Anal. Chem. 1997, 69, 3506-3512
Piezoelectric Immunosensors for Urine Specimens of Chlamydia trachomatis Employing Quartz Crystal Microbalance Microgravimetric Analyses Iddo Ben-Dov and Itamar Willner*
Institute of Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel Einat Zisman
Savyon Diagnostics Ltd., Kiryat Minrav, 3 Habosem Street, Ashdod 77101, Israel
The assembly of a biosensor for Chlamydia trachomatis based on the microgravimetric quartz crystal microbalance (QCM) analysis of the bacteria association to an antibody-functionalized electrode is described. The sensing interfaces consist of a primary cystamine monolayer assembled onto Au electrodes associated with the quartz crystal. The monolayer is further modified with sulfosuccinylimidyl 4-(p-maleimidophenyl)butyrate (sulfoSMPB) and the goat IgG-anti-mouse IgG Fc-specific Ab or the fragmented F(ab′)2 anti-mouse IgG Ab that act as sublayers for the association of the sensor-active anti-C. trachomatis LPS-Ab. Bacteria in the concentration range from 260 ng‚mL-1 to 7.8 µg‚mL-1 are sensed by the functionalized crystals. The association of C. trachomatis to the sensing interface can be confirmed and amplified via interaction of the crystal with various antiC. trachomatis antibodies. Urine-pretreated functionalized quartz crystals are applied in the analysis of C. trachomatis in urine samples. The sensitivity limits of the electrodes for sensing the bacteria in urine samples corresponds to ∼260 ng‚mL-1. The functionalized crystals assembled via association of anti-C. trachomatis LPS-Ab to the fragmented F(ab′)2 anti-mouse IgG Ab reveal long-term stability upon storage at 4 °C. The intracellular species, Chlamydia trachomatis, has been acknowledged as a widespread sexually transmitted microorganism.1 In women, various infections can lead to infertility, ectopic pregnancy, and chronic pelvic pain. In men, C. trachomatis causes urethritis and epididymitis, and newborns to infected mothers are at risk of chlamydial conjunctival and pulmonary disease.2 Cell culture offers a general means for the isolation and identification of C. trachomatis.3 Various immunoassays, enzyme-linked immunosorbent assays (ELISA), and fluorescence labeled antibody assay were developed for C. trachomatis, yet the assay is limited in its sensitivity, requires a long time (∼3 h), and is often subjective.4 Polymerase chain reaction (PCR) was found to be sensitive and specific in testing specimens from men with (1) Judson, F. J Reprod. Med. 1989, 30, 269. (2) Mardh, P. A.; Paavonen, J.; Puolakkainen, M. Chlamydia; Plenum Publishing Corp.: New York, NY, 1989; pp 127-249. (3) Taylor-Robinson, D.; Thomas, B. J. Genitourin. Med. 1991, 67, 256. (4) Taylor-Robinson, D. J. Infect. 1992, 25 (Suppl. 1), 61.
3506 Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
nongonococcal urethritis (NGU).5 This method is, however, timeconsuming (12-24 h) and is not applicable for routine screening and analysis. Detection of anti-C. trachomatis antibodies is problematic due to inefficient stimulation of antibodies formation, the existence of antibodies several years after infection,4 and the possible existence of anti-Chlamydia pneumoniae or anti-Chlamydia psittaci in the sera.6,7 Thus, the development of a highly sensitive detection means for C. trachomatis in urine specimens or urethral swabs could provide an important diagnostic method for this infection. The advance in biosensor technology, specifically in immunosensor systems, suggests that new means for the detection of infectious diseases, e.g., C. trachomatis, could be developed. Different immunosensor devices based on electrochemical8-11 or surface plasmon resonance12 transduction of the formation of antigen-antibody complexes were reported in recent years. A rapidly progressing analytical method to sense antigen-antibody interaction involves the microgravimetric quartz crystal microbalance (QCM) analyses employing piezoelectric crystals, i.e., quartz crystal.13,14 The crystal frequency is controlled by its mass, and any mass change associated with the crystal stimulates a frequency change. The Sauerbrey relation, eq 1, expresses the (5) Quinn, T. C.; Bobo, L.; Holland, S. M.; Gaydos, C. A.; Hook, E.; Viscidi, R. P. In Chlamydial Infections; Bowie, W. R., Ed.; Cambridge University Press: Cambridge, UK, 1990; pp 491-494. (6) Schachter, J. In Chlamydial Infections; Oriel, D., Ridgway, G., Schachter, J., Ward, M., Eds.; Cambridge University Press: Cambridge, UK, 1986; pp 311-320. (7) Puolakkainen, M.; Vesterinen, E.; Purola, E.; Saikku, P.; Paavonen, J. J. Clin. Microbiol. 