Introduction

Phosphatase is an important enzyme in recycling phosphate in marine environments as well as within living cells. Its distribution is dependent on environmental factors such as phosphate availability, dominant habitants, water depth, salinity, temperature, and the internal ratio of nitrogen to phosphorus in higher organisms and microbial communities (Hoppe 2003). Interestingly, high alkaline phosphatase (AP) activity occurs in eutrophic or deep water in the presence of relatively high phosphate concentrations (Dell’Anno and Danavaro 2005). Analysis of a marine community revealed that bacteria attached to particulates or associated with another marine habitat were frequently stronger producers of APs than free-living bacteria (Ivanova et al. 2005). In fact, the bacterial APs can account for a relatively higher uptake of Pi from degradable organic matters (Paerl and Merkel 1982). The activity of constitutive APs that is often irrespective of the environmental phosphate concentration is the result of AP expression under symbiotic conditions, and the AP may have a role in nutrient exchange with host organisms rather than in nutrient uptake from the host tissues (Aono et al. 2004; Plisova et al. 2005). Thus, the marine bacterium Cobetia marina strain KMM 296 (Collection of Marine Microorganisms PIBOC FEB RAS), which was recovered from mussels, appeared to have metabolic specialization by producing a highly active extracellular AP (specific activity of 15,000 U/mg protein). This enzyme may play an active role in the invertebrate’s shell mineralization during seasonal phosphate diminution, when the bone-like exoskeleton is broken down and continuously renewed (Leveque et al. 2004; Plisova et al. 2005; Dorozhkin 2009). The strain C. marina KMM 296 did not require Na+ for growth and could grow at the wide range of temperatures from 4 to 42 °C, which is in contrast to other seawater isolates of Halomonas-like strains (Ivanova et al. 2005). The AP from C. marina KMM 296 (CmAP) was found to be relatively heat-stable for a psychrophile and functioned independent of extraneous bivalent metal ions and high inorganic phosphorus (Pi) or EDTA concentrations in the incubation medium (Plisova et al. 2005). Only a limited number of marine bacterial APs have been reported to be similar to CmAP in that they have unusual structures and cold-adapted properties such as a low thermostability and high catalytic efficiency (Plisova et al. 2005; Nasu et al. 2012). There are a few experimentally studied marine bacterial cold-active APs from Antarctic strains (Kobori et al. 1984; Rina et al. 2000), Vibrio sp. G15-21 (Hauksson et al. 2000), Shewanella sp. (Murakawa et al. 2002), Shewanella sp. SIB1 (Suzuki et al. 2005), Holomonas sp. 593 (Ishibashi et al. 2005) and C. marina KMM 296 (Plisova et al. 2005). All of these enzymes belong to γ-Proteobacteria, with the exception of an unidentified Antarctic bacterium species TAB5 (Rina et al. 2000).

We have developed a method for the production of a highly active recombinant AP of C. marina (CmAP) in Escherichia coli. To gain insight into the function of this unique enzyme, we report the results of its biochemical and catalytic characterization, site-specific mutagenesis experiments and molecular homology modeling data. In addition, we show the distinctive features of CmAP that make it important to the commercial application.

Materials and Methods

Bacterial Strains and Plasmids

C. marina KMM 296 (Collection of Marine Microorganisms PIBOC FEB RAS) cells were the source of chromosomal DNA for the CmAP gene isolation. The strain was cultivated by the shake-flask method in a 1,000-mL Erlenmeyer flask on a rotary shaker (160 rpm) at 25 °C for 36 h. The bacterial cells were grown in a culture medium containing 5 g peptone, 2.5 g yeast extract, 1 g glucose, 0.05 g Mg2SO4, and 0.2 g K2HPO4 in 750 mL seawater and 250 mL distilled water. E. coli DH5α and Rosetta(DE3) (Novagen) cells were used as hosts for propagation and expression of the recombinant plasmid Pho40, respectively. The recombinant plasmid 40Ph of 7,699 bp for the synthesis of the recombinant protein CmAP consists of the NcoI/EcoRI-fragment of plasmid pET-40b (+) (Novagen) and the gene corresponding to the mature CmAP open reading frame.

Recombinant DNA Technique

For the CmAP gene amplification, Encyclo Taq Polymerase (Eurogen) and the gene-specific upstream primer, 5′-TTAACCATGGCAGAGATCAAGAATGTCATTCTGAT-3′, and the downstream primer, 5′-TTAAGAATTCCTTCGCTACCACTGTCTTCAGATACTGTCC-3′, were used. For preparing the CmAP mutants, the single amino acid substitutions Asp12Gly, Asp64Gly, Asp443Gly, Thr118Gly, Glu268Gly, Asp315Gly, Trp274Gly, His316Gly, and Asp273Gly were introduced into the 35-mer amino-modified oligonucleotides complemented to both DNA chains used in polymerase chain reaction (PCR). Then, the resultant PCR products appeared for both as DNA matrix and primers in equivalent ratio for the next PCR. The resultant mutant gene was amplified with the gene-specific primers as described above. PCR was carried out in automatic amplifier Eppendorf. Restriction endonucleases and T4 DNA ligase were purchased from Fermentas. The PCR product and plasmid pET-40b (+) restricted were purified on agarose gel using a Qiagen column. The recombinant insert of the resultant plasmid Pho40 was sequenced using the automated PE/ABI 310 DNA sequencer and the PE/ABI-ABI PRISM BigDye Terminator cycle sequencing Ready Reaction Kit (PE Applied Biosystems). Preparation of E. coli competent cells and heat shock transformation were carried out according to the standard method (Sambrook et al. 1989). All biochemicals and reagents were purchased from Fermentas and Sigma-Aldrich.

