Article pubs.acs.org/JPCB
Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX
Insights into the Recognition of Phosphate Groups by Peptidic Arginine from Action Spectroscopy and Quantum Chemical Computations Published as part of The Journal of Physical Chemistry virtual special issue “F. Javier Aoiz Festschrift”. Juan Ramón Avilés-Moreno,† Giel Berden,‡ Jos Oomens,‡ and Bruno Martínez-Haya*,† †
Department of Physical, Chemical and Natural Systems, Universidad Pablo de Olavide, 41013 Seville, Spain Radboud University, Institute for Molecules and Materials, FELIX Laboratory, Toernooiveld 7c, 6525ED Nijmegen, The Netherlands
Downloaded via UNIV NOTTINGHAM on August 27, 2019 at 01:59:17 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
S Supporting Information *
ABSTRACT: The side group of the amino acid arginine is typically in its guanidinium protonated form under physiological conditions and participates in a broad range of ligand binding and charge transfer processes of proteins. The recognition of phosphate moieties by guanidinium plays a particularly key role in the interactions of proteins with ATP and nucleic acids. Moreover, it has been recently identified as the driving force for the inhibition of kinase phosphorilation activity by guanidinium derivatives devised as potential anticancer agents. We report on a fundamental investigation of the interactions and coordination arrangements formed by guanidinium with phosphoric, phosphate, and pyrophosphate groups. Action vibrational spectroscopy and ab initio quantum chemical computations are employed to characterize the conformations of benchmark positively charged complexes isolated in an ion trap. The multidentate structure of guanidinium and of the phosphate groups gives rise to a rich conformational landscape with a particular relevance of tweezer−like configurations, where phosphate is effectively trapped by two guanidinium cations. The pyrophosphate complex incorporates a Na+ cation, which serves to compare the interactions associated with the localized versus diffuse charge distributions of the alkali cation and guanidinium, respectively, within a common supramolecular framework.
■
INTRODUCTION
of guanidinium to phosphates promotes the recognition of nucleic acids and has led to important developments in areas of biotechnology related to sensing and transmembrane transport of biomolecular and genetic material.8−16 Guanidinium− phosphate interactions have gained particular recognition due to their key role in the control of ATP-mediated phosphorylation reactions, with remarkable implications in the design of novel antitumor agents. Guanidinium derivatives have been shown to be capable of acting as inhibitors of kinase enzymatic action, which is often overactivated in cancer cells.17 Moreover, the phosphorylation of arginine residues has been identified as a key regulatory factor for protein degradation in Grampositive bacteria.18 This investigation exposes the versatile conformational landscape spanned by the low-energy coordination arrangements of guanidinium with phosphoric, phosphate, and pyrophosphate moieties in isolated supramolecular complexes. Figure 1 represents the structure of the benchmark species involved in this investigation. Favorable ionic interactions largely follow from the matching between the spacing of the
The differentiated interaction of guanidinium with molecular anions lies at the heart of the biochemical activity displayed by arginine-rich proteins in a broad range of enzymatic and metabolic processes.1 The planar geometry of guanidinium and the delocalization of positive charge among its five NHδ+ bonds are particularly suitable for coordination with the diffuse anionic electronic structure of phosphates, carboxylates, or sulfates. In fact, in most events of binding of these ions, guanidinium competes favorably with ubiquitous cationic species such as Na+, K+ or NH4+ and protonated amines. These features have served as inspiration for the development of synthetic guanidinium-based molecular materials for sensing and catalysis.2−4 Recent studies in our group have corroborated the favorable interactions of guanidinium with carboxylic side groups in a cyclic polyether.5−7 In those systems, guanidinium displayed a capability for binding both to the ether and to the carboxylate groups of the host macrocycle. Nevertheless, when guanidinium and protonated amine −NH3+ groups were both present in the guest cation (as in the free amino acid arginine), guanidinium−carboxylate and ammonium−ether interactions prevailed. Here, we focus on guanidinium−phosphate interactions, a topic of particular biochemical relevance. The efficient binding © XXXX American Chemical Society
Received: June 29, 2019 Revised: August 13, 2019 Published: August 15, 2019 A
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
acid. The resulting ions were pulse injected into the ion trap for storage at room temperature. Mass selection of the following three complexes with net charge +1 was performed in the present study: (1) dPa0·G+ (m/z = 228) (2) dP−·(G+)2 (m/z = 301) (3) pP2−·(G+)2·Na+ (m/z = 347) Here, G+, dPa0, dP−, and pP2− stand for methyl guanidinium, diethyl phosphoric acid, diethyl phosphate, and pyrophosphate, respectively (see Figure 1). In the IRMPD experiments, the mass-selected complexes are stored in the ion trap and thermalized with He background gas at room temperature. For this study, it will be assumed that thermal equilibrium is reached in the ion trap and Boltzmann population ratios associated with the energetics predicted by the quantum chemical computations will be indicated when discussing the potential contribution of the different conformers to the experimental signal. It has been pointed out that solvent effects or operating conditions of the ion source may have an influence on the relative population of conformers lying close in energy, occasionally