Ultrasensitive detection of biomolecules using functionalized multi-walled carbon nanotubes

Sensors and Actuators B 124 (2007) 161–166

Ultrasensitive detection of biomolecules using functionalized multi-walled carbon nanotubes PingAn Hu a,∗ , Takashi Tanii b , Guo-Jun Zhang a , Takumi Hosaka b , Iwao Ohdomari a,b a

b

Nanotechnology Research Center, Waseda University, Tokyo, Japan Department of Electronical Engineering and Bioscience, School of Science and Engineering, Waseda University, Tokyo, Japan Received 28 September 2006; received in revised form 20 November 2006; accepted 7 December 2006 Available online 20 December 2006

Abstract An ultrasensitive and facile bioassay approach is developed using biofunctional multi-walled carbon nanotubes (MWNTs). The biofunctional MWNTs are exploited to detect trace protein in the homogeneous solution using fluorescent detection. The effect of target concentration on the fluorescent intensity is systematically studied, revealing that the detected signal is strengthened with increase in target concentration. The homogenous target-capture process performed on the biofunctional MWNTs enhances the detection sensitivity largely. The detection limit of 100 pg/ml is achieved through this bioassay. © 2007 Elsevier B.V. All rights reserved. Keywords: Detection; Biomolecules; Carbon nanotubes

1. Introduction The biomolecular detection is of central importance to the diagnosis and treatment of diseases [1,2]. Recently, great efforts have been made to develop new technologies for highly sensitive bioanalysis. Such biological detection commonly relies on hybridization or antigen–antibody interactions [1–4]. Among them, the most widely used technique is patterning of biological macromolecules onto solid surfaces in the form of microarrays (“chips”). The target-capture process is performed on the substrates (e.g., silicon wafer, glass slide) via biological recognition. A signal probe (fluorescent dye molecules are used usually) is utilized to signal such biological interactions. Sensitivity is a central factor for bioanalytical technique. To achieve a high sensitivity, a large amount of research has focused on signal amplification by utilizing various nanomaterials (e.g., quantum dots, metal nanoparticles) as strong and photostable signal probes [5–7]. Although these approaches have made condsiderable progress in biomolecular detection, they still have several drawbacks: (1) these techniques involve a complex procedure ∗ Corresponding author. Present address: Centre for Advanced Photonics, Department of Engineering, Cambridge University, Cambridge, United Kingdom. Fax: +81 3 5286 9076. E-mail address: [email protected] (P. Hu).

0925-4005/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2006.12.025

for immobilization of the biomolecules on the flat substrate; (2) since the target-catching procedure is carried out on the flat surface of microarray or titer plate, such heterogeneous procedure increases assay time and decreases the sensitivity due to the slow target-binding kinetics; (3) some nanomaterials (e.g., nanoparticles of Ag, CdS, CdSe) used for signal amplification are sensitive to air, which causes reduced reproducibility. So far, only a few researches are about overcoming some of the above issues. For example, it is expected that the sensitivity can be enhanced by means of homogeneous target-binding kinetics, therefore, Mirkin and co-workers have developed a novel assay for ultrasensitive detection of protein via homogeneous target-binding process by using soluble magnetic microparticles [8]. Carbon nanotubes (CNTs) are well known for their high surface ratio and high chemical stability, indicating that the CNTs are ideal scaffolds for immobilization of biomolecules. Biofunctional MWNTs are soluble, which makes them suitable for development of homogeneous bioassay. In addition, very recently, with purpose to create miniature biological electronics and optical devices using CNTs as building blocks, much research interest has been focused on intersection of biology and carbon nanotubes as well as biological functionalizations of carbon nanotubes, interaction between biomolecular and CNTs and carbon nanotube biosensors [9–20]. Previous

