Anal. Chem. 2007, 79, 6042-6044
Production of Biosensors with Exchangeable Enzyme-Containing Threads Toonika Rinken,*,† Jaak Ja 1 rv,‡ and Ago Rinken‡
Institute of Physical Chemistry and Institute of Organic and Bioorganic Chemistry, University of Tartu, Jakobi 2, Tartu 51014, Estonia
We introduce a simple method for the construction of biosensors, based on coiling an enzyme-containing, threadshaped material around a cylindrical signal transducer in the form of winding stairs with a variable length of step and so forming a variable biocatalytic membrane on the sensor surface, which can be easily modified for particular purposes. In the model system, we immobilized glucose oxidase (GO) on a nylon thread, formatted from a sheaf of numerous minor filaments and used as a biorecognition element integrated with a Clark-type oxygen sensor. The immobilized enzyme was evenly distributed throughout the thread, and the activity of the enzyme could be measured in units of length. Appropriate pieces of the enzyme-containing thread with a certain amount of GO could be cut for a definite biosensor or bioreactor. The enzyme amount and substrate diffusion parameters, which together control the sensor’s working range and sensitivity, could be changed simultaneously with the change of the length of the thread. Besides glucose oxidase, experiments with other enzymes have confirmed the applicability of the proposed technological solution. Thus, the thread-type matrixes enable one to construct sensors with a required range of work, sensitivity, and selectivity, which can be easily customized within seconds. Biosensors are analytical devices composed of a biological recognition element interfaced to a signal transducer, which together convert the concentration of an analyte to a measurable response.1 In a more focused approach, biosensors are classified as a subtype of chemical sensors that rely on biochemical recognition.2 In theory, virtually any biological recognition element can be in any mode interfaced to any of the signal transduction technologies, which in respect of certain applications show inherent advantages and limitations. Bioanalytical systems, based on a single signal transducing unit and a set of different or exchangeable biorecognizing threads would be a good option for the construction of miniature multipurpose equipment for quick on-site analyses. The insoluble supporting material for enzyme immobilization to form the biosensor bioselective element should be easy to produce and * Corresponding author. E-mail:
[email protected]. † Institute of Physical Chemistry. ‡ Institute of Organic and Bioorganic Chemistry. (1) Turner, A. P. Sens. Actuators 1989, 17, 433-50. (2) Janata, J.; Josowicz, M. Anal. Chem. 1998, 70, 179R-208R.
6042 Analytical Chemistry, Vol. 79, No. 15, August 1, 2007
comfortable to handle, mechanically and chemically stable, but also reasonably priced.3 At the same time, this material has to carry immobilized enzymes with sufficient operational activity, which does not change substantially in the course of analyses. Dozens of different materials in different forms and shapes, such as membranes, films, capsules, etc., but also fibers have been used for the immobilization of different bioactive compounds.4 In addition, bioactive compounds are often bound directly onto the surface of the signal-transducing device, for example, by immersing the electrode into a solution of monomers and a bioactive component, yielding an invariable electrode coated with bioactive material containing polymer.5 The main disadvantage of such biosensors is the inflexibility of the system and the need to create separate specific sensing systems for every particular measuring task. As the inactivation of enzyme molecules in time is inevitable, it is necessary to renew regularly the whole biosensor system, including the signal transducer as well, although this unit itself may be in good working condition and there is no real need for its substitution. Here we propose a novel method for biosensor construction, easily enabling one to vary the range of determination, sensitivity, and selectivity of the biosensor and turning the bioselective part of the biosensor into an exchangeable unit. EXPERIMENTAL SECTION Glucose biosensor was used as a model biosensing system with glucose oxidase (GO) as the biorecognition enzyme. In the role of a basic physicochemical transducer, we used an amperometric oxygen sensor, which had a cylindrical cathode covered with 25-µm-thick polyethylene film with the area of 5.65 cm2 (Elke Sensor). Glucose oxidase (EC 1.1.3.4, from Penicillium vitale, Lvov Biotechnology Venture) was immobilized onto a commercial nylon-6.6 thread by activating the thread with CaCl2 and 3.6 M HCl and coupling GO with glutaraldehyde as described earlier.6 The obtained bioactive thread was kept in 0.1 M acetate buffer solution (pH 5.60) at 4 °C until use, and it had no considerable loss of enzymatic activity in 6 months. Besides glucose oxidase, we also used amine oxidase (EC 1.4.3.6, from Pisum sativum, (3) Cao, L.; Langen, L. v.; Sheldon, R. A. Curr. Opin. Biotechnol. 2003, 14, 387-94. (4) Cao, L. Curr. Opin. Chem. Biol. 2005, 9, 217-26. (5) Yacynych Alexander. Method for making electrochemical sensors and biosensors having a polymer modified surface. U.S. patent 5,540,828, 1996. (6) Hornby, W. E.; Goldstein, L. Methods Enzymol. 1976, 44, 118-34. 10.1021/ac070327j CCC: $37.00
© 2007 American Chemical Society Published on Web 06/28/2007
Figure 2. View of the nylon thread in a microscope. The actual diameter of a nylon filament is 25 µm, and the whole thread consists of 60 coiled filaments. Figure 1. Biologically active thread wound spirally around the cylindrical cathode of an oxygen sensor, forming “winding stairs” with a variable length of step. b, Biologically active thread. O, Cylindrical surface of a signal transducer.
