At present, protocols for nisin-induced gene expression only exist for acidifying batch cultures (see e.g. 719) but not for pH-controlled batch cultures. Therefore, a protocol that was previously developed at our laboratory for large-scale lysostaphin production 15 was used as a starting point. The cultivation and induction conditions were as follows: pH controlled (using NaOH) growth at pH 6.5, growth temperature 30°C, inoculum 1%, induction at OD₆₀₀ = 1 (= 0.3 g/L cell dry weight 20) with 10 ng/mL nisin, and harvest after 6 h. The initial medium composition was: 5% lactose, 1.5 % soy peptone, 1% yeast extract, 1 mM MgSO₄ and 0.1 mM MnSO₄.

In a series of 58 1-L fermentations the following parameters were tested and optimized: pH, neutralization agent, addition of phosphate and zinc, fermentation temperature, the time point of induction, i.e. cell density, the concentration of peptone, yeast extract and lactose in the medium and the lysostaphin production time after induction.

The main fermentation end product of L. lactis is lactate, which is toxic and retards growth above a certain concentration. The toxic effect of lactate depends on its undissociated form 21. Therefore, any increase of the pH can lead to a prolongation of growth in the presence of higher concentrations of lactate. NaOH is a more alkaline neutralizing agent than NH₄OH and could cause cell damage. Furthermore, NH₄OH could contribute to the ammonium metabolism of the cell 22. Figure 2 shows that neutralization with NH₄OH has a small effect, however the combination of NH₄OH with an increase of the fermentation pH to 7.0 leads to prolonged exponential growth (not shown) and to a higher final cell density. Increase of the pH to 7.5 leads to a retardation of the growth rate and has therefore not been further pursued.

0.01% of sodium phosphate (Na₂HPO₄ *2 H₂O) was added to evaluate the effect on growth performance. Furthermore, zinc was specifically added, because lysostaphin is a metallo-enzyme with zinc as co-factor 16. Elemental analysis of the basic medium had shown that it contained 1.2 mg/L Zn²⁺. This amount would be sufficient for a maximum of 500 mg/L lysostaphin if all the zinc were available. Therefore, 100 μM (16 mg/L) ZnSO₄ was added, which would allow the production of a maximum of 2.5 g/L active lysostaphin. Addition of these components had no significant effect on growth under the initial fermentation and induction conditions (Figure 3). However, in a later stage when induction was performed at a higher cell density and with more nisin (induction at OD₆₀₀ = 5 with 40 ng/ml nisin) considerably less lysostaphin was formed when phosphate was omitted (150 mg/L without phosphate versus 220 mg/L lysostaphin with phosphate).

After these initial experiments, it was decided to set the following basic fermentation conditions: pH at 7.0, to use NH₄OH as neutralizing agent, to add both 0.01 g/L sodium phosphate (Na₂HPO₄ *2 H₂O) and 100 μM ZnSO₄. To establish a baseline for lysostaphin production and to analyze the effect of phosphate or zinc on lysostaphin production, an induction experiment was carried out. Induction of lysostaphin production was initiated at OD₆₀₀ = 1.0 with 10 ng/mL nisin. Figure 3 shows the basic pattern of growth and induction. After the addition of nisin, product formation begins immediately, and occurs parallel to growth. When lysostaphin production was induced, growth of the culture slowed down considerably about 30 min after induction. This is accompanied by a steep drop of the viable counts on lactose M17 agar plates of 4 orders of magnitude. However, despite the growth retardation, lysostaphin production was linear for 8 h before it abruptly ended (Figure 3) 15.

Overproduction of proteins can trigger stress responses and thus e.g. degradation of the target protein in the cell 23. One of the strategies employed to prevent or minimize this effect is to lower the fermentation temperature and thereby lower the threshold for the induction of the stress response, and also slow down protein production. Figure 4 shows both growth and lysostaphin production during fermentation at 30°C, 25°C and 20°C. Induction was carried out at a cell density of OD₆₀₀ = 1 (0.3 g/L cell dry weight) with 10 ng/mL nisin. At 25°C and 20°C growth is, as expected, slower than at 30°C. Also lysostaphin production is slower and much lower at 20°C as compared to 30°C. However, these experiments also show that the NICE system works at these temperatures; especially at 25°C similar yields are obtained as at 30°C. However, because of the faster production and the slightly higher yields, 30°C was chosen as fermentation temperature for all subsequent experiments.

