Secretory production of a beta-mannanase and a chitosanase using a Lactobacillus plantarum expression system

Abstract

Background: Heterologous production of hydrolytic enzymes is important for green and white biotechnology since these enzymes serve as efficient biocatalysts for the conversion of a wide variety of raw materials into value-added products. Lactic acid bacteria are interesting cell factories for the expression of hydrolytic enzymes as many of them are generally recognized as safe and require only a simple cultivation process. We are studying a potentially food-grade expression system for secretion of hydrolytic enzymes into the culture medium, since this enables easy harvesting and purification, while allowing direct use of the enzymes in food applications.

Results: We studied overexpression of a chitosanase (CsnA) and a β-mannanase (ManB), from Bacillus licheniformis and Bacillus subtilis, respectively, in Lactobacillus plantarum, using the pSIP system for inducible expression. The enzymes were over-expressed in three forms: without a signal peptide, with their natural signal peptide and with the well-known OmpA signal peptide from Escherichia coli. The total production levels and secretion efficiencies of CsnA and ManB were highest when using the native signal peptides, and both were reduced considerably when using the OmpA signal. At 20 h after induction with 12.5 ng/mL of inducing peptide in MRS media containing 20 g/L glucose, the yields and secretion efficiencies of the proteins with their native signal peptides were 50 kU/L and 84% for ManB, and 79 kU/L and 56% for CsnA, respectively. In addition, to avoid using antibiotics, the erythromycin resistance gene was replaced on the expression plasmid with the alanine racemase (alr) gene, which led to comparable levels of protein production and secretion efficiency in a suitable, alr-deficient L. plantarum host.

Conclusions: ManB and CsnA were efficiently produced and secreted in L. plantarum using pSIP-based expression vectors containing either an erythromycin resistance or the alr gene as selection marker.

Keywords: Alanine racemase; Chitosanase; Food-grade, Bacillus; L. plantarum; OmpA; Secretion; Signal peptide; pSIP; β-Mannanase.

Fig. 1

Vector construction. Both pSIP409 and pSIP609 vectors were used for expression of B. licheniformis β-mannanase (BlManB) and B. subtilis chitosanase (BsCsnA). Enzyme expression was under the control of the P_orfx promoter (also known as P_sppQ) [32], which can be induced by the 19-residue peptide pheromone IP-673. The vectors contain an erythromycin resistance (erm ^R) or an alanine racemase (alr) gene as selection marker, for pSIP409 and pSIP609, respectively. Polyhistidine tags were incorporated C-terminally to facilitate one-step affinity purification. The 256_rep replicon allows DNA replication in L. plantarum. Each enzyme was cloned in three forms, two of which contain a signal peptide for secretion (native or OmpA). The genes marked in red constitute the two-component system needed for peptide-pheromone driven induction; the grey areas marked with a T are terminator sequences

Fig. 2

Production and secretion of B. licheniformis β-mannanase (BlManB, a) and B. subtilis chitosanase (BsCsnA, b) using pSIP409-type constructs containing the native or the OmpA signal peptide. The recombinant L. plantarum strains were batch-cultured in 3-L vessels with pH control at pH 6.5. After harvesting at various time points, volumetric enzyme activities in both the culture supernatant and the cell lysate were determined as described in the “Methods” section. Data given are the average of two independent experiments ± their standard deviation. The lines connecting the points are drawn for illustration purposes only. Key data, supplemented with data for the constructs without signal peptide are summarized in Table 1

Fig. 3

SDS-PAGE analysis of culture supernatants and purified BlManB and BsCsnA. The Coomassie-stained gels illustrate the purification of B. licheniformis ManB (a) and B. subtilis CsnA (b) from culture supernatants. For crude culture supernatants, 20 µL of sample was loaded, whereas for the purified enzymes a total of 20 and 5 µg protein of ManB and CsnA samples, respectively, were loaded. M indicates the Kaleidoscope protein standard (Bio-Rad). Detailed information on the protein contents and enzyme activities in these samples are provided in Table 1

Fig. 4

Signal peptide cleavage sites. N-terminal sequence analysis of purified secreted proteins was performed by Edman degradation. The sequences of the native Bacillus signal peptides and the E. coli OmpA signal peptide are underlined. Arrows indicate the cleavage sites as deduced from sequence analysis. The first five amino acids obtained by Edman degradation are colored and boxed. Amino acids in red indicate extra amino acids that were introduced during genetic engineering of the fusions between the Bacillus enzymes and the E. coli OmpA signal peptide

Fig. 5

Effects of the selection marker on production of BlManB (a) and BsCsnA (b). The figures show the volumetric activities in culture supernatants and cell lysates of L. plantarum strains harboring pSIP409-based or pSIP609-based expression vectors, which are based on using erm ^R or alr as selection marker, respectively. All conditions were as in Fig. 2. Data given are the average of two independent experiments ± their standard deviation. The lines connecting the points are drawn for illustration purposes only

Fig. 6

Production of BlManB and BsCsnA using a food-grade expression system. L. plantarum TLG02 strains, harboring pSIP609/BlManB_nt [(BlManB_nt(alr)] or pSIP609/BsCsnA_nt [(BsCsn_nt (alr)], were cultured in 3-L batch fermentations in MRS medium, containing 20 or 40 g/L glucose as indicated. The upper panels show Coomassie-stained SDS-PAGE gels with culture supernatants collected at various time points (20 µL of sample per lane). The bottom panels show the time course of the cultivation corresponding to the SDS-PAGE samples. The pH of the batch cultivation was kept constant at 6.5. Note that doubling the amount of glucose led to a doubling of OD₆₀₀, but only to a small increase in enzyme production levels (Table 2)