Development of Mevalonic Acid Biosensor Using Amperometric Technique Based on Nanocomposite of Nicotinamide Adenine Dinucleotide and Carbon Nanotubes

This study reports the development of an amperometric biosensor for quantiﬁcation of mevalonic acid (MA), which is the ﬁrst inter- mediate of HMG-CoA reductase in the isoprenoids biosynthesis pathway and therefore a useful indicator of HMGR activity. This method offers important advantages over previous reports because no radiolabeled substrates or expensive techniques are required, and time of analysis is relatively short. Self-assembled NAD + onto multiwall-carbon nanotubes (MWCNTs) was synthesized for a biosensing system. Adsorption of NAD + on MCWNTs was characterized by X-ray photoelectron spectroscopy (XPS) technique. This biosensor was constructed by modifying a screen-printed carbon electrode (SPCE) with NAD + /MWCNTs nanocomposite. The electrochemical and electrocatalytic behaviors of the modiﬁed electrode were studied using amperometry and cyclic voltam- metry (CV). The resulting biosensor demonstrated great electrocatalytic activity, good stability and fast response to MA. At the NAD + /MWCNTs-modiﬁed SPCE, the catalytic currents are linearly proportional to the concentrations of MA in the wide range from 10 nM to 140 nM with a limit of detection down to 5 nM (S/N = 3), and the biosensor exhibited a sensitivity of 18.3 μ A/mM. We measured the interference effect on the MA analysis and the results demonstrated its imperviousness to the effects of haemoglobin, bilirubin and serum albumin.

Measurement of mevalonic acid (MA) is of great interest in health monitoring as it is the first intermediate of 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGR) and appears to be a good indicator of HMGR activity. HMGR plays a crucial role in regulating sterol biosynthesis. Its activity is shown as the rate-limiting step in the mevalonate pathway in the isoprenoid metabolism pathway 1,2 and thus is a prime target for the drug development. Biofluid levels of MA can be used as a biomarker for many diseases. For example, the reduction of plasma concentrations and urinary excretion of MA levels is an indirect measure of decreased cholesterol levels. For treatment of hypercholesterolemia, the statin class of drugs are used to block HMGR activity resulting in inhibition of MA synthesis. 3,4 Moreover, cholesterol biosynthesis deficiencies cover a heterogeneous group of disorders. Mevalonate kinase is an enzyme located proximally in the pathway of cholesterol and nonsterol isoprene biosynthesis. Its allelic defects can cause hyperimmunoglobulinemia D syndrome (HIDS) and Mevalonic aciduria (MVA). The diagnosis of these diseases can be performed by determining MA concentration in urine followed by enzyme activity determination or mevalonate-5-phosphate (MVAP) detection by isotope dilution UPLC-MS/MS. 5 Interestingly, monitoring MA levels in biofluids allows a better understanding of the drug pharmacodynamics, especially with respect to response to statin therapy due to any interindividual variations, and may allow for improvements in patient management and treatment regimes. 6 Furthermore, inhibition of HMGR may be useful for treatment of many diseases other than hypercholesterolemia. [7][8][9][10] For example, elevated levels of HMGR activity was reported in rapidly proliferating human cancer cells 11 indicating that tumor malignancy may be regulated by this enzyme. MA is not only an intermediate in the biosynthesis of sterol, 12 but also isoprenoids, [13][14][15][16] which constitute a large class of secondary metabolites that are important in industries, for instance, steroids, anti-oxidants, flavors, fragrances, and pharmaceuticals including the anti-malarial artemisinin. 17 Therefore, detection of MA level is significant since it is involved with many applications. 18 z E-mail: rungtiva.pal@kmutt.ac.th A number of methods for measuring serum, plasma, or urinary MA have been reported. These methods involve enzymatic method, 19 enzyme immunoassay, 20 radioimmunoassay, 21,22 GC-MS methods 3,23-25 and LC-MS/MS assay. [26][27][28][29][30][31] However, their disadvantage is due to the expense of equipment, extensive sample preparation, time consuming, complications as well as radioactive materials. Hence, there is a need for developing a new method, which should be rapid, sensitive, specific, cost effective and preferable at the point of sample collection. From our preliminary report, 32 we demonstrated the cyclic voltammetry behavior of modified electrode which successfully showed characteristic of an efficient catalysis to MA. For this work, we aim to further develop a NAD + /MWCNTs-based amperometric biosensor through noncovalent attachment of NAD + /MWCNTs nanocomposite on a screen-printed carbon electrode (SPCE) to create innovative disposable biosensor, which is more reliable than the conventional techniques to measure MA concentration. Carbon nanotubes (CNTs) are useful as immobilization substances as they offer high surface area, superior electrical conductivity remarkable mechanical strength, and good chemical stability. 33 New types of sensor and biosensors based on nanotubes have been developed by immobilizing molecules and bimolecules on CNTs. A few papers reported the preparation of NAD + and CNTs composite modified electrode for biosensing [34][35][36] and biofuel cell applications. 37,38 To address these issues, electrochemical measurements are utilized to illustrate significant functions provided by nanocomposite, electrochemical reaction catalysis, and the electron transfer enhancement between SPCE surfaces and nanocomposite. The main challenge in developing a sensor for determining MA was the use of NAD + /MWCNTs nanocomposite modified screen-printed carbon electrode since NAD + acts as a cofactor in the reaction of MA with CoA which is then converted to HMG-CoA (see Scheme 1).

