520
Bioconjugate Chem. 2000, 11, 520−526
Mechanism of Intermolecular Purine-Purine-Pyrimidine Triple Helix Stabilization by Comb-Type Polylysine Graft Copolymer at Physiologic Potassium Concentration Anwarul Ferdous,† Toshihiro Akaike, and Atsushi Maruyama* Department of Biomolecular Engineering, Faculty of Bioscience and Biotechnology, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8501, Japan. Received November 30, 1999; Revised Manuscript Received March 31, 2000
We previously reported a novel strategy to stabilize purine motif triplex DNA within a mammalian gene promoter at physiologically relevant pH, temperature, and potassium (K+) concentrations by a comb-type poly(L-lysine)-graft-dextran copolymer [Ferdous et al., (1998) Nucleic Acids Res. 26, 39493954]. Here we describe the major contribution(s) of the copolymer to stabilize the purine motif triplex DNA at physiological K+ concentrations. Self-aggregation through guanine-quartet formation of guanine-rich (G-rich) triplex-forming oligonucleotides (TFOs) has long been proposed for K+-mediated inhibition of the purine motif triplex formation. However, this was not the case for the severe inhibitory effect of K+ observed under our reaction conditions. Rather significant decrease in rate of triplex formation involving a G-rich TFO was a major factor to confer K+ inhibition. Interestingly, in the presence of the copolymer the rate of triplex formation was tremendously increased and K+-induced dissociation of preformed triplexes was not observed. Moreover, the triplex-promoting/stabilizing efficiency of the copolymer was amazingly higher than that of physiological concentrations of spermine. An absolute increase in binding constant of the TFO to the target duplex could therefore be the predominant mechanistic source for the copolymer-mediated triplex stabilization under physiological conditions in vitro.
INTRODUCTION
Over a decade, oligonucleotide-based strategy by triple helix (triplex) formation (i.e., the antigene strategy) through sequence-specific interaction of triplex-forming oligonucleotides (TFOs)1 with the target duplex has attracted a great deal of attention for its possible role to alter gene functions (1-8). Unfortunately, extreme instability of both purine motif (2, 9, 10) and pyrimidine motif (1, 11) triplexes under physiological conditions severely limited their therapeutic applications. In the pyrimidine-motif, pyrimidine-rich TFOs bind parallel to the homopurine strand of the duplex by Hoogsteen hydrogen bonding to form T‚A:T and C+‚G:C triplets (1, 12). Since the cytosine residues in the pyrimidine-rich TFOs are to be protonated to bind with the guanine of the G:C duplex, an acidic condition (pH < 6) is needed for the pyrimidine-motif triplex formation and thus impedes its utility in vivo (1, 11, 12). On the other hand, purine-rich TFOs run antiparallel to the homopurine strand and bind by reverse Hoogsteen hydrogen bonding to form A‚A:T (or T‚A:T) and G‚G:C triplets (2, 9, 10). Despite the fact that physiological concentrations of various monovalent cations (M+), particularly K+ (13, 14), severely inhibit the purine motif triplex formation, its stability at physiologic pH makes it more suitable to use in vivo. * To whom correspondence should be addressed. Phone and Fax: +81-45-924-5122. E-mail:
[email protected]. † Present address: University of Texas, Southwestern Medical Center at Dallas, Department of Internal Medicine, TX. 1 Abbreviations: PLL-g-Dex, polylysine-graft-dextran copolymer; EMSA, electrophoretic mobility shift assay; G-rich, guaninerich; M+, monovalent cations; ODN, oligodeoxynucleotide; TFO, triplex-forming oligonucleotide; t-DNA, calf thymus DNA.
