Novel Glutamate–Putrescine Ligase Activity in Haloferax mediterranei: A New Function for glnA-2 Gene

The genome of the halophilic archaea Haloferax mediterranei contains three ORFs that show homology with glutamine synthetase (GS) (glnA-1, glnA-2, and glnA-3). Previous studies have focused on the role of GlnA-1, suggesting that proteins GlnA-2 and GlnA-3 could play a different role to that of GS. Glutamine synthetase (EC 6.3.1.2) belongs to the class of ligases, including 20 subclasses of other different enzymes, such as aspartate–ammonia ligase (EC 6.3.1.1), glutamate–ethylamine ligase (EC 6.3.1.6), and glutamate–putrescine ligase (EC 6.3.1.11). The reaction catalyzed by glutamate–putrescine ligase is comparable to the reaction catalyzed by glutamine synthetase (GS). Both enzymes can bind a glutamate molecule to an amino group: ammonium (GS) or putrescine (glutamate–putrescine ligase). In addition, they present the characteristic catalytic domain of GS, showing significant similarities in their structure. Although these proteins are annotated as GS, the bioinformatics and experimental results obtained in this work indicate that the GlnA-2 protein (HFX_1688) is a glutamate–putrescine ligase, involved in polyamine catabolism. The most significant results are those related to glutamate–putrescine ligase’s activity and the analysis of the transcriptional and translational expression of the glnA-2 gene in the presence of different nitrogen sources. This work confirms a new metabolic pathway in the Archaea domain which extends the knowledge regarding the utilization of alternative nitrogen sources in this domain.


Introduction
Glutamine synthetase (GS; EC 6.3.1.2), which belongs to the class of ligases, can form carbon-nitrogen bonds using ATP. This class includes 20 different subclasses of enzymes, including GS, aspartate-ammonia ligase (EC 6.3.1.1), glutamate-ethylamine ligase (EC 6.3.1.6), and glutamate-putrescine ligase (EC 6.3.1.11). GS is an essential enzyme in nitrogen metabolism that catalyzes the synthesis of glutamine from glutamate and ammonium. Glutamine biosynthesis occurs through the biosynthetic reaction that first involves the formation of γ-glutamyl-phosphate from ATP and glutamate and, later, the release of phosphate, resulting in L-glutamine [1]. GS is a metalloenzyme dependent on ATP and divalent metal ions such as magnesium (Mg 2+ ) or manganese (Mn 2+ ), obtaining greater effectiveness via its catalytic activity in vitro in the presence of Mn 2+ [2][3][4]. GS acts with the glutamate synthase enzyme (GOGAT; EC 1.4.7.1), catalyzing the reductive transfer of the amide group from L-glutamine to 2-oxoglutarate. This reaction, which is dependent on reducing power, generates two molecules of L-glutamate, one of which is recycled as a substrate for the GS reaction, while the other is exported or used to produce other amino acids [5]. GS genes show homology in different organisms, even in distant inosa has an extended polyamine degradative pathway that involves seven γ-glutamylpolyamine synthetases (PauA1-7), which are specific for the different polyamines, monoamines, or other substrates [31]. The distribution of polyamines among the different groups of Archaea is characteristic of each of them. Hyperthermophilic, acidophilic, and thermoacidophilic archaea contain a significant diversity of linear polyamines, whereas methanogenic archaea contain homospermidine, putrescine, and more commonly, spermidine [33][34][35]. Some of the first observations on polyamines related to archaea determined that halobacteria lacked polyamines [36,37].
This work shows that Hfx. mediterranei is able to grow in the presence of alternative nitrogen sources, such as putrescine, as the only source of nitrogen or carbon. In addition, this study provides new insight about the GlnA-2 role, which exhibits a novel glutamateputrescine ligase activity instead of GS activity, and represents the first time that this activity has been detected in Archaea domain.

Bioinformatic Analysis
For the selection of the amino acid sequences, three independent alignments for each Hfx. mediterranei glutamine synthetase-GlnA-1 (HFX_0245), GlnA-2 (HFX_01688), and GlnA-3 (HFX_01686)-were carried out by BLASTP [38] against the NCBI database of "nonredundant protein sequences (nr)". Sequences with a percent identity greater than 95% between them were eliminated. Phylogenetic inference was performed using the maximum likelihood method and the Le Gascuel model [39]. Finally, the phylogenetic tree with the highest logarithmic probability value was selected. A total of 500 re-samples were used by bootstrapping [40]. The initial tree for the heuristic search was automatically obtained using the Neighbor-Join (NJ) algorithm from a matrix of pairwise distances estimated using the JTT model and, subsequently, by selecting the topology corresponding to the highest log-likelihood value.
For the identification of conserved residues and domains present in the GlnA proteins from Hfx. mediterranei, the alignment against GS model protein of Salmonella typhimurium was carried out. Furthermore, an in silico analysis focused on the identification of conserved domains in the GS was performed using different tools: HMMER [41] and Prosite Scan [42].

Strains and Culture Conditions
Hfx. mediterranei R4 (ATCC 33500T) was grown at 42 • C with aeration at 250 rpm. The culture media contained 25% (w/v) seawater [43] with different nitrogen sources and, in the absence of a nitrogen source (Table 1), was supplemented with 5 g/L glucose, 0.0005 g/L FeCl 3 , and 0.5 g/L KH 2 PO 4 . The pH value was adjusted to 7.3 with NaOH. For the preparation of the medium in the absence of nitrogen, cells from a culture with nitrogen source were harvested by centrifugation at 13,000 rpm for 30 min, washed with 25% seawater, and then transferred to a medium without a nitrogen source to induce the nitrogen starvation. Cells were subjected to nitrogen starvation for 120 h.
E. coli strains DH5α for cloning and JM110 for efficient transformation of Hfx. mediterranei were grown overnight in Luria-Bertani medium with ampicillin (100 mg/mL) at 37 • C.

