Microbes live in dense and diverse communities. There are billions of bacteria from thousands of species living together in your gut or in the soil. Sequencing the total DNA of these communities can give us a catalog of the diversity that's there, but it can't tell us much about the relationship between those organisms, how the communities form, and how they evolve together. In a fascinating new paper, my friends Rachel Dutton and Ben Wolfe, along with their colleagues Julie Button and Marcela Santarelli from the Harvard FAS Center for Systems Biology, propose a unique experimental "organism" for studying the formation of microbial ecosystems: cheese.

Perhaps it's an understatement to say that I have a thing for cheese. I've been a huge fan of Ben and Rachel's work for years, and Ben was a star in the video I made last year about cheese microbes. Their new paper is awesome, summarizing their analysis of the bacteria and fungi living in 137 cheeses from 10 different countries, with many surprising results and important implications for how we might study microbial communities in the future.

Compared to other microbe-rich environments, these cheeses contain a relatively small number of microbial types. There were only 14 genera (the step above species in the classification hierarchy) of bacteria and 10 of fungi present in the cheese at more than 1% of the total population. On average, each individual cheese had only 6.5 bacterial and 3.2 fungal genera present in the rind. With such small numbers, you might think that these were just what the cheesemakers added as starter cultures, but 60% of the bacteria and 25% of the fungi weren't added on purpose.

Where did these other strains come from? There are a few ways besides starter cultures that microorganisms can enter the "microbial superhighway" of the cheese. Raw milk has plenty of microbes in it, as does the environment where the cheeses are aged, and many of the unexpected strains likely attached to the rind as the cheeses aged. After cheese has curdled but before it's aged, cheesemakers also add salt to enhance the flavors and slow down the growth of bacteria. Several of the strains found in the cheeses are salt-tolerant ocean bacteria, which likely hitched a ride in the sea salt that the cheesemakers added at that step.

Different types of cheeses have different kinds of bacteria and fungi that contribute to the unique flavors of each cheese. Comparing the populations of bacteria in each of the cheeses, Ben and his colleagues could identify three main clusters of microbial types, which correlated with the three main types of rind -- bloomy rind cheese made with fuzzy white mold like Brie and Camembert, drier natural rind cheese, and stinky washed rind cheese with a slimy coating of orange-colored bacteria. These same clusters also showed up when looking at microbial genes rather than simply types: genes coding for enzymes that make compounds with "putrid" or "sweaty" aromas, for examples, were highly correlated with the washed-rind cheeses.

Perhaps surprisingly, there was no correlation between the population of microbes in the cheese and the geographical location where the cheese was made. This doesn't mean that microbial terroir doesn't exist (more geographical correlations might have popped up if they looked at the species rather than genus level), but that the environment that the cheesemakers create during the aging process is much more important in deciding the eventual shape of the microbial community. Rachel says that when it comes to microbial communities, "if you build it, they will come."

So that's what they did: because there were only a small number of strains and all of those strains could grow on cheese curd, the team could build cheese "in vitro" and see what bacteria showed up. They created thousands of tiny cheeses in 96-well dishes and added the 6 bacteria and 5 fungi that most commonly appeared in their analysis. To recreate the environment of each kind of cheese, they either made drier curd (natural rind), added Galactomyces fungus (bloomy rind), or washed the curd in salt solution (washed rind). At the end of the experiment, even though all the wells started with the same population of bacteria and fungi, each type of environment -- natural, bloomy, or washed -- had a population that more closely matched the rind type of regular "in situ" cheese. With the in vitro cheese model, they could also track the changes in the microbial populations over time, observe how changing different environmental conditions like pH affected the population of bacteria, and track interactions between different pairs of bacteria and fungi.

We are only beginning to understand how microbial communities form and how they function. Cheese is an amazing microbial ecosystem, created by the interaction of microbes that have been domesticated over thousands of years and "wild" microbes from the farm, the sea, and the aging cave. They are "seminatural" ecosystems shaped by the environments that cheesemakers build to nurture particular flavors and textures. Now they are also experimental ecosystems, where we can learn how to design and shape new microbial relationships.

To learn more you can check out the "video abstract" from Cell:

and check out the paper:

Cheese Rind Communities Provide Tractable Systems for In Situ and In Vitro Studies of Microbial Diversity.

Wolfe, Button, Santarelli, and Dutton. Cell: 158(2), 422-433.