Materials and Methods
Reagents and materials.-β-Nicotinamide adenine dinucleotide ion (NAD + ), mevalonic acid (MA) and coenzyme A (CoA) were purchased from Sigma Aldrich (USA). Multi-wall carbon nanotubes Scheme 1. Conversion of MA to HMG-CoA by HMG-CoA reductase. 19 HMGR converts MA and CoA to HMG-CoA with NAD + as a cofactor. HMGR also converts HMG-CoA to MA with NADH as a cofactor. Apparatus.-X-ray photoelectron spectroscopy (XPS, PHI Ver-saProbe II) was carried out at the SUT-NANOTEC-SLRI Joint Facility, BL5.2: SUT-NANOTEC-SLRI, Synchrotron Light Research Institute. The excitation energy was 1486 eV using an Al K α source. The diameter of the irradiated area was 100 micron.
All electrochemical experiments were carried out with μ-Autolab modular electrochemical system (Eco Chemie Ultecht, The Netherlands). This instrument was operated using the GPES program (Eco Chemie). The electrochemical experiments were performed with a SPCE from Quasense. The SPCE consisted of a carbon working electrode and a reference electrode (an Ag/AgCl ink was used for screen-printed reference electrode). The three-electrode system was employed with a bare and modified carbon working electrode (GC) as working electrode (3 mm x 5 mm), an Ag/AgCl (saturated Sodium chloride) as reference electrode and counter electrode. A 0.1 M phosphate buffer was used as a supporting electrolyte. All solutions were deoxygenated by pure nitrogen for at least 15 min. The system can perform various electrochemical measurements including cyclic voltammetry and chronoamperometry.
Preparation of the NAD + /MWCNTs nanocomposite modified electrode.-Firstly, 10 mg NAD + and 2 mg MWCNTs were dispersed in 10 mL DI water. Then, NAD + was attached to the surface of MWCNTs by sonicating for 2 hours at room temperature. The mixture solution was then filtered with the membrane filter and washed with DI water several times to remove non-absorbed NAD + . After that, the NAD + /MWCNTs nanocomposite was redispersed in 250 μL DI water and kept at 4 • C. The NAD + /MWCNTs nanocomposite modified SPCE was fabricated by dropping 5 μL NAD + /MWCNTs nanocomposite on the surface of SPCE. The modified electrode was dried for 8 hours and stored at 4 • C until use.  recombinant HMG-CoA reductase. The rates of NAD(P)H consumed were monitored at room temperature every 20 sec for up to 30 mins by scanning spectrophotometrically the decrease in absorbance at 340 nm, using a Tecan Infinite M200 plate reader. Results were expressed as specific activity of the enzyme (μmol of MA produced/min/mg protein.

Results and Discussion
Adsorption of NAD + on MWCNTs.-To investigate NAD + /MWCNTs adsorption, the XPS measurement was focused on the nitrogen and phosphorous which represented the adenine, nicotinamide and phosphate groups in NAD + . XPS is a surface sensitive technique which allows the quantifying of elemental composition near the surface region. Thus, this technique is usually employed to identify oxidation state and type of the element in the interested materials. The P 2p and N 1s XPS spectra of the NAD + /MWCNTs nanocomposite with curve fittings are shown in Figure 1.
Fine scan for N 1s displayed a broad and asymmetric feature due to contributions from different nitrogen species ( Figure 1A). Peak fitting analysis exhibited two binding energies at 400.0 and 401.2 eV, which could be ascribed to the C-N and O=C-N groups. 39,40 Calculation of peak area revealed the fraction between C-N/O=C-N to be 80.9/19.1 which was consistent with NAD + molecular structure. The measurement of P 2p was further performed to confirm the NAD + /MWCNTs adsorption. The single type of phosphorous at the binding energies of 134.0 eV (2p3/2) and 134.8 eV (2p1/2) was consistent with phosphorous of the phosphate ( Figure 1B).

