Amylase 2017; 1: 26–34

Research Article Open Access

Jean-Luc Da Lage* An optional C-terminal domain is ancestral in α-amylases of bilaterian animals

DOI 10.1515/amylase-2017-0003 immunoglobulins and enzymes (for details, see, e.g. [1]). Received February 17, 2017; accepted March 28, 2017 The design and building of novel proteins often rely on Abstract: The modular structure and organization of exon shuffling, in which introns serve as linking regions most proteins is a fascinating aspect of their origin and between the pieces to be joined. This process, which had evolution. α-Amylases are known to be formed of at least been predicted by Walter Gilbert in his seminal article [2], three domains. In a number of bacterial α-amylases, is quite common in Eukaryotes, especially in Vertebrates, one or several additional domains may exist, which are but has also been exemplified, for example, as a recent carbohydrate binding modules, interacting with raw event in Drosophila [3]. The domains, joined together, substrates. In animal α-amylases, however, no additional may be tightly linked or attached to each other by a low domain has been described. Here we report the presence complexity protein region named a linker. However, of a C-terminal domain, previously described only in the introns are not necessary, since domain shuffling may bacterium Pseudoalteromonas haloplanktis. This domain also occur in Prokaryotes, which are devoid of introns (see is widely distributed in invertebrate α-amylases and below). must be ancestral, although it has been lost in important α-Amylase plays a crucial role to achieve the phyla or groups, such as vertebrates and insects. Its hydrolysis of starch and other polysaccharides from food function is still unknown. In a single genome, enzymes and nutrients into maltose and maltodextrines. Although with and without the terminal domain may coexist. In a amino acid sequences are highly variable among few instances, this domain has been recruited by other organisms, the general tertiary structure is well conserved. proteins in both and animals through domain It is made of a “(β/α)8 barrel” or “TIM barrel” (domain A), shuffling. a protruding few structured domain B involved in Ca2+ ion binding and catalysis at the interface with domain A, and Keywords: domain shuffling; evolution; intron; an all-β domain C adopting a Greek key conformation. horizontal gene transfer; carbohydrate binding module In about 10% of bacterial and fungal α-amylases and related enzymes, an additional terminal domain is Abbreviations: AHA, Pseudoalteromonas haloplanktis present, sometimes in N-terminal position, but most often α-amylase; CBM, carbohydrate binding module; SBD, in C-terminal position, with a function of raw or granular starch binding domain. starch-binding, hence its name of starch binding domain (SBD) [4-6]. SBDs belong mainly to the carbohydrate binding module (CBM) families 20, 21, 25, 26, 34, 41, 45, 48, 1 Introduction 53, 58, 68 and 69 ([7], http://www.cazy.org/). They consist of all-β domains forming an open, distorted β-barrel, with Protein evolution and innovation often goes about through binding sites for polysaccharides, which use conserved admixture of existing materials, such as pieces of existing tryptophan residues. proteins, often fully functional domains. Countless A different, unrelated additional C-terminal domain examples have been described, such as muscular proteins, has been described from a bacterium, Pseudoalteromonas haloplanktis, which is an antarctic, marine γ-proteobacterium (). Its unique *Corresponding author: Jean-Luc Da Lage, Evolution, Génomes, α-amylase (AHA), which is adapted to cold temperature, Comportement, Ecologie, CNRS, IRD, Univ Paris-Sud, Université has been extensively studied (e.g. [8-11]). Its amino acid Paris-Saclay, F-91198 Gif-sur-Yvette, France; sequence is strongly related to animal ones, and we have E-mail: [email protected] suggested that AHA and animal (bilaterian) α-amylases

