Hybridization Characteristics of Biomolecular Adaptors, Covalent DNA

Department of Biotechnology and Molecular Genetics, University of Bremen, FB 2-UFT, Leobener Strasse,. D-28359 Bremen, Germany, and Centre for ...
0 downloads 0 Views 205KB Size
168

Bioconjugate Chem. 1998, 9, 168−175

Hybridization Characteristics of Biomolecular Adaptors, Covalent DNA-Streptavidin Conjugates Christof M. Niemeyer,*,† Wolfgang Bu¨rger,† and Rein M. J. Hoedemakers‡ Department of Biotechnology and Molecular Genetics, University of Bremen, FB 2-UFT, Leobener Strasse, D-28359 Bremen, Germany, and Centre for Biomedical Technology, University of Groningen, Hanzeplein 1, 9713 HZ Groningen, The Netherlands. Received September 19, 1997; Revised Manuscript Received December 12, 1997

Semisynthetic, covalent streptavidin-DNA adducts are versatile molecular connectors for the fabrication of both nano- and microstructured protein arrays by use of DNA hybridization. In this study, the hybridization characteristics of six adduct species, each containing a different DNA sequence of 21 or 24 bases, have been compared. First, the adducts were conjugated to biotinylated alkaline phosphatase, and their binding to immobilized oligomer complements of similar lengths was quantified in a microplate assay. The binding efficiency observed varied to a great extent with the specific sequence of the oligonucleotide attached, and could not be predicted from affiliated thermodynamic data of duplex stability. To further elucidate the hybridization properties, the hybridization rate constants of association and dissociation (kassn and kdissn) have been determined for both unconjugated oligonucleotides and protein adducts, using a surface plasmon resonance biosensor. The kassn values observed for the oligonucleotides are in the range of 9 × 103 to 2 × 105 M-1 s-1 and correlate with structural properties of the probe strands. Up to 3-fold decreased kassn values were obtained for the corresponding protein adducts. Likewise, values were observed for kdissn ranging from 1.4 × 10-4 to 1.9 × 10-5 s-1 for the oligonucleotides. The dissociation of the analogous protein conjugates was reduced by up to 5-fold. The extent of this decrease correlates with the formation of homodimeric or intramolecular aggregation of probe strands. A mechanistic model for explaining these data is based on attractive intramolecular interaction between the nucleic acid and protein moiety.

INTRODUCTION

The generation of new biomaterials for analytical, industrial, or pharmaceutical applications is a central goal of biotechnology. Owing to its unique recognition capabilities, physicochemical stability, and mechanical rigidity, DNA is a promising candidate for the construction of nanometer-sized frameworks (1, 2) which are potentially useful for the selective positioning of molecular devices such as conducting polymers (3), proteins (4), or nanocrystal metal clusters (5, 6). Recently, hybrid molecules of short single-stranded DNA and streptavidin (STV) have been applied in the supramolecular assembly of novel bioconjugates (4). The covalent attachment of an oligonucleotide moiety to STV provides a specific recognition domain for a complementary nucleic acid sequence in addition to the four native biotin-binding sites. These bispecific binding capabilities allow the adducts to serve as versatile connectors in a variety of applications, such as the fabrication of nano- and microstructured protein arrays. While the nanostructures can be accessed by positioning several DNA-protein adducts along a single-stranded RNA or DNA carrier molecule containing a set of complementary sequences, the microstructures can be obtained from given surface-immobilized DNA arrays, such as microchip-based oligonucleotide arrangements currently under active investigation for integrated nucleic acid analysis (7-9). For example, * To whom correspondence should be addressed. Phone: (49) 421 218-4911. Fax: (49) 421 218-7578. E-mail: [email protected]. † University of Bremen. ‡ University of Groningen.

the “self-sorting translation” of a model DNA array into polyfunctional immunoglobulin G (IgG) arrays was attained by hybridization of a mixture of specific sequencecontaining DNA-STV adducts, each previously conjugated to a particular biotinylated IgG (4). Due to the sensitivity of most biomolecular devices, such as antibodies and enzymes, the effective employment of nucleic acid hybridization-based assembly strategies requires mild reaction conditions such as moderate temperatures and salt concentrations. Thus, binding efficiency of oligonucleotide-tagged compounds is controlled by hybridization rate constants for association, kassn, and dissociation, kdiss, rather than thermodynamic duplex stability. Kinetic data of DNA hybridization are scarce (10-17), and it is difficult a priori to design sequences of single-stranded oligonucleotides that are optimized for fast and efficient binding to their complementary target strands. As an experimental approach for selecting appropriate sequences, the hybridization characteristics of six covalent DNA-protein adducts have been analyzed in this study. Each of these hybrids (HA through HF; for sequences of A24 through F24, see Table 1) contains a different oligonucleotide sequence of 21 or 24 bases, chosen from the antisense strand of the multiple cloning site of the bacteriophage M13mp18. As a relevant test of near-surface hybridization properties, a binding assay of the DNA-STV adducts was carried out in microplates (Figure 1). The adducts were preconjugated to biotinylated alkaline phosphatase (AP), and their binding to immobilized single-stranded oligonucleotide targets was quantified by colorimetric detection of enzymatic activity. The hybridization efficiency observed varied to a great extent with the specific sequence of the

