Article pubs.acs.org/joc
Guanidinium Promoted Cleavage of Phosphoric Diesters: Kinetic Investigations and Calculations Provide Indications on the Operating Mechanism Riccardo Salvio*,† and Alessandro Casnati‡ †
Dipartimento di Chimica and IMC - CNR Sezione Meccanismi di Reazione, Università La Sapienza, 00185 Roma, Italy Dipartimento di Scienze Chimiche, della Vita e della Sostenibilità Ambientale, Università degli Studi di Parma, Viale delle Scienze 17/A, 43124 Parma, Italy
‡
S Supporting Information *
ABSTRACT: The catalytic activity of the guanidinium units toward the cleavage of phosphoric diesters is deeply investigated both with kinetic experiments and DFT calculations. The first part of the investigation aims to determine how the structure of the substrate (phenyl or alkyl esters) is able to influence the guanidinium-catalyzed hydrolysis changing the mechanism from ANDN to AN+DN. In the cleavage of the DNA model bis(4nitrophenyl)phospate (BNPP), experimental kinetic data highlight the operation of a guanidine−guanidinium catalytic dyad that can act both intermolecularly and intramolecularly on different molecular scaffolds exhibiting notable values of effective molarity. 31P NMR spectra and DFT investigation provide indication that the deprotonated guanidine involved in such a catalysis acts as a general base in the deprotonation of a water molecule involved in the cleavage, and not as nucleophilic unit. Moreover, DFT calculations were carried out to determine the guanidinium promoted activation energy of pseudorotation. The results indicate a remarkable drop in the activation energy of this process for dialkylphosphate esters explaining, in part, the higher sensitivity of diribonucleoside to the presence of guanidinium-based catalysts compared to the more activated RNA model HPNP.
■
INTRODUCTION
In recent papers, we reported about the synthesis and the catalytic activity of a series of artificial phosphodiesterases provided with guanidinium units.3b−d,4h These supramolecular systems have been deeply and exhaustively investigated in terms of catalytic efficiency, selectivity and operating mechanism. Such investigations were carried out mainly by kinetic data analysis, potentiometric titrations, and UV measurements. In the literature a number of contributions deals with quantum mechanical calculation about the cleavage of phosphoric diesters,8−10 mainly in the presence of catalysts provided with metal center(s), acting by themselves9 or in synergic action with other units.10 Here we present an experimental and computational study carried out with kinetic analysis and DFT calculations, respectively, on guanidinium promoted cleavage of phosphoric diesters, with the aim of complementing the experimental data collected so far in our previous work and of better understanding some mechanistic details. In particular, the first part of the work has been focused on the determination of the geometry of phosphateguanidinium complexes and the energies involved in such
The biological relevance of phosphodiester bonds, due to their central role in living systems, has pushed many researchers to design and synthesize small molecule enzyme mimics able to cleave efficiently and selectively DNA, RNA, and their model compounds.1−4 Most of these artificial catalysts contain metal cations2 that can play the role of binding sites or activators. A relevant number of artificial phosphodiesterases are provided with one or more guanidinium units that act as catalytic functions by themselves3 or in conjunction with other catalytically active components.4 The use of this function in the design of supramolecular systems is not surprising because in nature the role of guanidinium is crucial in the coordination or interaction with different oxoanions, such as carboxylates and phosphates.1b,5 In many enzymes arginine residues are involved in the activation of the substrate and in the subsequent stabilization of the transition states and intermediates by hydrogen bonds and electrostatic attractions.5,6 Another reason for the widespread use of this moiety, both in natural and synthetic systems, is the high pKa value (12−14) that ensures protonation over a wide pH range.5a,7 © 2017 American Chemical Society
Received: July 31, 2017 Published: September 1, 2017 10461
DOI: 10.1021/acs.joc.7b01925 J. Org. Chem. 2017, 82, 10461−10469
Article
The Journal of Organic Chemistry
Table 1. Calculated Relative Binding Energiesa (ΔG°, kcal mol−1) between Different Phosphates and Guanidinium Derivatives Together with the Corresponding Hydrogen Bond Lengths (Å)b
guanidinium (R3H) entry
phosphoric diester
energy
HB length
1 2 3
DMP (R1R2=CH3) DPP (R1R2Ph) BNPP (R1R2=p-nitroPh)
−3.60 −1.71 −0.22
1.731 1.764 1.812
methylguanidinium (R3Me) c
c
phenylguanidinium (R3Ph)
energy
HB length
energy
HB lengthc
−4.21 −2.70 0
1.740 1.771 1.818
−3.42 −2.27 −0.34
1.724 1.759 1.805
a The energy values are relative to the weaker interaction, i.e., bis(p-nitrophenyl)phosphate−methylguanidinium, ΔG = 0.043 kcal mol−1. bSee Experimental Section and SI for details about the DFT calculations. cThe indicated distance is the average value between the H1−O1 and H2−O2 distances.