demanding a statistical analysis of the computational structures more extended than the one performed here.22 The stored ions were irradiated with two free-electron laser infrared macro-pulses. Each macropulse is approximately 5 μs long, has an energy of 20−70 mJ, and consists of a train of micropulses with a repetition rate of 1 GHz. The spectral bandwidth of the FELIX radiation is 0.5% of the central IR frequency. The IRMPD spectrum is generated from the normalized IR dissociation yield (Y), as determined from six averaged mass spectra, by plotting the fragment fluence (defined as −ln(1 − Y)) as a function of the IR frequency, with linear corrections of the fragment fluence to account for changes in the laser pulse energy.23 Figure 2 depicts illustrative IRMPD mass spectra recorded for each of the three complexes. Fragment peaks are observed only when the laser frequency matches a vibrational transition of the complex. The dominant fragmentation products are in the three cases related to the release of guanidinium from the complex. The G+ cation (m/z = 74) was observed as dissociation product for the three systems, although it could be monitored in the spectral scans only for the dPa0·G+ complex. For the two heavier complexes, dP−·(G+)2 and pP2−·(G+)2·Na+, the spectrometer required a low mass cutoff for optimal operating conditions that precluded a systematic registration of the G+ cation. In these cases, the most intense IRMPD products involve a net loss of a G unit from the complex (i.e., the product retains a proton from guanidinium upon dissociation), leading to the cationic fragments dPa0·G+ (m/z = 228) and pP−·G+·Na+ (m/z = 274), respectively. Quantum Chemistry Calculations. Ab initio quantum chemical computations at the MP2/6-311++G(d,p) level were employed to characterize the low energy conformations of the ionic complexes. The calculations were carried out with the Gaussian 09 code.24 An initial ensemble of candidate molecular structures was produced by means of simulated annealing with the Universal force field, as implemented in the Materials Studio suite v7.0 (Accelrys Software). About 120−150 nonredundant structures were produced for each complex, which were initially optimized with density functional theory at the B3LYP-D3/6-311++G(d,p) level (B3LYP functional with Grimme’s D3 dispersion correction). The around 30−40 most
Figure 1. Schematic representation of methyl guanidinium (G+) and of the benchmark phosphate species involved in the present investigation, diethyl-phosphoric acid (dPa0), diethyl phosphate (dP−), and pyrophosphate (pP2−).
NH/NH2 groups of guanidinium and those of the PO, POO−, POH, and POP groups. This leads to a rich ensemble of low-energy configurations that includes concerted and bifurcated coordination arrangements. The incorporation of Na+ cations to the coordination structure is also considered in order to assess the preferential binding sites of the “soft” guanidinium cation and the “hard” alkali metal cation with the phosphate moieties. In this study, cationic complexes with an overall single charge are mass-selected and isolated in an ion trap at room temperature for interrogation by means of vibrational action spectroscopy. Details about the experimental and computational methodologies are provided below. It will be shown that this approach provides well-defined molecular benchmarks for the quantum chemical models, leading to fundamental insights into the intrinsic relative affinities and the electronic structures responsible for guanidinium−phosphate interactions. The characterization of the preferred coordination arrangements in isolated complexes is expected to guide predictions of supramolecular behavior in solution. Solvent effects on anion recognition actually constitute an active field of research,3,12,13 and a controversial issue that demands insights from experimental and computational contributions linking gasphase complexation with anion binding in solution and crystal phases.19 In this context, the ensemble of supramolecular conformations derived from this study should conform a valuable reference for the rationalization of guanidinium− phosphate recognition.
■
METHODS IRMPD Spectroscopy. The infrared multiple photon dissociation (IRMPD) spectroscopy experiments were carried out in a quadrupole ion trap mass spectrometer (Bruker AmaZon Speed), coupled to the free-electron laser FELIX.20 IRMPD is a type of action spectroscopy that is based on the detection of photofragments resulting from the sequential absorption of infrared photons at wavelengths resonant with vibrational transitions of the interrogated molecular ion.21 The ionic complexes were produced by means of electrospray ionization of water/methanol solutions of methyl guanidinium hydrochloride and either one of the phosphate presursors, namely, diethyl-phosphoric acid or pyrophosphoric B
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry B
■
Article
RESULTS The most salient conformational features of each of the complexes are discussed below, in light of the experimental and computational information obtained in our study. The multipodal structure of guanidinium has the potential to build concerted interactions with several oxygen atom sites. This yields a rich conformational landscape and a challenging scenario for the microscopic approach employed in this investigation. The five conformations of lowest energy obtained for the three benchmark complexes investigated in this work are depicted in Figures 3, 4, and 5. Those same
Figure 2. Illustrative IRMPD mass spectra obtained with the laser frequency tuned at the vibrational band C of each of the three complexes object of this study (see Figures 3−5). The most intense product peaks registered during the scans, namely, G+ (m/z = 74), dPa0·G+ (m/z = 228), and pP−·G+·Na+ (m/z = 274), are related to guanidinium loss from the complexes (see text for details).
