Anal. Chem. 2005, 77, 1771-1779
Optical Fiber Immunosensor Based on a Poly(pyrrole-benzophenone) Film for the Detection of Antibodies to Viral Antigen T. Konry,† A. Novoa,† Y. Shemer-Avni,† N. Hanuka,† S. Cosnier,‡ Arielle Lepellec,‡ and R. S. Marks*,†
The National Institute for Biotechnology in the Negev, and the Department of Biotechnology Engineering, Faculty of Engineering Science and the Department of Virology, Faculty of Health Science, Ben Gurion University of the Negev, P.O. Box 653, Beer Sheva, 84105, Israel, and Laboratoire d'Electrochimie Organique et de Photochimie Redox, UMR CNRS 5630, Universite´ Joseph Fourier Grenoble 1, 301 Rue de la Chimie, BP 53, Grenoble Cedex 9, 38041, France
We describe herein a newly developed optical microbiosensor for the diagnosis of hepatitis C virus (HCV) by using a novel photoimmobilization methodology based on a photoactivable electrogenerated polymer film deposited upon surface-conductive fiber optics, which are then used to link a biological receptor to the fiber tip through light mediation. This fiber-optic electroconductive surface modification is done by the deposition of a thin layer of indium tin oxide on the silica surface of the fiber optics. Monomers are then electropolymerized onto the conductive metal oxide surface; thereafter, the fibers are immersed in a solution containing HCV-E2 envelope protein antigen and illuminated with UV light (wavelength ∼345 nm). As a result of the photochemical reaction, a thin layer of the antigen becomes covalently bound to the benzophenonemodified surface. The photochemically modified fiber optics were tested as immunosensors for the detection of anti-E2 protein antibody analyte that was measured through chemiluminescence reaction. The biosensor was tested for sensitivity, specificity, and overall practicality. Our results suggest that the detection of anti-E2 antibodies with this microbiosensor may enhance significantly HCV serological standard testing especially among patients during dialysis, which were diagnosed as HCV negative, by standard immunological tests, but were known to carry the virus. If transformed into an easy to use procedure, this assay might be used in the future as an important clinical tool for HCV screening in blood banks. We present here a newly developed optical microbiosensor for the diagnosis of hepatitis C virus (HCV) by using a novel photoimmobilization methodology based on a photoactivable electrogenerated polymer film. In the field of optical fiber immunosensors, we have previously shown that optical fibers, which are made of doped silica and hence are electrically inert, can be * To whom correspondence should be addressed. E-mail: rsmarks@ bgumail.bgu.ac.il. Tel.: +972 (8) 6477182. Fax: +972 (8) 6472857. † Ben Gurion University of the Negev. ‡ Universite ´ Joseph Fourier Grenoble 1. 10.1021/ac048569w CCC: $30.25 Published on Web 02/17/2005
© 2005 American Chemical Society
modified by an electrically conductive layer. This fiber-optic electroconductive surface modification was done by the deposition of a thin layer of indium tin oxide (ITO) and may be used for the electrogeneration of polymerized films.1,2 For instance, surfaceconductive fiber optics were efficiently modified by electrochemically polymerized poly(pyrrole-biotin) films.1,2 The latter were then used for the immobilization of biomolecules by the biotinavidin linkage.1,2 The great advantage of this procedure is that one can now use the versatile electrochemical deposition procedures to precisely electrogenerate a polymer coating over conductive microsurfaces of complex geometry. In addition, the films can be prepared easily in a rapid, reproducible, and well-controlled, one-step procedure. We further improved the procedure by creating an electropolymerizable pyrrole-benzophenone molecule that allows the linking of a biological receptor to the fiber tip through light mediation (Figure 1A). The photoreaction process for benzophenone included a triplet-state excitation (A), a hydrogen abstraction (B), and a radical recombination (C) creating the covalent binding with, for instance, nucleic acids or proteins bearing amino acids with sterically accessible C-H bonds.3-7 Such an immobilization using an electropolymerized photoreactive group is easily applicable to a wide variety of biomolecules. This innovative photoelectrochemical method for the immobilization of biological macromolecules combines the advantages of photolithography with those of the electrochemical addressing of polymer films.8 HCV is widely spread and is a major cause of sever liver diseases, such as cirrhosis and hepatocellular carcinoma.9,10 (1) Konry, T.; Novoa, A.; Cosnier, S.; Marks, R. S. Anal. Chem. 2003, 75, 26332639. (2) Marks, R. S.; Novoa, A.; Konry, T.; Krais, R.; Cosnier, S. Mater. Sci. Eng. C-Biol. S 2002, 21, 189-194. (3) Dorman, G.; Prestwich, G. D. Biochemistry 1994, 33, 5661-5673. (4) Fleming, S. A. Tetrahedron 1995, 51, 12479-12520. (5) Prestwich, G. D.; Dorman, G.; Elliott, J. T.; Marecak, D. M.; Chaudhary, A. Photochem. Photobiol. 1997, 65, 222-234. (6) Berens, C.; Courtoy, P. J.; Sonveaux, E. Bioconjugate Chem. 1999, 10, 5661. (7) Turro, N. J. Molecular Photochemistry; W. A. Benjamin: New York, 2001. (8) Cosnier, S.; Senillou, A. Chem. Commun. 2003, 3, 414-415. (9) Hanuka, N.; Sikuler, E.; Tovbin, D.; Neville, L.; Nussbaum, O.; Mostoslavsky, M.; Orgel, M.; Yaari, A.; Manor, S.; Dagan, S.; Hilzenrat , N.; Shemer-Avni, Y. J. Med. Virol. 2004, 9999, 1-7.
