A Novel Bifunctional Hybrid with Marine Bacterium Alkaline Phosphatase and Far Eastern Holothurian Mannan-Binding Lectin Activities

A fusion between the genes encoding the marine bacterium Cobetia marina alkaline phosphatase (CmAP) and Far Eastern holothurian Apostichopus japonicus mannan-binding C-type lectin (MBL-AJ) was performed. Expression of the fusion gene in E. coli cells resulted in yield of soluble recombinant chimeric protein CmAP/MBL-AJ with the high alkaline phosphatase activity and specificity of the lectin MBL-AJ. The bifunctional hybrid CmAP/MBL-AJ was produced as a dimer with the molecular mass of 200 kDa. The CmAP/MBL-AJ dimer model showed the two-subunit lectin part that is associated with two molecules of alkaline phosphatase functioning independently from each other. The highly active CmAP label genetically linked to MBL-AJ has advantaged the lectin-binding assay in its sensitivity and time. The double substitution A156N/F159K in the lectin domain of CmAP/MBL-AJ has enhanced its lectin activity by 25±5%. The bifunctional hybrid holothurian's lectin could be promising tool for developing non-invasive methods for biological markers assessment, particularly for improving the MBL-AJ-based method for early detection of a malignant condition in cervical specimens.


Introduction
The peculiar characteristic to recognize and bind specific carbohydrates made lectins of animal and plant origin useful tools for detecting changes in carbohydrate profiles and identifying aberrant glycans in neoplastic cells with the aim of more precise diagnostics and more accurate prognosis. The technique most common and widespread is the use of lectins in immunohistochemical assays [1], [2], [3], [4]. Molecules with a narrow specificity, which are able to bind selectively to carbohydrates, have also a key importance in the development of research related with mechanism of cancerogenesis or inflammation at the molecular level as well as for designing drugs targeted to a relevant molecule. In consequence of the strong innate defense system and the absence of the adaptive immunity in marine invertebrates, a number of their lectins were found to play considerable biological recognition role and therefore have unique specific activities. Several methods regarding the use of marine invertebrate lectins, including mannan-binding C-type lectin from Far Eastern holothurian MBL-AJ, as tools for recognizing aberrant glycans or foreign microbial structures have been proposed in recent days [5], [6], [7], [8], [9], [10], [11], [12]. In addition, MBL-AJ was successfully applied for differential diagnostics of benign and malignant neoplasms of uterine cervix by the analysis of contents of lectin-binding carcinoembryonic antigen (CEA) in vaginal secretion [13]. The native MBL-AJ has oligomeric forms with identical 17-kDa subunits and natural ligands among extracellular low-branched bacterial a-D-mannans composed of a-1,2and a-1,6-linked D-mannose residues, participating in recognition of the altered structures in the invertebrate during embryogenesis, morphogenesis and the formation of the immune response [5], [10], [11]. Cross-reactivity of MBL-AJ and human serum lectin MBL detected by the antibodies against MBL-AJ suggested the presence of common antigenic determinants [5]. However, the MBL-AJ specificities resulted in the absence of the MBL-AJ interaction with the components of the healthy patient's serum have been found to facilitate the detection of the slight structural differences of glycans, excluding the wrong positive results in the assayed samples [13]. Although cancers of ovaries, cervix and uterus are regarded as difficult and expensive for the detection at an early stage [13], [14], [15], [16], the method with the use of MBL-AJ has allowed identifying statistically reliable differences between the levels of the lectin-binding CEA between healthy women and patients with cervical cancer, and between patients with benign and malignant neoplasm [13]. Moreover, it is important that MBL-AJ-based method gives a possibility to assess tumours noninvasively.
To improve the MBL-AJ-binding assay for detection of aberrant patterns of glycosylation a novel bifunctional hybrid protein CmAP/MBL-AJ based on the Far Eastern holothurian lectin MBL-AJ gene fused with the gene of the marine bacterium alkaline phosphatase CmAP has been synthesized with E. coli cells. CmAP/MBL-AJ has the binding activity of lectin improved by mutations and the high enzymatic activity of alkaline phosphatase facilitating the identification of MBL-AJ ligands without the additional steps of the MBL-AJ-specific antibodies production and conjugation with an enzyme.

