Science1 publisher2 min readPublished
A shape-first search of 214 million predicted proteins turned up a GPCR-family protein inside the cell
Sylvester researchers compared the shapes of more than 214 million predicted proteins, then followed one hit, TM184C, to vesicles riding microtubules into bridges between cells. The human protein rescued a yeast mutant.
The Scientist · Science desk
What happened
- A Sylvester Comprehensive Cancer Center team compared the shapes of more than 214 million predicted proteins with AI, and reported the results in Nature.
- Searching by three-dimensional form instead of genetic sequence returned hidden members of the G protein-coupled receptor family, which lets cells sense signals from outside.
- Disrupting TM184C left cells forming fewer of those connections, with changes in cell shape and in how vesicles were organised.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- capability Labs can now nominate members of a receptor family whose sequences do not declare the relationship. Confirming each candidate is still cell-biology work.
- constraint The paper gives no count of family hits and no ligand for TM184C, so this work cannot yet be used to size a drug-target list, only to show one protein was worth chasing.
- precedent A cross-species rescue is a stiffer test than a co-localisation image, and structure-first papers that follow will be read against it.
Structural similarity is a hypothesis about function, and the Miami group tested theirs in an organism that parted from us long ago. Deleting Hfl1, the yeast protein that resembles TM184C, left the cells with problems, and adding the human protein fixed them [10]. The study reports the function has stayed similar in yeast and humans across about a billion years of evolution [11]. A knockdown phenotype in one human cell line can arise by many indirect routes. A cross-species rescue asks the human protein to hold down a specific job in a cell that has never expressed it.
"For decades, we have largely explored protein biology using sequence as our guide," said Daniel G. Isom, the study's senior author and a faculty member in the Molecular and Cellular Pharmacology Department at Sylvester [3]. His team compared three-dimensional form across more than 214 million predicted proteins, and pulled out GPCR-family members that sequence work had not flagged [1][2]. "What we found suggests there is another layer of biology that has been hiding in plain sight," Isom said [4]. The account names one protein from that search, TM184C. It does not give a count of family hits, or a ligand for TM184C [14]. On this record the shape search is vindicated by a single protein that survived a functional test in two organisms.
Much of TM184C sits inside the cell, in the membranes of vesicles, which is not where a receptor is expected. Those vesicles travel along microtubules and collect in thin projections that join neighbouring cells, through which cells trade metabolites, vesicles and organelles, mitochondria included [5][6]. "When we saw TM184C-positive vesicles moving through connections between cells, we realized these structures could be routes for substantial material exchange," said Jenniffer Arcuri, a senior scientist at Sylvester and the paper's lead author [8].
Reducing the protein also raised autophagy markers. The team takes that as evidence that TM184C tunes the recycling process [9]. Tumour cells often grow with limited oxygen and nutrients, and a bridge that carries mitochondria is one more way for some of them to get by [12]. "As part of my doctoral work in the Isom lab, I am studying how these connections occur in normal cells and aggressive cancers like glioblastoma," said Bruno Colon, a graduate student in Molecular and Cellular Pharmacology [13].
What to watch
- Whether the Nature paper's supplementary data lists the other structure-only GPCR-family hits, and how many are expressed in human cells.
- A ligand or G-protein coupling for TM184C would move it from a family assignment to a pharmacological target.
- Whether the bridge phenotype reproduces in glioblastoma cells, which Colon's doctoral work is testing.