Evidence for paper review

Rapid appraisal

The preprint reports three CRISPR Cas13 based RPI reporters: BiFC-dCas13 for imaging endogenous RNA-protein proximity, NanoBiT-dCas13 for reversible luminescent monitoring, and Split-TurboID-dCas13 for RNA-centered proximity labeling; all are validated on NEAT1-NONO paraspeckles and identify YTHDC1 as a paraspeckle-associated m6A reader [preprint]. These tools appear robust for nucleus-targeted RPI work but have limitations (irreversibility of BiFC, nuclear bias, limited cell types tested, and preprint status) that reduce immediate translational confidence [preprint; Cas13 reviews].

Key evidentiary anchors

• Preprint on Cas13 RPI tools
• Cas13 applications review

Next actions: run targeted reproduction in other cell types and provide raw proteomics lists and scripts (see Run AI Biology Analysis below).

Full critical review and analysis

1) What the paper did (concise, evidence-linked)

• Designed three dCas13 (CasRX/dCas13d) based RPI reporters by fusing split reporters to dCas13 and to protein partners: BiFC-dCas13 (split mVenus), NanoBiT-dCas13 (LgBiT/SmBiT luciferase complementation), and Split-TurboID-dCas13 (split TurboID proximity biotinylase) and optimized crRNA arrays to increase signal-to-noise when targeting NEAT1 [Primary preprint methods and optimizations]
• Validated functionally in 293T and K562 cells: BiFC-dCas13_v2/v3 visualized nuclear foci consistent with paraspeckles; NanoBiT-dCas13 produced crRNA-dependent luminescence that was reversible with 1,6-hexanediol; Split-TurboID-dCas13 biotinylated known paraspeckle proteins (SFPQ PSPC1) and identified YTHDC1 as enriched by LC-MS/MS [Preprint key results]

2) Strengths

1. Practical integration of established split reporter systems with RNA-targeting dCas13 creates modular RPI assays with complementary properties: imaging, reversible kinetics, and proteomic proximity labeling—covering orthogonal experimental needs in RPI research [Preprint describing modular toolset]
2. Optimization steps are documented: iterating crRNA arrays (single/dual/triple), different split fragment placements, dCas13 truncations and addition of dsRBD and different NLS sequences to improve S/N and nuclear targeting—this matters for reproducibility and adaptation to other RNAs [Preprint optimization data]
3. Biological discovery: using Split-TurboID-dCas13 the authors nominated YTHDC1 (an m6A reader) as a paraspeckle-peripheral factor regulating NEAT1 stability—this is an interesting mechanistic lead linking m6A-modified RNA regulation to paraspeckle biology [Preprint identification of YTHDC1]
4. Concordance with literature: Cas13/dCas13 was previously used for RNA imaging and mapping approaches; this work fills a specific gap by designing reporters to detect RPIs directly rather than tracking RNA alone [Cas13 applications review]

3) Major limitations and critical caveats

• Preprint status: the manuscript is on bioRxiv and has not completed peer review; methods and negative/control data should be scrutinized and raw datasets (full MS lists, raw images, plasmid maps, code) must be released for community validation [Preprint status and data availability]
• Cell type and cellular compartment bias: all reporter constructs include multiple NLS sequences and experiments were performed in 293T and K562 cells—conclusions are currently limited to nuclear RNAs/MLOs and these immortalized cell lines; cytoplasmic MLOs (stress granules P-bodies) were not tested and may require different localization/stability considerations [Preprint note on nuclear targeting limitation]
• Potential off-target binding and background: early BiFC-dCas13_v1 showed mVenus signal with non-targeting crNT due to non-specific dCas13 binding; while v2 redesign reduced background additional orthogonal controls (RNA FISH colocalization, crRNA mismatch series, RNase treatment, rescue by NEAT1 depletion) are essential and only some controls are shown [Preprint non specific binding issue]
• Split reporter artifacts and proximity ambiguity: proximity-driven reconstitution (BiFC NanoBiT TurboID) reports molecular proximity not direct binding; in larger RNPs or dense MLOs, false positives from high local concentration are possible—orthogonal cross-validation (co-IP CLIP eCLIP RIP-MaP or orthogonal RNA-centric pull-downs) is required to demonstrate direct RPI versus colocalization within an MLO [Cas13 usage caveats summary]
• Quantitative proteomics depth and replication: LC MS/MS reported enriched paraspeckle proteins and YTHDC1 but reporting N=2 biotinylations and absence of full supplementary raw MS tables reduces confidence; recommended: provide peptide counts fold changes FDRs volcano plots and raw spectra in PRIDE or MassIVE for independent reanalysis [Preprint proteomics replication note]

