Highly Stable, Cold-Active Aldehyde Dehydrogenase from the Marine Antarctic Flavobacterium sp. PL002

: Stable aldehyde dehydrogenases (ALDH) from extremophilic microorganisms constitute efﬁcient catalysts in biotechnologies. In search of active ALDHs at low temperatures and of these enzymes from cold-adapted microorganisms, we cloned and characterized a novel recombinant ALDH from the psychrotrophic Flavobacterium PL002 isolated from Antarctic seawater. The recombinant enzyme (F-ALDH) from this cold-adapted strain was obtained by cloning and expressing of the PL002 aldH gene (1506 bp) in Escherichia coli BL21(DE3). Phylogeny and structural analyses showed a high amino acid sequence identity (89%) with Flavobacterium frigidimaris ALDH and conservation of all active site residues. The puriﬁed F-ALDH by afﬁnity chromatography was homotetrameric, preserving 80% activity at 4 ◦ C for 18 days. F-ALDH used both NAD + and NADP + and a broad range of aliphatic and aromatic substrates, showing cofactor-dependent compensatory K M and k cat values and the highest catalytic efﬁciency (0.50 µ M − 1 s − 1 ) for isovaleraldehyde. The enzyme was active in the 4–60 ◦ C-temperature interval, with an optimal pH of 9.5, and a preference for NAD + -dependent reactions. Arrhenius plots of both NAD(P) + -dependent reactions indicated conformational changes occurring at 30 ◦ C, with four(ﬁve)-fold lower activation energy at high temperatures. The high thermal stability and substrate-speciﬁc catalytic efﬁciency of this novel cold-active ALDH favoring aliphatic catalysis provided a promising catalyst for biotechnological and biosensing applications.


Introduction
Extremophilic bacteria are characterized by tolerance or the capability of growing under extreme conditions, such as hypersaline habitats, high pressures, and extreme temperatures. Many extremophiles inhabiting environments characterized by multiple extreme conditions are known as polyextremophiles [1,2]. The capability to overcome such stressing conditions present in extreme environments, is due to special characteristics of their extremophilic enzymes (extremozymes) [3][4][5].
In recent decades, research has exploited extreme environments looking for producers of a wide array of biotechnologically useful extremozymes that are one of the great interests for industrial processes, mainly in biocatalysis due to their particular stability and high catalytic activity over a wide range of various parameters (temperature, salinity, pH, etc.) and their coping ability with specific industrial pipeline requirements. Additionally, the use of enzymes in biological processes is considered part of the environmentally friendly chemicals, known as Green Chemistry, which aim to minimize or eliminate the use and production of dangerous substances for human health and environment [6]. The global DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). The aldH gene (1506 bp) was synthesized (ATG Biosynthetics GmbH, Merzhausen, Germany) and inserted in the pHAT2 expression vector (EMBL, Heidelberg, Germany) using NcoI/BamHI sites. The accuracy of the amplified nucleotide sequence was verified by Sanger sequencing (ATG Biosynthetics GmbH, Merzhausen, Germany). The recombinant plasmid pF-ALDH was used for gene expression in Escherichia coli BL21(DE3) (Thermo Fisher Scientific, Massachusetts, USA). The transformants were cultivated at 37 • C for 18 h in Luria broth medium containing 100 µg mL −1 ampicillin up to OD 600 = 0.750, and gene expression was performed by induction with 0.5 mM isopropyl-β-D-thiogalactoside (IPTG) at 25 • C for 16 h. The induced cells were collected by centrifugation for 10 min at 9000× g (4 • C) and stored at −80 • C.

Purification of Recombinant F-ALDH and Size Exclusion Chromatography
The recombinant F-ALDH was obtained as a fused protein with 6xHis polypeptide appended to the amino terminal end of the enzyme and purified by affinity chromatography [42]. The induced cells (400 mL culture) were resuspended in 20 mL buffer A (50 mM potassium phosphate buffer, pH 7, 200 mM KCl, 10 mM beta-mercaptoethanol (β-ME)) and disrupted by ultrasonic treatment for 5 min with 5 s pulses and 60 s pauses, using a Sonopuls ultrasonic homogenizer (Bandelin, Berlin, Germany). The cellular extract was centrifuged at 16,000× g for 30 min at 4 • C, and the soluble fraction was applied on a 1-mL Ni-NTA agarose (Qiagen, Hilden, Germany) gravity flow column equilibrated with buffer A. The column was washed with 10 mL buffer A, and 6 mL buffer A containing 100 mM imidazole, and the recombinant F-ALDH was eluted in the presence of 250 mM imidazole. The fractions were analyzed by electrophoresis on 12.5% SDS-PAGE gels [43]. After the buffer exchange to 50 mM potassium phosphate pH 7, 10 mM β-ME, using the 7K MWCO Zeba Spin Desalting columns (Thermo Fisher Scientific, Massachusetts, USA), the purified enzyme was stored at −20 • C in the presence of 15 % glycerol. Protein concentrations were measured using the Bio-Rad Protein Assay Kit II (BioRad, Wien, Austria).

