Surface Plasmon Resonance (SPR) Biosensor Development Charles T. Campbell

A. Background Surface plasmon resonance (SPR) sensing has been demonstrated in the past decade to be

an exceedingly powerful and quantitative probe of the interactions of a variety of biopolymers with various ligands, biopolymers, and membranes, including protein:ligand, protein:protein, protein:DNA and protein: membrane binding. It provides a means not only for identifying these interactions and quantifying their equilibrium constants, kinetic constants and underlying energetics, but also for employing them in very sensitive, label-free biochemical assays [1-36].

In a typical SPR biosensing experiment, one interactant in the interactant pair (i.e., a ligand or biomolecule) is immobilized on an SPR-active gold-coated glass slide which forms one wall of a thin flow-cell, and the other interactant in an aqueous buffer solution is induced to flow across this surface, by injecting it through this flow-cell (Fig. 1). When light (visible or near infrared) is shined through the glass slide and onto the gold surface at angles and wavelengths near the so-called “surface plasmon resonance” condition, the optical reflectivity of the gold changes very sensitively with the presence of biomolecules on the gold surface or in a thin coating on the gold. The high sensitivity of the optical response is due to the fact that it is a very efficient, collective excitation of conduction electrons near the gold surface. The extent of binding between the solution-phase interactant and the immobilized interactant is easily observed and quantified by monitoring this reflectivity change (Fig. 2) An advantage of SPR is its high sensitivity without any fluorescent or other labeling of the interactants.

Fig. 1. Basic components of an instrument for SPR biosensing: A glass slide with a thin gold coating is mounted on a prism. Light passes through the prism and slide, reflects off the gold and passes back through the prism to a detector. Changes in reflectivity versus angle or wavelength give a signal that is proportional to the volume of biopolymer bound near the surface. A flow cell allows solutions above the gold surface to be rapidly changed.

Fig. 2. A typical SPR biosensing experiment, showing the optical response versus time: The gold surface with immobilized interactant starts in pure buffer at time 0. At 100 s,solution containing the other interactant is introduced. At 300 s, the flow cell is flushed with pure buffer, and at 420-520 s, the starting surface is regenerated with a sequence of reagents.

Such SPR biosensing has found a very wide variety of applications, has contributed data to thousands of scientific publications on biomolecular interactions, and has enjoyed substantial commercial success. The leader of this SPR Core (Campbell) has published a number of papers using SPR biosensing [37-43], has developed and studied surface-functionalization strategies for SPR sensing [38-40, 43-48], and has developed widely-used data-analysis techniques for SPR sensing [49]. The most popular commercial instruments for SPR biosensing are those with trademark Biacore [2, 3].

Campbell’s group has been active in technology development for SPR biosensor applications, always using home-made SPR sensor systems. These studies include: 1). Development of the first simple method for absolute quantitative analysis of binding amounts based on SPR response (e.g., determination of the ratio of the number of bound protein molecules per immobilized target molecule), published as a cover article in Langmuir [49] and now widely used by those developing SPR technology. 2). Kinetic and equilibrium studies using SPR of the binding of wild-type (WT) and several mutants of streptavidin (SA) to biotin-terminated alkylthiols immobilized in a self-assembled monolayer on a gold surface in a mixture with oligoethyleneglycol-terminated alkylthiols [37, 38]. The kinetics of competitive dissociation of the bound SA (Fig. 7) proved that, when the biotins are at low concentration in the monolayer, SA and its mutants bind to the surface through a single biotin with a strength much like that for free biotin in homogeneous solution, but when present at moderate concentrations (10-30%), SA binds to the surface via two biotins with approximately twice the binding energy. Fig. 7. SPR curves showing the on- / off-rates of wild-type (WT) streptavidin and three mutants to biotins, immobilized to the gold sensor surface using alkylthiolate links (in a mixed monolayer). Right: Off-rate half lives for the different mutants plotted versus the fraction of biotin in the mixed monolayer. From [37, 38]. This optimized monolayer offers an even distribution of free biotin sites on the surface, since each adsorbed streptavidin has two free biotin-binding sites pointing away from the solid. (The other two sites point toward the solid and are blocked upon anchoring the monolayer to the solid.) This SA linker layer is extremely stable and offers a high density of free biotin binding sites (one site pointing toward the solution phase every 4.5 nm x 4.5 nm, on average [37, 38, 45]) which we have used very successfully for further SPR sensor biofunctionalization (see below). More recently, we developed a method for storing this linker layer in air and maintaining its functionality [43]. 3). Application of the SA linker layer described above to study the binding of peripheral-membrane proteins to intact lipid vesicles with SPR [39]. Here, 80 nm lipid vesicles prefunctionalized with a few percent biotinylated lipids were attached to the biotin sites in this SA monolayer. We have shown through studies with phospholipase A2 (PLA2) binding to these immobilized vesicles (Fig. 8) that they provide an excellent template for rapid and quantitative studies of peripheral membrane protein

