Mapping wine microbes: An introduction to microbial ecology for wine drinkers

Now if that demystifies wine, I can understand how some might consider that threatening; I actually consider it enabling and wonderful. – David A. Mills, PNAS podcast (Dec 2013)

Some consider it the holy grail of wine research: to answer the question, “What is terroir?” It’s probably not even a fair question to pose in that simplistic way, because it’s fairly clear that the answer is a complicated combination of every factor in the winemaking process, and each discipline of science will weigh in with a theory of its own. It’s local microclimate and weather; it’s the chemistry of the soil; it’s the interaction of the plant with its surroundings; in the variety and growth characteristics of the oak tree used to make the barrels; it’s all in the local tradition that governs the practices of the winemaker; it’s the perception of the drinker who sees a region written on the label (for further reading, some places to start include Heymann 2013, White 2009, Gougeon 2009. In fact, the scientific evidence for each element responsible for terroir merits several blog posts of its own...) Of course, as a microbiologist I’m here to tell the story of the single-celled organisms.


Hot new research

The latest chapter in this story is a super-cool study that was published a few months ago by the Mills Lab, a food microbiology group at UC Davis. Of course, there’s been previous work characterizing the microbiology both in the vineyard and in the fermentation tank. However, until recently, culture-based work and DNA sequencing has been tedious enough that researchers were only able to detect, at most, a few hundred microbes in any one sample—it was simply a lot of work to identify each organism. But in a world where there may be hundreds of different microbes inhabiting a single grape, and tens of thousands in a gram of soil, that kind of sequencing depth didn’t give us any confidence that we could really compare microbial communities between vineyards.

Now, the advent of super-sexy next-generation DNA sequencing makes it possible to detect millions of microbes all at once, and allowed Bokulich and colleagues (2014) to compare communities between several vineyards and varietals in a rigorous way, and finally give the beginning of a concrete answer to the question of whether microbes could contribute to a wine’s terroir. I think this is a really neat study—plus, many of the techniques here are exactly the ones I use in my own research—but it’s also a pretty dense paper with some seriously complex figures. So here’s my attempt at summarizing.

I should also point you to a podcast by PNAS with an interview by the principal investigator: “A microbial map for wine


Some background - basic microbiology that every wine drinker ought to know

Before discussing the paper, here's some context.

  • The microbe responsible for fermenting grape juice into wine is the yeast Saccharomyces cerevisiae. It’s a fungus, a eukaryote (which means that even though it’s single-celled it’s relatively closely related to us humans). The most notable thing it does in fermentation is to eat sugar and excrete alcohol, but while it’s in the wine it carries out other chemical processes that affect the flavor of the wine in subtle ways.
  • Yeast are found on grapes in the vineyard, to some extent. In the old days, winemakers used to harvest grapes, crush them, and then let the native yeast do the work. However, that can lead to unpredictable results, so nowadays it’s common practice to add cultured yeast so you know what you’re getting.  A lot of great winemakers still don’t do this, but a lot of great winemakers do, and I think almost all mediocre winemakers do. Winemakers can purchase “domesticated” yeast with all sorts of cool different wine-making properties. You can do some Saccharomyces window-shopping yourself at White Labs, Lallemand, or Wyeast.
  • There are a few bacteria that are known to affect wine, as well—for instance, some (the most famous, Oenococcus oeni) that carry out malolactic fermentation, adding buttery flavor and reducing the wine’s acidity; or Brettanomyces species, which furnish a controversial “funky” flavor. Sometimes winemakers purchase and add Oenococcus, or do things to encourage the naturally-occurring bugs to work, but for the most part bacteria take a backseat to yeast in winemaking.
  • In the vineyard, grapes have a veritable zoo of bacteria and fungi (including yeast that are not Saccharomyces) all over them. These inevitably make it into the first pressing, and once they have access to the sugar from inside the grapes they go to town, growing and dividing and also doing chemistry that can affect the flavor of the wine (Jolly 2013). However, the period of their influence is very short; either the winemaker adds enough yeast to crowd out the natural microbes, or else a few species of the natural yeast are usually able to outcompete everyone else, so wine fermentation quickly becomes dominated mostly by Saccharomyces cerevisiae. (Notably, winemakers do NOT usually pasteurize the grape must to kill the native microbes on purpose, the way many cheesemakers pasteurize their milk before adding the bacterial cultures they want.)

