Email updates

Keep up to date with the latest news and content from BMC Evolutionary Biology and BioMed Central.

Open Access Research article

Molecular evolution of the polyamine oxidase gene family in Metazoa

Fabio Polticelli12, Daniele Salvi3, Paolo Mariottini1, Roberto Amendola4 and Manuela Cervelli1*

Author Affiliations

1 Dipartimento di Biologia, Università “Roma Tre”, I-00146, Rome, Italy

2 National Institute of Nuclear Physics, Roma Tre Section, I-00146, Rome, Italy

3 CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, Campus Agrário de Vairão, 4485-661, Vairão, Portugal

4 ENEA, CR Casaccia, BAS.BIOTEC MED, I-00123, Rome, Italy

For all author emails, please log on.

BMC Evolutionary Biology 2012, 12:90  doi:10.1186/1471-2148-12-90


The electronic version of this article is the complete one and can be found online at: http://www.biomedcentral.com/1471-2148/12/90


Received:5 February 2012
Accepted:14 June 2012
Published:20 June 2012

© 2012 Polticelli et al.; licensee BioMed Central Ltd.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Polyamine oxidase enzymes catalyze the oxidation of polyamines and acetylpolyamines. Since polyamines are basic regulators of cell growth and proliferation, their homeostasis is crucial for cell life. Members of the polyamine oxidase gene family have been identified in a wide variety of animals, including vertebrates, arthropodes, nematodes, placozoa, as well as in plants and fungi. Polyamine oxidases (PAOs) from yeast can oxidize spermine, N1-acetylspermine, and N1-acetylspermidine, however, in vertebrates two different enzymes, namely spermine oxidase (SMO) and acetylpolyamine oxidase (APAO), specifically catalyze the oxidation of spermine, and N1-acetylspermine/N1-acetylspermidine, respectively. Little is known about the molecular evolutionary history of these enzymes. However, since the yeast PAO is able to catalyze the oxidation of both acetylated and non acetylated polyamines, and in vertebrates these functions are addressed by two specialized polyamine oxidase subfamilies (APAO and SMO), it can be hypothesized an ancestral reference for the former enzyme from which the latter would have been derived.

Results

We analysed 36 SMO, 26 APAO, and 14 PAO homologue protein sequences from 54 taxa including various vertebrates and invertebrates. The analysis of the full-length sequences and the principal domains of vertebrate and invertebrate PAOs yielded consensus primary protein sequences for vertebrate SMOs and APAOs, and invertebrate PAOs. This analysis, coupled to molecular modeling techniques, also unveiled sequence regions that confer specific structural and functional properties, including substrate specificity, by the different PAO subfamilies. Molecular phylogenetic trees revealed a basal position of all the invertebrates PAO enzymes relative to vertebrate SMOs and APAOs. PAOs from insects constitute a monophyletic clade. Two PAO variants sampled in the amphioxus are basal to the dichotomy between two well supported monophyletic clades including, respectively, all the SMOs and APAOs from vertebrates. The two vertebrate monophyletic clades clustered strictly mirroring the organismal phylogeny of fishes, amphibians, reptiles, birds, and mammals. Evidences from comparative genomic analysis, structural evolution and functional divergence in a phylogenetic framework across Metazoa suggested an evolutionary scenario where the ancestor PAO coding sequence, present in invertebrates as an orthologous gene, has been duplicated in the vertebrate branch to originate the paralogous SMO and APAO genes. A further genome evolution event concerns the SMO gene of placental, but not marsupial and monotremate, mammals which increased its functional variation following an alternative splicing (AS) mechanism.

Conclusions

In this study the explicit integration in a phylogenomic framework of phylogenetic tree construction, structure prediction, and biochemical function data/prediction, allowed inferring the molecular evolutionary history of the PAO gene family and to disambiguate paralogous genes related by duplication event (SMO and APAO) and orthologous genes related by speciation events (PAOs, SMOs/APAOs). Further, while in vertebrates experimental data corroborate SMO and APAO molecular function predictions, in invertebrates the finding of a supported phylogenetic clusters of insect PAOs and the co-occurrence of two PAO variants in the amphioxus urgently claim the need for future structure-function studies.

Background

Polyamines (PA), such as spermine (Spm), spermidine (Spd) and putrescine (Put), are polybasic molecules ubiquitous in living organisms, with many important biological functions. These molecules directly affect cell growth, differentiation, and apoptosis by reversibly interacting with nucleic acids, regulating chromatin status and gene expression, and modulating ion-channels’ function and stability [1,2]. Polyamine oxidases (PAO) are flavin adenine dinucleotide- (FAD-) containing enzymes that catalyze the oxidation of polyamines. The substrate specificity and the nature of the oxidation products depend on the source of the enzymes. Generally, PAOs oxidizes Spm, N1-acetylspermine (N1-acetylSpm) and N1-acetylspermidine (N1-acetylSpd), but not Spd. On the other hand, in vertebrates Spm is directly oxidized by the cytosolic enzyme spermine oxidase (SMO, EC number 1.5.3.16), a flavoprotein characterized in the past as a human polyamine oxidase (PAO-h1) [3] and then subsequently named SMO [4,5]. Spm oxidation leads to the production of Spd, 3-aminopropanal and hydrogen peroxide (H2O2). While N1-acetylSpm and N1-acetylSpd are oxidized by the peroxisomal FAD-dependent enzyme N1-acetylpolyamine oxidase (APAO, EC number 1.5.3.11) to produce respectively Spd and Put, 3-aceto-aminopropanal and H2O2. Substrate oxidation modes of SMO and APAO are summarized in Figure 1.

thumbnailFigure 1. Enzymatic reaction catalyzed by SMO and APAO proteins. SMO oxidises the carbon on the exo-side of the N5-nitrogen of Spm, producing Spd, 3-aminopropanal (3-AP) and hydrogen peroxide (H2O2). APAO oxidises the carbon on the exo-side of the N5-nitrogen of N1Ac-Spm and N1Ac-Spd producing Spd and Put respectively, in addition to 3-aceto-aminopropanal (3-aceto-AP) and H2O2.

