The Role of Hydrogen Bond in Catalytic Triad of Serine Proteases

In order to investigate the origin of catalytic power for serine proteases, the role of the hydrogen bond in the catalytic triad was studied in the proteolysis process of the peptides chymotrypsin inhibitor 2 (CI2), MCTI-A, and a hexapeptide (SUB), respectively. We first calculated the free energy profile of the proton transfer between His and Asp residues of the catalytic triad in the enzyme-substrate state and transition state by employing QM/MM molecular dynamics simulations. The results show that a low-barrier hydrogen bond (LBHB) only forms in the transition state of the acylation of CI2, while it is a normal hydrogen bond in the acylation of MCTI-A or SUB. In addition, the change of the hydrogen bond strength is much larger in CI2 and SUB systems than in MCTI-A system, which decreases the acylation energy barrier significantly for CI2 and SUB. Clearly, a LBHB formed in the transition state region helps accelerate the acylation reaction. But to our surprise, a normal hydrogen bond can also help to decrease the energy barrier. The key to reducing the reaction barrier is the increment of hydrogen bond strength in the transition state state, whether it is a LBHB or not. Our studies cast new light on the role of the hydrogen bond in the catalytic triad, and help to understand the catalytic triad of serine proteases.     

Strong N−H⋅⋅⋅O Hydrogen Bonding in a Model Compound of the Catalytic Triad in Serine Proteases

Low-barrier hydrogen bond (LBHB) involvement in enzyme catalysis is examined by analysis of experimental nuclear and electron densities of a model compound for the catalytic triad in serine proteases (shown schematically), which is based on a cocrystal of betaine, imidazole, and picric acid. The three short, strong N−H⋅⋅⋅O hydrogen bonds in the structure have varying degrees of covalent bonding contributions suggesting a gradual transition to the LBHB situation.     

A neutron scattering study of strong-symmetric hydrogen bonds in potassium and cesium hydrogen bistrifluoroacetates: Determination of the crystal structures and of the single-well potentials for protons

The crystal structures of potassium and cesium bistrifluoroacetates, KH(CF(3)COO)(2) and CsH(CF(3)COO)(2), respectively, were determined at room and cryogenic temperatures with the single crystal neutron diffraction technique. The crystals belong to the monoclinic space groups, I2a and A2a, respectively, and there is no evidence of any structural phase transition. In both crystals, trifluoroacetate entities in centrosymmetric dimers are linked by very short hydrogen bonds lying across a center of inversion. The thermal parameters provide no evidence of any double minimum potential for hydrogen bond protons. Single-minimum potentials were determined via best fitting to the inelastic neutron scattering spectral profiles of the stretching vibrations. They comprise a narrow well for the ground state and a very broad quasiharmonic well for excited states. The spread out of the wave functions of these states shows that protons are no longer confined between the oxygens. Presumably, they are attracted by the lone pairs of oxygen atoms. These potentials emphasize the covalent nature of the OO bond and the ionic character of the hydrogen bond proton.     

Quantum chemical calculations on structural models of the catalytic site of chymotrypsin: comparison of calculated results with experimental data from NMR spectroscopy

Hybrid density functional quantum mechanical calculations were used to study the strength of the hydrogen bond between His(57) N(delta)(1) and Asp(102) O(delta)(1) in chymotrypsin and how it changes along the reaction coordinate. Comparison of experimental shifts with the results of chemical shift calculations on a variety of small molecules, including species containing very strong hydrogen bonds, has validated the overall approach and provided the means for calibrating and correcting the calculated values. Models of the active site of chymotrypsin in its resting state and tetrahedral intermediate state were derived from high-resolution X-ray structures. The distance between His(57) N(delta)(1) and Asp(102) O(delta)(1) in each model was varied between 2.77 A (weak hydrogen bond) and 2.50 A (extremely strong hydrogen bond), and the one-dimensional potential energy surface of the hydrogen-bonded proton (or deuteron/triton) was determined. The zero-point energy, probability distribution, and chemical shift were determined for each distance. Calculated values for NMR chemical shifts, NMR chemical shift differences between (1)H and (3)H, and (2)H/(1)H fractionation factors were compared with published experimental values. Energies provided by the calculations indicated that the hydrogen bond between His(57) N(delta)(1) and Asp(102) O(delta)(1) in the chymotrypsin active site increases in strength by 11 kcal mol(-)(1) in going from the resting state of the enzyme to the tetrahedral intermediate state. This result confirms the hypothesis that the strengthened hydrogen bond plays an important role in lowering the energy of the transition state and, hence, in the catalytic efficiency of the enzyme. Models of the transition state that best fit the experimental data are consistent with a "strong" hydrogen bond between His(57) N(delta)(1) and Asp(102) O(delta)(1) but apparently not a "low-barrier" or "very strong" hydrogen bond.     