© 2017 Jean-Luc Da Lage, published by De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License. C-terminal α-amylase domain in animals 27 could be related to each others through a lateral transfer, gigas (Mollusca, Bivalvia). Sequence data were deposited whose direction is still not ascertained, but more likely in GenBank (accession numbers are indicated in Table S1). from bacteria to bilaterians [12,13]. AHA possesses an original C-terminal domain, unrelated to SBDs, and not 2.2 Searches in databases yet ascribed to any family, which has been shown to act possibly as a secretion helper, cleaved from the core enzyme Using the putative C-terminal domains of C. fluminea or by a non-specific protease after exportation across the AHA as a query, we searched by BLASTP and TBLASTN [14] outer cell wall [10]. Within the bacterial kingdom, we had in sequence databases for other occurrences of domains previously found this “propeptide” domain only in a few similar to the AHA C-terminal domain. The GenBank nr, species of the Pseudoalteromonas genus, and also in the GenBank EST and GenBank Bacterial genomes databases bacterium Saccharophagus degradans (Cellvibrionaceae, (http://blast.ncbi.nlm.nih.gov/Blast.cgi) were searched. formerly ascribed to Alteromonadaceae). In the latter We investigated traces archives at the NCBI for non- case, the C-terminal domain was attached to a plant- vertebrate metazoans. More specific genome servers were like α-amylase, but not to the orthologous animal-like also used: the EchinoBase server (http://www.echinobase. α-amylase, which also exists in this species, showing an org/) for echinoderms; server of the Joint Genome Institute example of domain translocation and shuffling without (http://genome.jgi.doe.gov/) for Daphnia pulex, Capitella the need of an intron [12]. In animals, however, no teleta, Lottia gigantea, Branchiostoma floridae, Helobdella extra domain had been characterized until now. In fact, robusta and Ciona intestinalis; specific servers for the sequence similarity between the AHA C-terminal domain Aplysia californica EST project (http://aplysia.cu-genome. and an animal sequence had been only noticed [10] in org/) and the nematode Pristionchus pacificus (http:// Caenorhabditis elegans. www.pristionchus.org/); nematode-specific servers In the present study, we show that this domain not only (http://www.nematode.net/); the “neglected genomes” exists in animals, but is widespread, and thus is probably database (http://www.nematodes.org/bioinformatics/ ancestral in bilaterians. The long-standing presence of this databases.shtml); also for marine organisms (http://www. domain in various non-vertebrate animals suggests that it marinegenomics.org/); for the Urochordate Oikopleura has a function. Although this hypothetic function cannot dioica (http://www.genoscope.cns.fr/); for the tick be reached for now, we draw an evolutionary history of Ixodes scapularis (http://www.vectorbase.org/); and this uncharacterized, yet quite common, protein domain. the InsectBase for various insects (http://www.insect- genome.com/). The accession numbers of sequences are given in Table S1. 2 Materials and methods Amino acid sequences were aligned with the program MUSCLE [15], and the alignment was manually curated for 2.1 Experimental data uncertainties and then served for building a maximum likelihood tree using the server Phylogeny.fr (http://www. We checked the presence of a C-terminal domain phylogeny.fr/), with default parameters [16]. experimentally in several molluscan species. The α-amylase genes were entirely sequenced in the bivalves Corbicula fluminea and Mytilus edulis, and almost 3 Results entirely in the limpet Patella vulgata, using the Genome walker Universal kit (Clontech). The C-terminal domains 3.1 Occurrence of the AHA-like C-terminal were identified by BLAST search [14] in the GenBank domain in Bacteria database. From the alignment of these domains with those of P. haloplanktis, S. degradans and C. elegans, we The C-terminal domain, around 185 amino acids designed PCR primers in conserved parts of the domain in length, was found only in a limited set of (Ctermdir: CARGAYCTNTTYATHCGNGG; Ctermrev: γ-, mostly , for instance TCNGCNCCRTACCARTC) and used various combinations Pseudoalteromonas atlantica (YP_662421) and a number for amplification of fragments, if possible showing of closely related species, e.g. P. tunicata (ZP_01134110) attachment to the core α-amylase sequence, i.e. also and Alteromonadales bacterium TW7 (ZP_01613200), using primers designed within the core enzyme. The all attached to an α-amylase orthologous of that of P. species assayed by PCR were the chiton Acantochitona sp. haloplanktis (subfamily GH13_32 in the classification of (Mollusca, Polyplacophora) and the oyster Crassostrea Stam et al. [17]). It was also found in two Cellvibrionaceae 28 Jean-Luc Da Lage

