December 04, 2013

400 thousand year old human mtDNA from Sima de los Huesos

It will come to no surprise to people who noticed an earlier paper on cave bear mtDNA from Atapuerca that the folks at the Max Planck Institute would try to do the same for the plentiful human remains found in the Pit of Bones.

A new paper in Nature reports their success, and overnight increases by an order of magnitude the time depth for which we now have human mtDNA from what is commonly designated as Homo heidelbergensis, from right in the middle of the Middle Pleistocene. Obviously, this opens new vistas for archaeogenetic research, making it possible to directly look at early pre-sapiens forms of humans, and not only on their final forms prior to their replacement, the Neandertals and Denisovans.

The most impressive aspect of the new paper is most likely the technical challenges that the researchers must've overcome to achieve this result. The cave bear DNA showed that this was possible, but human DNA adds an additional complication in the form of contamination by a closely related species, us.

But, the new evolutionary result which will interest those of us not interested in the minutiae of biomolecules will no doubt be the fact that the Sima hominin's mtDNA formed a clade with the much more recent Denisova girl.

Until now, we knew that Neandertal mtDNA grouped together and so did modern human mtDNA. The two groups shared a Middle Pleistocene common ancestor and a much more distant common ancestor (~1 million years) with the mtDNA found in Denisova. The new Sima specimen shares descent from Denisova. This is important because it shows that whatever archaic human population the Denisovan mtDNA belonged to also extended to western Europe. And, surprisingly, the Sima specimen did not group with Neandertals, as might be expected because of the incipient Neanderthaloid morphology of the Sima hominins which has been a matter of controversy as it pushes back the evolutionary lineage of H. neandertalensis deeper into the Middle Pleistocene that some researchers accept.

Before this paper, it was believed that H. heidelbergensis evolved somewhere (perhaps Near East or Africa), a subset of it evolved to H. sapiens in Africa, and a different subset evolved in Eurasia, leading up to H. neandertalensis in the west, and unknown forms in the east, of which the Denisova girl was a matrilineal descendant. The next question is: when did Neandertals and Neandertal mtDNA appear in Europe? 

It can now be hoped that such questions will be answered directly. The Sima individual studied in this paper is not some frozen specimen from the Arctic, preserved by a freak accident in pristine form for hundreds of thousands of years, but a person who lived in Southwestern Europe. I am fairly sure that this won't be the last really old human we see a paper about in the coming years. Human mtDNA used to present a simple picture at the time of the discovery of African mitochondrial Eve: the deepest splits were in Africa and Eurasians belonged to a subset of African variation. But, as more and more archaic Eurasian mtDNA is sampled, it now appears that modern human mtDNA is a subset of world human mtDNA whose deepest splits are in Eurasia, and the next deepest splits are in Africa. Obviously, this may be a consequence of the fact that archaic human mtDNA has only been sampled from Eurasia, for factors relating to DNA preservation. But, it is nonetheless interesting to wonder where on the tree the mtDNA of archaic Africans would fall.

Nature (2013) doi:10.1038/nature12788

A mitochondrial genome sequence of a hominin from Sima de los Huesos

Matthias Meyer et al.

Excavations of a complex of caves in the Sierra de Atapuerca in northern Spain have unearthed hominin fossils that range in age from the early Pleistocene to the Holocene1. One of these sites, the ‘Sima de los Huesos’ (‘pit of bones’), has yielded the world’s largest assemblage of Middle Pleistocene hominin fossils2, 3, consisting of at least 28 individuals4 dated to over 300,000 years ago5. The skeletal remains share a number of morphological features with fossils classified as Homo heidelbergensis and also display distinct Neanderthal-derived traits6, 7, 8. Here we determine an almost complete mitochondrial genome sequence of a hominin from Sima de los Huesos and show that it is closely related to the lineage leading to mitochondrial genomes of Denisovans9, 10, an eastern Eurasian sister group to Neanderthals. Our results pave the way for DNA research on hominins from the Middle Pleistocene.

Link

22 comments:

hairysteve20 said...

Fascinating that they seem to be more closely related to Denisovans. Of course further DNA tests may throw up a great deal of diversity in this group. Perhaps, despite their basic technological level, Ancient humans were very mobile. Maybe they didn't have homogenous, neat 'ranges' but were mixed all over the place.

Perhaps technological assemblages were consistent across areas and not species.

