NAD Enzyme Basics: Cofactor Binding, Specificity and Assays

An NAD enzyme is any enzyme that uses nicotinamide adenine dinucleotide as a cofactor or substrate, in either its oxidized (NAD+) or reduced (NADH) form. Most are oxidoreductases that move a hydride between the nicotinamide ring and a substrate. A smaller group consumes NAD+ to modify or demodify proteins, for example ADP-ribosyltransferases and sirtuins.

These enzymes are widely studied in protein research, in part because their activity can often be followed directly by spectrophotometry. This article covers cofactor chemistry, binding-site structure, engineered specificity, biocatalytic use, assay design and stability.

The chemistry of the cofactor

NAD+ is two nucleotides joined through a pyrophosphate bridge. One half is adenosine monophosphate. The other is nicotinamide mononucleotide. All redox chemistry happens at C4 of the nicotinamide ring, which accepts or donates a hydride ion (two electrons and one proton).

NADP+ differs from NAD+ by a single phosphate on the 2′ hydroxyl of the adenosine ribose. That group sits far from the reactive ring. Even so, it is the main handle enzymes use to tell the two cofactors apart. In cells, NADH mainly supports catabolic oxidation, while NADPH supplies reducing power for biosynthesis.

The reduced forms have a convenient optical property. NADH and NADPH absorb at 340 nm, while the oxidized forms do not. The molar absorption coefficient at 340 nm is close to 6.22 mM-1 cm-1, and most continuous assays for these enzymes rely on it.

Where the cofactor binds

Cofactor binding sites have been studied by crystallography for decades. Two features recur: a conserved binding fold, and in some enzymes, large movements of whole domains during catalysis.

The Rossmann fold

Many NAD-dependent dehydrogenases bind the cofactor in a Rossmann fold. This is a repeating beta-alpha-beta unit with a glycine-rich loop (often written GXGXXG) that contacts the pyrophosphate. The glycines let the loop pack closely against the cofactor. Several backbone NH groups in the loop, each part of a peptide bond, hydrogen bond to the phosphate oxygens, helped by the N-terminus of the following helix.

Specificity is usually set near the adenosine ribose. An acidic residue, often aspartate, at the end of the second beta strand hydrogen bonds to the 2′ and 3′ hydroxyls and repels the extra phosphate of NADP+. Enzymes that prefer NADP+ tend to place arginine or lysine there to pair with that phosphate. These are tendencies, not strict rules, and exceptions are common.

Domain movements in HMG-CoA reductase

Class II HMG-CoA reductase shows that binding is not always a static fit. This enzyme reduces HMG-CoA to mevalonate using NAD(P)H. Miller and Kung note that HMGRs from different organisms vary widely in their specificity for NADH or NADPH, and that the structural basis for this range was unknown [1]. They solved crystal structures of the enzyme from Streptococcus pneumoniae and paired them with kinetic data. The kinetically preferred cofactor was NADPH, and the NADPH-bound structure suggested how that cofactor is recognized [1].

The HMG-CoA-bound structure revealed large domain movements that had not been described before. The authors proposed that these movements may control substrate binding and allow cofactor exchange without releasing the reaction intermediate [1]. The overall reaction needs two hydride transfers, so a spent cofactor must leave and a fresh one must enter mid-cycle.

Engineering cofactor preference

Cofactor sites can be redesigned to accept analogues. Nicotinamide cytosine dinucleotide (NCD) replaces the adenine of NAD with cytosine. Liu and colleagues describe NCD as a biocompatible NAD analogue and note that several redox enzymes have been engineered to favor it and used to create NCD-dependent redox systems [2].

In the same study, the group solved the X-ray structure of an NCD-favoring malic enzyme variant from Escherichia coli. Residues at positions 346 and 401 acted as gatekeepers of the cavity that normally holds the adenine moiety [2]. When Arg346 was replaced with acidic or aromatic residues, the mutants showed much lower NCD preference [2].

The team then tested whether the finding transferred to a different enzyme. In Lactobacillus helveticus D-lactate dehydrogenase, Ile177 occupies the position equivalent to Arg346. Variants at Ile177 showed a similar trend in cofactor preference [2]. Because many NAD-dependent oxidoreductases share key structural features, the authors suggested these gatekeeper positions can guide further engineering.

A small number of residues around the nucleobase pocket can therefore shift preference. Structural alignment between distinct dehydrogenases is a reasonable first step for choosing mutation sites, though each new enzyme still needs kinetic testing.

NAD-dependent enzymes in biocatalytic cascades

Cofactor cost and recycling often limit preparative enzyme reactions. A well-matched pair of enzymes can recycle the cofactor internally. Tonin, Otten and Arends built such a pair for the epimerization of the 7-hydroxyl group in cholic acid and chenodeoxycholic acid [3].

The route uses two enantiocomplementary hydroxysteroid dehydrogenases. A 7α-HSDH oxidizes the 7α-OH to a ketone and produces NADH. A 7β-HSDH then reduces the ketone to the 7β-OH and returns NAD+ [3]. For this to be redox-neutral, both enzymes must use the same cofactor. The group identified an NADH-active 7β-HSDH from Lactobacillus spicheri, then expressed, purified and characterized it.

Combined with a 7α-HSDH from Stenotrophomonas maltophilia, the system converted both bile acids with only catalytic amounts of NAD+. Yields of the 7β epimers exceeded 90% [3]. Because the cascade is redox-neutral by design, it does not call for a separate regeneration enzyme or a sacrificial co-substrate.

