Common Coenzymes and Their Chemistry

The mechanistic logic of these five coenzymes underlies major swaths of modern biochemistry, pharmacology, and biotechnology. NAD⁺ metabolism is now a frontline target in aging and metabolic disease research: NAD⁺ precursors (nicotinamide riboside, nicotinamide mononucleotide) are marketed as supplements aiming to restore declining NAD⁺/NADH pools with age, and sirtuin deacetylases — which consume NAD⁺ stoichiometrically as a cosubstrate rather than a simple redox carrier — link cellular redox state directly to gene silencing and metabolic regulation. FAD-dependent monooxygenases and oxidases are exploited industrially (glucose oxidase in blood glucose test strips relies on FAD's ability to react directly with O₂, regenerating FADH₂ back to FAD while producing H₂O₂ that is then detected electrochemically). Coenzyme A chemistry sits at the heart of statin pharmacology and lipid metabolism: HMG-CoA reductase, the target of statins, uses a thioester substrate (HMG-CoA) exactly analogous to the thioester chemistry described here, and fatty acid synthase elongates acyl chains through repeated Claisen-condensation cycles chemically identical to the citrate synthase mechanism worked through above, but iterated with malonyl-CoA as the two-carbon donor. TPP chemistry is medically relevant in thiamine deficiency states (Wernicke-Korsakoff syndrome, beriberi), where impaired pyruvate dehydrogenase and α-ketoglutarate dehydrogenase activity — both TPP-dependent — cause severe neurological and cardiovascular pathology, and TPP-dependent enzymes are also industrial catalysts for asymmetric C–C bond-forming "carboligation" reactions in green chemistry (biocatalytic synthesis of chiral acyloins). PLP-dependent enzymes constitute one of the largest single mechanistic superfamilies in biology — essentially all amino acid metabolism (transamination, decarboxylation to neurotransmitters like GABA, dopamine, and serotonin, racemization for bacterial cell wall D-amino acids) runs through PLP chemistry, making PLP-dependent enzymes prime drug targets: vigabatrin inhibits GABA transaminase (a PLP enzyme) to raise brain GABA levels as an anticonvulsant, and antibacterial strategies increasingly target bacterial PLP-dependent alanine racemase (essential for peptidoglycan synthesis) to exploit chemistry absent from human cells. Across all five coenzymes, the shared thread — a small molecule supplying a functional group the polypeptide alphabet cannot — is also the design principle behind directed evolution and enzyme engineering efforts that swap or repurpose coenzyme-binding domains to create novel biocatalysts for pharmaceutical and industrial synthesis, and it is precisely this active-site logic that gets formalized kinetically once you move to the Michaelis-Menten Model treatment of enzyme rate laws.

Ask the tutor: How does the pyruvate dehydrogenase complex sequentially use TPP, lipoamide, CoA, FAD, and NAD⁺ in a single catalytic cycle to convert pyruvate to acetyl-CoA?

Think of an enzyme active site as a machine shop and a coenzyme as the one power tool the shop doesn't manufacture in-house. Amino acid side chains are excellent at acid–base catalysis (His, Asp, Glu, Lys, Cys, Tyr), at forming covalent intermediates with electrophilic carbons (Ser, Cys, Lys), and at binding metals and substrates with exquisite shape complementarity. What they cannot do is stabilize a genuine carbanion at a carbonyl carbon, accept a hydride ion cleanly, or form a delocalized electron sink that can absorb negative charge many bonds away from the catalytic residue. Coenzymes are recruited precisely to fill those gaps, and their chemical structures are shaped entirely around the one job. NAD⁺ and FAD both do redox chemistry but by different electron bookkeeping. NAD⁺'s nicotinamide ring is electron-poor at C4 because of the adjacent pyridinium-like nitrogen; a hydride (two electrons plus a proton, formally H⁻) delivered to C4 restores aromaticity of the ring in NADH and is thermodynamically very favorable. Because a hydride is a two-electron carrier, NAD⁺ chemistry is inherently a two-electron process — it only accepts H⁻, never a single electron in isolation, which is why NAD⁺-linked dehydrogenases always act on substrates that can give up a hydride (alcohols → ketones/aldehydes, aldehydes → acids). FAD, by contrast, has an isoalloxazine ring system that can accept electrons one or two at a time, passing through a resonance-stabilized flavin semiquinone radical. This makes FAD the coenzyme of choice whenever an enzyme must interconvert two-electron and one-electron chemistry, such as linking substrate oxidation to the one-electron carriers of the electron transport chain, or activating O₂ (whose ground state is a triplet, effectively a diradical). Coenzyme A and TPP tackle carbon skeleton chemistry from opposite directions. CoA's terminal –SH is a soft, polarizable nucleophile that attacks activated carbonyls (acyl-phosphates, acyl-adenylates) to form thioesters; the resulting C–S bond is a poor π-acceptor compared with C–O, so the thioester carbonyl is not resonance-stabilized the way an oxoester is, leaving a "high-energy" (large negative hydrolysis ΔG°′) bond that can drive both group transfer and (in a well-known Claisen-type condensation) C–C bond formation, as in the thiolase and citrate synthase reactions. TPP works from the electrophile side: it turns a proton on the thiazolium ring's C2 into a nucleophilic ylide carbon, which attacks a substrate carbonyl (pyruvate, α-ketoglutarate, or a ketose sugar) to form an adduct whose developing negative charge is delocalized into the ring's iminium nitrogen. That delocalization — a "carbanion sink" — is what makes possible the otherwise very difficult heterolytic cleavage of a C–C bond next to a carbonyl (decarboxylation of an α-keto acid). PLP is the most versatile of the five because it forms an actual covalent bond to substrate. At rest, PLP's aldehyde condenses with an active-site lysine ε-amino group to form an internal aldimine (a Schiff base). Substrate amino acid displaces the lysine via transaldimination, forming an external aldimine. Now any one of the three substituents around the substrate's α-carbon — the carboxylate (decarboxylation), the α-hydrogen (racemization, transamination), or the side chain (retro-aldol / β,γ-elimination reactions) — can be cleaved, because whichever bond breaks, the resulting carbanion delocalizes into the protonated pyridine nitrogen through the extended conjugated π system, exactly analogous to what the thiazolium ring does for TPP but with three possible bonds to activate instead of one.

