Ribo-STAMP Maps Which Proteins Brain Cells Actually Make, Not Just Which Genes They Express

Gene expression studies have transformed neuroscience over the past decade. Single-cell RNA sequencing can now catalog the messenger RNA molecules present in individual neurons, generating catalogs of thousands of cell types and their molecular identities. But measuring mRNA - the intermediate step between a gene and its protein product - tells an incomplete story in neurons, because neurons routinely decouple mRNA abundance from protein production.

Messenger RNA molecules in neurons are often stored in the long branching extensions of the cell, produced in advance and held in reserve until synaptic activity demands them. The result is a large gap between which mRNAs are present in a neuron and which proteins are actually being synthesized at any given moment. Neurological conditions including autism spectrum disorder, fragile X syndrome, and tuberous sclerosis complex are all thought to involve disruptions in this translation process - but measuring translation directly in individual cells had not been feasible until now.

Ribo-STAMP: Marking What Gets Translated

A team led by Gene Yeo, PhD, professor at UC San Diego School of Medicine, and Giordano Lippi, associate professor of neuroscience at Scripps Research, developed and applied a technology called Ribo-STAMP to solve this problem. The approach works by attaching a molecular editing enzyme to ribosomes - the cellular machines that carry out protein synthesis. As a ribosome translates each mRNA into protein, the attached enzyme makes specific chemical changes to that RNA strand. Standard sequencing techniques can then detect these modifications, identifying which mRNAs were actively translated rather than merely present.

In the study, published in Nature, the team applied Ribo-STAMP to nearly 20,000 individual cells in the mouse hippocampus, a brain region essential to learning and memory that has been studied intensively enough to allow verification of novel findings against existing knowledge. The result is the first atlas of translation activity - what the researchers call the "translatome" - at single-cell resolution in the brain.

Two Memory Neuron Types Produce Proteins at Very Different Rates

The most striking finding concerned two types of neurons that are both critical for hippocampal memory circuits: CA1 and CA3 pyramidal cells. Both cell types have well-characterized roles in encoding and retrieving memories. Despite this functional similarity and their neighboring location within the hippocampus, the translation data showed that CA3 neurons were producing proteins at substantially higher rates than CA1 neurons.

The magnitude of the difference was unexpected. Previous work based on mRNA levels had suggested the two cell types were fairly similar molecularly. The translation data indicates they are more distinct than previously believed - and that the difference involves not just which proteins they produce but how vigorously they synthesize them. The high translation rates in CA3 neurons may reflect their specific role in pattern completion, the process by which the hippocampus reconstructs full memories from partial cues, which is thought to require more active synaptic maintenance.

Isoform Length Predicts How Much Protein Gets Made

A second major finding concerned mRNA isoforms. Most genes can produce multiple variants of their mRNA through alternative splicing, each encoding slightly different versions of the same protein or carrying different regulatory sequences. The Ribo-STAMP data showed that in hippocampal neurons, isoforms with longer regulatory regions - specifically longer 3' untranslated regions, the sequences that follow the protein-coding portion of the mRNA - tended to be translated at higher rates.

This has implications for disease. Many neurological conditions are associated with shifts in which isoform of a gene is expressed. If isoform identity influences translation efficiency as strongly as this study suggests, a disease-associated change in isoform expression could alter protein levels even without changing the total amount of mRNA present. That mechanism would be invisible to conventional transcriptomics but visible in translation data of the kind Ribo-STAMP generates.

High and Low Translation States Within Individual Cell Types

Beyond differences between cell types, the researchers found that individual neurons of the same type can exist in two distinct states: high translation and low translation. Neurons in the high-translation state tended to make proteins involved in synaptic communication and in cellular energy production - suggesting these neurons are more active and require more protein synthesis to sustain their activity. The existence of these states within a single defined cell type implies that simple cell type classification does not fully capture the range of functional states a neuron can occupy.

The study was conducted in mice using a transgenic approach to express the Ribo-STAMP construct in neurons. While the mouse hippocampus provides a well-validated model for studying translation biology, whether the specific quantitative differences observed between CA1 and CA3 neurons translate directly to human hippocampal biology has not been established. The study also captures a snapshot of translation activity rather than a dynamic picture; how the translatome shifts with learning, memory consolidation, or disease progression are questions that future studies will need to address.

Co-first authors were Samantha Sison and Eric Kofman at UC San Diego School of Medicine, and Federico Zampa at Scripps Research. The study was funded in part by NIH grants including MH126719, NS121223, and HG011864.