Plant Cell Physiol. 39(7): 711-720 (1998)
JSPP © 1998
Crystalline Cellulose in Hydrated Primary Cell Walls of Three
Monocotyledons and One Dicotyledon
Bronwen G. Smith1-4, Philip J. Harris1-5, Laurence D. Melton 2 4 and Roger H. Newman 3
' School of Biological Sciences, The University of Auckland, Private Bag 920J 9, Auckland, New Zealand
Food Science, University of Otago, Dunedin, New Zealand
3
Industrial Research Limited, P.O. Box 31-310, Lower Hutt, New Zealand
2
The molecular ordering of cellulose, including its
crystallinity, in the unlignified primary cell walls of three
monocotyledons (Italian ryegrass, pineapple, and onion)
and one dicotyledon (cabbage) was characterized by solidstate U C NMR spectroscopy. These species were chosen because their primary cell walls have different non-cellulosic
polysaccharides and this may affect the molecular ordering
of cellulose. Values of the proton rotating-frame relaxation
[T1/((H)] and spin-spin relaxation [T2(H)] time constants
showed that the cellulose in the cell walls of all four species
was in a crystalline rather than an amorphous state. Furthermore, a resolution enhancement procedure showed
that the triclinic ( U and the monoclinic i\p) crystal forms
of cellulose were present in similar proportions in these cell
walls. However, the calculated cross-sectional dimensions
of the cellulose crystallites varied among the cell walls (in
the range 2-3 nm): the largest were in the Italian ryegrass,
the smallest were in the onion and cabbage, and those of intermediate size were in the pineapple. The crystallite dimensions may thus be affected by the non-cellulosic polysaccharide compositions of the cell walls.
Key words: Cellulose crystallites — Cell-wall polysaccharides — Monocotyledons — Non-cellulosic polysaccharides — Primary cell walls — Solid-state CP/MAS 13C
NMR spectroscopy.
Primary cell walls of angiosperms are composed of cellulose microfibrils embedded in a matrix phase consisting
mostly of polysaccharides which have a variety of structures (Bacic et al. 1988, Carpita and Gibeaut 1993). Although there have been numerous studies using electron microscopy to determine the orientation of the cellulose
microfibrils in the cell walls (e.g. Roland et al. 1975, Neville
and Levy 1984), relatively little is known about the molecu-
lar ordering of the cellulose, including its crystallinity. Xray and electron diffraction have been used to study this,
but the techniques are sensitive only to long-range order
and thus provide information only about the repetition of
cellulose molecules at regular intervals through a crystallite. For this reason, most published X-ray and electron
diffraction studies of cellulose have focused on sources in
which the cross-sectional dimensions of the cellulose microfibrils are relatively large, for example in the cell walls of
the alga Valonia (Revol 1982, Sugiyama et al. 1990, 1994),
although two exceptions were studies of the primary cell
walls of cotton fibres and suspension cultured rose cells
(Chanzy et al. 1978, 1979).
Another technique for examining the molecular ordering of cellulose in primary cell walls is solid-state carbon-13
nuclear magnetic resonance (NMR) spectroscopy, the principles of which have been reviewed by Harris (1986, 1993).
The most commonly used type of this technique, using
cross-polarisation, magic-angle spinning (CP/MAS 13C
NMR), is sensitive to short-range order and can distinguish
between a cellulose molecule that is exposed on a surface
and one that is surrounded by similar molecules (Newman
and Hemmingson 1995). CP/MAS I3C NMR is thus a promising tool for studying the ordering of polysaccharides in
primary plant cell walls. In particular, the technique can be
used to study the molecular ordering of cellulose in situ in
whole, hydrated, primary cell walls (Jarvis 1990, Jarvis and
Apperley 1990, Newman et al. 1994, 1996). Being able to examine molecular ordering in intact cell walls is an advantage because methods which involve the chemical separation of cellulose from non-cellulosic polymers can alter the
nature of the cellulose crystallinity (Martel and Taylor
1993, Hemmingson and Newman 1995).
CP/MAS I3C NMR spectroscopy has been used to examine the molecular ordering of cellulose in primary cellwall preparations from the fruit of apple (Newman et al.
1994) and the leaves of Arabidopsis thaliana (Newman et
al. 1996), both dicotyledons. In the latter study, five nuclear spin relaxation procedures were used and it was found
that proton rotating-frame relaxation, which has the time
constant T(/,(H), provided the clearest distinction between
cellulose and other wall components for the purposes
of editing CP/MAS 13C NMR spectra. The CP/MAS 13C
NMR spectrum was separated into subspectra containing
signals assigned to crystalline cellulose [long T1;)(H)] and
Abbreviations: CP/MAS 13C NMR, cross polarization/magic
angle spinning carbon-13 nuclear magnetic resonance; PSRE, proton spin relaxation editing; Tlp(H), proton rotating-frame relaxation time constant; T2(H), spin-spin relaxation time constant.
4
Present address: Food Science Postgraduate Programme,
Department of Chemistry, The University of Auckland, Private
Bag 92019, Auckland, New Zealand.
5
Corresponding author.
711
712
Crystalline cellulose in primary cell walls
disordered substances [short T ^ H ) ] . The percentage of cellulose molecules in the interior and on the surface of the cellulose crystallite was then determined from the relative
strengths of C-4 peaks at 89 (crystallite interior) and 84
ppm (crystallite surface) in the subspectrum associated
with relatively long values of T,^(H). The cross-sectional diameter of the cellulose crystallite can be estimated from
these percentages. Any disordered (noncrystalline) cellulose would have been indicated by NMR signals in the subspectrum associated with relatively short values of Tip(H)
(Newman et al. 1993, Newman and Hemmingson 1995).
However, Newman et al. (1994, 1996) found that in the primary cell walls of both apple and A. thaliana the cellulose
was apparently entirely crystalline, and composed of small
crystallites with a cross-sectional diameter of about 3 nm.
