B. licheniformis strain DSM13 was used as the source for the isolation of the mannan endo-1,4-β-mannosidase gene, manB since this strain has been used extensively for large-scale production of various industrial enzymes including serine protease (subtilisin) or α-amylase . The genome of strain DSM13 has recently been sequenced, and a number of new genes of potential biotechnological applications have been identified . The mannan endo-1,4-β-mannosidase gene was cloned by PCR cloning, using primers designed from the published genome database. This is the first report on the cloning, expression, and characterization of recombinant mannan endo-1,4-β-mannosidase from B. licheniformis. Other reports on recombinant Bacillus mannan endo-1,4-β-mannosidases were dealing with enzymes from B. subtilis [16, 19–23]. B. stearothermophilus , and B. circulans .
Mannan endo-1,4-β-mannosidases can be classified into two distinct families, glycosyl hydrolase (GH) family 5 and 26, based on amino acid sequence similarities and hydrophobic cluster analysis . Family GH5 was formerly known as cellulase family A and encompasses diverse enzymes , whereas glycosyl hydrolase family 26 comprises only members with mannan endo-1,4-β-mannosidase (EC 18.104.22.168) and β-1,3-xylanase (EC 22.214.171.124) activities . Amino acid sequence analysis of mannan endo-1,4-β-mannosidase from B. licheniformis revealed that the enzyme belongs to family GH26. In addition, we also cloned and expressed the mannan endo-1,4-β-mannosidase gene (manB) from B. licheniformis strain DSM 8785. The two enzymes have only one amino acid different, and the properties of these two heterologously expressed recombinant enzymes are identical (data not shown).
The expression and production of the recombinant mannan endo-1,4-β-mannosidase reported here is based on a previously published E. coli expression system . The mature mannan endo-1,4-β-mannosidase gene was fused to the E. coli ompA signal sequence and is under the control of tac promoter. Thus, the enzyme could be efficiently secreted, and harvested from the culture medium, periplasm, or cell lysate fraction, depending on the culture condition. When the gene was induced for over-expression by 1 mM IPTG for 3 - 4 h, we routinely obtained about 25 mg of recombinant enzyme from the cytoplasmic and periplasmic extracts of 1-liter cultures, which contain more than 40,000 units of purified enzyme. Under the induction with IPTG, a significant fraction of the enzyme was still found in the cytosol. This could indicate that the over-expressed enzyme possibly saturates the bacterial secretion system . It should be mentioned that no optimization aiming at increased enzyme yields was performed. Thus, by applying optimized culture and induction conditions together with a suitable fermentation strategy, considerably higher recombinant protein yields can be expected. Thus, our expression system is highly efficient for expression of bacterial β-mannanses and should be applicable for other enzymes as well. More importantly, the extracellular location of the enzyme might be of interest for large-scale cultivations as it circumvents the necessity of cell disruption.
Mannan endo-1,4-β-mannosidases are active on various mannans and substituted mannans, but display negligible to low activity towards other plant cell wall polysaccharides [3, 28]. The enzymes randomly hydrolyse β-1,4-linkages in diverse substrates such as pure mannans, galactomannans, glucomannans and galactoglucomannans . In this study, we found that B. licheniformis ManB shows the highest relative activity for glucomannan prepared from konjac followed by pure low-molecular mass 1,4-β-D-mannan of DP (degree of polymerization) < 15 and high-viscosity (high molecular mass) locust bean gum. However, we were not able to detect notable activity for guar gum and copra meal using the standard assay of 5-min incubation. Based on the kinetic characterization and judged from the specificity constant kcat/Km, the galactomannan locust bean gum (low viscosity) is the preferred substrate, however the differences in the specificity constant are not very pronounced when compared to konjac glucomannan and pure mannan. Apparently, B. licheniformis ManB prefers soluble and low-substituted mannan substrates. This is evident from a comparison of the relative activity on soluble LBG, a galactomannan from Ceratonia siliqua with a mannose-to-galactose ratio of 4:1, and soluble guar gum, a galactomannan from Cymopsis tetragonoloba with a mannose-to-galactose ratio of 2:1 . While the former is a good substrate, the activity on the latter is negligible during the 5-min standard assay. Similarly, activity on copra mannan, an insoluble galactomannan with a very low degree of galactosyl substitution, is very low .
There have been a number of reports on the characterization of mannan endo-1,4-β-mannosidases, both native and recombinant, from various organisms as summarized in Additional file 1. The pH and temperature optima as well as the stability of the enzymes are clearly varying, depending on the sources of the enzymes. Typically, the enzymes from non-bacterial sources show lower pH and temperature optima as well as lesser stability (See Additional file 1). The specific activity (from 3.8-8300 U/mg) and kinetic parameters (Km ranging from 0.3-10.2, Vmax from 3.8-2000) of the mannan endo-1,4-β-mannosidases from various sources, when using LBG as a substrate, vary greatly as shown in Additional file 1. This obviously reflects differences in the structure of the enzymes, for example highly thermostable mannan endo-1,4-β-mannosidase tend to have lower specific activity compared to their mesophilic counterparts [24, 30, 31]. In this respect, the B. licheniformis ManB described in our report is characterized by a very high specific activity of 1672 U/mg as well as by a relatively high stability. However, when comparing different mannan endo-1,4-β-mannosidases it is important to note that locust bean gum, which is a standard substrate for measuring mannan endo-1,4-β-mannosidase activity, is highly viscous and difficult to prepare. It can be assumed that the large discrepancy of enzyme activity in some of the reports can in part result from various techniques used in substrate preparation. For example we were not able to estimate with confidence the kinetic parameters when using high-viscosity, commercial LBG as a substrate. Thus, only the kinetic parameters when using low-viscosity LBG, low-viscosity glucomannan from konjac, and β-mannan are reported here.
TLC analysis of hydrolysis products confirmed that recombinant B. licheniformis mannanse is an endo-mannanase, which can efficiently and randomly cleave higher molecular weight mannans containing more than six mannose monomers. The enzyme could only cleaved mannopentaose after an extended incubation for 12 h and had no detectable activity against mannobiose, -triose or -tetraose. This property suggests that this enzyme could be applicable for the generation of prebiotic manno-oligosaccharides (MOS), as higher oligosaccharides formed will not be hydrolyzed further. Extensive hydrolysis of cheap and commercial available locust bean gum can therefore result in a mixture of MOS containing various oligosaccharides that may have a diverse prebiotic and anti-obesity  effects in different regions of the gut. Higher oligosaccharides are currently discussed as prebiotics with enhanced persistence that can reach more distal regions of the gut, and thus show their positive effect also in that region .