Sigma), immobilized onto nylon thread, as a bioselective compound for the determination of cadaverine and putrescine. The experimental conditions for the immobilizatin of amine oxidase were as described above for glucose oxidase, besides the pH of the reaction medium (phosphate buffer, 7.00). Fragments of the thread with a length varying from 20 to 130 cm, containing different amounts of enzyme (or no enzyme at all), were cut and used for the construction of biosensors by winding the thread spirally around the outer surface of the polyethylene membrane-covered cathode. The principal construction of the thread-covered biosensor is shown on Figure 1. Measurements with biosensors were carried out in an airtight and thermostated glass cell in air-saturated buffer solutions (at pH 5.60 and 7.00 in the case of glucose oxidase and amine oxidase, respectively) at 25 °C under continuous stirring at different substrate concentrations. Injection of an analyte into the reaction medium started the reaction, and the resulting decrease of the sensor output signal, proportional with oxygen concentration, was registered automatically until the signal change remained under 1%/min. The experimental data were analyzed with the help of the integrated model for membrane-covered biosensors, enabling us to calculate the signal steady-state parameters from the biosensor transient response.7 RESULTS AND DISCUSSION Enzymes immobilized onto thread-shaped carriers open new horizons and perspectives in the construction of enzyme-based biosensors. Although Clark-type oxygen sensors are too clumsy and big for microanalyses, these were used as a model support onto which the enzyme thread could be coiled in the form of winding stairs. For the immobilization of enzymes, we used different commercially available synthetic and natural threads. The nylon-6,6 thread, consisting of ∼60 filaments and being twisted together into a bigger sheaf to form the thread (Figure 2), proved (7) Rinken, T.; Rinken, A.; Tenno, T.; Ja¨rv, J. Biosens. Bioelectron. 1998, 13, 801-7.
to be the optimum thread for biosensor fabrication. The nylon filaments, having a diameter of 25 µm, were robust and there were no fragments in the thread throughout its whole length. As the dimensions of enzyme molecules normally do not exceed 100 Å (for glucose oxidase 50-80 Å), there was enough space to place the enzyme molecules onto the carrier surface during immobilization, as the diameter of filaments was more than 1000 times bigger in comparison with the attached molecules. There was also enough space between the filaments to enable sufficient transport of substrates to the enzyme by diffusion. On the other hand, the thread, coiled onto the cylindrical cathode surface, decreased significantly the steady-state output signal of the oxygen sensor as an additional diffusion barrier, also in solutions, where there was no glucose. The apparent oxygen concentration detected with the amperometric sensor decreased proportionally with increase of the length of the coiled thread, and there was no significant difference between the threads with and without the immobilized enzyme. The mechanical tensile strength also did not change noticeably in the course of enzyme immobilization: it remained above 5 MPa for threads with and without enzymes, which was fully enough for a convenient handling of the threads. In this system, the thread spiral can be considered as a net, in which the concentration of enzyme molecules and the diffusion parameters can be easily varied to obtain an optimal biosensor measurement range and sensitivity. In a net, the diffusion characteristics are determined by the size of meshes between the net threads; in our system, the diffusion barrier of the carrier is determined by the length and density of thread coiling. The measurable substrate and oxygen diffuse from the solution onto the sensor through the enzyme-containing layer, where part of the oxygen is consumed in the enzymatic reaction, catalyzed by the enzyme, and the remaining oxygen molecules are detected on the sensor. The amount of oxygen reaching the sensor in the steady state is proportional to the concentration of the second substrate. As the output signal of the sensor approaches its final value asymptotically, and in most cases the steady state cannot be fixed exactly, the steady-state parameters can be calculated according to an integrated biosensor model from the transient output of the sensor, which takes into account the reaction kinetics, diffusion, and the inertia of the sensor.8,9 Using thread (8) Rinken, T.; Rinken, A.; Tenno, T.; Ja¨rv, J. Anal. Lett. 1996, 29, 859-77.
