Bacterial strains, culture conditions, and plasmids
The used bacteria strains Pseudomonas aeruginosa PAO1 and Pseudomonas putida KT2440 [31, 58] were routinely cultivated in LB-medium (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl), while Escherichia coli DH5α , Bacillus subtilis TEB1030 , and Corynebacterium glutamicum ATCC 13032  were cultivated in TB-medium (12 g/L tryptone, 24 g/L yeast extract, 5 g/L glycerol, 12,54 g/L dipotassium phosphate, 2,31 g/L monopotassium phosphate). All bacteria were cultivated at 37°C except P. putida and C. glutamicum, which were grown at 30°C. P. putida and E. coli containing the vector pVLT33  and derivatives thereof were selected by adding 50 μg/mL kanamycin to LB-agar and liquid cultures. For selecting pVLT31 and derivates tetracycline with concentrations of 10 μg/mL for recombinant E. coli and 20 μg/mL for recombinant P. putida were added. Rhamnolipid production with P. aeruginosa and recombinant P. putida was carried out using LB-medium complemented with 10 g/L glucose.
Rhamnolipid toxicity determination
The experiments were carried out with a micro bioreactor system (BioLector, m2p-labs GmbH, Baesweiler, Germany), in 48 well plates (Flowerplates m2p-labs GmbH, Baesweiler, Germany). The biomass concentration was quantified by online light scattering. All bacteria apart from P. putida KT2440 were cultivated in 800 to 1000 μL TB-medium. P. putida KT2440 was grown in 500 μL LB-Medium supplemented with 10 g/L glucose and 90 mM potassium phosphate buffer (pH 7.4). The growth medium contained rhamnolipid concentrations between 0 g/L and 90 g/L. The cultures were shaken at 1,100 to 1,200 rpm (throw: 3 mm) and incubated at 37°C or 30°C, for E. coli DH5α, B. subtilis TEB1030, C. glutamicum ATCC13032, and P. putida KT2440, respectively.
Construction of the rhamnolipid production module
The rhlAB operon was amplified from the genomic DNA of P. aeruginosa PAO1 that was isolated with a DNA isolation Kit (DNeasy Blood and Tissue Kit, QIAGEN, Hilden, Germany), using DNA polymerase (PfuTurbo, Stratagene, Waldbronn, Germany) as described by the supplier. The used primer had the following sequences: sense 5'TTGAATTCCATCGGCTACGCGTGAACACGG'3, antisense 5'TTTTTCTAGATCAGGACGCAGCCTTCAGCC'3. The oligonucleotides were obtained from Eurofins MWG Operon (Ebersberg, Germany). The rhlAB PCR product was digested with EcoRI/Acc65I and subsequently ligated into pVLT33, which was digested with the same enzymes, creating the plasmid pVLT33_rhlAB. Restriction enzymes and T4 DNA ligase were obtained from Fermentas (St. Leon-Rot, Germany) and used as recommended. DNA manipulation was carried out as described in Sambrook and Russell . Ligations were transformed into competent E. coli DH5α using a standard protocol . Transformed cells were selected on LB-agar plates containing 50 μg/mL kanamycin. Experiments with the kanamycin resistant single gene deletion strain P. putida KT42C1, lacking the poly(3-hydroxyalkanoic acid) synthase 1 encoded by phaC1, required subcloning of rhlAB into pVLT31, which contains a gene for tetracycline resistance.
P. putida KT2440 was transformed using electroporation as described by Choi et al. . Cells containing plasmid pVLT33 or the derivate pVLT33_rhlAB were selected on LB-Agar plates or liquid cultures containing 50 μg/mL kanamycin. A resulting strain was denominated P. putida KT2440 pVLT33_rhlAB. This strain was utilized for all further works.
Characterization of rhamnolipid production by P. putida KT2440 pVLT33_rhlAB
For the production of rhamnolipids a main culture of 50 mL LB-medium supplemented with 10 g/L glucose and 50 μg/mL kanamycin in a 500 mL Erlenmeyer flask was inoculated with 1 ml from a starter culture and incubated at 30°C and 200 rpm (throw: 25 mm). The expression of rhl-genes was induced by adding IPTG (isopropyl β-D-1-thiogalactopyranoside) to a final concentration of 0.4 mM (guaranteeing full induction) from the beginning of the fermentation. Rhamnolipids were extracted 24 h after induction.
