Indirect development with an intermediate larva exists in all major animal lineages1, which makes larvae central to most scenarios of animal evolution2–11. Yet how larvae evolved remains disputed. Here we show that temporal shifts (that is, heterochronies) in trunk formation underpin the diversification of larvae and bilaterian life cycles. We performed chromosome-scale genome sequencing in the annelid Owenia fusiformis with transcriptomic and epigenomic profiling during the life cycles of this and two other annelids. We found that trunk development is deferred to pre-metamorphic stages in the feeding larva of O. fusiformis but starts after gastrulation in the non-feeding larva with gradual metamorphosis of Capitella teleta and the direct developing embryo of Dimorphilus gyrociliatus. Accordingly, the embryos of O. fusiformis develop first into an enlarged anterior domain that forms larval tissues and the adult head12. Notably, this also occurs in the so-called ‘head larvae’ of other bilaterians13–17, with which the O. fusiformis larva shows extensive transcriptomic similarities. Together, our findings suggest that the temporal decoupling of head and trunk formation, as maximally observed in head larvae, facilitated larval evolution in Bilateria. This diverges from prevailing scenarios that propose either co-option9,10 or innovation11 of gene regulatory programmes to explain larva and adult origins.

Indirect development with an intermediate larva exists in all major animal lineages, and thus larvae are central to most scenarios for animal evolution. Yet how larvae evolved remains disputed. Here we show that changes in the timing of trunk formation underpin the diversification of larvae and bilaterian life cycles. Combining chromosome-scale genome sequencing with transcriptomic and epigenomic profiling in the slow-evolving oweniid Owenia fusiformis, we found that different genes and genomic regulatory elements control the development of its feeding larva and adult stage. First, O. fusiformis embryos develop into an enlarged anterior domain that forms larval tissues and the adult head, as posterior growth and trunk patterning is deferred to pre-metamorphic stages. These traits also occur in the so-called "head larvae" of other bilaterians, with whom O. fusiformis larva shows extensive transcriptomic similarities. Conversely, animals with non-feeding larvae and gradual metamorphoses, such as the annelid Capitella teleta, start trunk differentiation during embryogenesis, like direct developers. Together, our findings suggest that the ancestral temporal decoupling of head and trunk formation, as retained in extant "head larvae", allowed larval evolution in Bilateria, questioning prevailing scenarios that propose either co-option or innovation of gene regulatory programmes to explain larva and adult origins.

Fox genes are a large and conserved family of transcription factors involved in many key biological processes, including embryogenesis and body patterning. Although the role of Fox genes has been studied in an array of model systems, comprehensive comparative studies in Spiralia—a large clade of invertebrate animals including molluscs and annelids—are scarce but much needed to better understand the evolutionary history of this gene family. Here, we reconstruct and functionally characterise the Fox gene complement in the annelid Owenia fusiformis, a slow evolving species and member of the sister group to all remaining annelids. The genome of O. fusiformis contains at least a single ortholog for 20 of the 22 Fox gene classes that are ancestral to Bilateria, including an ortholog of the recently discovered foxT class. Temporal and spatial expression dynamics reveal a conserved role of Fox genes in gut formation, mesoderm patterning, and apical organ and cilia formation in Annelida and Spiralia. Moreover, we uncover an ancestral expansion of foxQ2 genes in Spiralia, represented by 11 paralogs in O. fusiformis. Notably, although all foxQ2 copies have apical expression in O. fusiformis, they show variable spatial domains and staggered temporal activation, which suggest cooperation and sub-functionalisation among foxQ2 genes for the development of apical fates in this annelid. Altogether, our study informs the evolution and developmental roles of Fox genes in Annelida and Spiralia generally, providing the basis to explore how regulatory changes in Fox gene expression might have contributed to developmental and morphological diversification in Spiralia.

Bacterial symbioses allow annelids to colonise extreme ecological niches, such as hydrothermal vents and whale falls. Yet, the genetic principles sustaining these symbioses remain unclear. Here, we show that different genomic adaptations underpin the symbioses of phylogenetically related annelids with distinct nutritional strategies. Genome compaction and extensive gene losses distinguish the heterotrophic symbiosis of the bone-eating worm Osedax frankpressi from the chemoautotrophic symbiosis of deep-sea Vestimentifera. Osedax’s endosymbionts complement many of the host’s metabolic deficiencies, including the loss of pathways to recycle nitrogen and synthesise some amino acids. Osedax’s endosymbionts possess the glyoxylate cycle, which could allow more efficient catabolism of bone-derived nutrients and the production of carbohydrates from fatty acids. Unlike in most Vestimentifera, innate immunity genes are reduced in O. frankpressi, which, however, has an expansion of matrix metalloproteases to digest collagen. Our study supports that distinct nutritional interactions influence host genome evolution differently in highly specialised symbioses.

The mechanical properties of the extracellular matrix (ECM) determine cell differentiation, proliferation and migration through mechanoresponsive proteins including YAP. However, how different mechanical signals cooperate, synergize or compete to steer cell behavior remains poorly understood. Here, we have examined competition between the two major ECM mechanical cues, i.e. rigidity, which activates cell mechanosensing, and viscous energy dissipation, which reduces stiffness blunting cell mechanotransduction. To trigger competition, we have engineered protein hydrogels allowing concomitant modulation of stiffness and viscosity by mechanisms characteristic of native ECM. Culturing cells on these hydrogels, we have found that substrate energy dissipation attenuates YAP mechanosensing prevailing over stiffness cues. Hampered YAP activation on more dissipative substrates correlates with faster actin flow and smaller focal adhesions. Mechanistically, inhibition of actomyosin contractility reverses the outcome of the competition between rigidity and energy dissipation. Our results highlight the dominating contribution of substrate viscosity to the biology of the cell.