Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (114K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Shindo, H.
Right arrow Articles by Shimizu, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shindo, H.
Right arrow Articles by Shimizu, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 1997 Oxford University Press 4786-4791

Dynamics in the isomerization of intramolecular DNA triplexes in supercoiled plasmids

Dynamics in the isomerization of intramolecular DNA triplexes in supercoiled plasmids Heisaburo Shindo*, Nobuaki Matsumoto and Mitsuhiro Shimizu

School of Pharmacy, Tokyo University of Pharmacy and Life Science, 1432-1 Horinouchi, Hachioji, Tokyo 192-03, Japan

Received July 28, 1997; Revised and Accepted October 7, 1997

ABSTRACT

We report here kinetic and thermodynamic studies on differential isomerization of intramolecular Pyr*Pur·Pyr triplexes in supercoiled plasmids. Two structural isomers of the triplex exist: one with the 3'-half of the Pyr strand as the third strand (H-y3 form) and the other with the 5'

-half as the third strand (H-y5 form). The relative populations of the two triplex isomers was determined using the chemical probe with diethyl pryrocarbonate as a function of incubation time. The results demonstrated that triplexes were formed rapidly after a pH change from pH 8.0 to 5.0 and that the initial population of the two isomers exponentially changed with incubation time to reach true thermodynamic equilibrium with a time constant of 0.6-10 h, depending on temperature and the presence of Mg2+. The results clearly demonstrated that interconversion occurs between the two isomers and that the presence of Mg2+ generally retarded the interconversion rates. Kinetic and thermodynamic analyses of the relative populations of the two isomers revealed that the apparent energy barrier for transition from duplex to the H-y3 form is higher than that to the H-y5 form, but H-y3 is more stable in enthalpy terms than H-y5. Therefore, H-y3 is kinetically inferior but thermodynamically favored at higher supercoil levels in plasmids. The presence of Mg2+ resulted in both a kinetic and a thermodynamic preference for H-y5 formation, relative to the H-y3 form.

INTRODUCTION

A polypurine-polypyrimidine sequence with mirror repeat can adopt an intramolecular triplex structure (also called H-DNA) which consists of a triple-stranded stem in one half of the mirror sequence and a single purine strand (see 1 -3 for reviews). A major type of the triplex structure is composed of canonical base triplets via Hoogsteen hydrogen bonding, C*G·C and T*A·T, forming a Pyr*Pur·Pyr triplex (Pyr* strand as the third strand). Two structural isomers exist for an intramolecular Pyr*Pur·Pyr triplex (see 1 ,3 ,4 for reviews): one with the 3'-half of the Pyr strand as the third strand (H-y3 isomer) and the other with the 5'-half as the third strand (H-y5 isomer). Formation of H-y3 was greatly preferred over H-y5 in most of the sequences reported (5 -11 ).

Htun and Dahlberg (8 ) first demonstrated differential formation of the H-y3 and H-y5 isomers depending on the supercoil density, which was taken into account by an inequivalent topology of two isomers in the nucleation step of triplex formation. A model proposed by them could explain supercoil-dependent behavior of differentiation of triplex isomers in terms of the sense of pre-nucleation rotation at the center of Pur·Pyr tracts, but it was not sufficient to understand the other factors affecting triplex isomerization. Parniewski et al. (12 ) showed formation of a mixture of the two isomers for the sequence (AG)7ATCGATATATSTCG(AG)7. Kang et al. (13 ,14 ) demonstrated that the presence of divalent cations is one of the crucial factors in formation of H-y5 for the sequence (GAA)4TTCGC(GAA)4. We showed that a higher G+C content in the central region of Pur·Pyr tracts and divalent metal ions both caused a preference for the H-y5 isomer (15 ). All these results together led us to conclude that the size of the denaturation bubble or the melted region at the center of a Pur·Pyr tract is critical in determining differential preference for triplex isomers (15 ), i.e. larger opening of the base pairs is required to wrap the pyrimidine strand into the H-y3 form than into the H-y5 form. A recent kinetic study with oligonucleotide intramolecular triplexes by Roberts and Crothers (16 ) supported the above notions and, moreover, they explained the `isomer paradox of H-form DNA, i.e. a bias toward H-y3 formation' in terms of pre-nucleation geometry and accessibility of the major groove.