Identification and Expression of glnA-1 and glnA-2 Promoter Regions
The identification of the possible TATA boxes, BRE sequences, and transcription initiation sites for each of the genes was carried out using a combination of the different bioinformatic tools for bacteria and eukaryotes. In addition, a manual search was performed based on consensus sequences for TATA boxes and halophilic BRE sequences [44][45][46][47]. From Hfx. mediterranei R4 genomic DNA, the promoter regions identified of the glnA-1 and glnA-2 genes were amplified with the specific primers for each of these areas, including the cut-off points of the restriction enzymes HindIII and NcoI (Table S1). The cloning vector pGEM-T Easy (Promega, Barcelona, Spain) and the halophilic vector pVA513 (kindly provided by Dr Mike Dyall-Smith (University of Melbourne, Australia)) were used, in addition to the chemically competent cells E. coli DH5α and JM110 (Promega, Barcelona, Spain). The halophilic vector pVA513 has origins of replication for both E. coli (pBR322) and Haloferax sp. (pHK2), an ampicillin resistance gene to work in E. coli (Amp R ), a novobiocin resistance gene to work in Haloferax sp. (Nov R ), and HindIII and NcoI restriction enzymes cleavage targets followed by the Haloferax lucentense β-galactosidase gene. The Hfx. mediterranei transformants with the constructions pVA513-p-glnA-1 and pVA513-p-glnA-2 were grown in Hm-CM, in Hm-DM with 40 mM ammonium or with 40 mM nitrate as the nitrogen source and, in Hm-NS. All media were supplemented with 0.3 µg/mL novobiocin. The cultures were carried out in triplicate, and growth was monitored by measuring the OD 600nm throughout the entire growth period. The characterization of the promoters was carried out by measuring the β-galactosidase activity and the cell-free extracts were processed as described in previous work [48,49]. All measurements were made in triplicate. One unit of β-galactosidase activity was defined as the amount of enzyme that catalyzes the hydrolysis of 1 µmol of ONPG min −1 .

Gene Expression Analysis by Reverse Transcription PCR
RNA was isolated from Hfx. mediterranei R4 strain in Hm-CM cultures and in Hm-DM in the presence of two different nitrogen sources, 40 mM nitrate or 40 mM ammonium, in the middle of the exponential phase and the stationary phase. RNA was isolated after nitrogen starvation for 48, 96, and 120 h from the Hfx. mediterranei R4 strain in Hm-NS cultures. Total RNA isolation, quality, and quantity were analyzed as described in previous work [50]. Four independent biological replicates of each condition were performed. For cDNA synthesis, an RNA sample (0.5-0.6 µg) and M-MuLV Reverse Transcriptase (Thermo Scientific, Waltham, MA, USA) were used. Negative controls were performed without enzyme or RNA. The RT-PCR protocol was performed according to the manufacturer's instructions. The oligonucleotides used to perform the RT-PCR were designed based on the glnA-1 and glnA-2 genes (Table S1).

Proteins Expression Analysis by Western Blotting and Molecular Weight Determination
Western blotting was performed as described in the Western blotting principles and methods manual (GE Healthcare, Chicago, IL, USA) using 30 µg of protein extracts, obtained from Hfx. mediterranei R4 cultures in Hm-DM in the presence of different nitrogen sources (40 mM nitrate, 40 mM ammonium, or 5-80 mM putrescine). Protein extracts obtained from Hfx. mediterranei R4 cultures in the absence of a nitrogen source (Hm-NS) for 48, 96, and 120 h were also used.
Harvested cells were resuspended to 40% (w/v) in 20 mM Tris-HCl, 2 M NaCl, and 10 mM MgCl 2 buffer pH 7.5. As a primary antibody (polyclonal rabbit anti-GlnA antibody), two synthetic peptides (GenScript, NJ, USA) were used at a concentration of 0.5 µg/mL, and a secondary antibody was labeled with peroxidase 1:50,000 (Thermo Scientific, Waltham, MA, USA), which uses luminol as a chemiluminescent substrate (GE Healthcare, Chicago, IL, USA). Recombinant GlnA proteins obtained by heterologous expression in E. coli were used as the positive control [14,51]. Moreover, recombinant GlnA-2 quaternary structure verification was analyzed by gel filtration chromatography using a HiPrep 16/60 Sephacryl S-300 High-Resolution column (GE Healthcare, Chicago, IL, USA) previously equilibrated with 20 mM Tris-HCl pH 7.0, 2 M NaCl. Standard proteins for gel filtration chromatography from 12.5 to 678 kDa were used as markers to estimate the protein molecular mass. Protein was eluted in the presence of salt, at a flow rate of 0.5 mL/min in 20 mM Tris-HCl, 2 M NaCl and 10 mM MgCl 2 buffer pH 7.5. All of the purification steps were carried out at room temperature.

Assays for GlnA Protein: Enzymatic Activity
GS and glutamate-putrescine ligase activities were measured using the method described by Shapiro and Stadtman [52]. In both reactions, inorganic phosphate is produced from the consumption of ATP. The schematic GS biosynthetic reaction to produce glutamine is shown in Scheme 1a, and the glutamate-putrescine ligase reaction to produce γ-glutamylputrescine is shown in Scheme 1b.
after nitrogen starvation for 48, 96, and 120 h from the Hfx. mediterranei R4 strain in Hm-NS cultures. Total RNA isolation, quality, and quantity were analyzed as described in previous work [50]. Four independent biological replicates of each condition were performed. For cDNA synthesis, an RNA sample (0.5-0.6 μg) and M-MuLV Reverse Transcriptase (Thermo Scientific, Waltham, Massachusetts, United States) were used. Negative controls were performed without enzyme or RNA. The RT-PCR protocol was performed according to the manufacturer's instructions. The oligonucleotides used to perform the RT-PCR were designed based on the glnA-1 and glnA-2 genes (Table S1).

Proteins Expression Analysis by Western Blotting and Molecular Weight Determination
Western blotting was performed as described in the Western blotting principles and methods manual (GE Healthcare, Chicago, IL, USA) using 30 μg of protein extracts, obtained from Hfx. mediterranei R4 cultures in Hm-DM in the presence of different nitrogen sources (40 mM nitrate, 40 mM ammonium, or 5-80 mM putrescine). Protein extracts obtained from Hfx. mediterranei R4 cultures in the absence of a nitrogen source (Hm-NS) for 48, 96, and 120 h were also used.
Harvested cells were resuspended to 40% (w/v) in 20 mM Tris-HCl, 2 M NaCl, and 10 mM MgCl2 buffer pH 7.5. As a primary antibody (polyclonal rabbit anti-GlnA antibody), two synthetic peptides (GenScript, NJ, USA) were used at a concentration of 0.5 μg/mL, and a secondary antibody was labeled with peroxidase 1:50,000 (Thermo Scientific, Waltham, MA, USA), which uses luminol as a chemiluminescent substrate (GE Healthcare, Chicago, IL, USA). Recombinant GlnA proteins obtained by heterologous expression in E. coli were used as the positive control [14,51]. Moreover, recombinant GlnA-2 quaternary structure verification was analyzed by gel filtration chromatography using a HiPrep 16/60 Sephacryl S-300 High-Resolution column (GE Healthcare, Chicago, IL, USA) previously equilibrated with 20 mM Tris-HCl pH 7.0, 2 M NaCl. Standard proteins for gel filtration chromatography from 12.5 to 678 kDa were used as markers to estimate the protein molecular mass. Protein was eluted in the presence of salt, at a flow rate of 0.5 mL/min in 20 mM Tris-HCl, 2 M NaCl and 10 mM MgCl2 buffer pH 7.5. All of the purification steps were carried out at room temperature.

