Though extinct in the arctic today, endless herds of bison were common in the prehistoric northern grasslands of the Pleistocene era. Unremarkable to their contemporaries, two bison that inhabited these regions were very special animals. One breathed its last sometime over 58 000 years ago in Eva Creek, Alaska; the other met its end in the cold and remote Kolyma river region of Siberia around 3000 years later. The remains of both animals became covered with snow or soil, and were slowly buried in permafrost.

That's where their story would have ended, had their bones not been dug up 60 millennia later, carefully packaged and dispatched to Wellcome Trust bioarchaeology research fellow Cristina Nielsen-Marsh at the University of Newcastle upon Tyne for analysis.

Dr Nielsen-Marsh and colleagues caused a stir in palaeontological circles last winter by announcing in the journal Geology that they had successfully extracted the first entire protein sequence from fossilized bones. "Prior to this it was not widely believed that proteins could be retrieved intact, in fossils as old as 50 000 years," commented Dr Nielsen-Marsh's collaborator, geochemist Peggy Ostrom of Michigan State University in Wisconsin.

"The sequencing of ancient proteins has the potential to expand our understanding of ancient phylogeny and evolution," says Professor Ostrom. Phylogenetics is the study of the tree of life - a kind of map showing how species are related to one another - and Dr Nielsen-Marsh's innovative work may help us to work out the relationships between extinct species and their living counterparts as never before. As with DNA, differences and similarities in species' proteins can reveal which organisms - like the dodo and the pigeon - are close relations.

Complementing DNA

With Wellcome Trust funding since 2001, Dr Nielsen-Marsh has been able to combine her background in the study of how bones degrade, with groundbreaking new chemical techniques for analysing proteins. Working alongside teams in Oxford (see box below), Michigan and Newcastle she is attempting to find out how long biological material might last in well-preserved fossils.

Though DNA has been extracted from fossils (see Deep freeze DNA) only chunks of proteins have been found before. Proteins can be very useful, however, as they are much more durable than DNA in buried bones - perhaps lasting up to an astounding 100 million years, says Matthew Collins, who heads up York University's Bioarchaeology Centre. That leaves open the possibility that even dinosaur bones could retain fragments of information.

Osteocalcin, the protein found in the bison remains, is unusual in that it binds extremely strongly to bone minerals, which may account for its survivability in fossils. Though osteocalcin itself cannot tell us much about past life, comparing that protein between extinct and living species could show us how they are related. Surprisingly, morphology of fossils doesn't always give a clear picture of where they fit into the tree of life. That’s where comparisons of DNA and protein sequences come in.

Jumping hurdles

Back in the late 1970s some workers began to consider using proteins from fossils to compare the relatedness of species. Unfortunately, proteins are usually completely absent from fossils or present in vanishingly low quantities, and immunological methods - using antibodies to hunt down ancient protein - are unreliable.

Dr Nielsen-Marsh's work has circumvented these hurdles in two ways. First, in her previous work on the EU Bone Degradation project she studied the state of preservation of more than 500 animal and human bones from across Europe and elsewhere, ranging in age from 200 million years to the present day. With that knowledge, Dr Collins and Dr Nielsen-Marsh have been able to focus resources on bones most likely to contain ancient proteins. The best burial environments include those that are very cold (permafrost), ashfall sites (such as Pompeii) and tar pits (La Brea in California for example) - all of which limit the growth of bacteria, which eat the biomolecules in bone.

Secondly, working alongside Professor Ostrom in Michigan, Dr Nielsen-Marsh has been using high-resolution mass spectrometry methods to analyse tiny quantities of fossil protein. In this method, proteins are broken into chunks and the mass of each fragment is carefully measured. This information can be used to predict the amino acid sequence of the fragment, and combining the fragment sequence gives the full protein sequence.

"It's pretty remarkable to have found protein in 50 000-year-old bison bone," says Dr Collins, "but now we want to know just how far back these proteins go." The Newcastle lab has tested the durability of osteocalcin in cow bones heated for long periods (over a year in some cases). These findings have been extrapolated to give a 'guesstimate' of how long the protein would take to degrade over millennia in the burial environment. Dr Collins set out to prove that proteins couldn't possibly still survive in bones many millions of years old - but to no avail. "We tried, but the protein is so damn tough, we couldn't disprove that notion."

To his surprise the calculations could not disprove the idea that osteocalcin might survive in much older fossils. "The estimates are very crude," he says, but under very exceptional circumstances, it's not inconceivable that it might survive for a whopping 10 million years in temperate Europe, or even up to 100 million years if frozen. Sadly, says Dr Collins, no terrestrial region of earth has been frozen for so long.

For the moment, DNA still has many advantages - DNA is much easier to sequence than protein and contains more variation. But mass spectrometry is a fertile area of research, and protein sequencing is becoming more routine. So it could be studies of protein, not DNA, that let us look at evolution in action in the far distant past.

See also:

• Deep freeze DNA

External links

• Professor Peggy Ostrom: Research interests
• Dr Matthew Collins: Research interests
• Dr Christina Nielsen-Marsh: Research interests