Vibrational Relaxation of the Backbone and Base Modes in LacDNA

67-103 Donat Street, RO-400293 Cluj-Napoca, Romania. 2 UNED, Departamento de Ciencias y Técnicas Fisicoquímicas, Paseo de la Senda del Rey,...
0 downloads 0 Views 2MB Size
Subscriber access provided by NEW YORK UNIV

Article

Vibrational Relaxation of the Backbone and Bases Modes in LacDNA Complexes by UV Resonance Raman Spectroscopy Cristina M. Muntean, Ioan Bratu, and Antonio Hernanz J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.7b04271 • Publication Date (Web): 26 Jun 2017 Downloaded from http://pubs.acs.org on June 28, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Vibrational Relaxation of the Backbone and Bases Modes in LacDNA Complexes by UV Resonance Raman Spectroscopy Cristina M. Muntean 1*#, Ioan Bratu 1*# and Antonio Hernanz 2

1

National Institute for Research & Development of Isotopic and Molecular Technologies,

67-103 Donat Street, RO-400293 Cluj-Napoca, Romania 2

UNED, Departamento de Ciencias y Técnicas Fisicoquímicas, Paseo de la Senda del Rey,

9, E-28040 Madrid, Spain *

Corresponding authors:

Dr. Cristina M. Muntean [email protected] Dr. Ioan Bratu [email protected]

#

CMM and IB contributed equally to this work.

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

Abstract

Vibrational band shape analysis through time correlation function concept is widely used to obtain experimental information on the molecular dynamics of medium size molecules in different environments. Interesting details are revealed by extending this technique to biomolecules such as functional groups of the nucleic acids in media approaching the physiological conditions. In this work a study into the UV resonance Raman (UVRR) vibrational half bandwidths of functional groups in LacDNA, upon lowering the pH (pH 6.4, pH 3.45) and in the presence of Mn2+ and Ca2+ ions, respectively, was of interest. The corresponding global relaxation times have been derived. Also, the 793 cm-1 UVRR band, corresponding to ν (backbone O-P-O, dT) oscillator of LacDNA in aqueous solutions was selected for band shape analysis. Vibrational relaxation appears as the dominant relaxation process for this mode, vibrational dephasing being the most efficient for this oscillator. Current theories developed for vibrational dephasing have been applied to this profile and relevant relaxation parameters have been obtained and discussed. To our knowledge this is the first study on DNA oligomers vibrational band shape analysis through time correlation function concept.

2 ACS Paragon Plus Environment

Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Introduction

The dynamics of nucleic acids and their components is crucial for biological function of these biomolecules and particularly, plays a significant role in the molecular recognition of DNA-protein and DNA-ligand systems.1 The static structures of nucleic acids must function in dynamic processes in the living cell, through participation in specific interactions with other molecules. Relaxation processes of these biomolecules are thus a fundamental component of their behavior. Study of dynamics of the functional groups in nucleic acids can give information on their mobility and interactions. Particularly, the charged phosphate group in nucleotides and nucleic acids is responsible for interactions with counterions and for important features of the reactivity of these molecules.2 The vibrational spectra of biomolecules are commonly used to measure intensities and wavenumbers of absorption or light scattering. Nevertheless, these molecules are not isolated in the physiological media, where they develop their relevant biochemical activity. Timedependent forces act on them broadening their vibrational bands.3 Insight knowledge about the molecular dynamic processes involved in liquids mixtures can be obtained by Raman spectroscopy (4 and references therein). The macromolecular motion in fluids is generally too slow to be observed in the Raman time window that is accessible in the frequency domain. On contrary, the motion of functional groups can be fast enough.5,6 Previously, monitoring the changes in the normal Raman full widths at half-maximum (FWHMs) and, correspondingly, in the global relaxation times of the functional groups in DNA, upon varying the physico-chemical parameters, was of interest.7,8 In these cases, solution dynamics of nucleic acids was under study. Besides, the surface dynamics of plant

3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 33

genomic DNA on silver nanoparticles was reported by us for several cases, using surfaceenhanced Raman scattering (SERS) (9,10 and references therein). Furthermore, vibrational band shape analysis through time correlation function concept is widely used to obtain experimental information on the molecular dynamics of small and medium size molecules in different environments. Interesting details are revealed by extending this technique to biomolecules such as the components of the nucleic acids in media approaching the physiological conditions.3 The symmetric stretching mode, νs(PO32-), of several mononucleotides (5’-GMP, 5’-CMP and 5’-dCMP) have been studied in H2O and 2