A critical comparison of approximation methods and models for equilibrium properties of low-barrier hydrogen bonds

Recent experimental evidence has led to the conclusion that short, strong hydrogen bonds can stabilize transition states of enzyme catalyzed biochemical reactions. Evidence for such hydrogen bonds is the low value of the isotopic fractionation factor, phi, which is defined as the equilibrium constant for the generic reaction, R-H + DOH <--> R-D + HOH, where H is the hydrogen atom participating in the low-barrier hydrogen bond in a molecule R-H. In this work we assess two approximation methods for computing the isotopic fractionation factors for single and multidimensional systems containing a low-barrier hydrogen bond. These methods are WKB and an approach that corrects the classical partition function via a quantum correction factor. We find that the latter approach is universally accurate and applicable in both single and multidimensional systems containing a low-barrier hydrogen bond. We also assess two different models for the coupling of a molecule's low-barrier hydrogen bond to other degrees of freedom, both internal and external to the molecule, and show that each leads to a lowering of the fractionation factor.     

Contribution of multicentered short hydrogen bond arrays to potency of active site-directed serine protease inhibitors

We describe and compare the pH dependencies of the potencies and of the bound structures of two inhibitor isosteres that form multicentered short hydrogen bond arrays at the active sites of trypsin, thrombin, and urokinase type plasminogen activator (urokinase or uPA) over certain ranges of pH. Depending on the pH, short hydrogen bond arrays at the active site are mediated by two waters, one in the oxyanion hole (H(2)O(oxy)) and one on the other (S2) side of the inhibitor (H(2)O(S2)), by one water (H(2)O(oxy)), or by no water. The dramatic variation in the length of the active site hydrogen bonds as a function of pH, of inhibitor, and of enzyme, along with the involvement or absence of ordered water, produces a large structural manifold of active site hydrogen bond motifs. Diverse examples of multicentered and two-centered short hydrogen bond arrays, both at and away from the active site, recently discovered in several protein crystal systems, suggest that short hydrogen bonds in proteins may be more common than has been recognized. The short hydrogen bond arrays resemble one another with respect to ionic nature, highly polar environment, multitude of associated ordinary hydrogen bonds, and disparate pK(a) values of participating groups. Comparison of structures and K(i) values of trypsin complexes at pH values where the multicentered short hydrogen bond arrays mediating inhibitor binding are present or absent indicate that these arrays have a minor effect on inhibitor potency. These features suggest little covalent nature within the short hydrogen bonds, despite their extraordinary shortness (as short as 2.0 A).     

Correlated proton motion in hydrogen bonded systems: tuning proton affinities

The theorem of matching proton affinities (PA) has been widely used in the analysis of hydrogen bonds. However, most experimental and theoretical investigations have to cope with the problem that the variation of the PA of one partner in the hydrogen bond severely affects the properties of the interface between both molecules. The B3LYP/d95+(d,p) analysis of two hydrogen bonds coupled by a 5-methyl-1H-imidazole molecule showed that it is possible to change the PA of one partner of the hydrogen bond while maintaining the properties of the interface. This technique allowed us to correlate various properties of the hydrogen bond directly with the difference in the PAs between both partners: it is possible to tune the potential energy surface of the bonding hydrogen atom from that of an ordinary hydrogen bond (localized hydrogen atom) to that of a low barrier hydrogen bond (LBHB, delocalized hydrogen atom) just by varying the proton affinity of one partner. This correlation shows clearly that matching PAs are of lesser importance for the formation of a LBHB than the relative energy difference between the two tautomers of the hydrogen bond.     

