Correspondence: Lutz Fischer, Institute of Food Science and Biotechnology, Department of Biotechnology, University of Hohenheim, Garbenstr. 25, 70599 Stuttgart, Germany. Tel.: +49 711 459 23018; fax: +49 711 459 24267; e-mail: email@example.com
A novel expression system for Lactobacillus plantarum was developed. This system is based on the manganese starvation-inducible promoter from specific manganese transporter of L. plantarum NC8, which was cloned for the first time. The expression of a β-glucosidase from Pyrococcus furiosus (CelB) was achieved by cultivating L. plantarum NC8 at low manganese concentrations with MRS medium and the pmntH2-CelB expression vector. A CelB activity of 8.52 μkatoNPGal L−1 was produced in a bioreactor (4 L). The advantages of the novel expression system are that no addition of an external inducing agent was required, and additionally, no further introduction of regulatory genes was necessary. The new promoter meets the general demands of a food-grade expression system.
Lactic acid bacteria (LAB) are gram-positive, facultative anaerobic bacteria that are widely used in the food industry to produce various fermented foods (Konings et al., 2000). Because they do not generate endotoxins and are nonsporulating, they are generally recognized as safe (GRAS) by the U.S. Food and Drug Administration. There is widespread knowledge about the cultivation of LAB at bioreactor scale. Lactococcus lactis was considered as a model organism for a long time, in the last decade also other Lactobacilli became more deeply studied. For instance, Lactobacillus plantarum was established and used for metabolism studies as well as for the development of genetic tools (Siezen & van Hylckama Vlieg, 2011). For these reasons, interest in L. plantarum as a suitable production host for recombinant proteins has increased over the last decade. Here, the LAB have potential as delivery systems for valuable proteins like antibodies and antigens (Diep et al., 2009). To the particular demands of the food industry, Lactobacilli are desirable cell factories for the production of food-grade enzymes (Peterbauer et al., 2011).
Several different types of promoters from L. plantarum have been isolated to produce recombinant proteins. Some of them are constitutive, while a few of them are inducible (Rud et al., 2006). The most common expression systems are the NICE and pSIP expression systems (de Ruyter et al., 1996; Sørvig et al., 2003). The promoters of these systems are based on the regulatory system of antimicrobial peptides and the quorum sensing mechanism. Thus, for the induction of recombinant proteins in L. plantarum, an inducing peptide must be added during cultivation. The pSIP system is the most extensively used inducible system. It is derived from Lactobacillus sakei and uses the regulatory elements from the class IIb bacteriocins, either Sakacin A when using the pSIP403 vector or Sakacin P for the pSIP409 vector (Sørvig et al., 2003, 2005). The recombinant overproduction of heterologous proteins is induced by an externally added peptide pheromone. When these systems were employed in previous studies, different types of enzymes such as aminopeptidase N from L. lactis, β-glucuronidase from Escherichia coli, β-galactosidases from Lactobacillus sp., and a β-glucosidase from Pyrococcus furiosus were recombinantly expressed in L. plantarum and L. sakei with high yields (Sørvig et al., 2003; Halbmayr et al., 2008; Böhmer et al., 2012). Other types of auto-inducing expression systems were developed for L. lactis and Bacillus subtilis in recent years. These are based on phosphate or metal starvation transporter systems (Kerovuo et al., 2000; Sirén et al., 2008). Because phosphate is an essential component for bacterial growth, the promoters of these transport proteins are turned on under starvation conditions, and this can be used for recombinant protein expression. A similar system based on regulatory elements of high specific Zn2+ uptake during a Zn2+ starvation was developed for recombinant protein production in L. lactis (Llull & Poquet, 2004). Such systems can be more convenient for the recombinant production of proteins in some cases because they require no addition of any inducing agents. This will extend the tools for recombinant protein expression in Lactobacilli, and it completes the well-established systems with external inducers.