1986, 23, 924-928. (8) Doyl, J. M.; Wehmeyer, K. R.; Heineman, W. R.; Halsall, H. B. In Electrochemical Sensors in Immunological Analysis; Ngo, T. T., Ed.; Plenum Press: New York, NY, 1987; pp 87-102. (9) (a) DiGleria, K.; Hill, H. A. O.; McNeil, C. J.; Green, M. J. Anal. Chem. 1986, 58, 1203-1205. (b) DiGleria, K.; Hill, H. A. O.; Chambers, J. A. J. Electroanal. Chem. 1989, 267, 83-91. (c) Chambers, J. A.; Walton, N. J. J. Electroanal. Chem. 1988, 250, 417-425. (10) Wehmeyer, K. R.; Halsall, H. B.; Heineman, W. R.; Volle, C. P.; Chen, C. Anal. Chem. 1986, 58, 135-139. (11) (a) Aizawa, M. In Electrochemical Sensors in Immunological Analysis; Ngo, T. T., Ed.; Plenum Press: New York, NY, 1987; pp 269-291. (b) Gebauer, C. R. In Electrochemical Sensors in Immunological Analysis; Ngo, T. T.; Ed., Plenum Press: New York, 1987; pp 239-255. (12) (a) Liedberg, G.; Nylander, C.; Lundstro¨m, I. Biosens. Bioenerg. 1995, 10, i-ix. (b) Davis, J.; Roberts, C. J.; Dawkes, A. C.; Sefton, J.; Edward, J. C.; Glasbey, T. O.; Haymes, A. G.; Davies, M. C.; Jackson, D. E.; Lomas, M.; Shakesheff, K. M.; Tendler, S. J. B.; Wilkins, M. J.; Williams, P. M. Langmuir 1994, 10, 2654-2661. (13) (a) Suleiman, A. A.; Guilbault, G. G. Analyst 1994, 119, 2279-2282. (b) Ko ¨nig, B.; Gra¨tzel, M. Anal. Chem. 1994, 66, 341-344. S0003-2700(97)00216-3 CCC: $14.00
© 1997 American Chemical Society
change in the crystal frequency, ∆f, upon alteration of the crystal mass by ∆m,
( )
∆f ) -
fo2 ∆m ) -Cf∆m ) -1.83 × 108∆m NFq
(1)
where fo is the fundamental frequency of the crystal, Fq is the quartz density (2.648 g cm-3), and N is the shear modulus of quartz (167 kHz‚cm). Thus, mass changes occurring on the crystal as a result of an antigen association to the crystal surface, or an antibody dissociation from the crystal, are reflected by a frequency decrease or frequency increase of the crystal, respectively. Antigen-modified QCM crystals were applied for the detection of antigens by a competitive method by interacting the crystal with a mixture of the antigen and a predetermined amount of the respective antibody.15 Piezoelectric detection of human IgG was reported in the presence of anti-human IgG or protein A-modified crystals.16 Similarly, microbial cells such as Candida albicans, Escherichia coli, Salmonella, and Shigella were analyzed by the application of crystals modified with the respective complementary antibodies.17,18 In a series of reports, we addressed the assembly of bioactive monolayers on Au surfaces as sensing interfaces. Immobilization of redox relay modified enzymes on Au electrodes led to the development of a series of amperometric biosensors.19,20 This approach was extended to assemble antigen monolayers on Au surfaces for the amperometric detection of the complementary antibody or for the electrochemical analysis of the antigen by a competitive assay with a predetermined amount of the antibody.21,22 Recently, antigen monolayers assembled onto Au surfaces associated with quartz crystals were used for the microgravimetric detection of the complementary antibodies.23 The organization of the bioactive material as a monolayer on the transducing support (i.e., electrode or crystal) results in the rapid mass transport of the analyte to the sensing interface, and hence rapid response times of the sensor are achieved. That is, in contrast to other systems, where the antibody (or antigen) is immobilized onto the transducing support by polymer matrices that introduce diffusional barriers for the complementary component, no transport limitations are present in the monolayer assemblies. Furthermore, the synthetic methodologies developed (14) (a) Guilbault, G. G.; Hock, B.; Schmid, R. Biosens. Bioelectron. 1992, 7, 411-419. (b) Ko ¨sslinger, C.; Drost, S.; Aberl, F.; Wolf, H.; Koch, S.; Woias, P. Biosens. Bioelectron. 1992, 7, 397-404. (15) (a) Roederer, J. E.; Bastiaans, G. J. Anal. Chem. 1983, 55, 2333. (b) Roederer, J. E.; Bastiaans, G. J. U.S. Patent 4735906, 1988. (16) Muramatsu, H.; Dicks, J. M.; Ramiya, E.; Karube, I. Anal. Chem. 1987, 59, 2760. (17) Muramatsu, H.; Kajiwara, K.; Tamiya, E.; Karube, I. Anal. Chim. Acta 1986, 188, 257. (18) Muramatsu, H.; Watanabe, Y.; Hikuma, M.; Ataka, T.; Kubo, I.; Tamiya, E.; Karube, I. Anal. Lett. 1989, 22, 2155. (19) (a) Willner, I.; Katz, E.; Riklin, A.; Kasher, R. J. Am. Chem. Soc. 1992, 114, 10965-10966. (b) Willner, I.; Riklin, A. Anal. Chem. 1994, 66, 1535-1539. (20) (a) Riklin, A.; Willner, I. Anal. Chem. 1995, 67, 4118-4126. (b) Willner, I.; Riklin, A.; Shoham, B.; Rivenzon, D.; Katz, E. Adv. Mater. 1993, 5, 912915. (c) Katz, E.; Riklin, A.; Willner, I. J. Electroanal. Chem. 1993, 354, 129-144. (d) Willner, I.; Katz, E.; Riklin, A.; Kasher, R.; Shoham, B. U.S. Patent 5,443,701, 1995. (21) (a) Willner, I.; Blonder, R.; Dagan, A. J. Am. Chem. Soc. 1994, 116, 93659366. (b) Willner, I.; Rubin, S.; Cohen, Y. J. Am. Chem. Soc. 1993, 115, 4937-4938. (22) Blonder, R.; Katz, E.; Cohen, Y.; Itzhak, N.; Riklin, A.; Willner, I. Anal. Chem. 1996, 68, 3151-3157. (23) Cohen, Y.; Levi, S.; Rubin, S.; Willner, I. J. Electroanal. Chem. 1996, 417, 65-75.