Recombinant AP Expression

For the expression of CmAP and its mutants Asp12Gly, Asp64Gly, Asp443Gly, Thr118Gly, Glu268Gly, Asp315Gly, Trp274Gly, His316Gly, and Asp273Gly, the E. coli Rosetta(DE3) cells that transformed with the recombinant plasmid Pho40 (E. coli Rosetta(DE3)/Pho40) were grown on lysogeny broth (LB) plates containing 25 mg/mL kanamycin overnight at 37 °C. A single colony was selected and grown at 200 rpt in 20 mL of LB with 25 mg/mL kanamycin 37 °C for 12 h and then transferred to 1 L of fresh LB with 25 mg/mL kanamycin. When the cell density reached an A600 of 0.6–0.8, 0.2 mM IPTG was added to induce the expression of the protein, and the incubation continued at 16 °C up to 12 h at 200 rpt. The E. coli Rosetta(DE3) cells were transformed with the pET-40b (+) plasmid as a control.

Recombinant AP Purification

All purification steps were carried out at 6 °C. After harvesting, the E. coli Rosetta(DE3)/Pho40 cells were resuspended in 150 mL of a buffer containing 0.05 M Tris–HCl, pH 8.6, and 0.01 % NaN3 (buffer A) and disintegrated by ultrasonic treatment, then centrifuged at 10,000g for 30 min. The supernatant was applied to a column (4 × 25 cm) of DEAE-52 cellulose (Whatman). Elution of the protein was performed by a gradient of NaCl concentration (0.05–0.38 M) in 18 column volumes (CV) of the buffer A. The enzymatically active fractions were collected and applied onto a column (1 × 2.5 cm) of Ni-agarose (Qiagen). Elution of the protein was carried out by buffer B (50 mM Tris–HCl, pH 8.6, 50 mM EDTA, 0.01 % NaN3). The active fractions were collected and desalted by using a DEAE-Toyopearl 650 M (Toyo Soda) column (0.7 × 2.5 cm) and then incubated with enterokinase (Invitrogen) at 21 °C for 15 h. Then, the protein solution was applied onto a gel filtration column (1.5 × 170 cm) of Sephacryl S-100HR (Sigma) with the buffer C (50 mM Tris–HCl, pH 8.6, 100 mM NaCl, 0.01 % NaN3) at a flow of 0.13 mL/min, +6 °C. All purification steps were evaluated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The concentration of protein was determined according to Bradford (1976).

Enzyme Activity Assay

The standard assay for AP activity was carried out at 37 °C using 2 mM p-nitrophenylphosphate (pNPP) (Sigma Chemical Co.) in 0.1 M Tris–HCl buffer, pH 9.0, containing 0.2 M KCl. The release of p-nitrophenol (ε = 18.5 mM/cm) was monitored at 405 nm. One unit of AP activity was defined as the quantity of the enzyme required to release 1.0 μmol of p-nitrophenyl from pNPP in 1 min. The specific activity was calculated as units per 1 mg of protein.

Kinetic Parameters

Kinetic parameters of hydrolyzing reaction were verified by plotting reaction rates against concentrations of pNPP 0.1–0.6 and 0.2–3.0 mM for the enzyme assay in 0.1 M Tris–HCl buffer, 0.2 M KCl, at pH 8.0 and 10.3, respectively. For the enzyme activity assay carried out in 1 M diethanolamine (DEA) buffer at pH 8.0 and 10.3, the concentrations of pNPP were in the range of 0.1–1.0 and 0.5–15 mM, respectively. The K m, V max, and k cat values were determined by nonlinear regression analysis of the plots by using Origin 8.1. The extinction coefficients for p-nitrophenol were measured to be 18.5 (0.1 M Tris–HCl) and 18.7 mM/cm (1 M DEA) at pH 10.3 and 17.7 mM/cm at pH 8.0 in both buffers. All enzyme activity assays were carried out at 25 °C.

Molecular Mass Determination

The molecular size of the active enzyme was determined by gel filtration on a column of Sephacryl S-100HR (Sigma) (1.5 × 170 cm) in 50 mM Tris–HCl, pH 8.6, and 100 mM NaCl at a flow of 0.13 mL/min at 6 °C and calibrated using Bio-Rad standard molecular weight markers: thyroglobulin bovine (670 kDa), γ-globulin bovine (158 kDa), ovalbumin chicken (44 kDa), myoglobin horse (17 kDa), and vitamin B12 (1.35 kDa). The molecular mass of the enzyme was determined by 12.5 % PAGE. The protein preparation was mixed with Laemmli sample buffer in the absence of β-mercaptoethanol and with or without heat treatment at 95 °C and then applied to PAGE. The gels were stained according to Laemmli (1970). PAGE for the native CmAP was carried out in the absence or presence SDS in concentration from 0.1 to 1 % without heat treatment. The “native” gels were cased over with the standard enzyme assay buffer containing 2 mM pNPP and stained for AP. Both PAGEs were performed at 25 °C.

Temperature and pH Stability Assay

The apparent temperature optimum for the enzyme-specific activity was determined by measuring the activity at different temperatures in standard enzymatic assay buffer for 30 min. The reaction temperatures ranged from 20 to 65 °C. The half-life of the enzyme was determined as the temperature and the time required to half the initial activity. The enzyme was incubated at the appropriate temperatures for 10-min intervals, cooled in the ice bath, and the residual activity was measured using the standard enzyme assay at 25 °C for 30 min.

