eLife Episode 49: Pigeon patterning and stiff lungs
In this episode of the eLife Podcast, how early life copied its genetic information, the genes that give pigeons their plumage patterns, are the EU's expectations of field trials for a bovine TB vaccine realistic, and why a thinner, more heavily crosslinked collagen could hold the key to pulmonary fibrosis...
In this episode
00:36 - How early life copied itself
How early life copied itself
with Phil Holliger, MRC Laboratory of Molecular Biology, Cambridge
Arguably one of the most important questions in biology is how life got started on Earth. A popular theory is that it began with DNA’s single stranded molecular relative “RNA”. RNA can carry genetic information but it’s also capable of folding itself into complex three dimensional shapes that also endow it with enzymic - or catalytic - capabilities. The problem is how to make a tangled 3d structure copy itself. DNA and modern RNA systems do it by using proteins to temporarily straighten out the strands to make them readable, but early life wouldn’t have had those proteins. So how did it do it? Chris Smith heard what Phil Holliger has discovered…
Phil Holliger - So in present day biology, DNA stores information in the nucleus; proteins are the sort of bricks and mortar and the engine of the cell; and RNA is like a carrier of information from the information store to the machinery - to the proteins. But I think there's compelling evidence to suggest that early life must have been a lot simpler. So there must have been a primordial biology, which may have been based mainly on RNA, and that really raises to question how these RNA strands were copied in that primordial biology because in present biology both RNA and DNA strands are copied by proteins. So in that early biology RNA would have had to be able to copy itself.
Chris - There is a precedent for RNA having enzymic - and catalytic - function though isn't there, because there are things like ribozymes, which are machines in cells where they exploit the fact that RNA has a structure and it can guide chemical reactions; so there is a sort of precedent for having that view...
Phil Holliger - Exactly, so this is actually one of the reasons why people think that early biology might have been based on RNA that, unlike DNA which is more like an informational string, as a molecule RNA has this amazing capacity to both be an informational string and fold up into intricate three dimensional shapes which can have catalytic function including the capacity to actually copy other RNA strands.
Chris - So what's the problem with the hypothesis then, if RNA can do that it can be both informational and it can be biologically active as a catalyst, why couldn't early life of just use that?
Phil Holliger - When we were trying to build such a system where RNA could copy itself, we hit upon a snag. At the ribozyme, although it could copy RNA strands very well, if they were unfolded - if they were strings - once they had assumed a three dimensional shape they could no longer be copied. So you essentially hit a paradox where RNA needs to fold up in to a three dimensional shape to be able to copy itself, but once it has folded up it can no longer be copied! The paper is about how we have begun to find a solution to this paradox.
Chris - How did you?
Phil - The idea was the following: so maybe what we needed to do is to move away from trying to build something that was closely analogous to present day biology. In modern day biology, when DNA and RNA are copied they're copied in single letter steps. So single DNA or RNA letters are incorporated one by one by the copying enzymes. But our original ribozyme had exactly the same type of function like this - it would copy RNA in these single letter steps. But then when it when it hit a structure, kind of, it could not move forward. The solution we found is that if we move to three letters step - a triplet as we call it - and this is not used in nature at all to copy the DNA or RNA - when we move to triplets, this triplets - by virtue of binding much tighter to the RNA strand - they could actually straighten out the RNA template and as we approached a structure begin to invade and unravel the structure. And this way we found that eventually we could copy even extremely stable RNA structures - structures that have a melting temperature close to the boiling point of water.
Chris - So to put this another way it's almost like you've got a sort of a shirt with lots of wrinkles in it; and it goes through the iron and gets flattened out...
Phil Holliger - Yeah...
Chris - ...while it's momentarily flattened out, your enzyme can read that nice flat surface and make a parallel equivalent, or a copy, of that bit of the shirt and then it falls off the other side of the ironing board and gets all its wrinkles back again - so it goes back to its three dimensional shape, but you've got a copy in the meantime?
Phil - that's approximately right; unfortunately it doesn't actually wrinkle back - I think it will stay as a double strand, because that's probably the most stable form. So in fact this is one of the problems that we are tackling now; to set the system back to its beginning, we will now have to find a way to unravel that double strand again.
Chris - So you've solved one problem and created another!