binding equilibria and kinetics. The vesicles give more biologically relevant values than planar membrane models, which we show give inaccurate protein binding constants. Fig. 8. Determination of the equilibrium dissociation constant (6±2x10-7 M) by measuring the amount of phospholipase bound (f) to a phosphatidylcholine vesicle membrane using SPR (from [39]). 4). Application of the SA linker layer described above to study protein binding to dsDNA using [40]. The dsDNA was biotinylated on one end and attached to the free biotin sites in the SA monolayer. The binding of Mnt proteins to a 107 base-pair segment of dsDNA containing the 21 base-pair Mnt operator sequence and to unrelated sequences showed very sequence-specific binding. 5). Development of other surface-functionalization strategies for SPR biosensing [44, 46-48]. 6). Development of a framework for analyzing adsorption kinetics from liquid solutions (e.g., in protein binding to surface-immobilized species) to eliminate diffusion effects and determine the probability of binding per collision with the surface [50, 51]; and application of this method to a fundamental study of protein binding to an immobilized ligand using SPR [38]. B. Develop High Throughput SPR Biosensing in Array Format Usinf SPR Microscopy

The studies described in the previous section can be performed in the simplest manifestation of SPR, wherein only a single interaction is probed at any given time. However, our recent experiments show that SPR biosensing measurements can be made in array format with 120 interactions measured simultaneously. That is, we have proven that 120 SPR binding and dissociation curves like in Figs. 2, 3 and 6 can be measured simultaneously using SPR microscopy and a computer-interfaced video camera to probe the interactions of a protein with all 120 elements in a 10x12 array of spots on a sensor surface [42]. In our proof-of-principle experiments, the spots were functionalized with different sequences of dsDNA, and the probe was a transcription factor (i.e., a DNA-binding protein), but in principle the spots could be functionalized with many other types of interactants . This means that SPR measurements such as illustrated in Figs. 3-8 to determine equilibrium constants and kinetic constants can be made in two orders-of-magnitude higher throughput. Furthermore, the array format offers a very important advantage in that it allows a much more reliable method to remove spurious signals due to non-specific binding and changes in the index of refraction of the solvent (due either to concentration or temperature changes during

f

Glass A

injection). This is done by subtracting the signal from a nearby control spot on the array. Example data from such an array are shown in Figs. 11-12. A very important aspect of this array format is its sensitivity to a smaller number of adsorbed molecules than normal SPR biosensing. In the above example, we demonstrated sensitivity to <0.5 pg (<2x107 molecules) of bound protein (MW = 21 kD) in each array spot at a time resolution of 1 s. This improved sensitivity means that, in principle, much less sample is required to get the necessary signal. A further advantage of the array format is its potential for application for high-throughput analyses of the concentrations of biobolymers in complex mixtures. If half of the 10x12 array above were spotted with receptors for 60 different biopolymers, and half were used for control spots to subtract out background changes due to non-specific binding, this array could serve as a format for simultaneously measuring the concentrations of 60 different biopolymers in a complex sample mixture. For example, consider an array of dsDNAs wherein 60 spots have sequences offering the binding sites for 60 different DNA-binding proteins like transcription factors, and the other 60 spots are control sequences. Exposure of this array to a mixture containing many of these DNA-binding proteins (e.g., a pre-purified nuclear extract) would result in signals at the different array elements which could identify which proteins are present, and, if repeated at different levels of dilution, could tell their concentrations. This is similar to to a recent report wherein a modified Affymetrix ssDNA array chip was modified to make an array of dsDNAs [52]. When combined with fluorescence microscopy, this was used to probe multiple protein:DNA interactions simultaneously using fluorescently-tagged proteins. The major advantage of the proposed SPR array approach over this approach is the lack of need for fluorescent labeling when using SPR.

