CRISPR imaging system makes plant RNA visible one molecule at a time

For years, biologists have been able to watch individual RNA molecules move through living animal cells in exquisite detail, capturing everything from transcriptional bursts to the delivery of transcripts to distant synapses. Plants, however, have stubbornly resisted the same scrutiny. Now a team at Sichuan University reports in PLOS Biology a CRISPR-based imaging platform that finally brings single-transcript RNA visualization to living plant cells, and they have already used it to film the flowering signal FLOWERING LOCUS T (FT) as it docks at plasmodesmata, squeezes between cells, and travels long distances through grafted plants.

The core obstacle is architectural. Mature plant cells are dominated by a huge central vacuole that compresses the cytoplasm into a thin layer pressed against the cell membrane. Any fluorescent probe floating freely in that confined space becomes intensely concentrated, producing a diffuse background haze that drowns out the faint signals from individual RNA molecules. Chloroplast autofluorescence compounds the problem. Conventional live-cell RNA imaging methods, such as the MS2 and PP7 aptamer systems, also require stitching long repetitive tag sequences into the target RNA itself, raising concerns that the observed behavior reflects the artificial tag rather than the transcript’s native life.

The CRISPR–Cas13 family offered a tag-free alternative: catalytically inactive dCas13 proteins can be guided by CRISPR RNAs to bind endogenous transcripts without altering their sequence, and coupling them to multivalent scaffolds such as the SunTag amplifies the signal enough to see scarce RNAs. But when the Sichuan team, led by Jiayu Zhang and Jiuyuan Bai, transplanted a dCas13–SunTag system into tobacco leaf cells, the vacuole-driven background was so severe that single RNA puncta were effectively invisible. The system needed to be not just bright, they reasoned, but smart—able to tell the difference between probes that had found their target and probes that had not.

Their solution was a binding-activated stability switch built from a new plant-specific degron. The team first tested ten previously reported degron sequences in Nicotiana benthamiana, but none eliminated GFP fluorescence strongly enough. So they turned to quantitative proteomics, treating leaves with the proteasome inhibitor MG132 and using LC–MS/MS to find proteins that accumulate when proteasomal degradation is blocked. From 86 significantly upregulated candidates, they screened the C-terminal peptides of the top 30 and found six that drove GFP degradation. The most potent, derived from a tobacco proteinase inhibitor I-B-like protein, suppressed normalized GFP signal to less than a tenth of control levels. Truncation mapping pinned the activity to the final 21 amino acids, and the team named the motif cPIL, for C-terminal Proteinase Inhibitor-Like degron.

Installing cPIL as a switch required some protein engineering. The researchers circularly permuted the single-chain antibody fragment that binds the SunTag, relocating its N- and C-termini to the vicinity of the antigen-binding pocket so the degron could be fused close to the binding interface. Of ten permuted designs, only one—termini adjacent to Asn194—still produced fluorescent puncta when co-expressed with SunTag, and it bound the scaffold as efficiently as the wild-type scFv. In the unbound state, the exposed cPIL degron hands the free GFP–cpScFv probe to the proteasome; when the probe binds a SunTag array on an RNA-bound dCas13X complex, the degron is protected and the fluorescence persists at the target site. Cycloheximide chase experiments showed the cPIL fusion losing roughly 80 percent of its fluorescence within 30 minutes, while MG132 treatment rescued it, confirming proteasome-dependent turnover.

The payoff was dramatic. When the team co-expressed dCas13X–24×SunTag with either the cPIL-tagged or untagged reporter in cells carrying a target mRNA, the cPIL fusion cut diffuse background fluorescence by about 6.7-fold while leaving the brightness of individual RNA foci essentially unchanged, boosting the signal-to-noise ratio by roughly 6.7-fold. The team chose the compact Cas13X protein, only about 445 amino acids long, as the RNA-recognition module, an engineering decision aimed at keeping the overall complex small enough to have any hope of trafficking through plasmodesmata, which enforce strict size limits on molecular passage.