As indicated above, we assign the C-4 peaks of cellulose at 89 and 84 ppm to cellulose molecules in the crystallite interior and to cellulose molecules on the crystallite
surface respectively. However, the peak at 84 ppm was
originally assigned either to "C-4 carbons which are in
amorphous regions regions of the polymer" or to "glucose
units on the surface of the elementary fibrils" (Earl and
VanderHart 1980). Since this early study, further evidence
for the latter interpretation has been obtained. First, Earl
and VanderHart (1981) compared celluloses from different
sources and demonstrated a qualitative agreement between
the intensity at 84 ppm and the percentage of structural
units exposed on surfaces. Second, it was found that this
peak could be split into two components at 83.9 and 85
ppm by resolution enhancement (Newman 1994). This is
consistent with nonequivalent crystallographic sites being
exposed on surfaces, but it is difficult to explain in terms of
an amorphous state. More recently, it was shown that it is
possible to achieve at least partial resolution of these two
component signals without resorting to resolution enhancement (Newman 1998). Third, when cellulose was subjected
to mechanical damage, the strengths of the 84 and 89 ppm
peaks decreased in a correlated manner as would be expected for a conversion of crystalline to amorphous cellulose (Wormald et al. 1996). Fourth, it was found that the
peaks at 84 and 89 ppm shared the same value of the proton rotating-frame relaxation time constant T,P(H), as
would be expected for proton spin diffusion between adjacent cellulose molecules (Newman and Hemmingson 1995,
Newman et al. 1996). Fifth, a two dimensional NMR experiment has demonstrated 13C-13C spin exchange linking the
peaks at 84 and 89 ppm (Bardet et al. 1997). This would not
be possible if the structures were separated by distances
> 1 nm, but it is consistent with spin exchange between surface molecules and adjacent molecules within the interior.
CP/MAS I3C NMR can also provide information on
the crystalline form of cellulose in intact cell walls. Crystalline cellulose in plant cell walls is believed to be a composite
of two crystalline forms designated triclinic (Ia) and mono-
clinic (1^) (VanderHart and Atalla 1984, Sugiyama et al.
1991). Distinguishing between these two forms (triclinic
and monoclinic) by CP/MAS I3C NMR (VanderHart and
Atalla 1984) has been improved by resolution enhancement
methods (Cael et al. 1985, Belton et al. 1989, Debzi et al.
1991, Newman 1994, Newman et al. 1994) and by using
principal component analysis (Lennholm et al. 1994). The
two forms (triclinic and monoclinic) can also be distinguished using Fourier transform infrared (FT-IR) spectroscopy (Debzi et al. 1991, Sugiyama et al. 1991, Kataoka and
Kondo 1996). By using CP/MAS 13C NMR with resolution
enhancement, it was shown that the primary cell walls of
apples and A. thaliana contained a mixture of both triclinic
(IJ and monoclinic (1^) in approximately equal proportions
(Newman et al. 1994, 1996). It has also been shown recently
that in the cellulose produced by the bacterium Acetobacter xylinum the proportions of the triclinic (Io) and monoclinic (Iy,) forms can be altered by the addition to the incubation medium of polysaccharides, such as xyloglucans
and carboxymethylcelluloses, as the cellulose is being synthesised (Hackney et al. 1994, Yamamoto and Horii
1994, Whitney et al. 1995). In similar experiments with
A. xylinum, it was also shown using X-ray diffraction that
such polysaccharides modified cellulose aggregation (Atalla
et al. 1993).
Although CP/MAS 13C NMR spectroscopy has been
used to examine the molecular ordering of cellulose in the
primary cell walls of dicotyledons (Newman et al. 1994,
1996), it has not been used to examine the molecular ordering of cellulose in the primary cell walls of monocotyledons. In the present study, we examined the molecular
ordering of cellulose in the primary cell walls of three species of monocotyledons and, for comparison, one species
of dicotyledon. If the molecular ordering of cellulose synthesised by Acetobacter xylinum can be influenced by polysaccharides added to the incubation medium, the molecular ordering of cellulose in the primary cell walls of plants
may be influenced during synthesis by the types of noncellulosic polysaccharides present. In the present study,
we tested this hypothesis by examining the molecular ordering of cellulose in primary cell walls with different noncellulosic polysaccharide compositions. Primary cell walls
of dicotyledons and the monocotyledon family Poaceae
(grasses and cereals) have quite different non-cellulosic
polysaccharide compositions (Bacic et al. 1988, Carpita
and Gibeaut 1993). Primary cell walls of dicotyledons contain pectic polysaccharides (including rhamnogalacturonans, arabinans and galactans) and xyloglucans, whereas
primary cell walls of the Poaceae contain glucuronoarabinoxylans, variable amounts of (1 —> 3, 1 -* 4)-/?-D-glucans,
and small amounts of xyloglucans and pectic polysaccharides. However, not all monocotyledons have primary cell
walls with non-cellulosic polysaccharide compositions like
that of the Poaceae: some have compositions similar to
Crystalline cellulose in primary cell walls
those of dicotyledons, and others have compositions intermediate between those of the Poaceae and dicotyledons
(Harris et al. 1997).
In the present study we used CP/MAS "C NMR spectroscopy to compare the molecular ordering of cellulose in
the primary cell walls of three species of monocotyledons:
Italian ryegrass (Lolium multiflorutri) (family Poaceae),
pineapple {Ananas comosus) (family Bromeliaceae), and
onion (A Ilium cepa) (family Alliaceae). We also included
the primary cell walls of the dicotyledon cabbage (Brassica
oleracea var. capitatd) (family Brassicaceae) for comparison. The non-cellulosic polysaccharide composition of
the primary cell walls of each of these species is known
in detail. Italian ryegrass cell walls contain large amounts
of glucuronoarabinoxylans, and small amounts of (1 -* 3,
1 —* 4)-/?-D-glucans, xyloglucans and pectic polysaccharides (Chesson et al. 1985). Pineapple cell walls also contain large amounts of glucuronoarabinoxylans and small
amounts of pectic polysaccharides, but contain more xyloglucans and no (1 -> 3, 1 —• 4)-/?-D-glucans (Smith and Harris 1995); their composition is intermediate between that of
Italian ryegrass and dicotyledons. Onion cell walls contain
large amounts of pectic rhamnogalacturonans and galactans, and smaller amounts of xyloglucans; their composition is similar to that of dicotyledons (Mankarios et al.
1980, Redgwell and Selvendran 1986, Ryden et al. 1989,
Ohsumi and Hayashi 1994). Cabbage cell walls are similar
in composition to onion cell walls except arabinans rather
than galactans are the predominant neutral pectic polysaccharides (Stevens and Selvendran 1980, 1984a, b).
In the present study, we report in particular on the proportion of cellulose that is in a crystalline form, the crosssectional dimensions of the cellulose crystallites, and the
proportions of the triclinic (Ia) and monoclinic (lp) forms
of cellulose. The cell-wall preparations used were hydrated
to ensure good resolution of signals from cellulose crystal
surfaces (Newman et al. 1994) and to maintain a nearnative state of the isolated cell walls.
