Elena Schroeter is accustomed to being economical with her samples. A postdoctoral researcher at North Carolina State University, Schroeter analyzes pieces of ancient bone that have been preserved in the ground for millions of years—and in doing so, destroys them. So her collaborators rarely give her more than a gram or two of material to work with. “People don’t want you to grind up their dinosaurs,” she explains. “You have to learn how to do a lot with a little.”
But even just a pinch of dinosaur bone dust could help reveal the ancient animal’s secrets. In one recent project, for example, Schroeter and her advisor Mary Schweitzer extracted and analyzed collagen peptides from just 200 mg of an 80-million-year-old fossil of a Cretaceous-era herbivore, Brachylophosaurus canadensis, excavated in Montana. The amino acid sequences of those peptides, published last year, placed the dinosaur on a branch of the phylogenetic tree between crocodiles and basal birds such as ostriches.1 What’s more, the team’s collection of analyzable peptides from the ancient specimen suggests that there might be other fossils out there with similar molecular information hidden in them.
Although the findings were controversial—some researchers still doubt that proteins can resist degradation for tens of millions of years—Schroeter is one of a small but growing number of researchers specializing in the analysis of ancient proteins, or paleoproteomics, to learn about the biology of organisms past. It’s been a goal of scientists for some time now; in the 1950s, several researchers were already discussing the possibility of studying peptides preserved in fossils. But only in the last two decades have advances in techniques for protein analysis, such as mass spectrometry, made the feat practical.
The potential for learning about ancient life from paleoproteomics is substantial. Via their amino acid sequences, peptides offer many of the same insights as DNA about genomic makeup—information that can support new or existing phylogenetic trees, inform research on past migrations, and assist with species identifications, even amidst a jumble of ancient remains. (See “What’s Old Is New Again,” The Scientist, June 2015.) But proteins tend to last longer in the geological record than nucleic acids, thanks to both greater volumes at deposition and more-degradation-proof molecular structures. “Both DNA and proteins are chains of building blocks,” explains Enrico Cappellini, a paleoproteomics researcher at the Natural History Museum of Denmark. “But the bonds connecting those blocks are more stable in proteins than in DNA.” The oldest confirmed DNA samples, extracted from ice cores taken in southern Greenland, are less than 800,000 years old, while the oldest protein, even by conservative estimates, dates back several million years.
When you extract proteins from fossils, a lot of other gunky stuff from the fossil—humic acid and other kinds of organics—coextracts with them.—Elena Schroeter, North Carolina State University
In some respects, peptides offer even more biological insight than their genetic precursors. “Proteins are the functional representation of the genome,” says Tim Cleland, a physical scientist at the Smithsonian Institution and previously Schweitzer’s PhD student. Protein levels vary between tissues and change as an organism ages. And posttranslational modifications to the molecules could potentially offer information about an organism’s physiology or biochemistry that DNA sequences alone can’t provide.
To explore these possibilities, researchers are now prospecting for proteins in various ancient materials, from the dental plaque of medieval skeletons to the bones of dinosaurs that walked the Earth hundreds of millions of years ago. And although the field is grappling with its fair share of debates—over where to look for proteins and how to confirm their identity, for instance—the results of recent efforts are providing an unprecedented view of ancient life on Earth. “For me, it’s huge,” says Cappellini, whose team published 126 protein sequences from a 43,000-year-old woolly mammoth bone in 2011.2 The study of ancient proteins “opens a new chapter in paleontology.”
Panning for ancient proteins
Modern proteins are generally identified with the aid of a mass spectrometer, a machine that analyzes the mass and charge of fragments of a molecule to infer the compound’s makeup. Using mass spectrometry, researchers can reconstruct a protein’s amino acid sequence, and even potentially its posttranslational modifications. The technique has allowed rapid advances in the field of proteomics. Notably, the international Human Proteome Project is reportedly close to reaching its goal of mapping all human proteins. Yet there are several wrinkles to be ironed out in the method’s application to ancient peptides.
For a start, the abundance of intact proteins in samples retrieved from ancient remains tends to be low due to degradation. And extracting those proteins becomes messier the older the material gets. “When you extract proteins from fossils, a lot of other gunky stuff from the fossil—humic acid and other kinds of organics—coextracts with them,” says Schroeter. “Part of the struggle in trying to get protein out of fossils is trying to concentrate it, but at the same time clean it in a way that you just don’t have to do with modern tissue.” Researchers work with various compounds, including acids to remove organics, and resins that can draw peptides out of a mixture, but the field has yet to agree on the best approach, Schroeter says.
