Cofactors dependent Oxidoreductase enzymes are of great industrial use in biotransformation reaction [1,2]. A majority of these enzymes require coenzymes such as NAD (H), NADP (H) and ATP for their actions. In recent years, NAD (H) and NADP (H) based reactions have been examined extensively for their chemical processing applications [3]. Because of the stoichiometric requirement and prohibitively high costs of NAD (H), cofactor-linked enzymes have not found to be of much use at commercial scale. Efficient regeneration and reuse of cofactors are essential for large scale synthetic applications of cofactor based enzymes [4]. Therefore, to maintain cost effectiveness in the use of cofactor-based enzymes at commercial scales, it is necessary to develop an efficient method to recycle them in situ. In addition, cofactor regeneration can also drive the reaction to completion, simplify product isolation, and allow the removal of inhibitory cofactor byproducts, further reducing the cost of synthesis.

Chemical, electrochemical, photochemical, microbial and enzymatic reactions have all been developed for cofactor regeneration [4]. Enzymatic method is preferable due to its high efficiency and selectivity. There are two ways to achieve the enzymatic regeneration; one is through the use of substrate coupled reaction system and the second is through the use of second enzyme for recycling of cofactor. For majority of cofactor regeneration reaction, the use of second enzyme is preferred [3] because of its affordability of broader option of substrates. Cofactor regeneration with the help of immobilized enzyme systems which is preferred in case of industrial systems for recovery and reuse of enzyme is desirable.

Solid support attached insoluble cofactor and enzymes are much easier to reuse and may afford more flexible reactor design as compared to membrane systems.

However, most of the immobilized cofactors have been used with free enzymes, as it was difficult to achieve activity with systems that have cofactor and enzyme both immobilized [5]. In one of the study, silica nanoparticles supported enzymes: Lactate Dehydrogenase (LDH); Glutamate Dehydrogenase (GLDH) and NAD (H) were prepared to catalyze coupled reactions for production of α-ketoglutarate and lactate with cofactor regenerated within the reaction cycle [3].

Other similar studies for cofactor recycling have been done by Zahab et al. [6] and Luo et al. [7]. In the present work two enzymes alcohol dehydrogenase (ADH) from Baker’s yeast and formate dehydrogenase (FDH) from Candida boidinii were immobilized on Alumina nanoparticles and cofactor recycling reactions were carried out both by free and immobilized enzyme systems and efforts were made to develop an efficient cofactor recycling system. It was assumed that ADH will catalyze reaction propionaldehyde to n-propanol generating reduced form of NAD i.e., NADH which in turn will be oxidized by FDH using Sodium formate as substrate. Product of these reactions of Sodium formate will generate water and carbon dioxide which are non interfering with the reaction set up and can easily be removed.

Materials

Alcohol dehydrogenase (ADH) (EC 1.1.1.1, lyophilized powder) from Baker’s yeast, NAD (H), NAD, propionaldehyde and n-propanol were all purchased from SRL (Sisco Research Lab Private Ltd. Mumbai, India). Formate dehydrogenase (FDH) from C.boidinii (EC:1.2.1.2,lyophilized powder), 3-aminopropyl Trimethoxy silane (APS) and Bovine serum albumin (BSA) were purchased from Sigma Chemical Co. (St. Louis, USA) Sodium Formate was purchased from Hi Media Laboratories Pvt. Ltd. Mumbai, India. Glutaraldehyde was procured from Loba Chemie Pvt. Ltd., Bombay. Alumina and Zirconium nanoparticles were a kind gift from CGCRI (Centre for glass and Ceramics Research Institute, Kolkata, India).

Functionalization of nanoparticles

Prefabricated alumina nanoparticles were functionalized using two different methods, modified procedures for silanization of alumina sheets [8] and silica gel [9] respectively. 0.5 g of dry alumina nanoparticles were silanized using 10 ml of 5% APS (Aminopropyl trimethoxysilane) in dry acetone for 1 h (at RT and neutral pH). The nanoparticles were recovered by centrifugation at 8000 rpm for 30 min. Excess silane was washed off with acetone and the nanoparticles baked at 100°C overnight in an oven. The dried nanoparticles were then treated with 2.5% glutaraldehyde (in 0.1 M phosphate buffer, pH 7.05) for one hour, washed and stored as a wet cake till further use (method 1). In another method (method 2) same process was modified by treating dry nanoparticles with 10 ml of 10% aqueous APS (pH 3.45) at 75°C for 2.75 hours in a temperature controlled water bath and washing was done with distilled water. In case of prefabricated zirconium nanoparticles, same methods were performed as that followed by Chen et al. [10].

