# Wine microbiology & terroir

Yeast succession and ethanol tolerance, malolactic conversion by Oenococcus oeni, Brettanomyces as a defined fault, and how much of the microbial terroir hypothesis the evidence actually supports.

A spontaneous ferment is a relay: several yeasts start it, and the one that finishes it is the one that tolerates what the others produced.

Source: https://en.bioecon.ru/docs/agri-food/food-alt-protein/wine-terroir-microbiome/
Updated: 2026-08-25



Wine fermentation can be started with a selected yeast or left to the organisms already present on the fruit and in the winery. The second route is the interesting one microbiologically, because it is not one organism doing the work but a sequence.

## Succession, driven by ethanol

Grape must at the start carries a diverse population dominated by non-*Saccharomyces* yeasts — *Hanseniaspora*, *Metschnikowia*, *Candida*, *Torulaspora* — which are more abundant on grape skins than *Saccharomyces cerevisiae*, often by a wide margin.

These begin the fermentation and, in the first days, produce much of the wine's aromatic complexity: esters, higher alcohols, glycerol, and enzymatic release of aroma compounds bound to sugars in the grape.

Then they stop. Most non-*Saccharomyces* yeasts cannot tolerate ethanol much above 4–6 percent, and they die as the alcohol they helped create accumulates. *S. cerevisiae*, present initially as a minority, tolerates ethanol far better and finishes the fermentation to dryness.

**The succession is therefore driven by the product**: the early organisms create the conditions that eliminate them and select their successor. Inoculating with *S. cerevisiae* at the start does not merely add an organism — it overwhelms the early phase and removes the non-*Saccharomyces* contribution, which is precisely the trade-off between reliability and complexity that the two approaches represent.

## Malolactic conversion is bacterial, and not a fermentation of sugar

A second, separate transformation is usually carried out by *Oenococcus oeni*, a lactic acid bacterium that tolerates low pH and high ethanol.

It decarboxylates **malic acid to lactic acid**. Malic acid is diprotic and sharp; lactic acid is monoprotic and softer, so acidity falls and the wine's perceived harshness reduces — the primary reason the conversion is encouraged in most reds and some whites. It also produces **diacetyl** from citrate metabolism, which is the buttery note in some Chardonnay, and it consumes a nutrient that would otherwise support spoilage organisms, improving microbial stability.

## Brettanomyces is a defined fault

*Brettanomyces bruxellensis* converts hydroxycinnamic acids present in grapes into volatile phenols — **4-ethylphenol** and 4-ethylguaiacol — described as barnyard, medicinal or horse-like. It tolerates ethanol, low pH and low nutrients, and survives in barrel wood where it is difficult to eliminate.

Whether a trace is a flaw or a feature is a matter of style; that the compound and its origin are precisely identified is not. This is one of the few wine faults where the sensory descriptor maps to a named molecule from a named organism by a known pathway.

## What "microbial terroir" does and does not show

Sequencing surveys have repeatedly found that grape and must microbial communities differ by region, vineyard and vintage, in patterns that correlate with climate and geography. That much is well replicated.

The stronger claim — that these regional microbial differences cause the sensory characteristics attributed to terroir — is not established. The gap is causal: correlation between community composition and region does not demonstrate that the community produces the regional flavour, and the same regions differ in soil, climate, clone, viticulture and winemaking practice, all of which also shape both the microbiota and the wine. Controlled work showing that defined non-*Saccharomyces* strains produce measurable sensory differences supports plausibility, and plausibility is not the same as demonstration. This is an active research question, and it should be described as one.