| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Mathematical Sciences, Rensselaer Polytechnic Institute, Troy, New York 12180
Correspondence: Address reprint requests to Michael Zuker, Dept. of Mathematical Sciences, Rensselaer Polytechnic Institute, Troy, NY 12180. Tel.: 518-276-6902; Fax: 518-276-4824; E-mail: zukerm{at}rpi.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTS AND DISCUSSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
In particular we focus on the role of stacking between neighboring nucleotide residues of single unfolded strands as an important source of enthalpy change on helix formation which has not been modeled computationally thus far. Changes in the states of the single strands with temperature are shown to lead to a larger heat effect at higher temperature. An important consequence of this is that predictions of enthalpies, which are based on databases of nearest-neighbor energy parameters determined for molecules or duplexes with lower melting temperatures compared with the melting temperatures of the oligos for which they are used as a predictive tool, will be underestimated.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
; Shoemaker et al., 2001), PCR, sequencing by hybridization, and gene diagnostics, including SNP detection. These technologies require accurate prediction of hybridization thermodynamics to matched versus mismatched sites. The statistical thermodynamic theory of DNA/RNA hybridization has been well understood for a number of years (Zimm, 1960; Poland, 1974). However, despite the fact that a half century has passed since the discovery of the structure of the DNA double helix, major questions still remain regarding its thermodynamic behavior and stability. Thus, the parameters that determine the cooperativity of melting have been difficult to measure and also, it has been difficult to test some basic theoretical assumptions such as the justification of the Jacobson-Stockmayer loop-weighting function, the nearest-neighbor model for the interaction between nucleotides, and the role of heat capacity changes, 
volume changes, 
V, and compressibility changes, 
that accompany nucleic acid conformational transitions.
We have observed that in current theoretical treatments of DNA or RNA melting, little attention has been paid to the effects of concentrations of different conformational species in the solution to the overall equilibrium (Applequist and Damle, 1963). It is well known that melting of oligomeric DNA or RNA molecules is typically complicated by coupled equilibria between the different conformational species. Thus, the melting transitions of mono-molecular hairpins would be expected to be concentration independent, whereas double- or multiple-stranded complexes should melt at higher temperatures when strand concentration is increased. Over the past two decades, the structure and the conformational flexibility of several specially designed and synthesized oligonucleotides have been characterized (Breslauer et al., 1975; Albergo et al., 1981; Gralla and Crothers, 1973; Early et al., 1981). The goal of such studies has been to develop an understanding of the molecular forces that control the various sequence and solvent specific conformational forms found within DNA and RNA polymers.
Melting experiments have been the most useful way to measure the stabilities of RNA and DNA structures under different conditions. Thermodynamic parameters are easy to extract from UV absorbance versus temperature curves of simple RNA or DNA secondary structures, duplexes or hairpins, that melt in a single, two-state transition. Comparisons of RNAs and DNAs with different basepairs, loop sequences, bulges, etc. have yielded an extremely useful database from which the stabilities of larger structures can be estimated (Freier et al., 1983; Sugimoto et al., 1987; Hickey and Turner, 1985; Puglisi and Tinoco, 1989; Blake, 1972). The idea behind the experiments on such short model oligomers is that small but specific structural changes, such as loops, bulges, dangling ends, etc., can be detected by changes in the chemical potential.
The estimation of these parameters is based on nearest-neighbor approximations for inter-residue interactions (Borer et al., 1974). The major assumption is that the stability of a basepair is dependent only on the identity of adjacent basepairs because the major interactions involved in transformation between different conformations of the polynucleotide sequence are stacking and hydrogen bonding. Both are short-range interactions. SantaLucia has published a detailed article comparing the nearest-neighbor parameters from seven different laboratories based on data from natural polymers, synthetic polymers, oligonucleotide dumbbells, and oligonucleotide duplexes (SantaLucia, 1998). The analysis shows that the data are in very good agreement.
There have been several major improvements in the calculation of the partition function for a single-stranded species based on the McCaskill algorithm (McCaskill, 1990; Hofacker et al., 1994; Matzura and Wennborg, 1996) or estimation of the free energy based on free energy minimization and the corresponding sub-ensemble around the minimum free energy conformation (Sankoff et al., 1983; Zuker and Sankoff, 1984; Zuker and Stiegler, 1981; Zuker, 1989a; Williams and Tinoco, 1986; Waterman, 1983; Waterman and Byers, 1985; Zuker, 1989b). For more details about current energy rules and free energy minimization, see the article by Zuker et al. (1999). In this work, our primary applications are to relatively small oligonucleotide sequences, which makes appropriate the use of certain simplifications. For dimer formations, we have ignored the possibility of intramolecular basepairs. For single-stranded species, we have used the minimum free energy computed using the "nafold" program in the "mfold" package by Zuker et al. (1999). A new and more robust software package is being prepared to replace the initial development programs used for calculations in this work. This new software uses our own version of partition function calculations for single-stranded species (McCaskill, 1990; Hofacker et al., 1994).
