Altruism, Synonymous Sites and Gene Evolution - Professor Laurence Hurst

The Director of the Milner Centre for Evolution, Professor Turi King, talks to Professor Laurence Hurst about his career, his research, and his passion for all things evolution.

Episode Transcript

(00:02):
Hello and welcome, you are listening to a podcast by the Milner Centre for
Evolution at the University of Bath. I'm Professor Turi King, your host,
and today I'm talking to Laurence Hurst, who set up the Milner Centre for Evolution with
Jonathan Milner and became its first director. I'm going to be talking to him about his career,
his current research, and his passion for all things evolution.

(00:30):
Laurence, you're a professor of evolutionary genetics here at the University of Bath,
are you able to trace back your interest in evolution to a particular moment?
I can almost date it to second year lectures at university. They weren't about evolution;
they were about behavioural ecology. And it was about how one can use evolutionary thinking,

(00:53):
particularly optimization theory and game theory, to approach interesting
questions about animal behaviour. And it never even occurred to me that you could
think about behaviour in evolutionary terms. I thought, wow, that's unbelievably interesting.
And what I particularly liked, I think then about those sorts of ways of looking at evolution,

(01:16):
is that they come at the problem by posing really good, why is it like that, sort of questions.
So, for example, the classical problems of evolutionary genetics and behavioural ecology,
but particularly behaviour ecology for example, would be problems like altruism. And this was one

(01:37):
of the ones that I met in the second year, and I thought, that's really interesting
because it was such a well posed problem.So, the problem more or less is if organisms
are in it for their own benefit, how come some of them are acting to the benefit of others,
at a cost to themselves? I go, yeah, no, that is a good question actually. I can understand
why you would not do that, but I can't obviously understand why you would do that.

(02:00):
And likewise, the evolution of sex is always posed in these sorts of frameworks where it makes
perfect sense if everything was asexual, it makes no sense that most things pay a very heavy cost of
being sexual, and that that cost being, you have to produce sons and sons don't do any investing
into the next generation in most species. And because of that, why bother making sons? There

(02:23):
must be some big advantage to being sexual that is not captured by the initial intuition.
So, I was captured by that way of framing questions as, you'd expect this, but we see that,
how come? And I still think that's a great way of entertaining good brains.
What did you do for your PhD?So, I did my PhD, well DPhil really at

(02:45):
the Department of Zoology in Oxford. But I rapidly discovered that I am not smart enough to answer
complicated questions about real organisms. And I've always gravitated towards the simplest way
of asking a question, you might say. And to me, that meant really nice, interesting questions

(03:06):
about genes. For example, I remember actually I met, in my undergraduate lectures, the observation
that in unicellular algae like Chlamydomonas, the gametes are the same size, but when they fuse one
of the two sets of chloroplasts, and one of the two sets of mitochondria get destroyed.
So, they still have a uniparental inheritance, and I just scratched my head about that, and thought

(03:29):
that's really weird. Why do that?And so, I ended up doing the early part of my PhD
on why that might be the case and how that then might be related in turn to why there are just
two mating types in something like Chlamydomonas and most of those single cell things. When, again,
you can pose that question as, it would make a lot more sense to not have two.

(03:53):
So, if there are three mating types, for example, and the rule is you could mate with anybody who's
not you, then at the equilibrium situation, if there are three, then any individual can mate on
the average of two thirds of the population. If there are ten, you can make with nine tenths of
the population. If there's… in some ciliate single celled organisms, some of them have,
you know, a hundred and mushrooms go up to 30,000 or so. And that makes perfect sense to me.

(04:16):
If you were a mushroom, probably one of your biggest problems is having sex because you
can't move. You meet something on the ground and go, can we have sex? And if there's two types,
half the time go, not tonight dear. 30,000 types, you more or less have solved the problem. So that
makes perfect sense to me. The inverse problem is then, okay, why do so many things have two?
Two is not only a bad solution, it's actually the worst solution. It's one of those rare ones

(04:40):
where organisms are doing something that almost objectively, you can say is not just simply bad
solution, it is actually the worst solution.So yeah, but basically, it's thinking about
genes and genomes, but in the same sorts of terms that one might want to think, what's the best
thing for an organism to do? It's just what's the best thing for a gene or a genome to do?
Given this is how it's inherited, given this is what it can or cannot do.

