Enhancement of Reaction Specificity at Interfaces - The Journal of

Mar 12, 2008 - ... one reactant must move through the medium and reach the interface. ... interactions based on the configuration-dependent diffusion ...
0 downloads 0 Views 373KB Size
3948

J. Phys. Chem. B 2008, 112, 3948-3954

Enhancement of Reaction Specificity at Interfaces Bat Ami Gotliv,† Shirley S. Daube,‡ and Ron Naaman*,† Department of Chemical Physics and Chemical Research Support, Weizmann Institute of Science, RehoVot 76100, Israel ReceiVed: December 4, 2007; In Final Form: January 17, 2008

A realistic picture of a cell is that of a highly viscous, condensed gel-like substance, crowded with macromolecules that are mostly anchored to membranes and to intricate networks of cytoskeletal elements. Theoretical and experimental approaches to investigating crowding have not considered the role of diffusion through a crowded medium in affecting the selectivity and specificity of reactions. Such diffusion is especially important when one considers interfaces, where at least one reactant must move through the medium and reach the interface. Here, we address this issue by directly investigating how diffusion through a gel medium affects the competition between a single specific reaction and a large number of weak nonspecific interactions, a process that is typical of reactions occurring at interfaces. We present an approach for achieving orientationcontrolled interactions based on the configuration-dependent diffusion rate of the reacting molecule through a gel medium. The effectiveness of the method is demonstrated by the high selectivity obtained both in the adsorption of DNA to a surface and in DNA hybridization to preadsorbed single-strand oligomer on a surface.

Introduction A realistic picture of a cell is that of a highly viscous, condensed gel-like substance, crowded with macromolecules that are mostly anchored to membranes and to intricate networks of cytoskeletal elements.1,2 It is well-established that the crowding and confinement of macromolecules enhance the rate of biochemical reactions.3-6 For example, it has been found that macromolecular crowding increases the binding of DNA polymerase to DNA.7 The thermodynamic implications of crowding, in reference to DNA behavior, have also been investigated and discussed.8,9 The prevailing notion is that crowding renders the reactants confined to a restricted volume, thereby increasing their collision probability and therefore the reaction rate. The theoretical and experimental approaches to investigating crowding have not considered the role of diffusion through a crowded medium in affecting the selectivity and specificity of reactions. Such diffusion is especially important when one considers interfaces, where at least one reactant must move through the medium and reach the interface. Here, we address this issue by directly examining how diffusion through a gel medium affects the competition between a single specific reaction and a large number of weak nonspecific interactions, a process that is typical of reactions occurring at interfaces. We present an approach for achieving orientationcontrolled interactions based on the configuration-dependent diffusion rate of the reacting molecule through a gel medium. The effectiveness of the method is demonstrated by the high selectivity obtained both in the adsorption of DNA to a surface and in DNA hybridization to preadsorbed single-strand oligomer on a surface. Cellular processes occur very often at interfaces between a soluble molecule (e.g., protein, sugar) and its target site on a solid-like substance (e.g., endoplasmic reticulum membrane, * Corresponding author. E-mail: [email protected]. † Department of Chemical Physics. ‡ Chemical Research Support.

condensed chromosome). In biotechnology, artificial interfaces can also be found in such devices as DNA and proteomics microarrays and biosensors. Reactions at interfaces are unique in that one of the reactants is highly restricted from diffusion while the other has to reach the surface. Both in vivo, at a biological interface, and in an artificial system of a chip/solution interface, the challenge is to target a molecule approaching the interface or to achieve a reaction with an already adsorbed species, so that only the desired specific functional group on the molecule will interact. Gels are in widespread use in biotechnology, mainly as media for separating species differing by mass or charge. Gel-based chips were suggested as a replacement for surface-based chips in order to provide increasing concentrations of probe molecules.10 Combining surfaces with gel electrophoresis has been shown to improve the selectivity of DNA microarrays.11 Gels are also compatible with enzymatic reactions. The approach commonly used to achieve controlled DNA adsorption on surfaces is to attach chemically an end group to the 5′ or 3′ end of the DNA. This end group covalently attaches to the surface, whereas nonspecific interactions through the phosphates and bases in the DNA backbone can be minimized by introducing molecules that compete with the weaker nonspecific DNA interactions and displace the DNA backbone groups from the surface.12,13 The remaining DNA is then forced to be oriented above the surface, available for chemical modification. Alternatively, the nonspecific interactions can be reduced by blocking the surface before adsorbing the DNA.14 Despite these methodologies, microarray analyses are hampered by high background noise resulting from the remaining nonspecific binding, thus necessitating the use of various statistical models to overcome this problem.15 The adsorption-through-gel method applied in this study dramatically reduces nonspecific DNA adsorption by several orders of magnitude while increasing the specificity of hybridization as compared to the same process performed in solution. This highly specific grafting methodology eliminates the need

