Duplication and Diversification of the Spermidine/Spermine N1-acetyltransferase 1 Genes in Zebrafish

Spermidine/spermine N1-acetyltransferase 1 (Ssat1) is a key enzyme in the polyamine interconversion pathway, which maintains polyamine homeostasis. In addition, mammalian Ssat1 is also involved in many physiological and pathological events such as hypoxia, cell migration, and carcinogenesis. Using cross-genomic bioinformatic analysis in 10 deuterostomes, we found that ssat1 only exists in vertebrates. Comparing with mammalian, zebrafish, an evolutionarily distant vertebrate, contains 3 homologous ssat1 genes, named ssat1a, ssat1b, and ssat1c. All zebrafish homologues could be transcribed and produce active enzymes. Despite the long history since their evolutionary diversification, some features of human SSAT1 are conserved and subfunctionalized in the zebrafish family of Ssat1 proteins. The polyamine-dependent protein synthesis was only found in Ssat1b and Ssat1c, not in Ssat1a. Further study indicated that both 5′ and 3′ sequences of ssat1b mediate such kind of translational regulation inside the open reading frame (ORF). The polyamine-dependent protein stabilization was only observed in Ssat1b. The last 70 residues of Ssat1b were crucial for its rapid degradation and polyamine-induced stabilization. It is worth noting that only Ssat1b and Ssat1c, but not the polyamine-insensitive Ssat1a, were able to interact with integrin α9 and Hif-1α. Thus, Ssat1b and Ssat1c might not only be a polyamine metabolic enzyme but also simultaneously respond to polyamine levels and engage in cross-talk with other signaling pathways. Our data revealed some correlations between the sequences and functions of the zebrafish family of Ssat1 proteins, which may provide valuable information for studies of their translational regulatory mechanism, protein stability, and physiological functions.


Introduction
Polyamines, found in the cells of most species, play vital roles in cell proliferation and many physiological functions [1]. Thus, cellular polyamine homeostasis is strictly maintained by regulation of its anabolic and catabolic pathways [2]. In mammals, the interconversion pathway enhances control of cellular polyamine. Spermidine/spermine N 1 -acetyltransferase 1 (Ssat1) is the key enzyme in the rate-determining reactions of this pathway, by which spermine or spermidine accepts the acetyl group from acetyl-CoA to produce N 1 -acetylspermine or N 1 -acetylspermidine [3]. Ssat1 effectively reverses the biosynthetic reactions and alters cellular polyamine equivalence because the acetylated derivatives are readily excreted and are good substrates for acetylpolyamine oxidase (Apao), a peroxisomal enzyme that readily converts these molecules to smaller polyamines [4,5].
Accumulating evidence has indicated that mammalian Ssat1 is also involved various physiological and pathological events, including liver regeneration [6], ischemia-reperfusion injury [7,8,9], pancreatitis [10,11], lipid metabolism [12,13], carcinogenesis [14], cell migration [15], and hypoxia signaling [16], through its ability to modulate polyamine content or by directly interacting with other protein effecters, such as hypoxia inducible factor 1-a (Hif-1a) and integrin a9. Moreover, the regulation of Ssat1 is as versatile as its functions, occurring at multiple levels including transcription [17,18], mRNA processing [19,20,21], translation [22], and protein stabilization [23,24,25]. Curiously, we know rather little about the properties of Ssat1 from nonmammalian species. Here, we found that ssat1 is only present in the vertebrate lineage. Comparing with mammalians, zebrafish, an evolutionarily distant vertebrate, contains not one but three ssat1 genes. To understand whether these ssat1 genes have evolved distinct structural and functional properties, their spatial and temporal expression, translational regulation inside the ORF, protein stability, enzyme kinetics, and protein-protein interactions were extensively investigated. Our data suggest that these zebrafish ssat1 homologues might be paralogous genes which underwent subfunctionalization in their regulatory mechanisms and physiological functions.
