How to Build a Metalloenzyme: Lessons from a Protein-Based Model of Acetyl Coenzyme A Synthase.

"What I cannot create, I do not understand"─Richard Feynman. This sentiment motivates the entire field of artificial metalloenzymes. Naturally occurring enzymes catalyze reactions with efficiencies, rates, and selectivity that generally cannot be achieved in synthetic systems. Many of these processes represent vital building blocks for a sustainable society, including CO2 conversion, nitrogen fixation, water oxidation, and liquid fuel synthesis. Our inability as chemists to fully reproduce the functionality of naturally occurring enzymes implicates yet-unknown contributors to reactivity. To identify these properties, it is necessary to consider all of the components of naturally occurring metalloenzymes, from the active site metal(s) to large-scale dynamics. In this Account, we describe the holistic development of a metalloprotein-based model that functionally reproduces the acetyl coenzyme A synthase (ACS) enzyme.ACS catalyzes the synthesis of a thioester, acetyl coenzyme A, from gaseous carbon monoxide, a methyl group donated by a cobalt corrinoid protein, and coenzyme A. The active site of ACS contains a bimetallic nickel site coupled to a [4Fe-4S] cluster. This reaction mimics Monsanto's acetic acid synthesis and represents an ancient process for incorporating inorganic carbon into cellular biomass through the primordial Wood-Ljungdahl metabolic pathway. From a sustainability standpoint, the reversible conversion of C1 substrates into an acetyl group and selective downstream transfer to a thiolate nucleophile offer opportunities to expand this reactivity to the anthropogenic synthesis of liquid fuels. However, substantial gaps in our understanding of the ACS catalytic mechanism coupled with the enzyme's oxygen sensitivity and general instability have limited these applications. It is our hope that development of an artificial metalloenzyme that carries out ACS-like reactions will advance our mechanistic understanding and enable synthesis of robust compounds with the capacity for similar reactivity.To construct this model, we first focused on the catalytic proximal nickel (NiP) site, which has a single metal center bound by three bridging cysteine residues in a "Y"-shaped arrangement. With an initial emphasis on reproducing the general structure of a low-coordinate metal binding site, the type I cupredoxin, azurin, was selected as the protein scaffold, and a nickel center was incorporated into the mononuclear site. Using numerous spectroscopic and computational techniques, including electron paramagnetic resonance (EPR) spectroscopy, nickel-substituted azurin was shown to have similar electronic and geometric structures to the NiP center in ACS. A substrate access channel was installed, and both carbon monoxide and a methyl group were shown to bind individually to the reduced NiI center. The elusive EPR-active S = 1/2 Ni-CH3 species, which has never been detected in native ACS, was observed in the azurin-based model, establishing the capacity of a biological NiI species to support two-electron organometallic reactions. Pulsed EPR studies on the S = 1/2 Ni-CH3 species in azurin suggested a noncanonical electronic structure with an inverted ligand field, which was proposed to prevent irreversible site degradation. This model azurin protein was ultimately shown to perform carbon-carbon and carbon-sulfur bond formation using sequential, ordered substrate addition for selective, stoichiometric thioester synthesis. X-ray spectroscopic methods were used to provide characterization of the remaining catalytic intermediates, resolving some debate over key mechanistic details.The overall approach and strategies that we employed for the successful construction of a functional protein-based model of ACS are described in this Account. We anticipate that these principles can be adapted across diverse metalloenzyme classes, providing essential mechanistic details and guiding the development of next-generation, functional artificial metalloenzymes.

iPSC-derived neuronal models of PANK2-associated neurodegeneration reveal mitochondrial dysfunction contributing to early disease.

Mutations in PANK2 lead to neurodegeneration with brain iron accumulation. PANK2 has a role in the biosynthesis of coenzyme A (CoA) from dietary vitamin B5, but the neuropathological mechanism and reasons for iron accumulation remain unknown. In this study, atypical patient-derived fibroblasts were reprogrammed into induced pluripotent stem cells (iPSCs) and subsequently differentiated into cortical neuronal cells for studying disease mechanisms in human neurons. We observed no changes in PANK2 expression between control and patient cells, but a reduction in protein levels was apparent in patient cells. CoA homeostasis and cellular iron handling were normal, mitochondrial function was affected; displaying activated NADH-related and inhibited FADH-related respiration, resulting in increased mitochondrial membrane potential. This led to increased reactive oxygen species generation and lipid peroxidation in patient-derived neurons. These data suggest that mitochondrial deficiency is an early feature of the disease process and can be explained by altered NADH/FADH substrate supply to oxidative phosphorylation. Intriguingly, iron chelation appeared to exacerbate the mitochondrial phenotype in both control and patient neuronal cells. This raises caution for the use iron chelation therapy in general when iron accumulation is absent.

