Pichia angusta strain Y-1397, Arxula adeninovorans strain Y-78(6), Gluconobacter oxydans strain B-1280, Gluconobacter oxydans strain B-1227, Bacillus subtilis strain VKM B-434 and Rhodococcus sp. strain R-20 were obtained from the All-Russian Collection of Microorganisms, G. K. Skryabin Institute of Biochemistry and Physiology of Microorganisms (IBPM), Russian Academy of Science (RAS). Escherichia coli strain K-802 was kindly provided by the Laboratory of the Structure-functional Analysis of Genetic Systems of Microorganisms, IBPM, RAS.

Strains of G. oxydans were grown on medium containing (g/l): sorbitol, 20; yeast extract, 2. Cells were grown for 18 h in 750-ml Erlenmeyer flasks containing 100 ml of growth medium on a shaker (200 rpm, 28С). P. angusta strain VKM Y-1397 was grown for 24 h in 750-ml Erlenmeyer flasks (medium volume 100 ml) at 28°С on a shaker (220 rpm) in a liquid medium of the following composition (g/l): (NH₄)₂SO₄, 2.5; MgSO₄, 0.2; K₂HPO₄, 0.7; NaH₂PO₄, 3.0; yeast extract, 0.5; 100 l of vitamin solution per 100 ml of the medium (thiamin and biotin – 5 mg in 10 ml of methanol); 1 ml of solution of trace elements per 100 ml of the medium (CaCl₂ × 2H₂O, 0.11; FeSO₄ × 7H₂O, 0.017; ZnSO₄ × 6H₂O, 0.009; MnSO₄ × 5H₂O, 0.00023; CuSO₄ × 5H₂O, 0.0045); glycerol (1%, v/v). A. adeninovorans strain Y-78(6) was grown at 37°С for 24 h on medium containing (g/l): glucose, 10; peptone, 5; yeast extract, 0.5. Rhodococcus sp. R-20, E. coli K-802, and B. subtilis VKM B-434 were grown on the following nutrient medium of IBPM (g/l): aminopeptide from animal blood hydrolysate, 60.0; tryptone, 50.0; fodder yeast extract, 10.0; soybean extract, 30.0 (medium developed in G. K. Skryabin Institute of Biochemistry and Physiology of Microorganisms). Cells were grown at 28°С for 18 h. Strain E. coli K-802 was grown at 37°С.

Cells were separated by centrifugation at 10,000 x g for 5 min and twice washed with potassium-phosphate buffer (30 mM, pH 7.5). For G. oxydans strains, potassium-phosphate buffer (30 mM, pH 6.6) was used. Cells were immobilized by physical sorption on glass fiber filters (GF/A, Whatman). For this purpose, 5 l of cell suspension containing biomass in the concentration of 100 mg of wet weight/ml was applied to the filter and dried at room temperature for 20 min.

For the fabrication of biosensors, a bioreceptor of 3 × 3 mm² was fixed to the measuring surface of a Clark oxygen-type electrode by a capron net and a fitting ring. Measurements were performed in an open cuvette, and the sensor signal was recorded by an IPC2L amperometric potentiostat connected to a desktop computer. An analyte sample (5-100 l) was introduced into a 2-ml cuvette, and measurements were performed under constant stirring. Biosensor responses were recorded as the maximal rate of change in signal (nA/s).

Table 1 shows the responses of microbial sensors to 1 mM glucose and 1 mM disaccharides. Glucose serves as a positive control, and as expected, all strains responded most strongly to this sugar. Most of the strains responded to maltose and cellobiose. Both G. oxydans strains responded to melibiose. None of the strains was highly sensitive to lactose. Only B. subtilis strain VKM B-434 responded strongly to sucrose. Its response to maltose was about half that of sucrose. In contrast, E. coli strain K-802 responded strongly only to maltose. The complementary nature of these responses suggests that these strains might be useful for selective biosensor detection of disaccharides. Based on these results, the latter two strains were chosen for further testing as a model microbial sensor for differential detection of sucrose and maltose.

Figure 1

Table 1. Microbial sensor responses to 1 mM glucose and 1 mM disaccharides.

Fig. 1 shows the calibration dependence of a microbial sensor based on E. coli strain K-802 for sucrose and maltose. The range of linear dependence of the signal is 0.05-5.0 mM for both sucrose and maltose, although clearly the response is much stronger for maltose.

Figure 2

Fig. 1. Calibration dependence of the strain K-802-based sensor for detection of maltose (A) and sucrose (B).

By contrast, Fig. 2 shows the calibration dependence of a microbial sensor based on B. subtilis strain VKM B-434 for sucrose and maltose. Although the kinetics are more complex in this case, practical response ranges were determined to be 0.05-5.0 mM for maltose and 0.005-0.5 mM for sucrose. Clearly, this sensor responds more strongly to sucrose.

Figure 3

Fig. 2. Calibration dependences of sensor based on strain VKM B-434 for detection of maltose (A) and sucrose (B).

The additivity of microbial sensor signals was assessed within the initial part of the determined response ranges. Multi-component analysis requires that the sensor signal corresponding to one substance changes upon the addition of the second substance to the sample. The simplest case is complete (linear) additivity, in which the sensor response to a sample containing two substances is equal to the sum of sensor responses to each component separately. Nonlinearity usually requires a thorough study of weighted coefficients or the introduction of approximating dependences. For example, in our previous study on the selective detection of ethanol, we used the classical model of multi-component analysis (cluster analysis) suggesting the linear additivity of sensor responses: , where S_K is the total response of sensor K, being a linear combination of individual responses S_K(i) to m of the substrates present in the mixture with coefficients a_i (Lobanov et al., 2001). The range of linear additivity makes it possible to consider coefficients a_i before individual responses S_k(i) to be concentration-independent, which significantly simplifies calibration.

For the current study, additivity was tested by measuring the responses to 0.025 mM sucrose, 0.25 mM maltose, and a mixture of these two disaccharides at the same final concentration. As shown in Table 2, the sensor signals were linearly additive for selected samples.

Figure 4

Table 2. Additivity of microbial sensor signals.

These findings suggest that a system based on the bacteria B. subtilis strain VKM B-434 and E. coli strain K-802 can be considered as a model for microbial sensors for the selective detection of a mixture containing the disaccharides maltose and sucrose. Such sensors could find practical application in the food processing industry, particularly in the enzymatic production of glucan oligosaccharides. Similarly, results suggest that microbial sensors based on strains of G. oxydans could be developed for the selective detection of cellobiose or melibiose. Cellobiose sensors could be useful in the biorefining of cellulosic biomass. The general screening method employed in this study could be extended to identify microbial strains able to oxidize additional substrates, including lactose and oligosaccharides.