stable conformers were subsequently reoptimized with the MP2 method. The conformers were ranked according to their relative configurational free energies (zero-point corrected electronic energies with thermal and vibrational contributions to entropy at T = 298 K). Natural bond orbital (NBO) analysis25 was employed for a detailed characterization of the electronic orbitals involved in the redistribution of charge associated with the ionic interactions in the different coordination arrangements of the complexes. The main focus of the NBO analysis is set on the σ* antibonding orbitals of the −NH groups of guanidinium and on the lone-pair orbitals of the oxygen atoms of the PO, POH, and POC groups, that are involved in the coordination arrangements. Two qualitative types of prbitals are found among the oxygen lone pairs, namely, a set of globular shaped orbitals (labeled n1, n1′ , etc., for discussion) and a set of lobular shaped ones (labeled n2, n2′, etc.). The theoretical IR spectrum of each conformer was produced by convoluting the normal modes of vibration obtained in the MP2 computation with a line broadening of 25 cm−1 (full width at half-maximum) and a scaling of the MP2 harmonic vibrational frequencies by a factor 0.97.
Figure 3. Summary of the results obtained in this study for the dPa0· G+ complex. Top-left: Conformations of lowest energy dPa1−dPa5, with relative free energies indicated in kJ·mol−1. Top-right: IRMPD spectrum measured for the complex (upper black trace) and IR spectra predicted by the MP2 computation for the different conformers (blue traces). See Table 1 for a qualitative assignment of the vibrational bands A−G. Bottom: Dominant natural bond orbitals involved in the coordination arrangement of the complex (the relative contributions to the stabilization energy are indicated). Charge transfer occurs from globular (n1) and lobular (n2, n3) lonepair orbitals of the oxygen atoms to the antibonding σ* orbitals of the NH bonds of guanidinium.
figures also show the IRMPD spectrum measured for each complex, along with the computational IR spectra associated with the different low-energy conformers. Moreover, an illustration of the NBO analysis of the main electronic orbitals involved in the most relevant supramolecular coordination arrangements is also provided. C
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
dPa0·G+ Complex. The characterization of the dPa0·G+ complex serves to assess the preferred coordination achieved by the guanidinium NHδ+ bonds with the oxygen atom of the neutral PO group, in concurrence with the neighboring oxygen atoms of the POH and POC moieties. Figure 3 depicts the conformations of lowest energy predicted by the MP2 computation for the complex. The most stable configuration, dPa1, corresponds to a concerted coordination of two NH bonds with the PO group of the phosphoric acid. The conformers next in energy, dPa2 and dPa3, are associated with individual coordinations of NH bonds with the oxygen atoms of the PO group and of either the POH group or the phosphoester POC backbone. Finally, conformers dPa4 and dPa5 represent cases in which guanidinium coordinates with the oxygen atoms of the phosphoester and POH groups, with negligible interactions with the PO group. These two latter arragements are not favored by the MP2 computation, which assigns to them energies higher than 35 kJ·mol−1 relative to dPa1. Hence, the MP2 computation points to a preferred coordination of guanidinium with the PO bond, with alternative less stable coordinations with the POH or POC oxygen atom sites. The analysis of the natural bond orbitals involved in the different types of NH···O coordinations provides an intuitive framework for the rationalization of the stability of the low energy conformations of the complex. Figure 3 depicts the most relevant NBOs involved in the dPa1, dPa2, and dPa3 conformers and indicates their corresponding relative contribution to the stabilization energy. The formation of the complex involves the transfer of electronic density from lonepair orbitals of the oxygen atoms of the phosphoric and phosphoester acid moieties, to the antibonding σ* orbitals of the NHδ+ bonds of guanidinium. Three types of lone-pair orbitals participate in the coordination of the PO group with the guanidinium cation, namely, one of globular sp-like geometry, denoted n1, and two virtually p-like lobular orbitals, labeled n2 and n3. Electronic density in these orbitals partly drag charge from the bonding orbitals of the PO bond. The interaction of guanidinium with the POH and POC groups receives contributions through the p-like lone pairs exclusively. Interestingly, the PO n1 → σ* charge redistribution provides the largest individual contribution to the host−guest interactions in the complex, namely, 48% of the total stabilization energy in conformers dPa1, dPa2, and 40% in dPa3. These values are similar to the joint contributions from the PO n2/n3 → σ* charge redistributions, each of which amounts to 15−25% of the total stabilization energy. Figure 3 compares the IRMPD spectrum measured for the complex with the MP2 infrared spectra predicted for the low energy conformers dPa1−dPa5. The experimental IRMPD spectrum features a progression of bands in the 700−1800 cm−1 range that are assigned by the MP2 computation to vibrational modes of the guanidinium and diethyl phosphoric acid frameworks, as described in Table 1. The prominent band A and the weaker band B on the high energy flank of the spectrum correspond to methyl guanidinium vibrations. Diethyl phosphoric acid participates in the vibrational motions associated with the remaining bands C−G. In bands C and F, stretching vibrations of the PO bond and of the POH bond couple to stretching and rocking motions of the CN bonds and NH2 groups of guanidinium, respectively. Bands D, E, and G are associated with stretching modes of the PO CC backbone and to angular motions of the POH and
Figure 4. Summary of the results obtained in this study for the dP−· (G+)2 complex. The experimental and computational information shown is as described in the caption of Figure 3.