If the clade containing this specimen and Denisovans had an input from a hominid '??' is it looking more likely that it was Erectus, since traces of it may show here and in the Denisovan sample?

More questions than answers but I can't wait for more work in this area, it would be great if the Red Deer Cave people were sequenced.

WarLord said...

Now, what about the Mungo Man from Australia? Will they retest his DNA once more?

Mark D said...

This is getting a lot of press coverage today, eg.

http://www.nytimes.com/2013/12/05/science/at-400000-years-oldest-human-dna-yet-found-raises-new-mysteries.html?_r=0

If mtDNA from 400,000 years ago can be found, then the thousands of specimens from the Lower Palaeolithic, such as Boxgrove Man, now in storage everywhere can be analyzed.

Nguyen Van Phauc said...

From the NYT article on this:

"Beth Shapiro, an expert on ancient DNA at the University of California, Santa Cruz, favors an even more radical possibility: that the humans of Sima de los Huesos belong to yet another branch of humans. They might have been a species called Homo erectus..."

Any reason to think that this is not erectus DNA?

Interesting to think about the implications if it is.

terryt said...

"The new Sima specimen shares descent from Denisova. This is important because it shows that whatever archaic human population the Denisovan mtDNA belonged to also extended to western Europe. And, surprisingly, the Sima specimen did not group with Neandertals, as might be expected because of the incipient Neanderthaloid morphology of the Sima hominins which has been a matter of controversy as it pushes back the evolutionary lineage of H. neandertalensis deeper into the Middle Pleistocene that some researchers accept".

Yes. Not really what we were expecting at all. What it suggests to me is that ancient human movement was much more common and much more extreme than what is usually assumed. As demonstrated further:

"Human mtDNA used to present a simple picture at the time of the discovery of African mitochondrial Eve: the deepest splits were in Africa and Eurasians belonged to a subset of African variation. But, as more and more archaic Eurasian mtDNA is sampled, it now appears that modern human mtDNA is a subset of world human mtDNA whose deepest splits are in Eurasia, and the next deepest splits are in Africa".

Ancient movement includes both movement out of and movement into Africa.

eurologist said...

I am happy to see that heidelbergensis is again more widely seen as the main "solution" to this problem -- as I have suggested for a long, long time.

Basically, outside English speaking research areas, the accepted position for the past over half a century has been that pretty much any European fossils and stone tools from about 800kya to ~300kya were heidelbergensis. From ~300kya to 150kya Neanderthal features developed*, with after 150kya only Neanderthals left; but with a west-east cline in characteristics petering out in E Europe (no Neanderthals in the Levant or W Asia until ~60kya). But, since there are sporadic fossils and stone tools in W/C Asia and Siberia from that time, the most parsimonious explanation is that non-Neanderthalized heidelbergensis lived on, there, and actually at one point (~200kya?) made it all the way to E Asia (and likely acquiring E Asian erectus contribution, explaining some of the Denisova and extant E Asian oddities).

Now, since the climate was quite optimal during most of heidelbergensis span in Europe, and a very similar development in cranial features and volume is seen in Africa, it makes a lot of sense to assume that this was a common development with significant genetic exchange between Africa, the Levant, W Asia, and Europe.

In view of that, the mtDNA results can be best explained by assigning to (most of) heidelbergensis (and thus Sima de los Huesos and Denisova) late European erectus mtDNA (which is part of the European contribution to the heidelbergensis development), while sapiens and Neanderthal are part of the African erectus (ergaster) female lineage and document the gene flow from Africa into heidelbergensis shortly before Neanderthal divergence.

*of course, the Spanish finds seem to indicate first Neanderthal-like features a little earlier

terryt said...

"Now, what about the Mungo Man from Australia? Will they retest his DNA once more?"

I hope so, although I don't think they can. I'm fairly sure the remains have been handed back to the Aborigines.

"I am happy to see that heidelbergensis is again more widely seen as the main 'solution' to this problem -- as I have suggested for a long, long time".

I too think it's time to resurrect heidelbergensis. If the Denisova' element is heidelbergensis, though, it makes it very unlikely that the Denisova admixture in Papuans is Southeast Asian. Heidelbergensis made only a very minor impression in SE Asia, if any at all.

"the mtDNA results can be best explained by assigning to (most of) heidelbergensis (and thus Sima de los Huesos and Denisova) late European erectus mtDNA (which is part of the European contribution to the heidelbergensis development), while sapiens and Neanderthal are part of the African erectus (ergaster) female lineage and document the gene flow from Africa into heidelbergensis shortly before Neanderthal divergence".