Assays for GlnA Protein: Enzymatic Activity
GS and glutamate-putrescine ligase activities were measured using the method described by Shapiro and Stadtman [52]. In both reactions, inorganic phosphate is produced from the consumption of ATP. The schematic GS biosynthetic reaction to produce glutamine is shown in Scheme 1a, and the glutamate-putrescine ligase reaction to produce γglutamylputrescine is shown in Scheme 1b. The reaction mixtures were incubated at 42 °C for 15 min. The final product was detected at 660 nm. This assay was performed with both recombinant GlnA proteins and cell extracts from Hfx. mediterranei R4 growth in the presence of different nitrogen sources. All tests were carried out in triplicate. Negative control assays were also performed in parallel with the absence of the enzyme or the absence of ATP. One unit of GS activity was defined as the amount of protein that produces 1 μmol of phosphate/min. The reaction mixtures were incubated at 42 • C for 15 min. The final product was detected at 660 nm. This assay was performed with both recombinant GlnA proteins and cell extracts from Hfx. mediterranei R4 growth in the presence of different nitrogen sources. All tests were carried out in triplicate. Negative control assays were also performed in parallel with the absence of the enzyme or the absence of ATP. One unit of GS activity was defined as the amount of protein that produces 1 µmol of phosphate/min.

Construction and Characterization of the HM26-∆glnA-2 Mutants
In-frame deletion mutants of the glnA-2 gene (HM26-∆glnA-2) were obtained from the parental strain Hfx. mediterranei HM26 (R4-∆pyrE2) using the pop-in/pop-out method as described previously for Hfx. mediterranei [15,20,53]. For the characterization of the HM26-∆glnA-2 strain, the growth and stability of the mutant in Hm-CM and Hm-DM in the presence of different nitrogen sources, i.e., 5-80 mM nitrate, 5-80 mM ammonium, or 5-80 mM putrescine, were studied. The statistical analysis of the growth parameters of the HM26-∆glnA-2 strain was carried out compared to the parental strain Hfx. mediterranei HM26 (R4-∆pyrE2) obtained in previous work [15] under the same conditions. Three biological replicates were performed for each strain and culture medium. The stability of glnA-2 deletion during growth on different culture media was determined by PCR screening, Southern blotting, and Western blotting.

Bioinformatic Analysis of the GS Proteins
The phylogenetic analysis revealed that at the origin of the obtained phylogenetic tree (Figure 1), there is a bifurcation differentiating two central nodes. In the first node are all the GlnAs from species belonging to the Archaea domain, with the exception of two bacterial species (S. coelicolor and Streptomyces luteus). Two other well-differentiated secondary nodes emerge from the first node. Most of the proteins found in the different branches of the first secondary node are species from the Haloferacaceae family, which includes the GlnA-1 protein from Hfx. mediterranei. Most of the proteins found in the other secondary node are species from the Halobacteriaceae family and other representatives of the Haloferacaceae, Natrialbaceae, and Halorubraceae families (according to abundance order). This node contains the GlnA-2 and GlnA-3 proteins from Hfx. mediterranei, which are phylogenetically closer to each other than GlnA-1. Several GlnA proteins, which are present in the same organism, are located at different nodes, as occurs with the GlnA proteins of Hfx. mediterranei, where GlnA-1 is located at a different node than GlnA-2 and GlnA-3. These results suggest an ancient duplication of Hfx. mediterranei GlnA proteins and subsequent diversification of the paralogs on the two main branches of the tree.
To elucidate whether the glnA genes encode functional GS, the amino acid sequences were aligned against the sequence of the S. typhimurium GlnA, determining the presence or absence of the amino acid residues characteristic of the active site of GS. The sequence of S. typhimurium GlnA (P0A1P6) was established as a reference model, given that its structure is determined by X-ray crystallography. It shows 37.6% and 28.0% identity with GlnA-1 and GlnA-2, respectively. The alignment results show that GlnA-1 contains 18 conserved residues, whereas GlnA-2 contains 10 of these 18 conserved residues ( Figure 2).
There are 35 and 187 conserved amino acids in the GS of most species and prokaryotes, respectively. GlnA-1 contains 28 (80.0%) conserved residues and GlnA-2 contains 21 (60.0%), whereas 84 (44.9%) and 48 (25.7%) of the residues present in prokaryotes are conserved in GlnA-1 and GlnA-2, respectively. The residue substitutions could be due to the mechanisms of adaptation and divergence processes. The adenylylation box is composed of eight universally conserved residues (KNKPDKLY). GlnA-1 presents five residues of the adenylylation box, including the tyrosine residue (Y 397 ), whereas the Hfx. mediterranei GlnA-2 protein only has three residues, lacking the adenylylation residue (Y 397 ), which is replaced by the C 375 residue.
The conserved residues and their relative positions (numbering refers to S. typhimurium GlnA) present in each GlnA protein of Hfx. mediterranei are summarized in Table 2, confirming that GlnA-2 does not present essential residues for glutamine synthesis (E 327 and D 50 ). Furthermore, the Y 179 residue essential for binding to ammonium is replaced by alanine, providing significantly more space for binding with a bulky substrate. These results suggest that GlnA-1 protein from Hfx. mediterranei corresponds to functional GS as previously described [20], whereas GlnA-2 may be involved in other functions. By comparison, Hfx. mediterranei GlnA-1 and GlnA-2 protein sequence analysis by InterProScan1 predicted a beta-C-terminal catalytic domain corresponding to the characteristic GS enzymatic domain (IPR008146). This domain is present in many proteins of the GS family. To elucidate whether the glnA genes encode functional GS, the amino acid sequences were aligned against the sequence of the S. typhimurium GlnA, determining the presence or absence of the amino acid residues characteristic of the active site of GS. The sequence of S. typhimurium GlnA (P0A1P6) was established as a reference model, given that its structure is determined by X-ray crystallography. It shows 37.6% and 28.0% identity with GlnA-1 and GlnA-2, respectively. The alignment results show that GlnA-1 contains 18 conserved residues, whereas GlnA-2 contains 10 of these 18 conserved residues ( Figure  2). There are 35 and 187 conserved amino acids in the GS of most species and prokaryotes, respectively. GlnA-1 contains 28 (80.0%) conserved residues and GlnA-2 contains 21 (60.0%), whereas 84 (44.9%) and 48 (25.7%) of the residues present in prokaryotes are conserved in GlnA-1 and GlnA-2, respectively. The residue substitutions could be due to the mechanisms of adaptation and divergence processes. The adenylylation box is composed of eight universally conserved residues (KNKPDKLY). GlnA-1 presents five residues of the adenylylation box, including the tyrosine residue (Y397), whereas the Hfx. mediterranei GlnA-2 protein only has three residues, lacking the adenylylation residue (Y397), which is replaced by the C375 residue.
The conserved residues and their relative positions (numbering refers to S. typhimurium GlnA) present in each GlnA protein of Hfx. mediterranei are summarized in Table  2, confirming that GlnA-2 does not present essential residues for glutamine synthesis (E327 and D50). Furthermore, the Y179 residue essential for binding to ammonium is replaced by alanine, providing significantly more space for binding with a bulky substrate. These results suggest that GlnA-1 protein from Hfx. mediterranei corresponds to functional GS as previously described [20], whereas GlnA-2 may be involved in other functions. By comparison, Hfx. mediterranei GlnA-1 and GlnA-2 protein sequence analysis by InterProScan1