Electrochemical properties of NAD + /MWCNTs modified SPCE.-
The electrochemical behaviors of NAD + /MWCNTs nanocomposite were investigated using cyclic voltammetry. Figure  2 shows the cyclic voltammograms (CVs) of different modified SPCE in 0.1 M PBS pH 7.0 at a scan rate of 50 mV/s. The results show that no peaks appeared at bare SPCE (curve a) and MWCNTs-SPCE (curve b) in potential range between −0.4 to 0.4 V vs. Ag/AgCl. Comparing with the bare SPCE, the effect of the MWCNTs-SPCE exhibited higher capacitive background current, which can be attributed to the properties of carbon nanotubes which increases the conductivity and surface area of the modified electrode. Obviously, immobilized NAD + with MWCNTs (curve c), a pair of redox peaks was clearly observed at 0.122 V vs. Ag/AgCl and −0.092 V vs. Ag/AgCl corresponding to the oxidation and reduction peak, respectively with the formal potential (E 0 ') of 0.015 V. This could be ascribed to the redox reaction of NAD + /NADH on MWCNTs, indicating that NAD + has been successfully adsorbed onto MWCNTs due to π-π interaction between NAD + and MWCNTs structure and promotes electron transfer on electrode surface.
Under continuous cycling (100 cycles), the peak currents were not significant decrease (see Figure S1 of the Supplemental Material) indicating that the nanocomposite did not leach out from the electrode surface even after several runs. Then, the strong interaction of NAD + with the MWCNTs essentially endowed a high stability to the resulting electrode, which shows strong potentials for practical applications. The peak to peak potential separation ( E p ) was about 214 mV vs. Ag/AgCl and the ratio of the redox peak current I pa /I pc was about 1 suggesting that the electrochemical reaction is quasi-reversible as a result of the construction of the nanocomposite and the small peak to peak separation values indicate a fast electron transfer rate resulting from the strong interaction between NAD + and MWCNTs.
Effect of scan rates.-Effect of scan rates on peak current of the NAD + /MWCNTs nanocomposite modified SPCE was investigated in 0.1 M PBS pH 7.0 at different scan rates 32 (see Figure S2). The anodic peak currents (I pa ) and cathodic peak currents (I pc ) increased linearly with increasing scan rates from 10 to 100 mV/s. In addition, the anodic and cathodic peak potentials shifted in positive and negative directions, respectively, indicating a surface-controlled and quasi-reversible process. The inset A in Figure 2 shows the plot of oxidation and reduction peak currents as a linear function on the scan rates, the relationship of current and scan rates was linear with the linear equations as y = 0.188x + 1.077 (r 2 = 0.999), y = −0.155x − 0.875 (r 2 = 0.999) for anodic and cathodic peak current, respectively. This result indicated a surface-confined electron transfer of NAD + adsorbed onto the MWCNTs. The surface concentration ( ) of NAD + on MWCNTs modified electrode can be determined based on the Laviron's equation 41 as Equation 1: where I p is the peak current, v is the scan rate, R is the gas constant, T is the temperature, Q (9.36 × 10 −5 C) is the charge obtained by integrating the cathodic peak at low voltage scan rate (10 mV/s), A (0.045 cm 2 ) is the area of the working electrode, n is the number of electron transfer, F is the Faraday's constant. In the present case, the number of electron transfer involved in the electrochemical mechanism was calculated to be 2.19 (n = 2) suggesting that the electrode consisted of two-electron transfer reaction. The surface coverage was calculated to be 1.08 × 10 −8 mol/cm 2 which indicated that high amount of NAD + was loaded on MWCNTs.
The heterogeneous electron transfer rate constant (k s ) and the charge-transfer coefficient (α) could be calculated from the peakto-peak separation ( E p ) according to the Laviron's equations for diffusionless thin-layer voltammetry derived by Laviron 41 which are as follows: The plot of the E pa and E pc versus logarithm of the scan rates yielded two straight lines with the slopes equal to 2.3RT/(1-α)nF and −2.3RT/αnF for the cathodic and anodic peaks, respectively. From the slopes and using Equation 4, the values of α and k s were estimated to be 0.5 and 2.05 s −1 , respectively. It was suggested that this nanocomposite may facilitate the electron transfer reaction as MWCNTs promote electron transfer between NAD + and electrode surface.

