Cheese as a model microbial ecosystem – Rachel Dutton (Harvard)

IN TRANSCRIPT
00:00:07.00 My name is Rachel Dutton and
00:00:09.01 I am a Bauer fellow at the Harvard University
00:00:12.06 Center for Systems Biology and
00:00:14.23 I am going to be talking today about
00:00:16.17 work in my lab using cheese as a model
00:00:19.28 system for studying microbial communities.
00:00:26.09 I’d like to start with a question: why should we
00:00:29.15 even study microbial communities?
00:00:31.29 Fact is that most microbes in nature
00:00:35.02 exist as part of communities. However we
00:00:38.19 know incredibly little about what is actually
00:00:40.22 happening within these communities of microbes.
00:00:44.07 So often as a microbiologist, what we’ll
00:00:46.07 study is an organism in isolation in the lab.
00:00:50.13 For example E.coli, growing as a
00:00:53.01 population of cells, individually.
00:00:57.00 However in the real world what we have are
00:01:01.03 these communities of different species living together.
00:01:04.12 And so that is not to say that one of
00:01:06.16 these is necessarily a better
00:01:07.28 way to study microbes;
00:01:09.02 just that we can get at different
00:01:10.21 types of question about biology
00:01:12.25 under these two different
00:01:14.11 types of systems.
00:01:17.16 In the lab, by studying E.coli
00:01:19.18 over many decades we have a
00:01:22.08 really fundamental understanding
00:01:23.28 of how cells work, how they
00:01:25.26 respond to the environment,
00:01:26.24 how they use nutrients, how they grow.
00:01:29.01 However if we are able to
00:01:30.21 study microbes within
00:01:31.22 the context of community,
00:01:33.14 we might be able to get some
00:01:34.22 questions that will confer very
00:01:37.16 interesting biology such as how often
00:01:40.17 do species interact with each other,
00:01:42.24 what are the mechanisms they
00:01:43.28 use to do so and are there any
00:01:45.13 general principles or mechanisms
00:01:47.27 of community formation.
00:01:50.04 These microbial communities
00:01:54.03 are pretty much everywhere on the planet.
00:01:56.13 Whether we are thinking about the soil,
00:01:58.15 in which you have billions of
00:02:00.22 organisms in every gram or
00:02:03.03 the human body, so
00:02:03.29 for example in the gut we have
00:02:06.12 again billions of organisms
00:02:09.18 in every single one of us.
00:02:13.15 Now the challenge with
00:02:14.22 understanding how these
00:02:16.09 communities work is that these
00:02:18.17 communities are often incredibly complex.
00:02:20.20 So in the two systems for example
00:02:23.02 we have hundreds or thousands
00:02:25.07 of different species all living together.
00:02:27.25 If we want to be able to understand
00:02:30.24 how these communities work,
00:02:32.08 we might be able to take an approach
00:02:34.02 similar to be using E.coli as a model
00:02:36.03 organism where use model
00:02:38.14 communities or model ecosystems.
00:02:40.14 So, can we actually come up with
00:02:41.29 model ecosystems that will help us
00:02:43.17 understand how these complex
00:02:46.09 multi-species communities work?
00:02:49.15 What we ideally like to have is an
00:02:53.10 experimentally tractable ecosystem
00:02:55.07 that we can use to really link the
00:02:57.08 ecology of the system, what is
00:02:58.25 happening out in the natural
00:03:00.11 environment, to mechanisms of how things work.
00:03:03.20 So what we would like to be able to do
00:03:05.18 is have a microbial ecosystem where we can
00:03:08.14 study the microbial ecology in the
00:03:11.23 native setting and this will allow us to
00:03:13.19 generate hypotheses about how this
00:03:15.23 community functions and what it does.
00:03:18.09 Then what we would like to be able
00:03:20.08 to do is bring the system into a lab and
00:03:22.14 be able to deconstruct it through isolation
00:03:26.06 of the organisms that are present in
00:03:27.19 the community and then ideally reconstruct
00:03:30.09 it under conditions that resemble those
00:03:32.19 found in the environment so that we can start
00:03:34.02 to test hypotheses of how these
00:03:36.26 microbes form communities and
00:03:38.03 eventually get to a mechanistic
00:03:39.21 understanding of how the system works.
00:03:44.05 What kind of communities might we
00:03:46.20 look towards for finding these sort of
00:03:48.29 experimentally tractable systems?
00:03:53.00 We’d like in an ecosystem that is
00:03:55.23 relatively simple, so instead of
00:03:57.22 hundreds or thousands of species,
00:03:59.19 maybe on the order of tens of
00:04:00.24 species that we would work with.
00:04:04.03 We would like for this community
00:04:05.17 to be very reproducible.
