Search

Few other invertebrates captivate our attention as cephalopods do. Octopods, cuttlefish, and squids amaze with their behavior and sophisticated body plans that belong to the most intriguing among mollusks. Little is, however, known about their body plan formation and the role of Hox genes. The latter homeobox genes pattern the anterior–posterior body axis and have only been studied in a single decapod species so far. Here, we study developmental Hox and ParaHox gene expression in Octopus vulgaris. Hox genes are expressed in a near-to-staggered fashion, among others in homologous organs of cephalopods such as the stellate ganglia, the arms, or funnel. As in other mollusks Hox1 is expressed in the nascent octopod shell rudiment. While ParaHox genes are expressed in an evolutionarily conserved fashion, Hox genes are also expressed in some body regions that are considered homologous among mollusks such as the cephalopod arms and funnel with the molluscan foot. We argue that cephalopod Hox genes are recruited to a lesser extent into the formation of non-related organ systems than previously thought and emphasize that despite all morphological innovations molecular data still reveal the ancestral molluscan heritage of cephalopods

RNA-seq libraries were constructed with a Lexogen SENSE mRNA-Seq Library Prep Kit V2 and sequenced with an Illumina Hi-Seq 2500 generating paired-end, stranded 125 bp libraries resulting in 53,523,481 (ovu1) and 63,328,653 (ovu2) paired end reads. The overall transcriptome assembly follows the procedure performed by De Oliveira et al. (2016) [42]. The short‐read libraries were preprocessed using Trimmomatic v. 0.36 (ref. [43]) to remove known specific Illumina adapters from the paired‐end libraries (Illumina universal adapter). Filtering by quality and length was performed with a SLIDINGWINDOW:4:15 MINLEN:36. First and last nucleotides from reads with low quality score were clipped and the library file was converted into FASTA format using fq2fa from SeqKit version 0.11.0 (ref. [44]). Quality of the initial and filtered library was assessed with the software FastQC v.0.11.8 (ref. [45]) considering quality score of the bases, GC‐content, and read length. 13.33% (ovu1) and 16,23% (ovu2) of reads were excluded during the preprocessing procedure resulting in a total of 46,386,109 (ovu1) and 53,052,713 (ovu2) reads. The assemblies and all downstream analyses were conducted with a high‐quality and clean library. The filtered transcriptome was assembled into contiguous cDNA sequences with IDBA_tran v1.1.3 software (ref. [46]) using the default settings (except: −mink20 −maxk 80 −step5). The resulting assembly was assessed using the tool QUAST (available at: http://quast.bioinf.spbau.ru [47]). The number of contigs was 16,723 sequences (ovu1) and 18,551 (ovu2)