P. aeruginosa was cultivated in 10 mL phosphate-limited protease peptone-glucose-ammonium salt medium (PPGAS) at pH 7.2, which promotes the production of rhamnolipids , containing 5 g/L glucose, 10 g/L peptone 0.02 M NH4Cl, 0.02 M KCl, 0.12 M Tris-HCl, and 0.0016 M MgSO4. After 24 h at 37°C, with an agitation of 150 rpm, rhamnolipids were harvested.
Cultivations of P. putida KT2440 pVLT33_rhlAB carried out in order to supply rhamnolipid-characterization via thin layer chromatography (TLC) and HPLC-ESI-MS featured slightly different process parameters. Only 10 ml of LB-Medium, supplemented with 10 g/L glucose and 50 μg/mL kanamycin in a 100 mL Erlenmeyer flask, were inoculated with an OD580 of 0.05 from a starter culture and incubated at 30°C and 150 rpm (throw: 25 mm). IPTG was added to a final concentration of 0.4 mM at an OD580 of 0.5.
To gain further insight into the fermentation kinetics of rhamnolipid producing P. putida experiments in 300 mL bioreactors (RALF, Bioengineering AG, Wald, Switzerland) were carried out. During 22 hours, the off-gas was analyzed with an off-gas sensor (BlueSens GmbH, Herten, Germany) applying dual wavelength infrared light. This sensor facilitated the simultaneous quantification of 12CO2 and 13CO2 concentrations in the off-gas. The reactor was filled with 200 mL LB-medium complemented with 10 g/L 100% labeled 13C6-glucose, IPTG to a final concentration of 0.4 mM, and 20 μg/mL tetracycline. The temperature was adjusted to 30°C and the aeration rate to 12.1 NL/h. The stirrer speed was set to 500 rpm and after seven hours of fermentation increased to 750 rpm.
The scale-up of rhamnolipid production was tested in a 3.2-liter fermenter vessel (KLF 2000, Bioengineering AG, Wald, Switzerland) with a working volume of 2 liters. The fermenter contained two 6-blade turbine stirrers, a temperature control, a pH control, and a gas inlet. The operating conditions were set to pH 6.8 and a temperature of 30°C, a constant gassing rate of 0.5 vvm and a stirrer speed in the range from 300 to 900 rpm depending on the online-determined pO2 signal. Additional glucose was fed using a peristaltic pump.
Quantification of rhamnolipids
For analysis, rhamnolipids were extracted using 100 μL for orcinol-assay and 500 μL for TLC respectively of cell-free culture broth and 500 μL of ethyl acetate. Samples were mixed by vortexing, with a subsequent phase separation by centrifugation in a tabletop centrifuge at maximum speed (30 sec). The upper, rhamnolipid-containing phase was transferred to a new reaction tube. This procedure was repeated three times. Finally, the organic solvent was removed by evaporation in a vacuum centrifuge.
Thin layer chromatography of rhamnolipids
For detection of rhamnolipids using TLC, the dried rhamnolipids were dissolved in 10 μL ethanol. 5 μL of this solution were spotted on a silica 60 TLC-plate (Macherey-Nagel, Düren, Germany). In addition, 5 μL of a 0.1% commercial rhamnolipid extract (JBR425, Jeneil Biosurfactant Co., LCC, Saukville, USA) containing mono- and di-rhamnolipids was spotted. The running buffer was a mixture of chloroform, methanol, and acetic acid in a ratio of 65:15:2. To visualize the rhamnolipids on the TLC-plates, the plates were covered with a detection agent consisting of 0.15 g orcinol, 8.2 mL sulfuric acid (60%), and 42 mL deionized water. For preservation, dried plates were incubated at 110°C for 10 min.
Rhamnolipid quantification using orcinol assay
The total amount of rhamnolipids was determined using the orcinol assay [65, 66]. The evaporated rhamnolipids were dissolved in 100 μL deionized water. Subsequently 100 μL orcinol solution (1.6% orcinol in deionized water) and 800 μL sulphuric acid (60%) were added. The samples were incubated at 80°C for 30 min and 1000 rpm orbital shaking in a thermomixer (Eppendorf AG, Hamburg, Germany). After cooling to room temperature, the samples were measured at 421 nm in comparison to different concentrations of the commercial rhamnolipid extract using a Genesys 10 UV spectrophotometer (Thermo Fisher Scientific, Waltham, USA).