For a long time, it has been proposed that an intraand/or intermolecular self-association through guaninequartet (G-quartet) and parallel or antiparallel homoduplex formation of G-rich TFOs confer K+ inhibition of the purine motif triplex formation (15-18). But a decrease in rates of association and/or an increase in rates of dissociation of triplexes have also been suggested for K+-mediated inhibition of triplex formation (14, 17, 19, 20). We reported that a comb-type poly(L-lysine)-graftdextran copolymer (Figure 1A), which hereafter will be referred to as the PLL-g-Dex copolymer, significantly stabilizes triplexes (21-23) and abrogates M+ inhibition of the purine motif triplex formation involving a 30-mer target duplex from rat collagen R1 (I) gene promoter and its specific G-rich TFO, Pu (Figure 1B) (22, 23). Although a chemical modification strategy of guanine in G-rich TFOs has been shown to increase triplex stability partially by reducing G-quartet formation (20, 24-28), the putative mechanism(s) involved in the copolymer-mediated triplex stabilization with unmodified Pu is not yet clearly understood. The main interest of our present study is, therefore, to define a mechanistic explanation for the copolymermediated triplex stabilization under physiologic [K+]. Results from electrophoretic mobility shift assays (EMSAs) indicate that under our reaction conditions Pu is relatively insensitive to K+ to form G-quartet structure. On the other hand, kinetic studies strongly suggest that tremendous decrease in the rate of triplex formation (i.e., the binding rate of Pu to duplex). However, addition of the copolymer in the reactions significantly increased the binding rate of Pu to duplex DNA (i.e., promotes triplex formation) and also decreased dissociation rate of Pu. Finally, the triplex-promoting/stabilizing efficiency of the
10.1021/bc990166t CCC: $19.00 © 2000 American Chemical Society Published on Web 06/17/2000
Triplex Stabilization by Graft Copolymer
Figure 1. Schematic representations of the copolymer and ODNs used in this study. (A) Structural formula of the PLL-gDex copolymer. Preparation and characterization of the copolymer has been described previously in detail (21, 29). Degree of substitution (n) is 0.2. Number-averaged polymerization degrees of PLL and dextran are 200 and 36.5 (m), respectively. (B) Sequences of the ODNs. ODN sequences of the target duplex (T-1 and T-2) and TFOs used in this study have been described before (22). In brief, T-2 was end-labeled and mixed with T-1 to prepare the labeled duplex for triplex analysis. Purine-rich (Pu) and control (C) TFOs are shown aligned with the triplex-forming target region in duplex (shown in bold).
copolymer was remarkably higher than that of physiologic concentrations of spermine. EXPERIMENTAL PROCEDURES
Oligonucleotides and t-DNA. Oligodeoxynucleotides (ODNs) for target duplex DNA (T-1 and T-2), purine-rich (Pu) and control (C) TFOs used in this study were obtained and processed as described previously (22). In brief, all ODNs were purified by gel electrophoresis on a 15% denaturing polyacrylamide gel, dissolved in 10 mM tris-acetate (pH 7.0), and quantitated by UV spectroscopy. Preparation of stock solution of calf thymus DNA (t-DNA) (Sigma-Aldrich, Japan) and its use to eliminate copolymer-DNA interaction before electrophoresis were described in detail (22, 23). Preparation of PLL-g-Dex Copolymer. Poly(Llysine) HBr (PLL HBr, Mw ) 4.5 × 104) and dextran T-10 (Dex) (Mn ) 5.9 × 103) were obtained from Peptide Institute, Inc., Osaka, Japan and Pharmacia Biotech., Uppsala, Sweden, respectively. Preparation, isolation, and characterization of the PLL-g-Dex copolymer has been described in detail (21, 29). Briefly, the copolymer (Dex content ) 90 wt %, Mn ) 2.5 × 105, as free salt) was prepared by a reductive amination reaction between PLL HBr and Dex in dimethyl sulfoxide using NaBH3CN as a catalyst. Electrophoretic Mobility Shift Assays (EMSAs). To analyze G-Quartet Formation of Pu. Pu was endlabeled with [γ-32P]ATP (Amersham) and T4 polynucleotide kinase and labeled Pu free from unincorporated [γ-32P]ATP was purified on a Sephadex G-25 column (Pharmacia Biotech.) according to the manual. After freeze-drying, Pu was dissolved in 10 mM tris-acetate (pH 7.0) and was stored at -30 °C. All TFOs (Pu and C) were heated at 65 °C for 10 min to prevent self-aggregation and then quickly cooled on ice. To effect G-quartet formation, 10 nM labeled Pu was mixed with unlabeled Pu for the indicated final concentrations of Pu. Control TFO (C) was then added as carrier DNA to adjust the equimolar concentration (5 µM) of TFOs (Pu + C) as well
Bioconjugate Chem., Vol. 11, No. 4, 2000 521
Figure 2. EMSA of Pu aggregation. (A) Pu does not form selfaggregates in the presence or absence of KCl and the copolymer. To effect self-aggregation, Pu (0.17 µM) containing trace amount of labeled Pu was mixed with control TFO for final 5 µM concentration of DNA in buffer containing 50 mM Tris-acetate (pH 7) and 10 mM MgCl2 (buffer A). Reaction mixtures (9 µL) were incubated for 6 h at indicated temperatures (°C) either in the presence (+) or absence (-) of KCl and 2.5 µg of the copolymer (1 µM). Samples were resolved through 15% native polyacrylamide gel with (+) or without (-) adding thymus DNA (t-DNA). In lane D, formamide-denatured (10 min at 90 °C) sample was used without incubation. The gel was dried and exposed to Kodak Bio-max film at -80 °C. An arrow shows the position of monomeric form of Pu. (B) Effect of Pu concentration on K+-induced self-aggregation. Reactions were carried out identically as described in panel A, except that Pu at the indicated final concentrations was incubated with (+) or without (-) KCl and then analyzed as above.
