Detection of Inducible Nitric Oxide Synthase Using a Suite of Electrochemical, ﬂuorescence, and Surface Plasmon Resonance Biosensors ⇑ Naumih M

Analytical Biochemistry 413 (2011) 157–163

Contents lists available at ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier.com/locate/yabio

Detection of inducible nitric oxide synthase using a suite of electrochemical, ﬂuorescence, and surface plasmon resonance biosensors ⇑ Naumih M. Noah, Saamia Alam, Omowunmi A. Sadik

Center for Advanced Sensors and Environmental Systems (CASE), Department of Chemistry, State University of New York at Binghamton, Binghamton, NY 13902, USA article info abstract

Article history: A suite of biosensors for rapid detection of inducible nitric oxide synthase (iNOS) is described. First, a Received 17 December 2010 metal-enhanced electrochemical detection (MED) sensor, which relied on the redox properties of a silver Received in revised form 4 February 2011 monolayer, was developed. The linear detection range was between 8.64 102 and 5.4 101 ng/ml Accepted 7 February 2011 with a detection limit of 1.69 104 ng/ml. This method was compared with surface plasmon resonance Available online 1 April 2011 (SPR) biosensors in which polyclonal mouse anti-iNOS was covalently immobilized onto a gold surface using an iNOS antigen. The linear detection range recorded was between 3.37 101 and 5.4 102 ng/ Keywords: ml with a detection limit of 2 103 ng/ml. Finally, an ultrasensitive portable capillary (UPAC) ﬂuores- Inducible nitric oxide synthase cence immunosensor, in which a mouse anti-iNOS antibody was covalently immobilized onto the inner Metal-enhanced detection Surface plasmon resonance surface of a capillary and a rabbit anti-iNOS antibody was employed as the secondary antibody, was Ultrasensitive portable capillary sensor developed. The resulting signals were found to be directly proportional to iNOS concentrations between 1 2 3 Pain biomarkers 1.52 10 and 1.52 10 ng/ml with a detection limit of 1.05 10 ng/ml. These immunosensors exhibit low cross-reactivity toward potential interferents such as human serum albumin and ovalbumin. The SPR and UPAC biosensors were validated using simulated blood spiked with recombinant iNOS, resulting in recoveries of 85% and 88.5%, respectively. The research presented in this article could poten- tially provide new ways of detecting NO for diagnostic and biomarker purposes in medical research. Ó 2011 Elsevier Inc. All rights reserved.

Nitric oxide (NO)1 is a highly reactive free radical and an impor- other hand, overproduction of NO often leads to its role as a major tant intra- and intercellular signaling molecule used for the regula- cytotoxic mediator in pathological processes, especially in inflammation of diverse physiological and pathophysiological mechanisms tory disorders. Excess NO (>10 lM) interacts with oxygen radicals to such as cardiovascular, nervous, and immunological systems [1,2]. form highly reactive peroxynitrite, which in turn induces inflamma- Also, NO has contrasting roles in living organisms; it acts as a biolog- tory cellular cytokines and hence cytokine-induced cell death via ical mediator similar to neurotransmitters regulating the blood ves- apoptosis and necrosis [2,3]. sels as well as an important host defense effector in the immune NO is synthesized in vivo during the conversion of L-arginine to system. At small concentrations (62 lM), it acts as an intra- or inter- citrulline by nitric oxide synthase (NOS). It occurs in different iso- cellular second messenger, usually via guanylate cyclase [2]. On the forms with different properties depending on the role of NO to be synthesized. For example, endothelial NOS (eNOS) and neuronal NOS (nNOS) are constitutive enzymes and are constantly ex- ⇑ pressed, even at rest in the blood vessels and the brain neurons Corresponding author. Fax: +1 607 777 447. [4]. In contrast, inducible NOS (iNOS) is expressed in macrophages E-mail address: [email protected] (O.A. Sadik). and other tissues in response to infection or inflammation [5].Itis 1 Abbreviations used: NO, nitric oxide; NOS, nitric oxide synthase; iNOS, inducible NOS; UPAC, ultrasensitive portable capillary; MED, metal-enhanced electrochemical expressed after an exposure to diverse stimuli such as the detection; BSA, bovine serum albumin; NaH2PO4, sodium dihydrogen phosphate; inflammatory cytokines and lipopolysaccharides (LPSs) [6]. Once Na2HPO4, disodium hydrogen phosphate; Na2CO3, sodium carbonate; NaHCO3, expressed, iNOS generates significantly large and sustained sodium hydrogen carbonate; IgG, immunoglobulin G; Tris–HCl, Tris–hydrochloric amounts of NO in the blood, resulting in pathological effects acid; HSA, human serum albumin; DDAO, 9H-(1,3-dichloro-9,9-dimethylacridin- 2-one-7-yl); PNPP, paranitrophenyl phosphate; GMBS, c-maleimido butyryloxy suc- [6,7]. The production of iNOS is a nonspecific event because it cinimide ester; NaCl, sodium chloride; H2SO4, sulfuric acid; NaOH, sodium hydroxide; can occur in a variety of cell types. For example, an increased pro- DMSO, dimethyl sulfoxide; HCl, hydrochloric acid; EDC, 1-ethyl-3-(3-dimethyl duction of NO has been observed during inflammation and arthri- aminopropyl) carbodiimide; NHS, N-hydroxysuccinimide; MgCl2, magnesium chlo- tis; hence, iNOS can be considered as a pain biomarker. It has also ride; PBS, phosphate-buffered saline; EDTA, ethylenediaminetetraacetic acid; DEA, been found to overexpress cyclooxygenase-2 (COX-2), which is an- diethanolamine; ELISA, enzyme-linked immunosorbent assay; PBS-T, PBS containing 0.05% Tween 20; OD, optical density; Ab–Ag, antibody–antigen; SPR, surface plasmon other pain biomarker. These two major pain biomarkers have been resonance; UPD, underpotential deposition; SAM, self-assembled monolayer. found to be expressed in cancer cells [8–10].