Enzyme Assay
F-ALDH activity was determined spectrophotometrically by measuring the rate of NADH formation at OD 340 nm . The reaction mixture contained 1 mM aldehyde, 1 mM NAD(P) + , and 100 mM Glycine/KOH pH 9.5 buffer with 2 mM β-ME in a final reaction volume of 200 µL. The reaction initiated by the addition of 200 ng F-ALDH was continuously measured for 5 min at 25 • C using a FLUOstarOmega microplate reader (BMG Labtech, Offenburg, Germany). One unit of enzyme was defined as the amount of enzyme catalyzing the formation of 1 µmol NADH per minute, using the molar absorption coefficient

Cloning and Expression of F-ALDH Coding Gene
The Flavobacterium PL002 strain isolated from Antarctic seawaters [40] was cultivated for seven days at 15 °C on Reasoner's 2B broth (R2B, Melford Biolaboratories Ltd., UK) medium [41]. Total DNA was extracted from the collected cells using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). The aldH gene (1506 bp) was synthesized (ATG Biosynthetics GmbH, Merzhausen, Germany) and inserted in the pHAT2 expression vector (EMBL, Heidelberg, Germany) using NcoI/BamHI sites. The accuracy of the amplified nucleotide sequence was verified by Sanger sequencing (ATG Biosynthetics GmbH, Merzhausen, Germany). The recombinant plasmid pF-ALDH was used for gene expression in Escherichia coli BL21(DE3) (Thermo Fisher Scientific, Massachusetts, USA). The transformants were cultivated at 37 °C for 18 h in Luria broth medium containing 100 µg mL −1 ampicillin up to OD600 = 0.750, and gene expression was performed by induction with 0.5 mM isopropyl-β-D-thiogalactoside (IPTG) at 25 °C for 16 h. The induced cells were collected by centrifugation for 10 min at 9,000× g (4 °C) and stored at −80 °C.

Purification of Recombinant F-ALDH and Size Exclusion Chromatography
The recombinant F-ALDH was obtained as a fused protein with 6xHis polypeptide appended to the amino terminal end of the enzyme and purified by affinity chromatography [42]. The induced cells (400 mL culture) were resuspended in 20 mL buffer A (50 mM potassium phosphate buffer, pH 7, 200 mM KCl, 10 mM beta-mercaptoethanol (β-ME)) and disrupted by ultrasonic treatment for 5 min with 5 s pulses and 60 s pauses, using a Sonopuls ultrasonic homogenizer (Bandelin, Berlin, Germany). The cellular extract was centrifuged at 16,000× g for 30 min at 4 °C, and the soluble fraction was applied on a 1-mL Ni-NTA agarose (Qiagen, Hilden, Germany) gravity flow column equilibrated with buffer A. The column was washed with 10 mL buffer A, and 6 mL buffer A containing 100 mM imidazole, and the recombinant F-ALDH was eluted in the presence of 250 mM imidazole. The fractions were analyzed by electrophoresis on 12.5% SDS-PAGE gels [43]. After the buffer exchange to 50 mM potassium phosphate pH 7, 10 mM β-ME, using the 7K MWCO Zeba Spin Desalting columns (Thermo Fisher Scientific, Massachusetts, USA), the purified enzyme was stored at −20 °C in the presence of 15 % glycerol. Protein concentrations were measured using the Bio-Rad Protein Assay Kit II (BioRad, Wien, Austria).

Enzyme Assay
F-ALDH activity was determined spectrophotometrically by measuring the rate of NADH formation at OD340 nm. The reaction mixture contained 1 mM aldehyde, 1 mM NAD(P) + , and 100 mM Glycine/KOH pH 9.5 buffer with 2 mM β-ME in a final reaction volume of 200 µL. The reaction initiated by the addition of 200 ng F-ALDH was continuously measured for 5 min at 25 °C using a FLUOstarOmega microplate reader (BMG Labtech, Offenburg, Germany). One unit of enzyme was defined as the amount of enzyme catalyzing the formation of 1 µmol NADH per minute, using the molar absorption coefficient ꜪNADH = 6.22 × 10 3 L mol −1 cm −1 . NADH = 6.22 × 10 3 L mol −1 cm −1 .