To prove the system could see native transcripts, the researchers generated stable Arabidopsis thaliana lines carrying the complete imaging cassette on a single construct and targeted the cytoskeletal ACT2 mRNA. Discrete nuclear and cytoplasmic puncta appeared in leaves and roots, and benchmarking against single-molecule FISH in fixed tissue showed strong colocalization—87.0 plus or minus 3.8 percent—with punctum sizes and per-cell counts matching the fixed-tissue gold standard, while the live system actually achieved higher signal-to-noise. An independent MS2/MCP reporter assay confirmed guide-dependent specificity, with about 86 percent of GFP puncta overlapping mCherry-marked reporter RNAs when an ACT2-targeting guide was used, and two independent guides performed comparably. Crucially, the imaging system left target transcript abundance, protein levels, and mRNA decay rates undisturbed.

The platform generalized readily. Targeting CORTEX4, IRX9, and ATPP2-A1—transcripts enriched in cortex, xylem, and phloem respectively—the team resolved diffraction-limited foci in the correct cell files of living Arabidopsis roots at three, five, and seven days after germination, with spatial patterns matching published single-cell transcriptomic atlases and quantifiable developmental dynamics.

The biological payoff came from FT, the mobile mRNA behind the florigenic signal that leaves send to the shoot apex to trigger flowering. Whether FT mRNA itself moves systemically has long been debated, since FT protein is the established mobile signal. Time-lapse imaging showed FT mRNA puncta accumulating at plasmodesmata marked by PDCB1–BFP, dwelling there for roughly 20 seconds, then crossing the cell boundary in about 6 seconds, at an average of about 12 successful transfer events per minute per cell interface. In grafting experiments, ft mutant scions carrying the imaging system but no FT transcript lit up with fluorescent puncta two weeks after being grafted onto wild-type rootstocks—only when equipped with an FT-targeting guide—and RT-PCR and sequencing confirmed full-length FT mRNA in the scions, directly visualizing rootstock-to-scion transport.

The team also identified a trafficking factor. RNA pull-down with biotin-labeled FT sense RNA followed by mass spectrometry highlighted three candidates, including the well-known RNA-binding protein GRP7, also called AtSRBP1. Grafting the reporter scion onto grp7 mutant rootstocks sharply reduced FT signals in the scion, and time-lapse imaging in the grp7 background revealed a striking phenotype: FT puncta still reached plasmodesmata-associated regions but never crossed during the three-minute observation window. The data point to GRP7 acting at or after plasmodesmata docking, though the authors caution that GRP7 binds hundreds of RNAs with limited sequence selectivity, so it may promote a transport-competent FT ribonucleoprotein complex rather than recognize FT specifically. The system has limits—the complex remains large, guide design is partly empirical, and cPIL has so far been validated only in a GFP-based configuration—but as a first-generation toolkit for watching plant RNA biology in real time, it opens a window onto one of botany’s most consequential messages: the molecular courier that tells a plant when to flower.

Subject of Research: Live-cell single-molecule RNA imaging in plants using a binding-activated CRISPR–dCas13X system

Article Title: A binding-activated CRISPR/dCas13X imaging system enables precise RNA tracking in plants

Article References: Zhang, J., Tao, Y., Zhao, Y., Lei, M., Shen, Y., Li, Q., Liu, C., & Bai, J. (2026). A binding-activated CRISPR/dCas13X imaging system enables precise RNA tracking in plants. PLOS Biology, 24(10), e3004043. https://doi.org/10.1371/journal.pbio.3004043

Image Credits: AI Generated

DOI: 10.1371/journal.pbio.3004043

Keywords: CRISPR, dCas13X, RNA imaging, plasmodesmata, FT mRNA, florigen, Arabidopsis thaliana, degron, GRP7, signal-to-noise ratio, grafting, plant RNA biology