H2O solutions, at different concentrations and temperatures.2,11-12 The effect of K+ and Mg2+

counterions on the vibrational energy transfer has also been considered. A big effort is being devoted to study the hydration dynamics of biomolecules. Very recently, backbone vibrations of DNA were introduced as probes of ultrafast structural fluctuations of hydrated DNA oligomers.13 Dynamics at the DNA surface, in the region of the backbone have been studied in artificial DNA helices at different hydration levels. Nonlinear two-dimensional infrared spectra of backbone modes were found to be highly sensitive to interfacial interactions with the aqueous environment.14 All these advances are contributing to obtain a better knowledge on vibrational relaxation mechanisms and to build an interesting picture of the molecular dynamics of nucleic acids, their derivatives and components (3,11-12 and references therein). Particularly, resonance Raman spectroscopy (RRS) is a useful technique employed to obtain experimental data on molecular relaxation dynamics.15 It is specially appropriated to study biomolecules.16-17 Normal Raman spectroscopy suffers from small scattering cross section, which limits its use as a low-level bioanalytical technique (18 and references therein). In the UV resonance Raman (UVRR) approach, the laser-excitation wavelength is tuned into resonance with the target molecular subunit, and thereby the resonance-enhanced Raman

4 ACS Paragon Plus Environment

Page 5 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

spectrum of the subunit is selectively obtained.19 UVRR spectroscopy can be a source of more detailed information regarding local structure and dynamics of DNA as a function of sequence and conformation (20 and references therein). An additional advantage of the selectivity of UVRR is its sensitivity. Because of the large increase in the intensity of scattering due to the electronic resonance, sample concentrations as low as 10-5 M may be used.21 Different studies on DNA structure and interactions as probed with resonance Raman spectroscopy were presented.22-25 In the following, we report an UV Resonance Raman spectroscopic (UVRR) study on double-stranded and single-stranded DNA oligomers. We will concentrate on the vibrational half bandwidths and time correlation functions (CF) that give information on molecular dynamics behavior. Particularly, investigation of subpicosecond dynamical changes induced in a natural DNA recognition site, in the presence and absence of divalent metal ions (Mn2+, Ca2+) at two pH values (6.4 and 3.45), respectively, providing data about changes in the half bandwidths and in the global relaxation times of LacDNA vibrational modes, are of interest. The targeted DNA is a non-palindromic 22-mer duplex representing the primary cyclic AMP receptor protein

(CRP)

binding

site

of

the

E.

coli

lac

promoter,

d(TAATGTGAGTTAGCTCACTCAT)· d(ATGAGTGAGCTAACTCACATTA) (LacDNA) (26 and references therein). Also, its corresponding single-stranded oligonucleotides, d(TAATGTGAGTTAGCTCACTCAT) (SS1) and d(ATGAGTGAGCTAACTCACATTA) (SS2) were investigated.18 It is shown that changes in the subpicosecond dynamics of functional groups in LacDNA complexes can be monitored with UV Resonance Raman spectroscopic (UVRR). The low concentration of LacDNA in our aqueous systems, imposed the use of UVRR in our present study.

5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 33

We will present in the following the motivation for the selection of our aqueous systems. Several mechanisms involving DNA, e.g. S1-nuclease recognition in eukaryotic cells, might involve structural transitions from B-form DNA to protonated DNA forms, at low pH. Besides, in vitro studies show that the stability and specificity of protein-DNA complexes are remarkably dependent on the type and concentration of ions present in the solvent milieu. The sensitivities of these complexes to changes in the salt concentration derive mostly from the contributions of charge-charge interactions between the protein and DNA. Experimental findings indicate that the internal salt concentration may play a general role in modulating differential gene expression in E. coli (27 and references therein). Particularly, it is widely recognized that Ca2+ ions are central to a complex intracellular messenger system that is mediating a wide range of biological processes. Besides in the living cell, manganese(II) ions act as cofactors for a large variety of enzymes (proteins) with many functions.