So there is definitely reason to believe that local wild native microbes could contribute a certain je-ne-sais-quoi to the character of a wine, and that that character could be tied to the local terroir through the microbes.


What this study did

In order to test that hypothesis, here is what Bokulich and colleagues did, in summary:

  • They sampled grape must, the crushed destemmed berries right before fermentation begins. This way they were sure to sample all the microbes that managed to make it from the vineyard into the winery, before those populations started changing over the course of fermentation.
  • In 2010, to examine the effects of geography and varietal, they collected 235 samples, representing eight different wineries from four CA wine regions-- Napa, Sonoma, San Joaquin Valley, and Central Coast. Among these samples were musts from Chardonnay, some from Zinfandel, and some from Cabernet. (Unfortunately, the names of the wineries are not published.) In 2012 they went back to Napa to collect 39 more samples, to compare with the 2010 samples to measure the effect of vintage.
  • They extracted DNA from all these samples, then amplified and sequenced the DNA. In this process, they didn’t actually sequence the whole genomes of the microbes. Instead, they sequenced targeted genetic markers for bacteria and for fungi—markers that are both universal and varied enough that the sequence alone can be used to identify the organism it came from. (For those in the know: for bacteria they used the 16S V4 region, and for fungi the ITS region.) They obtained a total of about 5 million bacterial sequences and 3 million fungal sequences—which, their data indicate, pretty much covers all of the bacterial and fungal diversity that was there.
  • They mapped these sequences to databases of known bacterial and fungal sequences, so that they could identify what taxa (species/genus/family etc.) were present.
  • Then they did a whole lot of elaborate statistics—which is what one needs to do in order to make sense of 8 million sequences. These provided the content for the paper's complex technicolor figures, which are definitely worth a look if you care to spend some time thinking about what diversity metrics measure. But in short, the questions they sought to ask were:
             - How similar is each sample to every other sample?
             - Precisely what is similar and different about the microbial communities in each sample—that is, which organisms are more or less common in each sample? 
             - Do we see geographic trends? Temporal trends?
             - Are there particular environmental factors that might be responsible for the differences in the microbial communities?


What they found

  • Region makes a big difference. Microbial communities from different regions were significantly different, even when all varietals and all vintages were combined. When comparing only grapes from a certain varietal, the effect of environment was even stronger. This is true for both bacteria and fungi.
            Cabernet Sauvignon communities seem to follow a north-south axis: that is, Napa/Sonoma microbes are very different from the southernmost Central Coast microbes, and the northern Central Coast microbes fall somewhere in between. This is not true for Chardonnay. Go figure.
  • Varietal also makes a difference, though less so. Fungal communities seem to be more varietal-specific than bacterial communities are. However, if you compare grapes from within the same region, the effect of varietal is significant for both bacteria and fungi.
             This could be due to biological differences between varietals (“host-microbial interactions on the grape surface”). The authors give the example of Zinfandel, which has thin skins and a dense growth habit and therefore suffers berry breakage more often. This could be the reason it hosts more Gluconobacter and Lactobacillales, as well as more fermentative yeasts—these are all organisms commonly found on damaged fruits.
             On the other hand, different grape varietals are often grown differently, for instance with different trellising styles, creating microclimates which may be another reason for the difference in microbial population.
  • The study found several environmental factors which might be responsible for the differences in microbial communities, though it was difficult to identify any that show really strong effects. On top of which, many of the factors covary. The most convincing factors include relative humidity and total precipitation, each of which seems to influence a specific suite of bacteria and fungi. Overall, the authors report “noisy but significant trends.”
  • Vintage has only a minor effect on microbial populations, and only on very small spatial scales. That is, within a given vineyard, microbes do vary from year to year, but that variability is very small compared to the variability between vineyards. And the effect is only really in the bacteria; fungal populations are much steadier.


Why is this important?