The chemistry of each coenzyme reduces to one central bond-forming/bond-breaking event, whose thermodynamics can be written as a governing relation. For NAD⁺/NADH hydride transfer, the two-electron reduction potential sets the driving force:

where:

• $n = 2$ (a hydride carries two electrons),
• $F = 96{,}485\ \text{C/mol}$ (Faraday constant),
• $\Delta E^{\circ\prime} = E^{\circ\prime}_{\text{acceptor}} - E^{\circ\prime}_{\text{donor}}$, the difference in standard reduction potentials at pH 7.

Reading: a substrate couple with a more negative (i.e., more reducing) standard potential than the NAD⁺/NADH couple ($E^{\circ\prime} = -0.32\ \text{V}$) will spontaneously reduce NAD⁺, releasing free energy proportional to the potential gap. For coenzyme A thioester group-transfer chemistry, the driving force for acyl transfer is captured by the standard free energy of hydrolysis:

where this large negative value (compare −20 kJ/mol for a typical oxoester) reflects the poor resonance stabilization of the thioester carbonyl relative to its hydrolysis products. Reading: because thioester hydrolysis is so favorable, thioesters can thermodynamically drive otherwise unfavorable C–C bond-forming condensations (e.g., citrate synthase) or phosphorylation reactions (e.g., substrate-level phosphorylation in succinyl-CoA synthetase). For TPP-mediated decarboxylation, the key quantity is the acidity of the thiazolium C2–H, since this sets the rate of ylide (catalyst) formation:

Reading: although pKₐ ≈ 18 seems far too high for physiological deprotonation, the enzyme's active-site base and electrostatic environment shift the effective pKₐ down and stabilize the ylide, making catalytically relevant concentrations of the reactive carbanion accessible at pH 7. For PLP chemistry, the governing principle is not a single equilibrium constant but a qualitative electronic rule, often called the Dunathan stereoelectronic hypothesis:

where:

• $\theta$ is the angle between the bond being broken (C–H, C–COO⁻, or C–Cβ) and the plane of the conjugated PLP–imine π system,
• maximal orbital overlap (and hence maximal rate enhancement) occurs at $\theta = 0°$, minimal at $\theta = 90°$.

Reading: the active site orients the external aldimine so that exactly one of the three Cα substituents lies perpendicular to (in conjugation with) the ring π system, and that is the bond selectively cleaved — this is how one coenzyme catalyzes decarboxylation in one enzyme and transamination in another using the identical cofactor.

Example 1 — Foundational: computing ΔG°′ for lactate dehydrogenase (an NAD⁺-linked oxidation). Lactate dehydrogenase catalyzes:

Standard reduction potentials: pyruvate/lactate, $E^{\circ\prime} = -0.185\ \text{V}$; NAD⁺/NADH, $E^{\circ\prime} = -0.320\ \text{V}$. Here NADH is the electron donor (it is oxidized to NAD⁺) and pyruvate is the electron acceptor (reduced to lactate). So

The negative ΔG°′ confirms that reduction of pyruvate by NADH is strongly favorable under standard conditions — consistent with lactate dehydrogenase's physiological role in regenerating NAD⁺ for glycolysis during anaerobic (or fast-glycolytic) metabolism, where the reaction runs left-to-right using accumulated NADH. Note the asymmetry deliberately built into this example: the two half-reactions have different potentials (−0.185 V vs. −0.320 V, not equal or round numbers), and correctly tracking which species is oxidized vs. reduced is what determines the sign of ΔE°′ — swapping that assignment would flip the sign and give a qualitatively wrong conclusion.