There’s also an inherent risk that specimens will contain proteins from a contaminating source, a problem that has always plagued studies of ancient tissues. “It’s extremely easy for proteins to get into archaeological or ancient artifacts,” says Matthew Collins, a biomolecular archaeologist at the University of York in the U.K. Peptides from people handling the specimens, or from animals that have been near the dig site, may not be easily distinguishable from proteins endogenous to the specimens, particularly in the case of peptides that are well-conserved across the animal kingdom, such as collagen, he adds. “How can you tell something is definitely old?”
To try to avoid such complications, paleoproteomic researchers take extreme caution with their samples, notes Schweitzer. In addition to using different workspaces and equipment for ancient and modern material, labs collect controls at every step in the procedure to try to keep track of any contamination sources. “We’ve got the bone we’re interested in, and then we’ve got sediment the bones were embedded in, which shouldn’t have [the same] proteins,” Schweitzer says. Researchers also analyze blank samples to detect any rogue peptides, she adds. “Everything that we could possibly control is controlled for.”
Nevertheless, the specter of contamination hovers over much of the research in the field, particularly where surprising claims are involved. In fact, few of the researchers who spoke to The Scientist had not faced contamination challenges in the past. Collins highlights one of his own papers as an example: a 1992 study reporting the presence of the bone protein osteocalcin in the fossils of two Cretaceous dinosaurs, each more than 70 million years old.3 “Altogether, that paper was quite convincing. We did a pretty careful study,” he says. “I’m convinced that it was contamination now.”
One hundred thousands years told in ancient proteins
In spite of the technical challenges, scientists continue to apply the tools of paleoproteomics to the study of hominins and other animals that have walked the Earth over the past few hundred thousand years. In 2014, a research group that included Collins, Cappellini, and paleogeneticist Christina Warinner of the Max Planck Institute for the Science of Human History published an analysis of proteins extracted from the hardened dental plaque, or calculus, of several 1,000-year-old human skeletons that had been buried in a medieval German monastic site and excavated in 1990. The team characterized 239 bacterial proteins, shedding light on the medieval oral microbiome, along with 43 human proteins, more than half of which are involved in the innate immune system and one-third of which are common to modern calculus samples.4
Later that year, the same researchers reported finding the whey protein β-lactoglobulin in the plaque of human teeth dating back to the Bronze Age, around 3,000 BCE.5 The sequences provided the first ever direct evidence of milk consumption during the period. Using such approaches, “you can begin to explore who is consuming what” throughout human history, says Collins. “It’s a really interesting area.”
Several research groups are also using proteins for species identification thanks to a tool known as collagen fingerprinting, or zooarcheology by mass spectrometry (ZooMS). Developed in 2009 by the University of Manchester’s Michael Buckley, the method uses enzymes to break proteins into fragments that can be analyzed using mass spectrometry and compared with libraries of collagen sequences. “[It’s] a simple and cheap means to obtain species identification,” Buckley, previously a PhD student with Collins, writes in an email to The Scientist. Just last summer, his group used collagen fingerprinting to identify the 50,000-year-old remains of extinct kangaroos retrieved from caves in Tasmania.6
The technique has become a mainstay in Collins’s lab. He and Frido Welker, a postdoc at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, recently used ZooMS and other paleoproteomic analyses as part of a project to identify a collection of 40,000-year-old hominin remains pulled from a cave in France.7 Well-known for the array of jewelry, tools, and other artifacts discovered there, the Grotte du Renne site was assumed by many researchers to have been constructed by Homo sapiens. But an analysis of proteins in bone fragments found at the site indicated that the remains were in fact from Neanderthals—a finding that some anthropologists took as evidence that these archaic hominins were capable of greater creative expression than previously thought.
The same researchers have used a similar approach to solve non-hominin mysteries, too. A few years ago, Welker, Collins, and colleagues extracted collagen from the roughly 12,000-year-old fossils of two South American ungulates, Toxodon and Macrauchenia. The remains have puzzled evolutionary biologists since Darwin’s era because each genus shares traits with multiple extant mammalian lineages. Macrauchenia, for example, resembled a humpless camel with a trunk. Thanks to their collagen sequences, however, the pair has now been placed in a sister group to Perissodactyla, the animal order containing horses and rhinos.8
In general, these findings have been relatively uncontroversial, having involved specimens that are less than 100,000 years old. When it comes to more-ancient inquiries, there’s less agreement about the legitimacy of the results.
Do dinosaur fossils habor ancient proteins?