Analytical Chemistry, Vol. 79, No. 15, August 1, 2007
6043
Table 1. Characteristics of Calibration Curves of Biosensor with Different Lengths of GO Threada length of the GO thread, cm
K1/2, mmol/L
20 40 60 80 100
6.29 4.08 2.30 2.26 1.77
Amax
range of glucose determination, mmol/L
sensitivity, L/mol × 106
0.213 0.328 0.476 0.705 0.731
0.5-18.9 0.3-12.2 0.1-6.9 0.05-6. 8 0.03-5.3
33.87 80.22 207.36 311.85 413.46
a K 1/2 and Amax have been determined by fitting calibration curves (Figure 3) to the hyperbole A ) Amax[glucose]/(K1/2 + [glucose]). The sensitivity of the sensor is presented as the ratio Amax/K1/2.
Figure 3. Dependence of a biosensor output parameter A (calculated signal change at steady state; time t f ∞) on the concentration of glucose for sensors with threads of different length having glucose oxidase activity of 1.28 × 10-3 IU/cm.
as an enzyme carrier, we can regulate the system’s parameters without special efforts. For example, 1 cm of the thread with immobilized glucose oxidase prepared as described in the Experimental Section had an activity 1.28 × 10-3 IU. Combining threads with and without enzyme, the apparent activity of the enzyme in the thread decreased according to the ratio of active and blank thread. By using activated thread with a length of 20130 cm, we could provide an enzyme layer with the activity of 0.026-0.166 IU, corresponding to the amount of soluble enzyme from 0.22 to 1.43 µg. Increasing the length of the thread increased the enzyme concentration, but also the diffusion barrier, and therefore, the range of determination became smaller, although the sensitivity increased several times (Table 1). The dependences of the maximal signal change of biosensors on glucose concentration for 20-100-cm threads are shown on Figure 3. These hyperbolic dependences can be characterized with two coefficients, K1/2 (halfsignal effect constant corresponding to the substrate concentration causing half of the maximal signal change) and Amax (theoretical maximum of signal change), which are given in Table 1. K1/2 characterizes the first part of the curves and is the base for the estimation of the range of determination of these biosensors. The ratio Amax/K1/2 was used as a measure of biosensor sensitivity. For example, an 80-cm-long thread fragment, covering ∼60% of the cathode membrane surface had glucose oxidase activity of (9) Rinken, T.; Tenno, T. Biosens. Bioelectron. 2001, 16, 53-9. (10) Rinken, T., Ja¨rv, J., Rinken, A., and Tenno, T. Biosensor and method of its construction. Estonian patent No. EE 04250 B1, 2004.
6044 Analytical Chemistry, Vol. 79, No. 15, August 1, 2007
0.102 IU, which enabled us to determine glucose concentration in the range 0.05-6.8 mmol/l ((0.05 mmol/L), while with shorter threads, the determination range could be up to 18.9 mmol/L. However, the sensitivity with the latter system was almost 10 times lower (Table 1). In the case of amine oxidase-containing thread, the enzymatic activity was 0.55 × 10-3 IU per 1 cm of thread, and with an 80-cm-long thread, we could measure cadaverine concentrations from 0.02 to 0.15 mmol/L ((0.01 mmol/L). The optimal length of the thread to obtain the required range for substrate determination and sensitivity of the sensor can be achieved by modifying the length of the thread and the ratio of enzyme-containing and empty carriers. In the tested glucose oxidase system, the optimal length of the thread was found to be 60-80 cm, causing 45-70% decrease of the initial signal and not significantly increasing the inertia of the system in comparison with shorter (20-40 cm) threads. Usage of complementary blank thread without enzyme allowed us to change the diffusion parameters and so the determination range with a moderate loss in sensitivity (increase in K1/2 without substantial changes in Amax values). Here it should be taken into account that with the increase of determination range the required analysis time increases as the higher diffusion barrier decreases the speed of reaching the stationary phase. The proposed system also enables us to combine threads with different enzymes to modify the selectivity of the biosensing system, for example, to determine two different analytes with different enzymes with the same biosensor system.10 CONCLUSIONS The proposed method of biosensor construction opens a flexible and technologically simple prospect for novel solutions, where whatever basic signal transducing system can be used in combination with a set of different bioselective threads. The application of biocatalytically active threads allows the following: (1) to measure the amount (activity) and catalytic density of a biologically active component in the units of length (in the case of homogeneous distribution of the biological component throughout the thread); (2) low-cost maintenance and rapid production and application of pieces of biologically active material, which contain a certain amount of biological component; (3) flexible setup of assay with required determination range and sensitivity by using thread(s) of different length with or without a bioactive component; (4) operative setup of assay for different analytes using a single signal transducer and changing the thread with a bioactive component (modifying biosensor selectivity); (5) potentially construct biosensors for specific compounds by combining simultaneously several carriers with different biological components in one system. ACKNOWLEDGMENT This work was supported by the Estonian Science Foundation project 6492.
Received for review February 15, 2007. Accepted May 25, 2007. AC070327J