Construction of recombinant plasmids
Expression plasmid encoding the fusion protein CmAP/MBL-AJ has been constructed on the base of NcoI/SalI part of pET-40b(+) vector (Novagen) and the chimeric gene comprised the mature CmAP open reading frame (accession n. DQ435608) and cDNA fragment encoding mature lectin MBL-AJ (accession n. AAT42221) linked by the sequence encoding the peptide linker (G 4 S) 3 . For the CmAP gene amplification, marine bacterium Cobetia marina KMM 296 (Collection of Marine Microorganisms PIBOC FEB RAS) chromosomal DNA, Encyclo Taq Polymerase (Evrogen), and the gene-specific upstream primer, 59-TATTCCATGGCA-GAGATCAAGAATGTCATTCTGAT-39, and downstream primer including the linker (G 4 S) 3 , 59 -TTAAGAGCTCAGAACCAC-CACCACCAGAACCACCACCACCAGAACCACCACCACCC-TTCGCTACCACTGTCTTCAGATACTGTCCT -39, were used. The recombinant plasmid pET40CmAP was obtained by ligation of CmAP gene into pET-40b(+) and used for the subsequent cloning of the lectin MBL-AJ gene.
For the lectin gene amplification, cDNA of A. japonicus coelomocytes, the gene-specific upstream primer, Lect-X-dir: 59-AGCTGAGCTCTGTCTGACGGCTTGTCCGGAGTTTTG-39, and downstream primer, Lect-X-rev: 59-CAGTG-TCGACCTCCAAATGATACTCGATACAGGGAAG-39 (construction 1 -CmAP/MBL-AJ-Stop_codon), or X-Lect-pET40rev-TEV: 59-ACTGGTCGACTTAGTGGTGGTGGTGGTGGTG-GTGGTGACCCTGAAAATAAAGATTCTCCAAATGATAC-TCGATACAGGGAAG-39 (construction 2 -CmAP/MBL-AJ-TEV) were used. cDNA for MBL-AJ was produced as previously described [5]. The plasmid pET40CmAP/MBL-AJ was obtained by ligating the cDNA fragment encoding MBL-AJ into the recombinant vector pET40CmAP, linearized by SacI/ SalI site-specific scission. Restriction endonucleases and T4 DNA ligase were purchased from Fermentas. The PCR product and linearized plasmid were purified from agarose gel with Gel Extraction kit (Qiagen). Correct recombinant plasmid sequences were verified by DNA sequencing using an automated PE/ABI 310 DNA sequencer and PE/ABI-ABI PRISM BigDye Terminator cycle sequencing Ready Reaction Kit (PE Applied Biosystems). Preparation of E.coli DH5a competent cells for plasmid propagation and heat shock transformation were carried out according to the standard method [17]. All biochemicals and reagents were from Fermentas and Sigma-Aldrich. E. coli Rosetta (DE3) cells were used as a host for the expression of the recombinant hybrid CmAP/MBL-AJ.

Expression of the recombinant hybrid CmAP/MBL-AJ
For the expression of CmAP/MBL-AJ, the E.coli Rosetta(DE3) cells transformed with the recombinant plasmid pET40CmAP/ MBL-AJ were grown on LB agar plate containing 25 mg/ml kanamycin overnight at 37uC. A single colony was picked and grown at 200 prm in 20 mL of LB medium with 25 mg/ml kanamycin at 37uC for 12 h, then transferred to 1 L of fresh LB with 25 mg/ml kanamycin. When the cell density reached an A 600 of 0.6-0.8, 0.2 mM IPTG was added to induce the expression of the protein, and the incubation continued at 16uC up to 12 h at 200 rpm. E.coli Rosetta(DE3) cells were transformed with pET-40b (+) as a control.
