Biomacromolecules 2003, 4, 1013-1017
1013
Periodic Disorder along Ramie Cellulose Microfibrils Yoshiharu Nishiyama,*,† Ung-Jin Kim,† Dae-Young Kim,‡ Kyoko S. Katsumata,† Roland P. May,§ and Paul Langan| School of Agricultural and Life Science, The University of Tokyo, Tokyo 113-8657, Japan, College of Life Resources Science, Dongguk University, Seoul 100-715, Korea, Intitut Laue Langevin, BP 156, 38042 Grenoble Cedex 9, France, and Bioscience Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545 Received December 21, 2002; Revised Manuscript Received March 24, 2003
Small angle neutron scattering studies have been carried out on cellulose fibers from ramie and Populus maximowicii (cotton wood). Labile hydrogen atoms were replaced by deuterium atoms, in water-accessible disordered regions of the fibers, to increase the neutron scattering contrast between the disordered and crystalline regions. A meridional Bragg reflection, corresponding to a longitudinal periodicity of 150 nm, was observed when scattering collected from hydrogenated and deuterated dry ramie fibers was subtracted. No Bragg reflection was observed with the cotton wood fibers, probably because of lower orientation of the microfibrils in the cell wall. The ramie fibers were then subjected to electron microscopy, acid hydrolysis, gel permeation chromatography, and viscosity studies. The leveling off degree of polymerization (LODP) of the hydrolyzed samples matched exactly the periodicity observed in the diffraction studies. The weight loss related to the LODP was only about 1.5%, and thus, the microfibrils can be considered to have 4-5 disordered residues every 300 residues 1. Introduction Cellulose is biosynthesized by polymerization of glucose residues at the cell membrane by an ordered synthase complex,1 followed by assembly of the extended parallel chains into nanometer thick crystalline microfibrils2 often millimeters in length. In plant cell walls, these high tensile strength microfibrils are the fundamental structural unit. A matrix of other polysaccharides surrounds the microfibrils cross-linking them through hydrogen bonding. Direct observation by high-resolution microscopy of a continuous crystal lattice along individual microfibrils, first from highly crystalline algal cellulose3 and then from ramie cellulose,4 lead to the concept of single crystal microfibrils. Yet levelingoff degree of polymerization (LODP) behavior during acid hydrolysis of higher plant celluloses points to the presence of intrinsic disordered regions arranged at regular intervals along the length of the microfibrils: hydrolysis proceeds rapidly at first and then levels off at a chain length of 200300 residues.5-6 These observations raise a number of important questions. In particular, it is unclear whether the disordered regions in higher plant microfibrils are intrinsic or introduced during acid hydrolysis. If the former is true, a mechanism that introduces such defects at regular intervals during microfibril formation has to be found and also a reason for their presence in microfibrils from higher plant cell walls but not in the * To whom correspondence should be addressed. E-mail: nishiy@ sbp.fp.a.u-tokyo.ac.jp. † The University of Tokyo. ‡ Dongguk University. § Intitut Laue Langevin. | Los Alamos National Laboratory.