The effects of pH values on activity and stability of the enzyme were determined by measuring the activities at the different pH values by the standard method of the enzyme activity assay. The pH ranges varied from 4.0 to 11.0. The buffers used were 0.03 M NaCOOH3 (4.0–6.0), 0.05 M Tris–HCl (7.0–9.0), 0.1 M NaHCO3 (9.0–11.0), and 0.1 M glycine (9.0–11.0). The pH stability was evaluated by pre-incubation of the enzyme in each buffer for 1/2, 2, and 5 h and the residual activity measured using the standard enzyme assay.

Effects of Chemicals

The effects of Mg2+ on the temperature stability were evaluated by adding 2–10 mM MgCl2 to the enzyme pre-incubation mixture containing 50 mM Tris–HCl, pH 8.6, at each temperature for up to 5 h, then the remaining activity assayed by the standard method. The effect of EDTA on the activity and stability of the enzyme was evaluated by adding it to the standard reaction mixture for the enzyme activity assay at 37 °C for 30 min or pre-incubating the enzyme mixture containing 50 mM Tris–HCl, pH 8.6, with EDTA in different concentrations (0–100 mM), and then the remaining activity was assayed by the standard method. The effects of KCl, NaCl, and SDS on the activity of the enzyme were evaluated by adding the chemicals in the standard incubation mixture at different concentrations.

Metal Ion Analysis

For atomic absorption analysis, samples were desalted by gel filtration on a Sephadex G-25 column in Tris–HCl (pH 8.6) buffer. The concentration of protein for analysis was 0.22 mg/mL, and samples were acidified with 1 M HCl prior to injection. Shimadzu 6800 (Japan) atomic absorption spectrometer was employed to measure bivalent metal ions in flame atomization mode.

Statistics

All values presented in this article are representative of at least three independent experiments. Data were analyzed using the Student’s t test of the SigmaPlot 2000 version 6.0 program (SPSS Inc.). Differences from controls were considered significant at p ≤ 0.05.

Molecular Modeling

The target-template alignment customization of the modeling process and 3D model building of CmAP were carried out by the molecular graphics package Molecular Operating Environment version 2012.10 (Chemical Computing Group Inc., 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2012). The evaluation of structural parameters, contact structure analysis, physicochemical properties, and molecular docking and visualization of the results were carried out by the use of the modules of MOE v.2012. The molecular dynamics simulations were performed by using GROMACS v. 4.6.3 (Van der Spoel et al. 2005, 2013; Hess et al. 2008). The calculations of the pKa values were performed by PROPKA server (http://propka.ki.ku.dk/) (Li et al. 2005; Olsson et al. 2011).

Estimates of binding strengths and specificities were carried out for different ion types. Knowledge-based potentials for interactions of all nonhydrogen protein atoms with inorganic ions were obtained from a training set of Protein Data Bank (PDB) structures using Monte Carlo reference state (MCRS) as described previously (Rahmanov et al. 2010). The structure was covered by a cubic grid with a step of 0.2 Å and at grid nodes not sterically clashing with protein atoms. Probabilities of occupancy were calculated for different ions as a sum of inputs from all neighboring protein atoms. Conjugate gradient descent procedure was followed by random walk of decreasing steps to obtain the final maxima of ion probabilities of occurrence.

Results

Expression and Purification of CmAP

The CmAP production reached 2 mg of the final yield of the functionally active protein from 1 L of the E. coli Rosetta(DE3)/Pho40 cells culture carrying the expression recombinant vector 40Ph (Table 1). The optimal concentration of IPTG and temperature of the strain cultivation for the CmAP expression were 0.2 mM and 16 °C, respectively, for 12 h (data not shown). At the final step of purification, SDS-PAGE showed a single band of the protein with a molecular mass of approximately 55,000 Da corresponding to the mature CmAP (Fig. 1). The presence of 5 % mercaptoethanol, 10 % SDS, and 10 mM dithiothreitol or 6 M urea in the sample buffer did not have any effect on electrophoretic mobility of the CmAP band (data not shown).

Table 1 Isolation and purification of CmAP
Full size table
Fig. 1
figure 1

SDS-PAGE of CmAP: crude cell extract (A), enzymatic active fraction of hybrid recombinant protein 6xHis-DsbC-CmAP from DEAE-Cellulose-52 (B), protein fraction eluted from Ni-NTA agarose and processed by enterokinase (C), and CmAP eluted from Sephacryl S-100HR (D), Molecular weight marker (Bio-Rad) (M). Black cursor corresponds to hybrid recombinant protein 6xHis-DsbC-CmAP (87 kDa). Empty cursor corresponds to CmAP (55 kDa)

Full size image

Effect of pH and Temperature on CmAP Activity and Stability

The recombinant CmAP had optimal activity at pH 9.5–10.3 in 1 M DEA buffer with 15–20 mM of pNPP. CmAP was stable at pHs between 6.0 and 11.0 but was gradually inactivated at pHs below 6.0 (Fig. 2). CmAP was also gradually inhibited when the enzyme was exposed at above pH 8.5 maintained by 0.1 M glycine (Fig. 2).

