The Institute of Paper Science

and Technology
Atlanta, Georgia

Doctor's Dissertation

The Influence of Hemicelluloses on the Structure

of Bacterial Cellulose

Karin Ingegerd Uhlin

March, 1990
THE INFLUENCE OF HEMICELLULOSES ON THE STRUCTURE OF
BACTERIAL CELLULOSE

A thesis submitted by
Karin Ingegerd Uhlin
M.S. 1985, The Royal Institute of Technology

Stockholm, Sweden

in partial fulfillment of the requirements

for the degree of Doctor of Philosophy from
the Institute of Paper Science and Technology

Atlanta, Georgia

Publication rights reserved by the

Institute of Paper Science and Technology

March, 1990
 TABLE OF CONTENTS
Page
ABSTRACT 1
INTRODUCTION 3
BACKGROUND 6
Acetobacter xylinum 6

Disruption of bacterial cellulose aggregation 7

Sorption of hemicelluloses and gums by cellulose 10
Xyloglucans 10
Xylans 12
Glucomannans 13
Gums 14
Cellulose polymorphism and molecular conformations 14

Cellulose I and II 15
Hierarchical organization of cellulose structures 16
Analysis techniques for probing structural levels
of cellulose 18

The significance of a B-(1-4)-linked polysaccharide backbone

for sorption 19
The additives used in this thesis and their relation to
plant cell walls 19
Xyloglucan 19

Xylan 22
Mannan 23

Carboxymethyl cellulose (CMC) 23

Thesis objective 23
Experimental plan 24

i
 vi
Table page numbers
Table Page
number Caption number
1 Cellulose derivatives tested for alteration of
bacterial cellulose 20
2 Other compounds tested for alteration of
bacterial cellulose 20
3 Binding of polysaccharides and xyloglucan to
pea cellulose 21
4 Chemical composition of xyloglucan 27
5 Carbohydrate composition of crude mannan (CM) 44
6 Carbohydrate composition of xylan 45
7 Weights of celluloses after alkaline extractions 47
8 Peak widths (20-degrees): bacterial celluloses produced
in the absence or presence of additives, after
extraction with 1.0 N NaOH 51
9 Peak positions (20-degrees): bacterial celluloses
produced in the absence or presence of additives,
after extraction with 1.0 N NaOH 52
10 Peak intensities and relative intensities: bacterial
celluloses produced in the absence or presence of
additives, after extraction with 1.0 N NaOH 53
11 Conformational components of bacterial celluloses
produced in absence or presence of additives, after
the extraction with 0.1 N NaOH 61
12 Fibril widths (nm) of bacterial celluloses produced in
absence or presence of additives, at different concent-
rations (% w/v) of the additives in the growth medium 66
13 X-ray data for cotton (CF1) and ramie celluloses 70
14 13C NMR spectrum of mannan 97
15 13C NMR spectrum of xylan 99
16 X-ray diffractograms in Appendix B 101
17 Raman spectra in Appendix C 110
18 Conformational components of bacterial celluloses 118
 ABSTRACT

One assumption implicit in much research on cellulose structure is that

the crystal structure of higher-plant celluloses is the same as that of
relatively pure celluloses. Another assumption is that treatments which are
necessary to isolate celluloses from native material will not alter the
cellulose structure. The purpose of this thesis was to investigate the impact
of hemicellulose-like polysaccharides on the crystallinity and morphology of
nascent cellulose fibrils while they were aggregating, and the impact of
isolation procedures on the aggregation pattern of cellulose.

The approach was to build a model system of cellulose and hemicellulose.

The bacterium Acetobacter xylinum was the source of cellulose. Carboxymethyl

cellulose (CMC) and the hemicelluloses mannan, xylan, and xyloglucan were used
as additives in the bacterial growth medium. The celluloses produced in the
presence of additives were extracted with sodium hydroxide solutions of
increasing concentrations to observe the effect of isolation on the cellulose

structure. The alterations of cellulose aggregation were evaluated via x-ray

diffraction, Raman spectroscopy, and transmission electron microscopy. The

analyses were quantified by resolution of the x-ray diffractograms and Raman

spectra, and by image analysis of the photomicrographs.

Cellulose I was produced by A. xylinum in the presence of all additives.

Compared with control cellulose, celluloses produced with additives had

decreased degree of crystallinity, smaller crystallites, lower degree of

orientation, narrower fibrils, and changed dimensions of the crystal unit
cells. The influences of xylan and xyloglucan were stronger than those of
 2
mannan and CMC. The patterns of aggregation and fibril widths of the
bacterial celluloses produced with xylan and xyloglucan resemble those found
in celluloses from higher plants. Furthermore, comparisons of x-ray
diffractograms and Raman spectra from the extraction stages indicated changes
in the aggregation patterns induced by the isolation of the celluloses.

The major contributions from this study are the proposals that 1) matrix

polysaccharides in the cell walls of higher plants influence the aggregation

of cellulose; and 2) the aggregation pattern of cellulose is determined, in

part, by the process for isolation of the cellulose.

 INTRODUCTION

The structures of cellulose have been the subject of numerous

investigations. However, several assumptions have been implicit in many of
these studies. One common assumption is that the aggregation pattern of
cellulose is not influenced by other cell wall components present when the
cellulose is produced. Another assumption is that removal of noncellulosic
material during the isolation of cellulose will not alter the structure of

cellulose. The objectives of this thesis were to study the influence of

hemicelluloses and of isolation procedures on the structure of bacterial
cellulose.

Bacterial cellulose has been used frequently in cellulose studies as a

substitute for celluloses from higher plants, in part due to its high purity

and crystallinity and also because of the problems encountered in producing

cellulose I in vitro with plant enzyme preparations. In the first study of
the modification of bacterial cellulose, Ben-Hayyim and Ohad 1 found that the

presence of carboxymethyl cellulose could delay the aggregation of cellulose

synthesized by Acetobacter xylinum. Haigler and Brown 2-4 investigated the

effect of a wide variety of polysaccharides and other compounds on the

bacterial cellulose structure. These researchers found that the bacterial
cellulose responded in a variety of ways, depending on the characteristics of
the disrupting agent.

In another study of bacterial cellulose 5, newly-formed bacterial

cellulose could easily be modified, whereas the degree of transformation was

much lower when the cellulose was more mature. This result indicates that

3
 4
cellulose may be more susceptible to chemical or physical influences by other
compounds when the cellulose is newly polymerized.

On the basis of such studies, the question was posed in the present
investigation as to how celluloses in plant cell walls respond, in terms of
changes in aggregation pattern and morphology, to other compounds present
during deposition. In plant cell walls, hemicelluloses are deposited
simultaneously with cellulose 6, and, therefore, are the focus of this study.

The assumption that removal of noncellulosics has no influence on

cellulose structures is based upon the belief that higher-plant native-
cellulose structures are identical to those of highly purified celluloses
produced by bacteria or manufactured commercially. This disregards the fact
that most isolation methods involve extraction with strong acid or base. As
Preston 7 has written, "It had been shown ... that the removal from a cell wall

of pectic substances, hemicelluloses and lignin had no effect upon the x-ray
diagram of the wall (so long as the extraction procedures were not so severe

as materially to degrade the cellulose) except for some sharpening of the arcs
such as would indicate a small increase in crystallite size." Other studies
have demonstrated that wood pulps prepared by different processes display
subtle but consistent spectral and diffractometric differences.8-10 Such
discrepancies led to the second objective for this investigation.

To challenge these two assumptions, the approach was taken in this study
to use the cellulose-producing bacterium A. xylinum as a model system of

cellulose biosynthesis in cell walls of higher plants, excluding lignin and

similar wood components. In order to simulate the environment in plant cell

walls, hemicelluloses or other compounds could be added to the growth medium
 5
while the cellulose was synthesized and aggregated. The basis for this study
was the hypothesis that, depending on properties of the hemicellulose, it
could interact with the aggregation, and not with the polymerization, of
bacterial cellulose. To observe the effect of isolation procedures on the
cellulose structure, the celluloses produced in the presence of hemicelluloses
were subjected to alkaline extractions as simplified representations of
standard laboratory isolation procedures.
 +
BACKGROUND

Acetobacter xylinum

Acetic acid bacteria of the genus Acetobacter are important organisms in

industrial use as well as in research. They have been used for the production
of vinegar from wine, and for the production of gluconic acid, ketogluconic
acids, and sorbose. Glucose is a common substrate, but many other simple
carbohydrates, alcohols, or polyalcohols can be used as carbon sources. The
oxidative pathways in the acetic acid bacteria have been important models for
biochemical studies of fermentation. The cellulose-producing species of these
acetic acid bacteria have been useful in research on cellulose biogenesis. 11

Vinegar was produced for centuries by natural fermentation without the

nature of the process being understood. Gradually it was realized that living
organisms were involved. The first systematic studies of acetic acid
fermentation were carried out by Pasteur.12 Brown13 was among the first to

recognize cellulose formation by A. xylinum.

All strains of A. xylinum produce cellulose extracellularly in the form

of flat, twisting ribbons. Each cell produces one ribbon which is associated
with the surface of the bacterial cell.1 4,15 The ribbon is assembled from
fibrils, which themselves consist of a number of glucan chains. The fibrils
are in close association with a longitudinal row of pores in the bacterial
cell membrane. 16 The cellulose-synthesizing multienzyme complexes are thought

to be imbedded in the outer membranes of the cells where the glucan chains are
assembled for extrusion through the pores in the cell wall. When A. xylinum

is incubated in liquid medium, the ribbons intertwine to form a thick,

6
 7

leathery pellicle at the air-liquid interface on top of the growth medium.

It is not clear why A. xylinum produces cellulose. The pellicle is

suggested to function as a floating platform to position these obligately
aerobic organisms near the air-liquid interface. 17 The pellicles have been
shown to provide protection against UV-light, and against other microorganisms

competing for the same source of nutrition by enhancing colonization of A.

xylinum.18 One speculation is that cellulose synthesis is a mechanism for
mobility, since the force from the polymerization and crystallization actually
propels the cell forward. 15'19 The production of cellulose may, on the other
hand, be a remnant of metabolism without any direct function.

DISRUPTION OF BACTERIAL CELLULOSE AGGREGATION

The research on disruption of the bacterial cellulose ribbon assembly is

aimed at deepening the understanding of the formation of cellulose fibrils and
ribbons. The focus has been primarily on the morphology, and not the

aggregation, of the bacterial cellulose. The disruption has been induced by

compounds of two types: 1. Fluorescent brightening agents and direct dyes,

and 2. polysaccharides.