In search of more-ancient proteins, Collins, Cappellini, and colleagues recently analyzed fossilized pieces of ostrich eggshell, which are found across many African paleontological and archeological sites. The oldest shell fragment, retrieved from a dig in Tanzania, registered as 3.8 million years old. In this and other samples examined, the researchers discovered the structural protein struthiocalcin. Computer simulations suggested that regions of the protein with the strongest chemical binding to the surface of calcite crystals in the shell had survived in the best condition, indicating that minerals could play a key role in protein preservation.9
© GEORGE RETSECK
Other researchers claim to have plucked analyzable proteins from much older biological remains. In 2007, Schweitzer and her colleagues published a paper in Science that described soft tissue from a 68-million-year-old Tyrannosaurus rex fossil. Mass spectrometry, as well as immunological assays using antibodies, indicated the presence of collagen, the authors wrote.10 In 2009, Schweitzer published similar findings in B. canadensis, using the same fossil that she and Schroeter would go on to study in more detail over the next decade. This paper extended the age limit of ancient proteins to around 80 million years, and suggested that preserved peptides might regularly be found in Cretaceous dinosaurs.11
Recently, a team led by University of Toronto researcher Robert Reisz pushed that limit even further back. In 2017, the group made headlines when it reported finding collagen in a rib bone from the Jurassic plant eater Lufengosaurus. In this case, the researchers used a method that avoided destruction of their 195-million-year-old fossil: Fourier transform infrared (FTIR) spectroscopy, which measures how a sample absorbs radiation in order to infer the types of chemical bonds present. Although this technique cannot reliably determine the sequences of amino acids in a sample, Reisz and colleagues detected bonds typical of collagen.
This finding could stretch the age limit for ancient proteins, albeit probably in a very degraded form, into the realm of hundreds of millions of years.12 For now, “we’re not able to do anything more than identify the presence of collagen,” Reisz says. “But technology is advancing at a very rapid rate. What was not possible even 10 years ago is now possible. . . . With further work and further materials, who knows what we’ll be able to find.”
Claims of proteins this old have met skepticism, with other researchers in the field arguing that the results don’t jibe with theories about protein degradation. For some researchers, including Collins and many of his collaborators, the ostrich eggshell struthiocalcin, not peptides from the older dinosaur remains, currently represents the oldest analyzed protein on record. “I tend to take the standpoint that there are basic laws of chemistry and physics which are limiting the long-term persistence of these molecules,” Collins says. He suggests that there’s currently a split in the paleoproteomics community between researchers who, like him, operate under the assumption that proteins can’t survive more than a few million years (and so focus their efforts on specimens that fall within that limit), and those who search for proteins in more-degraded samples, some of which are tens of millions of years old. “It’s the difference between people who wade into deep water from the shore, and people who jump from a high place into deep water,” he says. “They’re different ways of approaching the problem.”
More-specific criticisms have been leveled at the dinosaur studies, too, including suggestions that methods other than mass spectrometry are unlikely to yield reliable information about degraded proteins, and that cross-contamination cannot be definitively ruled out. Following Schroeter’s recent analysis of B. canadensis, for example, Buckley’s group compared the published dino peptide sequences with those of modern animals. The researchers concluded that the sequences in both the 2007 and the 2009 studies could be matched to mass spectrometry data from ostriches, while the most recent sequences from B. canadensis matched those of alligators; they point out that both animals were sources of protein in the labs where the analyses of the fossils were conducted.13 Schroeter and Schweitzer note that, given the team’s cautious protocols, such double contamination is highly unlikely.
Cappellini, who told Science at the time of the North Carolina team’s 2017 publication that he was “convinced beyond reasonable doubt” by the most recent analysis of B. canadensis, has agreed to work with Schweitzer’s group to try to replicate the findings. “I don’t have a predefined position,” he tells The Scientist. “If they or, together with them, we will find that dinosaur proteins are there and can be retrieved, I will be super happy about that. It’ll be proof that we can go that far back in time in genetic reconstruction.”
Taking paleoproteomics into the future
In addition to stimulating debate on the limits of protein preservation, disagreement over the validity of multiple reports of very ancient proteins has highlighted a lack of consensus in the paleoproteomics community about which methods to use when. From relatively indirect techniques, such as FTIR, to a suite of varying mass spec protocols, multiple approaches are reported in the literature. And not all necessarily produce the same results. “[Paleoproteomics needs] a standard methodology that’s tested and used by everyone,” says Schweitzer. “The field is not going to get as strong as it could get if we’re all using different methods that say different things.”
To help facilitate the discussion, Collins is organizing an ancient protein conference for later this year. “I’d like people to meet together more, share ideas,” he says, adding that currently, researchers in paleoproteomics tend to be spread out at various conferences depending on the age of their samples. “The people who work on old things tend to go to paleontology conferences; the ones who work on young things go to archaeology conferences,” he says. The most recent meeting specifically built around an ancient protein theme was 20 years ago, he adds. “I think the time is ripe for the next one, to try to get all these different groups together.”
