Combinational Application of Surface Plasmon Resonance

To better understand the interactions between estrogen receptor (ER) subtypes and the estrogen response elements (EREs), we employed surface plasmon ...
0 downloads 0 Views 229KB Size
Anal. Chem. 2006, 78, 5552-5558

Combinational Application of Surface Plasmon Resonance Spectroscopy and Quartz Crystal Microbalance for Studying Nuclear Hormone Receptor-Response Element Interactions Xiaodi Su,*,† Chin-Yo Lin,‡,§ Sean J. O'Shea,† Huey Fang Teh,† Wendy Y. X. Peh,† and Jane S. Thomsen*,‡

Institute of Materials Research and Engineering, 3 Research Link, Singapore 117602, and Genome Institute of Singapore, 60 Biopolis Street, Singapore 138672

Conventional methodologies for studying protein-DNA complexes, such as electrophoretic mobility shift assays (EMSAs), lack the real-time sensitivity and precision to accurately characterize the complex dynamics of interactions between transcription factors and their binding sites. To better understand the interactions between estrogen receptor (ER) subtypes and the estrogen response elements (EREs), we employed surface plasmon resonance (SPR) spectroscopy and quartz crystal microbalance with dissipation measurement (QCM-D) and made the following observations: (1) base substitutions in ERE half-sites reduced binding affinity for both ERr and ERβ, (2) ERr has a higher sequence specificity than ERβ or there were more nonspecific interactions between ERβ and control DNA, and (3) ERr bound ERE as dimers and ERβ bound as tetramers. These findings highlight intrinsic differences in DNA-binding properties between receptor subtypes, which are not apparent based on the high degree of conservation (96% identity) in their DNA-binding domains and results from EMSA studies. With this study, we demonstrate the potential of utilizing SPR and QCM in combination for a comprehensive characterization of ERDNA interactions, including sequence-dependent binding mechanisms and structural differences in ERr-DNA and ERβ-DNA complexes. Estrogen receptors (ERs) are members of the nuclear receptor superfamily of transcription factors that regulate genes responsible for development and maintenance of reproductive tissues as well as general maintenance of many other physiological functions.1-3 The interaction of ERs with DNA sequences known as estrogen response elements (EREs) is required for estrogen regulation of * To whom all correspondence should be addressed. E-mail: [email protected]. Tel: 65-68748420. Fax: 65-68720785. E-mail: [email protected]. Tel: 65-64788179. Fax: 65-64789060. † Institute of Materials Research and Engineering. ‡ Genome Institute of Singapore. § Current address: Department of Microbiology and Molecular Biology, Brigham Young University, Provo, UT 84602. (1) Deroo, B. J.; Korach, K. S. J. Clin. Invest. 2006, 116, 561-570. (2) Green, P. S.; Simpkins, J. W. Int. J. Devl. Neurosci. 2000, 18, 347-358. (3) Jansson, L.; Holmdahl, R. Inflammation Res. 1998, 47, 290-301.