microfibrils from bacteria and algal plant cell walls. Direct evidence of the presence of intrinsic disorder will have important implications for our understanding of cellulose biogeneration and plant physiology. If the latter is true, then our understanding of the mechanism of acid hydrolysis will be challenged. In the currently accepted model, acid hydrolysis involves direct attack at the site of hydronium ioncatalyzed hydrolysis at the surface of the microfibrils.7 There have been several reports of periodic arrangements of crystalline and amorphous regions along synthetic polymer fibers.8 In many cases, meridional Bragg reflections resulting from longitudinal periodicity can be observed in Small-angle X-ray scattering (SAXS) experiments. Structural information such as the size of the crystalline regions, the difference in density between amorphous, and crystalline regions or the boundary between these regions can be studied by analyzing the scattering profile. However, the X-ray scattering density contrast between any crystalline and amorphous regions in cellulose is very small, and meridional Bragg reflections are observed in SAXS experiments on higher plant celluloses only after acid hydrolysis.9 In small angle neutron scattering (SANS) experiments, the neutron scattering density contrast between amorphous and crystalline regions is also small but can be significantly increased by using deuteration. In early SANS experiments by Fischer et al.,10 labile hydrogen atoms in disordered wateraccessible regions of regenerated cellulose fibers were replaced by deuterium atoms by vapor exchange with heavy water. Because of the large difference in scattering of neutrons by hydrogen and deuterium, this specific deuteration greatly altered the neutron scattering density in the amorphous regions. Meridional Bragg reflections corresponding
10.1021/bm025772x CCC: $25.00 © 2003 American Chemical Society Published on Web 05/13/2003
1014
Biomacromolecules, Vol. 4, No. 4, 2003
to longitudinal periodicities of 15-20 nm were observed. Although native higher plant cellulose (ramie) was also studied, the measured q range did not cover the length scale implied by LODP behavior in acid hydrolysis experiments. We report here SANS studies of periodic disorder in cellulose microfibrils from higher plants: ramie and Poplus maximowicii (cotton wood). The ramie samples were further studied by acid hydrolysis, scanning electron microscopy, and gel permeation chromatography. Our results represent direct evidence for small disordered regions that are periodically distributed along the cellulose microfibrils of higher plants, a result with fundamental implications for plant physiology. 2. Experimental Section Samples. Dewaxed ramie fibers were purchased from a textile dealer. The sugar composition was analyzed by the improved alditol-acetate method.11 The tension wood block was cut from cotton wood harvested in Hokkaido University Forest of the University of Tokyo. Small Angle Neutron Scattering. SANS measurements were carried out on samples of ramie fiber bundles and a tension wood block from cotton wood. In each case, hydrogenated (H) and deuterated (D) samples were prepared and measured in both a wet state and a dry state. All samples were sealed within a quartz capillary of 4 mm outer diameter and 1 mm wall thickness, with their fiber axes parallel to the capillary axis. The wet H or D samples were prepared by soaking in either normal or heavy water. Dry H and D samples were prepared from the wet samples by vacuumdrying at 40 °C. All SANS measurements were carried out on D1112,13 and D22 at the Institute Laue-Langevin, Grenoble. The q range covered extended from 0.002 Å-1 to 0.25 Å-1 corresponding to Bragg d-spacings of about 3-300 nm. Here, q is defined as (4π/λ) sin(θ), where θ is the Bragg angle and λ is the wavelength. A cadmium aperture of slit opening 1.5 mm wide and 20 mm high was placed before the capillary to avoid reflection from the capillary wall. Most measurements on D11 were carried out with a collimation length of 20 m and a sample to detector distance of 20 m. For each sample, the transmitted intensity was measured with the beam-stop removed and a beam-attenuator in place in order to protect the detector. The data were corrected for variations in detector response, background subtracted, and then visualized using the program GRASP developed by Charles Dewhurst at the ILL. Image rotation, scaling, and subtraction were accomplished using software written by the first author. Scanning Electron Microscopy. A single ramie fiber was swollen in water, and the secondary wall was partially delaminated with a pair of forceps. The fiber was dried at 105 °C for an hour and Pt coated with an ion sputter. A field-emission-type scanning electron microscope S-4000 (Hitachi) was operated at 6 kV acceleration voltage, and the secondary electron image was obtained as a 16 bit gray scale image. The image analysis was performed using the public domain program ImageJ, a program developed at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/ij/.
Nishiyama et al.
Figure 1. Small angle neutron scattering from dry D-ramie (left) and H-ramie (right). The fiber axis is vertical, and 1 pixel corresponds to 3.14 × 10-3 (nm-1) in q.
