Malolactic fermentation is far more than an acid adjustment. As lactic acid bacteria consume malic acid, they also release a cascade of metabolic byproducts that shape a wine’s aroma, mouthfeel, and stability. Diacetyl contributes buttery notes, esters modulate fruit expression, and the bacteria’s enzymatic machinery can either enhance or strip away desirable volatile compounds. The choice of which bacterium performs this transformation is therefore one of the most consequential decisions a winemaker makes, and the industry’s reliance on a narrow genetic pool has long struck researchers as both a limitation and an opportunity.

The research team, led by Carla Vargas-Luna of the Pontificia Universidad Católica de Chile, focused on two indigenous strains: Levilactobacillus brevis BCV-46 and Levilactobacillus brevis C18, both isolated from Chilean grapes. They compared these against a commercial Lactiplantibacillus plantarum strain, a species that has emerged in recent years as the leading alternative to Oenococcus oeni. The experiments were conducted at microvinification scale, using four-liter fermentations of Chardonnay must, with each bacterial strain co-inoculated alongside Saccharomyces cerevisiae, the yeast responsible for the primary alcoholic fermentation. This co-inoculation strategy, in which yeast and bacteria are introduced together rather than sequentially, is gaining favor because it shortens production schedules and reduces the risk of spontaneous bacterial growth between fermentations.

A central concern in any such experiment is whether the bacteria will interfere with the yeast’s work. Alcoholic fermentation depends on Saccharomyces cerevisiae converting sugars to ethanol efficiently and completely, and some lactic acid bacteria produce compounds that inhibit yeast metabolism. The researchers found encouraging news on this front: co-inoculation did not impair alcoholic fermentation in any of the treatments. The yeast completed its task regardless of which bacterial companion was present, confirming that these native strains are compatible with standard winemaking yeast at the timing and conditions tested.

The malolactic conversion itself, however, proved more uneven. Only partial malolactic fermentation occurred across the trials, and the degree of malic acid degradation varied substantially by strain. The commercial Lactiplantibacillus plantarum achieved the highest level of malic acid breakdown, demonstrating why it has become the benchmark alternative starter. Among the native strains, Levilactobacillus brevis BCV-46 outperformed its compatriot C18 in acid conversion. The incomplete fermentation suggests that these strains, while viable, may require further optimization of inoculation timing, cell density, or nutritional conditions to fully complete the acid conversion in white wines, which are generally a harsher environment for lactic acid bacteria than reds due to lower pH, higher acidity, and fewer nutrients.

Despite the partial conversion, the chemical differences among the wines were measurable and meaningful. The researchers documented variations in chemical composition across the treatments, though all wines remained within acceptable enological limits, meaning none of the bacterial interventions pushed the wines outside the ranges considered sound and stable for commercial production. This is a nontrivial finding, because lactic acid bacteria can consume residual sugars, produce acetic acid, or alter pH in ways that compromise wine integrity. The native strains passed this basic quality threshold, establishing their fundamental suitability for further development.

The safety dimension of the study produced the sharpest contrast between the two native isolates. Biogenic amines, including histamine, tyramine, and putrescine, are nitrogen-containing compounds that some lactic acid bacteria generate by decarboxylating amino acids. At elevated concentrations, these amines can trigger headaches, flushing, and blood pressure changes in sensitive individuals, and histamine in particular is regulated in some markets. Using molecular screening, the researchers searched for the amino acid decarboxylase genes that confer the ability to produce these compounds. The genes were detected only in L. brevis C18, and correspondingly, wines fermented with that strain showed higher concentrations of tyramine and putrescine. The other strains, including BCV-46 and the commercial L. plantarum, did not carry the decarboxylase genes and did not elevate these amines.

This single genetic distinction transforms the practical outlook for the two native strains. BCV-46 emerges as a candidate worth pursuing: it degraded malic acid respectably, left the wine chemically sound, and carried no detectable amine-producing potential. C18, by contrast, would require careful evaluation before any commercial use, since its amine production represents a genuine food safety consideration. The finding underscores a principle that wine microbiologists have emphasized repeatedly: safety and quality traits are strain-dependent rather than species-dependent, and each candidate starter must be individually characterized before deployment.

Perhaps the most appealing results came from the sensory analysis. The wines produced with different bacterial strains developed significantly different sensory profiles, and the selected native strains enhanced floral, fruity, and creamy attributes in the Chardonnay. These are precisely the aromatic dimensions that white wine consumers prize, and the fact that indigenous Latin American strains could amplify them hints at a strategy for producing wines with a genuine sense of place. Regional microbial biodiversity, long dismissed as a source of unpredictable spoilage, may instead be a reservoir of distinctive flavor signatures that commercial starter cultures cannot replicate.

