The laser of a Raman spectrometer shines on a Tyrannosaurus rex tooth

The laser of a Raman spectrometer shines on a Tyrannosaurus rex tooth, gathering data about the organic material still locked inside it.


Warm-blooded velociraptors? Fossilized proteins unravel dinosaur mysteries

In the bowels of Yale University's Peabody Museum of Natural History, Jasmina Wiemann yanks open a drawer in a floor-to-ceiling specimen cabinet. She lifts out a wickedly sharp, sickle-shaped dinosaur claw, black as coal. "This is the type specimen of Deinonychus—the basis for the Velociraptor in the Jurassic Park movies," she says. The black color signals something just as striking. The fossil isn't just a mineral replica of the original claw. It is likely two-thirds dinosaur residue, Wiemann says. "I bet this specimen is maybe 70% organic material by volume—more than we'd think!"

That fossils can harbor organic matter isn't new. Whole fields of science have sprung up to decipher ancient DNA and intact proteins. But most researchers think that in fossils as old as the Deinonychus claw, most of the useful sequences of those molecules have long vanished. Now, Wiemann and her Ph.D. adviser, Yale's Derek Briggs, have devised a way to extract information locked in degraded proteins, even in fossils hundreds of millions of years old. "This [kind of] molecular preservation is really common, and we just didn't know," Wiemann says.

She and Briggs have shown how, when conditions are right in the weeks and months after an animal dies, cellular proteins can react with lipids and sugars. The process transforms the proteins into a mix of hardy polymers that repel water, resist microbes, and are impervious to heat. The polymers are chemically similar to those formed when meat browns or toast burns—and they can apparently last for eons.

Other researchers have claimed to find intact proteins more than 100 million years ago, in the age of dinosaurs, but the field has remained skeptical of such ancient preservation. No one could explain how proteins managed to survive the degradations of time, says paleontologist Maria McNamara of University College Cork in Ireland. Wiemann "very nicely came up with this very, very clever mechanism" for how proteins could persist after all, in an altered form.

After a proof-of-principle paper last year, Wiemann and Briggs are applying their nondestructive technique—shining a laser on specimens to reveal ancient chemical bonds—to help solve paleontological mysteries. This week at a meeting in Australia, they planned to report how they used protein residue data to help resolve where turtles fit on the vertebrate family tree, and to support the idea that pterosaurs, the largest animals ever to fly, were warm-blooded.

The technique is new, so "it needs to be validated by more fossil and experimental work," McNamara says. And the method by itself can't resolve evolutionary puzzles with certainty. But she and others find the chemical mechanism convincing. "It's a completely new level of understanding of preservation. They are shedding light on the why and how," says Jingmai O'Connor, a paleontologist at the Institute of Vertebrate Paleontology and Paleoanthropology in Beijing. "It's incredible how [Wiemann] is revolutionizing our field, opening so many new doors by applying chemistry to a field where chemistry has rarely been applied."

For Wiemann, the first clues to how biomolecules might persist for hundreds of millions of years came from dinosaur eggs. As an undergraduate and master's student, she worked with a team led by Martin Sander at the University of Bonn in Germany that showed that 67-million-year-old dinosaur eggs were blue-green, not white. To isolate the color pigments, Wiemann dissolved pieces of fossil eggshell in a solution that removed the calcium. Sometimes she found pliable brown residues at the bottom of her test tubes. Under the microscope, the residues resembled the organic matrix of eggshells, and she wondered whether they were bits of original tissue. "It was quite exciting to see," she says. But she didn't have time to figure out exactly what she was seeing.

For her Ph.D., she went to Yale to join Briggs's lab, where she set out to identify those brown dregs. She found more residues when she decalcified pieces of fossil bone and teeth. "The residue you get is very different from fresh proteins," she says, but under the microscope it showed tantalizing hints of soft-tissue structures—blood vessels, cells, even nerve projections.