The recombinant hybrid CmAP/MBL-AJ purification All purification steps were carried out at +6uC. After harvesting, the E.coli cells were resuspended in 150 ml of the buffer A, containing 0.02 M Tris-HCl, pH 8.2, and 0.01% NaN 3 , and disintegrated by ultrasonic treatment, then centrifuged at 10000 g for 30 min. The supernatant was applied to a column (4625 cm) of DEAE-52 cellulose (Whatman). Elution of the protein was performed with NaCl gradient (0.05 M-0.38 M) in the buffer A. The enzymatically active fractions were collected and applied to a column (162.5 cm) of Ni-agarose (Qiagen). Elution of the protein was carried out in buffer B (0.02 M Tris-HCl, pH 8.2, 0.05 M EDTA, 0.01% NaN 3 ). The active fractions were collected and desalted with DEAE-Toyopearl 650 M (Toyo Soda) column (0.762.5 cm), then incubated with TEV protease (New England Biolabs) at 21uC for 12 hours. Then the protein solution was applied to a gel filtration column (1.56170 cm) of Superdex 200 (Sigma) equilibrated with buffer C (0.02 M Tris-HCl, pH 8.2, 0.1 M NaCl, 0.01% NaN 3 ). All purification steps were evaluated by SDS-PAGE according to Laemmli [18]. The concentration of the protein was determined according to Bradford [19]. Peptide sequencing of the N-terminal end was performed on Edman Automated Sequencing Apparatus Beckman 890C.

Enzymatic activity assay
The standard assay for alkaline phosphatase (AP) activity was carried out at 37uC using 2 mM p-nitrophenylphosphate (pNPP) (Sigma Chemical Co.) in 0.1 M Tris-HCl buffer, pH 9.0, containing 0.2 M KCl, or in 1 M diethanolamine (DEA) buffer, pH 10.3, containing 15-20 mM pNPP. The release of pnitrophenol (e = 18.5 mM 21 ?cm 21 ) was monitored at 405 nm. One unit of AP activity was defined as the quantity of the enzyme required to release 1.0 mmol of p-nitrophenol from pNPP in 1 min. The specific activity was calculated as units per 1 mg of protein.

Lectin activity assay
The embryonic alpha-1-acid glycoprotein (0.1 mg/mL) was dissolved in 0.1 M carbonate buffer, pH 9.5, containing 0.15 M NaCl, and 150 ml was added to the first well and was two-fold serially diluted in the same buffer in each well of a polystyrene 96well ELISA microtiter plate (Maxisorp, Nunc), and then incubated at 4uC overnight. After incubation, the plate was washed three times with the buffer containing 0.01 M Tris-HCl, pH 7.5, 0.15 M NaCl, 0.05% Triton X-100 (TBS-T) and then three times with water. After that, 160 mL of bovine serum albumin (1 mg/ mL) in TBS-T was added to each well to block non-specific binding sites. The plate was then incubated for 1 hour at room temperature and washed as described above. 150-mL aliquots of the samples containing chimeric lectin CmAP/MBL-AJ with concentration 100 mg/mL were added to each well. TBS-T was used as negative control, and interaction of MBL-AJ with adsorbed antibodies under the same conditions served as a positive control. After that, the plate was kept for 1 h at room temperature and washed again. Then 150 mL of pNPP in the buffer for AP assay was added to each well. After 2-5 min of incubation, 100 mL of 2 M sodium hydroxide was added to stop the reaction. Absorbance of the resultant solution was measured at 400 nm with a plate spectrophotometer (Bio-Tek Instruments, USA).

Molecular mass determination
The molecular size of the native hybrid CmAP/MBL-AJ was determined by gel filtration on the column (1.56170 cm) of Superdex 200 (Sigma) in 20 mM Tris-HCl, pH 8.2, 100 mM NaCl at a flow of 0.2 ml/min at 6uC and calibrated using Bio-Rad standard molecular weight markers: bovine thyroglobulin (670 kDa), bovine c-globulin (158 kDa), chicken ovalbumin (44 kDa), horse myoglobin (17 kDa), Vitamin B 12 (1.35 kDa). The molecular mass of CmAP/MBL-AJ was determined by 12.5% SDS polyacrylamide gel electrophoresis (SDS-PAGE). An aliquot of the protein solution was mixed with Laemmli sample buffer and heated at 95uC for 5 min, then applied to SDS-PAGE and stained according to Laemmli [18]. PAGE for the native CmAP/MBL-AJ was performed without SDS and heat treatment. The ''native'' gels were stained for AP by the casing over with the standard enzyme assay buffer containing 2 mM pNPP. Both PAGEs were performed at 25uC.