Fluorescent Brightening Agent 28 (FBA 28 or Calcofluor) and direct dyes

are compounds used for the whitening and coloring of cellulosic material. The

structures of these compounds are similar and the important characteristics of

the molecules are the planar and linear configurations. The molecules contain

between four and six aromatic rings substituted with hydrophilic groups to
increase the water-solubility. FBA 28 and most other brighteners are based on
a diaminostilbene skeleton, whereas the common feature in most direct dyes is
 8

one or more azo (N-N) groups. 20.21

The presence of fluorescent brightening agents and direct dyes modified

the morphology of the bacterial cellulose. When FBA 28 was present in the

incubation medium of A. xylinum, the cellulose synthesized was in the form of

a sheet of nonfibrillar cellulose, instead of the normal ribbon. 2'3 At lower
concentrations of FBA 28, a band with fibrillar substructure was produced.
The brightening agents and dyes prevented crystallization of the cellulose, as
shown by electron diffraction. 22 Both the sheet and band products lacked
crystal patterns, but cellulose I was recovered when the altered celluloses
were washed free of disrupting agent. Kai23 supported these results in an
x-ray study. The presence of direct dyes resulted in the formation of
cellulose products similar to those from FBA 28. The magnitude of the effect
of each of the direct dyes depended on the degree of linearity and planarity
of each of the respective molecules.

Carboxymethyl cellulose and other cellulose derivatives have also been

reported to alter the assembly of the bacterial cellulose ribbon.1.4 The
presence of CMC in the growth medium delayed the formation of visible

aggregates of bacterial cellulose.1 Instead of ribbons, bundles of separate

intertwining fibrils were synthesized in the presence of CMC. 4 The cellulose

produced in the presence of CMC was crystalline, indicating that the influence
of CMC on the bacterial cellulose is different than the influence of

brightening agents and dyes. In addition to CMC, many other polysaccharides

have been tested for their effects on ribbon assembly, including a xylan

(isolated from larch or oats?), a pectin (probably from citrus fruit), and a

yeast mannan. 4 Of these additives, only xylan disrupted the ribbon assembly.
 9
When the crystallization of bacterial cellulose is delayed or prevented
in the presence of CMC or FBA 28, the rate of cellulose synthesis
24
increases.1. The presence of CMC in the systemincreased the cellulose
polymerization by about 30%. 1 The incorporation of 14C-labeled glucose into
bacterial cellulose showed that the polymerization of cellulose was increased
by 200-400% in the presence of FBA 28.24 The degree of polymerization of FBA
28-altered cellulose was comparable to control cellulose, as measured by

viscosity.2

The inhibition of the crystallization, but not the polymerization, of

bacterial cellulose indicates that polymerization and crystallization are
sequential processes separated in time. Furthermore, the kinetic studies
referred to above suggest that the crystallization is the rate-limiting step.

Haigler, Benziman, and Brown have proposed a model of hierarchical cell-

directed self-assembly of the cellulose ribbon based on these results.19.24 In
this model, brightening agents and dyes interact with the ribbon assembly
earlier in the assembly process than the polysaccharides do.

The separation of cellulose polymerization and crystallization is also

shown in a study by Kai and Koseki. 5 A. xylinum was incubated for either 5 or
30 minutes and the cellulose was treated with sodium hydroxide solutions of

different concentrations. When cellulose from the 5-minute incubation was

extracted with 10-4 N alkali, electron diffraction revealed a crystallinity
typical of cellulose I. When alkali solutions of higher concentrations

(2.5x10-4 to 2.5x10-3 N) were used for extraction of products from the 5-

minute incubation, a cellulose II crystallinity was detected and the

conversion was estimated to be 50 to 100%. The report does not mention if any
noncrystalline cellulose was present before or after the alkaline treatments.
 10

When the cellulose from the 30-minute incubation was extracted with 1.15 N
alkali, about 75% of the material showed cellulose I crystallinity and the
remaining material showed cellulose II crystallinity. The reverse ratio of
cellulose I and II material was obtained when 2.4 N alkali was used. Such an
increased resistance to alkali with a prolonged period of incubation indicates
that nascent material has weaker bonds but that the bond strength increases
drastically during about 30 minutes, thereby supporting the idea of time-

separated polymerization and crystallization of bacterial cellulose.

Recently, pea xyloglucan has been shown to disrupt the assembly of the
bacterial cellulose ribbon.25 Fibrils of cellulose synthesized in the
presence of xyloglucan had a smaller diameter than the control fibrils. The

interference of xyloglucan was suggested to be similar to that of CMC. The

interference would be at a level in the assembly process where the fibril
formation was not influenced and the further assembly of fibrils to a ribbon
was prevented.

SORPTION OF HEMICELLULOSES AND GUMS BY CELLULOSE

To complement the discussion of associations with bacterial cellulose in

the preceding section, sorption by plant celluloses will be examined here.

Xvyloglucans

Hayashi et al. 25 studied the sorption of 125I-labeled pea xyloglucan by

pea cellulose. The xyloglucan used in the study had a molecular weight of
3.3x10 5 Daltons, determined via gel filtration. They found that sorption

equilibrium was reached within four hours at 400 C, showing saturation at

 11
approximately 5 ug xyloglucan per 100 ug cellulose (5% w/w). Sorption was
inhibited above pH 6; 50% inhibition was observed at pH 7.

In a study by Valent and Albersheim 26 , 14C-labeled seven- and nine-sugar

fragments of sycamore xyloglucan were sorbed by cellulose powder. This
sorption was induced by using aqueous ethanol or acetone as the medium.
Equilibrium-sorption was studied after 2.5 hr and determined from the amounts
of xyloglucan remaining in the supernatants. Increasing the temperature
caused a decrease in sorption. Changing the pH between 2 and 7 in 60-65%
solvent had no effect on the sorption of the sugar fragments by the cellulose.

Seed xyloglucan sorption by cotton linters, at constant pH and

temperature, was studied by Molinarolo at IPC. 27 The molecular weight of the

xyloglucan was 1.6x10 6 Daltons as determined with gel filtration. By
colorimetric analysis of the supernatants, it was determined that sorption
equilibrium was reached after approximately 24 hr at 25C. The xyloglucan was
sorbed in monolayers, with a maximum specific sorption of 3.9 mg xyloglucan
per gram of cotton (0.39%). The cotton linters had a hydrodynamic specific
surface area of 1.06 m2/g. The xyloglucan branches were thought to interfere
with any association between xyloglucan molecules, whereas the xyloglucan
backbone could associate with the cellulose molecules.

The sorption of narrower molecular weight fractions of xyloglucan were

also measured in the same study. 27 These three fractions had molecular

weights of 2x10 5, 7x10 5, and 1.9x10 6 Daltons. The effect of molecular weight
on the maximum specific sorption was insignificant over the range
investigated.

The experimental conditions for these three studies were very different
 12
in terms of cellulose substrates, temperatures, xyloglucans, etc., making it
difficult to compare the xyloglucan sorption behaviors. Hayashi et al.25
showed that pea cellulose sorbed 3.3 times as much pea xyloglucan as cotton
fibers did, explaining in part the differences in the sorption behaviors.

Xylans

Sorptions of xylans by cellulose substrates have been investigated by

several authors.25'2830 Walker28 studied the influence of uronic acid content
on cotton linter sorption by partially reducing a spruce 4-0-methyl-
glucurono-arabinoxylan. The sorption of unreduced xylan was 9.1 mg per gram
of cotton linters. A 40% and 60% reduction of the acids increased the
sorption to 12.0 and 16.1 mg xylan per gram of linters, respectively. The
cotton linters had a specific surface of 1.3 m2/g. Clayton and Phelps 29
investigated the rate of sorption of birch xylan on a-cellulose (see also the
next section about sorption of mannan). Both authors concluded that even a
small number of uronic acid groups present on the xylan would reduce the

sorption by cellulose.

In a study by Mora et a7.30, xylan was sorbed by a beaten kraft birch

pulp and a reed cellulose. Gold-enzyme labeling of the native xylans showed
an even distribution as a thin layer along the cellulose fibril, whereas the

redeposited xylan had formed heavily labeled clusters. The association

between xylan molecules was preferred over the cellulose-xylan association.

Hayashi et al. 25 compared the affinities of xylan and xyloglucan for

cellulose. The presence of a larch xylan (in a 10-fold excess over

xyloglucan) decreased the sorption of pea xyloglucan by a pea cellulose by

 13
about 50%.

Glucomannans

In two early studies29,31, mixtures of glucomannans and xylans were

sorbed by cellulose. Most 31 used four '4C-labeled fractions of slash pine for
sorption by a-cellulose. Each fraction was a mixture of glucomannans and
xylans. Sorption equilibrium was not reached within 10 days and the sorption
was considered irreversible. The fractions with increased sorption rates had
higher contents of mannose, lower uronic acid contents, and higher molecular
weights. Clayton and Phelps 29 sorbed a 14C-labeled spruce glucomannan and a
tritium-labeled birch xylan upon an a-cellulose, both individually and in a
mixture. In both cases, the glucomannan was about twice as readily sorbed as
the xylan, in spite of lower degree of polymerization and conformational
similarity of the xylan chain with cellulose. In a mixture, the
hemicelluloses competed for the sorption sites and the rate of sorption of
both hemicelluloses was reduced.

The influence of acetyl groups on glucomannan sorbed by cotton linters

was studied by Laffend and Swanson. 32 A reserve form of acetylated

glucomannan was isolated from orchid bulb flour, enzymatically degraded to the

molecular weight of a pine glucomannan, and sorbed by cotton linters. The

removal of the acetyl groups led to a decreased water-solubility, a higher

rate of sorption, an earlier mass equilibrium, and a higher specific sorption.

Monolayer sorption was theorized for the acetylated glucomannan with a maximal

specific sorption of 7.5%. Deacetylated glucomannan, on the other hand,

sorbed in ever-increasing amounts as its initial concentration was increased.
 14
Gums

The sorption of gums by cellulose has been studied extensively because

of their importance as beater additives and strength enhancers in papermaking.
Russo and Thode33 investigated the sorption of locust bean gum (a galacto-
mannan, with a ratio of galactose to mannose of approximately 1:4) by a
bleached sulfite pulp. The influence of the mannose-to-galactose ratio was
studied by Dugal and Swanson 34 using a modified guar gum sorbed by a cotton
comber pulp. Increasing the mannose-to-galactose ratio from an initial 1.49
to 2.82 improved the fiber bonding. Going beyond this optimum mannose-to-
galactose ratio did not increase the strength properties and actually
decreased them in some cases.

In summary, the sorption of hemicelluloses and gums by cellulose is a

multifaceted and complex association. The amount sorbed and the sorption
behavior of a polymer depend on parameters such as the polymer composition

(including branching pattern, size, and functional groups), the adsorbent

(e.g., surface area and charge), the solubility in the selected solvent, and
the temperature.

CELLULOSE POLYMORPHISM AND MOLECULAR CONFORMATIONS

Cellulose has been known for a long time to appear in four polymorphs
which are distinguishable by their different x-ray diffractogram patterns.