– first in Saccharophagus degradans; it had been found animal α-amylases is quite complex, because numerous previously linked to a plant-type (subfamily GH13_6) [12]. independent gene duplications and sequence divergences We found a gene orthologous to this plant-type gene, occurred, so that a tree drawn from these sequences does also with the C-terminal domain, in Cellvibrio japonicus not faithfully reflect animal phylogeny (for instance, (ACE84223). Moreover, in this species, we also found a nematode sequences are highly divergent, whereas duplicate of the C-terminal domain in C-terminal position these organisms should be clustered with arthropods). of the protein aqualysin (ACE84702). The Alteromonadales However, when reported on such an α-amylase gene tree Thalassomonas actiniarum (WP_044836227) and another (Fig. S1), the presence of the additional domain seems to γ-proteobacterium (Chromatiales), Rheinheimera be distributed along two main branches (which may reflect nanhaiensis, were found to have a C-terminal domain, a very old duplication), but with much more occurrences attached to a plant-type amylase (WP_008218869). We in one of the branches. Some molluscs and crustacea found no other occurrence among thousands of bacterial have sequences in each main branch, some with the genomes available at the NCBI, although the GH13_32 C-terminal domain, some being domainless. The domain α-amylase type (i.e. the most related to the animal types) thus seems to have been lost several times independently, occurs in a few dozens of various bacterial species, not and completely at least in Vertebrates + Tunicates and in only γ-proteobacteria [13]. insects. In the latter cases, data are yet too scarce to clarify its presence or absence in early diverging Hexapoda, such 3.2 Occurrence of the AHA-like C-terminal as Collembola, which would help to date the loss event domain in metazoans in insects. Genomes and individual sequences available to date in databases, such as GenBank, Flybase (http:// We searched the domain in genome databases, but also flybase.org/) or InsectBase (http://www.insect-genome. experimentally for several non-vertebrate species. The com/) do not exhibit the C-terminal domain. For the results are summarized on a species tree (Fig. 1). As a same reason, we cannot conclude firmly to a complete whole, we found sequences similar to the AHA C-terminal loss in the clade of Myriapoda + Chelicerates, although domain in various Protostomes and Deuterostomes. It complete amylase genes lacking the C-terminal domain was found in non-hexapod pancrustacea, in nematodes, have been sequenced. The former characterization of molluscs, annelids, echinoderms and cephalochordates, only C-terminal domainless α-amylases in some molluscs but not in insects, myriapods, arachnids, urochordates or crustacea, generally from mRNAs [18-20], suggested and vertebrates. Like in bacteria, the length of the that these C-terminal domainless genes were the most domain was around 185-190 residues. In some cases, active ones. However, EST libraries showed that the when the domain was detected, there was no evidence for α-amylases linked to the AHA-like C-terminal domain its attachment to the catalytic part of α-amylase. It was may be transcribed as well. the case with some EST databases, where no sequence was evidenced to show a continuity between the two 3.3 The Daphnia pulex case parts of the protein. Interestingly, as we will see below, the C-terminal domain may exist independently from Independent losses of the C-terminal domain may indeed an amylase. In several cases, two types of α-amylase occur, as illustrated by the situation in the genome of genes were found, with or without a C-terminal domain Daphnia pulex (Crustacea, Branchiopoda). In this species, (Fig. 1). Note that no α-amylase gene at all was found there are at least four α-amylase gene copies, two of which, in sequenced genomes of Platyhelminthes (Schmidtea Amy1 and Amy1’ are tandem duplicates. The duplication mediterranea, Schistosoma mansoni), of the leech has involved about 500 bp upstream of the start codon Helobdella robusta, of the aphid Acyrthosiphon pisum and all the usual coding sequence of an α-amylase. and of the louse Pediculus humanus. Our results clearly Sequence identity is almost complete in the second half demonstrate that the AHA-like C-terminal domain was part of the gene, probably due to gene conversion (Fig. S2). already present in the ancestor of all extant Bilateria. And yet, Amy1 has a C-terminal domain, whereas Amy1’ Indeed, one intron position is conserved across has not. It is likely that the domain was lost in Amy1’ Protostomes and Deuterostomes (Fig. 2). It is not quite during the duplication event, as the duplicated segment clear whether, at the time of the last bilaterian ancestor, stopped inside the linker. It is unlikely that the domain two types of genes already existed, or if the domain was gained in Amy1 (e.g. from Amy2), since the tree of was specifically lost independently a number of times C-terminal domains (Fig. 3) shows no closer relationship following gene duplications. The evolutionary history of of the C-terminal domain between Amy1 and Amy2. C-terminal α-amylase domain in animals 29

Figure 1. Animal species tree (consensus from the literature [27-33]) showing the occurrences of the optional C-terminal domain. Legend: “+”, domain present; “-”, domain absent; “+ & -”, occurrence of both gene copies with and without the domain; (1), no evidence of attach- ment to α-amylase; (2), very long subterminal intron inside domain C.

