Limits...
Parametrization of DFTB3/3OB for magnesium and zinc for chemical and biological applications.

Lu X, Gaus M, Elstner M, Cui Q - J Phys Chem B (2014)

Bottom Line: The gas-phase results are compared to DFT (mostly B3LYP), ab initio (MP2 and G3B3), and PM6, as well as to a previous DFTB parametrization (MIO).The results indicate that DFTB3/3OB is particularly successful at predicting structures, including rather complex dinuclear metalloenzyme active sites, while being semiquantitative (with a typical mean absolute deviation (MAD) of ∼3-5 kcal/mol) for energetics.The remaining limitations of DFTB3, such as the treatment of interaction between metal ions and highly charged/polarizable ligands, are also discussed.

View Article: PubMed Central - PubMed

Affiliation: Department of Chemistry and Theoretical Chemistry Institute, University of Wisconsin-Madison , 1101 University Avenue, Madison, Wisconsin 53706, United States.

ABSTRACT
We report the parametrization of the approximate density functional theory, DFTB3, for magnesium and zinc for chemical and biological applications. The parametrization strategy follows that established in previous work that parametrized several key main group elements (O, N, C, H, P, and S). This 3OB set of parameters can thus be used to study many chemical and biochemical systems. The parameters are benchmarked using both gas-phase and condensed-phase systems. The gas-phase results are compared to DFT (mostly B3LYP), ab initio (MP2 and G3B3), and PM6, as well as to a previous DFTB parametrization (MIO). The results indicate that DFTB3/3OB is particularly successful at predicting structures, including rather complex dinuclear metalloenzyme active sites, while being semiquantitative (with a typical mean absolute deviation (MAD) of ∼3-5 kcal/mol) for energetics. Single-point calculations with high-level quantum mechanics (QM) methods generally lead to very satisfying (a typical MAD of ∼1 kcal/mol) energetic properties. DFTB3/MM simulations for solution and two enzyme systems also lead to encouraging structural and energetic properties in comparison to available experimental data. The remaining limitations of DFTB3, such as the treatment of interaction between metal ions and highly charged/polarizable ligands, are also discussed.

Show MeSH
Radialdistribution function of water oxygen around Mg2+/Zn2+ in 20 Å water droplet. The pure MM model isbased on the CHARMM force field.105
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4306495&req=5

fig4: Radialdistribution function of water oxygen around Mg2+/Zn2+ in 20 Å water droplet. The pure MM model isbased on the CHARMM force field.105

Mentions: The metal ion-Oradial distribution function (g(r)) and coordination numbers are shown in Figure 4 for the two cationsin solution. A well-defined first solvation shell is observed in bothcases. The first peak in DFTB3/MM hydrated Mg2+ occursat 2.07 Å, in good agreement with X-ray diffraction data of 2.09Å ± 0.04 Å144 and with previousAIMD simulations in the range of 2.08–2.13 Å.145−149 The MM force field predicts the Mg2+–O distanceto be shorter than the DFTB3/MM model with a sharper peak at 2.0 Å.The inner coordination sphere remains similar to the gas-phase structure.For the cluster [Mg(H2O)6]2+ in thegas phase, the Mg2+–O distance is 2.10 Å inboth DFTB3 and B3LYP/aug-cc-pVTZ optimized structures. For Zn2+, the first maximum from DFTB3/MM is at 2.11 Å, consistentwith X-ray absorption measurements 2.06 ± 0.02 Å150 and earlier AIMD results 2.07–2.13 Å.145,146,151 The classical MM model givesthe first peak at 2.08 Å, although the peak is higher and narrowerthan in the DFTB3/MM model. The small difference in the location ofthe first peak in DFTB3/MM may be traced back to the gas-phase model:for the cluster [Zn(H2O)6]2+ in thegas phase, Zn2+–O is 2.16 Å in DFTB3 versus2.12 Å in B3LYP/aug-cc-pVTZ.


