Cap trimethylation as a gene-selective translation switch rather than a demonstrated export signal for mitochondrial mRNAs

Question: How does m2,2,7G cap trimethylation by TGS1 compare with other RNA cap modifications in regulating nuclear export and translation of mRNAs encoding mitochondrial proteins?

Answer

Across the retrieved papers, m2,2,7G is categorically different from the other cap and internal modifications examined, but not in the way the question's framing anticipates. TGS1 is the only factor in this corpus reported to convert an existing m7G cap into a hypermodified cap on nuclear transcripts, and it does so on more than 500 mRNAs in AML cells, with a functional concentration on nuclear mRNAs encoding TCA-cycle and oxidative-phosphorylation components whose translation is promoted at mitochondria-associated cytosolic ribosomes [1]. The comparison class behaves differently: m7G formation by RNMT, cap-adjacent 2'-O-ribose methylation, m6A, and m1A are all reported to act through defined readers or through the initiation machinery to tune translation and stability, gene-selectively by RNA context [5] [15] [2] [7]. Notably, no paper in this corpus demonstrates that any cap modification, including m2,2,7G, controls nuclear export of a mitochondrial-protein mRNA; the export literature here concerns general mRNP adaptors and exporter dosage [4], while mitochondrial mRNA targeting is attributed to motor-driven and 3'UTR-dependent mechanisms [14]. Taken together, these papers position TGS1 as a translational, not an export, regulator of mitochondrial gene expression.

TGS1's m2,2,7G mark is the only cap-terminal hypermodification in this corpus acting on nuclear OXPHOS transcripts, and its reported output is translation

The TGS1 study reports that the enzyme di-methylates the 5' cap 7-methylguanosine to generate m2,2,7G on nuclear mRNAs, including critical components of the TCA cycle and oxidative phosphorylation complexes, and that this promotes their translation mediated by mitochondria-associated cytosolic ribosomes [1]. Known TGS1 targets had previously been confined to snRNAs, snoRNAs, the telomeric RNA component and a limited number of selenoprotein mRNAs, so the >500 mRNA targets reported here substantially expand the substrate range [1]. Loss of TGS1 impairs mitochondrial respiration, raises oxidative stress, causes differentiation and cell cycle arrest in vitro and in vivo, and sensitizes AML cells to the ferroptosis inducer RSL3 [1]. Critically, the abstract describes the modified transcripts as nuclear and the translation event as occurring at mitochondria-associated cytosolic ribosomes, without attributing the phenotype to altered nucleocytoplasmic trafficking [1].

Cap architecture is decoded by structurally distinct readers, and no m2,2,7G reader has been reported here

The m7G cap is recognized by eIF4E, whose cap-binding pocket promotes recruitment of mRNAs to the initiation complex, and crystal structures of eIF4E with a series of synthetic cap analogues establish structure-activity relationships linking cap architecture to affinity and to protein expression in cells and mice [9]. Chemical probing with benzoxaborole photoaffinity probes confirms that the druggable cap pocket of eIF4E is the m7G-binding site itself [10]. A different decoding logic is described for cap-adjacent 2'-O-ribose methylation (cOMe), which stabilizes TOP ribosomal-protein mRNAs, enhances binding of the LARP1-DM15 domain to the TOP mRNA 5' end in concert with the tetratricopeptide protein Sgt/SGTA/B, and is required for localization of TOP ribosomal-protein mRNAs to Drosophila synapses [15]. By contrast, no reader, export receptor, or cap-binding protein with selectivity for m2,2,7G appears in this corpus, leaving the downstream effector of TGS1-dependent translation unidentified [1]. The only cap-recognition discrimination reported against 7-methylation is in a kinetoplastid decapping enzyme, where ALPH1 turns over cap analogues and RNAs lacking the m7G methyl group significantly faster than m7G-methylated substrates, and cleaves at the beta-gamma pyrophosphate bond to yield diphosphate-RNA [3].