Htun and Dahlberg (8 ) suggested that the two isomers are not interconvertible but that their populations would be a kinetic consequence of the topological difference. Our previous study (15 ) also emphasized that the rate limiting step in intramolecular formation involves partial melting of the duplex near the center of Pur·Pyr tracts and that a higher G+C content at the center requires a higher supercoil density for triplex formation. On the other hand, because more negative supercoils are relaxed when H-y3 is formed (8 ,10 ,15 ), this isomer is thermodynamically more stable than H-y5 under negative supercoil stress. As pointed out by Glover et al. (10 ) and Shimizu et al. (15 ), the folded back loop in the pyrimidine strand of the H-y5 form is less accessible to solvent than that of H-y3, suggesting some differences in the base stacking or base pairing mode in the two isomers. Thus differential isomerization may be understood as a combination of thermodynamic stability and the kinetic terms in triplex formation.

We now ask whether the populations of the two isomers are determined kinetically or energetically and if interconversion could occur. To this end we determined the kinetic and thermodynamic parameters on relative preference and lifetime for the two isomers by chemical probing. We have shown that the two isomers are indeed interconvertible, with characteristic rates depending on temperature and concentration of Mg2+ cation, and that true thermal equilibrium can be attained within 1-20 h at room temperature or higher.

MATERIALS AND METHODS

Construction of plasmid pMS215 was described previously (15 ). Plasmids were grown in Escherichia coli JM109 and prepared using Magic Max-prep (Promega).

Modification of plasmids by diethyl pyrocarbonate (DEPC; Sigma) was performed as described (9 ,17 ,18 ). Briefly, 3 µg plasmid in TE buffer (10 mM Tris-HCl, 0.1 mM EDTA, pH 8.0) were diluted into 100 µl 0.3 M Tris-acetate buffer, pH 5.0, in the presence or absence of 50 mM MgCl2 and incubated for the given periods of time. The reaction was initiated by addition of 10 µl DEPC and incubated with vortexing at the given temperature. After 30 min the reaction was stopped by precipitation of the DNA with ethanol. The DNA was digested with HindIII and SacI endonucleases and the purine strand of the HindIII-SacI fragments was radioactively labeled at the HindIII site by filling in the 5'-overhang with the Klenow fragment of DNA polymerase (New England Biolabs) in the presence of [[alpha]-32P]dATP (~3000 mCi/mmol; ICN), dGTP, dCTP and dTTP. The radioactively labeled inserts were isolated by gel electrophoresis and treated with piperidine at 90°C for 20 min. A sequencing gel was run and the DEPC modification sites were visualized by autoradiography.

RESULTS

We have previously shown that the sequence (GA)7TGGC(AG)7 in the supercoiled plasmid pMS215 used here predominantly forms the H-y3 isomer in the absence of Mg2+, while the H-y5 isomer is dominant in the presence of 50 mM Mg2+ (15 ). Therefore, this plasmid will provide a suitable example for both kinetic and thermodynamic studies of triplex isomerization. We used chemical probing with DEPC to quantify the relative populations of the two isomers in pMS215, since it preferentially reacts with purine residues (A > G) in the single-strand region of the 5'- and 3'-half of the purine strand in H-y3 and H-y5 respectively, but is much less reactive with the duplex form.

To elucidate dynamic isomerization of intramolecular triplex formation, fine mapping of chemical modification sites by DEPC in supercoiled plasmid pMS215 was performed as a function of incubation time under three different experimental procedures, as shown in Figure 1 . Supercoiled plasmid pMS215 dissolved in TE buffer (10 mM Tris-HCl, pH 8.0, 0.1 mM EDTA) was added to 0.3 M Tris-acetate buffer, pH 5.0, without Mg2+ at the given temperature and after appropriate time intervals chemical probing with DEPC was carried out for 30 min. This is denoted experiment A. In the second experiment, B, the supercoiled plasmid (pH 8.0) was added to 0.3 M Tris-acetate buffer, pH 5.0, in the presence of 50 mM Mg2+, followed by the same procedures as in experiment A. In the third experiment, C, the supercoiled plasmid was added to 0.3 M Tris-acetate buffer, pH 5.0, in the absence of Mg2+ and incubated for 30 min. After addition of MgCl2 solution to 50 mM, chemical probing was carried out at the given intervals of incubation time.


Figure 1. Flow charts in experiments A, B and C employed for a kinetic study of isomerzation of the intramolecular triplex in supercoiled plasmid pMS215.