0003-2697/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2011.02.010 158 Detection of inducible nitric oxide synthase / N.M. Noah et al. / Anal. Biochem. 413 (2011) 157–163

There are numerous approaches in the literature for studying and 1.0 mM ethylenediaminetetraacetic acid (EDTA), and PBSB NO. These include chemiluminescence, paramagnetic resonance consisted of PBS and 1% BSA. All PBS buffer solutions were adjusted spectrometry, resonance imaging, and bioassays. We previously re- to pH 7.6 using concentrated phosphoric acid or concentrated ported on the development of an inexpensive amperometric sensor NaOH. Tris–HCl buffer solution was prepared using 6.3509 g of for the determination of the inhibitory effects of NO on xanthine Tris–HCl and 7.2199 g of Tris base and the pH adjusted to 8.0 using oxidase, horseradish peroxidase, polyphenol oxidase, and glucose HCl or 10 M NaOH. Carbonate buffer (pH 9.6) was prepared using oxidase [11–13]. Recently, Njagi et al. [14] reported an electro- 0.1 M Na2CO3 and 0.1 M NaHCO3 and the pH adjusted to 9.6 by chemical sensor for monitoring NO in brain slices. Several micro- titrating the Na2CO3 with NaHCO3 via mixing. PNPP solution was electrodes based on either direct or catalytic oxidation of NO prepared by dissolving two PNPP tablets in 0.5 M diethanolamine have also been developed [15]. Other NO sensors developed have (DEA) buffer (pH 9.8) containing MgCl2. used metalloporphyrins and platinized platinum to provide an electrocatalytic oxidation of NO [16–18]. However, these methods Methods have been found to have poor sensitivity and selectivity and also lack stability in simulated biological samples [14,15,17]. In addi- ELISA tion, NO is highly reactive with a very short half-life (5 s), thereby The enzyme-linked immunosorbent assay (ELISA) protocol used making it difficult to be continuously monitored in real time and was based on a method described elsewhere [19]. Briefly, flat- in vivo. bottomed polystyrene 96-well microplates were coated overnight In this work, we propose to monitor iNOS as a viable surrogate at 4 °C with 100 ll/well of 1 lg/ml mouse anti-iNOS polyclonal for the short-lived species, NO. The goal of this study was to devel- antibody (prepared in carbonate buffer, pH 9.6) as the capture op and assess three different biosensors for rapid detection and antibody. The plate was washed three times with PBS containing characterization of iNOS. A suite of advanced biosensors that will 0.05% Tween 20 (PBS-T, pH 7.6), and the same washing procedure enable rapid, portable, autonomous, and real-time sampling of was followed at each subsequent stage of the assay. The plates iNOS detection capabilities is envisioned. Toward this end, we ini- were incubated overnight at 4 °C with 200 ll of 1 mg/ml BSA pre- tiated the development of portable immunosensors, ultrasensitive pared in PBS (pH 7.6). The plates were washed before incubating portable capillary (UPAC) and metal-enhanced electrochemical them at 4 °C with 100 ll of different concentrations of iNOS in detection (MED), for on-demand diagnostics. Comparative analysis PBS–BSA buffer overnight at 4 °C. After washing, 100 llof1lg/ of these biosensors is reported here. ml rabbit anti-iNOS polyclonal antibody was applied to the wells (except the blank). Following incubation for 2 h at ambient tem- peratures, the plates were washed and 100 ll of goat anti-rabbit Materials and methods IgG–alkaline phosphatase in Tris buffer (pH 8.0) was added (excluding the blanks), followed by an additional 1 h of incubation Materials at room temperature. After final rinsing with Tris buffer (pH 8.0), 100 ll of 1 mg/ml PNPP solution prepared in 10% DEA buffer was iNOS and rabbit anti-iNOS polyclonal antibody were purchased finally added and the plates were incubated at room temperature from Cayman Chemical (Ann Arbor, MI, USA) and