by us to assemble the monolayer allow us to construct highly ordered complex monolayers that include several bioactive materials in a tailored architecture. Here we report on the development of an immunosensor for the piezoelectric, microgravimetric detection of Chlamydia trachomatis in urine samples. We show the stepwise assembly of the bioactive sensing interface and the means to design specificity and confirmatory detection and discuss the advantages of the device as compared to commercially available ELISA assays. EXPERIMENTAL SECTION A quartz crystal microbalance (EG&G Model QCA 917) interfaced to a computer with home-made software was used in the studies. The analysis cell consisted of a Teflon housing for the electrode mounted on a home-manufactured plastic cell equipped with a septum for sample injection and inlet and outlet tubes for rinsing the cell. The flow system was linked to a peristatic pump that allowed the rinsing of the cell with the buffer solution. The quartz crystal was mounted in the cell in a perpendicular configuration, and only one electrode was exposed to the analyte solution. Quartz crystals, 9 MHz, sandwiched between two Au electrodes, geometrical area ∼0.2 cm2 (EG&G), were used in all the measurements. Sulfosuccinylimidyl 4-(p-maleimidophenyl)butyrate (sulfoSMPB, Pierce), cystamine dihydrochloride (Fluka), polyclonal, Fcspecific goat IgG-anti-mouse IgG and fragmented F(ab′)2 polyclonal, Fc-specific goat IgG-anti-mouse IgG, (Sigma), bovine serum albumin (BSA, Sigma), mouse IgG, and mouse IgE-antidinitrophenol (Sigma) were used as supplied with no further purification. All other chemicals were from Aldrich. The monoclonal mouse IgG-anti-Chlamydia trachomatis antibodies (antiLPS, anti-P60, anti-MOMP) and formaldehyde-inactivated C. trachomatis bacteria were supplied by Savyon Diagnostics (Ashdod, Israel). Quartz crystal electrodes were pretreated by soaking them in 1 M NaOH for 20 min, followed by rinsing with distilled water. The resulting crystals were then soaked in 1 M HCl for 2 min, rinsed with distilled water, and dried under a stream of argon. Concentrated hydrochloric acid, 50 µL, was placed on the Au surfaces for 2 min, followed by rinsing of the crystal with ethanol and distilled water. The cleaned electrodes were modified with the cystamine monolayer by treatment of the crystal, in a test tube, with a 0.02 M aqueous solution of cystamine dihydrochloride for 2 h. The resulting electrode was rinsed with ethanol and water. In the next step, the monolayer-modified crystal was reacted with a phosphate-buffered saline solution (PBS, 137 mM NaCl, 2.7 mM KCl, 10 mM Na2PO4, 1.8 mM KH2PO4, pH ) 7.4) that included sulfo-SMPB, 2 mM, for 1 h. The surface densities of the cystamine submonolayer and sulfo-SMPB were determined by following the crystal frequencies at each modification step in air and applying the Sauerbrey relation, eq 1. The resulting functionalized crystal was further modified by two alternative routes: (a) the electrode was reacted with polyclonal goat IgG-anti-mouse IgG, or (b) the electrode was reacted with fragmented polyclonal goat IgG-antimouse IgG F(ab′)2. The modifications were carried out by the interaction of the electrode with 17 µg mL-1 of the respective antibodies in 50 mM HEPES buffer solution that included 10 mM EDTA, pH ) 6.5, for 30 min. The resulting electrodes were rinsed with the BPS buffer solution and further interacted with a PBS solution that included the mouse IgG-anti-C. trachomatis LPS antibody, 1.7 µg mL-1, for 30 min. In the appropriate experiments, Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
3507
Scheme 1. Assembly of the C. trachomatis Sensing Interface on the Au Electrodes Associated with the Quartz Crystal by the Immobilization of Anti-IgG Ab or Fragmented Anti-IgG F(ab′)2 Ab and Schematic Sensing Principle of Chlamydia trachomatisa
a
HS, free thiol groups associated with the antibody; LPS, lipopolysaccharide.
