Explainers

Patreon for winemaking creators: ethanol fermentation EMP pathway, malolactic fermentation Oenococcus oeni, red wine anthocyanin chemistry, proanthocyanidin tannin polymerization, sulfite Henderson-Hasselbalch equilibrium, potassium bitartrate cold stabilization, oak lactone and ellagitannin chemistry, and the Apple Tax

2026-07-10 · ~3,900 words

The Patreon for winemaking creators guide covers creator subtypes and tier structures — what home winemakers, commercial vintners, and wine educators offer at each price point and which Patreon features suit a seasonal, harvest-driven content calendar. This post covers what that guide does not: the molecular, biochemical, and physical chemistry layer that winemaking creators who teach at an advanced level need to understand to explain what is actually happening in the fermentation vessel, the barrel, and the bottle. Why does yeast need thiamine pyrophosphate to decarboxylate pyruvate? What precise chemical cascade allows Oenococcus oeni to convert harsh malic acid to softer lactic acid, and why does diacetyl form from a non-enzymatic reaction outside the cell? Why does a wine’s color shift from red to purple depending on pH — and what are pyranoanthocyanins? How do tannins produce the sensation of astringency at the molecular interface with salivary protein? Why does a wine’s pH change the required sulfite dose by a factor of three? What is cis-β-methyl-γ-octalactone and why does American oak smell different from French oak? None of this appears in beginner Patreon tiers, but it is precisely the scientific depth that retains intermediate-to-advanced patrons who have been making wine for years and want to understand the chemistry underlying every sensory experience in their cellar.

Ethanol fermentation: the EMP pathway, pyruvate decarboxylase ThPP cofactor, and ADH1 NADH recycling

The conversion of grape sugar to ethanol proceeds through the Embden-Meyerhof-Parnas (EMP) glycolytic pathway, a ten-enzyme sequence conserved across nearly all living organisms. Glucose (and fructose, after phosphorylation) enters glycolysis and is processed through two phases: an investment phase (consuming 2 ATP to phosphorylate glucose-6-phosphate and fructose-1,6-bisphosphate) and a payoff phase (producing 4 ATP and 2 NADH per glucose from substrate-level phosphorylation and oxidative reactions). The net yield is 2 ATP, 2 NADH, and 2 pyruvate per glucose. In the aerobic setting, NADH would be oxidized back to NAD&sup+; by the mitochondrial electron transport chain with concurrent ATP synthesis. In the anaerobic environment of active wine fermentation, however, the electron transport chain is suppressed (Saccharomyces cerevisiae undergoes the Crabtree effect — glucose repression of respiration even in the presence of oxygen at high sugar concentrations), and NAD&sup+ regeneration must occur through fermentative routes.

The two-step fermentative pathway that converts pyruvate to ethanol begins with pyruvate decarboxylase (PDC, EC 4.1.1.1). PDC catalyzes the non-oxidative decarboxylation of pyruvate to acetaldehyde and CO₂, and it requires thiamine pyrophosphate (ThPP, also called thiamine diphosphate TPP) as an obligate cofactor. The mechanism is mechanistically elegant: the C-2 carbon of the thiazolium ring in ThPP, which is flanked by the positively charged sulfur and nitrogen atoms, is unusually acidic (apparent pKa approximately 18, lower than typical C-H bonds) because the ring heteroatoms stabilize the resulting carbanion. The C-2 carbanion attacks the electrophilic carbonyl carbon of pyruvate, forming a covalent lactyl-ThPP adduct. The enzymatic environment then facilitates decarboxylation of the C-1 carboxylate group of the lactyl adduct, releasing CO₂ and forming a hydroxyethyl-ThPP intermediate; the carbanion at the former C-2 of pyruvate is again stabilized by the thiazolium system (ylide resonance). The hydroxyethyl-ThPP intermediate then collapses to release acetaldehyde and regenerate free ThPP. Acetaldehyde (ethanal, CH₃CHO, MW 44.1 Da) is a reactive, volatile compound with a green/fruity aroma at low concentrations and a pungent, irritating smell at high concentrations (>100 mg/L in wine is considered a fault). During healthy fermentation, acetaldehyde accumulates transiently in early fermentation and is progressively reduced to ethanol by alcohol dehydrogenase.

Alcohol dehydrogenase 1 (ADH1, EC 1.1.1.1, the primary fermentative isozyme in S. cerevisiae; ADH2 is the aerobic catabolic isozyme expressed post-fermentation) catalyzes the reduction of acetaldehyde to ethanol using NADH as the electron donor: acetaldehyde + NADH + H&sup+ → ethanol + NAD&sup+;. ADH1 is a zinc metalloenzyme; the zinc ion in the active site (catalytic zinc, distinct from the structural zinc in the protein scaffold) coordinates the carbonyl oxygen of acetaldehyde, polarizing the C=O bond and positioning the acetaldehyde carbonyl carbon for direct hydride transfer from the C-4 pro-R hydrogen of NADH. The reaction is reversible but is driven strongly toward ethanol production under fermentation conditions by the high acetaldehyde concentration produced by PDC and the high NADH/NAD&sup+ ratio of anaerobic metabolism. The regeneration of NAD&sup+ is the central thermodynamic reason fermentation exists: without it, glycolytic flux would halt at glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which requires NAD&sup+ as an oxidant. Each mole of glucose processed through glycolysis and fermentation thus yields 2 mol ethanol + 2 mol CO₂ + 2 mol ATP (net) — the Gay-Lussac equation, with a theoretical ethanol mass yield of 0.511 g ethanol per gram of glucose.

Glycerol is the most significant fermentation byproduct after ethanol and CO₂, typically present at 5–10 g/L in dry wine. Glycerol is produced from dihydroxyacetone phosphate (DHAP, a glycolytic intermediate) by the sequential action of glycerol-3-phosphate dehydrogenase (GPD1 during osmotic stress/early fermentation; GPD2 under anaerobic conditions) and glycerol-3-phosphatase. The NADH consumed in glycerol biosynthesis serves as an alternative route for NAD&sup+ regeneration when acetaldehyde is not yet available (pre-fermentation, when SO₂ additions bind acetaldehyde and prevent its use as a NADH acceptor) or when redox imbalances develop. At concentrations above 4–5 g/L, glycerol contributes viscosity and a subtle sweetness perception that contributes to mouthfeel; this is why high-Brix fermentations producing more glycerol (and wines from botrytized fruit, where glycerol can reach 20–25 g/L) are often described as rounder and fuller-bodied.

