Hello, and thank you for listening to the MicroBinfeed podcast. Here, we will be
discussing topics in microbial bioinformatics. We hope that we can give you some
insights, tips, and tricks along the way. There is so much information we all
know from working in the field, but nobody writes it down. There is no manual,
and it's assumed you'll pick it up. We hope to fill in a few of these gaps. My
co-hosts are Dr. Nabil Ali Khan and Dr. Andrew Page. I am Dr. Lee Katz. Both
Andrew and Nabil work in the Quadram Institute in Norwich, UK, where they work
on microbes in food and the impact on human health. I work at Centers for
Disease Control and Prevention and am an adjunct member at the University of
Georgia in the U.S. So hello, and welcome to the Microbial Bioinformatics
podcast. Lee, Andrew, and myself, Nabil, are your hosts for today, and we are
joined again by Professor Mark Pallant, who is a professor of microbial genomics
in the University of East Anglia and a research group leader at the Quadram
Institute. We're continuing our trip down memory lane and reflecting on some of
the exciting bioinformatics moments and lessons learned for bioinformatics in
the 21st century. So Mark, welcome back. Great, thank you. Well, thanks for
having me. Last time round, we left it just around the turn of the millennium,
and it was an exciting time to be in microbiology because we were getting all
these genome sequences and we were seeing for the first time the big picture,
even with E. coli that people have been studying for the best part of a hundred
years, half of the genes were not discovered until its genome was sequenced. And
I got involved in that area, not just through the genome sequencing projects, I
mentioned Campylobacter, but I was also involved in genome sequencing projects
for Carinobacterium epitherii and an unusual organism called Tropharima whipli,
which is an organism that causes a disease called Whipple's disease and lives in
the small intestine and at that time could not be grown in the lab. There's a
guy called Didier Raoul in Marseille, his team had actually managed to grow it
in association with human cells. And we approached him and said, do you want to
be involved in a Sanger lead sequencing project? And he said, no, I don't get
involved in projects unless I lead them, would I be the lead? And the Sanger
said, no, we don't do things that way. And so I managed to persuade a guy, David
Relman, who actually was the one who, he was a hero of mine at the time. And one
of the great things about a career in academic life is you do get to meet your
heroes and sometimes you can get to work with your heroes as time goes on. And
David had been the first one to show that Tropharima whipli was a distinct
organism, it amplified 16S sequence and had actually given it that name based on
the 16S sequence. And so he was up for genome sequencing it. And it took them 18
months to grow enough biomass of this organism in the lab before they had enough
DNA to sequence it. So the actual growth phase of the project took longer than
the genome sequencing phase of the project. Luckily, in France, they set up
their own project and I'm not normally nationalistic, but in this occasion, I'd
like to say that the UK won and France came second place in that one. So, Mark,
is this the guy who there's a bit of a scandal around him with COVID and the
hydro? Indeed, it is the same guy. And I think that in life, there are often
these people that stir things up and sometimes they create a lot of trouble and
a lot of difficulty. But the other time they do create something new that wasn't
there. And Didier Raoul is one of those people who had, his reputation has been
severely tarnished during the COVID epidemic with his championing of treatments
that don't work. But he was the first person to work out, or his group, at
least, was the first group to grow this organism, Trophorhippoli, in the lab. He
was also the group that discovered the Mimivirus, these giant viruses with huge
genomes that have functions that overlap with bacteria, so completely blurs the
distinction between viruses and bacteria. You know, that's Microbiology 101 is,
oh, there's a clear difference between bacteria and viruses. It turns out there
isn't, and that's down to Didier Raoul. He's also invented Culturomics, high
throughput culture. And if you go to the databases and you look for the species
epithet applied to bacteria that is most commonly used, it is Massiliensis,
which is the Latin word for from Marseille, because his group have named
everything that comes out of Marseille Massiliensis or Massilia or whatever. And
there are hundreds and hundreds of bacteria that they've named. His contribution
to science is quite astonishing. It's just unfortunate that he's just had these
problems recently. And, you know, there are similarities with Craig Venter,
which was a person upset a lot of people as well, even though we we do look back
and say, well, he did he did progress the field by shaking things up. And yes,
so we sequenced Tropharima whipleri. Interestingly, just in recent year, in the
last year during lockdown, one of my old school friends from primary school, no
less, got in contact with me and I hadn't seen him for nearly 50 years. And I
said, how are you getting on? He said, oh, yeah, he's right. We'd had this heart
infection with this very unusual organism. I said, oh, what's that? He said,
it's called Tropharima whipleri. I said, oh, I know that organism. We sequenced
its genome. It was like it brought it home to me that it's a very rare
infection, but it affected someone I knew. And he'd actually he's just in the
last few weeks relapsed. So, you know, it is an important infection. Anyway,
what happened then was that I started getting interested in doing homology
searching, similar to what I've done with urease, but looking at these new
genomes. So suddenly we have this landscape of dozens and dozens of new genomes
and we could analyse those genomes. We could see what's in them and we could
start to make sense of it. One of the first such opportunities was I moved to
Ireland. I took up a chair in the Queen's University Belfast, Northern Ireland.