S1043-1802(97)00170-5 CCC: $15.00 © 1998 American Chemical Society Published on Web 02/07/1998

DNA−STV Kinetics

Bioconjugate Chem., Vol. 9, No. 2, 1998 169

Table 1. Oligonucleotides Used in This Study probe

descriptiona

TA16 TA21

5′-thiolated 16-mer; Tm ) 44.8 °C; ∆G ) -29.3 kcal/mol 5′-thiolated 21-mer; Tm ) 52.3 °C; contains mismatch (underlined) 5′-thiolated 24-mer; Tm ) 65.1 °C; ∆G ) -43.9 kcal/mol 5′-biotinylated 24-mer; complementary to TA24, TA21, and TA16 5′-thiolated 21-mer; Tm ) 52.1 °C; ∆G ) -34.2 kcal/mol 5′-biotinylated 21-mer; complementary to B21 5′-thiolated 21-mer; Tm ) 68.0 °C; ∆G ) -44.7 kcal/mol 5′-biotinylated 21-mer; complementary to TC21 5′-thiolated 21-mer; Tm ) 58.1 °C; ∆G ) -37.4 kcal/mol 5′-biotinylated 21-mer; complementary to TD21 5′-thiolated 21-mer; Tm ) 68.5 °C; ∆G ) -43.2 kcal/mol 5′-biotinylated 21-mer; complementary to TE21 5′-thiolated 21-mer; Tm ) 65.5 °C; ∆G ) -43.7 kcal/mol 5′-biotinylated 24-mer; complementary to TF24

TA24 bcA24 TB21 bcB21 TC21 bcC21 TD21 bcD21 TE21 bcE21 TF24 bcF24

sequence 5′-thiol-G AAA TTG TTA TCC GCT-3′ 5′-thiol-TCC TAT GTG AAA TTG TTA TCC-3′ 5′-thiol-TCC TGT GTG AAA TTG TTA TCC GCT-3′ 5′-biotin-AGC GGA TAA CAA TTT CAC ACA GGA-3′ 5′-thiol-GTA ATC ATG GTC ATA GCT GTT-3′ 5′-biotin-AAC AGC TAT GAC CAT GAT TAC-3′ 5′-thiol-CCG GGT ACC GAG CTC GAA TTC-3′ 5′-biotin-GAA TTC GAG CTC GGT ACC CGG-3′ 5′-thiol- CAG GTC GAC TCT AGA GGA TCC-3′ 5′-biotin- GGA TCC TCT AGA GTC GAC CTG-3′ 5′-thiol-AGT GCC AAG CTT GCA TGC CTG-3′ 5′-biotin- CAG GCA TGC AAG CTT GGC ACT-3′ 5′-thiol- GTT TTC CCA GTC ACG ACG TTG TAA-3′ 5′-biotin- TTA CAA CGT CGT GAC TGG GAA AAC-3′

a Melting temperatures (T ) and free energies of duplex stability (∆G) have been calculated for underivatized oligonucleotides at 25 °C m and 150 mM NaCl using Primer Select version 3.01a (DNA-Star Inc.), following a model based on adjacent dinucleotides (19, 20).

example, positively charged side chain groups of the protein moiety. MATERIALS AND METHODS

Figure 1. Schematic drawing of the enzyme-linked oligonucleotide sorbent assay (ELOSA) for the determination of hybridization efficiency of covalent DNA-STV adducts. The latter were coupled with a signal-generating enzyme by mixing with biotinylated alkaline phosphatase (AP) to form preconjugates (e.g., HA24-AP). Biotinylated capture oligonucleotides were immobilized on STV-coated microplates, and remaining biotinbinding sites of the surface-bound STV were blocked with free D-biotin, represented by shaded spheres. The preconjugates were then allowed to bind to their complement by formation of specific Watson-Crick base pairs. Subsequently, the amount of hybridized material was determined by AP-dependent color reaction using pNPP as a substrate.

oligonucleotide domain, thus, enabling a classification of strong or weak binders. To further elucidate the results observed for the enzyme conjugates, real-time biospecific interaction analysis (BIA) technology was used to quantify the hybridization kinetics of both the STV-DNA hybrid and their analogous, unconjugated oligonucleotides, TA through TF. BIA, a method based on surface plasmon resonance (SPR), has previously been used to study DNA hybridization (10, 13, 14, 18). The kinetic rate constants for association (kassn) and dissociation (kdiss), measured here for the six sequences under investigation, confirm the hybridization characteristics previously observed in the microplate assay, and correlate with structural features of the probe strands rather than thermodynamic duplex stability (19, 20). Owing to their elevated information contents, the BIA data also reveal influences of the protein moiety on the process of nucleic acid hybridization. Both annealing and dissociation rate constants are reduced relative to those of the analogous oligonucleotides. This effect may be due to intramolecular surface interaction between the nucleic acid and, for