00:04:08.00 It would be great if it were easy
00:04:09.11 to get a hold of samples, easy to access the
00:04:11.11 community it self and ideally
00:04:13.03 something we can manipulate the
00:04:14.19 community at many different levels
00:04:16.06 whether we’re thinking of the
00:04:17.16 environment in which the community is growing,
00:04:19.26 the substrates on which it’s growing,
00:04:21.25 or the individual organisms.
00:04:25.23 Now in my lab we are using fermented
00:04:28.05 foods as our model ecosystems
00:04:31.09 and fermented foods are actually one of
00:04:33.01 the most ancient forms of
00:04:34.27 biotechnology where humans have
00:04:36.02 intentionally manipulated an environment so
00:04:37.29 that we can control the growth of
00:04:39.13 certain types of microbes.
00:04:41.06 Many of these fermented foods actually
00:04:43.08 have these wonderful reproducible microbial
00:04:46.02 communities forming during
00:04:47.02 the process of fermentation.
00:04:49.01 So we will spend the rest of the time today
00:04:50.27 talking about cheese but there’s
00:04:52.07 many other examples. We have
00:04:54.11 fermented beers, dry salamis and
00:04:58.19 sausages, fermented vegetables such
00:05:02.03 as kimchi, which is fermented
00:05:04.09 cabbage. Chocolate and coffee are
00:05:07.14 both great examples of fermented foods.
00:05:11.03 Things like pickles, natto, which is a
00:05:14.16 really interesting Japanese fermentation,
00:05:16.16 and things like soy sauce.
00:05:18.01 Pretty much every culture on the
00:05:20.24 planet has come up with many
00:05:22.29 different examples of fermented
00:05:24.01 foods and all these fermentations
00:05:25.26 involve the metabolism of different
00:05:29.02 foods by microbial communities.
00:05:33.03 In the case of cheese, there are two
00:05:36.05 really important steps in cheese
00:05:37.17 making in which the microbial
00:05:39.13 communities are important. We have
00:05:43.29 microbes actually responsible
00:05:45.09 for the complete transformation
00:05:47.12 of milk into cheese
00:05:49.06 through a couple different steps.
00:05:50.17 We start with milk
00:05:52.29 which will come from some dairy
00:05:55.26 animal, whether it is a cow or goat or
00:05:58.11 sheep, and this milk, from the
00:06:00.19 perspective of the microbe,
00:06:01.25 really is just a great of food.
00:06:05.06 We drink milk because it is
00:06:06.19 nutritious but microbes also think
00:06:08.28 it is very nutritious so
00:06:09.29 milk is a great source of
00:06:11.21 protein in form of caseins,
00:06:13.08 sugar in the form of lactose and
00:06:17.29 fat in the form of different kinds triglycerides.
00:06:21.16 The way that this simple,
00:06:25.11 very common food source is
00:06:29.17 transformed into cheese is first
00:06:30.29 through the action of lactic acid
00:06:32.26 bacteria and the enzyme rennet.
00:06:34.12 Lactic acid bacteria are can
00:06:37.24 be commonly found in
00:06:39.07 many different environments
00:06:40.27 and they’re also found in milk,
00:06:45.00 or they can be added to
00:06:45.23 the milk to make cheese.
00:06:49.23 These lactic acid bacteria,
00:06:51.27 they are also known as
00:06:52.20 starter cultures, are really
00:06:54.05 what help turn the milk into
00:06:55.22 the curds, which is the first
00:06:57.01 stage of cheese making.
00:06:59.04 They do this through the
00:06:59.17 fermentation of the
00:07:00.21 milk sugar, lactose.
00:07:03.16 If we have lactose,
00:07:05.04 these bacteria have the
00:07:06.12 enzyme which is able to
00:07:08.24 break down lactose, which
00:07:10.14 is a disaccharide, into the simple sugars
00:07:12.22 and then take these simple
00:07:15.05 sugars and ferment
00:07:16.01 them for energy and this
00:07:18.00 results in the production of lactic acid.
00:07:20.14 Lactic acid serves many
00:07:22.29 different purposes. In
00:07:24.29 cheese making as well in
00:07:25.19 many different fermented foods,
00:07:27.28 some of these purposes
00:07:29.14 are the flavor, so it
00:07:31.03 gives the sort of
00:07:31.18 tartness to the food.
00:07:35.15 It helps in preservation – so the
00:07:36.29 acidity inhibits the growth
00:07:38.13 of many pathogens
00:07:39.04 that might otherwise be
00:07:40.15 a problem in food spoilage.
00:07:42.22 And it helps the protein,
00:07:45.00 or the caseins, in the milk
00:07:46.08 start to come together
00:07:47.24 into aggregates which
00:07:48.19 are called curds.
00:07:49.02 Now if you were to just use
00:07:52.14 lactic acid bacteria
00:07:53.29 during the production
00:07:55.17 or for the fermentation of milk, what
00:07:57.10 you would end up with is yogurt.
00:07:58.08 Yogurt is a very simple
00:08:00.11 fermented product which
00:08:02.19 just uses the action of
00:08:03.19 these lactic acid bacteria
00:08:05.12 to coagulate the milk into this gel
00:08:08.04 of yogurt.