Rhamnolipid quantification using RP-HPLC-CAD
Reversed phase high performance liquid chromatography corona charged aerosol detection (RP-HPLC-CAD) was used for rhamnolipid quantification. Culture samples were centrifuged at 17,700 × g for 30 minutes. 100 μL supernatant were added to 900 μL deionized water, mixed on a vortex shaker, and analyzed on a gradient quaternary reversed phase HPLC system (LaChrom, VWR- Hitachi, Darmstadt, Germany). The system was equipped with an integrated C8(2) silica based column (Luna C8(2), 4.6 × 150 mm, 5 μ, 100 Å, Phenomenex, Inc. Torrance, CA, USA) and a corona CAD (ESA Biosciences Inc., MA, USA). The sample volume was set to 20 μL. The sample was eluted at a flow rate of 800 μL per minute and the temperature of the column oven was set to 40°C. The mobile phase contained filtered water with 0.4% trifluoracetic acid (TFA) (solvent A), acetonitrile (solvent B), and methanol with 0.2% TFA (solvent C) in different ratios. The method started with 20:0:80 (Vol.-% of solvent A:Vol.-% of solvent B:Vol.-% of solvent C) and switched at 7.5 minutes to 2:18:80 during 2 minutes. After 20.6 minutes, the starting concentration was reestablished, again during 2 minutes. The method ended after 24.6 minutes.
Rhamnolipid composition characterization by HPLC-ESI-MS
High performance liquid chromatography electrospray ionization mass spectrometry (HPLC-ESI-MS) was used for rhamnolipid characterization (Central Division of Analytical Chemistry/BioSpec, Forschungszentrum Jülich, Jülich, Germany). Rhamnolipids were extracted from 1 L culture broth (5 L Erlenmeyer flask) as described by Déziel et al.  with small modifications. Cells were removed by centrifugation for 30 min at 9,000 × g and 10°C. The supernatant was acidified with 37% HCL to a pH of 3 and incubated overnight at 4°C. The precipitated rhamnolipids were recovered by centrifugation (9,000 × g, 45 min, 4°C) and resuspended in 15 mL acidified water (pH 3). This suspension was extracted three times with 15 mL ethyl acetate. The combined organic phases were evaporated in a vacuum centrifuge. The residue was dissolved in 15 mL of 0.05 M NaHCO3, acidified to pH 2 with 37% HCl, and incubated overnight at 4°C. The precipitate was finally recovered by centrifugation for 60 min at 13,000 × g and 4°C.
For characterization, a binary HPLC system (Agilent 1100 series, Agilent Technologies, Waldbronn, Germany), assembled with a diode array detector (DAD) (190-400 nm), coupled with a triple quadrupole mass spectrometer (4000QTRAP™, Applied Biosystem/MDS SCIEX, Foster City, CA, USA) assembled with a turbo ion spray source was used.
For rhamnolipid separation, normal phase chromatography was used with column dimensions of 150 × 2 mm i.d., 3 μm particle size (ProntoSIL 120-C8-SH, Bischoff Chromatography, Leonberg, Germany) at 20°C. The gradient elution was done with deionized water with 0.1% formic acid (solvent A) followed by different concentrations of acetonitrile with 0.1% formic acid (solvent B). The elution started with 60% B isocratic for 4 min, from 4 to 24 min a linear increase from 60% B to 90% B was applied, subsequently followed by a second isocratic step (90% B for 10 min), and ended by a return to 60% B in one min. The re-equilibration was done with 60% B isocratic for 10 min. All steps were performed at a constant flow rate of 300 μL/min. The injection volume was 20 μL.