Fits the facts as I see them.

"This is getting a lot of press coverage today, eg."

Thanks for the link. Quote from it:

"It is possible, for example, that there are many extinct human populations that scientists have yet to discover. They might have interbred, swapping DNA".

That has long been my belief. And:

"Denisovans were believed to be limited to East Asia, and they were not thought to look so Neanderthal-like".

Since the dicovery of the Denisova line I have had difficulty accepting it was a Southeast Asian line at all. Another quote:

"One alternative explanation is that the humans of Sima de los Huesos were not true Neanderthals, but belonged to the ancestors of both Denisovans and Neanderthals. It is also possible that the newly discovered DNA was passed to both Neanderthals and Denisovans, but eventually disappeared from Neanderthals, replaced by other variants".

My 'guess' is that the Denisova population is more ancient (as we sort of know) than the Neanderthals. And they did not originate in Spain but entered it some time after their first appearance. Where from? is the question that remains.

Rokus said...

Anyway, the people that brought Denisovan DNA to Australia and the rest of Oceania were very early AMH and not some H. Erectus. Also the Australian mtDNA of LM3 and the hominine that introduced the chr.11 Insert must have belonged to this same Sima de los Huesos/Denisova clade. This insert you can still find at saturation level all over Eurasia/America and also at high frequencies in Africa, ie. indicating a northern origin. Now parsimony makes it ever less likely that Denisovan DNA was also an East Asian phenomenon, the apparent admixture of Denisovans in MA-1 should be taken more serious as an indication of a northern origin. However, we might start to respectfully disagree on the origin of this Denisovan-like admixture in MA-1 and apparent close AMH kin that had a more abundant Denisovan admixture. Maybe the source was much more to the west, and much more “modern” than the Neanderthal traditionally attributed to this region. If Neanderthal really didn’t originate in Europe, then what happened to their predecessors? Northern Europeans still have less Neanderthal admixture than could be expected, and also carry some ancient genes that can’t be found in Neanderthal. Some European hominine seems to have survived longer than being recognized.

bicicleur said...

is the 'figure 4' phylogenetic tree included in this blog correct or is it still subject to further investigations? if it its correct, is there a time-frame for the the splits? when you look at the tree, many theories can allready be ruled out and we have to find a new explanation re the origins of Neanderthals

terryt said...

"the people that brought Denisovan DNA to Australia and the rest of Oceania were very early AMH and not some H. Erectus".

Yes.

"Now parsimony makes it ever less likely that Denisovan DNA was also an East Asian phenomenon, the apparent admixture of Denisovans in MA-1 should be taken more serious as an indication of a northern origin".

A position I have maintained since the Denisova element in SE Asian/Papauns was first discovered. Long before that I had come to expect that mt-DNA N and Y-DNA C had moved east through Central Asia rather than through South Asia. Consequently I saw them as having been the carriers of the Denisova element.

"when you look at the tree, many theories can allready be ruled out and we have to find a new explanation re the origins of Neanderthals"

John Hawks has a very interesting piece:

http://johnhawks.net/weblog/reviews/neandertals/neandertal_dna/sima-de-los-huesos-dna-meyer-2013.html

I found this particularly relevant:

"But the last five years have made clear that both groups -- the fossil scientists drawing straight lines of diverging fossil populations, and the geneticists drawing straight lines of diverging -- were wrong. Just look at the evidence. Humans today descend in part from Neandertals, even though Neandertal mtDNA is gone. Europeans today are largely different from the Europeans of 10,000 years ago, with a massive mtDNA replacement along with the introduction of Neolithic culture, and at least a second later large-scale replacement of genetic diversity. Earlier Neandertals in Europe have different mtDNA diversity than later Neandertals in Europe. Denisova cave was home to an earlier population of hominins with different mtDNA than the later Neandertals who lived there. Mitochondrial DNA has never been a straight line linking earlier and later populations within a single location. Whenever we look at ancient DNA in hominins, the earlier populations have different mtDNA diversity than the later ones. Moreover, wherever we have ancient mtDNA from other species -- bison, mammoths, cave bears, and others -- we find that later mtDNA sequences do not represent the earlier diversity".

