Ancient viral DNA helps human embryos develop

0:23Welcome to the 900th edition of The Nature Podcast. this time, how viral DNA seems central to human development. And a DNA-based computer that's powered by heat. I'm Nick Patrick Howe. And I'm Sharmini Bundel.

0:50This week, a study shows evidence that an ancient viral infection appears to be playing an important role in human embryo development, and perhaps even human evolution. Reporter Benjamin Thompson is here with more. When we think about our genome, we think of something that's uniquely us, human through and through. But in reality, a large portion of the human genome is made of DNA that originated elsewhere. Take retroviruses. These viruses infected our distant ancestors by slipping in their genetic material, and in some cases, it stuck around and was passed down through the generations. In fact, these endogenous retroviruses, as they're known, make up about 8 % of our genome.

1:39And if that sounds alarming, worry not, because these sequences have changed a fair bit over time, as Raquel Fuego from Stanford University in the US explains. Because there's been many years, they acquire a lot of mutations, so they are not able of functioning as regular viruses. And very interestingly, they form part of regulatory parts of our genome. Yes, you heard that right. These viruses now seem to have a function in our genomes. But how are ancient embedded viral DNA sequences regulating the activity of human genes? Well, because these viral genomes contain short DNA sequences that originally acted either as on-off switches or as volume controls, controlling the activity of viral genes back when they started out.

2:30But over countless generations, some of them have been co-opted to be used to control the activity of human genes. These short sequences are dotted all over the human genome and can play important roles. Evidence from other species suggests that these viral sequences are important in embryo development, but not a huge amount was understood about the role they play in humans. That's where a new paper by Raquel and her colleagues comes in. They have used CRISPR to selectively alter the activity of specific viral-associated sequences to see what their function is. Their work reveals that these genome insertions are vital for embryo development and might have played a role in making us, us.

3:16Now, studying human embryo development requires many practical and ethical considerations. So, in this case, the team made use of structures called blastoids that are derived from a type of stem cell. These stem cell-derived blastoids don't develop into embryos, but they mimic the ball of cells that forms around day five of development, and they allow researchers to investigate the human-specific processes going on shortly after fertilization. In this case, the team silenced the activity of a specific type of viral sequence inserted into the human genome called an LTR5HS. There are hundreds of these, but using CRISPR techniques, Raquel managed to target 90 % of them in one go.

4:06What we saw is that when we repressed these insertions, the blastoids can't form at all. But not only that, we see that it's actually dose dependent. If we silence more, we see less blastoid formation. If we silence less, they still have some capacity of forming blastoids, even though they are not normal. So this gave us a hint that they may be essential for blast information and, of course, potentially for human embryogenesis. Similar to what had been seen in non-human work, the team showed that these sequences are involved in controlling levels of gene expression. The team then deleted a handful of these LTR5HSs, one at a time, to see what effect they were having on specific genes.

4:51Deleting one led to a fairly drastic outcome. Just by deleting this LTR, we saw that the cells started growing very, very slow. And we said, OK, this is very interesting. What is this gene doing? It turns out that this specific LTR5HS in this location was essential for the expression of a gene called Zincfinger729. A gene that is, in evolutionary terms, quite young. Zincfinger729 is a primate-specific gene. So you can see it in macaques a little bit. but it's not present before that. The zinc finger 729 gene encodes a protein that switches on a bunch of other genes involved in basic cellular functions, things like metabolism and cell proliferation.

5:39So, important stuff. But usually, vital functions like these are evolutionarily ancient. So, they are controlled by evolutionarily ancient mechanisms. To have them under the control of a young gene doesn't make sense at first glance. And things get even weirder. The original viral LTR5HS sequence required for this zinc finger protein to be produced is an even more recent addition to our genome. So this one LTR5HS that had the phenotype that regulates this zinc finger 729 is only present in humans. So here we have this curious situation where critical cellular processes needed for early human embryo development are under the control of a relatively young gene only found in some primates, which is only switched on at high levels in humans thanks to a virus that left its mark on a specific place in our genomes, but not those of our closest ape cousins.

6:39Exactly how and when this system came to be is a puzzle, but Raquel and the team think that it might have endured because it gave cells in a developing embryo an advantage. It turns out that the ZincFinger729 protein is really good at binding to DNA and switching genes on. With its production turbocharged by the LTR5HS sequence, it may have been that this protein quickly outnumbered the original switches controlling the essential genes, resulting in higher gene activity and maybe more rapid cell growth and division. It's also relevant that this thin finger protein regulates cell proliferation genes.