162

P. Hu et al. / Sensors and Actuators B 124 (2007) 161–166

approaches of biological analysis using carbon nanotubes usually exploit electric detection and require complex procedure for operation and fabrication of the device. For example, CNTs electrodes or CNT field effect transistors are reported to detect the interaction of specific DNA pair protein-receptor in solution [16–20]. In this paper, we develop a facile CNT-based assay with homogeneous target-binding procedure by using fluorescent detection. The biofunctional MWNTs are allowed to efficiently bind the corresponding receptor in homogeneous solution. The fast target-binding kinetics increases the sensitivity largely. The detection concentration is down to 100 pg/ml. 2. Experimental 2.1. Materials The multi-walled carbon nanotubes (MWNTs) with a diameter of 10–60 nm and a length range of 10–200 ␮m, were purchased from Aldrich–Sigma Company. Rabbit immunoglobulin (IgG) was purchased from Sigama–Aldrich, anti-rabbit immunoglobulin labeled Cy5 (antiIgG) was purchased from Zymed Laboratories Inc., amino-PEO-biotin (biotin) was purchased from Molecular Probes and avidin FITC conjugate was purchased from Pierce. All reagents and solvents were obtained from commercial suppliers and used without further purification. The raw MWNTs are suspended in a concentrated sulfuric acid, nitric acid mixture (3:1, v/v) and sonicated for 2–5 h. To remove sulfuric acid and nitric acid, the oxidized CNTs are filtered through 1 ␮m pore sized hydrophilic polytetrafluo-

roethylene (PTFE) membrane and washed with de-ionized water until no acid is detected, followed by drying in vacuum at 70 ◦ C for 24 h. 2.2. Synthesis of MWNT-biotin and MWNT-IgG The process for covalently attaching amino-PEO-biotin or IgG to the nanotube surface is detailed as follows: in the first step, 1.0 mg of carboxylic MWNTs are suspended in 5.0 ml of DMF by sonicating the mixture for a short time. Then, 2 ml of a 4 mg 1-ethyl-3-(3-dimethyl aminopropyl)carbodiimide (EDAC) dimethylformamide (DMF) solution and 2 ml of a 15 mg Nhydroxy succinimide (NHS) DMF solution are added to the above suspension and mixed. The mixture is continually sonicated for 2 h. The suspension is then filtered through hydrophilic PTFE membrane, washed successively with ethanol and water to remove excess EDAC, NHS and byproduct urea. In the second step, 0.1 mg of the sufosuccimidyl CNTs PBS solution and 0.1 mg of amino-PEO-biotin or IgG are mixed in 10 ml M 2-(4-Morpholino)ethanesulfonic acid (MES) buffer solution (pH ∼ 6.1). The mixture shielded from light is stirred for 2 h. To remove unbound biomolecules completely, the resulting mixture is transferred to a membrane tube for dialysis against freshwater for 3 days. 2.3. X-ray photoelectron spectroscopy (XPS) XPS measurements were carried out on JPS90-MX (JEOL) system. Equipped with an angle-resolved hemispherical elec-

Fig. 1. Schematic drawing for bioassay using the functional MWNTs.

P. Hu et al. / Sensors and Actuators B 124 (2007) 161–166

tron energy analyzer and a base system pressure of 1–10−9 Torr. All measurements were performed using the Al KR line with a photon-energy of 1486.6 eV, the biofunctional nanotubes were characterized by XPS after treatment at UHV to remove any residual physically adsorbed materials. 2.4. TGA experiment The TGA experiment was conducted with a heating rate of 10 ◦ C/min in the presence of air. The onset of the weight loss is at 240 ◦ C/min, with bulk of the weight loss occurring before 400 ◦ C/min. 2.5. Biological detection using as-prepared biofunctional MWNTs The 50 ␮l of functional MWNT (2 ␮g/ml calculated by the weight of MWNT) solution is incubated with 50 ␮l target solution (avidin labeled with FITC and anti-rabbit IgG labeled with Cy5) with various concentrations ranging from 1 pg/ml to 0.2 ␮g/ml for about 10 min at room temperature. In order to confirm the successful detection, a drop of resulting solution was dispersed on the clean glass slide for observation under microscopy fluorescent.