Figure 2. Fine mapping of DEPC sites as a function of incubation time in supercoiled plasmid pMS215, performed under three experimental procedures A, B and C. (a) Autoradiograms. Lanes A, B and C represent experimental procedures A, B and C respectively. The chemical probe analysis was performed 0.25 (left panel) and 24 h (right panel) after triplex was formed. (b) Densitometric scans of the autoradiograms at 0.25 and 24 h in each experiment A, B and C as shown in (a).

A set of fine mappings of the (GA)7TGGC(AG)7 insert in pMS215 is shown in Figure 2 a, where lanes A, B and C correspond to experiments A, B and C respectively. Densitometric scans are shown in Figure 2 b. In experiment A the intensity of the 5'-half of the purine strand was high relative to that of the 3'-half at incubation time 0.25 h and it increased further at 24 h, indicating that the relative population of the H-y3 isomer increased with incubation time. In experiment B the intensity of the 3'-half was predominant over the 5'-half at t = 0.25 h and decreased with incubation time, indicating that the relative population of the H-y5 isomer was high at the initial time but decreased with incubation time. In experiment C the intensity of the 5'-half of the purine strand was nearly comparable with that of the 3'-half at t = 0.25 h and slightly decreased at 24 h, indicating that the population of the H-y5 form increased with time. Such a time dependence of the relative populations of the two isomers indicate that interconversion occurs between the two triplex isomers.

Table 1 . Relative population of H-y3 isomer and lifetimes (h) of intramolecular triplexes in supercoiled plasmid pMS215 in the presence and absence of Mg2+ at pH 5.0 and at various temperatures
Temperature (°C) 0 mM Mg2+ 50 mM Mg2+
  H-y3 (%) (t = 0) H-y3 (%) (t = [infin]) Lifetime of H-y5a H-y3 (%) (t = 0) H-y3 (%) (t = [infin]) Lifetime of H-y5a Lifetime of H-y3b
25 52.7 63.8 1.8 ± 0.4 24.2 38.8 15.8 ± 1.9 -
35 56.1 76.1 1.6 ± 0.2 29.3 44.1 4.1 ± 0.9 9.1 ± 1.8
45 59.7 79.5 0.9 ± 0.4 32.2 49.4 3.4 ± 0.9 -
aThe lifetimes of H-y5 were measured as the time constant in hours (or reciprocal of the rate constant) for conversion of the H-y5 to H-y3 form as shown in curves A and B in Figure 3.
bThe lifetime of H-y3 was measured only at 35°C with a large error because population changes due to conversion from H-y3 to H-y5 were generally very small, as shown in curve C in Figure 3.

As reported by Kang et al. (19 ) and Hanvey et al. (20 ), transition from the duplex to the triplex form under supercoil stress was found to be rapid, reaching completion in <10 min. Therefore, it is reasonable to assume that transition will be complete during the time of chemical probing (30 min) adopted here and that only triplexes are present.

To quantify the relative populations of the two isomers, the regions of three central gel bands of seven hyper-reactive adenine sites on each 5'-half and 3'-half of the purine strand were chosen to avoid flanking effects and the integrated intensities were measured from the densitometric scans in Figure 2 b. The relative population of H-y3 (%) was then calculated and plotted as a function of incubation time at 25 and 35°C in Figure 3 . This revealed an exponential decay to equilibrium at all temperatures studied. For example, as shown in Figure 3 b, the fraction of H-y3 increased with a time constant of 1.6 h in the case of experiment A and reached an equilibrium value of 76%, whereas it decreased with incubation time in the case of experiment C and reached an equilibrium value of 44% at 24 h. The former indicates conversion of H-y5 to H-y3 and the latter indicates the reverse conversion. As expected, the fraction of H-y3 was nearly identical between experiments A and C at t = 0 (actually t = 0.25 h, corresponding to half of the reaction time for chemical probing). Furthermore, the fraction at equilibrium in experiment B was nearly identical to that in experiment C. All these results are consistent with the idea that dynamic interconversion occurs between the two isomers.


Figure 3. Time course of relative population of the H-y3 isomer, performed in experiments A, B and C at 25 (a) and 35°C (b). The curves correspond to: solid line, experiment A; dotted line, experiment B; broken line, experiment C.