00:08:08.23 With cheese making, the enzyme
00:08:10.13 rennet helps stimulate the
00:08:12.05 further coagulation of milk
00:08:14.02 proteins into curds so
00:08:16.02 that you can separate the
00:08:17.13 whey from the curd and that
00:08:19.03 curd is what becomes the
00:08:20.10 starting point for many
00:08:21.19 different kinds of cheese.
00:08:23.16 The second stage,
00:08:25.19 which I think is the most
00:08:27.01 interesting part of
00:08:27.26 cheese making, is the
00:08:29.19 secondary stage, which is
00:08:32.01 aging of the cheese.
00:08:33.21 Initially you have lactic
00:08:34.23 acid bacteria growing,
00:08:35.23 you have the fresh
00:08:36.11 cheese formed which
00:08:37.28 basically just tastes a little bit like sour
00:08:40.20 milk; something very fresh and tart.
00:08:42.19 You can take this cheese now and age it
00:08:46.11 and depending on how you age it
00:08:48.28 you get many different kinds
00:08:50.27 of aged cheeses which
00:08:52.29 have different aromas, flavors,
00:08:54.28 textures, appearances and the key
00:08:58.27 factor on determining
00:09:00.12 what kind of aged cheese
00:09:01.26 you end up with are
00:09:03.02 the microbes that
00:09:04.03 are colonizing surface of the cheese.
00:09:06.07 We have different bacteria
00:09:08.02 and fungi, many of
00:09:09.06 which are coming in
00:09:09.26 from the environment.
00:09:10.26 They are not inoculated,
00:09:12.19 yet they form these
00:09:14.00 reproducible communities
00:09:15.16 on the surface of the cheese.
00:09:18.11 So this rind forms
00:09:22.14 during the aging process
00:09:23.25 and this rind is actually a
00:09:25.26 surface associated microbial community.
00:09:28.10 So if you look at this picture
00:09:30.24 here: this is a close up of a
00:09:33.04 cross section of a slice of cheese
00:09:35.00 and this top layer is the
00:09:37.10 rind, or this biofilm,
00:09:39.29 that is formed during the aging
00:09:41.06 process. And the first time I
00:09:44.01 saw a cheese up close like this
00:09:46.19 it reminded me of when I
00:09:49.02 was a student in the microbial diversity
00:09:51.16 course at Woods Hole, at the MBL and we
00:09:55.00 had spent a lot of time out
00:09:56.29 in the salt marshes looking
00:09:57.29 at microbial mats and it reminded me
00:10:02.03 exactly of a microbial mat.
00:10:03.24 You have this beautiful
00:10:06.12 community that is
00:10:08.28 forming on a surface.
00:10:11.00 And so I say this to make the point that
00:10:13.28 the communities we are studying
00:10:15.17 on the rind of cheese are not really
00:10:18.01 unique to cheese. They are
00:10:20.02 an example of a type of community
00:10:22.07 that you find in many different
00:10:24.04 environments. In the case of microbial mats,
00:10:26.03 we have fossil evidence suggesting that
00:10:28.04 this type of microbial community
00:10:29.20 can date back at least
00:10:30.29 3.2 billion years.
00:10:36.14 These rind communities
00:10:37.24 can take many different
00:10:38.21 forms. Here we have pictures of
00:10:41.09 many different kinds of aged cheeses.
00:10:44.02 There are hundreds if not
00:10:45.17 thousands of varieties of cheese
00:10:47.18 and we wanted to get an idea of
00:10:52.09 what type of diversity there
00:10:53.28 is within these communities.
00:10:55.24 When I started this project, we would
00:10:59.17 just do simple things like,
00:11:01.00 culture from the surface
00:11:03.24 of the cheese just to get an
00:11:05.01 idea of what types of microbes
00:11:06.20 we might find in this environment.
00:11:08.26 These are just some pictures that I took from
00:11:12.00 cultures that we grew in the
00:11:13.28 lab just on standard media
00:11:15.18 on petri dishes and these
00:11:17.10 are just close up shots
00:11:18.22 of some of the gorgeous
00:11:20.21 organisms that we have growing
00:11:22.06 together in these communities.
00:11:23.18 In this particular one,
00:11:24.27 we have filamentous fungi, or molds,
00:11:27.15 growing with what appeared to
00:11:30.05 be bacterial cells, bacterial colonies.
00:11:32.25 Here’s another where
00:11:34.21 we have a mixture of
00:11:36.16 different types of microbes
00:11:37.28 growing together. These are
00:11:39.29 all isolated from different types
00:11:41.12 of cheeses.
00:11:43.21 Another here where we have another mold
00:11:46.07 growing with other bacteria.
00:11:49.02 And again, just a really
00:11:52.24 wonderful microbial communities that
00:11:55.01 we have isolated from these cheeses.