In the last decades, these PA catabolic enzymes have been extensively characterized and it is well documented that both enzymes play an essential role in maintaining vertebrate PA homeostasis, which is mandatory for cellular life [2,6-9]. Unfortunately, repeated attempts from independent labs to obtain SMO and APAO crystals suitable for X-ray diffraction studies failed [A. Mattevi, personal communication; A. Fiorillo and A. Ilari, personal communication]. The only structural data available for these enzymes are those derived from molecular modeling studies of mouse SMO [5,10,11] and mouse APAO [12] which indicated a very similar active site environment for the two proteins, making it difficult to rationalize their different substrate specificity and sensitivity to small molecule inhibitors [13]. On the contrary, the crystal structure of the Saccharomyces cerevisiae PAO (FMS1) has been obtained and its biochemical characterization proved that it is able to oxidize Spm, N1-acetylSpm and N1-acetylSpd [14,15]. Since the yeast PAO is capable to catalyse the oxidation of both acetylated and non-acetylated polyamines, and in vertebrates these functions are addressed by two specialized polyamine oxidase subfamilies (APAO and SMO), it can be hypothesized an ancestral reference for PAO enzymes and a paralogous relationships between APAOs and SMOs. Yet, we still have a limited knowledge on the structural and functional diversity of metazoans polyamine oxidases and their evolutionary history has never been studied.

In this study we developed a phylogenomic framework [16] to investigate the phylogenetic relationships, the structural evolution and the functional divergence among polyamine oxidases proteins in animals. We identified, through an exhaustive BLASTP search strategy all the available proteins homologous to SMO and APAO enzymes and we assembled a comprehensive multiple amino acid sequences alignment including 76 polyamine oxidases from all the vertebrate classes and main invertebrate phyla. Phylogenetic analysis allowed inferring an evolutionary scenario where the ancestor PAO coding sequence, present in invertebrates as an orthologous gene, has been duplicated in the vertebrate branch to originate the paralogue SMO and APAO genes. Overlaying the tree topology with data from molecular structure modelling and biochemical function data/prediction, we traced along the evolutionary tree the processes behind the origin of the functional and structural diversity found in polyamine oxidase proteins. Finally, the presence of the alternative SMO protein isoform [10,17,18] was confirmed in all the placental mammals analysed, suggesting that in this group a mechanism of alternative splicing (AS) allowed a further increase of the structural and functional variation of the SMO proteins.

Results and discussion

Clustering SMO, APAO and PAO homologous proteins

An exhaustive species-specific BLASTP search of PAO homologs was carried out in the available databases using the following threshold values: E-value = 1×10-10, > 30% sequence identity over at least 50% of the ORF length (see section Methods for details). These searches enabled us to retrieve 76 homologous PAO proteins from the baker’s yeast (Saccharomyces cerevisiae) and 52 animal taxa (Table 1), including nematode (Caenorhabditis elegans), placozoa (Trichoplax adhaerens), echinoderms (Strongylocentrotus purpuratus and Nematostella vectensis), arthropods (Pediculus humanus corporis, Tribolium castaneum, Drosophila melanogaster, Anopheles gambia, Nasonia vitripennis, and Apis mellifera), and chordates such as urochordates (Ciona intestinalis), cephalochordates (Branchiostoma floridae), and vertebrates: fishes (Tetraodon nigroviridis, Takifugu rubripes, Oryzias latipes, Danio rerio, and Gasterosteus aculeatus), amphibians (Xenopus laevis and X. tropicalis), reptile (Anolis carolinensis), birds (Gallus gallus, Taeniopygia guttata, and Meleagris gallonavo), monotremate mammal (Ornithorhyncus anatinus), marsupial mammals (Monodelphis domestica, and Macropus eugenii), and placental mammals (Procavia capensis, (Loxodonta africana, Microcebus murinus, Rattus norvegicus, Mus musculus, Cavia porcellus, Dipodomys ordii, Tarsius syrinchtae, Callithrix jacchus, Macaca fascicularis, Nomascus leucogenys, Gorilla gorilla, Pan troglodytes, Pongo pygmaeus, Pongo abelii, Homo sapiens, Equus caballus, Pteropus vampyrus, Myosotis lucifugus, Felis catus, Canis familiaris, Ailuropoda melanoleuca, Sus scrofa, Tursiops truncatus, Bos taurus, and Oryctogalus cuniculus).

