Antibody Detection in Human Serum Using a Versatile Protein Chip Platform Constructed by Applying Nanoscale Self-Assembled Architectures on Gold Petra Pavlickova,*,† Nina Mejlhede Jensen,‡ Hubert Paul,† Michael Schaeferling,† Chiara Giammasi,† Margit Kruschina,† Wei-Dong Du,† Michael Theisen,‡,§ Michael Ibba,‡ Flavio Ortigao,† and Dev Kambhampati*,† ThermoHybaid, Sedanstrasse 18, 890 77 Ulm, Germany, and Statens Serum Institut, Departments of Tuberculosis Immunology and Clinical Biochemistry, Artillerivej 5, 2300 Copenhagen, Denmark Received January 29, 2002
We report a novel high-throughput (HTP) protein chip platform, constructed on gold using self-assembly techniques, for conducting high quality antigen-antibody interactions. Biotinylated monolayers were used to immobilize a streptavidin surface with high packing density. This biocompatible platform was then used for detection of serum IgM antibodies. Serum samples of patients suspected to suffer from Lyme borreliosis were used to validate the protein chip platform using biotinylated peptide AAOspC8 molecules as the test probes. Various experimental parameters such as the effect of concentration of probes, targets, temperature of incubation, and their effect on the resulting signal-to-noise ratio are described in detail. Highly specific protein interaction data with a high signal-to-noise ratio were obtained with serum sample solutions as low as 1 µL/spot (1/10 diluted). Keywords: protein chip • antibody • self-assembled monolayer • microarray • gold • Lyme borreliosis
Introduction The development and application of microarrays has led to significant advances in modern biotechnology and medicine. Using DNA microarrays, gene expression profiles were attained for Arabidopsis thaliana,1 Saccharomyces cerevisiae,2,3 and human lymphoid cells and tissues such as bone marrow, brain, prostate, and heart.4 The use of oligonucleotide arrays has recently been reviewed.5 Although DNA microchips are interesting tools for deciphering gene expression behavior, this technique has not been able to reliably monitor protein expression levels within the cells. In recent years, only a limited number of studies have been performed to analyze protein-protein, protein-DNA, or protein-RNA interactions in a high-throughput fashion.6-9 Protein (peptide) chip technology has several advantages over conventional methods. First, the chip-based assays enable rapid analysis of a large number of samples in a single experiment.10 Second, the amount of material needed is very small. Reaction volumes could be 20-40 times lower than the amount that is generally used in conventional 384-microwell (microtiter) plates. Third, the signal-to-noise ratio exhibited using microarrays is much better (>10-fold) than that observed for traditional microtiter plate assays.11 * To whom correspondence should be addressed. Tel.: +49 (0) 731 93579 670. Fax: +49 (0) 731 93579 291. E-mail: (P.P.)
[email protected], (D.K.)
[email protected]. † ThermoHybaid. ‡ Statens Serum Institut, Department of Tuberculosis Immunology, Statens Serum Institut. § Department of Clinical Biochemistry, Statens Serum Institut. 10.1021/pr0200036 CCC: $22.00
2002 American Chemical Society
Most of the current analyses of protein interactions on microarrays are based on epoxy, aldehyde, or poly-lysinecoated glass slides.12,13 Although, the glass slides have the advantage that they can be used directly using conventional assays and detection methods, they have several limitations. Microarray smearing is a common artifact found in most glassbased arrays, and it is one of the methods by which cross-talk between the adjacent probes occurs, leading to false results. The need to block the free reactive surface of the chips to minimize nonspecific interactions or an occurrence of nonhomogeneous spot areas could also pose a problem on such slides. The biochip platform developed here consists of a streptavidin sensor surface that is assembled on to a gold layer using nanoscale biotinylated self-assembled architectures. The selfassembled architectures are formed spontaneously upon the interaction of an active group from solution (thiols, disulfides, silanes, etc.) with an appropriate substrate (metal-gold, silver, or oxidic surfaces).14 Self-assembled monolayers (SAM)s on gold are easy to fabricate, chemically stable under laboratory conditions, and flexible in introducing various chemical functional groups, resulting in different surface properties. The applicability of this technology for detection of specific antibodies in human serum samples was analyzed using a small synthetic peptide antigen derived from the C-terminus of outer (1) Abbreviations used: OspC, outer surface protein C of B. burgdorferi; AAOspC8, 8-mer C-terminal peptide of OspC in B. burgdorferi; VlsE, variable surface antigen; FITC, fluorescent isothiocyanate; HTP, high throughput; rt, room temperature; SAM, self-assembled monolayer.
