Proteins

Researchers use generative AI to design novel proteins

Christopher.Sorensen — Fri, 05 May 2023 16:40:22 +0000

Researchers use generative AI to design novel proteins

Christopher.Sorensen Fri, 05/05/2023 - 12:40

Cutline

Professor Philip Kim and PhD student Jin Sub (Michael) Lee have developed a generative AI system that can create proteins not found in nature, promising to speed drug development (supplied images)

Jim Oldfield

Topic

Breaking Research

Temerty Faculty of Medicine

Donnelly Centre for Cellular & Biomolecular Research

Artificial Intelligence

Computer Science

Faculty of Arts & Science

Molecular Genetics

Proteins

Research & Innovation

Researchers at the University of Toronto have developed an artificial intelligence system that can create proteins not found in nature using generative diffusion – the same technology behind popular AI image-creation platforms such as Midjourney and OpenAI’s DALL-E.

The system will help advance the field of generative biology, which promises to speed up drug development by making the design and testing of entirely new therapeutic proteins more efficient and flexible.

“Our model learns from image representations to generate fully new proteins at a very high rate,” says Philip M. Kim, a professor in the Donnelly Centre for Cellular and Biomolecular Research at �鶹ֱ��app’s Temerty Faculty of Medicine. “All our proteins appear to be biophysically real, meaning they fold into configurations that enable them to carry out specific functions within cells.”

The findings were published in the journal Nature Computational Science and are the first of their kind in a peer-reviewed journal. Kim’s lab also published a pre-print on the model last summer through the open-access server bioRxiv ahead of two similar pre-prints from last December – RF Diffusion by the University of Washington and Chroma by Generate Biomedicines.

Proteins are made from chains of amino acids that fold into three-dimensional shapes, which in turn dictate protein function. Those shapes evolved over billions of years and are varied, complex and limited in number.

Now, with a better understanding of how existing proteins fold, researchers have begun to design folding patterns not produced in nature.

A major challenge, says Kim, has been to imagine folds that are both possible and functional.

“It’s been very hard to predict which folds will be real and work in a protein structure,” says Kim, who is also a professor in the departments of molecular genetics in the Temerty Faculty of Medicine and computer science in the Faculty of Arts & Science. “By combining biophysics-based representations of protein structure with diffusion methods from the image generation space, we can begin to address this problem.”

The new system, which the researchers call ProteinSGM, draws from a large set of image-like representations of existing proteins that encode their structure accurately. The researchers feed these images into a generative diffusion model that gradually adds noise until each image becomes all noise. The model tracks how the images become noisier and then runs the process in reverse, learning how to transform random pixels into clear images that correspond to fully novel proteins.

Jin Sub (Michael) Lee, a doctoral student in the Kim lab and first author on the paper, says that optimizing the early stage of this image generation process was one of the biggest challenges in creating ProteinSGM.

“A key idea was the proper image-like representation of protein structure, such that the diffusion model can learn how to generate novel proteins accurately,” says Lee, who is from Vancouver but did his undergraduate degree in South Korea and master’s degree in Switzerland before choosing �鶹ֱ��app for his doctorate.

Also difficult was validation of the proteins produced by ProteinSGM. The system generates many structures – often unlike anything found in nature. Almost all of them look real according to standard metrics, says Lee, but the researchers needed further proof.

To test their new proteins, Lee and his colleagues first turned to OmegaFold, an improved version of DeepMind’s software AlphaFold 2. Both platforms use AI to predict the structure of proteins based on amino acid sequences.

With OmegaFold, the team confirmed that almost all their novel sequences fold into the desired protein structures. They then chose a smaller number to create physically in test tubes, to confirm the structures were proteins and not just stray strings of chemical compounds.

“With matches in OmegaFold and experimental testing in the lab, we could be confident these were properly folded proteins. It was amazing to see validation of these fully new protein folds that don’t exist anywhere in nature,” Lee says.

Next steps based on this work include further development of ProteinSGM for antibodies and other proteins with the most therapeutic potential, Kim says. “This will be a very exciting area for research and entrepreneurship.”