Our primary aim in this article is to develop the statistical mechanical formalism appropriate to hybridization processes between finite length DNA and RNA sequences that takes into account the whole ensemble of single- and double-strand species in solution and the exchange of material between them. In particular, we focus also on the role of stacking between neighboring nucleotide residues of single strands as an important source of enthalpy change on helix formation (Poerschke et al., 1973). Thus, each oligomer has a stacked-unstacked transition that is superimposed upon the helix-coil transition. Therefore we should expect that the measured enthalpy for the helix-single strand transition will be less at low temperatures where the nearest-neighbor nucleotide residues in the single strands are partially stacked than at very high temperatures where the nearest-neighbor nucleotide residues are totally unstacked. Currently, the thermodynamic analysis of hybridization processes is based on Poland's method and its modifications (Poland, 1974; Fixman and Freire, 1977; Poland, 1981). This method can deal only with nearest-neighbor stacking between two strands of RNA or DNA. It can also handle symmetric internal loops caused by strand separation. However, Poland's method does not take into account the terminal mismatch stacking and no intrastrand basepairs are allowed. On the other hand, the folding model that we will develop here deals with matches, mismatches, symmetric and asymmetric interior loops, bulges, and single-base stacking that might exist at the ends. In addition, our model takes into account the whole ensemble of single- and double-stranded species in solution and the exchange of material between them. This leads to a statistical thermodynamic description of both self-folding and hybridization of RNA or DNA sequences. The only drawback is that whereas the time complexity of the Poland algorithm is 
and its improvement based on the Fixman and Freire approximation is O(n), ours is 
However, when a limitation for loop length is applied, we go from 
to 
We also have developed a set of computer programs that aid in the analysis of RNA or DNA melting with multiple unfolding transitions. The programs can simulate both UV hyperchromicity and scanning calorimetry melting curves by summing computed pairing probabilities multiplied by measured extinction coefficients and by numerical differentiation of the computed Gibbs free energy, respectively.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
and 
molecules of polynucleotide sequences A and B are added to a physiological solution at a given volume V and temperature T. It is assumed that the solution is sufficiently dilute so that the average distance between the molecules, 
is greater than the intermolecular distance necessary for each molecule to explore all possible conformations without overlapping with other molecules and bigger than the range of forces acting between the molecules. The molecules and are allowed to form chemical species in terms of hybridizations between all possible pairs; two homo-dimers, AA and BB, and the single hetero-dimer, AB. Based on the above assumptions, the solution can be described as an ensemble of ideally mixed species such as single-folded strands 
and 
single-unfolded strands 
and 
as well as the double-stranded hybridized forms of all possible pair combinations between the polynucleotide sequences 
and 
The species are characterized by their corresponding ensembles of possible conformational states. Letting 
and 
the partition function for such a system at a given temperature T, volume V, and all possible distributions of the initial material and between the corresponding species 
is computed as (Kubo, 1965):
 | (1) |
and 
are the partition functions for the corresponding species. The partition functions of the single-stranded species have the forms:
 | (2) |
 | (3) |
and 
and 
and 
are the partition functions of the corresponding self-folded species AF and BF. The "1" in Eqs. 2 and 3 pertains to the fact that by definition, a folded state must contain at least one basepair, whereas 
and 
include the unfolded states. The free energies 
and 
of the unfolded species AU and BU have been set to 0.
Taking into account that the free energy of a closed system at constant temperature, volume and pressure tends toward a minimum (Landau and Lifshitz, 1969) the equilibrium distributions of 
and 
are determined by the minimization of the free energy under the constraints that
 | (4) |
 | (5) |
 | (6) |
 | (7) |
 | (8) |
 | (9) |
and 
denote the corresponding chemical equilibrium constants.
These relations lead to the following system of nonlinear equations for the species concentrations:
 | (10) |
Given the 
and the first two equations can be solved with respect to 
and 
using, for example, Newton's method for solving two nonlinear functions. Replacing 
and 
in the rest of the equations, 
and are determined straightforwardly. Finally, taking into account that and 
we have:
 | (11) |
The chemical potentials of the species can be obtained by differentiating the free energy 
with respect to the concentrations of their corresponding molecules. Thus we have:
 | (12) |
 | (13) |
 | (14) |
 | (15) |
 | (16) |
Finally, the free energy of the whole ensemble of species can be represented in the form (Landau and Lifshitz, 1969):
 | (17) |
Extinction coefficients and melting curves
The transition between folded and unfolded structures, as well as the partial forms of their conformational intermediates, can be monitored as a function of the temperature by any physical property that is dependent on the number and type of basepairs formed. Fortunately, the absorption spectra as well as thermodynamics are physical properties that are consistent with the nearest-neighbor models (Puglisi and Tinoco, 1989; Blake, 1972; Petersheim and Turner, 1983). In other words, given nearest neighbors must have identical values of their absorptions or melting free energies regardless of their position in the interior or at the ends of the sequence. Thus, the property being monitored as a function of the temperature is proportional to the fraction of basepairs that are stacked as a nucleic acid molecule or duplex is melted. In this article we do not restrict ourselves to the case of two-state transitions where there are only two types of conformational species as the temperature changes: fully folded and fully unfolded. Rather, we consider the ensemble of all possible intermediate states, thus yielding the most detailed possible picture of the melting process between the folded and unfolded states of the single and double-stranded forms. The task we are going to solve can be formulated as follows in the next paragraph.