Figure 5. Summary of the results obtained in this study for the pP2−· (G+)2·Na+ complex. The experimental and computational information shown is as described in the caption of Figure 3. D
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
vibrational modes may deviate from the overall scaling, as it appears to be often the case for the PO stretching vibration in phosphate and related groups. Within the scope of this study, such deviation is found for the dPa0·G+ complex, as just discussed, and also for the pP2−·(G+)2 ·Na+ complex (see below). In contrast, the computation of the PO stretching modes for the dP−·(G+)2 complex leads to a band C in agreement with experiment within the framework of a general scaling factor. These findings suggest that the accurate modeling of the vibrational modes of phosphate moieties demands specific approaches, depending on the molecular system and possibly on its environment. In order to further assess the perfomance of the B3LYP and MP2 quantum chemical methods, a comparison of the IR spectra predicted by the two types of computations has been included as Supporting Information (Figure S1) for all the systems and conformers discussed in the manuscript. In general terms, the spectra predicted by both approaches are coincident, notably including the PO stretching modes that conform band C. The only significant differences found between the MP2 and B3LYP spectra are related to the splitting of the two components of band A (guanidinium C−N stretching), which is smaller in the B3LYP computation (in better agreement with experiment). For the dPa0·G+ complex, the B3LYP computation does not seem to describe accurately the position of band D (POCC backbone stretching vibrations), which is significantly red-shifted with respect to the MP2 computation and the experiment. Hence, at least for the dPa0·G+ complex, the MP2 approach is required for a reliable comparison wih the experimental IRMPD spectrum. dP−·(G+)2 Complex. Deprotonation of the phosphoric acid facilitates the binding of two guanidinium cations, leading in our study to a ternary dP−·(G+)2 complex with a net single positive charge. In the conformations of lowest energy, depicted in Figure 4, the phosphate moiety occupies the central part of the complex and the guanidinium cations provide and efficient tweezer-like trapping environment for the anion. Figure 4 shows that stable coordination arrangements arise from a variety of concerted interactions of the phosphate and phosphoester oxygen atom sites with the guanidinium cations. Rearrangements connecting the different configurations involve remarkably small net changes in energy, of less than ∼2 kJ·mol−1 (i.e., Boltzmann population ratios of more than 0.45) for the first four conformers. The IRMPD spectrum of the dP−·(G+)2 complex, shown in Figure 4, shares the main features of the analogous spectrum of the dPa0·G+ complex. It is, however, less congested as a consequence of the loss of vibrational modes associated with the POH group (see Table 1), in particular in the range below 1100 cm−1 where isolated bands are neatly observed, in contrast to the partially overlapping band structures of the phosphoric acid complex. The comparison of the IRMPD spectrum with the computational MP2 spectra sorts out the low energy conformers into two groups, depending on the level of agreement achieved. On the one hand, the MP2 spectra of conformers dP1, dP3, and dP4 provide a good match to the experimental bands throughout the whole spectral range investigated. On the other hand, those of conformers dP2 and dP5 display appreciable differences with the experiment, in particular for bands C−F. These two latter conformers have in common a direct interaction of one of the guanidinium moieties with the phosphate POO− group, through a concerted coordination of two NHδ+ bonds. Such a coordination
Table 1. Qualitative Assignment of the Main Bands Observed in the IRMPD Spectra of the Three Complexes Investigated in This Work, As Derived from the Dominant Vibrational Motions Predicted by the MP2 Computationa band A B C D E F G A B C D E F A B C D E F
mode assignment dPa0·G+ Complex CN stretching, NH2+ scissoring CH3 umbrella, CH2 scissoring, NH wagging PO and CN stretching, NH2 rocking stretching of POCC backbone POH bending, OPO and CC stretching POH stretching, CN3 symmetric stretching POC stretching, CH3 rocking dP−·(G+)2 Complex same as above same as above asymmetric POO stretching POC and symmetric POO stretching stretching of POCC backbone, NH wagging POC stretching, NH/NH2 wagging pP2−·(G+)2·Na+ Complex same as above same as above asymmetric POO stretching CN and symmetric POO stretching, N2 rocking POH bending stretching of HOPOPOH backbone, NH wagging
a
The bands are labelled A−G, as indicated in Figures 3−5.