But as for Neanderthals origins:

"Now that we know that the last 100,000 years of Neandertal evolution was complex and not centered in western Europe, I don't see why we should assume a straight line between Sima de los Huesos at more than 300,000 years ago and later Neandertals".

WarLord said...

"Also the Australian mtDNA of LM3 and the hominine that introduced the chr.11 Insert must have belonged to this same Sima de los Huesos/Denisova clade."

And I thought that Mungo Man's mtDNA was more related to modern humans than to Neanderthals? At least, this is what the study from 2001 indicated.

andrew said...

It is important to recognize that the Denisovan-heidelbergensis within their clade is older than the modern human-Neanderthal split.

The Denisovan-heidelbergensis split is about 800kya-900 kya give or take, around the time of the oldest heidelbergensis remains and also coincidentally, around the time of the oldest stone tools on the island of Flores, but about a 1 million years after H. Erectus leaves Africa and shows up in Java. The two may both derive from the same wave of hominin evolution and have a deep common ancestor, but these two species had a very long time in human evolutionary time scales to diverge from each other. These two branches had been diverged from each other for longer than the entire evolutionary history of modern humans before the Northern Spanish individuals whose mtDNA was recovered died.

Clearly, conventional wisdom was off when it comes to understanding Neanderthal origins. Non-African Homo now looks like it has an Erectus wave, a Denisovan-heidelbergensis wave, and a Neanderthal wave, at least, before modern humans leave Africa. But, I think Hawks is right in re-emphasizing our ignorance and the lack of grounds to make assumptions that we had made in the past.

eurologist said...

Terry,

John Hawks also wrote:
Europeans today are largely different from the Europeans of 10,000 years ago, with a massive mtDNA replacement along with the introduction of Neolithic culture, and at least a second later large-scale replacement of genetic diversity.

Sorry - but that is just completely ignorant American / Anglosaxon literature speak. SE Europeans, such as those living in the Balkans, still were Europeans. Copper and Bronze-working people around the Pontic were still Europeans. Proto-Indo-European folks from the Pontic region were still Europeans.

Rokus said...

I thought that Mungo Man's mtDNA was more related to modern humans than to Neanderthals? At least, this is what the study from 2001 indicated.
This sounds like an interesting axioma. The short mtDNA string identified for LM3 revealed three bitpairs uniquely being shared with the chr.11 insert (16264T, 16290T, 16301T), excluding an apparent shared outlier of Bonobo (16263C). IMO, enough to indicate some kind of close affinity other than ancestral or clade between the Insert hominine and LM3. However, how "modern" this clade could have been? The insert counts 19 defining bitpairs that deviate from rCRS, three of which are unique, IMO defining the hominine source as a divergent branch as well. Four insert bitpairs are shared by primates AND groups of modern mtDNA that apparently diverged earlier, implying that rCRS would thus probably include two uniquely derived mutations (16230G->A, 16278T->C) and two more derived mutations shared with LM3 (16129A->G, 16311C->T). The remaining 8 defining insert bitpairs are shared by primates (5), Denisovans (6) or both (3, including one Bonobo outlier); none are uniquely shared by Neanderthal, indicating that the insert-hominine-LM3 clade makes also a clade with the Denisovan hominine. Altogether, 10 out of 19 defining insert bitpairs were also shared by Denisovans but strange enough, only one (ancestral 16311C) is unequivocally shared also with Sima the los Huesos. Instead, the latter apparently groups with LM3 and rCRS at 16249T (outlier Bonobo), 16259C (ancestral) and 16368T (shared Neanderthal)- the rest of the shared Denisovan- insert bitpairs being no-calls for Sima de los Huesos. I am not sure why no calls pop up in Sima de los Huesos exactly where Denisovan bitpairs tend to match defining insert bitpairs, I am inclined to interpret this as a technical issue implying divergence from rCRS and 'thus' clade identity with Denisovan bitpairs.
So how come LM3 mtDNA is considered to be member of the clade of modern humans, while obviously they make an excellent match with the clade of Denisovan/Sima de los Huesos? Well, most probably this just isn't so very sure at all, after all Denisovan mtDNA wasn't known until recently and only a short string of LM3 mtDNA has been resolved, hardly enough to discount ancestral noise. Moreover, this result appears heavily distorted by the Out of Africa paradigm. Otherwise, the axioma may only be solved thus that modern humans - and Neanderthal alike - are a subclade of the Denisovan/Sima de los Huesos clade. At least Sima de los Huesos has all appearance to be closer to modern humans than Denisovans, though so far this seems due to its ancestral status and the unresolved portion of key bitpairs. Confined to the tract of bitpairs known for LM3, Denosovan divergence from rCRS still exceeds Neanderthal divergence by almost 50%, predominantly due to the additional insert-like bitpairs, while for just the resolved part Sima the los Huesos already misses more than 50% of the overall count of divergent Denisovan bitpairs, where it adds only one divergent bitpair of its own. One clade or two clades? I have the feeling this question may still be open, but if it isn't than LM3 is definitely a conservative member of "the other" clade.