Lee says he would like to see generative biology move toward joint design of protein sequences and structures, including protein side-chain conformations. Most research to date has focused on generation of backbones, the primary chemical structures that hold proteins together.

“Side-chain configurations ultimately determine protein function, and although designing them means an exponential increase in complexity, it may be possible with proper engineering,” Lee says. “We hope to find out.”

This research was funded by the Canadian Institutes of Health Research.

�鶹ֱ��app scientists develop custom-engineered protein to battle MERS virus

ullahnor — Tue, 23 May 2017 19:32:47 +0000

�鶹ֱ��app scientists develop custom-engineered protein to battle MERS virus

ullahnor Tue, 05/23/2017 - 15:32

Cutline

MERS particles attach to the surface of an infected human cell. The engineered protein can paralyze MERS by binding tightly to a key enzyme (image courtesy of the Donnelly Centre for Cellular and Biomolecular Research)

Jovana Drinjakovic

Author legacy

Jovana Drinjakovic

Topic

Subheadline

Could lead to anti-viral therapeutics for both humans and animals

In 2012, a 60-year-old man with flu-like symptoms died in Saudi Arabia, becoming the first victim of the Middle East Respiratory Syndrome or MERS.

Until now, there has been no vaccine or known treatment. That could change thanks to a new anti-viral tool, developed by University of Toronto researchers.

Scientists have crafted a custom-engineered protein that can be used to treat MERS and a wide range of pathogens, a potential game-changer in anti-viral therapeutics for both humans and the farming industry.

Writing in the journal PLOS Pathogens, a team led by Professor Sachdev Sidhu of the Donnelly Centre for Cellular and Biomolecular Research and department of molecular genetics, describes how to turn ubiquitin, a staple protein in every cell, into a drug capable of thwarting MERS in cultured human cells.

“Vaccines are important for prevention, but there is a great need for anti-viral medicines to treat people who have become infected,” says Wei Zhang, a postdoctoral researcher in Sidhu’s lab who did most of the work on the study.

“With our tool, we can quickly generate anti-viral medicine, and we hope that our method will inspire other researchers to try it out against diverse pathogens,” says Zhang.

MERS is similar to SARS, the virus that killed almost 800 people in a global epidemic. Both kill upwards of a third of people infected and, like many viruses, emerged from animals – bats and camels in the case of MERS. Although MERS has so far been detected in 27 countries, the outbreak has largely been contained within Saudi Arabia, according to the World Health Organization.

Like many viruses, MERS works by hijacking the ubiquitin system in human cells composed of hundreds of proteins that rely on ubiquitin to keep the cells alive and well. Upon infection, viral enzymes alter ubiquitin pathways in a way that allows the virus to evade the immune defense while multiplying and destroying the host tissue as it spreads in the body.

“Viruses have evolved proteins that allow them to hijack host proteins. We can now devise strategies to prevent this from happening,” says Zhang.

Zhang and colleagues engineered the human ubiquitin protein into a new form that paralyzes a key MERS enzyme, stopping the virus from replicating. These synthetic ubiquitin variants act quickly, completely eliminating MERS from cells in a dish within 24 hours.

The researchers also created UbVs that block the Crimean-Congo virus, the cause of a hemorrhagic fever that kills about 40 per cent of those infected.

And they’re designed to only target only the virus – hopefully minimizing side effects in any future drug.

But before these engineered proteins can be developed into medicine, researchers first must find a way to deliver them into the right part of the body. For this, Zhang and Sidhu are working with Roman Melnyk, an assistant professor at �鶹ֱ��app and a biochemist at the Hospital for Sick Children and a world expert in protein delivery.

The team is also investigating the possibility of finding drugs that work in a similar manner but can cross the cell membrane.

It is likely that the proteins will be tested first in plants and animals where regulatory approvals are less strict than they are for human drugs.

“We are also working on an engineered ubiquitin that targets a corn virus responsible for destroying large swaths of corn fields in North America, with colleagues in Manitoba,” says Zhang.

