Draft:TimeVault

Traditional bulk and single-cell RNA sequencing methods provide a static understanding of gene expression as cells must be lysed in order to extract RNA.^[2] As a result, these approaches capture gene expression only at the time of sampling, despite the dynamic nature of biological systems.

The chronological order of molecular events within a cell is crucial for understanding cellular responses to stressed states, their ability to make decisions, and adaptability to environmental changes. Such information provides insights into developmental processes, disease progression, and the influence of molecular cues on long-term cellular behaviour.^[3]

Several approaches have attempted to address the static limitations of transcriptomics. Live-seq^[4] utilizes fluidic force microscopy to perform small cytoplasmic biopsies from living cells, enabling repeated transcriptome profiling from the same cell over time. Other techniques, including CRISPR-based DNA "typewriters" and RNA Timestamp^[5] encode temporal information into DNA or RNA through ordered barcode insertion or time-dependent editing, allowing researchers to infer when particular genes or transcripts were expressed.^{[6][7][8][9][10][11][12]}

TimeVault was developed to enable passive intracellular storage of transcriptomic information across time within the same cellular lineage. The method was developed by Dr. Fei Chen and colleagues at the Broad Institute of Massachusetts Institute of Technology and Harvard University. Introduced in 2026, the system uses engineered vault particles to store polyadenylated RNA intracellularly, creating a durable molecular record of past gene expression states that can persist in daughter cells and later be linked to cellular fate.^[1]

Vaults^[13][15][16] are large, barrel-shaped ribonucleoprotein particles mostly composed of multiple copies of major vault protein (MVP). They are naturally occurring, abundant, and highly conserved across eukaryotes.^[17] TimeVault utilizes these structures by engineering the MVP to capture messenger RNA (mRNA) via a fused poly(A) binding molecule.^[18]

The TimeVault design couples proteins engineered by fusing the vault poly(ADP-ribose) polymerase (vPARP) vault-interacting (INT) domain to the poly(A) binding protein (PABP), which has high affinity for the poly(A) tails of cytosolic mRNA.^[19][20][21] When expressed, the MVP-PABP fusion assembles into vault particles that bind polyadenylated transcripts, and protect them from degradation.

To control when transcriptomes are recorded, TimeVault expression is placed under a tetracycline-inducible promoter, which creates a defined window during which mRNA is captured. During induction, cytosolic poly(A) transcripts are captured within vault particles; after induction is turned off, the vaults persist intracellularly and can later be recovered for downstream analysis, while excluding transcripts produced after the induction period.

TimeVault enables, transcriptome-wide recording of gene expression states within living cells with minimal impact on cellular physiology. Fei Chen and colleagues showed that mRNA captured within TimeVault particles remained stable inside living mammalian cells for more than seven days providing a durable record of past gene expression states.^[1]

Since TimeVault secures cytosolic RNA through PABP's high affinity for poly(A) tails, it can capture a broad and relatively unbiased recording of the transcriptome. The TimeVault system is genetically encoded and operates intracellularly, which eliminates the need for mechanical sampling of living cells, while capturing temporally dynamic expression states within cells.^[1]

A defining property is the temporal separation between the recorded and current transcriptomes. The vault stores RNA corresponding to the expression state during the tetracycline-inducible recording window, while the cytosolic RNA reflects the cell's current transcriptome. This allows for direct comparison of past and present transcriptional profiles for the same cellular lineage, which can help identify genes and pathways that are associated with later cellular behavior.^[1]

Cells are engineered to express an inducible TimeVault construct encoding MVP and the PABP-INT fusion capture molecule. At a chosen time point or under specific conditions, such as drug exposure or stress, the TimeVault inducer is turned on and transcriptome recording is initiated.^[1]

During the recording window, vault particles assemble and capture cytosolic poly(A) transcripts, storing a copy of the transcriptome within the vault particle. Once the inducible promoter is turned off, TimeVault expression stops, but the vault-stored RNA remains stable within the cells as they continue to divide, differentiate, and respond to stimuli.^[1]

To read out the stored transcriptome, cells are lysed and TimeVault particles are isolated from the lysate. The RNA is then extracted from these vaults and prepared for downstream analysis in parallel with the cytosolic RNA to capture both the previous and current transcriptomes from the sample.^[1]

TimeVault provides a way to link historical gene expression states to later cellular outcomes across a range of biological contexts. A major application of TimeVault is studying stress response and adaptation. TimeVault can capture the transcriptome induced by stress or other stimuli, and researchers can then use that expression to better understand outcomes such as survival or differentiation within the same cell lineage.^[1]

As of 2026, a major limitation of the TimeVault technology is that vaults are only able to capture polyadenylated transcripts, meaning that non-poly(A) RNAs and nuclear-restricted transcripts are not recorded in the “past” transcriptome. This prevents complete transcriptome coverage, omitting the RNA-mediated regulatory mechanisms which can influence cellular fate.

Fei et al. mention several other limitations with the current implementation of TimeVault^[1]. Firstly, this technology is restricted to capturing the past transcriptome from a single recording interval. Since the inducible gene encoding for the TimeVault components will be transcribed as long as the inducing factor is present, it is difficult to capture RNAs within precisely defined timepoints; this is particularly challenging when the inducing factor is a drug which diffuses through the cellular medium and takes time to metabolize.

Additionally, their analysis and measurement of TimeVault were conducted via bulk RNA-seq which excludes single cell resolution of past states. It was also noted that vault protein overexpression has little phenotypic effect on growth and the transcriptome of cells; however, because the biological function of vault complexes remains incompletely understood, overexpression of the vault protein itself could potentially perturb certain biological systems.

With continued engineering and research, TimeVault could be optimized to overcome several of these limitations. For example, engineering strategies that enable multiple independently controlled recording windows, integrating the device with single‑cell RNA‑seq readouts, and minimizing or eliminating vault protein overexpression could allow multi‑interval temporal recording, single‑cell resolution of historical transcriptomes, and reduced perturbation of the underlying biological systems.