Biomacromolecules 2001, 2, 1137-1147
13C
1137
Nuclear Magnetic Relaxation Study of Segmental Dynamics of the Heteropolysaccharide Pullulan in Dilute Solutions Photis Dais* and Soultana Vlachou NMR Laboratory, Department of Chemistry, University of Crete, 914 09 Iraklion, Crete, Greece
Francois R. Taravel Centre de Recherches sur les Macromolecules Vegetales (CERMAV), CNRS, BP 53, 38041 Grenoble Cedex 9, France (affiliated with the Joseph Fourier University of Grenoble) Received April 5, 2001 13C
spin-lattice relaxation times (T1) and nuclear Overhauser enhancements (NOE) were measured as a function of temperature and magnetic field strength for the heteropolysaccharide pullulan in two solvents, water and dimethyl sulfoxide. The relaxation data of the endocyclic ring carbons were successfully interpreted in terms of chain segmental motions by using the bimodal time-correlation function of Dejean de la Batie, Laupretre, and Monnerie. On the basis of the calculated correlation times for segmental motion, the flexibilities of the pullulan chain at a repeat-unit level have been studied and compared with the segmental mobility of the homopolysaccharides amylose and dextran in the same solvents. The internal rotation of the free hydroxymethyl groups about the exocyclic C-5-C-6 bonds superimposed on segmental motion has been described as a diffusion process of restricted amplitude. The rate and amplitude of the internal rotation of the free hydroxymethyl groups were not affected by the local geometry of the pullulan chain. Introduction It is well-known that NMR spectroscopy is a powerful technique to study the conformational dynamics of synthetic and natural polymers in dilute solutions (e10% w/v).1-3 In particular, 13C relaxation experiments can detect motion at several carbon sites in a repeat unit of the polymer chain, thus providing a detailed insight into molecular motion at a repeat-unit level. However, the quantification of these motions requires the knowledge of the time-correlation function (TCF), which is a projection of mechanisms and rates of motions occurring in the polymer chain. In principle, TCF function can be obtained by an inverse Fourier transform of the experimentally measured spectral density function. Unfortunately, nuclear magnetic relaxation measurements at one magnetic field strength sample motions at very narrow frequency range that involves the Larmor frequencies of proton and carbon nuclei, as well as the sum and differences of these frequencies, which are of the same magnitude as the individual frequencies. Even in the extreme case, where one can obtain the spectral density function at a large number of magnetic field strengths, the inverse Fourier transform of the spectral density functions will provide a few discrete points of the TCF, which cover a narrow time interval range of nanoseconds. As a result, fast internal motions in the range of 100-1 ps cannot be detected. Since the spectral density function is normally measured only at a few frequencies (two to three), and hence the TCF cannot be obtained from experiments, it is customary to use * To whom correspondence should be addressed. E-mail: dais@ chemistry.uch.gr.
theoretical TCFs (models) for a quantitative interpretation of the experimental relaxation data. Although theoretical TCFs are not expected to reflect faithfully the complexity of the polymer motions, their applications, in the past 2 decades, offer a satisfactory approximation in describing the rates and mechanisms of these motions. In our laboratory, we have undertaken a systematic study of the structure and conformational dynamics of polysaccharides in solution in an attempt to understand the relationship between structure and functional properties of these biopolymers. In this respect, variable temperature and multifield 13C relaxation experiments have been performed for a series of linear polysaccharides, such as amylose, dextran, inulin, R(1f3)-, and β(1f3)-linked glucans,4,5 as well as for the branched polysaccharide scleroglucan.6 Several TCFs have been used to interpret the relaxation data of these polysaccharides and describe the various modes of reorientations of the complex carbohydrate chains. Relaxation experiments as a function of temperature and magnetic field strength allowed the discrimination of the various models used. This was made on the basis of the ability of the models to reproduce the experimental relaxation parameters (T1, T2, and NOE) corresponding to the minimum of the curve of the T1 values vs the reciprocal of temperature, 1/T (K-1), or the capability of a particular model to predict some independent relaxation data not previously used for adjusting the model parameters.5,7-9 Pullulan is formed from yeast-like Aureobasidium pullulans microorganisms. It is a linear polysaccharide comprising maltotriose units connected through R(1f6) glycosidic linkages. Its structure has been determined unambiguously