stored at 20 °C. for 30 min. Optical densities (ODs) of the solutions were then mea- Mouse anti-iNOS polyclonal antibody was purchased from Abcam sured at 405 nm using a Synergy HTRDR multidetection microplate (Cambridge, MA, USA). Bovine serum albumin (BSA) was reader. purchased from Thermo Fisher Scientific (Pittsburgh, PA, USA). The following reagents were purchased from Sigma–Aldrich (St. MED Louis, MO, USA): sodium dihydrogen phosphate (NaH PO ), 2 4 All electrochemical measurements were performed in a conven- disodium hydrogen phosphate (Na HPO ), sodium carbonate 2 4 tional three-electrode setup using an EG&G PAR 263A potentiostat (Na CO ), sodium hydrogen carbonate (NaHCO ), goat anti-rabbit 2 3 3 equipped with EG&G M270 software for data acquisition. AT-cut immunoglobulin G (IgG), alkaline phosphatase enzyme conjugate, gold quartz crystals (9 MHz, area = 0.2 cm2) consisting of 1000 Å Tris–hydrochloric acid (Tris–HCl), and Tris base. Others were human gold film with a texture of 61 lm and a 50-Å chromium adhesion serum albumin (HSA), ovalbumin, acetic acid, acetone, and phos- layer between the electrode and quartz was purchased from Inter- phoric acid. 9H-(1,3-Dichloro-9,9-dimethylacridin-2-one-7-yl) national Crystal Manufacturing (Oklahoma City, OK, USA) and used (DDAO) phosphate, paranitrophenyl phosphate (PNPP), and c- as the working electrodes. All solutions were purged with nitrogen maleimido butyryloxy succinimide ester (GMBS) were purchased for 5 min before any electrochemical measurements using a well- from Pierce Chemicals (Rockford, IL, USA). Sodium chloride (NaCl), type, 300-ll cell made of Teflon. sulfuric acid (H SO ), and sodium hydroxide (NaOH) were pur- 2 4 The gold quartz crystals were cleaned in freshly prepared chased from J.T. Baker (Phillipsburg, NJ, USA). Dimethyl sulfoxide Piranha solution (30% H2O2/H2SO4, 1:3), rinsed with water fol- (DMSO) was purchased from EMD Chemical (Gibbstown, NJ, USA). lowed by ethanol, and finally blow-dried by flushing with nitrogen Hydrochloric acid (HCl), mercaptoundecanoic acid, anhydrous etha- gas. The deposition procedure was as shown in Scheme 1A [20–22]. nol, 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (EDC), N- Briefly, silver was underpotentially deposited onto the gold quartz hydroxysuccinimide (NHS), and sodium acetate trihydrate were crystals using 1 mM silver nitrate solution at the following purchased from Fluka (Allentown, PA, USA). Magnesium chloride potentiostat settings: initial potential 503 mV, potential step 1 of (MgCl ) was purchased from Aldrich (Milwaukee, WI, USA). Borosil- 2 303 mV (49.90 s), and potential step 2 of 503 mV (10.11 s). icate thin-walled capillaries were purchased from Warner Instru- The silver-modified electrode (UPD–Ag) was further treated by ments (Hamden, CT, USA). the attachment of thiolated protein G on the gold surface. To get a maximum antibody–antigen (Ab–Ag) interaction, the antibody Stock solution preparations should be immobilized in such a manner to provide the best orien- tation and reduce any steric hindrances. The protein G used in this All stock solutions were prepared in Nanopure water with resis- case enabled the oriented immobilization of the antibody [23–25]. tivity of 18 MX. A 0.01-M phosphate-buffered saline (PBS) solution Then 200 ll of the thiolated protein G was applied on the surface of was prepared by dissolving 1.9175 g of Na2HPO4, 0.314 g of NaH2- the electrode in the Teflon well and left for 1 h. The electrode was PO4, and 8 g of NaCl in Nanopure water. PBS–EDTA consisted of PBS thoroughly rinsed three times with PBS (pH 7.6). A solution of Detection of inducible nitric oxide synthase / N.M. Noah et al. / Anal. Biochem. 413 (2011) 157–163 159

Scheme 1. Surface attachment chemistries for antibody immobilization onto sensor surface, (i) MED, (ii) SPR, (iii) UPAC.