Example 2 — Applied: TPP-dependent decarboxylation flux in pyruvate dehydrogenase, given a shifted ylide pKₐ. Suppose an active-site environment lowers the effective pKₐ of the thiazolium C2–H from 18 (free thiamine in water) to 14 (a substantial but experimentally reasonable active-site perturbation from desolvation and electrostatic stabilization of the ylide). At intracellular pH 7.4, estimate the fraction of TPP present as the catalytically active ylide form, and compare it with the fraction at the unperturbed pKₐ of 18. Using the Henderson–Hasselbalch relation for the deprotonation equilibrium (thiazolium-H ⇌ ylide + H⁺):

For pKₐ = 14:

So the ylide fraction is

For the unperturbed pKₐ = 18 case:

The active-site perturbation of only 4 pKₐ units increases the equilibrium ylide fraction by a factor of $10^{4} = 10{,}000$-fold. This illustrates why even modest electrostatic/desolvation stabilization of the ylide by the protein has an enormous kinetic payoff: the enzyme does not need to make the ylide the dominant species (it never is — it remains a vanishingly small equilibrium fraction even in the best-case estimate) but only needs to shift that vanishing fraction upward by orders of magnitude to achieve catalytically useful turnover, because the steady-state flux through the ylide intermediate scales with its equilibrium population when the subsequent carbonyl-attack step is fast.

Example 3 — Exam/competition level: using the Dunathan stereoelectronic rule to predict PLP reaction specificity. An aminotransferase active site holds the PLP–glutamate external aldimine so that the Cα–H bond is measured (by comparison to a fixed reference crystal structure) at a dihedral angle of $\theta = 15°$ from the π system of the conjugated imine, while the Cα–COO⁻ bond sits at $\theta = 78°$. A separate PLP-dependent decarboxylase active site (e.g., glutamate decarboxylase) instead holds the Cα–COO⁻ bond at $\theta' = 5°$ and the Cα–H bond at $\theta' = 85°$. Using the qualitative Dunathan relation that cleavage rate scales as $\cos^2\theta$ (maximal orbital overlap with the π system at θ = 0°), estimate the relative rate enhancement for cleaving the aligned bond versus the perpendicular bond in each active site, and state which bond is selectively broken in each enzyme. For the aminotransferase:

The Cα–H bond, aligned nearly parallel to the π system, is cleaved roughly 20-fold faster (in this simplified orbital-overlap model) than the Cα–COO⁻ bond — consistent with the aminotransferase's actual chemistry, which proceeds by abstraction of the Cα–H proton (forming a resonance-stabilized quinonoid intermediate) as the first step of transamination, leaving the carboxylate untouched. For the decarboxylase active site:

Here the Cα–COO⁻ bond is aligned for cleavage and is preferentially broken (decarboxylation), releasing CO₂ and leaving behind a resonance-stabilized carbanion/quinonoid that is reprotonated at Cα to give the amine product (e.g., GABA from glutamate). This worked example shows the deep mechanistic point of Dunathan's hypothesis: identical PLP chemistry, identical cofactor, but the protein dictates reaction specificity purely by rotating the substrate around the Cα–N bond to place a different substituent in the reactive, π-aligned orientation — the coenzyme supplies the electron sink, the protein supplies the geometry that selects which bond feeds into it.

Much of the conceptual framework connecting coenzyme structure to catalytic mechanism traces to Fritz Lipmann, who in the 1940s identified and characterized coenzyme A and articulated the concept of "high-energy" thioester and phosphate bonds as central currency of metabolism — work that earned him the 1953 Nobel Prize in Physiology or Medicine (shared with Hans Krebs). Lipmann's insight was not merely structural: he recognized that the biological significance of a cofactor lies in the free-energy relationships it enables, reframing biochemistry around energy-coupled group-transfer chemistry rather than a static list of named reactions. This same energetic logic — that a coenzyme is valuable because of what thermodynamic and electronic properties it contributes to a reaction, not merely because it participates in one — runs through the mechanistic treatment of NAD⁺, FAD, TPP, and PLP given here, all of which were structurally elucidated and mechanistically dissected over the same mid-twentieth-century golden age of vitamin biochemistry (Warburg's work on flavins and nicotinamide dinucleotides; Esmond Snell and Alexander Braunstein's parallel elucidation of PLP-dependent transamination).