Evidence for a strong hydrogen bond in the catalytic dyad of transition-state analogue inhibitor complexes of chymotrypsin from proton-triton NMR isotope shifts

We present here the first accurate measurements of 1H (H) versus 3H (T) isotope shift (DeltadeltaT-H = deltaT - deltaH) in a protein. This approach was used to investigate the strength of the hydrogen bond between His57 and Asp102 in the catalytic dyad of chymotrypsin in three transition-state analogue inhibited complexes: N-acetyl-l-phenylalanyl trifluoromethyl ketone (N-AcF-CF3), N-acetyl-l-valyl-l-phenylalanyl trifluoromethyl ketone (N-AcVF-CF3), and N-acetyl-l-leucyl-l--phenylalanyl trifluoromethyl ketone (N-AcLF-CF3). The measured DeltadeltaT-H values for His57 Hdelta1 in these complexes were between -0.63 and -0.68 ppm. These values are consistent with a strong hydrogen bond in each of these complexes, but not with a very strong hydrogen bond, which would be expected to have a DeltadeltaT-H. value near or greater than zero.     

On the Possibility of Using UV Spectroscopy as a Measure of the Low-Barrier Hydrogen Bond

In many enzyme-catalyzed biochemical pathways, a short, strong hydrogen bond between an enzyme and substrate is an important structural feature. These bonds are termed low-barrier hydrogen bonds. In this paper, we show that UV spectra can be used as an experimental technique to determine if a system contains a low-barrier hydrogen bond (LBHB). We simulate, using the time-dependent view of UV spectroscopy, several different UV spectra: absorption, photodissociation, and emission, on systems containing a low-barrier hydrogen bond. We find several distinguishing spectral features in these UV spectra for systems that possess a LBHB.     

Hydrogen-bonding partner of the proton-conducting histidine in the influenza m2 proton channel revealed from (1)h chemical shifts

The influenza M2 protein conducts protons through a critical histidine (His) residue, His37. Whether His37 only interacts with water to relay protons into the virion or whether a low-barrier hydrogen bond (LBHB) also exists between the histidines to stabilize charges before proton conduction is actively debated. To address this question, we have measured the imidazole (1)H(N) chemical shifts of His37 at different temperatures and pH using 2D (15)N-(1)H correlation solid-state NMR. At low temperature, the H(N) chemical shifts are 8-15 ppm at all pH values, indicating that the His37 side chain forms conventional hydrogen bonds (H-bonds) instead of LBHBs. At ambient temperature, the dynamically averaged H(N) chemical shifts are 4.8 ppm, indicating that the H-bonding partner of the imidazole is water instead of another histidine in the tetrameric channel. These data show that His37 forms H-bonds only to water, with regular strength, thus supporting the His-water proton exchange model and ruling out the low-barrier H-bonded dimer model.     

Theoretical study of CH…O hydrogen bond in proton transfer reaction of glycine

Density functional theory (DFT) calculations are used to study the strength of the CH…O H‐bond in the proton transfer reaction of glycine. Comparison has been made between four proton transfer reactions (ZW1, ZW2, ZW3, SCRFZW) in glycine. The structural parameters of the zwitterionic, transition, and neutral states of glycine are strongly perturbed when the proton transfer takes place. It has been found that the interaction of water molecule at the side chain of glycine is high in the transition state, whereas it is low in the zwitterionic and neutral states. This strongest multiple hydrogen bond interaction in the transition state reduces the barrier for the proton transfer reaction. The natural bond orbital analysis have also been carried out for the ZW2 reaction path, it has been concluded that the amount of charge transfer between the neighboring atoms will decide the strength of H‐bond. © 2005 Wiley Periodicals, Inc. Int J Quantum Chem, 2006     

