Bond Rupture of Biomolecular Interactions by Resonant Quartz Crystal

Oct 26, 2007 - Industrial Research Limited, Crown Research Institutes, P.O. Box 31-310, Lower Hutt, New Zealand, and Institute of Technology and Engin...
0 downloads 10 Views 283KB Size
Anal. Chem. 2007, 79, 9039-9044

Bond Rupture of Biomolecular Interactions by Resonant Quartz Crystal Yong J. Yuan,*,† Matthew J. van der Werff,†,‡ Huoguang Chen,† Evan R. Hirst,†,‡ Wei L. Xu,‡ and John E. Bronlund‡

Industrial Research Limited, Crown Research Institutes, P.O. Box 31-310, Lower Hutt, New Zealand, and Institute of Technology and Engineering, Massey University, Private Bag 102 904, Auckland, New Zealand

Significant progress has been achieved in understanding affinity-based diagnostics, which use the highly specific “lock and key” recognition and binding between biomolecules, for example, an antibody and its antigen. These are the most specific of analytical tests. One of the most challenging issues is to distinguish between true binding and ever-present nonspecific binding in which more loosely bound proteinaceous material gives false results in conventional affinity methods. We have used bondrupture scanning to eliminate nonspecific binding by introducing energy mechanically through displacement of a resonant quartz crystal. The removal of the analyte was recorded with a simple all-electronic detection system quickly providing confirmation of the presence of the target molecule. The system can measure the resonant frequency difference and detect noise signals, respectively, due to mass changes and bond breaks between biotinylated self-assembled monolayer (SAM) and streptavidin-coated polystyrene microspheres (SCPM). Both static and dynamic scanning modes can reveal previously unrecognized desorption of streptavidin-coated polystyrene microspheres. An established framework of bondrupture scanning is a promising diagnostic tool for investigating the specific and nonspecific interactions by measuring the characteristic level of mechanical energy required to break the bond. The quartz crystal microbalance (QCM) was first proposed by Sauerbrey1 as a mass-sensitive device; the resonant frequency of a QCM is very sensitive to mass changes on its surface. Sauerbrey’s equation for the change in resonant frequency of the QCM is shown in eq 1. The change in resonant frequency depends on the first harmonic (f0), the properties of the quartz (cj66 and Fq), and most relevant to this exercise, the surface mass (m), and the area that the mass is spread over (A). The electrodes are oscillated by the piezoelectric effect which causes mechanic vibrations in response to changing voltage. The mechanical system has resonant modes of operation, the frequency of which are highly sensitive to mass changes on the electrode. With the use of this relationship, it is possible to detect a change in mass in * Corresponding author. E-mail: [email protected]. † Crown Research Institutes. ‡ Massey University. (1) Sauerbrey, G. Z. Phys. 1959, 155, 206-222. 10.1021/ac701717m CCC: $37.00 Published on Web 10/26/2007

© 2007 American Chemical Society

the order of nanograms by measuring the change in resonant frequency.

∆f ) -

2f02 Axjc66Fq

∆m ) -Sf∆m

(1)

The Sauerbrey equation presented a linear relationship between the frequency shift and the attached mass, i.e., the mass sensitivity Sf is constant. Although it was first developed for oscillation in air and only applies to rigid masses attached to the crystal, QCM measurements have also been performed in liquid, in which case a viscosity-related decrease in the resonant frequency was observed.2 By coating different biological layers on the QCM surface, the mass sensitivity of the elastic layer is obviously different from that derived from the Sauerbrey equation. In addition, the contact condition is also affected by the size of the attached biomolecules and the properties of the host environment. The bond rupture of strong covalent C-C bonds in brushlike macromolecules was induced by tension along the polymer backbone,3 due to adsorption physical interactions of the side chains with the substrate. The notion of hearing the rupture of antibody-antigen bonds was proposed by increasing the voltage applied to the crystal gradually.4-7 However, upon increasing the driving amplitude after adsorption of a significant amount of 200 nm streptavidin-coated polystyrene spheres, no desorption was observed.8 Motivated by these conflicting results, the purpose of this research is to distinguish the specific and nonspecific bonds by increasing the amplitude applied to the crystal gradually and monitoring the change in mass.9 Changes in resonant frequency of immobilized biotin-polyethylene glycol (PEG)-amine layer and sequentially interacted streptavidin molecules as results of (2) Kanazawa, K. K.; Gordon, J. G. Anal. Chem. 1985, 57, 1770-1771. (3) Sheiko, S. S.; Sun, F. C.; Randall, A.; Shirvanyants, D.; Rubinstein, M.; Lee, H.; Matyjaszewski, K. Nature 2006, 440, 191-194. (4) Dultsev, F. N.; Ostanin, V. P.; Klenerman, D. Langmuir 2000, 16, 50365040. (5) Dultsev, F. N.; Speight, R. E.; Fiorini, M. T.; Blackburn, J. M.; Abell, C.; Ostanin, V. P.; Klenerman, D. Anal. Chem. 2001, 73, 3935-3939. (6) Cooper, M. A.; Dultsev, F. N.; Minson, T.; Ostanin, V. P.; Abell, C.; Klenerman, D. Nat. Biotechnol. 2001, 19, 833-837. (7) Cooper, M. A. Meas. Sci. Technol. 2003, 14, 1888-1893. (8) Edvardsson, M.; Rodahl, M.; Kasemo, B.; Ho¨o ¨k, F. Anal. Chem. 2005, 77, 4918-4926. (9) van der Werff, M. J.; Yuan, Y. J.; Hirst, E. R.; Xu, W. L.; Chen, H.; Bronlund, J. E. IEEE Sens. J. 2007, 7, 762-769.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007 9039