NAD+ in protein acetylation and deacetylation

Not every NAD-dependent enzyme performs redox chemistry. The sirtuin family of histone deacetylases cleaves NAD+ during each deacetylation. They release nicotinamide and transfer the acetyl group to the ADP-ribose fragment, giving O-acetyl-ADP-ribose.

Shen, Wei and Zhou reviewed how this links metabolism to gene regulation. Histone acetyltransferases use acetyl-CoA, and deacetylation can depend on NAD+. The cellular levels of these two metabolites may therefore regulate acetyltransferase and deacetylase activity [4]. The review also notes that many acetylated proteins belong to metabolic pathways, including photosynthesis and respiration in plants, and that histone-modifying enzymes are emerging as regulators of metabolic enzyme acetylation [4].

The authors frame metabolite sensing as an open question. Whether and how these enzymes read cellular NAD+ or acetyl-CoA signals still needs direct testing [4].

Designing activity assays

The 340 nm signal makes continuous assays simple, but several details decide whether the rates are meaningful.

• Initial rates. Fit only the early linear portion of the trace, typically the first few percent of conversion.
• Absorbance range. Keep starting NADH low enough that total absorbance stays within the linear range of the instrument.
• Cofactor saturation. When studying the substrate, hold the cofactor well above its apparent Km, and the reverse when studying the cofactor.
• Background controls. Run no-substrate and no-enzyme blanks. Crude lysates often contain NADH oxidase activity.
• Specificity reporting. Compare kcat/Km for each cofactor rather than rates at a single concentration.

Coupled assays extend the range. In the bacterial luminescence system, NAD(P)H:FMN oxidoreductase reduces FMN, and luciferase uses the reduced flavin to emit light. Such coupled systems are sensitive but have more points of failure, so each component should be checked on its own.

The table below collects the cofactor findings from the studies discussed above.

Enzyme	Source organism	Cofactor point	Study
HMG-CoA reductase (class II)	S. pneumoniae	Kinetic preference for NADPH	[1]
Malic enzyme variant	E. coli	Favors NCD; residues 346 and 401 gate the adenine pocket	[2]
D-lactate dehydrogenase variants	L. helveticus	Ile177 changes alter NCD preference	[2]
7β-HSDH	L. spicheri	NADH-active, enables an NAD+-only cascade	[3]

Stability, storage and handling

The cofactor and the enzyme each degrade in their own way. NADH is labile in acid, while NAD+ is more labile in alkaline solution. Prepare cofactor stocks in a buffer suited to the form in use, aliquot them, and keep them frozen. Check NADH stocks by absorbance before use rather than trusting the weighed mass.

Short peptides often have little stable tertiary structure, while enzymes depend on a precise fold (an overview of the types of peptides covers the shorter chains). A folded enzyme can lose activity through unfolding and aggregation, not only through chemical breakdown. Crude or partly purified preparations may also carry proteases that clip the chain, so protease inhibitors are often added during purification.

Matrix choice can help. Esimbekova and colleagues studied thermal inactivation of butyrylcholinesterase, NAD(P)H:FMN oxidoreductase and the coupled oxidoreductase-luciferase system of luminous bacteria in viscous starch and gelatin environments [5]. From the kinetic curves they proposed inactivation mechanisms and identified conditions under which both polymers stabilized activity during storage and exposure to physical and chemical stresses. Removing water by drying the enzymes immobilized in these gels gave a significant further gain in stability [5].

Reading the literature on NAD enzymes

A few checks make papers in this field easier to judge.

1. Confirm how cofactor preference is defined. A ratio of catalytic efficiencies is more informative than a single activity ratio.
2. For crystal structures, note the resolution and whether ligand density is clear. Proposed domain motions from static structures are hypotheses until tested.
3. For engineering work, check whether mutants were tested in more than one enzyme scaffold, as in the malic enzyme and lactate dehydrogenase study.
4. For biocatalysis, distinguish conversion measured by chromatography from isolated yield.
5. For acetylation and signaling claims, separate correlation in metabolite levels from direct biochemical evidence of sensing.

Organism of origin also matters. Cofactor specificity within one enzyme class can vary between organisms, as noted in the HMG-CoA reductase work [1], so results from one source should not be assumed for another.

References

1. Miller BR, Kung Y. Structural Features and Domain Movements Controlling Substrate Binding and Cofactor Specificity in Class II HMG-CoA Reductase. Biochemistry. 2018. PMID: 29224355. doi:10.1021/acs.biochem.7b00999.
2. Liu Y, Guo X, Liu W, et al. Structural Insights into Malic Enzyme Variants Favoring an Unnatural Redox Cofactor. Chembiochem. 2021. PMID: 33523590. doi:10.1002/cbic.202000800.
3. Tonin F, Otten LG, Arends IWCE. NAD+ -Dependent Enzymatic Route for the Epimerization of Hydroxysteroids. ChemSusChem. 2019. PMID: 30265441. doi:10.1002/cssc.201801862.
4. Shen Y, Wei W, Zhou DX. Histone Acetylation Enzymes Coordinate Metabolism and Gene Expression. Trends Plant Sci. 2015. PMID: 26440431. doi:10.1016/j.tplants.2015.07.005.
5. Esimbekova EN, Govorun AE, Lonshakova-Mukina VI, et al. Gelatin and Starch: What Better Stabilizes the Enzyme Activity?. Dokl Biol Sci. 2020. PMID: 32483706. doi:10.1134/S0012496620020039.