Oligonucleotides. The 5′-thiol-modified oligonucleotides (TA24 through TF24; for sequences, see Table 1) were purchased from MWG-Biotech (Germany) or NAPS (Go¨ttingen, Germany). The oligonucleotides were synthesized by standard methods using C6 Thiolmodifier reagent (Glen Research) and were purified by HPLC. The purity of the samples was determined from anionexchange chromatography analysis on a Mono Q column (Pharmacia). The concentrations were determined by UV absorbance measurement. Biotinylated oligomer complements were also purchased from MWG-Biotech or NAPS (for sequences, see Table 1). Synthesis of DNA-Streptavidin Adducts and ELOSA. Synthesis and purification of covalent DNASTV adducts were carried out from thiolated oligonucleotides, TA24 through TF24, and recombinant streptavidin (Boehringer Mannheim), as previously described (4). In brief, STV (10 nmol) was derivatized with maleimido groups using a heterobispecific cross-linker (sulfo-SMPB, Pierce), reacted with a 5′-thiolated oligonucleotide (10 nmol) and subsequently purified by anion-exchange chromatography. A 1:1 molar ratio of nucleic acid and protein moiety of the hybrids was verified by gel electrophoretic and photometric analysis, and concentrations of the preparations were determined by absorbance measurements (4). Conjugates of the DNA-STV hybrids and AP were prepared by mixing 0.1 µM stock solutions of the hybrids (HA24 through HF24; for sequences, see Table 1) and a 4-fold molar excess (3.8 µM stock solution) of biotinylated AP (Sigma) in TETBS buffer [20 mM TrisHCl buffer (pH 7.3), 150 mM NaCl, 5 mM EDTA, and 0.01% (w/v) Tween-20] containing 0.1 mg/mL bovine serum albumin. After incubation for 60 min at room temperature, the mixtures were diluted to a final concentration of 10-1 nM for the conjugates with TETBS containing 0.1 mg/mL bovine serum albumin and 10 mM D-biotin (Fluka), and the dilutions were incubated for an additional 30 min. STV-coated microplates were prepared as previously described (21), and 50 µL of a 600 nM solution of the biotinylated oligomer complements (capture oligomers) in TETBS was allowed to bind for 60 min. After the plate was washed with TETBS containing 10 mM D-biotin, 50 µL of the AP conjugate solution described above was allowed to hybridize for 60 min at room temperature. Negative controls were run

170 Bioconjugate Chem., Vol. 9, No. 2, 1998

in parallel by applying preconjugates to wells, coated with noncomplementary oligonucleotides. After incubation, the plate was washed four times with TETBS and twice with TBS [Tris-buffered saline, 20 mM Tris-HCl buffer (pH 7.3) and 150 mM NaCl]. AP-dependent color reactions were carried out with 200 µL of 1 M diethanolamine (pH 9.8) and 0.5 mM MgCl2 containing 10 mM pnitrophenyl phosphate (pNPP). The absorbance at 405 nm was determined with a Victor Multilabel-Counter (Wallac) after incubation for 10 min at 37 °C. Relative intensities (signal to noise, S/N values) were calculated from the observed absolute signal intensities divided by the corresponding background signal obtained for binding to noncomplementary capture oligomers, and typically fell in the range of 2-25, depending upon the preconjugate concentration. Standard deviations were calculated from the signal intensities of interest. Typically, duplicate determination of a particular sample revealed a percent error of 15%. BIA Measurements. Biospecific interaction analysis was performed using a BIAcore-1000 instrument from Pharmacia Biosensor AB (Uppsala, Sweden). Immobilization of Capture Oligomers. A continuous flow of hybridization buffer [150 mM NaCl, 30 mM sodium phosphate, 3 mM EDTA, and 0.25% Triton X-100 (pH 7.4)] at 10 µL/min and 25 °C was maintained throughout the entire analysis. Biotinylated capture oligomer (10 µM in hybridization buffer) was injected for 4 min over STV-coated sensorchips (SA-Chips, Pharmacia Biosensor AB). Loosely attached material was removed by three cycles of 10 µL of 50 mM NaOH. An amount of 500-2000 RU of the biotinylated oligomer complements was immobilized for kinetic analysis. Kinetic Analysis. Oligonucleotides or DNA-STV adducts at various concentrations in hybridization buffer, typically in the range of 400-25 nM, were injected over the DNA-specific surface at 10 µL/min for 4 min. The interaction was monitored continuously as the change in SPR signal. Dissociation was effected by injecting hybridization buffer only and was followed for 15 min. The surface was regenerated by injection of 10 µL of 50 mM NaOH. The entire process was automated. Binding and dissociation curves were recorded automatically, and analytical cycles were preprogrammed. The data analysis was carried out using BIAcore Evaluation 2.1 software (Pharmacia). The pseudo-first-order kinetic model was used, and parts of the binding curves where mass transport was evident were excluded from the calculations. A series of controls was carried out (data not shown). Injection of probes across surfaces, modified with noncomplementary oligonucleotides, revealed no significant increase in RU response, suggesting that the binding observed in the kinetic experiments is due exclusively to specific single-stranded DNA hybridization. Lowering the amount of coupling density of the biotinylated oligomer did not result in differences in kinetic constants, suggesting that any influence of rebinding (22) is negligible in these experiments. RESULTS AND DISCUSSION

Semisynthetic hybrids of DNA and STV were synthesized from various 5′-thiolated oligonucleotides (Table 1) by means of chemical cross-linking as previously described (4). In brief, STV was derivatized with maleimido groups using a heterobispecific cross-linker, and subsequently reacted with a thiolated oligonucleotide. The chief products, hybrids containing a single oligonucleotide moiety per STV, were purified by anion-exchange chromatography.

Niemeyer et al.