3.4 Origin of the linker region dynamics elsewhere [21]. Although we have shown that the C-terminal domain is ancestral, it might have moved or The C-terminal domain has always been found to be been duplicated from one copy to another, as it happened linked to the enzyme by a low-complexity linker about in the bacterium S. degradans. If some C-terminal domains 20-50 residues in length (shorter in Nematodes), often rich have been recruited by previously domainless α-amylase in proline, threonine, serine and glycine. We defined the genes through exon shuffling, what would be the origin of linker region by the amino acid sequence lying between the amino acid linkers? In some species, the gene region the end of the usual, conserved α-amylase sequence, encoding the linker contained an intron (amphioxus AmyA that is, the end of domain C succeeding the TIM barrel, and AmyB, Daphnia Amy2). In the Annelide Capitella and the first conserved stretch RTVIF in the C-terminal teleta and all the molluscs, the linker was preceded by an domain. In the species studied, the linkers were highly intron, subterminal relative to the usual enzyme sequence variable in sequence, compared to both the core enzyme [22]. However, there is no indication in our sequence sequence and the C-terminal domain. Functionally, the analyses for partial or complete exonisation of introns, biased amino acid composition of the linkers endows nor internal duplications, which could have mediated the them with properties typical of linkers joining catalytic building of independent linker regions. One may notice domains and CBMs in various glycoside hydrolases. We in amphioxus AmyA a fourfold repeat of His-Pro-Thr. studied specifically the linkers and their mechanistic This motif, and other ones, such as poly-Thr (Capitella), 30 Jean-Luc Da Lage

Figure 2. Alignment of C-terminal domains of bacteria and animals, truncated to the most commonly found end. Colours show similarities as follows: red, 100% similarity; orange, 80-100% similarity; yellow, 60-80% similarity; white, below 60% similarity. Intron positions, when known, are indicated by vertical bars: pink, phase 0; blue, phase 1; green, phase 2. C-terminal α-amylase domain in animals 31 poly-Gly (sea urchin) may stem from three-nucleotide scallop Pecten maximus, the oyster Crassostrea gigas and microsatellites. At least this kind of mechanism may have the abalone Haliotis discus – genes devoid of C-terminal shaped very dissimilar linker regions, whatever their domain have been characterized in these species (P91778, common or independent origins. On the other hand, the AF321515, EF103352) – but careful examination shows linker region may enhance the probability of loss of the that they expand a bit downstream of the usual end of the C-terminal domain, as shown in Daphnia genes, where enzyme (end of the domain C), suggesting the presence of the deletion breakpoint seems to have occurred within a relictual part of a linker, and thus a loss of the C-terminal the linker, since a part of the linker remains at the end of domain. We have done a similar observation in sea urchin Amy1’ (Fig. S2). The same remark could also apply to the α-amylases (XP_011683732, XP_782889 and XP_794418).

Figure 3. Tree drawn from the alignment of C-terminal domains of bacteria and animals. Colour legend: orange, bacteria; pink, crustacea; blue, molluscs; green, nematodes; red, deuterostomes. Figures along branches are the posterior probabilities of the maximum likelihood computation. 32 Jean-Luc Da Lage