Caloric restriction mimetics: natural/physiological pharmacological autophagy inducers.

Nutrient depletion, which is one of the physiological triggers of autophagy, results in the depletion of intracellular acetyl coenzyme A (AcCoA) coupled to the deacetylation of cellular proteins. We surmise that there are 3 possibilities to mimic these effects, namely (i) the depletion of cytosolic AcCoA by interfering with its biosynthesis, (ii) the inhibition of acetyltransferases, which are enzymes that transfer acetyl groups from AcCoA to other molecules, mostly leucine residues in cellular proteins, or (iii) the stimulation of deacetylases, which catalyze the removal of acetyl groups from leucine residues. There are several examples of rather nontoxic natural compounds that act as AcCoA depleting agents (e.g., hydroxycitrate), acetyltransferase inhibitors (e.g., anacardic acid, curcumin, epigallocatechin-3-gallate, garcinol, spermidine) or deacetylase activators (e.g., nicotinamide, resveratrol), and that are highly efficient inducers of autophagy in vitro and in vivo, in rodents. Another common characteristic of these agents is their capacity to reduce aging-associated diseases and to confer protective responses against ischemia-induced organ damage. Hence, we classify them as "caloric restriction mimetics" (CRM). Here, we speculate that CRM may mediate their broad health-improving effects by triggering the same molecular pathways that usually are elicited by long-term caloric restriction or short-term starvation and that imply the induction of autophagy as an obligatory event conferring organismal, organ- or cytoprotection.

Noninvasive measurement of murine hepatic acetyl-CoA ¹³C-enrichment following overnight feeding with ¹³C-enriched fructose and glucose.

The (13)C-isotopomer enrichment of hepatic cytosolic acetyl-CoA of overnight-fed mice whose drinking water was supplemented with [U-(13)C]fructose, and [1-(13)C]glucose and p-amino benzoic acid (PABA) was quantified by (13)C NMR analysis of urinary N-acetyl-PABA. Four mice were given normal chow plus drinking water supplemented with 5% [1-(13)C]glucose, 2.5% [U-(13)C]fructose, and 2.5% fructose (Solution 1) overnight. Four were given chow and water containing 17.5% [1-(13)C]glucose, 8.75% [U-(13)C]fructose and 8.75% fructose (Solution 2). PABA (0.25%) was present in both studies. Urinary N-acetyl-PABA was analyzed by (13)C NMR. In addition to [2-(13)C]- and [1,2-(13)C]acetyl isotopomers from catabolism of [U-(13)C]fructose and [1-(13)C]glucose to acetyl-CoA, [1-(13)C]acetyl was also found indicating pyruvate recycling activity. This precluded precise estimates of [1-(13)C]glucose contribution to acetyl-CoA while that of [U-(13)C]fructose was unaffected. The fructose contribution to acetyl-CoA from Solutions 1 and 2 was 4.0 ± 0.4% and 10.6 ± 0.6%, respectively, indicating that it contributed to a minor fraction of lipogenic acetyl-CoA under these conditions.

Inactivation of thiolase by 2-alkynoyl-CoA via its intrinsic isomerase activity.

Selective inactivation of cytosolic thiolase by 2-alkynoyl-CoA via its intrinsic isomerase activity was studied, which provides an example for rationally developing mechanism-based inhibitors based on a side activity of the enzyme, and may become a supplemental method for better treatment of cardiovascular disease and cancer.

Enzymatic determination of free carnitine in milk and infant formulas.

A simple spectrophotometric assay for free carnitine in milk and supplemented infant formulas has been developed. After acid extraction, carnitine is measured enzymatically through its reaction with carnitine acetyltransferase coupled with acetyl coenzyme A and dithiobenzoate (DTNB). The manually performed method is rapid, accurate, and convenient for routine quality-control analysis of infant formulas. The chemistry is also suitable for automation, for greatly enhanced throughput.