Transcript selectivity is conferred by RNA context rather than by the modification chemistry alone

RNMT, the enzyme that installs the m7G cap, is recruited to RNA through direct interaction with RNA G-quadruplexes predominantly located in 5' UTRs, anchoring it adjacent to its guanosine cap substrate and biasing methylation toward transcripts encoding growth and proliferation proteins; this interaction is specific to the RNMT monomer rather than the RNMT-RAM complex [5]. This provides a general mechanism by which a cap-modifying enzyme achieves gene specificity without sequence-specific RNA binding, and the TGS1 observation that a defined subset of several hundred transcripts is targeted likewise implies context-dependent recruitment rather than global cap methylation [1]. For internal modifications the same principle holds with a different twist: the efficiency of cap-independent initiation correlates with the total m6A methylation of a transcript rather than with the precise localization of the mark, and m6A promotes association of mRNAs with eIF4F components when mTOR activity falls [2]. For m1A, TRMT61B methylates thousands of novel human mRNA sites in vitro and in cells with a YMRA single-stranded preference, but the effect of m1A on translation and ribosome recruitment varies by transcript, which the authors argue requires context-aware interpretation [7].

Localization and translation of mitochondrial-protein mRNAs are governed by motors and cis-elements that operate independently of cap methylation

Transcriptome-wide APEX-seq with minute-scale temporal resolution shows that the majority of nuclear-encoded RNAs encoding mitochondrial proteins localize to the outer mitochondrial membrane for local translation, that retrograde dynein-based motor transport is required for this targeting, and that translation efficiency of localized RNAs is itself a determinant of localization [14]. In polarized fungal hyphae, endosome-coupled transport via Rrm4 of mRNAs encoding ETC components such as Atp3 is required for efficient mitochondrial protein import at growth poles [12]. At the organelle surface, proximity labelling of TOMM20 and TOMM70 recovers distinct RNA-binding protein repertoires, including SYNJBP2, a membrane-bound RBP that binds and protects mRNAs encoding mitochondrial proteins [8], and TOMM22/MTX2 proximity proteomes identify additional RBPs supporting localized translation at the outer membrane [13]. Super-resolution STED and MINFLUX smFISH resolve individual mitochondrial mRNAs in relation to GRSF1 granules and mitoribosomes, including mRNA release during apoptosis [11]. None of these localization mechanisms is reported to depend on cap trimethylation, which constrains how much of the mitochondrial mRNA targeting phenotype TGS1 loss could plausibly explain [1].

Export is regulated at the level of mRNP adaptors and exporter dosage, a layer the cap-modification papers do not connect to

The export evidence in this corpus concerns the conserved Mex67-Mtr2 (NXF1-NXT1) exporter and its adaptors: excess Mex67 in nuclear mRNPs impairs export, and in Deltahpr1 cells increased Nab2, Yra1 and Mex67 levels accompany the export defect, with Nab2 or Yra1 overexpression suppressing it and returning Mex67 to wild-type levels [4]. Directionality in trypanosomes is attributed to asymmetric nucleoporin architecture and a Ran-based system in the absence of TREX-2 and Dbp5 homologues [6]. Neither study examines cap structure as an export determinant, and the TGS1 study reports nuclear methylation and ribosome-associated translation without measuring export kinetics of individual transcripts [1].

Conflicts and gaps

The central unresolved issue is causal ordering in the TGS1 study: because m2,2,7G is installed on nuclear mRNAs and the phenotype is read out as translation at mitochondria-associated ribosomes, the data as reported do not exclude a prior export or localization step that the cap modification facilitates [1]. No m2,2,7G-selective binding protein is identified anywhere in this corpus, so the mechanism connecting the mark to ribosome recruitment is unknown, in contrast to the structurally resolved eIF4E-m7G interaction [9] and the LARP1-DM15-cOMe reader [15]. A second disagreement concerns whether modification acts instructively or permissively: the m6A work argues that total transcript methylation, not site position, sets sensitivity to loss of cap-dependent initiation and that m6A compensates when canonical initiation is impaired [2], whereas the m1A work finds transcript-specific effects that resist a single global rule [7]. Third, cross-species generality is untested: the TGS1-mitochondrial axis is demonstrated in AML cells [1], while the mitochondrial mRNA localization machinery has been dissected in mammalian cells, fungal hyphae and Drosophila [14] [12] [15], and the two literatures do not intersect experimentally.

