542
Bioconjugate Chem. 2002, 13, 542−547
Mixed-Phase Hybridization of Short Oligodeoxyribonucleotides on Microscopic Polymer Particles: Effect of One-Base Mismatches on Duplex Stability Kaisa Ketoma¨ki, Harri Hakala,† and Harri Lo¨nnberg* Department of Chemistry, University of Turku, FIN-20014 Turku, Finland. Received September 5, 2001; Revised Manuscript Received December 27, 2001
Hybridization of short oligonucleotides (10- and 11-mers) to complementary probes immobilized to microscopic polymer particles was quantified by a sandwich type mixed-phase hybridization assay based on a time-resolved fluorometric measurement of a photoluminescent europium(III) chelate from the surface of a single particle. Among the 54 sequences that were studied, 21 were fully complementary to the particle-bound probes, while 33 contained an internal one-base mismatch. The observed affinities were compared to those predicted by the nearest-neighbor model. In addition, various factors, such as the pore size of the particle, the linker structure, the charge type of the probe, and the efficiency of agitation, that might be expected to affect the kinetics of mixed-phase hybridization have been examined.
INTRODUCTION
We have recently described a mixed-phase hybridization assay that is based on the use of a mixture of categorized polymer microparticles as the solid phase (1). The method allows simultaneous screening of several potential base mutations or deletions from a single PCRamplified sample, and it hence offers an alternative for the widely used oligonucleotide microarrays (2-6). Each particle bears, in addition to a given oligonucleotide probe, prompt fluorophores that allow recognition of the particle category when the particles after hybridization are subjected one by one to the fluorescence measurement. The hybridization is, in turn, quantified by timeresolved fluorometry. On application of a sandwich type protocol, the PCR-amplified oligonucleotide forms first in solution a duplex with an appropriately designed oligonucleotide probe labeled with a photoluminescent lanthanide chelate, and hybridizes then with the particlebound probe (Scheme 1). The lanthanide ion emission is measured directly from a single particle after complete quenching of the emission of the prompt fluorophores (7). The photoluminescent lanthanide chelates exhibit as markers one major advantage over organic dyes; the Stokes shift, i.e., the difference between the wavelengths of emission and excitation, is large, and hence, the concentration quenching is negligible. Accordingly, the intensity of the emission signal measured from an individual particle is strictly linearly related to the number of chelates on the particle over 5-6 orders of magnitude (8-10). For this reason, the method suits quantitative analysis of the concentration of oligonucleotides in the sample exceptionally well (1). All mixed-phase hybridization assays, using either microarrays or categorized microparticles, are eventually based on the dependence of the stability of solid-supported duplexes on the sequence. It has been well established (11-17) that the stability of oligodeoxyribo† Present address: Wallac Oy, P.O. Box 10, FIN-20101 Turku, Finland.
Scheme 1. Principle Hybridization Assaya
of
the
Sandwich
Type
a P1, particle-bound oligonucleotide probe; A1-A54, target oligonucleotides studied; O3, fluorescently tagged oligonucleotide probe.