Figure 11. Our SPR microscope’s CCD camera image and line profile of a dsDNA array under buffer solution. A) A 10x12 array of dsDNA fabricated on a streptavidin array on a biotin-functionalized, gold-coated glass slide using a robotic microspotter. “Operator” refers to rows of spots which have a dsDNA sequence containing the binding site for the Gal4 protein. “Control” refers to spots with other DNA sequences but no binding site. “SA only” refers to spots with only the streptavidin (SA) linker layer, but not functionalized with any biotinylated dsDNA. LEFT: initial array. RIGHT: intensity difference image, after Gal4 binding from solution. B) A line-profile of the SPR optical signal for a vertical line passing through the fourth row of spots in the left-hand image of (A). All data are the result of averaging for 1 s the camera frames collected by the CCD camera at a rate of 30 Hz. From [42, 48, 53].

A)

B)

Distance, µm0 2000 4000 6000

Cha

nge

in %

Ref

lect

ed In

tens

ity

0

5

10

15

20

25

30

35

1 23

4 56 7 8

9 10

SA only

BAT/OEG

SA + dsDNA

SA

BAT/OEG

SA + dsDNA

SA

BAT/OEG

SA + dsDNA

SA

A BControl

Control

Operator

Operator

Control

Control

Operator

Operator

SA only

SA only

A B A BControl

Control

Operator

Operator

Control

Control

Operator

Operator

SA only

SA only

Control

Control

Operator

Operator

Control

Control

Operator

Operator

SA only

SA only

A B

Figure 12. Demonstration of simultaneous, real-time measurement of the kinetics of the binding of a protein (the transcription factor Gal4) to a 120 spots on a dsDNA array using SPR microscopy, with sensitivity of 0.5 pg (107 molecules) and 1 s time resolution. For clarity, we show data for only six representative spots on the dsDNA array surface, but similar data were obtained simultaneously for all 48 spots of the array which were functionalized with the dsDNA operator sequence for Gal4 binding. Before Gal4 injection, the surface was in protein-free buffer. The adsorption / removal curves show the SPR response in each such array element after subtracting the response measured at a nearby Control-dsDNA spot, to eliminate signal contributions due to non-specific Gal4 binding and changes in the index of refraction of the buffer solutions. The curves were obtained by integrating the reflected intensity over 200 µm x 200 µm areas within each array spot on the sensor surface and camera-frame-averaging for 1 s. From [42] We will also develop a 10-channel microfabricated flow cell system that allows flowing 10 different solutions simultaneously across the surface and monitoring simultaneously the SPR response in all 10 channels with SPR microscopy. This will allow equilibrium constant and kinetic rate constant measurements such as in Figs. 3-6 to be performed with 10 different solutions concentrations simultaneously in one experiment. It also reduces the amount of solution required. When coupled to a 10x12 array such as described above, this will allow equilibrium constant and kinetic rate constant measurements such as in Figs. 3-6 to be performed on 12 different targets with 10 different solutions concentrations simultaneously in one experiment.

Note that this 10-channel microfluidics system also will offer an excellent way to identify and quantify the involvement of cofactors in binding. One simply repeats the above measurements at different concentrations of the cofactor to determine its influence on the binding constant, binding energy or rate parameters. Alternatively, one uses the same protein concentration but a different concentration of cofactor in each of the 10 flow cells, and tests in detail the effect of cofactor concentration in one experiment. In principle, this could be done simultaneously with a 12 different targets by coupling the 10 channels to a 10x12 array.

Our SPR microscope follows in many ways the design by Prof. W. Knoll (U. Mainz) [54-58].

0

2

4

6

8

10

12

14

0 200 400 600 800 1000Time, seconds

Gal

4 C

over

age,

1011

Gal

4/cm

2

Inject Gal4

Buffer rinse

Fig. 13. Design of our SPR microscope (from [42, 48, 53]).