Journal of Proteome Research 2002, 1, 227-231
227
Published on Web 04/18/2002
research articles
Pavlickova et al.
surface protein C (AAOspC8) of Borrelia burgdorferi,15,16 the causative agent of Lyme borreliosis. In the protein chip assay, biotinylated synthetic peptide was coupled to a streptavidin surface followed by interaction with serum IgM antibodies from Lyme borreliosis suspect patients. The use of the biotinylated derivate of the OspC protein has been described previously.17 The antigen-antibody biomolecular interactions on the protein chip were monitored using fluorescence detection schemes. A wide range of experimental parameters such as the dynamic range of the biotinylated probe, reproducibility, times of incubation, and temperature that influence the efficacy of the biochip assays were investigated and subsequently optimized for detection of serum IgM antibodies. This biochip strategy resulted in a rapid detection of specific serum antibodies.
Experimental Section Material. Biotinylated peptide AAOspC8 (Biotin-AAVAESPKKP) was from a previous study.18 Patient sera (n ) 41) were from a routine screen (Staten Serum Institut, Copenhagen, Denmark) of Lyme borreliosis suspect patients and were found to contain IgG against the Borrelia flagellum. Six serum samples from healthy donors served as negative controls. Protein Chip Platform. The protein chip platform, XNA on Gold (ThermoHybaid, Ulm, Germany) is based on a highly defined nanoscale biocompatible sensor architecture constructed by using the principles of self-assembly techniques on gold. Biotin termined self-assembled monolayers are used to immobilize a streptavidin layer with high packing density (Figure 1A). The interaction between biotin and streptavidin is strong (KD ∼1 × 1015 M-1), ensuring the integrity of the protein chip surface under a wide range of experimental conditions. The streptavidin sensor surface can be used to couple any biomolecule (nucleic acid, protein, lectin, sacharide, lipid, etc.) that is conjugated with biotin. The protein chips are designed on the basis of a standard microscope slide format. Two sets of 96 or 384 spots (i.e., 192 or 768 spots in total) are available for conducting biomolecular interactions. The individual gold spots are separated from each other by a Teflon layer. The hydrophobic Teflon surface prevents cross-talk between the adjacent spots and virtually eliminates microarray smearing. The chips are compatible with fluorescence, mass spectrometry and radio-imaging detection schemes. Immobilization of Biotinylated AAOspC8 Probe on Streptavidin-Coated XNA on a Gold Biochip. AAOspC8 probes were dissolved in PBS-T buffer (120 mM NaCl, 2.7 mM KCl, 10 mM phosphate buffer pH 7.4, 0.1% Tween-20 (v/v)) to obtain a series of concentrations in the range of 0.1-25 ng/mL (1 µL of biotinylated reagent/spot). The surface immobilization step was conducted within a plastic case with moistened tissue (humid chamber) to minimize evaporation effects. After spotting, the humidified case was covered with a tight lid and incubated at rt or at 37 °C for 30, 60, or 120 min or alternatively at 4 °C with overnight incubation. After the immobilization of AAOspC8 probes, the chip was rinsed with sterile water, followed by washing 3 × 5 min in PBS-T with gentle agitation. Thereafter, the chips were rinsed with sterile water and dried under nitrogen. The dried chips were immediately used for immunodetection of IgM antibodies present within the sera of Lyme borreliosis suspect patients. Immunodetection with IgM from Sera of Lyme Borreliosis Suspect Patients and Healthy Subjects. After the antigenimmobilized protein chip was dried, the human sera diluted 228
Journal of Proteome Research • Vol. 1, No. 3, 2002
Figure 1. Schematic of the protein chip platform (A). Biotinylated self-assembled architectures are used to immobilize a streptavidin sensor surface with high packing density. All kinds of biotinylated probes can then be immobilized onto this surface. Schematic of the detection step (B). A sandwich assay employing anti-human IgM-FITC antibody was used to detect indirectly the interaction between the biotinylated AAOspC8 probes and the IgM antibodies present within the serum samples.