the anti-LPS antibody solution included also BSA, 5 mg mL-1. The assembly of the two antibodies on the functionalized monolayer was characterized by following the crystal frequencies at each modification step in air. The resulting electrodes were mounted in the QCM cell, and the cell was filled with the PBS buffer solution. The crystal was allowed to stabilize to a constant frequency, (1 Hz. Samples of C. trachomatis in PBS of variable concentrations were injected into the cell, and the crystal frequency changes were monitored as a function of time. In the appropriate systems, C. trachomatis in urine samples were injected into the cell. The C. trachomatismodified crystals were then rinsed with the PBS buffer, 10 mL, at 2 mL min-1. The respective anti-C. trachomatis antibodies (antiLPS, anti-P60, or anti-MOMP) were injected into the cell, 175 ng mL-1, and the crystal frequency changes were followed as a function of time. In the appropriate experiments, the C. trachomatis-anti-C. trachomatis-functionalized electrodes were treated, after rinsing, with the polyclonal, Fc-specific, goat IgG-anti-mouse IgG, 175 ng mL-1, to obtain further amplification. The crystal frequencies were monitored as a function of time to follow the association of the antibodies. RESULTS AND DISCUSSION The interfaces sensing Chalmydia trachomatis were prepared as schematically outlined in Scheme 1. A cystamine monolayer was assembled on the Au electrodes associated with the quartz crystal. Sulfosuccinimidyl 4-(p-maleimidophenyl)butyrate (sulfoSMPB) was covalently linked to the base cystamine monolayer. Cysteine residues of anti-mouse IgG Ab or of fragmented F(ab′)2 anti-mouse IgG Ab were linked to the maleimide residues on the 3508
Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
monolayer. The resulting antibody-functionalized interfaces were treated with anti-C. trachomatis LPS (lipopolysaccharide) antibody to yield the active sensing interface. The Fc fragment of the latter antibody binds to the IgG base layer. The assembly of the layered sensing interface was characterized by following the crystal frequency changes after each modification step. Table 1 summarizes typical values of the crystal frequency changes at the different steps of modification upon assembly of the anti-C. trachomatis Ab to the anti-mouse IgG Ab or the F(ab′)2 fragmented anti-mouse IgG Ab. The surface coverages of the different components are also included in the table. It should be noted that, although the anti-mouse IgG Ab or the F(ab′)2 anti-mouse IgG includes two binding sites for the anti-C. trachomatis Ab, the surface coverage of the electrode by the latter antibody is only ∼50% of the surface coverage of the base antibody monolayers. This could originate from the nonspecific covalent binding of the base antibodies to the monolayer that results in the blocking of the Fab fragments toward binding the anti-C. trachomatis antibody. Figure 1 shows the frequency changes of crystals tailored by the two methods upon interaction of the electrode with different concentrations of C. trachomatis. In these experiments, the electrodes are consecutively treated with increasing concentrations of C. trachomatis. Figure 1A shows the frequency changes of the sensing interface of the anti-C. trachomatis antibody immobilized (24) The substantial differences in the frequency changes observed for the assembly of the primary cystamine monolayer for a series of electrodes are attributed to the different roughness factors of the respective Au surfaces. These differences in the surface roughness are screened upon the assembly of the secondary layers, and specifically, the high molecular weight antibodies.
