Trupej, A.; Bergant, V.; Novljan, J.; Dodel, M.; Klobucar, T.; Adamek, M.; Lee, F. C. Y.; Yap, K.; Makeyev, E.; Kokot, B.; Cehovin Zajc, L.; Pichlmair, A.; Urbancic, I.; Mardakheh, F. K.; Modic, M.

The spatial organisation of RNA condensates is fundamental for understanding of basic cellular functions, but may also provide pivotal insights into diseases. One of the major challenges to understanding the role of condensates is the lack of technologies to map condensate-scale protein architecture at subcompartmental or nanoscale resolution. To address this, we introduce HCR-Proxy, a proximity labelling technique that couples Hybridization Chain Reaction (HCR)-based signal amplification with in situ proximity biotinylation (Proxy), enabling proteomic profiling of RNA-proximal proteomes at subcompartmental resolution. We benchmarked HCR-Proxy using nascent pre-rRNA targets to investigate the distinct proteomic signatures of the nucleolar subcompartments and to uncover a spatial logic of protein partitioning shaped by RNA sequence. Our results demonstrate HCR-Proxy\'s ability to provide spatially-resolved maps of RNA interactomes within the nucleolus, offering new insights into the molecular organisation and compartmentalisation of condensates. This subcompartment-specific nucleolar proteome profiling enabled integration with deep learning frameworks, which effectively confirmed a sequence-encoded basis for protein partitioning across nested condensate subcompartments, characterised by antagonistic gradients in charge, length, and RNA-binding domains. HCR-Proxy thus provides a scalable platform for spatially resolved RNA interactome discovery, bridging transcript localisation with proteomic context in native cellular environments.