Table 1. Polyamine oxidase proteins sequences used in this study

Among the 75 homologous PAO sequences found in Metazoa, 13 sequences are annotated as invertebrate PAO proteins, 36 as vertebrate SMO proteins, and 26 as vertebrate APAO proteins. In few cases among arthropods we found additional sequences which showed some similarity with PAOs, but their overall identity with PAOs was lower than 30% and it was not possible to obtain a reliable alignment along the entire length. Thus, we consider these sequences as non-homologs to PAOs. The lower number of APAO sequences retrieved compared with SMO sequences is due to the availability of sequences in Genbank rather than to the absence in certain vertebrate taxa of APAO proteins.

Phylogeny and evolution of the polyamine oxidase family

The phylogenetic relationship among the polyamine oxidase proteins as inferred by the Maximum Likelihood and Bayesian trees were consistent at the main nodes and revealed a basal position of all the PAO enzymes present in invertebrates relative to vertebrate SMOs and APAOs (Figure 2). Polyamine oxidases from invertebrates such as placozoa, nematodes, echinoderms, and urochordates do not constitute a monophyletic assemblage, while PAOs from insects, although fairly differentiated from each other, all shared a common ancestor and constitute a derived and supported clade within the PAO group. However, the lack of support for other PAO groups can be due to a sparse taxon sampling [19]. The two PAO variants sampled in the cephalochordate amphioxus are basal to the dichotomy between two well supported monophyletic clades including, respectively, all the SMOs and APAOs from vertebrates.

thumbnailFigure 2. The evolutionaty tree of the the polyamine oxidase proteins. The evolutionary history tree of the SMO, PAO, and APAO polyamine oxidase proteins in Metazoa as inferred by using the Maximum Likelihood method based on the JTT model with a proportion of invariable sites (I) and gamma-distributed rates across sites (G) (Bayesian tree showed identical topology at the main nodes; data not shown). The analysis involved 76 amino acidic sequences with the yeast sequence used as outgroup. In correspondence of the main nodes (black circles) the bootstrap support (BS) and Bayesian posterior probabilities (BPP) values are reported (above: BS; below: BPP). The support for the secondary nodes is reported as white circles (BS = 100; BPP = 1.00) and grey circles (BS ranging from 90 to 99; BPP = 1.00). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Main monophyletic groups are indicated as follow: M = mammals; B = birds; R = reptiles; A = amphibians; F = fishes; I = insects.

These results provide evidence that SMO and APAO subfamilies originate from a duplication event from a PAO-like gene ancestor, while gene speciation accounts for their ubiquitous occurrence in vertebrates. Moreover, the obvious difference in substrate specificity between SMO and APAO enzymes ([3,20] and results from this study) indicates that after the duplication these gene subfamilies underwent divergent evolution and functional specialization.

Within both SMO and APAO subfamilies, ortholog sequences show substantial diversity and divergence among and within vertebrate classes. Fish polyamine oxidases constitute a distinct clade from that of tetrapods in which polyamine oxidases from amphibians, reptiles, birds and mammals cluster in different protein lineages, although phylogenetic relationships between these protein lineages are largely unresolved both in the SMO and APAO branches (Figure 2). Within the mammals a further evolution of both SMO and APAO polyamine oxidase proteins can be traced along the evolution of Monotremata, Marsupialia, and between major groups of Placentalia. Indeed, the APAOs of Afrotheria (Proboscidea and Hyracoidea), Cetartiodactyla (Artiodactyla and Cetacea), Rodentia and Primates, as well as SMOs of Carnivora, Cetartiodactyla (Artiodactyla and Cetacea), and Primates, constituted monophyletic supported clades ([21,22] for taxonomic reference). Thus, the phylogenetic relationships among both SMO and APAO orthologs strictly mirror the phylogenetic relationships within vertebrates, suggesting that these SMO and APAO proteins evolved throughout the speciation events in vertebrates.

The co-occurrence of SMO and APAO enzymes in all the vertebrates suggests that their specific functions evolved earlier than vertebrates. Besides, the presence of two related PAOs in the cephalochordate amphioxus suggests that the duplication event from which they originated could have even pre-dated the Chordate radiation. According to this hypothesis, we would have expected two PAO gene copies also in the urochordate taxon here analysed, Ciona intestinalis, as recent molecular phylogenetic studies place cephalochordates as the basal group within the phylum Chordata, with vertebrates and urochordates diverging later [23,24]. However, several studies demonstrated that the urochordate genomes have lost many genes respect to cephalochordate and vertebrate genomes [25-27], thus the presence of a single PAO gene copy in Ciona intestinalis could be due to a secondary gene loss. An alternative and equally parsimonious hypothesis could be that a lineage-specific duplication of the PAO gene occurred in the amphioxus and an independent duplication followed by functional divergence arose in the ancestor of the vertebrates. In amphioxus several genes have gone through lineage-specific duplications relative to the chordate ancestor, for example some genes belonging to the opsin and to the innate immunity receptor groups are organized in gene families that have dramatically expanded copy number of homologs (reviewed in [28]; see also [27,29,30]). These findings are not surprising given that amphioxus has been evolving from the common ancestor of cephalochordates and vertebrates for over 550 Myr. In the case of the two PAO gene copies of amphioxus, due to their low amino acid sequence similarity (37%), these genes likely arose by an ancient duplication event. However, a broad comparative genome analysis and structure-function studies on chordates PAOs are required before it will be possible to understand whether this duplication pre-dated the Chordate radiation or it is specific to the amphioxus lineage and how the functions of the two PAO genes found in such chordates were partitioned during evolution.