−CH3 groups. This assignment is supported by the excellent matching found between the vibrational bands of the IR spectrum of conformer dPa1 and those of the IRMPD measurement. The spectrum of dPa1 provides in fact the best agreement with experiment among the low energy conformers of the complex. Contributions to the recorded IRMPD signal from the dPa2 and dPa3 conformations are also plausible and would in fact serve to explain the appreciable broadening of some of the bands bands, in particular bands A, C, and D. The overall qualitative picture emerging from these results is a preferential coordination of guanidinium with the PO group (conformer dPa1) and a potential migration of one of the guanidinium NH bonds to the oxygen atoms of the POH group (dPa2) or the POC moiety (dPa3) with a moderate endothermicity (MP2 energies of 5 and 7 kJ·mol−1 , respectively; these relative energies would correspond to predicted Boltzmann population ratios of 0.13 and 0.06 for these conformrs with respect to dPa1 at the room temperature of our experiments). It seems timely to point out that the position of band C predicted by the MP2 computation is somewhat red-shifted (ca. 30−40 cm−1) with respect to the IRMPD experiment. This band is closely related to vibrational modes involving the stretching of the PO bond of the phosphoric acid group (Table 1). Interestingly, previous studies on related systems have reported similar red shifts of the PO stretching vibrational modes, leading to the conclusion that the computed frequencies of this type of mode did not require any scaling factor for comparison with experiment.26 The use of a single effective scaling parameter for the full vibrational spectrum, as applied here, is often reasonable within the moderate spectral resolution attained with a free-electron laser (in our case, up to 5 cm−1). Nevertheless, the computed frequency of particular E
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
found despite the fact that our conformational survey for this complex was particularly thorough (up to 150 and 40 candidate conformations were computed at the B3LYP and the MP2 levels, respectively). In pP1, the two guanidinium cations build identical concerted coordinations with PO bonds from the two phosphate ends. The spacing imposed by the phosphoanhydride P−O−P backbone leads to a separation between the oxygen atoms of the coordinating PO bonds of 3.1 Å. This distance is somewhat larger than the separation between the NH bonds of guanidinium (2.4 Å) but it leads to favorable coordination bonds with PO···H angles of ∼125° and an efficient overlap of the lone pair orbitals of the oxygen atoms with the antibonding orbitals of the guanidinium NH groups. The natural bond orbital representation of Figure 5 shows that the charge redistribution induced by these orbitals account for around 75% of the total stabilization energy of the pP1 conformer. The Na+ cation accommodates at a position equidistant and coplanar with the other two PO bonds, with a distance 2.2 Å from the O atoms and P−O···Na bond angles of 125°. Charge transfer occurs in this case dominantly from the lobular lone pair orbitals of the oxygen atoms to the antibonding σ* orbital of the Na+ cation. The relative contribution of the delocalization of charge toward the Na+ cation accounts for only 10% of the total stabilization energy. This finding stresses the more favorable interactions arising from the diffuse distributions of charge of the molecular ions, guanidinium and pyrophosphate, in comparison to those involving the more localized charge of the alkali cation. It should be noted that the coplanar coordination of Na+ with the PO bonds is in contrast with the V-shaped and more directional coordination of the two guanidinium cations. This feature poses an important constraint to the shared coordination of guanidinium and Na+ with the same pair of PO bonds, due to mutual repulsion between the two cations. Unlike in pP1, in none of the higher energy conformers pP2−pP5 do the two guanidinium cations bind to coincident oxygen atoms. In pP2, the guanidinium cations coordinate to complementary pairs of PO phosphate bonds across the POP backbone. In pP3 and pP4, one of the guanidinium cations moves partially away from the phosphate moieties and binds to an oxygen atom of either the POP backbone or the POH bond. In pP5, the two guanidinium cations move to a peripheral position and interact separately with single PO bonds of each of the phosphate groups. In pP4 and pP5, the Na+ cation displays a coordination arrangement, coplanar with two PO bonds, similar to the one described above for pP1. In contrast, in pP2 and pP3 the Na+ cation interacts with three oxygen atoms across the backbone of the anion, in a coordination arrangement facilitated by a relative rotation of the two phosphate groups. The analysis of the IRMPD experiment for the pyrophosphate complex is consistent with the prediction of pP1 as the most stable conformer. The computational IR spectrum for this conformer is in excellent agreement with the band sequence observed in the measurement and it may concluded that pP1 captures the fundamental structural features of the complex. A detailed inspection of the IRMPD spectrum also reveals an appreciable broadening of some of the vibrational bands, in particular in bands A and C, which points to the participation of additional conformers. The consideration of pP2−pP5 could in fact account for the observed broadenings. Unfortunately, some discrepancies are evident between their