Wiemann and Briggs—an expert in soft-tissue preservation—noticed that residues were most likely to come from fossils of a specific type: black or brown ones from lighter-colored rocks that formed in shallow seabeds or iron-rich sandstones. Those environments are oxidative, rich in reactive oxygen molecules and dissolved metal ions, and the water that enters decaying bones and tissues is alkaline—conditions that promote biochemical reactions called glycoxidation and lipoxidation. Food chemists know them well. Called Maillard reactions, they occur anytime something is toasted, grilled, caramelized, or "browned," turning proteins, fats, and sugars into tough, complex polymers that are brown and often taste delicious.

How ancient bonds can cheat time

In certain environments, a dead organism’s proteins degrade into hardy polymers. Different proteins form different polymers, so Raman spectroscopy, which can distinguish various kinds of bonds, can offer clues to proteins’ original chemistry, even after eons.

Time and degradation Time and degradation Raman spectra Light intensity After hundreds of millions of years, some bonds within proteins persist and can be traced with Raman spectroscopy. Different amino acids Proteins Protein residues Wavelength Wavelength Light intensity

As a teenager, Wiemann had taken university level classes in organic chemistry. "So I knew about the Maillard reactions and how you could go from a protein to a water-resistant thing," she says. She wondered whether something similar had happened to the proteins in the fossilized tissues.

To find out, Wiemann turned to Raman spectroscopy. Many biochemical techniques search for a specific compound, Wiemann says, "but Raman spectroscopy is more exploratory." It uses laser light to identify the types of chemical bonds in a sample. Different bonds absorb different wavelengths, leaving a fingerprint in the spectrum of the reflected light. The technique enabled Wiemann and Briggs to confirm that the brownish residues were indeed made of complex polymers, the end products of glycoxidation and lipoxidation. When the researchers artificially fossilized modern bones and eggshells, heating them under oxidative conditions in the lab, the modern tissues formed the same kinds of compounds, the team reported in Nature Communications in November 2018.

Different proteins form different polymers, so despite their transformation, ancient molecules retain some of their original chemistry. The result is a new kind of molecular tool for studying ancient life—one that complements ancient DNA and proteins, Briggs says.

Although ancient DNA carries the most detailed biological information, it degrades most quickly. Relatively intact proteins can persist for nearly 4 million years and can still distinguish between closely related species. With protein residues, "the information is again reduced," Wiemann says. The polymers don't preserve 3D structure or a complete sequence of amino acids. But the compounds are incredibly stable, preserved "through deep time," Briggs says. Wiemann says they have identified protein residues in 500-million-year-old fossils from Canada's Burgess Shale in British Columbia.

That claim is startling, but "the chemistry makes sense," says Evan Saitta, a paleontologist at the Field Museum in Chicago, Illinois. "If you have ever scrubbed a grill after cooking, you will know that these protein products are stable at high temperatures—and are clearly insoluble."

To understand the meaning of the spectra, though, Wiemann had to build a database with the spectral signatures of many fossils and modern samples. And she had to find out whether those spectra could tell her anything interesting about ancient life.

Luckily, Wiemann works at a museum with a "colossal" collection, Briggs says. The Peabody Museum holds "100,000 vertebrate fossils and about 4.5 million invertebrate fossils," he says. Wiemann spent her evenings raiding the tall cabinets, seeking fossils with the telltale dark color in light sediment. "No one ever comes here in the evenings," she says of the collection rooms. "It's actually fun."

The scans don't damage the specimens, so curators agreed to lend them for study. Wiemann's office, down the hall, is littered with specimens from many geologic periods and across the tree of life, each in its carefully labeled box. "Curators definitely appreciate that it's nondestructive," she says.

She has so far analyzed more than 100 specimens with the Raman spectrometer, which looks like a large microscope. She examined an Allosaurus specimen from Carbon county in Utah, so black it glistens; fish from the famed Green River deposit in Wyoming; the claw from Deinonychus, a species that inspired the first suggestions that some dinosaurs could run quickly and were the ancestors of birds; and her old favorites, fossil eggshells.

With this still-growing database, she first tested whether the spectral signatures could reveal relationships among ancient animals. By comparing slight differences in the number and height of spectral peaks, a computer could correctly slot eggshell samples from more than a dozen dinosaurs, including early birds, onto a phylogenetic tree. Samples of bones and teeth matched about 60% of known relationships, Wiemann planned to announce at the Society of Vertebrate Paleontology meeting in Brisbane, Australia, this week.