The best characterized forms are cellulose I, or crystalline native cellulose,

and cellulose II which can be produced from native cellulose by treatment with
alkali. The two other polymorphs are cellulose III and cellulose IV.
 15
Cellulose I and II

In the classical hypothesis of the structures of cellulose polymorphs,

the major difference between cellulose I and II is a difference in polarity of

the cellulose chains. The backbone conformations of the chains themselves are
essentially identical. Cellulose I has parallel chains whereas cellulose II
has alternating antiparallel chains. Evidence for the classical hypothesis is
taken mainly from studies using x-ray diffraction followed by calculations of
packing energies and conformations.35-38 Other authors claim that the
complexity of the cellulose molecule and the relatively small number of
reflections in a diffraction pattern make it inappropriate to solve the

structures of cellulose based solely on x-ray diffraction data.39,40 In a

comparison of different conformation models, the conclusions regarding one

packing mode or another depend on the set of diffraction data used. 4 1 By

using more flexible constraints for the calculations, the conformation of
cellulose II has been refined to a parallel chain model.42

Another hypothesis was developed when spectroscopy became more widely

used for cellulose characterizations. The techniques used are infrared and
Raman spectroscopy and solid state 13C nuclear magnetic resonance spectroscopy
enhanced with cross-polarization and magic angle spinning (CP/MAS).

Additional information from these techniques indicated that conformational

differences of the heavy atoms carbon and oxygen in the cellulose chains are
responsible for the cellulose I and II polymorphs. Atalla 4 3 concluded from
Raman spectra, in combination with potential energy calculations, that only
two stable conformations were allowed. More than one stable pattern of

intermolecular hydrogen bonding has been found to be consistent with each

conformation.44
 16
Hierarchical organization of cellulose structures

Atalla 8,45 has proposed a hierarchical organization of cellulose

structures, with different levels related to each other. The primary level of
organization is the chemical pattern of covalent bonds, in this case B-(1-4)-
linked anhydroglucose units. The conformation of the anhydroglucose chains in

space is the secondary level of organization. In the cellulose molecule, this

level includes variations in the position of the C6 primary alcohol group, in

the dihedral angles defining the glycosidic linkages, and in the internal
dihedral angles defining the conformation of the pyranose rings. The
arrangement of the glucan chains relative to each other is described by the
tertiary level of organization. The divisions between the levels are not
absolute because of an interdependence: the packing of chains depends on the
shape of individual chain molecules, and the forces of intermolecular
interactions determine in part the equilibrium conformation of individual
chains.

The two conformations of crystalline cellulose deduced from Raman

spectroscopy and energy calculations are examples of structures on the
secondary level of organization. They are called ki and kil after their
predominance in cellulose I and II, respectively. 9 The corresponding form for
the disordered portion of cellulose is represented by ko. Figure 1 shows
schematics of the conformations proposed for kI and kH
1. The key difference

between the conformations k1 and kii is the presence in k I of a bifurcated

intramolecular hydrogen bond associated with every other glucosidic bond in

addition to a simple hydrogen bond. kil in contrast, has two simple hydrogen
bonds per dimer.10 A corollary to this model is that adjacent anhydroglucose

units are nonequivalent, and, thus, that the anhydrocellobiose is the

 17
repeating unit in cellulose.

Legend:
* a Carbon
o I Oxygen
* a Hydrogen
- a Covalent bond
...... Hydrogen bond

Figure 1. Schematic representation of conformations k, and ki, (from

reference 10).

In an early investigation of native celluloses from a variety of

sources, it was found that the infrared spectra from algal and bacterial

celluloses differed from those of regenerated (from cellulose IIIi), ramie,

and cotton celluloses.46 Solid state 13C NMR has been used more recently for a
 18
closer investigation of celluloses from these sources. 474 8 Based on the NMR-
studies, native celluloses are proposed to be composites of two distinct
crystalline forms, I. and I. These forms are examples of structures on the
tertiary level. The two forms of cellulose I crystallinity have different
intermolecular hydrogen bonding patterns, but the conformations of the heavy
atoms are similar. The crystalline component of every native cellulose

studied was a mixture of these two forms, but I. was predominant in celluloses
from bacteria and noncharophycean algae, and IB was the major form in higher
plant celluloses. Celluloses rich in IQ that have been subjected to acid
hydrolysis 10, steam annealing 49 , or solid-state chemical transformation have a
crystalline structure similar to the I,-type, indicating that the I, form is
more stable or more resistant than the Ia form. Cellulose triacetate 50 and
cellulose IIII51 were the intermediates in these solid-state transformations.

In a study of the solubility of celluloses in an amine-based solvent, it

was found that mercerized cellulose (cellulose II) dissolved in this solvent,
but regenerated cellulose II did not. The x-ray diffractogram patterns of the
mercerized and regenerated celluloses indicated that the heavy atom lattices

were similar, if not the same. It was suggested that the patterns of
intermolecular hydrogen bonding are different in these two forms of cellulose
II, similar to the crystalline cellulose forms Ia and I. 44

Analysis techniques for probing structural levels of cellulose.

Raman spectroscopy is primarily sensitive to the vibrations of the

skeleton in the cellulose molecule and less sensitive to the packing of the
molecules in a three-dimensional lattice. Raman spectroscopy is therefore a

useful technique for probing the secondary level of organization. In

 19
contrast, solid state 13C NMR probes both the secondary level and the tertiary
level. Measurements using high resolution enhancements of NMR are sensitive
both to conformation of the molecules and to sites which are nonequivalent
within the repeating unit of the molecule. X-ray and electron diffraction
techniques are most sensitive to regularity in three dimensional lattices,
and, therefore, probe at the tertiary level of structural organization.

THE SIGNIFICANCE OF A 1-(1-4)-LINKED POLYSACCHARIDE BACKBONE FOR SORPTION

The association between cellulose and polysaccharides with different

backbone configurations has been tested in two studies.4 '25 Tables 1 and 2
show the results of incubating bacterial cellulose in the presence of
cellulose derivatives and other compounds. Of the bond types and
configurations studied, only the polysaccharides with a B-(1-4)-linked
backbone are capable of disrupting the cellulose ribbon assembly. Hayashi et
a7. 25 studied the sorption of radioiodinated pea xyloglucan onto pea cellulose
in the presence of a 10-fold excess of a variety of other polysaccharides.

Only a xylan with a B-(1-e4)-linked backbone competed with the xyloglucan for
the sorption sites, thereby decreasing the sorption of the xyloglucan (Table
3). The special backbone linkage was not the unique feature of the xylan in
this study; it was also the only pentosan, and, together with pectin, the only

polysaccharide with uronic acids. 52

THE ADDITIVES USED IN THIS THESIS AND THEIR RELATION TO PLANT CELL WALLS

Xvloglucan

Xyloglucans are found in immature and storage tissues in many plant

 20

Table 1. Cellulose derivatives tested for alteration of bacterial

cellulose.4

Additive Source Structural feature Effect

carboxymethyl 1 )-(1-4)-, R=-CH 2COOH loose coils of fibrils
methyl 2 p-(1-4)-, R=-CH 3 occasionally splayed,
coherent but no twist
hydroxy- 2 -(1-4)-, R=-CH3, coherent but no twist,
propylmethyl -CH2CH(CH )OCH3, occasionally splayed
or -CH CHtOH)CH
hydroxyethyl 1 4-(1 R=-CH2CH 2 OH
4)-, coherent but no twist,
or -CH CH2OCH 2CH 2OH or no alteration
hydroxypropyl 1 8-(1-4)-, R=-CH 2CH(OH)CH3 no alteration, or co-
or herent but no twist
-CH2CH(CH 3)OCH 2CH(OH)CH 3
Sources: (1) Hercules Inc., Wilmington, DE; (2) Dow Chemical Co.,
Indianapolis, IN.

Table 2. Other compounds tested for alteration of bacterial cellulose.4

Additive Source Structural feature Effect

xanthan gum 1 1-(1-4)-, R=-6-0-Ac- no alteration
Manp-(2,4)-GluA-fi-
(1-4)-Manp-2,4-Pyruvate
cellobiose 2 P-(1-4)-Glup coherent no twist,
or no alteration
xylan 2 P-(1-4)-Xylp highly splayed fibrils
pectin 2 a-(1-4)-GalA slight alteration
starch 2 a-(1-4)-Glup no alteration
laminarin 2 -- (1-3)-Glup no alteration
mannan 2 a-(1-6)-Manp slight alteration
agar 3 alternating a-(1-3)- no alteration
and P-(1-4)-bonds,
>3 sulfated D/L galactans
PEG 4 H(OCH2 CH2 )nOH normal ribbons and bent
and fractured ribbons
glycerol HOCH 2 CHOHCH 2OH no alteration
Sources: (1) Kelco Div., Merck & Co., Clark, NJ; (2) Sigma Chem. Co.,
St. Louis, MO; (3) Wilson Diagnostics; (4) Fischer Scientific.
 21

Table 3. Competitive binding of polysaccharides and xyloglucan to pea

cellulose.25

Additional Relative binding Structural feature of

polysaccha- activity of other polysaccharide
ride xyloglucan
None (only XG) 100%
Bacterial glucan 103% #-(12)-Glup
Laminarin 99% f-(1l3)-Glup
Pachyman 101% m-(1l3)-Glup
Pustulan 99% 0-(1l6)-Glup
Lichenan 92% mixed #-(1-3)- and
P-(1-4)-Glup
Xylan 47% 0-(1-4)-Xylp
Arabinogalactan 107% 3-(1-3)-Galp
Pectin 108% a-(14)-GalA

species. Two different kinds of xyloglucans have been identified, those from
seed walls and those from primary cell walls.

The seed xyloglucans function as reserve food for developing seeds and

occur in large amounts (40-50%). They are generally easily isolated.

Xyloglucans from different seed sources have similar repeating patterns and
sugar composition. The repeating units are often hepta- to decasaccharides.
D-glucose, D-xylose and o-galactose are the common sugars. Xyloglucan from
seeds of tamarind (Tamarindus indica) has the sugar composition glucose,
xylose, and galactose 4:3:1. The xyloglucan backbone is a B-(1-4)-linked
glucan substituted in the sixth position with a-xylose, either singly or with
terminal galactose residues, linked B-(1-2)- to the xylose.53 The xy-loglucan
used in this thesis was isolated from tamarind seeds.

The other group of xyloglucans has been isolated from primary cell walls
 22
of young plants. The levels of xyloglucan in primary walls are comparable to
those of cellulose.54 In contrast to the seed xyloglucans, the xyloglucans
from cell walls are very hard to isolate; strong acid or alkali is required.
Typical sources include pea stem55, cotton56, and suspension cultured cells of
sycamore57 and loblolly pine58 . Such cell-wall xyloglucans often contain
terminal fucose units on the side chains in addition to glucose, xylose and
galactose.

Much research has been devoted to understanding the functions of

xyloglucan in cell walls. Xyloglucan is thought to be involved in the cell
wall expansion necessary for cell growth.59 It has been speculated that
oligosaccharides from xyloglucan, especially a nonasaccharide, have biological
activity.60

XYlan

Angiosperms contain glucuronoxylan as the major hemicellulose (15-30% of

dry wood). The bulk of the xylan is present in the secondary cell walls. 6 1

The backbone of B-(1-4)-xylose units has side chains of methylglucuronic acid

units or acetyl groups. In birch, the unit next to the reducing xylose end

group is a galacturonic acid, which is linked to a a-L-rhamnose, which, in

turn, is connected to the xylan chain. Birch was the source for the xylan

used in this thesis.