As a result of interconversions between the single and double-stranded forms at each temperature, there is an equilibrium between the different conformational species; single-stranded A, single-stranded B, double-stranded AA, double-stranded BB, and double-stranded AB. Each of these forms is characterized with an ensemble of conformational states where each conformation is characterized by the fraction of its basepairs and their location along the sequences that are melted at any given temperature. Thus, along the sequence(s) we have alternating loops, single-stranded regions, and double-stranded regions. The locations and the lengths of these portions depend on their relative Boltzmann statistical weights. We set the double-stranded forms as our zero level from which the contribution of the melted single-stranded forms should be counted. This assumption is rather convenient; experiments have shown that the contribution from the double-stranded forms is 
75% from that of the melted single forms (Bloomfield et al., 2000). With this approximation, the absorption of the ensemble of all possible conformational species, taking into account their interconversions and their own ensemble of conformational changes, can be represented in the following form:
 | (18) |
(T) is the extinction of the ensemble of all possible species as a function of temperature, T; 
for i running among all the different species represent the species extinctions as a function of temperature; and [AU](T), [AF](T), [BU](T), [BF](T), [AA](T), [BB](T), and [AB](T) represent the mole fractions of the corresponding species as a function of the temperature which are calculated as described above. We will focus now on the extinctions of the species 
First we should take into account that the extinction is determined by the contribution of the melted or mismatch loop regions along the constituent sequences of the double-stranded species (Bloomfield et al., 2000). At each given temperature there is an ensemble of conformations with a narrow or broad distribution of such loops. The contribution of each of them is proportional to its relative Boltzmann statistical weight. It follows from here that the extinction for the AB species, for example, can be represented in the form 
where the contributions from sequences A and B are as follows
 | (19) |
 | (20) |
Here 
and 
stand for the lengths of the nucleic acid A and B, and 
is for extinction coefficients of single bases (one argument) or (by a slight abuse of notation) for dinucleotides (two arguments). 
and 
are the probabilities that an arbitrary single {i}, {i + 1} or double {i, i + 1} nearest-neighbor positions along the sequence 1 forms basepairs with the sequence 2 and vice versa. Using these probabilities we can express the probabilities that two closest along the sequence A or B nucleotides with positions i and i + 1 are melted (giving a contribution (i, i + 1) to the total absorbance) as 
and 
respectively.
Heat capacity and melting
With increasing temperature, the overwhelming majority of the single and double-stranded species conformations tend toward their corresponding unfolded states, as reflected by the large enthalpy and entropy gain associated with base pairing disruption and loop formation. The relative changes of the species concentration, which reflect the structural changes of the melting process, have a nonlinear character as shown by the mass equations derived above. This makes the analysis of the melting process rather complicated. Differential scanning calorimetry (DSC) is a widely used tool for investigating the conformational changes in the melting process (Breslauer et al., 1992; Sturtevant, 1987). It measures the difference in heat or the heat capacity required to raise the temperature of the solution. The advantage of this method is that thermodynamic parameters such as enthalpy, H, entropy, S and free energy, F = E – TS + PV, which can be obtained in this way, do not depend on a theoretical model for the underlying conformational changes occurring in the melting material. Usually in the melting experiments the change in V is negligible, so the term PV is constant and the free energy is estimated from F = E – TS, and the enthalpy H is equivalent to the internal energy E. From statistical thermodynamics it is well known that the heat capacity, 
is derived from the second derivative of the free energy, F, with respect to the temperature T.
 | (21) |
 | (22) |
 | (23) |
The summation over i includes all possible different species in the solution. One still has to make the above calculations for varying temperatures over the desired range. To compute H and 
for a particular temperature 
an approach based on the derivation made by the "Vienna group" is used (Hofacker et al., 1994). A least-squares parabola is fitted to F at 2m + 1 points: 
The second derivative of this polynomial is the estimate used for the second derivative of F with respect to a temperature 
Recursive calculation
Partition function
In any statistical thermodynamic model, all the thermodynamic information is contained in the partition function under consideration
 | (24) |
is the free energy of the 
state of the system. In our model, we wish to calculate the species partition functions, 
and that determine the thermodynamic properties of the single-stranded species 
and the double-stranded species and of the polynucleotide molecules A and B.