Fig. 2
figure 2

Stability of CmAP in different buffers. The enzyme activity was measured using the standard enzyme assay (0.1 M Tris–HCl buffer, pH 9.0) after 5-h incubation at different pH values. Squares indicate 0.03 M NaCOOH3 (4.0–6.0), circles indicate 0.05 M Tris–HCl (7.0–9.0), diamonds indicate 1 M DEA (8.0–10.3), down triangle indicates 0.1 M NaHCO3 (9.0–11.0), and up triangle indicates 0.1 M glycine (9.0–11.0) (n = 9, p ≤ 0.05)

Full size image

CmAP had temperature optimums at 40 °C and 50 °C for the manifestation of its higher specific activity (Fig. 3). CmAP was a relatively heat-stable enzyme with a half-life of 27 min at 40 °C and 16 min at 45 °C in the presence of 2 mM EDTA (Fig. 4), while 50 % activity was lost in 45 min at 37 °C. However, CmAP was stable up to 45 °C with a half-life of 15 min at 47 °C and 120 min at 45 °C, when the enzyme was incubated with 10 mM MgCl2 (n = 15) (data not shown).

Fig. 3
figure 3

Effect of temperature on activity of CmAP. The enzyme activity was standard assayed (0.1 M Tris–HCl buffer, pH 9.0) using each temperature as the reaction temperature for 30 min (n = 15, p ≤ 0.05)

Full size image
Fig. 4
figure 4

Heat stability of CmAP. The enzyme was incubated at the appropriate temperatures in the presence of 2 mM EDTA and the residual activity measured using the standard enzyme assay (0.1 M Tris–HCl buffer, pH 9.0) at 10 min intervals for 30 min. Black squares indicate 20 °C, empty up triangle indicates 25 °C, empty circles indicate 30 °C, down triangle indicates 37 °C, crossed out diamond indicates 40 °C, and crossed out triangle indicates 45 °C (n = 9, p ≤ 0.05)

Full size image

Upon dilution of a thousand fold, CmAP was fully stable at −20 °C during more than 1-year storage in the dilution buffer (0.05 M Tris–HCl, pH 8.6, 0.01 % NaN3).

Mg2+ Content of CmAP

According to the results of atomic absorption spectrometry, only zinc and magnesium were found in the protein samples (n = 4). The molar ratio of Mg2+ in CmAP was determined as 1.2 ± 0.5.

Effects of Chemicals on CmAP

The incubation of CmAP with 10–100 mM EDTA for 5 h decreased the activity only by 40 % (Fig. 5). The elution of CmAP from Ni-agarose with 50 mM EDTA (pH 8.6) did not have any effect on the CmAP activity, whereas the pre-incubation of the resin-bound CmAP with 100 mM EDTA required the addition of 2 mM Mg2+ to retain 100 % of its activity (n = 4) (data not shown). Incubation with 0.1–5 % of SDS decreased the CmAP activity also by 40 %, indicating that CmAP is SDS-resistant (n = 6) (data not shown). PAGE of the native protein in the presence or absence of SDS followed by staining of the gels with the substrate yielded a single active band of CmAP (data not shown). CmAP was activated 2-fold by addition of 0.1–0.2 M of NaCl or KCl. Increase of NaCl concentrations more than 0.2 M resulted in a partial inactivation of CmAP, whereas a slight increase of activation occurred even at 1 M KCl (Fig. 6).

Fig. 5
figure 5

Effect of EDTA on CmAP activity and stability. The enzyme activity was measured using the standard assay with the addition of EDTA in the incubation mixture (diamonds); the enzyme was pre-incubated with EDTA concentration from 0 to 100 mM at 25 °C for 5 h and the residual activity measured using the standard enzyme assay (0.1 M Tris–HCl buffer, pH 9.0) (n = 6, p ≤ 0.05) (squares)

Full size image
Fig. 6
figure 6

Effect of Na+ and K+ on the CmAP activity. Effect of NaCl (squares); effect of KCl (circles). The enzyme activity was standard assayed (0.1 M Tris–HCl buffer, pH 9.0) (n = 6, p ≤ 0.05)

Full size image

Kinetic Properties of CmAP

Kinetic factors for CmAP were determined under various conditions in 0.1 M Tris–HCl and 1 M DEA buffers at 25 °C, by varying pH values from 8.0 to 10.3. CmAP had turnover number k cat = 28,300 ± 2,100 s and Michaelis–Menten factor K m = 13.2 ± 1.48 mM in 1 M DEA buffer with catalytic efficiency k cat/K m of 2.1 × 106 s/M for pNPP (Table 2). Catalytic efficiency and K m value of CmAP at pH 8.0 in 1 M DEA buffer were 8.2 × 107 s/M and 0.3 mM ± 0.02, respectively. CmAP had lower catalytic parameters under all conditions in 0.1 M Tris–HCl buffers (Table 2).

Table 2 Kinetic factors of CmAP at 25 °C
Full size table

Primary Structure of CmAP and Comparison with Other AP Sequences

The single AP gene consisting of 1,605 bp and encoding 535 amino acid residues (aa) of the AP precursor (accession no. DQ435608) was found in C. marina. A 32 aa signal sequence was identified that would direct the enzyme to the periplasmic space. The mature protein has a calculated molecular mass of 54,370 Da and an isoelectric point (pI) of 4.56, which are in a good agreement with experimental data (Plisova et al. 2005). The ORF had a G + C content of 56.64 % and encoded a high content of nonpolar and uncharged polar aa, such as Ala (11.33 %), Gly (8.75 %), Leu (8.55 %), Ser (6.76 %), Val (6.56 %), Thr (6.36 %), and negatively charged Asp (7.36 %). A search of the GenBank database disclosed the high identity (53–75 %) and similarity (69–86 %) of CmAP to several explored and putative APs from marine γ-Proteobacteria, namely halophilic Halomonas sp. #593 (accession no. AB271127) and Halomonas elongate (FN869568), Vibrio parahemoliticus (BA000031), psychrophilic Vibrio sp. G15-21 (AF352014), Shewanella denitrificans (CP000302), Psychromonas ingrahamii (AAQS01000001), and Hahella chejuensis (CP000155) (data not shown). Only the Vibrio G15-21 AP was completely biochemically and structurally evaluated (Hauksson et al. 2000; Helland et al. 2009). A relatively lower identity (≤38 %) and similarity (≤52 %) were observed for other APs, including AP from Antarctic bacterium TAB5 and mammalian and marine invertebrate APs (Rina et al. 2000; Le Du et al. 2001; De Backer et al. 2002; Kozlenkov et al. 2002).