We have recently demonstrated quantitative, real-time measurement of kinetics of sequence-specific binding of DNA-binding proteins to double-stranded DNA (dsDNA) immobilized in a 10x12 array on a planar gold surface using this SPR microscope [42]. We studied the binding of the yeast transcription factor Gal4 to a 120-spot dsDNA array made with alternating 200-µm spots of its dsDNA operator sequence and an unrelated DNA sequence. Example data was shown in Figs. 11-12 above. The results proved that this method could be used to simultaneously monitor the kinetics of binding of proteins to 120 different dsDNA sequences with sensitivity to <0.5 pg (<2x107 molecules) of bound protein in each array spot at a time resolution of 1 s. The method is label free and also allows absolute quantitative determination of the binding stoichiometry (i.e., the number of proteins bound per dsDNA) at each time. The dsDNA array was fabricated using a robotic microspotting system to deliver nanoliter droplets of biotinylated dsDNA solutions onto a streptavidin linker layer immobilized with biotinylated alkylthiols on a thin gold film (see below). Simultaneous monitoring of binding to the many array elements allows the use of reference spots (i.e., array elements with unrelated dsDNA sequences) to correct the signal for non-specific protein-DNA binding and changes in bulk refractive index of the solutions in the SPR microscope’s flow cell. This allows high-throughput analyses of the kinetics and equilibrium of protein-dsDNA binding.

To perform the above study we developed strategies for microspotting arrays of double-stranded DNAs (dsDNAs) on a gold-coated glass slide for high-throughput studies of protein – DNA interactions by surface plasmon resonance (SPR) microscopy [48]. The methods use streptavidin (SA) as a linker between a biotin-containing mixed self-assembled monolayer (SAM) and biotinylated dsDNAs to produce arrays with high packing density (0.5-1.2 x 1012 dsDNA/cm2). The primary mixed SAM is produced from biotin- and oligo(ethylene glycol)-terminated thiols (BAT and OEG, resp.) bonded as thiolates onto the gold surface. First, a robotic microspotting system is used to deliver nanoliter droplets of dsDNA in solution on a uniform layer of SA pre-adsorbed onto this BAT/OEG SAM. SPR microscopy analysis shows that the binding ratio is ~0.2 dsDNA per SA for most of the

array elements. We discovered some limitations of this method and found an alternative strategy that uses instead a microspotted SA array created on the mixed SAM rather than the uniform monolayer. SPR microscopy analysis shows uniform, nearly close-packed SA coverage (2.1 to 2.4 x 1012 SA/cm2) in the spots. We used this SA array to immobilize dsDNA by microspotting dsDNA solution directly onto each SA array element and demonstrate a dsDNA packing density of 8.5±3.5 x 1011 dsDNA/cm2 and a binding ratio (0.5±0.2 dsDNA per SA) that is consistent for the length of dsDNA used.

Our group developed accurate but simple formulae for quantitative analysis of adsorption amounts based on SPR response [49] which are widely used today by those developing SPR technology. We recently extended these formulae to quantitative analysis in the SPR microscopy mode [53]. This method was necessary for the success of the above studies. It provides a very simple means to convert local reflectivity changes measured in SPR microscopy to effective adlayer thicknesses and absolute surface coverages of adsorbed species. For a range of angles near the SPR resonance where the local metal surface’s reflectivity changes linearly with angle, the reflectivity at fixed angle varies with time proportional to the change in effective refractive index (ηeff) near the surface. This change in ηeff can be converted to absolute adsorbate coverage versus time using methods we developed earlier for quantitative SPR spectroscopy [49]. The sensitivity factor for the system to bulk changes in refractive index, i.e., % reflectivity change per refractive index unit (RIU), is the only calibration required. Applications of this method to study protein adsorption from aqueous solution with an SPR microscope operating at 633 nm shows a detection limit of 0.072 % change in absolute reflectivity for simultaneous measurements of all 200 µm x 200 µm areas within the 16 mm2 light beam with 1 s time resolution. This corresponds to a change in effective refractive index of 1.8 x 10-5 and a detection limit for protein adsorption of 1.2 ng/cm2 (~0.5 pg in a 200 µm spot). For a 60 kD protein, this is <5x106 molecules. The linear dynamic range is ∆ηeff = ~0.011 RIU or ~720 ng/cm2 of adsorbed protein. Using a nearby spot as a reference channel one can correct for instrumental drift, which is important when measuring slow adsorption processes near the detection limit. In principle, the same method can be applied to measurements approaching the resolution limits of SPR microscopy (~4 µm). We have already developed a 9-channel microfabricated flow cell system that allows flowing 9 different solutions simultaneously across the surface and monitoring simultaneously the SPR response in all 9 channels with our SPR microscope. It is shown in Fig. 14, along with an SPR microscope image using it.

Fig. 14. Design of our 9-channel flow cell and SPR microscope image with 9-channel flow cell in operation. Each microchannel is only 100 micrometers across, and the total cell volume is only 30nL. In principle, this cell allows equilibrium constant and kinetic rate constant measurements such as in Figs. 3-6 to be performed with 9 different solutions concentrations simultaneously in one experiment.