Structural and functional properties of invertebrate PAOs

Polyamine oxidases from invertebrates do not constitute a monophyletic assemblage in our phylogenetic reconstruction and no clear pattern of residues conservation can be evidenced in the active site region (Additional file 1: Figure S1). This finding is consistent with the hypothesis that invertebrate PAOs have broad substrate specificity as indeed it has been demonstrated for the best characterized member of this proteins group, the Saccharomyces cerevisiae PAO (FMS1). In fact, FMS1 is capable of oxidizing Spm, N1-acetylSpm, N1-acetylSpd, and N8-acetylSpd, but not Spd [15].

Additional file 1. Figure S1. Amino acid sequence alignment of selected polyamine oxidases. The amino acid blocks are coloured according to BLOSUM62 score (dark blocks corresponding to higher score). Residues building up the active site are indicated by red stars, those forming the putative SMOs specificity pocket (Glu216 and Ser218 in mouse SMO) are indicated by green stars. For acronyms and isoform numbering see Tables 1 and 2. The amino acid sequence of the Zea mays PAO (ZmPAO) whose structure has been used for comparative modeling of the DmPAO is also shown for reference. The figure was generated using Jalview [32].

Format: DOC Size: 624KB Download file

This file can be viewed with: Microsoft Word ViewerOpen Data

Interestingly, according to the phylogenetic analysis (Figure 2), PAOs from insects are all derived from a common ancestor and constitute a well-defined and supported clade within the PAOs’ branch. To investigate the structural and functional properties of insect PAOs, a molecular model of Drosophila melanogaster PAO (DmPAO) has been built by homology using the three-dimensional structure of maize PAO, the closest homolog found in the Protein Data Bank [31], as a template. Unexpectedly, inspection of the active site region of DmPAO evidenced a stronger structural similarity with the SMOs active site than with that of the invertebrate PAO and of FMS1 (Figure 3). In particular, all the residues involved in substrate binding in SMOs are conserved in DmPAO with the exception of Glu224 substituted with a Thr residue which, nonetheless, can potentially interact with the substrate primary amino group (Figure 3). In addition, comparative analysis of the amino acid sequences of insects PAOs and mammalian SMOs and APAOs indicates that the polar active site pocket hypothesized to be responsible for the substrate specificity of SMOs (see the next section) has very similar properties in insect PAOs. This observation suggests that substrate specificity of insect PAOs may not be as broad as that of FMS1. However, our in silico analysis of DmPAO does not allow to draw a reliable conclusion on this respect and for a deep understanding of the functional properties of insects PAOs a structural and biochemical characterization of these enzymes is required.

thumbnailFigure 3. Molecular models of the active site region of SMO and PAO. Schematic representation of the active site region of mouse SMO (MmSMO), Drosophila melanogaster PAO (DmPAO) and Saccharomyces cerevisiae PAO (FMS1) in complex with the substrate Spm. MmSMO and DmPAO complexes with Spm have been obtained by molecular modelling ([11] and this work, respectively) while the structure of the FMS1-Spm has been determined experimentally [14]. For the sake of clarity, the FAD cofactor is coloured in purple, backbone atoms in orange and Spm carbon atoms in green.

In contrast with other invertebrates, the genome of the amphioxus encodes two PAO-isologs. As previously discussed, two PAO copies are only found in vertebrates and in the amphioxus suggesting that the duplication of the ancestral PAO gene predated the Chordate radiation. Thus, the knowledge of the structural and biochemical properties of the amphioxus’ PAO variants is crucial for understanding whether the duplication of the ancestral PAO gene and the functional divergence between the resulting paralogs were strictly coupled or not during the evolution of the polyamine oxidase subfamilies. Although the biochemical function of the two amphioxus’ PAOs is completely unknown, as they show a substantial sequence divergence (see Additional file 1: Figure S1) it is plausible the hypothesis of an early functional divergence between them from an ancestral PAO-like function.

Structural and functional properties of SMOs and APAOs

Molecular modeling allowed mapping the sequence conservation of SMOs and APAOs onto the three-dimensional structure of a previously published molecular model of these enzymes [11,12]; see Methods section and to identify critical amino acids of the active site regions. The three-dimensional models displayed in Figure 4, shows that the specificity of the SMO enzymes results in a very high degree of conservation of all the residues building up the active site cavity. In detail, polar residues (His82, Gln200, Glu224, Tyr482, Ser527, Thr528) and hydrophobic residues (Trp80 and Trp427) thought to interact with Spm, as well as the catalytically crucial Lys367 [11], are all conserved in more than 90% of the sequences analysed. Interestingly, also residues Glu216 and Ser218, which form a pocket on one side of the active site, are almost strictly conserved in all SMO sequences analysed. In the case of APAO, instead, there seems to be a lower evolutionary pressure towards strict conservation of the active site residues, the only residues conserved in more than 90% of the sequences analysed being Trp62, His64, Tyr430, Ser473 and Thr474 (ortholog to SMO residues Trp80, His82, Tyr482, Ser527 and Thr528), besides the catalytically important Lys315 (Figure 4). It is worthwhile to note that all the residues conserved in APAOs are shared with SMOs and thus the different specificity of the two enzymes appears difficult to rationalize. From this viewpoint, an interesting observation is that the SMOs strictly conserved residues Glu216 and Ser218 are substituted by hydrophobic residues in APAOs (typically with a Leu and an Ala residue, respectively) making the corresponding active site pocket of APAO fit to host a hydrophobic group rather than a charged one. Indeed, docking of N1-acetylSpm within the mouse SMO active site indicates that the methyl group of this molecule would be placed in the polar pocket made up by Glu216 and Ser218, making energetically unfavourable the binding of N1-acetylSpm within the SMO active site [11]. On the contrary, binding of Spm to APAO would lead to positioning of one of the charged terminal amino groups of this molecule within the Leu-Ala hydrophobic pocket, making energetically unfavourable the binding of Spm within the APAO active site [11].