Enzyme immobilization

Immobilization of enzyme on alumina nanoparticles were carried both with adsorption method and by covalent coupling. Physical adsorption of enzymes on alumina nanoparticles (not functionalized before use) were checked for suitability as a viable method for immobilization by allowing suitably diluted enzyme (1 mg/ml, 0.1 M phosphate buffer, pH 7.05) to react with alumina nanoparticles at 5-10°C for 2 hours. The nanoparticles were collected by centrifugation at 8000 rpm for 30 minutes and the supernatant stored for checking the presence of unbound enzyme. The particles were washed and stored. Covalent coupling was performed by same method except that particles used were functionalized by above mentioned Method 1 and 2. In case of zirconium, functionalized nanoparticles were added to phosphate buffer (10 ml) of same pH as that of enzyme solution, immobilized solid particles were collected by centrifugation, followed by repeated washing with buffer solution and stored after freeze drying.

Activity assay

ADH activity was measured by monitoring the decrease in NADH absorbance at 340 nm with time. The final reaction mixture consisted of 30 mM propanaldehyde, 0.3 mM NADH and suitably diluted enzyme solution in 0.1 M phosphate buffer (pH 7.05) containing 0.1% BSA. The reaction was started by the addition of 100 μl of the suitably diluted enzyme and the absorbance at 340 nm was monitored for 5 minutes. One Unit of enzyme activity was defined as enzyme required for oxidize one micromole of NADH/minute. FDH activity was measured by monitoring the increase in absorbance at 340 nm due to the generation of NADH. One unit of FDH activity was defined as the formation of 1.0 μmol of NADH per min under the standard assay conditions (pH 7.4, 25°C). The activities of enzyme loaded particles were measured in a batch reactor. Suitably diluted nanoparticles conjugated with enzyme suspensions were prepared and added to the reaction mixture containing respective substrates and cofactors, as the case may be. Separate blanks were used for each sample containing a particular amount/concentration of nanoparticles. Simultaneously, enzyme activity was also determined in the supernatant.

Enzyme loading and specific activity

Different amounts of enzymes (5, 10 and 20 mg) were incubated with the same amount of nanoparticles to determine the loading capacity of the particles. Enzyme loading was calculated as activity units retained/ amount of nanoparticles. Further, the protein loading was checked to ascertain the changes (if any) in the specific activity of the enzyme upon immobilization. The protein loading was determined based on the difference of the protein content in solution (including wash) before and after immobilization, by Folin-Lowry [11] method.

Characterization of immobilized enzymes on nanoparticles

Nanoparticulate systems morphology such as shape and occurrence of aggregation phenomena was studied by scanning electron microscopy (SEM). Samples of nanoparticles were mounted on metal stubs, gold coated under vacuum and then examined on a Carl Zeiss EVO Series Scanning Electron Microscope.

Effect of pH on immobilization of enzyme:

Enzyme and metal nanoparticle solution was prepared in buffer of different pH (4.0, 9.0, 7.5) and same procedure of immobilization was repeated in all the cases.

Reusability and stability assay

Reusability assay of enzyme immobilized nanoparticles were done by repeated washing with phosphate buffer before adding to second set of reactions. Stabilities of free and immobilized enzymes were checked at different time interval in same storage condition and compared.

