Heterologous Expression and Characterization of Thermostable Levansucrase (BsSacB) from Bacillus subtilis BB03

Corresponding Author: D. Theertha Prasad Department of Biotechnology, University of Agricultural Sciences, GKVK, Bengaluru 560065, India Email: dtprasadd@gmail.com Abstract: Thermostable levansucrase encoding gene (sacB) with its promoter was isolated from Bacillus subtilis BB03. BsSacB is composed of 1419 bp, encoding 473 amino acid residues and with promoter regions TTGCAA (-35) and TAGAAT (-10). The amino acid sequence analysis of BsSacB showed conserved motifs of microbial fructosyltransferase belonging to Glycoside Hydrolase family 68 (GH68). Comparative analysis of the protein structure was carried out with 3D model of BsSacB constructed using PyMol software. BsSacB gene was expressed in Escherichia coli and the purified enzyme demonstrated thermo stability up to 50°C with a high hydrolytic activity at pH 6.0 and presence of Ca 2+ promoted the activity by 15%. An attempt was made to further characterize the thermostable recombinant levansucrase from B.subtilis BB03 using crude sucrose rich substrates. Optimal levan production was seen with molasses (1.94±0.04 g/L) at 15 h as compared to cane juice (1.0±0.01 g/L) at 12 h. An increase in levan formation with gradual decrease in hydrolytic activity was distinctly evident in both the molasses and juice media.


Introduction
Levans are prebiotic agents with potential health benefits, selectively support intestinal health and act as low calorie sweeteners (Byun et al., 2014;Kang et al., 2009). They are used as blending component in cosmetics to provide cell-proliferating and skin moisturizing effects (Abdel-Fattah et al., 2012). Due to their physical and biological functions as antitumor, anti-inflammatory agent, they have potential application in food and pharmaceutical industry (Yoon et al., 2004;Kim et al., 2005). Levans are water soluble fructo-oligosaccharides which form viscous solutions, hence used as emulsifier or encapsulating agent in cosmetics, biodegradable plastics, textile coatings and detergents (Kang et al., 2009).
For industrial exploitation of the potential of these enzymes, it is essential to gain knowledge about their molecular and structural characteristics. Characterization, cloning and expression of levansucrase (SacB) from Bacillus sp. in a heterogeneous system like E.coli have been attempted. Studies on recombinant expression of levansucrase in gram positive bacterium like B. megaterium are in developing phase (Rairakhwada et al., 2010;Korneli et al., 2013). Several expression studies have been performed using E.coli system (Rairakhwada et al., 2010;Kang et al., 2005;Nakapong et al., 2013); however they require addition of IPTG to induce levansucrase production. Further, the availability of low cost media and sucrose substrate is a major constraint in levan production. The present study report the isolation of BsSacB from B.subtilis with its own promoter inducible in presence of sucrose that can effectively be expressed in E.coli.
The results of the studies on effect of temperature, pH, substrate concentration and utilization of alternate sucrose rich substrate on hydrolysis and trans fructoslysation activity using the recombinant BsSacB expressed in E.coli has been presented.

Genomic DNA Isolation and Cloning of BsSacB Gene
Genomic DNA of B.subtilis BB03, an isolate from the Banana peel, which produces thermostable levansucrase was isolated (Sambrook et al., 1989) and used as template for amplifying BsSacB. Initially, the primers for sacB (LevF-5'GCACTGTCGCAAACTATCAC-3' and LevR-5'ACGTGATGCCGTCAATCG3') were designed based on the reported sacB gene sequences of bacterial levansucrases (Velázquez-Hernández et al., 2009). PCR reaction was performed in a 25.0 µL reaction volume containing 2.5 µL of 10X Taq buffer B, 1.5 µL of 25.0 mM MgCl 2 , 1.0 µL of 10 mM dNTPs (2.5 mM each), 1.0 µL of each primer, 1.0 µL of Taq DNA polymerase (3.0 U/µL; Bangalore Genei, India), 0.5 µL of isolated genomic DNA (0.3 µg/µL) and reaction conditions of initial denaturation at 94°C for 5 min, followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 66°C for 1 min and extension at 72°C for 1 min and a final extension at 72°C for 7 min. The sequence of the amplicon obtained was analyzed on NCBI BLASTn, which showed homology with sacB of B.subtilis (X02730.1). Based on this, primers Lev 16F (5'GTTCTTTAGGCCCGTAGTCTG3') and Lev 19R (5'AATACGGTTAGCCATTTGCCTGC3') were designed and full length sacB along with the promoter was obtained. PCR amplification was performed and the products were sequenced. Amplicons were analyzed on 1% agarose/EtBr gel electrophoresis, purified using QIAEX II gel extraction kit (QIAGEN, India) and ligated into pTZ57R/T vector (Fermentas, India) as per the manufacturer's instructions. Recombinant pTZ:BsSacB was transformed into E.coli DH5α. Positive colonies were analyzed by colony PCR using Lev16F and Lev19R primers as well as internal primers (Lev46F-5'CCGATGAACATCAAAAAGTTTGC3'; Lev74R-5'ATCTCTCAGCGTATGGTTATCC3'). Recombinant plasmid was isolated using QIAprep mini spin kit (QIAGEN) and sequenced using automated DNA sequencer (Mega BASE 500, Applied Biosystems) and verified with BLASTn analysis.