slightly moved away from their equilibrium geometry and their orientation slightly changed and, in all cases, the two molecules got close and converged to a minimum that shows a two-point interaction (see Figure 1). The geometry of the guanidinium-
interaction. Other calculations have been carried out to determine whether and with which substrates the guanidinium is able to change the hydrolysis mechanism from ANDN to AN+DN with the formation of a phosphorane intermediate. Kinetic data for the cleavage of the DNA model bis(4nitrophenyl)phospate (BNPP) in the presence of guanidiniumbased catalysts were performed with an indication of a guanidine−guanidinium cooperation as catalytic mechanism. 31 P NMR and in silico study were used to detect and collect clues of the presence of a phosphoroamidate species that might be formed as a reaction intermediate in these conditions. Moreover, DFT calculations were carried out to study the guanidinium-promoted pseudorotation of the phosphorane intermediate.
■
RESULTS AND DISCUSSION Guanidinium−Phosphate Binding. A prerequisite for understanding the role of guanidinium in phosphoryl transfer reactions is the investigation of the interaction between the phosphate moiety and the guanidinium unit. The binding ability of the latter can be ascribed to its planar and rigid structure and to its geometrical complementarity to this oxoanion that allows the formation of a two-point hydrogen bonding chelate motif11 although other hydrogen-bonding patterns are possible.12 A number of calculations were carried out on the guanidinium derivatives, phosphate diesters, and their corresponding adducts as listed in Table 1. The phosphoric diesters taken into consideration are dimethylphosphate (DMP), diphenylphosphate (DPP), and bis(4-nitrophenyl)phosphate (BNPP). The choice of the investigated compounds was made on the basis of their occurrence in research papers dealing with the cleavage of phosphodiesters.1−4 The in silico experiments were carried out with the density functional theory method using Gaussian 09 package13 at B3LYP/6-311+g(d,p)//B3LYP/6-311+g(d,p) level of therory (see Experimental Section and Supporting Information for further details and for the coordinates of all the optimized structures). The polarized continuum model was used to take into account the solvent effect. The value of the dielectric constant was set to 72, that is the experimental value measured in bulk for a DMSO:H2O 80:20 mixture,14 hereafter referred to as 80% DMSO. This solvent mixture is well-known to be suitable for the investigation of phosphoryl transfer reactions 1b,4c,14,15 also in systems based on the sole guanidinium unit.3b−d,16 In the optimization procedure, the structures of the phosphoric diesters and the guanidinium derivatives were
Figure 1. DFT calculated structure of the dimethylphosphate− guanidinium complex showing a two-point hydrogen bonding chelate interaction.
phosphate group and the H1−O1 and H2−O2 distances (average value = 1.77 Å) indicate the presence of a chelate hydrogen bonding. This value is significantly lower than the typical distance between the hydrogen and the acceptor atom in a single hydrogen bond (2.0−2.2 Å)17 and this is probably due to the chelate effect operating in this binding. In Table 1 are reported the energies of the interactions between the indicated compounds calculated by the difference of the energies of the adduct and the isolated components. The energy values reported in the table were added with zero-point vibrational energies and corrections to enthalpy and entropy calculated with the thermochemical analysis tool of Gaussian 09 (see Experimental Section for details) and therefore can be considered as Gibbs free energy differences. Although these energies can not be used to determine absolute binding constants because of the relevant energetic effect due to the solvation, they allow a more direct comparison of values obtained with the different binding partners. For all the guanidinium derivatives the binding energies become less negative (weaker association) upon replacing R1 and R2 with electron withdrawing groups. Namely the dimethylphosphate 10462
DOI: 10.1021/acs.joc.7b01925 J. Org. Chem. 2017, 82, 10461−10469
Article
The Journal of Organic Chemistry
therefore reflect the depth, as well as the shape, of the potential energy surface. Also in the present case, there is no clear and evident trend as pointed out in Table 2. Therefore, a preferential binding of guanidinium or guanidinium-d6 is not predictable a priori. The thermodynamic isotope effect18c in guanidinium-phosphate binding can affect the reliability of proton inventory studies dealing with hydrolysis mechanism of phosphodiesters as pointed out by Yatsimirsky et al.14 On account of this effect the computational approach can be a valid alternative to figure out whether the guanidinium performs its activity delivering a proton, i.e., acting as a general acid (vide inf ra). Guanidinium Promoted Basic Hydrolysis of Phosphodiesters. A series of calculations on the basic hydrolysis of phosphodiesters promoted by guanidinium units were carried out in order to determine the amount of energy involved in the process and to gain information about the mechanism and the possible formation of a phosphorane intermediate induced by the interaction with guanidinium. Phosphoryl transfer reaction can proceed through the mechanisms illustrated in Scheme 1.1c,19 For phosphodiesters the generally accepted mechanisms