7:18So we also speculate that the cells that acquire these LTR5HS elements could cycle faster, maybe out competing the cells that didn't have this LTR5HS in the embryo. They may have had like a competitive advantage. Raquel says that while the essential processes involved in early embryogenesis are similar across mammalian species, the pattern of endogenous retrovirus sequences isn't. It could be that these patterns are driving species-specific differences, and those could help sculpt what our species is. She hopes these findings could help give new insights into early human embryo development and the issues that can arise during this time.

8:02Vincent Pasch from KU Leuven in Belgium works on the early development of human embryos. He's written a News and Views article about the research and also stresses the need to understand this developmental stage as a first step that could ultimately result in clinical interventions for when embryos develop incorrectly. A lot of what we know about this process comes from animal studies. But if we want to start to try to understand what is specific to humans, then we must rely on experiments in humans. And this paper really uses these stem cell-based models, blastoids, to gain insight into human-specific features of development.

8:44So that's what I really like about this study, is that it really tells us something about what is specific to humans. Vincent says that experiments like these wouldn't have been possible were it not for the development of blastoid technology in the last few years. But notes that it's important to remember that these stem cell-based blastoids are analogs, not the real thing. So one of the main limitations of the paper is that the work is not done in human embryos. We still don't know if these human-specific elements and ape-specific genes are essential for the development of the embryo itself. So it will still be needed to validate these findings in embryos.

9:31These experiments are much harder to do, but they will have to be done. Raquel readily agrees that there may be differences between their model system and human embryos. But she says their work does offer further insights into our early developments and the role that endogenous retroviruses, these echoes of ancient infections lodged in our genome, play in making us who we are. How many human-specific or ape-specific functions may be regulating is an open question that actually now I will continue investigating because we have many more endogenous retroviruses family and many more researchers that are active.

10:09So I think we just like open a box and you know now we have the functional tools to ask these questions. That was Raquel Fueyo from Stanford University in the US. You also heard from Vincent Pasch from KU Leuven in Belgium. To read Raquel's paper and Vincent's News and Views article look out for links in the show notes. Coming up, researchers have been building DNA computers and now they've found a way to power them using heat. Right now though, it's time for the Research Highlights with Dan Fox.

10:43Longer whale mothers are more likely to produce female offspring compared to shorter mums, according to new research. For species whose males compete for access to females, scientists disagree about whether mothers with ample resources would benefit more from producing a son, who would have a better chance when competing against other males, or a daughter who could give birth. To investigate whether whales manipulate the sex ratios of their offspring, researchers analysed archival whaling data describing over 200 ,000 whale foetuses from seven different species of whale. The team found that in the seven species they looked at, longer mothers, who have more stored fat to power the demanding task of carrying and looking after young, tended to produce more female offspring.

11:33This pattern could slow these species' recovery from whaling, because whalers tend to target larger animals. If the surviving, shorter females give birth to more males, there will be fewer females available for reproduction. Read that research in full in the Proceedings of the Royal Society B.

11:55Corpse flowers are mammoth plants that bloom about once every decade, and when they do flower, produce a stench like rotting meat, perfect for attracting carrion-loving pollinators. Now, researchers have found that this stinkiness peaks on the first night of their spectacular blooms. Researchers studied the chemicals that waffled off a corpse flower named Cosmo when it bloomed in May 2024. The emissions included a range of organic sulfur compounds. The sulfur emission rates rivaled those of landfills and wax and waned throughout the night. Most of the emissions came during the first night of the bloom.

12:34That's when the plant's female flowers blossom, luring pollinators that become entrapped until they can depart the following night. The male flowers, which bloom on the second night, emit a sweeter, milder scent. Knowing how and why corpse flowers produce their stink can help researchers to better preserve these rare plants. Don't kick up a stink trying to find that research. It's published in Geophysical Research Letters.

13:05Let's face it, computers are pretty central to our lives. I certainly spend most of my life looking at one screen or another. And for good reason. They allow us to do all sorts of things that would have been near impossible decades ago. Podcasts probably would have been a hard sell in 1970. Don't come at me, radio nerds. But what if computers could be much, much smaller, or maybe even integrated into our own bodies? Imagine medicine that self-diagnose complex diseases and release just the right therapeutics. This is Lulu Chin, a researcher working on computers made from DNA. Now, while self-diagnosing medicine is a long way off, different teams have been working on using DNA as a computer for a while, because this is exactly the kind of application it promises in the long term.

