Metal coordination by L-amino acid oxidase derived from flounder Platichthys stellatus is structurally essential and regulates antibacterial activity

L-amino acid oxidases (LAAOs) have antibacterial activity and play important roles in innate immunity. We have previously identified a LAAO of ~52 kDa in size from the mucus layer of the flounder Platichthys stellate (psLAAO1) and have successfully produced psLAAO1 as a secreted bioactive recombinant protein by using Pichia pastoris (P. pastoris). The recombinant psLAAO1 inhibited the growth of bacteria to the same levels as native psLAAO1 present in the mucus layer. In this study, homology modeling of psLAAO1 predicted metal coordination by residues Y241, H348, and D406. We show that the Michaelis constant (K_m) of psLAAO1 decreased and the catalytic constant (K_cat/K_m) value increased following pre-treatment of the protein with a chelating agent. In contrast to the non-chelated protein sample, enzymatic activity of EDTA-treated psLAAO1 gradually decreased or was absent after one or two freeze-thaw cycles. The H348A psLAAO1 mutant generated by site-directed mutagenesis and recombinantly produced by P. pastoris did not display antibacterial activity. The results of the metal detection assay revealed that for the non-metal coordinating histidine mutant (H209A, control), the levels of iron, zinc, and magnesium were similar to those of wild-type psLAAO1, whereas magnesium was not detected in the H348A mutant sample. A wild-type psLAAO1 sample treated with chelating agent did not contain zinc and magnesium ions. In conclusion, metal coordination by psLAAO1 affects enzymatic activity, and H348 is involved in the coordination of magnesium, and metal coordination by psLAAO1 provides essential structural stability. Key Points: •Homology modeling of psLAAO1 predicted metal coordination by residue H348 •The H348A psLAAO1 mutant showed no antibacterial activity or magnesium coordination •Metal coordination by H348 affects enzyme activity and structural stability

Keywords: L-amino acid oxidase; Antibacterial protein; Enzymatic activity; Metal coordination; Metalloprotein

L-amino acid oxidase (LAAO, EC1.4.3.2) is a flavoenzyme that catalyzes oxidative deamination of an L-amino acid substrate to an α-keto acid with the generation of ammonia and hydrogen peroxide (Wellner and Meister [31]). Hydrogen peroxide is a powerful oxidizer that acts as an intracellular signal and is involved in the oxidative burst of phagocytes, which leads to the elimination of invading microorganisms (Clifford and Repine [5]). The antibacterial activity of LAAOs is due to the production of hydrogen peroxide (Izidoro et al. [15]). We have examined the selectivity, specificities, directed oxidation activity, local effectivities, and production of intermediate products or high concentrations of hydrogen peroxide over a short period by LAAOs (Kasai et al. [18], [19]). The effects of low-dose exposure of hydrogen peroxide produced by LAAOs to the surface of bacteria membrane proteins and DNA and RNA synthesis have also been examined (Kasai et al. [18]). LAAOs are distributed widely and play important roles in innate immunity as antibacterial agents (Nagaoka et al. [28]; Fujii et al. [11]; Hughes [14]), apoptosis-inducing agents (Murakawa et al. [27]; Butzke et al. [4]; Kanazawa et al. [16]; Mukherjee et al. [26]), and the regulation of cell cycle arrest (de Melo Alves Paiva et al. [7]) and display anti-viral (Zhang et al. [33]) and anti-parasite properties (Deolindo et al. [8]). LAAOs are conserved in various animal fluid components (Nagaoka et al. [28]; Fujii et al. [11]; Hughes [14]; Murakawa et al. [27]; Butzke et al. [4]; Kanazawa et al. [16]; Mukherjee et al. [26]; de Melo Alves Paiva et al. [7]; Zhang et al. [33]; Deolindo et al. [8]; Kitani et al. [20]).