Sequence Analyses and Phylogeny
The primary structure analysis of F-ALDH was carried out using the ExPASy Prot-Param tool [44] to calculate the theoretical molecular weight (MW), isoelectric point (pI), aliphatic index, and grand average of hydropathicity (GRAVY). The identity and similarity percentages of F-ALDH with homologous enzymes were determined using the Emboss Needle pair alignment tool (http://www.ebi.ac.uk/Tools/psa/emboss_needle/ [45]. Protein multiple alignment was carried out using the CLUSTAL OMEGA EMBL-EBI (1.2.4) software http://www.ebi.ac.uk/Tools/msa/clustalo/) [45].
Phylogenetic analysis was performed using the MEGAX software [46]. The MUSCLE algorithm in the MEGA package was used to align the primary structure of ALDHs obtained from the GenBank database after eliminating all positions containing gaps and missing data. The consensus phylogenetic tree was obtained by maximum-likelihood analysis of 70 amino acid sequences with a total of 411 positions in the final dataset [47]. Initial trees for the heuristic search were obtained by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances using a JTT model [48] and selecting the topology with the superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites (five categories (+G, parameter = 0.6872)). The tree topography was evaluated using the bootstrap analysis of 1000 repetitions.

Phylogeny and Sequence Analyses of F-ALDH
BLAST screening of the Flavobacterium PL002 genome sequence [40] led to the identification of a 1506-bp aldH gene coding for an open reading frame homologous to bacterial aldehyde dehydrogenases (accession number MH542427). The hypothetical enzyme (F-ALDH) consisted of 501 amino acids, with calculated mass of 54.95 kDa and theoretical pI of 5.87.
The phylogenetic analysis (Figure 1) of the F-ALDH amino acid sequence in relation with the homologous aldehyde dehydrogenases from Flavobacterium spp. and other genera showed the Antarctic enzyme grouped with ALDSs from the Clade I of the genus Flavobacterium. F-ALDH appeared to be a close homolog (93.7%) of aldehyde dehydrogenases from the psychrotrophic F. frigoris [WP 007139093.1] originating from Antarctic lakes [49] and mesophilic F. granuli [WP 072945862.1] retrieved from wastewater [50]. pI of 5.87.
The phylogenetic analysis (Figure 1) of the F-ALDH amino acid sequence in relation with the homologous aldehyde dehydrogenases from Flavobacterium spp. and other genera showed the Antarctic enzyme grouped with ALDSs from the Clade I of the genus Flavobacterium. F-ALDH appeared to be a close homolog (93.7%) of aldehyde dehydrogenases from the psychrotrophic F. frigoris [WP 007139093.1] originating from Antarctic lakes [49] and mesophilic F. granuli [WP 072945862.1] retrieved from wastewater [50].    Figure 1). The archaeal ALDH [WP_010978003.1] from Sulfurisphaera tokadaii [56] was used as a rooting enzyme for the bacterial ALDH phylogenetic tree, marking a larger structural divergence based on the microbial kingdom.
The deduced amino acid sequence of F-ALDH is homologous with ALDH from the Antarctic isolate Flavobacterium frigidimaris [BAB96577.1] [39], showing 89% identity and 94% similarity with this psychrophilic counterpart and showing lower scores (38-39% identity, 56-60% similarity) with ALDHs from mesophilic bacteria Escherichia coli [Heim 1991] and Staphylococcus aureus [46], and thermophilic Thermus thermophilus [55] ( Table 1). The lowest identity (30%) and similarity (52%) was obtained with the hyperthermophilic Sulfurisphaera tokodaii archaeal ALDH [56] (Table 1). These differences in similarity levels of ALDHs primary structures suggested the presence of structural adaptation elements of enzymes from cold environments to ensure an equilibrium among their activity, stability, and flexibility required for catalysis low temperatures [57]. Multiple alignment of these aldehyde dehydrogenases ( Figure 2) highlighted the presence in F-ALDH of all residues involved in catalysis and substrate binding universally conserved in ALDHs [24,58,59], confirming that this Antarctic enzyme is a functional aldehyde dehydrogenase. Among these conserved residues, Cys296, Glu257, G293, and N160 (F-ALDH numbering) were shown to play a role as a catalytic nucleophile (Cys296) and base (Glu257), in the orientation of Cys residue (Gly293), and in stabilizing the aldehyde carbonyl oxygen (Asn160) during catalysis [60]. Gly259-Gly260 dipeptide appeared to stabilize the interface between the catalytic and cofactor binding domains [32]. The 12residue specific motif (marked with stars) comprising Arg75, Gly151, Asn160, Pro162, Gly177, Lys183, Gly259, Cys296, Pro401, Gly447, Asn452, and Gly465 is also conserved in F-ALDH. Glycine residues are in a high percentage (8.6%) and positioned in critical turns [60]. Also, among the 21 residues that could be involved in substrate binding, conserved residues are located in the substrate binding site (Asn157, Trp165, Lys166, and Leu418 in ALDH from S. aureus) and in the cofactor binding site (Trp156, Lys180, Gly233, Glu390, and Phe392 in ALDH from S. aureus), while Gly233 and Phe392 are responsible for binding the NAD nicotinamide ring [60]. Asn157 could stabilize the active cysteine during catalysis [61,62], and Trp156 and Lys180 situated in a hydrophobic pocket interacts with NADP + [59,62]. residues are located in the substrate binding site (Asn157, Trp165, Lys166, and Leu418 in ALDH from S. aureus) and in the cofactor binding site (Trp156, Lys180, Gly233, Glu390, and Phe392 in ALDH from S. aureus), while Gly233 and Phe392 are responsible for binding the NAD nicotinamide ring [60]. Asn157 could stabilize the active cysteine during catalysis [61,62], and Trp156 and Lys180 situated in a hydrophobic pocket interacts with NADP + [59,62].  