Bioinformatic Analysis of BsSacB
BLAST search was performed in NCBI (http://www.ncbi.nlm.nih.org) and protein prediction was performed in CBI (http://www.cbi.pku.edu.cn). Alignment of the BsSacB nucleotide and protein sequence with other structurally related levansucrase genes were performed using the ClustalW program. Phylogenetic tree was drawn with the Mega 4.0 program (Tamura et al., 2007). Coding region of BsSacB (Acc.no.CBI68350) was analyzed by PSI-BLAST program (Position specific interactive BLAST) to identify structure of related protein. Atomic coordinates of the 1OYG structure, solved at 1.5A° resolution, was used as a template for comparative modeling. Sequence and structural alignment of BsSacB with 1OYG Chain A was performed using ClustalW (http://www.ebi.ac.uk/Tools/msa/clustalw2/). Aligning was carried out without the insertion of gaps in the conserved secondary structural regions. Spatial restraints were derived and used in the three dimensional structure determination using PyMol molecular graphics software (www.pymol.org). The stereochemical quality of model was assessed by PROCHECK (Laskowski et al., 1993; www.ebi.ac.uk/thornton-srv/software/PROCHECK/ ) at the same resolution as in 1OYG structure. The overall stereo chemical quality of the model was assessed by Ramachandran plot analysis.

Expression of Recombinant Levansucrase
E.coli harboring pTZ:BsSacB was grown in LB medium at 37°C for 12 h and used for levansucrase expression analysis. After the 12 h growth, the cells were lyzed by freezing at -80°C followed by thawing in 37°C water bath and vortexed for 5 sec., the freeze-thaw cycle was repeated 4 times respectively (modified with reference to Gay et al., 1983). This was followed by centrifugation at 12000 × g for 10 min at 4°C. The freeze thaw method is found to separate the highly expressed recombinant protein from the bulk of endogenous E.coli proteins (Sangiliyandi and Gunasekaran, 2000). The supernatant obtained was brought to 60% (v/v) with respect to acetone. Protein precipitated was collected by centrifugation 17,000 × g for 15 min at 4°C, dissolved in 0.05M sodium phosphate buffer, pH 6.0. The extract was further purified using Zeba spin columns (Thermo scientific), centrifuged at 3,040 rpm for 2 min. Protein concentration was determined at 280 nm using BSA as standard and analyzed by Native PAGE (Sambrook et al., 1989). The unstained gel was sliced into 5 mm bits and extracted in 100 µL 50 mM sodium phosphate buffer, pH6.0 and assayed for levansucrase activity by standard procedure described.

Assay for Levansucrase and Levan Production
Levansucrase activity was assayed in a reaction mixture containing 6.5 µL of protein sample (100 µg/mL) and 50 mM sodium phosphate buffer, pH6.0, containing 8.5% (w/v) sucrose and 1.0 mM CaCl 2 followed by incubation at 50°C for 60 min and the amount of glucose released was measured using glucose oxidase kit (Merck, India; Vaidya and Prasad, 2012). For levan production, assay mixture was incubated at 40°C for 12 h. Levan produced was precipitated with 4 volume of ice-cold 70% aqueous ethanol and centrifuged at 17,400 × g at 5°C. The resulting precipitate was re-suspended in sterile distilled water and heated at 80°C for 10 min followed by hydrolysis in 0.1N HCl at 80°C for 10 min. After the hydrolysis, the levan content was assayed using 1% resorcinol reagent and expressed as fructose units (Viikari and Gisler, 1986).