We have reported that the LAAO from the epidermal mucus of the flounder Platichthys stellatus (psLAAO1) has strong antibacterial activity against various pathogenic bacterial species and strains (Kasai et al. [17]). Furthermore, we have successfully expressed the antibacterial protein psLAAO1 as a secretory bioactive recombinant protein in the methylotrophic yeast P. pastoris (Kasai et al. [18], [19]). Several studies have reported procedures for the recombinant production of LAAOs (Hahn et al. [13]; Bloess et al. [3]). These recombinant LAAOs display bactericidal (Kasai et al. [18], [19]; Li and Li [23]) and parasiticidal (Li et al. [24]) activity and bind to the surface of various species of bacteria as observed for native LAAOs (Kasai et al. [18], [19]), whereas several LAAOs have no cytotoxic activities against human monocytes (de Melo Alves Paiva et al. [7]), macrophages, and erythrocytes (Okubo et al. [29]). These results indicate that LAAOs are potential therapeutic agents for the treatment of various bacterial infections. Although previous studies have shown that the antibacterial activities of LAAOs are associated with the production of hydrogen peroxide, an understanding of the detailed regulation mechanisms and specificity toward bacteria remain missing.

Metals coordinated by proteins often play an integral role in the catalytic reaction of enzymes and provide structural stability (Bertini et al. [1]). In particular, zinc, magnesium, and calcium stabilize the active conformation of proteins (Williams [32]), and the recent crystal structure of a snake LAAO revealed that metal coordination stabilized the protein conformation (Feliciano et al. [10]). Moreover, the enzymatic activity of several snake LAAOs increased in the presence of manganese ions (El Hakim et al. [9]), whereas enzyme activity was inhibited by zinc (El Hakim et al. [9]), nickel (El Hakim et al. [9]; Costa et al. [6]), cobalt (El Hakim et al. [9]), copper (Costa et al. [6]), aluminum (El Hakim et al. [9]; Costa et al. [6]), sodium (Costa et al. [6]), and calcium ions (Costa et al. [6]). Therefore, coordination of metals by LAAOs appears to play a role in regulating enzyme activity.

In this study, we performed modeling analysis of psLAAO1 to predict the metal coordination site and examined the antibacterial and enzyme activity with or without chelating treatment. psLAAO1 mutants were prepared based on the predicted metal coordination results, and a metal detection assay was performed. The results revealed that metal coordination by psLAAO1 affected antibacterial and enzyme activities as well as structural stability.

The amino acid sequence of psLAAO1 (Accession Number: BAI66016.1) was retrieved from the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov). SWISS-MODEL (swissmodel.expasy.org) was used to search for templates and build homology models. The model quality of the template structure and target to template alignment were assessed by the global model quality estimate (GMQE) and the quaternary structure quality estimate (QSQE), and quantified modeling errors and estimates of the expected model accuracy were assessed by QMEAN scores, which use several statistical descriptors expressed as mean force potentials (Waterhouse et al. [30]; Biasini et al. [2]). The GMQE is expressed as a number between 0 and 1, and higher values are indicative of higher reliability. In general, a GMQE score greater than 0.7 is considered a reliable predictor of quaternary structure from the modeling process. The QSQE is given a score between 0 and 1 and reflects the expected accuracy of the inter-chain contacts for a model built using a given alignment and template. The QMEAN should be a score greater than −4.0. Data analysis of model building, scoring, and prediction of metal coordination was conducted by Altif Laboratories Inc., Tokyo, Japan.