10.1021/jp711441q CCC: $40.75 © 2008 American Chemical Society Published on Web 03/12/2008

Enhancement of Reaction Specificity at Interfaces for blocking reagents, long incubation times, and elevated temperatures during the procedures performed at the interface. Electrical current is not required, and the gel can be removed after DNA adsorption. A model is provided to explain the observed increase in specificity provided by the gel environment. Experimental Section Gold Substrate Preparation. Polished n-type single-crystal (111) silicon wafers were cleaned in a plasma asher (March, plasmod) for 5 min in an atmosphere of oxygen and argon. Immediately after having been cleaned, the wafers were placed in an electron beam evaporator. The deposition was performed at a base pressure of 5 × 10-6 mbar. A 10-nm chromium layer was deposited at a rate of 0.1 nm/s, and then a 150-nm gold layer was deposited at the same deposition rate. The wafer was cut into 1 × 4 cm2 samples using a diamond pen. These gold slides were then cleaned in an ultraviolet ozone cleaner (Uvocs, model T10X10/OES/E) for 20 min and subsequently incubated in absolute ethanol (Merck) with agitation. Finally, the slides were rinsed with ethanol and stored for no more than 24 h in ethanol solution before adsorption. Silicon Substrate Preparation. Polished n-type single-crystal (111) silicon wafers were cut with a diamond pen into 1 × 4 cm2 slides. The slides were then cleaned by immersion in acetone and then in absolute ethanol in a sonicator and subsequently dried with nitrogen. Self-assembled layers of (3mercaptopropyl)trimethoxysilane [(HS(CH2)3Si(OCH3)3, Fluka] were prepared by incubating the clean slides in 1% v/v (3-mercaptopropyl)trimethoxysilane in 0.1% v/v acetic acid and absolute ethanol solution for 30 min. The slides were then washed in acetic acid/ethanol solution and dried for 60 min in argon. Epoxy-coated glass slides (Nexterion Slide E) were purchased from SCHOTT (Jena, Germany). Preparation of Radioactively Labeled DNA. A 26-nucleotide(nt)DNAoligomer(5′-CTAAGATTTTCTGCATAGCATTAATGs SH) covalently attached to a thiol linker [SHd(CH2)3sSsSs (CH2)3sOH)] at its 3′ end and its complementary nonthiolated DNA (5′-CATTAATGCTATGCAGAAAATCTTAGsOH) were purchased from MWG Biotech (Ebersberg, Germany). A 50-nt single-stranded DNA (ssDNA) oligomer with a thiol linker at its 3′ end (5′-ACGCCGCTACATCCATAATTCAGCCTAAGA TTTsSH) and nonthiolated 50-nt DNA oligomer (5′-ACGCCG CTACATCCATAATTCAGCCTAAGATTTsOH) were purchased from Integrated DNA Technologies (IDT, Coralville, IA). The DNA oligomers used in this study were purchased in their lyophilized and oxidized protected form. One hundred picomoles of ssDNA was incubated with 20 units of the enzyme polynucleotide kinase (Fermentase), 12 pmol of γ-32P-ATP (3000 Ci/mmol, NEN Radiochemicals, Boston, MA), 10 mM MgCl2, and 70 mM Tris-HCl (pH 7.5), but without DTT (dithiothreitol) to prevent its competitive binding to the gold. After a 20-min incubation period at 37 °C, the enzyme was heat-inactivated for 10 min at 70 °C. The sample was then loaded onto a Bio-Spin 6 column (Bio-Rad Laboratories, Hercules, CA) to remove any unincorporated γ-32P-ATP. Purified radiolabeled ssDNA was then combined with 400 pmol of unlabeled ssDNA, and this mixture was incubated with 10 mM of the reducing agent tris(2-carboxyethyl) phosphine (TCEP) (Molecular Probes, Eugene, OR) in 100 mM Tris-HCl (pH 7.5) at 4 °C overnight to allow complete reduction of the disulfide bond just before adsorption onto the surface. The DNA samples were then loaded onto a Bio-Spin 6 column (Bio-Rad Laboratories) and equilibrated in the adsorption buffer [20 mM