Levansucrase Activity and Stability
Levansucrase activity and levan production profile was analyzed over a range of pH 5.0-9.0 (50 mM sodium acetate buffer of pH 5.0, 50 mM sodium phosphate buffer pH 6.0-7.0 and Tris-Cl buffer pH 8.0-9.0). Levansucrase was also analyzed for pH stability, (pH 5.0-9.0) by pre-incubating the enzyme sample at appropriate pH for 15 min prior to the assay. Optimum temperature for the activity and levan formation was determined by incubating reaction mixture at 4-80°C under standard assay conditions. Temperature stability of the enzyme was tested by pre-incubation at above mentioned temperatures for 15 min and residual activity was measured.
To examine the effect of metal ions, levansucrase was pre-incubated for 60 min at 4°C with 1.0 mM concentration of cations (Mg 2+ , Cu 2+ , Ca 2+ , K 1+ , Mn 2+ ; Fe 3+ and Hg 2+ ) prepared in 50 mM sodium phosphate buffer, pH 6.0. Levansucrase activity and levan production profile was analyzed at different sucrose concentrations (1.0-13.0% w/v) in 50 mM sodium phosphate buffer pH 6.0 and temperature (50°C for sucrose hydrolase activity and 40°C for levan production). Michaelis constant (Km) and velocity maximum (Vmax) was calculated. To determine the effect of enzyme concentration, enzyme solution (0-100 µg/10 µL) was spotted on agar plate (1.35%) containing 8.5% (w/v) sucrose and incubated overnight at 40°C. Whitish slimy spots formed were photographed under UV light for better visualization. Effect of incubation time on levansucrase activity and levan produced was measured at every 3 h intervals for 24 h.

Cane Juice and Molasses as Substrate
Cane molasses or cane juice as source of sucrose was used to assess levansucrase activity and levan production. Prior to analysis, cane molasses was clarified using 0.1N H 2 SO 4 and pH was adjusted to 6.0 using 0.1N NaOH. Cane juice was heated at 80°C for inactivating the innate invertase activity and filtered through Whatmann No. 3 filter paper. Clarified molasses or cane juice were analyzed for sucrose, glucose and fructose concentration by high performance liquid chromatography (Waters) using Ultra Amino column, Restek (150 mm length and 100A° pore size) with RI detector (Waters 410). E.coli -pTZ:BsSacB was grown in cane molasses and juice diluted to contain 1.0% sucrose, fortified with 0.2% yeast extract and 0.2% (NH 4 ) 2 SO 4 (Han and Watson, 1992) in 500 mL Erlenmeyer flasks at 37°C under shaking conditions at 150 rpm for overnight. Gay et al. (1985) reported use of sacB as a counter selectable marker for E.coli and other gram negative bacteria. It was seen that the production of levansucrase in those microbial systems was lethal in presence of 5.0% sucrose in the medium. Similarly, higher sucrose concentration did not resulted in efficient cell growth in this system thus limiting our study for using sucrose concentration compatible with the growth. Further, growth was monitored by recording the OD at 600 nm; enzyme from the harvested cells at regular intervals was assayed for levansucrase activity and levan production.

Statistical Analysis
All the parameter analyses were repeated with three independent biological samples. The data were analyzed by one-way Analysis of Variance (ANOVA) and the treatment means were compared by using Duncan's Multiple Range Test (DMRT) at p≤0.05. Data were expressed as Mean ± Standard Error (SE).
Three dimensional structure of BsSacB protein was compared with crystal structure of B.subtilis levansucrase (CBI68350.1, 1OYG_A chain) using PyMol Molecular Graphics Software and PROCHECK to quantify the residues in available zones of the Ramachandran plot. It was found that 87.8% of residues were located in the most favored zones, 10.7% in allowed regions, 1.0% in generously allowed regions and 0.5% (two residues) in disallowed region. Ramachandran plot characteristic and the goodness factors (quality of covalent and overall bond-angle distances) confirm this to be the best fit for the predicted model proposed for BsSacB (Fig. 4).