What would change the answer

Three experiments would settle the causal ordering. First, single-molecule FISH or nuclear/cytoplasmic fractionation of specific TGS1-target OXPHOS transcripts with and without an export block, in TGS1-depleted versus catalytically dead TGS1 knock-in cells, would separate export from translation effects [1]. Second, chemically synthesized reporters carrying either m7G or m2,2,7G caps, translated in the same cells and monitored by ribosome profiling, would test whether the mark acts directly on initiation, in the manner already applied to synthetic cap analogues for eIF4E [9] and to m1A-containing reporter pools [7]. Third, cap-analogue affinity capture or photoaffinity labelling with an m2,2,7G-specific probe, analogous to the benzoxaborole approach for the eIF4E pocket [10], would identify the reader; coupling such a reader's depletion to APEX-seq measurements of outer-mitochondrial-membrane RNA localization would directly test whether the cap mark feeds into the dynein-dependent targeting pathway [14].

References

[1] Miano, V., Carra', G., Bouvet, M., Miljkovic, J. L., Paupe, V., Shun Yu, C., Ignoti, D., Milioni, S., Manfreda, L., Viavattene, D., Benedetti, A. M., D'angelo, A., Luan, S.-L., Acquarone, D., Poli, V., Porporato, P. E., Gustincich, S., Turner, S. D., Martelli, M. P., Setti, A., Menga, A., Brunetti, L., Savino, A. M., Murphy, M. P., Pandolfini, L., Barbieri, I.. "TGS1 mediates mRNA 5'-cap trimethylation to promote oxidative phosphorylation in acute myeloid leukaemia.". bioRxiv. 2026-10-19T00:00:00.000Z. URL: https://doi.org/10.64898/2026.03.18.712389

[2] Smolin, E. A., Buyan, A. I., Buzdin, A. A., Lyabin, D. N., Kulakovskiy, I. V., Eliseeva, I. A.. "Transcript-wide m6A methylation defines the efficiency of the cap-independent translation initiation". bioRxiv. 2026-01-28T00:00:00.000Z. URL: https://doi.org/10.1101/2025.11.25.690160

[3] Pereira, L., Dzadz, D. A., Londono, P. A. C., Warminski, M., Kowalska, J., Seifert, N., Braune, S., Holetz, F. B., Zoltner, M., Kramer, S., Gorna, M.. "The trypanosome mRNA decapping enzyme ALPH1 prefers caps without m7G methylation and produces diphosphate-RNA". bioRxiv. 2026-01-29T00:00:00.000Z. URL: https://doi.org/10.64898/2026.01.29.702356

[4] Stefanyshena, N., Straesser, K.. "An elevated level of the mRNA exporter Mex67-Mtr2 in nuclear mRNPs impairs nuclear mRNA export". bioRxiv. 2025-02-28T00:00:00.000Z. URL: https://doi.org/10.1101/2025.02.26.640412

[5] Hepburn, L. A., Granneman, S., Cowling, V. H., Clara Silva, J.. "RNMT is recruited to RNA by interaction with RNA G-quadraplexes". bioRxiv. 2025-11-16T00:00:00.000Z. URL: https://doi.org/10.1101/2025.11.16.688685

[6] Gabiatti, B. P., Krenzer, J., Braune, S., Krüger, T., Zoltner, M., Kramer, S.. "A combination of expansion microscopy and proximity labelling reveals conserved and unique asymmetric functional hubs at the trypanosome nuclear pore". bioRxiv. 2024-10-26T00:00:00.000Z. URL: https://doi.org/10.1101/2024.10.04.616621

[7] Fang, D., Babich, J. M., Stanton, R., Vock, I. W., Abdallah, K., Zhu, M., Li, R., Simon, M. D., Gilbert, W. V., Nachtergaele, S.. "Investigation of TRMT61B methyltransferase activity on mRNA and its effects on translation". bioRxiv. 2025-12-08T00:00:00.000Z. URL: https://doi.org/10.64898/2025.12.08.692954