When I arrived, it was not long after the Good Friday Agreement that brought
peace to Northern Ireland. And the British and Irish governments were making it
very clear that they wanted there to be much more collaboration across the
border. And they set up a research fund where people could apply for grants for
research straddling Northern Ireland and Southern Ireland, the Republic of
Ireland. And soon after I arrived, there was a guy who worked in Trinity
College, Dublin, called Tim Foster, who was an Englishman who was famous in the
field. But he was one of my heroes in a sense in that he pioneered the
application of molecular biology to understand the way which Staph aureus causes
disease. I contacted Tim Foster and I said, they set up this scheme that's
supposed to be getting people working on either side of the border,
collaborating with each other, but they're not going to check our passports and
say, well, actually, we're British people working in Northern Ireland and in the
Republic who are using this money. So why don't we see if we can tap into that
funding? And he said, yeah, OK, why not? I said, well, what about we try and
data mine the recently released Staphylococcus aureus genome? He said, yeah, OK,
I suppose so. I said, anything you're particularly interested in? He said, well,
we know that there's this enzyme called SORTASE, which is involved in targeting
proteins to the surface of Staph aureus. These are so-called LPXTG proteins. And
it would be great if we could mine that genome because we know some of them, but
it would be good to know all of them that are in that genome. Do you think you
could go away and do some bioinformatics to find all these? So I said, all
right, I'll have a look for SORTASE and its substrates. And it turned out that
the substrates had this particular motif, LPXTG, and they usually in Staph
aureus were scattered around the chromosomes. And so I just took the SORTASE and
I used a program called CyBLAST, which had just come into common usage, which is
an iterative process. So you can do a BLAST search and say, oh, here's a load of
homologs. And then you say, right, let's now build a model out of all those
homologs to look even further into kind of homology space, if you like. To me at
that time, CyBLAST became the equivalent of crack cocaine because you run a
CyBLAST search and you see a load of things. You think, wow, yeah, that's good.
Those are things I kind of know about. Well, there's a couple of things I wasn't
expecting. You press the button and reiterate and then suddenly a whole load new
things come in. Wow, I found a load of new stuff. Then you run it again and you
find a load of new stuff. And it was like, wow, this is just amazing. And so I
did this with SORTASE and I did it with the substrates as well, with the
particular region of the substrates that is recognized by SORTASE. And I did it
on Staph aureus and I found, I can't remember the exact number, half a dozen or
more new SORTASE substrates. And Tim went on and patented all those and we wrote
up a nice paper on those new, they did some lab work characterizing them in
Staph aureus. But I thought, why stop at Staph aureus? We've got all these other
genomes out there, dozens and dozens of them. Why don't I just run the same
analysis across all of these? And I did. And that was one of those moments where
it brings to mind, there's a guy called Alfred North Whitehead, who in the early
20th century said the goal of every natural scientist is to seek simplicity, but
mistrust it. And this has been a theme that you kind of think you understand the
thing and it's all quite simple. And then it turns out when you look harder,
it's not that simple at all. And so what I found.  found was that when I looked
in other organisms, the sortase substrates weren't all scattered around the
chromosome like they are in Staph aureus, they're actually usually clustered in
a gene cluster with the sortase enzyme. In some cases, including in Staph
aureus, in fact, there was more than one sortase. We kind of thought, oh, it's a
single gene, single copy gene, it's doing a fundamental thing, there won't be
multiple copies of it, but it turned out in some organisms, including
Chloronibacterium diphtheriae, which I was working on the genome project for, it
turned out there were multiple sortases and multiple substrates, and they were
often multiple clusters with the two things together. And so it turned out that
this was one of these things where you suddenly, by looking at the much wider
genomic landscape, and carefully using these homology searches, you can come
across things that you never anticipated. And it turned out that the sortase
wasn't just sticking things on the cell surface. In some of the actinobacteria,
it was actually targeting things to the surface and then making them into
fimbri, gram positive fimbri. So it was part of the mechanism for building these
fimbrial appendages that were on the surface of that particular group of gram
positives. So that was one example where it was clear that if you do these
homology searches, and you do them carefully, and you use all the data that's
available to you, you can suddenly see a much wider landscape. I had a similar
experience with a guy came to, by the time I was in Birmingham, he came and he
said, oh, we've got this new test for TB that relies on an antigen called
ESAT-6. It's a T cell test, and it's much more sensitive than anything else. And
I said, so what is this ESAT-6? He said, well, it's the major antigen. I said,
well, yeah, the bacterium doesn't make it to make an antigen, to be an antigen.