Expression of BsSacB in E.coli
Proteins from the recombinant E.coli having pTZ:BsSacB were analyzed by native-PAGE and ingel levansucrase activity. The protein band with levansucrase activity (3.9U/mg) from the in-gel assay (Fig. 5) corresponded to the molecular weight of 52 kDa by SDS-PAGE which is in accordance with the predicted molecular weight. BsSacB encoded protein is typical to monomeric enzymes with the molecular weight ranging from of 46-73 kDa. However studies report that levansucrases from Rahnella aquatilis JMC-1683, Leuconostoc mesenteroides B-512 FMC and Actinomyces viscosus T-14V exist as dimmers (Ohtsuku et al., 1992;Kang et al., 2005;Pabst, 1977)

Effect of pH and Temperature on Levansucrase Activity and Levan Synthesis
Levansucrase activity of BsSacB protein was maximum (6.5 U/mg) at pH 6.0 and the activity significantly reduced above pH 7.0. Eighty-seven percent of the activity was retained at pH 5.0. Optimum yield of levan was at pH 6.0 which was observed to be 14.3 g/L and reduced to 6.7 g/L at pH 7.0 (Fig. 6a). Levansucrase activity was maximum at 50°C (10.6 U/mg) and showed a steep decrease with increase in temperature above 50°C. However, levan production was optimum at 40°C (10.8 g/L) and reduced drastically at 60°C (1.03 g/L; Fig. 6b). Similar difference was observed in optimum temperature for sucrose hydrolysis activity and levan formation with other recombinant levansucrase clones expressed in E.coli. Studies on R. aquatilis levansucrase showed hydrolase activity at 40°C and levan formation at 30°C . Levansucrase from B.amyloliquefaciens showed optimal activity at 30°C and levan production at 0-4°C (Rairakhwada et al., 2010). Studies by Hettwer et al. (1995) revealed that P.syringae levansucrase activity and levan formation was maximum at 60°C and 18°C respectively. Sangiliyandi and Gunasekaran (2000) reported the levansucrase activity at 60°C and levan synthesis at 30°C by Z.mobilis. However, BsSacB levansucrase was found to be tolerant over a wide pH and temperature range, stable over a pH range from 5.0 to 7.0 (Fig.7a) and temperature range from 4 to 50°C (Fig.7b). These results put together suggest that BsSacB levansucrase shows higher thermal stability in hydrolytic activity as compared to that in transfructosylation reaction. However, it has been observed that at saturated levels, levan formed was found to rapidly hydrolyze by levansucrase itself at higher temperatures .  Fig. 7. pH stability (a) and thermal stability (b) of the enzyme was examined by pre-incubating the enzyme for 15 min at respective pH and temperatures followed by standard assay conditions ----*reaction mixtures were incubated with 1.0 mM cation for 10 min **reaction mixtures were pre-incubated with 0.5 mmol/L EDTA for 10 min followed by 1.0 mmol/L cations and assayed for residual levansucrase activity

Effect of Metal Ions on Enzyme Activity and Levan Formation
Most of the levansucrases require metal ions as cofactors. Lactobacillus reuteri, Leu. mesenteroides and B.subtilis levansucrases require Ca 2+ (van Hijum et al., 2004;Ozimek et al., 2006;Morales-Arrieta et al., 2006;Velázquez-Hernández et al., 2009) whereas Fe 3+ can also act as an substitute cofactor for levansucrase by certain B.subtilis strains (Chambert et al., 1990). Table 1 shows the effects of metals ions on levansucrase activity and respective levan yield. Ca 2+ enhanced levansucrase activity by 15%. Na + and K + did not show any significant effect. Mg 2+ , Fe 3+ , Mn 2+ and Hg 2+ ions inhibited levansucrase activity more than 50%. The competition and replacement of a majority of smaller bivalent cations with Ca 2+ may result in a deformed tertiary structure and may be the reason for an inhibitory effect (Waldherr et al., 2008). EDTA blocked more than 70% of the enzyme activity. However, in presence of Ca 2+ , activity was restored to 44% of its initial activity. Levan production was 29% higher in presence of Ca 2+ compared to other metal ions. Sequence alignments of family GH68 proteins showed the presence of calcium ion-binding site and Asp339 was identified as key amino acid residue coordinating Ca 2+ binding (Meng and Futterer, 2003) in majority of levansucrases from Gram-positive bacteria (Ozimek et al., 2006) including BsSacB (Fig. 2).