Molecular Modeling and Site-Specific Mutagenesis of CmAP

Alignment scores and stereochemical quality of the CmAP models showed the Vibrio sp. G15-21 alkaline phosphatase (VAP) as the best template with 69.4 % identity and 82 % similarity (Fig. 7a). The CmAP model building process was completed on the base of the high-quality experimental structure of VAP at 1.4-Å resolution (PDB code: 3E2D_A), showing that the bimetallo core structure was conserved (Fig. 7b). The Cα root mean square deviation (RMSD) value between CmAP and VAP was calculated to be 0.43 Å. Although catalytic Ser65 and substrate-binding Arg129 and the residues bound with the metal ions were identical, the slight differences were observed within the secondary structure. CmAP has a higher content of α-helices and a lower content of β-strand elements than VAP does (Table 3). CmAP displays a three-layer β-sandwich fold and contains β-sheet of nine β-strands surrounded by α-helices at both sides (α-β-α type). The CmAP structure is stabilized by 148 hydrogen bonds, 180 hydrophobic contacts, and 20 ionic bonds with the absence of disulfide bonds.

Fig. 7
figure 7

a Sequence alignment of the alkaline phosphatase CmAP from marine bacterium C. marina (UniProt: Q1W622) with template sequence of alkaline phosphatase VAP from marine bacterium Vibrio sp. G-15 (PDB code: 3E2D_A). Figure prepared with program Maestro 9.3. Secondary structure elements were shown for VAP. Identical and similar residues are shown with red and orange colors. b The 3D structure model of the alkaline phosphatase CmAP monomer from marine bacterium C. marina built with the use of the crystal structure of VAP from marine bacterium Vibrio sp. G15-21 as template. Protein structure is shown as ribbon diagram. Secondary structure elements are β-strands (yellow color), α-helixes (red color), turns (blue color), and unordered structure (gray color). Metal ions Zn2+ (1), Zn2+ (2), and Mg2+ (3) are shown as blue and brown spheres (color figure online)

Full size image
Table 3 Content of the secondary structure elements in CmAP and VAP
Full size table

CmAP had the same large “crown” domain and the long extension loops that cover the active site and can participate in the dimer formation as found in VAP (Fig. 7b) (Helland et al. 2009). However, the loop for the dimer stabilization (325–351 aa) was very movable depending on the molecular environment conditions (Figs. 7a and 8). As in the case of VAP, the CmAP catalytic site was deeply buried in a crevice running between two loops. Although the overall molecular surface of CmAP was largely hydrophobic, the calculation of the electrostatic surface potentials near the active site revealed the presence of a large extended surface of the positively charged aa along the narrow entrance to the enzymatic cave (Fig. 9). It may effectively cumulate the substrate molecules from the solution in the active center direction. On the other hand, the negatively charged aa located opposite the bound negatively charged Pi, facilitating its release from the active site (data not shown). A distinctive feature of the enzyme was also found by analysis of CmAP protonation at various pH values. At the pH values ≥10.0, the substrate-binding Arg129 and the residue Tyr441 oriented toward the bound substrate become neutral and negatively charged, respectively (data not shown).

Fig. 8
figure 8

Conformational flexibility of CmAP. a Monomeric CmAP: active centre (1) and the subunit loop stabilizing the dimer (2). b Dimeric CmAP: active centers (1)

Full size image
Fig. 9
figure 9

Electrostatic neighbored surface potentials of CmAP active site bound with substrate pNNP. Positively charged amino acid residues are indicated by blue color. Negatively charged amino acid residues are indicated by red color (color figure online)