Table 1. Typical Frequency Changes (∆f) of Quartz Crystals upon Stepwise Assembly of the Components in the Sensing Interfaces and the Respective Calculated Surface Coverage (σ) of the Corresponding Componentsa sensing interfaceb modification step cystamine sulfo-SMPB untreated anti-mouse IgG Ab F(ab′)2 anti-mouse IgG anti-C. trachomatis Ab
sensing interfacec
∆f/Hz
σ/mol‚cm-2
∆f/Hz
σ/mol‚cm-2
-202 -105 -105
7.562 × 10-9 2.315 × 10-9 4.0 × 10-12
-195 -95
7.300 × 10-9 2.095 × 10-9
-63
2.4 × 10-12
-89 -74
5.1 × 10-12 2.8 × 10-12
a
Calculated using eq 1. The stated values represent an average of the values observed for six electrodes. The frequency changes of the primary cystamine monolayer vary within (20%, those of the sulfoSMPB monolayer within (10%, and those of the antibody monolayers within (5% for different electrode preparations.24 b Sensing interface consists of the goat IgG anti-mouse IgG, Fc-specific Ab as sublayer. c Sensing interface consists of the goat F(ab′) anti-mouse IgG Ab as 2 sublayer.
A
B
Figure 1. Frequency changes of functionalized quartz crystals as a function of time upon interaction with different concentrations of C. trachomatis. (A) Sensing interface is functionalized by Fc-specific, goat IgG-anti-mouse IgG Ab and anti-C. trachomatis LPS Ab. (B) Sensing interface is functionalized by fragmented F(ab′)2 anti-IgG and anti-C. trachomatis LPS Ab. In both systems, the concentrations of C. trachomatis correspond to (a) 78 ng mL-1, (b) 260 ng mL-1, (c) 780 ng mL-1, (d) 2.6 µg mL-1, and (e) 7.8 µg mL-1.
onto the intact anti-mouse IgG antibody layer. Figure 1B shows the frequency changes of the anti-C. trachomatis sensing layer associated with the fragmented F(ab′)2 anti-mouse IgG. In both systems, as the concentration of the bacteria increases, the frequency of the crystal decreases, implying the association of C. trachomatis to the sensing interface. For example, in the presence of bacteria at concentrations 0.26 (curve b) and 2.6 (curve d)
Figure 2. Frequency changes of a C. trachomatis-functionalized crystal as a function of time upon treatment with different antibodies. The electrode is composed of the goat IgG-anti-mouse, Fc-specific IgG Ab and anti-C. trachomatis LPS Ab as sensing interface. The electrode was treated with C. trachomatis, 2.6 µg mL-1, prior to interaction with the antibodies: (a) Goat IgG-anti-mouse IgG (Fcspecific), (b) mouse IgE-anti-dinitrophenol, (c) mouse IgG-antidinitrophenol, (d) mouse IgG-anti-C. trachomatis P60, (e) mouse IgG-anti-C. trachomatis MOMP, and (f) mouse IgG-anti-C. trachomatis LPS.
Figure 3. Sequential frequency changes of a quartz crystal as a function of time upon interaction with C. trachomatis and subsequently with anti-C. trachomatis LPS. The crystal is functionalized with the Fc-specific goat IgG-anti-mouse IgG Ab, anti-C. trachomatis LPS Ab, and C. trachomatis, 2.6 µg‚mL-1 and treated with (a) anti-C. trachomatis, 175 ng mL-1 and (b) Fc-specific goat IgG-anti-mouse IgG Ab, 175 ng‚mL-1. The quartz crystal was washed with the PBS buffer solution between steps (a) and (b).