That's just a side effect of interaction with the host. But what is its function
for the bacterium? He said, I don't know. I'm an immunologist. I don't care. So
I thought, well, I'll go and have a look and see what I can find. And I did side
blast searches. And what I found was that there was a particular ATPase
associated with the system in TB, in the gene cluster. And there were these
substrates. And when I carefully did the homology searches, I found that these
things were not just present in mycobacterium tuberculosis and other related
mycobacteria, not even just in the actinomyces, which other people had kind of
hinted, but they were very widely distributed. They were found in Staphylococcus
aureus, in Streptococcus agalactii. They were even found in some gram negatives,
which is a bit weird. Why on earth would they be there? And we still don't know
to this day what they're doing in some gram negatives. But this opened up a
whole new like, wow, these things are much more common than we anticipated. And
there are many more systems out there. And in fact, one of the nice ironies was
that the Sorte system I mentioned earlier, that was a great example of
predictive science in that there's a guy, Olaf Schneewind, who'd seen all the
LPXTG proteins and predicted that there was a specific enzyme involved in
putting them on the surface that he called Sorte. And his wife actually,
Dominique, was so interested in what I'd shown with the ESAT6 system, WXG100, as
we called it, that she started working on that in Staphylococcus aureus and
actually started researching that area. So that was quite a nice thing. I then
got more interested in E. coli and did lots of work on E. coli, started working
on type 3 secretion in E. coli, particular secretions around also flagella
secretion. One of the things I found when analysing E. coli genomes, we kind of
imagined that E. coli K12 was handed to us by God, if you like, as the model
organism. It's this perfect organism. And it's not going to show any kind of
hallmarks of degeneration or anything like that. What we found when we looked
was there was a type 3 secretion system cluster, which was called ETT2, which
appeared to be extremely widely prevalent in E. coli strains. But in every case
we looked at, there were frame shifts and deletions suggesting the thing didn't
function as a type 3 secretion system. And then we found, looking at the K12
genome, that was this funny little flagella cluster of two flagella genes that
were homologous to flagella genes from the main flagella cluster. But they were
incomplete. They had no start codons, they had no promoters. It's just like this
weird little scar. And then when we sequenced intra-aggregative E. coli genome,
we found that in that position in the genome, there was a whole complete new
flagella cluster, which we called FLAG2, that was present. And so the scar is
present in all the E. coli, and just a small number of them retain the ancestral
state. So as an example, where you look at K12, it's similar to what Darwin
spoke about, vestigial organs in the origin species, but you get vestigial
genes. And we'd seen that as people were sequencing genomes in Mycobacterium
leprae, they found lots and lots of degenerated genes that were present there.