In addition to phosphate and Zn2+, trace amounts of manganese are of physiological importance in a variety of ways for almost all bacteria, including its need as cofactors of enzymes. Therefore, manganese ions are present in microorganism in very low amounts. However, in L. plantarum and some other LAB, high amounts of manganese up to 30 mM accumulate intracellularly due to the absence of superoxide dismutase (SOD) in these species (Archibald & Fridovich, 1981a, b; Archibald & Duong, 1984). The Mn2+ ions are used instead of the SOD to scavenge the toxic product superoxide (O2−) when the bacteria are grown in the presence of oxygen or during fermentative growth (Horsburgh et al., 2002). In the presence of Mn2+ ions, superoxide is chemically converted to hydrogen peroxide (H2O2), which is further enzymatically converted by a catalase. Thus, L. plantarum requires high concentrations of manganese in the standard MRS medium to be able to grow successfully. The uptake of Mn2+ ions occurs by different types of cation transporters with an active Mn2+ ion transport system (Hantke, 2005). The transporters were identified and described after the genome sequencing of L. plantarum WCFS1 in previous work by in silico analysis as ABC (ATP-binding cassette) transporter type (called mntH2) and Nramp (natural resistance associated macrophage protein) transporter (called mtsCBA) (Groot et al., 2005). Expression of these transporters during manganese starvation was verified by Northern blot analysis. In our work presented here, the promoter of the specific transport protein MntH2 was cloned and investigated for its use as novel promoter for recombinant protein production.
Material and methods
Chemicals and enzymes
All chemicals were of analytical grade or higher and were purchased from Sigma-Aldrich (Seelze, Germany) or Carl Roth (Karlsruhe, Germany). T4 DNA ligase and hexokinase/glucose-6-phosphate dehydrogenase (HK/G6PDH) were purchased from Roche (Mannheim, Germany). All restriction enzymes were from New England Biolabs (Frankfurt, Germany). HotStar HiFidelity Polymerase was purchased from Qiagen (Hilden, Germany).
Bacterial strains and culture conditions
Escherichia coli XL1-blue was purchased from Stratagene (Santa Clara), and was grown in Luria-Bertani medium at 37 °C with shaking (120 r.p.m.). Lactobacillus plantarum NC8 was obtained from culture collection by the Norwegian University of Life Science (Ǻs, Norway) and grown in MRS medium according to De Man et al. (1960) or in MRS media with different MnSO4 concentrations at 30 °C. Agar plates were solidified by adding 1.5% (w/v) agar. When required, erythromycin was added as follows: 200 μg mL−1 for E. coli and 10 μg mL−1 for L. plantarum.
Cultivation at microtiter plate scale
A microtiter plate-based Bioscreen C automatic cultivator (LabSystems, Finland) was used for growth analysis of L. plantarum at different MnSO4 concentrations (0–296 μM). Cultivation was performed in microtiter plates at 30 °C with continuous shaking. Determination of growth was performed in 250 μL of solution, and 240 μL of each media was inoculated with 10 μL of an overnight culture in the respective medium. Four wells were used for quadruplicate testing of each medium. Optical density at 600 nm (OD600 nm) was measured and readings for each well were taken every 15 min for 16–24 h.
Preparation of plasmids, cloning and transformation
Construction of the expression vector pmntH2-CelB was performed using E. coli XL1-blue and standard molecular biology techniques (Sambrook & Russel, 2001).
The promoter was amplified using the genomic DNA of L. plantarum NC8 as template. Isolation of genomic DNA was performed according to Sørvig (Sørvig et al., 2005). Primers (Table 1) were constructed using the sequenced genome from L. plantarum WCFS1 (Kleerebezem et al., 2003, GenBank: AL935263.2). The resulting PCR product of approximately 500 bp in length was cloned in the pSIP409-CelB vector using SalI and NcoI restriction sites to remove the PorfX promoter and regulator elements of this induction system. This generated the expression plasmid pmntH2-CelB. The plasmid was transformed into E. coli. Plasmids used in this study are shown in Table 2. Plasmid DNA was purified from E. coli by the QIAprep Spin Miniprep Kit from Qiagen and subsequently transformed into L. plantarum by electroporation (Aukrust & Blom, 1992). Positive clones were selected by plating on MRS agar containing 10 μg mL−1 erythromycin and incubation for 24–48 h at 30 °C under anaerobic atmosphere using an anaerobic jar and Anaerocult A reagent from Merck (Darmstadt, Germany).
Table 1. Primers used in this study
Table 2. Plasmids used and constructed in this study
Initial expression experiments were performed aerobically in baffled shaking flasks at 30 °C. Freshly prepared L. plantarum clones were grown overnight in test tubes (5 mL scale) and transferred to 100 mL MRS media with erythromycin and different MnSO4 concentrations. The cells were harvested by centrifugation (10 min, 8000 g) after the stationary growth phase was reached, washed, and resuspended with sodium acetate buffer (50 mM, pH = 5.0). Cells (30% w/v) were disrupted by sonification using an ultrasonic processor UP 200S (Hielscher Ultrasonic, Teltow, Germany). Crude extract was obtained after separating the cell debris by centrifugation (10 min, 8000 g, 4 °C). The best performing MnSO4 concentrations for recombinant expression of CelB were determined by analysis of the CelB activity.