Phil - Well I think that's probably always there. I think that we always were aware that that was part of the replication cycle. Now, in the cell kind of like there are some intricate molecular machines which have exactly that function to unzip double strands. We will have to find an alternative solution, but we're working on it!
Chris - Does this mean though, that in Darwin's little pool where there must have been these early forms of life if they were using RNA the way you speculate they were, that the supply of molecules that they must have had to feed their RNA replication machinery must have been these triplets - letters three letters - joined together in a sequence?
Phil Holliger - I think the most likely scenario is that you know there was a whole you know mix of letters there: there were probably single letters, double, triple letters and maybe some quadruple, quintuple letters. Presumably the longer, more complicated letters were rarer; presumably the copying enzyme would use whatever as appropriate. So you know and a nice kind of extended strand, you know, you could probably quite happily use the single letters; but then once it hits a structure then more extended letters - the triplets etc - kind of become more useful.
Chris - How did you make the discovery in the first place though. Was this just brute force - trying loads and loads of variants of the RNA system - until it worked?
Phil Holliger - We do not have currently the molecular understanding of RNA to do bespoke engineering of let's say designed catalytic activity of copying function of exactly what we want. So what we use is an evolutionary process where we force the RNA to begin to use the substrates that we like it to use - like the triplets - by iterative cycles off of evolution in the test tube and this is a very very powerful process.
Chris - So if you've got this right., what are the implications for early life about four billion years ago?
Phil - So potentially there is really no conceptual problems with building such a system from scratch that could copy itself and with self replication presumably the copying process would make a few errors. So you have mutation as well and with self replication and mutation you would have evolution and that would really get the ball rolling towards ever more efficient self replication and you could see how that process would take off and lead to more and more complex things...
07:50 - Pigeon patterning
with Mike Shapiro, University of Utah
How did pigeons come by their plumage patterns, and what genes control the process? Speaking with Chris Smith, Mike Shapiro explains how he has taken advantage of the diversity created by natural evolution to find out...
Mike - We were interested in trying to understand the DNA level changes that lead to animal pigment patterning. We've known for a long time the types of genes that control that type of pigments that are generated: blacks versus browns versus reds versus yellows. But an outstanding mystery in pigmentation genetics is how those pigments know where to go and how patterns are generated - spots and stripes and so forth
Chris - In terms of actually how the pigments are added to the skin, or in the case of birds, feathers, how is that achieved?
Mike - There are specialised cells that are called melanocytes that produce the pigment granules and those granules are exported onto the surface of the skin or a feather. There is a fair bit known about the recipe for the granules themselves but where are those granules go has remained a mystery for quite a while. And now, by taking advantage of things that are not laboratory mice were able to take advantage of evolutionary diversity to try to answer those questions.
Chris - So what have you done?
Mike - We've focused on pigeons as a model to try to understand how pigment patterns are generated and in particular in pigeons there are a few basic colour pattern types; and we've known for a long time - mostly through the work of pigeon hobbiests - that the genetic basis of this is relatively simple. That is, very few genes - possibly even one - have a major effect on how stripes are laid down versus a pattern called checkers because it resembles a checkerboard.
Chris - So you're saying that people who've bred pigeons over decades knew how to breed for certain traits, but they had no idea what the molecular genetics of this was?
Mike - That's correct; and I would extend that beyond decades to millennia, pigeons domesticated probably 5000 years ago and people have been raising them in captivity ever since. So we know from a long history of pigeon breeding a lot about the number of genes that go into making certain traits. But, as you pointed out, we know very little about the molecules involved.
Chris - So how did you track them down?
Mike - One way that geneticists track down genes that are responsible for particular traits or particular diseases is to identify groups with and without the trait that we're interested in. So in the case of pigeons the ancestral colour pattern is a pattern called "bar" - it's basically two stripes on the wing. And we compared the genomes of birds that have that bar pattern to pigeons that have what we call a checker pattern where there is more pigment on the wings in sort of a checkerboard-like pattern, and we try to identify what was different about their DNA sequences. And when we did that we identified just one place in the entire genome in over a billion base pairs of DNA that was different between those two groups.
Chris - And in what way is it different?