(05:03):
So, what did you start doing next? So, you've done your PhD.
So, I sort of carried on in that sort of similar vein trying to understand gene and genome biology.
So, I was getting quite interested in genomic imprinting. That's this phenomenon whereby genes
are expressed either paternally or maternally but not both. But it was also at this point that

(05:24):
sequence data was starting to appear, and then started asking just questions about how genes
evolve, but thinking in, sort of, similar ish terms about which ones are fast evolving, which
ones are slow evolving, and so on and so forth.We found that one of the fairly good predictors
of how fast a protein evolves is how fast the proteins next to it evolve. And that's not

(05:46):
a trivial or obvious result at all.I remember reading a quote from Tom
Cavalier-Smith saying that, in eukaryotes there will be no selection on gene order,
there's no rhyme or reason why genes should be in one place or another. And that was a red rag to a
bull. And we just showed very straightforwardly that actually in the human genome there's real

(06:07):
rhyme and reason as to where our genes are, that our housekeeping genes are clustered together and
our tissue specific genes are typically not in those clusters, and so on and so forth.
And so, there was an underlying rationale for the structuring of the human genome. We considered it
also for the yeast genome. And in fact, in the yeast genome, we could show there's a really

(06:27):
striking effect. You can tell me nothing about the genes whatsoever, I don't need a single point
of annotation of the gene whatsoever, but I can predict, moderately accurately, what the knockout
growth effect of any given gene is going to be, simply by knowing the knockout fitness of the
genes on either side of it.Interesting.
There are clustering phenomena within genomes whereby the essential genes,

(06:49):
which in humans are the housekeeping genes tend to cluster together. So, if you tell me the
ABC and B's then I'm interested in, and I know nothing about. But if A and C are both essential,
you knock them out, the yeast can't grow. I can tell you that actually, there's a high likelihood
that if you knock out just B, the cell also can't grow. And I need to know nothing other than that.

(07:10):
Which is very strange because it's very different from our normal assumption. We
have this underlying presumption that genes are autonomous, even though we know they sit on DNA,
and we know they've got neighbours, but we assume that they're autonomous. And yeah, quite a lot of
the work after that was about, to what extent are genes actually interfering with each other?
In no small part, because one of the problems with gene therapy, where a lot of my interests lie,

(07:34):
is that if you insert a gene into DNA, we now know that it will be affected by the neighbours,
and the neighbours will be affected by it. Which means that it is not a simple, safe process.
And so, people have the concept of safe harbours, a place where we think you might be able to insert
a gene and it won't do any damage to neighbours. But the first gene therapy trials had to be

(07:55):
cancelled, because it was noticed that the gene therapy was killing the individuals that it had been
sent to save, and in no small part it was because of effects like this. The gene was inserting,
it was upregulating neighbouring oncogenes.And this is really, really crucial, this work
that you're doing then, which is showing that actually where we put a gene, if we
want to do gene therapy, it's actually really important to know what's going on either side.

(08:18):
And it's one of the big things of the Millner Centre, it’s using evolutionary thought to find
new kind of technological and clinical sort of applications to this work. And
that's exactly what you're doing with this.So, we discovered a few years ago, again,
one of the best predictors of the extent to which a gene has changed its expression profile. So,
you can take the expression profile in human, expression profile in chimps and A another

(08:42):
outgroup. From that you can work out the human - chimp ancestor expression profile. And so,
you can say well how much has a gene changed its expression from that ancestor through to
human or that ancestor through to chimp? And again, it turns out the best predictor of the
amount of change in gene expression, has it gone up, has it gone down, is the amount of
gene expression change in the neighbours again.So that's telling you exactly the same story as

(09:04):
we're learning from transgene insertions. If the neighbour changes gene expression, you change
gene expression. And it's true over evolutionary time and it's true in short immediate time. So,
you do a transgene insert and the best predictor of the expression level of a transgene insert
is the expression levels of the neighbours.So, this is this thing about how genes they

(09:26):
have expression levels. So, whether or not they are expressed, or read,
more frequently than the others. Does that mean that you could get to the level where if you want
to do gene therapy, you can go, right I want this gene to be expressed at this particular
kind of rate. So, I'm going to put it here because we know about these genes either side.
With respect to the insertion sites, the question is much more can we do it

(09:46):
safely. And safely here means not simply the expression level of my gene, but I
also don't want it messing with my neighbours.But we do also have to have the gene expressed.
And one of the things that's been known for a long time is that if you take, you know,
a normal human gene with lots of exons, most human genes are way too long for us to handle. So what