Molecular modeling
The structural models of CmAP and MBL-AJ were generated by comparative modeling approach for further modeling the putative spatial structure of the bifunctional hybrid CmAP/MBL-AJ. The target-template alignments and modeling of CmAP and MBL-AJ 3D-structures were carried out by the molecular modeling package Molecular Operating Environment version 2013.08 (Chemical Computing Group Inc., 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2013). The high-quality experimental structure of alkaline phosphatase from marine bacterium Vibrio sp. G-15 with 1.4 Å resolution (VAP; PDB code: 3E2D) was used as a template for homology modeling of CmAP 3D-structure. The crystal structure of C-type lectin CEL-IV from holothurian Cucumaria echinata with 1.65 Å resolution (PDB code: 3ALU) was used as a template for homology modeling of MBL-AJ 3D-structure. The CmAP/MBL-AJ chimeric model was manually generated by adding (G 4 S) 3 linker to the C-terminus of the CmAP model and by binding with N-terminus of MBL-AJ using program MOE. Structural optimization of CmAP, MBL-AJ and CmAP/MBL-AJ models were carried out by energy minimization method with force field Amber12:EHT using program MOE 2013.08. The structure of the model oligosaccha-  [20] to design the lectin configuration with a highest affinity to branched oligosaccharide. The binding energy or solvation (the molecular mechanics generalized Born interaction energy) was the non-bonded interaction energy between the receptor and the ligand and comprised van der Waals, Coulomb and generalized Born implicit solvent interaction energies [20]. The affinity (the estimated binding affinity) was that of the GBVI/WSA dG scoring function reported in units of kcal/mol [20]. The evaluation of structural parameters, analysis of protein-ligand contacts, physicochemical properties, molecular docking and visualization of the results were carried out with the modules of MOE 2013.08 program.
Site-specific mutagenesis of lectin-binding domain of CmAP/MBL-AJ The single W100A, Q103H, A156N, F159K and double A156N/F159K mutants of CmAP/MBL-AJ were genetically engineered by site-directed mutagenesis. The single or double amino acid substitutions were introduced into the forward and reverse gene-specific primers to the expected gene regions. The resultant PCR products were purified from agarose gel and annealed both the DNA matrix and primers in the equivalent ratio for the next PCR. The resultant mutant gene was amplified with the primers Lect-X-dir and Lect-X-rev, and fused to the linearized vector pET40CmAP and expressed as described above.
The lectin-binding activity assay of the mutant CmAP/MBL-AJ colonies was carried out by the modified method described above. Mucin (0.1 mg/mL) was dissolved in 0.1 M carbonate buffer, pH 9.5, containing 0.15 M NaCl, and 150 ml was added to each well of a polystyrene 96 -well ELISA microtiter plate (Maxisorp, Nunc), and then incubated at 4uC overnight. After incubation, the plate was washed three times with the buffer, containing 0.01 M Tris-HCl, pH 7.5, 0.15 M NaCl, 0.05% Triton X-100 (TBS-T) and then three times with water. After that, bovine serum albumin (1 mg/mL) in TBS-T was added as described above. Samples containing chimeric lectin CmAP/MBL-AJ (100 mg/mL) were two-fold serially diluted in TBS-T and added in 150 mL aliquots to each well. TBS-T was used as a negative control. After that, the plate was incubated, washed and visualized as described above. All of the lectin activity assays were performed in 3 independent parallels for 3 to 5 times.

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.

