Catalytic Zinc Complexes for Phosphate Diester Hydrolysis

Creating efficient artificial catalysts that can compete with biocatalysis has been an enduring challenge which has yet to be met. Reported herein is the synthesis and characterization of a series of zinc complexes designed to catalyze the hydrolysis of phosphate diesters. By introducing a hydrated aldehyde into the ligand we achieve turnover for DNA-like substrates which, combined with ligand methylation, increases reactivity by two orders of magnitude. In contrast to current orthodoxy and mechanistic explanations, we propose a mechanism where the nucleophile is not coordinated to the metal ion, but involves a tautomer with a more effective Lewis acid and more reactive nucleophile. This data suggests a new strategy for creating more efficient metal ion based catalysts, and highlights a possible mode of action for metalloenzymes.


3-(bis((6-methylpyridin-2-yl)methyl)amino)-2,2-dimethylpropane (L5)
To a solution of 2,2-dimethylpropan-1-amine (0.60 mL, 0.44 g, 4.9 mmol) in dry DMF (50 mL) was added 6methyl-2-(bromomethyl)-pyridine 1 (2.00 g, 10.8 mmol) and K 2 CO 3 (1.49 g, 10.8 mmol). The solution was stirred at 40 °C for 16 h. The organic component was extracted into CH 2 Cl 2 (3 × 40 mL), the extract dried over MgSO 4 , filtered and the solvent was removed under reduced pressure. The product was purified using silica gel chromatography (Hexane : EtOAc -10:1) to yield 1.28 g of L5 as a pale yellow solid (88%).   Titration data and speciation curves of L2 and Zn complex 2. Figure S3. left: titration of 10 mL of 2 mM L2 and 4 mM HCl with 10 µL aliquots of 0.2 mM NaOH; right: speciation diagram of 1mM L2. The theoretical fit and speciation were carried out with Hyperquad and HySS software 3 and gave pK a s of 3.85 and 6.99. Figure S4. left: titration of 10 mL of 2 mM Zn complex 2 and 4 mM HCl with 10 µL aliquots of 0.2 mM NaOH; right: speciation diagram of 1 mM 2. The theoretical fit and speciation were carried out with Hyperquad and HySS software 3 using formation constants for Zn(OH) (10 -7.84 ) and Zn(OH) 2 (10 -16.86 ) taken from literature. 4 The fit gives pK a s for the Zn complex 2 of 7.84 and 9.02, and log K d as 4.85.  Observed rate constants (k obs ) were obtained by the initial rate method over the first 30 minutes of the reaction. Equation S1 is derived from Scheme S1, and fitted to the data in Tables S1 to S8 to give the values of k 2 used in Figure S5.
Scheme S1 [3] mM k obs s -1 at pH 6       Observed rate constants (k obs ) were obtained by the initial rate method over the first 30 minutes of the reaction. Equation S1 derived from Scheme S1 was fitted to the data in Tables S10 to S17 to give the values of k 2 used in Figure S12.
[4] mM k obs s -1 at pH 6 The values of the conditional dissociation constant ! !"" differ by a factor of ~2.6 fold, which is in good agreement with variation expected at these pHs based on the pK a of L4 (6.91) and the first pK a of the Zn complex of 4 (8.09) established by potentiometric titration. The limiting second rate constants agree well with the data obtained at constant Zn:ligand ratio; the K d values are about 4 fold greater than obtained from potentiometric titration (0.12 mM).

Cleavage of BNPP in methanol by Zn complexes 2, 3 and 4'.
The reactions in dry methanol led to the loss of both leaving groups from BNPP, with the second reaction much slower than the first. Thus, a sequential reaction scheme (scheme 2) was used to fit the appearance of both equivalents of 4-nitrophenol (equation S3). This yielded the observed rate constant for the initial reaction between the complex and BNPP, given in tables S3, S4 and S5. Plots of k obs against [complex] show an upward curvature that is very similar to the behaviour of the complexes in water at lower pH. However, in methanol, it can be explained by cooperativity between two complexes as described by Mohamed et al. 5 This reaction scheme is given in scheme 3, and leads to equation S4. The k 2 values have been obtained by fitting equation S4 to the data in tables S3, S4 and S5 derived by Mohamed et al. 5 Although the active species may be involve a second metal ion complex, the variation in the nucleophile structure is the only change between each complex.