The Ehrlich pathway produces fusel alcohols when yeast encounters YAN deficiency and must catabolize amino acids for nitrogen. The pathway consists of three enzymatic steps: (1) transamination of the amino acid to its corresponding α-keto acid (catalyzed by branched-chain aminotransferase, BAT1/BAT2, or aromatic aminotransferase ARO8/ARO9, with α-ketoglutarate as the nitrogen acceptor); (2) decarboxylation of the α-keto acid to an aldehyde by PDC or a PDC-homologous enzyme; (3) reduction of the aldehyde to the fusel alcohol by alcohol dehydrogenase. Leucine yields isoamyl alcohol (3-methyl-1-butanol, banana/fusel aroma, sensory threshold approximately 50–65 mg/L in wine) via α-ketoisocaproate; valine yields isobutanol (2-methyl-1-propanol, harsh solvent, threshold approximately 200 mg/L) via α-ketoisovalerate; phenylalanine yields 2-phenylethanol (rose/honey/floral aroma, threshold approximately 10–75 mg/L, a desirable aromatic at low concentrations) via phenylpyruvate; threonine yields n-propanol via α-ketobutyrate. Adequate YAN (target 150–350 mg N/L depending on initial Brix) suppresses the Ehrlich pathway because yeast preferentially assimilates inorganic ammonium (from DAP) and α-amino nitrogen (from amino acids), reducing the pressure to catabolize amino acids for nitrogen extraction alone.

Malolactic fermentation: Oenococcus oeni, MLE enzyme, deacidification, and diacetyl chemistry

Malolactic fermentation (MLF) is a secondary biological conversion performed by lactic acid bacteria — predominantly Oenococcus oeni, the one species of the genus specifically adapted to wine’s extreme chemical environment. O. oeni is a heterofermentative (non-motile, gram-positive) coccus capable of growth at pH as low as 3.0, ethanol concentrations up to 13–15%, temperatures as low as 10°C, and in the presence of up to 30 mg/L total SO₂. The genome of O. oeni strain PSU-1 (the first fully sequenced wine strain) encodes a remarkably streamlined metabolic network: at only 1.84 Mb, it has one of the smallest gram-positive bacterial genomes, reflecting extensive gene loss during adaptation to the nutrient-limited, acidic wine environment. The organism grows on malic acid and hexoses (glucose, fructose) as primary carbon sources; in wine, it is adapted to subsisting on residual sugars, citric acid, and the malic acid it converts.

The central reaction of MLF is catalyzed by the malolactic enzyme (MLE, also called malate decarboxylase or NAD&sup+-dependent malolactic enzyme; EC 1.1.1.38). MLE is a single enzyme that converts L-malate directly to L-lactate plus CO₂ in what appears to be one step but actually involves two half-reactions: (1) oxidation of L-malate to oxaloacetate with NAD&sup+ as hydride acceptor; (2) decarboxylation of oxaloacetate to pyruvate and CO₂ (an oxaloacetate decarboxylase reaction); (3) reduction of pyruvate to L-lactate with the NADH produced in step (1). Steps (1)–(3) all occur within the single active site of MLE without release of intermediates, making MLE a mechanistic triad enzyme. The enzyme requires Mn²&sup+; as a divalent metal cofactor (not Mg²&sup+ as in many similar enzymes); Mn²&sup+ coordinates the C-2 hydroxyl and C-1 carboxylate of malate in the active site, orienting the substrate for the concerted dehydrogenation-decarboxylation. This Mn²&sup+ requirement has practical consequences: wine media very low in manganese can limit MLF velocity, and some winemaking additives inadvertently chelate Mn²&sup+;, slowing MLF completion.

The net chemical reaction of MLF is: L-malic acid + NAD&sup+ → L-lactic acid + CO₂ + NADH, but the intracellular NADH is immediately recycled by O. oeni’s central metabolism, making the overall process irreversible under wine conditions. The driving force for L-malate uptake into the cell and lactate export is provided by the malate/lactate antiporter in the cell membrane — a secondary active transporter that exchanges one malate²− (entering) for one lactate⁻ (exiting). The net charge movement (malate brings in 2 negative charges; lactate removes 1 negative charge; the decarboxylation in the cytoplasm consumes one proton from the intracellular environment, equivalent to one net proton exported across the membrane per malate consumed) generates a proton-motive force that drives ATP synthesis by the bacterial F₁F₀-ATPase in reverse, yielding approximately 0.5 ATP per malate converted. This makes MLF an energy-generating process for O. oeni — one of the reasons why it is energetically favored by the bacterium even in wine where energy sources are scarce.

The deacidification effect of MLF: L-malic acid has pKa1 3.46 and pKa2 5.10 (diprotic, two carboxyl groups). L-lactic acid has a single pKa 3.86 (monoprotic). At wine pH 3.2–3.6, the conversion of one dicarboxylic acid (malic) to one monocarboxylic acid (lactic) plus CO₂ reduces the total buffering capacity and titratable acidity of the wine. The pH rise from MLF depends on the initial malic acid concentration (high in cool-climate wines, Cabernet Sauvignon, Sauvignon Blanc — up to 4–6 g/L; lower in warm-climate wines) and on the wine’s buffering capacity (dominated by tartrate/bitartrate, sulfate, and residual amino acids). Typical MLF-induced pH increase is 0.1–0.5 units, reducing titratable acidity by 0.2–3.0 g/L tartaric acid equivalents.

Diacetyl formation and control is the most consequential flavor consequence of MLF after deacidification. Diacetyl (2,3-butanedione; CH₃CO-COCH₃; MW 86.1 Da; buttery/cream/rancid butter aroma) does not form via a bacterial enzymatic reaction in the MLE pathway. Instead, it arises from α-acetolactate (2-hydroxy-2-methyl-3-oxobutanoate), an intermediate in the acetoin biosynthesis branch of O. oeni’s metabolism. α-Acetolactate is produced inside bacterial cells from two acetaldehyde units via α-acetolactate synthase (IlvB); it is intended to be converted further to acetoin (3-hydroxy-2-butanone) by α-acetolactate decarboxylase (IlvD) inside the cell. However, some α-acetolactate leaks from the cells into the wine medium. Once outside the cell, α-acetolactate undergoes spontaneous non-enzymatic oxidative decarboxylation in the presence of dissolved oxygen — a chemical (not biochemical) reaction in which molecular oxygen accepts an electron, generating a peroxide radical intermediate that collapses to release diacetyl. The reaction rate depends on [O₂], temperature, and pH. The sensory threshold for diacetyl in white wines is approximately 8 µg/L; in red wines approximately 15 µg/L. Diacetyl reduction after MLF completion: active O. oeni cells reduce diacetyl to acetoin (sensory threshold >1 mg/L, not perceptible at wine concentrations) via diacetyl reductase. Extended lees contact under active bacteria, maintaining MLF completion before any SO₂ addition (which kills O. oeni and halts diacetyl reduction), and avoiding excessive oxygen exposure during MLF are the practical tools for minimizing residual diacetyl.

