Microorganism and medium
Clostridium beijerinckii NCIMB 8052 was used in this study, since it is known to utilize different lignocellulosic sugars for growth and butanol production (17). The inoculum was pre-grown for 14 hours on reinforced Clostridial medium (CM0149, Oxoid) in an incubator at 37 °C under anaerobic and static conditions. The fermentation medium composition was 5 g/l sugar, 2.5 g/l Na-acetate, 5 g/l yeast extract, 2 g/l (NH4)2SO4, 0.01 g/l NaCl, 0.75 g/l KH2PO4, 1.5 g/l K2HPO4, 0.2 g/l MgSO4.7H2O, 0.01 g/l MnSO4.H2O, 0.01 g/l FeSO4.7H2O, 0.01 g/l P-aminobenzoic acid, 0.01 g/l biotin and 0.1 g/l thiamine. Six different mixed sugar solutions of glucose and xylose were prepared at 6 different ratios with a total sugar concentration of 5 and 10 g/l. Consequently, in total 12 different sugar compositions using 2 different total sugar concentrations and 6 different glucose to xylose ratios were studied in parallel. The cultures containing 5 g/l total sugar, and glucose (G) to xylose (X) ratios of 100:0, 80:20, 60:40, 40:60, 20:80 and 0:100 are referred to as 5-G100:X0, 5-G80:X20, 5-G60:X40, 5-G40:X60, 5-G20:X80 and 5-G0:X100, respectively throughout the study. Similarly, the cultures containing 10 g/l total sugar and glucose to xylose ratios of 100:0, 80:20, 60:40, 40:60, 20:80 and 0:100 are referred to as 10-G100:X0, 10-G80:X20, 10-G60:X40, 10-G40:X60, 10-G20:X80 and 10-G0:X100, respectively.
Fermentations were performed in in batch mode in round-well plates with 48 x 3 ml microbioreactor wells with 1.5 ml working volume. Anaerobic conditions were sustained by flushing with nitrogen gas (Aga grade 6.0) at 37 °C and shaking at 400 rpm in a BioLector® instrument (m2p-labs GmbH, Baesweiler, Germany). Figure 5 shows the BioLector® setup together with a round-well plate and a single well with embedded sensors.
Cultures in the wells were started by adding 4% (v/v) inoculum prepared as described above. The BioLector® measures cell mass density by scattered light, applied in the present study in 20 minutes intervals. A gain of 20 (EX: 620 nm, EM:620 nm) was used for the experiments to avoid saturation at high cell mass. pH was measured every 20 minutes with a gain of 19 (EX: 470 nm, EM: 525 nm).
To benchmark the microbioreactor fermentations, fermentations in 120 ml serum flasks were performed with 50 ml working volume in an incubator at 37 oC under static and anaerobic conditions, with 12 different sugar compositions as explained above. A representation of the experimental design of fermentations performed both in microbioreactors and serum flasks is shown in Figure 6.
Both the microbioreactor and the serum flask experiments were performed using the same batch of medium and inoculum to minimize errors due to medium preparation and inoculation. Inoculum size was 4% (v/v) in all cases. There was no pH control applied. Experiments were terminated after 79 hours. Samples were taken at the start and the end of the fermentations for analysis of medium components and products, cell mass and pH. Data shown represent the mean values from experiments performed in quadruples, and error bars represent the standard deviations in microbioreactor results.
After cultivation, fermentation samples from both microbioreactors and serum flasks were used to determine residual sugars and fermentation products by high-performance liquid chromatography (HPLC). The samples were filtrated (Millipore filter, 0.2 µm) before HPLC analysis on an Agilent System LC1260 equipped with UV (210 nm) and RI detector and an Aminex HPX-87H column (BioRad). Samples were eluted with 5 mM H2SO4 at a flow rate of 0.6 ml/min at 45 °C. Quantification was performed using standards for each component.
Estimation of kinetic coefficients
Cell mass growth rates were estimated during exponential growth phase in which nutrients are non-limiting and thus the growth rate is independent of the nutrient concentration. Therefore, the rate of growth is
where X is the cell mass concentration (g/l), t is time (h), and µ is the specific growth rate (h-1). The specific growth rate is determined by estimating the slope of the cell mass concentration versus time plot.
The product yield based on substrate consumption is a commonly used kinetic coefficient that indicates how efficient the conversion of substrate to product of interest is (18). The product yield (g product/g sugar) is
where P is the product concentration, butanol (g/l), and S is the total substrate concentration, glucose and xylose (g/l).
Analysis of variance (ANOVA) is a procedure to assign sample variance to different sources and to decide whether the variation arises within or among different population groups. A two-way, nonparametric ANOVA method, Friedman's test is used in this study (19). The p value that the Friedman's test returns is used to determine significance. If the p value is near zero, this casts doubt on the null hypothesis. A sufficiently small p value suggests that at least one column-sample median is significantly different from the others. It is common to declare a result significant if the p value is less than 0.05 or 0.01, and we chose 0.05 in this study. Matlab function 'friedman' is used with replicate number of 4 for fermentations done in BioLector® and 1 for fermentations done in serum flasks. Two different effects are considered: total sugar concentration values of 5 and 10 g/l, and glucose to xylose ratios of 100:0, 80:20, 60:40, 40:60, 20:80.