Neutral and ionic hydrogen bonding in Schiff bases

Low-temperature, high-resolution X-ray studies of charge distributions in the three Schiff bases, the dianil of 2-hydroxy-5-methylisophthaldehyde, 3,5-dinitro-N-salicylidenoethylamine and 3-nitro-N-salicylidenocyclohexylamine, have been carried out. These structures exhibit interesting weak interactions, including two extreme cases of intramolecular hydrogen bonds that are ionic N(+)-H...O- and neutral O-H...N in nature. These two types of hydrogen bond reflect differences in geometrical parameters and electron density distribution. At the level of geometry, the neutral O-H...N hydrogen bond is accompanied by an increase in the length of the C(1)-O(1) bond, opening of the ipso-C(1) angle, elongation of the aromatic C-C bonds, shortening of the C(7)-N(2) bond and increased length of the C(1)-C(7) bond, relative to the ionic hydrogen bond type. According to the geometrical and critical point parameters, the neutral O-H...N hydrogen bond seems to be stronger than the ionic ones. There are also differences between charge density parameters of the aromatic rings consistent with the neutral hydrogen bond being stronger than the ionic ones, with a concomitant reduction in the aromaticity of the ring. Compounds with the ionic hydrogen bonds show a larger double-bond character in the C-O bond than appears in the compound containing a neutral hydrogen bond; this suggests that the electronic structure of the former pair of compounds includes a contribution from a zwitterionic canonical form. Furthermore, in the case of ionic hydrogen bonds, the corresponding interaction lines appear to be curved in the vicinity of the hydrogen atoms. In the 3-nitro-N-salicylidenocyclohexylamine crystal there exists, in addition to the intramolecular hydrogen bond, a pair of intermolecular O...H interactions in a centrosymmetric dimer unit.     

Theoretical studies on the deacylation step of serine protease catalysis in the gas phase, in solution, and in elastase

The deacylation step of serine protease catalysis is studied using DFT and ab initio QM/MM calculations combined with MD/umbrella sampling calculations. Free energies of the entire reaction are calculated in the gas phase, in a continuum solvent, and in the enzyme elastase. The calculations show that a concerted mechanism in the gas phase is replaced by a stepwise mechanism when solvent effects or an acetate ion are added to the reference system, with the tetrahedral intermediate being a shallow minimum on the free energy surface. In the enzyme, the tetrahedral intermediate is a relatively stable species ( approximately 7 kcal/mol lower in energy than the transition state), mainly due to the electrostatic effects of the oxyanion hole and Asp102. It is formed in the first step of the reaction, as a result of a proton transfer from the nucleophilic water to His57 and of an attack of the remaining hydroxyl on the ester carbonyl. This is the rate-determining step of the reaction, which requires approximately 22 kcal/mol for activation, approximately 5 kcal/mol less than the reference reaction in water. In the second stage of the reaction, only small energy barriers are detected to facilitate the proton transfer from His57 to Ser195 and the breakdown of the tetrahedral intermediate. Those are attributed mainly to a movement of Ser195 and to a rotation of the His57 side chain. During the rotation, the imidazolium ion is stabilized by a strong H-bond with Asp102, and the C(epsilon)(1)-H...O H-bond with Ser214 is replaced by one with Thr213, suggesting that a "ring-flip mechanism" is not necessary as a driving force for the reaction. The movements of His57 and Ser195 are highly correlated with rearrangements of the binding site, suggesting that product release may be implicated in the deacylation process.     

Role of Asp102 in the catalytic relay system of serine proteases: a theoretical study

The role of Asp102 in the catalytic relay system of serine proteases is studied theoretically by calculating the free energy profiles of the single proton-transfer reaction by the Asn102 mutant trypsin and the concerted double proton-transfer reaction (so-called the charge-relay mechanism) of the wild-type trypsin. For each reaction, the reaction free energy profile of the rate-determining step (the tetrahedral intermediate formation step) is calculated by using ab initio QM/MM electronic structure calculations combined with molecular dynamics-free energy perturbation method. In the mutant reaction, the free energy monotonically increases along the reaction path. The rate-determining step of the mutant reaction is the formation of tetrahedral intermediate complex, not the base (His57) abstraction of the proton from Ser195. In contrast to the single proton-transfer reaction of the wild-type, MD simulations of the enzyme-substrate complex show that the catalytically favorable alignment of the relay system (the hydrogen bonding network between the mutant triad, His57, Asn102, and Ser195) is rarely observed even in the presence of a substrate at the active site. In the double proton-transfer reaction, the energy barrier is observed at the proton abstraction step, which corresponds to the rate-determining step of the single proton-transfer reaction of the wild-type. Although both reaction profiles show an increase of the activation barrier by several kcals/mol, these increases have different energetic origins: a large energetic loss of the electrostatic stabilization between His57 and Asn102 in the mutant reaction, while the lack of stabilization by the protein environment in the double proton-transfer reaction. Comparing the present results with the single proton transfer of the wild-type, Asp102 is proven to play two important roles in the catalytic process. One is to stabilize the protonated His57, or ionic intermediate, formed during the acylation, and the other is to fix the configuration around the active site, which is favorable to promote the catalytic process. These two factors are closely related to each other and are indispensable for the efficient catalysis. Also the present calculations suggest the importance of the remote site interaction between His57 and Val213-Ser214 at the catalytic transition state.     