Fermentation experiments in the bioreactor
Expression experiments were continued in a Biostat E fermenter (B. Braun, Melsungen, Germany) with 4 L working volume. Lactobacillus plantarum was cultivated at 100 r.p.m., anaerobic with N2 gassing < 0.1 vvm or aerobic with O2 gassing with pO2 > 30%, 0.5 vvm, and pH = 6.45, controlled with 2 M NaOH at 30 °C. The experiments were run in batch mode. Samples were withdrawn throughout the fermentation to determine biomass, glucose concentration, CelB activity, and manganese concentrations. Before bioreactor inoculation, precultures were first grown overnight at 5 mL, followed by precultures at 50 mL scale overnight, and finally were grown overnight at 400 mL scale in baffled shaking flasks in the appropriate medium under aerobic conditions. The cells were harvested and disrupted as described above.
Cell growth was monitored by measuring the optical density at 600 nm (OD600 nm). Biomass was quantified gravimetrically as cell dry weight of cells (CDW). Samples were centrifuged, twice washed with saline, and dried in preweighted tubes at 40 °C at 10 mbar in a RVC 2-33 IR vacuum centrifuge from Christ (Osterode, Germany). Protein concentrations were determined by the method of Bradford (Bradford, 1976). The glucose concentrations were measured in microtiter plates by a photometric assay at 340 nm. HK/G6P-DH was used as the coupling enzyme for the reaction, and the technique was based on the manufacturers protocol for the D-glucose/D-fructose test kit (R-Biopharm AG, Darmstadt, Germany; product code 10 139 106 035). The total Mn2+ content was measured with a Perkin-Elmer model 2380 atomic absorption spectrophotometer. For intracellular Mn2+ determination, lyophilized cells were digested overnight in 70% nitric acid at 37 °C. The digestion mixture was diluted with water to a final nitric acid concentration of 10% before measurement with atomic absorption spectrophotometer.
Enzyme activity measurements
CelB activity (EA) was determined using o-nitrophenyl-β-D-galactopyranoside (oNPGal) as a substrate at 75 °C with 50 mM sodium acetate buffer, pH 5.0, in 1 mL of solution as described previously (Mayer et al., 2010). One nanokatal is defined as the amount of enzyme that catalyzes the release of 1 nmol of o-nitrophenol from oNPGal per second.
The native protein in cell-free crude extracts used for the enzyme solution was heat-denaturated for 15 min at 75 °C. After that the heat-denaturated, interfering native Lactobacillus protein was precipitated by centrifugation (10 min, 8000 g, 4 °C). CelB activity was calculated with an absorption coefficient of 0.495 L mmol−1 cm−1. All measurements were performed in triplicate.
Results and discussion
Growth analysis of L. plantarum at different Mn2+ concentrations
As previously reported, due to the absence of a SOD, L. plantarum needs a relative high concentration of manganese ions for optimal growth (Archibald & Fridovich, 1981a, b; Archibald & Duong, 1984). Lactobacillus plantarum NC8 was grown in MRS media supplemented with decreasing Mn2+ concentrations (0–296 μM) to determine the limiting concentration for this ion. The growth investigations by optical density (OD600 nm) were performed in microtiter plate-based cultivations as described above. The results are shown in Fig. 1. The highest OD600 nm of 1.78 was reached in microtiter scale with the highest Mn2+ ion concentration tested (296 μM). Up to Mn2+ ion concentrations of 100 μM, a clear limitation of growth was observed. These results illustrated that the growth of L. plantarum NC8 was directly correlated to the amount of Mn2+ ions in the medium. It was in accordance with previously published data, when the extracellular concentration of MnSO4 was below 150 μM, its intracellular concentration became dependent on the quantity of Mn2+ ions added in the culture medium (Archibald & Duong, 1984). In chemically defined media, no growth was observed without manganese salt supplementation (Hao et al., 1999; Groot et al., 2005), but in complex media, such as the MRS medium used in the present study, slight growth was detectable even without adding any MnSO4 into the medium. This observation was also described in previous literature (Watanabe et al., 2012) and is due to trace amounts of manganese in the complex compounds of MRS medium, such as yeast and meat extracts or tryptone.