00:11:58.15 Now this is looking
00:11:59.29 at colonies we have isolate from the
00:12:01.29 surface of a cheese but
00:12:03.04 what happens if we look
00:12:04.07 directly at these communities.
00:12:06.10 What does that look like?
00:12:07.21 We have done some
00:12:08.25 scanning electron microscopy
00:12:10.10 to visualize the communities
00:12:13.05 directly in situ and
00:12:15.25 this is looking top down
00:12:17.19 at the surface of a cheese.
00:12:20.05 This is the rind of the cheese
00:12:21.10 and conveniently a little chunk
00:12:23.29 has sort of been removed from the
00:12:26.13 top of this cheese so
00:12:27.20 we can start to see the
00:12:29.02 inside of this community
00:12:32.05 and if we go down deeper into the rind,
00:12:35.06 if you sort of imagine that you shrink
00:12:38.10 yourself down to the size of the bacterium,
00:12:41.17 and you are standing inside the rind of a cheese
00:12:43.27 and you look around,
00:12:45.12 this is what you would see.
00:12:46.28 So we see this incredible
00:12:49.29 collection of different microbial cells.
00:12:53.08 We have small cells of
00:12:55.10 different shapes which
00:12:56.23 look like they are bacterial
00:12:58.10 cells, different types of bacteria.
00:13:00.20 We have fungal cells.
00:13:02.01 So these look like they
00:13:03.23 could be spores from fungi and this
00:13:06.28 is an incredibly densely packed
00:13:08.18 environment so we estimate that
00:13:12.09 we have about 10 billion
00:13:13.21 cells for every gram of cheese
00:13:16.03 rind that you would eat.
00:13:18.19 What we would like to do is
00:13:21.02 use this wonderful fermented food
00:13:24.11 to try and dissect the
00:13:25.10 formation of microbial communities.
00:13:26.20 Could we actually use cheese
00:13:29.05 as a model ecosystem;
00:13:31.12 something where we can study the diversity
00:13:35.09 ecology of the system in its natural state.
00:13:37.11 We could potentially dissect it in the lab
00:13:40.16 and then recreate conditions
00:13:41.24 so we can start forming the
00:13:43.26 communities and understanding how they work.
00:13:48.14 And so the first step we had to
00:13:50.08 do is actually characterize
00:13:51.27 what types of microbes are present.
00:13:53.10 Despite the fact that cheese
00:13:55.16 has been made for thousands
00:13:57.14 of years, we don’t really
00:13:59.03 have a very complete understanding of
00:14:02.02 what types of microbes are present on cheese,
00:14:04.12 as well as on many of these
00:14:05.06 other fermented foods. So
00:14:08.00 we took a approach to
00:14:09.26 look at the organisms
00:14:11.17 using DNA sequencing technologies.
00:14:16.02 And I though I would go through,
00:14:17.20 very quickly, how we actually
00:14:20.19 approach this question when we
00:14:22.23 have an environmental sample.
00:14:24.13 How do we actually know which
00:14:26.01 species are present in this environment
00:14:27.27 when we can’t necessary
00:14:29.10 distinguish all the organisms by eye?
00:14:33.10 So the first thing we do is go out
00:14:34.29 to the environment and collect samples.
00:14:36.26 So in our case, we are collecting samples
00:14:38.18 from cheese. We collect
00:14:40.05 the surface of cheese,
00:14:41.17 or the rind, and then we
00:14:43.19 bring it back to the lab.
00:14:45.13 Once we get the sample back to the lab
00:14:48.14 we remove the DNA from
00:14:51.24 the rest of the sample,
00:14:52.12 and that is what we’re
00:14:53.09 going after, is the DNA, so the
00:14:54.25 genetic information, from each of these cells.
00:14:57.06 Every single microbial cell has its genome.
00:15:00.06 We extract the those genomes from the sample.
00:15:04.17 Now with this mixture of genomes
00:15:06.14 what we can do is use a method called PCR
00:15:10.26 to just target regions that
00:15:14.02 are somewhat like a fingerprint
00:15:15.24 for a microbe. Every different microbe
00:15:17.25 has a different sequence.
00:15:19.25 Based on what species it is
00:15:22.12 and by looking at this region, we can
00:15:24.15 identify what type of species are present.
00:15:27.28 Once we have amplified these samples using PCR
00:15:30.26 what we can do is sequence
00:15:34.00 each of these regions from
00:15:35.26 the different species.
00:15:37.13 We can use this next generation sequencing
00:15:41.22 to actually read out all
00:15:44.00 the sequences that present
00:15:45.11 in any sample that we have.
00:15:48.11 And then to be able to determine
00:15:49.17 what species are actually present,
00:15:51.21 we then compare the sequences,
00:15:53.26 or these fingerprints, to databases.
00:15:56.19 We can use several different
00:15:59.05 types of databases which
00:16:00.13 have collections of microbial
00:16:02.18 sequences that we use as references.
