Article Information

• PDF (124 kb)
• Full Text with Large Figures
• Cited by in Scopus (3)

PubMed

• Articles by Boualem Hammouda
• Articles by David Worcester

Related Articles

• Thermal Stability of Collagen Fibers in Ethylene Glycol

  Thermal Stability of Collagen Fibers in Ethylene Glycol Biophysical Journal, Volume 80, Issue 3, 1 March 2001, Pages 1480-1486 C.A. Miles and T.V. Burjanadze Abstract The mechanism that renders collagen molecules more stable when precipitated as fibers than the same molecules in solution is controversial. According to the polymer-melting mechanism the presence of a solvent depresses the melting point of the polymer due to a thermodynamic mechanism resembling the depression of the freezing point of a solvent due to the presence of a solute. On the other hand, according to the polymer-in-a-box mechanism, the change in configurational entropy of the collagen molecule on denaturation is reduced by its confinement by surrounding molecules in the fiber. Both mechanisms predict an approximately linear increase in the reciprocal of the denaturation temperature with the volume fraction () of solvent, but the polymer-melting mechanism predicts that the slope is inversely proportional to the molecular mass of the solvent (), whereas the polymer-in-a-box mechanism predicts a slope that is independent of . Differential scanning calorimetry was used to measure the denaturation temperature of collagen in different concentrations of ethylene glycol (=62) and the slope found to be (7.29±0.37)×10K, compared with (7.31±0.42)×10K for water (=18). This behavior was consistent with the polymer-in-a-box mechanism but conflicts with the polymer-melting mechanism. Calorimetry showed that the enthalpy of denaturation of collagen fibers in ethylene glycol was high, varied only slowly within the glycol volume fraction range 0.2 to 1, and fell rapidly at low . That this was caused by the disruption of a network of hydrogen-bonded glycol molecules surrounding the collagen is the most likely explanation. Abstract | Full Text | PDF (168 kb)

• Conformational Changes in Single-Strand DNA as a Function of...

  Conformational Changes in Single-Strand DNA as a Function of Temperature by SANS Biophysical Journal, Volume 90, Issue 2, 15 January 2006, Pages 544-551 J. Zhou, S.K. Gregurick, S. Krueger and F.P. Schwarz Abstract Small-angle neutron scattering (SANS) measurements were performed on a solution of single-strand DNA, 5′-ATGCTGATGC-3′, in sodium phosphate buffer solution at 10°C temperature increments from 25°C to 80°C. Cylindrical, helical, and random coil shape models were fitted to the SANS measurements at each temperature. All the shapes exhibited an expansion in the diameter direction causing a slightly shortened pitch from 25°C to 43°C, an expansion in the pitch direction with a slight decrease in the diameter from 43°C to 53°C, and finally a dramatic increase in the pitch and diameter from 53°C to 80°C. Differential scanning calorimeter scans of the sequence in solution exhibited a reversible two-state transition profile with a transition temperature of 47.5±0.5°C, the midpoint of the conformational changes observed in the SANS measurements, and a calorimetric transition enthalpy of 60±3kJ mol that indicates a broad transition as is observed in the SANS measurements. A transition temperature of 47±1°C was also obtained from ultraviolet optical density measurements of strand melting scans of the single-strand DNA. This transition corresponds to unstacking of the bases of the sequence and is responsible for the thermodynamic discrepancy between its binding stability to its complementary sequence determined directly at ambient temperatures and determined from extrapolated values of the melting of the duplex at high temperature. Abstract | Full Text | PDF (211 kb)

• Unique Properties of Purine/Pyrimidine Asymmetric PNA·DNA Du...