arrangement affects substantially the stretching modes of the phosphate moiety, and the resulting MP2 spectra display sizable discrepancies with the IRMPD measurement. This lack of agreement seems to rule out the concerted interaction of either one of the guanidinium moieties with the full POO− group. Such a result is somewhat unexpected, as the concerted guanidinium−-phosphate coordination interaction occurring in conformers dP2 and dP5 could be anticipated to be among the most stable ones, given the close matching between the distances between the oxygen atoms in the phosphate group (2.6 Å) and between the H atoms of the two NH groups of guanidinium (2.4 Å). In fact, conformation dP2 is roughly isoenergetic with dP1 (within 1 kJ·mol−1) in the MP2 computation. The large stability predicted for the dP2 conformer at the MP2 level relies on a strong charge transfer interactions between the lone pair orbitals of the oxygen atoms of the phosphate group and the antibonding orbitals of the guanidinium NH bonds. This is shown in Figure 4, which depicts the natural bond orbitals relevant to the coordination arrangaments of conformers dP1 and dP2. In dP2, a large part of the stabilization energy arises from the lobular p-like lone pair orbitals of the POO− group, which yield particularly strong interactions with guanidinium. The contribution from the more isotropic sp-like globular lone pair orbitals is somewhat less relevant but also quite significant. In light of the comparison of the MP2 infrared spectra with the IRMPD measurement, it appears that the MP2 computation overestimates these interactions. The present results support instead the individual coordination of each of the guanidinium cations with the oxygen atoms of the PO and POC groups of the phosphate anion observed in conformations dP1, dP3, and dP4. pP2−·(G+)2·Na+ Complex. The extended multipodal structure of the doubly charged pyrophosphate anion pP2− is capable of binding two guanidinium cations in tweezer-like coordination arrangements with distinct features with respect to those just described for the dP− anion. Moreover, since our study probes complexes with net positive charge, the incorporation of a third cation is required for detection. The complex with three guanidinium cations was not observed under the present working conditions; the electrospray ionization of the pyrophosphate/guanidinium mixture did not produce an appreciable signal at m/z = 398. However, it did yield an intense signal for the complex incorporating two methyl guanidinium units and one Na+ cation (m/z = 347). Since Na+ is ubiquitous in biochemical environments, the characterization of the pP2−·(G+)2·Na+ complex should be of interest to assess the potential competition between the alkali cation and guanidinium for binding with the pyrophosphate moiety, possibly weakening the efficiency of the peptidic “guanidinium tweezer”. From a fundamental point of view, it also serves to compare the diffuse interactions of guanidinium with the more localized charge distribution of Na+, within a common supramolecular framework. Figure 5 depicts the five most stable MP2 conformations of the complex. Favorable coordination arrangements arise from the efficient guanidinium−phosphate and sodium−phosphate intermolecular interactions and the optimum separation of the three cationic moieties in the complex. A singular configuration, namely, pP1, appears to be particularly stable; all other conformations are more than 15 kJ·mol−1 higher in energy (hence, a negligible Boltzmann population ratio, well below 0.01 at room temperature). Closer lying conformers were not F
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
coordination angles for the efficient overlap of the lone pair orbitals of the oxygen atoms and the antibonding orbitals of the guanidinium NH groups. This feature should be of general relevance for the coordination of guanidinium and pyrophosphate groups in complex systems. The Na+ cation may bind either in a coplanar configuration with two PO bonds or, alternatively, in interaction with three oxygen atoms across the pyrophosphate backbone assisted by a relative rotation of the phosphate groups. The MP2 computation predicts a distinctly stable conformation, associated with a coplanar coordination of Na+ with two of the POδ− bonds, that forces a concerted coordination of the guanidinium cations with the other pair of POδ− bonds in an overall V-shaped local configuration. The infrared spectrum of this conformer matches accurately the progression of vibrational bands displayed by the IRMPD measurement, which supports that it captures the main structural properties of the most stable coordination arrangements for the complex. The supramolecular conformations described in this study for isolated complexes should guide the microscopic rationalization of guanidinium−phosphate binding in more complex environments. The assessment of solvent effects in anionic chemistry is, however, a challenging and still unsolved issue.19 Interestingly, we find a qualitative similarity between the coordination arrangements found in our study with those derived from recent experiments on arginine−phosphate recognition in an aqueous environment.27 It may be tentatively inferred that, while ion solubility likely affects the equilibrium constants of the complexation process, the preferred guanidinium−phosphate coordination arrangements in solution seem to resemble qualitatively those found in isolated complexes. Work currently in progress in our group seeks to extend the present investigation to progressively more complex benchmark anionic substrates. These will incorporate sugar motifs (ribose or glucose) and will eventually include the AMP/ADP/ ATP family. The interaction of methyl guanidinium with these systems should provide valuable information to rationalize the coordination mechanisms and the conformational contraints underlying their recognition by Arg-containing peptides/ proteins in biological environments.
MP2 spectra and the experiment, for instance, band D in conformers pP2 and pP5, and band E in conformers pP3, pP4, and pP5, do not seem to be properly described. In view of these observations and despite the extensive conformational search perfomed in this study, the existence of additional lowenergy conformations that would not have been found in our survey seems possible. Such conformers would plausibly maintain the general features of the arrangements described here.