163

introduced onto the surfaces of MWNTs by a mixed acid treatment as reported elsewhere [11]. IgG or amino-PEO2-biotin is covalently attached to the oxidized MWNT via a process of diimide-activated amidation in presence of EDAC and NHS (the as-produced MWNT conjugates are briefed as MWNT-biotin and MWNT-IgG, respectively). Thermal gravimetric analysis (TGA) scanned at a slow-temperature scanning rate, because most of the organic functional groups are evaporated before the onset of carbon nanotube weight loss (>400 ◦ C in air). The amount of immobilized biomolecules (amino-PEO2-biotin or IgG) on MWNTs is determined by thermal gravimetric analysis (TGA). The analysis reveals that the sidewall coverage of MWNTs by immobilized biomolecules (biotin, IgG) is about 43%. As well known, the raw MWNTs are insoluble or poorly dispersive in water and most of organic solvents [21]. After the biomolecules are covalently attached to the sidewall, the functional MWNTs have fine solubility in physiological solutions and easily formed into homogeneous black solution [22]. X-ray

2.6. Atomic force microscopy experiments Atomic force microscopy (AFM) images for characterization of biological recognition on MWNTs by using a Nanoscope III Multimode scanning probe microscope in the tapping mode. Typically, the fresh solution of MWNT-biotin/avidin (or MWNT-IgG/antiIgG) was prepared using the procedure outlined above. A drop of MWNT-biotin/avidin (or MWNT-IgG/antiIgG) solution (100 ng/ml) was then placed onto a freshly cleaved glass surface. The sample was then air-dried and analyzed by AFM. 3. Results and discussion Our method for biomolecular detection using functionalized MWNT is depicted in Fig. 1. Briefly, carboxylic groups are

Fig. 2. X-ray photoelectron spectrum of N peak: (a) oxidized MWNTs, (b) MWNT-biotin and (c) MWNT-IgG.

Fig. 3. Fluorescent characterizations of bioassay using the functional MWNT: the trace biomelecules in sample are efficiently absorbed onto functional MWNT via biological recognition in homogenous solution. The fluorescent linear structures observed with fluorescent microscopy imply that the trace biomolecules have been successfully detected: (a) MWNT-biotin/avidin labeled with FITC and (b) MWNT-IgG/antiIgG labeled with Cy5.

164

P. Hu et al. / Sensors and Actuators B 124 (2007) 161–166

photoelectron spectrum (XPS) is utilized to confirm chemical immobilization of biomolecules onto MWNTs. As shown in Fig. 2, N (1s) signal is not detected for oxidized MWNTs a strong peak of N (1s) signal is obtained from MWNT-IgG, while a weak N (1s) signal is detectable from MWNT-biotin. The intensity of N peaks corresponds to the nitrogen components of IgG (N: ∼23%) and amino-PEO-biotin (N: 15%). The N (1s) spectrum (Fig. 2) shows a peak with a binding energy of 400.4 eV, which matches the previous studies that amides and amines both have binding energies in the range of 399.5–400.5 eV [23,24]. To show the excellent utility of the biofunctionalized MWNTs in biodetection, we employ ligand–receptor pairs of biotin/avidin labeled with FITC and IgG/antiIgG labeled with Cy5 as the research systems. MWNT-biotin (or MWNTs-IgG) is used to detect corresponding trace receptor in the mixed solution. When the biofunctional MWNT is incubated in a mixed solution for about 10 min, the trace protein in the solution (avidin labeled with FITC or antiIgG labeled with Cy5) is bound to the surface of the functional MWNT (MWNT-biotin or MWNT-IgG) via biomolecular recognition in homogenous solution and form another MWNT conjugate (MWNT-biotin/avidin or MWNT-IgG/antiIgG). Some research has shown that fluorescent microscopy is a powerful tool for visualization of