The psLAAO1 sequence was codon optimized for expression in yeast by artificial gene synthesis (GENEWIZE, South Plainfield, NJ, USA) using the full-length psLAAO1 mRNA (Accession number: AB495360). The codon-optimized psLAAO1 sequence (Accession number: LC556326) was custom cloned into the pUC57-Amp plasmid vector to yield the native protein with an N-terminal secretory signal sequence (MGAHVMKCEIYVVSVLLFTMSQSQTAA) to ensure that the recombinant protein was expressed in yeast as a soluble form. The 5′-UTR of the artificial psLAAO1 included an SnaBI site and an initiation site of the Kozak translation sequence, (A/Y)A(A/T)AATGTCT (start codon is underlined). A 6×His-tag sequence, stop codon, and AvrII (BlnI) site were located at the C-terminus. The insert was cleaved with the restriction enzymes, and the insert was sub-cloned into the pPIC3.5 expression vector (Thermo Fisher Scientific, Waltham, MA, USA) at the multiple cloning site located between the 5′ AOX1 promoter and 3′ AOX1 terminator using the same restriction enzymes. The DNA construct was sequenced using an ABI PRISM 3130 genetic analyzer (Thermo Fisher Scientific). For protein overexpression, the plasmid was linearized at the SalI site (HIS4 site) and transformed by a MicroPulser™ (Bio-Rad Laboratories, Inc. Hercules, CA, USA) into the P. pastoris (GS115) yeast strain (Thermo Fisher Scientific) according to the protocol provided by the manufacturer. The His+ clone was selected by using a histidine-free Regeneration Dextrose medium agar plate. Total DNA was isolated from the desired His+ recombinant, which is suitable for genomic PCR. Homologous recombination into the yeast genome was detected by genomic PCR using a 5′ AOX1 sequencing primer (5′-GACTGGTTCCAATTGACAAGC-3′) and 3′ AOX1 sequencing primer (5′-GCAAATGGCATTCTGACATCC-3′), according to the manufacturer's protocol. Yeast cells were grown in 1 L of buffered glycerol-complex medium at 28°C until the culture reached an OD600 of between 2.0 and 6.0. Protein expression was then induced in buffered minimal glycerol-complex medium with 0.5% methanol. The supernatant of the P. pastoris culture was collected, and ammonium sulfate added to the supernatants to a final concentration of 1 M. Supernatants was then loaded onto a hydrophobic interaction chromatography column containing Capto™ phenyl (high sub) resin (GE Healthcare UK Ltd., Little Chalfont, UK) that was equilibrated with 1 M (NH4)2SO4 in 50-mM phosphate buffer (pH 7.0). The column was washed with the same buffer. The recombinant protein was eluted from the column using 50-mM phosphate buffer (pH 7.0). The antibacterial activity of each fraction was examined using growth inhibition plates. The antibacterial fractions were collected, and the expressed His-tagged fusion protein was isolated by Ni-NTA agarose (Qiagen, Hilden, Germany) according to the manufacturer's instructions. After Ni-affinity elution, an ultrafiltration membrane column was used to exchange the buffer (20-mM Tris-HCl, pH 7.0). The Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific) was used to determine the concentration of the purified protein.

The antibacterial activity of recombinant psLAAO1 was determined using a growth inhibition plate assay. Staphylococcus aureus (S. aureus; ATCC25923), methicillin-resistant S. aureus (MRSA; clinical isolate 87–7927), and Vibrio parahaemolyticus (V. parahaemolyticus; RIMD2210001) were cultured in trypticase soy agar (TSA) for 16 h at 37 °C. To prepare plates, bacteria were suspended in TSA at a final concentration of 1 × 106 colony-forming units (CFU)/mL. A hole with a diameter of 2.8 mm was punched into the agar and filled with 10 μL of the purified protein or fractions from each of the column purification steps. After overnight incubation at 37 °C, the clear zone around the hole was measured. S. aureus was purchased from the ATCC. V. parahaemolyticus and MRSA (clinical isolate) were obtained from the Hirosaki University School of Medicine and Hospital.

Two Float-A-Lyzer G2 dialysis membrane devices (MWCO: 8–10 K, Spectrum Laboratories, Inc., Rancho Dominguez, CA, USA) were loaded with purified recombinant psLAAO1, and the samples were dialyzed overnight at 4 °C against a 500-fold volume of 20-mM Tris-HCl buffer (pH 7.0) in the presence or absence of 0.1% EDTA. After dialysis, each device was transferred to a 1000-fold volume of chelating agent-free 20-mM Tris-HCl buffer (pH 7.0), and dialysis was continued against this buffer at 4 °C overnight. This dialysis step was repeated at least three times. The concentration of the dialyzed protein samples was measured using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). Storing the chelated psLAAO1 at −80 °C caused a decrease or loss of both enzyme and antibacterial activity after one or two freeze-thaw steps. Therefore, samples used in all experiment were not freeze-thawed after chelation and used immediately.

L-amino acid oxidase activity of psLAAO1 was determined by measuring the production of hydrogen peroxide using an enzyme-coupled assay (Macheroux et al. [25]). Five μg/mL purified recombinant psLAAO1 (0.1% EDTA-treated or non-chelated) was prepared in 20-mM Tris-HCl buffer (pH 7.0) containing 1 U/mL peroxidase, 500-μg/mL o-phenylendiamine and 25-mM L-lysine as the substrate. Reactions were performed at 37 °C. The quantity of the generated hydrogen peroxide was measured by absorbance at 450 nm. Km, Vmax, Kcat, and Kcat/Km were calculated at different L-lysine concentrations (7.5–15 mM) and in the presence of 5 μg/mL purified recombinant psLAAO1 with or without 0.1% EDTA pre-treatment using the Michaelis–Menten equation and Lineweaver–Burk analysis.