In gymnosperms, xylan constitutes 5-10% of the wood. 62 It is similar to

the xylan in angiosperms, but the xylan backbone is also substituted with a-
L-arabino units. The xylan content is lowest in the middle layer of the
secondary wall (S2) and higher in the S1 and S3 layers.
 23
Mannan

Mannans occur in gymnosperms in the form of a galactoglucomannan which

constitutes about 20% of the wood. It has a main chain of alternating o-
mannose and D-glucose units and side chains of D-galactose units and O-acetyl
groups. Mannose-containing hemicelluloses are concentrated in the middle
layer of the secondary cell wall. 63 Angiosperms are composed of 2-5% of P-
(1-4)-glucomannan.62

Mannan homopolymers can be found in seeds such as ivory nut, which was

the source of the mannan used in this thesis. The main chain is B-(1-4)-
linked mannose units substituted with galactose units. 64 There are
indications that glucose is incorporated in the main chain and that some
mannose units are a-linked. 65 A few (1-6)-linkages presumably are also
present.66

Carboxvmethyl cellulose (CMC)

CMC or cellulose gum is not a natural product of plants. It is an

anionic, water-soluble cellulose carboxymethyl ether, produced by the reaction

of alkali cellulose with sodium monochloroacetate.67

THESIS OBJECTIVE

The main objective for this study was to qualitatively investigate the
influence of B-(1-4)-linked hemicellulose-like polysaccharides on the

structure of cellulose synthesized by Acetobacter xylinum. The second

objective was to investigate the influence of isolation procedures on the
bacterial cellulose produced in the presence of these additives.
 24
The objective was to study the influence on cellulose structure during
the early stage of cellulose aggregation. The bacterium Acetobacter xylinum,
rather than plant cells, was the producer of the cellulose, creating a model
system of the cellulose biogenesis in plants which excluded lignin and other
wood components. A plant cell-wall environment was simulated by having a
hemicellulose present in the medium while the bacterial cellulose was
synthesized and aggregated. The hemicellulose-like polysaccharides
investigated were mannan, xylan, and xyloglucan. Carboxymethyl cellulose, a

cellulose derivative, was included in this investigation to correlate this

study with work previously reported in the literature.

The investigation was based on the hypothesis that the presence of

hemicelluloses influences the aggregation, but not the polymerization, of
cellulose. The mode of interaction depends on properties of the

hemicellulose, such as chain length, branching, and electrical charges. The

effects on aggregation may be expressed as changes in degree and nature of
crystallinity, backbone conformation, and/or morphology. The anticipated
alterations of the cellulose structure were evaluated via x-ray diffraction,
Raman spectroscopy, and electron microscopy.

EXPERIMENTAL PLAN

The experimental work of this thesis consisted of three parts. The

first part involved the discovery and preparation of suitable water-soluble

polysaccharides for use as additives. It was essential that the additives

were completely water-soluble to control the conditions for the aggregation of

cellulose. A low molecular-weight fraction of mannan was isolated from ivory

nut. Water-soluble xylan was prepared from a previously isolated birch xylan.
 25
Carboxymethyl cellulose and xyloglucan were obtained in water-soluble forms.

In the second part of the experimental work, bulk samples of cellulose

were produced in the presence of an additive. A method with resting

nondividing cells was employed to prepare the cellulose bulk samples. The
medium was stirred gently during incubation, allowing the additive to be mixed

homogeneously. The cellulose-additive products were subjected to a series of

alkali extractions to get an estimate of the strength of the association
between the cellulose and additive. The second purpose of the series of

extractions was to investigate the influence of isolation on the structure of

cellulose. The cellulose samples were evaluated with x-ray diffraction and
Raman spectroscopy. Profile fitting and conformational component resolution

were performed on the diffractograms and Raman spectra, respectively, to

further quantify the results.

The influence of the additives on the morphology of cellulose fibrils

was investigated in the final part of the work. Specimens for transmission

electron microscopy were prepared from bacterial cellulose fibrils produced in

the presence of an additive. The dimensions of the fibrils were measured with

the help of an image analyzer.

 MATERIALS AND METHODS

THE ADDITIVES TESTED FOR THEIR INFLUENCES ON BACTERIAL CELLULOSE

In the routine experiments, the additives tested for their influences on

bacterial cellulose were CMC, xyloglucan, mannan, and xylan. Each additive
will be described in further detail in the following sections. In addition,
xanthan gum and salicin were used in two experiments.

CARBOXYMETHYL CELLULOSE

Carboxymethyl cellulose (CMC) was of the purest food grade (7MF)

supplied by Hercules, Inc. It was completely water-soluble and was used

without further purification. The degree of polymerization was approximately

500, the molecular weight was 105 , and the degree of substitution was 0.65-
0.85.67 X-ray diffraction of CMC gave an amorphous pattern.

XYLOGLUCAN

Xyloglucan was kindly provided by Dr. S.L. Molinarolo. 27 The xyloglucan

had been isolated from Tamarind seeds (Tamarindus indica) by hot-water

extraction, and extensively characterized.

The molecular weight of the xyloglucan was determined to be 1.6x10 6 by

size exclusion chromatography. Characterization of the xyloglucan identified

97% of the components (Table 4). Acetate and pyruvate groups could not be
detected in the xyloglucan but the methoxyl content was 0.44%. X-ray
diffraction showed the xyloglucan to be amorphous.

26
 27

Table 4. Chemical composition of xyloglucan. 27

Nitrogen, % <0.1
Carbohydrate Analysis, %
Glucose 49.8
Galactose 15.4
Mannose 1.5
Xylose 28.8
Arabinose 1.4
Total carbohydrate, % 96.9
Ash, % 0.2

MANNAN

Finely ground ivory nut (Phytelephas macrocarpa) was obtained from Mr.
J. Becher at IPC. A water-soluble fraction of mannan was isolated according
to the following procedure.68'69

Isolation and purification

The ivory nut flour (165 g) was extracted with chloroform-methanol (1:1,
2000 ml) in a Soxhlet apparatus for 12 hours. The extracted residue was air
dried overnight. Repeated extractions yielded 159-163 g (96-99%).

Extracted ivory nut flour was bleached by adding the flour (56 or 100 g)
to a warm (30'C) solution of sodium chlorite in water (250 g NaClO 2 in 2500 ml
H20 was used for 100 g nut flour). The temperature was raised to 35C and
glacial acetic acid (250 ml for 100 g nut flour) was added. The reaction
flask was placed in a 300C water bath for 24 hours. The bleached ivory nut
flour was collected by centrifugation at low speed, and filtration of the
 28
supernatants on a fine sintered-glass funnel. The solid phase was washed with

water on the funnel until free from yellow liquor and then washed with
methanol. The bleached ivory nut flour was dried in vacuo at 40C to yield
51 g (91%) or 87-89 g (87-89%), respectively, of a white solid.

The filtrate of acid-chlorite liquor was decomposed before it was

discarded, in order to reduce the odor of C10 in the liquor. Aqueous sodium
hydroxide solution was kept in the filtration flask to neutralize the chlorite
liquor after it was separated from the bleached ivory nut flour.

Alkaline extraction of the ivory nut flour was the next step in the
isolation procedure. The bleached flour (35 g) was placed in a solution of
sodium hydroxide (700 g of 6.0% NaOH) and nitrogen gas was bubbled through the

mixture for 30 minutes. The mixture was stored under refrigeration (31C)
for three days and was shaken occasionally. The mixture was centrifuged and
the supernatants were filtered through a fine sintered-glass filter. The
solids were washed three times with water by dispersing the pellets in
distilled water (400-500 ml), centrifuging, and filtering, before the solid
residue was discarded. The mannan was precipitated by adding four volumes of
ethanol to the alkali filtrate and the three wash-waters. The precipitates
were settled overnight at 31C and were collected by filtration the next day
(giving PEtl).

The combined precipitates (PEtl) were extracted four times with water

and the mannan was again precipitated from the water with ethanol (added
ethanol in four times the volume). The precipitates were settled overnight at

31C and were collected by filtration the next day (giving PEt2). The
precipitates were collected by filtration and extracted four times with water.
 29

The water-extracts were combined and neutralized with ion exchange resin
(Aldrich Amberlite IR 120+ washed with distilled water until no further change
of pH), and freeze-dried to yield a white fluffy material (UnFCM). A portion

of this material (UnFCM) was again fractionated by dispersing it in water and

centrifuging. The supernatants were filtered through a fine sintered-glass
funnel, and Millipore filters with 1.2 Am and 0.8 pm pore sizes. The filtrate

was freeze-dried to yield crude mannan (CM) (3.7 g).

Salts were removed from the crude mannan (CM) by ultrafiltration.

Portions of the crude mannan were dissolved in distilled water, and placed in
an ultrafiltration cell of 200 or 400 ml volume. A membrane with molecular-

weight cut-off of 500 was used (Amicon, YC05), and the cell was pressurized
with nitrogen gas to 55 psi. The inorganic material and the smallest mannan

oligomers were filtered through the membrane, leaving behind the higher
oligomers of mannan in the cell. The filtration was continued until the

conductivity of the filtrate had stabilized. This took approximately six

days. The fraction retained in the cell was filtered through a fine sintered-

glass funnel and Millipore filters with 1.2 jm and 0.8 im pore sizes, and the

filtrate was freeze-dried. The mannan was dissolved in water, centrifuged at

11,900g for 60 minutes, and freeze-dried to yield a water-,soluble mannan

(WSM1).

Another method for desalting the mannan was tested with improved

results. (However, this method was not employed for the mannan added to the
bacterial growth medium.) The crude mannan (CM) was extracted four times with

aqueous ethanol (76%) and the ethanol was separated by centrifugation. The

mannan pellets were dispersed in water, and the residual ethanol was removed
in vacuo. The precipitate was separated by centrifugation, and the
 30
supernatants were freeze-dried. The mannan was dissolved in distilled water
and additional salts were removed by ultrafiltration for 1.5 days. The

retained fraction was filtered through a fine sintered-glass funnel and

Millipore filters with 1.2 um and 0.8 um pore sizes, and the filtrate was
freeze-dried. The mannan was again dissolved, centrifuged at 26,800g for 60
minutes, and the supernatant was freeze-dried, yielding a water-soluble
fraction of mannan as a white fluffy material (WSM2). The advantage of this
second method of desalting was the time saved in the ultrafiltration step.
Another advantage of the shorter dwell time was the decreased risk for
contamination during ultrafiltration.

X-ray diffraction of the water-soluble mannan (WSM1) was performed.