in 1/10 PBS-T (1 µL of total volume/spot) were transferred on to the spots containing the immobilized AAOspC8 probes (Figure 1B). Incubation was carried out at rt or at 37 °C for 30, 60, or 120 min in a humid case. The chips were subsequently rinsed with sterile water, followed by washing 3 × 5 min in PBS-T with gentle agitation. Following the above-mentioned step, the chips were subsequently rinsed with sterile water and dried under nitrogen. The drying of the antibody-antigenimmobilized chip has no significant effect on the biomolecular interaction (data not shown). Antibody binding was detected with FITC-conjugated rabbit anti-human IgM antibody (DAKO,
Antibody Detection in Human Serum
research articles
Figure 2. Quantification of biotinylated peptide antigen AAOspC8 using the protein microarray platform. Biotinylated AAOspC8 antigen was immobilized on the streptavidin surface at different concentrations (0-25 ng/mL, 1 µL final volume) at rt for 2 h, followed by incubation with five different sera from Lyme borreliosis suspect patients (a-e) and two sera from healthy donors (f, g) (diluted 1/10 in PBS-T, 1 µL final volume). Human IgM antibodies were recognized by rabbit anti-human IgM antibody labeled with FITC (diluted 1/40 in PBS-T, 1 µL final volume).
Copenhagen, Denmark) diluted in 1/40 PBS-T immediately after the drying step. After incubation at rt for 90 min in a dark environment, the chips were rinsed with sterile water, washed 3 × 5 min in PBS-T, and scanned using a fluorescent scanner (XNA ScanPro 20 microarray scanner, ThermoHybaid, Ashford, U.K.). The microarray analysis software, AIDA 2.11, from Raytest (Straubenhard, Germany) was used to analyze the image files.
Results Quantification of Biotinylated Peptide Antigen AAOspC8 Probes Using the Protein Microarray Platform. The biotinylated 8-mer peptide (AAOspC8) from the C-terminus of the outer surface protein C (OspC) was used as an antigen for detection of specific antibodies in serum samples from patients suspected of Lyme borreliosis. The AAOspC8 samples at different concentrations (1 µL final volume) were spotted on the streptavidin coated biochips. The dynamic range was demonstrated to be 1-15 ng/mL corresponding to 1-15 fmol of AAOspC8 probes per spot (Figure 2). All samples were spotted in duplicate on each array. No blocking reagent was used either during antigen immobilization or prior to target incubation on the chips. AAOspC8 antigen probes were incubated with five different serum samples from Lyme borreliosis suspect patients and two serum samples from healthy subjects (1 µL final volume), which were used as negative controls. The patient serum samples and negative samples were chosen from the group of 41 sera of Lyme borreliosis suspect patients and six sera of healthy donors. The intensity of the FITC signal corresponding to the anti-human IgM antibody was found to vary in accordance with the concentration of the surfaceimmobilized AAOspC8 probe. Concentration higher than 25 ng/ mL had no significant effect on the final fluorescence signal intensity. Optimization of Time and Temperature Conditions for AAOspC8 Probe Immobilization and Incubation with Serum Antibodies. AAOspC8 samples (5 ng/mL, 1 µL final volume)
Figure 3. Optimization of time and temperature conditions for AAOspC8 probe immobilization and incubation with serum antibodies. Biotinylated AAOspC8 antigen (5 ng/mL, 1 µL final volume) was immobilized on the chips at rt or at 37 °C for 30, 60, or 120 min or alternatively at 4 °C with overnight incubation (A). The incubation with serum of Lyme borreliosis suspect patient was performed at rt for 2 h in this case. The optimization of the incubation time of the serum from Lyme borreliosis suspect patient was performed at rt or at 37 °C for 30, 60, or 120 min (B). AAOspC8 antigen (5 ng/mL, 1 µL final volume) probes were immobilized on the chips at rt for 30 min in this experiment. Human IgM antibodies were recognized by rabbit anti-human IgM antibody labeled with FITC (diluted 1/40 in PBS-T, 1 µL final volume).