(10:07):
we do is we say, let's get rid of the introns, or at least most of the introns, to make something
nice and compact, and we'll just simply take the gene as it is and get rid of the introns and you
put it back into the DNA and it doesn't work.But this then is the second branch where we're
looking at gene therapy and the like, because we noticed, amongst other things, that the

(10:28):
so-called synonymous sites, what were used to be called silent sites in the protein coding parts
of genes. We used to think they were completely irrelevant, but you can do this experiment. So, I
did this with my postdoc, Christine Mordstein, and collaborator Greg Kudla, in Edinburgh and you say
GFP, lovely green fluorescent protein because that gives us a really easy way to measure how much the

(10:49):
protein is being expressed. Same construct, same everything, except we're just going to
change these so-called synonymous sites.So, the genetic code has 64 KOH-dons,
we only need 20 amino acid. And what that means is that the same amino acid can be coded for by
more than one KOH-don. So, glycine, one of the most common amino acids, for example,

(11:10):
that we use, is coded by a KOH-don starting GG. So GGA, GGC, GGT and GGG all give you glycine.
And it used to be thought it wouldn't matter which one of those third positions matters.
They all give you glycine, what does it matter.But we can make these transgenes and we go,
we're just going to randomise those silent sites, in principle which should have absolutely no

(11:30):
effect whatsoever. In practice, we get orders of magnitude variation, the amount of protein coming
out, just by changing the synonymous sites.So that gives us a really, really good trick
to go, okay, I think I now know how I can go from my big gene to my little gene,
and all I have to do is tweak those silent sites and it works very, very nicely indeed. And then

(11:55):
there's of course a question of why? What's the underlying mechanism behind all that? And that's
what we're working on at the moment.But it's coming back to the evolution,
because actually the reason we got into that in the first place is there's been a funny little
natural experiment. So, when we saw that there's this generic problem, taking big genes make them
into little genes. Evolution has done that for us because we've got a couple of hundred genes

(12:17):
in our genome which have done the experiment already. They have gone from big to little,
and these are things known as retrogenes.So, these are daughter genes of our normal genes,
which are genes and pieces, but they're not genes in pieces. And what we found is when we
looked at these, they had some really interesting evolutionary patterns. But one thing they had
done is that they all had increased what we call the GC content, at the third site. So,

(12:40):
we wondered, why are they doing that? And that's why we actually did the experiment.
And it turns out if you increase the GC content at these third sites. So, for glycine for example,
use GGC, that will give you higher expression level than if you use the A or the T versions. But
we got at that by studying how retrogenes evolved.So, do you know why
that's important? The Gs and the Cs?So, anybody who studies genetics, the first

(13:04):
thing that you learn in molecular genetics is DNA makes an RNA, which makes protein, and that's the
function of DNA. And you go, okay, well what proportion of the human genome actually ends
up specifying protein? So, our best guess at the moment is 1.2%. Well, what's the rest of it doing?
So, what I think is actually going on, we have a general evolutionary theory. And that says

(13:28):
that what's going to happen in evolution to mutations that are a little bit harmful to
you. We're not talking they're going to kill you, not that, they're just a little bit harmful. Well,
it turns out, so this general theory written by Kimura and Ohta, called nearly neutral theory of
evolution, it’s very different from the Darwinian view of evolution. But it says that if you have

(13:48):
a slightly deleterious mutation, if you're in a very large population, yeast, bacteria,
something like that, they're single cell so its massive populations. Basically, selection is
enough time to get rid of this damn thing.So, it's not going to be, in the longer term,
a problem for them. But if you've got a smaller population, if you got a big body, you will have
a small population, then you've got a problem, and these things will carry on accumulating.

(14:12):
And what you find is that the amount of stuff you have extra in genes,
the amount of stuff you have between genes, both correlate when we look across all organisms,
and the really good predictor is what's your population size, just as Ohta would predict.
So, what we think is going on is our genome is bloated, it's rubbish, and it's full up with

(14:32):
dead transposable elements, not because we need dead transposable elements, not because they're
doing something brilliant for us, but because we simply can't get rid of them. Yeast can,
bacteria can, we can't because our ancestral effective population size was really, really small.
So, in the big view what we have is a problem with lots of junk DNA in it. But while most of

(14:54):
our DNA doesn't seem to be doing very much of any interest, the maximum estimate is usually
about 10% or so, nearly all of it does this DNA to RNA bit. Which then gives you a problem,
what is going on here, what are these?So, what I think is going on is we have a
fundamental problem, which is we're making unwanted transcripts all of the time. And

(15:14):
there's lots of evidence for this. So, you can do an experiment. This was recently reported
by two other groups where you take a bit of DNA, completely naive for an organism,
so it's never seen it before. It's not obviously protein coding or anything like that, it's just
load of random DNA. You put it into a cell, what does it do? More than 90% of the sequence gets
transcribed. And so, we're pushing out transcripts left, right and centre, most of which are rubbish.