Identification and Expression of glnA-1 and glnA-2 Promoter Regions
The genomic organization analysis determined that the glnA-1 gene (Scheme 2a) appears in the genome located downstream of the Lrp/AsnC transcriptional regulator [49] and upstream of SAM-methyltransferase genes. The glnA-2 gene (Scheme 2b) is located between two transporter genes (downstream of the MATE transporter and upstream of the ABC transporter family), and close to the glnA-3 gene; both are distant from the glnA-1 gene.

Identification and Expression of glnA-1 and glnA-2 Promoter Regions
The genomic organization analysis determined that the glnA-1 gene (Scheme 2a) appears in the genome located downstream of the Lrp/AsnC transcriptional regulator [49] and upstream of SAM-methyltransferase genes. The glnA-2 gene (Scheme 2b) is located between two transporter genes (downstream of the MATE transporter and upstream of the ABC transporter family), and close to the glnA-3 gene; both are distant from the glnA-1 gene. The promoter regions of the glnA-1 and glnA-2 genes were identified by bioinformatic analysis. The selected regions contain the possible TATA box and the BRE sequence, in addition to the transcriptional start site and the start of the gene ( Figure S1). These pro- The promoter regions of the glnA-1 and glnA-2 genes were identified by bioinformatic analysis. The selected regions contain the possible TATA box and the BRE sequence, in addition to the transcriptional start site and the start of the gene ( Figure S1). These promoter regions were cloned in the halophilic vector pVA513. The Hfx. mediterranei transformants with the pVA513-p-glnA-1 ( Figure S2) and pVA513-p-glnA-2 ( Figure S3) constructions were characterized. The growth of these transformants was similar to those described previously. The characterization of the promoter region of the glnA genes was carried out by measuring specific β-galactosidase activity using different cultures of Hfx mediterranei growth under several nitrogen sources. β-galactosidase activity was detected in all media analyzed at the different growth times (Figures S2 and S3).
In Hm-CM and Hm-DM with 40 mM ammonia or nitrate, the transformants with the pVA513-p-glnA-1 construction showed the maximum specific β-galactosidase activity in the middle of the exponential phase. From that point, the activity of β-galactosidase decreased, although it was detected until reaching the stationary phase. In the absence of nitrogen (Hm-NS), as expected, the transformants with the pVA513-p-glnA-1 construction showed higher β-galactosidase activities than in the presence of ammonium as a nitrogen source. Under nitrogen starvation conditions, the β-galactosidase activity was maintained for up to 120 h, obtaining the best values of specific activity at 24 h of nitrogen deficiency (Table 3, Figure S2).
The transformants with the pVA513-p-glnA-2 construction ( Figure S3) showed βgalactosidase activity values that were very similar in all the conditions tested independently of the nitrogen source. Maximum activity values were detected in the middle of the exponential growth phase in all the conditions. Under nitrogen starvation conditions, β-galactosidase activity values did not show changes. Regardless of the number of hours of starvation of the nitrogen source, these values were between 2.6 ± 0.3 and 2.9 ± 0.2 U/mg. Furthermore, in this condition, the β-galactosidase activity values were maintained up to 120 h. These results suggest that the glnA-1 expression appears to increase when the ammonium concentration decreases, whereas the glnA-2 expression appears to be similar in all of the conditions analyzed.
The β-galactosidase activity was also analyzed in the presence of 40 mM putrescine ( Figure S4). Surprisingly, Hfx. mediterranei can grow in Hm-DM with 40 mM putrescine as a nitrogen source. Under this culture condition, the transformants with the pVA513-p-glnA-1 and pVA513-p-glnA-2 constructions showed β-galactosidase activity at different growth phases. The maximum activity values were detected in the middle of the exponential growth phase in this condition (Table 3).

Gene Expression Analysis by Reverse Transcriptase PCR
The expression of glnA-1 and glnA-2 genes from Hfx. mediterranei was analyzed by RT-PCR using RNA isolated from cultures grown in the presence of different nitrogen sources.
The results obtained in this analysis show that the glnA-1 and glnA-2 genes are expressed in Hm-CM and in Hm-DM with different nitrogen sources (40 mM ammonium or 40 mM nitrate) at different growth phases (in the middle of the exponential phase and the stationary phase). As expected, under nitrogen starvation, glnA-1 and glnA-2 genes are expressed independently of the nitrogen deficiency time (48,96, and 120 h). The results showed that both genes are constitutively expressed in all conditions tested ( Figure 3) and are in agreement with those obtained in the characterization of the promoter regions, in which β-galactosidase activity was obtained in all of the analyzed media at different growth phases.