Coupled Enzymatic Reactions

Coupled enzymatic reactions were set up using enzymes (ADH & FDH), substrates (propanaldehyde and Sodium formate), cofactors and analysis was done both by spectrophotometer (Optizen 3220 UV^bio) and by Gas chromatography (HP Nucon, 6500). Effect of changing enzyme activity ratios on the steady state cofactor concentration was studied by using a constant ADH concentration and the FDH concentrations were suitably adjusted to get final ADH to FDH activity ratios of 2.5, 5 and 10.28. The final substrate concentrations in the reaction cocktail consisted of 30 mM propanaldehyde, 30 mM sodium formate and 0.3 mM NADH. Spectrophotometrically, at time zero, ADH was added and the reaction was monitored for 5 minutes by measuring the absorbance at 340 nm. Then, the addition of FDH was carried out and the coupled reaction was monitored for another 10 minutes (till steady state was reached w.r.t the cofactor concentrations in the reaction). Combined reactions were also monitored by gas chromatography; three different systems were tried, containing both free enzymes, one immobilized enzyme (ADH) and both enzymes immobilized respectively. The reactions were carried out in 0.1 M potassium phosphate buffer (pH 7.5) at room temperature with magnetic stirring, with the amount of each catalytic component controlled to achieve the desired molar ratio. A typical reaction included addition of 3 mg ADH-carrying particles and 30 mg FDH-carrying particles in 1mL reaction mixture containing 0.8 M propanaldehyde, 0.8 M sodium formate and 1 mM NADH. 50 μl aliquots were collected and centrifuged to remove the immobilized enzymes, and the supernatants used for GC analysis. The reactions were monitored by measuring the rate of disappearance of substrate which is propanaldehyde in the combined reaction. Controls without enzymes were used to check for loss of propanaldehyde due to other factors.

Enzyme immobilization on nanoparticles and mechanisms

Adsorption method of immobilization was ineffective in case of binding of enzymes to alumina nanoparticles while it worked for zirconium nanoparticles. No detectable activity was observed in either the supernatant or in the particles by adsorption method when enzymes were added to alumina. This suggests that direct contact with the surface of the alumina particles denatures the enzyme completely. Unsuccessful direct coupling of the enzyme with the alumina nanoparticles suggested the need for surface modification of the particles by molecules of desired properties. Biofunctionalization of metals with the use of silane precursors involves treating the surface hydroxyl groups with trialkoxyaminosilanes and consequent functionalization using different routes for modification of the primary amino groups. Modification is carried out for covalent coupling of enzymes with the help of glutaraldehyde, a bi-functional molecule with two aldehyde groups. Upon reaction, one of the aldehyde groups reacts with the amino group on the aminosilanes to form an amide bond, leaving the other aldehyde group free to react with the secondary amino groups present on the amino acids comprising the protein (Figure 1). Nearly equal activity, equivalent to 4.8 mg of the free enzyme, was retained in the supernatant in both methods of alumina functionalization. However, the enzyme loading on the particles varied (Figures 2 and 3). In case of zirconium, 7.5 pH, room temperature (28°C), contact time of 24 hours and enzyme concentration 1.2 mg/ml were considered optimum for immobilization. A typical SEM micrograph for the metal nanoparticles without and with bound enzymes has been shown in Figure 4a and b. Nanoparticle aggregates were clearly visible and had a mean diameter of 250 nm, after binding enzymes their mean diameter increases slightly almost similar to that of unbound ones, this reveals that binding process did not significantly result in change in size of particles this could be due to change occurred only on particles surface, same case was reported by Li et al. [12] in case of binding of SCAD on magnetic nanoparticles.

Figure 1: Schematic representation of the process of biofunctionalization of alumina nanoparticles.

Figure 2: Immobilized enzyme activity assays. M1: Method 1; M2: Method 2; Sup: Supernatant; Particles: Nanoparticle suspensions. The sample dilutions are also mentioned.

Figure 3: Activity retention on particles and supernatant post Immobilization.

Figure 4: Scanning electron micrographs of metallic nanoparticles (a) Alumina and Zirconium without enzymes (b) Alumina and Zirconium with bound enzymes.

