Hydroxycinnamic Acid Precursor System¶

Version 1 · Disclaimer

Consolidates the diagnostic framing, pathway chemistry, dose calculations and pitching/temperature/oxygen strategy from prior sessions. Collaborator notes, not primary literature; confidence flagged inline. Stock preparation lives in 08 (single source — duplicates removed). §11 (inhibitors) and §12 (dosing cadence) added after review; §1–§10 numbering unchanged so external cross-references still resolve.

1. Diagnosis: funk inconsistency¶

The empirical pattern (some funky early, some funky after 2-3y, some never; HCA-rich additions reliably enable funk — see 09) points at the substrate side of the pathway, not Brett capability. The clove→leather/hay fade proves VPR operates; the constraint is magnitude. Four bottlenecks:

1. Vinyl phenol scrubbing in primary — POF+ Sacch decarboxylates HCAs early; vigorous CO₂ off-gasses much of the vinyl phenol pool before Brett's VPR (upregulated post-sugar-depletion) can reduce it. (Medium-high)
2. Caffeic deficiency — malt supplies almost no caffeic acid; the funk-enabling additions are all caffeic-derivative-rich; the failures mostly aren't. (Medium; four-carboy experiment tests it directly)
3. Flat 22 °C secondary — below VPR optimum and free of the seasonal transitions that drive Brett through activity phases, while warm enough to keep LAB drifting acidity upward. (Medium)
4. Enzyme inhibition and intermediate interception — a wort can be precursor-replete and still make no funk if PAD/VPR are inhibited (SO₂, sorbate, metals, adverse pH/redox) or if the vinyl phenol intermediate is chemically trapped before VPR reaches it. Precursor availability is necessary, not sufficient. Full treatment in §11. (Medium; added after the original "entirely precursor-limited" framing was challenged)

The original framing — that the inconsistency is entirely precursor availability — is now treated as incomplete. Both axes are live, and they are distinguishable in practice: a precursor-limited carboy responds to the §9 rescue programme; an inhibited or intercepted one does not respond at any dose, which is exactly the §9 week-24 "no response at full dose" outcome. Run the §11 audit before concluding the strain is at fault.

2. The four pathways¶

Two steps everywhere: PAD (free acid → vinyl phenol; fast in POF+ Sacch, slow in Brett, absent in POF− Sacch and B. custersianus) then VPR (vinyl → ethyl; Brett only, plus some LAB weakly).

Complex traditional funk = 4-EP backbone + 4-EG layer + 4-EC depth. Aged-lambic abundance: 4-EP > 4-EG > 4-EC. Some strains stall at the catechol (4-VC→4-EC) step — if caffeic loading yields no 4-EC, suspect strain, not substrate.

The two steps have very different kinetics, and the gap between them is the root of most dosing decisions: PAD is fast and near its optimum in a fermenting wort; VPR is slow and, in finished beer, operating well below both its pH and its temperature optimum (§11). Vinyl phenol is therefore the compound that accumulates, and the compound that gets lost (§12).

3. Targets in finished beer¶

4-EP:4-EG ≈ 2:1–3:1. Overshoot risk sits just above the traditional range — err low first time; blending fixes overshoot (06), nothing un-does a band-aid solo bottle except dilution.

4. Wort baselines (what malt gives free)¶

Raw wheat > malted/flaked for bound ferulic (denser arabinoxylan). Medium confidence on absolute figures; wide literature range.

5. Conversion maths¶

Stoichiometric mass ratios ~1.3 mg precursor per 1 mg ethyl phenol for all three pathways (chlorogenic → caffeic ~2.0). Working efficiencies: 50% for free acids in active fermentation; 3–5% overall for bound forms (chlorogenic in green coffee extract — esterase liberation and then conversion, each of them partial). Efficiency order: p-coumaric > ferulic > caffeic (competing oxidation). Check: 110 mg chlorogenic ≈ 56 mg caffeic equivalent; in 5 L that is ~11 mg/L, and 3–5% of it lands as ~250–400 µg/L 4-EC, the yield quoted in 07 §2 and 08 §3.

Dose table (from 10 mg/mL stocks; per 5 L unless stated)¶

Interactive version of this table and the §12 schedule: HCA dosing bench.