(15:38):
Turns out also if you want to transcribe a gene normally, so you want to go, you know,
from A to B, as it were, it will also go from A and go backwards. So, it makes a transcript off
the other side. And we have a whole system of RNA chewing up, or don't allow it out of the nucleus,
that seems to be a big way that we control these things. So, we've got a load of transcripts, some

(15:59):
we want to make protein out of, some are rubbish.So, there's two ways to handle it, one is to go
don't make it in the first place. And indeed, we've got some nice systems, there's a system
called the hush system for example, which tries to recognise long A rich sequences because that
is what retroviruses look like. So, if you have a retrovirus, it goes inserts back into your DNA,

(16:23):
it's typically quite long and it’s typically rather A rich, just because of the nature of the
process. And this thing comes along, it recognises them, says, no, you're not going to do this, stop.
And so, then the other is you have this really intriguing problem, that you have this swirl,
like a sort of cloud of transcripts, some of which are rubbish, some of which you want. How are you
going to know the difference between the two?So, what we argued is that there's actually a

(16:46):
series of fingerprints, that work both in theory and in practice. So, if you look at what is the
nucleotide content of the real bonafide protein coding genes, what is the nucleotide content of
the rubbishy transcripts on the genome as a whole, there's an absolute standout. And that
is our protein coding genes are GC rich, and they have to be. Why? Because the underlying

(17:08):
amino acids sometimes have to be GC rich.So, if you just have a load of rubbish sequence,
there's no particularly good reason it would be GC rich, could just be anything. And in fact, this is
where the next bit of theory comes in, because the mutation bias in our genome is actually GC to AT.
Whether that's an evolved strategy of any flavour, we don't know. But we know that on the average,

(17:29):
when you get mutations, they tend to go GC to AT. If selection isn’t operating, the genome will be
about 28% G and C, which means that if you just make a random transcript, it will not be GC rich.
So, what we found is that there are filters, after filters, after filters, traps, after traps,
after traps, and they all work in the same direction. Without exception, they go, I

(17:50):
don't like you if you're AU rich, I do like you if you're CG rich. That makes perfect sense because
the nucleus itself is a great filter. I mean, if you think about what is a eukaryotic nucleus, it's
a structure with lots of little holes in it to let certain things out and certain things not be let
out. And it is passport control for transcripts.Your passport is to show I am not viral

(18:15):
transposable element, rubbish transcript etc. etc. I don't end halfway through an
intron. GC is telling you that's not chance, you’re probably one of mine.
So that's really interesting. So, you're looking at what we would ordinarily think of as being,
oh that's not going to make any difference. But you're actually finding it is making a difference…
And it's making a difference because what then seems to happen is that we have built in all

(18:38):
of these traps and filters. So, if you’re GC rich, fine go through, and what that
means that put selection on the third sites to go, okay, I need to be GC rich. And so,
I'm going to get through these filters. And that's why when we do our transgene experiments,
and we make some transgenes that change only at the synonymous sites, GC rich, some AT rich, the

(18:59):
AT rich ones don't work. They almost don't express at all, and the GC rich ones absolutely lovely.
And we could then actually show that the big difference between the two is,
the AT rich ones don't get out of the nucleus. They’re blocked in the nucleus
and the GC rich ones go, fine we we’ll accumulate and then wait for the ribosome to do its business.
So, you're making some quite big discoveries about the genome then?