An ion pair conserved in all tetrameric ALDHs could be formed in F-ALDH between the homologous Lys166-Glu467 (in S. aureus) in support of the determined oligomeric structure (tetramer) of the cold-active F-ALDH. Moreover, this multiple alignment revealed a partial conservation of the F-ALDH oligomerization domain, mainly with that of F. frigidimaris ALDH, consisting of two regions (I 127 -L 150 and Q 485 -F 501 ) ( Figure 2). Among these, the I 127 -L 150 stretch showed a high content of Ser (25%) and Glu (16%) residues, and a 5:1 ratio of negatively to positively charged residues. The calculated hydropathicity GRAVY index (0.208) of this region indicated the hydrophobic nature of the F-ALDH subdomain, similar to E. coli ALDH (0.133) and distinct from the rest of the investigated ALDH homologs (negative values). The amino acid composition corresponded to an aliphatic index of 130, indicating an increased thermal stability. The second region (Q 485 -F 501 ) was rich in positively charged residues Lys (23.5%) and hydrophobic residues Leu and Phe (11% each). However, the ratio of negatively to positively charged amino acids was reversed in this case (1:4), and the negative GRAVY index (−0.571) corresponded to a hydrophilic region with a low aliphatic index (68.8).

Cloning, Expression and Purification of F-ALDH
The gene coding for F-ALDH was inserted in the pHAT2 expression vector that appended a His-tag at the N-terminal of the recombinant enzyme to facilitate purification. The resulted construct (pF-ALDH) was expressed in E. coli at a level of 45 mg L −1 of culture, and the soluble fraction was purified to near homogeneity in one step by affinity chromatography on Ni-NTA agarose resin after elution with 250 mM imidazole (Figure 3). The estimated molecular mass of the recombinant F-ALDH by SDS-PAGE was~55 kDa, in accordance with the calculated one. Under nondenaturing conditions, the size of the molecule determined by size exclusion chromatography on a HiLoad Superdex 200 column corresponded to a single peak of 210 kDa (not shown), indicating a homotetrameric enzyme, a commonly found oligomeric structure in class 1 and class 2 members of ALDHs [63,64]. ; identical (stars) and conserved (dots) residues from all enzymes; active site residues (grey shade); NAD + /NADP + cofactor binding site residues (black shade); aldehyde dehydrogenase motifs (PROSITE entry PS00687and PS00070) (boxes).
An ion pair conserved in all tetrameric ALDHs could be formed in F-ALDH between the homologous Lys166-Glu467 (in S. aureus) in support of the determined oligomeric structure (tetramer) of the cold-active F-ALDH. Moreover, this multiple alignment revealed a partial conservation of the F-ALDH oligomerization domain, mainly with that of F. frigidimaris ALDH, consisting of two regions (I127-L150 and Q485-F501) (Figure 2). Among these, the I127-L150 stretch showed a high content of Ser (25%) and Glu (16%) residues, and a 5:1 ratio of negatively to positively charged residues. The calculated hydropathicity GRAVY index (0.208) of this region indicated the hydrophobic nature of the F-ALDH subdomain, similar to E. coli ALDH (0.133) and distinct from the rest of the investigated ALDH homologs (negative values). The amino acid composition corresponded to an aliphatic index of 130, indicating an increased thermal stability. The second region (Q485-F501) was rich in positively charged residues Lys (23.5%) and hydrophobic residues Leu and Phe (11% each). However, the ratio of negatively to positively charged amino acids was reversed in this case (1:4), and the negative GRAVY index (−0.571) corresponded to a hydrophilic region with a low aliphatic index (68.8).

Cloning, Expression and Purification of F-ALDH
The gene coding for F-ALDH was inserted in the pHAT2 expression vector that appended a His-tag at the N-terminal of the recombinant enzyme to facilitate purification. The resulted construct (pF-ALDH) was expressed in E. coli at a level of 45 mg L −1 of culture, and the soluble fraction was purified to near homogeneity in one step by affinity chromatography on Ni-NTA agarose resin after elution with 250 mM imidazole ( Figure 3). The estimated molecular mass of the recombinant F-ALDH by SDS-PAGE was ~55 kDa, in accordance with the calculated one. Under nondenaturing conditions, the size of the molecule determined by size exclusion chromatography on a HiLoad Superdex 200 column corresponded to a single peak of 210 kDa (not shown), indicating a homotetrameric enzyme, a commonly found oligomeric structure in class 1 and class 2 members of ALDHs [63,64].