Mike - When we scanned through that region of the genome, looking for mutations that might affect places that are coding for protein, we didn't find anything that was consistently associated with having a bar or a checker pattern. However the mutations around those genes were indeed consistently associated. There is one gene in particular that piqued our interests. It's a gene called NDP, which stands for Norrie Disease Protein, and even though the gene itself wasn't different between the birds, the different patterns, the region very close to it was very different between the birds that had checkers and bars; and through some follow up experiments what we found is that that region was associated with differences in the way the gene was expressed. So rather than changing the gene itself, what appears to have happened in these pigeons is the way that gene is controlled - the way its expression is regulated - is different. In addition to that, this same region that differs between the bar and the checker birds was duplicated in some of the birds with darker colour patterns, so it appears not only that this region differentiated checker birds from the bar birds, but the more copies of DNA that the checker birds had in this region the darker they were.
Chris - So that's like a gene dosage effect?
Mike - We think it's something like that, and again it's not the gene itself, but when we assay how the gene is expressed the more copies of DNA the birds have in this region the higher the expression is in the feathers.
Chris - Now why do you think that there are these different animals in the first place. And do they actually have any impact on how the birds fare in the wild?
Mike - It appears that they do. Interestingly there is some evidence from ecological studies not done by us that the birds that have the checker pattern fare better in specific places and those specific places, oddly enough, are urban environments. The natural habitats for these birds are rock cliffs. You can think of like the Cliffs of Dover in the UK. They also live in deserts in the Middle East - that's one of their native habitats and part of their native range. But in urban habitats they're obviously more recent introductions, and the birds with the checker pattern fare better reproductively: their offspring fledge better from the nest; they survive to adulthood. Interestingly the converse is true in some rural environments: so in harsh rural environments such as the Faroe Islands, there is not much food available. The birds of the checker pattern don't fare as well. And one reason that may underlie this is that that checker birds are not as good at putting on fat over the winter as the bar birds. So this is a huge disadvantage in places that have really harsh winters but in cities, where there are French fries and other things available around, these checker birds do just fine and, again, in some cases they will breed year-round whereas the ones with the bar pattern will stop...
13:54 - Bovine TB vaccination
Bovine TB vaccination
with Andrew Conlan, University of Cambridge
Mycobacterium bovis - which causes bovine TB - is a close relative of the TB bacterium that causes disease in humans; and because it can infect us, cattle are rigorously monitored. The problem is that there’s also an environmental reservoir of the infection: wild animals, including badgers, can carry it and - some suspect - transmit it to cows. For this reason there have been calls to cull badgers. But is this justified when there is, potentially, a vaccine we could use? Unfortunately no one knows how effective that vaccine would be, and before we can use it we need the EU to change the law. They’ll only do that if there’s adequate data to defend the decision. So are their demands feasible? Speaking with Chris Smith, Andrew Conlan, from the University of Cambridge Vet School, models disease dynamics...
Andrew - For the past 20 years, the long term strategy of the UK government for the control of bovine tuberculosis in cattle has been the development and use of an effective vaccine. Bovine tuberculosis, although primarily affecting cattle, is a disease that can affect anything warm and furry, including humans. However it's also known to infect many wildlife species - most controversially in this country, the badger. And it's the controversy over the culling of badgers to control the spread of bovine TB which was one of the main reasons that cattle vaccination has been seen as the final end game for control.
Chris - Why should developing a vaccine for an infectious disease be anything other than a good thing?
Andrew - So vaccines are, as we know, the most efficient way of controlling infectious diseases because they don't just protect the individuals that we vaccinate but they also reduce the rate at which the disease is spread in the population - so called herd immunity effects. The problem with bovine TB is the only vaccine that we have - which is the same BCG vaccine that we used in humans - has limited efficacy it's not quite clear if it can demonstrate herd immunity effects in populations, and more worryingly, it interferes with the basic diagnostic test that is currently the cornerstone of control; and in fact, legally, the definition of disease is reaction to this test which means if we were to vaccinate animals at the moment with the BCG vaccine they would test positive and then we would have to kill them.
Chris - What can we do about that?
Andrew - We could try and develop a vaccine which doesn't cause the reaction with the skin test. However, the problem with that is almost every vaccine which has shown any useful efficacy is in some way related to or based upon BCG. The second thing we can do is we can replace the skin test. So the current test is based on basically boiling up large amounts of the bacteria to get a soup of proteins which we then inject into animals and see how strongly the animal reacts to them. If we could replace that test with a more defined test which just uses proteins which we know are in the wild bovine tuberculosis bacterium but not in the vaccine strain then we could use the vaccine without interfering with the test and slaughter programmes that we use at the moment.