Gene Expression Analysis by Reverse Transcriptase PCR
The expression of glnA-1 and glnA-2 genes from Hfx. mediterranei was analyzed by RT-PCR using RNA isolated from cultures grown in the presence of different nitrogen sources.
The results obtained in this analysis show that the glnA-1 and glnA-2 genes are expressed in Hm-CM and in Hm-DM with different nitrogen sources (40 mM ammonium or 40 mM nitrate) at different growth phases (in the middle of the exponential phase and the stationary phase). As expected, under nitrogen starvation, glnA-1 and glnA-2 genes are expressed independently of the nitrogen deficiency time (48,96, and 120 h). The results showed that both genes are constitutively expressed in all conditions tested ( Figure 3) and are in agreement with those obtained in the characterization of the promoter regions, in which β-galactosidase activity was obtained in all of the analyzed media at different growth phases.

Proteins Expression Analysis by Western blotting and Molecular Weight Determination
Protein expression analysis was performed using Hfx. mediterranei R4 protein extracts from cultures grown under different conditions to analyze the GlnA-1 and GlnA-2 translational expression profiles depending on nitrogen source.
The Western blotting results show that GlnA-1 is expressed in Hm-CM independently of the growth phase. In Hm-DM, with 40 mM ammonium or 40 mM nitrate, the GlnA-1 protein also showed expression at the same growth phase (at initial exponential phase, at middle exponential phase, and in stationary phase). The GlnA-1 protein expression was detected in Hm-NS at 48, 96, and 120 h of nitrogen starvation (Figure 4).

Proteins Expression Analysis by Western blotting and Molecular Weight Determination
Protein expression analysis was performed using Hfx. mediterranei R4 protein extracts from cultures grown under different conditions to analyze the GlnA-1 and GlnA-2 translational expression profiles depending on nitrogen source.
The Western blotting results show that GlnA-1 is expressed in Hm-CM independently of the growth phase. In Hm-DM, with 40 mM ammonium or 40 mM nitrate, the GlnA-1 protein also showed expression at the same growth phase (at initial exponential phase, at middle exponential phase, and in stationary phase). The GlnA-1 protein expression was detected in Hm-NS at 48, 96, and 120 h of nitrogen starvation (Figure 4). Curiously, Hfx. mediterranei grew in Hm-DM with putrescine as the only sour nitrogen or carbon. These results indicate that Hfx. mediterranei can grow with polyam However, the maximum OD600nm values reached were lower than those obtained i presence of other nitrogen or carbon sources ( Figure S5).
Western blotting analysis carried out using Hfx. mediterranei protein extracts, i presence of different concentrations of putrescine (20-80 mM) as a nitrogen source, cate that the GlnA-1 and GlnA-2 proteins also showed expression in the middle o exponential growth phase ( Figure 5). In Hm-CM, in the presence of 40 mM putresci the same growth phase, the expression of the GlnA-1 and GlnA-2 proteins could al detected. Consequently, these results are an excellent starting point to determine the ence of a new pathway for putrescine degradation in Haloarchaea. In parallel, the molecular weight of recombinant GlnA-2 protein was calculated u size exclusion chromatography, being 667 kDa ( Figure S6). As the molecular weig GlnA-2 monomer was found to be 55 kDa by SDS-PAGE analysis, it suggests tha protein forms a dodecamer, as does GS. Although the GlnA-2 molecular weight pre a high degree of similarity with that of GS protein, it may not be a GS protein. Si Curiously, Hfx. mediterranei grew in Hm-DM with putrescine as the only source of nitrogen or carbon. These results indicate that Hfx. mediterranei can grow with polyamines. However, the maximum OD 600nm values reached were lower than those obtained in the presence of other nitrogen or carbon sources ( Figure S5).
Western blotting analysis carried out using Hfx. mediterranei protein extracts, in the presence of different concentrations of putrescine (20-80 mM) as a nitrogen source, indicate that the GlnA-1 and GlnA-2 proteins also showed expression in the middle of the exponential growth phase ( Figure 5). In Hm-CM, in the presence of 40 mM putrescine in the same growth phase, the expression of the GlnA-1 and GlnA-2 proteins could also be detected. Consequently, these results are an excellent starting point to determine the presence of a new pathway for putrescine degradation in Haloarchaea.
In parallel, the molecular weight of recombinant GlnA-2 protein was calculated using size exclusion chromatography, being 667 kDa ( Figure S6). As the molecular weight of GlnA-2 monomer was found to be 55 kDa by SDS-PAGE analysis, it suggests that this protein forms a dodecamer, as does GS. Although the GlnA-2 molecular weight presents a high degree of similarity with that of GS protein, it may not be a GS protein. Similar results have been described in E. coli, in which the GS protein and the glutamate-putrescine ligase (PuuA) share structural similarities [54].
Western blotting analysis carried out using Hfx. mediterranei protein extracts, i presence of different concentrations of putrescine (20-80 mM) as a nitrogen source, cate that the GlnA-1 and GlnA-2 proteins also showed expression in the middle o exponential growth phase ( Figure 5). In Hm-CM, in the presence of 40 mM putresci the same growth phase, the expression of the GlnA-1 and GlnA-2 proteins could al detected. Consequently, these results are an excellent starting point to determine the ence of a new pathway for putrescine degradation in Haloarchaea. In parallel, the molecular weight of recombinant GlnA-2 protein was calculated u size exclusion chromatography, being 667 kDa ( Figure S6). As the molecular weig GlnA-2 monomer was found to be 55 kDa by SDS-PAGE analysis, it suggests tha protein forms a dodecamer, as does GS. Although the GlnA-2 molecular weight pre a high degree of similarity with that of GS protein, it may not be a GS protein. Si

GS Activity and Glutamate-Putrescine Ligase Activity Assays
As expected, GlnA-1 recombinant protein shows GS activity (1.48 ± 0.11 U/mg) (Figure 6a). In Hm-CM, the GS activity obtained (0.22 ± 0.01 U/mg) is lower than those obtained in other conditions, because the GS-GOGAT pathway is more active under nitrogendeficient conditions. In Hm-DM with 40 mM nitrate or 40 mM putrescine, and in Hm-CM in the presence of putrescine, higher levels of GS activity were obtained (1.23 ± 0.06 U/mg, 2.99 ± 0.07 U/mg and 2.26 ± 0.06 U/mg, respectively) ( Figure 6b); however, surprisingly, GlnA-2 recombinant protein shows only glutamate-putrescine ligase activity (Figure 6a). In Hm-CM, no glutamate-putrescine ligase activity was obtained (0.11 ± 0.03 U/mg) because GlnA-2 protein is not expressed in this condition; whereas, in Hm-DM with 40 mM nitrate, glutamate-putrescine ligase activity was clearly detected (1.75 ± 0.07 U/mg). Independently of the medium composition, the highest values of glutamate-putrescine ligase activities were obtained in the presence of 40 mM putrescine (3.42 ± 0.09 U/mg in Hm-CM with putrescine or 4.47 ± 0.11 U/mg in Hm-DM with putrescine) (Figure 6b).
Biomolecules 2021, 11, x 1 results have been described in E. coli, in which the GS protein and the glutamate-pu cine ligase (PuuA) share structural similarities [54].