J. Phys. Chem. B, Vol. 112, No. 13, 2008 3949 Tris-HCl and 400 mM NaCl (pH 7.5) for adsorption on gold or 0.05 M sodium carbonate (pH 8.3) for adsorption on silicon]. In this way, the ssDNA molecules were collected during the flow through an adsorption buffer, whereas low-molecularweight species [TCEP and the reduction product HSs(CH2)s OH, the latter of which can compete with the DNA for binding to the surface) were captured on the spin columns. The ssDNA concentration was determined by adsorption at 260 nm (GeneQuant Pro, GE Healthcare, Chalfont St. Giles, U.K.). Amino-modified ssDNA and fluorescently labeled ssDNA oligomer were purchased from Integrated DNA Technologies (IDT). The 26-nt DNA oligomers (the same sequences as described above) were designed to contain amine modification at their 3′ ends and were fluorescently labeled with CY3 at their 5′ ends. DNA Adsorption. We developed a method for DNA adsorption from gels. Low-melting-point agarose (AMRESCO, Solon, OH) was prepared at different concentrations (0.5%, 1%, and 2% w/v) in doubly distilled water. The agarose was then heated until it was completely dissolved and then equilibrated at 60 °C in a water bath for a minimum of 15 min. A pipet tip was prewarmed by being immersed in warm doubly distilled water (60 °C) several times and then used immediately to spot 10 µL of warm agarose solution on clean gold or silicon slides. The agarose spot was then quickly polymerized and equilibrated for 30 min in a controlled-humidity environment. When acrylamide gels were used for the DNA adsorption, different concentrations of acrylamide/bisacrylamide (1:19, 40% w/v, Biological Industries, Bet-haemeek, Israel) (5%, 10%, and 20% v/v) were prepared in doubly distilled water. Polymerization was performed just before adsorption by mixing 0.5 mL of acrylamide/bisacrylamide at the desired concentration with 10 µL of 10% ammonium persulfate (APS, Bio Lab) and 1 µL of N,N,N′,N′-tetramethylethylenediamine (TEMED, Biosolve, Valkenswaard, The Netherlands). Aliquots of the acrylamide mixture (10 µL) were then spotted immediately on clean gold slides, and the gel spots were polymerized and equilibrated for 30 min in a controlled-humidity environment. Diffusion of radioactively labeled thiolated or nonthiolated ssDNA (either 26 or 50 nt) was performed by spotting 1 µL of 6.5 µM ssDNA in adsorption buffer above the polymerized gel spots. In addition, to evaluate the specificity of the adsorption through gel, ssDNA adsorption was also performed from solution (1 µL of 6.5 µM ssDNA diluted in 10 µL of doubly distilled water). The adsorption reaction was then performed for either 1 h or overnight at room temperature in sealed Petri dishes with humidity to avoid dryness. After the adsorption, the water solution from each slide was carefully removed with a pipet, and the gels were removed either by a flow of nitrogen to remove the acrylamide gel or by incubation with 30 U/mL agarase (Sigma) at 43 °C for 30 min to ensure agarose removal. Each slide was then rinsed in a 20 mM Tris-HCl/400 mM NaCl (pH 7.5) solution to wash away excessDNA. The slide was then soaked in 10 mL of the same buffer for 15 min with agitation, followed by 15 min of soaking in a 20 mM Tris-HCl/200 mM NaCl (pH 7.5) solution. Finally, the slides were quickly rinsed in doubly distilled water and then air-dried. Adsorption Characterization. Slides adsorbed with radioactive DNA were taped with double-sided tape to Whatman filter paper. The covered paper was exposed to a phosphorimaging screen (Fuji). After the appropriate exposure time, the screen was imaged with a phosphorimage scanner (BAS-2500, Fuji). Next, the number of pixels at a specific area on each slide was

3950 J. Phys. Chem. B, Vol. 112, No. 13, 2008

Gotliv et al.