00:16:05.15 This gives us a good idea
00:16:07.11 of the amount of diversity and
00:16:10.23 the particular organisms that
00:16:12.13 we have in each of our samples.
00:16:14.11 Using this method, we
00:16:16.16 wanted to look as broadly as possible
00:16:18.26 to see what type of diversity is present in
00:16:21.16 cheeses around the world.
00:16:23.26 To sample the microbial diversity of cheese,
00:16:26.17 what we did was collect
00:16:28.22 samples from 10 different countries,
00:16:30.24 137 different types of cheese,
00:16:34.09 and 362 wheels of cheese.
00:16:38.14 We took each of these samples through the
00:16:42.16 protocol that I just told you about
00:16:44.15 to look at the types of bacteria and fungi
00:16:48.18 that are present in each of these communities.
00:16:51.26 This is a representation of
00:16:54.17 the data that we have from this study.
00:16:58.00 We have on the top a
00:17:00.21 set of columns – the different
00:17:03.27 abundances of bacteria in each of the samples
00:17:06.15 and on the bottom set of columns,
00:17:07.18 the different abundances of fungi.
00:17:10.09 In this graph, each of the
00:17:12.25 columns is a different type of cheese.
00:17:16.13 We took this data, we clustered them
00:17:20.09 based on the similarity of species within
00:17:23.14 each of the communities, and you can start to see
00:17:25.08 some of these similar
00:17:26.29 clusters of cheeses that
00:17:28.15 have related communities
00:17:29.25 on this phylogenetic tree.
00:17:32.22 Now, when we started
00:17:33.29 looking at what organisms
00:17:36.08 were actually present
00:17:37.12 in these cheeses, we had a few surprises.
00:17:41.06 Now if we start to zoom
00:17:42.02 in on the bacteria that we
00:17:44.00 found in these samples,
00:17:45.10 what we see are 14
00:17:48.00 dominant genera of bacteria.
00:17:50.00 These are all of the genera that we find
00:17:52.09 greater than 1% abundance
00:17:54.09 across our dataset,
00:17:56.21 and while some of
00:17:58.03 these organisms had been studied
00:17:59.21 in various fermented foods before,
00:18:01.24 such as Staphylococcus
00:18:03.01 and Brevibacterium, we had
00:18:05.23 a couple organisms that
00:18:07.01 were new to fermented foods,
00:18:08.27 had never been described
00:18:09.24 in cheese or any other
00:18:10.26 fermented food before
00:18:11.28 such as Nocardiopsis and Yaniella.
00:18:14.16 We don’t really know what they
00:18:16.22 are contributing to the cheese environment
00:18:19.11 but they can be very
00:18:21.17 abundant in certain samples.
00:18:23.07 And the other surprising
00:18:24.20 thing about the bacterial samples
00:18:26.14 is that we actually have a lot of bacteria
00:18:28.09 that you would normally associate with
00:18:29.28 the marine environment.
00:18:31.14 So we have many sequences that
00:18:35.08 belong to genera Halomonas, Psychobacter,
00:18:39.18 Pseudoalteromonas, and Vibrio
00:18:41.27 which are very often associated
00:18:43.29 with marine environments and so
00:18:45.28 we are very interested in
00:18:46.22 understanding where these microbes are
00:18:48.08 coming from and what they might be
00:18:49.13 doing in this particular community.
00:18:52.06 If we look at the fungal
00:18:54.27 portion of the diversity,
00:18:57.10 we see a slightly smaller set of organisms:
00:19:03.10 ten dominant genera that we find
00:19:05.06 over and over again in different cheeses.
00:19:08.18 Again, some of these have been studied before:
00:19:11.19 Debaryomyces and Galactomyces for example but then
00:19:14.19 we have this wonderful collection of
00:19:16.07 other fungi which we know very little about.
00:19:18.22 For example, Scopulariopsis,
00:19:20.25 which you can see in brown,
00:19:22.19 is very abundant in many of
00:19:26.12 the samples and we know
00:19:28.14 virtually nothing about this organism,
00:19:30.20 yet we think it has some really
00:19:33.04 interesting roles in terms interactions
00:19:34.27 which I’ll show you in a moment.
00:19:39.08 This is looking at the
00:19:41.08 cheese at its sort of final stage,
00:19:43.06 right before it would be ready to eat.
00:19:44.28 These are the communities, after
00:19:46.13 they have already been established.
00:19:48.17 But one of the nice things about
00:19:49.29 working with fermented foods
00:19:51.18 is that you can sample the formation
00:19:54.07 of community as it’s actually happening.
00:19:57.02 This is something that can be
00:19:58.03 really challenging in many
00:19:59.22 environments where it’s not
00:20:01.12 clear where the beginning and end of
00:20:03.14 the community formation is.
00:20:06.11 So these communities we know are very dynamic.