Biochemical Characterization of F-ALDH
Substrate specificity of the cold-active F-ALDH was screened for a broad spectrum of aldehydes comprising six aliphatic and four aromatic aldehydes in the presence of NAD + or NADP + as cofactors (Figure 4). For all tested substrates, the activity of this cold-active enzyme indicated an overall preference for the NAD + -dependent reaction. Among aldehydes, the highest NAD + -dependent activity was measured when using 4-isopropilbenzaldehyde (18.9 ± 0.7 U mg −1 ) and isovaleraldehyde for the NADP +dependent reaction (22.6 ± 1.2 U mg −1 ) (Figure 4). A high (86.8%) NAD-dependent activity was also observed when using isovaleraldehyde and butiraldehyde (60.5%) for the NADPdependent reaction. These aldehydes are known as flavoring agents for the food and beverage industries, and 4-isopropilbenzaldehyde (cuminaldehyde) is also used in cosmetics [65]. Benzaldehyde is known to form covalent linkages to lysine amino groups with an inhibitory effect on this class of enzymes [66]. However, in the case of F-ALDH, the substrate specificity suggested an independent response on the type of aldehyde (aliphatic or aromatic) used in the reaction (Figure 4). Also, the major difference observed for the catalysis of benzaldehyde and the 2-isopropyl and 4-isopropil derivates of this aromatic substrate (Figure 4) is in accordance with the specificity variation within the structurally heterogeneous group of benzaldehyde compounds [67]. In this sense, further 3D modeling and molecular docking analyses could highlight subtle structural features responsible for the substrate specificity of both substrates and cofactors used by this cold-active ALDH.  The influence of metal ions and various additives on the NAD-dependent activity showed an overall stable response when using benzaldehyde as a substrate (Table 2). A moderate activity loss (18-22%) was observed in the presence of the protease inhibitor. Meanwhile, the addition of 10% ethanol or betamercaptoethanol (β-ME) induced a slight  The influence of metal ions and various additives on the NAD-dependent activity showed an overall stable response when using benzaldehyde as a substrate (Table 2). A moderate activity loss (18-22%) was observed in the presence of the protease inhibitor. Meanwhile, the addition of 10% ethanol or betamercaptoethanol (β-ME) induced a slight activation up to 144.7% and 124.1%, respectively (Table 2). Unlike potassium and sodium ions that had no impact on F-ALDH activity, even at 100 mM concentrations and at different pH values (8 and 9.5), the tested divalent ions appeared to inhibit this cold-active enzyme ( Table 2). At 1 mM concentration, Ca 2+ seems to have a lower inhibitory effect (7%) than Mg 2+ (9-15%), and Ni 2+ (22%), while HgCl 2 produced total inhibition independent of the concentration used (Table 2).
Reducing agents are important in ALDH activity since the catalytic site is a Cys residue. Therefore, β-ME was tested to determine the impact of this compound on F-ALDH, inducing a 16-24% activation. Moderate inhibition was also produced by 2 mM EDTA (20%) inhibition and triton X-100 (10 %) ( Table 2).
The effect of ethanol on the enzymatic activity was also tested since this solvent was used for solubilizing the aldehyde substrates. In this case, the presence of 1% ethanol induced a slight inhibition (15%), whereas the addition of 10% ethanol increased the activity by 40% (Table 2), an effect possibly due to a putative alcohol oxidative reaction catalyzed by F-ALDH.
Optimum pH for the F-ALDH reaction was 9.5 when using 50 mM glycine-KOH buffer, showing a 100-fold increase in the pH 6-9.5 interval and a decline for more alkaline values ( Figure 5). As expected, the enzyme was inactivated at pH 6, a value close to the calculated pI (5.8) of this enzyme. At the same pH (9), the nature of the buffer also played a role on the F-ALDH activity, while the activity was reduced by 25% in the presence of TrisHCl as compared to glycine-KOH at the same concentration (50 mM). No activity was measured at pH 2 and pH 3 when using 50 mM Glycine-HCl buffer (data not shown).