Analytical Chemistry, Vol. 77, No. 6, March 15, 2005 1771
Figure 1. (A) Mechanism for the photochemical reaction of benzophenone with a C-H bond of an amino acid side chain (adapted from Scheme 1 of ref 5). (B) pyrrole-benzophenone.
Cryptic HCV infection relates to patients chronically infected with HCV that are seronegative but have HCV-RNA.9,10 These patients are not identified in the standard serological tests for HCV, which are based on the detection of antibodies to core, NS3 and NS5 antigens of HCV, but not to the glycosylated E2 envelope protein.9,10 They will therefore be wrongly diagnosed as noninfected and are considered a potential risk for others.9,10 Cryptic HCV infection in dialysis units occurs frequently and, when unrecognized, can become a major factor for contracting the virus due to medical procedures. In previous studies, it has been shown that the use of the glycosylated E2 envelope protein in its native form, to capture antibodies to HCV-E2 envelope protein, in patients using a Western blot (Wb) and an enzyme-linked immunosorbent assay (ELISA)-based assay, apparently allows serological detection of the vast majority of cryptic HCV-positive sera.9,10 This study was conducted in order to assess the efficiency of our newly developed optical immunosensor to detect antibodies to the HCV E2-antigen in sera of patients chronically infected with cryptic HCV, which are not responsive to core, NS3 and NS5 proteins of the virus, but are HCV-RNA positive. We tested the occurrence of anti-E2 antibodies in the sera of patients from the liver clinic and from dialysis patients previously screened by standard methods (Western blot, ELISA) and RT-PCR. The presence of anti-E2 antibodies was found to be correlated with the presence of HCV RNA. Antibodies to E2 envelope protein in those samples could be detected by standard methods (ELISA and Western blot) in as many as 25-30% of the sera with cryptic HCV infection.9,10 Using a newly developed optical microbiosensor, we demonstrate an increase in the detection limit of anti-E2 antibodies to 50% of the sera with cryptic HCV infection. In addition, we should note that while our method is using only one antigen of the virus (E2 envelope protein) the standard test uses three different antigens of HCV (core, NS3 and NS5) to detect antibodies to HCV (anti-HCV antibodies third generation, Axym, (10) Hanuka, N.; Sikuler, E.; Tovbin, D.; Mostoslavsky, M.; Hausman, M.; Orgel, M.; Yaari, A.; Shemer-Avni Y. J. Viral Hepatitis 2002, 9, 141-145.
