Aqueous Micelles as Catalytic Nanoreactors for Dephosphorylation Reactions

A reação entre um novo derivado de imidazol, 3,5-bis((1H-imidazol-1-il)metil)anilina (BIm), e o triéster de fosfato dietil 2,4-dinitrofenil fosfato (DEDNPP), que procede por um mecanismo SN2(P), é substancialmente acelerada em presença do surfactante catiônico brometo de cetiltrimetrilamônio (CTAB) e do surfactante aniônico dodecil sulfato de sódio (SDS). Tal efeito micelar atípico ilustra o emprego de efeitos hidrofóbicos dos reagentes para otimizar o efeito catalítico de micelas aquosas. Tal observação é importante e útil no planejamento de modelos miméticos de reações biológicas, já que a melhoria da incorporação dos reagentes aumentará efetivamente a qualidade dos sistemas biomiméticos, sendo que ambos os surfactantes catiônico e aniônico poderão atuar como nanoreatores homogêneos, concentrando os reagentes neutros hidrofóbicos.


Introduction
Imidazole is a versatile group involved in highly efficient enzymatic reactions as a general acid, general base and/or nucleophile.The histidine imidazole plays a fundamental role in many phosphoryl transfer reactions such as the cleavage of RNA, catalyzed by RNAse A. 1 Additionally, cellular signaling processes involve rapid and reversible phosphohistidine formation, via nucleophilic attack by neutral imidazole on phosphorus. 2 Thus, considering the diverse pathways in which imidazole participates, different mechanisms have been proposed and biomimetic models have been developed to better elucidate these reactions. 3,4][7] Recently we reported the reactions of imidazole with activated phosphate di-and tri-esters, such as diethyl 2,4-dinitrophenyl phosphate (DEDNPP), Scheme 1. 8 Imidazole promoted rate enhancements of several thousand fold in the dephosphorylation reactions, compared to spontaneous hydrolysis, and nuclear magnetic resonance (NMR) and electrospray ionization mass spectrometry (ESI-MS) experiments showed that the reactions proceeded via nucleophilic attack on the phosphorus atom, forming phosphorylimidazole species stable enough in some cases to be observed directly in aqueous medium.

Microheterogeneous environments have been developed
6][17] Similarly, metallomicelles show significant rate increases and the results further comprehension of the role of metal ions in enzymes. 18owever, imidazole reactions are not particularly easy to catalyze by the addition of aggregates such as micelles, due to the high water solubility of imidazole, which disfavors its incorporation into the micellar phase.
To study the effect of micellar media on these particular reactions, we synthesized a new imidazole derivative, 3,5-bis((1H-imidazol-1-yl)methyl)aniline (BIm), which readily incorporates to aqueous micelles, and are known to efficiently accelerate nucleophilic reactions of phosphate esters. 19,20We report the reaction of DEDNPP with BIm (Scheme 2) in micellar environments and examine in detail the effect of the ionic surfactants cetyltrimethylammonium bromide (CTAB) and sodium dodecyl sulfate (SDS).

Materials
The phosphate ester DEDNPP was synthesized by standard methods as previously described. 21The surfactants SDS (Sigma) and CTAB (Sigma) were the best available reagent grade and purified as described elsewhere. 22All other inorganic and organic reagents were of the best available analytical grade from commercial sources (Merck, Aldrich, Fluka and Acros Organics) and used as received.Doubly deionized water, conductance < 5.6 × 10 -8 Ω -1 cm -1 and pH 6.0-7.0, from a NANOpure analytical deionization system (type D-4744) was used to prepare the standard and reagent solutions.The synthetic route to obtain the novel imidazole derivative is described in the Supplementary Information along with the characterization data.

Potentiometric titration
The pK a values of BIm in water and CTAB aqueous micellar medium at 25 °C were determined by potentiometric titration.The pH was measured with a Metrohm 713 pH meter and the glass electrode was calibrated against standard buffers, at 25.0 ± 0.1 ºC.Titrations were performed in a 150 mL thermostatted stirred cell under N 2 at 25 °C.The solutions were titrated with small increments of CO 2 -free KOH, and all precautions were taken to eliminate carbonate and CO 2 during the procedure.