Results
Expression and purification of the bifunctional hybrid CmAP/MBL-AJ A novel method for producing a highly carbohydrate-specific holothurian lectin MBL-AJ has been developed. The recombinant bifunctional hybrid CmAP/MBL-AJ expression and purification were monitored by the alkaline phosphatase CmAP activity facilitating the rapid and effective detection of the CmAP/MBL-AJ-containing fractions. Although the expression of all engineered the full-length mature CmAP and MBL-AJ genes fusions in E. coli cells resulted in a high-level production of AP activity, the construction (2) containing C-terminal 6XHis tags was more useful for the hybrid CmAP/MBL-AJ purification than the same construction (1) without C-terminal 6XHis tag. The additional His-tag drastically enhanced the binding of CmAP/MBL-AJ to Niagarose column. However, for increasing of the lectin-binding activity of CmAP/MBL-AJ, the step of the TEV protease cleavage of C-terminal 6XHis tag was preferable ( Table 1).
As a result, the CmAP/MBL-AJ production reached up to 1.1 mg of the final yield of the functionally active protein from 1 L of the E.coli Rosetta(DE3) cells culture carrying the expression recombinant vector pET40CmAP/MBL-AJ (Table 1, Fig. 1).
The optimal conditions for the CmAP/MBL-AJ expression were the 0.2 mM concentration of IPTG and the strain cultivation at 16uC for 12 h (Fig. 2). At the final step of purification, SDS-PAGE showed a single band of the protein with a molecular mass of approximately 100 kDa, corresponding to the calculated molecular mass of the hybrid protein CmAP/MBL-AJ consisted of the mature proteins of CmAP (55 kDa), MBL-AJ (17 kDa), peptide linker (G 4 S) 3 , the plasmid fusion protein DbsC (32.5 kDa) and Nand C-terminal 6 X His tags (Fig. 1, 3). The relative molecular size determined by gel filtration corresponded to 200 kDa, suggesting a homodimeric form of the enzymatically active hybrid CmAP/ MBL-AJ. CmAP/MBL-AJ appeared as tetramer at the highly alkaline pH$9 (data not shown).   MBL-AJ is stable at pH between 6.0-11.0, but CmAP is gradually inactivated below pH 6.0. CmAP was also gradually inhibited when the enzyme was exposed above pH 8.5 maintained by 0.1 M glycine (Fig. 4).
CmAP/MBL-AJ has temperature optimums 40uC-50uC for the manifestation of the highest AP activity of 5100 U/mg in 1 M DEA buffer or 1500 U/mg in 0.1 M Tris-HCl buffer, 0.2 M KCl, at pH 10.0 (Table 1). CmAP/MBL-AJ has a half-life of 27 min at 40uC and 16 min at 45uC in the presence of 2 mM EDTA (data not shown). Inactivation of 50% of AP activity occurred in 45 min at 37uC. However, CmAP/MBL-AJ was stable up to 45uC with a half-life of 15 min at 47uC and 120 min at 45uC, when the enzyme was incubated with 10 mM MgCl 2 (data not shown).
The level of lectin-binding activity of the bifunctional hybrid CmAP/MBL-AJ was estimated by its affinity to embryonic alpha-1-acid glycoprotein or mucin. The detailed carbohydrate-binding specificity of MBL-AJ determined by inhibition assay has been previously shown [5].
The highly active AP label genetically linked to the lectin MBL-AJ has advantaged the lectin-binding assay in its sensitivity and time in comparison with the two-step method of the wild-type MBL-AJ conjugation with the lowly active horse peroxidase (Fig. 5).
The lectin activity of the bifunctional hybrid CmAP/MBL-AJ is Ca 2+ and pH-dependent and reversibly inhibited by EDTA. The optimum buffer for the lectin activity of CmAP/MBL-AJ is 0.1 M    Tris-HCl, 0.15 M NaCl, 0.01 M CaCl 2 , pH 7.0-9.0 that is suitable for both the AP and lectin activities. The highest affinity of CmAP/MBL-AJ to ligands is observed at 4uC. However, CmAP can also exhibit an appreciable hydrolytic activity at the low temperature due to its cold-adaptability (data not shown). The diluted of the highly purified CmAP/MBL-AJ preparation up to thousand fold was fully stable at 220uC during more than oneyear storage in the dilution buffer (0.02 M Tris-HCl, pH 8.2, 0.01% NaNO 3 ).