nucleotide duplexes in aqueous salt solutions may be rather reliably predicted by applying the so-called nearest-neighbor model for helix propagation. The same approach has more recently been used to determine the nearest-neighbor parameters for all possible internal point mismatches (16, 18-21). Since the applicability of a mixed-phase hybridization assay is largely determined by the selectivity of hybridization, it is important to know how well the nearest-neighbor parameters referring to hybridization in solution may be used to predict the affinity for particle-bound probes. We have shown previously (1, 8) that the microparticle approach described above allows, at least in favorable cases, unequivocal detection of a point mutation or deletion, provided that the length of the complementary region is carefully optimized. For example, a single AG or CT mismatch in the middle of a 12-mer complementary sequence has been shown to decrease the affinity for the particle-bound probe by 2 orders of magnitude (8). Deletion of a single internal A or C has also been shown to retard the hybridization of a 12-14-mer probe by up to 2 orders of magnitude, although the destabilization
10.1021/bc0100859 CCC: $22.00 © 2002 American Chemical Society Published on Web 03/28/2002
Effect of One-Base Mismatches on Duplex Stability
appears to be rather sequence dependent (1, 8). It is, however, known that all base mismatches are not as strongly destabilizing in solution. In fact, the so-called linearly independent nearest-neighbor ∆G° parameters for dimeric 5′-NX-3′ fragments, where X stands for a mismatched nucleoside, range from -1.1 to 1.3 kcal/mol (16, 18-21), while those for the corresponding fully Watson-Crick bonded fragments fall between -2.2 and -0.6 kcal/mol (17). To find out how much various internal one-base mismatches in various molecular environments decrease the affinity of oligonucleotides for particle-bound probes, 21 fully complementary short oligonucleotides (10- and 11-mers) and 33 oligonucleotides having one internal point mismatch have been subjected to the sandwich type mixed-phase hybridization assay described in Scheme 1. The observed affinities are compared to those predicted by the nearest-neighbor model. In addition to these equilibrium studies, various factors, such as the pore size of the particle, the linker structure, the charge type of the probe, and the efficiency of agitation, that might be expected to affect the kinetics of mixedphase hybridization have been examined. EXPERIMENTAL PROCEDURES
Preparation of Oligonucleotide-Coated Microparticles. The polymer particles used to immobilize the oligonucleotide probes were the products of SINTEF. They were made of a copolymer of glycidyl methacrylate (40%) and ethylene dimethylacrylate (60%), and functionalized with primary amino groups (1 mmol/g) obtained by reacting the particle-bound epoxy groups with diethylenetriamine. The particles were uniformly sized, having a diameter of either 50 (P1) or 59 µm (P2). The smaller particles (P1) had an average pore size of 30 nm, a total pore volume of 0.822 mL/g, and a surface area of 137 m2/g, and the larger ones (P2) had a pore size of 50 nm, a total pore volume of 1.469 mL/g, and a surface area of 92 m2/g. A desired proportion (10-30 µmol/g, i.e., 1-3% ) of the amino groups of P1 and P2 was acylated with (4,4′-dimethoxytrityloxy)acetic acid, and the rest of the groups were carefully capped by repeated acetylations (22). The oligonucleotide probes were assembled on these supports via a normal phosphoramidite strategy, as described previously (10) in detail. Oligonucleotides and Their Fluorescently Tagged Conjugates. Oligonucleotides were assembled via a normal phosphoramidite strategy and purified by ionexchange HPLC. Preparation of the fluorescently tagged oligonucleotides, O1-O3, has been described previously (8, 9). Hybridization Assays. The hybridization assays in kinetic measurements were carried out by shaking 50 particles (P1 or P2) in 10 µL of Tris buffer (50 mM, pH 7.5, containing 0.5 M NaCl and 0.01% Tween 20) that contained O1 at a concentration of 17 nmol/L (9). The reaction temperature was 25 °C, and aliquots were withdrawn from 15 min to 24 h. The particles were thoroughly washed, and a single particle was subjected to a time-resolved fluorescence measurement (7). The sandwich type hybridization assays used to study the dependence of the duplex stability on sequence were carried out by shaking 50 particles (P1) in 10 µL of the Tris buffer described above. The concentration of both the target oligonucleotide (A1-A54) and the fluorescently tagged oligonucleotide (O3) was 17 nmol/L. The reaction temperature was 25 °C and the reaction time 24 h. The particles were washed, and the fluorescence emission was quantified by the single-particle technique (7). Typically,