The thermal stability of the recombinant F-ALDH was determined by incubating the enzyme at various temperatures up to 70 • C and assaying the activity at 25 • C. The enzyme exhibited a high stability up to 55 • C, with a 60% apparent activation in the 40 • C-50 • C interval, and a 60% residual activity at 60 • C ( Figure 6A). At 70 • C the enzyme was completely inactivated. At 50 • C, the highest temperature favoring an active conformation, a prolonged incubation of the purified F-ALDH (Supplementary Figure S1) for up to 3 h showed a progressive partial inactivation down to 48-50% after 2 h, possibly due to oligomers dissociation. The thermal stability of the recombinant F-ALDH was determined by incubating the enzyme at various temperatures up to 70 °C and assaying the activity at 25 °C. The enzyme exhibited a high stability up to 55 °C, with a 60% apparent activation in the 40 °C-50 °C interval, and a 60% residual activity at 60 °C ( Figure 6A). At 70 °C the enzyme was completely inactivated. At 50 °C, the highest temperature favoring an active conformation, a prolonged incubation of the purified F-ALDH (Supplementary Figure S1) for up to 3 h showed a progressive partial inactivation down to 48-50% after 2 h, possibly due to oligomers dissociation. When stored at 4 • C for a longer period of time, the NAD + -dependent reaction appeared to be completely stable for four days, following a slight decrease (of 8%) after seven days and of 25% after eighteen days of incubation ( Figure 6B). Meanwhile, the NADP-dependent reaction was more affected, with 28% and 35% activity loss after seven and eighteen days, respectively ( Figure 6B). This high stability of the enzyme, especially at low temperatures, was in accordance with the environmental conditions for the Antarctic Flavobacterium sp. PL002 growth [40].
The effect of temperature on the reaction rate of F-ALDH was determined by measuring the NAD(P)-dependent activity at various temperatures ( Figure 6C). The temperature appeared to favor the F-ALDH reaction in the presence of the NAD + cofactor of up to four-fold higher at 50 • C. However, the cold-active F-ALDH had the same activity at 13 • C, independent of the cofactor used, corresponding to the environmental temperature of this bacterial strain. The Arrhenius plot (inset- Figure 6C) was biphasic for both NADand NADP-dependent reactions, suggesting a conformational change occurring at 30 • C. The calculated activation energy using NAD + as a cofactor shifted from 76 kJ mol −1 at 4-30 • C to 19 kJ mol −1 at a higher temperature (30-60 • C) interval, and from 53 kJ mol −1 to 11 kJ mol −1 between the same temperature intervals in the case of the NADP + -dependent reaction. This shift, indicating temperature-dependent changes in the catalytic mechanism, was also observed in ALDH from other psychrophilic species, like F. frigidimaris, where this enzyme required a higher activation energy for catalysis (27 kJ mol −1 ) than PL002 at higher temperatures, while lower (57 kJ mol −1 ) at temperatures below 30 • C [39]. This difference could be related to particular molecular adaptations in the psychrophilic F. frigidimaris as compared to the psychrotolerant Flavobacterium sp. PL002. Fermentation 2022, 7, x FOR PEER REVIEW 13 of 20 °C for up to 18 days, and the activity when using NAD + (black) and NADP + (grey) as cofactors was measured at 25 °C. (C) The effect of temperature on the F-ALDH activity was determined by measuring the reaction at different temperatures. Isovaleraldehyde saturation curves were performed at various temperatures in the 4-60 °C interval in the presence of NAD + (•) and NADP + (○), and the maximum velocity (Vmax) was calculated as described in Methods. The inset shows the corresponding Arrhenius plots.
When stored at 4 °C for a longer period of time, the NAD + -dependent reaction appeared to be completely stable for four days, following a slight decrease (of 8%) after seven days and of 25% after eighteen days of incubation ( Figure 6B). Meanwhile, the NADPdependent reaction was more affected, with 28% and 35% activity loss after seven and eighteen days, respectively ( Figure 6B). This high stability of the enzyme, especially at low temperatures, was in accordance with the environmental conditions for the Antarctic Flavobacterium sp. PL002 growth [40].
The effect of temperature on the reaction rate of F-ALDH was determined by measuring the NAD(P)-dependent activity at various temperatures ( Figure 6C). The temperature appeared to favor the F-ALDH reaction in the presence of the NAD + cofactor of up to four-fold higher at 50 °C. However, the cold-active F-ALDH had the same activity at 13 °C, independent of the cofactor used, corresponding to the environmental temperature of this bacterial strain. The Arrhenius plot (inset- Figure 6C) was biphasic for both NADand NADP-dependent reactions, suggesting a conformational change occurring at 30 °C. The calculated activation energy using NAD + as a cofactor shifted from 76 kJ mol −1 at 4-30 °C to 19 kJ mol −1 at a higher temperature (30-60 °C) interval, and from 53 kJ mol −1 to 11 Long term stability of F-ALDH was evaluated after storing the purified enzyme at 4 • C for up to 18 days, and the activity when using NAD + (black) and NADP + (grey) as cofactors was measured at 25 • C. (C) The effect of temperature on the F-ALDH activity was determined by measuring the reaction at different temperatures. Isovaleraldehyde saturation curves were performed at various temperatures in the 4-60 • C interval in the presence of NAD + (•) and NADP + ( ), and the maximum velocity (V max ) was calculated as described in Methods. The inset shows the corresponding Arrhenius plots.