The time course of the relative populations of the two isomers is reasonably expressed by the following equation.
f3(t) = f3[infin] - (f3[infin] - f3°) exp ( - t/[tau]) 1

Here f3 is the fraction of H-y3 at time t, f3° is the fraction at t = 0, f3[infin] is the fraction at equilibrium and [tau] is the time constant or the lifetime of the H-y3 or H-y5 form. The fractions at t = 0 and at equilibrium and the time constants for conversion were estimated by non-linear least squares curve fitting to the data shown in Figure 3 a and b and the averaged parameters at various temperatures over several repeated experiments are listed in Table 1 .

To further verify the occurrence of interconversion between two isomers, we carried out another experiment, D (Fig. 4 a): the triplexes were initially formed at 50 mM Mg2+ in 0.3 M Tris-acetate, pH 5.0, and then EDTA, pH 5.0, was added to 50 mM, followed by passage through a spin column to remove Mg2+ ions. The eluted solution was subjected to DEPC treatment at appropriate time intervals after removal of Mg2+. Autoradiograms of the sites modified with DEPC (Fig. 4 b) clearly show a large change with time in the relative populations, from the H-y5 form to H-y3. The equilibrium populations in experiment D were nearly identical to those at corresponding temperatures in experiment A (Fig. 3 b), although the initial population of the isomers was quite different from that observed in experiment A. Plots of the relative population versus incubation time (Fig. 4 c) demonstrated that a large change with time in the fraction of H-y3 occurs with time constants of 2.0, 1.8 and 0.6 h at 25, 35 and 45°C respectively, the values of which were nearly identical to those obtained in experiment A within experimental error (see Table 1 ). Thus these results conform with the idea that interconversion from the H-y5 to the H-y3 form indeed occurs.


Figure 4. Time course of relative population of the H-y3 isomer as a function of incubation time in experiment D. (a) Procedures in experiment D, (b) fine mapping of modification sites with DEPC as a function of incubation time at 35°C and (c) time course of relative population of the H-y3 isomer. The experiment was performed at 25 (solid line), 35 (dotted line) and 45°C (broken line).

Several points should be noted from Table 1 and Figure 3 . (i) There are at least two processes involved, one is transition from duplex to two triplex isomers and the other is interconversion between the H-y5 and H-y3 forms; (ii) the former transition is completed within the time of chemical probing (30 min); (iii) the latter transition is slow, with a time constant of 1-10 h at 35°C, depending on Mg2+ concentration; (iv) Mg2+ ions have two effects, the divalent cation facilitates H-y5 formation relative to H-y3 in the duplex-triplex transition and retards the interconversion rate between the two isomers by 2-8 times (Table 1 ).

As mentioned above for experiments A and B, the relative populations of H-y3 and H-y5 at t = 0 represent the transition probabilities for duplex to two triplex isomers. Therefore, the ratio of the transition probabilities of the two isomers can be written as follows.
ln(f3°/f5°) = ln(k3/k5) - (Ea3 - Ea5)/RT 2

Here, Ea3 and Ea5 are the apparent activation energies for transition from the duplex to the H-y3 and H-y5 forms respectively and k3 and k5 are the pre-exponential factors which may correspond to the probabilities of H-y3 and H-y5 formation from each transition state respectively. Arrhenius plots for the fraction ratio f3°/f5° at t = 0 are shown in Figure 5 and the differences in the apparent activation energies of the two isomers from the duplex are listed in Table 2 . Positive values of activation energy differences (2.6-3.7 kcal/mol) indicate that formation of the H-y5 isomer is kinetically more favorable than the H-y3 isomer, but unfavorable in terms of k3/k5. The latter result would be consistent with higher accessibility of the third pyrimidine strand to the major groove of the duplex at the pre-nucleation step for the H-y3 than for the H-y5 form of oligonucleotides (16 ).


Figure 5. Arrhenius plots of relative transition probability, f3°/f5°, from duplex to triplex isomers at incubation time t = 0, in the presence (dotted line) and absence (solid line) of 50 mM Mg2+.