Chris - So what's the question that you've been trying to address with this paper?
Andrew - This paper came out of our call by the government to design field trials for the evaluation of BCG vaccine in cattle. This was prompted by a chain of events started by Brian May who is not only the lead guitarist of Queen but is now a very prominent animal rights campaigner in particular with respect to badger conservation. He visited the EU and made a lot of noise about why we weren't using vaccination. I don't know if these were actually connected but about a month later there was a letter from the EU which set out what they would require the UK to show in order for there to be a change in legislation to allow the use of vaccination. What they said is that we would have to do field trials. We have to demonstrate the efficacy of the vaccine on the farm in the field and we would also have to show that any replacement test to the tuberculin tests - or so-called DIVA tests - that stands for differentiate infected from vaccinated animals - those DIVA tests would also have to be validated at the same time under field conditions. As part of this, I had already developed some mathematical models which describe the spread of bovine TB, and the possible use of vaccination. So I was involved in the successful tender to design those field trials and show how large they would have to be in order to address all of the questions that the EU wanted to see satisfied.
Chris - And what factors have you taken into account when you've drawn up this model?
Andrew - The very basic assumption is that bovine tuberculosis is an infectious disease, and we built into the model which describes the transmission between cattle and also includes the possibility that infection comes in from other sources; and it also describes how testing is currently used in order to remove infectious animals from herds and how vaccination may interact with the testing regime.
Chris - Now if you take all of the assumptions you've made and you take the predictions of your model how realistic are the expectations and demands of the EU for the implementation of vaccination? Is this something we could do?
Andrew - The models suggest that trials along the lines that the EU is suggested would be extremely challenging. They would have to be of a scale that is almost equivalent to doing a deployment of the vaccine at the national level. So to understand the basic question of whether the vaccine works or not, that's actually quite straightforward. It would take model suggests about hundred herds. However to address this specific question of showing that the vaccine actually reduces transmission, that would require a much larger scales of the order of 800-1000 herds. The reason for this is quite simple it's that the models suggest that even if there was a very strong effect of the vaccine reducing transmission, we just wouldn't be able to measure that effect on a reasonable time scale of a field trial. So we only have around about five years to do a field trial and because of the very slow transmission of bovine TB, there just won't be enough data to quantify that effect.
Chris - So is our answer to the EU then that this is just not practical?
Andrew - We make a series of positive steps that could be done; so in particular if this question about what effect the vaccine has on transmission is necessary to be asked - and actually the EU has good reasons for asking for this, because the big concern with introducing vaccination is that if you nobble the skin test and you start missing truly infected animals but the vaccine doesn't work, you could actually see an increase in disease even though there's a reduction in the apparent level of disease, so there's good reasons for it - but we would argue that a field trial is not the best way of doing it. In fact we propose some simple natural transmission studies - experimental studies - under much more controlled settings, which can be used to address this specific question and these are experiments we're already planning to do in India and in Ethiopia in separately-funded projects; but we would argue that if this question is important - and it is important - that is the more appropriate way to ask it than to field trials...
21:35 - Thinner collagen linked to lung fibrosis
Thinner collagen linked to lung fibrosis
with Mark Jones, University of Southampton
Scientists at Southampton University have some good news about the problem of lung fibrosis, where patients’ lungs become stiff and gas exchange is severely impaired. Prevailing wisdom was that this was caused by a build up of excessive connective tissue and indeed you can see that in end-stage disease. But by studying biopsy specimens from patients early in the disease course, as Mark Jones explains to Chris Smith, he's discovered that the structure of the collagen in their lungs is different. And he thinks he can intervene with drugs to retard the progression…
Mark - I'm a lung doctor and one of the conditions that I see a lot of is lung fibrosis and in particular a type called idiopathic pulmonary fibrosis. Unfortunately we don't have very effective treatments and one of the questions I really wanted to start out to address was to understand exactly what's happening at the time that fibrosis develops. So how does it start and what progresses over time. And the thing that we chose to focus on was collagen.
Chris - Now if one looks at the lungs of a person diagnosed with fibrosis, what do you see, both macroscopically but also down a microscope?
