1
In-situ quantitative analysis of a prostate-specific antigen (PSA) using a nanomechanical PZT cantilever

2
We report on a novel technique of resonant frequency shift measurement based on a nanomechanical cantilever with a PZT actuating layer for label-free detection of a prostate-specific antigen (PSA) in a liquid environment.

3
The nanomechanical PZT thin film cantilever is composed of SiO2/Ta/Pt/PZT/Pt/SiO2 on a SiNx supporting layer for simultaneous self-exciting and sensing; it was fabricated using a standard MEMS (micro electromechanical system) process.

4
The specific binding characteristics of the PSA antigen to its antibody, which is immobilized with Calixcrown self-assembled monolayers (SAMs) on a gold surface deposited on a cantilever, are determined to a high sensitivity.

5
For the bioassay in a liquid environment, a liquid test cell with a 20 µl volume reaction chamber has been fabricated, using a bonding technique between poly(dimethyl siloxane) (PDMS) bilayers.

6
An observed trend of resonant frequency change with respect to time could be explained by the binding kinetics due to the Langmuir isotherm and diffusion and by the effects of a small volume reaction chamber.

7
In the saturated regimes, the resonant frequency of the cantilever increased with increase of the PSA concentration in the reaction chamber, showing that the trend of the resonance frequency change was similar to that of the fluorescence results.

8
The saturated resonance frequency shift of the cantilever was proportional to the PSA antigen concentration of analyte solution.

Introduction

9
Over the past few decades, a considerable number of studies have been made on high-throughput identification and the quantitative analysis of biological molecules for the diagnosis, monitoring, and evaluation of complex diseases, such as cancer, as well as for basic study.1

10
In recent years, there has been growing interest in bio- and immuno-sensors that use micro-electro-mechanical systems (MEMS) and nano-electro-mechanical systems (NEMS).

11
Microfabricated cantilevers have been proposed as nanomechanical transducers for various sensing applications, such as biological, physical and chemical sensors to detect the presence of specific compounds with selectivity and quantity.2–4

12
Nanomechanical cantilevers have a high sensitivity and a minimal detectable mass density, in comparison with electromechanical devices, such as the quartz crystal microbalance (QCM), the flexural plates wave (FPW) device, and the surface acoustic wave (SAW) device.5

13
In general, the cantilevers can detect specific compounds in both dynamic and static modes.

14
The dynamic mode is used to detect the resonance frequency shift caused by cantilever mechanical properties, such as mass and stiffness in the binding of the target molecules to the receptors on the cantilever surfaces.

15
The static mode is a method to measure the differences of surface stress in adsorption between the two opposite cantilever surfaces, in which one side (top) is functionalized with receptor molecules.6

16
Most of the current nanomechanical cantilever based sensors are focused on the deflection, i.e. the static mode, arising from a change in surface properties due to antigen–antibody binding using the optical detection method.

17
However, the optical stress measuring system has some limitations, such as a narrow dynamic range and a parasitic deflection.

18
These limitations in detection using the static mode are a serious problem for the development of feasible cantilever sensors with high sensitivity and reproducibility.

19
As a result, the dynamic detection method is considered more advantageous, as it is insensitive to the drift of deflection signal due to the parasitic deflection.

20
According to previous studies, the dynamic mode has been studied for the sensing of changes in medium viscoelasticity and for the monitoring of resonance frequency shift caused by mass change.6,7

21
Most of the relevant experiments have been performed in gaseous environments, where the surface stress produced after the adsorption of the target substance also induced a detectable shift of the resonance frequency with an external actuator.8,9

22
However, there has been no report yet on the dynamic mode of a biosensor for protein–protein interaction detection that uses a cantilever included in the PZT actuating layer, without any external driving actuator in the liquid environment.

23
In this paper, we report on the dynamic mode detection of a prostate specific antigen–antibody (PSA) binding in a liquid environment using a nanomechanical PZT cantilever.

24
The PSA is a clinical marker of prostate cancer, which is currently the most prevalent form of cancer in men and the second leading cause of male cancer death in the United States.

25
This PSA, which is detectable in serum, has proved to be an extremely useful marker for the early detection of prostate cancer and for monitoring disease progression and the effects of treatment.10–12

26
In this study, the dynamic detection of PSA antigen–antibody binding is accomplished by monitoring the resonant frequency shift of the nanomechanical cantilever that is caused by the specific binding between the PSA antigen–antibody onto the functionalized gold surface of cantilever.