Table 2 . Kinetic and thermodynamic parameters for the transition from duplex to triplexes and of interconversion from the H-y5 to the H-y3 isomer in supercoiled plasmid pMS215 at pH 5.0 and 35°C
  Ea (kcal/mol) ln k3/k5 [Delta]G° (kcal/mol) [Delta]H° (kcal/mol) [Delta]S° (cal/mol/K)
0 mM Mg2+ 2.6 ± 0.1 7.5 -0.58 -7.5 -22.5
50 mM Mg2+ 3.7 ± 0.1 10.2 0.14 -3.8 -12.8

Equilibrium constant K for the interconversion process H-y5 <-> H-y3 is given by the fraction ratio f3[infin]/f5[infin] at equilibrium and its temperature dependence gives the enthalpy change [Delta]H° for the reaction process.
d(ln K)/dT = d(ln f3[infin]/f5[infin])/dT = -[Delta]H°/RT2 3

The enthalpy change for conversion from H-y5 to H-y3 ([Delta]H°) can be estimated by van't Hoff plots of equation 3, as shown in Figure 6 . Other thermodynamic parameters, free energy ([Delta]G°) and entropy change ([Delta]S°), were calculated using the basic equation [Delta]G° = -RT ln K = [Delta]H° - T[Delta]S°, which are also listed in Table 2 . Enthalpy and entropy changes obtained here may not be temperature independent for such a complex system, so that these values should only be considered as apparent values.


Figure 6. van't Hoff plots of relative population f3[infin]/f5[infin] at equilibrium in the presence (dotted line) and absence (solid line) of 50 mM Mg2+.

Free energy difference [Delta]G° was small between the two isomers, although enthalpy change [Delta]H° was significant (Table 2 ). The relatively large enthalpy change can be explained by the fact that H-y3 formation relaxes supercoils in plasmids by one more turn than H-y5 (8 ,10 ,15 ). A reduced magnitude of [Delta]H° in the presence of Mg2+ indicates that Mg2+ ions stabilize H-y5 more than H-y3 (discussed later).

The activation barrier for conversion from H-y5 to H-y3 was generally high, as judged from the extremely slow conversion rates (see Table 1 ). The conversion rate became slower in the presence of 50 mM Mg2+. This result would suggest that conversion occurs through the duplex and that Mg2+ ions stabilize the triplex structure relative to the duplex as an intermediate in the conversion process.

DISCUSSION

The mechanism of intramolecular triplex formation includes a denaturation bubble at the dyad axis of the Pur·Pyr tract and nucleation, followed by elongation of the triplex stem with strand separation of the duplex (8 ,10 ,15 ). In this work we have studied the kinetic and thermodynamic properties of isomerization in intramolecular triplex formation and demonstrated that interconversion between the two isomers occurs.

The time dependence of the relative populations of the two triplex isomers may arise from two causes. One of the isomers may be generated as a result of interconversion from the other or it could be newly generated from the duplex form remaining unchanged in the plasmid at the initial time. The latter case is unlikely for the following reasons. Transition probabilities for transition from the duplex to the two triplex isomers are considered to be constant, independent of time under a given experimental condition. Assuming that the isomers are generated from the duplex fraction and that no conversion between two isomers occurs, the relative populations of the two isomers should be time independent according to kinetic theory for a first order competitive reaction. However, this was not the case, as can be seen in Figure 3 a and b. Furthermore, as mentioned in Results, it seems clear that triplex formation in supercoiled plasmids was rapid, being complete within ~10 min (19 ,20 ). Hanvey et al. (20 ) demonstrated that triplexes were formed within a few minutes at 25°C after a pH jump from pH 8.0 to 5.0, as judged from S1 nuclease sensitivity of the Pur·Pyr insert in supercoiled plasmids. Thus all these results verify that the duplex is short lived and that interconversion indeed occurs between the two isomers.

Interconversion could reasonably be assumed to occur only through the duplex as a short-lived intermediate because the process will take the energetic minimum pathway. Unwrapping of the pyrimidine strand from the triplex stem initiated at the duplex-triplex junction at the 3'- or 5'-end of the pyrimidine strand would be followed by duplex formation with the complementary single-stranded purine loop in a concerted manner. Thus the free energy cost due to an increase in negative supercoiling in the unwrapping process will be compensated for by the stacking energy due to duplex formation with the single-stranded purine loop. The rate of conversion from H-y5 to H-y3 is faster than that of the reverse conversion (see the last two columns in Table 1 ), which may be accounted for by the asymmetric nature of the energy barriers for transition from duplex to two triplex isomers, as mentioned below.

The barrier for transition from duplex to the H-y5 form is significantly lower by 2.6 kcal/mol in the absence of Mg2+ and 3.7 kcal/mol in the presence of Mg2+ than the barrier for H-y3, consistent with our notion (15 ) that H-y3 formation requires a wider region of base pair opening than does H-y5. Such a disadvantage in H-y3 formation would be overcome by excess free energy due to the difference in relaxation of supercoils between H-y3 and H-y5 in the transition states. Therefore, H-y3 has a low transition probability from the duplex relative to H-y5 at lower supercoil densities, while it is predominant at high levels of supercoiling.