But the idea that this happened in the most revered organism, E. coli K12,
seemed a bit shocking. At the time, it seemed shocking. Now looking back, well,
why do you expect E. coli K12 is just another organism? It's just another
strain. So what? But that was very interesting. So for K12, I'm not an E. coli
expert, but the experts have told me that it's a weird organism. And in
contrast, you said it was handed to you by God. Wouldn't you have preferred
something a little bit more representative of E. coli? Well, one of the things
that happened early in genome sequencing days was that people said we should
sequence the type strains, the strains that everyone's using, the model organism
strains. That brought problems because the same is true that Salmonella typhimum
LT2, many of these strains have been in the lab for a long period of time have
been subcultured. The K12 even have been irradiated in its ancestry. And so
there is this question, do they reflect the real world? It's a bit like the
first human genome was largely Craig Venter's genome. And we don't want Craig
Venter's genome to be the only representative for the whole of humanity. And so
it became clear that we really need to sequence lots of strains. And Gordon
Doogan, when he moved to the Sanger, he made that point that it was best to take
fresh isolates that have been in the wild only recently and be minimally
subcultured and sequence those alongside the lab strains so that you actually
get to see things as they really are rather than the degenerative forms you see
in the lab. And that was a big insight. And I did actually in those early days
with Brendan Wren, we wrote a review for Nature where we kind of made this point
that you've got to look at all this genome stuff in what we call the eco-evo
context, the evolutionary ancestry of the strains you're looking at and of the
species. So how does Yersinia pestis arise to become a very specific thing from
a general background? Had mycobacterium leprae arise, all that sort of thing to
become very specialised, but also the ecological context. So that many things
that we'd say, oh, that's to do with virulence, they turned up in things that
weren't virulent organisms, they were just commensals or environmental
organisms, these things like the sawtases or the WXG 100 systems. And it became
clear a lot of people at that time were recognising that you could use
invertebrates as models, animal models, and you could even use amoebas as
models. But actually, that's not artificial, because the struggle for life, as
Darwin called it, between bacteria and eukaryotes goes on everywhere. It goes on
in the soil, in the oceans, and most of it is between unicellular eukaryotes and
bacteria. We, you know, complex metazoans are kind of an afterthought, that
struggle. So that was an interesting kind of observation that came out of all
that. The other thing that happened around that time was a guy from New Zealand
called Scott Beetson approached me. He'd been working with a guy, Chris Ponting,
and I'd worked with Chris Ponting. Chris Ponting taught me a lot about how to to
do domain searches and homology searches in proteins. He said, oh, we've got
this guy who wants to come and work with you. And I said, well, I suppose he can
come and work with me. And I said, Scott, you know, okay, you can work. He said,
well, I've got to put in this proposal for a Dorothy Hodgkin Fellowship. I said,
well, our grants department can't work that quickly. He said, oh, well, I've got
to do it. This is my one opportunity. So I called him and I said, would you take
it? And they said, I suppose so. And I thought, can he actually write a proposal
in a week? But he did, and he got it. And he came to work with me. And he'd been
working on Pseudomonas aeruginosa in the past, with another one of the giants in
the field, John Matic in Australia. And I said, well, you could work on that if
you want, Scott, but you don't, now you've got the fellowship, you don't have to
do what you said you'd do. And this great opportunity has come up because I've
been nosing around on the back of this ET2 secretion system, and just looking
more generally at type 3 secretion in E. coli. And what I've noticed is that
there are these so-called effectors, which are the targets of these type 3
secretion system. And everyone's assumed that they're all present or clustered
in this one gene cluster called the locus for entrocyte effacement, because they
think it's kind of a modular system, you just plug that into a cell and it
becomes able to do type 3 secretion. But we've got hints that there are lots of
these effectors that are homologues of those things, elsewhere in the
chromosome. But you think you could have a look at this. And so he went and had
a look at it. And to cut a long story short, by using homology, we managed to
find dozens and dozens of new effectors. We tied up with Gary Frankel in
Imperial, and Toru Tobe in Japan, and we're able to even show that many of these
were actually secreted by the secretion system.  and this led to a vast
expansion and led to a P&S paper in fact as well. But for me that was one of the
most rewarding things where we actually tied the bioinformatics predictions into
laboratory confirmation in a very coherent way and it was really quite exciting.
So I had a question, a little bit before the E. coli you were talking about
looking at the antigen on tuberculosis and you had to look into the function and
everything. Do you feel like that might have like a tie-in or might be a
predecessor to the efforts for reverse vaccinology in that whole field? Oh
definitely. Reverse vaccinology was something that came in in the early part of
the 21st century as well. On the back of those first tied genomes people going
in saying well we can go out and we can do that. In fact that particular protein
had been discovered before the genome had ever been sequenced. It just turned
out to be one of the brightest antigens that was recognised by T-cells and
that's why it was picked for this T-cell antigen detection test, the so-called
ELISPOT test that was developed for TB. But to this day it's still not entirely
clear what the function of that particular protein is. It seems to have all
sorts of proteins and people have looked at it from many different angles and it
seems to do all sorts of things and it's probably part of the secretion system
as well. It turns out that that whole system is very complex and very
multifaceted. All right and on that note I think we'll draw to a close. I want
to thank our guest Professor Mark Palin for joining us today. Join us next time
in the Microbinview podcast where we continue our journey with Mark. Thank you
so much for listening to us at home. If you like this podcast please subscribe
and rate us on iTunes, Spotify, SoundCloud or the platform of your choice.
Follow us on Twitter at Microbinview and if you don't like this podcast please
don't do anything. This podcast was recorded by the Microbial Bioinformatics
Group. The opinions expressed here are our own and do not necessarily reflect
the views of CDC or the Quadram Institute.