Performance of the promoters in MRS medium without MnSO4
The promoter region of the gene encoding the manganese transporters MntH2 from L. plantarum WCFS1 has been studied (Groot et al., 2005). As the gene is induced by manganese starvation, we amplified a 496 bp DNA fragment encompassing of the mntH2 promoter region. This promoter sequence included the transcription start site and Shine–Dalgarno sequence. Additionally, parts of the sequence (19 bp) are similar to target regions for the metalloregulator MntR from B. subtilis (Que & Helmann, 2000). The binding site of the ScaR regulator protein from Streptococcus gordonii was a part of the promoter sequence of PmntH2, as well (Jakubovics et al., 2000). The promoter was cloned upstream from the CelB gene as reporter gene in the pSIP409-CelB vector replacing the PorfX promoter and regulator elements. This resulted in the plasmid pmntH2-CelB. The vector backbone of the pSIP409-CelB vector was used, which consists of replication origins for E. coli (pUCori) and Lactobacilli (256rep), an erythromycin resistance marker (ermL) and the pepN terminator (Sørvig et al., 2003; Böhmer et al., 2012). As described by Sørvig et al. (2005), the used minimal replicon 256rep results in a copy number of about 6 and is known to replicate via a theta mechanism.
The expression performance of PmntH2 was investigated using the thermophilic glucosidase (CelB) from P. furiosus (Voorhorst et al., 1995) as reporter gene. Initial expression experiments with L. plantarum transformants were performed in baffled shaking flasks in MRS medium without any MnSO4 (aerobically, pH 6.45, 30 °C). Lactobacillus plantarum pmntH2-CelB achieved a CelB activity of 17 nkatoNPGal mgprotein−1. Thus, the proof of principle for recombinant protein production using a manganese starvation–based promoter was demonstrated. For the next expression experiments, different amounts of MnSO4 were added to the MRS medium to analyze the dependency of PmntH2-celB expression of Mn2+ concentration in the growth medium (see Table 3). The highest specific CelB activity of 20.27 nkatoNPGal mgprotein−1 was reached at 10 μM MnSO4.
Table 3. Growth and CelB activity of Lactobacillus plantarum NC8 pmntH2-CelB at different concentrations of MnSO4 (baffled shaking flasks, anaerobic conditions, 100 mL medium, 30 °C)
MnSO4 concentration (μM)
Final OD600 (−)
CelB activity (nkatoNPGal mgprotein−1)
17.55 ± 0.43
19.04 ± 0.12
20.27 ± 0.51
16.36 ± 0.18
9.60 ± 0.10
7.67 ± 0.17
5.63 ± 0.15
2.63 ± 0.07
Our results demonstrated that in addition to the reported phosphate starvation promoter system for Lactococcus and Bacillus species (Kerovuo et al., 2000; Sirén et al., 2008), another starvation promoter system based on manganese ions can be applied to suitable hosts such as Lactobacilli. A similar system was developed for L. lactis using promoter and repressor protein of a zinc uptake system (Llull & Poquet, 2004). Using the Zn2+ starvation-inducible system, induction factors of ~ 50 were reached after zinc consumption during cell growth. In our study, an induction factor of ~ 10 was detected comparing the medium with 10 μM MnSO4 to the medium with 296 μM MnSO4. Induction factors obtained with some other promoters using LAB systems are quite in the same order of magnitude (de Vos, 1999). To the best of our knowledge, this is the first time a starvation promoter expression system was used successfully in a ‘food-grade’ Lactobacillus host that did not need any addition of an inducing agent. It may also be possible to use this food-grade promoter PmntH2 for the production of live vaccines in L. plantarum as delivery vehicle, as was previously discussed with different type of LAB as host organisms, and for the expression of therapeutic proteins or vaccines (Renault, 2002; Wells & Mercenier, 2008; Diep et al., 2009).