ELOSA Analysis of Conjugate Hybridization. As a relevant test for quantifying the hybridization characteristics of the DNA-streptavidin biomolecular adaptors, a microplate assay in the form of an enzyme-linked oligonucleotide sorbent assay (ELOSA) (23) was used (Figure 1). For this purpose, DNA-STV covalent adducts were conjugated to biotinylated AP. Initially, to determine the optimal conjugation stoichiometry, a fixed amount of the hybrid HA24 was mixed with varying quantities of the biotinylated enzyme. After incubation, the reaction was quenched by addition of free D-biotin. The resulting preconjugates (HA24-AP) were then allowed to bind to their corresponding oligomer complement (bcA24), previously immobilized on a STV-coated microplate (21), by using the STV-biotin interaction (see Materials and Methods for details). After hybridization, the amount of material immobilized was estimated from the enzymatic activity of AP, using colorimetric pNPP as a substrate. Negative controls were carried out by incubation of a similar preconjugate containing a noncomplementary oligonucleotide sequence, HC21-AP (prepared from 2.5 equiv of AP), as well as by incubation of biotinylated AP or buffer only. As shown in Figure 2, the low background signal obtained for the controls confirmed that binding occurred exclusively via specific single-stranded DNA hybridization. Relative intensities (signal-to-noise, S/N) of about 20 were obtained for a 1:0.25 molar ratio of the biotin binding sites of STV and the biotinylated enzyme. An increase of the amount of biotinylated AP did not lead to significantly stronger signals (Figure 2). These findings indicate that an approximately equimolar conjugation stoichiometry of the two protein components occurred despite the tetravalent binding capabilities of STV, and is likely due to steric hindrance of the voluminous AP. Furthermore, variations of the concentration of the biotinylated capture oligomer during immobilization on the microplate showed that saturation of the binding capacity of single microplate wells occurred at oligonucleotide concentrations of about 60 nM (Figure 2), which corresponds to an estimated biotin-binding capacity of about 1-2 pmol/well (24). Further experiments were carried out using DNASTV adducts containing the various oligonucleotide sequences of A24, B21, C21, D21, E21, and F24 as indicated in Table 1. The adducts were conjugated with biotinylated AP and allowed to bind to their immobilized complement at analyte concentrations ranging from 10 to 1 nM. Negative controls were carried out for all experiments by incubation of the 10 nM preconjugate solutions in microplate wells containing noncomplementary capture oligonucleotides. The S/N values observed were concentration-dependent and typically in the range of 25-2. Investigation of a large number of independently synthesized preconjugates gave percent errors of about 25%, thus satisfying the requirements for the reproducibility of chemical hybrid synthesis and AP conjugation as well as the microplate binding assay protocol. Time course experiments, carried out over varying periods of hybridization, revealed that maximum signal intensities are reached within 30 min (not shown). Therefore, the signal intensities obtained after 1 h of binding reflect equilibrium binding and, thus, the stability of a DNA duplex under the given reaction conditions. As indicated in Figure 3, the amount of hybridization varied to a great extent with the oligonucleotide sequence of the conjugates. The relative fraction of binding observed for particular sequences was linearly dependent on reagent concentration in the range of 10-1 nM. A

DNA−STV Kinetics

Bioconjugate Chem., Vol. 9, No. 2, 1998 171

Figure 2. Optimization of the preconjugation stoichiometry for ELOSA. Fixed amounts of DNA-STV adduct HA24 were mixed with varying amounts of biotinylated AP. The molar ratio of biotin-binding sites of the adduct to biotinylated AP was in the range of 1:2.5 to 1:0.025. Negative controls (NCs) were carried out with a comparable HC21-AP preconjugate, containing a noncomplementary oligonucleotide sequence (NC1), bAP only (NC2), and buffer only (NC3). The low signal intensities observed for controls indicate that binding is exclusively due to specific DNA base pairing.

Figure 3. Quantification of the sequence-dependent hybridization efficiency, as determined from ELOSA microplate analysis. The binding of several DNA-STV AP conjugates, containing oligonucleotide sequences A through F (see Table 1), was quantified in an ELOSA assay at analyte concentrations of 10, 5, and 1 nM, respectively. The relative amounts of preconjugate immobilized were calculated from enzyme-generated signals observed and normalized as a percentage fraction of the signal intensity obtained for 10 nM HA24-AP.

comparison of the hybridization efficiency, calculated from the average slope of concentration-dependent signal increase, allows a classification of the sequences studied. The 24-mer species HA24-AP exhibited the strongest binding (relative slope of 1.00). The preconjugates HF24-AP and HB21-AP bind less efficiently (0.53 and 0.56, respectively). Species HC21-AP and HD21-AP exhibited low hybridization efficiencies (0.19 and 0.20, respectively), and the binding of HE21-AP (0.04) was barely detectable. These results could not be predicted