Effect of Incubation Time
Maximum levansucrase activity was at 15 h of incubation time and decreased. However a steep increase in levan production was observed up to 6 h followed by gradual increase till 24 h with 8.5% (w/v) sucrose (Fig. 9). Similar study reports that recombinant levansucrase from B. amyloliquefaciens synthesized levan till 24 h (Rairakhwada et al., 2010). However, wild type sacB from Z. mobilis B14023 was expressed in E.coli produced levan from the initial period of fermentation and reached a maximum concentration at 16 h (6.8 g/L and 10.7 g/L respectively) (Ananthalakshmy and Gunasekaran, 1999). The production capacity of recombinant R. aquatilis levansucrase was found to reduce with prolonged incubation time at saturated levels of levan . Since levansucrase has both transfructosylation and hydrolase activity, hydrolysis of fructans can be expected at higher concentration of levans . Sucrose agar plate assay for levan production showed whitish slimy layer and was directly proportional with the enzyme concentration ( Fig. 10) which is on par with our earlier observations (Vaidya and Prasad, 2012). Similarly, levan production from B.subtilis NRC 33a levansucrase was also related to effective concentration of enzyme. The conversion of fructose to levan was 84% and the sucrose concentration was critical to levan synthesis (Abdel-Fattah et al., 2005).

Enzyme Synthesis and Levan Production by Levansucrase from E.coli -pTZ:BsSacB Grown in Cane Molasses and Cane Juice
Attempts have been made to commercially produce fructooligosaccharides utilizing agro-industrial byproducts rich in sucrose (Sangeetha et al., 2005). Levan are useful prebiotic compounds with enormous application in food and pharma industry and reduction in its production cost has high impact on its industrial production and usage (Kucukasik et al., 2011). Cane molasses and juice contained 23.0 and 18.0% sucrose respectively. HPLC analysis revealed about 2.8% fructose and 2.9% glucose in cane molasses. E. coli harboring BsSacB showed comparatively slow growth in fortified cane molasses and cane juice formulated to 1.0% sucrose (attained stationary phase in 9 and 15 h, respectively; Fig. 11a and b) as compared to wild type. In previous study, B.subtilis strain BB04 was grown in cane molasses and juice diluted to 6% sucrose, giving higher levan production (11.32 g/L and 4.81 g/L, respectively) after 12 h (Vaidya and Prasad, 2012). Hydrolase activity in cane molasses drastically decreased after 6 h, whereas levan production continued up to 15 h and remained almost constant (Fig. 11c). In cane juice, levan formation showed increasing trend till 12 h (Fig. 11d) and correlated with decrease in levansucrase activity. Thus, it was evident that with decreasing hydrolytic activity the transfructosylation activity increased which was also seen with sucrose as substrate. However, maximal levan production obtained using sucrose (8.5 g/L) was higher than that of cane molasses (1.94 g/L) and cane juice (1.0 g/L). Further clarification of media is necessary for production level levan yield. Halomonas sp. was grown in sugar beet and starch molasses pretreated by different methods. The levan yield in sugar beet after the clarification and pre-treatment (1.31 g/g of dry cell weight) was greater than before (0.72 g/g of dry cell weight) (Kucukasik et al., 2011). Microbacterium laevaniformans was used for levan production (10.4 g/L in 48 h) using date syrup as source of sucrose and the results suggest that increase in fermentation time caused a decrease in levan yield (Moosavi-Nasab et al., 2010). Higher levan formation in cane molasses might be due to the presence of readily available fructose in the medium unlike cane juice. Levansucrase from Halomonas sp. effectively synthesized levan from sugar beet (4.19 g/L) but not from starch molasses (3.68 g/L) as source of sucrose and was attributed to readily available sucrose content in sugar beet, a limitation in starch molasses (Kucukasik et al., 2011). (c) (d) Fig. 11. Study of E.coli expressed FTF activity using fortified cane molasses (a, c) and cane juice (b, d). Determination of OD 600nm and cell dry weight (a, b); levansucarse activity and levan content (c, d). Points with superscripts a-g differ from each other significantly at p<0.05

Conclusion
The present study reports the functional characterization of a thermostable levansucrase from B.subtilis BB03 expressed in E.coli. The recombinant enzyme was stable over a wide pH range and showed considerable activity in cane molasses and juice. Several studies reported lower levan yield using crude low-cost sources as compared to sucrose (Han and Watson, 1992;Moosavi-Nasab et al., 2010;Oliveira et al., 2007) with wild type system, this study was an attempt to obtain levan production in a heterologous system using sacB gene with its own promoter without IPTG induction. Further studies are needed for analyzing different media formulations which may include better clarification strategies for cane molasses and juice, in order to obtain higher levan production. However, this study provides useful knowledge about the functional and molecular properties of a thermostable levansucrase from Bacillus expressed in a heterologous system and demonstrate the potential use of low cost sucrose sources like cane molasses and juice for levan yield.