So, there probably really is such a thing as regional microbial terroir! That’s awfully cool, in and of itself-- especially for those of us who care about microbial ecology and hope to convince non-microbiologists to feel the same way. But in addition to the cool factor, the authors mention several potential applications:

  1. If we can get a better handle on how environmental factors influence microbial populations, we may be able to predict how microbial terroir will change with global climate change.
  2. If we can also figure out which microbes are associated with specific desirable or undesirable qualities in wine, growers may be able to tweak the environmental conditions in their vineyard microclimates to be favorable to the right bugs.
  3. It might be possible to develop a genetic fingerprinting technique that will help determine where a certain batch of grapes was grown, based on the microbes present on them. (This wasn’t mentioned in the paper but I remember having heard one of the authors mention it in a conference talk.)

I have to admit I believe we’re quite far from being able to implement 1) or 2)—notice that they require a deep understanding of the environment-microbe and microbe-flavor relationships, which we're still working on. And 3) sounds pretty fun, but will come about only if winemakers have a sizeable economic problem with grape provenance falsification (do they?), and if sequencing technology becomes cheap enough to make this method of analysis worthwhile (which it probably will, eventually).


A few other things to consider

Other than the broad conclusion that wine microbes care about geography (notable as that finding is), the results of this study are not, unfortunately, applicable to the rest of the wine-growing world. There’s not really much way to predict what populations you’ll find in French or Spanish vineyards until we sequence them. So I guess we’d better get started on that!

This study, like so many in microbial ecology, looks not at whole organisms but rather at just one genetic marker that is then interpreted to represent the identity of the whole organism. For now, this is a necessity—the discovery that we can even do this is what opened the door to modern environmental microbiology. But it’s worth acknowledging that that one marker gives us only a general sense of what that microbe can do. I'll bet that a lot of important traits with oenological consequences can be transferred among microbes in a modular way (by horizontal gene transfer) and we'll probably only really understand what these organisms can do to wine on if we can sequence their enrie genomes, or grow them in the lab.

By sampling the grape must, rather than the grapes in the vineyard, this study captures the microbes that make it into the wine after transport and crushing. I do agree that that’s the population that will have the greatest direct influence on the wine. However, note that it also integrates any bugs that were picked up in the harvesting bins into which the grapes were dropped, in the truck in which they were transported, on the people who handled them, and on the equipment by which they were crushed. So although the vineyard signal may still be strong, there’s no guarantee that that’s the only signal we’re seeing. The other things also contribute to terroir.

And what about all the microbes in the soil and on the plant that affect the way it grows, and therefore influence the quality of the grapes, even if they don’t make it into fermentation? This commentary by Jack Gilbert and colleagues, in the same issue of PNAS as the Bokulich et al. study, argues that they may be even more important.
For more reading on the microbes that can be found in the vineyard, see a recent paper by Setati (2012).

There’s so much more to study. Thank goodness for all the microbial ecology PhD students who are willing take up the challenge!




Bokulich, Nicholas A., John H. Thorngate, Paul M. Richardson, and David A. Mills. 2014. “Microbial Biogeography of Wine Grapes Is Conditioned by Cultivar, Vintage, and Climate.Proceedings of the National Academy of Sciences 111 (1): E139–E148.

Gilbert, Jack A., Daniel van der Lelie, and Iratxe Zarraonaindia. 2014. “Microbial Terroir for Wine Grapes.” Proceedings of the National Academy of Sciences 111 (1): 5–6.

Gougeon, Régis D., Marianna Lucio, Moritz Frommberger, Dominique Peyron, David Chassagne, Hervé Alexandre, François Feuillat, et al. 2009. “The Chemodiversity of Wines Can Reveal a Metabologeography Expression of Cooperage Oak Wood.” Proceedings of the National Academy of Sciences 106 (23): 9174–79.

Heymann, Hildegarde, and Dario Cantu. 2013. “Systems Biology of Vintage and Terroir: Adding Some Flavor to the Wine Grape Transcriptome.” Flavour 2 (1): 1–2.

Jolly, Neil P., Cristian Varela, and Isak S. Pretorius. 2013. “Not Your Ordinary Yeast: Non-Saccharomyces Yeasts in Wine Production Uncovered.” FEMS Yeast Research 14 (2): 215–37.

Setati, Mathabatha Evodia, Daniel Jacobson, Ursula-Claire Andong, and Florian Bauer. 2012. “The Vineyard Yeast Microbiome, a Mixed Model Microbial Map.” PLoS ONE 7 (12): e52609.

White, Michael A., Philip Whalen, and Gregory V. Jones. 2009. “Land and Wine.” Nature Geoscience 2 (2): 82–84.