14:07Rather than a competitor for regular computers, these molecular machines could be useful for doing complex calculations at the nanoscale. The way they work is instead of having physical switches that represent ones and zeros like a regular computer, DNA strands with different sequences and of different concentrations can represent those classic computing bits. Here single-stranded bits of DNA would be inputs, which would then hybridise as the computations occur, and then the outputs would be different strands of DNA. And because DNA is tiny and well at home inside cells, all this processing could happen inside your body or on the surfaces of devices.

14:54But before we get there, there's something else researchers need to resolve. Like all machines, these molecular machines need a source of power. So how can such small systems keep going, responding again and again to their surroundings without just quickly running out of energy? This is a problem that researchers have been working on. Such DNA computers can be run on chemical energy or powered using enzymes. But these methods only work until that energy supply runs out and can lead to the buildup of waste products that interfere with the computations. They're also hard to scale up, as each different kind of computation requires a different chemical environment to function properly.

15:39We wondered, could there be a universal power source for artificial molecular machines? Something as reliable as, say, ATP in living cells or electricity in our everyday devices? And the perhaps surprising answer we found is heat. In a paper in this week's Nature, Lulu and her team describe how to use heat to power DNA circuits. This method relies on two previous discoveries. The first is that DNA can be trapped in a state where it's primed, ready to start reactions and perform computations, but can't until it's released. So imagine two DNA strands that are meant to bind together, but one of them is being held back by a third strand that slows the reaction down.

16:30It's like a spring pressed down and held in place. The energy is there waiting. Now, by adding a catalyst strand that releases the block, the spring is suddenly let go. and then the DNA strands pair quickly. It's like unleashing the stored energy to drive the system forward. The second discovery is that heat could potentially reset this process back to its original state. So when you heat up a test tube of DNA and then cool it down, the molecules don't always settle into their most favorable arrangement. Instead, especially when they have strong folded structures, the heating and cooling can actually reset them back to the spring-loaded states, ready to release energy again.

17:18Getting this to actually work is quite tricky though, as each of the different strands of DNA, the inputs, the strands they interact with, and the outputs all have to work together in concert to perform their computations and reset without interfering with each other. So Lulu had to become a bit of a chemical conductor. We had to develop design rules so that different types of molecules could reset at different temperatures in the same heat-cool cycle without interfering with each other. This is a bit of like an orchestra tuning up. The violins, cellos and trumpets, they all need to land on the right notes without clashing, right?

18:01So the performance can begin in harmony. This involved a mixture of DNA editing and clever chemistry, but eventually the team were able to make groups of different DNA molecules that perform different functions. For example, some of the DNA molecules were able to amplify signals or annihilate them. Altogether, this soup of DNA could perform computations. To put it through its paces, the team decided to create a neural network, a computing system inspired by neurons to classify handwritten numbers, sixes and sevens. We showed that the DNA-based neuron network can indeed make the correct output classification decision and repeat that.

18:45The more important thing here is repeating that decision when the input pattern changes. For example, it changes from a six to a seven, so the system has to recompute and make the right decision again, right? So we showed that it was able to do that for 10 rounds and the performance was essentially the same. The first round and the tenth round, we could not really identify any drop in performance even after that many rounds of computation. DNA-based neural networks have been demonstrated before, but the key here was that it was repeatable and only powered by heat. That, Lulu thinks, could open the door to DNA computing that could recharge whenever needed.

19:26Well, I thought it was very elegant. This is John Reif, a computer scientist who works on DNA machines and wasn't associated with this paper. There have been a number of prior published papers on restoring DNA computations, including that I've published, but I would say that this is the most, I believe, scalable methods so far to spend. As this method relies on heat to recharge the DNA computations, it could potentially be scaled up. Lulu showed that many hundreds of different DNA strands could be mixed together, and as long as the chemical orchestra is tuned, they could all perform their reactions.

20:12In theory, you could do many more. John, though, did point out that this heat-based method of DNA computation is not quite as quick as ones that use enzymes. One could argue it would be nice to have systems that could execute a little faster, but it's beautiful work. This heat-based system also avoids the buildup of waste products, and Lulu was able to repeat computations many times. But there were limits. So we estimated that if we use this high temperature, then essentially the system will stop working at a maximum of a thousand cycles. High temperatures eventually lead the DNA molecules to break down.

