Microorganism and growth medium
Pseudomonas putida KT2440 (ATCC 47054) was maintained on nutrient agar plates at 4°C. The inoculum medium for all fermentations contained per liter: (NH4)2SO4 4.70 g, MgSO4 · 7 H2O 0.80 g, Na2HPO4 · 7 H2O 12.00 g, KH2PO4 2.70 g, nutrient broth 1.00 g, glucose 9.00 g. The initial culture medium contained per liter: (NH4)2SO4 4.70 g, MgSO4 · 7H2O 0.80 g, Na2HPO4 · 7H2O 18.0 g, KH2PO4 4.05 g, trace element solution 10 mL. The trace element solution contained per liter: FeSO4 · 7H2O 10.0 g, CaCl2 · 2H2O 3.0 g, ZnSO4 · 7H2O 2.2 g, MnSO4 · 4H2O 0.5 g, H3BO3 0.3 g, CoCl2 · 6H2O 0.2 g, Na2MoO4 · 2H2O 0.15 g, NiCl2 · 6H2O 0.02 g and CuSO4 · 5H2O 1.00 g. Nonanoic acid (98%, Spectrum Chemicals) was fed separately in its pure form as it is immiscible in aqueous media. Acrylic acid (Sigma-Aldrich) was added to a glucose (99.5%, Sigma-Aldrich) solution of 240 g L-1. Feeding ratios of nonanoic acid (NA), glucose (G) and acrylic acid (AA) at 1.25: 1: 0.01 and 1.25: 1: 0.05 (w/w) were tested. Nitrogen was provided as 14% (w/v) ammonia solution and also served as the base for pH control. In case of nutrient depletion, supplemental solutions of trace elements with the above composition and a phosphate solution containing 36 g L-1 Na2HPO4 · 7H2O and 8.1 g L-1 KH2PO4 were prepared. Antifoam 204 (Sigma-Aldrich) was added to nonanoic acid (1% v/v) and manually injected through a sterile septum when required.
The inoculum was grown in three 500 mL shake flasks (100 mL medium in each flask) at 28.0 ± 1°C and 200 rpm overnight. The first two fermentations were conducted in a 7 L MBR stirred tank bioreactor (Bioreactor-AG, Switzerland) with a 5 L working volume. The third fermentation was done in a 5 L Minifors bioreactor (Infors-HT, Bottmingen, Switzerland) with a 3 L working volume. The cultivation temperature was 28.5 ± 1°C and the pH was controlled at 6.85 ± 0.05 using 14% (w/v) ammonia solution. Dissolved oxygen was measured with an Ingold polarographic probe and maintained above 30% air saturation by adjusting the agitation speed and the mixture of air and oxygen flow via mass flow controllers to a total gas flow at 1 vvm. The dissolved oxygen data were acquired by a LabVIEW 6.1 (National Instrument) program. Nonanoic acid and glucose feeding was controlled via separate peristaltic pumps by the LabVIEW program based on the mass of each reservoir.
Substrate feeding and control methods
The specific growth rate was controlled at 0.25 h-1 or 0.15 h-1 by exponentially feeding the carbon sources to be the growth-limiting nutrient. It was estimated that 1 g L-1 biomass would be produced from 1.6 g L-1 total carbon sources in the initial fermentation medium. The cumulative mass of carbon substrates St (g) to be fed at time t (h) was calculated based on exponential cell growth (Xt, g) expressed in the Equation below.
where X0 (g) is the estimated biomass at the beginning of the feeding; μ (h-1) is the desired specific growth rate; and Y
is the yield (g g-1) of biomass from the mixture of carbon substrates which was 0.66 g g-1, experimentally determined from continuous fermentation by feeding nonanoic acid, glucose and acrylic acid at a mass ratio of 1.25: 1: 0.05 at a specific growth rate of 0.25 h-1 (Jiang et al. 2012).
The mass of each carbon source required at time t was calculated according to the pre-defined ratio of the two substrates as follows:
The feeding ratio of nonanoic acid to glucose in this study was 1.25: 1 (w/w). Therefore, the mass fraction of nonanoic acid (f
NA) and that of glucose (f
G) in the total carbon source were 0.56 and 0.44, respectively.
Exponential substrate feeding began after a lag phase of approximately 5 h. Fermentations with a specific growth rate of 0.25 h-1 were conducted only under exponential feeding. However, in an effort to avoid nonanoic acid and acrylic acid overfeeding, exponential feeding at 0.15 h-1 was conducted for 23.3 h before changing to a constant feed rate of 8 g L-1 h-1.
Biomass concentration was determined gravimetrically from duplicate samples of 10 mL culture broth which were centrifuged at 6,000 × g for 15 min, washed and lyophilized. Sample supernatants were analyzed for the concentrations of residual nutrients and acrylic acid. Glucose was measured colorimetrically after reacting with 4-hydroxybenzoic hydrazide under alkaline condition (Lever 1972). Nonanoic acid was methylated in acidified methanol (Ramsay et al. 1991) and analyzed by a CP3900 Varian GC equipped with a flame ionization detector. Phosphate was measured based on the reduction of phosphomolybdate to molybdene blue (Clesceri et al. 1999). Ammonium was determined by the phenol-hypochlorite method (Weatherburn 1967). Acrylic acid was assayed by Hewlett-Packard GC equipped with a Cabowax®-PEG column after acidification with one tenth volume of 2 N hydrochloric acid (Qi et al. 1998).
PHA content and composition in the dry biomass samples were determined by methanolysis in 2 mL chloroform and 1 mL methanol which contained sulfuric acid (15% v/v) as acidifying agent and benzoic acid (0.2% w/v) as internal standard at 100°C for 4 h. After which, 1 mL distilled water was vigorously mixed on a Fisher Vortex and left overnight for phase separation. One μ L of the chloroform phase was injected into CP3900 Varian GC at a split ratio of 20. The injector and detector were maintained at 250 and 275°C, respectively. The oven heating profile was: initial 90°C for 0.5 min, 5°C min-1 to 95°C and hold for 0.5 min, 30°C min-1 to 170°C and hold for 2.5 min. The PHA standard was prepared by acetone extraction and methanol precipitation followed by three cycles of extraction and precipitation, as described by Jiang et al. (2006) and the monomeric composition characterized by GC and proton nuclear magnetic resonance at room temperature in a Bruker Avance 200 spectrometer using deuterated-chloroform containing 20 mg mL-1 PHA.