Bioconjugate Chem., Vol. 13, No. 3, 2002 543
10 individual particles were measured, the standard deviation ranging from 5 to 20%. Calibration. To convert the fluorescence emission measured from a single particle to the number of fluorescently tagged oligomers attached to this particle, two sets of particles P1a were subjected to hybridization with tagged oligomer O1 or O3, the initial concentration of which was varied from 5.1 to 51 nmol/L. After incubation for 20 h, one set of the particles was subjected to fluorescence measurement by the single-particle technique (7), while from the other set, europium(III) ion was released to solution and quantified according to the DELFIA protocol (25). Additionally, the concentration of O1 or O3 in solution before and after the hybridization was determined by the DELFIA protocol. The calibration line was measured separately for every particle that was employed. RESULTS AND DISCUSSION
Preparation of Oligonucleotide-Coated Microparticles and Characterization of Their Hybridization Properties. Two different types of uniformly sized porous particles made of a copolymer of glycidyl methacrylate (40%) and ethylene dimethylacrylate (60%) and functionalized with amino groups were used as the solid support in the mixed-phase hybridization assays. Usually, 50 µm particles (P1) having an average pore size of 30 nm and a surface area of 137 m2/g were employed. However, slightly larger (59 µm) particles (P2) having a larger pore size (50 nm) and a smaller surface area (92 m2/g) were also used for comparative purposes. To allow the assembly of oligonucleotides on the particles according to the normal phosphoramidite protocol, a desired proportion of the amino functions was acylated with (4,4′-dimethoxytrityloxy)acetic acid (DMTrOCH2COOH), while the rest were carefully capped by repeated acetylations (22). The level of N-(DMTrOCH2CO) loading of the particles, when determined by the DMTr cation assay (23), ranged from 12 to 28 µmol/g. The oligonucleotide chains were then assembled on the deprotected hydroxy function. The amide linker that was employed turned out to be entirely resistant to deprotection of the base moieties by ammonolysis. We have previously shown that oligonucleotide synthesis on similar polymer particles gives full-length 16-mer sequences in an only 30% yield, the rest of the sequences being truncated (10). Despite this heterogeneity of the oligonucleotide probe, the hybridization efficiency and selectivity of such particles have been demonstrated to be comparable to those obtained by postsynthetic immobilization by purified fully homogeneous probes (10). A fluorescently tagged 42-mer oligodeoxyribonucleotide, 5′-d(CTATATTCATCATAGGAAACACCAAAGATGATAATTX5C)-3′ (O1), was used to elucidate the hybridization properties of the oligonucleotide-coated particles. Here, X stands for N4-(6-aminohexyl)-2′-deoxycytidine tethered to a photoluminescent europium(III) chelate, {2,2′,2′′,2′′′-[(4′-{4′′′-[(4,6-dichloro-1,3,5-triazin2-yl)amino]phenyl}-2,2′:6′,2′′-terpyridine-6,6′′-diyl)bis(methylenenitrilo)]}tetrakis(acetato)europium(III) (24). The sequence complementary to the particle-bound 16mer is indicated in bold. The kinetics of the binding of O1 to the particles was followed by shaking 50 particles in 10 µL of Tris buffer (50 mM, pH 7.5, containing 0.5 M NaCl and 0.01% Tween 20) that contained O1 at a concentration of 17 nmol/L at 25 °C. The particles were thoroughly washed, and a single particle was subjected to a time-resolved fluorescence measurement (7). To
544 Bioconjugate Chem., Vol. 13, No. 3, 2002
Ketoma¨ki et al.
Table 1. Effect of the Oligo(T) Linker on the Efficiency and Kinetics of the Hybridization of the Fluorescently Tagged Oligonucleotide (O1) to a Complementary Oligonucleotide Probe Assembled on Microparticles P1 and P2 particle
sequence
τ1/2 (h)a
n(bound)/ n(total)
P1a P1b P1c P1d P1e P1f P1g P2a P2b P2c
3′-T3-[oligo]-5′ b 3′-T15-[oligo]-5′ 3′-T30-[oligo]-5′ 3′-T50-[oligo]-5′ 3′-T3-[oligo]-5′ c 3′-T3-[oligo]-5′ d 3′-T3-[TATAAGTAGTATCC]-5′ 3′-T3-[oligo]-5′ 3′-T50-[oligo]-5′ 3′-T100-[oligo]-5′
7 7 7 7 7 7 7 7 7 7
0.6 0.6 0.6 0.3 0.05 0.6 0.6 0.6 0.6 0.6
The half-life for hybridization at 25 °C. b The abbreviation [oligo] stands for the common hybridizing sequence 3′-d(GATATAAGTAGTATCC)-5′. c A methylphosphonate oligodeoxyribonucleotide. d A2′-O-Methylribonucleotide oligomer.