I tend to take the standpoint that there are basic laws of chemistry and physics which are limiting the long-term persistence of these molecules.—Matthew Collins, University of York
Meanwhile, efforts are underway to open up the field to a wider community of scientists. Schroeter and Cleland have put together a review paper, published earlier this year, that explains mass spectrometry and its application to studying ancient proteins for a nonexpert audience.14 “I think there’s a gap in the community of paleo at large in our understanding of how these techniques work,” says Schroeter. A broader appreciation could both help resolve discussions about controversial findings, and lead to collaborations for people working on all sorts of ancient samples, she adds. “We need to be more accessible with our methods. Because the way to get more people involved in this work is to give them a grounding in that information, [and] make them see the value of it.”
The range of samples may soon be expanding too. Part of Cleland’s work at the Smithsonian involves developing methods to study samples kept in museum collections, rather than those obtained from expensive and time-consuming fieldwork. He and his colleagues recently showed that one specimen—a 12,000-year-old skull of an extinct giant beaver that has been kept in the New York State Museum since 1845—contains “quite a bit of protein,” Cleland says.15 “We have these wonderful museums around the world that have these huge collections of material. It opens up the possibility of looking at a large number of things.”
OILS AND PIGMENTS
Some pigments, such as melanin, a polymer derived from the amino acid tyrosine, also survive relatively well in the fossil record and can provide clues about the physical appearance of ancient organisms. In 2015, a team at Lund University in Sweden reported the discovery of pigment in a 150-million-year-old fossilized specimen of Anchiornis huxleyi, a small, four-winged Jurassic dinosaur (Sci Rep, 5:13520, 2015). Mass spectrometry revealed that the animal’s fossilized feathers contained eumelanin, the pigment responsible for brownish-black colorations. And just last August, researchers in California and Germany reported the presence of bluish-green pigment molecules in fossilized eggshells from a Cretaceous dinosaur, a possible indication of egg camouflage, the team suggests (PeerJ, 5:e3706, 2017).
Although ancient biomolecules such as these are unable to offer the degree of evolutionary insight afforded by amino acid or nucleic acid sequences, such discoveries exemplify “the growing field of molecular paleontology,” Montana State University paleontologist David Varricchio told National Geographic last September following the eggshell paper’s publication. “With new machines and new techniques, it’s very exciting what can potentially be found in fossils.”
- E. R. Schroeter et al., “Expansion for the Brachylophosaurus canadensis collagen I sequence and additional evidence of the preservation of Cretaceous protein,” J Proteome Res, 16:920-32, 2017.
- E. Cappellini et al., “Proteomic analysis of a Pleistocene mammoth femur reveals more than one hundred ancient bone proteins,” J Proteome Res, 11:917-26, 2012.
- G. Muyzer et al., “Preservation of the bone protein osteocalcin in dinosaurs,” Geology, 20:871-74, 1992.
- C. Warinner et al., “Pathogens and host immunity in the ancient human oral cavity,” Nat Genet, 46:336-44, 2014.
- C. Warinner et al., “Direct evidence of milk consumption from ancient human dental calculus,” Sci Rep, 4:7104, 2014.
- M. Buckley et al., “Identifying remains of extinct kangaroos in Late Pleistocene deposits using collagen fingerprinting,” J Quat Sci, 32:653-60, 2017.
- F. Welker et al., “Palaeoproteomic evidence identifies archaic hominins associated with the Châtelperronian at the Grotte du Renne,” PNAS, 113:11162-67, 2016.
- F. Welker et al., “Ancient proteins resolve the evolutionary history of Darwin’s South American ungulates,” Nature, 522:81-84, 2015.
- B. Demarchi et al., “Protein sequences bound to mineral surfaces persist into deep time,” eLife, 5:e17092, 2016.
- M.H. Schweitzer et al., “Analyses of soft tissue from Tyrannosaurus rex suggest the presence of protein,” Science, 316:277-80, 2007.
- M.H. Schweitzer et al., “Biomolecular characterization and protein sequences of the Campanian hadrosaur B. canadensis,” Science, 324:626-31, 2009.
- Y.C. Lee et al., “Evidence of preserved collagen in an Early Jurassic sauropodomorph dinosaur revealed by synchroton FTIR microspectroscopy,” Nat Commun, 8:14220, 2017.
- M. Buckley et al., “A fossil protein chimera; difficulties in discriminating dinosaur peptide sequences from modern cross-contamination,” Proc R Soc B, doi:10.1098/rspb.2017.0544, 2017.
- T.P. Cleland, E.R. Schroeter, “A comparison of common mass spectrometry approaches for paleoproteomics,” J Proteome Res, doi:10.1021/acs.jproteome.7b00703, 2018.
- T.P. Cleland et al., “Peptide sequences from the first Castoroides ohioensis skull and the utility of old museum collections for paleoproteomics,” Proc R Soc B, doi:10.1098/rspb.2016.0593, 2016.