The polynucleotide sequences of the double-stranded forms are described as follows. The sequence for A is represented by 
and sequence B is represented by 
where 
and 
stand for their corresponding lengths and 
and 
are the sequence coordinates of the corresponding nucleotides of sequences 1 and 2. In this article we will use some simplification concerning possible conformational states of the hybrid form AB. Thus, hybridization will account only for stacked pairs, interior loops, bulges, and, at the ends, dangling bases. At this stage, we do not consider intramolecular basepairs. Stacking between the loop regions of sequence 1 and sequence 2 are also not considered (Fig. 1).
|
, Doktycz et al. (1990), Blommers et al. (1989), LeBlanc and Morden (1991), Zieba et al. (1991), and Zuker et al. (1999), the energy rules allow us to ascribe to each stacked pair an energy dependent only on pairs under consideration and their nearest neighbors. The recursion calculation is based on the condition that there are at least two nucleotides along sequences 
and 
that are in contact. A contact, or basepair, is denoted by 
for 
Sequence enumeration is always from 5' to 3'. The contact includes an initiation free-energy term necessary to bring the two sequences together 
Each nucleotide pair, 
formally divides the hybridized form 
of sequences 1 and 2 into two parts; left, L and right, R, in such way that the free energy, 
of is a sum of the free energies of the left 
and right 
parts plus the initiation free energy 
which is assumed to be the same for all possible pairs 
Thus,
 | (25) |
This additive property of the energy rules based on nearest neighbor approximation forms the basis of the recursion calculations of the partition function 
The additivity of the free energy leads to a multiplication of the partition functions of the left, ZL and right, ZR, parts. Therefore, the recursions for the partition functions of the left and right parts are:
 | (26) |
 | (27) |
 | (28) |
 | (29) |
Here 
and 
correspond to cases where only the pair is formed in the left and right parts, respectively; whereas 
and 
correspond to the dangling free energies of the tails of sequence 1 and 2. The dangling free energies take into account the ensemble of all possible stackings between the nucleotides adjacent to the pair. When 
and 
the free energy 
represents a stacked pair that belongs to a secondary structure, when 
and or and 
we have a bulge. In general when 
the free energy, 
represents an asymmetric internal loop, whereas 
leads to a symmetric loop. For a detailed description of the free energies of bulges, symmetric and asymmetric internal loops, and dangling ends, we refer the reader to the articles by Zuker et al. (1999) and (Mathews et al., 1999). Based on the multiplication property of the partition functions for the left and the right parts of the hybridization form for the total partition function we have:
 | (30) |
 | (31) |
Pair probabilities
The calculated partition functions will allow us to derive the probabilities of various conformations. Our main interest here is to calculate the probabilities 
and 
of a single 
and double 
pair formation, where 
These probabilities play a major role in hybridization or melting processes.
 | (32) |
 | (33) |
It is now easy to calculate the probabilities and that an arbitrary single {i}, {i + 1} or double {i, i + 1} nearest-neighbor positions along the sequence 1 forms basepairs with the sequence 2 and vice versa. We have:
 | (34) |
 | (35) |
 | (36) |
 | (37) |
In particular the equilibrium fraction of bases paired 
can be calculated from
 | (38) |
|
RESULTS AND DISCUSSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
entropy 
and the melting temperature, 
for a number of oligonucleotides taken from the literature (Allawi and SantaLucia, 1997). There are two main groups of sequences in Table 1; those in which the transition equilibrium involves only two states, i.e., duplex (as complete as possible if mismatches occur) and random coils, and those in which the transition equilibrium involves more than two states. Sequences from the two-state group are designed to have a melting temperature, between 30°C and 60°C and to minimize the possibility of forming stable alternative secondary structures such as slipped duplexes or hairpins. Non-two-state sequences are designed to form both duplex and hairpin species. In addition, sequences are designed with uniform distribution of the 11 different G·T mismatch containing nearest neighbors. The data set for internal G·T mismatches does not contain sequences that have terminal G·T mismatches. As a result, there are only 10 uniquely determined parameters as linear combinations of 11 G·T nearest-neighbor dimers (Allawi and SantaLucia, 1997). This special design is a very good test for the nearest-neighbor model according to which the terminal nearest neighbors make the same contribution as internal neighbors.
|
). The usual assumption is that because of the enthalpy-entropy compensation, 
and per basepair or stack for double-stranded helixes and single-stranded helixes, respectively, can be considered as temperature independent for most practically important temperature ranges. However, when and represent the difference between the double helix and the single-stranded sequences, adjustments should be made for stacking present in the single-stranded sequences before and after the double helix is formed. Without taking into account the temperature dependence (and probable dependence of some other factors) of the reference state, the deviation between the calculated and experimentally determined values for and can be quite large. The reference state refers to unfolded single strand sequences in which stacking between nearest neighbor nucleotides are present.