1772
Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
Abbott). Despite this fact, we were able to detect all HCV positive sera and reduce by 50% the false negative sera. EXPERIMENTAL SECTION Reagents. The pyrrole monomer, functionalized with a photoreactive benzophenone group (Figure 1B) was prepared as previously described by the esterification of the 3-benzoylbenzoic acid with the 1-(3-hydroxypropyl)pyrrole using the carbodimide method.8 The pyrrolic benzophenone was purified by silica gel chromatography and characterized by HNMR and IR spectra as well as by EI mass spectrum. Bovine serum albumin (BSA, A4503, fraction V) and polyoxyethylene-sorbitan monolaurate (Tween 20, P7949) were purchased from Sigma. Hepatitis C purified E2 envelope protein (genotype 1b, 384-661aa) was produced in baculovirus (XTL, Rehovot, Israel). The marker antibody was a horseradish peroxidase-labeled polyclonal goat anti-human IgG (Amersham) used at a concentration of 0.08 µg/mL. Luminescence measurements were carried out using the Western Blot Chemiluminescence Reagent Plus kit from NENTM Life Science Products (NEL105, containing enhanced luminol reagent and oxidizing reagent). Acetonitrile 97% (Rathburn) and lithium perchlorate 99.9% (Fluka) were used as received. Study Population. The study was carried out in the Soroka University Medical Center, Beer-Sheva, Israel. The study population consisted of three groups: healthy volunteerssnegative controls (anti-HCV-/RNA-, 7 samples), and 21 patients attending the Dialysis Unit and patients attending the Liver Clinic (see Table 3 and Figure 9 for details), with various liver diseases. The Helsinki committee of the Soroka Academic Medical Center approved the study, and informed consent was obtained from all persons involved.The sera were collected and tested immediately for anti-HCV antibodies (anti-core, NS3 and NS5) and anti-E2 antibodies by standard methods.9,10 In parallel, aliquots were frozen at -70 °C for HCV-RNA monitoring. Principle of ELISA and Western Blot To Detect Antibodies to E2 Envelope Protein. Anti-HCV E2 envelope immunoglobulin
was measured by ELISA and Western blot (see refs 9 and 10 for detailed protocol). Optical Fiber Tip Preparation. All fibers used were SFS400/ 440B silica fibers (Fiberguide Industries) with an original numerical aperture of 0.22. Their core was 400 µm in diameter (refractive index of 1.457 at 633 nm), and these are surrounded by a 40-µm silica cladding (refractive index of 1.44 at 633 nm), followed by a 150-µm-thick silicon buffer, and finally a 210-µm-thick black Tefzel jacket. The typical length of any single fiber used in our experiments was 20 cm. The black Tefzel jacket was stripped away mechanically using a fiber stripping tool (Micro-Strip, from MicroElectronics Inc), leaving a 1-cm declad optical fiber tip. To remove the silicon cladding the fiber-optic tips were dipped a few minutes in sulfuric acid 93.1% (v/v), then rinsed with distilled water, and afterward rinsed with acetone and left to dry. The silica cladding was removed by dipping the fiber-optic tips in 5% (v/v) HF during 10 h with stirring. Next, the optical fibers were rinsed with distilled water and methanol and finally blow dried. Indium Tin Oxide Sputtering. The depositions were carried out in an rf-sputtering Varian system. The sputtering target was a 4-in.-diameter circular disk of hot pressed powder 99.999% purity ITO (90 wt % In2O3 + 10 wt % SnO2), made by Testbourne Ltd. (Hampshire, U.K.). The sputtering chamber was evacuated to a pressure lower than 10-6 Torr. Argon and oxygen gases were introduced into the sputtering chamber with partial pressures of 2 × 10-3 and 1 × 10-4 Torr, respectively. The rf supply was then switched on and stabilized to a power of 400 W. The fiber tips were exposed to the sputter beam for 1-11/2 h. Therefore, the fiber core end face (important area where electropolymerization must occur) and at least one side of the cylindrical surface (for contacting and voltage supply) were targeted by the ITO. Deposition of Pyrrole-Benzophenone onto the Optical Fiber Surface by Electropolymerization. All electrochemical experiments were performed with a Princeton Applied Research model 173 equipped with a model 179 digital coulometer and a model 175 universal programmer in conjunction with a Sefram TGM 164 X-Y/t recorder. The electropolymerization of pyrrolebenzophenone and the characterization of the resulting modified electrode were run at room temperature in a conventional threeelectrode cell. A 10 mM Ag/Ag+ in CH3CN electrode was used as a reference electrode in acetonitrile electrolyte. The working electrodes were the ITO-modified optical fibers, an electrical contact being established with a platinum wire (AVOCADOFrance; diameter 50 µm) above the acetonitrile solution. The polypyrrole-benzophenone films were prepared by controlled potential oxidation (0.85 V for 5 min) of the monomer (2 mM) in CH3CN + 0.1 M LiCl4. FT-IR Characterization of the Poly(pyrrole-benzophenone)-Coated Optical Fiber Tip. Infrared reflection-absorption spectra for surfaces of the bare, ITO-coated, and poly(pyrrolebenzophenone)-coated optical fibers were obtained using a Bruker Equinox 55 FT-IR spectrometer with a Bruker IRRAS device PME 37. The spectra were averaged over 128 scans at a resolution of 4 cm-1. IR spectra are shown by absorbance units [absorbance is defined as -log(R/Ro)], where R and Ro are the reflectance of the sample and of the background, respectively. Energy-Dispersive Spectrometry (EDS). EDS analyses were made with a scanning electron microscope (SEM) (JEOK JSM-
Figure 2. Biosensor scheme describing the various steps involved in the immunoassay using ITO-poly(pyrrole-benzophenone)-coated optical fibers for the detection of anti-HCV in human sera samples.