200 llof1lg/ml mouse anti-iNOS antibody was then applied to The gold disk was activated using EDC (0.4 M)/NHS (0.1 M) chem- the surface and left for 2 h. This was followed by rinsing the elec- istry, as shown in Scheme 1B. After activation, 0.1 lg/ml mouse trode three times with PBS before applying 1 mg/ml BSA for 1 h to anti-iNOS (prepared in acetate buffer, pH 5.0) was applied and block any nonspecific binding sites of the antibody. Finally, this allowed to react for 15 min. One channel was used as control. Fol- modified electrode was incubated with varying concentrations of lowing the immobilization of mouse anti-iNOS, the nonspecific iNOS enzyme for 10 min and then characterized using cyclic binding sites of the antibody were blocked for 10 min using 1% voltammetry. BSA. Subsequently, different concentrations of iNOS enzyme, prepared in PBS (pH 7.6), were applied to the antibody-bound surface SPR for 15 min. The surface was regenerated using 0.1 M HCl solution All surface plasmon resonance (SPR) measurements used a dou- before the next antigen concentration was applied. ble channel cuvette-based instrument equipped with an autosam- pler (Autolab ESPRIT, Eco Chemie, Utrecht, Netherlands). The UPAC response of the immunosensor was automatically monitored using Capillary cleaning and pretreatment. The procedure used for UPAC a PC with ESPRIT software version 4.3.1 and kinetic evaluation soft- was as detailed in the literature [26,27]. Briefly, the capillaries used ware version 4.1.0. were attached to the tips of 1-cc Norm-Ject syringes using 1-cm A gold disk was modified with thiol groups by incubating with lengths of 1/32-in. inner diameter Tygon tubing affixed to Luer 1 mM 11-mercaptoundecanoic acid (11-MUA)/ethanol solution adapters. The capillary ends were then inserted into the tubing overnight. The disk was then thoroughly rinsed with ethanol and and held by the friction between the capillary and the tubing. A distilled water to remove any unadsorbed thiol groups followed 2.0-cm length of the Tygon tubing was attached to the distal end by blow-drying under a stream of nitrogen. The thiol-covered of the capillary to allow this end to be dipped into treatment disk/sensor was then used immediately or stored for later use. solutions without making contact with the outer surface of the 160 Detection of inducible nitric oxide synthase / N.M. Noah et al. / Anal. Biochem. 413 (2011) 157–163 capillary. Using this serial assembly, the solutions were easily flushed through the capillary by drawing on the syringe plunger. The capillaries were sequentially cleaned by washing with metha- nol/hydrochloric acid (1:1) solution for 30 min and concentrated sulfuric acid for 30 min and then exhaustively rinsed with Nano- pure water followed by acetone. Finally, these were dried by purg- ing the interiors with a gentle flow of nitrogen gas. The capillaries were incubated for 1 h with 2% 3-mercaptopropyl trimethoxysi- lane in anhydrous toluene, followed by rinsing three times with anhydrous toluene, and finally blow-dried with nitrogen gas.

Silanization chemistry and immobilization. The silanized capillaries, hereafter referred to as ‘‘sensor templates,’’ were either used immediately or stored in an airtight tube containing a drying agent for future use. These were modified using GMBS (Scheme 1C). Then Fig.1. Standard calibration curve for the sandwich ELISA showing the linear detectable range of iNOS. A limit of detection of 0.19 ng/ml was obtained. 1 mM of GMBS was prepared by dissolving 3.13 mg of GMBS (Pierce Chemicals) in 50 ll of DMSO and then diluted in 10 ml using anhydrous ethanol. Using a fresh plastic syringe (1 cc), the molecular recognition of the iNOS enzyme by the iNOS antibodies solution was drawn three times through the capillary dismal, and, therefore, confirmed that the interaction of the iNOS enzyme allowed to sit for 30 min, and then rinsed with carbonate buffer and its antibodies was viable. (pH 9.6) followed by a sandwich assay. In the sandwich assay, 1 lg/ml mouse anti-iNOS polyclonal antibody (prepared in carbonate buffer, pH 9.6) was used as the capture antibody. Different con- MED centrations of iNOS enzyme (prepared in PBS–BSA buffer, pH 7.6) were used as the antigen, whereas rabbit anti-iNOS polyclonal The MED concept relies on the idea that immobilized metallic antibody (prepared in PBS–BSA buffer, pH 7.6) was used as the sec- films deposited as a continuous layer or monolayer onto a solid ondary antibody. The sensor templates were washed three times substrate could greatly amplify biomolecular recognition of bio- after each step, and the incubation for each step was carried out molecules [21,22,28]. Hence, the MED immunosensor relies on for 1 h. The plates were then incubated with goat anti-rabbit IgG the integration of advanced nanomaterials with electrochemical conjugated to alkaline phosphatase as the detector antibody for techniques such as cyclic voltammetry, differential pulse voltam- 1 h, after which they were rinsed with Tris–HCl buffer and then metry, and underpotential deposition (UPD) of silver. Basically, taken out for analysis with DDAO as the substrate. the MED phenomenon allows label-free detection of bioaffinity recognition and precise and reproducible control of surface cover- age while simultaneously increasing the sensitivity of electro- Selectivity studies and real sample analysis chemical detection. This concept has been used in many different The effects of potential interferents, such as HSA and ovalbu- bioaffinity studies, ranging from genetic mismatch of DNA to Ab– min, were tested using the immunosensors. Simulated blood ob- Ag interactions [21,22,27,20], and has been found to offer very tained from Forensics Source (Jacksonville, FL, USA) was used as low detection limits. a real sample model to investigate the efficiency of the immuno- UPD is an electrochemical process whereby a single metal sensors. A known concentration of iNOS was spiked in the simu- adlayer is electroplated onto a dissimilar metal [29,30]. This pro- lated blood and then detected using these immunosensors. The cess