Derek Briggs (left) and Jasmina Wiemann sought specimens still rich in protein residues in Yale University’s Peabody Museum, looking for dark fossils in pale sediment.


"This is not something that is going to tell you how 10 hadrosaurs you found in a quarry were related, but if you find bits and pieces of turtle, stem bird, and crocodile, it can probably help you tell them apart," she says.

In another talk at this week's meeting, Wiemann's and Briggs's colleague Dalton Meyer planned to describe how they compared Raman spectra from bones of many extinct reptiles to find where turtles might belong on the reptile tree—a persistent puzzle. The spectra suggest turtles' ancestors, which lived more than 200 million years ago, were more closely related to dinosaurs (and modern birds) than to the ancestors of crocodiles, snakes, and lizards.

Wiemann, Briggs, and colleagues have applied the technique to organic matter from even further back in time. Tullimonstrum, also known as the Tully Monster, is a mysterious creature from the Mazon Creek fossil beds in Illinois. It lived more than 300 million years ago and has stumped paleontologists for decades. A soft-bodied oval with a long appendage, it has sometimes been identified as a vertebrate, some sort of worm, or a strange swimming snail. But Raman spectra of the animal's teeth suggest they were made of keratin or collagen, proteins made only by vertebrates and their relatives, Victoria McCoy, a graduate of Briggs's lab now at the University of Wisconsin in Milwaukee, planned to report at the meeting. That finding suggests the creature was some kind of vertebrate, rather than a snail or worm.

McCoy, Briggs, and their colleagues had already reached a similar conclusion from analyzing the shape of more than 1200 Tully Monster specimens, which they published in 2016 in Nature. The protein residue analysis "is helping to confirm our original morphology conclusions with chemical data, which is less ambiguous—and quite exciting," Briggs says.

Wiemann also told the meeting how protein residues may point to whether an extinct animal was warm-blooded. The reactions a cell uses to produce energy also generate side products called free radicals, which can trigger Maillard-like reactions and produce polymers similar to those in the fossil protein residues. The cells of warm-blooded animals with faster metabolisms carry out more of those reactions than do cold-blooded ones. Wiemann realized that if she could correct for the Maillard reactions induced by fossilization, she might detect variations due to metabolic differences. She tracked fossilization reactions in her experiments with modern samples and concluded that the ratio of two key types of chemical bonds in a fossil's Raman spectrum might point to metabolic rate.

She tested the idea, and at the meeting she planned to report that her results fit with what's known or suspected about metabolism in fossil and living creatures. Mammals, pterosaurs, and two-legged dinosaurs such as the fast-moving Allosaurus and Deinonychus had the signature of warm-blooded animals. Mammallike reptiles that lived 300 million years ago turned up as barely warm-blooded, and quadripedal dinosaurs such as Triceratops appeared to have even slower metabolisms. The ancestors of lizards and snakes appeared to be cold-blooded.

"We've only been able to guess about metabolic rates using features like bone histology or inferred brain size," O'Connor says. The new method may offer a more direct route to metabolism. She suspects warm-bloodedness evolved several times in the ancient bird lineages she studies and says her lab will soon start a project using Raman spectra to explore metabolic rates in bird fossils.

"That these [residues] can tell us something about basal metabolic rate is sort of mind-blowing," says Lawrence Witmer, a paleontologist at Ohio University in Athens. "It takes a conceptual leap, really an outstanding creativity, to see these connections we haven't seen."

Some researchers caution that independent labs haven't yet replicated the work. And they say Wiemann hasn't definitively ruled out that some spectra might stem from contamination—for example, from bacteria that colonized the fossil. "You don't know whether you're measuring the original stuff, more modern stuff, or a mixture of both," says Johan Lindgren, a paleontologist at Lund University in Sweden.

Wiemann says bacterial residues and other contaminants have characteristic Raman signatures that she can rule out. But back among the Peabody's shelves, she agrees that plenty remains to be done. "We have to optimize these methods," she says. "That's not something one person can do alone. It's something that the whole field has to take on."

With reporting by Elizabeth Culotta in New Haven, Connecticut.