3.5 Analysis of the alignment 3.6 Occurrences of the C-terminal domain, not linked to an α-amylase The alignment of C-terminal domains of bacteria and animals (Fig. 2) shows that the sequences are In the bacterium Cellvibrio japonicus, we found an AHA- characterized by the presence of four highly conserved like C-terminal domain linked to a non-amylase enzyme, cysteines in bacteria, with two more cysteines in animals. the aqualysin, an extracellular serine protease. This At positions 107-114 of the alignment, a deletion occurred enzyme has been described in few bacteria, and possesses in Protostomes. In silico predictions (http://clavius. both N-terminal and C-terminal propeptides involved bc.edu/~clotelab/DiANNA/) suggest that two or three in extracellular translocation [24]. However, those disulphide bonds, respectively, may be formed, although domains have no relationship with the additional AHA- with moderate confidence. As already noticed for AHA like C-terminal domain found in C. japonicus. Although [10], the predicted secondary structure is dominated by this case of domain shuffling most likely stems from a β-strands (http://www.compbio.dundee.ac.uk/jpred/). De duplication of the α-amylase C-terminal domain, the tree novo structure predictions are not satisfactory. The CABS- (Fig. 3) shows a rather fast divergence with its parent FOLD server (http://biocomp.chem.uw.edu.pl/CABSfold/) domain. proposed several structures, the preferred of which showed In animals, we found cases of duplicates of the a very flat sheet shape, which is not consistent with the C-terminal domain independent from α-amylase in predicted secondary structures (not shown). The QUARK Daphnia pulex only. Three occurrences were detected in server (http://zhanglab.ccmb.med.umich.edu/QUARK/; the genome (http://wfleabase.org/): (i) on the scaffold 14 [23]) showed a structure, which is more consistent with (JGI_V11_100284, nuc. 1142036-1144649); (ii) scaffold 51 the secondary structure predictions, but it did not predict (JGI_V11_323078, nuc. 789341-790623); and (iii) scaffold disulphide bonds (Fig. 4). In addition, iTASSER (http:// 183, (JGI_V11_229423, nuc. 111544-112544). The latter of zhanglab.ccmb.med.umich.edu/I-TASSER/) predicted, those duplicates was physically close (ca. 7 kb apart) to with low confidence, the presence of an acarbose binding the α-amylase gene Amy2, which contains a C-terminal site for both C. fluminea and AHA, and also a maltose domain. However, the sequences of the three duplicates binding site for AHA (not shown). were clustered together rather than with a putative α-amylase donor (Fig. 3), which suggests a common origin of these three domains, their ancestor having already diverged significantly from the Amy counterparts. However, the three sequences nonetheless shared the same intron positions as the sequences of C-terminal domains linked to α-amylases (Fig. 2). All of these duplicates were predicted to be linked to putative polypeptides with a linker-like sequence, but the reality of such neogenes was not ascertained. ESTs suggested transcription for the putative genes on scaffold 183 (WFes0162386) and 51 (Wfes0148654), but there were slight differences between the ESTs and the genomic sequences, and, more importantly, a lack of the protein parts predicted to be upstream of the linkers. In addition, the putative genes linked to an AHA-like C-terminal domain in scaffolds 14 and 51 shared similarities in their N-terminal part. This part, in turn, seems to be present in several copies in the Daphnia genome (eight hits in a BLASTP search).