µg mL-1, the crystal frequency changes correspond to ∆f ) -6 and -34 Hz, respectively (Figure 1A). For different electrode preparations (four runs), the observed frequency changes vary within (10%. The response time for the detection of the bacteria is ∼350 s and is defined as the time after which the crystal frequency levels off at low concentrations of C. trachomatis. The specific association of C. trachomatis to the electrode can be confirmed and amplified as schematically shown in Scheme 2. The sensing interface with the associated C. trachomatis is treated with an anti-C. trachomatis Ab. Binding of the latter Ab results in a further frequency decrease of the crystal. This confirmatory test can be further amplified by treatment of the C. trachomatisantibody complex with goat anti-mouse IgG. Association of the secondary antibody is expected to yield an additional increase in the multilayer mass and, consequently, a further decrease in the crystal frequency. Figure 2 shows the frequency changes of the sensing crystal that was interacted with C. trachomatis 2.6 µg mL-1 and subsequently treated with a variety of antibodies: goat antimouse IgG (curve a), mouse IgE-anti-dinitrophenol (curve b), and mouse IgG-anti-dinitrophenol (curve c). Specifically, the Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
3509
Scheme 2. Schematic Principle for Amplification and Confirmation of the Association of C. trachomatis to the Sensing Interface via the Application of Anti-C. trachomatis LPS Ab
crystal frequency changes of the C. trachomatis-modified surface in the presence of three different Chlamydia monoclonal antibodies, are shown: anti-P60 (60 kD membrane protein, curve d), antiMOMP (major outer membrane protein, curve e), and anti-C. trachomatis LPS (lipopolysaccharide, curve f). With the unrelated antibodies IgG and IgE, anti-dinitrophenol, and anti-mouse IgG Ab, no frequency changes are observed. These control experiments clearly indicate that the base layer of the anti-mouse IgG is saturated with the anti-C. trachomatis Ab and no foreign IgG binds to the surface. The anti-C. trachomatis antibodies cause a frequency decrease of the crystal. Thus, these antibodies bind to the initially associated C. trachomatis bacteria. It should be noted that the initial perturbations in the frequency responses of the crystal, observed immediately after the respective injections (∼10 s), originate from disturbances in the solution. The antiP60 and anti-MOMP reveal similar frequency changes in the crystal, whereas the anti-LPS antibody results in a substantially higher frequency change. This is consistent with the higher avidity of the latter antibody for the antigen bacteria. Thus, the interaction of the antibodies with the sensing interface that was treated with C. trachomatis provides a confirmatory test for the initial association of the antigen to the monolayer. The association of the anti-C. trachomatis Ab to the primary analyte antigen can provide an interface for the subsequent confirmation that the Ab binds to the bacteria rather than to a defective sublayer of the anti-mouse IgG. Figure 3 shows the frequency changes of the sensing interface treated with the bacteria upon interaction with anti-C. trachomatis LPS (curve a) and after subsequent treatment with goat anti-mouse IgG (curve b). Interaction of the electrode 3510
Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
with the anti-LPS results in a frequency change corresponding to ∆f ) -16 Hz. Treatment of the resulting electrode with the goat anti-mouse IgG yields an additional decrease in the crystal frequency (curve b), ∆f ) -18 Hz, implying that the secondary antibody binds to the anti-LPS. These results clearly indicate that the anti-LPS, being an IgG antibody, does not associate to a defective sublayer site but interacts with the bacteria antigen. This interaction yields the exposed Fc site of the IgG Ab that permits the secondary binding of goat anti-mouse IgG Ab. It should be noted that the goat anti-mouse IgG Ab does not yield any frequency changes upon interaction with the base sensing interface consisting of the anti-LPS Ab (cf. Figure 2, curve a). Thus, the goat anti-mouse IgG Ab specifically binds to the anti-LPS Ab associated with C. trachomatis. The set of successive frequency changes of the crystal, upon interaction with C. trachomatis and the anti-LPS and anti-IgG antibodies, provides a route to confirm the initial binding of the bacteria. Yet, this process provides a further advantage as it allows the amplification of the initial binding of C. trachomatis to the sensing interface by a cumulative frequency decrease observed only upon initial association of the antigen. A major concern in the development of immunosensors relates to the specificity of the sensing interface. Although the anti-C. trachomatis sensing interface is unaffected by foreign proteins such as cytochrome oxidase, cytochorme c, or glucose oxidase, we find that the crystal is influenced by BSA. Interaction of the electrode with 5 mg mL-1 BSA results in a frequency change of about -100 Hz due to nonspecific adsorption of the protein to the surface. The BSA-treated sensing surface is, however, not
Figure 4. Time-dependent frequency changes of a quartz crystal where its sensing interface was prepared by the attachment of antiC. trachomatis LPS Ab to the goat IgG-anti-mouse IgG, Fc-specific Ab in the presence of (a) BSA, 5 mg mL-1, or (b) C. trachomatis, 0.26 µg mL-1. (c) Subsequent to the treatment with the bacteria as in (b), interaction with anti-C. trachomatis LPS Ab, 175 ng mL-1.