Experimental

One set of spectroscopic data18 is analyzed in this work from the dynamic point of view. The experimental details of the UVRR spectra, obtained for LacDNA complexes were given

in

Muntean

et

al.,

2013.18

Single-stranded

oligonucleotides,

d(TAATGTGAGTTAGCTCACTCAT) (SS1) and d(ATGAGTGAGCTAACTCACATTA) (SS2) were purchased from Eurogentec, Belgium. DNA sequences have been HPLC-purified by the vendor and were used as received. The LacDNA 22-mer duplex has been prepared by mixing SS1 and SS2 in a 1:1 molar ratio in a DNA hybridization buffer (10 mM Tris, 150 mM NaCl, pH 7), in particular incubation conditions. Aliquots of LacDNA in storage buffer (about 100 µM concentration) were dialyzed against appropriate buffers, respectively, to

6 ACS Paragon Plus Environment

Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

obtain the 22-mer duplex complexes, in the presence and absence of divalent metal ions (Mn2+, Ca2+) at two pH values (6.4 and 3.45), respectively.18 Different physico-chemical conditions have been used18 to gain a molecular picture of the LacDNA molecule, changing its dynamics upon electrophilic agent binding to it, in solution. The following abbreviations have been used: DNA1 [LacDNA, 0 mM M2+ (divalent metal ion), pH 6.4]; DNA2 [LacDNA, 0 mM M2+ (divalent metal ion), pH 3.45]; DNA3 (LacDNA, 10 mM Mn2+, pH 6.4); DNA4 (LacDNA, 10 mM Mn2+, pH 3.45); DNA5 (LacDNA, 10 mM Ca2+, pH 6.4); DNA6 (LacDNA, 10 mM Ca2+, pH 3.45). Extinction

spectra

d(ATGAGTGAGCTAACTCACATTA)

of (LacDNA

d(TAATGTGAGTTAGCTCACTCAT)· duplex),

in

the

physico-chemical

conditions investigated here, show a maximum around 260 nm.18 So, for UV resonance Raman spectroscopy, we have used an excitation wavelength near the UV-Vis absorption spectra maximum of nucleic acids. Continuous wave Raman equipment was used. UVRR measurements were done with an excitation line of 275 nm from an argon ion laser (Spectra Physics, model BeamLok 2085). Raman spectra were recorded employing a 90° scattering geometry, using a rotating quartz cuvette.18 We have used an incident power around 15 mW. The beam diameter at focus was 18.2 µm. The scattered light was focused on the entrance slit of a double monochromator (Spex, model 1404 with 2400 grooves mm-1 holographic gratings, focal length 0.85 m) and detected with a liquid nitrogen cooled CCD camera (Photometrics, model SDS 9000).18 All UV resonance Raman spectra of DNA samples were registered in the same experimental conditions.

7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 33

Data treatment

The band parameters (full widths at half-maximum, FWHMs and global relaxation times) were determined for the data set. For each band profile, an individual baseline was taken into account. Consequently, the FWHMs were evaluated from the half maximum UVRR bands. The correlation functions (CFs) were computed by Fourier transform of the experimental profiles and were theoretically modeled by employing self-made software (see next section).

Obtaining vibrational relaxation parameters from UV Resonance Raman band shapes

The fastest relaxation process determines the shape of the band contour. In large and very slowly rotating molecules, taking into account the steric hindrance of the rotational motion, the non orientational relaxation processes are favored, i.e. vibrational ones. Among vibrational relaxation processes, resonant relaxation (important in pure liquids), non resonant ones (by rotational-translational energy transfer or transfer to other vibrational modes) and vibrational dephasing can be mentioned. The last one seems to be the most important vibrational relaxation mechanism.28 Depending on the intermolecular potentials of the lattice, the processes which modulate vibrational frequency of the oscillator are called “fast” or “slow”, respectively. In the case of “slow” modulation the perturbation of the intermolecular potentials remains significant for a long time, i.e. the initial phase of individual oscillators is rapidly lost. In the case of “fast” modulation, the perturbations decrease rapidly and the original phase memory is maintained for a longer time. One can speak then about “motional narrowing” of the band contours.28 8 ACS Paragon Plus Environment

Page 9 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

A relatively isolated band in the Raman spectra of a compound, that does not overlap with other bands and is well assigned to a vibrational mode, is the ideal situation in order to obtain reliable experimental information on the vibrational relaxation of the corresponding oscillator.3 The measurement of full width at half-maximum (FWHM) of the band, ∆ν 1 / 2 , depends only on three points of the band shape, and it is commonly used. In case of small overlapping from other bands over one of the band wings, the semi-FWHM measured over the other wing could be used. Assuming a Lorentzian profile for the experimental band shape, a condition very often realized in solution, the FWHM may be expressed by:

∆ν1/2 = (cπτ)-1

[1]

where τ is a “relaxation time” relative to all the relaxation processes contributing to the observed band profile, and c is the velocity of light.3,29 Correctly, vibrational correlation function, Gv(t) can be obtained from isotropic Raman band Iiso as Fourier transform of the experimental vibrational profile: Gv(t)=∫ Iiso(ω) eiωtdω

[2]