from thermodynamic data of the particular oligomer sequences (19, 20), such as the calculated melting temperatures and free energies of duplex formation (TM and ∆G, see Table 1). Analysis of DNA Hybridization by SPR Measurements. To elucidate further the binding characteristics, the rate constants of association (kassn) and dissociation (kdiss) of single-stranded DNA interactions have been determined by biomolecular interaction analysis (BIA) using a biosensor based on surface plasmon resonance (SPR). With this technique, the refractive index within a gold sensor chip-immobilized dextran matrix is continuously monitored, plotted against time, and presented in a sensorgram. The change of SPR response, measured in resonance units (RU), allows a quantification of interaction between chip-immobilized and liquid phase-diluted binding partners. Here, the biotinylated oligomer complements, bcA through bcF (see Table 1), were immobilized on streptavidin-coated sensor chips, and real-time kinetic measurements at several analyte concentrations were carried out with the various oligonucleotides and the analogous DNA-STV hybrids. A series of controls was carried out by injecting oligonucleotides or DNA-STV adducts over sensor chips modified with noncomplementary capture oligomers (not shown). No significant change in SPR response indicated that the binding observed was due exclusively to specific single-stranded DNA hybridization, a result similar to that obtained from ELOSA. Although the data obtained from BIA for triplicate determinations of a particular sample were generally very reproducible (errors of 10%), deviations increased upon measuring identical samples after a certain period of storage and/or on a different sensor chip (percent error of up to 50%). As no degradation products of the probes could be detected by gel electrophoretic analysis, these alterations may reflect quality differences of the sensor surfaces or changes in, i.e., the salt concentration of the buffers that were used (10, 25). Implementation of a large number of measurements, including various (n g 5) independently synthesized analyte samples and statistical analysis of the data obtained, allowed elucidation of the trends discussed below. To confirm the applicability of BIA, a kinetic study of the hybridization of three oligonucleotides was carried

172 Bioconjugate Chem., Vol. 9, No. 2, 1998

Niemeyer et al.

Table 2. Kinetic Constants of Single-Stranded DNA Hybridization Obtained from BIA Measurements of the Binding of Thiolated Oligonucleotides TA16, TA21, and TA24 to the Immobilized Biotinylated Oligomer Complement bcA24a probe

103kassn (M-1 s-1)

10-7kdissn (s-1)

107KA (M-1)

TA16 TA21 TA24

148 ( 29 211 ( 13 216 ( 35

527 ( 41 446 ( 47 253 ( 63

2.8 4.7 8.6

a

For sequences, see Table 1.

Table 3. Kinetic Constants of Single-Stranded DNA Hybridization Obtained from BIA Measurements of the Binding of Thiolated Oligonucleotides (Tx) and the Corresponding DNA-STV Hybrid Molecules (Hx), Respectively, to Their Immobilized Biotinylated Oligomer Complementsa probe

immobilized material RUrelb

103kassn (M-1 s-1)

10-7kdissn (s-1)

TA24 HA24 TB21 HB21 TC21 HC21 TD21 HD21 TE21 HE21 TF24 HF24

0.90 ( 0.06 3.03 ( 0.44 0.60 ( 0.02 1.72 ( 0.27 0.14 ( 0.02 0.26 ( 0.05 0.13 ( 0.01 0.19 ( 0.03 0.03 ( 0.01 0.07 ( 0.03 0.65 ( 0.11 2.12 ( 0.71

216 ( 35 144 ( 44 69 ( 7 53 ( 16 13 ( 9 18 ( 9 14 ( 5 9(2 9(4 12 ( 1 149 ( 17 38 ( 13

253 ( 63 50 ( 17 307 ( 77 90 ( 69 456 ( 162 257 ( 62 386 ( 26 330 ( 63 1415 ( 105 966 ( 54 193 ( 83 98 ( 27

a For sequences of A24 through F24, see Table 1. b Relative response units (RUrel) were calculated from the RU value observed for immobilization of capture oligonucleotide and the RU value observed for analyte binding at a concentration of 200 nM.

Figure 4. Representative sensorgrams obtained from BIA. (A) Binding of covalent the STV-DNA adduct HA24 to sensorimmobilized oligomer complement, bcA24, carried out at several analyte concentrations. Computational analysis of binding and dissociation phases yields their reaction rate constants, kassn and kdiss, respectively. (B) Overlay of sensorgrams obtained for the binding of covalent STV adduct HA24 and oligonucleotide TA24 at a concentration of 0.4 µM to sensor-immobilized bcA24 (for sequences, see Table 1). The decrease in hybridization rate constants is detectable from eye inspection of the sensorgrams. The binding of HA24 proceeds over a period of more than 200 s, while the TA24 association curve has reached saturation after this period of time. The decreased kdissn of HA24 is clearly visible in a normalized overlay of sensorgrams, obtained for the dissociation phase (boxed inlay).

out. The affinity constants (KA ) kassn/kdiss) of the probes investigated were expected to increase in the order TA16 < TA21 < TA24, as they contain identical sequence stretches of increasing lengths and TA21 contains a single G f A mismatch (Table 1). The kinetic rate constants observed (Table 2) were 1.5 × 105, 2.1 × 105, and 2.2 × 105 M-1 s-1 for association (kassn) and 5.3 × 10-5, 4.5 × 10-5, and 2.5 × 10-5 s-1 for dissociation (kdiss) of TA16, TA21, and TA24, respectively. Although derived from near-surface hybridization, these values are in a range comparable to the range of those obtained in