The amino acid sequences of ssat1-related genes were aligned by ClustalW [26] and edited manually using the BioEdit software [27] to prepare for the phylogenetic analysis. The alignment is available upon request. Phylogenetic trees were constructed by the neighbor joining (NJ) method using the PHYLIP package [28] with the substitution model and represented with the FigTree software (http://tree.bio.ed.ac.uk/software/figtree/). Tree topology was evaluated by bootstrap resampling with 1000 replicates.

Reverse Transcription (RT)-PCR and Molecular Cloning
Total RNA was isolated from human HEK293T cells (CRL-11268), adult zebrafish, and embryos with TRIzol reagent (Invitrogen). After DNase treatment, total RNA (500 ng) was reverse-transcribed to single-strand cDNA using SuperScript II reverse transcriptase (Clontech) according to the manufacturer's instructions. Transcription of each gene was detected by RT-PCR with specific primers (see Table S1 in the supplemental material). The PCR reaction was initiated with DreamTaq (Fermentas) according to the manufacturer's protocol. The reactions were separated by 1% agarose gel electrophoresis and visualized by staining with ethidium bromide.
To prepare expression plasmids for zebrafish ZF4 (ATCC CRL-2050) or HEK293T cells, the full-length ORF of human SSAT1, zebrafish ssat1a, ssat1b, and ssat1c, or DNA fragments encoding the cytosolic domain of zebrafish integrin 9a and the PAS-B domain of hif-1a were amplified by RT-PCR with specific primer sets (Table S1) and Pfu DNA polymerase (Stratagene). These amplified inserts were ligated into the EcoRI/XhoI sites of pcDNA3.1/myc-His. The recombinant plasmids were transformed into E. coli for preservation and amplification.
The chimeric ssat1a and ssat1b genes were prepared by using the megaprimer PCR technique [29]. We designed 4 chimeric primers, which could pair to the same region of ssat1a and ssat1b at nucleotides 235-248, 319-332, 374-389 and 452-467 (Table S1). To make ssat1a248b, for example, the chimeric primer (primer 11 in Table S1) and ssat1a forward primer (primer 3 in Table S1) were used to amplify the target DNA fragment (nucleotides 1-248 of ssat1a). Then, the 2 strands of newly synthesized PCR fragments were used as megaprimers with Pfu DNA polymerase to synthesize the whole plasmid (pcDNA3.1/ myc-His-ssat1b) and incorporate the chimeric target DNA fragments. The original plasmid DNA was digested by DpnI, and the DpnI-resistant chimeric DNA was recovered by transformation into competent bacteria. The chimeric gene was named ssat1a248b, as it contains nucleotides 1-248 of ssat1a and the remainder of ssat1b. Chimeric genes ssat1a332b, ssat1a374b, ssat1a453b, ssat1b248a, ssat1b332a, ssat1b389a, and ssat1b467a were also prepared in this manner (Fig. 2B). The chimeric gene ssat1aba was prepared by using nucleotides 389-513 of ssat1a as the megaprimer and a plasmid containing ssat1a332b as template. Therefore, ssat1aba and ssat1bab contain most of the ssat1a and ssat1b sequences, but their 332-389 nucleotide regions are from ssat1b and ssat1a.

Preparation of Recombinant Protein and Enzyme Activity Assay
The coding sequence of human SSAT1, zebrafish ssat1 homologues, and the integrin a9 cytosolic, and hif-1a PAS-B domains were amplified (Table S1) from the cDNA of HEK293T cells or zebrafish embryos. PCR products were inserted into the EcoRI and XhoI sites of pET28a or pGEX2T and transformed into E. coli BL21 to express recombinant proteins with N-terminal His 6tag or GST fusions. Following cell lysis, the His-tagged and GSTfused enzymes were purified by affinity chromatography on a HisTrap FF column (GE Healthcare) and a GSTrap FF column (GE Healthcare), respectively. The SSAT1 activity assay was described in our previous work [30].