Kinetics of the Recombinant F-ALDH
The F-ALDH steady state kinetic parameters (Table 3) were calculated for the aliphatic and aromatic substrates, isovaleraldehyde and benzaldehyde, respectively, when using both cofactors, NAD + and NADP + . The saturation curves for these substrates (Supplementary Figure S2) were hyperbolic, except for an inhibition effect at high concentrations of benzaldehyde in the NADP-dependent reaction (panel B) and of isovaleraldehyde in the presence of both cofactors (panels C and D). The catalytic activity measured at 30 • C varied with both the aldehyde substrate and the cofactor used in the reaction, exhibiting higher values for the NAD + -dependent reactions (Table 3). Overall, kcat appeared to be three-fold higher in the case of isovaleraldehyde /NAD + than isovaleraldehyde/NADP + and benzaldehyde/NAD + , and 10-fold higher than the benzaldehyde/NADP + reaction rate. Nevertheless, the maximum initial rates for isovaleraldehyde/NADP + and benzaldehyde/NAD + are comparable. Meanwhile, the K M for benzaldehyde and isovaleraldehyde ranged in the 136.8 µM-197 µM interval when using NAD + as a cofactor, and a 2.4-to 2.8-fold higher apparent affinity in the presence of NADP (Table 3). For the aliphatic isovaleraldehyde substrate, similar catalytic efficiencies (k cat /K M ) were obtained independent of the cofactor, while it was double for benzaldehyde in the presence of NAD + . Moreover, this cold-adapted F-ALDH presented a two-fold higher catalytic efficiency when using the aliphatic substrate than the aromatic benzaldehyde (Table 3).
In comparison with other ALDHs (Supplementary Table S2), the K M of F-ALDH for benzaldehyde/NAD + indicated a similar apparent affinity for this substrate as the homologous enzyme from Homo sapiens (saliva) [68] and Euglena gracilis (mitochondria) [69]. A higher apparent affinity for this aromatic substrate was observed relative to Nocardia sp. (by two-fold) [70] and H. sapiens ALDH7A1 (by four-fold) [71] homologous enzymes. Meanwhile, the Km of the cold-active F-ALDH is two-fold higher than for the Pseudomonas putida BADH [72], H. sapiens ALDH1B1 (liver) [73], and F. frigidimaris ALDH [39]. Among the ALDHs from cold environments, the closest relative F. frigidimaris ALDH (89% identity) appeared to have a higher apparent affinity for this substrate than PL002 that could be associated with a different molecular adaptation to cold environments, defining it as a psychrophilic bacterium [39]. Interestingly, ALDH2 from human liver mitochondria has an apparent affinity approximately 8000 times higher than F-ALDH, suggesting a different metabolic role involving this substrate in the two organisms [74]. For the same substrate and cofactor, the turnover number of F-ALDH was eight-fold higher than that of ALDH2 from the mitochondrial liver of H. sapiens, and just three times lower than P. putida ALDH, while the corresponding catalytic efficiency wase only six-fold higher for P. putida and 1000-fold higher for H. sapiens ALDH2 from liver mitochondria [72,74], respectively.
In the case of the benzaldehyde/NADP + reaction, the apparent affinity of F-ALDH was comparable with that of the homologous enzyme from P. putida [72], was higher than the R. norvegicus mitochondrial ALDH1 from the liver [75] by 30-fold, and was higher than ALDHs from E. coli [27] and R. norvegicus (liver/microsome) [75] by 100-fold. Meanwhile, the turnover number was comparable with that of ALDH2 from E. coli, and four-fold lower than P. putida ALDH, resulting in a 60-fold higher catalytic efficiency as compared to E. coli ALDH [27], and five-fold lower than that of the P. putida enzyme [72]. A comparable Km value was also observed for the isovaleraldehyde/NAD + reaction of Flavobacterium PL002 and Klebsiella pneumoniae ALDH [19], while the apparent affinity for this substrate was three-fold lower than that of the E. coli enzyme [27]. However, considering that the turnover number of F-ALDH was greater (by five-fold) than in the case of E. coli and the K. pneumoniae homologs, the calculated catalytic efficiency of this cold-active enzyme is significantly higher, by six and seventeen times as compared to that of the two bacterial dehydrogenases, respectively [19,27].