Scheme 1. (a) Two-Dimensional Structure of Biotin-(PEG)8-NH2 on a Gold Surface and (b) Architectural Topography of the Resonator-SAM-SCMS System

the respective bindings of gold-amine and biotin-streptavidin have been directly measured. The mass changes measured on binding biotin-PEG-amine agree closely with anticipated in situ bond-rupture signals measured by the dynamic mode scanning. The results obtained from this approach are in agreement with photographic data for the morphological changes expected. This allows a spectrum of bond forces to be generated. Here, we present a novel system simultaneously measuring the loss of mass at the first harmonic and noise at the third overtone to demonstrate the bond rupture. It is done by continuously increasing the voltage applied to a resonant quartz crystal to induce every rupture of the biotin-streptavidin bonds one after another. EXPERIMENTAL SECTION Chemicals and Self-Assembled Monolayer. The 6.098 µm streptavidin-coated polystyrene microspheres (SCPM) were from Polysciences, Inc. (Warrington, PA), and biotin-(PEG)8-NH2 was from Polypure AS, Oslo, Norway. A 10 MHz AT-cut QCM from International Crystal Manufacturing (Oklahoma City, OK) was cleaned in a piranha solution (1:3 v/v H2O2/H2SO4) to remove any organics from the gold surface and then immersed in 5 mM biotin-(PEG)8-NH2 for 24 h to form a biotinylated self-assembled monolayer (SAM). The presence of surface-bound alkylamine molecules on gold was demonstrated,10 similar to that observed for alkanethiols. Single-stranded DNA immobilized on gold surfaces11 showed primary amine chemisorption more strongly than the secondary amine of thymine deoxyribonucleotide. The amine-gold link was proposed to be through the N lone pair, by adding steric bulk to the nitrogen center.12 Therefore, the watersoluble biotin-(PEG)8-NH2 was used to form a SAM on gold surfaces as illustrated in Scheme 1a. Binding energy of amine (10) Kumar, A.; Mandal, S.; Selvakannan, P. R.; Pasricha, R.; Mandale, A. B.; Sastry, M. Langmuir 2003, 19, 6277-6282. (11) Kimura-Suda, H.; Petrovykh, D. Y.; Tarlov, M. J.; Whitman, L. J. J. Am. Chem. Soc. 2003, 125, 9014-9015. (12) Venkataraman, L.; Klare, J. E.; Tam, I. W.; Nuckolls, C.; Hybertsen, M. S.; Steigerwald, M. L. Nano Lett. 2006, 6, 458-462.