thumbnailFigure 4. Three-dimensional view of SMOs and APAOs sequence conservation. Structure-based views of the amino acid sequence conservation in the active site regions of vertebrate SMOs and APAOs. Protein regions coloured in blue correspond to residues conserved in at least 90% of the amino acid sequences analysed. Top panel: Amino acid sequence conservation in the SMO family mapped onto the structural model of mouse SMO [11]. Middle panel: Amino acid sequence conservation in the APAO family mapped onto the structural model of mouse APAO [12]. Bottom panel: Amino acid sequence conservation in the vertebrate PAO family (SMO and APAO sequences combined) mapped onto the structural model of mouse SMO. The green ellipse indicates the location of the polar pocket made up SMOs by residues Glu216 and Ser218 (numbering of the mouse enzyme), which are substituted by aliphatic amino acids in APAOs. The figure was generated using Jalview [32].

Mammalian SMO “long” isoform

SMO genes are able to increase the structural and functional variation of the corresponding proteins following an alternative splicing (AS) mechanism. Indeed, alternative splicing isoforms of SMO have been described in human and mouse, defined as “major” (i.e., SMO1 and SMOα for human and mouse proteins, respectively) and “long” isoforms (i.e., SMO5 and SMOμ for human and mouse proteins, respectively) [17,18]. The long isoforms possess an additional exon (exon VIa in mouse SMO) which is responsible for their nuclear localization and thus defined as Nuclear Domain B (NDB) [10]. The nuclear localization of the long isoforms requires also a second domain termed Nuclear Domain A (NDA), which is an internal amino acid stretch of the common exon V also present in the major isoforms [10].

To investigate the presence in mammalian genomes of these alternative splicing SMO isoforms and to analyse their sequence conservation, we performed a search in the public databases based on exon VIa sequence annotation that enabled us to retrieve 22 sequences corresponding to the long SMO isoform Table 2. All these sequences were identified in genomes from placental mammals. Surprisingly, no corresponding alternative splicing variant was found in marsupials (wallaby and gray short-tailed opossum) and monotremates (platypus) SMO genes (Figure 5) which also showed a lack of sequence conservation in the NDA regions. The absence of sequence conservation of the NDA and the lacking of the NDB was also confirmed in the SMO genes of birds, reptiles, amphibians and fishes (Figure 5). On the contrary, both the NDA and the NDB displayed a high degree of sequence conservation across all the placental mammals analysed, suggesting common structural and functional properties. Given the ubiquitous occurrence of these SMO isoforms in placental mammals, their high degree of sequence conservation, and their absence in marsupials and monotremates, a role of SMO isoforms in placenta development can be postulated.

Table 2. SMO sequences used in the analysis of the additional exon VIa

thumbnailFigure 5. Amino acid sequence alignment and structure-based view of SMO isoforms. A) Amino acid sequence alignment of the regions corresponding to Nuclear Domain A (NDA) and B (NDB) of SMO long isoforms. For acronyms and isoform numbering see Table 2. B) Structure-based view of the amino acid sequence conservation in SMOs (left) and in the NDA of placental mammals as opposed to marsupials and monotremates (right). Protein regions coloured in blue correspond to residues conserved in at least 90% of the amino acid sequences analysed.

These results suggested that SMO isoforms can be considered as a relatively recent evolutionary acquisition of (placental) mammals, where SMO gene increased its functional variation following an AS mechanism. Therefore, in PAOs, besides gene duplication, the AS mechanism is a further evolutionary mechanism which efficiently amplifies the gene variation and relative functional differentiation by producing different transcripts through a splicing machinery, even though the gene remains a single copy [33,34].

Conclusions

In the present study the explicit integration of phylogenetic reconstruction, homology-based structure prediction, and biochemical function data/prediction allowed understanding the molecular evolution of the polyamine oxidase gene family in Metazoa and provided a deeper insight into the structure(s)-function(s) of individual members. Gene duplication and speciation were identified as the evolutionary processes by which the functional and structural diversity observed in this protein family originated. Through gene duplication from an ancestral PAO-like gene coding for a polyamine oxidase with broad substrate specificity, the vertebrate SMO and APAO subfamilies evolved related but distinct functions. On the other hand, gene speciation (and also alternative splicing in the case of SMO of placental mammals) accounts for the protein diversity observed within each one of these two subfamilies.

The phylogenomic approach here employed allowed to trace along the evolutionary tree the acquisition of specific biochemical functions by the SMO and APAO subfamilies. However, while for vertebrates a representative sampling of SMO and APAO protein sequences is available and experimental data corroborate their predicted molecular function, for invertebrates PAOs few protein sequences and no structural and biochemical data are available, thus preventing a conclusive inference of their function and relationships. In this respect, the finding of the monophyletic clade of insect PAOs, and the occurrence of two PAO variants in the cephalochordate amphioxus, provide a guide for future structure-function studies aimed at clarifying the biochemical function of PAOs in invertebrates and the timing of the duplication event which originated the vertebrate SMO and APAO subfamilies.