carbon nanotubes in solution by using flurorophores [25], fluorescent polymer [26] or semiconducting nanocrystals [27]. So, the results of detection in our study can be confirmed and characterized by using fluorescent microscopy. As shown in Fig. 3, the fluorescent images of MWNT-biotin/avidin labeled with FITC (blue linear structures in Fig. 3a) and MWNTIgG/antiIgG labeled with Cy5 (red linear structures in Fig. 3b) can be observed. These fluorescent linear structures observed with fluorescent microscopy reveals that the target is successfully detected. To confirm the fluorescent image of MWNTs caused by the interaction of biological recognition on surfaces rather than physical absorption, a control experiment with physical mixing MWNT and antiIgG labeled with Cy5 (or avidin labeled with FITC) is carried out. Line-like structure described above is not observed from such a sample by using fluorescent microscopy. The whole detection process can be finished in 10 min. Compared to the electric detection, this approach just involves fluorescence microscopy, so this bioanalytic technique is very fast and easily operated. The absence of quenching via energy transfer from absorbed molecules to CNTs has been demonstrated [28]. It is difficult to image the fluorescent species that directly immobilized on the sidewalls of CNTs under fluorescence microscope because

Fig. 4. AFM characterization of bioassay using MWNT-biotin: the heave-like structures represent the pair of biotin/avidin absorbed onto the MWNT. Line scans shown in the pictures reveal that a region on the MWNT that is attached with biotin/avidin pair has a height of 19.4 nm, while a region that does not contain biotin/avidin has a height of 14.1 nm. The difference of 5.3 nm is the height of immobilized biotin/avidin.

P. Hu et al. / Sensors and Actuators B 124 (2007) 161–166

165

Fig. 5. AFM characterization of bioassay using MWNT-IgG: (a) the globular structures represent the pair of IgG/antiIgG. Line scans shown in pictures reveal that a region on the MWNT that attached with IgG/antiIgG pair has a height of 32.4 nm, while a region that does not contain IgG/antiIgG has a height of 18.6 nm. The difference of 13.8 nm is the height of immobilized IgG/antiIgG.

of short lifetime and less quantum. In our case, the fluorescent species are attached to the surface of CNTs via a pair of IgG/antiIgG (or biotin/avidin), which acts as the insulation layer of energy transfer and thus allows for the successful fluorescent imaging. Previous research has shown that AFM is an efficient tool for quantitative measurement of the overall three-dimensional shape of the proteins immobilized on carbon nanotubes [29]. In order to provide additional evidence for successful trace detection, structural analysis of the immobilized biomolecules is performed with AFM characterization. As shown in Fig. 4, the heave-like structures represent the pair of biotin/avidin absorbed onto the MWNT. Line scans shown in the pictures reveal that a region on the MWNT that is attached with biotin/avidin pair has a height of 19.4 nm, while a region that does not contain biotin/avidin has a height of 14.1 nm. The difference of 5.3 nm is the height of immobilized biotin/avidin. All the detected height of biomolecules immobilized on the MWNT varies from 4 to 7 nm, which fits the molecular dimension of biotin (0.52 nm × l.00 nm × 2.10 nm) [30] and avidin (5.60 nm × 5.00 nm × 0.40 nm) [31], proving that avidin has been bound to MWNT-biotin. As for AFM characterization of assay using MWNT-IgG, the globular structures on the MWNT represent IgG/antiIgG molecules (Fig. 5). A region on the

MWNT that attached with IgG/antiIgG pair has a height of 32.4 nm, and a region that does not contain IgG/antiIgG has a height of 18.6 nm. The difference of 13.8 nm is the height of immobilized IgG/antiIgG. The height of the globular structures on MWNT determined by AFM has a range of 9–30 nm and matches the dimensions of IgG and antiIgG [32–34], which indicates that the antiIgG is attached onto the MWNT-IgG in solution via biological recognition. Based on our finding that the intensity of detecting signal is mainly determined by the target concentration, the effect of target concentration on the fluorescent intensity is systematically studied. The target concentration is varied from 1 pg/ml to 20 ng/ml. Fig. 6 shows a curve of the fluorescent intensity response to the various target concentration. There is a sizable increase in signal intensity from 5 to 10 ng/ml. Above 10 ng/ml, the curve levels off. Fluorescent images obtained from the target concentration of 10, 5 and 1 ng/ml, reveal directly that the signal intensity in the images is enhanced as the target concentration is increased. In order to explore the detection limit, various samples with a concentration in the range from 10 pg/ml to 1 ng/ml are detected by this assay. There is no fluorescent signal below 100 pg/ml and very weak fluorescent signal can be detected above 100 pg/ml (shown in the small figure inserted in Fig. 6). So the detection limit is expected to be about 100 pg/ml.