9040

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

adsorption on four different surface types of gold was reported to be between 10.8 and 16.8 kcal/mol.13 The interaction of amine and gold can be described by a weak covalent bond, since the covalent bond energy is between 15 and 170 kcal/mol. Biomolecular Interactions and Rupture. The association of streptavidin with biotin is the strongest of known noncovalent protein ligand interactions (Ka ≈ 2.5 × 1013 M-1), due to a hydrogen bond network.14 The specific binding between biotin and streptavidin was achieved by adding a drop of SCPM disperse, which was diluted from 10 µL of 1.37% solid-latex-based microspheres in 5 mL of water. The microspheres distribution on the QCM gold surface was photographed as shown in Figure 1a. The bond rupture of SCPM was conducted as increasing the amplitude applied from 0.01 to 20 V for 10 s at each step and then measuring the QCM resonant frequency at 0.04 V (except two points of 0.01 and 0.02 V amplitudes measured at 0.01 V). After bond-rupture scanning conducted by a static mode, the gold surface was examined, and the morphology is shown in Figure 1b. While in a dynamic bond-rupture scanning mode at 20 V/min, the concentration of microspheres used was diluted from 10 µL of 1.37% solidlatex-based particles in 500 µL of water to achieve high coverage of SCPM on the surface. Although the disassociation of enthalpic energy is approximately 32 kcal/mol in water,15 the energy of the biotin-streptavidin crystal structure by empirical force field was calculated as 8.86 kcal/mol.16 We conducted the bond-rupture experiments in ambient nitrogen. This prevents QCM from damping in liquid and further achieving the high Q factor of QCM measurements of a defined bond rupture between biotin and streptavidin. System. An integrated digital processing solution toward a smart sensor is reported in ref 9. It consists mainly of the following parts: a digital transmitter that excites the crystal, a digital receiver (13) Pong, B.-K.; Lee, J.-Y.; Trout, B. L. Langmuir 2005, 21, 11599-11603. (14) Grubmu ¨ ller, H.; Heymann, B.; Tavan, P. Science 1996, 271, 997-999. (15) Chilkoti, A.; Stayton, P. S. J. Am Chem. Soc. 1995, 117, 10622-10628. (16) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.; Belew, R. K.; Olson, A. J. J. Comput. Chem. 1998, 19, 1639-1662.

Figure 1. Photos (a) and (b) were taken, respectively, before and after bond-rupture scanning.

Scheme 2. System Overview

that measures piezoelectric responses, such as amplification and impedance matching for maximum power transfer to a resonator, and a digital signal processor (DSP) that communicates with the transmitter and receiver and is reprogrammable via a PC. The system, as illustrated in Scheme 2, is a versatile predominantly digital system that includes a DSP which communicates with scientific software “Prospa” on a PC. A resonator is driven and its resonant frequency change is captured as an indication of bond rupture. The driving circuitry uses a variable amplifier whose gain is ultimately controlled from a graphical user interface on the PC. A digital transceiver system is designed and fabricated that integrates transmitter, receiver, amplification, and impedance matching. A printed circuit board (PCB) is built to accommodate all electronic devices and connected to the PC and a resonator. There are a variety of analogous methods possible to detect the

resonant frequency of the quartz crystal or the noise of bond rupture, for example, impulse excitation (IE), oscillator circuit (OC), and network analysis (NA). IE requires exciting the crystal with an impulse and measuring the frequency and decay time of the resulting oscillations; it has been used to measure the crystal characteristics such as the Q factor or the resonant frequency. OCs have been used to drive the crystal to good effect; however, their operation in heavily damped environments is difficult due to the increased impedance. NA provides a wealth of information that can be determined about the QCM, but it does not facilitate bond rupture. The system is designed with the intention of overcoming the limitations of other instrumental technologies by characterizing the interference of the weaker bonds, and its functionalities are implemented in DSP software. This provides the state-of-the-art Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9041

Figure 2. Resonance-frequency profile of bond-rupture scanning of specific and nonspecific biomolecular interactions.