Materials and Methods
Plant material—Italian ryegrass [Lolium multiflorum Lam.
cv. Grasslands Paroa] plants were grown in potting mix at the
University of Auckland from seed obtained from AgResearch
Grasslands, Palmerston North, New Zealand. Pineapple [Ananas
comosus (L.) Merr. cv. Smooth Cayenne] fruit and onion [Allium
cepa L. brown skinned early variety] bulbs were purchased from a
store in Auckland. Cabbages [Brassica oleracea L. var. capitata
cv. Winter Cross] were kindly donated by Mr. M. Kelderman,
Auckland, New Zealand.
Fresh sections were cut from the plant material of each species and examined by bright-field microscopy after treatment with
phloroglucinol-HCl which reacts with lignin to give a red colour
(Harris et al. 1980). On the basis of this colour reaction, tissues
which contained predominantly cells with unlignified primary
walls were selected for cell-wall isolation as follows. Parenchyma
713
cells with unlignified walls were obtained from thefleshof pineapple fruit (as described by Smith and Harris 1995) and from the
pith in the stem apex of the enlarged terminal bud of the cabbage.
Parenchyma cells with unlignified walls were obtained from the
scales of onion bulbs after first removing the epidermis and underlying vascular bundles from the convex surface and monolayer
of cells on the concave surface (McCann et al. 1990). Fully expanded leaf blades were cut from Italian ryegrass plants at the stage of
growth when the third leaf had just emerged. These leaves contained predominantly cells with unlignified walls (mostly epidermal
and mesophyll cells), but small numbers of tracheary elements
with lignified walls were also present. In all of the plants, the cells
used for the isolation of cell walls were assumed to be fully expanded.
Isolation of cell walls—The plant tissue was chopped into 3
mm3 pieces (pineapple and cabbage) or 3 mm2 pieces (onion and
Italian ryegrass) and immediately frozen in liquid nitrogen. The
frozen tissue was ground using a mortar and pestle to a fine
powder and 50 g of this was added to a phenol-HEPES solution
(100 ml) prepared by adding phenol (80 g) to 0.5 M HEPES-KOH
buffer pH 6.5 (20 ml). The cold mixture was homogenised using
a Polytron homogeniser (Kinematica GmbH, Kriens-Luzern,
Switzerland) set on full power for 8-11 min (for periods of 1 min)
to form a thick slurry. This was centrifuged (4,600 xg, 10 min at
20°C), the supernatant discarded, and the pellet resuspended in
80% (v/v) ethanol (40 ml) and centrifuged as before. The pellet
was washed by centrifugation three more times in 80% ethanol
(40 ml), then resuspended in 80% ethanol and stored at 4°C. Aliquots were removed for estimation of the dry weight. To ensure
that cytoplasmic protein was removed from the cell-wall preparations, breakage of the cells was monitored using bright-field microscopy after staining with Ponceau 2R which stains cytoplasmic
protein red (Harris 1983). No starch was detected histochemically
in any of the cell-wall preparations using I2 in KI (Jensen 1962).
Cell walls were also isolated from Italian ryegrass leaf blades
by mechanically breaking the cells in a buffer solution and
washing the cell contents away using the method of Smith and Harris (1995). After washing with water, the cell walls were then resuspended in 80% ethanol, centrifuged (500 x g, 5 min) and the supernatant removed. The pellet was then washed twice more by
resuspending in 80% ethanol and centrifuging as before. The cell
walls were then resuspended in 80% ethanol and stored as above.
An aliquot was removed for estimation of the dry weight.
CP/MAS "C NMR spectroscopy—A portion of each cellwall preparation was filtered from 80% ethanol and air-dried for
2 h at 20° C. The cell walls were then packed in a 7 mm diameter cylindrical sapphire rotor and retained with Kel-F end caps (Thermatech Engineering Corporation, Anaheim, CA). Moisture contents, determined by oven drying after the NMR experiments had
been completed, were all in the range 31% to 43% by weight.
Earlier experiments on apple cell walls (Newman et al. 1994) had
shown that moisture was essential for good resolution of signals
from cellulose crystal surfaces. The rotor was spun at frequencies
between 2.6 and 4.0 kHz in a Doty Scientific magic-angle spinning
probe for CP/MAS 13C NMR spectroscopy (Doty Scientific, Columbia, SC) at 50.3 MHz on a Varian XL-200 spectrometer
(Varian, Palo Alto, CA).
Spectra were obtained with a cross-polarization pulse sequence involving a 6 /is proton preparation pulse, a 1 ms contact
time, 30 ms of data acquisition and a recovery delay of 0.6 s
before the sequence was repeated. The proton decoupler field
strength corresponded to a rotating-frame precession frequency in
the range 53 to 58 kHz during data acquisition. Resolution-en-
714
Crystalline cellulose in primary cell walls
Table 1 Combinations of coefficients and corresponding time constants used in PSRE experiments (subspectra A and B)
Cell-wall
preparation
Combinations of
coefficients
Proton rotating-frame
time constant
(ms)
Italian ryegrass
A=-1.5S+4.8S
B=2.5S-4.8S'
A = - 1 . 1 S + 3.4S
B=2.1S-3.4S'
6.3
3.5
Onion
A = - 0 . 9 S + 3.2S
B = 1.9S-3.2S'
7.8
3.2
Cabbage
A = - 1 . 1 S + 3.5S
B=2.1S-3.5S'
8.4
3.6
Pineapple
hanced spectra were obtained by convoluting NMR free induction
decays with a function of the form (Newman et al. 1994, 1996):
F(t) = exp {(t/T x ) 2 /2 - (t/T Y ) 3 /3}
This function was obtained by incrementing the exponents in
a function originally designed for use in solution NMR spectroscopy (Ferrige and Linden 1978), but more recently used in solid-state
NMR studies of cellulose (Cael et al. 1985, Belton et al. 1989, Debzi et al. 1991). Incrementing the exponent in the first term generates a function that counteracts the Gaussian line broadening
commonly encountered in solid-state NMR, and incrementing the
exponent in the second term ensures that noise in the tail of the
free-induction decay is suppressed. Values of T x =6.0ms and
TY = 9.5 ms were chosen by trial and error.
A different procedure was used to distinguish between
domains with different molecular rigidities. In this procedure,
differences in the proton relaxation time constants [T^H)]
between different domains were exploited to generate proton
spin relaxation edited (PSRE) CP/MAS 13C NMR subspectra
(Newman and Hemmingson 1990, 1995, Newman 1992, Newman
et al. 1993, 1994, 1996). In the present study, normal and delayed
contact spectra, designated S and S', were acquired as interleaved
blocks of 64 transient signals, with proton spin locking during
a delay t = 4 m s for S' in the delayed-contact experiment. Preliminary values of T,,(H) were used to calculate initial values
of k and k', which were then refined to eliminate a signal at 89
ppm (assigned to crystalline cellulose) from subspectrum B and to
suppress signals associated with non-cellulosic substances from
subspectrum A. The calculations were then reversed to obtain improved estimates of T1()(H) for crystalline cellulose and for non-cellulosic substances (Table 1).