Methods

Protein and gene sequence homology search and retrieval

Annotated protein databases such as UniProt (http://www.ebi.ac.uk/uniprot/ webcite) and PFAM (http://www.sanger.ac.uk/Sofware/Pfam/ webcite) and major sequence repositories such as the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/ webcite) provide direct access to many known, full-length SMO, APAO and PAO protein sequences. Amino acid sequences from animal taxa and from the yeast were obtained using a combination of queries based on key terms and BLASTP [35] searches. Whenever possible, sequences whose enzymatic activities had been previously verified have been used as query sequences in all blast searches. Additional SMO, APAO and PAO predicted protein sequences were retrieved from currently sequenced and unfinished genomes at the Ensembl (http://www.ensembl.org/index.html webcite) database. A BLAST search was conducted (search with the same dataset used as query and subject database) on mouse representative ORF databases separately to detect homology. An E-value of 1×10-10 was used in the BLAST search at the Ensembl (http://www.ensembl.org/index.html webcite). Individual BLAST hits found for the same pair of ORFs were combined in the following way: when two ORFs have multiple BLAST hits with overlapping alignable regions, a non-overlapping combination of those hits with the longest alignable regions was selected out of all possible combinations. This combination was then treated as a single BLAST hit. Next the BLAST hits were filtered to keep only those with at least 30% identity, bitscore of at least 50, aligning at least 50% of the length of both ORFs, and excluding self-hits. Then duplicates, or multi-loci genes, were defined as pairs of ORFs having two-way hits in the filtered set of BLAST hits. Thus all ORFs were classified as singletons or duplicates. Gene families were then identified using single-linkage clustering: Step 1. Initially all genes are in their own families. Step 2. When two genes A and B are found to have a two-way hit, their whole families are merged together. Step 3. Repeat step 2 until no further merging can be done.

Additionally, to investigate the presence in mammalian genomes of alternative splicing SMO isoforms, a search in the public databases was performed based on exon VIa sequence annotation [10].

Phylogenic analysis

Phylogenetic analyses were performed by the Maximum Likelihood and the Bayesian inference methods using the amino acid sequences with the yeast as outgroup.

Maximum Likelihood analysis was carried out in PALM [36], which incorporates several programs in an integrated framework for phylogenetic reconstruction with automatic likelihood model selectors. Multiple amino acidic sequences alignment was performed by Clustal W [37]. The fitness among 112 models of protein evolution was estimated trough PhyML/ProtTest [38,39] and the optimal model was selected under the Bayesian Information criterion (BIC). The JTT model with a proportion of invariable sites (I) and gamma-distributed rates across sites (G) was selected for the phylogenetic inference performed via PhyML. Nodes support for the resulting phylogenetic tree was evaluated by 500 bootstrap replications.

Bayesian inference analysis was performed with MrBayes 3.2 [40] under the JTT model of protein evolution with gamma-distributed rates across sites and a proportion of invariable sites (rates = invgamma) suggested by ProtTest. Two independent runs of four Markov chain Monte Carlo (MCMC) chains each were executed in parallel for two million generations, sampling every 100 generations. Posterior probabilities for nodes were derived from a majority rule consensus of the trees sampled after convergence (25% was setted as burnin for sumt and sump).

Homology modeling and identification of critical amino acids

The multiple alignment of the retrieved PAO sequences, obtained using Clustal W 2.0 [41], was visualized using Jalview alignment editor [32] (Additional file 1: Figure S1).

The Jalview tool, which allows mapping of the sequence conservation data onto the three-dimensional structure of a reference protein, was used to generate a three-dimensional view of the sequence conservation data in the active site region of SMOs and APAOs, using previously published molecular models of the mouse enzymes [11,12].

The structural model of Drosophila melanogaster PAO (DmPAO) has been built by homology using the three dimensional structure of maize PAO (ZmPAO; PDB code 1B5Q) [31], the closest homologue found in the Protein Data Bank, as a template. In detail, the template structure was chosen through two PSI-Blast [42] iterations against the PDB sequence database using the sequence coded Swiss-Prot Q9VHN8 (corresponding to DmPAO) as a bait. Pairwise sequence alignment between DmPAO and ZmPAO was extracted from a multiple sequence alignment of all known PAO sequences obtained using Clustal W 2.0. Based on this alignment, the structural model of DmPAO was built using the homology modeling program Nest [43]. The ‘alignment tuning’ option of nest was used to refine the sequence alignment, to avoid the unlikely occurrence of insertions and deletions within secondary structure elements.

Abbreviations

APAO: Acetylpolyamine oxidase; AS: Alternative splicing; FMS1: Saccharomyces cerevisiae polyamine oxidase protein; NDA: Nuclear Domain A; NDB: Nuclear Domain B; PA: Polyamines; PAO: Polyamine oxidase; SMO: Spermine oxidase.

Authors’ contribution

This work was planned by FP and PM and carried out in collaboration with DS and MC. DS performed the phylogenetic analyses and participated in discussing the evolutionary aspects and drafting the manuscript. RA critically revised the manuscript for important intellectual content and data analysis. All authors read and approved the final manuscript.

