Genome and transcriptome mechanisms driving cephalopod evolution

Abstract

Cephalopods are known for their large nervous systems, complex behaviors and morphological innovations. To investigate the genomic underpinnings of these features, we assembled the chromosomes of the Boston market squid, Doryteuthis (Loligo) pealeii, and the California two-spot octopus, Octopus bimaculoides, and compared them with those of the Hawaiian bobtail squid, Euprymna scolopes. The genomes of the soft-bodied (coleoid) cephalopods are highly rearranged relative to other extant molluscs, indicating an intense, early burst of genome restructuring. The coleoid genomes feature multi-megabase, tandem arrays of genes associated with brain development and cephalopod-specific innovations. We find that a known coleoid hallmark, extensive A-to-I mRNA editing, displays two fundamentally distinct patterns: one exclusive to the nervous system and concentrated in genic sequences, the other widespread and directed toward repetitive elements. We conclude that coleoid novelty is mediated in part by substantial genome reorganization, gene family expansion, and tissue-dependent mRNA editing.

Fig. 1. Doryteuthis pealeii anatomy and phylogeny.

a Adult D. pealeii (image: Roger Hanlon). b Phylogeny of coleoid cephalopods derived from a single complete mitochondrial genome per species, with Nautilus as outgroup (not shown). Date ranges at nodes indicate minimum and maximum node ages in millions of years as estimated by a strict molecular clock. c Tissues collected from D. pealeii for RNA sequencing, classified as “Neural” (blue), “Non-Neural” (orange), and “Mixed” (purple) tissues. “Mixed” tissues correspond to axial nerve cord (ANC) and Retina (Ret) for containing heterogeneous cell types derived from neural and non-neural tissues. Blood (Blo—not pictured) and posterior salivary gland (PSG) were obtained from a non-reference D. pealeii individual.

Fig. 2. Conserved synteny across coleoid cephalopods.

Dotplots of orthologous gene content. a The scallop M. yessoensis and the African snail A. fulica. The chromosomes of these two molluscs are conserved both in regard to each other and to their linkage group identities. b M. yessoensis and N. pompilius, a non-coleoid cephalopod, show conservation of macrosynteny between early branching cephalopods and other molluscs. c M. yessoensis and D. pealeii illustrate derived rearrangements in squid genomes. d Comparisons of D. pealeii and N. pompilius suggest chromosomal rearrangements occurred after the split between nautiloids and coleoids. e D. pealeii and O. bimaculoides. Squid and octopus chromosomes show higher levels of conservation. f D. pealeii and E. scolopes. The chromosomes show near 1:1 correspondence between the two squid species. Axes are labeled with chromosome or contig IDs and gene indices.

Fig. 3. Bilaterian linkage group (BLG) orthologs on amphioxus, scallop, snail, and cephalopod chromosomes.

a Top: Amphioxus (B. floridae) chromosomes correspond 1:1 to BLGs with some exceptions: Bfl1—chordate fusion of BLGA1 and A2; Bfl2—recent amphioxus fusion of BLGC1 and BLGJ1; Bfl3—recent amphioxus fusion of BLGC2 and BLGQ; Bfl4—recent amphioxus fusion of BLGO1 and BLGI. Middle: M. yessoensis chromosomes show some mixing of BLGs, but most chromosomes primarily correspond to one BLG. Bottom: A. fulica chromosomes follow similar patterns as M. yessoensis chromosomes, except that A. fulica underwent a whole genome duplication resulting in several duplicate chromosomes. b BLG orthologues on cephalopod chromosomes show extensive mixing of multiple BLGs throughout. Top: D. pealeii chromosomes. Middle: E. scolopes chromosomes. Bottom: O. bimaculoides chromosomes. c BLG color assignments.

Fig. 4. Disruption of colinearity in squid, but not bird, genomes.

Mutual best hit dotplots between squids (a) and birds (b). c Megablast alignment in 10 kb windows of E. scolopes and D. pealeii (left) and O. bimaculoides and D. pealeii (right). The squid show some retention of colinearity while colinearity is lost between octopus and squid genomes. d Quantification of microsyntenic cluster sizes. Run length corresponds to the number of genes in a detected microsyntenic linkage (maximum number of intervening genes = 5), and cumulative genes (y-axis) corresponds to the total sum of genes in the run of a certain size or larger. This allows us to define an “N50” measure: 50% of squid genes are in microsyntenic runs of 4 or fewer genes and 50% of bird genes are in microsyntenic runs of 23 or fewer genes.