Grognard said...

Eurologist, he didn't say they were not europeans, but they are not the same europeans.

eurologist said...

he didn't say they were not europeans, but they are not the same europeans.

Grognard,

That's either splitting hairs or allowing inappropriate polemics into a scientific discussion:

Europeans today are largely different from the Europeans of 10,000 years ago, with a massive mtDNA replacement along with the introduction of Neolithic culture, and at least a second later large-scale replacement of genetic diversity.

All of what I bolded is absolutely wrong, and perfectly known to be absolutely wrong, from a pan-European perspective. I have no idea why Hawks would write such nonsense - but he did.

eurologist said...

Let me rephrase Hawks' statement so that it makes sense:

With the increased mobility of the Neolithic and Bronze and Iron Ages, and the ensuing establishment of widely-recruited standing armies, Europeans today, with important exceptions at rather wide fringes, are much more like each other than they ever were, before (after 33kya).

While most people would call that simply "stating the obvious," I am sure Europeans of the Balkans or the Pontic or the European Urals or steppe regions, including western Kazakhstan, are much more comfortable with this version than with Hawks' wild one.

terryt said...

"The new Sima specimen shares descent from Denisova. This is important because it shows that whatever archaic human population the Denisovan mtDNA belonged to also extended to western Europe. And, surprisingly, the Sima specimen did not group with Neandertals, as might be expected because of the incipient Neanderthaloid morphology of the Sima hominins which has been a matter of controversy as it pushes back the evolutionary lineage of H. neandertalensis deeper into the Middle Pleistocene that some researchers accept".

What this study actually shows is, once again, that haplogroups are not a reliable guide to aDNA. In other words, not a reliable guide to 'race' or even 'species'. The evolution toward Neanderthal was obviously happening in Western Europe long before 'Neanderthal mt-DNA' arrived.

Max said...

Who do you believe? It has been proven scientifically that the half-life of DNA under normal conditions (e.g., bury a body) is about 500 years. Now, to believe that you can find DNA longuer than a dozen base pairs from what is basically a fosilized bone, is to stretch far all possibilities. Consistently, it is very difficult in many cases to get DNA from a 2000 year old bone (frequently, the result is negative). What we can conclude? Who is wrong? The ones who said that such bones are that old, since it contradicts the observations, that DNA is a very fragile molecule. But you will not listen, you will not question the result, because that puts "everything" in conflict. I am a scientist, and find that claim quite ridiculous to say the list. In order to be able to extract DNA from that bone, it must be much less old than thought. That is a scientific fact.

nyymi said...

Is it possible that that the older samples (Denisovan, SDLH) appear more similar because of more extensive C to T substitutions, i.e. the inferred similarity is artificial. Well, it's easy enough to study what kind of SNPs make them more similar to each other..

eurologist said...

Max,

You say you are a scientist? It's not just one bone. It is a huge stockpile of assemblages in Europe and elsewhere, with clear evolution of of fossil and assemblage lineages, dated through various strata and numerous different techniques. Even if this bone could not be dated directly, the associated features place it at the very beginning of the heidelbergensis -> Neanderthal development, which dates it to around >~300kya from a wealth of other data. Of course, it can be a bit earlier if this location was the starting point for this transition.

Chad Rohlfsen said...

How many Neandertals, out of all regional variations have a full genome published? I see conflicting reports on the amount of admixture into modern Eurasians. One report says 3-7.9% admixture. Does this take into effect the amount that flowed back into Africa; then possibly drowning out additional admixture from Neandertal, as just a common ancestor?
The 1-2% in other papers on Europeans seems way too low considering the time spent side by side in Paleolithic Europe. Could additional alleles from other Neandertal groups, Erectus,Denisovan, and unknown archaics be hiding in there as well? Thanks!