A new study in Nature has mapped, at atomic resolution, how the enzyme DICER slices precursor microRNAs at exactly the right spot, with accuracy down to a single nucleotide. The research, which combined massively parallel biochemical assays with cryo-electron microscopy, identifies two distinct binding pockets at the 5-prime end of RNA substrates that separately control where DICER cuts and how reliably it does so. Because microRNAs regulate thousands of human genes, even a one-nucleotide shift in processing can redirect which messenger RNAs get silenced, with direct consequences for normal cell function and disease.
Two Pockets, Two Jobs
For years, the dominant model held that DICER acts as a molecular ruler, measuring a fixed distance from the end of a precursor RNA hairpin to decide where to cut. That picture turns out to be incomplete. The 2026 Nature paper demonstrates that DICER’s cleavage position and fidelity depend on separate 5-prime binding sites. One pocket anchors the RNA and sets the general cut site; the other fine-tunes precision so the enzyme reproducibly generates the same product length from a given substrate.
The distinction matters because it explains a long-standing puzzle: why DICER can be extremely precise on some precursors yet produce a mixture of product lengths on others. Earlier structural work had already captured a human DICER–pre-miRNA complex in a catalytic conformation, showing how the precursor sits inside the enzyme during cleavage. The new study reinterprets those structures, arguing that the two pockets operate semi-independently and that their relative grip on a given RNA substrate determines whether the cut lands cleanly or drifts by one nucleotide.
In this revised framework, DICER still behaves as a ruler, but a more nuanced one. The enzyme gauges distance from the 5-prime end, yet the exact register of the cut depends on how tightly the substrate’s terminus is clamped in each pocket. When both pockets engage strongly, the active site is positioned in a single, preferred configuration. When one pocket releases or tolerates alternative contacts, the same precursor can be diced into slightly different products, generating microRNA isoforms that differ by a nucleotide at either end.
A Guanine Boost to Cutting Accuracy
One of the study’s more surprising findings is that a guanine (G) at the 5-prime end of the precursor RNA can increase DICER’s precision across many substrates. This runs counter to earlier proposals about which nucleotides the enzyme prefers at that position. The researchers used massively parallel dicing assays, processing thousands of RNA variants simultaneously, to quantify how each 5-prime nucleotide identity affected product-length distributions. The cryo-EM data then showed the structural basis: a 5-prime G appears to lock into the relevant pocket more tightly, reducing wobble at the active site.
That observation connects to foundational work showing that anchoring the 5-prime terminus modulates both efficiency and accuracy of miRNA production. In that earlier study, small-RNA sequencing from cells expressing wild-type or pocket-mutant Dicer revealed that substantial fractions of microRNAs had altered cleavage sites when the pocket was disrupted. The new data now offer a mechanistic explanation for those shifts: when the pocket cannot grip the 5-prime nucleotide properly, DICER loses its single-nucleotide accuracy and generates broader distributions of product lengths.
Guanine’s stabilizing effect also helps rationalize why certain endogenous microRNA precursors show remarkably tight length control. Many highly conserved precursors carry 5-prime nucleotides that favor strong pocket engagement, biasing the enzyme toward a single dominant product. Conversely, precursors that lack those preferred identities, or that present unusual terminal chemistries, are more prone to heterogeneous processing and to the generation of multiple isoforms with distinct regulatory capacities.
Internal RNA Features Add Another Layer
The 5-prime pocket is not the whole story. Independent high-throughput studies have shown that internal RNA elements such as loops and bulges reshape DICER cleavage patterns, meaning the enzyme does not rely on end-counting alone. Internal mismatches, asymmetric bulges, and loop positioning all shift where the cut lands, sometimes by exactly one nucleotide.
Separate work on the combined sequence and structural features that guide dicing selectivity reinforces this point. That study found that substrate architecture and specific DICER residues interact to steer cleavage, with end preferences and internal motifs jointly directing the outcome. Local nucleotide identity at or near the cleavage site can also bias the cut position, as demonstrated by experiments in which single-base substitutions in precursor hairpins shifted the preferred product. Together, these results paint a picture of DICER as a context-sensitive processor, not a simple ruler.
This context dependence is biologically useful. By embedding regulatory information in both the ends and the interior of precursor hairpins, cells can tune not only whether a given microRNA is produced but also which exact isoform predominates. Small changes in stem stability or loop geometry (arising from RNA editing, alternative transcription start sites, or polymorphisms) can subtly change how the hairpin is presented to DICER’s active site, nudging the enzyme toward one cleavage register or another.
Cofactors Fine-Tune the Final Product
Even after DICER’s own pockets and the RNA’s internal features have set the stage, protein cofactors add a final adjustment. The transactivation response RNA-binding protein TRBP forms a complex with DICER and can shift the cleavage position for selected pre-miRNAs and alter length distributions. For certain substrates, the presence or absence of TRBP moves the dominant product between 20, 21, and 22 nucleotides in length.
This cofactor layer helps explain why single-nucleotide precision is biologically variable and substrate-dependent rather than absolute. DICER alone sets a baseline, but TRBP and related partners can nudge the cut by one position on specific precursors, likely by stabilizing alternative conformations of the DICER–RNA complex. The practical implication is that predicting DICER’s output for any given microRNA requires knowing not just the RNA sequence and structure but also which cofactors are present, in what stoichiometry, and under which cellular conditions.
Because cofactor expression is regulated by signaling pathways and stress responses, cells can dynamically reprogram microRNA isoform profiles without changing the underlying DNA sequence. In principle, this offers a rapid mechanism to adjust gene-silencing networks. By up- or downregulating DICER partners, cells could bias processing toward isoforms that favor one set of targets over another.
Why One Nucleotide Reshapes Gene Silencing
A single-nucleotide shift in where DICER cuts changes the terminal bases of the mature microRNA and can alter which strand is loaded into Argonaute. That alteration may redirect which messenger RNAs the microRNA targets, because recognition depends heavily on a short “seed” sequence at the 5-prime end of the guide strand. Changes in cleavage register can therefore swap one seed for another or subtly extend or truncate it, reconfiguring target networks even when the precursor hairpin itself is unchanged.
Computational and experimental analyses show that such “isomiRs” frequently have distinct sets of predicted and validated targets, and that their expression can vary across tissues and disease states. In cancer, for example, small shifts in seed usage may tip the balance between repression of tumor suppressors and oncogenes. In the nervous system, where microRNA regulation is tightly layered on developmental timing and synaptic plasticity, altered processing precision could contribute to pathological gene-expression programs.
The new structural view of DICER’s dual 5-prime pockets therefore has implications beyond basic enzymology. It suggests concrete strategies for therapeutic intervention: engineering precursor mimics with favored 5-prime nucleotides, optimized internal structures, and defined cofactor environments could steer DICER toward a desired isoform with high fidelity. Conversely, selectively perturbing one pocket or its interaction partners might be used to deliberately broaden product distributions, dampening the impact of any single microRNA species.
As more high-resolution structures and large-scale cleavage datasets accumulate, an integrated model is emerging in which DICER reads multiple layers of information (terminal chemistry, local sequence, global hairpin shape, and cofactor context) to position its molecular ruler. The discovery that two distinct pockets at the 5-prime end separately control cut placement and precision turns what once seemed like a simple measurement problem into a nuanced decoding of RNA features, with each nucleotide and structural element contributing to the final gene-silencing outcome.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