Fig. 5. Expansion of gene families.

a Protocadherin gene clusters in cephalopod genomes. The protocadherin- and C2H2-rich chromosomes for D. pealeii, E. scolopes, and O. bimaculoides are shown to scale. 285/288 D. pealeii PCDHs are located within a 50 Mb cluster on chromosome 15 (box). Of these, 163 are found in 5 tight subclusters of 40 (D^a, yellow), 37 (D^b, grass green), 20 (D^c, green), 36 (D^d, teal), and 30 (D^e, blue) genes. All but the D^e are facing in the same transcriptional direction. E. scolopes also demonstrates multiple clusters of PCDHs spanning chromosome 15. The orthologous chromosome in O. bimaculoides contains 149 of 168 PCDHs found in the genome, with two notable subclusters of 34 and 27 genes. Major clusters of C2H2 genes are noted in gray. b Phylogeny of coleoid (octopus: blue, decapodiform in black), snail (Lottia gigantea, teal), bivalve (Crassostrea gigas, sky blue), annelid (Capitella teleta, green) and human (red) PCDHs demonstrates lineage-specific expansions. A handful of very long branches in the decapodiform protocadherins correspond to truncated sequences that may represent pseudogenes. Notably, genomic clusters (indicated above the phylogeny by different color bars) also cluster on the tree. c Arrangement of S-crystallins/Glutathione S-transferases (GSTs) in cephalopod genomes. Purple bars indicate the location of GST genes, gray gradient indicates gene density. D. pealeii has 139 GSTs in a single cluster spanning 60 Mb on chromosome 39. The orthologous chromosome in E. scolopes contains 77 GSTs distributed in multiple clusters, while in octopus, 26 GSTs are found spanning chromosome 5.

Fig. 6. RNA editing profiles in D. pealeii.

a Edit frequencies of target sites (y-axis) per tissue sample (x-axis) from constitutively expressed edit sites. b The correlation matrix illustrates how squid tissues cluster by their edit frequencies. Clustering of tissues shows distinct groups, neural tissues to the left (blue), non-neural tissues in the center (yellow), and mixed tissues on the right (heterogeneous - “H”, dark purple): retina (Ret) and axial nerve cord (ANC). 13,578 constitutively expressed sites that have more than 3 reads in each of the samples with at least 5% or more edit frequency in at least one sample (a and b). c Frequency distribution of recoding edit sites discriminated by neural (blue) and non-neural (orange) samples. The majority (54%) of the recoding edit sites in neural samples have an edit frequency below 1%; in contrast, most of the recoding sites (94%) in non-neural samples are below 1%. d Scatterplot of the weighted average edit frequencies of neural samples (WN) against the weighted average edit frequencies of non-neural samples (WNN) classified by edit type: recoding (Rec), synonymous (Syn), intronic (Intron), splice junction (SJ), or in the 5′ or 3′ UTR. The weighted averages were used to classify edit sites as: Neural with differential editing between neural and non-neural samples where the ratio between WN and WNN is above 2.75 (blue); Ubiquitous Low with edit frequencies below 5% (light gray); Ubiquitous High with editing frequency rates >60% for neural and >40% for non-neural tissues (red); and Ubiquitous Medium with edit frequencies between 5–40% in WN and 5–60% in WNN (gray). e 197,549 sites with at least 10 reads of depth in neural and non-neural samples classified by genic locations and overlap with repetitive sequence (as indicated by the + and −). Coding edits are found predominantly in Neural and Ubiquitous Low edit types while repetitive sequences are frequently edited in the 3′ UTR, regardless of edit type. f–i Analysis of edit frequencies of neural-type edits that are robustly edited (>25% edit frequency in at least one sample). f The edit frequency per tissue highlights the GFL as the tissue with the highest distribution if edit frequency. Side-by-side comparison of weighted edit frequencies  of g recoding and synonymous sites, and h sites overlapping conserved protein domains. i Same as (h), showing the WN values segregated on the x-axis by the amino acid substitution score. Heatmap of RNA editing profiles for the constitutively expressed Dynein Cytoplasmic 1 Heavy Chain 1(DYNC1H1) gene (j), and the ATPase Na+/K+ Transporting Subunit Beta 1 (ATP1B1) gene (k), which is expressed in all neural tissues.

Fig. 7. Expression and editing of ADAR transcripts.

a Phylogenetic tree of ADAR homologs. The colors highlight ADAR1, ADAR2, and ADAR-like families. b Cartoon representation indicating conserved domains present in squid ADAR proteins: double-stranded RNA binding Domains (dsmr), Z-binding domain binds (Z-a), Adenosine deaminase (A-deaminase). c Expression of genes with PFAM domains that interact with RNA that are enriched in neural samples. Tissue abbreviations as in Fig. 1 and color code in top row as in Fig. 6. d mRNA editing profile of ADAR genes.

Fig. 8. Cephalopod-specific gene families in the D. pealeii genome.

a Reflectins in the D. pealeii genome. Top: three clusters of reflectins were identified on two chromosomes, with a single reflectin found on chromosome 26 (not shown). Bottom: Reflectin expression profiles across D. pealeii transcriptomes indicate these genes are deployed in iridescent tissues, including the iridophore layer of the skin, the tissue surrounding the eye (retina), and the ink sac. Cells are colored according to standard deviation from mean expression levels. b Suckerin genes in D. pealeii. Top: the D. pealeii genome contains 13 suckerin genes distributed in two clusters on chromosome 2. Bottom: heatmap of the expression profiles of the suckerins across D. pealeii transcriptomes demonstrate that the suckerins are most highly expressed in the club of the tentacle. c Histidine-rich beak proteins in D. pealeii. Top: Cluster of 10 histidine-rich beak proteins on chromosome 12. Bottom: Heatmap of expression profiles of histidine-rich beak proteins in D. pealeii transcriptomes demonstrate high expression in the buccal mass. Abbreviations as in Fig. 1 except: BucL buccal lobe, BraL brachial lobe, BucM buccal mass, FDCI dorsal fin skin.