Kinetics
Reactions were followed spectrophotometrically by monitoring the appearance of the product 2,4-dinitrophenolate.Reactions were started by adding 20 µL of stock solution of 5.0 × 10 -3 mol L -1 DEDNPP (in acetonitrile) to 3 mL of solution in quartz cuvettes to give the final substrate concentration of 3.33 × 10 -5 mol L -1 .In the reactions with BIm, the imidazole derivative was in excess to guarantee pseudo first-order kinetics with respect to the substrate.The temperature of reaction was maintained with a thermostatted water-jacketed cell holder, and the pH was maintained with 0.01 mol L -1 of the buffers citric acid (pH 3-3.5), acetic acid (pH 4-5.5), sodium dihydrogen phosphate (pH 6-7.5), 2-amino-2-hydroxymethylpropane-1,3-diol (TRIS, pH 8-9) and potassium hydrogen carbonate (pH 9.5) in the reactions in aqueous medium, and 0.01 mol L -1 of biological buffers 2-[bis(2-hydroxyethyl) amino]-2-(hydroxymethyl)-1,3-propanediol (BIS TRIS, pH 6-7) and TRIS (pH 7-9) in the reactions in micellar media.Absorbance vs. time data were stored directly on a microcomputer, and observed first-order rate constants, k obs , were calculated, using absorbance changes obtained for at least 90% of the reaction, using an iterative leastsquares program; correlation coefficients were > 0.999 for all kinetic runs.For the deuterium solvent isotope effect, pD was corrected by the relation pD = pH + 0.4, at 25 °C.

Titration studies
BIm has three amino groups that form four species with potentially different reactivities and potentiometric titration The pK a values given in Scheme 3 were obtained by fitting the titration curve given in Supplementary Information Figure S1 with the program BEST7.The aniline amino group was assigned to the highest acidity (pK a1 = 2.5), and the pK a values of the two imidazole groups (6.07 and 7.26) are close to 7. Based on these pK a values, the species distribution of BIm as function of pH can be conveniently calculated (see Figure S1) and allows us to understand its reactivity in the dephosphorylation reaction of DEDNPP.

Reaction of BIm with DEDNPP in aqueous solution
The results show that BIm promotes the DEDNPP dephosphorylation effectively.Figure 1 shows the pH-rate constant profile for the reaction, compared with that observed for the spontaneous hydrolysis of the triester, and shows that the monoprotonated and neutral species of BIm are reactive, and indicates that the dephosphorylation reaction involves the neutral forms of the imidazole groups.The similarities between the observed rate constants in the region of pH 3 to 4 with the uncatalyzed reaction at pH 7, indicates the absence of a significant contribution of the arylamino moiety, which is both sterically hindered and considerably less basic than the imidazole substituents.
The observed kinetics are consistent with Scheme 4 and allows the derivation of equation 1.

[ ( ) ]
Equation 1 is a description of the reaction of DEDNPP with water (k 0 ), and χ BImH + and χ BIm are the molar fractions of the monoprotonated and neutral species of BIm, respectively, which were calculated using the acid dissociation constants pK a2 and pK a3 described above (Scheme 3).The kinetic parameters obtained by fitting the pH-rate profile with equation 1 are given in Table 1, and show that the rate enhancements promoted by the reactions of monoprotonated and neutral BIm are of the order of 755-and 1700-fold, respectively.The calculated k 2 value is consistent with that obtained by the linear plot of k obs vs.
[BIm] at pH 8.5 (see Table S6).It is interesting to note that the second-order rate constant for reaction with the neutral BIm is about twice as large as that for monoprotonated BIm, as could be expected for a statistical effect due to the presence of two imidazole groups.
The activation parameters calculated at pH 8.5 (data and equations shown in Figure S4 and equations S5-S7), are similar to those reported for the reaction of DEDNPP with imidazole, with values of ∆H ‡ = +10.60 ± 0.46 kcal mol -1 and ∆S ‡ = -131.14± 6.49 e.u..The large and negative entropy of activation is typical of bimolecular associative or concerted mechanisms, and thus is consistent with both nucleophilic and general-base paths.Significantly, the observed solvent