10.1021/bm010073q CCC: $20.00 © 2001 American Chemical Society Published on Web 09/07/2001
1138
Biomacromolecules, Vol. 2, No. 4, 2001
by employing proton and carbon-13 NMR spectroscopy.10,11 Since pullulan is water-soluble and biodegradable, it has attracted much attention for industrial applications. A variety of pullulan derivatives have been prepared to examine their potential applications in the food, cosmetics, pharmaceutical and electronics fields.12,13 The molecular characteristics of pullulan and the dimensions of its molecules have been investigated and discussed in a number of papers from an experimental14-19 and theoretical standpoint.20,21 The general perception is that pullulan behaves as a random-coil polymer in aqueous solutions presumably related to the greater conformational freedom inherent in the R(1f6) glycosidic linkages. In the this study, we present the variable-temperature 13C relaxation measurements of pullulan in dilute water and dimethyl sulfoxide solutions at three magnetic field strengths. We will show that the T1 and NOE data can be accounted for by the theoretical DLM model and provide a quantitative picture of the local motions in pullulan. In particular, we will examine the contributions of both R(1f6) and R(1f4) linkages to the conformational dynamics of pullulan and compare these motional characteristics with those of dextran, and amylose, which represent the two extremes in the conformational freedom and constraint. Experimental Section Materials. Pullulan was purchased from Sigma. Deuterated water (D2O) and dimethyl sulfoxide (DMSO-d6) solvents were purchased from Aldrich. The weight-average molecular weights of pullulan, Mw, was determined by light scattering in DMSO solutions and it was found to be 1.45 × 105. NMR Relaxation Measurements. 13C NMR relaxation measurements were conducted at three 13C Larmor frequencies, 75.4, 100.5, and 125.7 MHz on Bruker AC300, AM400, and AMX500 spectrometers, respectively, under broadband proton decoupling. The temperature was controlled to within (1° C by means of precalibrated thermocouples in the probe inserts. Spin-lattice relaxation times (T1) were measured by the standard IRFT method with a repetition time longer than 5 × T1. A total of 256-720 transients were accumulated, for a set of 8-12 arrayed t values, depending on the type of spectrometer and temperature. The 180° pulse width ranged from 27 to 29 µs in the three spectrometers. Values of T1 were determined by a three-parameter nonlinear procedure with a rms error of (8%, or better. The reproducibility of each T1 value was (5% or better. NOE experiments were carried out by the inverse gated decoupling technique. At least three experiments have been performed at each temperature value. Delays of at least 10 times the longest T1 were used between 90° pulses. NOE values were estimated to be accurate to within (15%. Undegassed samples of pullulan in D2O and DMSO-d6 solutions (5% w/v) were used for the 13C relaxation experiments. Measurements with degassed samples did not show any measurable difference in the 13C relaxation parameters relative to those of the undegassed samples in agreement with earlier reports5,6 in polysaccharide systems that show relaxation parameters on a millisecond time scale. Viscosity. Viscosity measurements were performed using Ubbelohde type dilution viscometers in a bath thermostated at 30 ( 0.01° C. Intrinsic viscosities, [η], and Huggins’ constants, k′,
Dais et al.
Figure 1. Structure of the repeat-unit of pullulan.
were calculated through eq 1 by plotting reduced viscosity ηsp/C vs C.
ηsp/C ) [η] + k′[η]2C
(1)
In this equation, ηsp is the specific viscosity. The intrinsic viscosities of pullulan in D2O and DMSO-d6 solutions were 0.429 and 0.721 dL g-1, respectively. Numerical Calculations. The relaxation data were analyzed by using the MOLDYN program22 modified by us to include the spectral density functions used in the present study. The parameters of a given model were optimized until the sum of the squares of deviations of the difference between theoretical and experimental relaxation data reached a minimum. Details of the program and calculation procedure by employing various models have been given elsewhere.8,9