A quantum description of the proton movement in an idealized NHN+ bridge

A series of model calculations was done to analyze the delocalization of the proton in the linking hydrogen bond of the (Dih)(2)H(+) cation (Dih: 4,5-dihydro-1H-imidazole). Standard quantum chemical calculations (B3LYP/D95+(d,p)) predict a low barrier hydrogen bond (LBHB) and thereby a delocalized proton in the NHN(+) hydrogen bridge. Explicit quantum calculations on the proton indicate that the delocalization of the proton does not provide enough energy to stabilize a permanent LBHB. Additional Born-Oppenheimer Molecular Dynamics (BOMD) simulations indicate further that the proton is localized at either side of the NHN(+) bridge and that a central proton position is the result of temporal averaging. The possibility of the proton to tunnel from one side to the other side of the NHN(+) bridge increases with the temperature as the trajectory of the (Dih)(2)H(+) cation runs through regions where the thermal excitation of Dih ring vibrations creates equal bonding opportunities for the proton on both sides of the bridge (vibrationally assisted proton tunneling). The quantum calculations for the proton in (Dih)(2)H(+) suggest further a broad peak for the 1 ← 0 transition with a maximum at 938 cm(-1) similar to that observed for LBHBs. Moreover, the asymmetric NHN(+) bridge in a thermally fluctuating environment is strong enough to create a significant peak at 1828 cm(-1) for the 2 ← 0 transition, while contributions from the 2 ← 1 are expected to be weak for the same reason.     

Testing geometrical discrimination within an enzyme active site: constrained hydrogen bonding in the ketosteroid isomerase oxyanion hole

Enzymes are classically proposed to accelerate reactions by binding substrates within active-site environments that are structurally preorganized to optimize binding interactions with reaction transition states rather than ground states. This is a remarkably formidable task considering the limited 0.1-1 A scale of most substrate rearrangements. The flexibility of active-site functional groups along the coordinate of substrate rearrangement, the distance scale on which enzymes can distinguish structural rearrangement, and the energetic significance of discrimination on that scale remain open questions that are fundamental to a basic physical understanding of enzyme active sites and catalysis. We bring together 1.2-1.5 A resolution X-ray crystallography, (1)H and (19)F NMR spectroscopy, quantum mechanical calculations, and transition-state analogue binding measurements to test the distance scale on which noncovalent forces can constrain the structural relaxation or translation of side chains and ligands along a specific coordinate and the energetic consequences of such geometric constraints within the active site of bacterial ketosteroid isomerase (KSI). Our results strongly suggest that packing and binding interactions within the KSI active site can constrain local side-chain reorientation and prevent hydrogen bond shortening by 0.1 A or less. Further, this constraint has substantial energetic effects on ligand binding and stabilization of negative charge within the oxyanion hole. These results provide evidence that subtle geometric effects, indistinguishable in most X-ray crystallographic structures, can have significant energetic consequences and highlight the importance of using synergistic experimental approaches to dissect enzyme function.     

Radiative relaxation and fragmentation dynamics of S 2p-excited hydrogen sulfide

Radiative relaxation of S 2p-excited hydrogen sulfide (H(2)S) is investigated by dispersed ultraviolet and visible fluorescence spectroscopies. We observe distinct changes in the fluorescence spectra as a function of excitation energy. Excitation to Rydberg states below the S 2p ionization threshold yields intense fluorescence from neutral and ionic atomic fragments (H, S(+), and S(2+)). In addition to the atomic emission, fluorescence of the molecular fragment ion HS(+) is preferably found after excitation of the S 2p electron into the unoccupied 6a(1) and 3b(2) orbitals with sigma(*) character. This is interpreted as evidence for ultrafast dissociation of the core-excited molecule prior to electronic relaxation. The rotationally resolved fluorescence spectra of the A (3)Pi-->X (3)Sigma(-) transition are analyzed in terms of the fragmentation dynamics leading to the formation of the excited molecular fragment ion, where changes in bond angle are discussed in terms of the rotational population.     