Full size image

Estimates of binding strengths and specificities carried out for different ion types defined localization of two Zn2+ and one Mg2+ (data not show). In silico mutagenesis of CmAP Mg2+-binding site was carried out. The point mutations of the residues Asp12, Asp273, Asp315, His316, Thr118, Glu268, and Trp274 to Gly generated with program MOE showed that the Asp12Gly mutation should lead to the loss of ionic bonds of the residue with metal ions Zn1, Zn2, and Mg, hydrogen bonds Asp12.OD2-Ser65.OG, Asp12.OD2-Thr.OG1, and Asp12.OD1-His316.NE2 and hydrogen bonds with three water molecules in the active site, which are present in the Asp12 contacts (Fig. 10a, b). The Asp64Gly mutation led to the loss of ionic bond and hydrogen bond with Arg129 and hydrogen bonds with two water molecules in the active site, which are present in the Asp64 contacts (Fig. 10a, c). The Asp443Gly mutation led to the loss of hydrogen bond Asp443.OD2-Ser278.OG1 and hydrogen bonds with two water molecules, which are present in the Asp443 contacts (Fig. 10a, d). The Thr118Gly mutation led to the loss of hydrogen bond Thr118.OG1-Glu268.OE2 and hydrogen bonds with three water molecules in the active site, which are present in the Thr118 contacts (Fig. 10a, e). The Glu268Gly mutation led to the loss of ionic bond with Mg2+ and hydrogen bonds with two water molecules in the active site, which are present in the Glu268 contacts (Fig. 10a, f). The Asp315Gly mutation led to the loss of ionic bond Gly315 with ion Zn2 and hydrogen bond with Gly13.N, which are present in the Asp315 contacts (Fig. 10a, g). The Trp274Gly mutation led to the loss of hydrogen bond with Ser278 and hydrogen bond with water molecule in the active site, which are present in the Trp274 contacts (Fig. 10a, h). The His316Gly mutation led to the loss of hydrogen bond His316.NE2-Asp12.OD1, which is present in the His316 contacts (Fig. 10a, i). The Asp273Gly mutation led to the loss of ionic bonds with ions Zn1 and Zn2 and hydrogen bond with Thr318.OG1, which are present in the Asp273 contacts (Fig. 10a, j). The substitution of Gly for the residues Asp12, Asp273, Asp315, His316, Tre118, Glu268, and Trp274 was also experimentally carried out in order to elucidate their role for the CmAP activity. Where the interruption of the residue interaction with Mg2+ occurred, the enzymatic activity was completely lost, although the integrity of the molecular structure remained and only one or two hydrogen bounds were broken in the entire CmAP hydrogen bond network (Fig. 10b, e–j). Only the enzymatic activity of Asp64Gly and Asp443Gly mutants, where the breaking of nonessential for Mg2+-binding hydrogen bonds occurred in the periphery of the structure, was reduced 7-fold and 1/2-fold, respectively (Fig. 10c, d).

Fig. 10
figure 10figure 10

a, b, c, d, e, f Atomic detailed structures of CmAP active center and metal-binding sites (a) and its mutants Asp12Gly (b), Asp64Gly (c), Asp443Gly (d), Thr118Gly (e), and Glu268Gly (f) in silico generated with MOE 2012.10. The coordinates of the metal ions and sulfate ion produced by superposition model CmAP with crystal structure of VAP. Water is shown as red sphere. Metal ions Zn2+ (1), Zn2+ (2), and Mg2+ (3) were shown as blue and brown spheres. g, h, i, j Atomic detailed structures of the active center and metal-binding sites of CmAP mutants Asp315Gly (g), Trp274Gly (h), His316Gly (i), and Asp273Gly (j) in silico generated with MOE 2012.10. The coordinates of the metal ions and sulfate ion produced by superposition CmAP model with crystal structure of VAP. Water is shown as red sphere. Metal ions Zn2+ (1), Zn2+ (2) and Mg2+ (3) were shown as blue and brown spheres (color figure online)

Full size image

Molecular docking CmAP with AMP has shown that the residue Trp274 localized near the active site entrance forms stacking interaction with the substrate ring that may facilitate the optimal space orientation of the substrate toward the active site, indicating specificity to DNA (Fig. 11). While docking with pNNP, the only PO4 − connection in the CmAP active site occurs (Figs. 9 and 11).

Fig. 11
figure 11

Diagram of contacts of the substrate AMP interactions in CmAP active site

Full size image

Discussion

It is evident that marine bacterial phosphatases possess unique properties that contribute greatly to living and functioning in oceanic waters. These enzymes appear to be distinct from their terrestrial counterparts in substrate preference, metal binding ligands, Pi affinity, thermostability, and catalytic efficiency. The marine bacterial APs such as the Antarctic psychrophile TAB5 (Rina et al. 2000), the psychrophile Vibrio G15-21 (Hauksson et al. 2000), the psychrophiles Shewanella sp. (Murakawa et al. 2002) and Shewanella sp. SIB1 (Suzuki et al. 2005), and the moderately halophilic Halomonas sp. 593 (Ishibashi et al. 2005) have been purified and biochemically studied. The activities of these enzymes are in the range of 427–3,707 U/mg. There is no comparison of the specific activity between the terrestrial and marine γ-Proteobacterial APs. The specific activity of native E. coli AP was approximately 60 U/mg, whereas the native CmAP activity was approximately 14,000–15,000 U/mg (Plisova et al. 2005). Of the commercially available APs, the most active is the recombinant AP from bovine calf intestinal mucosa; it showed a specific activity of approximately 7,800 U/mg (Biozyme Laboratories). Currently, recombinant C. marina AP (CmAPase) was expressed by Nasu et al. (2012) using the synthetic gene adapted for E. coli and inserted in pET-16b(+) vector and had a specific activity of 12,700 U/mg, which is the highest activity reported for any bacterial or mammalian AP. The specific activity and catalytic efficiency of our recombinant enzyme (CmAP) produced in E. coli on the base of the nature gene of C. marina and pET-40b(+) vector were also the highest among any of the APs studied to date (Tables 1 and 2). The kinetic parameters of CmAP were 0.3–13.2 mM (K m) and 24,000–28,300 s (k cat) with the catalytic efficiency of 2.1 × 106–8.2 × 107 s/M (Table 2). For comparison, the K m of E. coli AP was 0.06 mM and a k cat of 62 s with the catalytic efficiency of 1.1 × 106, whereas the K m of bovine intestinal AP, including its catalytically improved mutants, ranged from 1.3 to 3.9 mM and the k cat ranged from 1,800 to 6,100 s with the catalytic efficiency of 1.4–1.6 × 106 s/M (Hauksson et al. 2000; Kozlenkov et al. 2002).