The broader significance of the work lies in its demonstration that the next generation of malolactic starter cultures may come from vineyards far removed from the traditional wine heartlands of Europe. As climate change alters growing conditions and winemakers worldwide seek to differentiate their products, native bacterial strains offer a dual advantage: adaptation to local conditions and the potential for unique sensory contributions. The Chilean and Mexican research groups behind this study have shown that careful screening, combining fermentation performance testing, chemical analysis, sensory evaluation, and genetic safety checks, can identify which native strains deserve investment and which should be left in the vineyard. For now, the commercial Lactiplantibacillus plantarum still leads in acid conversion, but the native BCV-46 strain has staked its claim as a safe, flavor-enhancing alternative, and the race to domesticate Latin America’s wine microbiome has only begun.

Subject of Research: Non-Oenococcus lactic acid bacteria as malolactic fermentation starters in Chardonnay wine

Article Title: Latin-American non-Oenococcus lactic acid bacteria in Chardonnay wine: effects on chemical composition, sensory profile, and biogenic amine production

Article References: Vargas-Luna, C., Franco, W., Arvizu-Medrano, S., Cortés-Herrera, C., Redondo-Solano, M., Barboza-Vargas, N., Godoy-Olivares, L., & Miranda-Castilleja, D. (2026). Latin-American non-Oenococcus lactic acid bacteria in Chardonnay wine: effects on chemical composition, sensory profile, and biogenic amine production. Food Science and Biotechnology. https://doi.org/10.1007/s10068-026-02330-6

Image Credits: AI Generated

DOI: 10.1007/s10068-026-02330-6

Keywords: malolactic fermentation, Chardonnay, lactic acid bacteria, Lactiplantibacillus plantarum, Levilactobacillus brevis, biogenic amines, wine microbiology, co-inoculation, Saccharomyces cerevisiae, sensory profile, food safety, Latin America

The study, published in Frontiers in Industrial Microbiology by researchers from Vioryl Chemical and Agricultural Industry and the University of West Attica, focused on Oenococcus oeni, a lactic acid bacterium that dominates malolactic fermentation in winemaking. The team isolated thirty bacterial strains from five spontaneously fermented Xinomavro wines from different Greek regions and confirmed their identity through 16S rRNA gene sequencing. Strain-level typing by multiplex PCR revealed three distinct profiles, designated MD00, MD02, and MD05. Intriguingly, the same strains appeared in wines from different vintages, suggesting that resilient resident populations persist in the winery environment, likely protected by biofilms on cellar surfaces.

The researchers fed these strains two different precursor molecules. The first was ferulic acid, a hydroxycinnamic acid abundant in plant biomass and agro-industrial waste, which is already a well-established vanillin precursor. The second was vanillylmandelic acid, or VMA, a vanilloid that is produced industrially as an intermediate in the classic two-step Rhône-Poulenc chemical synthesis of vanillin from guaiacol. VMA had never before been demonstrated as a microbial substrate for vanillin biosynthesis, making this aspect of the study a genuine first.

The results revealed a striking divergence in how the bacterium handles each substrate. When given ferulic acid, the O. oeni strains rapidly depleted it within forty-eight hours, converting it primarily into 4-vinylguaiacol, a smoky, clove-like aroma compound valued in its own right, with only modest amounts of vanillin appearing alongside. The researchers attribute this pattern to phenolic acid decarboxylases, cofactor-independent enzymes that cleave the ferulic acid side chain through a quinone methide intermediate. Vanillin formation from ferulic acid likely requires a slower, less efficient route involving hydration of the double bond and subsequent redox reactions, and under the tested conditions only about 4.3 milligrams per liter of vanillin was produced from 776 milligrams per liter of ferulic acid.

VMA told a very different story. Because it lacks the conjugated acrylic side chain of ferulic acid and carries an alpha-hydroxyl group instead, the decarboxylase route is blocked, and the bacterium channels it almost exclusively toward vanillin. The vanillin concentrations achieved from VMA were roughly nine times higher than those from ferulic acid, and only trace amounts of 4-vinylguaiacol appeared, confirming that the two substrates follow fundamentally separate biochemical pathways. The authors hypothesize that a cofactor-dependent decarboxylase or reductase system is at work, possibly analogous to thiamine pyrophosphate-dependent enzymes such as benzoylformate decarboxylase or prenylated flavin mononucleotide-dependent decarboxylases, though the exact enzymatic machinery remains to be identified through future genomic and biochemical screening.