4 Discussion

We have shown here that an extra protein domain, Figure 4. De novo three-dimensional model of the C-terminal domain linked in C-terminal position to the domain C of animal of Corbicula fluminea obtained with the QUARK server [23] and α-amylases, is widespread in bilaterian animals. Indeed, drawn with the Swiss-PdbViewer [34]. C-terminal α-amylase domain in animals 33 it has long remained unnoticed because the most Acknowledgments: I am grateful to Georges Feller for studied animal α-amylases were those of mammals fruitful advice and discussion. The work was funded by and insects, which both lack this domain. We call it an regular CNRS funding to my laboratory. optional domain also because, even in species where it is found, it may be present or absent if several gene copies Conflict of interest: The author declares no conflict of occur, which is quite common. The domain is ancestral interest. in bilaterian animals and may have accompanied the transfer of a bacterial amylase to a bilaterian ancestor [13]. Interestingly, only a small number of bacteria possess this References domain and are eligible for a donor status. It is possible [1] Patthy L., Protein Evolution, Blackwell Science, Oxford, 1999. that the donor was a symbiont, since horizontal gene [2] Gilbert W., Why genes-in-pieces? Nature, 1978, 271, 501. transfer from symbiotic bacteria to their hosts has been [3] Long M.Y., Langley C.H., Natural selection and the origin of often reported, endowing the receptor organism with new jingwei, a chimeric processed functional gene in Drosophila, abilities, and rendering it independent from its symbiont Science, 1993, 260, 91–95. [4] Boraston A.B., Bolam D.N., Gilbert H.J., Davies G.J., [25]. Another noteworthy fact is that the C-terminal Carbohydrate-binding modules: fine-tuning polysaccharide domain, although not frequently, has been recruited by recognition, Biochem. J., 2004, 382, 769–781. other genes, in bacteria as well as in animals (in Daphnia, [5] Janecek S., Svensson B., MacGregor E.A., α-Amylase: an namely). The presence of a linker or an intron between enzyme specificity found in various families of glycoside the core sequence and the terminal domain increases hydrolases, Cell. Mol. Life Sci., 2014, 71, 1149–1170. the likelihood of exon or domain shuffling. But for what [6] Rodriguez-Sanoja R., Oviedo N., Sanchez S., Microbial starch- binding domain, Curr. Opin. Microbiol., 2005, 8, 260–267. function? No interaction of the AHA C-terminal domain [7] Majzlova K., Janecek S., Two structurally related starch-binding with insoluble polysaccharides was noticed (Georges domain families CBM25 and CBM26, Biologia, 2014, 69, Feller, Université de Liège, personal communication). 1087–1096. The sequence of the domain is quite different from that [8] Feller G., Lonhienne T., Deroanne C., Libioulle C., Van Beeumen of SBDs or other CBMs, except that it shows a structure J., Gerday C., Purification, characterization and nucleotide sequence of the thermolabile α-amylase from the antarctic that is rich in β-sheets. The β-sheet-rich structure was psychrotroph Alteromonas haloplanktis A23, J. Biol. Chem., confirmed in AHA using infra-red spectroscopy, but 1992, 267, 5217–5221. this experimental method also retrieved 29% helices, [9] Feller G., Payan F., Theys F., Quian M., Haser R., Gerday C., which is not proposed by in vitro modelling. In addition, Stability and structural analysis of α-amylase from the antarctic two disulphide bonds were found experimentally [26]. psychrophile Alteromonas haloplanctis A23, Eur. J. Biochem., In P. haloplanktis, where it was first described and 1994, 222, 441–447. [10] Feller G., D’Amico S., Benotmane A.M., Joly F., Van Beeumen best characterized, it has been proposed to serve as a J., Gerday C., Characterization of the C-terminal propeptide “secretion helper” for export across the periplasmic involved in bacterial wall spanning of α-amylase from the membrane of bacteria [10], but this hypothesis remains psychrophile Alteromonas haloplanktis, J. Biol. Chem., 1998, debatable because it also occurs in animals. In animals, 273, 12109–12115. it might be used as an anchor to cell membranes, but not [11] Cipolla A., Delbrassine F., Da Lage J.-L., Feller G., Temperature adaptations in psychrophilic, mesophilic and thermophilic through transmembrane anchoring, since it shows no chloride-dependent α-amylases, Biochimie, 2012, 94, transmembrane domains (http://www.ch.embnet.org/ 1943–1950. software/TMPRED_form.html). However, to date, we do [12] Da Lage J.-L., Feller G., Janecek S., Horizontal gene transfer not know whether, in animals, the C-terminal domain from Eukarya to Bacteria and domain shuffling: the α-amylase remains attached to the core enzyme or if it is cleaved like model, Cell. Mol. Life Sci., 2004, 61, 97–109. in AHA. In vitro experiments using an antibody raised [13] Da Lage J.-L., Danchin E.G.J., Casane D., Where do animal α-amylases come from? An interkingdom trip, FEBS Lett., 2007, against the AHA “propeptide”[10] on mussel or daphnia 581, 3927–3935. extracts on Western blots gave unclear results, because [14] Altschul S.F., Gish W., Miller W., Myers E.W., Lipman D.J., Basic although the antibody would react well with animal local alignment search tool, J. Mol. Biol., 1990, 215, 403–410. domains, it seems that artifactual cleavages occurred, [15] Edgar R.C., MUSCLE: multiple sequence alignment with high making interpretations difficult (not shown). Further accuracy and high throughput, Nucleic Acids. Res., 2004, 32, 1792–1797. studies are needed to assess the in situ localization of [16] Dereeper A., Guignon V., Blanc G., Audic S., Buffet S., Chevenet the C-terminal domain in animal tissues and its possible F., et al., Phylogeny.fr: robust phylogenetic analysis for the co-localization with the core enzyme. non-specialist, Nucleic Acids Res., 2008, 36, W465–W469. 34 Jean-Luc Da Lage

