35393–35401,  DOI:10.1074/jbc.M206126200.
  36. E. Aguado, A. Longarte, E. Alejandro, J. A. Fernández and F. Castaño, J. Phys. Chem. A, 2006, 110, 6010–6015 CrossRef CAS.
  37. I. Usabiaga, J. Gonzalez, P. F. Arnaiz, I. León, E. J. Cocinero and J. A. Fernández, Phys. Chem. Chem. Phys., 2016, 18, 12457–12465,  10.1039/C6CP00560H.
  38. P. S. Shenkin and D. Q. McDonald, J. Comput. Chem., 1994, 15, 899–916,  DOI:10.1002/jcc.540150811.
  39. See ref. S1 in ESI.
  40. J. Contreras-García, E. R. Johnson, S. Keinan, R. Chaudret, J. Piquemal, D. N. Beratan and W. Yang, J. Chem. Theory Comput., 2011, 7, 625–632,  DOI:10.1021/ct100641a.
  41. E. R. Johnson, S. Keinan, P. Mori-Sánchez, J. Contreras-García, A. J. Cohen and W. Yang, J. Am. Chem. Soc., 2011, 132, 6498–6506,  DOI:10.1021/ja100936w.
  42. B. D. Howells, J. McCombie, T. F. Palmer, J. P. Simons and A. Walters, J. Chem. Soc., Faraday Trans., 1992, 88, 2595–2601 RSC.
  43. E. Aguado, I. León, J. Millán, E. J. Cocinero, S. Jaeqx, A. M. Rijs, A. Lesarri and J. A. Fernández, J. Phys. Chem. B, 2013, 117, 13472–13480,  DOI:10.1021/jp4068944.
  44. E. Aguado, I. León, E. J. Cocinero, A. Lesarri, J. A. Fernández and F. Castaño, Phys. Chem. Chem. Phys., 2009, 11, 11608–11616 RSC.
  45. I. León, J. Millán, F. Castaño and J. A. Fernández, ChemPhysChem, 2012, 13, 3819–3826 CrossRef PubMed.
  46. I. León, E. J. Cocinero, J. Millán, A. M. Rijs, I. Usabiaga, A. Lesarri, F. Castaño and J. A. Fernández, J. Chem. Phys., 2012, 137, 074303 CrossRef PubMed.
  47. I. León, R. Montero, A. Longarte and J. A. Fernández, J. Chem. Phys., 2013, 139 Search PubMed.
  48. I. León, R. Montero, A. Longarte and J. A. Fernández, Phys. Chem. Chem. Phys., 2015, 17, 2241–2245,  10.1039/C4CP03667K.
  49. I. León, P. F. Arnaiz, I. Usabiaga and J. A. Fernández, Phys. Chem. Chem. Phys., 2016, 18, 27336–27341,  10.1039/C6CP04373A.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c8cp04386h

This journal is © the Owner Societies 2019