Molecular modeling of the bifunctional hybrid CmAP/ MBL-AJ
Search for a template of CmAP showed that the crystal structure of the cold-active Vibrio sp. G15-21 alkaline phosphatase (VAP; PDB ID 3E2D) obtained with 1.4 Å resolution [21] and the crystal structure of alkaline phosphatase from the moderate halophilic bacterium Halomonas sp. 593 (HaALP) determined with 2.1 Å resolution recently [22] can be used to generate a homology model.
Amino acid sequences of CmAP, HaALLP and VAP have 75% and 69% of identity and 86% and 82% of similarity, respectively. The 3D-structure of cold-active alkaline phosphatase CmAP from marine bacterium C. marina (Uniprot Q1W622) was modeled using as a template the crystal structure of the cold-active Vibrio sp. G15-21 alkaline phosphatase (Fig. 6).
Superposition of all Ca atoms of CmAP model and template showed that the value of root mean square deviation (RMSD) is 0.43 Å . The structure of the CmAP monomer is stabilized by 148 hydrogen bonds, 20 ionic bonds and 180 hydrophobic contacts. The disulfide bonds are absent in CmAP, HaALP and VAP structures. Superposition of the CmAP structural model and HaALP and VAP crystal structures showed that the value RMSD is 0.87 Å for 497 Ca atoms. The CmAP monomer structure consists of a catalytic domain and a ''crown''-domain by analogy with VAP and HaALP. Catalytic domain displays a three-layer bsandwich fold and contains b-sheet of nine b-strands surrounded by a-helices at both sides. The structure of the catalytic domain has a fold of (a-b-a)-type that is AP characteristic. The CmAP, HaALP and VAP ''crown''-domains are the largest observed in known APs. The cold-active CmAP, HaALP and VAP are the AP variants with the highest known k(cat) value. Analysis of the CmAP 3D-structure models and crystal structures of VAP and HaALP showed that the structures of the active sites and metal binding sites of CmAP, HaALP and VAP are identical. The critical Nterminal region for the stability of dimeric ECAP from E. coli K12 is absent in CmAP as in VAP and HaALP. The VAP or HaALP crystal structures are the dimers but it was experimentally showed that the functional structure of the CmAP is a monomer [23].
Structural model of lectin MBL-AJ was generated using as a template of the crystal structure of sea cucumber lectin CEL-IV (PDB ID 3ALU) (Fig. 7 (A)) [24].
Amino acid sequence analysis of the lectin MBL-AJ and the template CEL-IV revealed that they have 65% identity and 79% similarity. Spatial structure of the lectin MBL-AJ is characteristic of the packed C-type lectins and contains two a-helix and 12 bstrands, disordered loops and ion Ca 2+ in carbohydrate-recognizing domain. 3D-superposition of Ca-atoms of the structural model and prototype showed the value of RMSD to be 0.4 Å . The lectin MBL-AJ has a longer loop near the carbohydraterecognizing domain than CEL-IV lectin (Fig. 7 (B)). The lectins MBL-AJ and CEL-IV have conservative position of six cystein residues that form two disulfide bonds stabilizing the monomer structure ( Fig. 7 (A)). MBL-AJ dimer and tetramer are the functionally active forms of the lectin [5]. Therefore, the MBL-AJ dimer structure is generated using the crystal structure of the CEL-IV lectin dimer ( Fig. 7 (C)). For the lectin CEL-IV it has been shown that two cysteine residues are involved in the formation of S-S bonds between monomers in the dimeric form of the lectin. Similar disulfide bond stabilizes the dimer structure of MBL-AJ ( Fig. 7 (C)). The structure of the bifunctional hybrid CmAP/MBL-AJ monomer consisting of the alkaline phosphatase CmAP, linker (G 4 S) 3 and lectin MBL-AJ was generated using MOE 2013.08 (Fig. 8).
