Primitive duplicate Hox clusters in the European eel's genome

Abstract

The enigmatic life cycle and elongated body of the European eel (Anguilla anguilla L., 1758) have long motivated scientific enquiry. Recently, eel research has gained in urgency, as the population has dwindled to the point of critical endangerment. We have assembled a draft genome in order to facilitate advances in all provinces of eel biology. Here, we use the genome to investigate the eel's complement of the Hox developmental transcription factors. We show that unlike any other teleost fish, the eel retains fully populated, duplicate Hox clusters, which originated at the teleost-specific genome duplication. Using mRNA-sequencing and in situ hybridizations, we demonstrate that all copies are expressed in early embryos. Theories of vertebrate evolution predict that the retention of functional, duplicate Hox genes can give rise to additional developmental complexity, which is not immediately apparent in the adult. However, the key morphological innovation elsewhere in the eel's life history coincides with the evolutionary origin of its Hox repertoire.

Figure 1. The life cycle of the European eel.

After hatching, presumably in the Sargasso Sea, cylindrical larvae develop into leaf-shaped leptocephalus larvae, which after drifting on the Gulf Stream for approximately one year metamorphose into glass eels close to the European coast. The glass eels may stay at the coast or migrate upriver, where they stay as juveniles (elvers and yellow eel) for many years (depending on the region: males 4–6 years, females 8–12 years). Finally, they develop into migrating silver eels; the cause and timing of silvering is not well understood. They mature during or after migration to the spawning grounds.

Figure 2. Genomic organization of the Hox gene clusters of the European eel.

Scaffolds are indicated by black lines and asterisks represent two gaps between scaffolds. Hox genes are indicated by colored arrows that are numbered according to their paralogous groups. Three pseudogenes are indicated by the symbol ψ. Neighboring genes are indicated by grey arrows. Conserved microRNA genes are indicated by triangles: miR-196 (closed triangles) and miR-10 (open triangles).

Figure 3. Classification of the European eel Hox clusters.

An unrooted phylogenetic tree showing the relationships between A. anguilla and fish Hox9 paralogues. Numbers indicate bootstrap support.

Figure 4. Phylogeny of Hox clusters of the European eel.

Unrooted phylogenetic trees based on alignments combining multiple Hox genes per cluster. A) Cluster A relationships, based on HoxA9, HoxA11 and HoxA13 genes. B) Cluster B relationships, based on HoxB1, HoxB5 and HoxB6 genes. C) Cluster C relationships, based on HoxC6, HoxC11, HoxC12 and HoxC13 genes. D) Cluster D relationships, based on HoxD4 and HoxD9 genes. Species included: A. anguilla, Salmo salar (Atlantic salmon), Danio rerio (zebrafish), Oryzias latipes (medaka), and Tetraodon nigroviridis (green spotted puffer). Asterisks indicate bootstrap support >90%.

Figure 5. Synteny around Hox clusters.

Conservation of flanking genes supports the classification of A. anguilla clusters into different orthologous subgroups. The eel clusters and up to seven flanking genes are compared with the genomic organization in zebrafish (Danio rerio) and medaka (Oryzias latipes). Coloured box outlines indicate preserved synteny between eel and the two other species, dotted outlines denote flanking genes found on extended eel scaffolds (see Methods). Interpretation should take into account residual synteny between a and b paralogous clusters. Limited data is available on HoxCb (lost in O. latipes, possibly misassembled in D. rerio) and HoxDb (lost in D. rerio) clusters.

Figure 6. Hox gene expression in A. australis embryos.

A) mRNA-seq-based gene expression in a 27 hpf embryo. B) Whole mount in situ hybridizations showing the expression of HoxB9a, HoxD12b and HoxC13a. HoxB9a expression can be detected in 26 hpf (dorsal view) and 48 hpf (lateral tail region view) embryos. HoxD12b and HoxC13a display expression in the tail region (lateral views) at 96 hpf, but not at 48 hpf. White arrowheads indicate anterior expression boundaries. Scale bars correspond to 100 µm.

Figure 7. Model for the evolution of teleost Hox gene organization.

Schematic Hox clusters , , are superimposed on a species phylogeny with estimates of divergence times , – which vary considerably between studies . Ancestral architectures are inferred on the basis of maximum parsimony, i.e. the number of cluster duplications and gene loss events is minimized. Salmo salar (Atlantic salmon) has presumably lost several duplicate clusters (not shown). Deduced gene loss in a lineage is illustrated by a cross, question marks denote uncertainty about cluster gene content in the pre-TSGD actinopterygian Polypterus senegalus (bichir). Arrows indicate the possible origins of the leptocephalus body plan.