Red wine color chemistry: anthocyanin pH equilibria, copigmentation, and pyranoanthocyanin formation

Red wine color emerges from anthocyanins, flavonoid pigments extracted from grape skins during maceration and fermentation. The dominant anthocyanin in Vitis vinifera red wines is malvidin-3-glucoside (oenin), the 3-glucoside of malvidin (3,5,7,4’-tetrahydroxy-3’,5’-dimethoxyflavylium cation). Malvidin carries methoxy substituents at the 3’ and 5’ positions of the B ring, making it the most methylated (and most oxidation-resistant) of the five major V. vinifera anthocyanidins (cyanidin, delphinidin, petunidin, peonidin, malvidin); oenin typically accounts for 50–85% of total anthocyanins in varieties such as Cabernet Sauvignon, Merlot, Pinot Noir, and Tempranillo, with the precise proportion depending on variety, climate, and ripeness. Total anthocyanin in red wine at harvest ranges from 50–300 mg/L depending on grape variety, vintage conditions, maceration time, and temperature.

Anthocyanins exist in a pH-dependent equilibrium among four molecular forms: the flavylium cation (AH&sup+;, the red/purple form with a full aromatic pyrylium ring carrying a delocalized positive charge; λmax approximately 510–525 nm; molar absorptivity approximately 28,000 L mol−¹ cm−¹ for oenin; dominant below pH 2.5), the carbinol pseudobase (A-OH, the colorless hemiketal form produced by water addition at C-2 of the flavylium ring, which disrupts the aromatic conjugation responsible for visible-light absorption; dominant at wine pH 3.0–4.0), the quinoidal base (A, the blue-purple form produced by deprotonation of the 4’-OH of the flavylium cation; λmax shifts bathochromically to 570–590 nm; dominant above pH 5 in solution), and the open-chain chalcone isomer (B, pale yellow; formed by slow tautomerization from the carbinol pseudobase via ring opening; minor contributor in wine but represents a slowly interconverting reservoir). At typical wine pH 3.2–3.6, the equilibrium lies predominantly toward the colorless carbinol pseudobase (approximately 80–95% carbinol) with only 5–20% in the red flavylium form. Despite this, wines appear deeply red because the molar absorptivity of the flavylium cation is extremely high and because total anthocyanin concentrations in young red wine are substantial.

Copigmentation is the non-covalent association between anthocyanin molecules and copigment molecules that dramatically increases the effective color intensity. Copigments are colorless or pale phenolic compounds present in wine that stack face-to-face with the planar aromatic system of the flavylium cation through π–π intermolecular interactions. Effective copigments in wine include hydroxycinnamic acids (caffeic acid, p-coumaric acid, ferulic acid and their tartrate esters: caftaric, coutaric, fertaric acids), hydroxycinnamic acid ethyl esters, flavonols (quercetin-3-glucuronide, quercetin-3-glucoside, kaempferol-3-glucoside), and other flavonoids. The copigmentation complex has two measurable effects: the hyperchromic effect (the molar absorptivity at λmax of the flavylium cation increases by 30–200% within the complex, producing deeper color than expected from the free anthocyanin concentration alone) and the bathochromic shift (the λmax of the complex shifts 5–20 nm to longer wavelengths, deepening the red toward purple-blue). Copigmentation also stabilizes the flavylium cation against water addition — the copigment physically blocks water access to C-2 of the flavylium ring, shifting the equilibrium toward the colored form. This explains why freshly pressed red wine with intact copigments is more intensely colored than the same wine after phenolic polymerization and polyphenol oxidase activity in aged or heat-exposed wine.

Pyranoanthocyanins are a distinct class of anthocyanin derivatives that accumulate progressively during wine aging. They form when anthocyanins react with small molecules (pyruvic acid, acetaldehyde, vinyl phenols, vinyl catechol) through a [4+2] cycloaddition or similar cyclization at C-4 and C-5 of the anthocyanin, creating a new pyrano ring that is fused across the C-4/C-5 bond of the flavylium nucleus. The C-4 position, now incorporated into the pyrano ring, is blocked from water addition — the key reaction that produces the colorless carbinol pseudobase in unmodified anthocyanins. As a result, pyranoanthocyanins are substantially less pH-sensitive than their anthocyanin precursors: they retain color at pH values where normal anthocyanins are predominantly colorless. Vitisin A (malvidin-3-glucoside-pyruvate adduct, formed by condensation with pyruvic acid) and vitisin B (malvidin-3-glucoside-acetaldehyde adduct) are the most commonly detected pyranoanthocyanins in wine; acetaldehyde-mediated adducts bridge between one anthocyanin and one flavanol unit to produce polymeric pigments with further modified color characteristics. The accumulation of pyranoanthocyanins over bottle aging contributes to the characteristic color evolution of aged red wines — from the vivid purple-red of youth (high anthocyanin, high copigmentation) to the more stable, less pH-sensitive brick-red and garnet tones of maturity (pyranoanthocyanins and polymeric pigments dominant).

Proanthocyanidin tannin polymerization and the molecular mechanism of astringency

Wine tannins are predominantly condensed tannins (proanthocyanidins) — oligomers and polymers of flavan-3-ol monomers joined by carbon-carbon interflavanic bonds. The two principal monomers in Vitis vinifera are (+)-catechin (the 2R,3S stereoisomer; MW 290.3 Da; colorless; abundant in seeds and woody tissue; the predominant terminal unit in proanthocyanidin chains) and (−)-epicatechin (the 2R,3R stereoisomer; MW 290.3 Da; the predominant extension unit in proanthocyanidin chains from seeds and skins). Epigallocatechin (with an additional 5’-OH on the B ring, producing a pyrogallol B ring instead of a catechol B ring) is the monomeric precursor for prodelphinidins (skin tannins) and is also present as a minor monomer in seed tannins.

The interflavanic bond connecting flavan-3-ol units is a carbon-carbon bond between C-4 of the upper (terminal or upper extension) unit and C-8 (predominantly) or C-6 of the lower (extension) unit. The C4→C8 bond is the most common in wine grape proanthocyanidins (B-type linkage); C4→C6 linkages also occur at lower frequency. These C–C bonds are highly resistant to hydrolysis — unlike the ester bonds in hydrolyzable tannins (gallotannins and ellagitannins from oak) that are cleaved by acid or alkali. This resistance to hydrolysis means that condensed tannins survive intact in wine for years and decades, undergoing further polymerization but not breaking down. The degree of polymerization (DP) in wine tannins ranges from DP-2 (dimer, MW approximately 578 Da) to DP-12 and beyond (dodecamer, MW approximately 3,500 Da or higher), with the average DP in young red wine typically DP-4 to DP-8 depending on extraction method and variety.