Figure 1. Experimental protocol: adsorption of thiolated and nonthiolated DNA onto a surface either through a gel (left) or from solution (right).

determined using the software Image Gauge (Fuji). The number of pixels per area was converted to the number of picomoles per unit area using the specific activity. The specific activity of each DNA sample was defined as the amount of radiation produced by 1 mol of that type of DNA. This number was determined empirically by spotting 1 µL samples of radioactive DNA with a known concentration on a gold slide. Finally, the samples were allowed to air-dry without any washing and were then phosphorimaged. The number of pixels at each spot was divided by the number of picomoles spotted, yielding the specific activity of the samples. This value already takes into account gold quenching. Hybridization. The hybridization experiments were performed on gold slides or epoxy-coated glass slides (SCHOTT) inside an array of wells separated from one another by RTV (room-temperature vulcanized) polymer. The RTV matrix setup was molded by polymerization of RTV615 silicone rubber compounds (GE Bayer silicone) at 70 °C onto an array of beads to create a matrix of empty wells. After 4 h of polymerization, the RTV polymer was removed, and the array of wells was cut to the desired shape and placed on a clean gold slide. Selfassembled layers of thiolated nonradioactive 26-nt ssDNA were prepared by spotting 10 µL (5 µM) of ssDNA in adsorption buffer on each well and allowing the samples to stand overnight. The monolayers in two of the wells contained radioactively labeled ssDNA and were used for calculating hybridization efficiency. When hybridization was performed on an epoxycoated glass slide, self-assembled layers of amino-modified unlabeled 26-nt ssDNA were prepared by spotting 10 µL (20 µM) of ssDNA in SSCx3 buffer on each well and allowing the samples to stand for 3 h. After target adsorption, the wells on the gold surface were washed several times with a 20 mM TrisHCl/400 mM NaCl (pH 7.5) solution to wash away unbound DNA and then air-dried. The wells on the glass slides were washed with SSCx3 buffer and then incubated for 1 min with doubly distilled water. One percent agarose gel was polymerized in some of the RTV wells (as described above), and the polymerization was equilibrated for 15 min under controlled humidity to prevent dryness. Four different radioactively labeled nonthiolated ssDNAs were used for hybridization on gold: (1)

the ssDNA complementary to that adsorbed on the surface (5′CATTAATGCTATGCAGAAAATCTTAG) without any mismatch, (2) an ssDNA that contained a mismatch of a 5-nt cluster in the middle of the sequence, (3) an ssDNA that contained a mismatch of five nucleotides distributed along the sequence, and (4) an ssDNA that was noncomplementary to the adsorbed ssDNA (5′-CCAAAGGGCAACGGGTGGCGGCAGGCTC). The hybridizations on the epoxy-coated glass slides were performed with 5′ CY3 fluorescently labeled oligomer dissolved in SSCx3 buffer. Each ssDNA was spotted above the gel (1 µL of 6.5 µM ssDNA in adsorption solution for hybridization on gold or 1 µL of 20 µM ssDNA in SSCx3 solution for hybridization on epoxy-coated glass slide). The probe ssDNA was allowed to diffuse and adsorb on the surface at room temperature for 1 h in a sealed Petri dish with humidity to avoid dryness. To estimate the specificity of the adsorption through the gel, an experiment with ssDNA in adsorption buffer was performed in parallel to the gel experiments. After 1 h of adsorption, the buffer solution was removed from the wells, and the RTV polymer was peeled from the gold surface. In most of the hybridization experiments, the gel spots were attached to the sides of the wells and thus were completely removed, together with the RTV. In other cases, the gel was easily removed from the gold surface with nitrogen flow. Excess unhybridized ssDNA was removed from the gold surface with several washes as described above. The epoxy-coated glass slides were washed for 10 min with SSCx2 buffer containing 0.2% SDS, for 10 min with SSCx2, and finally for 10 min with SSCx0.2. The hybridization on the gold surface was characterized and quantified with a phosphorimaging detector as described above. The epoxy-coated glass slides were imaged in a scanner (BAS-2500, Fuji) supplied with 532-nm laser for the excitation of CY3 molecules and an LPG (575-nm) filter. The number of molecules in each florescent spot was calculated by correlation with a known amount of fluorescently labeled ssDNA oligomer adsorbed on an epoxy-coated glass slide. Results Effect of Gel Medium on ssDNA Adsorption Specificity. To test the effect of diffusion through a gel on biochemical