■
SUMMARY AND CONCLUSIONS The microscopic background that supports the recognition of phosphate moieties by guanidinium side groups has been investigated. For this purpose, the benchmark complexes formed by methyl guanidinium with diethyl phosphoric acid, dPa0·G+, diethyl phosphate, dP−·(G+)2, and pyrophosphate, pP2−·(G+)2·Na+, have been characterized under isolated conditions with action spectroscopy and quantum chemistry modeling. Favorable coordination arrangements and conformations have been assigned for each of the supramolecular complexes, on the basis of ab initio MP2 computations for the structure and interactions, and the comparison of the predicted vibrational spectrum with the IRMPD bands observed experimentally. The study of the binary dPa0·G+ complex has exposed a preferential coordination of guanidinium with the PO group. A propensity for conformational transitions, leading to a partial binding of guanidinium with the oxygen atoms of the POH or POC groups, is predicted at a moderate energetic cost. Such transitions may confer the complex adaptibility to changing environmental conditions and steric effects. The stabilization of these low energy coordination arrangements has been found to be sustained by the redistribution of charge between sp-like lobular and p-like globular lone pair orbitals of the oxygen atoms and the antibonding σ* orbitals of the NH bonds of guanidinium. The characterization of the ternary dP−·(G+)2 complex has revealed a general coordination landscape dominated by tweezer-like conformations in which the guanidinium cations bind at roughly opposite sites of the phosphate group and provide an efficient “trapping environment” for the anion. An ensemble of roughly isoenergetic stable conformations are predicted by the MP2 computation for this complex. The best agreement with the IRMPD measurement is found for the conformers involving a coordination of guanidinium with the oxygen atoms of the POδ− and POC groups. Conformations in which a single guanidinum cation coordinates with the full POO− anionic group in a concerted arrangement of two of its NHδ+ groups, while rendered stable by the MP2 computation, seem to be ruled out due to discrepancies with the IRMPD spectrum, in particular for the vibrational band associated with the asymmetric stretching of the POO− group. The quaternary pP2−·(G+)2·Na+ complex also displays a variety of stable tweezer-like coordination configurations. The extended bianionic structure of pyrophosphate and the presence of the Na+ cation generate a rich configurational space in which the three cations compete for coordination while trying to minimize mutual repulsions. The guanidinium cations tend to bind to a pair of POδ− bonds of different phosphate groups across the phosphoanhydride P−O−P backbone. The separation between the oxygen atoms of the coordinating PO bonds (3.1 Å) and that between the NH bonds in guanidinium (2.4 Å) lead to favorable PO···H
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.9b06201. Comparison of the experimental IRMPD spectra with the infrared spectra predicted by the B3LYP and MP2 computations for the five low energy conformers of the three complexes investigated in this study (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Giel Berden: 0000-0003-1500-922X Bruno Martínez-Haya: 0000-0003-2682-3286 Notes
The authors declare no competing financial interest. G
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
■
(17) Diez-Cecilia, E.; Kelly, B.; Perez, C.; Zisterer, D. M.; Nevin, D. K.; Lloyd, D. G.; Rozas, I. Guanidinium-based derivatives: searching for new kinase inhibitors. Eur. J. Med. Chem. 2014, 81, 427−441. (18) Suskiewicz, M. J.; Clausen, T. Chemical biology interrogates protein arginine phosphorylation. Cell Chem. Biol. 2016, 23, 888−890. (19) Liu, Y.; Sengupta, A.; Raghavachari, K.; Flood, A. H. Anion binding in solution: beyond the electrostatic regime. Chem. 2017, 3, 411−427. (20) Martens, J.; Berden, G.; Gebhardt, Ch. R.; Oomens, J. Infrared ion spectroscopy in a modified quadrupole ion trap mass spectrometer at the FELIX free electron laser laboratory. Rev. Sci. Instrum. 2016, 87, 103108. (21) Rijs, A.; Oomens, J. Gas phase IR spectroscopy and structure of biological molecules. Top. Curr. Chem. 2014, 364, 1−42. (22) Ma, L.; Ren, J.; Feng, R.; Zhang, K.; Kong, X. Structural characterizations of protonated homodimers of amino acids: Revealed by infrared multiple photon dissociation (IRMPD) spectroscopy and theoretical calculations. Chin. Chem. Lett. 2018, 29, 1333−1339. (23) Berden, G.; Derksen, M.; Houthuijs, K. J.; Martens, J.; Oomens, J. An automatic variable laser attenuator for IRMPD spectroscopy and analysis of power-dependence in fragmentation spectra. Int. J. Mass Spectrom. 2019, 443, 1−8. (24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, Revision E.01; Gaussian, Inc.: Wallingford, CT, 2009. (25) Foster, J. P.; Weinhold, F. Natural hybrid orbitals. J. Am. Chem. Soc. 1980, 102, 7211−7218. (26) van Outersterp, R. E.; Martens, J.; Berden, G.; Steill, J. D.; Oomens, J.; Rijs, A. M. Structural characterization of nucleotide 5′triphosphates by infrared ion spectroscopy and theoretical studies. Phys. Chem. Chem. Phys. 2018, 20, 28319−28330. (27) Neal, J. F.; Zhao, W.; Grooms, A. J.; Flood, A. H.; Allen, H. C. Arginine-phosphate recognition enhanced in phospholipid monolayers at aqueous interfaces. J. Phys. Chem. C 2018, 122, 26362− 26371.
