What the krill microbiome can tell us about Southern Ocean ecosystems

Thursday 7 June 2018, 11:30am–12:30pm

This weeks seminar is presented by Dr Laurence Clarke (ACE-CRC). Laurence is a molecular ecologist who uses genetic approaches to examine threats to biodiversity, such as invasive species, habitat fragmentation and climate change. In this presentation, Laurence informs us about how microbes can be used to determine the health of krill and the Southern Ocean ecosystem.


Research into the bacteria that live inside and on our bodies (the human microbiome) has provided amazing insights into how our resident bacteria influence our lives. Bacteria associated with super-abundant Antarctic krill likely make a major contribution to marine bacterial diversity in the Southern Ocean, but this has been largely overlooked to date.

We used high-throughput DNA sequencing to characterise the bacteria in seawater and krill tissue samples from the Kerguelen Axis and Totten voyages, with samples from swarms separated by over 3000 km.

We found krill-associated bacterial communities are distinct from the surrounding seawater, with different communities inhabiting the carapace and digestive tract. What’s more, surface (carapace) bacterial communities were distinct between swarms, providing a biological marker of swarm fidelity. In addition, many individuals were found to harbour a potential symbiont in their digestive tracts known to help other crustaceans survive on low quality diets, which could provide a selective advantage over winter for krill possessing these bacteria.

Much like the human microbiome, the krill microbiome is likely to have important implications for both krill health and Southern Ocean ecosystems that is only now starting to be revealed.

The presentation will take place in the AAD theatrette. All welcome!

AAD Seminar Team

What the krill microbiome can tell us about Southern Ocean ecosystems

Video transcript

Bruce: Thank you everybody for coming today! I’ll just do an introduction of today’s speaker while we wait for the last people to show up. Today’s speaker – the third speaker in the AAD seminar series – Laurence Clarke who is an ACE-CRC [Antarctic Climate and Ecosystem Cooperate Research Centre] postdoc and he is based here at the AAD in the Ecological Genetics group. He did a PhD – he finished it in 2008 – at the University of Woollongong; he was actually working on the reserves of Antarctic moss on climate change. After his PhD, he was a research fellow at the University of Adelaide on a couple of different projects working in terrestrial ecology and he also started to get more into genetics and ended up working at the Australian Centre for Ancient DNA. After that position, he applied for a research fellow position at the ACE-CRC and that was in 2015 so he’s been here since then. He’s done work on a whole pile of projects one about which you will hear today. He’s worked on biofilm in the Free Ocean Carbon Enrichment experiment – the FOCE experiment. He was looking at biofilm communities, he has looked at biodiversity of proteists on the K-Axis voyage and he went on the K-Axis trip a couple of years ago. He has also looked at biodiversity in fish stomachs, so looking at fish diet and recently published a paper on that. And if that’s not enough, he has also recently become a father and bravely agreed to give a talk in his sleep-deprived state… So I let Laurence go ahead.

Laurence: Thanks for that introduction, Bruce. My first slide is an introduction slide but I reckon Bruce has covered pretty much everything on here. The only one he missed was work that Bruce did more of than me so I can’t really blame him, and that was applying genetics to look at the samples from the Continuous Plankton Recorder. Other than that he pretty much covered it. What I am here today to talk to you about are some quite fresh results looking at krill-associated bacteria or the krill microbiome.

Today I’ll give you a bit of an introduction about microbiomes and why they are important. Then I shall give you a bit of rundown on how we do it. Then I’ll take you through three results. The first one is the bacteria that we found in the krill stomachs, then the bacteria that we found on the krill surface or the krill moults, and finally how these microbial communities or microbiomes influence nutrient cycling in the Southern Ocean.    

So let’s take it away! Chances are that by now you’ve heard of the human microbiome. Basically it’s… we’re a little bit less human than we like to think. There are three times as many bacterial cells in our bodies than human cells for starters. Now these cells are very small so it’s a bit of an abstract thing to think about. But what drove this point home to me was that finding out that each one of our poos a third of it is food, a third of it is water and the other third is bacteria. So these cells are tiny and that is a lot of bacteria.

Our genome is made up of genes and best guesses are that we’ve got about 20,000 genes. Now in the human biome, we’ve got all these little environments on us - our mouth, our skin, our guts – and each of these have different bacterial communities and these communities are made up of many bacterial species. All these species have genomes of their own. So compared to the human genome, our microbiome is actually made up of 2 million or more genes. So 99% of our genes by numbers are bacterial genes. We are getting an increased understanding in the last 10 years of how important our microbiome actually is to us. It’s now being linked to everything from obesity to depression. The list is ever growing.