interacts more strongly with all the three investigated guanidinium derivatives (Table 1, entry 1) compared with the other two phosphodiesters (R1,R2Ph, p-nitroPh, entries 2 and 3) and the diphenyl phosphate more strongly (entry 2, Table 1) than the corresponding nitroderivative BNPP (entry 3). A consistent trend can be noted in O1−H1 and O2−H2 distances, that become longer for phosphodiesters with electron withdrawing substituents. This effect can be ascribed to a partial delocalization of the negative charge located on the phosphate unit operated by these substituents that results in a weaker binding to the guanidinium moiety. On the other hand, the effect of the substituents of the guanidinium unit (R3) on the binding energy is modest and no clear trend can be recognized in spite of the relevant experimental pKa value differences observed between phenylguanidinium (pKa = 11.5)3c and the other two derivatives (pKa guanidinium: 13.7, pKa methylguanidinium = 13.6).4e,14 This result may suggest that the guanidinium-phosphate interaction has mainly an electrostatic character. It is remarkable to note that the phosphoric acid diesters considered as binding partners have similar acidities and are essentially strong acids in this medium (pKa < 2).3d A second set of calculations was carried out to evaluate the interaction between guanidinium-d6 and the same phosphodiesters. Since hydrogen and deuterium are electronically identical, the electronic part of the total energies is equal. The differences are associated with the masses, and therefore to the vibrational energies. In Table 2 are reported the binding
Scheme 1. Limit Reaction Pathways for the Basic Hydrolysis of Phosphodiesters
Table 2. Binding Energies (kcal mol−1) between Guanidinium-d6 and the Listed Phosphodiesters and Comparison with the Energies with Protonated Guanidinium entry
binding partner
relative ΔG°a
ΔΔG°D‑Hb
1 2 3
dimethylphosphate (R1R2=CH3) diphenylphosphate (R1R2Ph) BNPP (R1R2=p-nitroPh)
−3.22 −1.93 0.00
0.15 −0.46 −0.015
a
Association energy between the deuterated guanidinium and the given phosphodiesters defined as in Table 1. The energy values are relative to the weaker interaction, i.e., BNPP/guanidinium-d6 (entry 3): ΔG°BNPP= −0.195 kcal mol−1. bDifference between the calculated binding energies of the given phosphodiester with guanidinium-d6 and guanidinium (ΔG°D−ΔG°H), see data in Table 1.
are (i) the addition−elimination pathway, designated as AN+DN, in which the attack of the nucleophile on phosphorus leads to a pentacoordinate phosphorane intermediate that affords the products in a second step. An alternative to this pathway there is (ii) the concerted mechanism, designated as ANDN, in which the phosphorane is a transition state and not an intermediate. The latter pathway has also been labeled by Kirby as SN2(P),20 for its close resemblance to the aliphatic nucleophilic substitutions. Scans of potential energy surfaces (SPES) calculations have been carried out for the same phosphodiesters considered in Tables 1 and 2. The attack of the hydroxide ion has been considered, fixing and keeping constant, in each calculation, the distance between the oxygen atom of HO− and the phosphorus atom. All the other geometric parameters are optimized in local minimum search. The same calculations were carried out in presence and in absence of the guanidinium unit interacting with the phosphate moiety using the coordination geometries optimized in the first part of the present work. In Figure 2 are reported the plots of the calculated energy versus the reaction coordinate that was defined as the difference between the distance between the phosphorus and the oxygen atom of the leaving group and the distance between the oxygen atom of the hydroxide anion and the phosphorus atom. For all the three cases (DMP, DPP, and
energies of deuterated guanidinium with DMP, DPP, and BNPP and the differences in binding energies between guanidinium and guanidinium-d6. The interaction of dimethylphosphate with the deuterated guanidinium is slightly more unfavorable than that with protonated guanidinium (entry 1, Table 2), whereas is a little more favorable in the case of BNPP (entry 3). For DPP the difference is more noticeable because it approaches half a kcal mol−1 (entry 2, Table 2). This difference can be traced to two particular vibrational modes: the symmetric and asymmetric stretching of the N−H(D) interacting with the oxygen atoms of the phosphate unit, and the ones which displaces the bridging H or D atoms away from the N−O axis (see Figure S1 in SI). In literature researchers try to address the issue whether a hydrogen bond is weaker or stronger than the equivalent deuterium bond.18 It is concluded that there is no simple and direct answer to this question18a since the equilibrium properties of systems containing hydrogen or deuterium bonds are dependent on vibrational motion of the nuclei and 10463
DOI: 10.1021/acs.joc.7b01925 J. Org. Chem. 2017, 82, 10461−10469
Article
The Journal of Organic Chemistry
should be considered significant. On the other hand, in the uncatalyzed process there is no appreciable minimum of energy. It should be actually noted that the structure of the phosphorane formed in the uncatalyzed basic hydrolysis of DMP does not show negative spring constant and therefore it is a minimum of energy by the mathematical point of view. However, this minimum of energy is so little deep (