27
It is demonstrated that the nanomechanical cantilevers allow for a rapid direct detection of PSA with a high sensitivity, but without the need of fluorescent, radioactive molecule labeling and any external actuator for driving.

28
The resonance shift responses of the cantilever were a function of the PSA antigen–antibody reaction time and the PSA concentration.

Theory

29
First resonant frequency can be expressed approximately as where k is the spring constant and m* is the effective mass of cantilever.

30
This equation is based on Newton's 2nd law and Hooke's law.

31
The relevant literature provides theoretical analysis in properties of cantilever materials and the results of such theoretical calculations are important parameters in mechanical evaluation of cantilevers.13–17

32
Lee et al. monitored the specific binding of the C-reactive protein (CRP) antigen–antibody with a PZT thin film cantilever by means of an electrical dynamic measurement system in air.18,19

33
This study showed that when a specific antigen was interacted with an antibody that was immobilized on the functionalized gold surface of the cantilever, a resonant frequency shift arose; this shift arose from not only the induced mass, but also from the change of the cantilever spring constant caused by the change of the surface stress.

34
When a specific antigen is interacted with its antibody on the functionalized cantilever surface, the resonant frequency of the cantilever can be written as where k is the spring constant, Δk is the change of the spring constant due to the surface stress change, Δm is the induced mass (antigen mass), and m* is the effective mass of the cantilever.

35
Accordingly, from eqns. (1) and (2), the resonant frequency shift is mainly caused by the mass increase and the spring constant change in the protein–protein interaction.

36
Therefore, in this paper, we discuss the assay using these effects.

Experimental procedure

Fabrication of the PZT cantilever for a bio-assay in an aqueous environment

37
We have fabricated micromachined cantilevers with PZT layers that can perform self-actuating and sensing by means of simultaneous direct and indirect piezoelectric effects.18,19

38
Fig. 1 shows an SEM photograph of PZT thin film cantilevers composed of SiO2/Ta/Pt/PZT/Pt/SiO2 on a SiNx supporting layer for electrical detection.

39
In order to operate the cantilevers in a liquid environment, all the cantilevers were coated with parylene-c, which provided an electrically insulating biocompatible barrier against moisture and biofluids.20

40
The parylene-c film was deposited on the cantilever by using the chemical vapor deposition technique.

41
This deposition was accomplished by a process of vapor deposition and polymerization, in which the dimeric parylene-c is vaporized under a vacuum of 0.1 Torr at 150 °C, pyrolized under a vacuum of 0.5 Torr at 680 °C to form a reactive monomer, and then pumped into a chamber containing the component to be coated at 25 °C.

42
At a low chamber temperature, the monomeric parylene was deposited on the substrate, where it was immediately polymerized via a free-radical process.

43
The film thickness of parylene-c on the cantilever was approximately 500 nm.

44
After parylene-c coating on the cantilever, a Au/Cr layer of 150/30 nm thickness was deposited on the bottom side of the cantilever using an e-beam evaporator.

45
The Au/Cr layer served as an immobilization layer for the PSA antigen–antibody interaction.

46
However, due to the surface energy difference between parylene-c and the Au/Cr layer, the adhesion between them was very poor.

47
In order to adjust the surface energy difference, we treated the parylene-c surface using an atmospheric or oxygen plasma.

48
These treatments led to a remarkable enhancement of the adhesion between parylene-c and the Au/Cr layer, by increasing the surface energy of the parylene-c surface and by utilizing the mechanical interlocking effect21,22.

SAMs formation and PSA antibody immobilization

49
After the Au/Cr deposition on the bottom side of the parylene-c coated cantilevers, the cantilevers were cleaned in a fresh piranha solution (a 4∶1 ratio of H2SO4 (98.08%) and H2O2 (34.01%)) for 1 min, in order to remove the experimental contamination of the Au surface.

50
The cantilevers were then rinsed with deionized water before the preparation of SAMs.

51
The rinsed cantilevers were dried under a stream nitrogen gas.

52
A monoclonal anti-PSA (anti-prostate specific antigen, Fitzgerald Industries International Inc., Concord, MA, USA) was immobilized using Calixcrown (a calixarene derivative) SAMs, which have an efficient immobilization property owing to their recognition of ammonium ions of proteins23 enabling proteins to immobilize in SAMs, as shown in Fig. 2.

53
The formation of Calixcrown SAMs was accomplished by immersing the parylene-c and Au-coated cantilever for 3 h in a solution of Calixcrown in CHCl3, at room temperature.