5552 Analytical Chemistry, Vol. 78, No. 15, August 1, 2006

target gene expression.4 Two human ER subtypes, estrogen receptor R (ERR) and estrogen receptor β (ERβ), have been identified. They are arranged into similar domains, where the degree of homology varies widely among the regions (being highest in the DNA-binding domain, 96% amino acid identity).5 While many of the biochemical properties are similar, they differ substantially in their tissue distribution. Understanding their DNAbinding ability would allow us to gain insight if they have differential biological function and tissue-selective actions. To characterize the binding behavior of ERs with various ERE targets, e.g., consensus ERE that contains palindromic repeats separated by a three-base spacer (5′-GGTCAnnnTGACC-3′) and imperfect ERE containing base substitution(s), people have largely relied on an electrophoretic mobility shift assay,6-11 a gel filtration chromatography assay,6,7 a fluorescence anisotropy assay,12-14 and recently an integrated genomewide molecular and computational approach.15 Most of these methods, however, use equilibrium binding measurements and involve tedious assay procedures, e.g., sample labeling using hazardous materials and a long period of sample incubations, and uses special laboratory conditions (e.g., radioactivity) for final data collection. (4) Klinge, C. M. Nucleic Acids Res. 2001, 29, 2905-2919. (5) Muramatsu, M.; Inoue, S. Biochem. Biophys. Res. Commun. 2000, 270, 1-10. (6) Melamed, M.; Arnold, S. F.; Notidess, A. C.; Sasson, S. J. Steroid Biochem. Mol. Biol. 1996, 57, 153-159. (7) Cheskis, B. J.; Karathanasis, S.; Lyttle, C. R. J. Biol. Chem. 1997, 272, 11384. (8) Cowley, S. M.; Hoare, S.; Mosselman, S.; Parker, M. G. J. Biol. Chem. 1997, 272, 19858-19862. (9) Hyder, S. M.; Chiappeta, C.; Stancel, G. M. Biochem. Pharmacol. 1999, 57, 597-601. (10) Loven, M. A.; Wood, J. R.; Nardulli, A. M. Mol. Cell. Endocrinol. 2001, 181, 151-163. (11) Yi, P.; Driscoll, M. D.; Huang, J.; Bhahat, S.; Hilf, R.; Bambara, R. A.; Muyan, M. Mol. Endocrinol. 2002, 16, 674-693. (12) Ozers, M. S.; Hill, J. J.; Erivin, K.; Wood, J. R.; Nardulli, A. M.; Royer, C. A.; Gorski, J. J. Biol. Chem. 1997, 272, 30405-30411. (13) Boyer, M.; Poujol, N.; Margeat, E.; Royer, C. A. Nucleic Acids Res. 2000, 28, 2494-2502. (14) Margeat, E.; Bourdoncle, A.; Margueron, R.; Poujol, N.; Cavaille`s, V.; Royer, C. A. J. Mol. Biol. 2003, 326, 77-92. (15) Lin, C. Y.; Stro ¨m, A.; Vega, V. B.; Kong, S. L.; Yeo, A. L.; Thomsen, J. S.; Chan, W. C.; Doray, B.; Bangarusamy, D. K.; Ramasamy, A.; Vergara, L. A.; Tang, S.; Chong, A.; Bajic, V. B.; Miller, L. D.; Gustafsson, J.-Å.; Liu, E. T. Genome Biol. 2004, 5, R66. 10.1021/ac0606103 CCC: $33.50

© 2006 American Chemical Society Published on Web 06/24/2006

Table 1. Sequences of the DNA Samples Involved in This Study name (denoted as)

sequence

wild-type ERE (wt-ERE) mutant ERE (mut-ERE) non ERE (non-ERE)

5′-biotin-GTCCAAAGTCAGGTCACAGTGACCTGATCAAAGT-3′ 5′-biotin-GTCCAAAGTCAGTTCACAGTGATCTGATCAAAGT-3′ 5′-biotin-GTCCAAAGTCAATCGCCAGCACGATGATCAAAGT-3′