Liquid state 13C NMR was performed by Spectral Data Services, Inc., Champaign,
IL. Carbohydrate analysis was performed on the mannan by the IPC Analytical
Department using the TAPPI method 70, and by the Analytical Department at the
U.S.D.A. Forest Products Laboratory, Madison, WI. At FPL, a method
(unpublished) was used where the carbohydrates were hydrolyzed with 72% H2SO4,
the acid solution was diluted, and the monomers were separated on a HPLC-
column.

XYLAN

Xylan isolated from birch was kindly provided by Dr. N.S. Thompson. The

xylan had been characterized earlier. 7 1 In order to obtain a water-soluble

fraction, the xylan was extracted with sodium hydroxide solution (102 ml of

0.01 N NaOH for 1.97 g xylan) overnight at 31C. The xylan was centrifuged
at 26,800g for 30 minutes. The pellets from centrifugation were dispersed in

sodium hydroxide solution (80 ml of 0.01 N NaOH) and centrifuged at 26,800g

 31
for 45 minutes. The pellets were again dispersed in sodium hydroxide solution
(50 ml of 0.01 N NaOH) and centrifuged at 26,800g for 45 minutes. The
combined supernatants were neutralized with ion exchange resin (Aldrich
Amberlite IR 120+, washed with distilled water until no further change of pH),

and freeze-dried. The freeze-dried material was extracted with water three
times, with centrifugation and filtration of the supernatants through a fine
sintered-glass funnel after each time. The filtrates were freeze-dried and
the combined water-soluble products (WSX) yielded 1.15 g (58.3% of 1.97 g) of
a white fluffy material.

Carbohydrate analysis of the xylan (WSX) was performed by the IPC

Analytical Department using the TAPPI method.70 Liquid state 13C NMR of xylan
(WSX) dissolved in D20/NaOD was performed by Spectral Data Services, Inc.,
Champaign, IL, and by Dr. L.L. Landucci at the U.S.D.A. Forest Products

Laboratory, Madison, WI, on xylan (WSX) dissolved in DMSO-d6 .

STRAINS OF Acetobacter xylinum

The strain of Acetobacter xylinum used in this study was kindly provided
by Dr. C.H. Haigler. The strain was originally of A. aceti ss. xylinum from
the American Type Culture Collection (ATCC 23769).22 It was a faster
cellulose producer than a strain of A. xylinum used initially, which had been
obtained directly from the American Type Culture Collection (ATCC 23769).

GROWTH MEDIUM

The medium developed by Hestrin 72 was used for the production of

bacterial cells. It consisted of 2.0% (w/v) D-glucose (Mallinckrodt), 0.5%

 32
bactopeptone (Difco), 0.5% yeast extract (Difco), and 0.1% dibasic potassium
phosphate (Baker). The pH was adjusted to 6.5 with hydrochloric acid. Solid
medium was prepared by adding 1.5% (w/v) Bacto-agar (Difco) to the liquid
medium before autoclaving. The medium was sterilized by autoclaving for 20
minutes at 120C and 15 psi.

INOCULATION PROCEDURES AND INCUBATION CONDITIONS

Aliquots of 5 ml of growth medium were used in 4 dram (z16 ml) screw-

cap vials for culture storage and transfers. The aliquots were inoculated
with colonies from solid medium or with 0.5 ml of cell suspension from a
culture previously incubated. Petri dishes (100 x 10 mm) with 25.ml growth
medium were used for growing cellulose pellicles and larger quantities of
cells. The dishes were inoculated with 2-5 ml of cell suspension. The
suspensions were prepared from cellulose pellicles which were shaken
vigorously and squeezed with forceps before the desired volume of cell

suspension was withdrawn from the medium. The vials and dishes were incubated
at 281 0 C for 24 hr to seven days.

In order to maintain a healthy and fast cellulose-producing culture, the

culture was streaked periodically on solid medium in Petri dishes. Colonies
growing high above the medium surface were selected for further inoculations.

Large flat colonies developed occasionally in the strain. These types of

colonies formed a soft pellicle.

PRELIMINARY WORK ON INCUBATION CONDITIONS

Initially, the influence of additives was studied using static

 33

conditions of incubation. The incubation lasted for three to seven days.

Cellulose pellicles were formed on these static cultures and the yield of
cellulose was usually good. However, no effect on the cellulose structure

could be detected. Because the production of new cellulose occurs on the

upper surface of the pellicle 17 , the possibility that the additive would be
depleted in the immediate surrounding of productive bacterial cells was a
concern. To provide for a more homogeneous growth environment, the method
using resting cells in shaking cultures was adopted.

INCUBATION SOLUTION FOR RESTING CELLS

The procedure and incubation solutions for resting cell production of

cellulose followed Haigler 73 and Benziman 74. The incubation medium for resting

cells was 50 mM phosphate buffer with 2% (w/v) D-glucose. Equal volumes of 50

mM monobasic (MCB) and dibasic (Baker) potassium phosphate solutions were

mixed, and the pH was adjusted to 6.8-7.0 with hydrochloric acid or sodium
hydroxide. The buffer was sterilized by autoclaving for 20 minutes at 120C

and 15 psi before glucose was added. The buffer turned bright yellow if

glucose had been added before autoclaving.

The additive was dissolved in double concentration in buffer. Glucose

was added in double concentration (4%) after the additive was dissolved and

right before inoculation. The standard protocol for a resting-cell incubation

consisted of six 125 ml Erlenmeyer flasks, each with 25 ml buffer containing

additive and glucose in double concentrations before addition of an equal
volume of cell suspension. To prevent the additives from being metabolized,

an excess of glucose was used in the buffer. The final concentration of

glucose was 2% instead of 40 mM which is commonly used in the literature. The

 34
final concentrations of the additives were 0.05%, 0.1%, 0.4% or 0.8%, after
adding the cell suspensions.

CELLULOSE PRODUCTS FROM RESTING CELLS

Each cellulose sample was prepared from the bacterial cells collected
from nine Petri dishes which had been incubated for 36 hr. The cultures used
to inoculate the nine Petri dishes had been transferred not more than three
times after being selected and picked from a solid medium. The pellicles were
gently placed in a big beaker and soaked in four changes of cold phosphate
buffer (50 mM) for 20-30 minutes each time until the yellowish color of the
growth medium had disappeared. The pellicles were kept at +2C between
changes of buffer. The pellicles were carefully blotted on paper towels

between each change in order to remove additional medium.

When the growth medium had been removed, the cells were dislodged from
the pellicles by shaking the pellicles vigorously in an Erlenmeyer flask.
Liquid with additional cells was squeezed out of the pellicles with the help

of forceps. The liquids from the shaking and squeezing were combined and
filtered through two layers of wet cheese cloth to retain any cellulose
remaining with the cells. The cell suspension was kept on ice, especially
after the filtration, to halt the cellulose production. The cell suspension
was diluted with phosphate buffer to 150 ml and divided among the six

Erlenmeyer flasks with dissolved additive, giving a final volume of 50 ml of

buffer and cell suspension in each flask. The flasks with resting cells were
incubated for 3 hr at 28'C and were gentle shaken at 50 rpm. The cellulose
product was collected by centrifugation at 11,900g, 45 minutes at 150C, and
then washed twice by dispersing the cellulose product in distilled water and
 35
centrifuging for 35 minutes before freeze-drying. Repeated preparations

yielded 4-8 mg control cellulose (6 preparations), 6-180 mg of CMC-cellulose

(3 preparations, 1x0.1%, 1x0.4%, and 1x0.8%), 8-12 mg of xyloglucan-cellulose
(7 preparations, 3x0.1%, 3x0.4%, and 1x0.8%), 7-10 mg of xylan-cellulose (4
preparations, 1x0.05% and 3x0.1%), and 5-11 mg of mannan-cellulose (4
preparations, lx0.05% and 3x0.1%). Raman spectra and x-ray diffractograms

were recorded on the freeze-dried cellulose samples.

The amount of cellulose produced by resting cells was very small. It

appeared that the cells quit producing cellulose after two to three hours.
Experiments were performed in which small volumes of the growth medium were
left in the cellulose pellicles providing a possibility for the cells to
divide. This certainly promoted a higher yield of cellulose. However, the

Raman spectra and x-ray diffractogram of the product indicated that cellulose

I was not the exclusive product, and the method was therefore abandoned.

oH of resting cell medium and binding of xyloglucan to cellulose

Hayashi et al.25 has reported that pea cellulose binds 40-50% more pea
xyloglucan at pH 5 than at pH 7. In order to determine if the effect of

xyloglucan on bacterial cellulose structure was influenced by pH, a sample was

prepared in phosphate buffer of pH 5 and xyloglucan (0.1%). The x-ray

diffractogram was recorded after washing the sample with water and freeze-
drying. Comparison of diffractograms of the samples prepared at pH 5 and pH 7
showed no noticeable difference.

The synthesis of cellulose in A. xylinum is accompanied by the formation

of organic acids; acetic acid from ethanol, gluconic acid from glucose, etc.,
 36
and the pH in the medium is thereby reduced." In this experiment, the pH
decreased from pH 5 to 3.9 during the incubation period of 3 hr. In another
sample preparation (with 0.05% mannan in the medium) the pH decreased from pH
6.8 to 6.5.

Celluloses produced in xyloalucan-salicin and xanthan aum

The resting cell method was used to produce cellulose samples from a
medium containing xanthan gum (from Sigma) and a mixture of xyloglucan and
salicin (courtesy of Dr. I.A. Pearl). All three compounds were used in
concentrations of 0.1%. Xanthan gum is produced by the bacterium Xanthomonas
campestris. This polysaccharide has a B-(1-4)-o-glucose backbone.75 Every
other glucose in the main chain has a side chain consisting of an acetyl-o-
mannose unit, a potassium D-glucuronate unit, and a terminal D-mannose unit of
which about half contain a pyruvate unit. Xanthan gum was tested as an
additive because of its structural similarities with CMC. The x-ray
diffraction pattern of the cellulose product from xanthan gum showed the
effect of xanthan gum to be very similar to the effect of CMC.

The second set of additives (xyloglucan and salicin) was tested for any
synergistic effect of a hemicellulose and a lignin-precursor type additive.

Salicin, or 2-(hydroxymethyl)phenyl-B-D-glucopyranoside, is structurally

related to the P-glucosides of the lignin precursors; they are water-soluble

and thought to be involved in the lignification of plant cell walls. The

cellulose from xyloglucan-salicin gave an x-ray diffraction pattern very
similar to the cellulose from xyloglucan. The Raman spectrum did not have any
intense bands in the region corresponding to conjugated or aromatic C-C, C-H
or C-OH bonds (1595 to 1650 cm-1 ) indicating that the salicin was probably
 37

removed during the washing of the cellulose product.