Unlike other ALDHs (Supplementary Table S2), the extremozyme, F-ALDH, showed substrate inhibition by isovaleraldehyde at high concentrations (> 2 mM) that was not observed for ALDHs from K. pneumoniae [19] and E. coli K-12 [27] when using the same substrate. Meanwhile, several aldehydes were reported to have an inhibitory effect, such as betaine aldehyde (K i = 0.34 mM) for the S. aureus ALDH [62], acetaldehyde for the Saccharomyces cerevisiae (K i = 2.55 mM) [76] and Nocardia sp. (K i = 5 mM) [70] ALDHs, formaldehyde (K i = 6 mM) and benzaldehyde (K i = 2 mM) for the Nocardia sp. ALDH [70], whereas there was a lower response for butyraldehyde for the Lachnoclostridium phytofermentans dehydrogenase (K i = 144.3 mM) [77] and isobutanal for Rattus norvegicus ALDH (K i = 100.1 mM) [78]. Interestingly, in the case of F-ALDH, the NAD + -dependent oxidation of benzaldehyde was not inhibited by the substrate excess, unlike the NADP + -dependent reaction that was inhibited by concentrations of this substrate above 15 mM.
Furthermore, based on the good affinity of this cold-active ALDH for benzaldehyde, an electrochemical assay using the enzyme in solution and a stable biosensor for benzaldehyde using immobilized F-ALDH were developed [79]. The biosensor was suitable for the determination of benzaldehyde impurities in benzyl alcohol, a non-active ingredient in liquid drugs. The biosensor provides a simpler, more economical alternative to the corresponding standard chromatographic tests for the quality control of benzyl alcohol. In addition, a patent application was reported for an electrochemical assay for the detection of thiram exploiting the inhibitory action of this fungicide on F-ALDH [80]. Thus, the high apparent affinity for aldehyde substrates like benzaldehyde and the high stability are also important assets for using this Antarctic bacterial enzyme in biosensors for practical applications [81].
It is well known that immobilization improves the enzyme stability under harsh conditions, enabling its repeated use for cost-effective industrial applications [82,83]. While immobilization could lead to a certain loss in enzymatic activity and to changes in the substrate specificity or substrate inhibition, the advantages of immobilization for bioreactors for many industrial applications or for the development of analytical devices, such as biosensors, are undeniable. For example, immobilization of F-ALDH by cross-linking with glutaraldehyde in a matrix of bovine serum albumin (a largely used method in the biosensing field) led to a change in Km from 145 µmol/L to 386 µmol/L [79]. Nonetheless, the immobilization enabled repeatable measurements in a flow injection system for over three days of extensive use at 60 tests/day [79]. Other methods enabling controlled quantitative enzyme immobilization, such as exploiting the nickel-histidine affinity in the case of His-tagged enzymes [84], immobilization on nanomaterials with high loading capacities [85], enzyme nanoflowers [86], CLEAs, or other recent strategies [83] could lead to better performance of immobilized F-ALDH.
Regarding industrial applications, the recycling of the soluble cofactor NAD + is one of the important aspects to consider [87]. For cost-effective processes, this is typically achieved by coupled enzymatic reactions. In this respect, the high apparent affinity for NAD + (Km ranging from 32.5 to 74 µM depending on the aldehyde substrate used), the ability to catalyze aldehydes oxidation at low temperatures with a high catalytic efficiency (0.89-1.06 µM −1 s −1 ), and the important stability determined over a broad range of temperatures of up to 50 • C of the recombinant Antarctic F-ALDH constitute valuable characteristics of a promising catalyst for cofactor regeneration by enzymatic processes at industrial scale [87].

Conclusions
In search of cold-active aldehyde dehydrogenases as potent and stable extremozymes for catalyzing applicative aldehyde conversion reactions at low temperatures, a novel ALDH was isolated from an Antarctic psychrotrophic bacterium and characterized. The high heterologous expression yield of this recombinant enzyme and the one-step purifi-cation were useful prerequisites for putative-enhanced catalysts. The synopsis of the structural and functional properties highlighted a NAD(P) + -dependent dehydrogenase with a preference towards the NAD + cofactor catalyzing a broad range of aliphatic and aromatic aldehyde oxidation is of interest in biotechnologies and biosensing. In addition, a particularly high thermal stability at temperatures up to 55 • C, rather common for (hyper)thermophiles-originating enzymes, together with activity preservation in the presence of various ions and compounds, constitute clear advantages for developing such an extremozyme-based biocatalyst. Moreover, the ability to carry out low-temperature catalysis characterized by high apparent affinity for benzaldehyde and isovaleraldehyde added to the benefit of using this cold-active F-ALDH in industrial processes with energy savings. Hence, these characteristics of the novel ALDH from the Antarctic Flavobacterium PL002 strain constitute promising leads for using F-ALDH as a stable catalyst in monitoring processes at low temperatures. Alternatively, this cold-active enzyme already demonstrated its potential for biosensor development exploiting both the direct affinity for aldehyde substrates and the inhibition by dithiocarbamate fungicides.