Surface plasmon resonance (SPR) spectroscopy and quartz crystal microbalance (QCM) are surface-sensitive technologies, capable of label-free and real-time monitoring of biomolecular interactions. One of the current trends in SPR and QCM research is to use a combined data collection and analysis from parallel SPR and QCM measurements to obtain complementary details of a particular binding event.16-22 This becomes possible as SPR spectroscopy and QCM are based on different physical principles, with each method being sensitive to different properties of the materials studied. SPR spectroscopy detects changes in the refractive index of thin films assembled on a noble-metal surface and can quantitate the amount/capacity of adsorbed materials.7,23,24 QCM, particularly QCM with dissipation monitoring (QCM-D), monitors both mass and structure properties of adsorbed molecules. New opportunities include the study of conformational changes/differences enabling, for example, distinction between two similar binding events or recording of protein unfolding.25 Although SPR spectroscopy and QCM techniques have been widely used for studying biomolecular interactions, including protein-DNA interactions,18,23,24,26,27 their usefulness in nuclear hormone receptor biology research has not been extensively demonstrated. There are a few examples of SPR characterization of ERR-consensus ERE interactions,7,28 ERR- and ERβ-coactivator interactions,29 and ligand-ERR interactions.30,31 There are no reports of QCM (or QCM-D) studies of ER-DNA interactions. In this present work, we further explore the potentials of SPR spectroscopy and the QCM-D technique for a comprehensive characterization of ERR- and ERβ-DNA interactions. We aim (16) Ho ¨o ¨k, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796-5804. (17) Reimhult, E.; Larsson, C.; Kasemo, B.; Ho ¨o ¨k, F. Anal. Chem. 2004, 76, 7211-7220. (18) Su, X. D.; Wu, Y.-J.; Knoll, W. Front. Biosci. 2005, 10, 268-274. (19) Su, X. D.; Wu, Y.-J.; Knoll, W. Biosens. Bioelectrons. 2005, 21, 719-726. (20) Larsson, C.; Rodahl, M.; Ho ¨o ¨k, F. Anal. Chem. 2003, 75, 5080-5087. (21) Zhou, C.; Friedt, J. M.; Angelova, A.; Choi, K. H.; Laureyn, W.; Frederix, F.; Francis, L. A.; Campitelli, A.; Engelborghs, Y.; Borghs, G. Langmuir 2004, 20, 5870-5878. (22) Carrigan, S. D.; Scott, G.; Tabrizian, M. Biomaterials 2005, 26, 7514-7523. (23) Smith, E. A.; Erickson, M. G.; Ulijasz, A. T.; Weisblum, B.; Corn, R. M. Langmuir 2003, 19, 1486-1492. (24) Maillart, E.; Brengel-Pesce, K.; Capela, D.; Roget, A.; Livache, T.; Canva, M.; Levy, Y.; Soussi, T. Oncogene 2004, 23, 5543-5550. (25) Zelander, G. Nat. Methods 2006, 41-42 (Suppl). (26) Su, X. D.; Robelek, R.; Wu, Y.-J.; Knoll, W. Anal. Chem. 2004, 76, 489494. (27) Shumaker-Parry, J. S.; Aebersold, R.; Campbell, C. T. Anal. Chem. 2004, 76, 2071-2082. (28) Asano, K.; Ono, A.; Hashimoto, S.; Inoue, T.; Kanno, J. Anal. Sci. 2004, 20, 611-616. (29) Wa¨rnmark, A.; Almlo¨f, T.; Leers, J.; Gustafsson, J-°A.; Treuter, E. J. Biol. Chem. 2001, 276, 23397-23404. (30) Rich, R. L.; Hoth, L. R.; Geoghegan, K. F.; Brown, T. A.; LeMotte, P. K.; Simons, S. P.; Hensley, P.; Myszka, D. G. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 8562-8567. (31) Usami, M.; Mitsunaga, K.; Ohno, Y. J. Steroid Biochem. Mol. Biol. 2002, 81, 47-55.