Cell Culture, Gene Transfection, and Cellular Protein Extraction
Zebrafish ZF4 cells were cultured in 45% DMEM (Gibco), 45% Ham's F12 medium (Gibco), and 10% FBS (Gibco) at 28uC with humidified air/CO 2 (19:1 v/v). Human HEK293T cells (CRL-11268) were cultured in DMEM with 10% FBS at 37uC with humidified air/CO 2 (19:1 v/v). The HEK293T cells were transfected by using standard calcium phosphate precipitation and transfection of ZF4 cells was performed as described previously [30]. To prepare cellular proteins for analysis, cells were collected by centrifugation at 3006g, washed twice with PBS, and extracted with M-PER reagent (Thermo Scientific) according to the manufacturer's instructions. The crude extract was dialyzed against 50 mM Tris-HCl (pH 7.5) with 16 Protease inhibitor (complete protease inhibitor cocktail, Roche) with an Amicon ultra centrifugal filter device (Millipore).

Detection of Protein Expression and Degradation by Western Blotting
HEK293T cells were seeded at a density of 3 6 10 5 /well (6-well plate) and transfected with 2 mg pcDNA3.1/myc-His plasmids containing the ORFs of human SSAT1, zebrafish ssat1a, ssat1b, ssat1c, or chimeric genes. After 24 h culture, cells were treated with 10 mM DENSPM, 5 mM MG132 or vehicle (DMSO), and incubated for another 24 h before harvest and detection of translated proteins. To assess protein stability, cells were transfected with 2 mg plasmid encoding Ssat1a, Ssat1aba, Ssat1a453b, or Ssat1b332a, or with 4 mg plasmid encoding Ssat1b, Ssat1c, Ssat1bab, Ssat1b332a, or Ssat1b467a. After 12 h culture, cells were treated with 200 mM cycloheximide or left untreated. Before harvest, cells were treated with 2 mM spermidine for 2-6 h in the presence of 10 mM MG132 or vehicle (DMSO).
Cell lysates were resolved by 12% SDS-PAGE and transferred to a PVDF membrane. Proteins were immunodetected with antimyc primary (1:2000, Cell Signaling) and anti-mouse IgG secondary antibodies (1:5000, Promega). Signals were detected with ECL Plus chemiluminescence reagent (GE Healthcare) and an imaging system (UVP Biospectrum).

GST Pull-down Assay
HEK293T cells (3610 6 cells/10-cm plate) were transiently transfected with expression vectors encoding myc-tagged Ssat1a, Ssat1b, Ssat1c, or the PAS-B domain of Hif-1a. Cells transfected with Ssat1a, Ssat1b and Ssat1c expression vectors were cultured in the medium with 10 mM DENSPM. After 48 h culture, cell lysates (100 mg) were harvested and mixed with 10 mg GST or GST fusion proteins in 500 ml PBS buffer at 4uC for 2 h, followed by addition of 20 ml of glutathione-Sepharose 4B beads (GE Healthcare). After mixing for 30 min, the beads were washed with PBS. The proteins were eluted in Laemmli sample buffer and analyzed by SDS-PAGE and western blotting.

Identification of ssat1-like Genes in Deuterostomes
Several ssat-like genes were found across the deuterostomia, including sea urchin, sea squirt, amphioxus, mouse, human, and 5 kinds of ray-finned fish. Phylogenetic analysis divided these genes into 3 clusters (Fig. 1). Cluster 1 was composed of human SSAT2, mouse Ssat2, and invertebrate ssat-like genes. Genes in cluster 2 were ssat2 orthologues from ray-finned fish. At least 2 ssat2 homologous genes were found in all ray-finned fish species analyzed in this study. These ssat2 homologues were further divided into 2 sub-groups that suggested an early duplication event at ssat2 in the common ancestor of ray-finned fishes. Cluster 3 included human SSAT1 and its cognate genes from vertebrates. Compared with the first 2 clusters, the ssat1 orthologues were more closely conserved. No ssat1 orthologue was found in invertebrates and only 1 ssat1 was found in most vertebrates except that there were 3 ssat1 homologues in zebrafish. The encoded amino acid sequences ( Fig. 2A) and cDNA sequences (Fig. S2) of zebrafish ssat1 homologous genes were highly similar to each other, and they were clustered together in the phylogenitic analysis ( Fig. 1).