Electrocatalytic behavior of NAD + /MWCNTs-SPCE to MA.-
Cyclic voltammetry was used to investigate the electrocatalytic responses of NAD + /MWCNTs-SPCE in the absence and presence of MA and CoA in 0.1 M PBS pH 7.0 at a scan rate of 20 mV/s. Figure 3A shows the CVs response to none, MA, CoA and MA+CoA obtained from the bare SPCE (curve a, b, c and d, respectively) and MWCNTs-SPCE (curve e, f, g and h, respectively). Unsurprisingly, there was no electrocatalytic activity observed after addition of MA (curve b and f), CoA (curve c and g) and MA+CoA (curve d and h) to the bare and MWCNTs-SPCE. Figure 3B shows the CVs of NAD + /MWCNTs-SPCE in the absence (curve i), the presence of MA (curve j), of CoA (curve k) and of MA+CoA at different concentrations (curve l, m and n). A typical pair of redox peak of NAD + /MWCNTs-SPCE (curve i) was obtained in the absence of MA and CoA. When only MA (curve j) or CoA (curve k) was presented, no catalytic current response to MA was observed and the redox peaks showed almost the same intensity. After addition of 20 μM (curve l), 50 μM (curve m) and 145 μM (curve n) of MA to the solution containing 150 μM CoA, the significant oxidation peak currents increased proportional with the increasing concentration of MA while the reduction peak current decreased, demon-strating the electrocatalytic activity of nanocomposite toward MA concentration. As mentioned, NAD + can be reduced to NADH using CoA as cofactor. This catalytic reaction cannot occur without NAD + . However, the oxidation peak potential shifted positively may be due to oxidation product from a two-step reaction of enzyme cycling which occurred at an electrode surface.
In the first reaction (Equation 5a), the reduction of NAD + to NADH help converting MA to mevaldehyde as an intermediate. Then, mevaldehyde reacts with CoA to form mevaldyl CoA in the Equation 5b. In the second reaction, mevaldyl CoA reacts with NAD + and, consequently, HMG-CoA and NADH are produced, respectively (Eq. 6). The process is illustrated as follows: 42 Optimized experimental conditions.-Effect of applied potential.-The applied potential was investigated in order to determine an optimal operational condition for MA measurement. Plot of chronoamperometric current versus working potential of NAD + /MWCNTs-SPCE to 50 nM MA in 0.1 M PBS pH 7.0 containing 150 μM of CoA (%RSD ≤ 6.99, n = 3) is shown in Figure 4A. Steady-state currents changed with increasing applied potentials from +0.1 to +0.3 V vs. Ag/AgCl. However, further increase of applied potentials from +0.3 to +0.6 V vs. Ag/AgCl led to reduction of steady-state currents due to the decrease in driving force of the oxidation of NADH at higher potentials. Hence, the potential of +0.3 V vs. Ag/AgCl was selected as the optimized monitoring potential due to its good sensitivity, also avoiding enzyme inactivation. 43 Effect of CoA concentration.-The amperometric current responses of the sensor depended on CoA concentrations because it is used as a mediating cofactor for MA detection. The cofactor might play an important role as an acceptor of electrons generated and transformed to NAD + . Therefore, the influence of CoA concentrations on the NAD + /MWCNTs-SPCE to MA was evaluated. The amperometric responses for various concentrations of CoA ranging from 37.5 to 300 μM were investigated when 50 nM MA was used and the results are shown in Figure 4B.
With increasing CoA concentrations, the amperometric responses of the sensor increased from 37.5 μM to 150 μM. Higher CoA concentrations gave rise to a decrease of the measured current that can be attributed to the inhibitory effect produced by high cofactor concentrations. Hence, the CoA concentration at 150 μM was chosen as the optimal cofactor concentration in this experiment.
Amperometric response of MA detection.-The analytical performance of NAD + /MWCNTs-SPCE and the determination of MA was carried out with amperometric method. The current response of the sensor was investigated under the optimal experimental conditions with injections of MA to the PBS containing 150 μM CoA at the applied potential of +0.3 V vs. Ag/AgCl ( Figure 5).  The reaction mechanism of the sensor summarized as follows: MA + CoA + 2NAD + → HMG-Co A + 2NADH [7] 2NADH → 2NAD + + 2e − + 2H + [8] MA reacts with NAD + and CoA to form HMG-CoA and NADH. The NAD + can then be recycled at the electrode leading to an increase in its oxidation current. The sensors showed excellent analytical performance, high stability, high sensitivity and high catalytic activity due to the properties of these nanocomposites which could promote the electron transfer to electrode surface as shown in Figure 6.   Figure 7A shows the current response of 50 nM MA compared with 100 μM hemoglobin, 100 μM bilirubin and 0.5 mM serum albumin to the NAD + /MWCNTs modified electrodes. In the presence of Hb, bilirubin and serum albumin, the current signal had no significant changes, indicating that the presence of Hb, bilirubin and serum albumin did not interfere with MA determination. These results confirmed that this sensor displayed good selectivity for the determination of MA.
Reproducibility and stability analysis of the MA sensor.-The reproducibility of the electrode was determined using ten electrodes made independently. After each determination, the sensor showed an acceptable reproducibility with the relative standard deviation (RSD) of 5.93% for the current response of 100 nM MA concentration indicating that the sensor had good reproducibility ( Figure S3). The long-term stability was also investigated. The storage stability of the sensor was examined by checking relative response currents of 100 nM MA every week over the test period of 4 weeks. After 4-week

