Esters, Phenols, Enzymes¶

Version 1 · Disclaimer

What makes the flavour compounds, and what removes them: the two ester families and their precursors, the oxygen switch, esterase decay, temperature and pitch-composition effects, and the Brett enzymes (β-glucosidase, β-lyase) that release glycoside-bound aromatics and thiols.

Literature synthesis: Tyrawa et al. 2019 (JIB, doi:10.1002/jib.565), Menoncin et al. 2019 (JIB), Spaepen & Van Oevelen 1982, Rodríguez et al. 2008 (PMC2223249), Daenen et al. 2016, Verstrepen et al. 2003 (ATF1/ATF2 regulation), Meilgaard 1975 (flavour thresholds), Shaner/Preiss MTF secondary-fermentation experiment, White Labs/Lallemand technical notes, MTF wiki. Confidence table at end.

1. The two ester families¶

Both families are condensation products formed intracellularly and excreted; both are substrate-limited rather than enzyme-limited in practice (§3). They differ in which substrate runs out first — acetate esters are limited by the higher-alcohol supply (amino-acid derived), MCFA ethyl esters by the acyl-CoA supply (lipid-metabolism derived).

Acetate esters (alcohol + acetyl-CoA, AATase — ATF1/ATF2 homologues) — the "classic Belgian" top notes:

MCFA ethyl esters (ethanol + acyl-CoA, AEATase — EEB1/EHT1 homologues) — the tropical carriers:

Brett accumulates C8–C12 fatty acids via elevated β-oxidation and esterifies them — the core mechanism of Brett's tropical output. The Brett ester-synthase genes are uncharacterised at molecular level; pathway inferred from end-products (high confidence on outputs, medium on Sacch EEB1/EHT1 analogy). Brett's AATase capability is comparatively weak: acetate esters are largely a Saccharomyces contribution and Brett actively removes them over time (§4, §6).

Threshold figures are Meilgaard-lineage beer values; treat as order-of-magnitude — matrix, ethanol and sweetness shift them, and the acetate esters are strongly suppressed perceptually by residual sweetness.

2. Mechanism and the oxygen switch¶

MCFA esters derive from acyl-CoA intermediates released prematurely from fatty acid synthase; acetyl-CoA carboxylase is the regulatory node.

• Anaerobic: no UFA synthesis → carboxylase stays feedback-inhibited → FAS stalls → MCFA acyl-CoA pools up → high MCFA ester output.
• Aerobic: UFAs made → inhibition relieved → elongation to long-chain lipids → MCFA pool drained.
• Exogenous UFAs (lipid-rich adjuncts, esp. oats' C18 linoleic/linolenic) drain the pool the same way. This is why oats are excluded from ester-forward grists — the definitive resolution of the earlier contradiction (an older note recommended oats for ester beers; that was wrong by mechanism and is superseded). Oats remain acceptable in funk-forward beers.

Acetate esters answer to the same oxygen switch, by a different route: ATF1 transcription is directly repressed by oxygen and by unsaturated fatty acids/ergosterol. So aerobic conditions and lipid-rich adjuncts suppress both families — the oxygen-early/anaerobic-late shape below is not a trade-off between them. Where they diverge is the second substrate: acetate esters also need a higher-alcohol pool from the Ehrlich pathway (leucine → isoamyl alcohol, valine → isobutanol, phenylalanine → 2-phenylethanol), which comes from wort amino acids, not from lipids. Acetyl-CoA is the shared currency: heavy aerobic respiration or vigorous biomass growth drains it away from both AATase and the acyl-CoA pool.

Practical shape: oxygen early (biomass) / anaerobic late (ester accumulation). Late O₂ also drives acetic acid and THP.

3. Precursor availability is rate-limiting¶

Rodríguez 2008: adding C6/C8 precursors raised MCFA ester output substantially; enzyme expression didn't change. The ceiling is substrate, not effort.

MCFA levers: preserve wort lipids (no kettle finings, generous trub carry — 05 §1), caprylic acid spikes (direct C8 substrate; internal acyl-CoA generation means it augments rather than enables), bicarbonate <50 ppm (>100 mg/L suppresses octanoic/decanoic acids and vinyl phenols).

Acetate levers (the parallel set, and the one the current protocols under-use): - Amino acid supply, not just FAN quantity. Leucine, valine and phenylalanine are the specific precursors. Organic nitrogen (nutritional yeast, yeast hulls, autolysate) supplies them; DAP supplies none and suppresses ester formation outright — the 05 §1 "organic nitrogen only" rule is an acetate-ester rule as much as a general one. - Protein-rich, lightly-converted wort: the turbid-pull/protein-rest regime in 05 §2 already loads the right amino acid profile; over-long protein rests that drive FAN very high push fast bland attenuation instead. - Temperature: acetate esters are the more temperature-responsive family (§5). - Pitch rate and growth: low-to-moderate growth favours acetate esters (acetyl-CoA not diverted into biomass); very high pitch rates flatten them. - Ceiling caveat: ethyl acetate rises with the same levers and turns solventy first. Acetate-ester pushes have a narrow window that MCFA pushes do not.