Surprisingly, the native C. marina AP signal peptide was found to be appropriate for E. coli system to produce periplasmic recombinant protein (Nasu et al. 2012). Alternatively, the use of E. coli signal peptide included in the expression plasmid pET40b(+) resulted in overproduction of the soluble highly active CmAP (Table 1). It should be noted that removing of the N-terminal His-tag and 32.5 kDa DsbC parts from the hybrid protein 6xHis-DsbC-CmAP could be optional due to the absence of their effect on the CmAP activity (Fig. 1). However, the recombinant plasmid pET22b(+) containing the mature CmAP sequence and a C-terminal 6xHis tag produced CmAP in body inclusion despite of the presence of N-terminal E. coli leader (data not shown). The additional C-terminal His-tag sequence probably decreased the CmAP solubility. The hydrophobic tail of native CmAP may be required to prevent secretion of the enzyme into the sea medium or to participate in protein folding (Kim et al. 1998). It was reported that natural marine bacteria have a cell surface phosphatase-like enzyme that rapidly hydrolyses 5′-nucleotides and regenerates Pi, increasing its uptake while decreasing the concentration of Pi in the environment (Ammerman and Azam 1985).

From our data, the relative molecular mass of CmAP evaluated by native-PAGE, SDS-PAGE, and gel filtration column chromatography under pH 8.6 was to be 55,000 Da, suggesting the one-subunit molecular organization of the enzymatically active CmAP under the experimental conditions. However, the gel permeation of a recombinant C. marina AP (CmAPase) carried out by Nasu et al. (2012) at pH 10.0 resulted in the formation of the enzyme dimer. It seems probable that more flexible links at the dimer interface of the cold-adapted enzymes may ease the structural rearrangement and thereby facilitating more rapid release of phosphate and thus catalytic turnover (Asgeirsson et al. 2003). Most known APs are active as homodimeric metalloenzymes that contain zinc and magnesium ions. Their active site pockets are conserved with the exception of a few key residues directly or indirectly involved in the coordination of the metal ion in the third ion binding site (M3), which is normally occupied by magnesium (Zalatan et al. 2008). A few APs have been reported to be monomeric, and some have been reported that were stable as a monomer but not active (Hulett et al. 1991; Hauksson et al. 2000; Zappa et al. 2001). Interestingly, Sphingomonas sp. AP (SPAP) was eluted as an active monomer in size exclusion chromatography experiments, although crystal formation results in visualization of a dimer (Bihani et al. 2011). It has been suggested that the dimerization found in the crystals is a crystallization artifact because surface area buried upon dimerization is lower than in other AP’s. Distinct from the majority of known APs, the marine γ-Proteobacterial AP dimer formation was archived by another mechanism. Introduction of bulky side chains in the subunit interface on two additional inserts contributed to VAP dimer formation (Helland et al. 2009). The absence of the stabilizing N-terminal helix, which is found in other APs, was compensated for by additional aa inserts. CmAP model showed the same large crown domain and the long extension loops that cover the active site and can connect the monomers. The CmAP 3D structure was predicted by the homology modeling approach owing to availability of a highly homologous VAP crystal structure as the only 3D structure of a marine γ-Proteobacterial AP (Fig. 7). Current approach to refining protein structure models with templates from the Protein Data Bank (PDB) can significantly improve the accuracy of the structures providing often better solutions to the X-ray crystallographic phase problem (Qian et al. 2007; Moult 2007). Although catalytic Ser65 and substrate-binding Arg129 and the residues bound with the metal ions Zn1, Zn2 and Mg in CmAP and VAP were identical, the slight differences were observed within the primary and secondary structures. The analysis of nonidentical aa region that was predominantly found on the molecular surfaces of the compared proteins showed an increase from 2- to 10-fold in Ala, Val, Arg, Pro, Gly, and Phe and otherwise decrease in Lys, Ile, Asn, and Tyr in CmAP (Fig. 7a). In addition, CmAP had 2 % difference in the content of the secondary structure elements and hydrophobic aa covering molecular surface (Table 3, Fig. 7a). The minor structural differences in CmAP and VAP probably define the apparent differences in their properties. CmAP exhibited 4- and 1/2-fold higher thermal stability with a half-life of 27 min versus the VAP half-life of only 6 min at 40 °C (Fig. 4) (Hauksson et al. 2000). Heat stability was increased up to 120 min at 45 °C for the loss of 50 % of the CmAP initial activity in the presence of stabilizer Mg2+ (data not shown). Although CmAPase had lower activity than CmAP that could be the result of the recombinant production method (Nasu et al. 2012). CmAP is more close to AP from moderate halophile Halomonas sp. 593 (HaALP) as judged by the aa sequence (75 % identity, 86 % similarity) as well as thermostability and halotolerance (Ishibashi et al. 2005). However, in distinction from HaALP, KCl was more effective than NaCl for activation and stabilization of CmAP, suggesting the importance of K+ ions for the CmAP functioning environment (Fig. 6). The important role of K+ in apoptose and proliferation of cells in mineralized tissues that are under continuous renewal was recently determined (Page and Di Sera 2006; Henney et al. 2009).