Now a quick introduction to everyone’s favourite crustacean, the Antarctic krill or Euphausia superba. Best guesses is there are around 380 million tonnes of krill in the Southern Ocean. It’s a massive population. They’ve got a circumpolar distribution – they are found all around Antarctica. Partially because of this huge population – when we look at the genes or genomes of krill, there is no genetic structure. What that means is if you take a krill from the West Antarctic Peninsula and a krill off Mawson station in East Antarctica, just by looking at their DNA you couldn’t tell that they are from those two different areas. As far as their genome is concerned, it looks like one big population. Now because there are huge numbers of krill, chances are there are huge numbers of krill-associated bacteria. If we do a back-of-the-napkin calculation, say that 0.1% of each krill is bacterial biomass that still gives you 380,000 tonnes of krill-associated bacteria in the Southern Ocean. So chances are that they are fairly important. This has pretty much been overlooked until now. I figured it’s about time to get in there and figure out what’s happening.  

I am just going to give you a quick rundown on the different parts of krill that I’ll talk about today. The first one is their moults, also known as their carapaces. It’s kind of the equivalent of our skin. They shed that on average every 15 to 30 days in summer. It’s made of chiton; I’ll come back to that near the end. Just like us they also poo. The krill faeces are those tiny green threads in here, and – just like us – I suspect that they are full of bacteria from their gut. As far as their digestive tract goes, they’ve got two main parts. The first one is their stomach, this kind of half-moon shape… hopefully can you see what the stomach is. Behind the stomach, there is the digestive gland, much bigger and green from all the phytoplankton they’ve been eating.

How do you go about sampling a krill microbiome? This is where I love the krill group! They have a growth experiment that they’ve set up. So they make use of the fact that they [the krill] moult every 15 to 30 days, and by comparing the moults with the actual krill they can work out a growth rate. To do this, they jar up the krill; that means from each individual krill you can collect the moult and the faeces, but also the stomach and digestive gland and whatever else from that one individual.

Let’s go through step by step; how does one get a krill microbiome. First, you’re out on a ship, in this case the Aurora Australis on the Southern Ocean. You wait for the acoustics to show up a big blob that is a krill swarm. You put your net in and hopefully you come up with some krill. At the same time, I took a seawater sample and filtered that so that I could look at the bacteria in the seawater, and I could compare that with the bacteria that I found on the krill. Once you’ve got your krill trawl, you set up your krill IGR experiment. You jar up about 300 individual krill and pop them into a tank on the back of the Aurora Australis, and you wait for six to twelve of them to moult. That takes 24 to 48 h hours. Once they moult, you collect that moult and also the krill and then you bring it back to the Australia. You can dissect them to get the stomach and digestive gland and whatever else you’re after.

Once you’ve got your samples, you know want to extract the DNA and sequence it. What’s made all this microbiome research possible in the last ten years has been the invention of high-throughput DNA sequencing. Just as an indication of how far we have come, the draft human genome took about ten years to finish. It started in the 1990s and the draft was finished in 2000. A massive international collaboration! They were using machines that would sequence one piece of DNA at a time. Now high-throughput DNA sequencers sequence literally millions of pieces of DNA in parallel in a single run. You can put hundreds of samples on one run, you get tens of thousands of samples per sample. A high-throughput sequencer can now give you the equivalent of 48 human genomes in one run that takes about two days. It was a major leap forward there.  

That’s where the headaches come in. The final step is the bioinformatics and the analysis. Just in this dataset that I talk to you about today, 14.8 million sequences and that is after getting rid of any low quality DNA sequences, errors and things like that. Within that let’s call it 15 million sequences there were 6000 unique sequences, and as far as bacteria are concerned, we call them Operational Taxonomic Units or OTUs, basically the equivalent of bacterial species. That’s our best way to tell them apart.

To start with I had 24 krill from four trawls on the K-Axis voyage. This plot – and there will be a few of them – basically shows firstly the seawater and the jar water, which is the jars they were in – is very distinct from the krill-associated bacteria, the krill microbiome. Just to walk you through this plot, points that are closer together have more similar bacterial communities. Hopefully you can see that sea water and the jar water at the very top of the plot are quite separate from our four different krill tissues, and hopefully you can also see that the four different krill tissues are also quite distinct. So the moults are quite distinct from all the digestive samples.

Getting on to the gut bacteria, the bulk of our microbiome resides in our guts. But the diversity in the krill gut was much lower than that in the sea water or on their moults. Sea water and moults have 4000 or more of these OTUs or species whereas the stomach and the digestive gland somewhere between 50 to a 100, so much, much lower.