Totals above are unchanged; how they are distributed in time is revised in §12 — in short, less at Day 0 in the funk-forward path, more at the secondary sampling points.

Donor design targets 4-EP 2,000–3,500 µg/L; perceptual ceiling ~5,000 µg/L beyond which the character reads "wrong" even diluted (medium confidence). Donor at 24–25 °C, no fruit, same Brett blend as recipients.

6. Pitching strategy — reconciled decision logic¶

The 10% POF+ co-pitch figure was only ever a funk-protocol number; it was never an ester-forward recommendation.

7. Temperature strategy¶

VPR optimum 25–28 °C; activity falls steeply below ~23 °C (22 °C runs ~60–75% of 25 °C rate). Options ranked:

1. Activation → maturation split (best): 24–25 °C months 2–9, then 22 °C. Mirrors seasonal cellars; more effective than any steady state. (Medium-high)
2. Steady 23–24 °C: ~80% of optimum, simple.
3. Steady 22 °C (status quo): works; explains the 3–5 year timeline.
4. Zone separation: warm shelf for months 3–12 carboys, cool cabinet for year-2+ — approximates (1) without cycling.

Acetic-prone strains (WBC type) need tight seals before any warm phase.

Note the purified-enzyme figure: VPR optimum is ~30 °C in vitro (Tchobanov 2008). The 25–28 °C working window is a compromise with Brett's own upper bound (03 §5) and with acetic/THP risk, not the enzyme's own ceiling.

8. Oxygen management¶

• Too anaerobic → sluggish Brett, THP lingers (clearing THP needs some O₂ + time).
• Too aerobic → AAB acetic, strain-dependent Brett acetic, THP formation.
• Sampling every 4–6 weeks (brief airlock lift) is the calibrated micro-oxygenation dose; the failed cotton-bung experiment (09) marks the upper bound. Dextrose at sampling is unnecessary while fermentation is ongoing (CO₂ still evolving) — reserve it for near-terminal carboys.
• Procyanidins (oak, pine bark, apple extracts) buffer O₂; oak-aged vessels tolerate more exposure.

There is a redox argument for the same cadence, independent of THP: VPR is NADH-dependent (Tchobanov 2008). Excess O₂ reoxidises NADH and starves the reduction step; strict anaerobiosis triggers the Custers effect (NADH/NAD⁺ imbalance stalls Brett's glycolysis). The 4–6 week lift sits between the two. (Medium)

9. Late-stage rescue protocol (stalled funk, 6–18 month carboys)¶

Preconditions: Brett established, primary long done, no active spoilage, and the §11 inhibitor audit passed — dosing an inhibited carboy wastes stock and reads as a false negative.

At 22 °C use 6-week intervals (24-week programme); at 24–25 °C compress to 4-week intervals.

Method: sample first; dose via syringe/vodka slurry under CO₂ blanket; swirl gently; re-seat airlock.

10. Diagnostic sequence when funk fails¶

1. Pathway check: clove→leather fade present? If no Brett character ever → repitch fresh known-funky Brett (Boon dregs, Funk Weapon #1).
2. Inhibitor audit: run §11 before spending stock. Cheap, retrospective, and it is the step that distinguishes "no substrate" from "substrate blocked".
3. Substrate: run §9. Response ⇒ build spiking into standard protocol.
4. Temperature: 24–25 °C for 6–9 months.
5. Strain diversity: add a different slant/dregs.
6. Accept and blend: ~40% of batches serve better as blend components than solo projects.

11. The inhibitor side: when precursors are present but funk still fails¶

Right precursors, wrong conditions, no funk. The pathway can be blocked at three places: the enzymes (PAD/VPR inhibited), the intermediate (vinyl phenol chemically trapped or stripped before reduction), and the liberation step (cinnamoyl esterase inhibited, so bound HCAs never become free acids).