In Table 1, the experimental data show good agreement both with the calculations by our group and that of SantaLucia. It is important to note that the calculations of the SantaLucia group are based on the approximation that folded species do not explore all possible conformations in their folded state but are represented by the conformations with minimum energy. In our calculations we do explore all possible conformations of the folded species within the limitations specified above. In both the SantaLucia and our groups, the entropy loss by the single-stranded unfolded sequences in forming the first basepair of the double helix is taken into account by an experimentally determined initiation parameter. Since the conformation of denatured single strands is unknown, this is a reasonable approximation. The enthalpy and entropy contribution per basepair relative to the unfolded sequences are taken from experiments at 37°C. As a result, the entropy values of our calculations are consistently greater than those of the SantaLucia group. The differences are relatively small which should be expected because of the way the sequences are designed. Nevertheless, our calculations show that there is a variability around the minimum energy conformations of the folded species, and of course this variability will increase with the sequence length. Of particular interest are the last six sequences in Table 1 which melt with non-two-state thermodynamics regardless of the way they are designed. For two of them NTS-1, (5'-CGTTGCGTAACG-3')2 and NTS-2, 5'-GCGTACGCATGCG-3'/3'-CGCATGTGTACGC-5' (Plum et al., 1995), the non-two-state melting is in contrast with the good agreement between the experimentally determined van't Hoff enthalpies derived from 
versus 
plots and from the fits of individual melting curves. Melting simulation of NTS-1 from SantaLucia and our group and NTS-2 from our group are represented in Figs. 2–4, respectively.
|
|
|
Similarly (Fig. 3), our simulations show the presence of the duplex from sequence 2, CGCATGTGTACGC, near the melting temperature for the NTS-2 duplex, GCGTACGCATGCG/CGCATGTGTACGC.
Fig. 4 shows the calculated heat capacity for the NTS-2 duplex. It is consistent with the agreement between the UV melting data and calorimetric data from the experimental results of Plum et al. (1995) and SantaLucia. Our calculations show that heat capacity is almost perfectly symmetric as one would expect from a two-state melting process. Thus, it is well known that the van't Hoff enthalpy change, 
calculated from the temperature dependence of the equilibrium constant from spectroscopic data equals the total calorimetric enthalpy change, 
only if the melting reaction follows a two-state transition between free and bound molecules and if the change in the spectroscopic signal used to calculate the equilibrium constant reflects the entire population of free and bound molecules. Otherwise, 
and 
are different. Therefore, when and are equal, this can be taken to indicate a two-state transition. However, our mole fraction calculations show that the apparent two-state shape for the heat capacity follows from the fact that the maximum of the CGCATGTGTACGC duplex fraction is centered at the melting temperature of the NTS-2 duplex. As a result, the melting of CGCATGTGTACGC duplex is superimposed on that of the NTS-2 duplex which leads to almost symmetric shape of the heat capacity curve. Our simulations show clearly the advantage of the general statistical mechanical approach which explores all possible conformations of single hairpins and double helixes as well as the conformational transformations among them. In such way, we avoid confusions from the assumption of two-state melting based on the agreement between the van't Hoff and the total enthalpy change during the melting process.
Next we analyze the role of stacking between neighboring nucleotide residues of single strands as an important source of enthalpy change on helix formation. Thus, each oligomer has a stacked-unstacked transition that is superimposed upon the helix-coil transition. Therefore we should expect that the measured enthalpy for the helix-single strands transition will be less at low temperatures where the nearest-neighbor nucleotide residues in the single strands are partially stacked than at very high temperature where the nearest-neighbor nucleotide residues are totally unstacked. On the other hand, in both our group and the SantaLucia group, the difference in heat capacities 
between the unfolded and folded states of the species is taken to be zero. The calculated enthalpy contributions and the melting temperatures (Table 1) between the SantaLucia group and our group, as should be expected, based on the same 
approximation), are in very good agreement. However, the existence of heat capacity differences between the unfolded and folded states has been demonstrated experimentally in a few groups (Holbrook et al., 1999; Freier et al., 1981; Chalikian et al., 1999). Thus, in the work of Holbrook et al. (1999), it has been found that the contributions to the coming from changes in nonpolar and polar surfaces in single helix formation and the docking of the single strand helixes in hybridization and self-folding processes largely offset each other. As a result, the observed heat capacity changes in double-strand helix formation must arise primarily from temperature-dependent coupled processes in the unfolded strands. From the analysis of DSC and UV thermal-scan data, the values of enthalpies of ordering and folding of the single strands together with their relative fractional extents at a given temperature allowed the authors in combination with the observed DSC and isothermal titration calorimetry (ITC) enthalpies to extract the for double helix formation. Changes in the states of the single strands with temperature are shown to lead to a larger heat effect at higher temperature. Our calculations together with the experimental ITC data on the enthalpies of association of two 14 bp complementary sequences 5'-GCGTCATACAGTGC-3' and 5'-GCACTGTATGACGC-3' taken from the work of Holbrook et al. (1999) are shown in Fig. 5.
