September 2015 Issue of Wines & Vines

How SO2 Additions Influence Microbial Diversity During Fermentation

by Michael L. Swadener and David A. Mills

The use of sulfur dioxide (sulfites or SO2) in winemaking is hardly a new technology-it has been used since the Romans started burning sulfur candles in emptied barrels to keep them from turning sour. The general antimicrobial and antioxidant effects of SO2 in wine have also been well known for several decades, and many modern winemakers find SO2 to be a key additive for producing (and preserving) premium quality wine.

At crushing/pressing, the addition of SO2 is thought to inhibit growth of certain non-Saccharomyces yeasts and lactic acid bacteria (LAB), which can contribute to spoilage and problematic fermentations. At adequate levels (approximately 50 ppm) SO2 also reduces the activity of the oxidative enzyme polyphenol oxidase, which can rapidly deplete oxygen in juice/must at the expense of the growing yeast population during the early stages of fermentation.

practical winery vineyard

What is not well established is precisely how SO2added early on shapes the microbial diversity of a wine fermentation. N.A. Bokulich et al. (2015) addressed this issue using some of the latest methods in microbial ecology. Traditional culture-based methods for examining microbial profiles of wines, while powerful and useful for many investigations, are limited in both their sensitivity and scope. These techniques are slow, laborious and not sensitive to the presence of very small or difficult-to-cultivate populations.

For this investigation, high-throughput marker gene sequencing was used to examine the changes in both bacterial and fungal communities throughout the fermentation of a Chardonnay wine. The method begins with environmental samples-in this case fermenting must/wine-whereby the total DNA was extracted and two separate sets of primers (one bacterial and one fungal) were used in separate polymerase chain reactions (PCR) to amplify the target regions of all microbes present.

For bacteria, this region was the V4 domain of the 16S rRNA genes, and for fungi the internal transcribed spacer (ITS) 1 loci were amplified. These two "amplicons" were then sequenced by modern techniques (in this case an Illumina MiSeq device), resulting in hundreds of thousands of short DNA sequences that were then quality filtered and classified into operational taxonomical units (OTUs) prior to identification and cataloging of the various microbial taxa by comparison to existing databases.

What this yields, in the end, is the relative frequencies of dozens of different types of microbes in each sample-a model of the microbial communities. Further statistical analyses allowed comparison of samples to find the significant differences that exist among them. It is important to note that this process is based on DNA and does not differentiate between live or dead cells, nor those that might lie somewhere in between-in a viable but nonculturable (VBNC) state.

However, its great strength lies in its sensitivity to even small, and otherwise difficult to detect, populations. Even more important, it is a rapid and affordable way to obtain a much more complete description of the microbial communities of hundreds (or thousands) of samples at once.

How was the experiment conducted?

The actual Chardonnay fermentations were conducted in triplicate 5-gallon fermentations, uninoculated at 0, 15, 20, 25, 35, 50, 75, 100 and 150 mg/L SO2, as well as 0 and 50 mg/L SO2 inoculated with commercial Saccharomyces cerevisiae EC-1118. The fermentations were conducted rather conventionally and sampled after one, two, three, five, seven, 10, 14 and 21 days. That is 11 treatments, in triplicate, at eight sampling times, for a total of 264 unique samples. Simply put, no other current method would allow for such rapid and in-depth analysis of so many treatments/time points as was possible here with marker gene sequencing.

What did all this sequencing reveal?

For one, SO2 mainly impacted bacterial diversity, as fungal populations across the range of sulfite additions did not appear to differ significantly from one another throughout the course of the fermentations. This lack of significant differences may have been partly due to high variation within replicate fermentations. Bacterial diversity, on the other hand, was found to be significantly impacted by different doses of SO2.

For uninoculated fermentations, 25 mg/L or more of SO2 seemed to stabilize bacterial populations. When SO2 was less than 25 mg/L, however, bacterial populations shifted in ways that were somewhat surprising. In no-sulfite-added musts, the average number of bacterial species decreased rapidly after 10 days and remained lower than the sulfited musts. This was likely caused by the growth of a few species, especially those in the genus Lactobacillus and the larger family, Lactobacillaceae, which dominated the bacterial communities and caused the less-abundant taxa to fall below detectable levels.