Seed tannins and skin tannins differ in subunit composition, galloylation, and molecular weight distribution. Seed procyanidins consist entirely of catechin and epicatechin units, many with gallic acid esterified at the C-3 position (3-O-galloyl-epicatechin extension units; epicatechin-3-O-gallate). Galloylation increases the number of hydrogen-bond donor groups per tannin molecule and increases tannin hydrophobicity, both of which strengthen tannin-protein interactions and increase perceived astringency and bitterness relative to non-galloylated tannins of the same DP. Seed tannins are released into wine primarily through mechanical disruption during crushing; more aggressive pressing extracts higher proportions of seed tannin. Skin tannins are a mixture of procyanidins and prodelphinidins (containing epigallocatechin and epigallocatechin gallate units; the catechol B ring of catechin/epicatechin replaced by a pyrogallol B ring with three hydroxyls at 3’, 4’, 5’); the pyrogallol B ring provides an additional hydrogen-bond donor and affects the geometry of tannin-protein interaction. Skin tannins are extracted during maceration; longer maceration (extended skin contact, cap management frequency, pump-over vs. punch-down technique) increases skin tannin extraction.

Astringency at the molecular level is the result of tannin-salivary proline-rich protein (PRP) interaction. Saliva contains a high concentration (1–3 mg/mL) of PRPs — a family of natively disordered proteins (IB-5, PIF-f, statherin, and related isoforms) that are exceptionally rich in proline (35–40% of all residues). The function of PRPs is debated but includes binding dietary tannins to prevent them from reaching and complexing with mucosal proteins. The tannin-PRP interaction has two major driving forces. First, hydrophobic stacking: the pyrrolidine ring of proline presents a flat hydrophobic face that is ideally shaped for π-stacking with the aromatic rings of catechol/pyrogallol groups on the flavan-3-ol B and A rings. Second, hydrogen bonding: tannin phenolic OH groups donate hydrogen bonds to backbone carbonyl groups and side chain amide and amine groups of the protein. The multi-point, bidentate nature of tannin binding (a single procyanidin dimer can engage in 4–8 simultaneous H-bonds and one or two π-stacking interactions) makes the complex thermodynamically favorable even though each individual interaction is weak. At sufficient tannin:protein stoichiometry (above the critical tannin:protein ratio that produces insoluble aggregates, approximately 1–3 mg tannin per mg PRP), the multi-point tannin-PRP complexes begin to cross-link and aggregate, forming a precipitate or gel network in the oral mucosa that physically and tactilely produces the puckering, gripping, drying sensation perceived as astringency.

During barrel aging and bottle aging, wine tannins undergo oxidative polymerization: copper and iron ions (trace metals from the must, equipment, and fining additions) catalyze the one-electron oxidation of catechol groups on tannin B rings to semiquinone radicals, which then couple with the C-8 position of adjacent catechin/epicatechin rings to form new C–C bonds, increasing the average DP of the tannin population. Simultaneously, tannin-anthocyanin adducts form (via acetaldehyde bridging or direct oxidative coupling), producing polymeric pigments that contribute both color and tannin-like tactile properties. The resulting aged wine tannins are less reactive with PRPs per unit mass (higher MW tannins have lower solubility and slower diffusion to salivary protein binding sites) but may produce a different quality of astringency — often described as softer, smoother, or more integrated — compared to the angular, grippy astringency of young, seed-tannin-rich wines with smaller-MW tannin oligomers.

Sulfite chemistry: Henderson-Hasselbalch equilibrium, molecular SO₂, and bound forms

Sulfur dioxide (SO₂) is the universal wine preservative, functioning both as an antioxidant and as an antimicrobial agent. When potassium metabisulfite (K₂S₂O₅) is dissolved in wine (or sulfur dioxide gas is dissolved directly), it dissociates and hydrates to produce a mixture of three species in equilibrium: molecular SO₂ (also written H₂SO₃, the un-ionized form), bisulfite ion (HSO₃−), and sulfite ion (SO₃²−). The two ionization equilibria are: SO₂ + H₂O ⇌ HSO₃− + H&sup+ (pKa1 = 1.81 at 20°C; note that some sources use the apparent pKa in the range 1.5–1.9 depending on ionic strength); HSO₃− ⇌ SO₃²− + H&sup+ (pKa2 = 7.2 at 20°C).

Applying the Henderson-Hasselbalch equation at wine pH 3.2–3.6: the fraction of free SO₂ present as molecular SO₂ = 1/(1 + 10(pH − pKa1)) = 1/(1 + 10(pH − 1.81)). At pH 3.2: molecular fraction = 1/(1 + 101.39) = 1/(1 + 24.5) ≈ 3.9%. At pH 3.4: molecular fraction = 1/(1 + 101.59) = 1/(1 + 38.9) ≈ 2.5%. At pH 3.6: molecular fraction = 1/(1 + 101.79) = 1/(1 + 61.7) ≈ 1.6%. This means that achieving 0.8 mg/L molecular SO₂ (a common bottling target for wines with potential microbial risk) requires approximately 20–24 mg/L free SO₂ at pH 3.2, approximately 32 mg/L at pH 3.4, and approximately 50–60 mg/L at pH 3.6 — a 2.5-fold difference between pH 3.2 and 3.6 for the same antimicrobial protection. This is why high-pH red wines require dramatically higher SO₂ additions than crisp, low-pH white wines and is the primary quantitative tool in winemaking sulfite management.

Only molecular SO₂ (not bisulfite or sulfite) has direct antimicrobial activity, because the un-ionized H₂SO₃ can diffuse freely across the lipid bilayer of microbial cell membranes. The bisulfite and sulfite ions, carrying negative charges, cannot pass through the lipophilic membrane interior. Once molecular SO₂ enters the cell and encounters the higher intracellular pH (approximately 6.5–7.0 in yeast and bacteria), it dissociates to bisulfite and sulfite, which then react with intracellular targets: bisulfite reacts with NAD&sup+ (forming a covalent NADH₂-bisulfite adduct that inactivates NAD&sup+-requiring dehydrogenases); sulfite reacts with disulfide bonds in proteins (cleaving S–S bridges); both react electrophilically with electron-rich sites in DNA bases. The antioxidant function of SO₂ involves direct reduction of hydrogen peroxide: HSO₃− + H₂O₂ → SO₄²− + 2H&sup+ + H₂O, consuming the H₂O₂ produced during the enzymatic and non-enzymatic oxidation of wine polyphenols and preventing its participation in Fenton chemistry (Fe²&sup+; + H₂O₂ → Fe³&sup+; + ⋅OH) that generates hydroxyl radicals for irreversible oxidative browning.