as to avoid nonspecific adsorption of Pu to the tube walls and subsequent losses during processing. Required amount of control TFO was also added in all other experiments for final 5 µM concentration of TFOs as described (22). Reaction mixtures (9 µL) in 50 mM tris-acetate (pH 7.0) and 10 mM MgCl2 (buffer A) were incubated for 6 h at different temperatures either in the presence or absence of KCl (150 mM) and the copolymer at a copolymer/DNA charge (P/D) ratio of 2 as described (21, 22). After incubation, 2 µL of 50% glycerol solution containing bromophenol blue and 1 µL of reaction buffer with or without t-DNA (6 µg) were added. Samples were immediately resolved at room temperature by electrophoresis at 8 V/cm for 5 h through a 15% native polyacrylamide gel containing 2% glycerol, 45 mM tris-borate, 1 mM EDTA, and 10 mM KCl. This minimal K+ concentration was also present in the electrophoresis buffer to maintain G-quartet integrity (30). Dried gels were then exposed to Kodak Biomax film at -80 °C. For Triplex Assays in Which Pu Was Added with the Target Duplex with or without Preincubation in the Presence or Absence of K+ and the Copolymer. Pu (0.34 µM) in 6 µL of buffer A was preincubated for 6 h at 37 °C in the presence or absence of 140 mM KCl and the copolymer as described above. Target duplex (3 ng) in buffer A (6 µL) with or without KCl (140 mM) was then
522 Bioconjugate Chem., Vol. 11, No. 4, 2000
Ferdous et al.
Figure 3. Coanalysis of self-aggregation and triplex formation by Pu. Reaction protocol that allowed coanalysis of KCl effects on Pu aggregation and triplex formation is shown diagrammatically above the gel. Reaction mixtures (6 µL) in buffer A containing Pu (0.34 µM) alone (lanes 5-7, 10, and 11), were preincubated (PI) for 6 h at 37 °C in the presence (+) or absence (-) of the copolymer and KCl (140 mM). Labeled duplex (3 ng) with (lanes 7 and 11) or without (lanes 1, 5, 6, and 10) 140 mM KCl in 6 µL of buffer A was then added and incubated for an additional 6 h to effect triplex formation at the indicated final concentrations of KCl. Triplex formation was also performed under the same conditions but without preincubating Pu (lanes 2-4, 8, and 9). t-DNA was added in reactions containing the copolymer and samples were electrophoresed to separate the duplex (D) and triplex (T) DNAs as described before (22).