If induction could be performed at higher cell densities a higher yield per volume of culture fluid could be expected. Since at higher cell densities more nisin may also be needed for complete induction, these parameters were tested together. Initial experiments showed that induction could be performed at an OD₆₀₀ of 3 or even 5. Figure 5 shows an experiment were induction was tested at OD₆₀₀ values of 1, 5 and 7 (equal to a cell dry weight of 0.3 g/L, 1.5 g/L and 2.1 g/L, respectively) with nisin concentrations of up to 80 ng/mL. Clearly lysostaphin production can be increased when the culture is induced at higher cell densities. At the same time, more nisin needs to be added for maximum induction. When induction was performed at OD₆₀₀ = 5, 160 mg/L lysostaphin was formed with 20 ng/mL nisin and 220 mg/L lysostaphin was produced when 40 ng/mL nisin was used for induction. This means that there was a clear correlation between the cell density at induction and the amount of nisin that is needed for maximal induction. Induction at even higher cell densities, OD₆₀₀ = 7 with 40 and 80 ng/mL nisin, was also tried; however this did not lead to higher lysostaphin yields, indicating that there is a limit in the growth cycle beyond which induction becomes ineffective. Exponential growth without induction continues until about OD₆₀₀ = 8, after which growth slows down as the culture slowly approaches the final cell density of about 15. This is due to the accumulating lactate that increasingly interferes with the energy metabolism of the cell 2425. The same mechanism also affects and limits the production of the recombinant protein. Induction at a cell density of OD₆₀₀ = 5 with 40 ng/mL nisin resulted in a maximum yield of the recombinant protein.

Although induction was successful at a five times higher cell density than the original induction conditions, we observed only about 2 – 2.5 times more product formation. This could be due to either a shortage in building blocks, i.e. amino acids or a limitation of the energy supply, i.e. the fermentable sugar. Therefore, the next step was to supply more nitrogen and carbohydrate sources. Initial experiments showed that the lysostaphin yield increases when either more yeast extract or more peptone was added and increases even more when both nutrients were increased simultaneously (results not shown). Figure 6 shows a series of fermentations in which the influence of a combination of nitrogen and carbon sources on product formation was investigated. Lysostaphin production was induced at OD₆₀₀ = 5 (1.5 g/L cell dry weight) with 40 ng/mL nisin. The addition of extra carbon source, 7% lactose versus 5% lactose, did not significantly increase the production of lysostaphin. However, doubling of the carbon source did raise lysostaphin yield from 225 mg/L to 290 mg/L. An increase in the supply of the nitrogen sources peptone and yeast extract also raised the lysostaphin yield form 225 mg/L to 290 mg/L. In a final experiment we used the best medium (2.5% peptone, 2% yeast extract and 7% lactose) and looked again at the combination of cell density and the amount of nisin for induction (Figure 7). Clearly, either lower or higher cell densities (OD₆₀₀ = 4 or 7) lead to slower production and lower yields. The same is true for the addition of more nisin (60 ng/mL), which at that concentration probably has a detrimental effect on the production of lysostaphin.

The highest production of lysostaphin was observed under the following conditions: the pH is fixed at 7.0, the temperature is 30°C, the neutralizing agent is NH₄OH, medium components are 7% lactose, 2.5% peptone, 2% yeast extract, 0.01% sodium phosphate (Na₂PO₄ *2 H₂O), 100 μM ZnSO₄, 1 mM MgSO₄ and 0.1 mM MnSO₄, cells are grown to an OD₆₀₀ = 5 and induced with 40 ng/mL nisin. Lysostaphin production proceeded for 6 to 8 hours and then abruptly stopped. After that time there may be a small decrease of lysostaphin concentration over time. However, this depended on the medium composition and the fermentation conditions (Figures 4, 6 and 7).