further influenced upon interaction with BSA but reveals the original affinity for C. trachomatis. Figure 4 shows the frequency changes of the crystal pretreated with BSA upon interaction with a new BSA sample (curve a) and upon treatment with C. trachomatis 0.26 µg mL-1. The BSA does not influence the crystal frequency, but addition of the antigen bacteria results in a frequency decrease, ∆f ) -6 Hz, implying the association of the analyte antigen to the surface (curves b). Thus, the treatment of the sensing interface with BSA results in the nonspecific adsorption of the protein but blocks any further nonspecific adsorption phenomena. The BSA-blocked interface, however, retains its affinity for sensing C. trachomatis. As an alternative protocol to generate specific sensing interfaces, we tried to block the nonspecific adsorption within the process of manufacturing of the electrodes. The anti-C. trachomatis LPS Ab was associated to the bare antibody layer in the presence of 5 mg mL-1 BSA (see Scheme 1). Indeed, the resulting sensing interface was unaffected in the presence of BSA but revealed analogous activity for sensing C. trachomatis. Figure 4 (curve c) shows the crystal frequency changes of a BSA-pretreated electrode after treatment with 0.26 µg mL-1 C. trachomatis and subsequent incubation with the antiC. trachomatis LPS Ab. The observed frequency decrease confirms the primary association of the antigen bacteria to the sensing interface and amplifies the transduced signal. The C. trachomatis sensing interfaces are aimed to detect the bacteria in urine samples. Accordingly, it is important to examine the sensor performance in real urine samples. The BSA-treated sensing interfaces were treated with urine lacking the C. trachomatis. This results in a frequency change of the crystal of about -50 Hz due to the nonspecific association of the urine ingredients. The resulting electrodes were insensitive toward pure urine samples but revealed activities in sensing the antigen in urine samples. Inactivated C. trachomatis bacteria were dispersed in urine, and the crystal was challenged with different concentrations of the antigen in urine. Figure 5 shows the responses of the sensing interface in the presence of different samples of urine containing variable concentrations of the antigen. The crystal frequency decreases as the concentration of the C. trachomatis in the analyte sample increases. We note, however, that the extent of frequency decrease is substantially lower than the values observed when the antigen was dissolved in a pure PBS buffer solution. For example, in the presence of C. trachomatis at
Figure 5. Time-dependent frequency changes of a functionalized crystal pretreated with urine upon interaction with fresh urine samples that include variable concentrations of C. trachomatis. Sensing interface consists of goat-IgG-anti-mouse IgG, Fc-specific Ab and anti-C. trachomatis LPS Ab. Concentrations of C. trachomatis in urine samples correspond to (a) 78 ng mL-1, (b) 260 ng mL-1 , (c) 780 ng mL-1, and (d) 2.6 µg mL-1.
concentrations corresponding to 0.26 and 0.78 µg mL-1, the frequency changes of the electrode in urine are -2 and -9 Hz, whereas in PBS solutions, the frequency changes are -6 and -17 Hz, respectively (cf. Figure 1A). This implies that the sensitivity of the sensor decreases in the presence of native urine. The decrease in the frequency responses of the crystal in the presence of urine could be attributed to viscosity changes of the urine solution which oppose the frequency changes induced by the mass changes occurring on the crystal. To overcome this difficulty, and to retain the sensitivities observed in the PBS buffer solution, we combined the confirmatory test that involves the anti-LPS Ab and the measurement performed with urine. In this experiment, the sensing interface is interacted with C. trachomatis-containing urine to yield the low frequency change as discussed above. After association of the antigen bacteria to the sensing interface, the cell is rinsed with the PBS buffer. The anti-LPS Ab is then injected to the cell, and the high-value frequency change as a result of the association of the antibody to C. trachomatis in PBS is observed. Figure 6 shows the crystal frequency changes upon interaction of the sensing interface with urine containing C. trachomatis at concentrations corresponding to 0.26 and 2.6 µg mL-1 (curves a and c, respectively). With the sample containing 0.26 µg mL-1 of the antigen bacteria, a very low frequency change, ∆f ) -2 Hz, is observed, and the analytic determination of the bacteria is inconclusive. Upon rinsing the cell with the PBS buffer solution and treating of the electrode with the anti-LPS antibody, high frequency changes of the electrode are observed. For the systems that were interacted with urine containing C. trachomatis at 0.26 and 2.6 µg mL-1, frequency changes corresponding to ∆f ) -5 (curve b) and -15 Hz (curve d) are observed after rinsing with PBS buffer and interaction with anti-LPS Ab, respectively. Figure 6 (curve e) shows the response of the sensing interface upon treatment with C. trachomatis-lacking urine, and curve f shows the response of the electrode after rinsing with the BPS buffer solution and injection of the anti-LPS Ab. No frequency changes in the electrode response are observed, consistent with the fact that no antigen bacteria were present in the analyzed sample. Thus, the recommended protocol for the analysis of C. trachomatis involves the interaction of the sensing interface with the analyte urine sample, followed by rinsing the cell with the PBS buffer solution and injecting the anti-LPS Ab. A frequency Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
3511
Figure 6. Time-dependent frequency changes of a functionalized crystal, pretreated with urine, upon interaction with fresh urine samples that include variable concentrations of C. trachomatis and subsequent amplification/confirmation of the primary analysis step by interaction of the crystal with anti-C. trachomatis LPS Ab. Sensing interface is composed of goat IgG-anti-mouse IgG, Fc-specific Ab and anti-C. trachomatis LPS Ab. (a) Treatment of electrode with C. trachomatis in urine, 260 ng mL-1. (b) Subsequent treatment of electrode formed in (a) with anti-C. trachomatis LPS Ab, 175 ng mL-1 in PBS. (c) Treatment of crystal with C. trachomatis, 2.6 µg mL-1 in urine. (d) Treatment of electrode produced in (c) with anti-C. trachomatis LPS Ab, 175 ng mL-1 in PBS. (e) Treatment of sensing crystal with a sterile urine sample. (f) Treatment of the crystal resulting in (e) with the antiC. trachomatis LPS Ab, 175 ng mL-1 in PBS. The crystals were rinsed with the PBS buffer solution, 10 mL, at 2 mL min-1, after the analysis of the urine samples and prior to interaction with the anti-C. trachomatis LPS Ab.