ALKALINE EXTRACTION OF CELLULOSE PRODUCTS

A series of extractions with sodium hydroxide was performed on cellulose

samples produced by resting cells in the presence of 0% or 0.1% of CMC,
mannan, xylan, or xyloglucan. The five samples were extracted once with

distilled water, centrifuged, and freeze-dried. X-ray diffractograms and

Raman spectra were recorded on these samples before the alkaline extractions.
The celluloses were then extracted with alkali according to the following
procedure: The cellulose pellets were wetted with three drops of distilled
water for 10-15 minutes in plastic centrifuge tubes. Sodium hydroxide

solution (8.0 ml) was added, and the celluloses were dispersed with a Pasteur

pipet and stirred occasionally during the 8-9 hr extraction period. The

celluloses were pelleted by centrifugation at 4,300g, 60 minutes at 15C, in a

Sorvall high speed centrifuge. The celluloses were washed twice with about 35
ml distilled water, centrifuged after each wash at 26,800g for 60-90 minutes,

and freeze-dried. The extractions were performed with 0.01 N NaOH, 0.1 N
NaOH, and 1.0 N NaOH diluted from an Acculute solution. X-ray diffractograms

and Raman spectra were recorded after each extraction.

Before the series of alkaline extractions was performed, two sets of

preliminary experiments were conducted. Three control samples were extracted

with 1.0 N NaOH. One sample of cellulose produced in the presence of 0.1%

xyloglucan was extracted first with 0.1% NH40H and then with 1.0 N NaOH. The

extraction was performed with ammonium hydroxide instead of sodium hydroxide

because the ammonium hydroxide could be removed by evaporation. However, the

concentration of ammonium hydroxide in solution could be changing because of

 38
its volatility. Therefore, the subsequent extractions were performed

exclusively with sodium hydroxide. The Raman spectrum of the extracted

xyloglucan-cellulose and the x-ray diffractograms of all extracted samples
were very similar to the spectrum and diffractograms of corresponding samples
in the alkaline extraction series.

X-RAY DIFFRACTION

X-ray diffractograms were recorded on a Philips powder diffractometer

(APD 3720) using CuK. radiation filtered with a crystal monochromator. The
samples were analyzed as pellets pressed at about 800 psi. The pellets were
analyzed on a low background holder. The radiation was generated using

conditions of 45 kV and 40 mA. The diffractograms were recorded from 5 to 30

degrees 20. The step size was 0.02 degrees and the count time 1.00 second,
giving a scanning rate of 1.2 degrees of 20 per minute.

PROFILE FITTING OF X-RAY DIFFRACTOGRAMS

The x-ray diffractograms were resolved into individual peaks using a

profile-fitting computer program obtained from Philips. 76 This resolution

model fitted profiles to the experimental diffraction profile by iteration,

using a nonlinear least-squares algorithm. Four peaks were used in the

fitting of each diffractogram, three peaks for the cellulose crystalline peaks
(101, 101, and 002 peaks) and one peak accounting for the amorphous background
of the diffractogram. The step size in the fitting was 0.04 degrees which,

being twice the size of the scanning step, was by default the smallest step
possible.
 39
Results from a profile fitting included a fitted total profile for the
diffractogram, profiles for the resolved peaks, a fitted background, and a
residual indicating the accuracy of the fit. In addition, the full width at
half maximum (FWHM), the intensity, and the position were obtained for each
fitted peak. The peak width was measured at half maximum intensity of the
fitted individual peaks which had been corrected for the background intensity.

RAMAN SPECTROSCOPY

The Raman system was a Jobin Yvon Ramanor HG2S monochromator with 514.5
nm radiation of an argon ion laser (Spectra-Physics model 2025) as the

excitation source. The incident laser power was 40-70 mW. The slit width was
400 um. The samples were analyzed as pressed pellets and the scattered Raman
emission was collected perpendicular to the excitation radiation. This
instrumental set-up provided for a macro-mode acquisition of the Raman light
from the samples which decreased effects attributable to orientation of the
cellulose samples. The data acquisition system consisted of a photo

multiplier detector (RCA C31034-A05) and a Tracor Northern TN-1500 digital

signal analyzer. The scanning motor in the monochromator was computer
controlled. The scan rate was 60 cm-1 per minute, and the scanning range was

250-3700 cm-1. Multiple (15-25) scans were recorded for each spectrum to
decrease distortion of the spectrum due to any drift in the laser power during

a single scan. The samples were highly fluorescent in the region scanned, and
the spectra were flattened to correct for contributions from the fluorescent

background.
 40

CONFORMATIONAL COMPONENT RESOLUTION OF RAMAN SPECTRA

The low frequency regions (250-550 cm-1)

of the Raman spectra were
resolved into linear combinations of spectra representative of the

conformations kI, ki, and ko. The original Raman spectra of bacterial
cellulose samples after the extractions with 0.01 N, 0.1 N, and 1.0 N NaOH
were subjected to this resolution. The standard procedure of preparing the
spectra for analysis included smoothing, flattening, and normalizing.77 The
spectra of celluloses extracted with 0.01 N NaOH were smoothed ten times. The
spectra recorded after the second and third alkali extractions were smoothed

five times because they had higher signal-to-noise ratios. The smoothed
spectra were flattened in the 250-750 cm-' region. The spectra were
integrated between 250 and 550 cm-1, and the integrated area was normalized to
600,000 by multiplying the spectra with a correction factor.

The resolution of the spectra was accomplished by an iterative search

procedure for the least-squares deviation of a linear combination of the
spectra of standards from the experimental spectra. The computer program used
for this search can be found in Appendix D. The reference spectrum used for

kI was of a bacterial cellulose pellicle purified by boiling in 1% NaOH for

one hour. The spectrum of regenerated microcrystalline cotton cellulose (CF1)

was the reference for kit. For ko, the spectrum of ball-milled
microcrystalline cotton cellulose was used as the standard.

SPECIMENS FOR TRANSMISSION ELECTRON MICROSCOPY

The procedure for preparation of specimens followed Haigler. 1973

'

Bacterial cellulose was produced directly on the grids by floating the grid
 41
upside down in the incubation medium. The specimens were prepared by the
following method: A cellulose pellicle was blotted on a paper towel to remove
excess liquid. The remaining liquid with bacterial cells was squeezed out of

the pellicle with forceps. The pellicle must be freshly produced (one or two
days old) otherwise the cells did not produce cellulose as readily and
elongated cells were formed. The cell suspension was centrifuged in a bench-
top centrifuge and the supernatant was pipetted off, leaving about 0.5 ml
liquid. The pellet of cells was stirred up and the cell suspension was stored
on ice.

Beforehand, the additive had been dissolved in 50 mM phosphate buffer to

a concentration of 0.8%, and glucose was added to a final concentration of 2%.
This solution was mixed with appropriate volumes of 50 mM phosphate buffer
containing 2% glucose to give incubation solutions of 0.1%, 0.4%, and 0.8%
additive. Mannan did not dissolve in buffer but was dissolved in sterile H20,

mixed with an equal volume of 100 mM phosphate buffer, and glucose was added
to 2%. Drops of the mixtures were placed in a spot dish. Bacterial cells

were attached to the grids by touching the coated side of a grid to the cell

suspension. The grids were floated with the coated side down on the drops of
additive mixture and incubated between 17 and 30 minutes at 281C. The grids

used for the specimens were 300 mesh copper grids that had been coated with
Formvar and carbon. 78 After incubation, the grids were rinsed in sterile
distilled water, and negatively stained with uranyl acetate (1% in H20)
containing Bacitracin (Aldrich Chem. Co., 0.1 mg/ml) as a spreading agent.

The grids were gently wiped on the backside and along the edge with a piece of
filter paper, and air dried.

The grids were observed in a JEOL JEM-100CX II electron microscope

 42
operating in the transmission mode, with an accelerating voltage of 80 kV.
All grids were placed in the same specimen holder. Photomicrographs were

taken at magnifications of 36,000, 48,000, and 72,000. The Kodak 4489 film

plates used for micrographs were developed in Kodak developer D-19 (diluted
1:3) for 3 minutes.

MICROGRAPH ENHANCEMENT IN THE IMAGE ANALYZER

Twenty-six representative micrographs, containing many ribbons, were

selected for further analyses. The selection consisted of five micrographs of

the control samples, and, for each additive, two micrographs from the use of
0.1% additive, one micrograph from the use of 0.4% additive, and two
micrographs from the use of 0.8% additive. One extra micrograph was selected

from the use of 0.4% xyloglucan.

The micrographs were digitized, enhanced and enlarged using a Tracor

Northern TN-8502 image analyzer equipped with a 1024x1024 high-resolution

camera (DAGE-MTI model 81). The enhancement consisted of contrast sharpening

with a seven-by-seven digital filter, and grey-scale contrast stretching which

eliminated the extreme grey shades and optimized the grey scale.

After the enhancement, the numbers of fibrils in the ribbons were

estimated. The widths of the ribbons and fibrils were measured on the monitor
using the distance measurement function of the image analyzer, after suitable
calibration. Each width was measured three times and averaged before further
processing. The standard deviations of these measurements averaged 7%, and
ranged from 0% to 25%.

The distance calibration was performed by digitizing micrographs of a

 43
line grating replica and measuring the line spacing on the screen. These
micrographs were taken at the same magnifications and with the specimen holder
in the same position in the microscope as for the cellulose specimens. The

line spacing on the grating replica was 462.9 nm (Polyscience, Inc.

Warrington, PA).
 RESULTS

CHARACTERIZATION OF MANNAN

The total carbohydrate content of the water-soluble mannan preparations

ranged from 83 to 99%. The mannan that was extracted with ethanol to remove
inorganics had the higher carbohydrate content. Mannose was the major
species, constituting 77-91% of the carbohydrates. Galactose, glucose,

xylose, and arabinose were all present to varying degrees. Table 5 shows the
composition of the crude mannan fraction before the ultrafiltration step in
the preparation of the water-soluble mannan used in the resting cell medium.
The purified material was consumed in a faulty analysis, and the crude
fraction was used in its stead. Except for the inorganic content, the samples
were expected to have the same composition, therefore, the material not
accounted for in Table 5 is primarily inorganics.

Table 5. Carbohydrate composition of crude mannan (CM).

Fraction Fraction of
of CM carbohydrates

Glucose 0.4 0.5

Galactose 3.3 4.5
Mannose 66.5 90.5
Xylose 1.7 2.3
Arabinose 1.6 2.2
Total carbohydrate, % 73.5 100.0

X-ray diffraction showed the mannan to be crystalline, with reflection

peaks indicating a mixture of mannan I and II.79 In a liquid state 13C NMR
44
 45
spectrum of the mannan (see Appendix A), the resonance peaks of the
nonreducing C4 end group and of the C4 internal moiety were identified.80 81
Using the integrated intensities of these resonances, the average degree of
polymerization (DP) of the water-soluble fraction of mannan was calculated to
be 5.6. The response from the carbons in different positions in the molecule
can be slightly different for such large molecules as penta- or hexa-
saccharides which could result in an underestimation of the DP.

CHARACTERIZATION OF XYLAN

X-ray diffraction showed the water-soluble xylan to be amorphous. The

carbohydrate composition of the xylan is shown in Table 6. The material not

accounted for in Table 6 is likely to be uronic acids and ash.

Table 6. Carbohydrate composition of xylan.