20:53So Lulu and the team are investigating whether they could get this to work at lower temperatures. The other problem is that none of this happens automatically. It requires someone there to carefully calibrate the temperature over time. In the future, Lulu hopes that they can find a consistent cycle of temperatures to power the reactions and allow the computations to repeat. But for now, the team have shown that heat can, in principle, power DNA computers. Something that Lulu hopes others will build on to help create the DNA-powered future she imagines. What I hope will happen is that other researchers could utilise the principle of recharging and extend it from heat to other forms of universal energy sources such like light, salt, or even acid gradients like those across cell membranes.

21:47Because in principle, any of these energy sources could serve for the same purpose. So this way, molecular machines could be designed to recharge themselves using, let's say, one of a few options, whichever is available, and functioning well in a changing environment. That's my dream. that was lulu chin from the california institute of technology in the u.s you also heard from john rife from duke university also in the u.s for more on that study check out the show notes for some links finally on the show it's time for the briefing chat where we discuss a couple of articles that have been highlighted in the nature briefing which is of course nature's daily roundup of science news so nick you go first what have you got for us this time so this week i was reading an article in Nature about some big news.

22:37It's a trial that shows a way to potentially treat Huntington's disease. Oh right, well that sounds pretty important. Can you give us a Huntington's disease sort of recap primer? Yeah, so Huntington's disease is a neurological condition that tends to affect people between the ages of 35 to 55. That's when it starts its progression and it progresses over time and symptoms become worse and worse. It may start with just some slight motor issues and a bit of forgetfulness and then eventually it progresses to confusion and mood swings and eventually death. So it's a really terrible condition for the people with effects and their families and this story is about the clearest evidence yet that the disease progression can be slowed with a one-time gene therapy treatment.

23:26Oh right, so is Huntingdon's well understood in terms of its sort of genetic causes? Yeah so this is a condition that is caused just by a single gene so this is a gene called Huntington and in people with Huntington's it produces a protein that builds up and basically causes all these problems that we've discussed. So it's been a target of gene therapies and things like it for a while. There was a lot of interest in something called antisense treatment and this is where you put short strands of dna or rna to bind with the messenger rna in the cell to stop the proteins that cause all these problems from being made in the first place but that didn't end up panning out in a trial it ended up performing worse than placebo after showing some initial promise oh gosh okay which has now led researchers to instead look to gene therapy and the idea with gene therapy is you insert a bit of DNA into people's cells and then this represents a permanent change that can essentially tackle the condition.

24:34So in this case they used a harmless virus to deliver instructions to cells to make microRNAs, These are tiny little RNAs that then bind to the messenger RNA and they stop this dysfunctional protein being produced in the first place. OK, so it sounds simple enough in a way. And so this was a trial of how well this worked? Yes, this was a trial of 29 people to see how this works. And what I'll say is this isn't a cure. It doesn't stop it altogether. They've shown that it slows the disease progression over the course of three years. So compared to the control group, it slowed progression by 75%.

25:16So this was a small trial of 29 people performed by a company called Unicure. That's spelt with a Q, by the way, which is a gene therapy company based in Amsterdam. And so they've released a press release that has given the details of this trial, but this has not been peer reviewed yet. There's not a full study as of yet. and obviously there's only a few participants but the researchers interviewed in this article think that this is a big step forward so far as i've said there is no treatment for huntington's so if this is effective and is able to slow disease progression then that could be really impactful for the people that huntington's affects and also based on the strength of this data the company are looking to get regulatory approval next year so this could be something that has an impact quite soon.

26:05And because this was a short-term trial that's not necessarily to say that it can't work to slow the disease over a longer time or that you know further changes couldn't have bigger or more long-lasting effects? Yeah there's a lot more to be understood about how exactly this works and at what points it's best to do this. So this was done in people with early signs of Huntington's and so you could imagine that you could try at different points to see how it would affect the disease progression and as i say it's not a cure either like the disease is still progressing just much slower the issues for this i would say is this is quite an intense procedure so i've simplified it somewhat by saying you put the dna in the virus just whack it in there but it's actually and it's true that most people tolerated the therapy itself there weren't the major concerns that have plagued gene therapy for other conditions but you had to undergo surgery to do this so they had to drill holes in the skull and insert a cannula to then you know direct the treatment to the parts of the brain that were most effective and some people had headaches pain and other complications from the surgery but not from the gene therapy itself and because of this too this means that this treatment is very expensive.