to discover whether there are differences in the DNA-binding behavior between the two ER subtypes, despite the high degree of conservation in their DNA-binding domain.5 Using the ERDNA interactions system, we demonstrate how SPR and QCM analysis of protein-DNA binding alleviates many of the problems associated with the gel shifts method and provide additional characterizations. EXPERIMENTAL SECTION Materials. Purified recombinant human ERR and ERβ were purchased from PanVera Corp. (Madison, WI). They were stocked in HEPES buffer containing 10% glycerol. The stock concentration is 2800 nM for ERR and 4500 nM for ERβ. For long-term storage, the ER proteins were stored in aliquots of 10 µL at -80 °C. Before use, they were thawed in a room-temperature water bath and returned to 4 °C to maintain the activity. Three 34-bp, double-stranded oligonucleotides with a biotin label at the 5′ end of one strand were synthesized by Proligo Primers & Probes (Boulder, CO). Table 1 shows the sequences of the biotinylated strands of these DNA. The wild-type ERE (denoted as wt-ERE) carries the palindromic GGTCA half-site (underlined) with a 3-bp separation. The imperfect (or mutant) ERE (denoted as mut-ERE) contains a symmetric base substitution (boldface) in each of the ERE half-sites. The non-ERE has the sequence in the ERE arms scrambled. The flanking sequence remains unchanged to ensure a similar GC content of ∼40% as the wt- and mut-ERE DNA. The biotinylated strands and the antistrands were They were annealed in phosphate-buffered saline (PBS, pH 7.4, Sigma P-4417), containing 10 mM EDTA, pH 7.5, and stored at -27 °C. SPR Measurement. The SPR measurements were conducted using the AutoLab ESPR (Eco Chemie, The Netherlands), a double-channel and prism coupling-based instrument. A goldcoated glass disk mounted on a prism through a thin layer of index-matching oil form the base of a two-channel cuvette. Different samples can be added into the two independent channels. In a kinetic measurement mode, molecular adsorption on gold-coated glass disks is followed by monitoring SPR angle (θ) or angle shifts (∆θ) over time. The measured ∆θ values correspond to the amount of adsorbed material with a mass sensitivity factor of 120 mdeg per 100 ng/cm2. The measurements were conducted at room temperature, and the noise level in ∆θ measurement was 0.5 mdeg. Quartz Crystal Microbalance with Energy Dissipation. The QCM-D measurements were conducted using a Q-sense instrument (Q-Sense AB, Go¨teborg, Sweden). This instrument allows for a simultaneous measurement of resonance frequency change (∆f) and energy dissipation change (∆D) by periodically switching off the driving power of the oscillation of the sensor crystal and recording the decay of the damped oscillation. The 5-MHz, ATcut quartz crystals (Q-Sense AB) were used as the reaction Analytical Chemistry, Vol. 78, No. 15, August 1, 2006

5553

Scheme 1. Schematic Illustration of the Assay Proceduresa

a SA is first immobilized on a biotin-containing thiol treated surface for biotinylated ERE assembly. ER proteins are then applied to bind to the immobilized DNA repeatedly upon regeneration of the immobilized DNA using 0.1% SDS.

carriers. The QCM-D setup allows for measurements of f and D at four harmonics (fundamental frequency and 15, 25, and 35 MHz, corresponding to the overtones, n ) 3, 5, 7, respectively) of the 5-MHz crystal. For clarity, only the normalized frequency shift (∆fnormalized ) ∆fn/n) and the dissipation shift, ∆D, for the third overtone is presented. The measurements were conducted at room temperature, and the noise in f and D under the liquid load was 0.3 Hz and 0.2 × 10-6, respectively. Sensor Disk Preparation and Assay Procedures. Scheme 1 illustrates the DNA assembly and protein-binding procedures involved in this study. Biotin-streptavidin (SA)-biotin bridge chemistry was used for the assembly of the biotinylated EREs. For this purpose, freshly cleaned SPR and QCM gold disks (by UV/ozone for 10 min followed by hot piranha solution for 2 min; be cautious!) were immersed overnight in a binary biotincontaining thiol mixture.15,21 After rinsing with ethanol followed by a drying step using nitrogen, the disks were ready to use. Streptavidin immobilization was from a 0.1 mg/mL SA solution in PBS buffer. The DNA assembly was from a 50 nM or 1 µM solution, to achieve a lower or a maximal DNA surface density, respectively. ER protein binding buffer is 40 mM HEPES-KOH, pH 7.4, containing 10 mM MgCl2, 0.2% Triton X-100, 2 mM DTT, and 100 or 200 mM KCl. At the end of each reaction, representative buffer solutions were introduced to replace the reaction solutions and to rinse the liquid cell. After one cycle of protein binding, 0.1% SDS (w/v in water) was applied to disassociate the protein-DNA complex and to expose the immobilized DNA for new cycles of ER binding. RESULTS AND DISCUSSION SPR Determination of Sequence-Dependent ER-DNA Interactions. Protein-DNA interactions play a central role in transcriptional regulation and other biological processes. Investigating sequence-dependent binding affinity and specificity in protein-DNA complexes is thus an important goal. Figure 1A shows the binding curves of ERR and ERβ to wt- and mut-ERE immobilized in different SPR channels. The SPR responses were recorded from SA immobilization and ERE assembly. At the end of the SA immobilization and ERE assembly, rinsing the liquid cell with PBS buffer leads to no detectable angle shift, showing that there are no loosely attached molecules. Application of ER protein solutions, after resetting the baseline in HEPES buffer (protein binding buffer), leads to a rapid SPR angle shift in the 5554 Analytical Chemistry, Vol. 78, No. 15, August 1, 2006