In the low-sulfite musts (15 and 20 mg/L) there were significant increases in the populations of acetic acid bacteria of the genus Gluconobacter that were not observed elsewhere. These low- and no-sulfite musts were also slower to begin active fermentation and had higher Brix values (2.2º, 0.6º and -0.3º Brix) after 21 days than the greater than 25 mg/L sulfite treatments (-1.0º to -1.5º Brix). This suggests that the growth of these bacteria may have impacted fermentation completion or, conversely, the conditions that restricted fermentation led to their population increases.

Impact of yeast inoculation on bacterial populations

A stabilizing effect similar to adding more than 25 mg/L SO2 was achieved by inoculation with a strong S. cerevisiae starter culture, even with zero SO2 added. Fermentation also proceeded more rapidly and to a slightly greater degree of completion when inoculated with yeast. Thus, inoculation alone may provide a reasonable level of protection and contribute to microbial stability during fermentation of non-sulfited wines. Emphasis should be placed on the fact that this observation applies to the period of active fermentation, and such a protective effect may not persist into later stages of winemaking.

Some additional results of the study, the significance of which might be less obvious to most winemakers, include the abundant presence of the bacterial genus Erwinia and other members of the same Enterorbacteriaceae family, regardless of sulfite levels. Traditionally, these bacteria have not been considered to be participatory members of wine fermentation, as they have not been isolated from wine by culture-based methods.

The high relative abundance of Erwinia from the very beginning of fermentation supports the thought that they are environmental microbes, coming from the surface of the grapes at pressing, and are not necessarily active during the fermentation. Their apparent persisting presence through fermentation and repeated detection in this and similar studies, however, suggests that they might warrant further study as to what sort of metabolic state they are in-live, dead or VBNC. If they are not merely dead and floating around, then there is a question of what role they might play in fermentation.

The future of marker gene sequencing in winemaking

Although this specific work may generate more questions than it answers with regard to SO2, it clearly demonstrates that modern marker gene sequencing creates a far fuller and higher definition picture of the constellation of microbes present in wine fermentations and how they respond to different treatments.

Indeed, as these techniques are increasingly used in combination with other "omics" approaches, such as metabolomics, we will begin to draw connections between particular microbial taxa and key positive or negative sensory attributes that define wines.

If the regional microbial taxa are considered truly part of the terroir of the ensuing wine, as suggested by N.A. Bokulich et al. 2014, perhaps these new tools will help researchers define which specific taxa play a key role in that terroir and how winemaking decisions, including SO2 usage, influence that expression.

This text was first published in the May/June 2015 Wine & Viticulture Journal and is edited and reproduced here with permission of the publisher,

Editor's Note: For more information go to Back Issues, January 2014: "Sulfur Dioxide-The science behind this anti-microbial, anti-oxidant wine additive," by Pat Henderson.


1. Bokulich, N.A., M. Swadener, K. Sakamoto, D.A. Mills and L.F. Bisson. 2015 "Sulfur Dioxide Treatment Alters Wine Microbial Diversity and Fermentation Progression in a Dose-Dependent Fashion." Am. J. of Enol. & Vitic., 66 (1), 73-79.

2. Bokulich, N.A., J.H. Thorngate, P.M. Richardson and D.A. Mills. 2014 "Microbial biogeography of wine grapes is conditioned by cultivar, vintage and climate." Proceedings of the National Academy of Sciences of the United States of America, 111 (1), E139-48.


Michael L. Swadener's research was supported by an Adolf L. & Richie C. Heck Research Fellowship, Horace O. Lanza Scholarship, Mario P. Tribuno Memorial Research Scholarship and Wine SpectatorScholarship.

Print this page   PRINTER-FRIENDLY VERSION   »
E-mail this article   E-MAIL THIS ARTICLE   »
Currently no comments posted for this article.