Values of the proton spin-spin relaxation constant T2(H) were
measured using a pulse sequence in which the proton decoupler
output was gated off for a period t between the proton-preparation pulse and the cross-polarization contact time (Alia and Lippmaa 1976). Heights were measured for two peaks assigned to cellulose (84 and 89 ppm) for values / in the range 0 to 12,us, and the
relaxation curves were interpreted as Gaussian functions.
Results
CP/MAS
I3
C NMR spectroscopy—The CP/MAS 13C
8.9
3.7
NMR spectra (total I3C NMR spectra; Fig. 1-4) obtained
from the primary cell-wall preparations of the three monocotyledons (Italian ryegrass, pineapple, and onion) and the
one dicotyledon (cabbage) are dominated by signals assigned to cellulose. Other, weaker signals are associated with
the non-cellulosic polysaccharides of the primary cell wall.
They are discussed here only in the context of possible
overlap with signals associated with cellulose. The cell-wall
preparations of Italian ryegrass obtained by the two different methods gave similar spectra. The results shown were
obtained with the cell-wall preparation obtained by the second method.
Crystalline and noncrystalline cellulose—A distinction
between ordered (rigid) and disordered (mobile) domains
of the wall was achieved by exploiting differences in proton
spin relaxation parameters [T,,(H)] (Table 1). The PSRE
subspectra show signals associated with relatively long (subspectrum A) (Fig. 1-4) and relatively short (subspectrum B)
(Fig. 1-4) values of Tlp(H) from polysaccharides contained
in ordered and disordered regions of the cell wall respectively. These subspectra (A and B) were obtained by identifying peaks associated with ordered (A) or alternatively
disordered (B) material and subtracting them each separately from the total CP/MAS 13C NMR spectrum using the
mathematical process described in Newman et al. (1996).
Proton spin diffusion mixes spin information over
dimensions of nanometres during a T^(H) experiment
(Zumbulyadis 1983), so all carbon atoms in the basic glucose structural unit must have the same apparent value of
T lp (H). It therefore follows that if the C-4 signal is excluded from a subspectrum, signals assigned to the other five
carbons must also be excluded. Signals assigned to crystalline cellulose were excluded from subspectrum B (Fig. 1-4)
by the editing process, i.e. by estimating values of T1/;(H) associated with the signal at 89 ppm (C-4 in crystal interiors),
calculating the consequent values of the spin relaxation factors and using these values to calculate the coefficients used
Crystalline cellulose in primary cell walls
715
Onion
Italian ryegrass
C-2.3.5
•Total
200
180
ISO
140
120
100
80
60
40
20
0
200
180
160
140
PPM from TMS
120
100
80
40
PPM from TMS
Fig. 1 CP/MAS 13C NMR spectra of the primary cell-wall preparation of Italian ryegrass. Total=Total CP/MAS 13C NMR; A =
Subspectrum A which corresponds to an ordered domain in which
the values of Ti,(H) are relatively long, compared to the values associated with subspectrum B; B=Subspectrum B which corresponds to a disordered domain in which values of Tlp(H) are relatively short, compared to the values associated with subspectrum
A. Labels refer to carbon atoms for glucosyl units of cellulose.
in the editing process (Table 1).
Exclusion of the signal at 89 ppm from subspectrum B
(Fig. 1-4) also resulted in exclusion of the signal at 105 ppm
in the spectra of the cell walls of Italian ryegrass, pineapple, and cabbage (Fig. 1, 2, 4). A signal at 105 ppm can be
assigned to the C-l of cellulose, whether crystalline or
disordered. If disordered (noncrystalline) cellulose had
been present, we would expect it to be associated with a
relatively short value of T lp (H). A signal at 105 ppm could
also be assigned to the C-l of the galactose residues of galactans or to C-l of galactose residues of xyloglucans. Pectic galactans in particular are known to occur in large
amounts in the cell walls of onion. Thus, in subspectrum B
Fig. 3 CP/MAS 13C NMR spectra of the primary cell-wall preparation of onion. For explanations of Total, A and B see legend to
Figure 1.
of onion cell walls (Fig. 3) the peak at 105 ppm could arise
from galactans, xyloglucans or possibly disordered cellulose or a mixture of these. In contrast, in subspectra B
(Fig. 1, 2, 4) of the cell walls of Italian ryegrass, pineapple,
and cabbage there is no peak at 105 ppm. This suggests,
that if the peak at 105 ppm in the spectrum of onion cell
walls is associated with xyloglucans or galactans then a
peak at 105 ppm might also be expected in the spectrum of
cabbage cell walls, which contain large amounts of xyloglucans (Stevens and Selvendran 1984b). This was not observed. We conclude that although the cell walls of Italian
ryegrass, pineapple, and cabbage are free from detectable
levels of disordered cellulose, small amounts may be
present in the cell walls of onion.
Additional evidence for crystalline cellulose was obtained from proton spin-spin relaxation values [T2(H)].
Cabbage
Pineapple
J C-2,3,5
200
180
160
140
120
100
80
20
PPM from TMS
Fig. 2 CP/MAS I3C NMR spectra of the primary cell-wall preparation of pineapple. For explanations of Total, A and B see legend
to Figure 1.
200
180
160
140
120
100
80
60
PPM from TMS
Fig. 4 CP/MAS 1JC NMR spectra of the primary cell-wall preparation of cabbage. For explanations of Total, A and B see legend
to Figure 1.
716
Crystalline cellulose in primary cell walls
Table 2 Values of the proton spin-spin relaxation time constant T2(H) for signals at 89
and 84 ppm
Cell-wall
preparation
Proton spin-spin relaxation time constant [T2(H)]
89 ppm
84 ppm
Italian ryegrass
8.5
9.3
Pineapple
8.7
10.0
Onion
9.3
9.3
Cabbage
9.0
10.2
The T2 relaxation processes are different from Tlp(H) in
that short values of T2 indicate rigidity rather than mobility. The values of T2(H) (Table 2) for the peak at 89 ppm are
from 8.5 to 9.0 /is and are similar to values for crystalline
cellulose, e.g. a value of T2(H) = 8.8±0.3 /us was found for
Avicel microcrystalline cellulose (Newman et al. 1996).