Since the lectin MBL-AJ is a functionally active dimer, the hybrid CmAP/MBL-AJ model was constructed based on the MBL-AJ dimer. The CmAP/MBL-AJ dimer model has the twosubunit lectin part that is associated with two molecules of alkaline phosphatase functioning independently from each other (Fig. 9). The MBL-AJ lectin exhibits high specificity for branched mannan [5]. Molecular docking predicted the structural model of MBL-AJ lectin binding with a model oligosaccharide (Fig. 10). The mannose residues of the main chain Man2 and from branching chain Man4 that are not the terminal mannose residues form contacts with the MBL-AJ carbohydrate-recognition domain. Analysis of the contacts in MBL-AJ carbohydrate-recognition domain showed that the model oligosaccharide forms hydrogen bonds with the lectin and Ca 2+ ion (Fig. 11). According to the lectin-oligosaccharide complex model, the residue Glu134 forms hydrogen bonds with O3 and O4 of the mannose residue Man2 from the main chain; Ca 2+ ion forms ionic and metal bonds with O3 of the mannose residue Man4; residue Asn137 forms hydrogen bond with O2 of the mannose residue Man5.
In silico mutation and site-specific mutagenesis of the lectin-binding domain of the bifunctional hybrid CmAP/

MBL-AJ
In silico design of the MBL-AJ lectin with an improved affinity and activity was carried out using point mutations and molecular docking approach. A high quality of the MBL-AJ model allowed estimating the role of its essential amino acid residues for the interaction of MBL-AJ with ligands by in silico mutation. The 3Dmodels of mutants W100A, Q103H, A156N, F159K, A156N/ F159K were generated. Molecular dockings of mutants with a model oligosaccharide were carried out and oligosaccharide binding energies with mutants and affinities were predicted ( Table 2). The models of W100A, Q103H, A156N and F159K mutants of the CmAP/MBL-AJ lectin module were built to study the molecular basis of its lectin-binding interaction with the model branched oligomannan. The residues W100 and Q103 were selected for substitutions by the reason of their homology with the template CEL-IV residues W79 and Q82 participating in the binding of the mannose residues [24]. The residue substitution A156N was selected by the reason of the homologous residue D ionic-binding with Ca2+ and H-binding with carbohydrates in the majority of the C-type lectins. The substitution of nonpolar F159 located near the mannose residue in the oligomannan for the positively charged residue K could facilitate the additional formation of hydrogen bond between the MBL-AJ lectin and the model oligosaccharide.
Calculation revealed that the double mutants A156N/F159K exhibit a lower binding energy and higher affinity than the wildtype lectin ( Table 2). Replacing enhances the interaction of the oligosaccharide with the lectin by additional energetically more favorable hydrogen and ionic bonds (Fig. 12).
In the case of the A156N, F159K mutants, their binding energies are lower than the wild-type lectin's one, but higher than the energy of the double mutant A156N/F159K. Mutations W100A and Q103H decrease lectin binding affinity to the carbohydrate moiety. Thus, in silico it was demonstrated that the specificity and high affinity of the MBL-AJ lectin are determined by its interaction with the monosaccharides from the main and branching chains of the model oligosaccharide.
For the experimental testing of the lectin-binding activity, W100A, H103Q, A156N, F159K and A156N/F159K mutants were expressed in E.coli. The chimeric mutants of CmAP/MBL-AJ remained the lectin activity in the range of the increasing of the level of affinity W100A-H103Q-A156N-F159K to mucin that was in accord with the calculated data ( Table 2, Fig. 13). The W100A, Q103H, A156N mutations decreased the lectin-binding activity in comparison with the wild-type CmAP/MBL-AJ by 1463%, 1262% and 862%, respectively.
The substitution F159K enhanced the lectin binding of CmAP/ MBL-AJ by 1063% due to the formation of the additional hydrogen bond with the OH-group of the mannose residue of the substrate (Fig. 12, 13). Surprisingly, the double A137N/F159K mutation had cumulative positive effect on the lectin affinity of CmAP/MBL-AJ, increasing its activity by 2565%, 4062% and 2863% in comparison with the wild-type CmAP/MBL-AJ and single A137N and F159K mutants, respectively (Fig. 13).