54
To immobilize the PSA antibody onto the functionalized Au surface of the cantilever with Calixcrown SAMs formation, the cantilever was immersed in a monoclonal PSA antibody diluted PBS solution, with a concentration of 10 µg ml−1, at room temperature for 1 h.

55
The cantilever was then washed with 30 ml of PBST solution (PBS with 0.5% Tween 20, pH 7.8) and dried under nitrogen gas.

56
The major binding force between the Cailxcrown and the anti-PSA (PSA antibody) could be attributed to the ionized amine groups of the anti-PSA, which bind to the crown moiety of the linker molecule via a host–guest interaction.

57
Hydrophobic interactions between the hydrophobic residues of the anti-PSA and methoxy groups of the Calixcrown may also be involved in protein immobilization.

58
This immobilization technique onto Calixcrown SAMs could be performed by minimized nonspecific antibody binding and increased antigen binding to the antibody caused by the correct orientation of the PSA antibody immobilized in high density on the Calixcrown SAMs.23

59
Subsequently, in order to prevent non-specific binding, the cantilevers were immersed in dissolved bovine serum albumin (BSA, Sigma, St. Louis, MO, USA) in a phosphate buffered saline (PBS) with a 10 mg ml−1 concentration, for 1 h at room temperature.

60
Then, the cantilevers were rinsed with PBS (pH 7.4) containing polyoxyethylenesorbitan monolaurate (Tween 20, St. Louis, MO, USA), and a final washing was performed with PBS solution alone.

61
To confirm the specific binding after the PSA antigen (Fitzgerald Industries International Inc., Concord, MA, USA)–antibody interaction, we performed fluorescence labeling, using FluoroLinker Cy3 Mono Reactive Dye (Amersham Biosciences, piscataway, NJ, USA).

62
After the PSA antigen–antibody interaction, a laser confocal scanner (GSI Lumonics Co.) was used to observe the fluorescent scanning images.

63
The fluorescence images of the cantilevers were used to compare with the resonant frequency change results in a liquid environment.

Liquid cell using poly(dimethylsiloxane) (PDMS)

64
In order to conduct a dynamic measurement in a liquid environment, we have made a liquid test cell with a 200 µm width channel and a 20 µl volume reaction chamber, by bonding poly(dimethylsiloxane) (PDMS) bilayers, as shown in Fig. 3.

65
A curing agent and a PDMS prepolymer (SYLGARD 184 Silicone Elastomer Kit, Dow Corning, Midland, MI) were mixed in a ratio of 1∶10.

66
The mixed prepolymer was degassed at 30 mTorr for 30 min in a vacuum dessicator with a mechanical pump to remove any air bubbles.

67
The prepolymer mixture was poured onto two metal molds for the top and the bottom layers of the liquid cell, and then the molds with the PDMS prepolymers were cured for 90 min in an oven at 80 °C.

68
After curing, the top and bottom layers of the liquid cell were peeled off from the metal molds.

69
Subsequently, the two PDMS layers of the liquid cell were cleaned in methanol with an ultrasonicator.

70
To bond the two PDMS layers, oxygen plasma treatment for surface modification was performed using a reactive ion etcher (RIE) system (Plasmatherm.

71
790 series) for 20 s.

72
The working pressure of the oxygen plasma was 200 mTorr, with a constant oxygen gas flow of 40 sccm.

73
Before the bonding of two layers, anti-PSA immobilized cantilevers were inserted between the top and bottom PDMS layers.

74
Oxygen plasma treated layers were immediately bonded to the modified surfaces facing each other.

75
Subsequently, the integrated chips consisting of a PDMS liquid cell and a cantilever device were created through the integration of both chips.

76
For biological passivation, we injected a BSA solution with a 10 mg ml−1 concentration into the liquid cell, in order to prevent non-specific adsorption of desired proteins onto the channel wall and the inactive part of cantilever.

77
All further biological assays were performed using these integrated chips with BSA treated cantilevers.

Measurements of the cantilever resonant frequency in a liquid environment

78
The resonant frequency of the cantilever was measured using an optical heterodyne laser Doppler vibrometer (MLD211D, Neo Ark Co. Japan) in a liquid environment, as shown in Fig. 4.

79
To measure the resonant frequency, the cantilever in liquid cell was excited by the indirect piezoelectric effect of the PZT cantilever using a function waveform generator (33120A, Agilent).