Bound SO₂ is the portion of total added SO₂ that has reacted with carbonyl-containing wine components to form covalent or strong electrostatic adducts. The most important binding partner is acetaldehyde: the acetaldehyde-bisulfite adduct (1-hydroxyethanesulfonate, the sodium or potassium salt) has a binding constant K ≈ 1.5×10⁶ M−¹ at wine temperature and pH. This binding constant is so large that essentially all bisulfite in the presence of equimolar acetaldehyde will bind; the resulting adduct is analytically “fixed” and does not release meaningful bisulfite under wine conditions. A wine with 50 mg/L bound acetaldehyde (not uncommon in oxidatively handled wines, wines fermented at high temperature, or wines with extended yeast contact) would bind approximately 170 mg/L SO₂ in the acetaldehyde adduct form, rendering that SO₂ inert for both antimicrobial and antioxidant protection. Other significant SO₂-binding compounds include pyruvic acid (K ≈ 1.3×10⁵ M−¹), α-ketoglutaric acid (K ≈ 5.3×10⁴ M−¹), and galacturonic acid (K ≈ 200 M−¹, minor). During MLF, Oenococcus oeni’s metabolism produces additional pyruvate and α-ketoglutarate, so post-MLF wines often show increased SO₂ binding demand — a practical consideration when calculating post-MLF SO₂ additions.

Potassium bitartrate cold stabilization: KHT solubility, nucleation, and CMC inhibition

Tartaric acid is the dominant organic acid in grape must and wine, present at 2.5–7 g/L depending on variety, climate, and winemaking practice. Tartaric acid (2,3-dihydroxysuccinic acid; MW 150.1 Da; pKa1 2.99, pKa2 4.34) is unusual among common organic acids in that it is not metabolized by Saccharomyces cerevisiae during alcoholic fermentation (unlike malic and citric acids, which are partially consumed) and is essentially stable to biological degradation under typical wine conditions. At wine pH 3.2–3.6, tartaric acid exists predominantly as the bitartrate anion (HT−, singly deprotonated at pKa1), with smaller proportions of un-ionized H₂T and tartrate dianion T₂−.

Potassium ions (released from grape pulp cells, skins, and seeds during fermentation at concentrations of typically 1–2 g/L in red wines, 0.8–1.5 g/L in whites) combine with bitartrate anion to form potassium hydrogen tartrate (KHT, cream of tartar; MW 188.2 Da). KHT has a solubility product Ksp of approximately 4.9×10−³ mol₂·L−² at 20°C (varies with alcohol content and ionic strength; higher ethanol decreases KHT solubility significantly). In newly fermented wine, K&sup+ and HT− concentrations together often exceed the Ksp for KHT at cellar temperatures (5–15°C) and certainly at cold temperatures — making the wine supersaturated with respect to KHT. The practical consequence is KHT crystallization in the bottle if the wine is cold-stabilized insufficiently, producing harmless but aesthetically undesirable tartrate crystals (sometimes called “wine diamonds”). Cold stabilization at −4°C (just above the wine’s freezing point, approximately −5 to −6°C for 12% ABV wines) for 1–2 weeks allows KHT to crystallize and precipitate in the tank or barrel; the clarified wine is then racked off the crystalline deposit.

The kinetics of KHT crystallization depend on two phases: nucleation (the formation of the first crystal nuclei from the supersaturated solution, requiring a free energy barrier to be overcome because the small surface-area-to-volume ratio of nascent nuclei is energetically unfavorable) and crystal growth (accretion of KHT from solution onto existing crystal faces, proceeding faster than nucleation and producing the visible macroscopic crystals). Cold stabilization treatment accelerates both phases by increasing KHT supersaturation at low temperature (driving the thermodynamic force for crystallization) and by the long holding time at temperature. Seeding with KHT crystals (contact cold stabilization, KHT seed crystals added to accelerate nucleation) or ultrasonic stimulation can dramatically shorten the stabilization time.

Carboxymethyl cellulose (CMC; sodium salt of cellulose esterified with carboxymethyl groups; DS typically 0.6–0.9; MW variable) is an approved KHT crystal growth inhibitor in the EU (Regulation EC 606/2009) and OIV member countries. CMC is a water-soluble anionic polymer that adsorbs selectively to the KHT crystal growth faces through electrostatic interaction between CMC carboxylate groups and the K&sup+ sites on the crystal face. By blocking the crystal growth sites, CMC prevents further KHT accretion even when the wine remains supersaturated at refrigerator temperature. The practical result is that CMC-treated wines can be bottled without cold stabilization at −4°C, remaining clear and crystal-free at the consumer’s refrigerator temperature. CMC is added at 100–200 mg/L before bottling; it is not metabolized or removed by typical fining and does not affect wine flavor or aroma at recommended doses, though it can interfere with bentonite fining if added before bentonite treatment is complete (CMC adsorbs to bentonite clay, reducing its protein-fining capacity).

Oak chemistry: β-methyl-γ-octalactone, vanillin from lignin, ellagitannins, and furfural

Oak barrels (and oak alternatives — staves, cubes, chips, spirals) contribute a chemically complex suite of volatile and non-volatile compounds to wine during aging. The major classes include lactones (primary species-specific aromatic compounds), phenolic aldehydes and phenols (from lignin thermal degradation), hydrolyzable tannins (ellagitannins from wood cell walls), and furan aldehydes (from hemicellulose dehydration during toasting). Understanding each class at the molecular level enables winemakers and educators to predict how barrel choice, toast level, and aging time interact to produce specific flavor and structural outcomes.

β-Methyl-γ-octalactone (also called the whisky lactone or oak lactone; IUPAC: 3-methyl-γ-octalactone; a bicyclic compound with a γ-butyrolactone ring fused to a cyclohexane ring through the 3,4 bond; MW 156.2 Da; two diastereomers: cis and trans) is the primary species-specific aromatic compound distinguishing oak from other wood species used in barrel making. It is present at substantially higher concentrations in Quercus alba (American white oak) than in Q. petraea (sessile oak, the primary French cooperage oak) or Q. robur (pedunculate oak, also French). The cis isomer of β-methyl-γ-octalactone has a sensory threshold of approximately 50 µg/L in wine (perceived as coconut, woody, vanilla-adjacent) and the trans isomer has a threshold of approximately 170 µg/L (woody, spicy). American oak cooperage can contain 300–1,000 µg/L total oak lactone (predominantly cis) in wine aged 12–24 months, far above threshold; French oak cooperage typically contributes 20–100 µg/L, near or below threshold. This fundamental chemical difference explains the pronounced coconut/vanilla character of wines aged in American oak (e.g., Spanish Rioja Reserva, many California Zinfandels) versus the subtler toasty/spicy character of wines aged in French oak. Toast level further modifies oak lactone: high-toast barrels produce more cis-β-methyl-γ-octalactone from pyrolytic reactions with precursor fatty acid and terpenoid compounds in the wood.