Parametrization of DFTB3/3OB for magnesium and zinc for chemical and biological applications.

Lu X, Gaus M, Elstner M, Cui Q - J Phys Chem B (2014)

Radialdistribution function of water oxygen around Mg2+/Zn2+ in 20 Å water droplet. The pure MM model isbased on the CHARMM force field.105
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4306495&req=5

fig4: Radialdistribution function of water oxygen around Mg2+/Zn2+ in 20 Å water droplet. The pure MM model isbased on the CHARMM force field.105
Mentions: The metal ion-Oradial distribution function (g(r)) and coordination numbers are shown in Figure 4 for the two cationsin solution. A well-defined first solvation shell is observed in bothcases. The first peak in DFTB3/MM hydrated Mg2+ occursat 2.07 Å, in good agreement with X-ray diffraction data of 2.09Å ± 0.04 Å144 and with previousAIMD simulations in the range of 2.08–2.13 Å.145−149 The MM force field predicts the Mg2+–O distanceto be shorter than the DFTB3/MM model with a sharper peak at 2.0 Å.The inner coordination sphere remains similar to the gas-phase structure.For the cluster [Mg(H2O)6]2+ in thegas phase, the Mg2+–O distance is 2.10 Å inboth DFTB3 and B3LYP/aug-cc-pVTZ optimized structures. For Zn2+, the first maximum from DFTB3/MM is at 2.11 Å, consistentwith X-ray absorption measurements 2.06 ± 0.02 Å150 and earlier AIMD results 2.07–2.13 Å.145,146,151 The classical MM model givesthe first peak at 2.08 Å, although the peak is higher and narrowerthan in the DFTB3/MM model. The small difference in the location ofthe first peak in DFTB3/MM may be traced back to the gas-phase model:for the cluster [Zn(H2O)6]2+ in thegas phase, Zn2+–O is 2.16 Å in DFTB3 versus2.12 Å in B3LYP/aug-cc-pVTZ.

Bottom Line: The gas-phase results are compared to DFT (mostly B3LYP), ab initio (MP2 and G3B3), and PM6, as well as to a previous DFTB parametrization (MIO).The results indicate that DFTB3/3OB is particularly successful at predicting structures, including rather complex dinuclear metalloenzyme active sites, while being semiquantitative (with a typical mean absolute deviation (MAD) of ∼3-5 kcal/mol) for energetics.The remaining limitations of DFTB3, such as the treatment of interaction between metal ions and highly charged/polarizable ligands, are also discussed.

View Article: PubMed Central - PubMed

Affiliation: Department of Chemistry and Theoretical Chemistry Institute, University of Wisconsin-Madison , 1101 University Avenue, Madison, Wisconsin 53706, United States.

ABSTRACT
We report the parametrization of the approximate density functional theory, DFTB3, for magnesium and zinc for chemical and biological applications. The parametrization strategy follows that established in previous work that parametrized several key main group elements (O, N, C, H, P, and S). This 3OB set of parameters can thus be used to study many chemical and biochemical systems. The parameters are benchmarked using both gas-phase and condensed-phase systems. The gas-phase results are compared to DFT (mostly B3LYP), ab initio (MP2 and G3B3), and PM6, as well as to a previous DFTB parametrization (MIO). The results indicate that DFTB3/3OB is particularly successful at predicting structures, including rather complex dinuclear metalloenzyme active sites, while being semiquantitative (with a typical mean absolute deviation (MAD) of ∼3-5 kcal/mol) for energetics. Single-point calculations with high-level quantum mechanics (QM) methods generally lead to very satisfying (a typical MAD of ∼1 kcal/mol) energetic properties. DFTB3/MM simulations for solution and two enzyme systems also lead to encouraging structural and energetic properties in comparison to available experimental data. The remaining limitations of DFTB3, such as the treatment of interaction between metal ions and highly charged/polarizable ligands, are also discussed.

Show MeSH