It can also be used to prefunctionalize the surface with 9 different target solutions. Its low volume (90 pL) also greatly reduces the amount of solution required, both for targets and probes.

To make using this flow cell with 9 different concentrations more convenient, we have also fabricated a laminar microfluidic diffusion diluter following earlier designs [59, 60]. This device allows us to generate 9 different protein concentrations as output with only two input streams (e.g., pure buffer and a solution of high protein concentration in buffer). This will greatly simplify quantitative determinations of equilibrium constants and rate constants, or cofactor effects on the same. REFERENCES: 1. Liedberg, B., I. Lundstrom, and E. Stenberg, Principles of Biosensing with an Extended

Coupling Matrix and SPR. Sensors and Actuators, 1993. B 11: p. 63. 2. Rich, R.L. and D.G. Myszka, Survey of the year 2001 commercial optical biosensor literature.

Journal of Molecular Recognition, 2002. 15(6): p. 352-376. 3. Rich, R.L. and D.G. Myszka, Survey of the year 2000 commercial optical biosensor literature.

Journal of Molecular Recognition, 2001. 14(5): p. 273-294. 4. Rich, R.L. and D.G. Myszka, Advances in surface plasmon resonance biosensor analysis.

Current Opinion in Biotechnology, 2000. 11(1): p. 54-61. 5. McDonnell, J.M., Surface plasmon resonance: towards an understanding of the mechanisms of

biological molecular recognition. Current Opinion in Chemical Biology, 2001. 5(5): p. 572-577. 6. Rich, R.L., et al., High-resolution and high-throughput protocols for measuring drug/human

serum albumin interactions using BIACORE. Analytical Biochemistry, 2001. 296(2): p. 197-207.

7. Heierhorst, J., et al., Synapsins as major neuronal Ca2+/S100A1-interacting proteins. Biochemical Journal, 1999. 344: p. 577-583.

8. Kuwahara, H., et al., A novel NE-dlg/SAP102-associated protein, p51-nedasin, related to the amidohydrolase superfamily, interferes with the association between NE-dlg/SAP102 and N-methyl-D-aspartate receptor. Journal of Biological Chemistry, 1999. 274(45): p. 32204-32214.

9. Cotman, S.L., W. Halfter, and G.J. Cole, Identification of extracellular matrix ligands for the heparan sulfate proteoglycan agrin. Experimental Cell Research, 1999. 249(1): p. 54-64.

10. Hutchings, N.J., et al., Linking the T cell surface protein CD2 to the actin-capping protein CAPZ via CMS and CIN85. Journal of Biological Chemistry, 2003. 278(25): p. 22396-22403.

11. Forbes, B.E., et al., Characteristics of binding of insulin-like growth factor (IGF)-I and IGF-II analogues to the type 1 IGF receptor determined by BIAcore analysis - Correlation of binding affinity with ability to prevent apoptosis. European Journal of Biochemistry, 2002. 269(3): p. 961-968.

12. Luo, J., et al., Antibody-antigen interactions measured by surface plasmon resonance: Global fitting of numerical integration algorithms. Journal of Biochemistry, 2001. 130(4): p. 553-559.

13. Roy, M.O., et al., Study of hydrophobic interactions between acylated proteins and phospholipid bilayers using BIACORE. Journal of Molecular Recognition, 2001. 14(1): p. 72-78.

14. Quinn, J.G. and R. O'Kennedy, Biosensor-based estimation of kinetic and equilibrium constants. Analytical Biochemistry, 2001. 290(1): p. 36-46.

15. Bamdad, C., A DNA self-assembled monolayer for specific attachment of unmodified double- or single-stranded DNA. Biophysical Journal, 1998. 75: p. 1997-2003.

16. Buckle, M., et al., Real time measurements of elongation by a reverse transcriptase using surface plasmon resonance. Proc. Natl. Acad. Sci., 1996. 93: p. 889-894.

17. Kok, R.G., D.A. D'Argenio, and L.N. Ornston, Mutation analysis of PobR and PcaU, closely related transcriptional activators in acinetobacter. J. Bacteriology, 1998. 180(19): p. 5058-5069.

18. Leeuwen, H.C.v. and M.J. Strating, Linker length and composition influence the flexibility of Oct 1 DNA binding. Embo Journal, 1997. 16(8): p. 2043-2053.