Vanillin (4-hydroxy-3-methoxybenzaldehyde; MW 152.1 Da; vanilla/sweet aroma; sensory threshold approximately 320 µg/L in wine) is released from oak lignin during toasting through the thermal degradation of the guaiacyl lignin polymer. Hardwood lignin (H/G/S type) contains three types of phenylpropanoid units: p-hydroxyphenyl (H), guaiacyl (G, with a 4-OH and 3-OCH₃ on the aromatic ring), and syringyl (S, with 4-OH and both 3-OCH₃ and 5-OCH₃). The dominant linkage in lignin is the β-O-4 aryl ether bond, connecting the β-carbon of one phenylpropanoid unit to the 4-oxygen of the adjacent unit. At toasting temperatures (180–220°C), β-O-4 bonds are thermally cleaved via homolytic or heterolytic mechanisms, releasing guaiacol (from the aryl ether oxygen terminus) and generating aldehyde end-groups on the phenylpropanoid unit — the guaiacyl propan-3-al (guaiacylaldehyde) β-oxidizes further to vanillin under oxidative conditions. The yield of vanillin from lignin toasting increases with toast level (higher temperature and/or longer duration) but decreases at very high toast levels (vanillin itself is thermally degraded at >200°C to guaiacol and other phenols through decarbonylation).

Ellagitannins from oak are hydrolyzable tannins — polyol esters of gallic acid and hexahydroxydiphenic acid (HHDP) that release ellagic acid upon acid hydrolysis. The principal ellagitannins in oak wood are castalagin and vescalagin (the C-1 epimers of the same pentagalloylglucose-derived heptamer; both with MW approximately 933 Da), and the grandinin and roburin series. These are the “oak tannins” that are distinct in structure, mechanism, and sensory effects from the condensed procyanidin tannins of the grapes themselves. Ellagitannins are hydrolyzable (their ester bonds are cleaved by acid and water), producing glucose, gallic acid, and the ellagic acid dimers in the wine over time. They contribute a distinct astringency quality (often described as grippy but not harsh, as opposed to the linear green grip of seed procyanidins), and they provide antioxidant protection through their phenolic hydroxyl groups (contributing to the micro-oxygenation buffering effect of barrel aging — ellagitannins react preferentially with dissolved oxygen from barrel breathing before the wine polyphenols are oxidized). Ellagitannins are present in new French oak at approximately 6–10 mg/g dry wood; American oak contains fewer (approximately 1–3 mg/g). Wines aged in new oak can acquire 100–400 mg/L ellagitannin over 12 months of barrel contact.

Furfural (furan-2-carbaldehyde; MW 96.1 Da; almond, bread, caramel aroma; sensory threshold approximately 1,000 µg/L in wine) and 5-methylfurfural (MW 110.1 Da) are produced during oak toasting from the dehydration of hemicellulose pentose sugars. Hemicellulose (primarily arabinoxylans and glucuronoxylans in oak wood) contains arabinose and xylose residues that undergo acid-catalyzed triple dehydration at toasting temperatures (180–220°C) through a 3-deoxypentosulose intermediate to produce furfural. The amount of furfural transferred to wine increases with toast level and contact time. At low concentrations in wine (<500 µg/L), furfural contributes subtle caramel and toasty notes; at concentrations above threshold (achieved in heavily charred American oak or in wines with extended barrel contact under high-toast conditions), furfural contributes overt almond and bread crust notes that some palates find positive and others find oakey or over-extracted.

Volatile acidity, Acetobacter oxidation, and the PQQ-ADH pathway

Volatile acidity (VA) refers to the acidic wine components that can be separated from the fixed (non-volatile) acids by steam distillation, primarily acetic acid (pKa 4.76) with small contributions from propionic acid, butyric acid, and trace amounts of other short-chain fatty acids. Acetic acid is the primary volatile acid in wine (comprising >95% of VA) and has a vinegar aroma above its sensory threshold of approximately 0.6–0.9 g/L in most wines. The associated ester, ethyl acetate (CH₃COOC₂H₅; MW 88.1 Da; nail varnish/solvent descriptor; sensory threshold 160–200 mg/L in wine), is produced by the chemical esterification of acetic acid with ethanol under wine conditions and contributes the characteristic solvent character of high-VA wines before the acetic acid itself is perceptible. EU limits for volatile acidity (expressed as acetic acid equivalents) are 1.08 g/L for dry white wines and 1.20 g/L for dry red wines (higher legal limits exist for high-alcohol, late-harvest, and botrytized wines).

The dominant mechanism of acetic acid production in wine is ethanol oxidation by Acetobacter aceti and A. pasteurianus, obligate aerobic gram-negative bacteria (alpha-Proteobacteria, family Acetobacteraceae). The Acetobacter pathway for acetic acid production involves two sequential oxidations of ethanol. The first step is catalyzed by the membrane-bound pyrroloquinoline quinone-dependent alcohol dehydrogenase (PQQ-ADH, also called quinoprotein ADH or ADH–1; a transmembrane enzyme complex with its active site in the periplasm, facing outward). PQQ-ADH oxidizes ethanol to acetaldehyde using PQQ (pyrroloquinoline quinone, a redox cofactor derived from tyrosine and glutamate) as the immediate electron acceptor; the reduced PQQH₂ is then reoxidized by a cytochrome c relay in the membrane that passes electrons to oxygen as the terminal acceptor. The second step converts acetaldehyde to acetate via NAD&sup+-dependent aldehyde dehydrogenase (ALDH) in the Acetobacter cytoplasm. The net reaction is: ethanol + O₂ → acetic acid + H₂O.

Because Acetobacter are obligate aerobes, volatile acidity production requires dissolved oxygen. Wines maintained under protective inert gas (nitrogen, argon, or CO₂ blanketing), stored in full vessels without headspace, and managed with adequate free SO₂ (molecular SO₂ inhibits Acetobacter growth at 0.5 mg/L; complete inhibition at 0.8 mg/L) are at low risk for VA development. VA problems arise when barrels are not topped (ullage, air headspace above the wine), when tanks are poorly blanketed, when bung seals fail, or when wines are stored at high temperatures (accelerating both Acetobacter growth and the rate of oxygen diffusion). Late-harvest and botrytized wines often have elevated VA (>0.8 g/L) because Botrytis cinerea itself produces acetic acid and gluconic acid through its own metabolism during the bunch rot phase; the EU exempts these wines from standard VA limits.

