Synthetic organelle records fleeting gene expression in single human cells

A burst of gene activity that helps a cell respond to a drug, an immune signal or a change in its environment often fades within hours, long before a researcher has a chance to look. Standard RNA sequencing destroys the cell to read its messenger RNAs, so it can only ever deliver a snapshot of the moment of lysis. A team at ETH Zurich has now built a way to let the cell keep its own diary. In a study published in Nature Biotechnology, Christopher Reinkemeier, Marcos Manero Carranza, Randall Platt and colleagues describe the recordanelle, a synthetic organelle made from a single fusion protein that sequesters RNAs inside living human cells and preserves them for days, so that the transcriptional history of a cell can be read out long after the triggering event has passed.

Transcriptomes are dynamic: cells constantly adjust their RNA repertoires to respond to stimuli, to cope with environmental changes and to communicate with neighbouring cells. Many of these adjustments are transient, and destructive readout methods such as RNA sequencing cannot follow them over long periods at transcriptome-wide scale. Molecular recording technologies have been invented to tackle exactly this problem. In bacteria, CRISPR spacer acquisition from RNA can write a permanent DNA-encoded record of past transcriptional events. In mammalian cells, prime editing has been adapted for a similar purpose. RNA itself can also serve as the storage medium, either by exporting it from the cell for continuous sampling or by storing it inside protein capsules. Yet each of these strategies carries a limitation: some work only in prokaryotes, some infer transcriptional activity indirectly, some sample only a small number of RNA species, and some lack single-cell sensitivity.

The recordanelle takes a deliberately minimalist route. It is built with one fusion protein that carries three functionalities in a single chain. A membrane anchor tethers the compartment to the surface of mitochondria, which has the practical benefit of aiding isolation of the recordanelle from the cell at the end of an experiment. A phase-separating protein forms the organelle inside the cell. And an RNA-binding protein captures transcripts, pulling cellular RNAs into the compartment as they are produced. The construct is expressed in cells, RNAs accumulate inside the organelle during the recording window, and at the end of the experiment the stored transcript information is retrieved to reveal past gene-expression events.

The first demonstrations were carried out in HEK293T cells, a widely used human cell line. The researchers showed that RNAs are bound by the recordanelle and that the RNA content is retained during the isolation procedure, a nontrivial requirement because any loss during purification would distort the record. Specificity is tunable: by swapping in different RNA-binding proteins, the compartment can be directed at selected transcripts. When the team equipped the recordanelle with a variant of human poly-A binding protein 4 to sample the full-cell transcriptome, it captured a representative part of 90 percent of all expressed genes.

The key test was whether the recordanelle could hold onto information about an event that is genuinely fleeting. The researchers turned to interferon-β, an immune signalling molecule that triggers a well-characterized transcriptional program. Cells were treated with a mild dose of interferon-β for 20 hours, after which the stimulus was removed. Both control cells and recordanelle-containing cells responded with the characteristic interferon response, confirming that the synthetic organelle does not blunt the biology it is meant to observe. In control cells, that interferon signature disappeared after just one day. In recordanelle-containing cells, however, the transient information could be retrieved with quantitative PCR or RNA sequencing for at least five days.

Just as important is what the recordanelle did not do. The team reports that the organelle had no measurable effect on the proteome, on cellular viability or on mitochondrial function. The authors also note that the technology works not only at the bulk population level but is sensitive enough to capture the transient interferon response in single cells. In single-cell RNA-sequencing analysis, cells expressing the recordanelle and cells carrying a negative control both showed an expression module of interferon-stimulated genes at day zero, immediately before the stimulus was removed. Two days later, the signal had vanished in the control cells but could still be faithfully detected in the recordanelle-expressing cells, shown as an interferon-stimulated gene expression module on a uniform manifold approximation and projection plot.

The Platt laboratory had previously developed transcriptome recording based on CRISPR spacer acquisition from RNA, but that approach is native to prokaryotes and translating it to mammalian cells proved difficult. The team instead chose a strategy tailored to the molecular biology of eukaryotes. Reinkemeier had earlier developed orthogonally translating organelles, compartments that require the selective handling of engineered messenger RNAs. The hypothesis was that a similar compartmental principle could be extended to endogenous RNAs. Work began in earnest when Manero Carranza joined the laboratory, and the researchers report that it was initially not obvious that the recordanelle could be purified while maintaining its RNA cargo. The process, they write, worked surprisingly well, and systematic optimization of the organelle's functionalities resulted in a robust technology that records information for much longer and more efficiently than they had hoped.

An independent expert weighs in on the work in the same journal. Steven Boeynaems of Baylor College of Medicine writes that the experiments provide evidence that the recordanelle is overall tolerated by cells and that the authors propose control experiments to ensure the tool can be adapted by others in the field and explored in additional cellular and organismal contexts. He states that he strongly supports publication of the work, calling it a breakthrough for the field that will undoubtedly have multiple applications. The Nature Biotechnology editorial team, in a note accompanying the paper, describes the mitochondria-anchored synthetic organelle as capturing and preserving RNA expression histories, an approach it characterizes as unique among transcriptional recording technologies.

The practical implications reach into several areas of biology and medicine. Because recordanelles capture and store transiently expressed RNAs and allow that information to be retrieved days after the causative event has subsided, the method can indicate what a cell has experienced in the past. The authors point to drug resistance in cancer cells as a candidate application: it is currently hard to determine whether cells had a certain transcriptome that primes them to resistance or whether their resistance arises during drug treatment, and a recorder that preserves early transcriptional events could help disentangle the two. By keeping long-term transcriptional records inside cells, the technology could also replace multi-timepoint RNA-sequencing experiments and thereby reduce the cost of temporal transcription studies. Looking further ahead, the team lists development, differentiation, aging, tissue regeneration and tumorigenesis as processes whose underlying gene-expression events could become accessible if the technology can be extended.

Significant caveats remain, and the authors are explicit about them. So far, recordanelles have been shown to work only in HEK293T cells, and it remains to be seen whether they will perform equally in other systems. Because the organelle operates by enriching and protecting RNAs, recording the downregulation of transcripts will be difficult, meaning the device is inherently better at logging what a cell turned on than what it switched off. The stated future directions are to translate the technology to in vivo models and to extend the recording time beyond the current several-day window. Those goals, if achieved, could make it possible to obtain a detailed transcriptional past of cells during complex biological processes.

Subject of Research: A synthetic mitochondria-anchored organelle that records transient gene expression in single human cells

Article Title: Recordanelles capture transient gene expression in single human cells

Article References: Recordanelles capture transient gene expression in single human cells. (2026). Nature Biotechnology. https://doi.org/10.1038/s41587-026-03350-7 (Author names are not included in the publisher’s citation export.)

Image Credits: AI Generated

DOI: 10.1038/s41587-026-03350-7

Keywords: recordanelle, synthetic organelle, RNA recording, single-cell RNA sequencing, transient gene expression, interferon-β, HEK293T cells, transcriptome, molecular recording, phase separation, poly-A binding protein, drug resistance