7400F). Micrographs were taken at 129-eV resolution of detector (Si) with an exposure time of 100 s. The specimens were analyzed using energy-dispersive X-ray spectrometry microanalysis analytical equipment installed in the scanning electron microscope. Atomic Force Microscopy (AFM). AFM measurements were performed at ambient conditions using a Digital Instrument Dimension 3100 mounted on an active antivibration table. A 100µm scanner was used. Microfabricated silicon oxide NSC11\50 type ultrasharp cantilevers with an integrated pyramidal tip were used. The 512 × 512 pixel images were taken in tapping mode with a scan size of up to 5 µm at a scan rate of 1 Hz. Photoimmobilization of the Antigen onto the Optical Fiber. To produce the desired activation radiation, we used a 1000-W Xe lamp mount (Oriel 6271) connected to a light condenser (Oriel 66021). The light was reflected through a dichroic mirror (Oriel 66226). The large spectrum radiation was then condensed into the monochromator (Oriel 77250) using the appropriate lens (Oriel, plano-convex lenses). Thereafter, a 345nm wavelength light output, with a light intensity of 80 mW cm-2, was projected for 7 min into the far end of the optical fiber that was previously coated at its near end with poly(pyrrole-benzophenone) and soaked in 120 µL of a 4 µg/mL HCV E2 envelope protein solution during irradiation. The light intensity was measured by an Ophir Optronics power meter Nova reader, PD300-UV. The excited polymerized benzophenone radicals could then bind to neighboring E2 envelope protein in the solution (4 µg/mL). Thereafter, the optical fibers were profusely rinsed and washed with phosphate buffer saline (pH 7.4) or sodium carbonate buffer (pH 9.4). Immunoassay Rationale and Design. The steps carried out for the construction of the biosensor are visualized in Figure 2. The blocking treatment was carried out with BSA (5% w/v) for 1 h in order to prevent the putative nonspecific binding of antibodies onto the glass fiber tips. Then, the fibers were rinsed and washed twice with PBS-Tween for 5 min, before being incubated in 120 µL of analyte solution for 20 min. The analyte consisted of blind human sera isolated from patients in Soroka hospital. Except for Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
1773
the positive and negative controls that were used for the construction of a calibration curve, the samples in Table 3 were tested blindly. The serological markers of HCV (anti-core, NS3, NS5, and E2 antibodies) obtained by standard methods and RT-PCR results were not disclosed to the student who performed the test with the microbiosensor until after the study was finished. The test was performed in a biological safety hood (level P2). The procedure followed the following pattern; E2 envelope proteinconjugated fibers were incubated for 20 min in 120 µL of several diluted serum solutions at dilutions ranging from 1:1 024 000 up to 1:500. Sets of three replicas were prepared for each concentration measured, and a series of fibers were set aside with the purpose of using them as blanks. Thereafter, the fibers were rinsed and washed once with PBST for 15 min and then three times for 5 min. Subsequently, the optical fiber tips were dipped into 120 µL of a solution containing the secondary antibody, horseradish peroxidase-labeled goat anti-human IgG immunoglobulin for 20 min, and finally, rinsed and washed once with PBST for 15 min, and then three times for 5 min. Chemiluminescence Measurements. The instrument setup for the chemiluminescence measurements has been previously described.11,12 A Hamamatsu HC135-01 photomultiplier tube sensor module was used for chemiluminescence measurements, combining the sensitivity of a photomultiplier tube with the intelligence of a microcontroller. The detector is optimized to the blue light region and includes a 21-mm-diameter active area convenient to gather light radiation without any optical focusing elements for luminescence measurements. The instrument setup was placed in a light-tight box. The far end of the fiber was held by a fiber holder (FPH-DJ, Newport) and placed into an adjustable single-fiber mount (77837, Oriel). To prevent damage to the photon-counting unit by environmental light, a manual shutter (71430, Oriel) was placed in front of the detector. To move the slide shutter, a workshop-made lever was placed outside the box. To receive and treat data, a specific driver was developed using LabView (version 3.1, National Instruments Corp.), which allowed monitoring of the chemiluminescence signal and data handling in real time. The immunosensor optical fibers were placed in a 400-µL sample tube containing the combined oxidizing reagent and enhanced luminol reagent solutions (NEN chemiluminescence reagent kit) in a 1:1 (v/v) ratio. The chemiluminescence readings were integrated for 1 s, and each measurement was obtained by taking a mean value of photon counts during 10 s. The responses of blank controls were also checked by the same procedure. RESULTS AND DISCUSSION Deposition and Characterization of the Photoreactive Poly(pyrrole-benzophenone) Layer. The immobilization of the photoactivable pyrrole derivative onto ITO-coated optic fibers was carried out via the electrochemical formation of a polymeric coating. The electropolymerization of the pyrrolic monomer (2 mM in CH3CN + 0.1 M LiClO4) was achieved by controlled potential oxidation at 0.85V for 5 min (Figure 3). The current (11) Polyak, B.; Bassis, E.; Novodvorets, A.; Belkin, S.; Marks, R. S. Water Sci. Technol. 2000, 42, 305-311. (12) Polyak, B.; Bassis, E.; Novodvorets, A.; Belkin, S.; Marks, R. S. Sens. Actuators, B 2001, 74, 18-26.