These results are further evidence for a rigid, crystalline
lattice in the cellulose of the four species examined. The
values for the peak at 84 ppm (resolved into two components under resolution enhancement) are similar to, or
slightly longer than, those for the peak at 89 ppm and are
assigned to crystallite surfaces (Newman et al. 1994).
Cross-sectional dimensions of cellulose crystallites—
The subspectra of crystalline cellulose (subspectrum A)
(Fig. 1-4) showed the C-4 signal split into two components
at 89 and 84 ppm which we assigned to crystal-interior and
crystal-surface cellulose. Relative signal areas were estimated by drawing vertical boundary lines at 80, 87, and 93
ppm (Fig. 5). These boundaries are similar to those used in
the studies of the cellulose in apple and A.thaliana cell
walls (Newman et al. 1994, 1996) except that the boundary
at 79 ppm was moved to 80 ppm to reflect an improved
estimate of the position of a dip between the signal assigned to crystal-surface C-4 and signals assigned to C-2, C-3,
and C-5. The percentages of cellulose molecules assigned to
the crystal-interior cellulose are 44%, 36%, 34%, and 32%
for Italian ryegrass, pineapple, onion, and cabbage respec-
87
85
PPM from TM3
Fig. 5 Arbitrary boundaries chosen for the distinction between
signal areas associated with crystal-interior and crystal-surface C4, illustrated for a portion of Fig. 4a (cabbage cell walls).
tively. These percentages have a standard error of approximately±l% (R.H. Newman unpublished).
For each percentage of cellulose molecules assigned to
crystal-interior cellulose, a limited number of models for
cross sections through cellulose crystallites are theoretically possible. However, it must be stressed that the NMR
results cannot be used to distinguish among these models.
For instance, one cannot distinguish between square and
rectangular cross sections of crystallites. Neither can they
be used to test the uniformity of cross-sectional dimensions. In Fig. 6, we show examples of theoretical models
which are consistent with the percentages of crystal-interior
cellulose. Fig. 6a is an example of a model for a cross section of a cellulose crystallite of Italian ryegrass. The model
shows 12 cellulose molecules in the interior of the crystallite, each hydrogen-bonded to two adjacent molecules, and
16 cellulose molecules exposed on the surface of the crystallite, i.e. hydrogen-bonded to just one adjacent molecule.
Fig. 6b shows an example of a model for a cross section of
a cellulose crystallite of pineapple. Similarly, Fig. 6c shows
an example of a model of a cellulose crystallite of onion or
cabbage. These models have crystallite dimensions for each
species as follows: Italian ryegrass 3.2x3.1 nm; pineapple
3.2 x 2.4 nm; and onion and cabbage 2.7 x 2.4 nm.
In the subspectra of crystalline cellulose (subspectrum
A) (Fig. 1-4) signals assigned to C-6 are split into components at 65 and 63 ppm, assigned to cellulose molecules
in the interior and surface of the crystallite respectively.
The signal at 65 ppm is relatively strong for Italian ryegrass
(Fig. 1), which supports evidence for relatively large crosssectional dimensions of the crystallites. Due to overlapping
signals, we were unable to assign unambiguously signals in
the 63-65 ppm region of the subspectra A of pineapple,
onion, and cabbage (Fig. 2-4) and consequently we were unable to obtain additional information on cross-sectional
dimensions of the cellulose crystallites for these species.
Crystal forms—The signals observed in the resolutionenhanced spectra (Fig. 7) are all consistent with cellulose I
being the dominant form of cellulose in all the primary
walls examined. Many of the signals are common to both
the triclinic (Ia) and monoclinic (1^) forms, but plot expansions of chemical-shift ranges associated with C-l (Fig. 8)
Crystalline cellulose in primary cell walls
717
o ooo o
00000O
00O0OO
OO0 00O
o dodo
a
65
PPM from TMS
OOOOO
ddddo'o
oddddo
o o'o'o 0
b
OOOO
000 0O
o dodo
'
00
*
s
'
00
Fig. 6 Examples of theoretical models of cross sections through
cellulose crystallites which are consistent with the percentages of
crystal-interior cellulose found in the present study: (a) model (28
cellulose molecules) with 43% crystal-interior cellulose, consistent
with the results for Italian ryegrass; (b) model (22 cellulose molecules) with 38% crystal-interior cellulose, consistent with the
results for pineapple; and (c) model (18 cellulose molecules) with
33% crystal-interior cellulose, consistent with the results for onion
and cabbage. Long axes of the cellulose molecules are oriented
normal to the plane of the paper. Dotted lines represent hydrogen
bonds. A limited number of other models are theoretically possible for the same percentages of cellulose molecules assigned to the
crystal-interior cellulose.
or C-4 (Fig. 9) show signals (marked by vertical broken
lines) that are unique to the triclinic (Ia) or monoclinic (Lj)
form. Plot expansions of the resolution-enhanced spectra
(Fig. 9) also show the fine structure of the crystal surfaces
of cellulose I. The signals at 84 and 85 ppm (labelled "s")
are assigned to surfaces of cellulose I crystallites, regardless
of whether the molecules are packed in the triclinic (IJ or
monoclinic (lfi) crystalline forms (Newman et al. 1994). The
splitting of about 1 ppm is attributed to the fact that the
repeating unit of cellulose is cellobiose and not glucose.
From these results we conclude that the cellulose in all of
the cell walls examined contained similar proportions of
Fig. 7 Resolution-enhanced solid-state 13C NMR spectra of primary cell-wall preparations of (a) Italian ryegrass, (b) pineapple,
(c) onion and (d) cabbage.
both the triclinic (Io) and the monoclinic (I/j) crystal forms.
We found little evidence for the presence of cellulose
II in the cell walls examined. Cellulose II is often identified
by a signal from C-l at 107-108 ppm because other signals
in the spectrum reported for cellulose II (Hemmingson and
Newman 1995) are too close to cellulose I signals for positive identification. The resolution-enhanced spectra of all
the cell-wall preparations examined show a signal at or
near 108 ppm (Fig. 7, 8). However, the detection of a signal
at 107.8 ppm does not prove the presence of cellulose II,
since other substances might contribute a signal at similar
chemical shifts. In particular, the signal is likely to be associated with C-l signals from arabinose which is a constituent of all the cell walls of all the species examined. The
signal at 108 ppm is very weak but is useful in providing an
109
105
PPM from TMS
Fig. 8 Plot expansions from Fig. 7 of the resolution-enhanced
solid-state 13C NMR spectra of primary cell-wall preparations of
(a) Italian ryegrass, (b) pineapple, (c) onion and (d) cabbage showing the chemical-shift range assigned to C-l. Labels "m" and "t"
refer to the monoclinic (fy) and triclinic (Ia) crystal forms of cellulose, and label "s" refers to molecules exposed on surfaces of both
crystal forms.