On the possibility of detecting low barrier hydrogen bonds with kinetic measurements

Recent experimental evidence has pointed to the possible presence of a short, strong hydrogen bond in the enzyme-substrate transition states in some biochemical reactions. To date, most experimental measures of these short, strong hydrogen bonds have monitored their equilibrium properties. In this work we show that kinetic measurements can also be used to detect the presence of short, strong hydrogen bonds. In particular, we find nontrivial differences among rate constant ratios of protonated to deuterated hydrogen bonds between strong and weak hydrogen bonds for proton transfer between donor-acceptor sites. We quantify this kinetic isotope effect by performing dynamical calculations of these rate constants by computing reactive flux through a dividing surface. This reactive flux is computed by evolving trajectories on an effective quantum mechanical potential energy surface.     

Very strong C-H...O, N-H...O, and O-H...O hydrogen bonds involving a cyclic phosphate.

Very short C-H...O, N-H...O, and O-H...O hydrogen bonds have been generated utilizing the cyclic phosphate [CH2(6-t-Bu-4-Me-C6H2O)2]P(O)OH (1). X-ray structures of (i) 1 (unsolvated, two polymorphs), 1...EtOH, and 1...MeOH, (ii) [imidazolium](+)[CH2(6-t-Bu-4-Me-C6H2O)2PO2](-)...MeOH [2], (iii) [HNC5H4-N=N-C5H4NH](2+)[(CH2(6-t-Bu-4-Me-C6H2O)2PO2)2](2-)...4CH3CN...H2O [3], (v) [K, 18-crown-6](+)[(CH2(6-t-Bu-4-Me-C6H2O)2P(O)OH)(CH2(6-t-Bu-4-Me-C6H2O)2PO2)](-)...2THF [4], (vi) 1...cytosine...MeOH [5], (vii) 1...adenine...1/2MeOH [6], and (viii) 1...S-(-)-proline [7] have been determined. The phosphate 1 in both its forms is a hydrogen-bonded dimer with a short O-H...O distance of 2.481(2) [triclinic form] or 2.507(3) A [monoclinic form]. Compound 2 has a helical structure with a very short C-H...O hydrogen bond involving an imidazolyl C-H and methanol in addition to N-H...O hydrogen bonds. A helical motif is also seen in 5. In 3, an extremely short N-H...O hydrogen bond [N...O 2.558(4) A] is observed. Compounds 6 and 7 also exhibit short N-H...O hydrogen bonds. In 1...EtOH, a 12-membered hydrogen-bonded ring motif, with one of the shortest known O-H...O hydrogen bonds [O...O 2.368(4) A], is present. 1...MeOH is a similar dimer with a very short O(-H)...O bond [2.429(3) A]. In 4, the deprotonated phosphate (anion) and the parent acid are held together by a hydrogen bond on one side and a coordinate/covalent bond to potassium on the other; the O-H...O bond is symmetrical and very strong [O...O 2.397(3) A].     

Bonding and stability of the hydrogen storage material Mg(2)NiH(4)

Structural stability and bonding properties of the hydrogen storage material Mg(2)NiH(4) (monoclinic, C2/c, Z = 8) were investigated and compared to those of Ba(2)PdH(4) (orthorhombic, Pnma, Z = 8) using ab initio density functional calculations. Both compounds belong to the family of complex transition metal hydrides. Their crystal structures contain discrete tetrahedral 18 electron complexes T(0)H(4)(4-) (T = Ni, Pd). However, the bonding situation in the two systems was found to be quite different. For Ba(2)PdH(4), the electronic density of states mirrors perfectly the molecular states of the complex PdH(4)(4-), whereas for Mg(2)NiH(4) a clear relation between molecular states of TH(4)(4-) and the density of states of the solid-state compound is missing. Differences in bonding of Ba(2)PdH(4) and Mg(2)NiH(4) originate in the different strength of the T-H interactions (Pd[bond]H interactions are considerably stronger than Ni[bond]H ones) and in the different strength of the interaction between the alkaline-earth metal component and H (Ba[bond]H interactions are substantially weaker than Mg[bond]H ones). To lower the hydrogen desorption temperature of Mg(2)NiH(4), it is suggested to destabilize this compound by introducing defects in the counterion matrix surrounding the tetrahedral Ni(0)H(4)(4-) complexes. This might be achieved by substituting Mg for Al.