(19:22):
Well, I would like to believe so, but then I'm quite conceited.
Okay, so just a small thing that you've been working on there. What else are you working on?
And then the other sorts of things that I do, I'm quite passionate about evolution
education and communication more generally.So, we have this other branch of what we're doing.
So, we started off doing large scale randomised control trials of different teaching methods,

(19:45):
to teaching evolution, and we came up with some very nice simple policies.
So, it turns out that if you do genetics and evolution, simply in that order,
just change the order, not what you’re teaching, nothing of that. Just change
the order in which you're teaching, it has a huge effect on the understanding of evolution,
but it also increases understanding of genetics.So, a very simple policy, just do genetics then

(20:06):
evolution. And in fact, I would say if you're doing a biology course you should start with
genetics, then go to evolution, and the rest is detail. Why? Because what makes biology different,
why are we not just a subbranch of chemistry? Because genetics gives us a really strange
set of properties. These things we call organisms have a very, very odd property,

(20:28):
which is that they propagate. And the things that they produce are remarkably similar to the thing
that produced them. That is a very strange property, we don't see that anywhere else.
And the consequence of that property is evolution. You don't get evolution, not biological evolution,
without this strange property of inheritance.You're clearly really passionate about

(20:51):
evolution. So why do you feel it's so important for people to learn about it?
I think you need to understand evolution because it is the best explanation we got for why we're
here. It is just fascinating as a process; I can say all of these things. But certainly within
the western tradition, there is a large skein of thought that sort of puts humans different, above,

(21:16):
and other. And you can sort of trace that back to, you know, Christian influences and you can say,
well, it's not greatly dissimilar to the notion that the Earth is the centre of the
universe and the Sun goes around the Earth, etc., etc. it's a sort of centrality and
importance to us, that puts us above nature.And certainly, if you look at problems like

(21:37):
climate change, biodiversity destruction, etc. etc., one of the most powerful tools
I think we could have is if people actually understand that we aren't special. We are one
small twig on a branch, but we are not simply part of this phylogenetic tree,

(21:57):
we are dependent on this tree. We are dependent on all the other species, as every other species
on the planet is dependent on other species on the planet. So, they sit within ecosystems and
these things are fragile and so on and so forth.But to see ourselves as a product of evolution,
rather than the product of evolution, I think is actually quite an important transition in people's

(22:23):
thinking. And you can see it happening already, I mean, I talk to my tutees and they clearly get it,
you know, the ones who are passionate about conservation see that we've put ourselves
above nature, whereas I think it's rather important to see ourselves as part of nature.
But when you just think about an image of evolution, what image comes to your mind about

(22:43):
evolution? Most then will give you a version of this so-called march of progress. So, we have the
chimp knuckle walking and then eventually standing up. But the biases in that are so huge. I mean,
it's implying that there's a directionality in the process, it's implying there's a sense of
progress to the whole process. You very obviously read left to right, going primitive to advanced,

(23:05):
you know. But actually, that's not primitive to advanced, that is simply evolutionary change.
In fact, if we look at our DNA, there's actually been more change from the ancestor
of humans and chimps to chimps, than there has been from that ancestor to humans.
But those sorts of assumptions seem to sit at the back of so many minds. But that idea that we are

(23:25):
sort of the fairy on top of the evolutionary Christmas tree, the end product, the perfect
thing, just seems to me completely wrong.But I think you can also think that actually
humans are not particularly good. So, if you look at early embryogenesis, so this is just
fertilised egg and you look at a fish, you look at amphibians, most of them are absolutely fine,
no problems whatsoever. You look at the human ones, and from what we know, about 50% of those

(23:50):
initially fertilised eggs die, and they die before the mother knows she's even pregnant.
And we have a quite good idea why, it's because they have the wrong number of chromosomes. Fish,
amphibians do not have this problem.Likewise, then when you get past six weeks,
20% or so then miscarry, and then the ones that do pop out, current estimates, because we've got

(24:10):
such a high mutation rate, 5 to 10% of us have some rare genetic disease. So, you look at that
and go, that doesn't look like a fairy on the top of a Christmas tree. That's not perfection.
And if that's true, and if you can teach that that's true, does that then change your mindset
about what humans actually are? We're just another species, we're not actually all that great,

(24:31):
in these terms. These are real enigmas to try and explain why we do have such a high mutation rate,
why we do have so many failing embryos, but also then this connection to nature,
that we're not separate from, but part of.So, I think when you teach evolution,
what I would say is the overriding concern should be, is not simply that it's our best
scientific explanation. And it is the most fascinating project to try and understand

(24:54):
the history of life on earth, how we got here, and so on and so forth. But actually,
it is about this, humans are not special, we are connected with nature, we need nature,
and its high time we started respecting nature.Laurence, thank you so much for talking with me.
This was a podcast by the Milner Centre for Evolution at the University of Bath.

(25:16):
I'm Turi King and thank you for listening. If you have any thoughts or comments on this
or any other episodes, please contact us via our X channel @MilnerCentre.
For more information about the Milner Centre for Evolution, you can visit our website.