Enhancement of Reaction Specificity at Interfaces

J. Phys. Chem. B, Vol. 112, No. 13, 2008 3951

Figure 2. Specificity and selectivity of adsorption of single-stranded DNA (ssDNA) on a gold surface. (A) Spot array (right) showing the signal from radioactively labeled ssDNA on a gold surface (columns 1-4) and on a silicon surface (column 5), as imaged by phosphorimaging. ssDNA diffused to the surface through either prepolymerized 1% agarose gel (rows A and B) or from a water drop (rows C and D). The 26-nt (columns 1 and 2) or 50-nt (columns 3 and 4) ssDNA was either thiolated at its 3′ end (DNAsSH, rows A and C) or nonthiolated (DNAsOH, rows B and D) and was adsorbed to the surface either for 1 h (columns 1 and 3) or overnight (columns 2 and 4). Column 5: 26-nt ssDNA was adsorbed overnight onto a mercapto-silanated silicon surface. The bar graphs on the left represent, on a logarithmic scale, the selectivity ratios, calculated as the ratio (DNAsSH/DNAsOH in gel)/(DNAsSH/DNAsOH in solution). (B) Adsorption through acrylamide gels. The spot array on the right shows the phosphorimaging signal of adsorbed 26-nt DNAsSH or DNAsOH through different concentrations of prepolymerized acrylamide/ bisacrylamide (aa) gel. The bar graphs on the left represent, on a logarithmic scale, the selectivity ratios, calculated as the ratio (DNAsSH/DNAs OH in gel)/(DNAsSH/DNAsOH in solution).

reactions occurring at interfaces, we developed the experimental protocol presented schematically in Figure 1. In this experimental arrangement, a thin layer (a few hundreds of micron thick) of gel is placed on top of the substrate. The solution of reacting molecules (either 26- or 50-nt-long DNA oligomers, in our particular case) is applied to the top of the gel, and the molecules diffuse through the gel and reach the substrate. The substrate used was either a gold surface or a silanated silicon surface (Figure 2, upper spot array, columns 1-4 and 5, respectively). The DNA oligomers were radioactively labeled so that their binding to the surface could be followed. The adsorption process was stopped by completely removing the gel, and the adsorption specificity was determined by scanning the radioactive signal left on the surface after it had been washed. Figure 2 (upper spot array, rows A and B) presents spots of radioactively labeled thiolated DNA (chemisorbed to the surface) and nonthiolated DNA (representing nonspecific physisorption), respectively, that had adsorbed to the substrate through a thin layer of 1% agarose gel prepolymerized on the surface. The specificity ratio was calculated by dividing the signal from bound thiolated DNA by that obtained from the nonthiolated DNA (Figure 2, upper spot array, rows A and B, respectively). The effect of the gel was evaluated by dividing the specificity obtained in the gel by the specificity obtained through adsorption from a water solution (Figure 2, upper spot array, rows C and D). This ratio is termed the “selectivity ratio”. Graph A in Figure 2 quantitatively summarizes the selectivity ratios derived from the upper spot array for two adsorption times, 1 h and overnight, and for two lengths of DNA oligomers, 26and 50-nt. For a 26-nt-long ssDNA adsorbed through gel in 1

h, the specificity ratio was 61 ( 10. This high selectivity ratio reflects the high specificity of adsorption achieved by the agarose gel. The specificity of the adsorption reaction through gel was reduced, however, when the adsorption reaction was performed overnight and was very similar to that obtained for adsorption from water (1 ( 2). This result suggests that the gel exerts a kinetic effect. Controlling the adsorption specificity for longer ssDNA chains is a much greater challenge, given that longer ssDNA chains have more conformational possibilities for nonspecific interactions. Indeed, as shown in the Figure 2, adsorption of a 50-nt ssDNA for 1 h resulted in a specificity ratio of 4 ( 1, markedly reduced in comparison to that for the shorter oligomers. However, the selectivity ratio was still above 1, suggesting that, even for this longer DNA oligomer, the gel had a significant effect on the specificity of the adsorption. Similarly to the behavior of the 26-nt ssDNA, this selectivity ratio was reduced to 1 after overnight adsorption. To explore the generality of the gel effect, we investigated the adsorption of the 26-nt ssDNA on a silicon surface modified with a mercaptoalkyl silane (Figure 2, upper spot array, column 5). The adsorption was performed overnight, as it involved the formation of a disulfide bond on the surface, a much slower process than the binding of a thiol group to a gold surface. The overnight adsorption from a water solution resulted in the reaction of substantial amounts of both thiolated and nonthiolated DNA and therefore yielded a specificity ratio approaching 1. Adsorption from agarose gel yielded significantly reduced amounts in comparison to adsorption from solution, although

3952 J. Phys. Chem. B, Vol. 112, No. 13, 2008

Gotliv et al.