Acid Hydrolysis. Samples of ramie cellulose, 100 mg in mass, were treated with 4 N HCl at 80 °C for different periods of time, up to 7 h. At the end of the time period, the sample was filtered using a glass filter, rinsed with distilled water, and then collected using a 0.10 mm membrane. The weight loss was determined from the weight of the remaining solid substance after oven-drying at 105 °C. The sample was dissolved into a 1 M cuen solution at 0.4% (w/v) cellulose concentration and measured for viscosity using a CanonFenske type viscosimeter. The viscosity average degree of polymerization DP(v) was obtained from the intrinsic viscosity according to the following formula14 [η] ) 0.57 × DPv1.0 Gel Permeation Chromatography (GPC). The hydrolyzed ramie samples were converted to cellulose trinitrate by treating with a mixture of nitric acid and phosphorus pentoxide. Derivatized cellulose was dissolved in tetrahydrofuran and analyzed by size exclusion chromatography (TSK GMHXL and TSK G2500H6 connected in this order, each 300 × 7.5 mm i.d.). The elution was detected by UV adsorption at 254 nm. The degree of polymerization was obtained from the chromatogram by calibration using polystyrene standards, as described previsouly.15 3. Results and Discussion The dominant feature in the SANS diagrams from all samples is a central diamond shape of diffuse scattering, although the exact shape and strength of this feature varies for different samples and different states, i.e., wet, dry, deuterated, and hydrogenated. SANS diagrams of the dried H and D ramie samples are shown in Figure 1. The diffuse scattering covers a region that includes the LODP length scales observed in acid hydrolysis studies. Fischer et al. has pointed out that the diffuse diamond shape can be interpreted as being due to a dilute system of submicron voids.9 Scanning electron miscroscopy was used in order to investigate the presence of such voids in the ramie fibers. Figure 2 shows the fibrillar structure of the delaminated surface of a ramie fiber together with its Fourier transform image. Although the fibrils show some small wiggling, they are mainly wellaligned with no voids present that could lead to a diamond shape of diffuse scattering in the appropriate q range. In fact, the FFT image did not reproduce the diamond shape,
Periodic Disorder along Ramie Cellulose Microfibrils
Biomacromolecules, Vol. 4, No. 4, 2003 1015
Figure 3. Difference diagram of small angle neutron scattering from dry D-ramie and H-ramie in Figure 1 (D-ramie - 3 × H-ramie). Figure 2. Scanning electron micrograph of the peeled surface of ramie fiber (scale bar ) 1 micron) and its FFT image (bottom left, scale bar ) 0.1 nm-1). The squared region in the FFT corresponds to the range observed in Figure 1.
although the equatorial feature agrees with the observed SANS. In order for a model to explain the diffuse scattering, it must incorporate scattering edges that are almost perpendicular to the fiber axis. In the model proposed by Fischer et al., these edges come from laterally elongated diamond shaped voids. Kinks often observed in highly crystalline cellulose after harsh mechanical treatments could generate such types of voids. However, the fact that we do not observe these voids in ramie fibers indicates that something else must be responsible for the diffuse scattering observed in our studies. One possibility is a dilute system of regions where the packing density is either particularly high or low, with boundaries almost perpendicular to the fiber axis. The nature of this system is not clear at the moment. If these dilute regions are not related to the internal structure of the microfibrils, their contribution should be similar in both H and D ramie. Figure 3 shows the result of subtracting the scattering collected from dry H ramie, scaled by a factor of 3, from the scattering collected from dry D ramie. A Bragg reflection corresponding to a periodic interval along the microfibrils of 150 nm on the meridian and a strong streak on the equator can be clearly seen. The scale factor of 3.0 was the highest value that could be used without the difference diagram becoming negative and reflects the fact that deuterated cellulose has a higher scattering length density than hydrogenated cellulose. The cotton wood sample gave similar diamond shapes, but we did not obtain a Bragg peak in the difference diagram. Thus, we did not consider the cotton wood any further. The sugar composition analysis of the ramie cellulose samples gave a major component of 88% glucose with minor components of rhamnose, arabinose, xylose, mannose, and galactose each representing between 2 and 4% of the whole
Figure 4. Vertical trace of Figure 3 plotted together with gel permiation chromatogram of hydrolyzed ramie on the same length scale.
material. At various stages of acid hydrolysis, only glucose was present in the solubilized fraction, the minor sugars tending to remain as an insoluble fraction. The minor sugars seem to exists in hydrolysis resistant domains of hemicellulose quite distinct from the cellulose microfibrils. The GPC profile of the hydrolyzed ramie fibers is shown in Figure 4 together with the integrated meridional projection of the difference diagram in Figure 3. There is a striking agreement between the scattering intensity distribution and the molecular weight distribution, both maximum at 150 nm. The residual weight of the hydrolyized ramie fibers is shown in Figure 5, and the corresponding degree of polymerization is shown in Figure 6. Two degradation rates can be seen in the plot of Figure 5, the faster component ending after ∼ 1.5 h. Because the extrapolation of the residual weight of the faster component does not give 100%, there would appear to be a very fast component that disappears before 20 min. This component represents ∼1.5% of the total weight and
1016
Biomacromolecules, Vol. 4, No. 4, 2003
Nishiyama et al.