80
During the sweeping of the desired frequency range, sensing signals of the vibrometer could be measured, demonstrating a first resonant frequency of the cantilever, at which the cantilever showed a maximum displacement value in liquid (e.g. PBS, BSA).

81
The ac sine wave form of 0.5 Vpp (peak to peak) with a superimposed dc voltage (0.25 V) was applied to the actuation (top) electrode, while the bottom electrode was grounded.

82
From this in-situ measurement in liquid, we detected the resonant frequency change of the cantilever due to the PSA antigen–antibody interaction.

83
Then, the resonant frequency shifts of the cantilevers were monitored with respect to the reaction time and the PSA concentration, respectively.

84
The same experiments were simultaneously performed by means of fluorescence detection using a fluorescent scanner.

85
Based on our comparison with the fluorescence results, we discuss the feasibility of in-situ biosensing in a liquid environment using the cantilevers.

Results and discussion

Resonant frequency change of a cantilever caused by PSA antigen–antibody specific binding in a liquid environment

86
In order to confirm the blocking effect of BSA on the non-specific adsorption of PSA, the cantilevers in the PDMS liquid cell were specially prepared without immersing in BSA solution after anti-PSA immobilization on the Calixcrown SAMs functionalized surface.

87
The cantilevers used in this set of experiments (Fig. 5) were made of SiNx supported PZT cantilever, with lengths of 300 µm and widths of 100 µm.

88
The high specificity and sensitivity of the PSA antigen–antibody could be confirmed by injecting another protein (e.g. CRP antigen, BSA) into the reaction chamber.10,24

89
Fig. 5 shows the typical resonant frequency shift of the cantilever after the first injection of BSA and the subsequent injection of the PSA antigen.

90
We performed an initial injection of BSA (10 mg ml−1) into the liquid cell, the PBS wash-out, and then performed an injection of the PSA antigen (100 ng ml−1).

91
We then measured the resonance frequency change of the cantilever with respect to time.

92
After the injection of BSA (10 mg ml−1) through the PDMS liquid cell using a syringe pump, there was a slight change of resonant frequency due to BSA absorption.

93
It is evident that over a period of 20 min, the resonant frequency decreased and then saturated to a steady-state value.

94
The slight shift of resonant frequency indicated that the non-specific binding of BSA onto the anti-PSA immobilized cantilever was relatively minimal.

95
When the PSA antigen with a 100 ng ml−1 concentration was injected in the liquid cell, using a syringe pump, a clear change in the resonant frequency of the cantilever was observed.

96
This was evidently caused by mass loading due to the PSA antigen–antibody interaction and a change of the spring constant, due to the compressive stress between the PSA antigen–antibody complexes on the functionalized cantilever surface.

97
In the case of the antigen–antibody interaction, compressive surface stress occurs on the functionalized cantilever due to repulsive electrostatic, steric intermolecular interactions, or changes of the hydrophobicity of surface.4,19

98
After the resonant frequency of cantilever had attained a steady state in a BSA (10 mg ml−1) solution and a PSA antigen (100 ng ml−1), the value of the resonant frequency shift was about 50 Hz in a BSA with a 100 mg ml−1 concentration.

99
In contrast, 150 Hz was the value of resonant frequency shift due to interaction between the PSA antibody and a PSA antigen with a 100 ng ml−1 concentration.

100
This result indicates that specific binding occurs during the interaction of the PSA antigen–antibody.

101
Therefore, the resonant frequency shift of the cantilever in a liquid environment happened only because of the PSA antigen specific binding with its antibody on the functionalized Au surface of the cantilever.

Resonant frequency change of the cantilever caused by three different PSA antigen concentrations in a liquid environment

102
To evaluate the in-situ resonant frequency change as a function of the PSA antigen concentration (1, 10 and 100 ng ml−1), cantilevers with the same dimensions in each PDMS liquid cell were used, and the PSA antibody was immobilized on the Calixcrown SAM coated cantilever.

103
BSA passivation was performed in order to prevent the non-specific adsorption of anti-PSA onto the microchannel wall and the inactive part (without immobilized anti-PSA) of the cantilever.

104
The cantilevers used in this experiment (Fig. 6) were made of SiNx supported PZT and were 50 × 150 µm.

105
Cy3 labeled PSA antigen solutions with different concentrations (1, 10 and 100 ng ml−1) were injected into each PDMS liquid cell with the cantilever using a syringe pump.