The study was done in collaboration with Marjolein Kikkert of Leiden University Medical Centre in the Netherlands and Brian Mark at the University of Manitoba.

�鶹ֱ��app engineering discovery potential game-changer for chronic illness treatment

katie.fong — Tue, 07 Jun 2016 15:54:14 +0000

�鶹ֱ��app engineering discovery potential game-changer for chronic illness treatment katie.fong Tue, 06/07/2016 - 11:54

Cutline

Left to right: Elliott Donaghue, Malgosia Pakulska, Jaclyn Obermeyer (Marit Mitchell photo)

Marit Mitchell

Author legacy

Marit Mitchell

Topic

Breaking Research

Molly Shoichet

Faculty of Applied Science & Engineering

Proteins

Chronic Illness

Health

A �鶹ֱ��app Engineering team has designed a simpler way to keep therapeutic proteins where they are needed for long periods of time. The discovery is a potential game-changer for the treatment of chronic illnesses or injuries that often require multiple injections or daily pills.

For decades, biomedical engineers have been painstakingly encapsulating proteins in nanoparticles to control their release. Now, a research team, all from the department of Chemical Engineering and the Institute of Biomaterials and Biomedical Engineering, led by University Professor Molly Shoichet has shown that proteins can be released over several weeks, even months, without ever being encapsulated. In this case the team looked specifically at therapeutic proteins relevant to tissue regeneration after stroke and spinal cord injury.

“It was such a surprising and unexpected discovery,” said co-lead author and recent PhD graduate Irja Elliott Donaghue, who first found that the therapeutic protein NT3, a factor that promotes the growth of nerve cells, was slowly released when just mixed into a Jello-like substance that also contained nanoparticles. “Our first thought was, ‘What could be happening to cause this?’”

Proteins hold enormous promise to treat chronic conditions and irreversible injuries — for example, human growth hormone is encapsulated in these tiny polymeric particles, and used to treat children with stunted growth. In order to avoid repeated injections or daily pills, researchers use complicated strategies both to deliver proteins to their site of action, and to ensure they’re released over a long enough period of time to have a beneficial effect.

This has long been a major challenge for protein-based therapies, especially because proteins are large and often fragile molecules. Until now, investigators have been treating proteins the same way as small drug molecules and encapsulating them in polymeric nanoparticles, often made of a material called poly(lactic-co-glycolic acid) or PLGA.

As the nanoparticles break down, the drug molecules escape. The same process is true for proteins; however, the encapsulating process itself often damages or denatures some of the encapsulated proteins, rendering them useless for treatment. Skipping encapsulation altogether means fewer denatured proteins, making for more consistent protein therapeutics that are easier to make and store.

“This is really exciting from a translational perspective,” said PhD student Jaclyn Obermeyer. “Having a simpler, more reliable fabrication process leaves less room for complications with scale-up for clinical use.”

The three lead authors, Elliott Donoghue, Obermeyer and 2016 PhD graduate Malgosia Pakulska, have shown that to get the desired controlled release, proteins only need to be alongside the PLGA nanoparticles, not inside them. Their work was published recently in the journal Science Advances.

“We think that this could speed up the path for protein-based drugs to get to the clinic,” said Elliott Donaghue.

The mechanism for this encapsulation-free controlled release is surprisingly elegant. Shoichet’s group mixes the proteins and nanoparticles in a Jello-like substance called a hydrogel, which keeps them localized when injected at the site of injury. The positively charged proteins and negatively charged nanoparticles naturally stick together. As the nanoparticles break down they make the solution more acidic, weakening the attraction and letting the proteins break free.

“We are particularly excited to show long-term, controlled protein release by simply controlling the electrostatic interactions between proteins and polymeric nanobeads,” said Shoichet. “By manipulating the pH of the solution, the size and number of nanoparticles, we can control release of bioactive proteins. This has already changed and simplified the protein release strategies that we are pursuing in pre-clinical models of disease in the brain and spinal cord.”

“We’ve learned how to control this simple phenomena,” Pakulska said. “Our next question is whether we can do the opposite—design a similar release system for positively charged nanoparticles and negatively charged proteins.”