Initially, total RNA was prepared from 80 mg of the mantle muscle of each species using 0.8 ml ISOGEN II solution (Nippon Gene, Tokyo, Japan) according to the manufacturer’s instructions. cDNA was synthesized using total RNA and oligo dT primers with a reverse transcriptase (Takara Bio, Shiga, Japan) according to the manufacturer’s instructions. cDNA synthesis parameters were as follows: 42˚C for 60 min, followed 70˚C for 15 min. Primers were designed from the conserved sequences of the paramyosins of oyster, mussels, abalone, and octopus previously reported. PCR, 3' RACE, and 5' RACE methods were performed to obtain a DNA fragment, and this fragment was then sequenced. PCR for first-strand cDNA synthesis was performed using a Go Taq Master Mix (Promega Corporation, Tokyo, Japan). PCR parameters for the first PCR were as follows: Initial denaturation at 95˚C for 30 s, followed by 35 cycles of 95˚C for 30 s, 55˚C for 30 s, and 72˚C for 1 min. Nested PCR was performed using the same PCR parameters, except that samples of the first PCR products were 10-fold diluted. PCR parameters for the 3' RACE analyses were as follows: Initial denaturation at 95˚C for 30 s, followed by 30 cycles of 95˚C for 30 s, 55˚C for 30 s, and 72˚C for 1 min. The 5' RACE analyses were performed using kits provided by Invitrogen (Carlsbad, CA) according to the manufacturer’s instructions. PCR for full-length cDNA cloning was performed using platinum® Pfx Taq DNA polymerase (Invitrogen). The full-length PCR parameters were as follows: 95˚C for 2 min, followed by 30 cycles of 95˚C for 30 s, 51.5˚C for 30 s, and 72˚C for 3 min. DNA analysis was performed using BigDye Terminator v3.1 (Thermo Fisher Scientific, Kanagawa, Japan).

Sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed by using 12.5% commercial polyacrylamide gels (e-PAGEL, ATTO Corporation, Tokyo, Japan) according to the method of [14] . Mantle muscle (0.2 g) and protein dissolution solution (3.75 µl) of 20 mM Tris-HCl (pH 8.0)-2% SDS-8 M urea-2% β-mercaptoethanol were mixed and boiled for 2 min. And then muscle was dissolved by stirring overnight. Prestained molecular weight markers (Bio-rad, Hercules, CA) were used as standard. The gel was stained with Coomassie Brilliant Blue (CBB) R-250 after electrophoresis for 1.5 hours and then destained overnight with 30% methanol containing 10% acetic acid.

The polyacrylamide gel applied to 1/50 (v/v) samples to CBB stained one was stained with AE-1360 Ez Stain Silver (ATTO, Tokyo, Japan). A gel slice was cut into small pieces and destained with destaining solution (15 mM K₃ [Fe(CN)₆], 50 mM Na₂S₂O₃). Destained gel pieces were trypsinized as described in the manual for the In-Gel Tryptic Digestion Kit (Thermo Scientific, Waltham, MA). The peptide mixtures obtained were subjected to a nanoscale liquid chromatography-electrospray (Thermo Scientific) equipped with a captive spray ionization source (Michrom Bioresources, Auburn, CA) and an Advance UHPLC System (Michrom Bioresources).

The deduced amino acid sequences were aligned using the ClustalW program (EMBL-EBI: The European Bioinformatics Institute, European Molecular Biology Laboratory). Subsequently, a phylogenetic tree was constructed based on the neighbor-joining method using software Mega6 software [15] .

Initially, cDNA was synthesized from frozen muscle RNA. Using this cDNA as a template, PCR was performed with the primers (pmF-1,2 and pmR-1,2) designed from paramyosin sequences of oyster, octopus, abalone, and mussels, as previously reported, to obtain the internal sequence of squid paramyosins (Figure 1, Table 1). The obtained sequences were then used to design the internal primer for amplification of the 5' end and the 3' RACE primer to obtain the partial sequence. For S. esculenta (Figure 1(c)), 5' RACE was attempted, in which the partial sequence containing the start codon was successfully amplified. The obtained N-terminal untranslated region of S. esculenta paramyosin was then used to design the forward primer (pm5'-F1) and the reverse primers (pm5'-R1,2) for full-length amplification. Using the designed primers for full-length amplification, PCR was performed for the amplification of the whole length of the S. esculenta paramyosin sequence. The primers for full-length amplification were subsequently used for amplification of the whole paramyosin sequence of O. bartramii, D. gigas, and G. onyx.

The full-length amplified fragments were inserted into the T-easy vector

(Promega Corporation) for cloning. The entire base sequence of paramyosin was determined through sequence analysis for these clones. The O. bartramii paramyosin gene was cloned as a 2605-bp fragment containing a 2574-bp ORF, encoding a polypeptide of 858 amino acids. The D. gigas paramyosin gene was cloned as a 2691-bp fragment containing a 2640-bp ORF, encoding a polypeptide of 880 amino acids. The G. onyx paramyosin gene was cloned as a 2609-bp fragment containing a 2574-bp ORF, encoding a polypeptide of 858 amino acids. The S. esculenta paramyosin gene was cloned as a 2631-bp fragment containing a 2574 ORF, encoding a polypeptide of 858 amino acids. The entire base sequence and the deduced amino acid sequence obtained for O. bartramii, D. gigas,

S. esculenta, and G. onyx paramyosins were submitted to the DNA Data Bank of Japan (DDBJ) (Accession numbers: LC272578, LC272579, LC272580, and LC272083).