Discussion
Currently, construction of the fusion or tagged proteins is a useful method to obtain chimeric molecules with an improved functionality or, moreover, dual mode of action such as enzymes with specific binding activities [25], [26], [27], [28]. Here we report on genetic engineering and overexpression in E.coli of the bifunctional hybrid protein CmAP/MBL-AJ with the alkaline phosphatase CmAP of marine bacterium and the Far Eastern holothurian lectin MBL-AJ activities in the purpose of further improving the enzyme-linked lectin assay (ELLA) for diagnosing of cervical cancer [5], [13], [29]. Although the holothurian MBL-AJ and mammalian lectins, including human MBL, had common properties to recognize bacterial mannans, the differences in the carbohydrate-binding specificities were found to be significant [5]. It has been previously shown that holothurian lectin MBL-AJ interacts with the promising cancer biomarkers in the following order of the lectin-binding affinity: CEA, embryonic a-1-acid glycoprotein, trophoblast-specific b1-glycoprotein and a-fetoprotein isolated from the abortive and retroplacental blood [13], [30], [31], [32]. MBL-AJ is not inhibited by monosaccharide, which is in the content of the normal human glycoconjugate's carbohydrate chains, and does not interact with the human blood serum components. On the base of the unique carbohydrate-binding domain that defines the MBL-AJ ability to distinguish microheterogeneity of the malignant or normal cell glycoconjugates, a novel method of cervical cancer diagnosis has been developed [13]. The method was adapted to determine the quantitative level of the lectin-binding forms of CEA in the vaginal secretions of patients. The vaginal secretions were collected from the proximal uterine cervix surface, irrespectively on the age and physiological state of patients. The method with the use of MBL-AJ has allowed identifying statistically reliable differences between the levels of lectin-binding CEA between healthy women and patients with cervical cancer, and between patients with benign and malignant neoplasms [13]. Moreover, it is important that MBL-AJ-based method give a possibility to assess tumours noninvasively. Loss of expression of the MBL-AJ ligands was found in cervical cancer. According to the results, the concentration of the MBL-AJ-binding CEA of the healthy patients and the patients with the benign neoplasm were determined to be 48.5611.8 U/ml versus the concentration of 11.467.5 U/ml of the patients with the cervical cancer diagnosis. The total CEA, which can be synthesized by both malignant and normal cells, were excluded from analysis because the cervical specimens were collected from the local source of the cancer CEA biosynthesis. A high sensitivity and   specificity of the MBL-AJ-CEA interaction allowed detecting the lectin-binding structures in the vaginal secretion in the concentration of CEA 3-50 ng/mL [13]. The method specificity and sensitivity were 93.6 and 87.8%, respectively, for patients with cervical cancer with the cut-off level of 12.74 ng/mL. Whereas the method prognostic valuation was calculated to be 87% and 95.2% for the positive and negative diagnosis, respectively. Thus, this method has advantages compared to those associated with determining of concentration of CEA and squamous cell carcinoma antigen (SCC) in blood serum [2], [13], [14].
However, the wild-type lectin MBL-AJ derivation from the holothurian A. japonicus coelomic liquid has many restrictions, namely: low concentration of MBL-AJ in the native source, preservation of the wild life of the Far Eastern holothurian A. japonicus, the narrow habitat of the endemic species of the holothurian. A recombinant analogue of MBL-AJ was attempted to be produced in the E. coli Top10/pQE_80L expression system [29]. However, the protein was expressed in the body inclusion, and its refolding resulted in 20%-yield of the soluble recombinant lectin of 69% of homology with the wild-type MBL-AJ by the level of interaction with the antibodies against MBL-AJ. The recombinant production of the bifunctional hybrid CmAP/MBL-AJ reached the full solubility and functionality of the lectin MBL-AJ. Moreover, the alkaline phosphatase CmAP fused with the lectin MBL-AJ by the long flexible linker facilitated the recombinant protein purification monitoring as well as the lectin binding assay, eliminating the additional steps, namely: production of the antibodies against MBL-AJ, labeling the MBL-AJ antibodies by horsh peroxidase or alkaline phosphatase and incubation of the labeled MBL-AJ antibodies with the ligand-binding lectin MBL-AJ.  *-The molecular mechanics generalized Born interaction energy is the non-bonded interaction energy between the receptor and the ligand and comprises van der Waals, Coulomb and generalized Born implicit solvent interaction energies [Labute, 2008]. **-The estimated binding affinity is that of the GBVI/WSA dG scoring function reported in units of kcal/mol. doi:10.1371/journal.pone.0112729.t002