A system that changes color upon contact with FA in a gaseous phase was constructed, in order to design a simple low-cost semi-quantitative biosensor for determination of airborne FA. The system was based on a membrane that separated the aqueous phase, which contains a chromogen and the enzymes AOX and HRP, and the FA-containing gaseous phase. FA diffused through the membrane to the aqueous phase and participated in the following chain process, which is catalyzed by the enzymes [34]:

The rate of color appearance in the second reaction actually depended on the FA concentration in the aqueous phase, which in turn depended on the FA concentration in the gaseous phase. Several experiments were carried out using a closed vessel that contained a layer of the FA aqueous solution at a known concentration, in order to study this effect. Henry’s law enables the calculation of the airborne FA concentration [35], which was also proven by measurement with the FA gas detector. Membrane devices containing various concentrations of AOX, HRP and a chromogen were placed above the aqueous layer, in contact solely with the gaseous phase. Reactions (1) and (2) took place upon diffusion of FA through the membrane into the aqueous phase, and the time of the clearly visually detectable appearance of color was registered. The membrane devices were then removed from the vessel, opened, and the color intensity was measured spectrophotometrically at the appropriate wavelength [36]. The color intensity was found to correspond to an absorbance A (450 nm) range from 1 to 1.2 in all cases. The clearest results were obtained when HRP was used at a concentration of 3.5 mg·mL⁻¹ (1000 U·mL⁻¹) and the chromogen—at a concentration of 2 mg·mL⁻¹. According to the Material Safety Data Sheet (MSDS), the threshold level value (TLV) of airborne FA is 0.3 ppm [37]. Three FA rates, 0.1, 0.3 and 18.5 ppm, were therefore chosen as fixed points for evaluating the reaction time. The first point corresponds to the safe FA level, the second to the maximum permissible level and the third is considered to be a level dangerous for humans. An optimal biosensor would react immediately to the dangerous FA level, would react moderately to the permissible level and would not react at all to the safe level. The effect of AOX activity on the time of color development was studied in order to achieve this aim (Figure 5). The reaction time was found to be inversely dependent on the AOX concentration, and once the color developed it remained stable for at least half an hour (Figure 5).

In the next stage, the effect of the airborne FA concentration on the time of color development was studied. For this aim, the AOX concentration in the membrane devices was kept constant, and the devices were exposed to various concentrations of airborne FA (Figure 6).

As shown in Figure 6, the time of color development depended on the FA concentration in the gaseous phase, and the higher the FA concentration, the shorter was the reaction time. This fact can easily be explained by the dependence of the FA diffusion rate through the membrane from the gaseous into the aqueous phase on the airborne FA concentration.

Experiments with varying FA and AOX concentrations were performed in order to determine and choose conditions for a selective reaction of the membrane device to safe and dangerous FA concentrations in the gaseous phase. The results of this series are presented in Figure 7. As shown in Figure 7, there was almost no difference between the color development time under exposure to 0.1 and 18.5 ppm at high AOX concentrations (5 to 10 U·mL⁻¹)—the color appeared already after 12–20 min in the first case and after several minutes in the second. With the medium AOX concentration (2.5 U·mL⁻¹), the reaction times to high and low FA levels were clearly distinguished—the device changed color within 3–5 min when exposed to 18.5 ppm and within 40 min when exposed to 0.3 ppm. However, the reaction to 0.1 ppm took place after 60 min so that the difference between the safe and the threshold levels was not sufficiently clear. The results for the lowest applied concentration of AOX (1.25 U·mL⁻¹) were very satisfactory. The device reacted very quickly to the dangerous 18.5 ppm concentration of FA, and very slowly (1 h) to the TLV concentration, but did not react at all to the safe 0.1 ppm concentration (the color did not develop during 24 h of exposure). This result enabled unambiguous differentiation between safe, threshold and dangerous concentrations of airborne FA. It also enabled reaching the conclusion that the proposed membrane device could be applied as a semi-quantitative biosensor for detection of FA in indoor air. The simplicity of the device opens prospects for its wide application in workstations related to the use of FA, from workshops to laboratories. The device showed good reproducibility and reliability of the results. In the experiments in which light absorption was measured at the chosen time periods, the relative standard errors (δ) did not exceed 3%. In the experiments in which the time of visual color development was measured, the relative standard errors varied from 20% for short periods (several minutes) to 2% for long periods (more than an hour).

The proposed biosensor for semi-quantitative determination of airborne FA has several advantages over existing instrumental methods of detection, such as chromatography or selective gas analyzers [37]. The biosensor does not require any instrumental support, but at the same time supplies an almost immediate reaction to hazardous levels of FA, when the developed visual signal can be clearly observed. The only disadvantage of this biosensor is its irreversibility, because once developed, the color does not disappear. However, this issue is compensated by the low cost and simplicity of the device, which can be produced wholesale and can be stored for a long time, since AOX and HRP are known for their high stability [38,39].

DEAE-Toyopearl 650M was purchased from Toyo Soda (Tokyo, Japan); phenylmethylsulfonyl fluoride (PMSF), EDTA and Triton X-100 were obtained from Sigma (Deisenhofen, Germany). Ethanol absolute was purchased from Riedel-de Haën (Seelze, Germany); hexachloroplatinum(IV)-acid-hexahydrate and methanol were obtained from Merck-Schuchardt (Hohenbrunn, Germany); HRP with a specific activity of 250–330 units per 1 mg solid (RZ 3.0) and o-dianisidine were purchased from Sigma Aldrich Chemie (Steinheim, Germany).

