Early evolution of conserved regulatory sequences associated with development in vertebrates

Abstract

Comparisons between diverse vertebrate genomes have uncovered thousands of highly conserved non-coding sequences, an increasing number of which have been shown to function as enhancers during early development. Despite their extreme conservation over 500 million years from humans to cartilaginous fish, these elements appear to be largely absent in invertebrates, and, to date, there has been little understanding of their mode of action or the evolutionary processes that have modelled them. We have now exploited emerging genomic sequence data for the sea lamprey, Petromyzon marinus, to explore the depth of conservation of this type of element in the earliest diverging extant vertebrate lineage, the jawless fish (agnathans). We searched for conserved non-coding elements (CNEs) at 13 human gene loci and identified lamprey elements associated with all but two of these gene regions. Although markedly shorter and less well conserved than within jawed vertebrates, identified lamprey CNEs are able to drive specific patterns of expression in zebrafish embryos, which are almost identical to those driven by the equivalent human elements. These CNEs are therefore a unique and defining characteristic of all vertebrates. Furthermore, alignment of lamprey and other vertebrate CNEs should permit the identification of persistent sequence signatures that are responsible for common patterns of expression and contribute to the elucidation of the regulatory language in CNEs. Identifying the core regulatory code for development, common to all vertebrates, provides a foundation upon which regulatory networks can be constructed and might also illuminate how large conserved regulatory sequence blocks evolve and become fixed in genomic DNA.

Figure 1. Conservation of non-coding sequences across the Meis2/c15orf41 locus in vertebrates.

(A) Plot of non-coding sequence conservation between mammals and fish across a 3.5 Mb region of human chromosome 15q14, encompassing the Meis2 and c15orf41 genes. Each vertical bar within the blue panel represents a CNE. (B) MLAGAN alignment of the c15orf41 locus enclosed by rectangle in (A). Human (Hu), mouse (Mu) and lamprey (La) genomic regions are aligned with the orthologous reigon in the Fugu genome. Exons are annotated and represented by mauve peaks (black arrows) and are detectable in all species. Pink peaks represent non-coding conservation. A number of these are conserved in lamprey (blue arrowheads) but a number are also absent (grey arrowheads).

Figure 2. Schematic representations of GFP expression patterns driven by core CNEs.

GFP-positive cells are marked onto camera lucida drawings of a zebrafish embryo on day 2 (24–30hpf) and day 3 (48–54hpf) of embryonic development. At each stage, the results are collated from all embryos with expression and overlaid to give a composite depiction of the GFP expression pattern. The number of GFP-positive embryos are noted under each schematic (n = ). The charts show the percentage (y axis) of GFP positive embryos with expression in each domain 1–14 (see below). In both the charts and the schematics, broad domain categories are colour coded. 1, forebrain; 2, midbrain; 3, hindbrain; 4, spinal cord; 5, other neurons (CNS, pink); 6, eye; 7, ear (sensory organs, purple); 8, notochord (dark green); 9, muscle (light green); 10, blood (red); 11, heart/pericardium (orange); 12, epidermal; 13, fin (both light blue); 14, other (dark blue).

Figure 3. Up-regulation of GFP by lamprey and human derived CNEs.

Images of live zebrafish embryos 48–54 hpf (A–C) and 24–30 hpf (D), lateral views, anterior to left. Images are shown as GFP fluorescent (A,B,D) and (C) merged fluorescent and bright field views. (A,C) GFP expression in the hindbrain, driven by an EBF3 CNE derived from lamprey and human respectively. (B,D) GFP expression in the spinal cord driven by a PAX2 CNE derived from lamprey and human respectively. e, eye; hb, hindbrain; mb, midbrain; nc, notochord; ov, otic vesicle; sc, spinal cord; y, yolk. Scale bars represent 100 µm.

Figure 4. Proposed model of early vertebrate evolution, genome duplications, and expansion of CNE repertoire.

Evolution of CNE repertoire occurred early in the history of vertebrates, coinciding with proposed genome duplication events and the emergence of agnathans. CNEs are barely detectable in invertebrate genomes and therefore must have evolved very early in the ancestral vertebrate, coinciding with whole genome duplication (WGD) events. The lamprey genome possesses a much smaller set of CNEs than sharks and other jawed vertebrates. This suggests that a large number of CNEs evolved and became fixed in all gnathostomes within a relatively short time period, between the emergence of agnathans and Chondrichthyes, coincident with the second proposed large scale genome duplication in jawed vertebrates. The proportion of CNEs that are found to be duplicated in the human genome is shown for each species (proportion of dCNEs); over half of the human CNEs that have matches to lamprey are found to be duplicated in the human genome. Timescale and divergence times taken from . The elephant shark genome has a maximum coverage of 75% and the light blue area represents the unsequenced portion.