  Unique Properties of Purine/Pyrimidine Asymmetric PNA·DNA Duplexes: Differential Stabilization of PNA·DNA Duplexes by Purines in the PNA Strand Biophysical Journal, Volume 90, Issue 4, 15 February 2006, Pages 1329-1337 Anjana Sen and Peter E. Nielsen Abstract PNA·DNA duplexes are significantly stabilized by purine nucleobases in the PNA strand. To elucidate and understand the effect of switching the backbone in a nucleic acid duplex, we now report a thermodynamics study along with a solution conformations study of two purine/pyrimidine strand asymmetric duplexes and a strand symmetrical control by comparing the behavior of all four possible PNA/DNA combinations. In essence, we are comparing an identical basepair stack connected by either an aminoethyl glycine PNA or a deoxyribose DNA backbone. We show that the PNA·DNA duplexes containing purine-rich PNA strands are stabilized with regard to the thermal melting temperature and free energy as well as enthalpy (and concomitantly relatively less entropically disfavored). Based on our data, we find it unlikely that differences in counterion binding (identical ionic-strength dependence was observed), hydration (identical and insignificant water release was observed), or single-strand conformation can be responsible for the difference in duplex stability. The only consistent difference observed between the purine-rich PNA versus the pyrimidine-rich PNA in isosequential PNA·DNA duplexes is the significant increase in both binding enthalpy and entropy for the PNA·DNA duplexes containing pyrimidine-rich PNA in organic solvent, which would indicate that these duplexes are relatively enthalpically disfavored in water. Although our results so far do not allow us to identify the origin of the different stabilities of homopurine/homopyrimidine PNA·DNA duplexes, the evidence does point to a significant structural component, which involves enthalpic contributions both within the duplex structure and also from bound water molecules. Abstract | Full Text | PDF (169 kb)

• …more

Copyright © 2006 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 91, Issue 6, 2237-2242, 15 September 2006

doi:10.1529/biophysj.106.083691

Nucleic Acids

The Denaturation Transition of DNA in Mixed Solvents

Boualem Hammouda^*, ,   and David Worcester^†, ‡

^* National Institute of Standards and Technology Center for Neutron Research, Gaithersburg, Maryland
^† Cold Neutrons for Biology and Technology Project, Department of Physiology and Biophysics, University of California-Irvine, Irvine, California
^‡ Biology Division, University of Missouri, Columbia, Missouri

Address reprint requests to Boualem Hammouda.

DNA is a biopolymer formed of backbone phosphates, linked to desoxyribose sugars and side group amine bases 1. The charged phosphate groups are hydrophilic, the desoxyribose sugar groups are mostly hydrophobic, and the amine bases contain both hydrophobic and hydrophilic groups. This is based on water binding ability as referenced below. DNA forms a helical structure which is stable because of the stacking of the amine bases and of the hydrogen bonding between them. The helix phase melts into disordered coils under various conditions including temperature increase.

Most research on DNA has been in aqueous media at low concentration 2,3,4,5,6,7,8,9,10. Crystallographic measurements give precise sizes for the DNA helical structure: 3.4Å average distance between adjacent phosphates, 34Å for the helix pitch, and 20Å for the helix diameter 3. The ultraviolet (UV) absorption spectroscopy method 3,4 has proven valuable for the characterization of the helix-to-coil transition. Some investigations used non-aqueous solvents such as ethylene glycol 6,9. Other analytical methods such as NMR or infrared (IR) spectroscopy 8,10 have been used to estimate hydration properties. It was found for instance that a specific number of water molecules are bound to each DNA nucleotide 10 (5 molecules per amine basepair, 4 molecules per phosphate, and 2 molecules per desoxyribose sugar group). X-ray and neutron fiber diffraction yielded information about water hydration as well 11. In Fuller et al. 11, it is stated “challenges in developing alternatives to a water environment can be expected to be very severe”. This article reports a study of the ethylene glycol alternative.

The helix-to-coil melting transition in DNA has been the subject of a large number of investigations in the literature, among which are review articles 12,13,14. Cheng and Pettitt 13 contains a selected literature review of experimental and modeling efforts. Stability of the helix structure is governed by the amine base stacking and the base-pairing through hydrogen-bonding. Factors like temperature, DNA concentration, pH, salt concentration and solvent mixtures affect the helix-to-coil transition 13. The helix-to-coil transition occurs in transcription and replication of DNA, and is also is a key aspect of the polymerase chain reaction in biotechnology.

The focus of our investigations is on the characterization of the helix-to-coil denaturation transition of DNA 1 in ethylene glycol and 2 in water/ethylene glycol mixtures. Two measurement methods are used: UV light absorption spectroscopy to characterize the helix-to-coil melting temperature and small-angle neutron scattering (SANS) to monitor structural changes across the melting transition.