Figure 3. (i) Hybridizations on a gold surface. Spot array on the right: Different radioactively labeled nonthiolated target ssDNA samples were hybridized for 1 h at room temperature to thiolated unlabeled ssDNA (26-nt) probes that had been preadsorbed overnight (control panel). The hybridization reactions were performed either through 1% prepolymerized agarose gel (row 1) or from solution (row 2). Column A: Target DNA fully complementary to the adsorbed DNA. Columns B and C: Target DNA contained a 5-nt mismatch that was either clustered in the middle of the oligomer (column B) or randomly distributed in the sequence (column C). Column D: the target DNA had no complementarity with the adsorbed DNA. Bar graph on the left: Hybridization yields in solution (i) and in gel (ii) obtained on a gold surface, calculated as the ratio between each target ssDNA to the preadsorbed probe DNA (control panel). (iii) Spot array on the right: hybridization reaction on epoxy-coated glass slide. Amine-modified unlabeled ssDNA (26-nt) probes were adsorbed for 3 h (control panel). Different CY3 fluorescently labeled unmodified ssDNA targets were hybridized with this probe for 1 h at room temperature either through 1% prepolymerized agarose gel (row 1) or from solution (row 2), with the same complementarities as described above. Bar graphs on the left: Hybridization yields in gel (iii) and in solution (iv) obtained on epoxy-coated glass slides and calculated as above.

nonthiolated ssDNA could hardly be detected on the surface. Hence, the selectivity ratio in this case was very large. Because the average pore size of 1% agarose gel is 408 nm,16 which is more than 10 times larger than the length of the DNA, one would expect to see no effect of the gel. Surprised by the dramatic effect that we observed, we repeated the experiments, replacing the 1% agarose gel by 0.5% and 2% agarose gel; however, no significant differences were detected (data not shown). We therefore postulate that the effect of agarose gel is conferred by the heterogeneity in pore sizes within the agarose, as the pores at the gel-surface interface are very small, much smaller than the “average size” of the pores in the gel bulk.17 Another explanation is based on the findings that there is extensive solute exclusion in the vicinity of hydrophilic or charged surfaces,1,18 thereby leaving unexpectedly small diffusion channels between surface pores. To have better control over the gel pore size in inducing adsorption selectivity, we performed the adsorption of a 26-ntlong DNA oligo onto the gold surface through acrylamide (aa) gels of varying concentrations (Figure 2B and lower spot array). The average pore size within an aa gel varies from 13 nm for 20% gel to 29 nm for 5% gel.19 The specificity ratio was found to be 15 ( 7.3 in 10% aa gels and 1 ( 0.6 in 5% aa gels. Interestingly, the specificity ratio in 20% aa was found to be very large (66 ( 46). Hence the results in Figure 2B led to a clear conclusion: the higher the concentration of the gel and the smaller the pores, the higher the selectivity of the reactions. DNA Hybridization through Gel Medium. DNA hybridizations on the surface were also performed through a gel. Different sequences of either radioactively or fluorescently labeled target ssDNA (26-28-nt) were hybridized to unlabeled 3′-modified (either thiolated or amino-modified) probe ssDNA preadsorbed on gold or an epoxy-coated glass surface, respectively. The labeled target DNA was introduced either by diffusion through 1% agarose gel or from solution (Figure 3, upper spot array, rows 1 and 2, respectively) for 1 h at room temperature. The

hybridization yields on the gold surface and on the epoxy-coated glass slides (as deduced from the upper spot array in Figure 3) were calculated as the ratio between the target DNA to the probe DNA on the surface (deduced from a control adsorption of a labeled sample of probe DNA). To evaluate the selectivity of the hybridization, four different sequences were compared. A fully complementary target to the probe ssDNA (maximum ∆G ) -44.09 kcal/mol for hybridization interaction) yielded a 7.7 ( 0.1% hybridization efficiency through agarose gel and a 9.30 ( 0.15% hybridization efficiency from solution when the interaction was performed on the gold surface (Figure 3i and 3ii, bar A). The hybridization efficiency on the epoxy-coated glass slide was 31 ( 0.27% through agarose gel and 20 ( 0.17% through solution (Figure 3iii and 3iv, bar A). For an ssDNA with a 5-nt mismatch cluster in the middle of the sequence and a reduced stability of the duplex (maximum ∆G ) -16.56), the hybridization efficiency dropped to a very low value of 0.170 ( 0.002% when hybridization on gold was performed from gel, whereas a higher value was obtained for hybridization from solution (Figure 3i and 3ii, bar B). For an ssDNA with five mismatched nucleotides randomly spread in the sequence (maximum ∆G ) -12.27), no hybridization could be detected on a gold substrate when this ssDNA was introduced from gel (Figure 3ii, bar C). A similar and very low signal was observed when the hybridization was performed from gel on an epoxy-coated glass slide (Figure 3iii and 3iv), yielding a hybridization efficiency of 0.89 ( 0.01%. When the same target was hybridized in solution, the hybridization efficiency was much higher, 11.7 ( 0.1%. A fully noncomplementary ssDNA (maximum ∆G ) -6.69 kcal/mol) did not interact with the adsorbed ssDNA when the interaction was performed by diffusion from agarose gel on gold (Figure 3ii, bar D), whereas some DNA remained on the surface when the oligomer was introduced from solution (Figure 3i, bar D). Similar results were observed when hybridizations were performed on epoxy-coated glass slides (Figure 3ii and iv, bar D).

