Potentiometric Enzyme Sensors

Because enzymes present such an attractive possibility for achieving chemical selectivity, enzyme electrodes were the first enzymatic chemical sensors (or first biosensors) made. The early designs used any available method of immobilization of the enzyme at the surface of the electrode. Thus, physical entrapment using dialysis membranes, meshes, and various covalent immobilization schemes have been 

The choice of the ion sensor clearly depends on the type of the enzymatic reaction, namely on the products and reagents of that reaction and on the conditions of the sample. Thus, for example, there are several possibilities of the choice of the ion sensor for enzyme electrode for urea. 

One could immobilize the urease layer on top of a Severinghaus electrode for CO2 or NH3 (Section 6.3.2) and use the device as an enzymatic-potentiometric gas sensor. The primary disadvantage of such an arrangement would be its slow response time. A more direct way is through the detection of the ionic species resulting from the hydrolysis of ammonia and carbon dioxide. 

An ionic sensor for the ions involved in the above reactions can be used. The choice is usually dictated by the conditions of the sample. Thus, at low pH, the species of choice would be NHj) and at high pH it would be HCO. For the neutral pH range the best performance would be obtained from monitoring the changes of the pH itself. 

Substrate Enzyme Sensor Stability Response Range [M] 

---

Among potentiometric enzyme sensors, the urea enzyme electrode is the oldest (and the most important). The original version consisted of an enzyme layer immobilized in a polyacrylamide hydrophilic gel and fixed in a nylon netting attached to a Beckman 39137 glass electrode, sensitive to the alkali metal and NHj ions [19, 2A Because of the poor selectivity of this glass electrode, later versions contained a nonactin electrode [20,22] (cf. p. 187) and especially an ammonia gas probe [25] (cf. p. 72). This type of urea electrode is suitable for the determination of urea in blood and serum, at concentrations from 5 to 0.05 mM. Figure 8.2 shows the dependence of the electrode response... 

The second limitation comes from the fact that, despite the exquisite selectivity of enzymatic reactions, the selectivity of the potentiometric enzyme sensors is very poor. The reason for this lies in the detection mechanism itself, when factors such as the buffer capacity of the sample seriously interfere. 

Enzymatic reactions coupled to optical detection of the product of the enzymatic reaction have been developed and successfully used as reversible optical biosensors. By definition, these are again steady-state sensors in which the information about the concentration of the analyte is derived from the measurement of the steady-state value of a product or a substrate involved in highly selective enzymatic reaction. Unlike the amperometric counterpart, the sensor itself does not consume or produce any of the species involved in the enzymatic reaction it is a zero-flux boundary sensor. In other words, it operates as, and suffers from, the same problems as the potentiometric enzyme sensor (Section 6.2.1) or the enzyme thermistor (Section 3.1). It is governed by the same diffusion-reaction mechanism (Chapter 2) and suffers from similar limitations. 

A new development in the field of potentiometric enzyme sensors came in the 1980s from the work of Caras and Janata (72). They describe a penicillin-responsive device which consists of a pH-sensitive, ion-selective field effect transistor (ISFET) and an enzyme-immobilized ISFET (ENFET). Determining urea with ISFETs covered with immobilized urease is also possible (73). Current research is focused on the construction and characterization of ENFETs (27,73). Although ISFETs have several interesting features, the need to compensate for variations in the pH and buffering capacity of the sample is a serious hurdle for the rapid development of ENFETs. For detailed information on the principles and applications of ENFETs, the reader is referred to several recent reviews (27, 74) and Chapter 8. 

Decarboxylases of phenylalanine, tyrosine, and lysine and ammonia lyases of histidine, glutamine, and asparagine are also highly selective. Guilbault et al. (1988) described a potentiometric enzyme sensor for the determination of the artificial sweetener aspartame (L-aspartyl-L-phen-ylalanine methylester) based on L-aspartase (EC 4.3.1.1). The ammonia liberated in the enzyme reaction created a slope of 30 mV/decade for the enzyme-covered ammonia sensitive electrode. The specificity of the sensor was excellent however, the measuring time of 40 min per sample appears not to be acceptable. The measuring time has been decreased to about 20 min by coimmobilizing carboxypeptidase A with L-aspartase (Fatibello-Filho et al., 1988). 

Hamann (1987) employed a potentiometric urea electrode in an enzyme difference analyzer for urea determination in serum. The difference between the potential changes of a urease-covered and a bare pH glass electrode is evaluated 30 s after sample injection. This fixed-time regime provides a measuring frequency of 20-25/h the linear range for 1 120 diluted samples is 1-20 mmol/l. These results are better than those of common potentiometric enzyme sensors. 

Next we have to define the boundary and the initial conditions. For so called zero flux sensors there is no transport of any of the participating species across the sensor/enzyme layer boundary. Such condition would apply to, e.g., optical, thermal or potentiometric enzyme sensors. In that case the first space derivatives of all variables at point x are zero. On the other hand amperometric sensors would fall into the category of non-zero-flux sensors by this definition and the flux of at least one of the species (product or substrate) would be given by the current through the electrode. 

The formation of a stationary concentration of reaction product at the electrode surface is a prerequisite for proper fimctioning of potentiometric enzyme sensors. Stable conditions are estabhshed after a certain settling time where reaction speed is equal to the diffusional transport of the analyte to the electrode (Fig. 7.30). This time is represented by the response time of the sensor, hi stationary state, the maximum concentration of reaction product is located at the electrode surface (Vadgama 1990). The enzyme layer should be designed to ensure that this optimum state is reached again following a change in sample composition. Response times of potentiometric biosensors are different. Commonly they assume values of some minutes. 