[8] Akram, S., Zittlau, K. I., Macek, B., Jansen, R.-P.. "Proximity labeling reveals differential interaction partners of the human mitochondrial import receptor proteins TOMM20 and TOMM70". bioRxiv. 2024-10-31T00:00:00.000Z. URL: https://doi.org/10.1101/2024.10.25.620316

[9] Schneider, S., Moreno, S., Thran, M., Schwarz, R. J.. "Structural basis of eIF4E recognition of synthetic mRNA 5' cap analogues links cap architecture to affinity and biological activity". bioRxiv. 2026-10-06T00:00:00.000Z. URL: https://doi.org/10.64898/2026.10.05.756672

[10] Combs, J. B., Peacock, D. M., Craven, G. B., Jung, S., Chen, Y., Le, S. M., Taunton, J., Shokat, K.. "Benzoxaboroles are structurally unique binders of eukaryotic translation initiation factor 4E". bioRxiv. 2026-07-09T00:00:00.000Z. URL: https://doi.org/10.64898/2026.02.24.707563

[11] Stoldt, S., Maass, F., Weber, M., Dennerlein, S., Ilgen, P., Gaertner, J., Canfes, A., Schweighofer, S. V., Jans, D. C., Rehling, P., Jakobs, S.. "Super-resolution microscopy of mitochondrial mRNAs". bioRxiv. 2025-02-27T00:00:00.000Z. URL: https://doi.org/10.1101/2025.01.23.634455

[12] Postma, J., Kuenzel, P., Wegmann, S. U., Muentjes, K., Devan, S. K., Sankaranarayanan, S., Krueger, S., Westhoff, P., Wierckx, N., Wijlick, L. v., Feldbruegge, M.. "Endosomal mRNA transport coordinates local mitochondrial bioenergetics during polar fungal growth". bioRxiv. 2026-06-12T00:00:00.000Z. URL: https://doi.org/10.64898/2026.06.12.731371

[13] Wang, X., Cavarischia-Rega, C., Macek, B., Jansen, R.-P.. "Proximity Labelling Reveals Spatial Organisation of Mitochondrial Protein Import Complexes". bioRxiv. 2026-07-21T00:00:00.000Z. URL: https://doi.org/10.64898/2026.07.20.739561

[14] Sharma, S., Wang, X., Nguyen, S., Rasband, M. E., Tat, T. T., Chupal, P., Chang, J.-Y., Van Nostrand, E. L., Kiss, D. L., Brown, A. I., Fazal, F. M.. "Kinetic Control of Nuclear-encoded Mitochondrial mRNA Localization and Local Translation". bioRxiv. 2026-08-24T00:00:00.000Z. URL: https://doi.org/10.64898/2026.08.22.746448

[15] Tian, Y. W., Hadzhiev, Y., Abis, G., Dodel, M., Alard, E., McQuarrie, D., Mardakheh, F. K., Conte, M., Mueller, F., Soller, M.. "LARP1-DM15/Sgt is a reader for cap-adjacent 2`-O-ribose methylation in TOP ribosomal protein mRNAs required for localization to synapses". bioRxiv. 2026-09-18T00:00:00.000Z. URL: https://doi.org/10.64898/2026.09.17.752351

How this survey was made

This survey answers one question: “How does m2,2,7G cap trimethylation by TGS1 compare with other RNA cap modifications in regulating nuclear export and translation of mRNAs encoding mitochondrial proteins?”

We keep our own collection of recent preprints and refresh it every day from bioRxiv, medRxiv, arXiv, PsyArXiv, SocArXiv and ChemRxiv. For each survey, we search that collection for recent work and keep the strongest matches: papers that use the same terms as the question, and papers on the same topic in different words.

The text was written from the titles and abstracts of those papers; the full texts were not read. Statements that draw on a particular paper are marked with a number in brackets. The reference list just above is the 15 papers whose work this text draws on, chosen from the 106 papers those searches returned; the rest of the shortlist is not printed. It is not a complete review of everything published on the question.

Treat this as a starting point and a reading list, not a peer-reviewed review. The searches behind this survey ran on October 21, 2026 and covered work from the previous two years.