All chemicals were of analytical reagent grade and all solutions were prepared using high performance liquid chromatography (HPLC)-grade water. The FA solution (1 M) was prepared by hydrolysis of the corresponding amount of para-formaldehyde in water (300 mg; 10 mL water) under heating the suspension in a sealed ampoule at 105°C for 6 h.

A mutant strain of the thermotolerant methylotrophic yeast H. polymorpha C-105 (gcr1 catX) selected in the Institute of Cell Biology (Lviv, Ukraine) [40] was used as a producer of AOX. This strain has impairment in glucose catabolite repression of AOX synthesis, is catalase-defective, and has the ability to over-produce AOX in glucose medium. Mutant H. polymorpha C-105 cells were cultivated in flasks on a shaker (200 rpm) at 30°C up to the middle of the exponential growth phase (~24 h) in medium containing (g/L) the following: glucose, 10; (NH₄)₂SO₄, 3.5; KH₂PO₄, 1.0; MgSO₄·7H₂O, 0.5; CaCl₂, 0.1; yeast extract, 3.0. The pH of the medium was adjusted to 5.5 with KOH. After washing, freshly grown cells (about 15 g wet weight) were re-suspended in two volumes of 0.05 M phosphate buffer, pH 7.5. The cells were disrupted in the presence of glass beads (d = 0.45–0.50 mm) in a planetary disintegrator at 1000 rpm (r_av 10 cm) and 4°C for 6 min. Whole cells and cell debris were removed by centrifugation at 15,000 g (r_av 8 cm) for 40 min. The supernatant cell-free extract was used for isolation of AOX.

Highly purified AOX was isolated by means of a two-step ammonium sulfate fractionation (at 40% and 60% of saturation), followed by ion exchange chromatography on DEAE-Toyopearl 650M [41]. The first stages of the purification process were carried out in the presence of protease inhibitors (1 mM EDTA and 0.4 mM phenylmethylsulfonyl fluoride, PMSF). The final preparation of the enzyme had a specific activity of 30 U per mg protein and was homogeneous during electrophoresis in a polyacrylamide gel in the presence of SDS. AOX preparations were stored as a suspension in 70% saturated ammonium sulfate (with an activity 200 U·mL⁻¹ of suspension) at −10°C without any remarkable decrease in activity over a period of several months. Before use, the enzyme suspension was centrifuged and the enzyme precipitate was dissolved in 0.1 M phosphate buffer, pH 7.6.

AOX activity was determined by the rate of hydrogen peroxide formation in reaction with methanol as monitored by the peroxidative oxidation of o-dianisidine in the presence of HRP [42]. The millimolar extinction coefficient of the colored product in an acidic solution (2.5 M HCl) at 525 nm was 13.38 mM^–1·cm^–1. Protein concentration was determined by the Lowry method.

The chemosensors were evaluated by constant-potential amperometry in a three-electrode configuration with an Ag/AgCl reference electrode, a Pt-wire counter electrode, and working electrodes—the 3.05 mm diameter carbon rod (type RW001, Ringsdorff Werke, Bonn, Germany) sealed in a glass tube using epoxy glue, the 4 mm planar carbon electrode DRP-150, and the planar gold electrode DRP-C220AT from DropSens [Llanera (Asturias), Spain].

Platinized electrodes were prepared by electrochemical Pt deposition from a 2% solution of H₂PtCl₆ according to [28]. Calibration was performed by stepwise addition of a standard FA solution.

The membrane biosensor device for detection of airborne FA was prepared based on dialysis tubes (Sigma-Aldrich) containing 1 mL of a solution containing 3.5 mg·mL⁻¹ of HRP with a specific activity of 250–330 units (RZ 3.0) per 1 mg solid, 2 mg·mL⁻¹ o-dianisidine as a chromogen (both purchased from Sigma Aldrich Chemie (Steinheim, Germany)), and 1.2–10 U·mL⁻¹ AOX in 0.05 M sodium phosphate buffer, pH 7.5. The device was exposed to various concentrations of airborne FA—from 0.1 to 18.5 ppm, obtained by saturating the gaseous phase with FA vapors which were in equilibrium with aqueous FA solutions of various concentrations (0.54–97.5 mM). The airborne FA concentration was calculated based on Henry’s law, where the Henry constants were calculated by the following equation [35]:

and also tested in the range of 0.1–0.6 ppm by the FA Gas Detector (Model FP-40 Riken Keiki, Japan). The FA solutions were placed in closed prism-form glass vessels supplied with stoppered flat covers, which enabled quick opening and closing of the vessel, in order to achieve FA equilibrium between the gaseous and aqueous phases. The above-described membrane devices were placed in the vessels on racks, so that the device was in contact solely with the gaseous phase. The time of the color development was registered, and immediately thereafter the devices were opened and the color intensity was measured using the UV-Vis Cary-50 (Varian, Australia) spectrophotometer.

The results obtained from at least three independent experiments fulfilled with duplicates were statistically analyzed by Anova single factor analyses or by Students’ t-test. The difference between the results was considered significant if the p-value was less than 0.05. Relative standard errors were calculated on the basis of the Descriptive Statistics data.