The metals coordinated to psLAAO1 were detected using the Metallo Assay LS (Metallogenics Co., Ltd., Chiba, Japan), which quantitatively determines the amount of metals in biological samples by using a microplate reader, in accordance with the manufacturer's instructions. Six molar HCl was added to the purified recombinant psLAAO1 solution to adjust the pH to 2.0–3.0 at room temperature. After centrifugation, the supernatant was collected and used for the assay. Metal concentrations from the collected supernatant were measured by a microplate reader at each maximum absorbance.

The pPIC3.5 vector with the inserted artificial psLAAO1 gene as the template and the KOD-plus-Mutagenesis kit (TOYOBO, Tokyo, Japan) were used to generate the psLAAO1 H348A and H209A mutants. The following mutagenic primers and its complementary sequence were designed to change the histidine codon to alanine codon at positions 348 (H348A) and 209 (H209A) of the psLAAO1 gene: H348A forward primer (point mutation site is underlined) 5′-CGCTTACGACTCTTCTACCAAGATT-3′; H348A reverse primer 5′-ACAACACGAAGGGCCTCCATCTTCT-3′; H209A forward primer 5′-CGCCTATACTGTTAAGGAATACTTG-3′; and H209A reverse primer 5′-TCGTACTTCTTCAAGGCTTCGGAGC-3′. Amplification was conducted according to the KOD-plus-Mutagenesis kit protocol. The inverse PCR products were treated with DpnI and self-ligated. The sequence was determined using an ABI PRISM 3130 Genetic Analyzer (Thermo Fisher Scientific). For protein expression, the plasmid was linearized at the SalI site and transformed into the P. pastoris (GS115) yeast strain (Thermo Fisher Scientific). Total DNA was isolated from the desired His+ recombinant, which is suitable for genomic PCR.

SDS-PAGE was performed according to the method of Laemmli (Laemmli [21]). Protein samples were heated in 10% glycerol, 2% SDS, 6% 2-mercaptoethanol, and 50-mM Tris-HCl buffer (pH 6.8) for 5 min in a boiling water bath and subjected to SDS-PAGE with 7.5% Mini-PROTEAN® TGX™ Precast Gels (Bio-Rad). Proteins were stained with Coomassie Brilliant Blue R-250 (CBBR-250). After electrophoresis, the proteins were electrically transferred from the gel onto a polyvinylidene difluoride (PVDF) membrane (GE Healthcare). The membrane was blocked with 20-mM Tris-HCl (pH 7.4), 125-mM NaCl, 0.2% Tween 20, and 5% skim milk (Morinaga, Tokyo, Japan). Recombinant psLAAO1 was detected using a rabbit polyclonal anti-psLAAO antibody (1:1000 dilution). Horseradish peroxidase-conjugated secondary antibody donkey anti-rabbit IgG (1:5000 dilution; GE Healthcare) was used for detection, followed by enhanced Amersham™ ECL™ Prime Western Blotting Detection Reagents (GE Healthcare).

The amino acid sequence of psLAAO1 was used as the template to search for models using SWISS-MODEL, and the models were assessed by composite scoring functions (Waterhouse et al. [30]; Biasini et al. [2]). Templates from the search with GMQE and QMEAN values above 0.7 and −4.0, respectively, are shown in Table 1. We then built homology models using these template structures, which were L-amino acid oxidases from snake venom and were monomer (PDB ID: 1tdo), homodimeric (PDB ID: 5ts5, 5z2g, 1f8s and 4e0v), and homotetrameric (PDB ID: 3kve) flavoproteins that catalyze the deamination of L-amino acids. These template structures contain the highly conserved structural motif of LAAOs and oligomeric forms are stabilized by metal ions (Feliciano et al. [10]; Georgieva et al. [12]). Predicted structures of the monomer, homodimer, and homotetramer psLAAO1s are presented in Fig. 1.

Modeling scores and information about template LAAOs

aAll template structures are X-ray crystal structures. b PDB ID; the 4-character unique identifier of every entry in the Protein Data Bank (https://www.rcsb.org/). c GMQE, d QSQE and e QMEAN scores are described in the section of materials and methods.

