Exploring the defective proteins associated with cancer and autism.

AlphaFold is aiding researchers in understanding how protein mutations lead to diseases and exploring ways to prevent them. Luigi Vitagliano is a Research Director at the Institute of Biostructures and Bioimaging in Naples, Italy, and he shares his experience with AlphaFold.

Being a structural biologist in the age of AlphaFold feels akin to the early days of gold mining. Before this groundbreaking technology, researchers were painstakingly sifting for individual gold nuggets, meticulously cleaning and examining them one by one. Then, suddenly, an entire gold mine materialized. It was a stroke of luck we couldn’t have anticipated.

For 30 years, I’ve delved into the proteins encoded by our DNA. Within most human cells, there exists a range of 20,000 to 100,000 different proteins. In some cases, the manner in which the amino acids in a protein fold can exhibit irregularities that are linked to various diseases.

Recently, I have been investigating a particular family of proteins known as potassium channel tetramerisation domain (KCTD) proteins, which remain largely enigmatic. Interestingly, mutations in these proteins—driven by genetic changes—are associated with a variety of diseases, including schizophrenia, autism, leukemia, colorectal cancers, and several brain and movement disorders.

Cells continuously produce new proteins, necessitating the removal of old or defective ones. Humans possess 25 types of KCTD proteins, with the majority tasked with identifying other proteins and marking them for degradation through a process called ubiquitination. This mechanism is vital for maintaining cellular health and preventing disease.

When KCTD proteins malfunction, the repercussions can significantly impact our health. Nevertheless, a considerable amount of their functions still eludes our understanding. Approximately one-fifth of the KCTD proteins within cells remained mysteries to scientists like myself, as we were unaware of their roles and how to prevent mutations that might lead to disease. Until now, a lack of structural information has obstructed research into these proteins.

The structures predicted by AlphaFold have demonstrated that, throughout evolution, the KCTD proteins’ forms have retained striking similarities, despite drastic differences in their genetic codes. This represents a significant advancement. Previously, we relied on genetic data to evaluate the similarities or discrepancies between proteins, which led us to believe that these proteins would vary greatly.

Using AlphaFold, we developed a new evolutionary family tree based on the structural characteristics of these proteins, rather than their genetic sequences. Evolutionary trees have typically been constructed using genetic information solely, often neglecting structural similarities. Given that structure relates to function, this fresh approach is exhilarating—it has the potential to uncover numerous insights into which KCTD proteins share similar functions and how these functions have evolved through time.

I employed AlphaFold to investigate and compare the structures of all 25 KCTD proteins to determine their similarities and differences, focusing on identifying critical structural components. To our delight, the predicted structures from AlphaFold have proven remarkably accurate.

For instance, we were already aware that a specific region of the KCTD proteins—the BTB domain—was conserved among all family members, leading us to believe it was the most crucial part. AlphaFold has unveiled many additional structural similarities among these proteins, paving the way for an entirely new realm of exploration.

For 60 years, including the 30 years of my own work in this field, we have grappled with the challenge of linking sequences and structures. Generations of esteemed scientists have attempted to address this issue without success. Then, unexpectedly, this solution emerged. All of our data and the structural information for the KCTD family members have come from AlphaFold, making this study possible.

I considered AlphaFold to be quite extraordinary. If someone had told me that we would amass over 200 million protein structures within two years, I would have found it hard to believe. What lies ahead in the coming decades is a quest to uncover the precise functions of these proteins. Excitement and discovery await us in abundance.