homogeneous reactions. For example, in a recent study, diffusional fluorescence correlation analysis was used to measure the hybridization of a set of six oligodeoxyribonucleotides to a complementary 101-mer target RNA (17). Different probes, designed for equal thermodynamic stability, revealed association constants between 3 × 104 and 1.5 × 106 M-1 s-1. The similar association rates of TA24, TA21, and TA16 presented in Table 2 are consistent with the general principle of nucleic acid hybridization. The low activation barrier of duplex formation is associated with rate-limiting nucleation of a small number of base pairs, which is then followed by rapid hybridization of the remaining base pairs (26-28). As a consequence, oligonucleotides containing identical sequence stretches may reveal similar association rates. The decrease in stability, observed in the order TA24 > TA21 > TA16, is also affected by variations in dissociation, particularly in the case of TA24 versus TA21. This result is in agreement with a recent study of homogeneous nucleic acid hybridization carried out by Kool and co-workers (15). Sequence-Specific Hybridization Characteristics. Binding efficiency and kinetic rate constants have been determined for the hybridization of oligonucleotides and the analogous covalent STV adducts, respectively (Figure 4). The results obtained are summarized in Table 3. The amount of analyte bound to the sensor chip increased linearly with the amount of immobilized capture probe. Therefore, relative RU values (RUrel; see column 2 of Table 3) can be calculated for each experiment by dividing the RU obtained from analyte by the RU obtained from capture probe binding. Owing to the higher molecular mass, the RUrel values observed for the hybrids are generally greater than those of the corresponding oligonucleotide (Figure 4B). Normalization of the RUrel of the particular sequences as a percentage of the binding efficiency of sequence A24 allowed a comparison of the amount of hybridization measured by ELOSA or BIA. As shown in Figure 5, the similarity of the resulting patterns of sequence-dependent binding indicates good agreement between both methods. Furthermore, attachment of a protein moiety evidently results in only minor changes of sequence-specific hybridization efficiencies. Upon comparison of the kinetic data determined for the oligonucleotides of sequences A through F, differences

DNA−STV Kinetics

Figure 5. Comparison of the sequence-dependent hybridization efficiency obtained from ELOSA (left group of bars) and biospecific interaction analysis (BIA) of oligonucleotides (right group) or covalent DNA-STV adducts (middle group). The signal intensities observed were normalized as a percentage fraction of the signal intensities obtained from species containing the A24 sequence. The amount of material immobilized in BIA was calculated from relative RU values (see Table 3 and the text for details).

of up to 2 orders of magnitude could be observed for both kassn and kdissn with no obvious correlation with thermodynamic values Tm and ∆G (see Table 1). A distinct dependency of kassn and the formation of secondary structure via intramolecular hydrogen bonds competing with intermolecular hybridization has previously been reported (17, 29). The kassn values of the binding of oligonucleotides to target RNA correlated with secondary structures of probes and target sites. The lowest rate constants were observed for probes exhibiting tight stem-loop structures. Within this context, the structural characteristics of the DNA sequences studied here have

Bioconjugate Chem., Vol. 9, No. 2, 1998 173

been examined by software analysis, and the free energies (∆G) for intramolecular loop formation and intermolecular self-association were calculated, following a model based on adjacent dinucleotides (19, 20). As indicated in Figure 6, sequences A24, B21, and F24 form less stable intermolecular aggregates, while the G/C-rich sequences of C21, D21, and E21 produce more stable hairpin loops and dimeric duplexes. Although no aggregates could be detected by photometric, gel electrophoretic, or chromatographic analysis (not shown), dimerization or intramolecular folding of the probes will be present in equilibrium and bias the kinetics of DNA hybridization. However, a simplistic correlation with energies of homodimeric duplex or loop formation is evidently not sufficient to explain subtle differences, such as those observed in the hybridization characteristics of, e.g., B21 and F24 or C21 and E21. More complex models, combined with computational methods (8), may assist in predicting hybridization characteristics from mere sequence information. Influences of the STV Moiety on Hybridization Kinetics. A comparison of oligonucleotides and the analogous STV adducts (Table 3) exhibits some clearly perceptible differences. The latter exhibited reduced rate constants of hybridization which could already be detected from eye inspection of the sensorgrams (Figure 4B). The amount of decrease in the dissociation rate constant correlated with structural features of the oligonucleotide. The kdissn of linear sequences, as in HA24, HB21, and HF24, decreased approximately 2-5-fold, while those of dimer-forming sequences, as in HC21, HD21, and HE21, were only affected 1-2-fold. A similar trend was evident from the comparison of kassn values. The decrease was 2-3-fold for linear sequences, as in HA24, HB21, and HF24, while the annealing of HD21 was almost unaffected. The rates of structurally rich sequences, HC21 and HE21, were even slightly acceler-

Figure 6. Secondary structures of the oligonucleotides used in this study. Oligonucleotide duplexes and hairpin loops of the sequences have been calculated with Primer Select version 3.01a (DNA-Star Inc.). The numbers indicate the free energy of structure formation at 25 °C.