Graph: Fig. 1 Modeling the structure of psLAAO1. (a) Monomer structure of psLAAO1 (template: PDB ID 1tdo). (b) Homodimeric structure of psLAAO1 (template: PDB ID 5ts5). (c) Homotetrameric structure of psLAAO1 (template: PDB ID 3kve). Structures are presented as ribbon representation, which were prepared by using the Waals software (Altif Laboratories Inc., Tokyo, Japan). FAD molecules are indicated as ball and stick models. Secondary structures of the helical and sheet domains (a) and each chain (b, c) are shown in different colors.

The modeling results using the 5ts5 structure, which had the highest QMEAN score, predicted that psLAAO1 formed a homodimer containing two flavin adenine dinucleotides (FADs) and coordination of two metal ions by several active residues from both chains (Fig. 2a). The coordination bond lengths between the metal ion and the oxygen atoms of Y241 (chain A) and D406 (chain A) and the nitrogen atom of H348 (chain B) were 2.61, 2.67, and 2.57 Å, respectively (Fig. 2b). Similarly, the coordination bond lengths between the second metal ion and the nitrogen atom of H348 (chain A) and the oxygen atoms of Y241 (chain B) and D406 (chain B) were 2.51, 2.60, and 2.45 Å, respectively (Fig. 2c).

Graph: Fig. 2 Coordination geometry of the predicted metal binding sites of psLAAO1. (a) Coordination models and interaction between metal ions and active residues from different chains. (b) Coordination bond lengths between the metal ion and the oxygen atoms of Y241 (chain A) and D406 (chain A) and the nitrogen atom of H348 (chain B). (c) Coordination bond lengths between the metal ions and the nitrogen atom of H348 (chain A) and the oxygen atoms of Y241 (chain B) and D406 (chain B). Ribbon representation and ball and stick models of the key side chains were prepared by Waals software (Altif Laboratories Inc., Tokyo, Japan)

We produced bioactive recombinant psLAAO1 using P. pastoris to investigate the effect of metal coordination on antibacterial activity. Purified recombinant psLAAO1 was dialyzed extensively against a buffer that contained EDTA as the metal-chelating agent. Recombinant psLAAO1 was also dialyzed extensively against the same buffer without EDTA as the control. Both samples were then dialyzed extensively against an EDTA-free buffer and subjected to antibacterial analysis by the plate inhibition assay. Pre-chelated psLAAO1 was more active than the non-EDTA treated (control) psLAAO1 sample against S. aureus ATCC 25923, MRSA clinical isolate 87–7927, and V. parahaemolyticus RIMD2210001 (Fig. 3).

Graph: Fig. 3 Antibacterial activity of wild-type psLAAO1 with or without pre-chelating treatment against (a) S. aureus ATCC 25923, (b) MRSA clinical isolate 87–7927, and (c) V. parahaemolyticus RIMD2210001. Each bacterial strain was suspended in TSA at a final concentration of 1 × 106 CFU/mL. Each sample was applied to the holes in the agar, and the antibacterial activity was measured after overnight incubation at 37 °C. N: Non-treated psLAAO1; E: EDTA pre-treated psLAAO1

We used the enzyme-coupled assay to determine the effect of metal coordination by psLAAO1 on kinetic activity. Purified recombinant psLAAO1 was treated with or without EDTA. These psLAAO1 samples were then buffer exchanged to an EDTA-free buffer and subjected to the enzyme-coupled assay. Enzyme kinetics using L-lysine as the substrate (7.5–15 mM) was performed with Lineweaver-Burk analysis (Fig. 4). The Km, Kcat, and Kcat/Km values with or without chelating treatment are shown in Table 2. Interestingly, the enzymatic and antibacterial activity of pre-chelated psLAAO1 were high immediately after EDTA treatment; however, the activity of pre-chelated psLAAO1 gradually decreased or was absent after one or two freeze-thaw cycles, whereas non-EDTA-treated psLAAO1 maintained its activity even after repeated freeze-thaw cycles.