From these experimental CFs and the calculated CFs fitting different theoretical models developed for vibrational and rotational mechanisms, it is possible to obtain values of significant relaxation parameters: relaxation time (τUVRRCF), time between collisions or modulation time (τc), modulation amplitude (〈 ω2〉), modulation speed (1/2·τc), etc.3 From various vibrational relaxation mechanisms, energy relaxation, resonant energy exchange and vibrational dephasing, the latter seems to be the most important.3,30 In biological molecules like nucleic acids with a large molecular weight, the contribution of an intramolecular near-resonant energy transfer process can, in principle, appear. From this point of view, vibrational relaxation cannot be due to the dephasing mechanism exclusively (12 and 9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 33

references therein). Particularly, it was shown that the main process responsible for the formation of the isotropic contour of Raman symmetrical stretching vibration bands of anions in solutions of electrolytes was vibrational dephasing, because of the formation of ionmolecular H-bonds.30 Some models have been developed for vibrational dephasing, based on GaussianMarkovian theory.28,31-33 According to Kubo-Rothschild’s model the logarithm of the vibrational CF is expressed by:

{

ln Gv (t ) = − ω 2 τ c2 [exp(− t / τ c ) − 1] + τ ct

}

[3]

where τc is the modulation time, which may be roughly considered as the time between collisions with the solvent and 〈 ω2〉 = 4πc2M2 is the mean squared frequency displacement around the band centre caused by the different environments “seen” by the oscillator (the average frequency fluctuation). Its magnitude is determined by the strength of the coupling between the oscillator and its environment; hence, it is called modulation amplitude. In the above relation M2 is the second band moment. The Kubo-Rothschild model stochastic lineshape theory has been applied successfully to different systems.34-36 However, Oxtoby37 proposed another equation for this CF: ln Gv (t ) = − ω 2 τ c2 ln cosh (− t / τ c )

[4]

Both equations behave at long times as straight lines with slopes -〈 ω2〉 τc. These equations have been fitted to the experimental CFs, obtained from the real part of the Fourier transform of the UVRR band profiles (793 cm-1) of LacDNA in aqueous solutions. The observed values of M2 might be used to evaluate 〈 ω2〉. The modulation time, τc and the quantity 〈 ω2〉1/2 τc, often named modulation speed, corresponding to the best fit to each model, were deternined.3 In the following we will present the vibrational dephasing models.

10 ACS Paragon Plus Environment

Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The Markovian character of the process takes into account the hypothesis that every next frequency depends only of one previous frequency. By employing the simple exponential cumulant expansion, one obtains the well-known Kubo equation28,32: -ln Gv(t)/M2τω2 = exp(-t/ τω) -1 + t/ τω

[5]

The use of stretched exponential Gω (t) = exp (-t/ τω)α

[6]

where 0< α ≤1 gives the following time correlation function (Sigma model)38: -ln Gv(t)/M2τω2=∑n=0∞[(-1)n(t/ τω)2+nα / n!(1+nα)(2+nα)]

[7]

as to describe non-Markovian modulation.39 Eq [7] transforms into Eq [5] at α=1. η is the modulation speed (η= 1/2 τc) for Sigma model, whereas T is a parameter similar to τω which varies between 0 and 1. α and T are also fit parameters without dimension, which result from Sigma model. The stretched exponential model (Sigma) is best for molted salts or aqueous solutions (38 and references therein). Measurements of the band moments and time correlation functions (CFs) as well as fittings to Kubo-Rothschild’s,28 Oxtoby’s37 and Sigma’s (Kirillov) equations were performed using home written programs. The Fourier transformation was computed by direct numerical integration of the transform integral at each time point. The global relaxation time was evaluated using the value of FWHM. The modulation time τc was obtained from the fittings of Kubo-Rothschild’s, Oxtoby’s and Sigma’s equations to the experimental CF.

Results and discussion

UV resonance Raman spectra of DNA1-DNA6 systems18, analyzed in this work from the dynamic point of view, are presented in Figure 1.18

11 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 33

Spectral changes and their correlations with LacDNA solution structure are given in detail in Muntean et al., 2013.18 For comparison, normal Raman spectra of calf thymus DNA at reduced and acidic pH values, in the presence of 10 mM Mn2+, excited with the 488-nm laser line, are presented elsewhere.40