19. Malmborg, A.C., et al., Real-time analysis of Oct protein-octamer interaction and transcription complex assembly. Molecular Immunology, 1995. 32(17-18): p. 1429-1442.

20. Jakimowicz, D., et al., Structural elements of the streptomyces oriC region and the interactions with DnaA protein. Microbiology, 1998. 144(Pt. 5): p. 1281-1291.

21. Bondeson, K., O. Ronn, and G. Magnusson, DNA binding of polyomavirus large T-antigen: kinetics of interactions with different types of binding sites. FEBS Letters, 1998. 423(6): p. 307-313.

22. Fisher, R.J., et al., Sequence-specific binding of human immunodeficiency virus type 1 nucleocapsid protein to short oligonucleotides. Journal of Virology, 1998. 72(3): p. 1902-1909.

23. Yamamoto, A., et al., Difference in affinity for DNA between HMG proteins 1 and 2 determined by surface plasmon resonance measurements. J. Biochem-Tokyo, 1997. 122(3): p. 586-94.

24. Galio, L., et al., Analysis of interactions between huGATA-3 transcription factor and three GATA regulatory elements of HIV-1 long terminal repeat, by surface plasmon resonance. Anal. Biochem., 1997. 253(1): p. 70-77.

25. Gotoh, M., et al., Rapid method for detection of point mutations using mismatch binding protein (MutS) and an optical biosensor. Genetic Analysis, 1997. 14(2): p. 47-50.

26. Haruki, M., et al., Kinetic and stoichiometric analysis for the binding of Escherichia coli ribonuclease HI to RNA-DNA hybrids using surface plasmon resonance. J. Biol. Chem., 1997. 272(5): p. 22015-22022.

27. Pond, C.D., et al., Surface plasmon resonance analysis of topoisomerase I-DNA binding; effect of Mg2+ and DNA sequence. Anti-Cancer Drugs, 1997. 8(4): p. 336-344.

28. Seimiya, M. and Y. Kurosawa, Kinetics of binding of Antp homeodomain to DNA analyzed by measurements of surface plasmon resonance. FEBS Letters, 1996. 398(2-3): p. 279-284.

29. Caruso, F., E. Rodda, and D.N. Furlong, Quartz Crystal Microbalance Study of DNA Immobilization and Hydridization for Nucleic Acid Senor Development. Anal. Chem., 1997. 69(11): p. 2043-2049.

30. Caruso, F., et al., DNA Binding and Hybridization on Gold and Derive Surfaces. Sensors and Actuators B Chem., 1997. 41: p. 189-197.

31. Persson, B. and e. al., Analysis of Oligonucleotide Probe Affinities Using Surface Plasmon Resonance: A Means for Mutational Scanning. Anal. Biochem., 1997. 246: p. 34-44.

32. Jordan, C.E., et al., Surface Plasmon Resonance Imaging Measurements of DNA Hybridization Adsorption and Streptavidin / DNA Multilayer Formation at Chemically Modified Gold Surfaces. Anal. Chem., 1997. 69: p. 4939-4947.

33. Watts, H.J., D. Yeung, and H. Parkes, Real-Time Detection and Quantification of DNA Hybridization by Optical Biosensor. Anal. Chem., 1995. 67: p. 4283-4289.

34. Lucosz, W., Principles and Sensitivities of Integrated-optical and SPR Sensors for Direct Affinity Sensing and Immunoassay. Biosensors and Bioelectronics, 1991. 6: p. 215.

35. Lucosz, W., Integrated-optical and SPR Sensors for Direct Affinity Sensing: II. Anisotropy of Adsorbed Adlayers. Biosensors and Bioelectronics, 1997. 12: p. 175.

36. Lundstrom, I., Real-time Biospecific Interaction Analysis. Biosensors and Bioelectronics, 1994. 9: p. 725.

37. Jung, L.S., et al., SPR Measurements of Binding and Disassociation of Wild-Type and Mutant Streptavidin on Mixed Biotin-Containing Alkyl-Thiolate Monolayers. Sensors and Actuators B, 1999. 54: p. 137-144.

38. Jung, L.S., et al., Binding and dissociation kinetics of wild-type and mutant streptavidins on mixed biotin-containing alkylthiolate monolayers. Langmuir, 2000. 16(24): p. 9421-9432.