174 Bioconjugate Chem., Vol. 9, No. 2, 1998

Figure 7. Schematic model of covalent DNA-STV adducts. The putative attractive interaction between surfaces of the protein and nucleic acid moiety is decreased by secondary structures of the DNA strand, e.g., in HE21. For linear sequences (e.g., in HA24), a closer contact of the surfaces influences the process of intermolecular hybridization to a greater extent.

ated by the protein attached. These phenomena were further investigated by melting temperature analysis of the duplex structures formed by DNA-STV conjugates or the analogous oligonucleotides with their oligomer complements (data not shown). Tm values in agreement with the predicted thermodynamic parameters were obtained for sequences A through F. However, a slight decrease in Tm of approximately 1-2 °C was observed for the protein conjugates, indicating a weak, duplexdestabilizig effect of the STV moiety under conditions in which the influences of DNA native structure as well as intramolecular DNA-protein interactions are reduced. A possible model mechanism for explaining these data is based on intramolecular attractive interactions between the protein and nucleic acid moiety of the hybrid (Figure 7). An attraction of the surfaces may be mediated by positively charged lysine or arginine residues of STV and the negatively charged phospate backbone of the DNA. This would restrict the degrees of freedom of the nucleic acid, and thus reduce the number of potential nucleation sites along the DNA strand. As a consequence, both association and dissociation velocities would be reduced. Additional steric hindrance of the protein may also affect the hybridization process. The fact that the influence of the STV molecule is lower on dimer- or loop-forming sequences is in agreement with this model; linear strands of nucleic acids might be affected to a greater extent, due to a more intimate interaction of their surface with the STV moiety (see HA24 in Figure 7). The slight increase observed for association of HC21 and HE21 suggests influences of an additional mechanism, recently suggested by Corey and co-workers, for explaining an accelerated hybridization of oligonucleotides that are conjugated with positively charged peptides or a basic protein, staphylococcal nuclease, to double-stranded DNA targets (30-32). As a result of Coulomb attraction of the protein moiety, the DNA domain of the probe is held in close proximity to the DNA template, thereby increas-

Niemeyer et al.

ing its effective concentration near its target site. This mechanism, however, appears to be of minor importance in the case of rapid single-stranded DNA hybridization but may be significant for slow, single-stranded DNA hybridization reactions. We have investigated the sequence-specific near-surface hybridization of both covalent DNA-protein conjugates and their analogous oligonucleotides at nondenaturing temperatures. The hybridization efficiencies, determined for binding to surface-immobilized complements, varied to a great extent with the base composition of the nucleic acid and correlated with DNA native structure rather than standard thermodynamic parameters. As demonstrated in a microplate assay, covalent DNA-STV adducts are convenient reagents for the quantification of sequence-dependent binding efficiency. The kinetic rate constants of the covalent adducts and the analogous unmodified oligonucleotides have been determined by SPR measurements. A comparison of the data suggests intramolecular interactions between the protein and nucleic acid moiety within the adducts. This study emphasizes the importance of structural influences on hybridization reactions proceeding with kinetic control, such as the supramolecular assembly of delicate DNA-tagged molecular devices. It is difficult to draw systematic conclusions at this stage since much remains to be learned about hybridization of nucleic acids near surfaces or tagged with macromolecules. Nevertheless, binding will profit from sequences specially optimized for fast and efficient hybridization. ACKNOWLEDGMENT

We thank D. Blohm for generous support and valuable discussion, S. Rohlfing, S. Behrens, and M. Adler for experimental help with the microplate assay, and T. Ferdelman for critical reading of the manuscript. Financial support for this work was obtained from Deutsche Forschungsgemeinschaft, Fonds der Chemischen Industrie, and To¨njes-Vagt Stiftung Bremen. LITERATURE CITED (1) Seeman, N. C. (1982) Nucleic acid junctions and lattices. J. Theor. Biol. 99, 237-247. (2) Seeman, N. C. (1997) DNA components for molecular architecture. Acc. Chem. Res. 30, 357-363. (3) Robinson, B. H., and Seeman, N. C. (1987) The design of a biochip: a self-assembling molecular-scale memory device. Protein Eng. 1, 295-300. (4) Niemeyer, C. M., Sano, T., Smith, C. L., and Cantor, C. R. (1994) Oligonucleotide-directed self-assembly of proteins: Semisynthetic DNA-streptavidin hybrid molecules as connectors for the generation of macroscopic arrays and the construction of supramolecular bioconjugates. Nucleic Acids Res. 22, 5530-5539. (5) Mirkin, C. A., Letsinger, R. L., Mucic, R. C., and Storhoff, J. J. (1996) A DNA-based method for rationally assembling nanoparticles into macroscopic materials. Nature 382, 607609. (6) Alivisatos, A. P., Johnsson, K. P., Peng, X., Wilson, T. E., Loweth, C. J., Bruchez, M. P., Jr., and Schultz, P. G. (1996) Organization of ‘nanocrystal molecules’ using DNA. Nature 382, 609-611. (7) Lipshutz, R. J., Morris, D., Chee, M., Hubbell, E., Kozal, M. J., Shah, N., Shen, N., Yang, R., and Fodor, S. P. A. (1995) Using oligonucleotide probe arrays to access genetic diversity. BioTechniques 19, 442-447. (8) Lockhart, D. J., Dong, H., Byrne, M. C., Follettie, M. T., Gallo, M. V., Chee, M. S., Mittmann, M., Wang, C., Kobayashi, M., Horton, H., and Brown, E. L. (1996) Expression monitoring by hybridization to high-density oligonucleotide arrays. Nat. Biotechnol. 14, 1675-1680.