|
The results indicate that 
decreases strongly with increasing temperature. In the past few years, a series of experimental and theoretical papers were published on the differences between the DNA and RNA polymer and oligonucleotide nearest-neighbor thermodynamics. These differences led to the notion that there is a length dependency in DNA thermodynamics. An important paper of SantaLucia (1998) showed that most probably, this length dependence is only for the salt effect but not for the nearest-neighbor propagation energies. Thus, increasing sequence lengths lead to increasing melting temperatures as a result of the increase of the total number of basepair contacts whereas the nearest neighbor propagation energies, to a good approximation, do not change. Moreover, the melting temperature can be increased by increasing the number of energy rich basepair contacts without increasing the sequence length. Thus, the difference between the predicted and experimentally determined enthalpies is a direct consequence of the change in the melting temperature and as a result the changes in the heat effect from the stacking in the unfolded single strand sequences rather than the sequence length itself. The above analysis indicates that predictions of enthalpies based on databases of nearest-neighbor energy parameters determined for sequences with lower melting temperatures compared with the melting temperature of the sequences for which they are used as a predictive tool will be underestimated. In the work of SantaLucia, all sequences are designed to have a melting temperature between 30°C and 60°C, whereas the database of nearest-neighbor parameters which they used is determined at 37°C. Our calculations demonstrate that the predicted enthalpies for the sequences in the work of SantaLucia are in good agreement with their experimental data because the sequences are designed to melt near the database melting temperature. In conclusion, we present here a general statistical mechanical approach appropriate to describe the hybridization processes between finite length DNA and RNA sequences that takes into account the whole ensemble of single and double strand species in the solution and their fractional extents at different temperatures. The folding models for both duplexes and self-folding of single strands developed here deals with matches, mismatches, symmetric and asymmetric interior loops, bulges and single base stacking that might exist at the ends and explores all possible conformations of the single and double strand species. The advantage of such a general approach is most clearly demonstrated in the cases where the melting of the different species are superimposed onto each other, leading to an agreement between the van't Hoff and the total enthalpy change during the melting process. As a result, it is not clear whether such a melting process is two-state or a multi-state that involves intermediates. In particular, we focused also on the role of stacking between neighboring nucleotide residues of single unfolded strands as an important source of enthalpy change on helix formation which has not been distinguished thus far. Changes in the states of the single strands with temperature are shown to lead to a larger heat effect at higher temperature. An important consequence of this is that predictions of enthalpies based on databases of nearest-neighbor energy parameters determined for sequences with lower melting temperatures compared with the melting temperature of the sequences for which they are used as a predictive tool, will be underestimated. Lastly, this article demonstrates the need for an accurate statistical mechanical description of the single-stranded unfolded sequences that can still preserve some nearest-neighbor stacking contacts. The development of such a method is in progress in our laboratory. Further information about the programs used in this article can be found at http://www.bioinfo.rpi.edu/applications/hybrid/.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This work has been supported by a grant to M. Zuker, number GM54250, from the National Institutes of Health.
Submitted on December 4, 2002; accepted for publication June 5, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTS AND DISCUSSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Allawi, H. T., and J. SantaLucia, Jr. 1997. Thermodynamics and NMR of internal G·T mismatches in DNA. Biochemistry. 97:10581–10594.
Applequist, J., and V. Damle. 1963. Theory of effects of concentration and chain length on helix-coil equilibria in two-stranded nucleic acids. J. Chem. Phys. 39:2719–2721.[CrossRef][Medline]
Blake, R. D. 1972. Thermodynamics of Oligo(A)N·2Poly(U)
from the dependence of the temperature of the helix-coil transition on oligomer concentrations. Biopolymers. 11:913–933.[CrossRef][Medline]
Blommers, M. J., J. A. Walters, C. A. Haasnoot, J. M. Aelen, G. A. van der Marel, J. H. van Boom, and C. W. Hilbers. 1989. Effects of base sequence on the loop folding in DNA hairpins. Biochemistry. 28:7491–7498.[CrossRef][Medline]
Bloomfield, V. A., D. M. Crothers, and I. Tinoco, Jr. 2000. Nucleic Acids. Structure, Properties and Functions. University Science Books, Sausalito, CA.