Figure 1. Discriminative ER interactions with different ERE sequences. (A) SPR responses from two channels were recorded for SA immobilization (0.1 mg/mL in PBS) and wt-ERE (solid line) and mut-ERE (dashed line) assembly (1 µM in PBS). After resetting the baseline using HEPES buffer containing 100 mM KCl, ERR or ERβ at varied concentrations were applied to react for ∼30 min. The V arrows indicate the time when the reaction solutions were replaced by corresponding buffer solutions. A short treatment of the surface using 0.1% SDS (dashed V arrows) dissociates the protein-DNA complex, leaving the DNA layer exposed for new cycles of protein binding. (B) A display of summary of ER binding amount (∆θ in mdeg) to the wt- and mut-ERE targets. Data were collected from a series of experiments conduced at varied DNA surface density and ER concentrations ranging from 25 to 300 nM.

first few seconds, followed by a much slower response. The rapid angle shift is due mainly to the change of bulk refractive index, as the ER protein solution contains a certain amount of glycerol (see Materials). At the end of protein binding, the application of HEPES buffer (containing no glycerol) washes away loosely attached ER protein (if there is any) and corrects the bulk buffer effect. Control experiments were conducted to introduce ER solutions to the SA-modified surface with no immobilized ERE (curve not shown). A similar bulk refractive index effect was observed, and the effect was found entirely reversible, which

Figure 2. Salt concentration effects on specific and nonspecific ER-DNA interactions. Overlaid sensorgrams were obtained by applications of ERR (A, B) and ERβ (C, D) over immobilized wt-, mut-, and non-ERE targets. Experiments were conduced in HEPES buffer containing either 100 (A, C) or 200 mM KCl (B, D). DNA immobilization was from 1 µM solutions. ER concentration was fixed at 110 nM.

indicates that there is no nonspecific protein adsorption on the surface. Thus, the ∆θ values recorded before and after ER binding are specifically related to receptor binding to DNA. We found that, in contrast to the SA immobilization and ERE DNA assembly steps, where identical SPR angle shift (∆θ) from the two channels are obtained (variation 70% for ERR and ∼50% for ERβ, whereas for the perfect wt-ERE, the protein binding (both ERR and ERβ) reduces by only ∼20%. The stronger salt concentration dependence for the interactions with the mut- and nonERE DNA indicates that these interactions are primarily due to sequence-independent electrostatic contacts between the negatively charged sugar-phosphate backbones of the DNA and the basic amino acid residues on the protein (at pH 7.4 both ERR and ERβ are slightly net positively charged). At a higher salt concentration, the salt ions compete with the proteins for the charge-charge interactions, thus disrupting the interactions.13,14 As for the specific ER-wt-ERE interactions, the complexes are formed mainly through multiple hydrogen bonds and van der Waals forces.35,36 The binding may be supported and stabilized by a smaller number of charge-charge interactions.37,38 Thus, the depletion in the interaction when increasing salt concentration occurs to a smaller degree. The destabilization of nonspecific electrostatic interactions at 200 mM KCl enhances the ER binding stringency and leads to a higher ER specificity, giving a ∆θER-wt-ERE/ ∆θER-mut-ERE ratio of ∼8 for ERR and ∼3 for ERβ. Results in Figure 2 also shows that there are more nonspecific interactions between ERβ and the control ERE targets (mut- and non-ERE) than ERR, both at 100 and 200 mM salt concentration. Understanding nonspecific ER-DNA interactions is important as they are more than just experiment nuisance that introduces complicity in sequence specificity. They also play roles in facilitating protein to search for its specific target sequence at a faster rate.37,38 SPR Determination of ER-ERE Binding Stoichiometry. Protein-DNA binding stoichiometry is an important parameter related to the binding mechanisms. To determine ER-ERE binding stoichiometry, we monitored protein amount that saturated immobilized DNA through titration experiments. Figure 3 shows the binding of ERR to immobilized wt-ERE at concentrations ranging from 54 to 700 nM. To ensure the protein successfully saturate the immobilized DNA at a reasonable concentration, a lower DNA surface density (49 mdeg, obtained from the assembly of ERE from a 50 nM solution) was used. Control (35) Jen-Jacobson, L.; Engler, L. E.; Ames, J. T.; Kurpiewski, M. R.; Grigorescu, A. Supramol. Chem. 2000, 12, 143. (36) Schwabe, J. W. R.; Chapman, L.; Finch, J. T.; Rhodes, D. Cell 1993, 75, 567-578. (37) Kalodimos, C. G.; Biris, N.; Bonvin, A. M. J. J.; Levandoski, M. M.; Guennuegues, M.; Boelens, R.; Kaptein, R. Science 2004, 305, 386-387. (38) Hippel, P. H. Science 2004, 305, 350-352.