Enhancement of Reaction Specificity at Interfaces

J. Phys. Chem. B, Vol. 112, No. 13, 2008 3953

Figure 4. (a) Free energy diagram of specific and nonspecific adsorptions from gel and from solution. (b) Depiction of the effect of gel on the configuration of the molecules as they reach the surface.

Clearly, the gel increased the selectivity of DNA hybridization in a manner proportional to the stability of the duplex. Note that all of the hybridization reactions were performed at room temperature for 1 h without any blocking reagents. This is in comparison to a regular blocking procedure on commercial epoxy-coated glass that is performed with at least 16 h of hybridization at 42 °C.20 Discussion The results from the current study can be summarized as follows: (i) Adsorption of functionalized DNA (thiolated) through a gel is more selective by orders of magnitude than adsorption from solution. (ii) DNA hybridizations are more selective when one of the reacting oligomers diffuses through gel as opposed to solution. (iii) The effect of the gel is a kinetic one, most likely due to an increase in the rate of the specific adsorption compared to a reduction in the rate of nonspecific adsorption; however, the medium (gel versus solution) does not affect the thermodynamic equilibrium. (iv) The pore size defines the rate ratio between the specific and nonspecific interactions. The smaller the pores, the faster the specific process relative to the nonspecific process. The present study demonstrates that the specificity of oligonuclotide adsorption on different substrates and their hybridization to DNA at the interface is significantly improved when the interaction is performed by diffusion of DNA from gel, compared to the same processes occurring in solution. Note that the hybridization between mismatched DNA pairs performed through gel on the gold surface was uniform, in contrast to the hybridization signal obtained from solution. This nonuniform signal, known as the “doughnut effect”,21 is one of the reasons for false signals in readouts of DNA microarrays. In addition, the hybridizations performed through gel did not require any blocking reagents. Nonideal blocking is another reason for false signals. The enhancement in selectivity can be rationalized by realizing that the mobility of the DNA molecules through the gel must be conformation-dependent (see Figure 4b). Specific interactions are characterized by high absolute enthalpies and low entropies, whereas nonspecific interactions are related to lower absolute enthalpies but higher entropies. Hence, for reactions occurring in solution, there is an energy barrier for the specific reaction, whereas there is no barrier or a low barrier for the nonspecific interaction. Therefore, the reaction rate for the specific process is lower than that for the nonspecific one.

Because of the configuration constraints of the molecules in the crowded gel, the entropy of the molecules is lower than thath in solution. As a result, the free energy state of the system is higher, and the barrier for the specific reaction becomes smaller, resulting in an increased rate of the specific process as compared to the rates of the nonspecific interactions (Figure 4a). Indeed, the differences in adsorption specificity between the gel environment and the solution environment on a gold surface were significant only when the reactions were performed for 1 h. After overnight adsorption, these differences essentially vanished. This implies a kinetic effect, which is consistent with the proposed mechanism. By a kinetic effect, we mean that the conformation-dependent rate of diffusion through the gel does not affect the system when it reaches equilibrium, but it is important in defining the ratio between specific and nonspecific products, before the equilibrium is reached. The microscopic manifestation of the model is given by the following picture: The gel can be viewed as a network of bonded polymers forming pores of variable sizes. The transmission of extended DNA through the pores would be different from that of a compact globular DNA (see Figure 4b). As a result of this “conformation-dependent mobility”, extended DNA is envisioned to move more quickly through the gel and reach the substrate in its extended configuration, whereas those strands that are only partly extended move much more slowly through the pores. Hence, the pores induce conformation-dependent kinetics, and oligomers with the most suitable orientation reach the substrate earlier than others. This leads to preferential binding of these molecules through the specific functional group at their end (thiol in our case) and, hence, to enhanced rates for the specific interaction. That is, when the DNA molecules diffuse through the gel, they reach the substrate in a configuration that matches the requirements for the specific interaction. The present findings not only suggest a new avenue by which gels can promote reaction specificity on surfaces, they also provide new insights into the effects of macromolecular crowding at surfaces and in solution. The results suggest that macromolecular crowding might not only enhance the rate of macromolecular reactions, but might also eliminate nonspecific interactions that otherwise lead to dead-end products. In the crowded environment, when the mobility of macromolecules depends on their conformation, the entropy associated with the molecules before reaction is necessarily low, as compared to the same molecules in ideal solutions. Because specific reactions are associated with low entropy, the change in entropy in a