Figure 1. Logarithmic equilibrium constants, log{n(bound)/ [n(total) - n(bound)]}, for the binding of oligonucleotides A1A54 to their complementary particle-bound probes (see Table 2) plotted vs the logarithmic stability constants of the respective duplexes in solution.
estimate the contribution of unspecific adsorption of the labeled oligomer to the particles, a comparative study with a noncomplementary fluorescently tagged 32-mer, 5′-d(X20ATCATCTTTGGT)-3′ (O2), was carried out. Consistent with previous results (8-10), the unspecific binding was in all cases almost negligible; less than 1% of the O2 was attached to the particles. The fluorescence emission of a single particle was converted to the number of fluorescently tagged oligomers (O1) attached to the particle with the aid of a calibration line, which was obtained by plotting the emission intensities measured from a single particle against the amount of europium(III) released to solution and quantified by the DELFIA protocol (25). The amount of europium(III) in solution before and after the hybridization, i.e., the concentration of O1, was determined similarly. The decrease in the concentration of O1 upon hybridization always agreed with the amount of europium(III) released from the particles after hybridization, which lent considerable support to the reliability of the determination. The calibration line that was obtained was essentially similar to that reported earlier for a closely related system (8). We have shown previously that the mixed-phase hybridization assays described above suffer from one major shortcoming, viz., slow binding of oligonucleotides from the solution phase to the particle-bound probes (810). On using five particles per 1 µL sample, the halflife of the equilibration is on the order of hours at 25 °C. The half-life is inversely proportional to the number of particles in a unit volume of the sample, and the process may hence simply be accelerated by using more particles (9). This, however, leads to decreased sensitivity; the analyte is distributed among a larger number of particles, and the emission from an individual particle is reduced. Accordingly, faster kinetics of hybridization is highly desirable. One might a priori assume that insertion of a long hydrophilic spacer between the hybridizing oligonucleotide sequence and the solid support accelerates the hybridization. To test this assumption, oligonucleotide probes consisting of a 16-nucleotide 5′-terminal hybridizing sequence, 3′-d(GATATAAGTAGTATCC)-5′, and a 3-100-nucleotide 3′-terminal oligo(T) spacer were assembled on P1 and P2. Unfortunately, introduction of such a spacer did not result in the desired acceleration. As seen from Table 1, the kinetics was not markedly affected either by the length of the spacer or by the pore size of the particle. Within the limits of experimental
errors, the half-life remained constant (7 h) with both particles. In contrast, the hybridization efficiency, i.e., the mole fraction of O1 bound to the particle, seemed to be moderately affected by these factors. With porous particles P2, having a large pore size (50 nm), the hybridization efficiency (60%) did not depend on the length of the spacer, but with particles P1 having smaller pores (30 nm), the efficiency was decreased to less than half of the original value when the spacer was longer than 30 nucleotide units. Addition of a detergent to the hybridization buffer also had some effect on the efficiency but not on the kinetics of hybridization. In the presence of 0.01% Tween 20, the hybridization efficiency was slightly increased, when the concentration of the target oligonucleotide was low. At high concentrations, the influence remained negligible. The chemical nature of the support-bound probe did not have any effect on the hybridization kinetics. In addition to oligodeoxyribonucleotide probes, deoxyribonucleotide methylphosphonate and 2′-O-methylribonucleotide oligomers were assembled on particles P1, and their ability to bind O1 was studied. The methylphosphonate probes exhibited a markedly reduced hybridization efficiency (5%) compared to the unmodified deoxynucleotide probes (60%), while the 2′-O-methylribonucleotide oligomers exhibited a slightly enhanced affinity for O1. Unfortunately, no positive effect on the hybridization kinetics could be observed in either case. The only factor that, in addition to the number of particles in a unit volume, had any influence on the hybridization kinetics was the efficiency of agitation. When the solution of O1 containing five particles (P1) in 1 µL was mixed by rotating a sealed tube at a rate of 20 rpm, the half-life of hybridization was ∼7 h at 25 °C. Ultrasonic mixing reduced the half-life to ∼3 h and efficient agitation with a planar shaker (Vortex) to ∼1 h (data not shown). This observation, together with the fact that the hybridization is accelerated with the increasing number of particles in the sample, strongly suggests that the rate of hybridization is limited by the diffusion of the oligonucleotide in solution, and not by the binding event itself. Dehybridization of O1 from support P1g, bearing a 14mer sequence complementary to O1, was also studied. At 25 °C, the half-life for the release of O1 from P1g was ∼15 h. Accordingly, the particles may well be rapidly washed without any loss of the hybridized oligonucleotide. However, at 40 °C, the dehybridization was surprisingly fast, the half-life being