5556 Analytical Chemistry, Vol. 78, No. 15, August 1, 2006

Figure 4. Determination of binding stoichiometry. SPR angle shifts caused by protein binding (ERR and ERβ to wt- and non-ERE) are plotted as a function of protein concentration. The net ∆θ for ERR and ERβ at saturation or near saturation is used to calculate the binding stoichiometry. The inset shows the normalized binding signals of ERR (O) and ERβ (9) to wt-ERE, from which we observe the relative affinity of the two receptors.

experiments were carried out in the reference channel that carries the non-ERE DNA. Similar experiments were conducted for ERβ (binding curves not shown). SPR angle shifts caused by protein binding are plotted as a function of protein concentration in Figure 4. Results show that the immobilized wt-ERE is saturated with ERR at >400 nM and is nearly saturated with ERβ at 900 nM. The net ∆θ at saturation or near saturation is 340 mdeg for ERR and 460 mdeg for ERβ (the net ∆θ is obtained by subtracting the signals in the control channel from the signals in wt-ERE channel). Using the SPR mass sensitivity of 120 mdeg per 100 ng/cm2 for protein and DNA,19 the binding stoichiometry is found to be 2.2:1 and 3.8:1 for ERR/wt-ERE and ERβ/wt-ERE complexes, respectively (molecular mass is 21.4 kDa for the ERE, 66.3 kDa for ERR, and 53.4 kDa for ERβ). The results confirm that ERR binds the ERE as a homodimer7,13 and suggest that ERβ binds to ERE at a higher order, probably as a tetramer. The formation of an ERERE complex with possibly greater than 2 orders has been proposed through estimation of the relative size of the DNA and protein-DNA complex in fluorescent anisotropy studies.12,13 With the SPR quantification of protein- and DNA-binding amount, we can determine the binding order with a higher certainty. Additional information obtainable from the protein titration curves is the relative affinity of the ER-DNA interactions. In the inset of Figure 4, we replot the titration curves of ERR and ERβ to wt-ERE by normalizing the binding signals. It shows more clearly that ERR saturates the DNA faster, at a lower receptor concentration than ERβ. We thus conclude, also as reported earlier using gel shift assays,8,10,11,13 that ERR has a higher affinity to the DNA than ERβ. Understanding the affinity difference is important as it is speculated that binding affinity plays a role in the ability of these receptors to regulate transcription activation.11,14 QCM-D Study of ER-DNA Interactions. QCM-D technology allows for simultaneous measurements of mass loading through frequency shift (∆f) and viscoelastic properties of the adsorbed materials through dissipation shift (∆D). Large dissipation changes are commonly associated with extended, flexible conformation of the attached biomolecules20,26 or loose bindings between interacting molecules.21 On the other hand, a small dissipation is usually reflective of dehydrated, well-structured