Enzyme loading and specific activity

Alumina particles functionalized by Method 1 showed retention of 470 units of enzyme activity whereas the corresponding value for nanoparticles prepared by Method 2 was 14.44 units and enzyme loading in Method 1 was 11.77 U (mg particles) as compared to 0.35 in Method 2. It was nearly 30 times more than that in Method 2. This could be attributed to the limited stability of covalently bonded molecules on alumina due to the activation of the Al-O-Si bond breakage due to a local increase in pH near the surface. This increase in pH has been attributed to the presence of amino groups in the APS molecule, and becomes significant when the particles are exposed to aqueous solutions before the conversion of the amine groups to amides by glutaraldehyde. Silanization in Method 1 was carried out in dry acetone, whereas Method 2 used aqueous solutions. Hence, the lower loading in Method 2 can be attributed to the breakage of the Al-OSi bond, resulting in lesser glutaraldehyde molecules available at the surface for coupling. Therefore, Method 1 was chosen as the method of choice for immobilization of ADH/FDH on alumina nanoparticles. The maximum loading of 10.7 ± 0.5 U/mg was achieved with 10 mg of ADH and hence this amount was used in all further experiments. For coupling of FDH, using an initial amount of 10 mg FDH, a loading of 0.077 U/mg was achieved. However, optimization could not be carried out due to limited quantity of FDH available and the above preparation was used for the experiments. The protein loading on the nanoparticles was determined as described previously to assess the changes in specific activity of the enzyme upon immobilization on nanoparticle supports. The protein content of the commercial enzyme preparation for ADH was about 30% (i.e., 0.3 mg protein/mg solid) while that for FDH was 43%. An increase in the specific activity (Vmax) for the enzymes is obtained upon the immobilization of the enzymes upon nanoparticle supports. Such a trend has been seen in other studies as well [13]. The reason for this behavior is not clear. However, in the aforementioned study; a corresponding decrease in activation energy for the enzyme was noted upon immobilization, which may be responsible for the increase in activity. In case of zirconium loading of enzyme (ADH) was calculated as 1.0 U/mg of nanoparticles.

Stability of immobilized enzyme

The free enzyme (ADH) solutions lost 24 and 72% of their activities upon storage at 4-8°C for 24 and 48 hours, respectively. In contrast, the immobilized enzymes lost about 3 and 25% of their activity after 24 hours and one week, respectively. Therefore, the stability of the enzymes increased due to their covalent coupling with the functionalized alumina particles. The reutilization capacity of the immobilized enzymes on zirconium particles were up to three cycles with minor loss of activity.

Influence of pH during enzyme immobilization

It has been reported that the pH value of the buffer solution during immobilization has a great influence on the activity and enantioselectivity of the biocatalyst [14]. In this work, the immobilization of enzyme (ADH) on zirconium was carried out at pH 4.0, 5.0, 7.5 and 9.0. Similar enzyme loadings within the range of 0.0040-0.0095 mg/mg of matrix are obtained, but the activities are quite different with optimal results achieved at pH values of 7.5.

Coupled enzyme reactions

Effect of ratio of enzyme activities on the steady state cofactor concentrations: Addition of second enzyme (FDH), led to gradual increase in absorbance due to production of NADH by FDH, with the absorbances indicating the attainment of steady state within 10 minutes (Figure 5). Increasing FDH concentrations (or decreasing the ADH/FDH activity ratios) leads to higher steady state concentrations of NADH. For example, when ADH/FDH activity ratio=10.28, nearly 88% of the NADH added initially is in the form of its oxidized form NAD+ at steady state (Figure 6). Therefore, we can control the steady state concentrations of the two forms of the cofactor (oxidized and reduced) by altering the activities of the enzymes in the reactions. Since the reduced form of the NAD (H) cofactor is less stable in solution, in subsequent experiments, high ADH/FDH ratios were used to retain greater percentage of the cofactor in its oxidized form.

Figure 5: Effect of enzyme activity ratio on steady state cofactor concentrations.

Figure 6: Fraction of cofactor existing as NADH at steady state as a function of the enzyme activity ratio.