39. Jung, L.S., et al., Quantification of tight binding to surface-immobilized phospholipid vesicles using surface plasmon resonance: Binding constant of phospholipase A(2). Journal of the American Chemical Society, 2000. 122(17): p. 4177-4184.

40. Shumaker-Parry, J.S., et al. Probing Protein:DNA Interactions Using a Uniform Monolayer of DNA and Surface Plasmon Resonance. in Proceedings of SPIE Photonics West Conference, International Biomedical Optics Symposium. 2000. San Jose, CA: SPIE.

41. Lu, H.B., et al., Protein contact printing for a surface plasmon resonance biosensor with on-chip referencing. Sensors and Actuators B-Chemical, 2001. 74(1-3): p. 91-99.

42. Shumaker-Parry, J., R.A. Aebersold, and C.T. Campbell, Parallel, Quantitative Measurement of Protein Binding to a 120-Element Double-Stranded DNA Array in Real Time Using SPR Microscopy. Anal. Chem., subm.

43. Xia, N., et al., A Streptavidin Linker Layer that Functions after Drying,. Langmuir, subm. 44. Lu, H.B., C.T. Campbell, and D.G. Castner, Attachment of functionalized poly(ethylene glycol)

films to gold surfaces. Langmuir, 2000. 16(4): p. 1711-1718. 45. Nelson, K.E., et al., Surface characterization of mixed self-assembled monolayers designed for

streptavidin immobilization. Langmuir, 2001. 17(9): p. 2807-2816. 46. Lu, H.B., et al. Apatite Coated Surface Plasmon Resonance Biosensors. in Sixth World

Biomaterials Congress. 2000. Kamuela, Hawaii, USA: Society for Biomaterials, Minneapolis, MN.

47. Lu, H.B., et al. Surface Functionalization for Self-referencing Surface Plasmon Resonance Biosensors by RF-plasma-deposited Thin Films and Self-assembled Monolayers. in Sixth World Biomaterials Congress. 2000. Kamuela, Hawaii, USA: Society for Biomaterials, Minneapolis, MN.

48. Shumaker-Parry, J., et al., Microspotting Streptavidin and Double-Stranded DNA Arrays on Gold for High-Throughput Studies of Protein - DNA Interactions by SPR Microscopy. Anal. Chem., subm.

49. Jung, L.S., et al., Quantitative Interpretation of the Response of Surface Plasmon Resonance Sensors to Adsorbed Films,. Langmuir, 1998. 14: p. 5636-5648.

50. Jung, L.S. and C.T. Campbell, Sticking probabilities in adsorption of alkanethiols from liquid ethanol solution onto gold. Journal of Physical Chemistry B, 2000. 104(47): p. 11168-11178.

51. Jung, L.S. and C.T. Campbell, Sticking probabilities in adsorption from liquid solutions: Alkylthiols on gold. Physical Review Letters, 2000. 84(22): p. 5164-5167.

52. Bulyk, M.L., et al., Quantifying DNA-protein interactions by double-stranded DNA arrays. Nature Biotechnology, 1999. 17: p. 573-7.

53. Shumaker-Parry, J. and C.T. Campbell, Quantitative Methods for Spatially-Resolved Adsorption / Desorption Measurements in Real Time by SPR Microscopy. Anal. Chem., subm.

54. Zizlsperger, M. and W. Knoll, Multispot parellel on-line monitoring of interfacial binding reactions by surface plasmon microscopy. Progr. Colloid Polym. Sci., 1998. 109: p. 244-253.

55. Knoll, W., Interfaces and thin films as seen by bound electromagnetic waves. Ann. Rev. Phys. Chem., 1998. 49: p. 1z.

56. Aust, E.F., et al., Surface plasmon and guided optical wave microscopies. Scanning, 1994. 16: p. 353-361.

57. Piscevic, D., et al., Oligonucleotide hydridization observed by surface plasmon optical techniques. Appl. Surf. Sci., 1995. 90: p. 425-436.

58. Hickel, W. and W. Knoll, Surface plasmon optical characterization of lipid monolayers at 5 µm lateral resolution. J. Appl. Phys., 1990. 67(8): p. 3572-3575.

59. Holden, M.A., et al., Generating fixed concentration arrays in a microfluidic device. Sensors and Actuators B-Chemical, 2003. 92(1-2): p. 199-207.

60. Jeon, N.L., et al., Generation of solution and surface gradients using microfluidic systems. Langmuir, 2000. 16(22): p. 8311-8316.