Bioreactor experiments with L. plantarum pmntH2-CelB
The performance of the CelB production of L. plantarum pmntH2-CelB in the bioreactor using MRS medium with 20 μM MnSO4 is shown in Fig. 2. To enhance the amount of biomass formation, and still have a good induction level of PmntH2, 20 μM MnSO4 was chosen as limiting concentration of Mn2+ . Contrary to the shaking flasks experiments described above, the cultivation was performed anaerobic (N2 gassing) to avoid unwanted oxidative damage by superoxide under Mn2+ limiting conditions. Fig. 2 illustrates the bioreactor cultivation of L. plantarum pmntH2-CelB under N2 gassing conditions. The maximal achieved biomass of 4.0 g L−1 cell dry weight is equal to an OD600 nm = 10.8 under manganese starvation conditions (20 μM MnSO4). This is about 11% lower than the biomass obtained by bioreactor cultivation in standard MRS medium (296 μM MnSO4), where a cell dry weight of 4.5 g L−1, equal to an OD600 nm = 12.3, was reached. A bioreactor cultivation was also performed under aerobic (air gassing) conditions using the same medium. Using this technique, we determined whether the expression rate of the promoter PmntH2 was influenced by oxygen. The biomass as well as the expression performance of PmntH2 by aerobic cultivation (data not shown) was quite the same to the cultivation with N2 gassing. Therefore, the presented system may be easily applicable due to less technical demands in kind of N2 gassing. In other studies, it was shown that L. plantarum grown aerobically in standard MRS medium (with 296 μM MnSO4), resulted in higher OD600 nm values than when grown under anaerobic conditions (Brooijmans et al., 2009). However, Watanabe et al. (2012) ascertained that this growth difference did not occur when the cultivation was performed without the addition of MnSO4 into the MRS medium. This was also the case in our studies in which we used a very low MnSO4 supplementation of 20 μM.
Maximal specific CelB activities were obtained in aerobic and anaerobic bioreactor cultivations with 22.4 ± 0.9 nkatoNPGal mgprotein−1 and 20.9 ± 0.5 nkatoNPGal mgprotein−1, respectively (anaerobic see Fig. 2). So, the induction of PmntH2 depended not on oxygen. In the bioreactor cultivations, the specific CelB activities were approximately as high as in the shaking flask experiments (see above). The biomass was four times higher in the bioreactor, and a maximal volumetric activity of 8.52 μkatoNPGal L−1 was obtained after 14 h of cultivation under N2 gassing conditions.
Kinetic of PmntH2 induction during manganese depletion
Analysis of the induction kinetics of L. plantarum pmntH2-CelB in medium with 20 μM MnSO4 is shown in Fig. 3 and was performed by quantification of the extracellular and intracellular manganese concentrations by AAS. In the beginning of the cultivation, the manganese accumulates intracellular up to concentrations of about 40 μmol gcdw−1 due to an uptake of manganese. In standard MRS medium with 296 μM MnSO4, the intracellular manganese accumulates with 76 μmol gcdw−1 almost twice as high. An intracellular enrichment of manganese due to transport systems as protection mechanism against the damaging effect of oxygen radicals in L. plantarum is described in the literature (Archibald & Fridovich, 1981a, b; Archibald & Duong, 1984; Groot et al., 2005). The mntH2-promoter was induced by manganese depletion due to the bacterial growth in the medium with 20 μM MnSO4 by auto-induction. An increase in CelB activity was visible after 8 h of cultivation in the mid-exponential phase of growth when extracellular manganese decreases less than 1.5 μM and the intracellular manganese decreases due to starvation conditions less than 10 μmol gCDW−1. No increase of CelB activity was recognized in standard MRS medium, where a 20 times lower activity of 1.1 nkatoNPGal mgprotein−1 was detected. The intracellular manganese concentration did not reduce below values of 31 μmol gcdw−1, also the extracellular concentration did not decrease below 19.7 μM. So, no inducing concentrations were reached due to an excess of MnSO4 resulting in low induction of the mntH2 promoter. In the literature, no expression of the MntH2 transporter was described also at Mn2+ concentrations of about 100 μM or higher (Groot et al., 2005).
The proof of principle of the recombinant protein expression by manganese starvation in L. plantarum was successfully demonstrated, although further improvement to the expression system will be needed before an industrial application will become economically feasible. Nevertheless it has to be mentioned, that in comparison with the established pSIP409-CelB system the activity was 60 times lower, in this system a specific CelB activity of 675 μkatpNPGal L−1 was achieved (Böhmer et al., 2012). The presented auto-inducing pmntH2 expression construct may be beneficial in applications, were no high yields of recombinant proteins are necessary, for example Lactobacilli as food-grade live vaccines.
The authors wish to thank Dr. Lars Axelsson for providing the plasmid pSIP409 and L. plantarum NC8. Further thanks to the group of Prof. de Vos and Dr. Kengen, University of Wageningen (NL), for the gift of the CelB gene. Grateful thanks to Prof. Dr. Schwack, University of Hohenheim, Institute of Food Chemistry, for technical support with the Perkin-Elmer model 2380 atomic absorption spectrophotometer. Parts of the research were financed by the German Federal Ministry of Economics and Technology (AiF/FEI Project No. 15801N), which is greatly appreciated.