Human SSAT1 is located on the X chromosome between the genes for peroxiredoxin 4 (PRDX4), acryl-CoA thioesterase 9 (ACOT9), and apolipoprotein O (APOO) (Fig. S1). The Ssat1 genes of evolutionarily distant vertebrates including medaka, stickleback, takifugu, and tetraodon are located between acot9 and apoo (data not shown). One of the zebrafish ssat1-like genes (NM_001093748) is also located between acot9 and apoo on chromosome 24; we therefore named it ssat1a. The other zebrafish genes (NM_001030199 and NM_001002169) are closely clustered together and located next to prdx4 on chromosome 5. We named them ssat1b and ssat1c, respectively. The ssat-like genes of invertebrates and ssat2 homologous genes of vertebrates are not grouped like ssat1 (Fig. 1) and their genomic localization also differ (data not shown).

The Expression Pattern of Zebrafish ssat1 Homologous Genes
The expression patterns of zebrafish ssat1 genes were analyzed by RT-PCR. During normal embryogenesis, ssat1c mRNA was the most abundant in every stage and was stably expressed from 12 to 96 hours post fertilization (hpf). The mRNA of ssat1a and ssat1b were not detected until 24 hpf (Fig. 3A, control). A previous study indicated that treatment of human cells with DENSPM, a spermine analog, enhances SSAT1 expression up to 20 fold [31]. Another group of zebrafish embryos were developed with 10 mM DENSPM added immediately after fertilization. All embryos survived and displayed no obvious abnormalities through 96 hpf. Neither the expression nor mRNA abundance of these ssat1 genes was changed (Fig. 3A, DENSPM).
The expression profiles of zebrafish ssat1 genes in the major organs of adult fish were also studied. ssat1a mRNA was mainly expressed in the heart, spleen and kidney, and was also detectable in the brain, eye, liver, and fin. ssat1b mRNA was detectable in the heart, liver, gut, and kidney, and weakly in the brain, eye, gill, spleen, muscle, fin and testis. ssat1c mRNA was detectable in every organ we tested and was the most abundant among these three homologues (Fig. 3B).

The Translational Regulation Inside the ORF of Zebrafish ssat1 Homologues
It has been reported that the translation of human SSAT1 is strictly controlled by polyamine and the SSAT1 ORF region is responsible for such regulation [22,32]. To test this mechanism, the ORF of each gene was ligated into pcDNA3.1/myc-His, which allows the mRNA of each gene to be stably and abundantly expressed. Human SSAT1 protein was extensively expressed in transfected HEK293T cells incubated with DENSPM but not in cells cultured in normal medium or medium with MG132 (Fig. 4A). Addition of MG132, a proteasome inhibitor, may increase SSAT1 protein stability but did not increase protein abundance in the absence of DENSPM. This is similar to previous studies, which suggested that the activity of SSAT1 is mainly regulated by polyamine and polyamine analogs in the translation level [22,23].
We investigated the translational regulation inside the ORF of zebrafish ssat1 under the same experimental conditions. It is interesting to note that background expression of Ssat1a was detectable in cells cultured without DENSPM and its expression increased by approximately 3-fold in the presence of DENSPM. On the other hand, the translational regulation inside the ORF of ssat1b and ssat1c were as stringent as that of human SSAT1 (Fig. 4A). We performed the same experiment in zebrafish ZF4 cells that also obtained the same result (Fig. S3). Thus, the ssat1 translational regulation machinery appears to be conserved in human and fish cells.