Figure 1. The 275 nm-excited UV resonance Raman spectra of dissolved 22-mer duplex d(TAATGTGAGTTAGCTCACTCAT)·d(ATGAGTGAGCTAACTCACATTA) (LacDNA) at about 100 µM concentration, in different physico-chemical conditions, respectively, as labeled in the figure: DNA1 (pH 6.4); DNA2 (pH 3.45); DNA3 (10 mM Mn2+, pH 6.4); DNA4 (10 mM Mn2+, pH 3.45); DNA5 (10 mM, Ca2+, pH 6.4); DNA6 (10 mM, Ca2+, pH 3.45). The presented wavenumber range is 620-1800 cm−1. The spectra were scaled to have equal intensity in the 793 cm-1 DNA band [C. M. Muntean, M. Salehi, S. Niebling, B. Walkenfort, The influence of divalent metal ions on low pH induced LacDNA structural changes as probed with UV resonance Raman spectroscopy, J. Raman Spectrosc., 2013, 44(12): 1693-1699. Copyright © 2013 John Wiley & Sons, Ltd. Reproduced with permission]. This choice of low pH was based on the observation that the midpoint of transition of Watson-Crick GC base pairs to protonated GC base pairs lies at around pH 3 (analyzing the guanine 681 cm-1 line from the Raman spectrum of calf-thymus DNA). Appropriated 12 ACS Paragon Plus Environment

Page 13 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

experimental conditions (solvent type, DNA concentration, pH, concentration of divalent metal ions, respectively) have been selected in relation with the hypothesis established in the theoretical models, e.g. the corresponding conditions to apply the vibrational dephasing models (diluted concentrations to avoid intermolecular interactions). We have chosen the value of 3.45 for low acidic pH and also low concentrations of divalent metal ions, in order to avoid DNA condensation in the samples (to maintain the stability of DNA). BIS-TRIS buffers used to obtain LacDNA complexes in aqueous solutions at pH 6.4, require reduced pH in order to keep their buffer properties.

Full-widths at half - maximum (FWHM) and global relaxation times

In this subsection a study into the Raman vibrational bandwidths of functional groups in double-stranded LacDNA, upon lowering the pH (6.4, 3.45) and in the presence of Mn2+ and Ca2+ ions, respectively, is of interest (see Table 1). Also, vibrational Raman profiles of the corresponding single-stranded DNAs are analyzed in this part. It is shown that changes in the subpicosecond dynamics of functional groups in LacDNA can be monitored with UV resonance Raman spectroscopy. Resonance Raman scattering simplifies the highly complex and congested vibrational pattern of DNA as exclusively vibrations coupled to the electronic transition are resonantly enhanced (41 and references therein). For the case of aqueous solutions of LacDNA molecules we can suppose that the dominant relaxation mechanism is the vibrational one. The values of the global relaxation time suggest also the existence of a vibrational relaxation time, because the reorientational movement is much slower for the 22-mer DNAs in aqueous solutions, respectively. Particularly, the absence of reorientational broadening in polynucleotides indicates that the

13 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 33

bases in polynucleotides reorient through an angle of 41º in times slower than 21 ps (7 and references therein). The half bandwidths of LacDNA UVRR vibrations have been measured at reduced (pH 6.4) and low (pH 3.45) pH values for DNA, MnDNA and CaDNA systems (Table 1). The corresponding global relaxation times evaluated on the basis of Eq. 1 are also presented in the Table 1. The full-widths at half-maximum (FWHM) of the UVRR bands studied in this work are in the wavenumber range 13.5- 39.5 cm-1 for double-stranded LacDNA and between 14 and 44 cm-1 for single-stranded DNA, respectively. The limit values of the interval 13.5- 39.5 cm-1 were obtained for CaDNA UVRR vibrations (pH 6.4). Limit values of FWHM range of metal-DNA bands were also observed in the case of Ca2+ ions, in a previous work of us dealing with normal Raman spectroscopy of calf-thymus DNA.29 The molecular relaxation processes corresponding to double-stranded LacDNA have a global relaxation time smaller than 0.786 ps and larger than 0.269 ps. For single-stranded oligonucleotides, a global relaxation time between 0.241-0.758 ps have been found. The vibrational energy transfer, corresponding to the modes around 1385 cm-1 ( dA, dT), 1430 cm-1 (C2’H2 scissor, dA, dG), 1493 cm-1 [dG(N7), dA] are faster in the case of double-stranded DNA (DNA1), as compared with the single-stranded DNA (SS1), respectively. On the contrary, the global relaxation times corresponding to the bands near 1590 cm-1 (dA, dG) and 1669 cm-1 [dT(C=O)] are increased for double-stranded DNA (DNA1), as compared with the single-stranded DNA (SS1), respectively, due to a larger molecule in the case of nucleic acid duplex, with an increased number of oscillators. The molecular relaxation processes are slower for the bands at 793 cm-1 (bk O-P-O, dT) as compared with the bands near 1490 cm-1 [dG(N7), dA] in the case of DNA1-DNA6 samples, respectively and this is explained by the rigidity of the DNA backbone, as compared

