Proteins from rabbit skin glue, chicken egg yolk, and wool fibers. A new type of analysis can determine that kind of biological species these materials were produced from. It can also help stewards of art choose effective conservation methods, decide how best to display or store a painting, and perhaps even sniff out fakes.
Looking at The Virgin and Child with Saint John and an Angel, a 15th-century painting attributed to the workshop of Italian master painter Sandro Botticelli, you are likely to admire the luminous quality of baby Jesus’ skin or Mary’s serene gaze.
You would be forgiven if you did not notice the chicken and ungulate in the painting. They are, after all, in the painting.
By applying techniques developed to study proteins in archaeological finds, a team of researchers from the University of Copenhagen and the National Gallery in London are learning to tease out the secret histories of paintings.
Their research, published in June in Scientific Reports, presents a way to analyse tiny animal proteins trapped in a painting’s pigments and base layer to learn about the materials and techniques the artist used and the conditions the painting has experienced in the decades or centuries since being created.
Scientists say that this kind of analysis can help stewards of art choose effective conservation methods, decide how best to display or store a painting and perhaps even sniff out fakes.
The results were “even better than we were hoping for – almost,” says Fabiana Di Gianvincenzo, a PhD student at the University of Copenhagen and lead author of the article.
History is a mystery
“When you walk around something like the National Gallery and you look at the paintings on the wall, they all look as if they are immaculate,” says David Peggie, chemical analyst and senior scientist at the National Gallery and co-author of the study. In reality, paintings like The Virgin and Child have been through a lot – centuries’ worth of cleaning and retouching, light damage and smoke accumulation. And nearly all of that is a mystery to the people charged with safeguarding the paintings.
“If we are lucky, we might have documentation for the last couple of hundred years where they have entered into a collection that has been recorded,” Peggie says. “We’re really trying to kind of start from the object itself and go okay, what can the object tell us? And that can be everything from something about the artist and how it was made all the way through to that history of what has happened to it in subsequent years.”
Anything that can be gleaned about the materials the artist used or the painting’s subsequent history can help its stewards decide the best methods for conservation, how the painting should be displayed or stored and learn about how art was produced at the time for its own intrinsic value.
We do know that the old masters’ work is lousy with animal proteins – egg yolk and animal glue were popular binding material for pigments. But little analysis has been conducted on the proteins themselves.
“Proteomics is a term used to describe a range of techniques that study the whole protein composition of a sample,” Di Gianvincenzo says. “It started with modern materials – you can study, for example, all the proteins contained in a cell or in an organism. But with palaeoproteomics, the same principle was then applied to study the proteins contained in ancient samples, a few hundred years old to millions of years old.”
An important tool in proteomics is mass spectrometry, in which samples are vaporised and then bombarded until they are reduced to their smallest constituent parts – for proteins, that is amino acids, which can then be sequenced to determine the structure of the protein.
Proteomic analysis can reveal not only the identity of the proteins, “but also what kind of biological species these materials were produced from, what tissue was used to produce this material and, finally, characterisation of the protein damage,” Di Gianvincenzo says.
“It allows us to know exactly what was used, try and understand why it was used and then what happened to it in time with the damage that we can see,” she explains.
The researchers were uncertain that proteomic analysis would be possible at all for the samples – protein binders make up less than 10% of the paint, and the pigments themselves make recovering proteins more difficult.
But the biggest hurdle was size – previous proteomics studies on paintings have only been able to identify the source of proteins with relatively large samples up to a few hundred micrograms (for reference, a poppy seed weighs 300 micrograms). Di Gianvincenzo’s team was aiming for samples in the tens of micrograms.
Before touching the Botticelli, Di Gianvincenzo and her colleagues wanted to verify that the system would work on mock-up paintings made using the same techniques as the old masters, but with known components.
The team scraped tiny samples of paint or paint and binder from the surface of 10 mock-up paintings. For each sample, they collected one larger sample – about the size of samples collected in previous studies, or 300 micrograms – and one smaller sample about a third of that weight.
They were pleasantly surprised to conclusively identify the proteins present in all 10 of the large samples and eight of the 10 small samples, finding proteins from rabbit skin glue, chicken egg yolk and wool fibres.
Tackling the Botticelli
After perfecting their protocol, it was time to put it through its paces with something six times older than the oldest mock-up paintings. While the Botticelli painting was undergoing routine restoration, minute samples were scraped off two points near existing damage near the edge of the painting – yellow paint from Saint John’s sleeve and blue paint and a bit of the grounding layer from Mary’s cloak.
The samples were so small the team did not have access to a fine enough scale to weigh them precisely, but the authors eyeballed them to be between 10 and 20 micrograms.
The scientists were encouraged to learn that the analysis of these micro-samples was able to identify the origin of the proteins present, sometimes down to the tissue type.
Both paint samples contained egg yolk from chickens – which contemporary sources confirm was a staple in Botticelli’s work – and animal glue from either sheep or goat (the proteins are very similar and difficult to distinguish).
The researchers were also able to assess the level of damage to the proteins by checking for deamidation, when parts of certain amino acids are removed or changed.
Peggie and Di Gianvincenzo say the knowledge that micro-samples from paintings can be enough for proteomic analysis opens many doors.
The first step will be sampling more “normal” paintings like the Botticelli to establish a baseline for comparison for different media, time periods and artists. “Then we can apply it when we have a sort of head-scratcher,” Peggie explains – paintings with areas of unusual damage that may be due to the use of an unconventional material or potential forgeries that may have used materials unavailable at the time or have protein damage inconsistent with their supposed age.
Next into the mass spectrometer for Di Gianvincenzo’s team will be samples from the Raphael Cartoons lent to the Victoria and Albert Museum from the Royal Collection by the late Queen Elizabeth II . “Raphael is a very interesting artist because he worked across a lot of media and had lots of collaborations,” Peggie says. He’s hopeful that, with the new proteomic techniques, Di Gianvincenzo and her colleagues can “try and unpack exactly how they were made”.