Results and Discussion NMR Spectra of Pullulan. The structure of the present polysaccharide has been verified from its 1H and 13C NMR spectra.10,11 The intensities of the three signals of the anomeric protons of the present pullulan sample at δ 5.42, 5.38, and 4.98 were equal, indicating that the sample has a 1:2 ratio of R(1f6) and R(1f4) linkages.11 This means that pullulan comprises solely R(1f6)-linked maltotriose units (Figure 1). The 125.7 MHz 13C NMR spectrum of pullulan in D2O has been presented and fully analyzed in previous studies.12,11 All carbon resonances have been assigned except perhaps the C-5B and C-5C carbons of rings B and C, respectively, whose chemical shifts can be interchanged. The 13C NMR spectrum of pullulan provides direct evidence that the R(1f6) linkage occurs between every third glucose ring.11 The carbon spectrum in DMSO-d6 is similar to that observed in D2O. However, the chemical shifts of the carbon resonances in this solvent were affected by temperature. Fewer carbon resonances are resolved in the spectra at lower temperatures, even at 125.7 MHz, owing to peak overlaps. Therefore, the relaxation parameters of the overlapped carbon signals over the temperature range of relaxation measurements (303-393 K) are not taken into consideration in describing the dynamics of pullulan. 13C NMR Relaxation Data. The relaxation parameters NT1 (N is the number of directly attached protons to carbons) and NOE of the resolved endocyclic carbons for each of the rings A, B, and C of pullulan in both D2O and DMSO-d6 solvents, at each temperature and magnetic field strength,
Biomacromolecules, Vol. 2, No. 4, 2001 1139
Pullulan in Dilute Solutions
Table 1. 13C Spin-Lattice Relaxation Times (NT1, ms) and NOE Values of Ring Carbon Atoms of Pullulan in D2O Solutions As a Function of Temperature and Magnetic Field Strengths (Bo, MHz) ring A
ring B
ring C
T (K)/Bo
75.4
100.5
125.7
75.4
100.5
125.7
75.4
100.5
125.7
308
206 (1.68) 211 (1.73) 219 (1.81) 236 (1.90) 252 (1.96) 273 (2.06)
277 (1.59) 280 (1.68) 286 (1.73) 292 (1.81) 309 (1.88) 321 (1.99)
346 (1.54) 341 (1.63) 343 (1.67) 358 (1.76) 376 (1.79) 395 (1.93)
210 (1.66) 211 (1.75) 222 (1.87) 241 (1.92) 254 (1.99) 278 (2.14)
276 (1.61) 276 (1.67) 281 (1.77) 291 (1.82) 305 (1.92) 332 (2.00)
345 (1.53) 349 (1.57) 350 (1.64) 364 (1.72) 380 (1.83) 403 (1.92)
208 (1.65) 207 (1.73) 214 (1.81) 232 (1.94) 247 (2.00) 270 (2.10)
271 (1.59) 270 (1.68) 275 (1.72) 284 (1.80) 299 (1.88) 316 (1.99)
338 (1.53) 338 (1.62) 340 (1.70) 350 (1.77) 368 (1.82) 381 (1.96)
318 328 338 348 358
Table 2. 13C Spin-Lattice Relaxation Times (NT1, ms) and NOE Values of Ring Carbon Atoms of Pullulan in DMSO-d6 Solutions As a Function of Temperature and Magnetic Field Strengths (Bo, MHz) ring A
ring B
ring C
T (K)/Bo
75.4
100.5
125.7
75.4
100.5
125.7
75.4
100.5
125.7
303
225 (1.44) 221 (1.53) 205 (1.62) 207 (1.78) 223 (1.98) 249 (2.12)
324 (1.39) 298 (1.46) 276 (1.50) 272 (1.64) 277 (1.79) 298 (1.91)
399 (1.35) 378 (1.43) 335 (1.46) 333 (1.56) 337 (1.68) 369 (1.85)
250 (1.41) 229 (1.49) 211 (1.58) 205 (1.67) 210 (1.77) 230 (1.99)
356 (1.37) 329 (1.45) 288 (1.52) 268 (1.56) 268 (1.65) 282 (1.88)
479 (1.36) 436 (1.42) 353 (1.50) 342 (1.55) 328 (1.63) 345 (1.77)
250 (1.43) 229 (1.49) 215 (1.56) 202 (1.69) 210 (1.73) 230 (1.97)
365 (1.42) 337 (1.46) 293 (1.51) 277 (1.57) 272 (1.68) 282 (1.83)
494 (1.39) 437 (1.40) 369 (1.45) 351 (1.52) 341 (1.65) 352 (1.76)
313 333 353 373 393
were similar within experimental error. Therefore, average values will be used to describe the dynamics of the present polysaccharide in these solvents. These values are summarized in Tables 1 and 2 as a function of temperature and magnetic field strength for D2O and DMSO-d6 solvents, respectively. Inspection of Table 1 reveals that the relaxation parameters of rings A and B are similar within experimental error, indicating that these two rings are characterized by similar local dynamics. This is also shown graphically in Figure 2, which presents the temperature dependence of the average experimental NT1 values of the three rings at a magnetic field strength of 75.4 MHz in D2O solvent. Figure 3 illustrates the temperature dependence of the NT1 values of the endocyclic carbons of rings A, B, and C in DMSO-d6 at 100.5 MHz. Two types of differences are evident in Figures 2 and 3: (1) differences in the position of the NT1 minimum vs temperature and (2) differences in the value of NT1 at the minimum. The position of the NT1 minimum vs temperature reflects differences in the rate of the local segmental motions in different rings. The lower the temperature at which the NT1 minimum is observed for a monomer unit, the higher its rate of the backbone segmental motion. Figure 2 shows clearly that the temperature at which the NT1 minimum is observed in D2O solvent is the same for rings A and B (