14 ACS Paragon Plus Environment

Page 15 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

with purinic nitrogenous bases of nucleic acids. The global relaxation time corresponding to the 793 cm-1 vibrational mode has an increasing value for DNA1, DNA3 and DNA5 (pH 6.4) and also a decreasing value for DNA2, DNA4 and DNA6 (pH 3.45), respectively. This behavior is caused by the binding of divalent metal ions to LacDNA molecules, either at N7 of guanine and to a lesser extent to adenine or by electrostatic interactions to phosphate groups. This behavior seems to be pH dependent, being influenced by the presence of bound protons to the 22-mer double helix, at low pH. Changes in pH reflect also structural changes involved in the hydrogen bonding strength between nitrogenous bases of DNA and may cause large changes in the electronic structure and delocalization within the molecule (18 and references therein). Particularly, adenine protonation leads to disruption of AT base pairs and to appearance of single-stranded regions in the double helical chain. The 1385 cm-1 band parameters attributed to dA, dT vibrations seems to be very stable in the case of DNA1, DNA3, DNA6 (0.544 ps) and DNA2, DNA4, DNA5 (0.531 ps), respectively. Molecular dynamics associated with the 1432 cm-1 vibration is faster in the metal-LacDNA complexes at low pH, as compared with that at reduced pH value. At low pH, there is a competition in the binding of proton and metallic cations to DNA. We appreciate that the number of the metal ions bound to LacDNA is increased in the case of reduced pH value, as compared to the low pH one, contributing to a slower vibrational dynamics at pH 6.4. Also, vibrational energy transfer processes characteristic to this band (C2’H2 scissor, dA, dG) is slower for CaLacDNA complex, as compared with Mn-LacDNA one, both at pH 6.4 and pH 3.45 respectively. Besides, the vibrational energy transfer is identical for the 1490 cm-1 vibration [dG(N7), dA], belonging to DNA1, DNA3 and DNA4 (0.472 ps). Vibrational relaxation corresponding to the dA, dG band near 1587 cm-1 is slower in the case of metal-LacDNA complexes at pH 3.45, as compared with metal-LacDNA complexes at pH 6.4, respectively. Also, molecular dynamics associated with this mode, is faster for LacDNA, in the presence of

15 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 33

Mn2+ ions as compared with LacDNA in the presence of Ca2+ ions, at pH 6.4 and pH 3.45, respectively. The carbonyl group of dT (C=O) manifests a decrease in the global relaxation time upon lowering the pH, in the absence of divalent metal ions and in the case of 10 mM Mn2+ ions, respectively. Besides, a slight increase in the global relaxation time upon lowering the pH, in the case of 10 mM Ca2+ ions, has been found for the 1665 cm-1 band. For double stranded LacDNA (pH 6.4), the characteristic global relaxation times of DNA functional groups decrease for the bands at 1385 (dA, dT), 1431 (C2’H2 scissor, dA, dG) and 1490 cm-1[dG(N7), dA], as compared to those corresponding to single-stranded DNA samples (SS1, SS2, pH 7). Besides, the vibrational energy transfer is slower for the 1587 (dA, dG) and 1668 cm-1 [dT(C=O)] vibrations, in the case of double-stranded LacDNA (pH 6.4), as compared to those of ssDNAs.

Table 1. Total half bandwidths (cm-1) of UVRR vibrational markers and global relaxation times of functional groups, characterizing LacDNA complexes in different physico-chemical conditions and the corresponding single-stranded oligonucleotides. Raman spectra were excited with a 275 nm laser line (see text for details). νmax(cm-1)

∆ν1/2 (cm-1)

1385 1430 1493 1590 1669

18±1 14±1 20±1 15.5±1 37.5±1

1386 1432 1493 1589 1668

19±1 15±1 21±1 15.5±1 44±1

793 1385 1431 1490 1587

17±1 19.5±1 15±1 22.5±1 14.7±1

τUVRR (ps) SS1 0.589±0.033 0.758±0.054 0.531±0.027 0.684±0.044 0.283±0.008 SS2 0.558±0.029 0.707±0.047 0.505±0.024 0.684±0.044 0.241±0.005 DNA1 (pH 6.4) 0.624±0.037 0.544±0.028 0.707±0.047 0.472±0.021 0.722±0.049