B354
Journal of The Electrochemical Society, 164 (7) B349-B355 (2017) storage, the sensor showed retained activity of 76.97% of its original response as shown in Figure S4.

Measuring of HMGR specific activity at various HMG-CoA
concentrations.-The HMGR assay kit from Sigma was used to determine HMGR specific activity at HMG-CoA concentrations ranging from 0.2-100 μM. The kit is designed for the detection of HMGR activity and based on the spectrophotometric measurement of the decrease in absorbance at 340 nm, which represents the oxidation of NAD(P)H by the catalytic subunit of HMGR in the presence of the substrate HMG-CoA.
HMGR converts 1 mole of HMG-CoA using 2 moles of NAD(P)H into 1 mole of MA therefore MA quantification can be used to determine HMGR specific activity. Figure 7B illustrates HMGR specific activity at various substrate concentrations measured at room temperature. At 200 nM MA concentration, the specific activity was calculated to be 4.87 nmol MA/min/mgP, which means 1 mg of HMGR can convert 4.87 nmol of HMG-CoA to 4.87 nmol of MA. This value is the lowest measurable specific activity obtained from this method. However, the detection range of our method is much lower, i.e. 10-120 nM, if our method is used to measure a sample containing unknown MA concentration and the I A = 23.6 μA which falls in the plateau range (I reaches steady state), we can only conclude that the specific activity is approximately 4.87 nmol MA/min/mgP at MA concentrations range from 140-200 nM. This is because the kit is based on spectrophotometric detection thus cannot detect such small change in absorbance due to oxidation of NAD(P)H by the catalytic subunit of HMGR at very low substrate HMG-CoA concentrations and resulting in such high error at low MA concentrations.
The HMGR Assay Kit is an important tool for the basic research of cholesterol and other related metabolic pathways. The assay is typically used to screen for different inhibitors/activators of the purified catalytic subunit of the enzyme. The ability to measure MA concentrations at a lower range may be crucial to the development of new inhibitors/activators of the enzyme as it allows detection near the normal physiological range which are reported to be 1.0-11.2 ng/mL (6.7-75.6 nM) in plasma 29 and 12.28 ± 2.54 ng/mL (82.9 ± 17.1 nM) in serum. 30 However, the commercially available enzymatic assay is based on spectrophotometric detection thus inherits a limitation common to other spectrophotometric-based methods such as the detection range and the limit of detection are much higher (in the μM range) than our method which is based on amperometric technique (in the nM range).

Conclusions
This research is the first report of using amperometric method to measure the MA, which can represent the HMGR activity. The disposable biosensor was developed and prepared based on noncovalentlinking of NAD + /MWCNTs nanocomposite coated on SPCE, with NAD + acting as a cofactor. Importantly, the fabrication of this biosensor offers many advantages compared to other complicated immobilization procedures such as no requirements for specific reagents, very simple preparation and modification of electrode. This biosensor showed good electrochemical characteristics and performances with a specificity, sufficiently sensitive method for the detection and quantification of MA. Finally, as the detection range of our technique is also within the physiological range, it may be useful for not only clinical applications but also for improvements in treatment regimes and patient management.