Graph: Fig. 4 Kinetic analysis of the L-amino acid oxidase activity with or without pre-chelating treatment. Lineweaver-Burk plot of psLAAO1 (5 μg/mL) pre-treated with 0.1% EDTA (open circles) or non-EDTA treated (filled circles) in 7.5–15-mM L-lysine and 20-mM Tris-HCl (pH 7.0) at 37 °C for 30 min. Absorbance at 450 nm was measured

Kinetic parameters for psLAAO1 treated with or without EDTA as a chelating agent

We produced a H348A psLAAO1 mutant using the P. pastoris expression system to investigate the role of H348 in coordination of metal ions. We produced wild-type (WT) psLAAO1 as a control to compare with the mutation procedure and the H209A mutant as a control to compare with the H348A mutation, in which H209 was predicted to be located at the surface of the protein and solvent exposed between both chains. Each inserted gene (1806 bp) by homologous recombination into the yeast genome was detected by genomic PCR (Fig. 5a). After chromatography purification, the recombinant WT, H209A, and H348A proteins were detected by western blot analysis using the polyclonal anti-psLAAO IgG (Fig. 5b). Antibacterial activity was observed for the WT and H209A psLAAO1 proteins, whereas the H348A mutant did not have antibacterial activity (Fig. 6).

Graph: Fig. 5 Genomic PCR and western blot analysis of wild-type psLAAO1 and the H209A and H348A psLAAO1 mutants. (a) Genomic PCR of mutagenesis constructs using the 5′ AOX1 sequence primer and 3′ AOX1 sequence primer. (b) Western blot using an anti-psLAAO IgG. M1: Wide-Range DNA Ladder (Takara Bio Inc.); WT: wild-type psLAAO1; 209: H209A psLAAO1; 348: H348A psLAAO1; M2: molecular weight marker of Precision Plus ProteinTM All Blue Pre-stained Protein Standards (Bio-Rad). Arrowheads indicate the specific bands representing psLAAO1

Graph: Fig. 6 Antibacterial activity of (a) wild-type psLAAO1, and (b) H209A, and (c) H348A psLAAO1 mutants against MRSA clinical isolate 87–7927. Bacteria were suspended in TSA to a final concentration of 1 × 106 CFU/mL. Each sample was applied to the holes in the agar, and the antibacterial activity was measured after overnight incubation at 37 °C

We performed a metal detection assay to determine the effect of metal coordination on enzyme activity. After purification of recombinant WT, H209A, and H348A psLAAO1s, these protein samples were buffer exchanged by extensive dialysis against a Tris-HCl (pH 7.0) buffer to remove salts and then subjected to the metal detection assay (Table 3). The mole ratios of one WT protein to iron, zinc, and magnesium were 0.12 ± 0.003, 0.50 ± 0.04, and 3.61 ± 1.09, respectively. Similarly, the mole ratios of one H209A protein to iron, zinc, and magnesium were 0.04 ± 0.001, 0.33 ± 0.03, and 3.48 ± 1.45, respectively, and the values are similar to those obtained for the WT protein. The mole ratios of one H348A protein to iron and zinc were 0.07 ± 0.004 and 0.43 ± 0.06, which are similar values to those obtained for the WT and H209A psLAAO1s. Conversely, magnesium was not detected in the H348A mutant sample. Notably, only iron was detected for the pre-chelated psLAAO1 sample. Calcium was not detected in all constructs and chelated protein samples.

The one molecule psLAAO1 to divalent metal ratios

*ND: not detect

In this study, we found that homology modeling of psLAAO1 predicted multiple structures (Fig. 1) with metal ions coordinated by residues Y241, H348, and D406 (Fig. 2). Moreover, we identified that the Michaelis constant and the catalytic constant values changed following pre-treatment of psLAAO1 with a metal chelating agent (Fig. 4, Table 2). Furthermore, the recombinant psLAAO1 H348A mutant displayed no antibacterial activity (Fig. 6). The metal detection assay detected iron, zinc, and magnesium ions in the WT psLAAO1 and H209A mutant samples. The iron and zinc levels for the H348A mutant were similar to those of the WT and H209A psLAAO1 proteins, but magnesium was not detected (Table 3). Only iron was detected in the pre-chelated psLAAO1 sample.