Acknowledgements

FP, MC and PM wish to thank the University of Roma Tre for financial support. DS is supported by FCT post-doctoral grants SFRH/BPD/66592/2009.

References

  1. Cohen SS: Polyamine oxidases and dehydrogenases. In A guide to the polyamines. Edited by Cohen SS. Oxford University Press, New York; 1998:1-82. OpenURL

  2. Wallace HM, Fraser AV, Hughes A: A perspective of polyamine metabolism.

    Biochem J 2003, 376:1-14. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  3. Wang Y, Devereux W, Woster PM, Stewart TM, Hacker A, Casero RA Jr: Cloning and characterization of a human polyamine oxidase that is inducible by polyamine analogue exposure.

    Cancer Res 2001, 61:5370-5373. PubMed Abstract | Publisher Full Text OpenURL

  4. Vujcic S, Diegelman P, Bacchi CJ, Kramer DL, Porter CW: Identification and characterization of a novel flavin-containing spermine oxidase of mammalian cell origin.

    Biochem J 2002, 367:665-675. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  5. Cervelli M, Polticelli F, Federico F, Mariottini P: Heterologous expression and characterization of mouse spermine oxidase.

    J Biol Chem 2003, 278:5271-5276. PubMed Abstract | Publisher Full Text OpenURL

  6. Thomas T, Thomas TJ: Polyamines in cell growth and cell death: molecular mechanisms and therapeutic applications.

    Cell Mol Life Sci 2001, 2:244-258. OpenURL

  7. Agostinelli E, Arancia G, Dalla Vedova L, Belli F, Marra M, Salvi M, Toninello A: The biological functions of polyamine oxidation products by amine oxidises: perspectives of clinical applications.

    Amino Acids 2004, 27:347-358. PubMed Abstract | Publisher Full Text OpenURL

  8. Casero RA Jr, Marton LJ: Targeting polyamine metabolism and function in cancer and other hyperproliferative diseases.

    Nature Rev Drug Disc 2007, 6:373-390. Publisher Full Text OpenURL

  9. Amendola R, Cervelli M, Fratini E, Polticelli F, Sallustio DE, Mariottini P: Spermine metabolism and anticancer therapy.

    Curr Cancer Drug Targets 2009, 9:118-130. PubMed Abstract | Publisher Full Text OpenURL

  10. Bianchi M, Amendola R, Federico R, Polticelli F, Mariottini P: Two short protein domains are responsible for the nuclear localization of mouse spermine oxidase (mSMO)μ isoform.

    FEBS J 2005, 272:3052-3059. PubMed Abstract | Publisher Full Text OpenURL

  11. Tavladoraki P, Cervelli M, Antonangeli F, Minervini G, Stano P, Federico R, Mariottini P, Polticelli F: Probing mammalian spermine oxidase enzyme–substrate complex through molecular modeling, site-directed mutagenesis and biochemical characterization.

    Amino Acids 2011, 840:1115-1126. OpenURL

  12. Bianchi M, Polticelli F, Ascenzi P, Botta M, Federico R, Mariottini P, Cona A: Inhibition of polyamine and spermine oxidases by polyamine analogues.

    FEBS J 2006, 273:1115-1123. PubMed Abstract | Publisher Full Text OpenURL

  13. Cervelli M, Amendola R, Polticelli F, Mariottini P: Spermine oxidase: ten years after.

    Amino Acids 2012, 42:441-450. PubMed Abstract | Publisher Full Text OpenURL

  14. Huang Q, Liu Q, Hao Q: Crystal structures of Fms1 and its complex with spermine reveal substrate specificity.

    J Mol Biol 2005, 348:951-959. PubMed Abstract | Publisher Full Text OpenURL

  15. Landry J, Sternglanz R: Yeast Fms1 is a FAD-utilizing polyamine oxidase.

    Biochem Biophys Res Commun 2003, 303(3):771-776. PubMed Abstract | Publisher Full Text OpenURL

  16. Sjölander K: Phylogenomic inference of protein molecular function: advances and challenges.

    Bioinformatic 2004, 20:170-179. Publisher Full Text OpenURL

  17. Murray-Stewart T, Wang Y, Devereux W, Casero RA Jr: Cloning and characterization of multiple human polyamine oxidase splice variants that code for isonzymes with different biochemical characteristics.

    Biochem J 2002, 368:673-677. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  18. Cervelli M, Bellini A, Bianchi M, Marcocci L, Nocera S, Polticelli F, Federico R, Amendola R, Mariottini P: Mouse spermine oxidase gene splice variants: nuclear sub-cellular localization of a novel active isoform.

    Eur J Biochem 2004, 271:760-770. PubMed Abstract | Publisher Full Text OpenURL

  19. Zwickl DJ, Hillis DM: Increased taxon sampling greatly reduces phylogenetic error.

    Syst Biol 2002, 51:588-598. PubMed Abstract | Publisher Full Text OpenURL

  20. Vujcic S, Liang P, Diegelman P, Kramer DL, Porter CW: Genomic identification and biochemical characterization of the mammalian polyamine oxidase involved in polyamine back-conversion.

    Biochem J 2003, 370:19-28. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  21. Eizirik E, Murphy WJ, O'Brien SJ: Molecular dating and biogeography of the early placental mammal radiation.