were immobilized on to a microarray for 30 min, 60 min, 120 min, or overnight. The immobilization of AAOspC8 probes for 30, 60, or 120 min was performed at two different temperature conditions: rt and 37 °C while the AAOspC8 overnight capture was carried out only at 4 °C. The results are shown in Figure 3A. The time required for AAOspC8 probe immobilization was found to be independent of the temperatures (rt, 37 °C). In contrast, the effect of target incubation times was found to be quite substantial. The maximum fluorescence signal was observed when the serum IgM target was incubated on the AAOspC8-immobilized chip for 2 h at rt conditions. A dramatic drop in fluorescence intensity was found in all experiments that were performed at 37 °C, Figure 3B. On the basis of the above observation, all subsequent probe immobilizations were performed at rt for 30 min while target incubation steps were carried out at rt for 2 h. Protein Chip Reproducibility. The effect of the chip-to-chip variability was tested by repeating each experiment four times on four different days. The experiment was based on incubation Journal of Proteome Research • Vol. 1, No. 3, 2002 229
research articles
Pavlickova et al.
Table 1. Reproducibility Analysis of the Protein Chipa experiment no. (LAU/mm2)
human serum
PS1 PS2 PS3 PS4 PS5 PS6 PS7 PS8 PS9 PS10 NS1 NS2 NS3 NS4 PBS-T
1
2
3
4
avg
SD
CV (%)
59.9 14.3 51.6 14.2 57.3 17.5 32.3 14.1 36.6 35.8 5.4 8.0 6.1 8.6 4.3
57.5 14.3 49.2 11.6 60.1 16.9 29.6 13.6 39.5 39.7 6.4 7.6 5.8 7.9 4.6
61.0 12.3 57.1 12.6 64.2 16.0 28.6 13.3 40.0 41.9 6.3 7.4 6.8 8.4 5.0
60.2 15.0 54.8 14.5 60.4 18.5 35.2 15.3 37.3 39.1 6.4 9.2 6.1 8.5 5.1
59.6 14.0 53.2 13.2 60.5 17.2 31.4 14.1 38.4 39.1 6.1 8.1 6.2 8.4 4.7
1.3 1.0 3.0 1.2 2.4 0.9 2.6 0.8 1.5 2.2 0.5 0.7 0.4 0.3 0.3
2.2 7.3 5.7 9.0 4.0 5.2 8.2 5.3 3.8 5.6 7.3 8.7 6.0 3.2 6.8
a The effect of the chip-to-chip variability was tested by repeating each experiment four times on 4 different days. Ten different serum samples from Lyme borreliosis suspect patients (PS1-10) and four serum samples from healthy donors (NS1-4) were used. The CV for the fluorescence signals ranged from 2.2 to 9.0%.
of IgM antibodies from serum samples of Lyme borreliosis suspect patients with AAOspC8 antigen (4 µg/mL, 1 µL final volume) probe. In this assay, 10 serum samples of Lyme borreliosis suspect patients and four negative controls were chosen from the group of 41 positive and six negative serum samples. All dilutions and sample solutions were freshly prepared on the day of experiment. On the same chip, all test analyses were performed in duplicate and the average fluorescence signals and standard deviations were calculated for each sample. The CV (coefficient of variation) for the fluorescence signals obtained from the values on the different biochips ranged between 2.2% and 9.0% (Table 1). Screening of 41 Sera from Lyme Borreliosis Suspect Patients Using a Protein Chip. The ability of the microarray to monitor the presence of antibodies against synthetic peptide antigen AAOspC8 was investigated using 41 different sera of patients suspected with Lyme borreliosis. Six sera of healthy people and a placebo test sample using PBS-T buffer were used as negative controls. A high fluorescence signal was obtained in approximately 30% of investigated patient serum samples compared to the fluorescence signals of the negative controls (Figure 4).