00:20:09.21 This is a picture from aging
00:20:12.16 cheese in a cave in Vermont
00:20:14.08 and over a 2 months aging period,
00:20:17.00 you can see this very dramatic change of
00:20:21.13 the microbial community over time
00:20:23.15 just by eye. And so we wanted
00:20:25.11 to go in and use these sequencing
00:20:26.25 methods to actually tell us what was
00:20:29.08 happening within the formation of these communities
00:20:31.26 How did the communities get to the
00:20:33.18 point that we saw in the previous study.
00:20:38.04 So we ended up sampling 3 different batches of cheese.
00:20:43.13 We sampled the cheese every week
00:20:45.08 a they were aging and what
00:20:47.07 you can see is this absolutely
00:20:49.02 beautiful and reproducible succession
00:20:51.14 of species. So this is very consistent
00:20:57.04 change over time of species within these communities.
00:21:00.22 At early timepoints, you see
00:21:02.11 that the community is dominated in the
00:21:05.01 bacterial portion of the community
00:21:06.13 by the bacterium, Staphylococcus,
00:21:08.27 the green bacterium in this graph.
00:21:11.15 And then over time,
00:21:13.05 the community shifts to one,
00:21:14.21 in which is dominated by the bacterium, Brevibacterium.
00:21:18.02 In the fungal portion of the community,
00:21:20.14 we also see a secession where the community
00:21:23.01 early on is dominated by the yeast,
00:21:25.02 Candida, in light green and then that is
00:21:28.05 followed by the filamentous fungi, or
00:21:29.29 molds, Penicillium and then
00:21:31.29 eventually Scopulariopsis.
00:21:34.12 From this type of data, we think
00:21:36.05 that this is a nice way of generating hypotheses
00:21:39.15 about which organisms might be important
00:21:41.20 in the establishment of a community
00:21:44.08 and in directing what type of organisms
00:21:47.02 eventually end up in the specific communities.
00:21:49.23 So this generates a lot of questions for us
00:21:51.15 about, for example, whether
00:21:53.15 Staphylococcus growth is required
00:21:55.28 for the later growth of the species such as
00:21:58.05 Brevibacterium or for the fungi as well
00:22:01.03 and this is something we are
00:22:01.29 actively following up within the lab.
00:22:05.15 What we have done so
00:22:08.05 far is try to get a picture of the
00:22:10.28 microbial ecology of the system.
00:22:13.00 We’ve taken these molecular, microbial ecology
00:22:17.00 and sequencing based approaches
00:22:18.17 to get a good picture of the diversity
00:22:20.19 of the microbes in these communities.
00:22:24.05 So now we have an idea of what
00:22:25.20 these communities look like,
00:22:27.15 can we actually bring them into
00:22:28.20 the lab and start to dissect them.
00:22:30.07 So are we able to pull apart the
00:22:33.14 communities into their individual species components.
00:22:38.12 What I am showing you here
00:22:43.12 are from our sequencing, the most
00:22:46.07 abundant bacteria we find in the system-
00:22:49.06 the average relative abundance
00:22:52.04 of the different bacteria that
00:22:53.25 we see across the cheeses.
00:22:55.21 We went in and actively went
00:22:59.28 after each of these organisms and
00:23:01.16 tried to culture them in the lab
00:23:03.17 and it turns out that we are able to
00:23:05.09 culture representatives of every
00:23:07.25 single one of these bacterial groups.
00:23:10.27 So we have this wonderful culture
00:23:13.01 collection of different bacteria.
00:23:14.12 So we have different Actinobacteria,
00:23:17.04 different proteobacteria,
00:23:18.29 different Bacteroidetes and different Firmicutes.
00:23:23.05 For the fungi, we did the same thing
00:23:25.28 and again we are able to culture
00:23:28.24 representatives of every single
00:23:30.08 one of the major genera
00:23:31.14 that we found in our cheese.
00:23:33.19 We now have a mixture of phylogenetically
00:23:36.22 diverse molds and yeast
00:23:38.22 that makeup the fungal
00:23:41.16 portion of these communities.
00:23:43.14 What we have been able to
00:23:46.01 do up to this point is have
00:23:47.23 an understanding of the ecology
00:23:49.01 and we’ve been able to establish a culture
00:23:52.00 collection with all of the dominant organisms.
00:23:54.09 And this can be really challenging
00:23:56.01 part of studying microbial communities
00:23:58.09 because it is often very difficult
00:23:59.26 to be able to culture these organisms
00:24:02.03 in isolation from the rest of the community.
00:24:06.06 What do we do with this culture collection?
00:24:08.28 We wanted to know: can we actually start
00:24:10.18 to reconstruct communities in the lab
00:24:12.18 so that we can start ask questions
00:24:15.06 about what the roles of these different microbes are
00:24:18.26 in community formation.
00:24:23.11 What we’ve tried to do now is rebuild cheese.
00:24:27.03 And we do this by doing what
00:24:30.18 we refer to as in vitro cheese making.