Figure 5. QCM-D determination of ER-ERE interactions. QCM-D ∆f (solid line) and ∆D (dashed line) responses versus time were recorded for SA immobilization, wt-ERE assembly, and ERβ (110 nM) and ERR (110 nM) binding. The ERE immobilization was from a 50 nM solution in PBS buffer. ER bindings were conducted in HEPES buffer containing 100 mM KCl. 0.1% SDS was used to regenerate the immobilized DNA.

Figure 6. Specific and nonspecific ERR-DNA interactions following different mechanisms. ∆D-∆f plots of ERR (110 nM) binding to wtand non-ERE targets.

Table 2. Summary of QCM-D Dataa steps

∆f

∆D (Hz) (×10-6)

∆D/∆f (×10-9 Hz-1)

SA EREb ERR-wt-ERE -mut-ERE -non-ERE ERβ-wt-ERE -non-ERE

25.1 16.1 19.4 15.3 10.6 32.3 19.3

0.31 2.37 0.76 0.75 1.16 1.12 0.98

12 147 39 49 109 35 50

a ∆f, related to mass loading; ∆D, related to the induced energy loss; and ∆D/∆f, energy loss per unit coupled mass. For clarity, the standard deviation error is omitted. All data are the average of at least three experiments. The standard deviation is ∼5%. b Average values for wt-ERE, mut-ERE, and non-ERE targets.

biomolecules packed to form rigid layers.16,22,26 In this study, we applied QCM-D to monitor ERE assembly followed by ER binding. Through a combinational analysis of the ∆f and ∆D responses, we evaluate the viscoelastic properties of the bound ERs and discuss their correlations with the conformation parameters such as flexibility and extent of water coupled, in seeking for understanding of sequence-dependent protein-DNA binding mechanisms. Figure 5 shows the ∆f and ∆D responses versus time for the DNA assembly and protein-binding reactions outlined in Scheme 1. In this example, SA immobilization was followed by the assembly of the wt-ERE target. Similar experiments were conducted for the mut-and non-ERE targets. Table 2 summarizes the ∆f, ∆D, and ∆D/∆f recorded for the SA immobilization, DNA assembly, and ERR and ERβ bindings. Analysis of the structure property of the SA film and the end-attached DNA molecules using ∆f and ∆D has been reported in our early studies.18,19 The SA layer is found highly rigid (low ∆D value) and the DNA layer highly flexible and water rich (high ∆D value). For the ER binding, particularly ERR, there is a clear trend of increasing dissipation (i.e., increasing ∆D/∆f ratio in Table 2) when it binds to the wt-, mut-, and non-ERE targets. This trend can be linked to an increase of overall flexibility of the ERRDNA complexes or a possible increase of internal friction in the protein-DNA interface. Our salt concentration experiments using SPR, as well as previous studies,12-14,34 suggest that the sequenceindependent ERR-non-ERE interaction is dominated by electrostatic interactions. Previous studies further indicated that loosely