1774 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
Figure 3. Current variation during the electrpolymerization process of an ITO-coated optical fiber carried out in 2 mM pyrrole-benzophenone in 0.1 M LiClO4/CH3CN. Applied potential, 0.85 V (vs Ag/Ag+).
evolution can be divided into two main regions.13-15 The first region (from 30 to 60 s) shows a slight increase in current due to pyrrole oxidation and oligomer formation. In the second region (from 60 to 175 s), the increase of current indicates the growth of a conductive polymer backbone, followed by a current stabilization, and finally by a decrease of current intensity due to the growth of the polymer layer thickness at the ITO-coated optic fiber end face and tip. The characterization of the deposited polymer film was checked by cyclic voltammetry (Figure 4A). In the positive region, a reversible peak system is recorded at + 0.31 V, reflecting the electroactivity of the polypyrrolic skeleton. The observed reversible oxidation wave is in good agreement with the reported E1/2 values for the poly(pyrrole-benzophenone),8 thus corroborating the electrodeposition of a poly(pyrrole-benzophenone) coating onto the optical fiber. The comparison with the cyclic voltammogram obtained for an ITO-coated optical fiber, not functionalized with poly(pyrrolebenzophenone), clearly shows the absence of the oxidation peak of the polypyrrolic film (Figure 4B). The amount of benzophenone derivative immobilized on the optical fiber surface (apparent surface coverage of electropolymerized benzophenone derivative) can be evaluated from the charge integrated under the polypyrrole oxidation wave corresponding to the electroactivity of the conducting polypyrrole backbone. The following surface coverage, Γ) 1.95 × 10-9 mol cm-2, was estimated by taking into account that 0.33 electron molecule-1 is involved in the electrooxidation of the polypyrrolic chain.16 Morphology and Characterization of Poly(pyrrole-benzophenone). The morphology of the polymer surface is of high interest in the design and the optimization of biosensing devices. To obtain information about the morphology of the polymeric coatings, SEM and AFM measurements were performed on polypyrrolic films synthesized electrochemically on ITO-coated optical fibers and compared to the polymer films obtained on flat surfaces.17-20 (13) Willicut, R. J.; McCarley, R. T. Langmuir 1995, 11, 296-301. (14) Mekhalif, Z.; Lang, P.; Garnier, F.; Re´gis, A. J. Chim. Phys. 1995, 92, 831834. (15) Cossement, D.; Plumier, F.; Delhalle, J.; Hevesi, L.; Mekhalif Z. Synth. Met. 2003, 138, 529-536. (16) Deronzier, A.; Moutet, J. C. Acc. Chem. Res. 1989, 22, 249-255.
Figure 4. Cyclic voltammograms recorded (A) on an ITO-modified optical fiber functionalized by an electrogenerated poly(pyrrolebenzophenone) film and (B) on an ITO-modified optical fiber in 0.1 M LiClO4/CH3CN; potentials measured vs Ag/Ag+ 10 mM, scan rate, 100 mV s-1.
Figure 5. (A, D) SEM micrograph of poly(pyrrole-benzophenone) film formed on an ITO-coated fiber optic. (B, E) shows the border between the polymerized pyrrole-benzophenone on ITO-coated fiber and bare ITO (at the top and from the right bare ITO, at the bottom and left polymerized film on ITO-coated fiber). (C, F) ITO-coated fiber optic. (G) Film thickness estimation. (H, I) Polymerized tip of an optical fiber with the observed polymer cauliflower-like arrangement.