GS Activity and Glutamate-Putrescine Ligase Activity Assays
As expected, GlnA-1 recombinant protein shows GS activity (1.48 ± 0.11 U/mg) ure 6a). In Hm-CM, the GS activity obtained (0.22 ± 0.01 U/mg) is lower than thos tained in other conditions, because the GS-GOGAT pathway is more active under gen-deficient conditions. In Hm-DM with 40 mM nitrate or 40 mM putrescine, and in CM in the presence of putrescine, higher levels of GS activity were obtained (1.23 ± U/mg, 2.99 ± 0.07 U/mg and 2.26 ± 0.06 U/mg, respectively) ( Figure 6b); however, su ingly, GlnA-2 recombinant protein shows only glutamate-putrescine ligase activity ure 6a). In Hm-CM, no glutamate-putrescine ligase activity was obtained (0.11 ± U/mg) because GlnA-2 protein is not expressed in this condition; whereas, in Hm with 40 mM nitrate, glutamate-putrescine ligase activity was clearly detected (1.75 ± U/mg). Independently of the medium composition, the highest values of glutamate trescine ligase activities were obtained in the presence of 40 mM putrescine (3.42 ± U/mg in Hm-CM with putrescine or 4.47 ± 0.11 U/mg in Hm-DM with putrescine) (F 6b).
At other concentrations of putrescine used (5-80 mM), both the GS activity an glutamate-putrescine ligase activity were also obtained (data not shown). These re suggest that GlnA-2 protein from Hfx. mediterranei may be responsible for glutamate trescine ligase activity and is the first time that this activity has been evidenced in Ar However, to date, the role that this activity plays in the archaea is still unknown be glutamate-putrescine ligase activity has not been described in Hfx. mediterranei or in other Haloarchaea.

Construction and Characterization of the HM26-∆glnA-2 Mutants
To determine the GlnA-2 protein role in Hfx. mediterranei metabolism, deletion tants of the glnA-2 gene were generated and characterized depending on nitrogen so The Southern blot results confirm that the gene encoding GlnA-2 was deleted (Figur The fact that the knockout mutants of the glnA-2 gene were obtained without adding tamine to the culture medium, in contrast to the glnA-1 gene deletion mutants [20], cates that the glnA-2 gene is not an essential gene for glutamine synthesis in Hfx. me ranei. The deletion mutants HM26-ΔglnA-2 were characterized in Hm-DM with diff ammonium, nitrate, or putrescine concentrations as the sole nitrogen source (dat shown). Statistical analysis revealed that the HM26-ΔglnA-2 strain showed no signi differences in growth rate in Hm-CM. On the contrary, in Hm-CM with 40 mM putre the differences were significant ( Figure S8a). Moreover, the HM26-ΔglnA-2 strain At other concentrations of putrescine used (5-80 mM), both the GS activity and the glutamate-putrescine ligase activity were also obtained (data not shown). These results suggest that GlnA-2 protein from Hfx. mediterranei may be responsible for glutamateputrescine ligase activity and is the first time that this activity has been evidenced in Archaea. However, to date, the role that this activity plays in the archaea is still unknown because glutamate-putrescine ligase activity has not been described in Hfx. mediterranei or in any other Haloarchaea.