    J Hered 2001, 92:212-219. PubMed Abstract | Publisher Full Text OpenURL

  22. Prasad AB, Allard MW: NISC Comparative Sequencing Program, Green ED: Confirming the Phylogeny of Mammals by Use of Large Comparative Sequence.

    Mol Biol Evol 2008, 25:1795-1808. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  23. Bourlat S, Juliusdottir T, Lowe CJ, Freeman R, Aronowicz J, Kirschner M, Lander ES, Thorndyke M, Nakano H, Kohn AB, et al.: Deuterostome phylogeny reveals monophyletic chordates and the new phylum Xenoturbellida.

    Nature 2006, 444:85-88. PubMed Abstract | Publisher Full Text OpenURL

  24. Delsuc F, Brinkmann H, Chourrout D, Philippe H: Tunicates and not cephalochordates are the closest living relatives of vertebrates.

    Nature 2006, 439:965-968. PubMed Abstract | Publisher Full Text OpenURL

  25. Holland LZ, Gibson-Brown JJ: The Ciona intestinalis genome: when the constraints are off.

    Bioessays 2003, 25:529-532. PubMed Abstract | Publisher Full Text OpenURL

  26. Vienne A, Pontarotti P: Metaphylogeny of 82 gene families sheds a new light on chordate evolution.

    Int J Biol Sci 2006, 2(2):32-37. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  27. Holland LZ, Albalat R, Azumi K, et al.: The amphioxus genome illuminates vertebrate origins and cephalochordate biology.

    Genome Res 2008, 18(7):1100-1111. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  28. Minguillón C, Ferrier DEK, Cebrián C, Garcia-Fernàndez J: Gene duplications in the prototypical cephalochordate amphioxus.

    Gene 2002, 287:121-128. PubMed Abstract | Publisher Full Text OpenURL

  29. Meulemans D, Bronner-Fraser M: The Amphioxus SoxB Family: Implications for the Evolution of Vertebrate Placodes.

    Int J Biol Sci 2007, 3(6):356-364. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  30. Putnam NH, et al.: The amphioxus genome and the evolution of the chordate karyotype.

    Nature 2008, 453(7198):1064-1071. PubMed Abstract | Publisher Full Text OpenURL

  31. Binda C, Coda A, Angelini R, Federico R, Ascenzi P, Mattevi A: A 30-angstrom-long U-shaped catalytic tunnel in the crystal structure of polyamine oxidase.

    Structure 1999, 7(3):265-276. PubMed Abstract | Publisher Full Text OpenURL

  32. Waterhouse AM, Procter JB, Martin DMA, Clamp M, Barton GJ: Jalview Version 2 - a multiple sequence alignment editor and analysis workbench.

    Bioinformatics 2009, 25(9):1189-1191. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  33. Jin L, Kryukov K, Clemente JC, Komiyama T, Suzuki Y, Imanishi T, Ikeo K, Gojobori T: The evolutionary relationship between gene duplication and alternative splicing.

    Gene 2008, 427:19-31. PubMed Abstract | Publisher Full Text OpenURL

  34. Kondrashov FA, Koonin EV: Evolution of alternative splicing: deletions, insertions and origin of functional parts of proteins from intron sequences.

    Trends Genet 2003, 19:115-119. PubMed Abstract | Publisher Full Text OpenURL

  35. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ: Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.

    Nucl Acids Res 1997, 25:3389-3402. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  36. Chen SH, Su SY, Lo CZ, Chen KH, Huang TJ, Kuo BH, Lin CY: PALM: A Paralleled and Integrated Framework for Phylogenetic Inference with Automatic Likelihood Model Selectors.

    PLoS One 2009, 4(12):e8116. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  37. Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignments through sequence weighting, position specific gap penalties and weight matrix choice.

    Nucl Acids Res 1994, 22:4673-4680. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  38. Guindon S, Gascuel O: A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood.

    Syst Biol 2003, 52(5):696-704. PubMed Abstract | Publisher Full Text OpenURL

  39. Abascal F, Zardoya R, Posada D: ProtTest: selection of best-fit models of protein evolution.

    Bioinformatics 2005, 21:2104-2105. PubMed Abstract | Publisher Full Text OpenURL

  40. Ronquist F, Teslenko M, van der Mark P, Ayres D, Darling A, Höhna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP: MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space.

    Syst Biol 2012, 61:539-542. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  41. Larkin M, Blackshields G, Brown N, Chenna R, McGettigan P, McWilliam H, Valentin F, Wallace I, Wilm A, Lopez R, et al.: Clustal W and Clustal X version 2.0.

    Bioinformatics 2007, 23:2947-2948. PubMed Abstract | Publisher Full Text OpenURL

  42. Schäffer A, Aravind L, Madden T, Shavirin S, Spouge J, Wolf Y, Koonin E, Altschul S: Improving the accuracy of PSI-BLAST protein database searches with composition-based statistics and other refinements.

    Nucleic Acids Res 2001, 29:2994-3005. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  43. Petrey D, Xiang Z, Tang CL, Xie L, Gimpelev M, Mitros T, Soto CS, Goldsmith-Fischman S, Kernytsky A, Schlessinger A, Koh IY, Alexov E, Honig B: Using multiple structure alignments, fast model building, and energetic analysis in fold recognition and homology modeling.

    Proteins 2003, 53(S6):430-435. PubMed Abstract | Publisher Full Text OpenURL