Borer, P. N., B. Dengler, I. Tinoco, Jr., and O. C. Uhlenbeck. 1974. Stability of ribonucleic acid double-stranded helices. J. Mol. Biol. 86:843–853.[CrossRef][Medline]
Breslauer, K. J., E. Freire, and M. Straume. 1992. Calorimetry: a tool for DNA and ligand-DNA studies. Methods Enzymol. 211:533–567.[Medline]
Breslauer, K. J., J. M. Sturtevant, and I. Tinoco, Jr. 1975. Calorimetric and spectroscopic investigation of the helix-to-coil transition of a ribo-oligonucleotide: rA7U7. J. Mol. Biol. 99:549–565.[CrossRef][Medline]
Chalikian, T. V., J. Volker, G. E. Plum, and K. J. Breslauer. 1999. A more unified picture for the thermodynamics of nucleic acid duplex melting: a characterization by calorimetric and volumetric techniques. Proc. Natl. Acad. Sci. USA. 96:7853–7858.
Doktycz, M. J., T. M. Paner, M. Amaratunga, and A. S. Benight. 1990. Thermodynamic stability of the 5' dangling-ended DNA hairpins formed from sequences 5'-(XY)2GGATAC(T)4GTATCC-3', where X,Y=A,T,G,C. Biopolymers. 30:829–845.[CrossRef][Medline]
Early, T. A., D. R. Kearns, W. Hillen, and R. D. Wells. 1981. A 300-MHz proton nuclear magnetic resonance investigation of deoxyribonucleic acid restriction fragments: dynamic properties. Biochemistry. 20:3764–3769.[CrossRef][Medline]
Fixman, M., and J. J. Freire. 1977. Theory of DNA melting curves. Biopolymers. 16:2693–2704.[CrossRef][Medline]
Freier, S. M., D. Alkema, A. Sinclair, T. Neilson, and D. H. Turner. 1983. Contribution of dangling end stacking and terminal base-pair formation to the stabilities of XGGCCp, XCCGGp, XGGCCYp, and XCCGGYp helixes. Biochemistry. 22:6198–6206.[CrossRef]
Freier, S. M., K. O. Hill, T. G. Dewey, L. A. Marky, K. J. Breslauer, and D. H. Turner. 1981. Solvent effects on the kinetics and thermodynamics of stacking in poly (cytidylic acid). Biochemistry. 20:1419–1426.[CrossRef][Medline]
Gralla, J., and D. M. Crothers. 1973. Free energy of imperfect nucleic acid helices. II. Small hairpin loops. J. Mol. Biol. 73:497–511.[CrossRef][Medline]
Hickey, D. R., and D. H. Turner. 1985. Solvent effects on the stability of A7U7p. Biochemistry. 24:2086–2094.[CrossRef][Medline]
Hofacker, I. L., W. Fontana, P. F. Stadler, S. Bonhöffer, M. Tacker, and P. Schuster. 1994. Fast folding and comparison of RNA secondary structures. Monatsh. Chem. 125:167–188.[CrossRef]
Holbrook, J. A., M. W. Capp, R. M. Saecker, and M. T. Record. 1999. Enthalpy and heat capacity changes for formation of an oligomeric DNA duplex: interpretation in terms of coupled processes of formation and association of single-stranded helices. Biochemistry. 38:8409–8422.[CrossRef][Medline]
Kubo, R. 1965. Statistical Mechanics. An advanced course with problems and solutions. North-Holland Publishing Company. Amsterdam, the Netherlands.
Landau, L. D., and E. M. Lifshitz. 1969. Statistical Physics. Pergamon, Oxford, UK.
LeBlanc, D. A., and K. M. Morden. 1991. Thermodynamic characterization of deoxyribooligonucleotide duplexes containing bulges. Biochemistry. 30:4042–4047.[CrossRef][Medline]
Mathews, D. H., J. Sabina, M. Zuker, and D. H. Turner. 1999. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure. J. Mol. Biol. 87:911–940.
Matzura, O., and A. Wennborg. 1996. RNAdraw: an integrated program for RNA secondary structure calculation and analysis under 32-bit Microsoft Windows. Comput. Appl. Biosci. 12:247–249.
McCaskill, J. S. 1990. The equilibrium partition function and base pair binding probabilities for RNA secondary structure. Biopolymers. 29:1105–1119.[CrossRef][Medline]
Owczarzy, R., P. M. Vallone, F. J. Gallo, T. M. Paner, M. J. Lane, and A. S. Benight. 1997. Predicting sequence-dependent melting stability of short duplex DNA oligomers. Biopolymers. 44:217–239.[CrossRef][Medline]
Petersheim, M., and D. H. Turner. 1983. Base-stacking and base-pairing contribution to helix stability: Thermodynamic of double-helix formation with CCGG, CCGGp, CCGGAp, CCGGUp, and ACCGGUp. Biochemistry. 22:256–263.[CrossRef][Medline]
Plum, G. E., A. P. Grollman, F. Johnson, and K. J. Breslauer. 1995. Influence of the oxidatively damaged adduct 8-oxodeoxyguanosine on the conformation, energetics, and thermodynamic stability of a DNA duplex. Biochemistry. 34:16148–16160.[CrossRef][Medline]
Poerschke, D., O. C. Uhlenbeck, and F. H. Martin. 1973. Thermodynamics and kinetics of the helix-coil transition of oligomers containing GC basepairs. Biopolymers. 12:1313–1335.[CrossRef]
Poland, D. 1974. Recursion relation generation of probability profiles for specific-sequence macromolecules with long-range correlations. Biopolymers. 13:1859–1871.[CrossRef][Medline]
Poland, D. 1981. Cooperative Equilibria in Physical Biochemistry. Clarendon Press. Oxford, UK.