DNA−STV Kinetics (9) Burke, D. T., Burns, M. A., and Mastrangelo, C. (1997) Microfabrication technologies for integrated nucleic acid analysis. Genome Res., 189-197. (10) Wood, S. J. (1992) DNA-DNA hybridization in real time using BIAcore. Microchem. J. 47, 330-337. (11) Parkhurst, K. M., and Parkhurst, L. J. (1993) KineticStudies of Oligonucleotide-DNA Hybridization in Solution by Fluorescence Resonance Energy-Transfer. Biophys. J. 64, A266-A266. (12) Morrison, L. E., and Stols, L. M. (1993) Sensitive Fluorescence-Based Thermodynamic and Kinetic Measurements of DNA Hybridization in Solution. Biochemistry 32, 3095-3104. (13) Application Note 306 (1994) Kinetic characterization of DNA hybridization using real time BIA. Pharmacia Biosensor AB, Uppsala, Sweden. (14) Gotoh, M., Saito, K., Shimizu, M., and Tosu, M. (1994) A novel approach for determination of mismatch formation effects on DNA hybridization kinetics using a biosensor. Nucleic Acids Symp. Ser. 31, 121-122. (15) Wang, S. H., Friedman, A. E., and Kool, E. T. (1995) Origins of High Sequence Selectivity: A Stopped-Flow Kinetics Study of DNA/RNA Hybridization by Duplex-Forming and Triplex-Forming Oligonucleotides. Biochemistry 34, 97749784. (16) Parkhurst, K. M., and Parkhurst, L. J. (1995) KineticStudies by Fluorescence Resonance Energy-Transfer Employing a Double-Labeled Oligonucleotide: Hybridization to the Oligonucleotide Complement and to Single-Stranded DNA. Biochemistry 34, 285-292. (17) Schwille, P., Oehlenschlager, F., and Walter, N. G. (1996) Quantitative Hybridization Kinetics of DNA Probes to RNA in Solution Followed by Diffusional Fluorescence Correlation Analysis. Biochemistry 35, 10182-10193. (18) Nilsson, P., Persson, B., Uhlen, M., and Nygren, P. A. (1995) Real-Time Monitoring of DNA Manipulations Using Biosensor Technology. Anal. Biochem. 224, 400-408. (19) Breslauer, K. J., Frank, R., Blo¨cker, H., and Marky, L. A. (1986) Predicting DNA duplex stability from the base sequence. Proc. Natl. Acad. Sci. U.S.A. 83, 3746. (20) Freier, S. M., Kierzek, R., Jaeger, J. A., Sugimoto, N., Caruthers, M. H., Neilson, T., and Turner, D. H. (1986) Improved free-energy parameters for predictions of RNA duplex stability. Proc. Natl. Acad. Sci. U.S.A. 83, 93739377.

Bioconjugate Chem., Vol. 9, No. 2, 1998 175 (21) Niemeyer, C. M., Adler, M., and Blohm, D. (1997) Fluorometric Polymerase Chain Reaction (PCR) Enzyme-Linked Immunosorbent Assay for Quantification of Immuno-PCR Products in Microplates. Anal. Biochem. 246, 140-145. (22) Nieba, L., Krebber, A., and Plu¨ckthun, A. (1996) Competition BIAcore for measuring true affinities: Large differences from values determined from binding kinetics. Anal. Biochem. 234, 155-165. (23) Whetsell, A. J., Drew, J. B., Milman, G., Hoff, R., Dragon, E. A., Adler, K., Hui, J., Otto, P., Gupta, P., Farzadegan, H., and Wolinsky, S. M. (1992) Comparison of three nonradioisotopic polymerase chain reaction-based methods for detection of human immunodeficiency virus type 1. J. Clin. Microbiol. 30, 845-853. (24) Adler, M. (1996) Auswahl und Etablierung alternativer Methoden zur Detektion von Immuno-PCR Produkten. Diploma Thesis, Universita¨t Bremen, Bremen, Germany. (25) Keller, G. H., and Manak, M. M. (1989) DNA Probes, Stockton, New York. (26) Wetmur, J., and Davidson, N. (1968) Kinetics of renaturation of DNA. J. Mol. Biol. 31, 349-370. (27) Po¨rschke, D., and Eigen, M. (1971) Co-operative non-enzymic base recognition. 3. Kinetics of the helix-coil transition of the oligoribouridylic-oligoriboadenylic acid system and of oligoriboadenylic acid alone at acidic pH. J. Mol. Biol. 62, 361-381. (28) Kool, E. T. (1997) Preorganization of DNA: Design principles for improving nucleic acid recognition by synthetic oligonucleotides. Chem. Rev. 97, 1473-1487. (29) Lima, W. F., Monia, B. P., Ecker, D. J., and Freier, S. M. (1992) Implication of RNA Structure on Antisense Oligonucleotide Hybridization Kinetics. Biochemistry 31, 2055-2061. (30) Corey, D. R. (1995) 48000-fold Acceleration of Hybridization by chemically modified Oligonucleotides. J. Chem. Am. Soc. 117, 9373-9374. (31) Iyer, M., Norton, J. C., and Corey, D. R. (1995) Accelerated Hybridization of Oligonucleotides to Duplex DNA. J. Biol. Chem. 270, 14712-14717. (32) Smulevitch, S. V., Simmons, C. G., Norton, J. C., Wise, T. W., and Corey, D. R. (1996) Enhancement of strand invasion by oligonucleotides through manipulation of backbone charge. Nat. Biotechnol. 14, 1700-1704.

BC970170+