Tentative assignment a,18 dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O) dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O) bk O–P–O b, dT dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG 16

ACS Paragon Plus Environment

Page 17 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

1668

35.5±1

795 1382 1432 1492 1586 1670

14±1 20±1 19±1 37.5±1

794 1385 1432 1491 1588 1664

16±1 19.5±1 14.5±1 22.5±1 15.5±1 33.5±1

793 1384 1430 1491 1585 1665

15.5±1 20±1 15.5±1 22.5±1 14±1 38±1

791 1384 1430 1489 1587 1665

14.5±1 20±1 13.5±1 22±1 14.8±1 39.5±1

793 1384 1430 1489 1586 1665

16±1 19.5±1 14.5±1 24.5±1 13.5±1 35.5±1

0.299±0.008 DNA2 (pH 3.45) 0.758±0.054 0.531±0.027 0.558±0.029 0.283±0.008 DNA3 (10 mM Mn2+, pH 6.4) 0.663±0.041 0.544±0.028 0.732±0.050 0.472±0.021 0.685±0.044 0.317±0.009 DNA4 (10 mM Mn2+, pH 3.45) 0.684±0.044 0.531±0.027 0.685±0.044 0.472±0.021 0.758±0.054 0.279±0.007 DNA5 (10 mM, Ca2+, pH 6.4) 0.732±0.050 0.531±0.027 0.786±0.058 0.483±0.022 0.717±0.048 0.269±0.007 DNA6 (10 mM, Ca2+, pH 3.45) 0.663±0.041 0.544±0.028 0.732±0.050 0.433±0.018 0.786±0.058 0.299±0.008

dT(C=O) bk O–P–O b, dT dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O) bk O–P–O b, dT dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O) bk O–P–O b, dT dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O) bk O–P–O b, dT dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O) bk O–P–O b, dT dA, dT C2’H2 scissor, dA, dG dG(N7), dA dA, dG dT(C=O)

a

Abbreviations: dA - deoxyadenosine; dG - deoxyguanosine; dC - deoxycytidine; dT deoxythymidine. b bk - backbone.

17 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 33

Relaxation of the 793 cm-1 (backbone O-P-O, dT) oscillator of LacDNA in aqueous solutions. Experimental CFs and their theoretical modelling The mode near 793 cm-1 (backbone O-P-O, dT) gives rise to a relatively isolated band in the UV resonance Raman spectra of LacDNA in aqueous solution, allowing additional dynamics analysis, e.g. vibrational band shape analysis through time correlation function concept. A basic condition to study the dynamical properties of an oscillator using this concept, is the presence of a spectral region of at least 3 half bandwidths free of overlapping, before and after the vibrational profile of the selected band and these conditions are fulfilled by this mode. Nevertheless, the 793 cm-1 (backbone O-P-O, dT) mode is a weak UV resonance Raman scatterer, that generates a weak UV resonance Raman band. However, for the facts mentioned above a band shape analysis was possible to be done. Particularly, backbone vibrations involve molecular elongations at the DNA-water interface and, thus, are particularly sensitive to dynamics originating from interactions between the charged and polar regions of the DNA surface and the water dipoles and counterion atmosphere at the interface.13 The corresponding experimental UV resonance Raman CFs for DNA1-DNA4, DNA6 samples are presented in Figure 2. Data are based on the 275 nm-excited UV resonance Raman spectra of dissolved 22-mer duplex d(TAATGTGAGTTAGCTCACTCAT)· d(ATGAGTGAGCTAACTCACATTA) (LacDNA) at about 100 µM concentration, in different physico-chemical conditions, respectively, as labeled in the Figure 1.18 CF decayed more rapidly starting from DNA2 up to DNA6 as compared to the corresponding one for DNA1; the relaxation processes are more efficient on going from DNA1 to DNA6. The 793 cm-1 band shows a fairly symmetric profile without shoulders or splitting (Figure 1). The main contribution to the UVRR CFs is of vibrational type, due to an adequate

18 ACS Paragon Plus Environment

Page 19 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

selection of the investigated systems (22-mer DNAs in aqueous solutions, respectively). The values of the relaxation time (τUVRRCF) suggest also the existence of a vibrational relaxation time, because the reorientational movement is much slower for these molecules or for their functional groups in aqueous systems.

Figure 2. Natural logarithm of the UV resonance Raman correlation functions (CFs) for the 793 cm-1 (backbone O-P-O, dT) band of LacDNA at about 100 µM concentration in aqueous solution, at 0 mM M2+ (divalent metal ion), pH 6.4 (DNA1); 0 mM M2+(divalent metal ion), pH 3.45 (DNA2); 10 mM Mn2+, pH 6.4 (DNA3); 10 mM Mn2+, pH 3.45 (DNA4) and 10 mM Ca2+, pH 3.45 (DNA6).