was used in our experiments for the deposition of a thin film amount of iNOS recovered was compared with the amount spiked of silver onto the gold electrode surface. A self-assembled mono- in the samples to obtain the recovery and efficiency of the layer (SAM) of thiolated protein G was then immobilized on the immunosensors. UPD silver. Protein G is known to bind to the antibody through its nonspecific binding sites (Fc), leaving the specific binding sites Results and discussion of the immobilized antibody available to bind to the target antigen. This format allowed an oriented immobilization of the antibody ELISA while eliminating nonspecific recognition [23,24]. The silver layer was reported to have enhanced the stability of the alkanethiol The ELISA test is the standard method used for detecting the SAM [31]. Mouse anti-iNOS antibody was immobilized on the presence of antibody or antigen in a sample. Using this protocol, modified electrode via the thiol groups on the thiolated protein G we performed a confirmatory test for the molecular recognition to form the sensor surface (Scheme 1A). The resulting modified of the iNOS enzyme for the mouse and rabbit anti-iNOS antibodies. electrode was characterized using cyclic voltammetry. In the pres- Briefly, the ELISA comprised a sandwich mode using iNOS as the ence of the capture antibody, the silver monolayer was oxidized, antigen, mouse anti-iNOS as the capture antibody, and rabbit and this resulted in reactive silver oxides accompanied by a revers- anti-iNOS as the secondary antibody. Goat anti-rabbit IgG (which ible redox signal (anodic and cathodic peaks at 135 and 2 mV, was conjugated to alkaline phosphatase) were used to generate respectively). The redox signals remained essentially constant after the analytical signal. The alkaline phosphatase acted on the sub- several scans, thereby indicating that the presence of Ab–Ag inter- strate PNPP to give a yellow water-soluble product that absorbed action was essential for the predicted insulation of the electrode light at 405 nm. The intensity/absorbance of the yellow product surface. was directly proportional to the amount of antigen (iNOS) present. On incubation of the mouse anti-iNOS antibody-modified elec- The results in Fig. 1 show an increase in the absorbance as the con- trode with different concentrations of iNOS enzyme, a decrease in centration of iNOS was increased. A linear detection range of be- the redox signals was observed. This indicated the formation of a tween 6.75 101 and 5.74 102 ng/ml was observed with a limit biospecific molecular recognition of the Ab–Ag complex, and hence of detection of 0.19 ng/ml (calculated based on three times the the insulation of the electrode surface, as was predicted by the standard deviation of the blank). These results indicated a MED concept. The surface insulation was found to be concentration Detection of inducible nitric oxide synthase / N.M. Noah et al. / Anal. Biochem. 413 (2011) 157–163 161 dependent. Both cathodic and anodic peak currents decreased as decrease in response with no particular trend, as opposed to HCl, the concentration of the iNOS enzyme was increased. A plot of which showed an increase as the concentration of iNOS increased. the changes in anodic current versus the iNOS concentration Thus, HCl was adopted as the regeneration solution for the subse- (Fig. 2B) showed a linear detection range between 8.64 102 quent experiments. and 5.4 101 ng/ml with a detection limit of 1.69 104 ng/ml To detect iNOS enzyme using the mouse anti-iNOS sensor sur- (calculated based on three times the standard deviation of the face, different concentrations of iNOS were used. Fig. 3A shows blank). To ascertain whether the change in the peak currents was the sensorgrams at various concentrations of iNOS enzyme concen- due to the interaction between the iNOS antibody and the enzyme, tration after subtracting the signal from the control channel. Based a control experiment that showed insignificant change in the peak on the examined iNOS concentration range (3.375 101– currents was run (Fig. 2B). 5.4 102 ng/ml), there was an increase in the SPR signal with iNOS concentration. This increase is believed to be associated with the SPR mouse anti-iNOS and the iNOS enzyme binding interaction on the sensor surface. Following the binding of the iNOS, the signal The immobilization of the mouse anti-iNOS used as the capture decreased due to the dissociation of the enzyme from the surface antibody with the SPR sensor was based on the covalent attach- by the HCl regeneration solution. A control experiment where only ment of the antibody on the mercaptoundecanoic acid SAM on buffer was used did not show any increase in the SPR signal. This the gold disk. EDC/NHS chemistry was used to activate the carbox- was a further confirmation of the ELISA data. The SPR signals in ylic group of the SAM, as depicted in Scheme 1B. One channel that Fig. 3A were used to construct a standard calibration curve for was used as control was subtracted from the signal of the sample the quantification of iNOS (Fig. 3B). The sensor response was found 1 2 channels. After the immobilization of the capture antibody, BSA to be linear between 3.38 10 and 5.4 10 ng/ml with a detec- 3 solution was introduced on the surface to block any nonspecific tion limit of 2 10 ng/ml (calculated based on three times the binding sites of the antibody [32,33]. One of the major advantages standard deviation of blank). As the data showed, the direct detec- of the SPR immunosensor over other immunological methods such tion method offered by SPR was advantageous for its low detection as ELISA pertained to the reusability of the surface [34]. The regen- limit. These experimental results indicated that the SPR immuno- eration of the sensor surface was tested using two solutions (0.1 M sensor was capable of detecting iNOS at lower concentrations. HCl and glycine–HCl, pH 2.7). By monitoring the change in the response signal due to the binding of the iNOS enzyme onto the mouse anti-iNOS antibody before and after regeneration, the effi- UPAC ciency of each regeneration solution was established. With glycine–HCl as the regeneration solution, the SPR signal showed a The UPAC device uses capillary sensor templates onto which biomolecules are covalently attached to the inner wall. Its operation