Fraction of Fraction of
sample carbohydrates
Glucose 0.7 0.9
Galactose 0.6 0.8
Mannose 1.5 2.0
Xylose 69.8 92.6
Arabinose 2.0 2.7
Ribose 0.4 0.5
Fucose 0.4 0.5

Total carbohydrate, % 75.4 100.0

Liquid state 13C NMR in D2 0/NaOD (see Appendix A) indicated that the
xylan contained carboxylic acid carbons (6=176.8 ppm). The integrated peak

areas from the C4 end group and internal moiety, respectively, gave an average

DP of 8.2. The NMR-spectrum of xylan in DMSO-d6 showed peaks at approximately

 46
the same magnetic shifts as in the spectrum of xylan in D20/NaOD. The
spectrum from DMSO-d6 had a lower resolution of the peaks. The DP was
estimated to 8.3 from the integrated peak areas of the C1 end groups in this

spectrum. As with the 13C NMR analysis of mannan, these values for xylan DP
based on integration of 13C NMR spectra may be underestimations. The DPN of a
birch xylan comparable to the original birch xylan was 215 as measured with
osmometry.71 However, a liquid state 13C NMR spectrum of the original xylan
recorded in D20/NaOD gave a DP of the same magnitude as the water-soluble
xylan, suggesting nonlinear responses from the carbons in different positions.

APPEARANCE OF THE CELLULOSE PRODUCTS IN THE PRODUCTION VESSELS

When the control samples of cellulose were produced with the resting

cell method, the cellulose was formed as a coherent aggregate in the medium.
Cellulose produced in presence of an additive was dispersed throughout the

medium as small flocculent particles. CMC at a concentration of 0.8% caused

the medium to become almost clear even though cellulosic material was isolated
later.

ALKALINE EXTRACTION OF CELLULOSE PRODUCTS

The series of extractions removed a large fraction of the cellulose

samples. Table 7 shows the weights of the celluloses after the extractions.
On average, only 17% of the starting material remained after the extractions.

The smaller the sample was, the more difficult it was to isolate the material
completely. Any difference in the final weights may therefore have been
exaggerated.
 47

Table 7. Weights of celluloses after alkaline extractions.

Start mtrl Wtr ext. 0.01 N 0.1 N 1.0 N NaOH

(mg) (mg) (mg) (mg) (mg)
Control 8.3 5.4 65% 1.6 30% 1.1 79% 0.6 67%
CMC 5.7 5.4 95% 2.5 45% 1.9 79% 0.9 50%
Xyloglu 11.3 9.3 82% 4.2 45% 3.2 80% 2.1 64%
Xylan 10.5 9.3 89% 3.5 37% 2.3 68% 1.5 61%
Mannan 5.0 3.9 78% 1.3 32% 1.0 80% 0.7 88%
The percentages represent amounts of samples remaining after
extractions compared to the weights before that extraction.

X-RAY DIFFRACTION AND PROFILE FITTING

The "old" terminology for the cellulose unit cell has been used
throughout this thesis. With this notation, the planes of the anhydroglucose
units lie in the ab plane and the axes of the cellobiose units are parallel to
the b axis of the unit cell. In an x-ray diffractogram of cellulose I, the
main diffractions are then called (101), (101), and (002), using Miller
indices based on this notation.

The x-ray diffractograms of bacterial cellulose samples were resolved

into individual peaks using a profile-fitting computer program. The profile-
fitting result of each diffractogram consisted of a fitted total profile,
profiles for each of the resolved peaks, a fitted background, and a residual
portion indicating the accuracy of the fit. A good fit was obtained when four
peaks were used for the fitting. The fitting gave more stable results for the
(101) and (002) peaks. The accuracy of the profile fitting will be addressed
further in the Discussion section. For each fitted peak, the resolution also
 48
gave the full width at half maximum (FWHM), the position, and the intensity.

The information for the resolved peaks will be discussed in more detail in the
following subsections.

Figures 2 shows the fitted diffractogram profiles of control cellulose

and celluloses produced in presence of CMC, mannan, xylan, and xyloglucan,
after the last extraction with 1.0 N NaOH. The result of fitting an x-ray
diffractogram is shown in Figure 3, which contains the experimental
diffractogram, the fitted total profile, background, and residual, together
with the fitted profiles of the four individual peaks. The total profile
equals the sum of the individual profiles and the background. (The original

x-ray diffractograms after all extraction steps can be found in Appendix B.)
Cellulose I was the dominant polymorph in all samples. In Figure 2, the
diffractogram of the control sample has an appearance typical of a bacterial
cellulose, with a separation of the (101) and (101) peaks. The intensity of
the (10T) peak in bacterial cellulose is about half of the intensity of the
(101) peak. The diffractograms of cellulose produced in the presence of an
additive were different because the intensities of the (101) and (10T) peaks

were closer to equal. This difference in appearance was most noticeable in

the diffractogram of cellulose produced in the presence of xylan or
xyloglucan.

Peak Full Widths at Half Maximum (FWHM)

The peak width at half maximum is a measurement of the degree of

crystallinity in a material. The diffractogram peaks are wider when the
lateral dimensions of the cellulose crystallites are smaller. The widths of

the fitted crystalline peaks are listed in Table 8 for control cellulose and
 49

MAN

XYL

XGL

20, degrees
Figure 2. Fitted x-ray diffractogram profiles of celluloses. CON:
Control cellulose without additive; CMC: Cellulose produced with CMC; MAN:
Cellulose produced with mannan; XYL: Cellulose produced with xylan; XGL:
Cellulose produced with xyloglucan.
Figure 3. The experimental diffractogram of xyloglucan-cellulose
extracted with 1.0 N NaOH, the fitted total profile, background, and
residual, and the fitted profiles of the four individual peaks.
 51

celluloses produced in the presence of xylan, xyloglucan, mannan, and CMC.

(Appendix B contains the complete set of peak widths for all samples after
each extraction step.)

Table 8. Peak widths (20-degrees): bacterial celluloses produced in the

absence or presence of additives, after extraction with 1.0 N NaOH.

CON CMC MAN XYL XGL

101 1.29 1.83 1.64 1.70 2.07

o1T 0.82 2.03 1.04 1.89 1.73
002 1.32 1.77 1.64 1.67 2.07
Increase (%) - 38 26 30 60

Average of the increase in percent for the (101) and (002) peaks
over a control sample.
CON: Control cellulose without additive; CMC: Cellulose produced with
CMC; MAN: Cellulose produced with mannan; XYL: Cellulose produced with
xylan; XGL: Cellulose produced with xyloglucan.

The presence of additives in the growth medium increased the peak widths
of the bacterial cellulose, and, hence, reduced the crystallite size. Average

values for the (101) and (002) peaks show that xyloglucan increased the widths
most effectively when compared to the control. The differences attributable

to influences of CMC, xylan, and mannan are of approximately the same

magnitude and are smaller than the effect of xyloglucan.

Peak Positions

The positions of the fitted peaks in the diffractograms of control

cellulose and celluloses produced in the presence of additives are summarized

in Table 9, and the complete peak-position information can be found in
Appendix B.
 52

Table 9. Peak positions (20-degrees): bacterial celluloses produced in

the absence or presence of additives, after extraction with 1.0 N NaOH.

CON CMC MAN XYL XGL

101 14.52 14.62 14.62 14.67 14.65
101 16.81 16.91 16.72 16.77 16.69
002 22.67 22.66 22.67 22.61 22.59
CON: Control cellulose without additive; CMC: Cellulose produced with
CMC; MAN: Cellulose produced with mannan; XYL: Cellulose produced with
xylan; XGL: Cellulose produced with xyloglucan.

The presence of any of the additives shifted the position of the (101)
peaks to higher values relative to the control sample. The position of the
(002) peaks shifted to lower values when xylan and xyloglucan were present in
the growth medium, and remained the same in the presence of CMC and mannan.

The position of the (101) peaks shifted to lower values in the presence of
xyloglucan, mannan, and xylan. CMC caused the position of this peak to shift
to a higher value.

The additives influenced the cellulose aggregation pattern resulting in

narrower ranges of the diffractogram peaks. The positions of the peaks in an
x-ray diffractogram reflect the dimensions of the crystalline unit cell. When
the peaks were closer together in the x-ray diffractograms, the angle between
the a and c axes in the unit cell became closer to 90 degrees. 82 Therefore,

the ac plane perpendicular to the cell axis (b) changed from a diamond-shape

to more of a rectangle.

Peak Intensities

Table 10 lists the fitted peak intensities from the diffractograms of

 53
control cellulose and celluloses produced in the presence of additives (see
Appendix B for the remaining data on peak intensities).

Table 10. Peak intensities and relative intensities: bacterial celluloses

produced in the absence or presence of additives, after extraction with
1.0 N NaOH.

CON CMC MAN XYL XGL

101 (cts) 34 29 29 38 77
101/101 0.44 0.72 0.69 0.87 0.60
002/101 2.50 3.93 3.21 3.05 3.51

CON: Control cellulose without additive; CMC: Cellulose produced with

CMC; MAN: Cellulose produced with mannan; XYL: Cellulose produced with
xylan; XGL: Cellulose produced with xyloglucan.

The intensities of the (101) and (002) peaks were normalized against the
(101) peak intensity. In bacterial cellulose, the intensity of the (101) peak

is approximately half of the (101) peak intensity. When additives were

present in the growth medium, the intensities of the (101) and (101) peaks

changed and became more equal. Bacterial cellulose was highly oriented, as

indicated by the nonequal intensities of the (101) and (101) peaks, whereas

the crystallites of celluloses produced in additives were more randomly

oriented.

RAMAN SPECTROSCOPY

The Raman spectra of the control cellulose and celluloses produced in

presence of CMC, mannan, xylan, and xyloglucan are shown in Figures 4-8.
These spectra were recorded after the extraction with 0.1 N NaOH. In the low-

frequency region of the spectra (250-600 cm-1), the bands are primarily
Figure 4. Raman spectrum of control cellulose after the extraction
with 0.1 N NaOH.
56
57
58
 59
sensitive to vibrations in the molecular skeleton. In all Raman spectra, the
positions of the bands shifted very little, never more than 2 cm-1. All the
spectra in Figures 4-8 compare favorably with a cellulose I spectrum.
However, the ratio of intensities and the resolution of bands varied among the
spectra. Appendix C contains the Raman spectra of the cellulose samples after

each extraction step.

Visual comparison of band resolution

A comparison was made of the resolution of adjacent bands in the Raman

spectra of samples after the extractions with 0.1 N and 1.0 N NaOH. The
spectra were visually compared with respect to the resolution of pairs or
triplets of bands after expansion of the 250-650 cm-1 and 250-1700 cm-1
regions. The bands used for these comparisons were at 330-346-380 cm-1, 436-
460 cm"1, 1094-1120 cm-1, and 1338-1376 cm-1 . The spectra of xyloglucan-

cellulose had the bands least resolved, followed by the spectra of cellulose
produced in the presence of xylan, CMC, and mannan. The control cellulose

gave the spectrum with the best resolved bands. The sharper and more resolved

bands are indications of larger sizes of the cellulose crystallites. 8 3 The

molecules are in a more homogeneous environment when the crystallites are

larger, therefore, the frequencies corresponding to the Raman bands have a

more narrow distribution resulting in narrower bands.