Reactions in micellar media
Since reactions between relatively hydrophobic organic substrates can be favored by incorporation in micelles with different properties, we decided to examine the pH-rate constant profiles for the reaction of DEDNPP with BIm in the presence of cationic and anionic micelles, CTAB and SDS, respectively, and the experimental results are shown in Figure 2.
In SDS micellar medium, using equimolar amounts of BIm and SDS preclude the determination of rate constants in lower pH because the increase in concentration of the protonated forms of BIm leads to precipitation of the anionic surfactant.So we used lower concentrations of BIm and higher surfactant concentration to obtain the profile in SDS micellar medium, compared to that of CTAB.Interestingly, despite the differences in the experimental conditions, the results show that both micelles promote similar increases in k obs , which is also confirmed by the plots of k obs vs.
[BIm] at constant CTAB and SDS concentrations (see Figure S5).In fact, the observed micellar effect can be rationalized in terms of the neutral character of both reactants (considering the most reactive neutral species of BIm), and the similar enhancements in both micellar media are indicative of quantitative incorporation of the reactants (DEDNPP and BIm) in both micellar phases.
The pH-rate profiles of the BIm reactions in CTAB and SDS make clear that the neutral form is the most reactive species in both cases.BIm was titrated in the presence of CTAB at 25 °C, and the apparent pK a HA values of the imidazole groups in CTAB micellar medium (6.05 and 7.10) are closely related to and slightly lower than those determined in aqueous solutions, because the binding constant of BIm, as discussed below, is not very high and, therefore, a significant fraction of the nucleophile is in the aqueous phase (titration curve, composition graph as function of pH and pK a values are given in the Supplementary Information).Furthermore, a decrease in the incorporation of BIm is expected at lower pH values, due to repulsive interactions with the cationic surfactant.The pH profile in SDS is slightly shifted to higher pH, compared to that of CTAB, suggesting that pK a values in SDS medium may be slightly higher, consistent with attractive interactions between cationic BIm and the anionic surfactant.In any event, the effects are very small and further examination is experimentally difficult because the observed changes are close to our limits of experimental error for titration in colloidal systems.
Figure 3 presents the effect of surfactant concentration on k obs in the reactions studied, and the observed profiles are fully consistent with the experimental pH-rate constant profiles.In fact, the reaction of 0.01 mol L -1 of BIm with DEDNPP in the presence of both CTAB and SDS shows very similar rate constant-surfactant concentration profiles at pH 8.5 and 25 ºC.
In general for bimolecular reactions, catalysis by micelles is largely dependent on the incorporation of the reactants.Thus, to fully understand the data, we treated them quantitatively in terms of the pseudophase model for micellar systems.
Quantitative treatment on the rate constant Micellar effects upon bimolecular organic reactions are largely explained in terms of incorporation of the reactants in the micellar phase, followed by differential reactivity in the aqueous and micellar phases, 23,24 as shown in Scheme 5.
The partitioning equilibria of substrate (S) DEDNPP and nucleophile (N) BIm can be described in terms of binding constants K DEDNPP and K BIm , respectively, and the overall reaction rate constant (k obs ) can be described as the sum of the rates in each pseudophase, equation 2.
The scripts M and W refer to the micellar and aqueous phases, respectively, and the terms k 2M and k 2W are second order rate constants, expressed in units of L mol -1 s -1 , in the micellar and aqueous pseudophases, respectively.The term [BIm] represents the total concentration of the nucleophile and is corrected by the molar fraction in each pseudophase (χ W BIm and χ M BIm ) and each term of the equation is multiplied by the molar fraction of the substrate in each phase (χ W DEDNPP and χ M DEDNPP ).The term C d is the concentration of micellized surfactant, and V M , the molar volume of the micelle in which the reaction takes place, which in general is considered either the micelle volume or that of the Stern layer. 25As usual, C d is assumed to be C T -CMC, where C T is the stoichiometric surfactant concentration, and CMC, the critical micelle concentration.The fits of the rate-surfactant profiles for the nucleophilic reactions of BIm in CTAB and SDS were performed considering that the contributions of reactions with water and hydroxide ion are unimportant, except in very dilute surfactant.The kinetic fits to equation 2 are reasonable, and relate second-order rate constants to local molarities in the micellar pseudophase and k 2M can be compared to second-order rate constant in the absence of surfactants (Table 2).
Values of CMC are lower under the reaction conditions than in water (CMC CTAB = 9 × 10 -4 mol L -1 and CMC SDS = 8 × 10 -3 mol L -1 ). 27This behavior is probably related to the effect of BIm in the aggregation properties of the surfactants. 23In fact, precipitation of SDS at lower pH values (using equimolar quantities) is strongly indicative of a system which spontaneously aggregates in aqueous solutions, and this allowed us to neglect CMC values in fitting data, since the stoichiometric surfactant is found almost exclusively in the form of micelles.
The magnitudes of the binding constants (K DEDNPP and K BIm ) are typical of moderately hydrophobic compounds, and the lower values of K BIm are indicative of greater incorporation of DEDNPP into both micelles compared to the nucleophile, which presents a slight preference for the anionic micelle.Note that at pH 8.5, DEDNPP seems to interact similarly with both micelles, while BIm apparently shows a slight preference for the anionic micelle.
Values of k 2M are in the range 4-13-fold lower than k 2W for both micelles, and this relation (k 2M < k 2W ) is commonly found in the literature.In fact, values of k 2W and k 2M depend upon the pseudophase model adopted, the equations derived and the parameters used in fitting data, and a great variety of values is reported. 28,29In our case, k 2M is about the same magnitude of k 2W and the catalytic effect will increase significantly with higher BIm incorporation.The similarity between k 2M and k 2W indicates that the nucleophilic reaction itself has similar reactivity in both micellar media and the observed micellar effect is a result of concentrating the reactants in the micellar pseudophase.The results show that either cationic or anionic micelles behave as homogeneous nanoreactors in the dephosphorylation reaction, irrespective of the charge of the head group.This observation is unusual, since reactions catalyzed by cationic aggregates are typically inhibited by anionic ones and vice versa. 29In fact, the micellar effects in the dephosphorylation of DEDNPP by neutral BIm seem to be governed essentially by the hydrophobic character of the reactants.Thus, as a guide for prospective studies, the results show that rates of dephosphorylation of triesters by nucleophilic imidazole groups may be significantly optimized through the addition of cationic and anionic surfactants, whose aggregates act as homogeneous nanoreactors which concentrate the neutral, hydrophobic reactants.