3954 J. Phys. Chem. B, Vol. 112, No. 13, 2008 crowded environment is small relative to the change for the same process when occurring in ideal solution. Although the present results were obtained for a gel-surface interface, we believe that they can provide new insights into the present understanding of the crowding effect by highlighting the contribution of configuration-dependent diffusion rates. Concluding Remarks The gel environment described in this study provides a relatively simple and reliable method for controlling DNA adsorption on a variety of surfaces. Not only is the DNA within the gel environment available for hybridization and biochemical manipulations, but the gel medium provides a dense environment for future studies on the effect of molecular crowding at a surface-solution interface. In addition, by exploring different gel properties, we were able to shed more light on the mechanism of DNA adsorption. We propose that the gel medium is the method of choice for grafting DNA, and most likely other biomolecules, on a variety of surfaces, given that the specificity of both DNA adsorption and hybridization were significantly improved within the gel and the experimental protocol was significantly simplified. Acknowledgment. This work was partially supported by the Grand Center for Sensors and Security. References and Notes (1) Pollack, G. H. Cells, Gels and the Engines of Life; Ebner & Sons: Seattle, WA, 2001.

Gotliv et al. (2) Konopka, M. C.; Shkel, I. A.; Cayley, S.; Record, M. T.; Weisshaar, J. C. J. Bacteriol. 2006, 188, 6115. (3) Sikorav, J. L.; Church, G. M. J. Mol. Biol. 1991, 222, 1085. (4) Minton, A. P. J. Biol. Chem. 2001, 276, 10577. (5) Minton, A. P. J. Cell Sci. 2006, 119, 2863. (6) Ellis, R. J. Trends Biochem. Sci. 2001, 26, 597. (7) Zimmerman, S. B.; Harrison, B. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 1871. (8) Goobes, R.; Kahana, N.; Cohen, O.; Minsky, A. Biochemistry 2003, 42, 2431. (9) Goobes, R.; Minsky, A. J. Am. Chem. Soc. 2001, 123, 12692. (10) Kolchinsky, A.; Mirzabekov, A. Hum. Mutat. 2002, 19, 343. (11) Xiao, P. F.; Cheng, L.; Wan, Y.; Sun, B. L.; Chen, Z. Z.; Zhang, S. Y.; Zhang, C. Z.; Zhou, G. H.; Lu, Z. H. Electrophoresis 2006, 27, 3904. (12) Herne, T. M.; Tarlov, M. J. J. Am. Chem. Soc. 1997, 119, 8916. (13) Ladd, J.; Boozer, C.; Yu, Q. M.; Chen, S. F.; Homola, J.; Jiang, S. Langmuir 2004, 20, 8090. (14) Aqua, T.; Naaman, R.; Daube, S. S. Langmuir 2003, 19, 10573. (15) Abdueva, D.; Skvortsov, D.; Tavare, S. Nucleic Acids Res. 2006, 34, E105. (16) Pernodet, N.; Maaloum, M.; Tinland, B. Electrophoresis 1997, 18, 55. (17) Pluen, A.; Netti, P. A.; Jain, R. K.; Berk, D. A. Biophys. J. 1999, 77, 542. (18) Zheng, J. M.; Chin, W. C.; Khijniak, E.; Khijniak, E.; Pollack, G. H., Jr. AdV. Colloid Interface Sci. 2006, 127, 19. (19) Holmes, D. L.; Stellwagen, N. C. Electrophoresis 1991, 12, 612. (20) Schott Protocol for Nexterion Slide E; SCHOTT: Jena, Germany, July 2007. (21) Heim, T.; Preuss, S.; Gerstmayer, B.; Bosio, A.; Blossey, R. J. Phys.-Condens. Matter 2005, 17, S703.