added to the preincubated Pu to reach the final Pu concentration of 0.17 µM and further incubated for an additional 6 h for triplex formation at the indicated final concentrations of KCl. Triplex formation was also analyzed under identical conditions but by adding the target duplex and Pu with or without K+ and the PLL-g-Dex copolymer simultaneously. Samples were resolved on a 15% native polyacrylamide gel prepared in buffer A containing 2% glycerol and triplex and duplex DNA bands were analyzed as described before (22). To Monitor the Association Rates of Triplex Formation in the Presence or Absence of K+ and Triplex Stabilizers. Triplex-forming reaction mixtures (9 µL) in buffer A containing 3 ng of target duplex, Pu (0.17 µM), and control TFO with or without KCl (140 mM), 1 mM spermine, and 1 µM of the copolymer (at P/D ratio of 2) were incubated at 37 °C for the indicated time period as shown in the figure. After each respective incubation time, triplex formation was analyzed as described above. For Assays Monitoring K+-Induced Dissociation of Preformed Triplexes. Triplex-forming reaction mixtures (6 µL) in buffer A containing target duplex, Pu, and control TFO with or without triplex stabilizers were incubated for 6 h at 37 °C to facilitate equivalent level of triplex formation (see, Figure 4). Then 3 µL of buffer A with or without KCl was added and dissociation of the preformed triplex DNA was monitored thereafter for the indicated time. Calculation of Triplex Formation by EMSAs. The amount of radioactivity present in the duplex and triplex forms was determined with the Ambis System. The apparent fraction, θ, of the target duplex bound by Pu was calculated using the following equation:
θ ) Striplex/(Striplex + Sduplex)
(1)
where Striplex and Sduplex represent the radioactive signal for the triplex and duplex bands, respectively, as described previously (26).
Figure 4. Kinetics of the purine motif triplex formation in the absence of KCl. (A) Analysis of triplex formation with or without triplex stabilizer. Kinetic analysis of triplex formation was performed according to the protocol shown above the gels. Labeled duplex (3 ng), Pu (0.17 µM), and control TFO were added together in 9 µL of buffer A and then incubated at 37 °C in the absence (None) or in the presence of either 2.5 µg of PLLg-Dex copolymer (1 µM and P/D ratio is 2) or 1 mM spermine (Spm) as shown on the right side of each gel. Triplex formation after indicated time period (h) was analyzed as described in Figure 3. Arrows indicate the position of duplex (D) and triplex (T) DNAs. (B) Triplex formation curves in the presence or absence of stabilizers. The fraction of duplex in triplex form (θ) was calculated from eq 1 as described in Experimental Procedures and plotted against the incubation time. Each point represents an average of two independent experiments.
The apparent dissociation rate constant (koff) for triplexes involving Pu was determined by taking the leastsquares fitting of kinetic data to the eq 2 (17):
θt/θ0 ) e-kofft
(2)
where t is the time after K+ addition, θt is the apparent fraction of target duplex bound by Pu at time t, and θ0 is the apparent fraction of target duplex bound by Pu at time zero. The relative koff values were obtained from the following equation:
rel koff ) koff(K+)/koff(0)
(3)
where koff(K+) is the value of koff of the reaction mixture supplemented with KCl, and koff(0) is the value of koff for the reaction mixture supplemented with buffer A. RESULTS
Correlation between Self-Aggregation of Pu and Triplex Inhibition by K+. Self-aggregation through
Triplex Stabilization by Graft Copolymer
Figure 5. EMSA to analyze the effect of K+ on association rates of triplex formation. (A) The copolymer accelerates the association rate of triplex formation in the presence of K+. Kinetic analysis of triplex formation with or without triplex stabilizers was carried out identically as described in Figure 4 but in the presence of 140 mM KCl. Autoradiography of the gel for triplex formation in the presence (+) or absence (-) of PLL-g-Dex copolymer and K+ after indicated time intervals (h) is shown. Identical experiments were performed with or without spermine (gels are not shown). (B) Triplex formation curves at 140 mM KCl in the presence or absence of triplex stabilizer. The fraction of duplex in triplex form (θ) was calculated and plotted against the time periods (h) as described in Figure 4B, except that θ in the absence of KCl under each condition (e.g., lane 10 for the copolymer) was considered to be 1 and calculated the values in the presence of KCl at each time point in the absence (None) or presence of 1 µM PLL-g-Dex copolymer and spermine (Spm). Each point represents an average of two independent experiments.
G-quartet, parallel and antiparallel homoduplex formation by G-rich ODNs has long been proposed to confer K+ inhibition of the purine motif triplex formation (1518). To explore this possibility, self-aggregation of Pu at physiologic [K+] was tested. Figure 2A shows that Pu at 0.17 µM containing labeled Pu (∼10 nM) almost remains in monomeric form regardless of the presence or absence of KCl and the incubating temperature (cf. lane D with 1, 5, 9, and 13). Note that no retarded band was also observed in the presence of the copolymer and t-DNA (lanes 4, 8, and 12). Since higher concentrations of G-rich ODNs did favor G-quartet formation (15, 17), selfaggregation of Pu at different concentrations was also analyzed but still failed to detect any retarded band (Figure 2B). These data suggest that K+-mediated inhibition of triplex formation that we reported previously (22, 23) and also observed in this study (see Figures 3 and 5) does not correlate with the self-aggregation hypothesis. However, the presence of the copolymer at higher concentrations (1.7-5 µM) of Pu with KCl slightly induced aggregation of Pu (data not shown).
