Quick Explanation
Long Explanation
DHX29 is a DExH-box RNA helicase that acts at the mRNA entry channel of the 40S ribosomal subunit. In mammals it is an initiation factor; in budding yeast, the homolog Dhx29 (Ylr419w) additionally functions during elongation.
In a fully reconstituted mammalian system, DHX29 stimulates 48S complex formation on upstream AUG codons regardless of Kozak context, does not affect near-cognate CUG recognition, and reduces leaky scanning. Mutational dissection separated two functions: RNA unwinding (retained by the 181-210 deletion mutant) from AUG-selection dominance (lost in that mutant, which fails to bind the eIF1A OB domain) .
Three independent lines of evidence now constrain the mechanism. First, in the fully reconstituted mammalian system, DHX29 binds the 40S subunit and its NTPase (ATP, GTP, UTP, CTP) activity is strongly stimulated by 43S complexes β yet DHX29 alone is a weak, non-processive helicase that unwinds only small stem-loop duplexes . Second, targeted UV cross-linking with 4-thioU mRNAs showed DHX29 makes only weak contacts with mRNA at the entry channel, the mRNA path is unchanged, and disrupting DHX29βeIF3 contacts (N terminusβeIF3a C terminus; polyQβeIF3b RRM) impairs scanning β leading the authors to conclude DHX29 acts by remodeling ribosomal complexes rather than by direct unwinding . Third, a 6.0 Γ cryo-EM structure of a mammalian DHX29-bound 43S pre-initiation complex (PDB 9CPA / EMDB EMD-3057; rabbit 40S with human DHX29 and eIF3, 49-chain heteromer, 2.25 MDa) provides a revised structural framework: the 40S subunit regulates DHX29's NTPase through an activation mechanism resembling G-patch protein regulation of DEAH helicases, and the model supports a 3'β5' translocase mechanism in which DHX29 transiently pulls mRNA opposite to the scanning direction to destabilize stem-loops trapped in the mRNA channel . Note that the transient-pulling and ribosome-remodeling interpretations are not mutually exclusive β remodeling of the entry-channel architecture could enable brief translocase-like pulling β but direct time-resolved unwinding kinetics of DHX29 on ribosome-bound structured mRNA have not yet been measured, so the pulling model remains a hypothesis awaiting falsification . In vivo, RNAi knockdown of DHX29 (>90% depletion in HeLa cells) halves protein synthesis ([35S]-methionine incorporation), disassembles polysomes with accumulation of mRNA-free 80S monomers, selectively impairs translation of mRNAs with moderately to extensively structured 5'UTRs (Cdc25C ΞGββ85 kcal/mol and XIAP ΞGββ47 kcal/mol shift to light polysomes and drop >3-fold in protein, whereas Ξ²-actin/GAPDH are minimally affected), and reduces cancer cell proliferation ~3-fold and xenograft tumor growth (25 mg vs 106 mg) . Because ~66% of human 5'UTRs have ΞG < β40 kcal/mol, these structured transcripts plausibly account for the global effect
DHX29 also aids initiation on structured endogenous circular RNAs: for circSfl, DHX29 improves efficiency and fidelity of start-codon recognition after 43S-PIC docking at AU-rich UTR patches .
Native cryo-EM of yeast translating ribosomes (2.4 Γ map from 56,757 POST-state particles) shows Dhx29 density at the mRNA entry channel in ~75% of pulled-down 80S, with its N-terminal dsRBD gripping an mRNA hairpin in the A-site. Walker A (K633A) and Walker B (E730Q) ATPase mutants enrich POST and collided ribosomes, showing hairpin resolution requires ATP hydrolysis; the ΞN dsRBD mutant loses hairpin recognition. Ribo-seq in dhx29Ξ cells shows increased occupancy upstream of structured loop motifs (GNRA, UNCG, U-turns, CUYG) .
Recruitment correlates with low-tRNA adaptation index codons, yet stalling at those codons is unchanged in dhx29Ξ (R = 0.07, P = 0.6) β evidence that Dhx29 targets structure, not codon optimality .
DHX29 has dual, context-dependent roles in cancer. On one side, RNAi knockdown data support an oncogene-like dependency: DHX29 depletion cut HeLa soft-agar colony formation at least 10-fold and reduced xenograft tumors ~4-fold (25 vs 106 mg), and ONCOMINE data mining revealed significant DHX29 mRNA overexpression in glioblastoma multiforme (P = 1.91eβ07), metastatic melanoma (P = 6.25eβ07), ovarian endometrioid carcinoma (P = 4.1eβ13), and ovarian serous adenocarcinoma (P = 7.86eβ11) . On the other side, a genome-wide CRISPR screen ranked DHX29 as a Wnt repressor (ranks 14/6), DHX29 KO raised active Ξ²-catenin and ~3-fold Cyclin-D1, DHX29 mRNA is reduced in FAP polyps, and CRISPRa activation dampens Wnt signaling β consistent with a tumor-suppressor role, though the mechanism may be indirect via its translational function . This apparent tension β pro-tumorigenic dependency in HeLa xenografts vs tumor-suppressor behavior in the Wnt screen β has not been resolved; tissue context, transcript-structure landscapes, and screen-specific artifacts could all contribute. DHX29 is also exploited by viruses: SARS-CoV-2 genomic mRNA uses DHX29-dependent leaky scanning to resist Nsp1-mediated host shut-off , and depletion of DHX29 reduces HCMV immediate-early protein synthesis and viral replication
Several key uncertainties remain. (1) The pulling model is still unvalidated dynamically: the 9CPA cryo-EM structure and AlphaFold-based modeling now provide a concrete structural framework, but no time-resolved or single-molecule unwinding experiments on ribosome-bound structured mRNA exist, so the 3'β5' translocase interpretation remains a hypothesis . Notably, an earlier UV cross-linking study reached the opposite conclusion β that DHX29 remodels ribosomal complexes rather than directly unwinding mRNA β because DHX29 only weakly cross-linked to mRNA and the mRNA path was unchanged; the two models remain in tension and are not yet distinguished experimentally . (2) The 6.0 Γ resolution of 9CPA means DHX29 domain orientations are approximate and some conformations are unresolved; a single 43S-bound state cannot capture the translocase cycle . (3) The yeast hairpin model is a consensus placeholder from mixed stem-loops; no direct in vitro unwinding assay or in vivo elongation-rate measurement exists for the yeast Dhx29 . (4) Species divergence: mammalian DHX29 is a dedicated initiation factor requiring 40S/eIF3 contacts for function, whereas yeast Dhx29 acts mainly during elongation β mechanistic conclusions from one system cannot be assumed to transfer to the other. (5) The Wnt findings are in vitro and correlative in polyps; hits vary across CRISPR screens, translational roles of DHX29 could confound signaling attribution, and the apparent oncogene-vs-suppressor tension across studies remains unresolved .
Top Data Sources
ExportMCP
Keep Exploring
Analysis Wizard
Hypothesis Graveyard
Potential Experiments
DiscussionKnow what changed, what holds up, and what remains uncertain. Every Friday. No ads.
Explore by Goal
Ask a Follow-Up
Key Insight