Fig.2. (A) Cyclic voltammograms obtained after incubating the mouse anti-iNOS- modified electrode with different concentrations of iNOS. (B) Standard calibration Fig.3. (A) Sensorgrams showing the dependence of the SPR response on iNOS curve using a plot of the anodic peak current change versus iNOS concentration. A concentration. (B) Standard calibration curve for the SPR immunosensor under the detection limit of 1.69 104 ng/ml was recorded. optimal conditions. A detection limit of 2 103 ng/ml was obtained. 162 Detection of inducible nitric oxide synthase / N.M. Noah et al. / Anal. Biochem. 413 (2011) 157–163 is based on the excitation of the optical waveguide and subsequent Table 1 collection of the emitted fluorescence from one end of the wave- Performance characteristics and comparison of the immunosensors. guide [26,27,35]. A fluorescent reagent provides the recognition sig- ELISA MED SPR UPAC nal after Ab–Ag binding with a photomultiplier tube (PMT) fitted at LOD (ng/ml) 0.19 1.69 104 2 103 1.05 103 the end, which is used to collect the integrated fluorescence. This LDR (ng/ml) 6.75e1– 8.64e2– 3.37e0– 1.52e1– configuration takes advantage of the waveguide properties of the 5.4e2 5.4e1 5.4e2 1.52e2 capillary to integrate the signal over an increased surface area with- % recovery – – 85 88.47 Assay time Days 5 min 30 min 5 min out simultaneously increasing the background noise from the Automation No No Yes Yes detector [24,32]. It uses the selectivity of the immobilized antibody Surface No No Yes No on a ‘‘plug and play’’ basis for molecular recognition of various regeneration bioaffinity reagents, including proteins, nucleic acid, cells, and bac- Note: LOD, limit of detection; LDR, linear detection range. teria [26]. Previous immunoassays using the UPAC sensor resulted in a significant enhancement in sensitivity. The protocol used for the UPAC immunosensor was explained in The percentage selectivities for HSA and ovalbumin using the UPAC Materials and methods. A mouse anti-iNOS antibody was used as method were found to be 8.7% and 9.4%, respectively. A better selec- the capture antibody immobilized onto the entire capillary surface tivity of approximately 3% was recorded from SPR. Therefore, the using a cross-linker (GMBS) (Scheme 1C). The iNOS enzyme was immunosensors offered good selectivity toward iNOS and its poten- then sandwiched by a rabbit anti-iNOS as the secondary antibody tial interferents. and a detector antibody (goat anti-rabbit IgG conjugated to alka- The efficiency of the immunosensors was tested using simu- line phosphatase). A fluorescent reagent (DDAO) was used as a lated blood spiked with recombinant iNOS (SBSR–iNOS) as a model substrate for recognition signal after the Ab–Ag binding. An in- real sample (Forensic Sources). Blood is a commonly used sample crease in the response signal was observed as the concentration in clinical testing. Different concentrations of iNOS were spiked of iNOS enzyme was increased (Fig. 4). This indicates a biorecogni- in SBSR–iNOS and then analyzed using the immunosensors. Cali- tion of the enzyme by its antibodies. A linear detection range of be- bration curves were first prepared using standard solutions of tween 1.52 101 and 1.52 102 ng/ml was obtained, beyond iNOS; the responses compared with the standard and the percent- which the curve leveled off, thereby indicating the saturation point age recovery were calculated. The results showed an increase in of the Ab–Ag binding. A limit of detection of 1.05 103 ng/ml the immunosensor response as the concentration of SBSR–iNOS (calculated as three times the standard deviation of blank) was ob- was increased. The percentage recovery was calculated based on served. These results indicated that UPAC was a viable and sensi- the response signals of the spiked sample and the standard (Eq. tive method for detecting iNOS. (2)). After blank subtraction, the average recoveries obtained were 85% for SPR and 88.5% for UPAC. These results demonstrated the Selectivity studies and analysis in simulated blood spiked with feasibility of using these immunosensors for complex samples: recombinant iNOS response signal of spiked sample % recovery ¼ 100: ð2Þ HSA and ovalbumin were used to evaluate the selectivity of the response signal of standard immunosensors because of their protein nature that may interfere with the detection of iNOS. From the tested immunosensors, the results showed that the interferents gave much lower responses Performance characteristics and comparative analysis as compared with iNOS, which increased with concentration. The percentage of selectivity was calculated using the iNOS response A comparison of the performance characteristics of the immu- as 100% (Eq. (1)): nosensors is summarized in Table 1. The immunosensors rely on the binding affinity of the iNOS antibody and the iNOS enzyme interferent response % selectivity ¼ 100: ð1Þ complex to influence the assay time and detection limits. The iNOS response MED protocol offered the lowest detection limit, which can be attributed to the signal enhancement provided by the silver monolayer and the direct self-assembly of the antibodies onto the electrode. This was followed by the UPAC device, where the high sensitivity is due to the signal integration over the length of the capillary via the waveguiding properties of the capillary while maintaining a constant electronic background. The immunosensor is found to have a three orders of magnitude lower detection limit and response time than the ELISA method. The immunosensor also shows comparable selectivities and recoveries. The SPR method of- fers label-free detection, automation, and surface regenerations compared with other immunosensors. Overall, these immunosensors can be used to validate each other, thereby offering promising tools for the detection of iNOS and other pain biomarkers.