In order to identify the key region for translational regulation inside the ORF, a series of chimeric genes (ssat1a248b, ssat1a332b, ssat1a374b and ssat1a453b), which contained the 59 region of ssat1a ORF and the 39 region of ssat1b, were prepared and tested (Fig. 2B). The results indicated that the last 181 nucleotides of ssat1b are important for translational inhibition, since chimeric genes containing more than 181 nucleotides from the 39 region of ssat1b retained the translational regulation pattern of ssat1b, such as ssat1a248b and ssat1a332b (Fig. 4B, lanes 1 and 2). The results of another series of chimeric mRNA with the 59 region of ssat1b and the 39 region of ssat1a (ssat1b248a, ssat1b332a, ssat1b389a and ssat1b467a shown in Fig. 2B) showed that the first 389 nucleotides of ssat1b are also important for regulation (Fig. 4B, lanes 3 and 4). Because the 332,389 nucleotide region of ssat1b is present in both ssat1a332b and ssat1b389a, the importance of this region was further investigated.
The chimera ssat1aba, which contains the 332,389 region of ssat1b and the remainder of ssat1a, still retained the ssat1a regulatory pattern that suggests the 332,389 region of ssat1b alone is not sufficient to inhibit the background protein translation ( Figure 4B). Further, ssat1bab, which contains the 332,374 nucleotide region of ssat1a and the remainder of ssat1b, behaved like ssat1b. These results, which are summarized in table 1, indicate

The Stability of Zebrafish Ssat1 Proteins
By increasing the transfected plasmid by 2-fold and protein loading by 10-fold, the expression of zebrafish Ssat1b and Ssat1c in cells cultured in normal medium could be observed by western blotting. We used these conditions to observe protein stability inside cells. After adding cycloheximine to the culture medium, translation of Ssat1 was stopped and the protein was subsequently turned over. As shown in Figure 4A, the Ssat1b protein was quickly turned over after translation was stopped (Fig. 5A, lanes 3-5). Protein degradation was mediated by proteasome, since the addition of MG132 prevented degradation (Fig. 5A, lane 2). In addition, the presence of spermidine improved the stability of Ssat1b up to 6 hr (Fig. 5A, lanes 6-8). On the other hand, Ssat1a and Ssat1c were more stable and no obvious protein degradation occurred within 6 h (Fig. 5).
The 10 ssat1a and ssat1b chimeric genes were also applied to identify the critical regions responsible for rapid degradation (Fig. 2B). Our results indicated that none of the chimeric proteins, which contain the C-terminal regions of Ssat1a, turned over rapidly without spermidine treatment, such as Ssat1b467a, Ssat1b332a, and Ssat1aba shown here (Fig. 5B, lanes 1-3). Moreover, the chimeric proteins, which contain the C-terminal regions of Ssat1b, did not undergo rapid degradation without spermidine treatment, unless they contain more than 70 residues from the Ssat1b C-terminal region, such as Ssat1a332b (Fig. 5B, lanes 4-6) or Ssat1a248b (data not shown). The results, which are summarized in Table 1, indicated the last 70 residues of Ssat1b are important for the regulation of protein stability.

The Enzyme Activities of Zebrafish Ssat1
The kinetic studies indicated all 3 Ssat1 isoenzymes were bioactive and could use both spermidine and spermine as substrates. However, the substrate preference of these isoenzymes was different. Ssat1b had similar K m values for spermidine and spermine, while Ssat1a had a smaller K m toward spermidine and Ssat1c had a smaller K m for spermine ( Table 2). Ssat1a and Ssatb had a better k cat /K m value for spermidine than that for spermine,  indicating these enzymes were more efficient in spermidine catabolism. In contrast, Ssat1c had a similar k cat /K m ratio for spermidine and spermine.