[17] Stam M.R., Danchin E.G.J., Rancurel C., Coutinho P.M., [26] Claverie P., Vigano C., Ruysschaert J.-M., Gerday C., Feller G., Henrissat B., Dividing the large glycoside hydrolase family 13 The precursor of a psychrophilic α-amylase: structural charac- into subfamilies: towards improved functional annotations of terization and insights into cold adaptation, Biochim. Biophys. α-amylase-related proteins, Protein Eng. Design Sel., 2006, 19, Acta, 2003, 1649, 119–122 555–562. [27] Blair J.E., Hedges S.B., Molecular phylogeny and divergence [18] Van Wormhoudt A., Sellos D., Cloning and sequencing analysis times of deuterostome animals, Mol. Biol. Evol., 2005, 22, of three amylase cDNAs in the Shrimp Penaeus vannamei 2275–2284. (Crustacea decapoda): evolutionary aspects, J. Mol. Evol., 1996, [28] Adamkewicz S.L., Harasewych M.G., Blake J., Saudek D., Bult 42, 543–551. C.J., A molecular phylogeny of the bivalve mollusks, Mol. Biol. [19] Le Moine S., Sellos D., Moal J., Daniel J.Y., San Juan Serrano F., Evol., 1997, 14, 619–629. Samain J.F., Van Wormhoudt A., Amylase in Pecten maximus [29] Delsuc F., Brinkmann H., Chourrout D., Philippe H., Tunicates (Mollusca, Bivalves): protein and cDNA characterization; and not cephalochordates are the closest living relatives of quantification of the expression in the digestive gland, Mol. vertebrates. Nature, 2006, 439, 965–968. Mar. Biol. Biotechnol., 1997, 6, 228–237. [30] Giribet G., Wheeler W., On bivalve phylogeny: a high-level [20] Moal J., Daniel J.Y., Sellos D., Van Wormhoudt A., Samain J.F., analysis of the Bivalvia (Mollusca) based on combined Amylase mRNA expression in Crassostrea gigas during feeding morphology and DNA sequence data, Invertebrate Biol., 2002, cycles, J. Comp. Physiol., 2000, 170, 21–26. 121, 271–324. [21] Feller G., Dehareng D., Da Lage J.-L., How to remain non-folded [31] Lecointre G., Le Guyader H., Classification phylogénétique du and pliable: the linkers in modular α-amylases as a case study, vivant, 3rd Ed., Belin, Paris, 2006. FEBS J., 2011, 278, 2333–2340. [32] Mallatt J., Giribet G., Further use of nearly complete 28S and [22] Da Lage J.-L., Maczowiak F., Cariou M.-L., Phylogenetic 18S rRNA genes to classify Ecdysozoa: 37 more arthropods and distribution of intron positions in α-amylase genes of Bilateria a kinorhynch, Mol. Phylogenet. Evol., 2006, 40, 772–794. suggests numerous gains and losses, PLoS One, 2011, 6, [33] Savard J., Tautz D., Richards S., Weinstock G.M., Gibbs R.A., e19673. Werren J.H., et al., Phylogenomic analysis reveals bees and [23] Xu D., Zhang Y., Ab initio protein structure assembly using wasps (Hymenoptera) at the base of the radiation of Holome- continuous structure fragments and optimized knowledge- tabolous insects, Genome Res., 2006, 16, 1334–1338. based force field, Proteins, 2012, 80, 1715–1735. [34] Johansson M.U., Zoete V., Michielin O., Guex N., Defining [24] Kim J.Y., Choi Y.L., Cho Y.S., Sung N.C., Kim C.H., Lee Y.C., Role and searching for structural motifs using DeepView/Swiss- of the C-terminal prodomain of aqualysin I in the maturation PdbViewer, BMC Bioinformatics, 2012, 13, 173. and translocation of the precursor across the cytoplasmic membrane in Escherichia coli, Biotechnol. Lett., 2002, 24, 433–439. Supplemental Material: The online version of this article [25] Boto L., Horizontal gene transfer in the acquisition of novel (DOI: 10.1515/amylase-2017-0003) offers supplementary material. traits by metazoans, Proc. Royal Soc. B, 2014, 281, 20132450.