DNA from salmon testes of molecular weight 1.3×10⁶g/mol was purchased from Sigma and used at a mass fraction of 4% mass fraction in all cases. This DNA has 41.2% (number fraction) GC content. To screen charge interactions, 0.1M (or 0.0058 g/g mass fraction) concentration of NaCl salt was added in all cases.

The conventional method for characterizing the helix-to-coil melting transition in DNA is UV absorption spectroscopy. The 260-nm absorbance is a strong and reliable indicator of amine base stacking (or unstacking). A Cary 50 instrument was used with a temperature control system. Absorbance from the 4% DNA mass fraction samples was so strong that sample thicknesses around 50μm were used to avoid signal saturation (i.e., to keep the absorbance below 3).

Signals from 4% mass fraction DNA/d-water/0.1M NaCl and 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl were measured. Deuterated solvents (d-water and d-ethylene glycol) were used to keep consistency with the SANS measurements to be described later. Fig. 1 shows the melting curves and transition temperatures in both cases. The transition temperature with the d-ethylene glycol solvent is conveniently located at 38°C±0.5°C, well below the transition temperature with the d-water solvent at 94°C±0.5°C. The melting curves are characterized by a sharp increase of the 260-nm absorption (referred to as “first temperature” in Fig. 1), then a leveling off (referred to as “second temperature”). The transition temperature is chosen halfway between these two temperatures.

Display large version of this figure

Figure 1

UV absorbance of the 260-nm line from 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl and from 4% mass fraction DNA/d-water/0.1M NaCl and for a wide temperature range. The helix-to-coil denaturation transition is estimated to be 38±0.5°C and 94±0.5°C, respectively. Vertical error bars are smaller than the symbols.

UV absorption spectroscopy measurements were made on a series of samples containing various fractions of d-water/d-ethylene glycol. All samples contained 4% mass fraction DNA and 0.1M NaCl. Fig. 2 shows the variation of the helix-to-coil melting temperature with solvent content together with the two characteristic temperatures (the first and second temperatures that give an indication of the transition range). It is seen that the melting temperature follows an almost linear variation (with a slight curvature around 80% mass fraction d-ethylene glycol) and that the transition range is fairly uniform. The reported transition temperatures were obtained upon heating. This melting transition is partially reversible although with strong hysteresis. The monotonic linear variation is attributed to the fact that the transition was approached from the helix side whereby solvents mix randomly (ideal solvent mixing behavior).

Display large version of this figure

Figure 2

Variation of the denaturation melting temperature with increasing d-ethylene fraction for 4% mass fraction DNA in mixtures of d-water/d-ethylene glycol. 0.1M NaCl salt was added in each case. The bottom left-half region corresponds to the helix phase and the top right-half region corresponds to the coil phase.

The fact that the melting temperature shown in Fig. 2 decreases with d-ethylene glycol fraction points to the conclusion that the hydrophobic groups CD₂ in d-ethylene glycol play an important role in the melting transition. They help solvent molecules cross the hydrophobic zone of the desoxyribose sugar groups thereby loosening the helical structure (unstacking of the amine bases and breaking of the hydrogen bonds between these bases).

DNA dissolves in pure d-water or pure d-ethylene glycol. It also dissolves in 50% mass fraction d-water mixture with either one of the following solvents: d-methanol D[CD(OD)]D, d-ethylene glycol D[CD(OD)]₂D, or d-glycerol D[CD(OD)]₃D. These three samples provide a convenient series whereby the number of [CD(OD)] groups in the solvent molecule is increased. For these three samples, the melting temperature increases from 63.1°C for 4% mass fraction DNA/d-water/d-methanol/0.1M NaCl to 71.8°C for 4% mass fraction DNA/d-water/d-ethylene glycol/0.1M NaCl to 78.5°C for 4% mass fraction DNA/d-water/d-glycerol/0.1M NaCl. This result shows that even though both the CD and the OD groups in the solvent molecule are increasing equally, the melting temperature increases. A possible reason for this is that the extra OD groups tighten the hydration layer around the phosphate groups thereby stabilizing the helix phase. The extra CD groups cannot counteract this driving force.