166

P. Hu et al. / Sensors and Actuators B 124 (2007) 161–166

Fig. 6. The signal intensity of detection response to the different target concentration.

4. Conclusion An ultrasensitive and facile bioassay is developed using biofunctional MWNTs. The biofunctional MWNTs are exploited to detect trace protein in the homogeneous solution using fluorescent detection. The process of homogenous target-capture kinetics increases the detection sensitivity largely. The detection limit of around 100 pg/ml is achieved using this bioassay. Acknowledgements The authors gratefully acknowledge financial supports from Japan Society for the Promotion of Science (JSPS), and a Grant-in-Aid for Center of Excellence (COE) Research from the Ministry of Education, Culture, Sports, Science and Technology. References [1] L. He, M.D. Musick, S.R. Nicewarner, F.G. Salinas, S.J. Benkovic, M.J. Natan, C.D. Keating, J. Am. Chem. Soc. 122 (2000) 9071–9077. [2] J. Wang, D.K. Xu, R. Polsky, J. Am. Chem. Soc. 124 (2002) 4208. [3] X. Zhao, D.R. Tapec, W. Tan, J. Am. Chem. Soc. 125 (2003) 11474. [4] W.C.W. Chan, S.M. Me, Science 281 (1998) 2016–2018. [5] Y.W.C. Cao, R.C. Jin, C.A. Mirkin, Science 297 (2002) 1536–1540. [6] D.J. Maxwel, J.R. Taylor, S.M. Me, J. Am. Chem. Soc. 124 (2002) 9606–9612. [7] Z. Li, R.C. Jin, C.A. Mirkin, R.L. Letsinger, Nucleic Acids Res. 30 (2002) 1558–1562. [8] S.J. Park, T.A. Taton, C.A. Mirkin, Science 295 (2002) 1503–1506. [9] K. Jiang, L.S. Schadler, R.W. Siegel, X. Zhang, H. Zhang, M. Terrones, J. Mater. Chem. 14 (2004) 37. [10] W. Huang, S. Taylor, K. Fu, Y. Lin, D. Zhang, T.W. Hanks, A.M. Rao, Y.P. Sun, Nano Lett. 2 (2002) 311. [11] D. Pantarotto, C.D. Partidos, R. Graff, J. Hoebeke, J.-P. Briand, M. Prato, A. Bianco, J. Am. Chem. Soc. 125 (2003) 6160. [12] S.E. Baker, W. Cai, T.L. Lasseter, K.P. Weidkamp, R.J. Harriers, Nano Lett. 2 (2002) 1413. [13] R.J. Chen, H.C. Choi, S. Bangsaruntip, E.Y. Enilmez, X. Tang, Q. Wang, Y.-L. Chang, H. Dai, J. Am. Chem. Soc. 126 (2004) 1563. [14] K. Bradley, M. Briman, A. Star, G. Grulner, Nano Lett. 4 (2004) 253. [15] P.P. Pompa, L. Blasi, L. Longo, R. Cingolani, G. Ciccarella, G. Vasapollo, R. Rinaldi, A. Rizzello, C. Storelli, M. Maffia, Phys. Rev. E 67 (2003) 41902.