Protein-protein Interactions of Zebrafish Family of Ssat1 Proteins
The structures of mammalian Ssat1s indicate the homodimer structure is essential for enzyme activity [33,34]. Since Ssat1a, Ssat1b, and Ssat1c are co-expressed in several zebrafish organs and their primary sequences are largely identical, the formation of heterodimers is possible. Here, GST pull-down experiments were applied to test this hypothesis. The GST-Ssat1a fusion protein was able to pull down Ssat1b-myc and Ssat1c-myc while GST itself could not (Fig. 6A). In addition, GST-Ssat1b interacted with Ssat1c-myc, suggesting the 3 zebrafish family of Ssat1 proteins could assemble into homodimers or heterodimers.
It has been reported that human SSAT1 interacts with the cytosolic domain of integrin a9 to enhance cell migration [15]. A recombinant GST fusion to the cytosolic domain of zebrafish integrin a9 (GST-Intg a9) was used to study interactions with zebrafish family of Ssat1 proteins. Ssat1b and Ssat1c interacted with integrin a9, but Ssat1a did not (Fig. 6B).
Human SSAT1 also binds to the PAS-B (Per-ARNT-Sim) domain of HIF-1a, facilitating its degradation [16]. A DNA fragment encoding zebrafish Hif-1a PAS-B was cloned into pcDNA3.1/myc-His and transfected into HEK293T cells. Cell lysates were extracted and incubated with GST-fused Ssat1 proteins. The results of GST pull-down experiments indicated that Ssat1b and Ssat1c interacted with the Hif-1a PAS-B domain but Ssat1a did not (Fig. 6C).

Discussion
Despite reports of several polyamine acetyltransferases in microbes, their sequences were neither similar to each other nor to animal ssat-like genes [35,36,37,38,39], suggesting different evolutionary origins. In this work, 29 ssat homologs from 10 deuterostomia species were identified and analyzed. Comparing with ssat2 genes and invertebrates' ssat-like genes, ssat1 genes are present only in vertebrates and highly conserved to each other (Fig. 1). Previous studies indicated that human SSAT2 and Ssatlike enzymes from several organisms do not involve polyamine catabolism, because these enzymes prefer to use thialysine as their substrate [40,41,42]. Therefore, the polyamine interconversion pathway might only be evolved in the vertebrate lineage.  In humans, SSAT1 mRNA is transcribed in almost every tissue with 2 splicing forms [20]. In addition to the normal mRNA, one transcriptional variant (SSAT-X mRNA), which cannot produce bioactive SSAT1 due to its incorporation of an extra exon from intron 3, is accumulated in hypoxic or iron-deficient cells [20]. When cellular polyamine level increasing, the transcription of SSAT1 is induced [17,18] while the alternative spliced SSAT-X mRNA is reduced [19]. In zebrafish, none of these ssat1 genes was induced by polyamine analog (Fig. 3A) and RT-PCR found no sign of alternative transcript splicing (Fig. 3). Zebrafish thus does not appear to regulate Ssat1 activity through mRNA transcription and alternative splicing.
A previous report indicated that an RNA binding protein, which might bind to the stem-loop structures in the 59 and 39 regions of the human SSAT1 ORF, strictly inhibits SSAT1 mRNA translation [22]. A nucleolin isoform was recently identified as a key factor to stabilize the 59 stem-loop structure of SSAT1 ORF. Addition of polyamine enhances the turnover of nucleolin thus significantly restore translation [43]. In zebrafish, the stringent inhibition of background translation was only observed in ssat1b and ssat1c (Fig. 4A). By assaying a series of chimeric zebrafish ssat1 genes, we found that the both 59 and 39 regions of ssat1b were important for such regulation (Fig. 4B) that seems to be in accordance with the previous study [22]. However, it is difficult to explain that altering the 59 or 39 mRNA structures by replacing the first 332 or the last 124 nucleotides of ssat1b with that of ssat1a did not affect the regulatory pattern of ssat1b (Fig. 4B, a332b and  b389a). Moreover, attempts to search similar stem-loop structures of human SSAT1 in zebrafish ssat1b and ssat1c were also failed.