Feeding strategy proved critical to productivity. In a one-time feeding approach, where all the VMA was added at the start of fermentation, vanillin production surged initially but plateaued as the substrate was exhausted, a classic signature of substrate-limited kinetics. In contrast, a continuous feeding regime, in which VMA was supplied incrementally at 1.818 millimoles per liter on multiple days, maintained a near-constant substrate concentration and sustained enzymatic activity over extended periods of up to twenty-four days. The team also tested higher substrate loadings of 10.9 and 16.36 millimoles per liter, monitoring the process kinetics through periodic sampling.

Strain differences emerged as a decisive factor. MD05 consistently outperformed its counterparts, exhibiting a steep, linear increase in vanillin production rate even at the highest VMA concentration, with no plateau or decline over fourteen days. Peak cell viability and maximum vanillin production coincided on day fourteen, underscoring how tightly biosynthesis depends on microbial health. MD02, by comparison, showed a peak in production followed by a decline at elevated substrate levels, which the authors attribute to metabolic stress, substrate overloading, or product toxicity. MD00 produced the lowest yields and was excluded from prolonged trials. The maximum metabolic efficiency reached approximately 25.7 percent yield relative to the theoretical biocatalytic conversion, a figure the researchers consider a promising starting point for optimization.

A subtle but important methodological challenge involved thermal decomposition. Autoclaving the fermentation medium, which was necessary because the precursors were insufficiently soluble at room temperature to allow membrane sterilization, caused some VMA to break down chemically into vanillin without any microbial help. To disentangle this abiotic contribution from genuine biotransformation, the team ran uninoculated controls subjected to identical sterilization and incubation conditions, then subtracted the background vanillin from the microbial cultures. Across all experiments, the uninoculated controls showed negligible vanillin formation, confirming that the compound in the inoculated flasks was overwhelmingly of microbial origin.

Downstream processing revealed both the promise and the practical hurdles of scaling the process. After fermentation, the cultures were centrifuged and the supernatant extracted twice with diethyl ether, then concentrated under vacuum. Quantification by ultra-high-performance liquid chromatography before extraction and gas chromatography-mass spectrometry afterward showed that between 31 and 46 percent of the produced vanillin was lost during recovery, likely through incomplete phase partitioning, surface adsorption, volatilization, and degradation. Strain MD05 again fared best, with the smallest absolute gap between measured and recovered vanillin. The GC-MS analysis of the final extract detected only expected fermentation byproducts, including lactic, butyric, isobutyric, and isovaleric acids and benzaldehyde, with no toxic or unexpected compounds, supporting the biosafety of the process.

The broader significance lies in regulation and sustainability. Under both European Regulation (EC) No. 1334/2008 and U.S. food labeling standards, vanillin can only be marketed as natural if it derives from natural precursors through processes free of synthetic chemical reagents and catalysts. Microbial conversion of natural precursors like ferulic acid satisfies these criteria, which is why biovanillin commands premium market value. O. oeni brings additional advantages: it is generally recognized as safe, uniquely adapted to harsh conditions of low pH, high ethanol, and nutrient limitation, and already deeply embedded in food fermentation practice. The authors suggest that future work should focus on identifying the specific enzymes behind the VMA-to-vanillin conversion, using comparative genomics to select high-performing strains like MD05, optimizing fermentation parameters such as pH and temperature, and exploring alternative downstream approaches such as adsorption resins to cut recovery losses. If those steps succeed, the wine cellar’s favorite bacterium could become an industrial workhorse for natural flavor production, turning agricultural byproducts into one of the world’s most beloved aromas.

Subject of Research: Microbial bioconversion of ferulic acid and vanillylmandelic acid to vanillin and 4-vinylguaiacol by Oenococcus oeni for sustainable natural flavor production

Article Title: Bioconversion of ferulic acid and vanillylmandelic acid to 4-vinyl guaiacol and vanillin mediated by Oenococcus oeni: towards sustainable flavor production

Article References: Dermitzaki, E., Gkanatsios, N., Tzamourani, A., Vamvakias, M., Bardouki, C., Zormpa, O. G., Dourtoglou, V. G., & Dimopoulou, M. (2026). Bioconversion of ferulic acid and vanillylmandelic acid to 4-vinyl guaiacol and vanillin mediated by Oenococcus oeni: towards sustainable flavor production. Frontiers in Industrial Microbiology, 4, Article 1868633. https://doi.org/10.3389/finmi.2026.1868633

Image Credits: AI Generated

DOI: 10.3389/finmi.2026.1868633

Keywords: Oenococcus oeni, vanillin, biovanillin, ferulic acid, vanillylmandelic acid, 4-vinylguaiacol, biotransformation, natural flavors, lactic acid bacteria, fermentation, wine microbiology, sustainable biotechnology