Figure 5 shows SEM micrographs of poly(pyrrole-benzophenone) films exhibiting homogeneous, thin (0.26 µm, Figure 5G) and compact structures, their surface being characterized by a (17) Martins, J. I.; Bazzaoui, M.; Reis, T. C.; Bazzaoui, E. A.; Martins, L. Synth. Met. 2002, 129, 221-228. (18) Bazzaoui, M.;.Bazzaoui, E. A.; Martins, L.; Martins, J. I. Synth. Met. 2002, 128, 103-114. (19) Miles, M. J.; Smith, W. T.; Shapiro, J. S. Polymer 2002, 41, 3349-3356. (20) Shapiro, J. S.; Smith, W. T.; MacRae, C. Polymer 1995, 36, 1133-1140.
cauliflower-like arrangement17,18 constituted by microspherical grains of approximately 0.1-1 µm in length (Figure 5A). In addition, the formation of nucleating centers is observed in Figure 5D, with large morphological units forming, wrinkles19,20 from 5 µm to hundreds of micrometers. The quality of the polymer was illustrated by the absence of cracks or detachment of the film. Panels B and E in Figure 5 show the border between the polymerized pyrrole-benzophenone on the ITO-coated fiber as Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
1775
Figure 6. Elemental analysis of the poly(pyrrole-benzophenone) polymer (points 1-4 on fiber-optic surface).
Figure 7. Typical two- and three-dimensional perspective view of 5 µm × 5 µm AFM images of conductive poly(pyrrole-benzophenone) film on an ITO-coated fiber optic. The estimated rms and Ra are 110 and 89 nm, respectively. Rms is the root-mean-square roughness, which is defined as the standard deviation of the distribution of surface heights; Ra is the arithmetic average height parameter and is defined as the average absolute deviation of the roughness irregularities from the mean line over one sampling length.22
Table 1. Absorption Peaks Recorded by Reflection FT-IR Spectrometry of the Poly(pyrrole-benzophenone) Functionalized ITO-Coated Optical Fibers peak position/cm-1
assignment
673 630 1750 1500-1550 740-770
ν b (C-H) benzene ν b (C-H) monosub benzene deriv ketone stretch ring vibration (aromatic hydrocarbon) γ-CH or β-ring modes of pyrrole ring
well as the bare ITO. The polymerized tip of the optical fiber can be observed in Figure 5G-I with a polymer cauliflower-like arrangement. An elemental analysis shown in Figure 6 indicates that the combined polymer covers the surface of the fiber (points 1-4) and has the basic compositional structure elements N, C, O. The polymer layer is thin enough to also detect via elemental analysis both layers of indium and silica that reside beneath them. Figure 7 shows a typical two- and three-dimensional perspective view of a 5 µm × 5 µm AFM image of the functionalized polypyrrole films. As previously observed for SEM images, the polymeric films present a very homogeneous surface composed of small clusters or granular structures. An additional confirmation of the occurrence of a layer of poly(pyrrole-benzophenone) on the ITO-coated optical fiber tips was obtained by recording the infrared reflection spectra on the functionalized fiber tip surface. Because of the small quantities of poly(pyrrole-benzophenone) deposited by electropolymeriza1776
Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
tion, the absorption intensities were low but sufficient enough to identify some absorption bands that are characteristic of the benzophenone and pyrrole moieties that compose the polymer. The absorption peaks and probable assignments are shown in Table 1. Immobilization of the E2 Envelope Protein. The optical fibers coated with the respective poly(pyrrole-benzophenone) were then soaked in diluted E2 envelope protein solutions (4 µg/ mL) and irradiated with UV light. Benzophenone and most of its derivatives absorb a photon at around 345 nm resulting in the promotion of one electron from a nonbonding sp2-like n-orbital on oxygen to an antibonding π*-orbital of the carbonyl group. The actual electron-deficient oxygen n-orbital becomes electrophilic and therefore interacts with weak C-H σ-bonds, resulting in a hydrogen abstraction to complete the half-filled n-orbital. When amines or similar heteroatoms are in the vicinity of the excited carbonyl, an electron-transfer step may occur, followed by proton abstraction from an adjacent group. In biological systems, the most effective H-donors include backbone C-H bonds in amino acids; thus, methylene groups of amino acid side chains are good candidates providing abstractable hydrogens through the general mechanism shown in Figure 1A.5 Immunosensor Behavior. The detection of anti-HCV E2 antibodies was achieved using an indirect fiber-optic immunoassay (Figure 2) after the electropolymerized deposition of a poly(pyrrole-benzophenone) thin film onto an ITO-coated optical fiber tip and photoimmobilization of viral E2 envelope protein. The analyte (anti-E2 antibodies), putatively located in blind clinical
Figure 8. (A) Calibration curve obtained using the same immunoassay procedure rationale in the detection of anti-E2 antibodies as in ELISA but using the ITO-poly(pyrrole-benzophenone)-coated optical fibers. Titers of human sera from 1:500 to 1:1 024 000 were tested in triplicate on individual fibers. (B) The linear range of the calibration curve obtained for titer 1:64 000 and lower. The curve was fitted according to the equation y ) A + B(x), where is the human sera (anti-E2 antibodies) dilution value and y is the chemiluminescence response. The correlation coefficient obtained was R2 ) 0.988.