Construction and Characterization of the HM26-∆glnA-2 Mutants
To determine the GlnA-2 protein role in Hfx. mediterranei metabolism, deletion mutants of the glnA-2 gene were generated and characterized depending on nitrogen source. The Southern blot results confirm that the gene encoding GlnA-2 was deleted ( Figure S7). The fact that the knockout mutants of the glnA-2 gene were obtained without adding glutamine to the culture medium, in contrast to the glnA-1 gene deletion mutants [20], indicates that the glnA-2 gene is not an essential gene for glutamine synthesis in Hfx. mediterranei. The deletion mutants HM26-∆glnA-2 were characterized in Hm-DM with different ammonium, nitrate, or putrescine concentrations as the sole nitrogen source (data not shown). Statistical analysis revealed that the HM26-∆glnA-2 strain showed no significant differences in growth rate in Hm-CM. On the contrary, in Hm-CM with 40 mM putrescine, the differences were significant ( Figure S8a). Moreover, the HM26-∆glnA-2 strain only showed significant differences in growth rate at high ammonium (20-80 mM) ( Figure S8b), and nitrate concentrations (40-80 mM) ( Figure S8c), whereas in the presence of putrescine as a nitrogen source, significant differences were observed at all concentrations tested (5-80 mM) ( Figure S8d).
To prove whether deletion of the glnA-2 gene could affect enzyme activity, an activity analysis was performed using extracts obtained from HM26-∆glnA-2 strain cultures in different nitrogen sources (Hm-CM, Hm-CM with 40 mM putrescine, Hm-DM with 40 mM nitrate, or putrescine). Both GS biosynthetic activity and glutamate-putrescine ligase activity were measured to carry out these analyses.
As expected, in Hm-CM, similar GS activity was obtained in both the parental Hfx. mediterranei strain (HM26) and the mutant Hfx. mediterranei strain (HM26-∆glnA-2). The same results were also obtained with the glutamate-putrescine ligase activity.
In nitrate-defined or putrescine-defined medium, GS activity was obtained (1.25 ± 0.01 U/mg and 2.23 ± 0.07 U/mg respectively) in the parental strain (HM26), indicating that this activity is due to GlnA-1. However, although GS activity was also detected in the mutant strain (HM26-∆glnA-2), the values showed a remarkable decrease of 30-40%. In the same conditions, glutamate-putrescine ligase activity was obtained in the parental strain (1.75 ± 0.07 U/mg with nitrate or 3.47 ± 0.11 U/mg with putrescine).
Surprisingly, no glutamate-putrescine ligase activity was obtained in the HM26-∆glnA-2 mutant in these conditions. The same GS and glutamate-putrescine ligase activities results were obtained in Hm-CM with 40 mM putrescine, which did not show HM26-∆glnA-2 mutant strain glutamate-putrescine ligase activity. These results confirm that GlnA-2 exhibits putrescine activity rather than GS activity and, therefore, is a glutamateputrescine ligase protein that is erroneously annotated as a GS protein (Figure 7).  (Figure S8c), whereas in the presence of putrescine as a nitrogen source, significant differences were observed at all concentrations tested (5-80 mM) ( Figure S8d).
To prove whether deletion of the glnA-2 gene could affect enzyme activity, an activity analysis was performed using extracts obtained from HM26-ΔglnA-2 strain cultures in different nitrogen sources (Hm-CM, Hm-CM with 40 mM putrescine, Hm-DM with 40 mM nitrate, or putrescine). Both GS biosynthetic activity and glutamate-putrescine ligase activity were measured to carry out these analyses.
As expected, in Hm-CM, similar GS activity was obtained in both the parental Hfx. mediterranei strain (HM26) and the mutant Hfx. mediterranei strain (HM26-ΔglnA-2). The same results were also obtained with the glutamate-putrescine ligase activity.
In nitrate-defined or putrescine-defined medium, GS activity was obtained (1.25 ± 0.01 U/mg and 2.23 ± 0.07 U/mg respectively) in the parental strain (HM26), indicating that this activity is due to GlnA-1. However, although GS activity was also detected in the mutant strain (HM26-ΔglnA-2), the values showed a remarkable decrease of 30-40%. In the same conditions, glutamate-putrescine ligase activity was obtained in the parental strain (1.75 ± 0.07 U/mg with nitrate or 3.47 ± 0.11 U/mg with putrescine).
Surprisingly, no glutamate-putrescine ligase activity was obtained in the HM26-ΔglnA-2 mutant in these conditions. The same GS and glutamate-putrescine ligase activities results were obtained in Hm-CM with 40 mM putrescine, which did not show HM26-ΔglnA-2 mutant strain glutamate-putrescine ligase activity. These results confirm that GlnA-2 exhibits putrescine activity rather than GS activity and, therefore, is a glutamateputrescine ligase protein that is erroneously annotated as a GS protein (Figure 7).