27:29So according to this article, it would cost upwards of a million US dollars per person. So that would make it probably out of reach for a lot of people and for many health services. But this is a start. Perhaps in the future that could change as more research is done and there are different to try and give people this treatment. Well, that does sound promising, at least for future Huntington's treatment. thanks Nick I actually also have a story about DNA it's the theme of the week we've accidentally had a really DNA heavy week it's just these DNA scientists out there publishing their important DNA work that we really wanted to cover but mine is about mitochondrial DNA so I've been reading this article in Nature about a Nature paper about how or why exactly mitochondria do this thing where they sometimes kick some of their DNA out out of them and into the cell that they're in so we have mitochondria in all our cells they're there being the powerhouse of the cell as the saying goes but yeah it's been known for some time that there's this phenomenon where mitochondria will just sort of eject some dna from inside them this is their mitochondrial dna into the cell and that can kick off inflammatory pathways associated with aging which we don't love although you know this is it's an immune response so it is supposed to be defending against pathogens but you know we know that there's this chronic inflammation issue and these researchers wanted to figure out kind of why the mitochondria were doing this and sort of figure out a little bit more about exactly what's going on there yeah i mean it's a good question because not only is this a weird thing that they seem to be doing but also it seems to be bad for us so understanding that seems like a good step to me well it's not bad for the mitochondria so the mitochondria are getting rid of sort of damaged or problematic DNA that, you know, they want out.

29:24And, you know, this might be caused by certain drugs and the mitochondria being stressed. And it seems to happen in aging cells. So, you know, maybe this happens naturally during aging and it's just one of these things. But these researchers had a mouse model that they basically used to kind of look into exactly what was going on here. And it's a mouse model of these genetically engineered mice that lack a certain enzyme right this enzyme is the catchly named mgme1 and this enzyme actually is important for the mitochondria because it helps them make accurate copies of their genome so you can already see the link here but what is also known is that mice without these enzyme as they age their kidneys become inflamed so they're actually really useful models for inflammation.

30:17But looking into exactly what was going on, the researchers found that these enzyme-lacking mice had loads of these loose DNA fragments in the kidney cells. And they saw that these loose pieces of DNA fragments bound to and turned on an enzyme that's a known contributor to inflammation in aged tissues. So you can see that sort of connection there but they also then looking closer found out kind of what was wrong with these dna fragments like like why why is it why is the yeah exactly and basically the cells were lacking in the building blocks of dna so deoxyribonucleotides you sort of get a sequence of them put them together and then you make dna so obviously if you're replicating your genome you need the building blocks in order to make another copy of the dna but because they were lacking these certain building blocks they instead were using rna building blocks oh like rna ribonucleotides incorporating that it still works but it's not as good it sort of hinders future dna replication and it could be that this is why the mitochondria are basically getting rid of this dna because it has way too many of the wrong stuff in essentially.

31:39And so now they've figured out sort of why this happens. Do they think there's a way to prevent it or reverse it? Well, so the one thing to say is they were looking in a mouse model that was engineered to lack a certain enzyme and thus have more kidney inflammation, right? So we don't know for sure, is this relevant to normal physiology? And also, is this normal, I'll say in quotation marks, is this a natural part of aging? Or is there sort of specific conditions that then trigger this whole sequence that then is contributing to aging and inflammation, or as it's sometimes called, inflammation, which is apparently the chronic inflammation that occurs as people get older.

32:16Portmanteau there. But no, this is still, it's a cliche, but there's a lot to still figure out what's going on here. So for example, they don't really fully understand yet exactly how this discarded mitochondrial DNA then contributes to cellular aging and inflammation. so it looks like there's a lot of potential targets for if we wanted to try and slow down aging and prevent inflammation in future this study you know provides a lot more steps that we could look at and maybe potentially try to target but yes i think we're a bit of a way away from that at the moment no well fascinating stuff shaman i always love hearing about mitochondria it seems to be one of the staples of the podcast as is dna apparently it is now dna all dna all the times dna podcast now dna podcast now but i think that's all we've got time for on the briefing chat this week.

33:03If you've enjoyed those stories and you want to get more like them, you can sign up to Nature Briefing to get more of them sent to your inbox and you can check out the show notes for some links to do that. And as always, you can reach out to us, you can say hi to us, send us a message. We're on X, we're on Blue Sky, we're at Nature Podcast is our handle there, or you can just send us an email, podcast at nature.com. I'm Sharmini Bandel. And I'm Nick Petra-Ciao. Thanks for listening.

33:34Thank you.