RNA writes the nucleolus

Two studies recast nucleolar architecture as an emergent product of pre-rRNA transcription and processing, with the resulting layers feeding back to improve efficiency, quality control, and evolutionary fitness.

For decades, depictions of ribosome biogenesis in the nucleolus have encouraged a structure-first narrative: transcription at the periphery of the fibrillar center (FC), early processing in the dense fibrillar component (DFC), and later maturation in the granular component (GC). That map has been enormously useful, but it also invited the causal assumption that architecture comes first and function follows. Recent studies by Quinodoz et al.¹ and Pan et al.² force a revision. The emergent model is dynamic and RNA-centered. Nascently transcribed ribosomal RNA (pre-rRNA) is not simply processed inside nucleolar layers. Instead, the synthesis and maturation stated of pre-rRNAs help define those layers, their order, and the rate at which pre-ribosomal particles move through them. The nucleolus emerges as a self-organizing reaction-partitioning device whose topology is continuously rewritten by the maturation state of the product under construction (Fig. 1).

Fig. 1: Ribosome biogenesis builds its own factory.

Left: rRNA sequence elements contribute to the formation and spatial organization of nucleolar phases. Right: classical models posited that pre-existing nucleolar subcompartments guide ordered ribosome assembly. Newer evidence supports a reciprocal model: rRNA transcription, processing, and ribonucleoprotein assemblies establish nucleolar phase organization. “Cap” denotes the nucleolar caps formed when rRNA transcription is inhibited; “inside-out” nucleolus refers to the organization seen after disruption of U3-dependent early pre-rRNA processing.

Quinodoz et al. support this case with strong mechanistic evidence. They mapped pre-rRNA processing in both time and space, showing that cleavage and modification steps are spatially segregated across nucleolar phases and that sequential maturation is linked to the outward flux of pre-rRNA through the nucleolus. Crucially, they tested causality directly. When early small-subunit (SSU) processing was impaired, either by depleting U3 snoRNA or by disrupting U3 base pairing with the 5′ external transcribed spacer (5′ ETS), the nucleolus did not simply slow down; it reorganized into an “inside-out” state. FC and DFC elements relocalized toward the GC periphery, normal phase ordering was inverted, and SSU precursors failed to undergo their normal outward flux. Their synthetic ribosomal DNA (rDNA) system went one step further: it showed that large-subunit (LSU) precursors are required to build the GC, while SSU-only constructs can still support 18S rRNA maturation.¹ Nucleolar architecture therefore tracks the identity and processing state of rRNA intermediates.

Pan et al. arrive at a complementary conclusion from a different experimental starting point. They map endogenous nucleolar organization across diverse human cellular contexts and extend the analysis evolutionarily to anamniotes. Their spatiotemporal study sharpens the substructural geography of the human nucleolus, showing that SSU processomes are maintained across the FC–DFC interface and peripheral dense fibrillar component (PDFC) domains, whereas LSU pre-rRNAs localize predominantly to PDFC–GC regions. In slowly proliferating cells, inefficient 5′ ETS-centered processing of SSU precursors is accompanied by remodeling of FC–DFC units and by reduced SSU precursor outflux; direct perturbation of 5′ ETS processing partially recapitulates these architectural changes, supporting a functional interdependence between processing state and nucleolar structure. The evolutionary comparison reinforces the same principle: anamniote bipartite nucleoli, which lack the FC, display an altered 5′ ETS distribution and slower pre-rRNA flux than multilayered amniote nucleoli. Consistently, experimental introduction of an FC–DFC interface into bipartite nucleoli enhances processing efficiency and, in bipartite zebrafish cells, accelerates pre-rRNA outflux, consistent with the idea that the emergence of nested FC–DFC units conferred a functional advantage for SSU pre-rRNA processing.²

Together, these studies challenge the structure-first model. They do not support the idea that a preassembled nucleolar architecture is the primary upstream cause of ribosome biogenesis. Instead, they support a reaction-driven model in which transcription continuously feeds pre-rRNA into the nucleolus, and processing progressively changes the interaction landscape of that RNA–protein material. Cleavage, modification, factor loading, and factor release likely alter valency, residence time, and phase preference. In the study of Quinodoz et al., this logic is made explicit by a phase-field model in which changes in pre-rRNA intermediates shift interfacial tensions and thereby reorder nucleolar phases. The broad message is that RNA does not merely travel through the nucleolus; it actively shapes its structure — much like a river carves its own bed.

Yet these papers are equally important for showing why the phrase “structure does not drive function” requires refinement rather than repetition. The most accurate interpretation is not that structure is irrelevant. Once established, architecture clearly matters. Pan et al. show that a multilayered organization with a nested FC–DFC interface can improve processing efficiency relative to bipartite nucleoli. Quinodoz et al. show that phases can act as retention checkpoints that prevent incompletely processed intermediates from escaping too early. The causal arrow is therefore asymmetric but reciprocal: transcription and processing are the primary drivers that generate nucleolar structure, and the resulting structure feeds back to sharpen, buffer, and ensure quality control for the same reactions. Function builds structure first; structure then optimizes function.

This conceptual shift extends beyond nucleolar biology: condensates are not merely passive compartments but can be sculpted by the pathways they organize, as changing intermediates alter local interactions over time. This view fits work showing that the nucleolus is a multiphase condensate, that nascent pre-rRNA sorting helps build the DFC, that maturation gradients can be visualized structurally, and that rRNA transcription and flow contribute to nucleolar material properties.^3,4,5,6,7,8

Several important questions now come into focus. How broadly do these design principles apply across species, developmental states, and stress conditions? Can synthetic nucleoli now be programmed to test how rRNA modifications or specific processing checkpoints reshape phase behavior? And might nucleolar states ultimately be classified not only by morphology but also by the landscape of pre-rRNA intermediates that gives rise to them? The nucleolus is an acidic condensate with local pH gradients;^9,10,11 how do the rRNA-processing steps described by Quinodoz et al. and Pan et al. shape local charge, pH, factor recruitment, and reaction kinetics within its subcompartments?

All in all, the deeper advances of these two studies move the field from a static picture of the nucleolus as a preexisting container for ribosome biogenesis to a dynamic picture of the nucleolus as a material state continuously produced by ribosome biogenesis itself. During cell growth, the nucleolus is not first assembled as a perfect scaffold and then used to process pre-rRNA. Rather, rDNA transcription, nascent pre-rRNA, and processing factors co-organize one another: the same activity that produces and matures pre-rRNA also helps maintain the nucleolus as a layered condensate, spatially concentrating factors in ways that support efficient ongoing ribosome biogenesis. In that sense, the nucleolus is both the factory and the product of ribosome biogenesis: a structure whose form follows, and then reinforces, function.