00:24:33.03 In traditional cheese making,
00:24:37.00 a cheese maker will make a cheese
00:24:39.00 place it into a cave like environment,
00:24:40.27 this cheese can be churned
00:24:43.23 or washed or treated in different ways
00:24:47.20 to end up with a certain type of cheese
00:24:50.01 and so you end up with a diversity of
00:24:52.26 cheeses depending on how treat the cheese.
00:24:54.27 Now we can do this in the lab.
00:24:57.26 So we make cheese curd based medium,
00:25:01.11 we pour this into plates and we can actually
00:25:04.03 manipulate the species that
00:25:06.14 are going into our cheese
00:25:08.03 and the way we treat the
00:25:09.14 cheese in the laboratory.
00:25:11.14 We go from our collection of species
00:25:13.05 we reconstruct communities and
00:25:15.03 follow their behavior over time.
00:25:17.29 We wanted to try and
00:25:19.06 reconstruct community formation
00:25:20.25 so what we did was a very
00:25:22.10 simple experiment where we took
00:25:24.27 the dominant microbes that we found
00:25:27.02 in the natural succession and
00:25:29.04 we mixed them together in equal numbers,
00:25:32.01 so all the bacteria and fungi, at the same time.
00:25:36.15 We basically took them together in
00:25:39.01 equal numbers, put them onto our in vitro media
00:25:42.08 and asked what happens when
00:25:44.26 you mix these organisms together.
00:25:47.29 And what we saw was, remarkably,
00:25:50.14 that succession in vitro follows a very
00:25:53.21 similar course to what we see in situ.
00:25:56.21 So this is showing the changes in
00:26:00.16 the population sizes of different species
00:26:03.00 over time. Again we started with
00:26:07.12 approximately equal numbers of
00:26:09.04 the different bacteria and fungi
00:26:10.12 and mixed them together and
00:26:11.27 what you see is that you get
00:26:14.01 similar patterns succession.
00:26:15.24 So the bacteria, initially Straphylococcus,
00:26:18.13 dominates the community that’s followed by
00:26:20.08 Brevibacterium and in the fungi,
00:26:22.28 initially the yeast, Candida,
00:26:25.07 dominates the community
00:26:26.02 and that’s then followed by filamentous
00:26:28.19 fungi, Pencillium and then Scopulariopsis.
00:26:31.04 What this tells us is that the succession
00:26:35.06 that we are seeing in these communities
00:26:37.12 is not in any way random.
00:26:41.05 It’s not dependent on the order
00:26:43.04 of arrival of the different organisms.
00:26:45.14 There is something intrinsic
00:26:46.17 about the growth of these organisms
00:26:48.12 in this particular environment
00:26:49.15 that leads to this particular
00:26:52.13 type of community formation.
00:26:53.29 And so what we are doing in the lab now
00:26:56.01 is trying to know remove species and look at
00:26:59.03 if we remove, say Straphylococcus,
00:27:01.29 what happens to the rest of the community.
00:27:03.21 And that is something we can do by having this very
00:27:06.02 experimentally tractable system.
00:27:09.24 Now one of the other things we’re
00:27:11.10 starting to look at in a lot more detail
00:27:14.11 are interactions between species.
00:27:16.23 We now from working with the
00:27:21.06 system that we have observed many
00:27:23.12 different interactions between
00:27:24.29 microbes isolated from cheese.
00:27:26.28 So we have examples of pretty
00:27:28.10 much every type of
00:27:29.24 interaction: positive and negative.
00:27:32.28 We have examples of interactions
00:27:35.03 between bacteria, between fungi and
00:27:38.15 between bacteria and fungi.
00:27:41.09 And so these are pictures of
00:27:42.26 some of the very dramatic looking interactions
00:27:46.23 where we have a stimulation of bacteria by fungi
00:27:50.26 for example or these dramatic
00:27:53.11 zones of inhibition that we see
00:27:56.13 upon growing fungi in the
00:27:58.01 presence of certain bacteria.
00:28:00.00 So we wanted to develop a way
00:28:02.07 to look at this in a more systematic
00:28:04.23 way and so what we’ve done is use
00:28:07.29 96 well plate cheese curd agar to
00:28:11.02 look at series of pairwise interactions.
00:28:14.07 What we’ve done so far is look
00:28:16.18 at interactions between bacterium
00:28:19.14 and fungi, which we know are both
00:28:22.02 important players in this system
00:28:23.16 and so we can take from our
00:28:25.01 collection of cultured organisms
00:28:27.09 and grow them in any
00:28:28.20 combination we want and
00:28:30.05 look at how they grow,
00:28:31.18 either by themselves or
00:28:33.02 in combination with another organism.
00:28:36.20 And so I’ll show you some
00:28:37.13 data from some of these
00:28:39.24 types of these interactions which has revealed to us
00:28:43.24 that interactions between species
00:28:46.00 in the system are incredibly widespread.