Discussion The main goal of this work was to examine the applicability of protein chips for detection of serum antibodies. One of the main advantages of the biochip technology is the minute amount of peptide antigen and human serum sample that is required to perform a single test. Only 1 µL of reagent, instead of 100 µL that is normally used for a standard ELISA test, is required for performing a biochip assay. The usual concentration of protein and peptide antigens for serodiagnosis of Lyme borreliosis in standard ELISA ranges between 0.4 and 5 µg/mL16,19 using total volumes of 100-200 µL per assay. Our results show that high-quality results can be obtained by using the AAOspC8 probe in nanogram concentrations (Figure 2). Increasing the concentration of AAOspC8 to more than 25 ng/ mL had no significant effect on the intensity of the fluorescence signal, indicating that saturation levels were reached even at such low concentrations. Since the probe is localized to a small 230
Journal of Proteome Research • Vol. 1, No. 3, 2002
Figure 4. Screening of 41 sera of patient suspected with Lyme borreliosis using a protein chip. The AAOspC8 antigen probes (4 µg/mL, 1 µL final volume) were immobilized on the chips at rt for 30 min. Immunodetection was performed by 41 different serum samples from Lyme borreliosis suspect patients (nos. 1-41) and six serum samples of healthy persons (nos. 43-48) (diluted 1/10 in PBS-T, 1 µL final volume) at rt for 2 h. A single PBS-T placebo test (no. 42) was used in this experiment. Human IgM antibodies were recognized by rabbit anti-human IgM-FITC conjugate (diluted 1/40 in PBS-T, 1 µL final volume).
area on the protein chip, soluble antibodies are able to interact rapidly with the probe without any mass transport or steric hindrance effects. In conventional ELISA, the dimensions of the microtiter plates inhibit fast biomolecular interaction, and as a result, the time required for reaching saturation is significantly higher than those observed in the case of microarrays. The binding of antibody to AAOspC8 antigen probe was temperature dependent since a dramatic drop in fluorescence signals was observed when the immobilized probe was incubated with the serum samples at 37 °C as compared to incubation at rt (Figure 3B). This difference in reactivity might be related to temperature-induced structural changes in either the antigen or the antibody. On the contrary, the AAOspC8 antigen probe immobilization can be performed either at rt or at 37 °C without any significant difference in the fluorescence signal (Figure 3A). We have used 8-mer peptide antigen AAOspC8 to detect IgM antibodies from serum samples of patients suspected to suffer from Lyme borreliosis since the AAOspC8 is specifically recognized by patient antibodies and may thus be used for serodiagnosis of this disease.16 In earlier reports, it was shown that the humoral immune response against OspC in patients with Lyme borreliosis is primarily of the IgM type, and very few patients have IgG antibodies against the C-terminal region
research articles
Antibody Detection in Human Serum
of OspC.16,20 The results presented here corroborate this fact as only IgM antibodies were detected in the human sera samples. A serological test for Lyme borreliosis should be based on more than one antigen probe to ensure higher specificity.21 Recently, a variable surface antigen (VlsE) from B. burgdorferi containing a conserved immunodominant region named IR6 (26-mer peptide) was described to be a suitable marker for the serodiagnosis of Lyme borreliosis.22 However, the potential benefit of combining AAOspC8 with other antigens derived from B. burgdorferi in a single assay remains to be investigated. Diagnostic assays based on peptides have several advantages over assays based on protein antigens. Peptides can be synthesized at relatively low costs even in large scale, while proteins must be purified from biological sources, which is a very expensive and time-consuming process. The use of biotin-streptavidin technology on SAM enables highly specific interaction and reproducible results. The CV was found to be between only 2.2 and 9.0% in a chip-top-chip variability test (Table 1). SAMs on gold also have the advantage that the surface properties can be efficiently characterized using optical, mechanical, and electrochemical analytical tools.14 Protein microarray technology provides a powerful and versatile tool for the genome-scale analysis of gene function, such as enzyme activity, protein-protein, and protein-nucleic acid interaction and small-molecule drug interactions, directly on the protein level. The technology has advantage of tiny amount of protein in each assay without increasing of material concentration. We believe that for routine use, serology is likely to be the most useful and popular approach for the laboratory diagnosis of Lyme borreliosis and other diseases. For most routine clinical diagnostics, only a few hundred spots are required for analysis. Our microarray is well suited to this application. Highly reproducible data with high signal-to-noise values can be obtained on the streptavidin coated chips. Protein microarrays are poised to play an important role in the diagnosis of diseases and will be routinely applied in HTP protein interaction studies soon.
Acknowledgment. This work was supported by Grant No. QLG-CT-99-00660 within the Fifth Framework Programm of the European Commission.