718
Crystalline cellulose in primary cell walls
87
85
PPM from TMS
83
81
79
Fig. 9 Plot expansions from Fig. 7, of the resolution-enhanced
solid-state 13C NMR spectra of primary cell-wall preparations of
(a) Italian ryegrass, (b) pineapple, (c) onion and (d) cabbage showing the chemical-shift range assigned to C-4. Labels "m" and "t"
refer to the monoclinic (Iyj) and triclinic (Ia) crystal forms of cellulose, and label "s" refers to molecules exposed on surfaces of both
crystal forms.
upper limit on the content of cellulose II.
We also found little evidence for the presence of molecules, such as xyloglucans or glucuronoarabinoxylans,
adhering to the surfaces of cellulose I crystallites. This was
based on the ability to resolve partly the two signal components at 84 and 85 ppm which suggests that in all four
primary cell walls examined, the cellulose crystal-surface
molecules are well-ordered and mostly free from adhering
molecules. If xyloglucans, glucuronoarabinoxylans, or any
other cell-wall polysaccharide were adhering to cellulose,
we would have expected a broadening of the peaks associated with crystallite surfaces. However, these peaks at 84 and
85 ppm were sharp indicating that most of the cellulose
crystallite surface was clean (Newman and Hemmingson
1995).
Discussion
The results showed that in all the primary cell walls examined the cellulose was present in a crystalline form, cellulose I, and the proportions of the two crystalline forms triclinic (Ia) and monoclinic (1^) were approximately equal
despite differences in the non-cellulosic polysaccharide compositions. Approximately equal proportions of these two
crystalline forms were also reported for the cellulose in the
primary cell walls of apple (Newman et al. 1994) and Arabidopsis thaliana (Newman et al. 1996), both dicotyledons.
However, in contrast to the cellulose in primary cell walls,
the cellulose in the secondary cell walls of the dicotyledons
cotton and ramie contained more of the monoclinic (1^)
form than the triclinic (Ia) form (Atalla and VanderHart
1994). A similar predominance of the monoclinic (1^) form
was also found in the secondary cell walls of the dicotyle-
dons Castanea sativa, Beilschmiedia tawa, Eucalyptus
delegatensis, and Quercus robur (Newman 1994). Nevertheless, it is too early to formulate general rules concerning
the proportions of the monoclinic (Lj) form and the triclinic (Ia) form of cellulose from different sources.
Although the proportions of monoclinic (Lj) and triclinic (Io) cellulose did not vary, the cross-sectional dimensions of the cellulose crystallites in the cell walls of the
three monocotyledons examined differed in the range 2-3
nm. The largest crystallites were found in the cell walls of
Italian ryegrass, the smallest were found in the cell walls of
onion and crystallites of intermediate dimensions were
found in the cell walls of pineapple. These three monocotyledons have cell walls with different non-cellulosic compositions: the Italian ryegrass cell walls containing predominantly glucuronoarabinoxylans, the onion cell walls
containing predominantly pectic polysaccharides and xyloglucans, and the cell walls of pineapple having an intermediate composition. The crystallites in the cell walls of the dicotyledon cabbage were similar in size to those in onion
and both cell walls have similar non-cellulosic polysaccharide compositions. Thus, the non-cellulosic polysaccharides of primary cell walls may influence the molecular
ordering of cellulose. However, further work is required to
determine if this relationship between cellulose crystallite
size and the composition of the non-cellulosic polysaccharides holds for other monocotyledon species. However, it is
interesting that the size of the cellulose crystallites in the primary cell walls of the dicotyledon cabbage were smaller
than those in the primary cell walls of the dicotyledons apple and A. thaliana (Newman et al. 1994, 1996).
Some of the studies on the crystallinity of cellulose in
primary cell walls in which X-ray and electron diffraction
were used can also be interpreted in terms of cellulose crystallites of similar sizes to those found by us. For example,
Chanzy et al. (1978, 1979), in a study of the primary cell
walls of cotton and rose, explained their results in terms of
cellulose crystallites that had small cross-sectional dimensions in the order of 2 to 3 nm implying that they were composed of 12 to 25 cellulose molecules. In contrast to X-ray
and electron diffraction, CP/MAS 13C NMR is much more
sensitive to short range order and provides a more sensitive
test for variations in the relative proportions of crystal-surface and crystal-interior molecules. Our results indicate
that crystallites with small cross-sectional dimensions (in
the range 2-3 nm) occur in primary cell walls of a number
of angiosperm species and suggest that the results of Chanzy et al. (1978, 1979) can most likely be explained in terms
of crystallites with small cross-sectional dimensions. The
cross-sectional dimensions of the cellulose crystallites are
similar to the width of microfibrils often observed by using
negative staining techniques (e.g. Reis et al. 1991).
In the present study we found surprisingly little evidence for a substantial interaction between cellulose and xy-
Crystalline cellulose in primary cell walls
loglucans or glucuronoarabinoxylans or any other polymer. A similar finding was also reported for apple and
A. thaliana primary cell walls (Newman et al. 1994, 1996).
This is inconsistent with current models of dicotyledon and
Poaceae cell walls (e.g. Talbott and Ray 1992, Carpita and
Gibeaut 1993) in which xyloglucans and glucuronoarabinoxylans respectively are postulated to coat and cross-link the
cellulose microfibrils. Evidence in support of a substantial
interaction between cellulose and xyloglucans or xylans
comes from studies which show that xyloglucans and
xylans bind to cellulose in vitro (e.g. Mitikka et al. 1995,
Vincken et al. 1995, Whitney et al. 1995). Current cell-wall
models have also led to the hypothesis that expansins (proteins that catalyse the extension of isolated plant cell walls)
act by inducing slippage between cellulose microfibrils
and adhering xyloglucans or xylans (McQueen-Mason and
Cosgrove 1994). We suggest that a relatively small number
of xyloglucans or xylans molecules would be sufficient to
cross-link the cellulose microfibrils.
In conclusion, we found using CP/MAS 13C NMR
spectroscopy that the cellulose in the primary cell walls of
three monocotyledons and a dicotyledon was all in a crystalline state with similar proportions of both the triclinic (Ia)
and the monoclinic (1^) crystal forms; there was no evidence of substantial amounts of adhering non-cellulosic
polysaccharides. However, the cross-sectional dimensions
of the cellulose crystallites varied among the plant species
examined and may be affected by the non-cellulosic components of the cell walls.