A recent study of yeast antizyme (Az) might provide another explanation. Like Ssat1, Az is also a homeostatic feedback regulator of polyamines, since the activity of ornithine decarboxylase, a key enzyme in polyamine biosynthesis pathway, is readily down-regulated in the presence of Az [2]. It is interesting to note that the mRNA translation of Az is regulated in response to the cellular polyamine level, too. A unique polyamine-dependent +1 ribosomal frameshifting regulatory mechanism involving the 59 region of Az mRNA is conserved in many Az orthologues [44]. In addition, Kurian et al. recently found another regulatory mechanism of Az synthesis involving the 39 region of its mRNA. In fact, it is the encoded amino acid sequence responsible for such kind of regulation, since the nascent Az will act as a sensor of polyamine and regulate the completion of its own translation [45]. Until now, it has been unclear whether nascent Ssat1 serves as a polyamine sensor and regulates its own translation. However, it was known that the amino acid residues in the C-terminus of human SSAT1 are also important for translational regulation [23]. Deleting more than 5 residues from the C-terminus has a significant impact, but silent mutations, in which several nucleotides in the 39 end were modified without changing the coding sequence, retain polyamine responsiveness [22]. Nevertheless, further study is needed to test this hypothesis.
The fourth regulatory mechanism of human SSAT1 is modulation of protein stability. Binding of polyamines or polyamine analogs changes the configuration of the SSAT1 protein. Therefore, SSAT1 is prevented from ubiquitination and becomes stable [23,25]. Extensive effort has led to the identification of key residues responsible for this regulation [24]. The MATEE sequence in the C-terminus is important. SSAT1 stabilizes in the absence of polyamine when M 167 , E 170 , and E 171 are substituted or the last 2 residues (E 170 and E 171 ) are deleted. SSAT1, however, carries on rapid degradation even in the presence of polyamine when point mutations are made at R 7 , C 14 , R 19 , H 126 , K 141 , E 152 , or R 155 . These residues in human SSAT1 are largely conserved in zebrafish Ssat1b, except the MATEE sequence is replaced by MESEE in the C-terminus ( Fig. 2A). Note that M 167 , E 170 , and E 171 are conserved in zebrafish Ssat1a and Ssat1b, but Ssat1a does not turn over rapidly (Fig. 5A). The chimeric enzyme Ssat1a453b, which preserved the last 20 residues of Ssat1b, was as stable as Ssat1a, suggesting more residues may be involved in this mechanism. The results of Ssat1a332b indicated that the last 70 residues of Ssat1b were important for its rapid degradation (Fig. 5B). Only 14 residues differ in the last 70 amino acids of Ssat1a and Ssat1b. Further study is needed to identify the residues responsible for regulating protein stability. All zebrafish family of Ssat1 proteins were active enzymes, though their substrate preferences and catalytic efficiencies differed. Moreover, our data also indicated these zebrafish Ssat1 proteins form heterodimers (Fig. 6A). Their kinetics may be more complex due to heterodimerization. Considering the differences of these Ssat1 proteins in their expression profile, regulation patterns and enzyme activities, zebrafish might be able to fine-tune the metabolism of polyamine to fit the physiological requirements in different organs by expressing different Ssat1 proteins.
Integrins are cell surface proteins that mediate cell-cell communication and cell morphology. Integrin a9, a mammalian specific form [46], is stimulated by extracellular signals, such as tenascin C [47], osteopontin [48], and vascular cell adhesion molecules-1 [49], and involved in embryogenesis [50], lymphangiogenesis, and wound healing [51]. It has been reported that overexpression of human SSAT1 enhances cell migration mediated by integrin a9 [15]. The first 20 amino acids of SSAT1 is crucial, since they could bind to the cytosolic domain of integrin a9 thus regulates the migration signaling [52]. In this study, we identified an integrin a9 orthologue in zebrafish. Although the length of the extracellular region differs significantly between human and zebrafish integrin a9, the sequences of their cytosolic domains are largely identical (Fig. S4). By using GST-pull down experiments, we confirmed that zebrafish integrin a9 interacts with Ssat1b and Ssat1c, but not Ssat1a (Fig. 6B). It is worth noting that Ser 15 of Ssat1b, Ssat1c, and human SSAT1 was replaced by Pro 15 in Ssat1a. The structures of SSAT1 reveal a conserved ahelix located between residues 12 and 28 ( Fig. 2A) [33]. Pro 15 in Ssat1a may break the helix structure and thus interfere with the interaction between Ssat1a and integrin a9.
