3.8 million year old protein

When analysing human remains, what if we could look at the genetics themselves? Well that may actually be possible because Matthew Collins and his team have discovered that vestiges of the original tissues can still remain inside a fossil for millions of years, stabilised by their surroundings as he explained to Chris Smith...

Matthew - For me that's the really exciting new findings that have come out in a very short period of time in fact. It's only really in the last decade that we've been able to systematically start to recover molecular data from these fossils.

Chris - By molecular data, just explain what you mean by that?

Matthew - Well, I guess we're all familiar with the idea of DNA. We are all made of our DNA and the discovery was made more like 20 years ago now that some DNA was surviving in some fossils. They were relatively young, so that was from the latest Neanderthals and short, short fragments were surviving.

But then really dramatic changes in technology have developed so much that we can begin to recover far more detailed genetic information. So we've now got complete genomes of Neanderthals and of Denisovans, and what's really remarkable about the Denisovan genome is it comes from nearly a fingernail bone and a tooth. So tiny amounts of material are releasing this wonderful rich molecular data.

The problem with that is that DNA record isn't going that far back in time. So we've already talked about the long 6 million years of evolution. Sadly the DNA record seems to fall away within about four hundred thousand years. So you've got the neanderthals, you've got moderns, but you don't go any further back.

Chris - Why?

Matthew - Well the problem with DNA is that it's a very fragile molecule. Not that fragile because it survives for quite a long time, but it's quite a complex molecule and some of the bonds that create DNA are relatively high energy bonds that are quite easy to break. So what we've been trying to do is look at other molecular sequences. DNA makes a thing called proteins. When you look around, your hair, you skin, that is all protein.  So what we have been trying to do is get information, sequence data, from the proteins.

Chris - So just to summarise this then. DNA is a recipe book in a cell. That recipe tells cells how to make proteins so, if you know what proteins a cell is making you can sort of work out what genes must have made it. And you're saying because DNA is more flimsy than proteins, instead of just looking for just DNA, which we can only go about half a million years or so, if we go looking for the proteins, actually we can go back potentially a lot further?

Matthew - Yes, and half a million years is caution because most of the DNA work that's been done has been done in very cold places, and that's not where humans evolved.

Chris - Tell us then, where have you been working and trying to pursue this?

Matthew - So where we've is gone to the classic hominid sites, but we haven't been looking for hominids themselves. We've gone to the humble ostrich eggshell. And the reason for that is because there's a lot of them about, so for developing new techniques it's relatively easy to get hold of the material. Because these sites are so well studied, we have very good dating control on those sites and the eggshells are fairly large. They have proteins trapped inside them and we've been able to develop protocols to get those proteins out.

Chris - You've been to places where people like Lee dig up the remains of our human relatives and they date them very carefully. So, if you find a piece of ostrich egg there in the same sort of context as Lee's digging up stuff he's dated to within an inch of it's life, you could be reasonably confident that your piece of ostrich egg is that date?

Matthew - That's right. The only places that people care to date that kind of material where hominids are found and we've gone to some of the classic hominid sites.

Chris - And how old are these bits of ostrich egg you're looking at?

Matthew - Well, we've deliberately gone progressively back further and further in time because we did not know at what point the sequences would run out. And that was the big surprise, as I say, they went back, and back, and back, and back and, in fact, we still have sequences in the oldest thing that we've so far analysed...

Chris - Which is how old?

Matthew - 3.8 million years.

Chris - So you've got a bit of ostrich egg from 3.8 million years ago?

Matthew - Yes we have.

Chris - And you can get the proteins out of them?

Matthew - And we can get partial protein sequences out. We don't get the full proteome of the ostrich, we get partial protein sequences.

Chris - So, putting this in perspective, when that ostrich laid that egg three point eight million years ago, material that it deposited in the egg is still there and you can extract it?

Matthew - It's DNA told it to make proteins, it made the proteins, the proteins made the eggshell, and some of that sequence still there. I mean it's really remarkable, really remarkable.

Chris - And you can get it out, and can you read the amino acid building blocks that make up the protein so you can work out the genetic code of the ostrich that laid the egg?

Matthew - Yeah. We can only read part of it because not all the protein survives and that's one of the things we've been trying to understand which parts survive and which don't. But, absolutely, we can read the sequence and then we can compare the sequence with the modern ostrich.

Chris - So Lee - are the materials that you are extracting from sites in South Africa of sufficiently high quality in terms of their preservation that you could have a conversation with Matthew, for example, and potentially apply his techniques because the overlap in time with your material?

Lee - Why, we've already had that conversation just this afternoon.

Chris - Oh, right.

Lee - I mean it sort of remarkable doing this in Cambridge. I was at the Eagle pub last night and that's where this all actually started, You know the idea that we're now combining what we thought to be separate streams that would never reach each other. DNA was like a railroad track running next to a river which was the hominid fossil record and protein was off on the other side. That we're now seeing these incredible advances that are bringing all three of these critical sciences, each needing each other, into building a picture of the complexity of where ostriches come from, but also where our lineage comes from and what it did at any one time.

Chris - The way I was getting at this is that if Matthew can work out at the least the partial genetic code of an ostrich, can we pull the same stunt for Homo naledi, one of the species that you have uncovered? Do you have the material that could be amenable to this?

Lee - I can tell you we will try it. I mean I think all of us are concerned about the destruction of relatively precious material, for these sort of states particularly in the early stages. But the fantastic point is that as we are beginning to push people further and further by saying "no you can't until you get the technique on smaller amounts." But it will happen and we are going to learn extraordinary things I would predict.

Chris - How much stuff do you need Matthew to do this?

Matthew - Well I mean Lee's quite right, we should not be destroying samples until we know exactly what we're doing. We're working with pretty small samples of ostrich eggshell at the moment considerably less than a teaspoon full of salt, but still I think we can get better. And what we want to try and do is get as much information as we can out of these samples. We're still in the stages of developing the techniques before we work on these really valuable fossils.

Chris - And what has changed in recent years that means that you are now able to do this where you couldn't before - how do you do it?

Matthew - We have ridden on the back of medical technology in a sense. People, once they'd sequenced the genome, they became interested in what the DNA was doing, so looking for the proteins, and so the medical world has been developing more rapid, and more sensitive instrumentation.  And what's been remarkable for us is it wasn't possible with the instruments from the last generation, it's only the last couple of years the machines have been good enough to detect these tiny amounts of proteins, and also to map the damage on the proteins so we can actually tell and ancient from a modern sample.