This is another complex graph. What was cool was we could kind of see something that is highly likely to be a symbiont that is helping the krill to survive. Each of the vertical bars in this graph represents a different one of our samples, and the different colours represent different bacterial phyla. The light grey that dominates most of the samples across there is Proteobacteria, and the blue that you can see in the digestive samples is Actinobacteria. Though what is most interesting to me is that hot pink number at the top. That’s in a phylum called Tenericutes, but that phylum was made up of one bacterium called Candidatus hepatoplasma. This bacterium is interesting because in other crustaceans it’s been shown to basically give an improved fitness. The crustaceans that have this bacterium are better able to survive on low quality diet. So there is every chance that krill with this bacterium, well, basically this bacterium is helping the krill to absorb nutrients that otherwise wouldn’t be available. We can also see is that the distribution of this bacterium is fairly random. It wasn’t in krill from a particular trawl, it wasn’t just in adult krill or juveniles, wasn’t in males or females. It was quite stochastic as to whether any given krill had this bacterium or not. So there is a question of how did it end up here?

I am going to duck over to the humans for a moment as a possible explanation. Now that we are getting a better understanding of how important the human microbiome is, a key question is: can we restore a healthy gut microbiome in people that have microbiome related diseases. One way you can potentially do this is by faecal transplant and without going into the details, you basically want to get a sample from a healthy donor and use it re-seed the gut microbiome of your unhealthy recipient. This has been shown to have some pretty amazing success rates. C. difficile is a particularly nasty gut bacterium that often can’t be resolved with antibiotics. So people have this incredibly long-lasting debilitating reaction that is difficult to treat. Faecal transplants have had a 90% success rate in resolving C. difficile infections which is pretty amazing.

Now, marine snow could be the faecal transplant for krill. You may or may not know, marine snow is quite different from the pure white stuff that us terrestrial folks are used to. It’s actually made of a lot of detritus, dead cells but also plenty of faecal matter. Krill are fairly indiscriminant feeders; anything that is in the right size range is going down the hatch. So it might be that the krill is able to ingest this Candidatus hepatoplasma basically to re-seed its own guts. Marine snow might be the answer for the random distribution of this bacterium in the krill.

Now moving on to the moults… As I said before we started out with four trawls and looking at the moults – effectively the skin microbiome of the krill – we got this pretty amazing result that those four trawls, those four different populations, all had completely distinct bacterial communities. You can see on the plot that they are perfectly cluster together. You don’t often see that kind of resolution in this kind of studies. So that was pretty amazing.

To sort of crank this up and figure out what was going on here, we wanted to ramp it up. So we took a total of 17 populations – or trawls – from both the K-Axis and Totten voyages. These trawls were anywhere from 4 km to 3500 km apart. They were also in different water masses, what I figured would be different environment, so south of the boundary of the Antarctic Circumpolar Current, as well as a bit further north.

And sure enough, moults from different trawls were quite distinct. Interestingly, the further apart your krill swarm was the more distinct your bacterial population was.

I wanted to get to the bottom of this. Is this an environmental effect, is it something else? One way to try and tease this apart is to look at krill that is presumably living in identical environments. One thing that was handy was that on the Kerguelen Axis voyage, we did a krill box; we sampled quite a few krill swarms within quite a narrow size range - sorry, distant apart. In this case, we had three swarms that were four to eleven kilometres apart. As far as the Southern Ocean is concerned this is a pretty tiny distance. Rob King did make the very good point though you wouldn’t want to through a cricket ball between them. Hopefully you can see on this plot, even though these populations were much more similar than what we saw in the other plot, but statistically speaking they are still very much distinct. Basically they are in the same environment but they still have distinct bacterial communities.

One other way we can look at this is in the krill aquarium. I don’t know how Rob ended up doing this – thank goodness he did! – basically they took one swarm from the Totten voyage , one single trawl, and split it across four tanks in the krill aquarium. These krill were getting the same food, the same light environment, and the same water for nine months but importantly that water was being filtered and sterilized between tanks. So it was the same water but the bacteria shouldn’t have been transferred between them. After nine months, we took a few krill out of each tank, 12 krill in this case, and we swabbed them. This was a different way of getting the moult microbiome; we didn’t really know if it would work but luckily enough it did. It felt like giving the krill a little massage; it’s quite cute. Sure enough, four different tanks, once again totally distinct bacterial communities had developed over nine months in the krill aquarium. You start off with what was presumably a homogenous swarm; over nine months time they ended up with distinct communities.

How is this happening? One answer could be roller derby. Doesn’t make much sense… Before I get into how this might help us understand, I give you a quick intro into what roller derby is. [video playing] Alright. Now that you all know what roller derby is - and probably woken up those two sleeping babies at the back of the room – bit of an explanation. There was a very cool study in the US looking at the skin microbiomes of roller derby players. They started off with three different teams from different parts of the States and each team at the start had a distinct skin microbiome, same as the krill. But they found that after each bout, and this is obviously a contact sport, that the microbiomes of the different teams has started to converge. The contact is having a homogenizing effect on the skin microbiomes of the roller derby players. If you want to see some roller derby action, I’ve got a little plug here for the Convict City Rollers, that is the local Hobart roller derby association.