It is probable that the structural features of the CmAP active site provide a barrier to prevent contact of denaturants, including chelators, with the active site. Metal ions released from the active site by heat or pre-incubation with EDTA in high concentrations for long time only led to decrease of enzymatic activity (Figs. 3, 4, and 5). These may coincide with the loss of the magnesium ion from the active site because of increased mobility of the binding ligands (Gudjónsdóttir and Asgeirsson 2008). However, CmAP retained its stability and activity after pre-incubation with EDTA in concentration of 50–100 mM in the presence of Mg2+ (data not shown). It has been suggested that CmAP does not require an excess of Mg2+ for binding in the M3 site, which has been generally noticed in the structural reorganization that occurs upon metal binding during enzyme dimerization and activation (Janeway et al. 1993; Poltorak et al. 1999; Babor et al. 2005). The high working efficacy and stability of a monomeric CmAP can be attributed to the intrinsically tight binding of both catalytic and structural metal ions. Strong metal–enzyme interactions in the AP active site have been reported (Zappa et al. 2001; Zaheer et al. 2009). Estimates of binding strengths and specificities of metal ions using the CmAP 3D structure model and atomic absorption analysis confirmed two zinc and one magnesium ions and their ligands in the active site (Figs. 7 and 10). However, the existence of the additional Mg2+-binding sites with possible functional role in CmAP may well be true considering the effect of exogenous Mg2+ on the CmAP thermal stability.

Amino acid residues that interact with the zinc ions in the M1 and M2 sites of APs are conserved in all known sequences. His116 and Trp274 occurred near the Mg2+-binding site (M3) and identified in CmAP have been mutated in VAP (Gudjónsdóttir and Asgeirsson 2008). The structural study of the CmAP mutants with the predicted removal of the Mg2+ site revealed differences in metal coordination. Remarkably, given its large detrimental effect on the substrate hydrolysis, mutation of the CmAP that predicted Mg2+-binding ligands to Gly appears to be linked to more general structural factors (Fig. 10). Each intermolecular Mg2+ contact is directly or indirectly implicated in all metal ion coordination and included in the unified hydrogen bond network penetrated through the entire molecular structure. The alteration in the metal coordination is directly related to alterations of Pi affinity in the active site of the enzyme (Wang et al. 2007). Furthermore, a little or no cooperativity between the substrate-binding Arg and the Mg2+ site in their contributions to catalysis has been found (Zalatan et al. 2008). It could explain the drastic increase in catalytic turnover and K m at pH ≥ 10.0 in CmAP (Table 2). When the Mg2+ ion stabilization effect on the transferred phosphoryl group and the Zn2+ bimetallo site interactions remained, the residue Arg129 became neutral (Figs. 10 and 11). As the release of Pi is the slow step in the mechanism at alkaline pH, the weakened binding of Pi together with the additional negatively charged Tyr441 at a high pH should contribute to rapid release of negatively charged product from the CmAP active center. This increases not only K m but also catalytic turnover (k cat) many times. However, catalytic efficiency of CmAP was higher at pH 8.0 in 1 M DEA buffer because the enzyme had a significantly lower K m value (Table 2). Remarkably, the residue Tyr367 is part of the subunit interface in the human placenta AP (PLAP) dimer, where it protrudes from one subunit and its hydroxyl group is located 6.1 Å from the phosphate and 3.1 Å from His432, which in turn chelates the zinc atom Zn1 in the active site of the other subunit, playing a crucial role in determining the allosteric character of PLAP (Millan 2006). Tyr325 from the VAP B-subunit extends into the A-subunit substrate-binding site with the hydroxyl group 4.69 Å from Zn at the M1 position (Helland et al. 2009). However, the residue Tyr441 also protrudes above the CmAP active site entrance and may play analogous role for functioning as monomer (Fig. 9).

Although APs catalyze the hydrolysis of a wide variety of phosphomonoesters at similar rates, CmAP prefers substrates such as pNPP, 5′-AMP, 5′-CMP, α-glycerophosphate (Plisova et al. 2005). The narrowest specificity was found in APs from mammalian sources. Distinct from bacterial APs, the calf intestinal AP only catalyzes the hydrolysis of pNPP and 5′-AMP at a high rate, whereas all other substrates are cleaved much slower (McComb et al. 1979). It is evident that the rate of hydrolysis with CmAP decreases with the increase of the size of the substrate molecular due to the narrow entrance into the active site (Figs. 9 and 11). However, CmAP was found to hydrolyze DNA effectively as well as the current luminescent substrates used in the majority of the AP methods (Plisova et al. 2005; Nasu et al. 2012).

It is evident that CmAP is an example of structural alterations in APs adapted to function in the marine environment. Taking advantage of enzymatic properties that nature has evolved, CmAP can be used in a number of biotechnology approaches. A relatively low thermostability can be useful for AP removing from the reaction mixture without additional purification in DNA techniques (Plisova et al. 2005). A high specific activity value under transphorylation conditions and the metal-cofactor independence as well as the monomeric station of AP can be useful for rapid monitoring and optimization of various aspects of recombinant protein production by tagging AP to the protein of interest via genetic fusion (unpublished data). Similarly, a broad pH stability and a high stability against chelate agents can be useful in recombinant protein purification. The feasibility of using the genetic fusions of AP monomer with peptides and proteins as a dual functional tag, as both an affinity tag and as a label in the development of an enzymatic assay to detect a ligand of interest, can be used in clinical diagnostics (unpublished data).

In summary, the detailed experimental and theoretical/computational investigations of marine CmAP suggest that their advances in structural arrangement, biochemical properties, and catalytic efficacy can be successfully used in such practical applications as immunoassay systems, Southern blotting, Western blotting, reporter gene assays, and ELISAs. Based on the understanding of the enzyme structure and function, it may be possible to design more efficient enzymes or enzymes with novel functionalities. This knowledge may have wide implications in enzyme engineering as well as drug design.