The MS was used in negative enhanced mass spectrum mode scanning from 200 - 1000 Da. A flow injection analysis with a standard was used at first to optimize the following parameters: IS -4500 V, declustering potential -100 V, curtain gas (N2) 10 arbitrary units (au), source temperature 500°C, nebulizer gas (N2) 50 au, and heater gas (N2) 20 au. Collision energy (CE) and third quadrupole-entry barrier were set to -5 V and 8 V, respectively. The negative enhanced product ion scan mode was used for structural elucidation MS/MS experiments, in which product ions are generated in the second quadrupole by collision-activated dissociation of selected precursor ions of the first quadrupole and mass analyzed in a linear ion trap. The CE ranged from 30 to 70 V.
The di-rhamnolipid standard (Rha-Rha-C10-C10) for HPLC analysis was a gift from Sanofi-Aventis Deutschland GmbH, former Hoechst AG (Frankfurt, Germany). Mono-rhamnolipid standard (Rha-C10-C10) was prepared as described before .
Rhamnolipid purification by adsorption
The medium was centrifuged in 200 mL cups for 60 min at 4,000 rpm (5810R Eppendorf AG, Hamburg, Germany) to remove cells and cell debris. The cell-free medium was loaded with five times the bed volume per hour by a peristaltic pump (MP-3 Micro Tube Pump, Eyela Inc., Tokyo, Japan) as specified by the manufacturer to a column packed with 90 g of conditioned hydrophobic polymeric adsorbent (Amberlite XAD-2, Sigma-Aldrich, St. Louis, MO, USA). After washing with bidistilled water, rhamnolipids were eluted with 99% isopropanol using a continuous flow (HPLC pump 114 M, Beckman Coulter, Inc., Brea, CA, USA). The organic solvent was evaporated in a freeze dryer (Alpha I-5, Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, Germany).
Theoretical capacity estimation
The flux balance analysis was carried out using the software Insilico Discovery (version 3.2.0, Insilico Biotechnology AG, Stuttgart, Germany). The provided metabolic network used for simulations was modified to represent the reaction network of P. putida (Additional file 1).
The following reactions were added to the P. putida
A linear optimization for rhamnolipid-production with simultaneous minimization of total fluxes was carried out. The rhamnolipid production rate was simulated with different carbon substrates (glucose, glycerol, sucrose, and octanoate). To ensure comparability of the results, the unit Cmol, which normalizes the rhamnolipid production rate to the amount of carbon atoms present in the carbon substrate was chosen. The substrate uptake was varied between 0 and 120 mCmol/(gCDW h). The maintenance metabolism, characterized through the simple reaction of ATP to ADP, was varied in the range of 0 to 50 mmol/(gCDW h). Blank et al.  described a value for the non-growth associated maintenance of 10.2 mmol ATP/(gCDW h) for P. putida DOT-T1E. The considerably higher upper limit of 50 mmol ATP/(gCDW h) accounts for scenarios of extra stress, e.g., for metabolic cost of handling high rhamnolipid concentrations. The chosen values for the growth rates were 0 1/h, 0.4 1/h and 0.8 1/h, reflecting ideal production condition, growth observed during rhamnolipid production, and maximal growth of P. putida on glucose . Additionally all occurring fluxes were limited to a maximal value of 120 mCmol/(gCDW h). Furthermore, variation of the fluxes through the pathways ED pathway, TCA cycle, and PP pathway were examined. In addition, an alternative glucose uptake system, the phosphotransferase system, and a complemented EMP pathway (insertion of a phosphofructokinase reaction for example encoded on a fructose utilization operon by fruK (PP0794), catalyzing the conversion of glucose-6P to glucose-1,6P) were simulated.
Determination of fermentation kinetics
The growth kinetic was described mathematically using a logistic growth model. Logistic growth of pseudomonads had been previously reported for rhamnolipid producing wild type P. aeruginosa growing on sunflower oil .
The biomass concentration X
was described using equation 1, where X
is the initial biomass concentration, X
the additional biomass concentration, t
the time after which half of X
is formed, and b
is a curve form coefficient.
The experimental data for the rhamnolipid and glucose concentrations could be described with equations 2 and 3, where r
is the specific rhamnolipid production rate [grhamnolipid
h)] and r
is the specific glucose uptake rate [gglucose
A multivariable least squares fit was used to illustrate the development of all three fermentation parameters depending on each other.
Prior to utilizing the described procedure, to fit glucose and rhamnolipid concentrations, two more attempts applying different models were carried out (Additional file 2).