Discussion
The Hfx. mediterranei genome has three genes annotated as GS (glnA-1, glnA-2, and glnA-3). Recently, it has been determined that the glnA-1 gene is essential for its growth, whereas proteins GlnA-2 and GlnA-3 could play a different role [20]. In particular, Hfx. mediterranei GlnA-2 presents substitutions in crucial amino acids for glutamine synthesis and the lack of the typical adenylylation residue. Consequently, it may be possible that it plays a catalytic role in other types of reactions. Similar substitutions have been described in other proteins annotated as GS [55,56]. Therefore, this protein would not be subject to control by covalent modification mediated by adenylyltransferases, as also occurs in GlnA4 from Myxococcus xanthus and GlnA2, GlnA3, and GlnA4 from S. coelicolor [22,55,56]. Other examples of GS-like proteins have been described in Pseudomonas sp. KIE171, in which the IpuC protein annotated as GS-like, is involved in the degradation of isopropylamine [57]; in E. coli K-12, the puuA gene was initially annotated as GS, subsequently demonstrating that this gene encodes a glutamate-putrescine ligase [54]; and in S. coelicolor M145, the GlnA3 protein is involved in the degradation of polyamines encoding a γ-glutamylpolyamine synthetase [22], and the GlnA4 protein in the degradation of ethanolamine [23].
Based on β-galactosidase activity results, the expression of the glnA-1 gene is indirectly related to the ammonium concentration, as occurs in S. coelicolor [58][59][60]. Furthermore, these results are in agreement with previous studies in which RT-PCR, qPCR, and microarray demonstrated that the transcription of the GS/GOGAT pathway genes and other genes involved in the nitrate assimilation pathway increased in nitrogen limiting conditions [13,16,17,61]. By comparison, these results indicate that the glnA-1 gene in Hfx. mediterranei is constitutively expressed, showing basal β-galactosidase activity. Opposite results have been reported for the Hfx. mediterranei nasA promoter, which presents a maximum activity at the beginning of the exponential growth phase, and shows a remarkable decrease in activity during the exponential phase in Hm-DM with 40 mM nitrate. Very low levels of β-galactosidase activity under the p-nas promoter control were detected in Hm-DM with 40 mM ammonia; values were close to zero throughout, indicating that the promoter was inactive under these conditions and only presented a very low basal activity [48]. The expression of glnA-2 appears not to depend on nitrogen limitation, unlike that of the glnA-1 gene.
At the translational level, the GlnA-1 protein shows expression in all nitrogen sources analyzed. By comparison, although the glnA-2 gene shows expression in all nitrogen sources analyzed, the GlnA-2 protein does not show expression at the translational level in Hm-CM. This could be because the glnA-2 gene is controlled by some type of regulation at the transcriptional level. At present, sRNAs are known to play an essential role in the post-transcriptional regulation of many processes in the three domains of life [62][63][64]. Recently sRNAs (sRNA274 and sRNA310) were identified in Hfx. mediterranei whose expression patterns change according to the nitrogen source [18]. This study shows that the glnA-2 gene is one of the potential target genes of these sRNAs, and it is possible that the expression of GlnA-2 would be regulated by this mechanism [19].
To elucidate the role of GlnA-2 in Hfx. mediterranei primary metabolism, its growth and expression were analyzed in the presence of other nitrogen sources, specifically in the presence of putrescine. This work shows that Hfx. mediterranei can grow in the presence of polyamine putrescine as the sole nitrogen or the sole carbon source. However, its growth is less than that using other nitrogen sources, such as ammonium or nitrate, or another carbon source, such as glucose. In addition, the β-galactosidase assay and Western blotting showed that GlnA-1 and GlnA-2 are expressed in Hm-DM in the presence of putrescine (20-80 mM). To test whether GlnA-2 can metabolize putrescine, GS activity and glutamate-putrescine ligase activity were assayed from recombinant GlnA-1 and GlnA-2 proteins. Results showed that only the GlnA-1 protein presents GS activity, whereas only the GlnA-2 protein presents glutamate-putrescine activity. Furthermore, in the mutant strain HM26-∆glnA-2, the GS activity decreased approximately 30-40%, whereas the glutamate-putrescine ligase activity was almost undetectable. Higher activity values were obtained under nitrogen-limiting conditions for GS and glutamate-putrescine activities. At high concentrations of ammonium or nitrate, the growth of the mutant strain HM26-∆glnA-2 showed significant differences concerning the growth of the parental strain HM26. This may be because the products of the pathway in which the GlnA-2 protein is involved may serve as a substrate of the pathway in which the GlnA-1 protein is involved, or some of its products may share both pathways (polyamine degradation and ammonium assimilation). However, in Hm-CM, the growth of both strains was similar because, in this condition, the GlnA-2 protein is not expressed, because it is absent in the mutant strain, it has no effect. The results of this work, both bioinformatic and experimental, show that the glnA-2 gene-previously annotated as a GS-encodes a glutamate-putrescine ligase protein.
Hfx. mediterranei may present a putrescine degradation reaction similar to that of E coli, in which the first step of glutamination would be catalyzed by GlnA-2 ( Figure 8). The first step of the γ-glutamylation pathway, or Puu pathway in E. coli, consists of the glutamylation of putrescine to γ-glutamylputrescine catalyzed by γ-glutamylputrescine synthetase or glutamate-putrescine ligase (PuuA) (EC 6.3.1.11). The γ-glutamylputrescine is completely oxidized to γ-glutamylaminobutyrate by γ-glutamylputrescine oxidase (PuuB) and γ-butyraldehyde dehydrogenase (PuuC). The γ-glutamylaminobutyrate is then hydrolyzed to glutamate and γ-aminobutyrate (γ-GABA) by γ-GABA hydrolase (PuuD) [30,54,65]. However, the aminotransferase pathway requires, first, the transamination of putrescine with 2-oxoglutarate to generate glutamate and γ-aminobutyraldehyde by putrescine aminotransferase (PatA) and, second, the oxidation of this γ-aminobutyraldehyde to γ-aminobutyrate by γ-aminobutyraldehyde dehydrogenase (PatD/PuuC). In addition, different genes related to polyamine metabolism in this bacterium have been identified in the Hfx. mediterranei genome, such as putA, aldY2, aldY5, hat2 speE, speB1, and pdaD. Figure 8. Representation of the possible metabolic pathways of putrescine degradation and other polyamines in Hfx. mediterranei R4. Image based on the metabolic maps included in the KEGG da tabase (organism code: hme). In blue, the EC numbers of each of the enzymes whose coding genes are present in Hfx. mediterranei are indicated. In grey, the enzymes that catalyze these reactions whose genes are not present in the genome of this halophilic archaea. In red, the two genes tha could encode glutamate-putrescine ligase. The white cylinder represents the different amino acid and polyamine transporters. Moreover, various polyamide transporter genes (HFX_0019, HFX_ 1687, HFX_6077 HFX_6079, and HFX_6080) homologous to bacterial species such as in S. coelicolor SCO2780, SCO3453 SCO5668, SCO5669, and SC05670 [22] and in halophilic archaea Hfx volcanii potA1 (HVO_A0293), potA2 (HVO_A0294), potB (HVO999), and potD Figure 8. Representation of the possible metabolic pathways of putrescine degradation and other polyamines in Hfx. mediterranei R4. Image based on the metabolic maps included in the KEGG database (organism code: hme). In blue, the EC numbers of each of the enzymes whose coding genes are present in Hfx. mediterranei are indicated. In grey, the enzymes that catalyze these reactions whose genes are not present in the genome of this halophilic archaea. In red, the two genes that could encode glutamate-putrescine ligase. The white cylinder represents the different amino acid and polyamine transporters. Moreover, various polyamide transporter genes (HFX_0019, HFX_ 1687, HFX_6077, HFX_6079, and HFX_6080) homologous to bacterial species such as in S. coelicolor SCO2780, SCO3453 SCO5668, SCO5669, and SC05670 [22] and in halophilic archaea Hfx. volcanii potA1 (HVO_A0293), potA2 (HVO_A0294), potB (HVO999), and potD (HVO_A0300) [66] are present in the Hfx. mediterranei genome.
Although the presence of intracellular polyamides in Hfx. mediterranei has not yet been determined, the results described show the presence of enzymes capable of catalyzing the first step of the γ-glutamylation pathway in Hfx. mediterranei. For the first time, this provides evidence of the presence of these enzymes in Archaea.

Conclusions
This work successfully demonstrated that the GlnA-2 protein, in contrast to the GlnA-1 protein, shows glutamate-putrescine ligase activity. Although these halophilic proteins share structural similarities, the substitutions in essential residues for glutamine biosynthesis, which are conserved in the three domains of life, could be the reason why GlnA-2 does not present GS activity. Similarly, the Hfx. mediterranei growth in the presence of putrescine represents the first attempt to reveal a novel polyamine utilization pathway in the Archaea domain, which remains complex and unexplored. This contribution is a milestone in understanding the ecology of Hfx. mediterranei and the use of alternative nitrogen sources, which allows this microorganism to be competitive in its habitat and to survive under stress conditions. Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/biom11081156/s1. Table S1: Primers sequences used in the promoter region cloning and in the analysis of expression at the transcriptional level by RT-PCR. Figure Figure S4: Cell growth and β-galactosidase activity of Hfx. mediterranei transformants pVA513-p-glnA in Hm-DM with 40 mM putrescine. (a) Hfx. mediterranei transformants pVA513-p-glnA-1. (b) Hfx. mediterranei transformants pVA513-p-glnA-2. Figure S5: Growth curves of the Hfx. mediterranei R4 in the presence of putrescine. All conditions are represented in different colors: • Hm-DM with putrescine as the sole nitrogen source (with 0.5% glucose). • Hm-DM with putrescine as the sole carbon source (without glucose, with ammonia as nitrogen source). • Hm-CM with putrescine. Figure S6: Determination of the molecular mass of the recombinant protein GlnA-2 by chromatography on Sephacryl S-300. The standard proteins are represented in gray whereas the GlnA-2 protein is represented in blue. Figure

Data Availability Statement:
The data presented in this study are available within the article.