human sera, was searched for its recognition of E2 envelope protein epitopes. The subsequent binding of a marker peroxidaselabeled IgG anti-human antibody allowed us to then carry out a chemiluminescence reaction by adding an enhanced luminol solution.21 The light, emitted as a side reaction, is transduced by the optical fiber to the measurement instrument setup. The standard curve for the human anti-E2 envelope protein antibodies sandwich immunoassay obtained by collecting data in triplicate at antibody titers ranging from 1:500 to 1:1 024 000. Figure 8A shows a typical behavior for the standard calibration
curve with an exponential growth, as seen from the curve fit, which results in a linear range shown in Figure 8B. The linear range curve fit was carried out using an equation of the form y ) A + B(x), where x is the dilution of human serum obtained from patients and y is the corresponding response signal obtained. The standard curve most useful in the quantization of titers was found to be from 1:64 000 and lower (Figure 8B), showing in this range an acceptable square correlation coefficient, R2 ) 0.988 and a satisfactory sensitivity of 2 × 108 photocounts/s (determined within the linear concentration range of the biosensor as the slope, B, of the calibration curve). At higher concentrations, the curve levels off with a response saturation observed from titers 1:4000 and up. As small inaccuracies in the measurement occur at such low dilutions, measurements may lead to large errors in prediction, so such samples would have to be diluted for accurate quantization. The detection limit of the immunosensor, defined as the amount (or concentration) of the analyte that gives a response (YDL), that is significantly different (three standard deviations (SDBR) from the background analysis of the analysis that is itself obtained from negative sera (YBR). Therefore, the lower detection limit (YLDL) was calculated using eqs 1 and 2.
YLDL > 3(SDBR) + YBR
(1)
YDL ) 3(SDBR) + YBR
(2)
The background signal recorded for the blank (in absence of the analyte, or using a known negative serum) and its calculated standard deviation, allowed us to reach a limit of quantization for a dilution as high as 1:1 024 000. To check the influence of possible nonspecific interactions on the overall chemiluminescent responses, a set of reference experiments were carried out. Table 2 shows the conditions used for the set of different experiments that were carried out. The steps shown in the table correspond to those illustrated in Figure 2.
Table 2. Reference Experiments Carried Out in Order To Determine the Possible Influence of Nonspecific Effects on the Obtained Responsea step exp
ITO sputtering
poly(pyrrole-benzophenone) electropolymerization
E2 protein photoattachment
positive sera
anti-human HRP
normalized responseb
1 2 3 4
+ + + +
+ + +
+ + +
+ + + -
+ + + +
1.000 0.0760 0.0198 0.0296
a The (+) and (-) signs indicate the steps accomplished or avoided respectively for each experiment. The response was recorded in all the experiments for human sera (anti-E2 protein antibodies) titer at 1:500. b Calculated as a ratio between the response obtained from experiments 2 to 4, and the one obtained from experiment 1.
Table 3. Sera of Patients, Defined Serologically by Western Blot and Screened by RT-PCR for Viral Load in the Blood, Tested by the Immunosensor for the Presence of Anti-E2 Antibodiesa serological state of HCV patients
total tested
anti-E2+ Western blot
anti-E2+ immunosensor
Western blot detection %
immunosensor detection %
anti-HCV+/RNA+ anti-HCV-/RNA+b anti-HCV-/RNA-
13 8 7
9 2 0
13 4 0
69 25 0
100 50 0
a The (+) and (-) signs indicate the occurrence or absence respectively of the antibodies and viral RNA. b Patients with cryptic HCV. Anti HCV: anti-core, NS5, and NS3 antibodies.
Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
1777
The results in Table 2 show that the responses from experiments 2-4 (obtained in absence of the polymer, E2 envelope protein, or human sera, respectively) are all relatively negligible. This study focused on developing a highly sensitive assay that will assist in improving the diagnosis of cryptic HCV. The assay is based on a biosensor that can detect anti-E2 antibodies. RTPCR, ELISA, and Western blot were performed by Soroka hospital.9,10 The study described in refs 9 and 10 demonstrates that Western blot is more sensitive than ELISA (detection limit used there for Western blot method was sera dilution of 1:4000). However, both systems were less sensitive for anti-E2 antibody detection than the biosensor. We were able to detect anti-E2 antibodies in all sera positive for anti-HCV (anti-core, NS3, and NS5) and HCV RNA (100% sensitivity), while anti-E2 antibodies, tested by Western blot, were detected in only 69% of anti-HCV (anti-core, NS3, and NS5) positive samples (Table 3, line 1). In addition, standard test for anti-HCV antibodies requires the use of the three cited antigens (core, NS3, and NS5) because of its lack of sensitivity, while thanks to the high sensitivity of the opticfiber immunosensor, using E2 envelope protein alone was sufficient to detect all RNA positive patients (Table 3, line 1). None of the negative sera reacted (100% specificity) positive in the immunosensor test (Table 3, line 3). Furthermore, 50% of the cryptic HCV sera, positive HCV-RNA that were negative for anticore, NS3, and NS5, were positive in our test to the anti-E2 antibodies (Table 3, line 2), while Western blot detected only 2530% of these sera.9,10 The improvements in sensitivity could possibly be attributed to the distinctive biosensor design using electrochemical and photochemical immobilization methods for capturing E2 envelope protein on the end of the optical fiber tip, as well as the general configuration of the sensor bringing the luminescence in the vicinity of the fiber-optic end face for greater light coupling efficiency. In addition, the results indicate that the prevalence of antibodies to E2-antigen in sera depends on and correlates with the presence of HCV-RNA (Table 3, lines 1 and 2). Negative controls that were anti-HCV-/RNA- were negative in both the immunosensor and the Western blot (Table 3, line 3). To demonstrate further the practical application of the immunosensor, we chose three clinical cases of HCV positive patients from the dialysis unit, which were tested during HCV infection and serum conversion (Figure 9). From patient 1, five samples were obtained at various times of infection. The immunosensor could detect anti-E2 antibodies in the first serum that tested positive for viral RNA, while Western blot could detect anti-E2 antibodies only in the next sample, when the immune response was stronger (antibody concentration was higher). If this patient would be a blood donor at the time of the first test, the presence of HCV would be missed by the current standard methods used by blood banks; thus, a person who is negative for HCV would become HCV infected. Patient 2 showed RNA presence and antiHCV (anti-core, NS3, and NS5 antibodies) in all three tests while detection of anti-E2 antibody was shown only by the immunosensor due to the higher sensitivity of the method. Patient 3 showed a late response; anti-E2 antibodies were detected only by the immunosensor and only after viral clearance. (21) Freeman, T. M.; Seltz, W. R. Anal. Chem. 1978, 50, 1242-1246. (22) Sree, U.; Yamamoto, Y.; Deore, B.; Shiigi, H.; Nagaoka, T. Synth. Met. 2002, 13, 161-165.
1778 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
Figure 9. Results of serum tests for three different patients. Test numbers indicate the order of the blood donation from patients. Anti - HCV: anti core, NS5, and NS3. The (+) and (-) signs indicate the occurrence or absence respectively of the antibodies and viral RNA.
Our results suggest that the detection of anti-E2 antibodies may enhance significantly HCV serological standard testing; especially among patients on dialysis who were diagnosed as HCV negative by standard immunological tests, especially since the immunosensor showed higher sensitivity for the detection of antiE2 antibodies than Western blot and ELISA. The sensitivity of the immunosensor might prove, in the future, once the technology would be applied to an automatic instrument, an important clinical tool for HCV screening. If transformed into a an easy to use procedure, it might replace RT-PCR for monitoring HCV during interferon treatment, as well as HCV-RNA and antigen detection in blood banks. CONCLUSIONS We have demonstrated herein that one can electropolymerize the conductive surface of a fiber optic with a photosensitive layer, which will enable a one-step immobilization procedure of proteinaceous biospecific entities such as immunoglobulins. The model analyte anti-E2 immunoglobulin was detected at a low titer of 1:1 024 000. Our results suggest that the detection of anti-E2 antibodies by the microbiosensor may enhance significantly HCV serological standard testing; especially among patients on dialysis, who were diagnosed as HCV negative by standard immunological tests. ACKNOWLEDGMENT We thank Yakov Yuzhelevski from the Department of Physics at (BGU) for his technical assistance in running the ITO rf sputtering instrumentation, Roxana Golan from the SPM facility of the R. Stadler Minerva center for mesoscale macromolecular
engineering at (BGU) that carried out the AFM measurement, and Luba Burlaka from the Center for Meso, Nanoscale Science and Technology, Department of Chemical Engineering (BGU) that carried out the SEM measurement and Alexandra Petrosov for technical assistance. We thank XTL (Rehovot, Israel) for the E2 antigen. The Western blot and the ELISA results form part of the Ph,D, thesis of N.H., while the rest of this work forms part of the
M.Sc. thesis of T.K. R.S.M. and H.S.C. thank NATO for providing the Collaboration Linkage Award (CLG 981086).
Received for review September 26, 2004. Accepted December 28, 2004. AC048569W
Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
1779