In general, paramyosin forms a coiled coil structure that contains a heptad repeat consisting of hydrophobic amino acids and a 28-residue repeat consisting of charged amino acids [2] [6] [16] [17] [18] . In the C-terminal region, it also contains an assembly competence domain (ACD) consisting of 29 residues [19] , through which the paramyosin molecules are believed to polymerize. Heptad repeats were identified throughout almost the entire sequences of the four squid paramyosins (Figures 2-5), in which hydrophobic amino acids were found in many of the a and d positions [20] [21] and charged amino acids were found in the e and g positions, displaying a structure characteristic of a coiled coil. Skip residues were also identified, resulting in the modification of heptad repeats into octad repeats. The skip residues further contribute to form 28-residue repeats containing repeated sequences [21] [22] . As shown in the figures, almost the entire region of the paramyosin sequences, excluding parts of the N- and C-terminal regions, was identified to form a coiled-coil structure. The sequences were also found to end in C-terminal tail pieces.

The percentage similarity among paramyosin sequences was found to be 96% between O. bartramii and D. gigas, 88% between O. bartramii and G. onyx, 90% between O. bartramii and S. esculenta, 86% between D. gigas and G. onyx, 87% between D. gigas and S. esculenta, and 89% between G. onyx and S. esculenta

(Table 2). No clear difference was observed between the three species in the order Teuthida (O. bartramii, D. gigas, and G. onyx) and the one species in the order Sepiida (S. esculenta). The N-terminal region was found to be the region showing the least homology among the sequences of the four squid paramyosins. The amino acids in the ACD were highly conserved, showing an exact match between O. bartramii and D. gigas, and between S. esculenta and G. onyx. Only one-residue differences were observed among the sequences that did not show

an exact match.

SDS-PAGE analysis showed that four squid mantle muscles was constituted three major myofibrillar proteins; myosin heavy chain around 250 kDa, paramyosin around 90 kDa and actin around 40 kDa (Figure 6). The bands of paramyosin were subjected to the analysis for amino acid sequences after destained

and trypsinized. As an example, deduced amino acid sequence of O. bartramii paramyosin and the covered regions of the peptide fragments was shown in Figure 7. The peptide fragments exhibited high coverage of 56.18% (484 residues). In addition, either end the tripsinized sites was K or R. Therefore, the tripsin treatment was identified as successful. The other three squids, D. gigas, S. esculenta and G. onyxalso showed high coverage of 46.36% (407 residues), 68.53% (587 residues) and 43.47% (372 residues), respectively (date not shown). From these results, it was suggested that paramyosin genes determined in this

study was actually expressed in each mantle muscles and existed as protein.

Phylogenetic analysis of paramyosin had been reported in fluke species and tapeworm species [23] [24] . An unrooted phylogenetic tree was constructed using a total of 26 paramyosin sequences, including one sequence from one octopus species, 13 sequences from seven shellfish species, five sequences from five fluke species, and three sequences from three tapeworm species, which are all known,

in addition to the sequences from the four squid species determined in this study, in order to examine the characteristics of the paramyosin structure (Figure 8, Table 3). The four squid paramyosins formed a group independent from the paramyosins of other species, to which the octopus paramyosin was most closely positioned. Fluke and tapeworm species also formed a group, whereas shellfish species remained dispersed. Shellfish species vary in shape such as univalves and bivalves, which may form independent groups. Nonetheless, these

shellfish species showed greater diversity compared to the other groups of species. Fluke [25] paramyosins showed high sequence similarities within the group, with percentages of 93% or greater. Tapeworm [26] [27] paramyosins also showed high sequence similarities with percentages of 98% or greater. Shellfish paramyosins showed lower sequence similarities compared to other groups of species, with percentages from 71% to 86%. The underlined species in Figure 6 have multiple isoforms of paramyosin, and the respective isoforms were found close to each other. For instance, oyster is known to have four paramyosin isoforms, which all have high similarities to each other with percentages of 96% or greater. Similarly, paramyosin isoforms from two species of fresh water snail, two species of scallops, and two species of Lingula, whose sequences are registered in the database, have high similarities among their respective isoforms, with percentages exceeding 96%. For these species that are known to have multiple isoforms, the sequences were suggested to encode paramyosin genes through genome analyses. In contrast, in species whose paramyosin sequences have been determined by means of cDNA cloning, such as those in the present study (including O. bartramii, D. gigas, S. esculenta, G. onyx, and mussel species), only a single gene has been identified. Although genome databases collect the sequences of all genes from entire genome sequences, cDNA cloning identifies only those genes that are actually expressed. In the present study, a single gene was identified in each species, suggesting that only one gene is expressed in the respective organisms. That is, although two or more paramyosin genes may be present on the genome, some are considered to be pseudogenes. Pseudogenes have base sequences similar to those of normal genes, although they are not expressed in vivo. For the species that are considered to have two or more paramyosin genes, the expression of these genes will be further investigated in vivo.