00:28:47.26 So what you are seeing here on the top
00:28:50.22 panel is the growth of different bacteria,
00:28:55.02 some of which are
00:28:57.11 the bacteria we find in the system
00:28:58.29 and how their growth responds
00:29:03.06 in the presences of different fungi.
00:29:05.13 So in black bars we have the
00:29:08.19 growth of bacterium alone and in the colored bars
00:29:11.11 are their growth in the presence of different fungi.
00:29:14.11 So what you can see is that the
00:29:16.22 bacteria are hugely influenced
00:29:19.03 by the presence of fungi.
00:29:20.22 So they can either influenced
00:29:22.17 in a positive way, so their growth stimulated,
00:29:24.29 or in a negative way, so their growth is inhibited.
00:29:28.23 In contrast, the fungi which you
00:29:30.19 see on the bottom panel,
00:29:32.07 don’t really seem to be impacted
00:29:36.00 by the presence of bacteria
00:29:37.22 with a few exceptions.
00:29:39.17 One very noticeable exception is
00:29:42.20 the growth of Scopulariopsis which
00:29:45.04 is inhibited very strongly by the presence
00:29:47.25 of one of the bacteria Arthrobacter.
00:29:50.02 This is one of the ways we are
00:29:52.23 starting to generate information
00:29:54.20 about the types of interactions that
00:29:57.05 we have in our system,
00:29:58.20 and that eventually what we would like to
00:30:00.03 know is what is the genetic basis
00:30:02.24 of these interactions and what’s the
00:30:05.09 chemical basis of these interactions.
00:30:09.05 What I’ve told you so far is that we’ve
00:30:13.01 been able to go in, look at the microbial
00:30:15.15 ecology, dissect the system
00:30:17.16 and now have in vitro system
00:30:20.07 for reconstructing the community
00:30:22.05 in which we see communities that form
00:30:24.04 that resemble the natural system
00:30:26.13 and that we can now start to dissect
00:30:28.11 and understand the interactions
00:30:30.26 that are at the base of this community formation.
00:30:34.21 Beyond understanding these
00:30:37.20 particular microbial communities
00:30:39.22 we think that this work will have
00:30:41.22 broad impact in the field of microbiology.
00:30:45.17 For example, we think that having
00:30:47.22 this very experimentally
00:30:49.09 tractable microbial ecosystem
00:30:50.29 we can start to get at some
00:30:52.06 of the general principles
00:30:54.02 of microbial community formation
00:30:55.17 and we think in doing so, we’ll have an impact
00:30:58.20 on advancing the conceptual, practical
00:31:01.12 and mechanistic understanding of communities
00:31:04.00 either in these simple systems
00:31:05.16 or hopefully more complex systems.
00:31:07.25 We think that these more
00:31:10.17 complex systems, which are often very
00:31:12.01 difficult to manipulate, we might
00:31:13.17 be able to gain some
00:31:14.09 insight, specifically into
00:31:15.29 the types of interactions or
00:31:17.29 mechanisms that are important
00:31:20.06 for community formation.
00:31:21.15 One of the potentially
00:31:24.03 interesting connections would be to
00:31:26.03 the human skin. We
00:31:27.03 see actually a lot of phylogenetic
00:31:28.18 overlap between the
00:31:30.00 species found on our skin
00:31:31.21 and the species on cheese.
00:31:34.12 We also think that because
00:31:36.20 they are widespread interactions,
00:31:39.10 especially between bacteria
00:31:40.20 and fungi in the system
00:31:42.08 we might be able to
00:31:43.08 do things like discover molecules
00:31:45.22 to manipulate the
00:31:46.18 growth of individual species
00:31:48.23 or manipulate the formation
00:31:50.00 of communities which could
00:31:51.06 be important for human health
00:31:53.00 or the environment or
00:31:54.13 even more applied systems.
00:31:56.20 So I would like to thank
00:31:58.05 a number of people for
00:31:59.14 helping with the data that I presented in this
00:32:03.29 seminar. I have a great lab
00:32:08.03 at Harvard. 2 wonderful postdocs,
00:32:11.02 Benjamin Wolfe and Julie Button
00:32:12.25 who’ve been involved in every aspect of this work.
00:32:15.16 I want to thank the Bauer Fellows
00:32:18.24 Program in the FAS Center
00:32:20.27 for System Biology at Harvard,
00:32:22.08 for supporting my lab
00:32:23.19 and funding from the NIGMS
00:32:26.06 Centers for Systems Biology.
00:32:28.15 I’d like to thank Marcela Santarelli
00:32:30.16 who is was visiting
00:32:31.18 graduate student who helped with
00:32:33.00 some of this work as well as
00:32:34.18 Jasper Hill Farms and Cellars
00:32:36.02 for collaboration and providing
00:32:38.18 us with many samples and
00:32:39.29 access to different cheeses.
00:32:41.03 Thank you.