Many proteins contain a metal essential for function. Approximately one third of all known enzymes are estimated to contain a metal ion as a functional factor. In particular, metalloproteins require a metal cofactor for structural stability or to carry out their function. Recently, several snake LAAO crystal structures have revealed that metal ions stabilize the enzymatically active homodimeric (Feliciano et al. [10]) and homotetrameric (Georgieva et al. [12]) quaternary structures, and these metal ion coordination sites are important for the biological activity of LAAOs. In general, metal ion coordination is essential for regulating activity and structural stability of various enzymes. In many studies, magnesium, zinc, and calcium ions stabilize the structural fold and physiologically active conformation of proteins (Bertini et al. [1]). Metal ions are an integral part of many enzymes and indispensable in the redox catalyzed reactions, and they also play an important role in formation of specific complexes such as a hemoglobin, which contains iron for oxygen binding. Therefore, metal coordination by LAAOs may play an important function in enhancing structural stability and indirectly the oxidation reaction.

Interestingly, a previous report suggested that LAAOs require magnesium ions for enzymatic activity, whereas zinc ions inhibit the activity LAAOs (Lazo et al. [22]). In general, magnesium stabilizes the interfacial interaction between subunits to give active oligomeric forms of many proteins (Williams [32]). Conversely, the inhibitory action of metal ions may be related to their ability to reversibly bind to the active site of the protein, thereby compromising substrate binding and reducing the activity of the enzyme by changing the conformation of the protein (Bertini et al. [1]). Therefore, we hypothesize that metal coordination at the surface of LAAO may regulate enzyme activity and structural stability.

The results of this study suggest that H348 of psLAAO1 mediates magnesium coordination by forming a metal coordination bond, because magnesium was not detected in the metal detection assay for the H348A mutant (Table 3). Moreover, the recombinant H348A mutant displayed no antibacterial activity (Fig. 6). Therefore, we hypothesize that H348 of psLAAO1 irreversibly coordinates magnesium, which is essential for the formation of the active site of the enzyme and protein stability (also during protein expression), because treatment of the sample with a chelating agent that removes magnesium caused a gradual reduction in stability after one or two freeze-thaw cycles. The observation that approximately 0.5 mole zinc coordinates with one mole protein for all constructs examined suggests that zinc does not coordinate with monomeric psLAAO1 (Table 3). Interestingly, we initially tried to determine enzyme kinetics using a greater substrate concentration range, but enzyme activity was observed to decrease significantly at low and high substrate concentrations. Consequently, an accurate Michaelis–Menten analysis of the data could not be obtained. Thus, we used a much narrower range of substrate concentrations to create Lineweaver-Burk plots. This decrease in enzyme activity at low and high substrate concentrations may be an allosteric effect due to reversible complex formation. Therefore, in the presence of zinc, psLAAO1 may form a dimer or higher oligomeric species. Furthermore, as metal chelating treatment increased the antibacterial activity (Fig. 3) and the affinity of psLAAO1 for the substrate (Fig. 4, Table 2), we postulate that reversibly bound zinc may inhibit enzyme activity, as described in a previous report (El Hakim et al. [9]).

Although LAAO is a primary antibacterial agent of innate immunity that produces hydrogen peroxide, the regulatory mechanism that specifically targets bacteria remains poorly understood. Understanding the regulatory mechanism of psLAAO1 in detail should provide a potential path for using psLAAO1 as an antimicrobial agent that treats infections caused by clinical multiple drug resistance pathogens because psLAAO1 shows high sensitivity and binding potency against bacteria. The metal-mediated regulation presented herein provides insight into the regulatory mechanism of psLAAO1, and our future work aims to further decipher the antibacterial activity, specificity toward bacteria and structural conformation of this metalloprotein.

This work was supported in part by a Grant for Priority Research Designated by the Japan Science and Technology Agency (Grant No. JPMJTM19AL) and a fund from SNOWDEN Co., Ltd.

We are grateful to Dr. Ishikawa and Mr. Ohishi for advice and discussion. We thank Ms. Tabata, Mr. Ishizuka, Ms. Takayama, Ms. Uraguchi, Mr. Honma, and Ms. Sakuyama for helpful assistance. We thank Edanz Group (https://en-author-services.edanzgroup.com/ac) for editing a draft of this manuscript.

Study design: KK, TN, and TM. Experimental: KK, YI, and AN. Data analysis: KK. Manuscript writing and revision: KK, KA, TN, and TM.

All authors have no conflicts of interest to declare.

This article does not contain any studies with human participants or animals performed by any of the authors.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.