These observations point to the importance of hydrophilic interactions (around the phosphate groups) and hydrophobic interactions (around the desoxyribose sugar groups) in the melting transition. When in the helix form, DNA is acting as a “micellar” system. The ability of the solvent to cross the hydrophobic region controls the stability of the helix phase.

Many models are available in the literature to describe the helix-to-coil transition. Efforts by Zimm 15,16 and Flory 17,18, using a configuration matrix method, presented predictions of the helix-to-coil transition temperature for the single strand and the double strand helical structures. The method summarized in Flory’s book 18 is followed here because of its intrinsic simplicity and its analytical form. Consider single strands consisting of N residues in the dilute regime and assume that the helical structure consists of ν helical sequences separated by coil sequences. Defining N_H as the total number of helical residues and N_C=N-N_H as the total number of coil residues, the relative helical fraction is p_H=N_H/N and the relative fraction of helical sequences is p_σ=ν/N. Introducing the enthalpy of melting of one helical residue as ΔH_m and the enthalpy of formation of one helical sequence as ΔH_σ, the associated partition function parameters are defined as s=exp(ΔH_m/RT) and σ=exp(ΔH_σ/RT), where R is the molar gas constant and T is the absolute temperature. The solution for this configuration matrix approach was presented 18 in terms of the following eigenvalues:

(1)

(2)

The enthalpy parameter s is related to the temperature T as s=1+ΔH_m(T_m-T)/RTT_m, where T_m is the melting temperature. A reasonable value for the enthalpy of melting ΔH_m=−6kcal/mol was used from the literature 19. The rescaled UV absorption data for the 4% DNA/0.1M NaCl/d-ethylene glycol are plotted in Fig. 3 along with the model’s best fit to the data corresponding to σ=0.0037. This value implies ∼29 helical sequences per 1000 residue at the melting transition point (at T=T_m) and an enthalpy of helical sequence formation of ΔH_σ=3,463 cal/mole. Zimm 16 had predicted a value between 1.5 and 3kcal/mol.

Display large version of this figure

Figure 3

The UV absorption data for the 4% DNA/0.1M NaCl/d-ethylene glycol sample across the helix-to-coil transition is compared to model prediction for the coil fraction p_C=1–p_H with the best fit parameter σ=0.0037. An enthalpy of melting of ΔH_m=−6kcal/mol has been used.

Other more involved models are available for the prediction of the helix-to-coil melting transition 19,20,21. These are based on numerical solutions and will not be pursued here.

SANS is a valuable measurement method for investigating structural changes and phase transitions in macromolecular systems. The use of deuterated solvents enhances the neutron contrast and therefore the measured signal. This technique probes length scales from the near atomic scale (a few angstroms) to the near micrometer scale. Focus in these investigations was put on the local structure measured in the high-Q region. Q is the magnitude of the scattering vector; it is proportional to the scattering angle (at the small angle approximation) and inversely proportional to characteristic inter-distances between scatterers within the sample.

The SANS technique uses a highly collimated monochromatic neutron beam incident on the sample. The scattered beam is detected by a position sensitive neutron area detector which records the scattering intensity for increasing scattering vector Q2πθ/λ, where θ is the scattering angle and λ is the neutron wavelength. SANS measurements were performed at the NIST Center for Neutron Research (NG3 SANS instrument). Standard data collection and reduction methods were used to obtain scattering intensity I(Q) on an absolute scale. Optimal sample thicknesses of 1mm were used in all cases. A series of measurements were performed from a 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl sample at temperatures ranging from 10°C to 80°C (at 5°C intervals). Fig. 4 shows a typical SANS spectrum at two temperatures; one below (25°C) and the other one above (50°C) the helix-to-coil transition temperature. The high-Q SANS signal is distinctively different in the two cases. The data show an abrupt decrease in the high-Q intensity for the helix phase but a gradual decrease for the coil phase.