11.1 Enzyme inhibitors¶

11.2 Interception of the intermediate¶

• Anthocyanin trapping. 4-vinylphenol and 4-vinylguaiacol react with anthocyanins to form vinylphenolic pyranoanthocyanins — stable pigments. Every molecule trapped this way is permanently removed from the VPR substrate pool. Anthocyanin-heavy dark fruit added while the vinyl pool is still building can therefore divert precursors into colour instead of funk. Consistent with the elderberry / mulberry failures in 09; cassis and aronia are counterexamples, so this is at best a partial explanation. Action: add anthocyanin-heavy fruit after ethyl phenols are established, not before. (Low-medium)
• CO₂ stripping. Already the §1 bottleneck 1 — worth restating here as interception rather than deficiency: the substrate was present and converted, then left the vessel.
• Lees and oak adsorption. Volatile phenols adsorb to yeast cell walls and to oak; lyophilised yeast is used commercially as a 4-EP bioadsorbent. Heavy lees contact and heavy oak both shave the finished figure. (Medium)

11.3 Blocked liberation of bound forms¶

• High procyanidin loads (pine bark, heavy new oak, apple polyphenol extract at high rate) bind proteins, including secreted cinnamoyl esterase, slowing release of the bound HCA long tail that 07 §1 depends on. This trades against their O₂-buffering benefit — use one heavy tannin source per vessel, not several. (Low-medium)
• Cyanogenic and other glycoside ballast (elderberry) — the existing suspicion in 07 §3 stands, unresolved.

11.4 Re-reading the 09 empirical failures¶

The "did not funk" list is better explained by inhibition/interception than by precursor content alone:

Action — record the form, not just the ingredient. For every future addition log: fresh / frozen / dried / concentrate / tea / extract, and whether the label declares sulfite, sorbate or benzoate. Retrofit this to the 09 entries where it can still be recalled. That single field would settle most of the table above, and it costs nothing.

12. Dosing cadence: micro-dosing vs full spikes¶

12.1 What the toxicity data actually says¶

Substrate toxicity is not a reason to micro-dose at homebrew doses. Harris et al. 2015 (Foods 4(4):581) measured growth inhibition across 12 Brett strains:

Toxicity order ferulic > p-coumaric >> caffeic; lags at 4 mM recovered to control growth within 72 h in all but two strains. A 25 mg spike into 25 L is 1 mg/L ≈ 0.006 mM — roughly three orders of magnitude below the lag threshold, four below MIC. Any local plume from a 1 mL bolus dissipates on the swirl. So: don't micro-dose for the culture's sake.

12.2 The three reasons that do hold¶

1. Flux matching (the main one). PAD is fast and VPR is slow and sub-optimal (§11). A bolus is decarboxylated to vinyl phenol faster than it can be reduced, and the vinyl pool then leaks — CO₂ stripping (worst in primary), oxidation, anthocyanin adducts, lees adsorption. Only the fraction reduced in situ survives as ethyl phenol. Staging matches input to VPR throughput. (Medium)
2. Residence-time losses of the free acid. Unconverted free HCAs oxidise, polymerise with wort polyphenols, and bind protein and lees over weeks. Mass delivered when conversion capacity exists yields more than the same mass parked in the vessel for months. (Medium)
3. Overshoot control. 4-EP overshoot is irreversible except by dilution (§3). Staged doses are taste-gateable; a single bolus is a bet. This argument stands on its own regardless of the chemistry. (High)

12.3 The reasons not to go finer¶

1. Every dosing event is an airlock opening. The 4–6 week sampling lift is already the calibrated O₂ dose (§8). Dose events must be co-located with sampling events, never added on top. This is the binding constraint and it sets the floor at roughly 4–6 doses across a six-month window.
2. Pipetting resolution. 10 mg/mL stock with a 1 mL syringe reads to ~0.5 mg. Below ~2 mg per 5 L, measurement error dominates the dose. (The 20 mg/mL stock noted in 08 §1 is for volume-constrained additions, not for finer splitting.)
3. Stock shelf life ~6 months (08 §1). A programme longer than that needs a fresh stock, not more doses.

12.4 Revised distribution (totals from §5 unchanged)¶

Per-acid rules:

• p-Coumaric — the acid worth staging. Front-load least, gate hardest.
• Ferulic — micro-dose or skip entirely. Most growth-inhibitory of the three, the 44 °C rest already delivers 8–15 mg/L, and 4-EG overshoot reads smoky/medicinal. Never bolus ferulic.
• Caffeic via green coffee — bound chlorogenic is already a slow-release micro-dose; esterase liberates it over weeks. Splitting capsules below a half adds handling risk for no gain. Dose whole or half capsules at sampling points.

Net change from the previous practice: same total mass, fewer molecules lost, and every dose after the first is taste-gated.