Three main factors affecting the relative populations of the two isomers can be considered at the equilibrium state (15 ): negative supercoil density, G+C content in the central region of the Pur·Pyr tract and divalent cations. Free energy change at equilibrium for conversion from H-y5 to H-y3 may contain the following supercoil-dependent and structural energy term
[Delta]G° = [Delta]Gsc° + [Delta]Gst° 4

Here [Delta]Gsc° is the free energy change associated with supercoil density at equilibrium and [Delta]Gst° is a structural term that may be related to the hairpin loop in the central region of the Pur·Pyr tract of the triplex structure. The supercoil-dependent term is given as follows (21 -23 ),
[Delta]Gsc = 1100 RTn [([Delta]Lk + [Delta]L3r)2 - ([Delta]Lk + [Delta]L5r)2] 5

Here, [Delta]L3r and [Delta]L5r are the amount of relaxation of supercoils due to H-y3 and H-y5 formation respectively, R is the gas constant, T is absolute temperature and n is the number of base pairs in the plasmid used (n = 2686 for pMS215).

Knowing that the amount of maximum relaxation of supercoil turns produced by triplex formation in pMS215 was [Delta]L3r = 3.9 ± 0.1 and [Delta]L5r = 3.0 ± 0.1 for the H-y3 and H-y5 isomers respectively (15 ), [Delta]Gsc° was calculated to be 4.74 kcal/mol from equation 5, using the fact that the supercoil density of the plasmid was [sigma] = -0.055 (corresponding to [Delta]Lk = -14). Substituting the observed value, -0.58 kcal/mol (Table 2 ), for free energy difference [Delta]G° in equation 4 gave an estimated free energy change, [Delta]Gst°, of ~4.16 kcal/mol at 35°C. This positive value of [Delta]Gst° indicates that the H-y5 form is thermodynamically more stable than H-y3 at low supercoil stress.

Thermodynamic data on oligonucleotide intramolecular triplexes are abundant. In contrast to the result obtained above, most of the data available indicate that the two triplex isomers have similar thermal stability (16 ,24 -26 ). There is some evidence, however, showing high stability of H-y5 relative to the H-y3 form of intramolecular triplex in supercoiled plasmids. First, the H-y5 form is dominant at lower supercoil densities (8 ,27 ) and it was formed even in the linearized plasmid at pH 4.0 (10 ). Second, as pointed out by Glover et al. (10 ) and Shimizu et al. (15 ), the hairpin loop region in the H-y5 form is much less accessible to solvent than that of H-y3, indicating a relatively ordered structure for the H-y5 loop. Such a relatively high reactivity of the H-y3 form can be seen in the autoradiograms shown in Figure 4 b: two guanines of the loop region are insensitive at t = 0, while they become sensitive at 12 h, when the H-y3 form is dominant. Less accessibility of the hairpin loop to solvent may arise from an extra Watson-Crick base pairing of the residue at the 3'-end of the folded back loop of the homopyrimidine strand: H-y5 would form such a Watson-Crick base pairing, while H-y3 would not (10 ). If this were the case, the residues near the 3'-end of the four-member loop in the center of the pyrimidine strand might play a key role in stabilizing the hairpin loop of the H-y5 form.

In conclusion, kinetic and thermodynamic studies revealed that a differential preference among intramolecular triplex isomers is initially determined by inequivalent kinetic barriers for the transition from the duplex to the triplex isomers and their relative populations change exponentially toward an increase in the H-y3 isomer to reach equilibrium. Such inequivalency of the kinetic barriers arises from differences in topology and size of a denaturation bubble in the transition states for triplex formation. The H-y3 form is thermodynamically less stable than H-y5 at lower supercoil level, which may arise from instability of the hairpin loop in H-y3 relative to that in H-y5. However, its disadvantage in triplex formation is overcome at higher negative supercoil level, because H-y3 formation results in relaxation of negative supercoils by one more turn than does H-y5. These will be general features for isomerization of intramolecular triplexes in supercoiled plasmids.

ACKNOWLEDGEMENTS

We thank Professor U.Matsumoto for his constant interest and support. This work was supported by a Grant-in Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan to H.S. and M.S.