4) Specific methodological concerns and recommended controls

1. Demonstrate RNA targeting specificity: use crRNA mismatch series and tiling with scrambled guides; show loss of reporter signal after NEAT1 knockdown by orthogonal method (antisense LNA gapmer) and restore signal with NEAT1 rescue or NEAT1 mutant [Preprint guide optimization note]
2. Distinguish direct binding from proximity: combine Split-TurboID hits with crosslinking immunoprecipitation (eCLIP) for YTHDC1 and NONO to ask whether YTHDC1 crosslinks to NEAT1 sequence sites; complementary sarID or RNA-RNA mapping methods can be informative for RNA interactors [Cas13 methods need orthogonal validation]
3. Address TurboID labeling radius and temporal resolution: include negative split fragment placements and no-guide controls and map labeling kinetics (shorter biotin pulses) to reduce labeling of distal bystanders and better resolve direct neighbors [Preprint TurboID time course note]
4. Report raw data and reagents: deposit raw confocal images, plate reader luminescence traces, plasmids sequences and MS raw files to public repositories to allow reanalysis and to increase reproducibility score [Preprint data availability]

5) Interpretation of YTHDC1 result and biological plausibility

The identification of YTHDC1 as associated with NEAT1-containing paraspeckles is plausible because YTHDC1 is a nuclear m6A reader implicated in RNA processing and nuclear export and has been linked to condensate behaviors in other studies; the preprint shows YTHDC1 localizes at the periphery of NONO-NEAT1 speckles and influences NEAT1 stability per their experiments—but causality remains to be firmly established by mapping YTHDC1 binding sites on NEAT1 (eCLIP) and showing that perturbing YTHDC1 m6A recognition alters paraspeckle assembly [Preprint YTHDC1 observationsCas13 methods contextual review

6) How convincing is the evidence overall?

The technology development and proof-of-concept on NEAT1-NONO is convincing as a methods paper: multiple orthogonal reporter classes converge on paraspeckle biology and recover expected paraspeckle proteins, which argues the approach works at least in the nuclear context and in the tested cell lines. However the absence of peer review, limited cell types, N sizes for proteomics, and incomplete public data deposition reduce the current level of confidence and demand independent reproduction and expanded controls to elevate claims about direct RPIs and biological function of newly identified factors.

7) Concrete suggestions to the authors (actionable)

1. Deposit raw proteomics files PRIDE, provide peptide counts PSMs unique peptides fold changes and FDRs and upload plasmid sequences to Addgene or GenBank.
2. Provide expanded controls: NEAT1 knockdown/rescue, crRNA mismatch series, RNaseA treatment to collapse RNA-protein interactions, and eCLIP for YTHDC1 and NONO to show direct crosslinking to NEAT1.
3. Shorten TurboID biotin pulse times (1 hour 0.5 hour) and repeat with more replicates N>=3 to distinguish near neighbors from bystanders.
4. Test cytoplasmic MLOs by creating cytoplasm-localized dCas13 variants (remove NLS, add NES) and validate in stress granule models (arsenite stress) to demonstrate broader generality.
5. Release analysis scripts and raw imaging/luminescence traces and include negative controls in figures and full statistical reporting in supplements.

8) Where this fits in the field and likely impact

If validated and broadly adopted these dCas13 split reporter platforms could become standard RPI toolkits analogous to PPI split systems but RNA-centric—enabling live imaging quantification reversible HTS friendly NanoBiT assays and RNA-centered proximity proteomics. That will open routes for CRISPR-based screens to find small molecules or genetic perturbations that alter endogenous RPIs in MLO biology and RNA regulation [Cas13 field context].

9) Final balanced verdict

The manuscript presents a thoughtful and technically competent set of RNA-centric RPI tools with solid proof-of-concept on paraspeckles that recover expected components and nominate a plausible new factor YTHDC1. The work is technically creative and near-ready for utility by labs focused on nuclear RNPs; however important validation, expanded controls, and full data release are needed before firm mechanistic conclusions or broad adoption. Confidence conditional on additional data and peer review: moderate.

Author review links