Display large version of this figure

Figure 4

Small-angle neutron scattering from a 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl sample measured at temperatures below (25°C) and above (50°C) the helix-to-coil melting temperature. The high-Q signal clearly shows structural change across the melting transition. Error bars are very small.

SANS is sensitive to composition fluctuations in the sample. The helix structure is more compact and therefore characterized by a higher SANS intensity I(Q). SANS intensities are therefore characteristically different in the helical and coil phases. The SANS intensity data were fit to the following functional form 22,23 that reproduces the main data features:

(3)

The term A/Qⁿ represents the low-Q network scattering part and the term C/[1+(QL)^m] represents the high-Q solvation part. B represents a Q-independent (mostly incoherent) background. The low-Q part represents scattering from a large gel network structure. As shown in Fig. 4, the low-Q part does not change much across the melting transition. Our focus here is on the high-Q signal exclusively. Nonlinear least-squares fits were performed on the SANS data to obtain the C, L, m, and B parameters.

Fig. 5 shows the variation of the “solvation intensity” (the fitted quantity C in Eq. (1)) for increasing temperature. The intensity drop between 25°C and 40°C characterizes the helix melting transition. Lowering temperature shows that this transition is weakly reversible with substantial hysteresis. Further temperature increase beyond the melting transition increases the solvation intensity as shown in Fig. 5. This result is typical of water-soluble polymers and was observed for a 4% mass fraction poly(ethylene oxide)/d-water solution 22. In the case of PEO, the solvation intensity increased until a Lower Critical Solution Temperature (LCST) of 105.4°C was reached.

Display large version of this figure

Figure 5

Variation of the SANS solvation intensity (the quantity C in Eq. (1)) for increasing temperature for a 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl sample. For temperatures beyond the melting transition, the solvation intensity increases. Error bars correspond to three standard deviations.

Fig. 6 summarizes the variation of the correlation length (the quantity L in Eq. (1)) across the melting transition. Neutrons “see” hydrogen more than any other atoms. This correlation length represents a weighted-average interdistance between the hydrogen-containing (mostly sugar-amine base) groups. It is ∼8.5±0.2Å in the helix phase and increases to 12.3±0.2Å in the coil phase. In the helix phase the sugar-amine base groups are closer together than in the coil phase. This increase in L is due to the opening of the tight helix structure into a loser coil configuration. This correlation length is not a measure of the DNA radius (literature value of 10Å in the helix phase). It is a measure of the correlations between hydrogen atoms. Raising the temperature further in the coil phase increases the correlation length even more. A similar trend was also observed in PEO 22. The correlation length increases in the coil phase because correlations are transmitted more easily through the contour of flexible coils than they are through the stiff helical structures. In polymer coils, the correlation length is proportional to the entanglement length (average length between entanglement points along the polymer chain).

Display large version of this figure

Figure 6

Variation of the correlation length (the quantity L in Eq. (1)) for increasing temperature for a 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl sample. A jump in the variation of the correlation length L across the helix-to-coil transition is observed (see the Figure insets). After melting, DNA coils swell with further temperature increase.

Fig. 7 represents the variation of the high-Q Porod exponent (the quantity m in Eq. (3)). This exponent is seen to vary between values around 3.7±0.1 in the helix phase to values close to 1.7±0.1 in the coil phase. DNA helices are appearing like cylinders with fairly tight surfaces (Porod exponents close to 4) and DNA coils behave like polymer chains in good solvent conditions or in a fully swollen configuration (Porod exponent of 5/3=1.67) 22.

Display large version of this figure

Figure 7

Variation of the high-Q Porod exponent (the quantity m in Eq. (1)) for increasing temperature for a 4% mass fraction DNA/d-ethylene glycol/0.1M NaCl sample. This exponent varies from 3.7 (cylinder) to 1.7 (swollen coil).

It is noted that once the melting transition has taken place, DNA coils behave like water-soluble synthetic polymer chains.