Hif-1a, a key regulator of oxygen homeostasis in all metazoans, is mainly regulated by an oxygen-sensing prolyl hydroxylase, which facilitates its rapidly degradation in proteosome [53]. A previous study has shown another oxygen-independent Hif-1a regulation mechanism that is triggered by the binding of human SSAT1 with the PAS-B domain of HIF-1a [16]. PAS domains, found in many proteins in all kingdoms of life, are structurally conserved protein-protein interaction modules [54]. Although the amino acid sequences in PAS-B domains of human and zebrafish Hif-1a are highly conserved (Fig. S5), it is interesting to note that only Ssat1b and Ssat1c, but not Ssat1a, were able to interact with the PAS-B domain of zebrafish Hif-1a (Fig. 6C). The sequence variants between these homologues may provide clues to identify the critical regions responsible for Hif-1a binding in the future.
Gene duplication is considered to be the major force of evolution [55], because new copies may acquire new functions by mutation (known as neofunctionalization) [56]. However, the fates of redundant genes might also include becoming pseudogenes (nonfunctionalization) or being preserved in a complementary partitioning of subfunctions (subfunctionalization) [56]. It is generally believed that 2 rounds of whole-genome duplication occurred during the intergradation of vertebrates from their deuterostome ancestors [57]. Interestingly, we noticed that not only ssat1 but also hif-1a [53] and integrin a9 [46] were evolved simultaneously in the vertebrate lineage. They might experience neofunctionalization to meet the physiological requirements of vertebrates.
In comparison with mammals, the ray-finned fishes underwent an extra round of whole-genome duplication, which caused the teleost radiation [58]. It seems that nonfunctionalization is the fate of the majority of duplicated ssat1 genes, as is the case in medaka, stickleback, takifugu, and tetraodon. Thus there is only one ssat1 left in these species. However, zebrafish, which contains not one but three ssat1 homologues, is an exception. The phylogenetic analysis indicates that all zebrafish ssat1 homologues are derived from a common ancestor (Fig. 1). The remnants of vertebrates' ssat1 syntenic genes are found scattering in the loci of each homologue. For example, acot9 and apoo are clustered with ssat1a in chromosome 24, and prdx4 is closely located near ssat1b and ssat1c in chromosome 5 (Fig. S1). It is worth to note that the sequences between ssat1b and ssat1c are more similar to each other (Fig. S2). Further, Ssat1b and Ssat1c also have similar translational regulation pattern and protein-protein interaction relationships with Hif-1a and Integrin a9. These observations suggest that ssat1a might be one of the products from the teleosts' whole genome duplication while the other one underwent a local duplication to form ssat1b and ssat1c later on.
Our results suggest that zebrafish ssat1 homologues are paralogous genes which experienced subfunctionalization in their function and regulation. It is worth noting that only Ssat1b and Ssat1c, but not the polyamine-insensitive Ssat1a, are able to interact with integrin a9 and Hif-1a. Thus these signal pathways could be regulated by Ssat1 in response to cellular polyamine level. Besides polyamine catabolism, it might be the key feature that allows Ssat1 to coordinate certain physiological responses in vertebrates, such as fine-tuning the advanced immune system and the homeostasis of polyamine and hypoxia.
By characterizing properties of zebrafish family of Ssat1 proteins and the artificial chimeric enzymes, our data revealed some correlations between their sequences and functions that may provide valuable information for studies of the translational regulatory mechanism, protein stability, and physiological functions of Ssat1 in the future.