Conclusion

We have developed a suite of characterization methods for detection of iNOS, a major pain and cancer biomarker. Both direct and sandwich methods were tested. ELISA was used as a confirmatory test for the biorecognition of iNOS enzyme by its antibodies. Fig.4. Calibration curve for iNOS using the UPAC immunosensor. A detection limit The results confirmed that the interaction between iNOS and its of 1.05 103 ng/ml was obtained. antibodies was viable, and a portable device in which different Detection of inducible nitric oxide synthase / N.M. Noah et al. / Anal. Biochem. 413 (2011) 157–163 163 sensing recognition chemistries are integrated could be envisaged. [12] E. Kilinc, M. Ozsoz, O.A. Sadik, Electrochemical detection of NO by inhibition of Rather than detecting the NO directly, these sensors enabled rapid oxidase activity, Electroanalysis 12 (2000) 1467–1471. [13] X. Zhang, H. Ju, J. Wang, Nitric oxide electrochemical sensors, in: and quantitative detection of iNOS as a surrogate sample for the Electrochemical sensors, biosensors, and their biomedical applications, short-lived NO species. The response time of the MED immunosen- Academic Press, San Diego, 2008. sor was approximately 5 min and the device required minimal [14] J. Njagi, J.S. Erlichman, J.W. Aston, J.C. Leiter, S. Andreescu, A sensitive electrochemical sensor based on chitosan and electropolymerized Meldola sample preparation, whereas the response time of the SPR method blue for monitoring NO in brain slices, Sens. Actuators B 143 (2010) 673–680. was approximately 30 min per sample. One advantage of the SPR [15] F. Bedioui, N. Villeneuve, Electrochemical nitric oxide sensors for biological method over the MED method was that the SPR device can be auto- samples: principle, selected examples, and applications, Electroanalysis 15 (2003) 5–18. mated and the surface can be regenerated. This enabled us to ana- [16] N. Diab, J. Oni, W. Schuhmann, Electrochemical nitric oxide sensor lyze several samples on the same sensor disk. The UPAC method preparation: a comparison of two electrochemical methods of electrode provided a portable approach with a linear detection range of be- surface modification, Bioelectrochemistry 66 (2005) 105–110. 1 2 [17] Y. Lee, J. Kim, Simultaneous electrochemical detection of nitric oxide and tween 1.52 10 and 1.52 10 ng/ml with a detection limit of carbon monoxide generated from mouse kidney organ tissues, Anal. Chem. 79 3 1.05 10 ng/ml. It also allowed ultrasensitive detection with (2007) 7669–7675. the signals integrated over the length of the capillary while main- [18] X. Zhang, J. Lin, L. Cardoso, M. Broderick, V. Darley-Usmar, A novel microchip taining a constant background noise. The MED method offered the nitric oxide sensor with sub-nM detection limit, Electroanalysis 14 (2002) 697–703. lowest detection limit. This was expected because the use of silver [19] J.T. Paweska, F.J. Burt, R. Swanepoel, Validation of IgG-sandwich and IgM- nanoparticle monolayer surface allowed signal amplification. The capture ELISA for the detection of antibody to Rift Valley fever virus in recoveries obtained using simulated blood spiked with recombi- humans, J. Virol. Methods 124 (2005) 173–181. [20] A.O. Aluoch, O.A. Sadik, G. Bedi, Development of an oral biosensor for salivary nant iNOS samples ranged from 85% (for SPR) to 88.5% (for UPAC). amylase using a monodispersed silver for signal amplification, Anal. Biochem. This suggests that the immunosensors fabricated in this study can 340 (2005) 136–144. be used to detect iNOS in complicated samples such as blood and [21] I.O. K’Owino, S.K. Mwilu, O.A. Sadik, Metal-enhanced biosensor for genetic mismatch detection, Anal. Biochem. 369 (2007) 8–17. saliva. The range of immunosensors developed can be integrated [22] I.O. K’Owino, R. Agarwal, O.A. Sadik, Novel electrochemical detection scheme to detect iNOS and other important pain and cancer biomarkers, for DNA binding interactions using monodispersed reactivity of silver ions, environmental field settings, and point-of-care diagnostics. Langmuir 19 (2003) 4344–4350. [23] J.M. Fowler, M.C. Stuart, D.K.Y. Wong, Self-assembled layer of thiolated protein G as an immunosensor scaffold, Anal. Chem. 79 (2007) 350–354. Acknowledgments [24] J.M. Fowler, M.C. Stuart, D.K.Y. Wong, Comparative study of thiolated protein G scaffolds and signal antibody conjugates in the development of electrochemical immunosensors, Biosens. Bioelectron. 