References (1) Schena, M.; Shalon, D.; Davis, R. W.; Brown, P. O. Science 1995, 270, 467-470. (2) Wodicka, L.; Dong, H.; Mittmann, M.; Ho, M. H.; Lockhart, D. J. Nat. Biotechnol. 1997, 15, 1359-1367. (3) Chu, S.; DeRisi, J.; Eisen, M.; Mulholland, J.; Botstein, D.; Brown, P. O.; Herskowitz, I. Science 1998, 282, 699-705. (4) Schena, M.; Shalon, D.; Heller, R.; Chai, A.; Brown, P. O.; Davis, R. W. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 10614-10619. (5) Arcellana-Panlilio, M.; Robbins, S. M. Am. J. Physiol. Gastrointest. Liver Physiol. 2002, 282, G397-402. (6) Bi, C.; Ernst, O. P.; Heyse, S.; Hofmann, K. P.; Vogel, H. Nat. Biotechnol. 1999, 17, 1105-1108. (7) Arenkov, P.; Kukhtin, A.; Gemmell, A.; Voloshchuk, S.; Chupeeva, V.; Mirzabekov, A. Anal. Biochem. 2000, 278, 123-131. (8) Fung, E. T.; Thulasiraman, V.; Weinberger, S. R.; Dalmasso, E. A. Curr. Opin. Biotechnol. 2001, 12, 65-69. (9) Zhu, H.; Snyder, M. Curr. Opin. Chem. Biol. 2001, 5, 40-45. (10) Uetz, P.; Giot, L.; Cagney, G.; Mansfield, T. A.; Judson, R. S.; Knight, J. R.; Lockshon, D.; Narayan, V.; Srinivasan, M.; Pochart, P.; Qureshi-Emili, A.; Li, Y.; Godwin, B.; Conover, D.; Kalbfleisch, T.; Vijayadamodar G.; Yang, M.; Johnston, M.; Fields, S.; Rothberg, J. M. Nature 2000, 403, 623-627. (11) Zhu, H.; Klemic, J. F.; Chang, S.; Bertone, P.; Casamayor, A.; Klemic, K. G.; Smith, D.; Gerstein, M.; Reed, M. A.; Snyder, M. Nat. Gen. 2000, 26, 283-289. (12) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760-1763. (13) Haab, B. B.; Dunham, M. J.; Brown, P. O. Gen. Biol. 2001, 2, 4.113. (14) Bain, C. D.; Whitesides, G. M. Angew. Chem., Int. Ed. 1989, 101, 522-528. (15) Bockenstedt, L. K.; Hodzic, E.; Feng, S.; Bourrel, K. W.; Silva, A.; Montgomery, R. R.; Fikrig, E.; Radolf, J. D.; Barthold, S. W. Infect. Immun. 1997, 65, 4661-4667. (16) Mathiesen, M. J.; Holm, A.; Christiansen, M.; Blom, J.; Hansen, K.; Ostergaard, S.; Theisen, M. Infect. Immun. 1998, 66, 40734079. (17) Mathiesen, M. J.; Christiansen, M.; Hansen, K.; Holm, A.; Asbrink, E.; Theisen, M. J. Clin. Microbiol. 1998, 36, 3474-3479. (18) Gregorius, K.; Theisen, M. Anal. Biochem. 2001, 299, 84-91. (19) Liang, F. T.; Steere, A. C.; Marques, A. R.; Johnson, B. J. B.; Miller, J. N.; Philipp, M. T. J. Clin. Microbiol. 1999, 37, 3990-3996. (20) Mathiesen, M. J.; Hansen, K.; Axelsen, N.; Halkier-Sorensen, L.; Thiesen, M. Med. Microbiol. Immunol. (Berlin) 1996, 185, 121129. (21) Nowakowski, J.; Schwartz, I.; Liveris, D.; Wang, G.; AgueroRosenfeld, M. E.; Girao, G.; McKenna, D.; Nadelman, R. B.; Cavaliere, L. F.; Wormser, G. P. Clin. Infect. Dis. 2001, 33, 20232027. (22) Jones, K.; Guidry, J.; Wittung-Stafshede, P. Biochem. Biophys. Res. Commun. 2001, 289, 389-394.
PR0200036
Journal of Proteome Research • Vol. 1, No. 3, 2002 231