Vibrational correlation function (CF) was assumed to be quite well represented by Fourier transform of UV resonance Raman profile of the 793 cm-1 band, corresponding to 22mer LacDNAs in aqueous solutions, respectively. Modelling of the experimental CF with Kubo-Rothschild, Oxtoby and Sigma theoretical CF models is presented in Figures 3-7 for different experimental conditions, respectively.

19 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 33

Figure 3. Natural logarithm of the UV resonance Raman correlation function (CF) for the 793 cm-1 (backbone O-P-O, dT) band of LacDNA, at about 100 µM concentration in aqueous solution, at A: 0 mM M2+ (divalent metal ion), pH 6.4 (dash). Kubo-Rothschild’s (solid line) fitted equation and B: 0 mM M2+ (divalent metal ion), pH 6.4 (dash). Oxtoby’s (solid line) fitted equation.

20 ACS Paragon Plus Environment

Page 21 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 4. Natural logarithm of the UV resonance Raman correlation function (CF) for the 795 cm-1 (backbone O-P-O, dT) band of LacDNA, at about 100 µM concentration in aqueous solution, at A: 0 mM M2+ (divalent metal ion), pH 3.45 (dot). Kubo-Rothschild’s (solid line) fitted equation. Kubo-Rothschild’s and stretched exponential (Sigma) functions coincide (α ∼ 1) and B: 0 mM M2+ (divalent metal ion), pH 3.45 (dot). Oxtoby’s (solid line) fitted equation.

21 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 33

Figure 5. Natural logarithm of the UV resonance Raman correlation function (CF) for the 794 cm-1 (backbone O-P-O, dT) band of LacDNA, at about 100 µM concentration in aqueous solution, at A:10 mM Mn2+, pH 6.4 (black solid line). Kubo-Rothschild’s (red solid line) fitted equation. B: 10 mM Mn2+, pH 6.4 (dot). Oxtoby’s (red solid line) fitted equation. C: 10 mM Mn2+, pH 6.4 (dot). Sigma’s (red solid line) fitted equation. Kubo-Rothschild’s and stretched exponential (Sigma) functions coincide (α ∼ 1). 22 ACS Paragon Plus Environment

Page 23 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 6. Natural logarithm of the UV resonance Raman correlation function (CF) for the 793 cm-1 (backbone O-P-O, dT) band of LacDNA, at about 100 µM concentration in aqueous solution, at A: 10 mM Mn2+, pH 3.45 (dot). Kubo-Rothschild’s (red solid line) fitted equation. B: 10 mM Mn2+, pH 3.45 (dot). Oxtoby’s (red solid line) fitted equation. C: 10 mM Mn2+, pH 3.45 (dot). Sigma’s (red solid line) fitted equation. α = 0.6; Sigma model CF is the best correlation function that can describe vibrational relaxation in this case. 23 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 33

Figure 7. Natural logarithm of the UV resonance Raman correlation function (CF) for the 793 cm-1 (backbone O-P-O, dT) band of LacDNA, at about 100 µM concentration in aqueous solution, at A: 10 mM Ca2+, pH 3.45 (dot). Kubo-Rothschild’s (red solid line) fitted equation. B: 10 mM Ca2+, pH 3.45 (dot). Oxtoby’s (red solid line) fitted equation. C: 10 mM Ca2+, pH 3.45 (dot). Sigma’s (red solid line) fitted equation. 24 ACS Paragon Plus Environment

Page 25 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The experimental CF for DNA5 could not be modeled with theoretical CFs. The experimental UVRR profile of DNA5 near 795 cm-1 has contributions in both sides of the maximum, being not free of overlapping, contrary to the other cases (DNA1, DNA2, DNA3, DNA4 and DNA6). So, it is not possible to obtain the whole profile by doubling one half of the band, either in the low wavenumbers region or in the high wavenumbers one. Besides, oscillations in the high wavenumbers profile side of the maximum were found. Due to the fact, that the experimental correlation function has oscillations for DNA5, it could not be modeled with theoretical correlation functions. The best model of each experimental vibrational CF can be selected as in the followings: -Oxtoby model fits better the experimental CF than Kubo-Rothschild one for DNA1; -Kubo-Rothschild model fits well at long times (>1 ps) for DNA2; in this case, Oxtoby model seems to be more suitable at short times (