ACKNOWLEDGMENTS B.M.H. is particularly thankful for the mentorship by F. Javier Aoiz during important stages of his research carreer. This study is part of project P12-FQM-4938 of Junta de AndaluciaFEDER (Spain) and has support from the project CALIPSOplus (Grant Agreement 730872, EU Framework Programme for Research and Innovation HORIZON 2020). We thank C3UPO for the HPC facilities and the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) for the support of the FELIX Laboratory.
■
REFERENCES
(1) Chandana, Th.; Venkatesh, Y. P. Occurrence, functions and biological significance of arginine-rich proteins. Curr. Protein Pept. Sci. 2016, 17, 507−516. (2) Houk, R. J. T.; Tobey, S. L.; Anslyn, E. V. Abiotic guanidinium receptors for anion molecular recognition and sensing. Top. Curr. Chem. 2005, 255, 199−229. (3) Schug, K. A.; Lindner, W. Noncovalent binding between guanidinium and anionic groups: focus on biological- and syntheticbased arginine/guanidinium interactions with phosphonate and sulfonate residues. Chem. Rev. 2005, 105, 67−104. (4) Blondeau, P.; Segura, M.; Pérez-Fernández, R.; de Mendoza, J. Molecular recognition of oxoanions based on guanidinium receptors. Chem. Soc. Rev. 2007, 36, 198−210. (5) Avilés-Moreno, J. R.; Berden, G.; Oomens, J.; Martínez-Haya, B. Isolated complexes of the amino acid arginine with polyether and polyamine macrocycles,the role of proton transfer. Phys. Chem. Chem. Phys. 2017, 19, 31345−31351. (6) Avilés-Moreno, J. R.; Berden, G.; Oomens, J.; Martínez-Haya, B. Guanidinium/ammonium competition and proton transfer in the interaction of the amino acid arginine with the tetracarboxylic 18crown-6 ionophore. Phys. Chem. Chem. Phys. 2018, 20, 4067−4073. (7) Avilés-Moreno, J. R.; Berden, G.; Oomens, J.; Martínez-Haya, B. Complexes of crown ether macrocycles with methyl guanidinium: Insights into the capture of charge in peptides. ChemPhysChem 2018, 19, 2169−2175. (8) Wexselblatt, E.; Esko, J. D.; Tor, Y. On guanidinium and cellular uptake. J. Org. Chem. 2014, 79, 6766−6774. (9) Hargrove, A. E.; Nieto, S.; Zhang, T.; Sessler, J. L.; Anslyn, E. V. Artificial receptors for the recognition of phosphorylated molecules. Chem. Rev. 2011, 111, 6603−6782. (10) Berlicki, L.; Rudzinska, E.; Mlynarz, P.; Kafarski, P. Organophosphorus supramolecular chemistry part 1. Receptors for organophosphorus compounds. Curr. Org. Chem. 2006, 10, 2285−2306. (11) He, J.; Hu, P.; Wang, Y. J.; Tong, M. L.; Sun, H.; Mao, Z. W.; Ji, L. N. Double-strand DNA cleavage by copper complexes of 2,2dipyridyl with guanidinium/ammonium pendants. Dalton Trans. 2008, 24, 3207−3214. (12) Kataev, E. A.; Müller, Ch.; Kolesnikov, G. V.; Khrustalev, V. N. Guanidinium-based artificial receptors for binding orthophosphate in aqueous solution. Eur. J. Org. Chem. 2014, 2014, 2747−2753. (13) Trujillo, C.; Previtali, V.; Rozas, I. A theoretical model of the interaction between phosphates in the ATP molecule and guanidinium systems. Theor. Chem. Acc. 2016, 135, 260. (14) Yusufaly, T. I.; Li, Y.; Singh, G.; Olson, W. K. Argininephosphate salt bridges between histones and DNA: Intermolecular actuators that control nucleosome architecture. J. Chem. Phys. 2014, 141, 165102. (15) Herce, H. D.; Garcia, A. E.; Cardoso, M. C. Fundamental molecular mechanism for the cellular uptake of guanidinium-rich molecules. J. Am. Chem. Soc. 2014, 136, 17459−17467. (16) Vargas, J. R.; Stanzl, E. G.; Teng, N. N. H.; Wender, P. A. Cellpenetrating, guanidinium-rich molecular transporters for overcoming efflux-mediated multidrug resistance. Mol. Pharmaceutics 2014, 11, 2553−2565. H
DOI: 10.1021/acs.jpcb.9b06201 J. Phys. Chem. B XXXX, XXX, XXX−XXX