Mark - We can look with a CT scan. We see that the bottoms of the lungs are scarred so you see lots of little white dots on the scan which shouldn't be there and there's just far too much protein there. And we think that that's extracellular matrix. If you look even closer down with a microscope you see more and more of this, these changes, but we really don't understand why they start.
Chris - And when you look at the composition of those proteins the building up is it just more of what we should have there, or is the distribution and the different concentrations of the different components wrong?
Mark - So we don't really know. There's a concept that collagen, which is really thought to be the main building block of the lungs, from a structure point of view, is increased over time. But one of the challenges is how you relate that to other components. We don't have a very good way of measuring and comparing.
Chris - So what did you actually do then who did you study, and how did you try and get to the bottom of what is actually causing this?
Mark - We wanted to look at collagen, and it's quite complicated in the way that it can assemble. So you can start with an individual "brick" - or collagen fibril as we can call it - and each of those is around one 1500th of a human hair - when they then join together they get bigger and bigger. And so when we want to look at it, one of the questions we can ask is "well, is there more collagen there, or is actually how the collagen are joining together changed? And one of the things you can look at is cross-linking.
Chris - Because, obviously, these diseases are complicated and what starts the process may not actually be where you end up, as in you could end up with lots of collagen in someone who has end-stage disease but that wasn't what caused the disease to start happening in the first place. So can you get a handle on that?
Mark - Exactly. So one of the ways that we can diagnose fibrosis is by taking a surgical lung biopsy we were able to use the tissue which wasn't needed for diagnosis to actually look at the changes which had happened in the lungs at time of diagnosis. Normally studies have looked at very much end stage tissue. So we were very fortunate to be able to look at this tissue at time of diagnosis. What we found actually the amount of collagen hadn't increased, but when we looked at each collagen fibril there seemed to be fundamental differences. They're smaller and the way that a join together seems to be more intricate - so they have more cross-links. One of the reasons that that's important is actually that can affect how stiff the lungs become; and we know that the stiffer an area is, the more likely fibrosis is to progress over time. And so the question then is what's driving those changes?
Chris - That's interesting isn't it? So the collagen - almost the recipe that's being used in the way it's being assembled - has changed between a healthy lung, or someone at the very beginning of this disease and someone who's got end stage disease. It's not the overall amount of collagen, it's the way it's wired up?
Mark - Exactly and that's kind of a really important concept to think about, because actually when we then look at treatments for fibrosis very often in the studies that we're doing before we looking at trying a drug in patients we need to think well what are the measures we need to look at? Very often we actually look at the amount of collagen as being proof that yes there's a possible drug and it possibly might actually work in patients. What we saw was that actually that amount of collagen seems far less important and what we really need to focus on is how is the colalgen structure changed. Because we think that might actually be something which might have far more benefit from a treatment point of view.
Chris - Apart from finding that the strands are physically different - in the sense they're are a bit smaller and they are stiffer - how have you managed to unpick - or untwist if you like - the strands to work out why they are like that?
Mark - One of the kind of challenges when it comes to studying something like fibrosis is that we can look in the tissue. But we know that's just one time point and really to try and unpick what might be the underlying mechanisms of this we need to be able to look in models where we can look at changes over time and so we were able to use cells from patients and actually start to grow little areas of fibrosis and we can look at the colalgen over time. We were then able to work with pharmaceutical companies with drugs which targeted some of the changes that we identified, and were able to then start to unpick how those changes might relate back to the patient.
Chris - So do you think this is almost a bit like a prion disease in the sense that you get an abnormal folding of a protein that then makes more of the same and that once the collagen starts to abnormally behave it then encourages more colleges to behave badly?
Mark - Exactly yes. I think this type of feedback loop which is self-perpetuating process is definitely something that we saw from our data and other groups have shown and pharmacology that you think you can throw at this.
Chris - How does it work and is it very effective?
Mark - We were able to look at specific types of cross-links that we thought were altered and we can use a specific drug which is able to inhibit those cross-links and I think the studies that we did it works very well and we seemed to get a resetting of fibrosis. But the next step is really to translate those treatments from our studies into human studies. And so the very first part of that - the Phase 1 trials - are in progress at the moment. So hopefully there is the potential for this to be of real benefit to patients in the future...