4. Esterase: why tropical fades¶

Brett esterase is bidirectional (same enzyme synthesises and hydrolyses). Spaepen & Van Oevelen: all Brett strains express it; most active months 6–12 and during bottle refermentation; prefers acetate esters (isoamyl acetate goes first); pH optimum 7.6, so beer pH ~3.5 suppresses but doesn't stop it.

Trajectory: esters peak with active growth (weeks 2–6 at 5 L/22–24 °C); acetate esters decay first and fastest — in a Brett-containing vessel treat any banana/honey/rose note as a months-scale asset, not a years-scale one; MCFA esters decay slower but decay; free C8/C10 acids accumulate (goaty). Ester peak and funk peak cannot coexist in one vessel — bottle at the ester peak, or blend an ester batch with a funk batch at packaging (05, 06).

Corollary for blending (06): an acetate-ester top note is best carried by the young/Sacch component of a blend and added late, because the Brett component will strip it. A high Brett pitch into a finished beer is a deliberate acetate-ester eraser (§6).

Harmonised timing (supersedes the scattered figures in older files): peak window weeks 2–6, sensory-gated bottling weeks 6–14, hard ceiling ~14 weeks.

5. Temperature (Tyrawa 2019)¶

15 vs 22.5 °C, nine B. bruxellensis strains: - Esters: strongly temperature-sensitive; 22.5 °C significantly higher for most beer strains; strain spread widens when warm. Wine strains under-threshold at both — strain choice matters more than process here. - Phenols: weakly temperature-sensitive; above-threshold at both temps; some strains left 4-VG unreduced at 15 °C. - 30 °C ferments "smelled terrible" — hard upper bound. Working windows: esters 23–24 °C; VPR optimum 25–28 °C (see 04 §7); nothing above 25 °C without a specific reason.

Family split (Sacch literature, transfers by mechanism — medium confidence in Brett): acetate esters are the more temperature-responsive of the two, roughly doubling over a 6–8 °C rise, because warmth raises both Ehrlich flux and AATase activity. MCFA ethyl esters respond less steeply. Practical read: if the goal is a banana/honey lift from a Sacch co-pitch, the temperature lever is worth using in primary; if the goal is pineapple from Brett, strain choice outranks it.

6. Pitch composition effects¶

• 100% Brett primary → highest MCFA esters (no competition for precursor pool, less CO₂ scrubbing). Phenols ~20% higher too vs co-ferment with EC-1118 (which consumes p-coumaric via its own PAD). Acetate esters stay low — a 100% Brett beer is tropical/pineapple without the banana-pear top layer.
• Sacch co-pitch → more acetate esters, fewer MCFA esters; the larger the Sacch fraction, the more the Brett ester advantage erodes. This is the only practical route to a real acetate-ester layer, and it costs MCFA output — a deliberate trade, not a free addition.
• Brett pitched into finished beer restructures esters in weeks: 4-VG→4-EG conversion by week 3; isoamyl acetate scrubbed dose-dependently (high Brett pitch deliberately strips banana); ethyl nonanoate/caprylate rise. Net effect is a shift from the acetate family to the MCFA family, not a general ester increase.
• Brett can make phenols alone (own PAD, slower); POF+ Sacch enlarges the vinyl phenol pool but competes and off-gasses it — full decision logic in 04 §6.

7. β-glucosidase (glycoside-bound aromatics)¶

Strain-specific (BbBGL2 functional gene); pH optimum ~5.75, weak below pH 4.5 — activity in finished beer is marginal, so expect slow/partial liberation. Relevant for whole hops, cherry/fruit glycosides, monoterpene release. Sensory-detectable differences confirmed for cherry beers between active and inactive strains. Test a given isolate empirically before relying on it.

8. β-lyase (thiols)¶

Cleaves cysteinylated precursors → 3MH (passionfruit/grapefruit, ~60 ng/L threshold), 3MHA (~4 ng/L), 4MMP (blackcurrant, ~1 ng/L). Precursor pools in Nelson Sauvin, Citra, Galaxy are ~1,000× free thiol levels — a small late hop tea can unlock large aroma if, and only if, the strain carries functional β-lyase (IRC7 analogue; strain-variable, less characterised in Brett). Test vs unhopped control; add late in secondary when CO₂ purging is low.

9. THP (mousy)¶

ETHP/ATHP production is strain-specific, stimulated by oxygen plus lysine and ethanol. Second reason (after acetic) to seal after the aerobic establishment window. EC-1118 at bottling scavenges residual THP; time at low pH slowly clears it. Perception threshold is very low — dilution via blending does not rescue a mousy donor.

10. Confidence table¶