Figure 5. Weight loss during hydrolysis of ramie fibers with hydrochloric acid.
Figure 7. Guinier plot of the Bragg peak in Figure 3 in the lateral direction. The filled circles are observed intensity with error bars, and the thick straight line is a least-squares fit of a function y ) a exp(-b2x) where b was evaluated to be 12.9 ( 0.3.
a binary system of scattering length density F1 and F2 with one component having length x at interval l the structure factor for the meridional Bragg spot reduces to F)
Figure 6. Change in viscosity average degree of polymerization of the residual during hydrolysis.
should be responsible for the LODP behavior as the LODP is already achieved within 20 min or less in Figure 6. Thus, there are only 4-5 residues every 300 that are misaligned in the case of ramie fibers. DP ) 300 corresponds to a length of 300 × 0.5 nm. After the initial rapid degradation, there would then appear to be two distinct linear degradations rates, possibly because of a difference in susceptibility of the smooth face and edges of the microfibrils. Helbert et al. report that the edges of the parallelogram shaped crosssection of microfibrils from tunicate become rounded after hydrolysis.16 In Figure 1, about 2.0 × 108 neutrons passed through the sample during the measurement, and the integrated intensity of the Bragg peak was about 450 counts. The expected intensity of this peak can be calculated from the macroscopic cross section as follows: The total differential scattering cross section of a system, with sample volume V and a repeating lattice of unit cell volume V0, is given by d
∑
dΩ
(2π)3 N0 | bj exp(iq b‚b r j)|2 V0N0 j
δΩ ) V
∑
(1)
where N0 is the number of nuclei per unit cell, bj is the scattering length of an individual atom, and rj is its position, with the summation, j, over all atoms in the unit cell. In this particular case, the sample can be thought of as consisting of microfibrils of a square cross section a2 that have a repeating structure along their length with periodicity l. For
∫
a2(F2 - F1) exp(iq b‚b) r dr )
a2(F2 - F1) qx 2 sin q 2
( )
(2)
where, q ) 2π/l (λ ) 2lsin θ). Thus, the integrated intensity of the Bragg spot is given by I ) I0t
∑
2 d (2π)3 a (F2 - F1) πx 2 δΩ ) I0tV 2sin | | dΩ V 0N 0 2π l
( )
(2π)3 x |(F2 - F1) |2 N l
) I0Vt
()
(3)
when x/l , 1, where t is the thickness of the sample, I0 is the incident neutron beam intensity, and N is the number of nuclei per unit volume. Substituting the appropriate values in eq 3 (I ) 450 counts; I0 ) 2.0 × 108 counts; t ) 0.1 cm; V ) 4.0 × 10-2 cm3; N ) 21 × 5.948 × 1021 (number/ cm3); (F1 - F2)2 ) 3.0 × 10-20 cm-4) results in a value for x/l of 0.016. This indicates that the size of the deuteriumexchanged regions along the microfibril direction is about 1.6% of the periodicity of these domains. Because the surface of microfibrils may also be deuterated, (F1 - F2)2 is most likely over-evaluated; therefore, the length of deuterium exchanged domains will probably be between 1.6 and 2% of the periodicity. This result agrees quite well with the value obtained from the acid hydrolysis studies. The lateral dimension of the disordered region was evaluated from a Guinier plot of the Bragg peak in the lateral direction,10 Figure 7. The resulting value of 12.9 nm is about twice the width of the individual microfibrils in the ramie fibers. One possible explanation for this is that the disordered regions occur at points where microfibils come into contact with each other. These results indicate that disordered regions are periodically arranged in ramie cellulose microfibrils. Ramie is one of a few plants in which the cellulose microfibrils are almost
Periodic Disorder along Ramie Cellulose Microfibrils
perfectly aligned along the plant fibers. Most higher plant cell walls have more complicated structures, and diffraction from periodic disorder will be buried in signals arising from the microfibril arrangement. Because lateral features are dominant even in the ramie difference diagram, a small amount of microfibrils oriented in the lateral direction will bury the signal in the meridian direction even if most of the microfibrils are well aligned parallel to the long axis as in the tension wood of cotton wood. We assume that this is the reason that no Bragg reflection from longitudinal periodicities was observed in SANS studies of cotton wood. However, similar LODP behavior in higher plant celluloses points to a fundamental phenomenon in higher plants, although further neutron scattering studies on a greater range of samples will be required in order to confirm this beyond doubt. A typical LODP behavior gives a relatively narrow molecular weight distribution after hydrolysis, but bacterial and algal cellulose rather give broader molecular weight distributions after hydrolysis.17 The absence of regular break points in bacterial and algal cellulose suggests that this phenomenon may have been introduced in higher plant cellulose biogenesis through an evolutionary process, although the reason for its introduction is unclear. Disordered cellulose is more susceptible to chemical and enzymatic attack than crystalline cellulose. Disordered regions can also act as anchoring points for matrix substances. However, recent studies of in vitro synthesis of cellulose by plant free cell extracts suggests another possible explanation.18 Whereas cellulose produced at the cell wall of rubus fruticosus (blackberry) does show LODP behavior, cellulose produced in vitro, without the constraints of a cell wall, does not. Although there are several possible explanations for this, one is that the in vitro microfibrils are able to grow independently in a neighbor-free environment, whereas microfibrils growing in the parent cell walls have to push their way through the existing cell wall, leading to a number of structural defects.