[16] A. Star, J.-C.P. Gabriel, K. Bradleyand, G. Grulne, Nano Lett. 3 (2003) 459. [17] K. Besteman, J.-O. Lee, F.G.M. Wiertz, H.A. Heering, C. Dekker, Nano Lett. 3 (2003) 727. [18] Y. Lin, F. Lu, Y. Tu, Z. Ren, Nano Lett. 4 (2004) 192. [19] R.J. Chen, S. Bangsaruntip, K.A. Drouvalakis, N.W.S. Kam, M. Shim, Y. Li, W. Kim, P.J. Utz, H. Dai, Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 4984. [20] J. Wang, G. Liu, M.R. Jan, J. Am. Chem. Soc. 126 (2004) 3010. [21] T.A. Taton, C.A. Mirkin, R.L. Letsinger, Science 289 (2000) 1757. [22] S.M. Santra, P. Zhang, K. Wang, R. Tapec, W. Tan, Anal. Chem. 73 (2001) 4988–4993. [23] K. Jiang, A. Eitan, L.S. Schadler, P.M. Ajayan, R.W. Siegel, N. Grobert, M. Mayne, M. Reyes-Reyes, H. Terrones, M. Terrones, Nano Lett. 3 (2003) 275–277. [24] Y. Qin, L. Liu, J. Shi, W. Wu, J. Zhang, Z. Guo, Y. Li, D. Zhu, Chem. Mater. 15 (2003) 3256. [25] R. Prakash, S. Ishburn, R. Superfine, R.E. Cheney, M.R. Falvoa, Appl. Phys. Lett. 83 (2003) 1219. [26] K. Otobe, H. Nakao, H. Hayashi, F. Nihey, M. Yudasaka, S. Iijima, Nano Lett. 2 (2002) 1157. [27] S. Chaudhary, J. Kim, H. Singh, K.V.M. Ozkan, Nano Lett. 4 (2004) 2415. [28] L.W. Qu, R.B. Martin, W.J. Huang, K.F. Fu, D. Zweifel, Y. Lin, Y.P. Sun, C.E. Bunker, B.A. Harruff, J.R. Gord, L.F. Allard, J. Chem. Phys. 117 (2002) 8089. [29] T.L. Ascah, K.M.R. Kallury, C.A. Szafranski, S.D. Corman, F. Liu, J. Liq. Chromatogr. Relat. Technol. 19 (1996) 3049. [30] Z. Lin, T. Strother, W. Cai, X. Cao, L.M. Smith, R.J. Hamers, Langmiur 18 (2002) 788. [31] S.S. Karajanagi, A.A. Vertegel, R.S. Kane, J.S. Dordick, Langmiur 20 (2004) 11594. [32] G.T. DeTitta, J.W. Edmonds, W. Stallings, J. Donohue, J. Am. Chem. Soc. 98 (1976) 1920. [33] Z. Liu, K. Otsuka, S. Terabe, M. Motokawa, N. Tanaka, Electrophoresis 23 (2002) 297. [34] L. Harris, E. Skaletsky, A. McPherson, J. Mol. Biol. 275 (1998) 861.

Biographies PingAn Hu received his PhD in Physical Chemistry in 2004 from Institute of Chemistry, Chinese Academy of Science. During 2004–2006, he worked at Waseda University in Japan as a JSPS postdoctoral fellow. Currently, he is working at the Department of Engineering, University of Cambridge, and focus on the research regarding nanoscale biosensor. Takashi Tanii received his PhD in Engineering in 2002 from School of Science and Engineering, Waseda University in Japan. He is now serving as a visiting Associate Professor at Department of Electronical Engineering and Bioscience, Waseda University. His research interests are in nano-electronics. Guo-Jun Zhang received his PhD in Medicine in 1999 from Tongji Medical College, Huazhong University of Science and Technology in China. During 1999–2001, he had worked as a postdoc at Wuhan University in China. Then, he moved to Institute for Physical High Technology Jena in Germany as a postdoc for another one year. Subsequently, he worked at Waseda University in Japan as a JSPS postdoctoral fellow. Now he is working at Institute of Microelectronics, Singapore as a researcher. Takumi Hosaka is a Graduate student at Department of Electronical Engineering, Waseda University. Iwao Ohdomari received his PhD in Engineering in 1972 from School of Science and Engineering, Waseda University in Japan. He became a full professor in 1979 at Waseda University. Since then, he has been focusing on the research on semiconductor electronics, radiation effects induced by single ion and its application, as well as fabrication, characterization and application of nano-structures.