This research was supported by a University of Auckland
Doctoral Scholarship (to BGS), The University of Auckland
Graduate Research Fund, and The Miss EL Hellaby Indigenous
Grasslands Research Trust.
References
Alia, M. and Lippmaa, E. (1976) High resolution broad line I3C NMR and
relaxation in solid nobornadiene. Chem. Phys. Letts. 37: 260-264.
Atalla, R.H., Hackney, J.M., Uhlin, I. and Thompson, N.S. (1993) Hemicelluloses as structure regulators in the aggregation of native cellulose.
Int. J. Biol. Macromot. 15: 109-112.
Atalla, R.H. and VanderHart, D.L. (1994) Native cellulose: a composite of
two distinct crystalline forms. Science 223: 283-285.
Bacic, A., Harris, P.J. and Stone, B.A. (1988) Structure and function of
plant cell walls. In The Biochemistry of Plants, Vol. 14. Edited by
Preiss, J. pp. 297-371. Academic Press, San Diego, CA, U.S.A.
Bardet, M., Emsley, L. and Vincendon, M. (1997) Two-dimensional spinexchange solid-state NMR studies of "C-enriched wood. Solid State
Nucl. Magn. Reson. 8: 25-32.
Belton, P.S., Tanner, S.F., Cartier, N. and Chanzy, H. (1989) High-resolution solid-state "C nuclear magnetic resonance spectroscopy of tunicin,
an animal cellulose. Macromolecules 22: 1615-1617.
Cael, J.J., Kwoh, D.W.L., Bhattacharjee, S.S. and Pan, S.L. (1985) Cellulose crystallites: a perspective from solid state I3C NMR. Macromolecules 18: 819-821.
Carpita, N.C. and Gibeaut, D.M. (1993) Structural models of primary cell
walls in flowering plants: consistency of molecular structure with the
physical properties of the walls during growth. Plant J. 3: 1-30.
Chanzy, H., Imada, K., Mollard, R., Vuong, R. and Barnoud, F. (1979)
719
Crystallographic aspects of sub-elementary cellulose fibrils occurring in
the walls of rose cells cultured in vitro. Protoplasma 100: 303-316.
Chanzy, H., Imada, K. and Vuong, R. (1978) Electron diffraction from the
primary cell wall of cotton fibres. Protoplasma 94: 299-306.
Chesson, A., Gordon, A.H. and Lomax, J.A. (1985) Methylation analysis
of mesophyll, epidermis, and fibre cell-walls isolated from the leaves of
perennial and Italian ryegrass. Carbohyr. Res. 141: 137-147.
Debzi, E., Chanzy, H., Sugiyama, J., Tekely, P. and Excoffier, G. (1991)
The I, and l0 transformation of highly crystalline cellulose by annealing
in various mediums. Macromolecules 24: 6816-6822.
Earl, W.L. and VanderHart, D.L. (1980) High resolution, magic angle sample spinning I3C NMR of solid cellulose I. J. Amer. Chem. Soc. 102:
3251-3252.
Earl, W.L. and VanderHart, D.L. (1981) Observations by high-resolution
carbon-13 nuclear magnetic resonance of cellulose I related to morphology and crystal structure. Macromolecules 14: 570-574.
Ferrige, A.G. and Linden, J.C. (1978) Resolution enhancement in FT
NMR through the use of a double exponential function. J. Magn.
Reson. 31: 331-340.
Hackney, J.M., Atalla, R.H. and VanderHart, D.L. (1994) Modification
of crystallinity and crystalline structure of Acetobacter xylinum cellulose
in the presence of water-soluble /M,4-linked polysaccharides: I3C-NMR
evidence. Int. J. Biol. Macromol. 16: 215-218.
Harris, P.J. (1983) Cell walls. In Isolation of Membranes and Organelles
from Plant Cells. Edited by Hall, J.L. and Moore, A.L. pp. 25-53.
Academic Press, London.
Harris, P.J., Hartley, R.D. and Lowry, K.H. (1980) Phenolic constituents
of mesophyll and non-mesophyll cell walls from leaf laminae of Lolium
perenne. J. Sci. Food Agric. 31: 959-962.
Harris, P.J., Kelderman, M.R., Kendon, M.F. and McKenzie, R.J. (1997)
Monosaccharide compositions of unlignified cell walls of monocotyledons in relation to the occurrence of wall-bound ferulic acid. Biochem.
Syst. Ecol. 25: 167-179.
Harris, R.K. (1986) Nuclear Magnetic Resonance Spectroscopy. A Physicochemical View. Revised Edition, Longman Scientific and Technical,
Harlow, Essex, U.K.
Harris, R.K. (1993) State of the art for solids. Chemistry in Britain 29:
601-604.
Hemmingson, J.A. and Newman, R.H. (1995) Changes in molecular ordering associated with alkali treated and vacuum drying of cellulose. Cellulose 2: 71-82.
Jarvis, M.C. (1990) Solid state "C-n.m.r. spectra of Vigna primary cell
walls and their polysaccharide components. Carbohydr. Res. 201: 327333.
Jarvis, M.C. and Apperley, D. (1990) Direct observation of cell wall structure in living plant tissues by solid-state "C NMR spectroscopy. Plant
Physiol. 92: 61-65.
Jensen, W.A. (1962) Botanical Histochemistry. W.H. Freeman, San Francisco.
Kataoka, Y. and Kondo, T. (1996) Changing cellulose crystalline structure
in forming wood cell walls. Macromolecules 29: 6356-6358.
Lennholm, H., Larsson, T. and Iversen, T. (1994) Determination of cellulose I,, and lf in lignocellulosic materials. Carbohydr. Res. 261: 119-131.
Mankarios, A.T., Hall, M.A., Jarvis, M . C , Threlfall, D.R. and Friend, J.
(1980) Cell wall polysaccharides from onions. Phytochemistry 19: 1731 —
1733.
Martel, P. and Taylor, I.E.P. (1993) Neutron diffraction and proton nuclear magnetic resonance: complementary probes of in situ cellulose dimensions and primary plant cell wall structure. Can. J. Bot. 71: 1375-1380.
McCann, M . C , Wells, B. and Roberts, K. (1990) Direct visualization of
cross-links in the primary plant cell wall. J. Cell Sci. 96: 323-334.
McQueen-Mason, S. and Cosgrove, D.J. (1994) Disruption of hydrogen
bonding between wall polymers by proteins that induce plant wall extension. Proc. Natl. Acad. Sci. USA 91: 6574-6578.