23 (2007) 633–639. The following agencies are acknowledged for funding: National [25] J.M. Fowler, M.C. Stuart, D.K.Y. Wong, An amperometric immunosensor with a Science Foundation (NSF) and the U.S. Environmental Protection thiolated protein G scaffold, Electrochem. Commun. 10 (2008) 1020–1023. Agency (EPA) through STAR funding. [26] M.A. Breimer, Y. Gelfand, O.A. Sadik, Integrated capillary fluorescence DNA biosensor, Biosens. Bioelectron. 18 (2003) 1135–1147. [27] S.K. Mwilu, A.O. Aluoch, S. Miller, P. Wong, O.A. Sadik, A.A. Fatah, R.D. Arcilesi, References Identification and quantitation of Bacillus globigii using metal enhanced electrochemical detection and capillary biosensor, Anal. Chem. 81 (2009) [1] F. Aktan, INOS-mediated nitric oxide production and its regulation, Life Sci. 75 7561–7570. (2004) 639–653. [28] F. Yan, O.A. Sadik, Enzyme-modulated cleavage of dsDNA for studying [2] H. Cho, C.-W. Yun, W.-K. Park, J.-Y. Kong, K.S. Kim, Y. Park, S. Lee, B.-K. Kim, interfacial biomolecular interactions, J. Am. Chem. Soc. 123 (2001) 11335– Modulation of the activity of pro-inflammatory enzymes, COX-2 and iNOS, by 11340. chrysin derivatives, Pharmacol. Res. 49 (2004) 37–43. [29] C.H. Chen, A.A. Gewirth, Correlation of electrode surface-structure with [3] M. Tannous, R. Amin, M.R. Popoff, C. Fiorentini, A. Kowluru, Positive activity toward H2O2 electroreduction for Bi monolayers on Au(III), J. Am. modulation by Ras of interleukin-1b-mediated nitric oxide generation in Chem. Soc. 114 (1992) 5439–5440. insulin-secreting clonal b (HIT–T15) cells, Biochem. Pharmacol. 62 (2001) [30] G.K. Jennings, P.E. Laibinis, Underpotentially deposited metal layers of silver 1459–1468. provide enhanced stability to self-assembled alkanethiol monolayers on gold, [4] X.Q. Wei, I.G. Charles, A. Smith, J. Ure, G.J. Feng, F.P. Huang, D. Xu, W. Muller, Langmuir 12 (1996) 6173–6175. S. Moncada, F.Y. Liew, Altered immune responses in mice lacking inducible [31] G.K. Jennings, P.E. Laibinis, Self-assembled n-alkanethiolate monolayers on nitric oxide synthase, Nature 375 (1995) 408–411. underpotentially deposited adlayers of silver and copper on gold, J. Am. Chem. [5] J.W. Coleman, Nitric oxide in immunity and inflammation, Int. Soc. 119 (1997) 5208–5214. Immunopharmacol. 1 (2001) 1397–1406. [32] D.R. Shankaran, K.V. Gobi, T. Sakai, K. Matsumoto, T. Imato, K. Toko, N. Miura, A [6] L. Franco, D. Doria, E. Bertazzoni, A. Benini, C. Bassi, Increased expression of novel surface plasmon resonance immunosensor for 2,4,6-trinitrotoluene inducible nitric oxide synthase and cyclooxygenase-2 in pancreatic cancer, (TNT) based on indirect competitive immunoreaction: a promising approach Prostaglandins Other Lipid Mediat. 73 (2004) 51–58. for on-site landmine detection, IEEE Sens. J. 5 (2005) 616–621. [7] C. Nathan, Q.W. Xie, Nitric-oxide synthases: roles, tolls, and controls, Cell 78 [33] D.R. Shankaran, K.V. Gobi, T. Sakai, K. Matsumoto, K. Toko, N. Miura, Surface (1994) 915–918. plasmon resonance immunosensor for highly sensitive detection of 2,4,6- [8] S.A. Wardlaw, T.H. March, S.A. Belinsky, Cyclooxygenase-2 expression is trinitrotoluene, Biosens. Bioelectron. 20 (2005) 1750–1756. abundant in alveolar type II cells in lung cancer-sensitive mouse strains and [34] S. Suwansa-ard, P. Kanatharana, P. Asawatreratanakul, B. Wongkittisuksa, in premalignant lesions, Carcinogenesis 21 (2000) 1371–1377. C. Limsakul, P. Thavarungkul, Comparison of surface plasmon resonance and [9] S.F. Kim, D.A. Huri, S.H. Snyder, Inducible nitric oxide synthase binds, S- capacitive immunosensors for cancer antigen 125 detection in human serum nitrosylates, and activates cyclooxygenase-2, Science 310 (2005) 1966–1970. samples, Biosens. Bioelectron. 24 (2009) 3436–3441. [10] M.M. Taketo, COX-2 and colon cancer, Inflamm. Res. 47 (1998) 112–116. [35] S. Koch, H. Wolf, C. Danapel, K.A. Feller, Optical flow-cell multichannel [11] C.D. Geddes, J.R. Lakowicz (Eds.), Advanced Concepts in Fluorescence Sensing: immunosensor for the detection of biological warfare agents, Biosens. Part A. Small Molecule Sensing, Springer, New York, 2005. Bioelectron. 14 (2000) 779–784.