So, krill roller derby. Krill swarms can contain 1000 up to 2000 individuals per cubic metre. They are constantly swimming into their friends and relatives and into their bacteria as well. So there is every chance that like a roller derby this is having a homogenizing effect.

That explains why a swarm ends up with a homogenous moult microbiome. But why do the different swarms end up with different microbiomes? It doesn’t seem like it’s environment based on the krill aquarium and the krill box populations. Maybe it’s what is called ‘ecological drift’. This is basically random changes in the population that accumulate over time. If you imagine in the krill aquarium that they’d been in there for nine months, if the krill bacteria are reproducing say at one generation per day, that is 270 generations over that nine month period. That is quite a bit of time to accumulate differences potentially. This might be incredibly handy for krill fishery management. We’ve seen from the genomics that we couldn’t tell the different krill populations apart from their own DNA. But it might be that the moult microbiome gives us a chance to look at the mixing of different krill populations, whether we are looking at different stocks in different parts of Antarctica.

Just before I move on, there was one time when we did see an environmental effect. This was an experiment where krill in two tanks in the krill aquarium were put on a winter diet for six months. Those krill were fed once a week instead of six times a week for that six month period and then they went back to a standard diet for four months after that. We swabbed them just like the other krill and what we ended up with – the krill from tanks H2 and 5 are the two in the top of the graph – so they are totally distinct from all the other tanks, and they are also slightly distinct from each other. Because of that diet they seem to have converged on that some microbiome. Now it turns out that these communities were particularly enriched in a species called Arenicella – this was Leonie Suter’s experiment to look at regression of krill sexual regression in winter – and she also managed to find out that these Arenicella are associated with diseases of lobster shells. So the picture down the bottom is actual bacterial lesions that appear on lobsters that are infected with this Arenicella species. So it likes environment can alter microbiomes as long as that environmental effect is strong enough.

Moving on to the last part of my talk, now that we know the community, we know the bacteria that are present. If we know their genomes, we can also know the genes that are present and we can use that to look at the function of these communities in the different krill populations. There is a computer program with an incredibly awkward acronym – it’s called PICRUSt – stands for Phylogenetic Investigation of Communities by Reconstruction of Unobserved States – I’m kind of glad that they do have the acronym. We start off with our OTU table of different bacterial species of our different samples. We then take all the known bacterial genomes – about 61,000 bacterial genomes have been sequenced by now. It’s quite a bit of information out there. All the bacteria in our samples that have already their genome sequenced we now know the genes that are present. For all the bacteria that don’t have genome sequence we can predict the genomes based on the closest relative. After that we end up with a table of genes by samples.

When you do that for the krill microbiome, you end up with 7000 genes. And for me, I don’t do that much of this genomic work, it is very easy to get lost in this genomic space. It’s tricky to get anything sensible out of this. So I decided to focus on one particular gene that was probably going to be important for the krill and their bacteria, and that was the chitinase genes. Chitin is the second most abundant compound on Earth after cellulose. They reckon there is about 1 billion tonnes produced each year. Krill moults are pretty much entirely made of chitin and each krill is on average about 7 per cent chitin. It’s a really important carbon source both for krill-associated bacteria and the Southern Ocean in general.

These are really preliminary results. Looking at these different moult microbiomes, it did seem the different communities did have different levels of this chitinase gene. But was particularly interesting – I am not really sure what it means yet – is that we had four trawl that were more northerly than the others, and all four of these trawls had low chitinase gene content whereas the southern trawls had a much broader range. Some of them much higher chitinase gene contents. So the different bacterial communities do effect the gene presence and the function of these communities. It will be interesting when we get into it and can figure out more of what this means.

Just to wrap up… Hopefully you’ve got a new appreciation of how many krill-associated bacteria are in the Southern Ocean and how important they might be for the krill and Southern Ocean ecosystems. We found a krill gut bacterium that is potentially helping them to survive on a low quality diet and to absorb nutrients. It could be important for their survival over winter. We’ve also seen that the moult microbiomes could be very useful markers for the krill fisheries to see whether we are looking at different effective stocks in different parts of Antarctica, and lastly by looking at the genes we can see that different communities potentially have different functional traits that could influence carbon and nutrient cycling in the Southern Ocean.

Thanks very much for listening and special thanks to Rob, Jessica Melvin, Leonie Suter, Andrew Bissett at CSIRO, all the K-Axis expeditioners, Bruce Deagle and also the krill STS folk who I haven’t stuck on the slide but they definitely deserve a thank-you, too. Thanks very much.

[end transcript]