Urea electrodes were among the Orst potentiometric enzyme sensors. They determine urea in blood and urine, and are based on the enzymadc decompK)sition of urea by urease ... 

Table 4.2(a) Some examples of potentiometric enzyme sensors... 

Enzyme sensors are based primarily on the immobilization of an enzyme onto an electrode, either a metallic electrode used in amperometry (e.g., detection of the enzyme-catalyzed oxidation of glucose) or an ISE employed in potentiometry (e.g., detection of the enzyme-catalyzed liberation of hydronium or ammonium ions). The first potentiometric enzyme electrode, which appeared in 1969 due to Guilbault and Montalvo [140], was a probe for urea with immobilized urease on a glass electrode. Hill and co-workers [141] described in 1986 the second-generation biosensor using ferrocene as a mediator. This device was later marketed as the glucose pen . The development of enzyme-based sensors for the detection of glucose in blood represents a major area of biosensor research. 

The classic potentiometric enzyme electrode is a combination of an ion-selective electrode-based sensor and an immobilized (insolubilized) enzyme. Few of the many enzyme electrodes based on potentiometric ion- and gas-selective membrane electrode transducers have been included in commercially available instruments for routine measurements of biomolecules in complex samples such as blood, urine or bioreactor media. The main practical limitation of potentiometric enzyme electrodes for this purpose is their poor selectivity, which does not arise from the biocatalytic reaction, but from the response of the base ion or gas transducer to endogenous ionic and gaseous species in the sample. 

Because each enzyme sensor has its own unique response, it is necessary to construct the calibration curve for each sensor separately. In other words, there is no general theoretical response relationship, in the same sense as the Nernst equation is. As always, the best way to reduce interferences is to use two sensors and measure them differentially. Thus, it is possible to prepare two identical enzyme sensors and either omit or deactivate the enzyme in one of them. This sensor then acts as a reference. If the calibration curve is constructed by plotting the difference of the two outputs as the function of concentration of the substrate, the effects of variations in the composition of the sample as well as temperature and light variations can be substantially reduced. Examples of potentiometric enzyme electrodes are listed in Table 6.5. 

In the potentiometric-type sensor, a membrane (glass, solid state, liquid) selectively extracts a charged species into the membrane phase, generating a potential difference between the internal filling solution and the sample solution (enzyme layer). This potential is proportional to the logarithm of the analyte concentration (activity) following the well-known Nemst equation. 

Meyerhoff and Rechnitz (1976) developed a potentiometric creatinine sensor by inclusion of creatinine iminohydrolase between the gas-permeable membrane of an ammonia electrode and a dialysis membrane. Since the specific activity of the enzyme used was very low, 0.1 U/mg, only 43 mU could be entrapped at the electrode. Therefore the sensor was kinetically controlled and reacted to addition of the enzyme activator tripolyphosphate by an increase in sensitivity from 44 mV to 49 mV per concentration decade and a corresponding decrease of the detection limit. These effects agree with theoretical considerations of reaction-transport coupling. The samples were treated with a cation exchanger to remove endogenous serum ammonia. 

Potentiometric penicillin sensors are mainly based on glass electrodes working with free enzyme or with enzyme included in an electrode cover. In the first penicillin electrode, (5-lactamase was directly photopolymer-ized in acrylamide on the electrode (Papariello et al., 1973). Enfors and Nilsson (1979) designed a sterilizable penicillin sensor where a solution containing (5-lactamase was pumped into a reaction chamber in front of a flat glass electrode, after sterilization of the electrode (Fig. 75). 

In potentiometric enzyme electrodes lyases producing carbon dioxide or ammonia are used as terminal enzymes of sequences. In fact, the term enzyme sequence electrode was introduced on the occasion of the design of a potentiometric D-gluconate sensor containing gluconate kinase (EC 2.7.1.12) and 6-phosphogluconate dehydrogenase (EC 1.1.1.44) (Jensen and Rechnitz, 1979). The authors found that for such a sensor to function the optimal pH values of the enzymes and the transducer should be close to each other. Furthermore, cofactors, if necessary, must not react with one another nor with constituents of the sample. It was concluded that the rate of substance conversion in multiple steps cannot exceed that of the terminal enzyme reaction. A linear concentration dependence is obtained when an excess of all enzymes of the sequence is provided, i.e. complete conversion occurs of all substrates within the enzyme membrane. Different permeabilities of the different substrates results in different sensitivities. This is particularly important with combinations of disaccharidases and oxidases, where the substrate is cleaved to two monosaccharides of approximately the same molecular size. The above... 

The above authors coimmobilized choline oxidase and AChE on a nylon net which was fixed to a hydrogen peroxide probe so that the esterase was adjacent to the solution. The apparent activities were 200-400 mU/cm2 for choline oxidase and 50-100 mU/cm2 for AChE. The sensitivity of the sequence electrode for ACh was about 90% of that for choline, resulting in a detection limit of 1 pmol/l ACh. The response time was 1-2 min. The parameters of this amperometric sensor surpass those of potentiometric enzyme electrodes for ACh (see Section 3.1.25). Application to brain extract analysis has been announced. 

FIG. 23 A schematic of the processes involved in osteoclastic resorption of bone. Two SECM tips, a potentiometric Ca2+ sensor, and an amperometric superoxide anion sensor are shown. The osteoclast resorbs bone through secretion of protons and hydrolytic enzymes, enz, which break down the bone matrix. The disposal of Ca2+ released in this process may, in principle, occur via intracellular (i) or extracellular (ii) pathways. 

---