YAN management, hydrogen sulfide production, and the TOSNA nutrient protocol

Yeast assimilable nitrogen (YAN) is the collective measure of nitrogen forms available for yeast nutrition during alcoholic fermentation. YAN consists of two components: free amino nitrogen (FAN, also called alpha-amino nitrogen; all α-amino acids that yeast can take up and use, which includes all proteinogenic amino acids except proline — proline cannot be catabolized anaerobically because its oxidation requires mitochondrial proline oxidase that is suppressed under fermentation conditions) and ammoniacal nitrogen (free NH₄&sup+ plus a small contribution from urea). Normal grape juice provides 150–400 mg N/L YAN depending on variety, vintage, and viticultural practices (nitrogen stress during vine growth, leaf removal, crop load management); white wines tend toward the lower end and red musts toward the higher. The target YAN for a complete, clean fermentation to dryness depends on initial Brix: at 22°Brix, 150–200 mg N/L is typically adequate; at 27°Brix, 300–350 mg N/L is recommended.

When YAN falls below approximately 150 mg N/L (or when the ratio of initial sugar to YAN — the sugar:YAN ratio — is high), yeast cannot synthesize adequate amounts of cysteine and methionine through the sulfur assimilation pathway. The enzyme sulfite reductase (MET10 gene product in yeast, a flavin-iron enzyme using NADPH as electron donor) reduces sulfite (SO₃²−, derived from inorganic sulfate assimilation or added winemaking SO₂) to sulfide (S₂−) as the first step in sulfur amino acid biosynthesis. S₂− is then used to produce cysteine and methionine in subsequent steps. Under YAN deficiency, the downstream nitrogen-requiring steps that would incorporate S₂− into cysteine/methionine cannot proceed, and excess S₂− is released from the cell as H₂S (hydrogen sulfide), producing the characteristic rotten-egg/sewer aroma (sensory threshold approximately 10–80 µg/L in wine). Copper ion (Cu₂&sup+;) can precipitate H₂S as CuS, which is why copper sulfate fining or copper mesh contact (traditional technique) reduces H₂S post-fermentation; however, copper fining leaves residual copper in wine and does not address the underlying YAN deficiency.

DAP (diammonium phosphate; (NH₄)₂HPO₄; containing 21% N by weight as NH₄&sup+;) is the most widely used YAN supplement in winemaking. DAP addition of 1 g/L provides approximately 210 mg N/L ammoniacal YAN. However, DAP added at excessive concentrations (>300 mg N/L total from DAP alone) can produce urea as a byproduct of yeast arginine metabolism (the arginase-urea cycle in yeast produces urea when arginine is catabolized using nitrogen from excess ammonium); urea can subsequently react with ethanol to produce ethyl carbamate (urethane, a potential carcinogen; EU limit 30 µg/L in wine), particularly at elevated temperatures during aging. This risk has driven adoption of organic nitrogen supplements (autolyzed yeast products such as Fermaid-O, which provide α-amino nitrogen from protein hydrolysates rather than ammonium) that reduce the urethane formation risk.

The TOSNA protocol (Tailored Organic Staggered Nutrient Additions, adapted from meadmaking to winemaking) for high-Brix wine fermentations divides YAN additions into multiple staggered doses across the fermentation rather than a single front-loaded dose. The rationale: a single large DAP addition at inoculation provides ammoniacal nitrogen in excess during early fermentation but leaves cells nitrogen-starved during the latter half when cell populations are high and sugar concentrations are still substantial. Staggered additions at the 24-hour mark (to support the exponential growth phase), at the 1/4-sugar-depletion point, and at the 1/3-sugar-depletion point maintain a more even supply of assimilable nitrogen throughout fermentation, reducing the severity of nitrogen starvation-induced H₂S and improving fermentation kinetics and completion. Combined use of DAP (ammoniacal nitrogen, immediately available) with organic nitrogen supplements (amino acids and peptides from yeast autolysates) provides both the rapid-response ammonium fraction and the sustained-release amino acid fraction, approaching the nutrient delivery profile of naturally nitrogen-replete musts.

iOS rates and the Apple Tax for winemaking creators

Winemaking creator audiences are heavily iOS across all major discovery and engagement platforms. YouTube winemaking content — fermentation process documentation, cellar management walkthroughs, sensory evaluation tutorials, vintage-by-vintage comparative tastings, grape-to-bottle production chronicles — tracks at 65–75% iOS, reflecting an affluent, food-and-beverage-enthusiast demographic that consumes long-form how-to video predominantly on phone and tablet. Instagram winemaking content — harvest photography, barrel hall imagery, bottling day documentation, label photography, wine tasting notes with bottle shots — tracks at 70–82% iOS; the wine enthusiast, food-and-wine, and home-winemaker communities on Instagram are iOS-concentrated. Pinterest winemaking boards — vineyard design guides, cellar equipment lists, wine pairing infographics, label design inspiration — track at 68–80% iOS. Substack wine newsletters, where many advanced winemaking educators have migrated their long-form educational content, show iOS email open rates above 70% in the food-and-drink newsletter category.

Beginning November 1, 2026, Apple charges Patreon 30% on every subscription payment processed through the iOS app.

At $150/month with 65% iOS: approximately $29.25/month ($351/year) in Apple fees. At $250/month with 72% iOS: approximately $54/month ($648/year). At $400/month with 78% iOS: approximately $93.60/month ($1,123.20/year). Enable Patreon’s web-only billing toggle before October 31, 2026 and update all subscription CTAs — Instagram bio link, YouTube About page URL, Substack footer link — to the direct Patreon web URL. Verify with a test subscription from Safari on an iPhone before November 1. Any subscription processed through the Patreon iOS app on or after November 1 will incur the 30% Apple commission regardless of the creator’s settings.

KeepTier is a self-hosted membership page for creators who want 100% of their tier revenue and zero Apple Tax. Plans from $9/month.

Frequently asked questions

How does alcoholic fermentation convert grape sugar to ethanol?

Ethanol fermentation in wine follows the Embden-Meyerhof-Parnas (EMP) glycolytic pathway: glucose is converted through ten enzymatic steps to two pyruvate molecules, yielding 2 ATP and 2 NADH. Under anaerobic fermentation conditions, pyruvate cannot be oxidized via the TCA cycle or electron transport chain, so yeast uses a two-step fermentative pathway. Pyruvate decarboxylase (PDC, EC 4.1.1.1) decarboxylates pyruvate to acetaldehyde, an enzyme that requires thiamine pyrophosphate (ThPP) as its cofactor: the thiazolium ring C-2 carbanion forms a covalent lactyl-ThPP adduct with pyruvate, stabilizing the transition state for decarboxylation and releasing CO₂ and the hydroxyethyl-ThPP intermediate, which then collapses to release acetaldehyde and regenerate ThPP. Alcohol dehydrogenase 1 (ADH1, EC 1.1.1.1) then reduces acetaldehyde to ethanol using NADH as the electron donor (oxidizing NADH to NAD&sup+;), recycling the NAD&sup+; required for glycolysis to continue. The net reaction per glucose is: glucose → 2 ethanol + 2 CO₂ + 2 ATP (Gay-Lussac equation; theoretical yield 0.511 g ethanol/g glucose). Glycerol (5–10 g/L in wine) is produced as a secondary NADH sink from dihydroxyacetone phosphate via GPD1/GPD2. The Ehrlich pathway produces fusel alcohols (isoamyl alcohol from leucine, isobutanol from valine, 2-phenylethanol from phenylalanine) when YAN is deficient and yeast catabolizes amino acids for nitrogen.