Total Turnover Number and Cofactor recycling rate: Coupled reactions were first carried out with free enzymes as that of Wykes et al. [5] in 250 μl reaction volumes in ELISA plates with continuous shaking. Using 0.8 M of substrates, 21.54 and 2.09 units of ADH and FDH respectively (ADH/FDH=10.28) and different concentrations of the cofactor, the reaction was carried out for 1 hour. A maximum Total Turnover Number (TTN=moles of product formed/moles of cofactor added) of 1320 was achieved. The cofactor recycling rate, i.e., the moles of product formed per hour normalized by the moles of cofactor added, was same as the TTN in this case (Table 1). When similar experiments were carried out with free FDH and immobilized ADH added in the same activity ratios, results were similar to the free enzyme systems, indicating that such systems perform equally well. Further, the examination of TTN in systems containing both enzymes immobilized indicated similar dynamic cofactor-enzyme interactions. Using 0.8 M substrates, 0.5 mg ADH loaded particles and 7.5 mg FDH loaded particles, the reaction was carried out for 5 hours and a maximum TTN of 2900 was reached (equivalent to 580 h^-1) with the lowest concentration of cofactor used being 140 μM (Table 2). Also, increasing the cofactor concentration lowered the TTN by the same ratio, which can be attributed to the low Km values for NADH that maximize the reaction rates at very low concentrations of the cofactor. Cofactor recycling experiments were not carried out by zirconium immobilized enzymes due to lack of resources and time.

Table 1: TTN and Cofactor Recycling Rate (CRR) at different initial cofactor concentrations for coupled reactions with free enzymes; CRR was calculated as number of moles of product formed per hour per mole of cofactor added.

Table 2: TTN and cofactor recycling rates at different initial cofactor concentrations for coupled reactions with immobilized enzymes.

In cofactor regeneration experiments, cofactor recycling rate is a very important parameter since it can affect the recycling capacity and hence the TTN achievable by the system. Higher cofactor recycle rates can be achieved by increasing the concentration of the substrates and/ or the amount of enzymes.

Effect of substrate concentration on cofactor recycle: The substrate concentrations are expected to affect the TTNs by affecting the rate of reaction of the enzyme. Two different concentrations of propanaldehyde (0.5 and 0.8 M) were tested and no difference was found in the TTNs, with the absolute molar conversion of propanaldehyde being similar at comparable time points (the total reaction time for the smaller concentration being proportionally smaller). These concentrations were well above the Km for propanaldehyde, however, the effect of lowering the concentration of propanaldehyde below its Km value could not be determined, because of the concentration range in contention being well below the limit of detection by gas chromatography.

Effect of enzyme concentration on cofactor recycle rates: Increasing the amount of enzymes in the system can improve the recycle rates as depicted above. A maximum rate of 6650 cycles/hour was achieved with 0.14 mM cofactor and the highest amounts of enzymes used in the experiment. Increasing the amount of enzymes may further increase the recycle rate; however, care has to be taken to maintain a high liquid-to-solid ratio in the system. This is because particle mobility is an important criterion in determining the efficiency of cofactor recycling in such systems. Higher particle concentrations are required to improve the collision frequency of the particles, which would facilitate the reactions. However, at higher concentrations, aggregation of the particles may be a problem. Therefore, a balance between the two is necessary, and hence it is expected that below a certain liquid-to-solid (w/w) ratio, increasing the amount of particles (consequently the enzymes) would result in a decrease in the cofactor recycle rates. The recycle rates can be further improved by optimizing the factors influencing the performance of the enzymes in the system - the buffer composition, its pH, the immobilization system and by increasing the stability of the cofactor in the system - by controlling the form of the cofactor at steady state and/or by immobilization of the cofactor itself.

ADH and FDH enzymes were immobilized on metals nanoparticles (i.e., Alumina and Zirconium) and a cofactor recycling system was developed consisting of both enzymes immobilized separately on Alumina nanoparticles and free cofactors. As expected, mobility of particles is critical to realizing dynamic enzyme–cofactor interactions, and efficient recycling of cofactor can be achieved when the particles well dispersed in the solution. The use of particle-attached systems will allow easy recovery through methods such as filtration and precipitation, and can be recycled and reused, thus substantially improving the potential of cofactor-dependent biochemical reactions for large-scale industrial bioprocessing applications.

Financial support from Lockheed Martin Corporation USA is gratefully acknowledged. We thank CGCRI Kolkata for providing nano particles and Dr Sanjay Dhakate, NPL for SEM and TEM facility for nanoparticle characterization during the study.