Equation S3
Scheme S3.     Initial geometry optimizations were performed at the Hartree Fock (HF) level of theory using the 6-31+G* basis set. Hay and Wadt's effective core potential with the double-ζ valence basis set (LANL2DZ 6,7 ) was used to describe zinc. The resulting structures were characterised by frequency calculations. Single point energy calculations on the optimized structures were performed at the DFT level of theory using the dispersion corrected M06-2X functional 8 with a larger (6-311+G**) basis set, and the metal center described as before. For both optimization and single point calculations, solvation was implicitly accounted for by the SMD continuum solvation model. 9 All quantum chemical calculations were performed using the Gaussian 09 simulation package. 10 The computational methodology used here is at a similar level to that used by Ohanessian et al. for their study of biomimetic Zn complexes. 11 In this study, it was demonstrated that reliable results in terms of geometry and chemical accuracy could be reached by simply performing a HF geometry optimization, followed by a B3LYP energy calculation with a larger basis set. In this work, we used the M06-2X functional (instead of the popular B3LYP) as it also includes dispersion correction.
Using the crystal structure of 4' as a starting point, three different Zn-complex structures were generated (Fig.  S23). Complex 1-1 is closely related to the crystal structure, with a water molecule replacing the nitrate group coordinated to the Zn ion in the crystal structure, and the methoxy group converted to a hydroxyl. From this optimized structure, two further conformations were generated. By rotating the nucleophile side chain dihedral angles N 1 -C 1 -C 2 -C 3 and C 1 -C 2 -C 3 -O 2 , complex 1-2 was generated. Complex 1-3 was obtained by changing the propeller like arrangement 12 of the ligand around the zinc. It was found that all structures are very similar in energy (Table S7). Conformations 1-1 and 1-3 show that the non-coordinated oxygen can readily be positioned close to the free coordination site on the Zn ion, and are almost mirror images as the coordinated water leads to a more symmetrical coordination sphere than when nitrate is coordinated. Figure S23: Optimised structures of water-bound zinc-complexes using HF method and 6-31+G* / LANL2DZ basis set.  Table S7: Selected distances (in Å) of the optimised structure and relative energy (in kcal mol -1 ) of the three conformers 1-1, 1-2 and 1-3. The energies were obtained by single point calculation on the optimised structure at M062X and 6-311+G**/ LANL2DZ level of theory using SMD solvent model.

S19
Similar calculations were performed on the monodeprotonated forms (at the non-coordinated hydroxyl) of these complexes. Once again, all the structures show similar energies. In these structures, the deprotonated non-bound oxygen shows a very similar distance to the substrate binding site compared to the protonated forms.   Table S8: Selected distances (in Å) of the optimised structure and relative energy (in kcal mol -1 ) of the three monodeprotonated conformers 2-1, 2-2 and 2-3. The energies were obtained by single point calculation on the optimised structure using M062X and 6-311+G**/ LANL2DZ basis set using SMD solvent model.

S20
Finally, to explore the possibility of the non-coordinated oxygen acting as a nucleophile towards a coordinated phosphate diester, we replaced the water in the monodeprotonated complexes with methyl-p-nitrophenylphosphate. As before, we explored the same three different conformations of the complex system. The presence of the phosphate diester reintroduces the asymmetry between 3-1 and 3-3, where it is evident that the noncoordinated oxygen can be either remote or close to the substrate.
3-1 reflects the original crystal structure, and shows a large distance between the non-bound oxygen atom and the phosphate centre (5.60Å). However, once the rotation of the side chain and the propellor inversion have taken place, the non-bound oxygen is much closer to the phosphate centre, reaching a value of 4.06Å with an angle to the leaving group PO bond of 164°. Figure S25 Optimised structure of monodeprotonated zinc complexes with phosphate bound using HF method and 6-31+G* / LANL2DZ basis set.

S21
Finally we performed a transition state optimization for the nucleophile attack reaction of the non-coordinated oxygen for conformation 3-2 ( Figure S26). The resulting structure was characterized by frequency calculations, and the minimum energy path connecting reactants to products through this transition state was evaluated by calculating the intrinsic reaction coordinate 13,14 (IRC=ξ) in both the forward and reverse directions in order to identify the corresponding reactant. Our results not only confirm this is a viable pathway for the phosphoryl transfer reaction, but comparison of the reactant stated optimized from the IRC calculation (RS) and the optimized complexes (3-1, 3-2 and 3-3) show almost no difference in energy.