Mitikka, M., Teeaar, R., Tenkanen, M., Laine, J. and Vuorinen, T. (1995)
Sorption of xylans on cellulose fibres. In The 8th International Symposium on Wood and Pulping Chemistry, Helsinki, Finland. Proceedings Vol. III. pp. 231-236. Gummerus Kirjapaino Oy Jyvaskyla,
Finland.
Neville, A.C. and Levy, S. (1984) Helicoidal orientation of cellulose micro-
720
Crystalline cellulose in primary cell walls
fibrils in Niteila opaca internode cells: ultrastructure and computed theoretical effects of strain re-orientation during wall growth. Planta 162:
370-384.
Newman, R.H. (1992) Nuclear magnetic resonance study of spatial
relationships between chemical components in wood cell walls.
Holzforschung 46: 205-210.
Newman, R.H. (1994) Crystalline forms of cellulose in softwoods and hardwoods. J. Wood Chem. Technol. 14: 451-466.
Newman, R.H. (1998) Evidence for assignment of "C NMR signals to
cellulose crystallite surfaces in wood, pulp and isolated celluloses.
Holzforschung 52: 157-159.
Newman, R.H., Davies, L.M. and Harris, P.J. (1996) Solid-state I]C nuclear magnetic resonance characterization of cellulose in the cell walls of
Arabidopsis thaliana leaves. Plant Physiol. 112: 475-485.
Newman, R.H., Ha, M-A. and Melton, L.D. (1994) Solid-state 13C NMR
investigation of molecular ordering in the cellulose of apple cell walls. J.
Agric. Food Chem. 42: 1402-1406.
Newman, R.H. and Hemmingson, J.A. (1990) Determination of the degree
of cellulose crystallinity in wood by carbon-13 nuclear magnetic resonance spectroscopy. Holzforschung 44: 351-355.
Newman, R.H. and Hemmingson, J.A. (1995) Carbon-13 NMR distinction between categories of molecular order and disorder in cellulose. Cellulose 2: 95-110.
Newman, R.H., Hemmingson, J.A. and Suckling, I.D. (1993) Carbon-13
nuclear magnetic resonance studies of kraft pulping. Holzforschung 47:
234-238.
Ohsumi, C. and Hayashi, T. (1994) The oligosaccharide units of the xyloglucans in the cell walls of bulbs of onion, garlic and their hybrid. Plant
Cell Physiol. 35: 963-967.
Redgwell, R.J. and Selvendran, R.R. (1986) Structural features of cell-wall
polysaccharides of onion Allium cepa. Carbohydr. Res. 157: 183-199.
Reis, D., Vian, B., Chanzy, H. and Roland, J-C. (1991) Liquid crystal-type assembly of native cellulose-glucuronoxylans extracted from plant
cell wall. Biol. Cell. 73: 173-178.
Revol, J.-F. (1982) On the cross-sectional shape of cellulose crystallites in
Valonia ventricosa. Carbohydr. Polym. 2: 123-134.
Roland, J.C., Vian, B. and Reis, D. (1975) Observations with cytochemistry and ultracryotomy on the fine structure of the expanding cell walls in
actively elongating plant cells. J. Cell Sci. 19: 239-259.
Ryden, P., Colquhoun, I.J. and Selvendran, R.R. (1989) Investigation of
structural features of the pectic polysaccharides of onion by "C-N.M.R.
spectroscopy. Carbohydr. Res. 185: 233-237.
Smith, B.G. and Harris, P.J. (1995) Polysaccharide composition of un-
lignified cell walls of pineapple [Ananas comosus (L.) Merr.] fruit. Plant
Physiol. 107: 1399-1409.
Stevens, B.J.H. and Selvendran, R.R. (1980) Structural investigation of an
arabinan from cabbage (Brassica oleracea var. capitata). Phytochemistry
19: 559-561.
Stevens, B.J.H. and Selvendran, R.R. (1984a) Pectic polysaccharides of
cabbage (Brassica oleracea). Phytochemistry 23: 107-115.
Stevens, B.J.H. and Selvendran, R.R. (1984b) Hemicellulosic polymers of
cabbage leaves. Phytochemistry 23: 339-347.
Sugiyama, J., Chanzy, H. and Revol, J.F. (1994) On the polarity of cellulose in the cell walls of Valonia. Planta 193: 260-265.
Sugiyama, J., Okano, T., Yamamoto, H. and Horii, F. (1990) Transformation of Valonia cellulose crystals by an alkaline hydrothermal treatment.
Macromolecules 23: 3196-3198.
Sugiyama, J., Persson, J. and Chanzy, H. (1991) Combined infrared and
electron diffraction study of the polymorphism of native celluloses.
Macromolecules 24: 2461-2466.
Sugiyama, J., Vuong, R. and Chanzy, H. (1991) Electron diffraction study
on the two crystalline phases occurring in native cellulose from an algal
cell wall. Macromolecules 24: 4168-4175.
Talbott, L.D. and Ray, P.M. (1992) Molecular size and separability
features of pea cell wall polysaccharides. Implications for models of pri. mary wall structure. Plant Physiol. 98: 357-368.
VanderHart, D.L. and Atalla, R.H. (1984) Studies of microstructure in
native cellulose using solid-state "CNMR. Macromolecules 17: 14651472.
Vincken, J-P., de Keizer, A., Beldman, G. and Voragen, A.G.J. (1995)
Fractionation of xyloglucan fragments and their interaction with cellulose. Plant Physiol. 108: 1579-1585.
Whitney, S.E.C., Brigham, J.E., Darke, A.H., Reid, J.S.G. and Gidley,
M.J. (1995) In vitro assembly of cellulose/xyloglucan networks: ultrastructural and molecular aspects. Plant J. 8: 491-504.
Wormald, P., Wickholm, K., Larsson, P.T. and Iversen, T. (1996) Conversions between ordered and disorderd cellulose. Effects of mechanical
treatment followed by cyclic wetting and drying. Cellulose 3: 141-152.
Yamamoto, H. and Horii, F. (1994) In situ crystallization of bacterial cellulose I. Influences of polymeric additives, stirring and temperature on the
formation of celluloses Ia and \f as revealed by cross polarization/magic
angle spinning (CP/MAS) "C NMR spectroscopy. Cellulose 1: 57-66.
Zumbulyadis, N. (1983) Selective carbon excitation and the detection of
spatial heterogeneity in cross-polarization magic angle-spinning NMR.
J. Magn. Reson. 53: 486-494.
(Received November 4, 1997; Accepted April 14, 1998)