Biomacromolecules, Vol. 4, No. 4, 2003 1017
In our experiment, we did not observe any peak corresponding to a 6-7 nm repeat reported recently on flax19 even in a dry state. Acknowledgment. The authors thank the Institut Laue Langevin for the provision of beamtime. P.L. thanks the office of Science and the Office of Biological and Environmental Research of the U.S. Department of Energy for financial support. References and Notes (1) Itoh, T.; Kimura, S. J. Plant Res. 2001, 114, 483-489. (2) Brown, R. M., Jr. J. Macromol. Sci.sPure Appl. Chem. 1996, A33, 1345-1373. (3) Sugiyama, J.; Harada, H.; Fujiyoshi, Y.; Uyeda, N. Planta 1985, 166 (2), 161-168. (4) Kuga, S.; Brown, R. M., Jr. Polym. Commun. 1987, 28, 311-314. (5) Nickerson, R. F.; Habrle, J. A. Ind. Eng. Chem. 1947, 39, 15071512. (6) Nelson, M. L.; Tripp, V. W. J. Polym. Sci. 1953, 10, 577-586. (7) Grohmann, K.; Torget, R.; Himmel, M. Biotechnol. Bioeng. Symp. 1986, 15, 59-80. (8) Alexander, L. E. X-ray Diffraction Methods in Polymer Science; Robert E. Krieger Pub.: New York, 1979; pp 319-356. (9) Takahashi, M.; Takenaka, H. Sen’I Gakkaishi 1979, 35, T99-T104. (10) Fischer, E. W.; Herchenroder, P.; Manley, R. St. J.; Stamm, M. Macromolecules 1978, 11, 213-217. (11) Blankeney, A. B.; Harris, P. J.; Henry, R. J.; Stone, B. A. Carbohydr. Res. 1983, 113, 291-299. (12) Ibel, K. J. Appl. Crystallogr. 1976, 9, 296-309. (13) Lindner, P.; May, R. P.; Timmins, P. A. Physica B 1992, 180-181, 967-972. (14) Sihtola, H.; Kyrklund, B.; Laamanen, L.; Palenius, I. Paperi ja puu 1963, 45, 225-232. (15) Shibazaki, H.; Kuga, S.; Okano, T. Cellulose 1997, 4, 75-87. (16) Helbert, W.; Nishiyama, Y.; Okano, T.; Sugiyama, J. J. Struct. Biol. 1998, 124, 42-50. (17) Kuga, S.; Mutoh, H.; Isogai, A.; Usuda, M.; Brown, R. M., Jr. in Cellulose, structural and functional aspects; Kennedy, J. F., Phillips, G. O., Williams, P. A., Eds.; Ellis Horwood: Chichester, U.K., 1989; pp 81-86. (18) Lai-Kee-Him, J.; Chanzy, H.; Muller, M.; Putaux, J.-L.; Imai, T.; Bulone, V. J. Biol. Chem. 2002, 277, 36931-36939. (19) Astley, O. M.; Donald, A. M. Biomacromolecules 2001, 2, 672-680.
BM025772X
