AI Identifies a ‘Hidden Protein’ and Reveals a Pathway for Material Exchange Between Cells

U.S. researchers used artificial intelligence (AI) to identify a previously uncharacterized “hidden protein” and determine that it forms pathways for material exchange between cells.

According to Azure Life Sciences and the University of Miami on the 12th (local time), a research team at the Sylvester Comprehensive Cancer Center at the University of Miami Miller School of Medicine used AI to compare more than 214 million predicted protein structures.

Searching by three-dimensional structure rather than gene sequence, the team discovered a previously hidden member of the G protein-coupled receptor (GPCR) family. GPCRs are a group of proteins that help cells detect and respond to external signals.

The researchers focused in particular on a protein called “TM184C.” Although it has a structure similar to GPCRs, it was found mainly in the membrane of the endoplasmic reticulum inside cells.

Vesicles containing TM184C moved along microtubules, the transport pathways inside cells, and gathered at thin protrusions that connect neighboring cells. These protrusions served as bridges and pathways through which cells exchanged metabolites, vesicles and organelles such as mitochondria.

When the researchers inhibited TM184C, connections between cells decreased, and cell shape and vesicle arrangement changed. This indicates that TM184C is involved in forming and maintaining pathways between cells.

Jennifer Arcuri, Ph.D., the study’s first author, said, “When we saw TM184C-positive vesicles moving through connections between cells, we realized that this structure could be a substantial route for material exchange.”

The discovery raises questions about which cells benefit when cells share resources. In healthy tissue, cells may be helped to withstand stress by moving fuel or damaged components to where they are needed. But if the exchange is imbalanced, one cell could benefit at another’s expense.

TM184C was also found to be involved in regulating autophagy, the process by which cells break down and recycle old or damaged components. In the study, reducing TM184C increased autophagy markers.

The researchers also studied Hfl1, a yeast protein similar to human TM184C. Removing Hfl1 in yeast caused problems in the cells, but introducing human TM184C resolved them. This means the function has remained similar in yeast and humans despite about 1 billion years of evolution.

Daniel Isom, Ph.D., the study’s senior author, said, “For decades, biomedical research has focused on proteins that we can identify and understand. AI provides a way to begin searching systematically.”

He emphasized, “AI should not be trusted blindly, but in the hands of experts and prepared people, it can lead to major achievements.”

The findings were published in the journal Nature.