REFERENCES

1 Sinden,R.R. (1994) DNA Structure and Function, Ch. 6. Academic Press, San Diego, CA.

2 Mirkin,S.M. and Frank-Kamenetskii,M.D. (1994) Annu. Rev. Biophys. Biomol. Struct., 23, 541-576. MEDLINE Abstract

3 Frank-Kamenetskii,M.D. and Mirkin,S.M. (1995) Annu. Rev. Biochem., 64, 65-95. MEDLINE Abstract

4 Wells,R.D., Collier,D.A., Hanvey,J.C., Shimizu,M. and Wohlrab,F. (1988) FASEB J., 2, 2939-2949.

5 Voloshin,O.N., Mirkin,S.M., Lyamichev,V.L., Belotserkovskii,B.P. and Frank-Kamenetskii,M.D. (1988) Nature, 333, 475-476. MEDLINE Abstract

6 Johnston,B.H. (1988) Science, 241, 1800-1804. MEDLINE Abstract

7 Htun,H. and Dahlberg,J.E. (1988) Science, 241, 1791-1796. MEDLINE Abstract

8 Htun,H. and Dahlberg,J.E. (1989) Science, 243, 1571-1576. MEDLINE Abstract

9 Hanvey,J.C., Shimizu,M. and Wells,R.D. (1988) Proc. Natl. Acad. Sci. USA, 85, 6292-6296. MEDLINE Abstract

10 Glover,J.N.M. and Pulleyblank,D.E. (1990) J. Mol. Biol., 215, 653-663.

11 Kohwi,Y. and Kohwi-Shigematsu,T. (1993) J. Mol. Biol., 231, 1090-1101. MEDLINE Abstract

12 Parniewski,P., Galazka,G., Wilk,A. and Klysik,J. (1989) Nucleic Acids Res., 17, 617-629. MEDLINE Abstract

13 Kang,S., Wohlrab,F. and Wells,R.D. (1992) J. Biol. Chem., 267, 1259-1264. MEDLINE Abstract

14 Kang,S. and Wells,R.D. (1992) J. Biol. Chem., 267, 20887-20891. MEDLINE Abstract

15 Shimizu,M., Kubo,K., Matsumoto,U. and Shindo,H. (1994) J. Mol. Biol., 235, 185-197. MEDLINE Abstract

16 Roberts,R.W. and Crothers,D.M. (1996) J. Mol. Biol., 260, 135-146. MEDLINE Abstract

17 Shimizu,M., Hanvey,J.C. and Wells,R.D. (1989) J. Biol. Chem., 264, 5944-5949. MEDLINE Abstract

18 Hanvey,J.C., Klysik,J. and Wells,R.D. (1988) J. Biol. Chem., 263, 7386-7396. MEDLINE Abstract

19 Kang,S., Wohlrab,F. and Wells,R.D. (1992) J. Biol. Chem., 267, 19435-19442. MEDLINE Abstract

20 Hanvey,J.C., Shimizu,M and Wells,R.D. (1989) J. Biol. Chem., 264, 5950-5956 MEDLINE Abstract

21 Depew,R.E. and Wang,J.C. (1975) Proc. Natl. Acad. Sci. USA, 72, 4280-4284.

22 Pulleyblank,D.E., Shure,M., Tang,D., Vinograd,J. and Vosberg,H.P. (1975) Proc. Natl. Acad. Sci. USA, 72, 4280-4284. MEDLINE Abstract

23 Frank-Kamenetskii,M.D. and Vologodskii,A.V. (1984) Nature, 307, 481-482. MEDLINE Abstract

24 Haner,R. and Dervan,P.B. (1990) Biochemistry, 29, 9761-9765. MEDLINE Abstract

25 Booher,M.A., Wang,S. and Kool,E.T. (1994) Biochemistry, 33, 4645-4651. MEDLINE Abstract

26 Wang,S., Booher,M.A. and Kool,E.T. (1994) Biochemistry, 33, 4639-4644. MEDLINE Abstract

27 Glover,J.N.M., Farah,C.S. and Pulleyblank,D.E. (1990) Biochemistry, 29, 11110-11115.

*To whom correspondence should be addressed. Tel: +81 426 76 4542; Fax: +81 426 76 4542; Email: shindo@ps.toyaku.ac.jp
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Print PDF (114K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Shindo, H.
Right arrow Articles by Shimizu, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shindo, H.
Right arrow Articles by Shimizu, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?