adsorbed proteins binding through electrostatic interactions tend to diffuse along the DNA strands. 34,37-39 Spolar and Record predicted that water of hydration located at the protein-DNA interface should not be appreciably displaced in the nonspecific complex.40 All these factors suggest the larger energy dissipation (∆D/∆f) measured for the ERR-non-ERE interaction could indicate increased friction, from vibration energy lost in driving weakly bound molecules and in moving a substantial water load in the protein-DNA matrix. In contrast, for the sequence-specific ERR-wt-ERE interaction, the protein and DNA often undergo conformational changes11,41 to form a tight and well-structured complex, and dehydration occurs.40 Both of these effects (a tightly bound film and less water) reduce dissipation in comparison with the case of the ERR-non-ERE interaction. In the above discussion, the ∆D/∆f values are given for the end point data, i.e., at saturation. It is also instructive to plot ∆D as a function of ∆f to eliminate time as an explicit parameter.20,21,42 The slope, and the change in slope, of ∆D-∆f curves provide quantitative information on the kinetic and conformational changes occurring during binding. Figure 6 shows the ∆D-∆f plots for ERR binding with the wt- and non-ERE. ERR binding with nonERE displays a linear ∆D-∆f plot, indicating that the protein experiences no large conformational changes dependent on surface coverage, whereas the sequence-specific binding with wtERE shows two distinguishable linear phases, with an initial larger slope being close to that of the nonspecific adsorption. This indicates that the protein binding is initially largely dissipative. At a sufficient high surface coverage (∆f more than ∼3 Hz), the bound protein becomes less dissipative, as detected by the sudden decrease of ∆D/∆f value. Using lac repressor-DNA interactions as a model system, Kalodimos and co-workers provided a structural view of how sequence-specific protein-DNA recognition may be initiated by nonspecific electrostatic interactions and how nonspecific and specific complexes might interconvert.37,38 This has also been a common hypothesis of how specific DNA is recognized by (39) Halford, S. E.; Marko, J. F. Nucleic Acids Res. 2004, 32, 3040-3052. (40) Spolar, R. S.; Record Jr, M. T. Science 1994, 263, 777-784. (41) Greenfield, N.; Vijayanathan, V.; Thomas, T. J.; Gallo, M. A.; Thomas, T. Biochemistry 2001, 40, 6646-6652. (42) Ho¨o ¨k, F.; Ray, A.; Norde`n, B.; Kasemo, B. Langmuir 2001, 17, 8305-8312.

Analytical Chemistry, Vol. 78, No. 15, August 1, 2006

5557

transcription factors.43 The QCM-D ∆D-∆f plots seem to suggest that the ER-DNA interactions follow this mechanism. The kink in the ∆D-∆f plot for the specific binding may reflect the conversion of a dissipative, nonspecific complex to a less dissipative, well-structured specific complex. The QCM-D data also reveal structural differences in ERRDNA and ERβ-DNA complexes. Bound ERR always shows a higher dissipation (i.e., bigger ∆D/∆f values) when compared to ERβ (Table 2). Taking the example of ERR and ERβ binding with nonERE DNA, ∆D/∆f values are 109 × 10-9 Hz-1 for ERR and 50 × 10-9 Hz-1 for ERβ. Since these bindings involve no sequence recognition, the bound ERR and ERβ are likely in their natural conformations.34,43 The higher dissipation measured for ERR binding (more energy losses) may suggest that this ER subtype is more dynamic and fluidlike in structure than ERβ. Whether or not the higher ERR structural flexibility makes it easier for ERR to undergo conformational change and thus lead to a higher affinity to specific ERE compared with ERβ8,10,11,13 would be of interest for further studies.

CONCLUSION We have conducted SPR and QCM studies of ERR- and ERβDNA interactions. Using the successful immobilization of EREs combined with the control of DNA orientation and activity, the SPR method allows for a quantitative determination of proteinbinding capacity and thus provides more readily evaluation of ER binding specificity, stoichiometry, and affinity to various ERE sequences when compared to traditional methods. Salt concentration experiments using SPR combined with QCM-D analysis elucidate the mechanisms of ER binding with DNA of difference sequences. With the combinational SPR and QCM analysis, we have observed intrinsic differences in DNA-binding properties of the two receptor subtypes, which were not apparent based on the results from studies using traditional methods. We believe that these findings will be of particular interest to researchers in the hormone receptor field to understand how estrogen-responsive genes are differently regulated by the receptors.

Received for review April 3, 2006. Accepted May 22, 2006. (43) Va´zquez, M. E.; Caamano, A. M.; Mascarenas, J. L. Chem. Sci. Rev. 2003, 32, 338-349.

5558

Analytical Chemistry, Vol. 78, No. 15, August 1, 2006

AC0606103