Figure 7. Frequency changes of a series of crystals functionalized by two different routes as a function of storage time: (b) goat IgGanti-mouse IgG Ab/anti-C. trachomatis LPS Ab, and (9) fragmented F(ab′)2 anti-mouse IgG Ab/anti-C. trachomatis LPS Ab. Crystals were stored in the dry state at 4 °C. Frequency measurements were performed by interaction of the crystals with C. trachomatis in PBS, 260 ng mL-1.
change of ∆f ≈ -5 Hz in the second step should be considered as a positive indication for the presence of the antigen bacteria in the analyte urine sample at a concentration of ∼260 ng mL-1. This sensitivity limit is superior to that of the present ELISA tests. Also, the detection of the bacteria in urine samples, and the detection time, represent important advantages of the described method.
3512
Analytical Chemistry, Vol. 69, No. 17, September 1, 1997
Finally, we assessed the stability of the tailored electrodes. The performance of the two types of immobilization of the anti-C. trachomatis sensing interfaces, namely, linkage of the antibody to the goat anti-mouse IgG antibody, or association to the F(ab′)2fragmented goat anti-mouse IgG antibody (cf. Scheme 1), were examined as a function of storage time. The electrodes were stored in sealed vials at 4 °C. Figure 7 shows the crystal frequency changes of the electrode in the presence of C. trachomatis at 2.6 µg mL-1, at different durations of storage. The electrodes that contain the base layer of goat anti-mouse IgG antibody show a degradation with time, while the electrodes where the base antibody consists of fragmented F(ab′)2 goat anti-mouse IgG antibody reveal unaltered activity for a period of 90 days of storage. CONCLUSIONS The present study assembled immunosensor interfaces for the microgravimetric, quartz crystal microbalance detection of Chlamydia trachomatis specimens in urine. A specific sensing interface for the antigen bacteria was designed, and an appropriate analysis protocol was evaluated. The preferred configuration of the sensing interface involves the covalent attachment of the fragmented F(ab′)2 goat anti-mouse Ab to a base maleimide function monolayer assembled onto the Au electrodes associated with the quartz crystal. Subsequent attachment of the anti-C. trachomatis LPS Ab to the base antibody layer, in the presence of 0.5% BSA (5 mg mL-1), followed by treatment of the electrode with pure urine, yields the specific interface for the antigen bacteria. The analysis protocol includes the injection of the analyte urine sample to the sensor, followed by rinsing the cell with PBS buffer and treating the electrode with the anti-LPS Ab. The frequency changes of the crystal are followed in the two steps, i.e., interaction with the urine sample and with anti-LPS Ab. A frequency decrease in the second step, i.e., addition of the anti-LPS Ab, that corresponds to ∆f ) -5 Hz, indicates the presence of the C. trachomatis in the analyte sample at a concentration of ∼260 ng mL-1, which is considered as the lower detection limit of the present devices. The microgravimetric method represents a relatively rapid and sensitive method for the analysis of C. trachomatis. The degradation of the sensing interface only upon its interaction with positive analyte samples represents a further advantage of this biosensor device. ACKNOWLEDGMENT This research is supported by Savyon Diagnostics, Ltd., Ashdod, Israel, and the Ministry of Commerce and Industry, Israel.
Received for review February 24, 1997. Accepted May 27, 1997.X AC970216S X
Abstract published in Advance ACS Abstracts, August 1, 1997.