of a resonator operation. It not only in situ monitors changes of the first-harmonic resonance frequency, but also measures signal levels of the third overtone simultaneously, due to the bond rupture on a resonant surface. RESULTS AND DISCUSSION Static Mode Bond Rupture. Changes of resonant frequency were first investigated discretely by the static mode bond rupture. The amplitude was first applied at a fundamental frequency for a certain period of time, and then the resonant frequency was measured at 0.04 V, before the next amplitude taking it in turns. From eq 1, the resonant frequency of QCM oscillation depends on surface mass, which decreases and increases as SCPM bind to the surface and rupture with binding. As shown in Figure 2, some loosely bound SCPM without biotin association were shaken off immediately before 0.2 V of amplitude applied, resulting in an abrupt increase in resonant frequency. The resonant frequency stays reasonable stable until 3.5 V, and then the bond-rupture phenomenon occurs from 3.5 to 12 V by dramatic increases of resonant frequency with voltage amplitude. As indicated in the resonant frequency response, the increases due to the nonspecific and specific bonds were, respectively, 70 and 890 Hz. Figure 2 demonstrates how nonspecific bonds can be distinguished by increasing the voltage applied to the crystal resonator gradually and monitoring the change in mass. To obtain a comprehensive understanding of the role of the amplitude applied to eliminate the nonspecific bonds between biotinylated SAM and SCPM, the detailed resonance-frequency changes are required. As shown in Figure 3, the frequency increased 120 Hz from 0.01 to 0.1 V of applied voltage amplitude, and then there is a 120 Hz plateau from 0.1 to 0.2 V of amplitude. This indicates SCPM without biotin association are shaken off easily at low amplitude. Interestingly, the frequency decreased with increasing amplitude applied from 0.3 to 1.5 V. A significant fraction (≈40%) of frequency decreased from 120 to 70 Hz in the time frame of the observation. The resonant frequency is reduced due to a reflecting shear wave from SCPM. Interfacial Slip Phenomenon. A key to the investigation of a bond-rupture system able to quantify the binding energy level of an analyte is understanding how significant shear deformation (as shown in Figure 3) is induced in the immobilized layer causing elastic energy to be stored and dissipated. 9042

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

Figure 3. Resonance-frequency profile of bond-rupture scanning of nonspecific biomolecular interactions.

Scheme 3. Mechanical Illustration Representing a Local Spring-Damper-Mass Modela

a m is the particle mass, and m is the mass of the particle on 1 2 the quartz surface. uvh and uqh are the lateral displacement of the contact particles and quartz contact surface, respectively, where δ is the distance between uvh and uqh.

The contact properties and the size of a molecule were elucidated by the slip parameter17 on the interface, which reflects the vibration transmission between the quartz crystal top surface and liquid particles. The interfacial slip phenomena were described by the transfer function between displacements of the sensor surface and that of particles at the bottom of the attached layer in contact with the sensor. When an electric field is applied across the quartz crystal between two metal electrodes, the mechanical displacement in the lateral direction is generated. The slip friction force is proportional to the relative velocity between the contact surfaces,18 i.e., the sensor surface of one of the metal electrodes and the attached layer in contact with the sensor. In this case, the interfacial slip phenomena occur in the interface between the gold electrode surface and the biotinylated SAM and SCPM layers. The layer is not infinite, and there is a reflecting shear wave from the particles. As illustrated in Scheme 3, the interfacial slip between the quartz surface and particles is modeled as a local mass and interaction element between the masses (i.e., a particle and a surface). The interaction between SCPM and the surface via biotinylated SAM is represented as a spring and damper with complex parameters G*, a force when the shear displacement occurs. The distance δ between velocities uvh and uqh is supposed (17) Lu, F.; Lee, H. P.; Lim, S. P. Smart Mater. Struct. 2003, 12, 881-888. (18) McHale, G.; Lucklum, R.; Newton, M. I.; Cowen, J. A. J. Appl. Phys. 2000, 88, 7304-7312.