106
During the interaction of the PSA antigen–antibody, we repeatedly measured the resonant frequency of the cantilever using a laser Doppler vibrometer until the resonant frequency of the cantilever attained a saturated value.

107
Fig. 6 shows the in-situ resonant frequency changes and the fluorescence scanner images of the functionalized PZT cantilevers as a function of the PSA concentration.

108
In all cases with three different PSA antigen concentrations, the resonant frequency of the cantilevers changed suddenly due to the PSA antigen–antibody interaction on the functionalized cantilevers surface in a liquid environment.

109
Early in the PSA antigen/antibody interaction, the resonant frequency of the cantilever changed drastically, in accordance with the binding kinetics of Langmuir isotherm effect and diffusion.25

110
This phenomenon due to the Langmuir isotherm and diffusion effect was similarly observed in other studies of protein–protein interactions using surface plasmon resonance (SPR) and QCMs.23,26

111
It is evident that over a period of 30–40 min, the resonant frequency decreased and then reached a saturated value.

112
The time for the resonant frequency to reach a steady state was somewhat shorter than the saturation times of the static mode in other reports.10,27

113
The rapid measuring time of the dynamic mode, in comparison with the dc mode, is a result of either the diffusion of molecules in a reaction chamber with a small volume (20 µl) or the conformational relaxation of the antigen–antibody complex on the functionalized cantilever surface.28

114
The diffusion can be significantly enhanced by a proper liquid cell design, which is currently being developed.

115
In the saturated regimes, increasing PSA antigen concentrations of 1, 10 and 100 ng ml−1 led to respective increasing values of the experimental resonant frequency change of 94, 166 and 203 Hz.

116
To confirm whether the resonant frequency change was caused by the PSA antigen–antibody specific binding, Cy3 labeled PSA antigen was used to observe specific binding after the antigen–antibody interaction.

117
After the PSA bioassays were performed in the liquid cells, all the cantilevers were separated from the cells, rinsed with PBS, and then dried at room temperature.

118
Fluorescence images of all cantilevers were taken using a laser confocal scanner.

119
Fig. 6 shows the fluorescent images of the cantilevers after the PSA antigen–antibody binding.

120
Fig. 6(b)–(d) show that the fluorescence images indicate the presence of Cy3 labeled PSA antigen (with three different concentrations) binding to the immobilized antibodies on the Calixcrown SAMs formed cantilever.

121
The brightness of the cantilever image increased with an increase of the PSA concentration, indicating that the trend of the resonance frequency shift was very similar to that of the fluorescence results.

122
From both the fluorescence results and the results of the resonance frequency change, it is evident that the ac mode-based PZT cantilever can be used as a nanomechanical biosensor with a high sensitivity in a liquid environment with no labeling being necessary.

123
In this experiment three functionalized cantilever arrays were used.

124
Because the heterodyne laser Doppler vibrometer (MLD211D, Neo Ark Co. Japan) has one laser beam, the measurement of the in-situ resonant frequency was performed by using one cantilever from among the three functionalized cantilevers.

125
However, the other cantilevers were used to confirm the final resonant frequency shift value, after reaching a steady state, for comparison with the three functionalized cantilevers.

126
After reaching a steady state, the resonant frequency shift values of the three cantilevers have practically the same value.

127
We found that the in-situ monitoring of the antigen–antibody bioassay could be realized using the dynamic mode detection of a cantilever, instead of using the widely published static deflection method.

Conclusions

128
In this work, we have presented a new in-situ detection method of an antigen–antibody bioassay in a liquid environment, using the dynamic mode of a cantilever to measure the resonant frequency change of a PZT cantilever.

129
The specific binding of a PSA antigen–antibody was confirmed by using a BSA solution with a high concentration and fluorescence images after the PSA antigen–antibody interaction.

130
The trend of the resonant frequency change with respect to time could be explained by binding kinetics due to Langmuir isotherm behavior and diffusion, and by the effect of a small volume reaction chamber.

131
In the saturated regimes, the resonant frequency of the cantilever increased with an increase of the PSA concentration in the reaction chamber, indicating that the trend of the resonance frequency change was similar to that of the fluorescence results.

132
The presented data demonstrate that this PSA detection system, based on a PZT cantilever using the ac mode in a liquid, is very feasible to adopt for a lab-on-a-chip (LOC).

133
We anticipate that the successful demonstration of this PSA detection system will lead to the rapid development of many protein detection systems using the resonant frequency of a cantilever in a liquid environment.