Benchmarking Density Functional Theory Methods for Metalloenzyme Reactions: The Introduction of the MME55 Set

Abstract

We present a new benchmark set of metalloenzyme model reaction energies and barrier heights that we call MME55. The set contains 10 different enzymes, representing eight transition metals, both open and closed shell systems, and system sizes of up to 116 atoms. We use four DLPNO–CCSD(T)-based approaches to calculate reference values against which we then benchmark the performance of a range of density functional approximations with and without dispersion corrections. Dispersion corrections improve the results across the board, and triple-ζ basis sets provide the best balance of efficiency and accuracy. Jacob’s ladder is reproduced for the whole set based on averaged mean absolute (percent) deviations, with the double hybrids SOS0-PBE0-2-D3(BJ) and revDOD-PBEP86-D4 standing out as the most accurate methods for the MME55 set. The range-separated hybrids ωB97M-V and ωB97X-V also perform well here and can be recommended as a reliable compromise between accuracy and efficiency; they have already been shown to be robust across many other types of chemical problems, as well. Despite the popularity of B3LYP in computational enzymology, it is not a strong performer on our benchmark set, and we discourage its use for enzyme energetics.

Exploring CPS-Extrapolated DLPNO–CCSD(T1) Reference Values for Benchmarking DFT Methods on Enzymatically Catalyzed Reactions

“…While we do not consider these lower-level approaches as potential benchmark strategies in the subsequent DFA tests, they are useful for observing trends across the REs and BHs with the T CutPNO values, and the CBS­(2,3) results allow us to compare CPS­(6,7) to the tightened value of T CutPNO = 10 –8 , which could not be done for CBS­(3,4) as the def2-QZVPP calculations were computationally too expensive. We also note that the CPS­(5,6) and TightPNO results at the CBS­(3,4) level for the metalloenzymes CDO, Hc, and NiSOD have been taken from the brief tests we conducted when selecting the reference values for our new MME55 set, where CPS­(5,6) was referred to as “estimated TightPNO”.…”

“…DLPNO–CCSD­(T)/TightPNO is regularly used to calculate reference values in benchmark studies of both organic and organometallic systems, − including our work on models of enzymatically catalyzed reactions. − Recent work has shown that the efficiency and accuracy of TightPNO calculations can be improved by altering the T CutPNO parameter, which determines the number of PNOs used in the correlation treatment for each electron pairPNOs with an occupation number below the threshold value are discarded. The default value of T CutPNO for TightPNO is 10 –7 , but this can be approximated with extrapolation of values calculated with T CutPNO = 10 –5 and 10 –6 (using the TightPNO default values for all other thresholds), at a significantly reduced computational cost.…”

“…Previous analysis has already focused on the accuracy of CPS results in comparison to canonical CCSD­(T) or other ab initio methods, , so here we look at how much the outcomes of benchmark studies change when CPS-extrapolated reference values are used instead of standard TightPNO ones. For this we use the small test set described in Figure , which contains 10 reaction energies (REs) and 4 barrier heights (BHs) across active site models of seven enzymesthese are cysteine dioxygenase (CDO), hemocyanin (Hc), and superoxide dismutase (NiSOD) from our MME55 set of metalloenzyme models and 4-oxalocrotonate tautomerase (4-OT), histone-lysine methyltransferase (HKMT), l -aspartate α-decarboxylase (AspDC), and haloalcohol dehydrogenase (HheC) from our set of mainly organic enzymatically catalyzed reactions . All of these are closed shell species.…”

“…We point out that all of these models contain only parts of the active site, not the entire enzyme; our aim in these previous benchmark studies has been to recommend reliable DFAs for the active site region that can then be used for cluster model studies or combined with force field methods in layered QM/MM approaches. Geometries for all models have previously been optimized at the PBEh-3c level of theory, which contains the DFT-D3­(BJ) , correction to improve the description of London dispersion interactions, and we direct the reader to the Supporting Information of refs and for the provided structures in xyz format.…”

“…Schemes of the enzymatically catalyzed reaction steps used to test CPS extrapolations in this studythe test set includes the reaction energies of all steps as well as the forward and reverse barriers of 4-OT and HKMT. The models of 4-OT, HKMT, AspDC, and HheC have been taken from the ENZYMES28 set (see ref for specific details of each model system), while CDO, Hc, and NiSOD have been taken from our MME55 set …”