Conclusions
The hydrolysis of phosphate triesters with good leaving groups in the presence of imidazole can a priori be considered to occur via an associative mechanism. 8his conclusion should apply to the reaction of BIm with DEDNPP, where the results are consistent with an S N 2(P) mechanism, with a zwitterionic delocalized transition state expected in the rate-limiting BIm attack.Interestingly, cationic and anionic micelles exert similar effects, which indicate that in the transition state a negative charge on the oxygen atoms and a positive charge on the nucleophilic imidazole group are developing.Thus, the rate enhancement may occur for this kind of reaction with both cationic and anionic head groups, in which both have a reasonable chance to interact with the transition state, and this interpretation is in agreement with our results.
In addition, notice that differences in reactant hydrophobicity may also promote micellar effects.Compare the reaction of the substrate with imidazole in CTAB (see Figure S6) to that obtained with BIm.In fact, the imidazole reaction is not catalyzed in CTAB medium, mainly because the nucleophile mostly remains in the aqueous phase.Thus, more hydrophobic BIm binds to the micelles more efficiently and supports the idea that hydrophobic effects also contribute to micellar effects.This is an important and useful guideline in planning mimics for biological reactions, because improving incorporation will effectively increase the quality of the biomimetic systems.

Table 1 .
Kinetic parameters for the reaction of BIm with DEDNPP, at 25 °Ca SKIE), calculated at pH(D) 8.5 in the plateau region shows a SKIE (k H2O /k D2O ) = 0.996, a result which is strongly indicative of rate-limiting nucleophilic attack at the phosphorus center by the neutral BIm species.Thus, in excess of nucleophile, we propose the dephosphorylation of DEDNPP involves one of the BIm imidazole groups available.

Table 2 .
Rate and binding constants obtained for the reactions of BIm with DEDNPP in CTAB and SDS aqueous micelles, at 25 °Ca,b in treating data, it was considered the molar fraction of neutral BIm; values of pK a for the acid dissociation of the imidazole groups in CTAB micellar medium (6.05 and 7.10) were equal to those obtained by potentiometric titration at 25 °C (see Supplementary Information), and values used in the fit for SDS micellar medium were the same as those of aqueous medium.Values used for V M were 0.37 and 0.25 L mol -1 for CTAB and SDS, respectively, obtained from the literature; 26 c obtained by fitting the pH-rate profile of reaction of BIm with DEDNPP in aqueous medium (Figure1, Table1). b