to be a dominant factor and depends upon the mechanical displacement in the lateral direction generated by an electric field across the quartz crystal between two metal electrodes. The biomolecular vibration is not synchronous with the driving resonator surface, due to the spring force plus the damping force, and varies across the biotinylated SAM and SCPM. The surface topography of the current biotinylated SAM and SCPM layer is illustrated in Scheme 1 architecturally. A biotin-PEG-amine molecule binds onto the resonator gold surface as shown in Scheme 1a. Furthermore, a biotinylated SAM layer bridges SCPM to the resonator via the streptavidin-biotin complex, detailed in Scheme 1b. It experimentally resulted in the resonant frequency decreasing by approximately 50 Hz as the applied amplitude was increased from 0.2 to 1.5 V. On the other hand, a shear modulus19 is represented by G* ) G′ + jG′′, where G′ is the storage modulus and G′′ the loss modulus. A layer immobilized on the thickness-shear mode (TSM) resonator surface is subject to an oscillatory driving force at a TSM resonator/layer interface. Typically, a layer is bonded to the TSM resonator surface sufficiently well that the base of the layer moves synchronously with the resonator surface. As shown in Scheme 1b, the upper portions of the layer, especially SCPM at a microscale, may lag behind the driving surface. Significant shear deformation is induced in the layer causing elastic energy to be stored and dissipated. In this regime, layer displacement is not synchronous with the driving surface and varies across its thickness. Although it is difficult to directly observe the dynamic behavior of the immobilized SAM and SCPM, which varies across its thickness, its influence on the TSM resonator electrical characteristics can be more readily determined. In addition, in the regime where the layer is deformed, measurement of the electrical characteristics of a layer-coated resonator can be interpreted to obtain its shear storage and loss moduli. Dynamic Mode Bond Rupture. A continually swept voltage is applied to verify that interactions of kinetic association/ disassociation without discrete steps are, in fact, capable of screening libraries of biomolecules. The dynamic mode bondrupture scanning provides an in situ approach to validate immobilized biotin interaction with streptavidin on a solid surface. It is not only a mass-sensing technique, employed for the detailed study of a model protein system, but can simultaneously measure some electronic signals (pulse or noise) generated, due to the molecules shaken off from surface distorting a resonant harmonic. The rapid increase of the resonant frequency before 0.2 V represented in a black curve in Figure 4 suggested that some loosely bound SCPM were shaken off immediately. Similar to the static mode as mentioned above, the resonator surface appears fairly stable from 1.5 to 3 V approximately. For higher voltages, the resonant frequency increased dramatically due to breaking off specific bonds of biotin and streptavidin. Figure 4 demonstrates how nonspecific bonds can be eliminated by increasing the voltage applied to the crystal gradually and monitoring the change in mass. This allows a dynamic spectrum of bond forces to be generated. An invaluable addition to this is to “listen” for the bond rupture as indicated in noise levels of the third overtone by a red curve. Remarkably, both noise signal zones (amplitudes from 0 (19) Martin, S. J.; Frye, G. C. Proc. IEEE Ultrason. Symp. 1991, 1, 393-398.

Figure 4. Profiles of specific and nonspecific biomolecular interactions by a dynamic bond-rupture scanning mode. The black line is the resonance-frequency change of the first-harmonic frequency, while the red line is the signal level of the third overtone. Table 1. Comparison of Driving Amplitude by Different Measurements of Bond Rupture amplitude range of bond rupture/V measurement of bond rupture

nonspecific

specific

dynamic mode frequency dynamic mode noise static mode frequency

0 to ∼1.4 0 to ∼0.8 0 to ∼0.2

3.0 to ∼4.5 3.5 to ∼5.0 3.5 to ∼12

to 0.8 V and from 3.5 to 5 V) correspond precisely to the nonspecific and specific bonds, respectively, closely matching the resonant frequency spectrum. The detailed amplitude range of frequency and noise measurements are summarized in Table 1. In contrast, the amplitude span of a dynamic mode from 3 to 4.5 V was very narrow, compared to the span of a static mode from 3.5 to 12 V as observed in the specific bond breakage of biotin-streptavidin. This means that the dynamic mode scanning is much more efficient way to break a bond, due to effective energy accumulation. The results presented here demonstrate that the in situ bond-rupture spectrum is a powerful tool for distinguishing the specific and nonspecific bonds of biomolecular interactions. The ability to detect frequency changes and noise signals within a minute, under a dynamic mode, also provides a powerful tool for medical diagnostics. Advantages of the method used are that the sensing signal processing can be reprogrammed and thus automated in real time, that versatile experiments can be done using the same PCB, and that the digital signals can be processed in software. As indicated in a dynamic mode of bond rupture, an average of the thirdovertone magnitude is recorded during driving cycles. This allows monitoring the resonator for bond-rupture noise around the thirdharmonic frequency. As evidenced in Figure 4, SCPM shaken off from the biotinylated SAM surface results in both frequency changes and noise signals being generated. Detecting bond rupture by both monitoring frequency change and bond-rupture noise detection allows specific and nonspecific bonds to be distinguished by their relative bond strength. CONCLUSIONS The study of solution-phase interactions between a SAM and proteins is of intense interest, especially to point-of-care medical Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9043

diagnostics. A mass-sensing technique, QCM, has been employed for the detailed study of a model protein system, namely, immobilized biotin interactions with streptavidin on a solid gold surface. Changes in resonant frequency of specific and nonspecific interactions have been directly measured. We present a newly

ACKNOWLEDGMENT This work was performed under the auspices of the Foundation for Research, Science and Technology of New Zealand by Industrial Research Ltd. under New Economy Research Fund C08X0408.

developed system to aid in experimenting with a resonant quartz

Received for review August 14, 2007. Accepted September 24, 2007.

crystal and bond rupture.

AC701717M

9044

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007