Bioconjugate Chem., Vol. 11, No. 4, 2000 523
Cheng and van Dyke (31) proposed that, in the presence of K+, certain classes of self-associated species of G-rich ODNs had no apparent change in electrophoretic mobility to their monomeric form. Keeping this possibility in mind, an experiment for coanalysis of self-aggregation and triplex formation by Pu was designed in the presence or absence of K+ and the copolymer. We found that triplex formation involving preincubated Pu in KCl, indeed, reduced slightly to that formed without preincubation (Figure 3, cf. lanes 3 and 4 with 6 and 7, respectively). Therefore, possibility of self-aggregation and/or unusual structure formation of Pu in the presence of K+ cannot be totally overruled. However, this inhibitory effect was unaffected in the presence of the copolymer (cf. lanes 4, 7, 9, and 11) and spermine (data not shown). This suggests that the copolymer-mediated triplex stabilization at physiological [K+] (cf. lanes 2, 4, and 9) may have nothing to do with self-aggregation and/or G-quartet formation of Pu. Effect of the Copolymer on Kinetics of Triplex Formation. Since preincubation of Pu with K+ neither formed aggregates nor significantly inhibited triplex formation (Figures 2 and 3), we, like others (14, 17, 19), presumed that K+ affects either association or dissociation rates of triplex DNA to confer triplex inhibition. First, kinetics of triplex formation in the presence or absence of K+ and triplex stabilizers were analyzed. As shown in Figure 4A, in the absence of K+, the purine motif triplex DNA formed very rapidly (i.e., a matter of minutes) even at submicromolar concentration of Pu (0.17 µM). In addition, triplex formation reached to almost equilibrium with or without triplex stabilizers after 4-6 h incubation (Figure 4B). Note that the rate of triplex formation either in the presence of the copolymer or at the physiological concentration (1 mM) of spermine was indistinguishable. Next we examined the kinetics of triplex formation at physiologic [K+] either in the presence (see the gel) or absence (gels not shown) of the copolymer (Figure 5A). In all cases, the fraction of duplex in triplex form (θ) after 6 h incubation without KCl (e.g., lane 10) was considered to be 1 and compared to that formed in the presence of KCl after indicated incubation time. As shown graphically in Figure 5B, the rate of triplex formation was tremendously decreased and only 0.22 of θ was obtained after 6 h incubation of Pu with the target duplex (compare Figures 4 and 5). On the contrary, the rate of triplex formation in the presence of the copolymer was quite surprising under the same conditions. Triplex formation (θ ≈ 0.9) reached to plateau within 15 min of incubation (Figure 5A, lane 3) and remained fairly constant throughout the time course (lanes 3-8). Whereas, with 1 mM spermine only a modest increase in rate was observed after 1 h (Figure 5B), having consistency with a previous study (32). Therefore, even at 20-fold less in stabilizer/DNA charge ratio the triplex-promoting efficiency (i.e., increase the binding rate of Pu) of the copolymer at physiologic [K+] is amazingly higher than that of spermine. These data are in good agreement with our previous observations (22, 23). Effect of the Copolymer on K+-Induced Dissociation of Preformed Triplexes. It has been reported that triplex inhibition by K+ occurs not only by slowing the rate of triplex formation but also by increasing the rate of triplex dissociation (i.e., destabilization of preformed triplexes) (17, 19). Although the copolymer significantly abrogates the inhibitory effect of K+ on rates of triplex formation (Figure 5), K+-induced destabilization of preformed triplex DNA and the effect of the copolymer were
524 Bioconjugate Chem., Vol. 11, No. 4, 2000
Ferdous et al. Table 1. Comparison of Apparent koff Values Obtained from Figure 6Ba stabilizer
koff (s-1)
rel koff
none spermine PLL-g-Dex
6.5 × 10-5 2.8 × 10-5