SANS measurements were made for a series of 4% mass fraction DNA sample in mixed d-water/d-ethylene glycol solvent mixtures (with 0.1M NaCl salt added) at three temperatures: 1), at 25°C where all samples are in the helix phase (based on the UV data shown on Fig. 2); 2), at 50°C where most samples are in the helix phase and some are in the coil phase; and 3), at 75°C where most samples are in the coil phase. Fig. 8 shows variation of the solvation intensity (the quantity C in Eq. (1)) with increasing d-ethylene glycol fraction and for the three measured temperatures. Linear variation is observed in the helix phase and parabolic variation is observed in the coil phase. In the helix phase, solvents mix randomly around the helical structures, whereas in the coil phase nonideal solvent mixing is observed. Similar results were obtained in the case of a series of 4% mass fraction poly(ethylene oxide) (or PEO) solutions in mixtures of d-water and d-ethylene glycol (B. Hammouda, unpublished data). When in the presence of hydrophobic and hydrophilic groups, mixed solvents tend to arrange themselves efficiently so as to minimize conformational “stress” around the polymer coils in the solvation shell. Solvent mixtures are better solvation agents than either of the individual solvents. This is manifested as a lower solvation intensity as shown in the parabola portions of the graphs in Fig. 8. The SANS technique cannot resolve the orientational conformations of the solvent molecules around the DNA coils. Such a task is very difficult for noncrystalline systems like the ones investigated here, but SANS can monitor chain conformation fluctuations reliably. Our results point to the fact that when in the coil state, DNA behaves like the simplest water-soluble polymer (PEO) despite its chemical complexity.

Display large version of this figure

Figure 8

Variation of the SANS solvation intensity (the quantity C in Eq. (1)) for increasing d-ethylene glycol fraction in a series of 4% mass fraction DNA samples in d-water/d-ethylene glycol solvent mixtures measured at three temperatures (25°C, 50°C, and 75°C). 0.1M NaCl salt was added in all cases. Ideal solvent mixing is observed around DNA helices (linear variation), whereas nonideal solvent mixing is observed around DNA coils (parabolic variation).

Our results show that the salmon DNA helical structures are less stable in d-ethylene glycol than in d-water which is seen as a decrease of the helix-to-coil transition temperature. This conclusion is in agreement with an earlier report 9 for 21-mer DNA at low concentration. This is partly due to water’s greater ability to form hydrogen bonds. Ethylene glycol’s ability to form hydrogen bonds is lower because of the hydrophobic-D₂ groups. Ethylene glycol is the only nonaqueous solvent found to dissolve DNA in its pure form. Other solvents such as alcohols (for example ethanol) dissolve DNA only when they are mixed with water. Some of these solvents reported in the literature 9 are formamide, dimethyl sulfoxide, methanol, ethanol, and glycerol.

UV and SANS measurements have been conducted to characterize the helix-to-coil transition for 4% mass fraction DNA in d-water (94°C) and in d-ethylene glycol (38°C) with 0.1M NaCl salt content. Transition temperatures obtained from the two analytical methods are consistent. Measurements have also been made on DNA in mixtures of these two solvents over the entire mixing range. Linear variation of the transition temperature was found. The coil phase can be reached either though heating or by varying the d-water/d-ethylene glycol solvent content.

Structural information on the helix and coil phases has been obtained by SANS. Helices behave like cylinders with fairly tight surfaces. Coils are in a fully swollen configuration. A measured correlation length was found to increase from 8.5Å to 12.3Å across the DNA melting transition for the DNA/d-ethylene glycol/0.1M NaCl system. This correlation length is a characteristic interdistance between the hydrogen-containing (sugar-amine base) groups. In the helix phase, these groups are close together (inside the helix), whereas in the coil phase they are disordered side groups on the randomly distributed DNA coils. Beyond the melting transition, DNA coils behave like the simplest water-soluble polymer (PEO) chains. They swell with further temperature increase. A significance of these findings is that DNA with its complicated structure behaves like the simplest water soluble polymer when in the coil phase. The main significance is the substantial lowering of the helix-to-coil transition temperature when ethylene glycol is used. Ethylene glycol (prime ingredient in antifreeze) is a toxic poison and is of interest in biology research.

In the case of mixed solvents, it was found that solvent molecules mix randomly around the helical structures but not around the melted coils. In the coil phase, solvent mixtures were found to be more effective at solvating the polymer chains than any of the individual solvents.