What is malolactic fermentation and why does it change wine flavor?

Malolactic fermentation (MLF) is a secondary biological conversion primarily carried out by Oenococcus oeni, a gram-positive lactic acid bacterium adapted to wine’s acid-ethanol environment. The central reaction is catalyzed by the malolactic enzyme (MLE, EC 1.1.1.38), which converts L-malate to L-lactate plus CO₂ through a concerted dehydrogenation-decarboxylation mechanism, requiring Mn²&sup+; as a divalent metal cofactor and NAD&sup+; as hydride acceptor. L-malate (pKa1 3.46, pKa2 5.10, diprotic, sharp/tart character) is replaced by L-lactic acid (pKa 3.86, monoprotic, softer/dairy character), raising wine pH by 0.1–0.5 units and reducing titratable acidity by 0.2–3 g/L. This deacidification softens mouthfeel and reduces green-apple and sharp tartness. The secondary flavor consequence is diacetyl (2,3-butanedione; butter/cream aroma; threshold 8 µg/L in white wines, 15 µg/L in reds), which does not form via the MLE reaction. Instead, diacetyl arises from α-acetolactate that leaks from O. oeni cells into the wine medium and undergoes spontaneous non-enzymatic oxidative decarboxylation in the presence of dissolved oxygen. Diacetyl production and timing is managed by controlling oxygen exposure during MLF, completing MLF fully before SO₂ addition (active bacteria reduce diacetyl to innocuous acetoin), and timing MLF relative to primary fermentation.

Why does wine turn purple vs red depending on pH?

Red wine color derives primarily from malvidin-3-glucoside (oenin), the dominant anthocyanin in Vitis vinifera, which exists in a pH-dependent equilibrium among four molecular forms. The flavylium cation (AH&sup+;) is the red/purple form (λmax approximately 515–525 nm; molar absorptivity ~28,000 L mol−¹ cm−¹), dominant at pH <3.0. The carbinol pseudobase (hemiketal, A-OH) is colorless and dominant at wine pH 3.0–4.0; water addition at the C-2 position of the flavylium ring disrupts the aromatic conjugation that produces visible color. The quinoidal base (A) is blue-purple (λmax shifted bathochromically to ~570–590 nm) and dominant above pH 5. At typical wine pH 3.2–3.6, approximately 80–95% of oenin is in the colorless carbinol form, but the remaining 5–20% flavylium cation produces intense color due to its very high molar absorptivity. Lower wine pH means a higher proportion of flavylium cation and redder color; higher wine pH shifts the hue toward purple-blue. Copigmentation — non-covalent stacking of anthocyanins with co-pigment molecules (caffeic acid, quercetin glycosides) — stabilizes the flavylium cation against water addition and increases effective absorbance by 30–200% (hyperchromic effect) with a bathochromic shift of 5–20 nm, deepening color further. Pyranoanthocyanins (vitisin A/B) form during barrel and bottle aging when acetaldehyde or pyruvate bridge across the C-4/C-5 positions of the anthocyanin, blocking water addition and producing pH-stable pigments that retain color at wine pH without relying on the flavylium/carbinol equilibrium.

What is the chemistry of wine tannins and astringency?

Wine tannins are predominantly condensed tannins (proanthocyanidins) — oligomers and polymers of flavan-3-ol monomers joined by C4→C8 carbon-carbon interflavanic bonds (B-type). The principal monomers are (+)-catechin and (−)-epicatechin; skin tannins additionally contain epigallocatechin (EGC) units, producing prodelphinidins with a pyrogallol B ring. Seed tannins are galloylated at C-3 (gallic acid ester linkage on epicatechin), increasing tannin-protein binding affinity and astringency/bitterness. Molecular weights range from approximately 578 Da (dimer) to 3,000 Da and above for hexamers and larger oligomers. Astringency results from tannin binding to salivary proline-rich proteins (PRPs) through two forces: (1) hydrophobic stacking of tannin aromatic rings against the pyrrolidine ring face of proline residues; (2) hydrogen bonding between tannin phenolic OH groups and protein backbone carbonyls and side chains. At sufficient tannin:PRP ratios, multi-point tannin bridging produces cross-linked PRP aggregates or precipitates in the oral cavity, perceived tactilely as puckering, drying astringency. During barrel and bottle aging, tannins undergo oxidative polymerization (increasing average DP) and form tannin-anthocyanin polymeric pigments, gradually changing the quality of astringency from the angular, grippy character of young seed tannins to the rounder, more integrated texture of aged polymeric tannins.

How does sulfite protect wine from oxidation and microbes?

Sulfur dioxide in wine exists in an aqueous equilibrium among three species: molecular SO₂ (H₂SO₃, pKa1 1.81), bisulfite ion (HSO₃−, dominant at wine pH 3.0–4.0), and sulfite ion (SO₃²−, pKa2 7.2, negligible at wine pH). Only molecular SO₂ is antimicrobially active: the un-ionized form diffuses freely across microbial cell membranes, dissociates at intracellular pH (~7.0) to bisulfite and sulfite, and reacts with NAD&sup+;, protein disulfide bonds, and DNA. The Henderson-Hasselbalch equation determines the molecular SO₂ fraction: at pH 3.2, approximately 3.9% of free SO₂ is molecular; at pH 3.6, only 1.6% is molecular. To achieve the target 0.5–0.8 mg/L molecular SO₂, winemakers must add 20–24 mg/L free SO₂ at pH 3.2 but 50–60 mg/L at pH 3.6 — a 2.5-fold difference. Antioxidant protection comes from direct reduction of hydrogen peroxide (HSO₃− + H₂O₂ → SO₄²− + 2H&sup+; + H₂O) and reduction of ortho-quinone intermediates from polyphenol oxidation. Bound SO₂ is inactivated by carbonyl binding: the acetaldehyde-bisulfite adduct (K ≈ 1.5×10⁶ M−¹) is effectively non-releasable, so wines high in acetaldehyde require substantially higher total SO₂ additions to achieve equivalent free SO₂ and target molecular SO₂.

Apple takes 30% of your Patreon revenue from iOS subscribers starting November 1, 2026. KeepTier is a self-hosted membership page that keeps 100% of your tier revenue out of Apple’s hands. Switch before the deadline.

See KeepTier plans from $9/month →

Patreon for winemaking creators · Patreon for beekeeping creators · Patreon for mead making creators · Eight Patreon alternatives compared