Lactococcus lactis subsp. cremoris NZ3900 26 carrying plasmid pNZ1710 (plasmid with gene for mature lysostaphin under control of the nisA promoter; 15) was stored as frozen stock in M17 medium 27 containing 0.5% lactose as carbon source and plasmid selection agent. The basic fermentation medium contained 5% lactose (Lactochem, Borculo Domo Ingredients, Zwolle, The Netherlands), 1.0% yeast extract (Biospringer, Maisons-Alfort, France), 1.5% soy peptone (Merck, VWR International, Amsterdam, The Netherlands), 1 mM MgSO₄ and 0.1 mM MnSO₄. All components were dissolved in water and sterilized at 110°C for 20 min. Additional components such as Na₂HPO₄ and ZnSO₄ were filter sterilized and added separately.

Nisin for induction was prepared as follows: 0.04% nisin powder (Sigma-Aldrich Chemie, Zwijndrecht, The Netherlands) was dissolved in 0.05% acetic acid and precipitated proteins were removed by centrifugation. Nisin was added for induction as indicated in the Results section.

The strain was taken from stock and sub-cultured twice before inoculation. 1-L Applicon fermenters were used coupled to an Applicon Biocontroller ADI 1030 (Applicon, Frederiksberg, Denmark) for pH and temperature control. The pH of the culture was controlled at the indicated values (see Results) with either 2.5 M NaOH or 2.5 M NH₄OH. The cell density was measured in samples that were taken at regular intervals, by determining the absorbance at 600 nm with a path length of 1 cm [OD₆₀₀]. For lysostaphin measurements, cells of appropriate samples were sedimented by centrifugation and stored at -20°C. After thawing, cell density was adjusted to OD₆₀₀ = 10 (according to the initial cell density reading) and 1 ml of the cell suspension was mixed with 1 g glass beads (0.1 mm Zirconia/Silica beads from Biospec products, Bartlesville, OK., U.S.A.) in screw-cap Eppendorf tubes. Subsequently, cells were subjected to bead-beating in a FastPrep beadbeater (FP120, QBiogene, Irivne, CA, U.S.A.), adjustment 4, 4 × 30 s. After the beads were allowed to sediment, lysostaphin concentration was determined in the whole cell extract.

Lysostaphin was quantified using SDS-capillary zone electrophoresis (SDS-CZE). The SDS sample buffer was prepared by dissolving in ca. 80 mL of water 606 mg of tris(hydroxymethyl)aminomethane (Tris), 1.00 g of sodium dodecylsulphate (SDS) and 37 mg of EDTA. Hydrochloric acid (0.1 M, 14.7 mL) was added and the solution was made up to 100 mL. Daily, 25 mg of DTT was added to 10 mL SDS sample buffer. This solution was used to dissolve cell extract samples (see below).

The SDS-CZE separation was performed using a Beckman Coulter P/ACE MDQ capillary electrophoresis system (CA, U.S.A.), equipped with a UV-detector operating at 214 nm using a 30 cm coated capillary and the separation buffer of the Beckman Coulter SDS 14–200 kit. The capillary temperature was set at 20°C.

Before each analysis, the 30 cm capillary was rinsed in the reverse direction for 1 min at 20 psi using 0.1 M HCl and subsequently for 3 min at 20 psi with SDS separation buffer. The sample solution was injected for 30 s at 1 psi, followed by forward rinse for 30 s at 0.5 psi of SDS sample buffer/water (1:1). The separation voltage was ramped in 1 min to 9 kV (ground at detector outlet) and held constant for 18 min.

A standard solution of lysostaphin was prepared as follows. From a solution of a known concentration of lysostaphin (ca. 1 mg/mL) 40 μL was pipetted into an Eppendorf vial of 0.5 mL and 120 μL of SDS sample buffer was added. The closed vial was incubated for 30 min at 80°C and subsequently rapidly cooled using ice water. From this solution 80 μL was transferred to the sample vial (of 0.2 mL Eppendorf vial).

For the preparation of samples of cell extracts, the same procedure as that for the standard solution of lysostaphin was followed, except that after cooling in ice water the solution was centrifuged for 5 min at 3000 g to remove any possible traces of precipitated proteins.

Analysis of minerals and trace elements in the basic medium was carried out by ICP-AES (Inductively Coupled Plasma – Atomic Emission Spectrometry) using the Vista Axial ICP of Varian (Palo Alto, CA, U.S.A.). The growth medium sample was prepared for ICP by dry-ashing, dissolved in nitric acid and measured against standards of Ca, Mg, Na, K, P, Fe, Zn, Mn and Cu.