Puglisi, J. D., and I. Tinoco, Jr. 1989. Absorbance melting curves of RNA. Methods Enzymol. 180:304–325.[Medline]
Sankoff, D., J. B. Kruskal, S. Mainville, and R. J. Cedergren. 1983. Fast algorithms to determine RNA secondary structures containing multiple loops. In Time Warps, String Edits, and Macromolecules: The Theory and Practice of Sequence Comparison. D. Sankoff and J. B. Kruskal, editors. Addison-Wesley, Reading, MA. 93–120.
SantaLucia, J., Jr. 1998. A unified view of polymer, dumbbell, and oligonucleotide DNA nearest-neighbor thermodynamics. Proc. Natl. Acad. Sci. USA. 95:1460–1465.
Seetharaman, S., M. Zivarats, N. Sudarsan, and R. R. Braker. 2001. Immobilized RNA switches for the analysis of complex chemical and biological mixture. Nat. Biotechnol. 19:336–341.[CrossRef][Medline]
Shoemaker, D. D., E. E. Schadt, C. D. Armour, Y. D. He, P. Garrett-Engele, P. D. McDonagh, and P. M. Loer. 2001. Experimental annotation of the human genome using microarray technology. Nature. 409:922–927.[CrossRef][Medline]
Sturtevant, J. M. 1987. Biochemical applications of differential scanning calorimetry. Annu. Rev. Phys. Chem. 38:463–488.[CrossRef]
Sugimoto, N., R. Kierzek, and D. H. Turner. 1987. Sequence dependence for the energetics of dangling ends terminal base pairs in ribonucleic acid. Biochemistry. 26:4554–4558.[CrossRef][Medline]
Waterman, M. S. 1983. Sequence alignment in the neighborhood of the optimum with general application to dynamic programming. Proc. Natl. Acad. Sci. USA. 80:3123–3124.
Waterman, M. S., and T. H. Byers. 1985. A dynamic programming algorithm to find all solutions in a neighborhood of the optimum. Math. Biosci. 77:179–188.[CrossRef]
Williams, A. L., and I. Tinoco, Jr. 1986. A dynamic programing algorithm for finding alternate RNA secondary structures. Nucleic Acids Res. 14:299–315.
Zieba, K., T. M. Chu, D. W. Kupke, and L. A. Marky. 1991. Differential hydration of dA-dT base pairing and dA and dT bulges in deoxyoligonucleotides. Biochemistry. 30:8018–8026.[CrossRef][Medline]
Zimm, B. H. 1960. "Theory of melting" of the helical form in double chains of the DNA type. J. Chem. Phys. 33:1349–1356.[CrossRef]
Zuker, M. 1989a. Computer prediction of RNA structure. Methods Enzymol. 180:262–288.[Medline]
Zuker, M. 1989b. On finding all suboptimal foldings of an RNA molecule. Science. 244:48–52.
Zuker, M., and D. Sankoff. 1984. RNA secondary structures and their prediction. Bull. Math. Biol. 46:591–621.
Zuker, M., and P. Stiegler. 1981. Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucleic Acids Res. 9:133–148.
Zuker, M., D. H. Mathews, and D. H. Turner. 1999. Algorithms and thermodynamics for RNA secondary structure prediction: a practical guide. In RNA Biochemistry and Biotechnology. J. Barciszewski, and B. F. C. Clark, editors. NATO ASI Series. Kluwer Academic Publishers. Dordrecht, the Netherlands. 11–43.
This article has been cited by other articles:
 |
|
 |
|
  G. P. Brewood, Y. Rangineni, D. J. Fish, A. S. Bhandiwad, D. R. Evans, R. Solanki, and A. S. Benight Electrical detection of the temperature induced melting transition of a DNA hairpin covalently attached to gold interdigitated microelectrodes Nucleic Acids Res., September 1, 2008; 36(15): e98 - e98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-T. Li, Y. Zhang, L. Kong, Q.-R. Liu, and L. Wei Trans-natural antisense transcripts including noncoding RNAs in 10 species: implications for expression regulation Nucleic Acids Res., September 1, 2008; 36(15): 4833 - 4844. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 |
) Random coil (SantaLucia); (+) duplex(SantaLucia); (
) hairpin (SantaLucia); (x) random coil (Dimitrov and Zuker); 
duplex (Dimitrov and Zuker); (*) hairpin (Dimitrov and Zuker).
1000 mM; 
0 mM. (
) 22 (duplex).