Conformational component resolution of Raman spectra

The low frequency regions (250-550 cm-1) of the Raman spectra of the

cellulose samples were resolved into spectra of kI, k1

1 and ko , the

conformational forms responsible for the polymorphs cellulose I, cellulose II,

 60
and amorphous cellulose. Table 11 describes the fractions of the
conformational components in the bacterial celluloses produced in the absence
and presence of additives. (The component fractions of the bacterial
celluloses after all the alkaline extractions can be found in Appendix D.)
The conformational component resolution resulted in k1-contents of the samples
which were at three levels: the control sample had the highest amount of k,;
celluloses produced in CMC or mannan had a medium content of k,; and
celluloses produced in xylan or xyloglucan had thelowest content of kp.

The conformational components given in Table 11 were based on spectra of

the celluloses after the extractions with 0.1 N NaOH. The resolution of the
samples after the extraction with 1.0 N NaOH gave lower fractions of the k X
and higher fractions of the ko conformations than after the preceding
extraction with 0.1 N NaOH, which would imply that the degree of crystallinity

was lower. However, the x-ray diffraction showed the amorphous fraction of
these samples to be lower after the last extraction. One explanation for this
discrepancy is that the degradation products possibly combined with the

cellulose (see also in the Discussion section about Raman bands at 1152 and

1506 cm-1). This could have led to a shift in the positions of the resonance
bands in the low-frequency regions from the positions of the bands in kI, in
addition to the unknown bands at 1152 and 1506 cm-1. Even small changes in

the band positions can shift the fractions of the different conformations in a
sample.
 61

Table 11. Conformational components of bacterial celluloses produced in

the absence or presence of additives, after the extraction with 0.1 N
NaOH.

CON CMC MAN XYL XGL

kI 0.770.03 0.690.04 0.670.02 0.510.04 0.51+0.03
kil 0.040.11 -.120.12 -.060.08 0.160.13 0.030.09
ko 0.170.10 0.410.12 0.370.08 0.300.12 0.450.09
CON: Control cellulose without additive; CMC: Cellulose produced with
CMC; MAN: Cellulose produced with mannan; XYL: Cellulose produced with
xylan; XGL: Cellulose produced with xyloglucan.

TRANSMISSION ELECTRON MICROSCOPY

Appearance of bacterial cellulose fibrils

Two photomicrographs of the control cellulose and celluloses produced in

the presence of each additive are shown in Figures 9-13. The remaining
micrographs which were used for the measurements of fibril widths can be found
in Appendix E. Photomicrographs showed the normal cellulose ribbons to be
flat and twisting (Figure 9). Some ribbons were very tightly twisted, whereas

other ribbons had fibrils which were more splayed or separated from each
other. The ribbons had a range of widths, varying from 40 to 150 nm.
Enlarging the micrographs using an image analyzer made it possible to estimate

the number of fibrils in a ribbon. This was easier at higher concentrations

of additive when the fibrils were more widely splayed; it was more difficult

to distinguish the subunits in micrographs of the control samples. The

ribbons consisted of about 15 fibrils and ranged between 5 and 25 fibrils.

The distance between twists varied significantly; the distances between two
twists measured typically within the range of 0.3 im to 0.8 sm.
 62

Figure 9. Micrographs of control cellulose, no. 2480 and 2536; scale bar:
0.3 um, X48,000.

Figure 10. Micrographs of cellulose produced in 0.1% CMC; no. 2448 and
2449; scale bar: 0.3 um, X48,000.
 63

produced in mannan; no. 2481: 0.1%

Figure 11. of cellulose
Micrographs
mannan; scale bar: 0.2 um, X72,000.
mannan, no. 2490: 0.8%

Figure 12. Micrographs of cellulose produced in xylan; no. 2454: 0.8%

xylan, no. 2470: 0.1% xylan; scale bar: 0.3um, X48,000.
 64

and 2532; scale bar: 0.3 um, X48,000.

Ribbons produced in the presence of CMC at all concentrations (0.1%,

0.4%, 0.8%) looked very similar to normal ribbons but were slightly more
splayed (Figure 10).

Mannan caused the ribbons to separate but they still had more of a flat
twisting appearance (Figure 11).

Xylan at 0.1% and 0.4% addition levels caused a separation of the

fibrils. At 0.8% addition level, the fibrils in most cases formed a coherent
bundle rather than a flat ribbon (Figure 12, no. 2454). The bundles were wavy
and not twisting like the ribbons. In some instances, fuzzy fine fibrils were
produced at a low addition level of xylan (Figure 12, no. 2470).
 65
Figure 13 shows photomicrographs of cellulose produced in the presence
of xyloglucan. The presence of xyloglucan splayed the ribbons. Nontwisting
coherent bundles were also formed at all addition levels.

The micrographs of specimens produced in the presence of xylan and

mannan also contained what looked like short fibrils. The widths of these
shorter fibrils were comparable with the fibrils in the ribbons, 4 and 5 nm,
respectively. The lengths of the short fibrils produced in the presence of
xylan were roughly 36 nm. On the micrographs of specimens produced in the
presence of mannan, the fibrils were considerably longer and averaged about

217 nm. The origin of these shorter fibrils was uncertain. One possibility
is that the presence of hemicellulose caused pieces of cellulose fibril to

break off. Another explanation of the short fibrils is that they were
aggregates of hemicellulose. In one study of redeposition of xylan onto
cellulose, the xylan formed aggregates at the surface of the microfibrils.30
It is feasible that uneven spots on the Formvar film covering the grids could
act as nucleation sites for aggregation of the hemicellulose.

Fibril width measurements with an image analyzer

After the photomicrographs had been digitized, the widths of the fibrils
were measured on the computer screen using a distance function in the image
analyzer. All widths of the long fibrils were averaged for each additive and

each addition level, and the widths were plotted against the addition levels

(Figure 14). The fibril widths are summarized in Table 12. For each
additive, the widths were also averaged for all fibrils over the concentration

range. In the sample of control cellulose, averaging the fibril widths gave

as a mean 5.9 nm 1.5 SD, which agrees with a bacterial fibril width of 5.3-
 66

5.5 nm reported in the literature. 84

Table 12. Fibril widths (nm) of bacterial celluloses produced in the

absence or presence of additives, at different concentrations (% w/v) of
the additives in the growth medium.

CON CMC MAN XYL XGL

0.0 % 5.91.5 - -
0.1 % - .6.01.5 3.20.4 4.41.2 4.11.1
0.4 % - 4.70.9 2.80.3 3.80.6 4.71.1
0.8 % - 5.61.6 3.40.5 3.60.6 5.21.2
Average 5.91.5 5.51.5 3.20.5 4.11.0 4.71.2

Number of 54 55 52 56 103
fibrils for averaging
Magnifi- 48,000 48,000 72,000 48,000 48,000
cation of micrographs (36,000)

CON: Control cellulose without additive; CMC: Cellulose produced with

CMC; MAN: Cellulose produced with mannan; XYL: Cellulose produced with
xylan; XGL: Cellulose produced with xyloglucan.

The fibril widths decreased in the presence of the additives over the
concentration range of 0.1-0.8% additive in the medium used in this study.

The presence of xylan, xyloglucan, and CMC decreased the average fibril widths
to 4.1 1.0 nm, 4.7 1.2 nm, and 5.5 1.5 nm, respectively. The presence

of mannan decreased the fibril width to an average of 3.2 0.5 nm.

The photomicrographs were taken at the highest magnification where it

was possible to obtain a focused image. This was determined in part by the

thickness of the Formvar film coating the grids. The majority of the

photomicrographs were taken at a magnification of 48,000; one micrograph of

xylan-cellulose was taken at 36,000, and the photomicrographs of cellulose

produced in the presence of mannan were taken at 72,000. The pixel sizes of

the digitized images were 0.69, 1.02, and 1.31 nm at magnifications of 72,000,
 67
48,000, and 36,000, respectively. The higher magnification of 72,000 is a
likely explanation for the lower standard deviation of the average mannan-
cellulose fibril width. The magnitude of the magnification should not have
any impact on the fibril-width mean values.

The standard deviations obtained for some of the average fibril widths
were of significant size; the largest was 28%. The standard deviations

reflected variations from different sources such as variations in the width of

one fibril, variations among fibrils in one ribbon or among ribbons at one
addition level of a certain additive. No test of confidence limits were

performed for the influences of the presence of an additive on the average

fibril widths. As an estimation, a 95% confidence interval corresponds to

1.96*u (standard deviation) which would result in no significant difference
between any of the cases where cellulose was produced in the absence or
presence of an additive. In spite of the lack of significant difference on

the 95%-confidence level, the presence of the additives xylan, xyloglucan, and
mannan resulted in trends of decreasing fibril widths. On the other hand, the

trend of influence from the presence of CMC was weaker.

The method of measuring cellulose ribbon subunits with a cursor on the

computer monitor was rather crude and included sources of mechanical error and
subjectivity. This can be seen in the rather large ranges of values obtained

for the widths of the cellulose fibrils. However, considering that fibril

measurements with errors ranging from 10 to 25% can be found in the

literature 2'14, and that the data in this study are in the order of nanometers,
the method can be considered satisfactory.
 68

Figure 14. Fibril widths of bacterial celluloses produced in the presence

of xylan (magnification of micrographs: 48,000 and 36,000), xyloglucan
(X48,000), mannan (X72,000), and CMC (X48,000). The bars indicate plus-
minus one standard deviation.
 DISCUSSION

The influences on the aggregation pattern and morphology of bacterial

cellulose were reported in the preceding section. Two implications of these
results will be discussed in this section, namely the similarities between
modified bacterial celluloses and celluloses from higher plants, and

influences of the cellulose isolation on the aggregation pattern.

Furthermore, the effect of the additives used in this study will be compared
with hemicellulose-cellulose sorption in the literature. Finally, some
aspects of profile fitting of x-ray diffractograms will be discussed, as will
an unresolved problem with the source of two unknown Raman bands.

TRENDS OF SIMILARITIES BETWEEN MODIFIED BACTERIAL CELLULOSES

AND HIGHER PLANT CELLULOSES

Comparison of the x-ray diffraction results

The celluloses from cotton and ramie were used as examples of celluloses

from higher plants. The x-ray diffractograms of these celluloses are shown in
Figures 15 and 16. These celluloses have a high crystallinity, but the
appearances of the diffraction profiles are different from a typical bacterial

cellulose which can be seen in Figure 2. The (101) and (l01) peaks are closer
together in cotton and ramie celluloses than in bacterial cellulose, and the

intensity ratios are different. Table 13 lists data from x-ray diffractograms

of cotton and ramie celluloses.

Peak widths, positions, and intensity ratios of the bacterial cellulose

samples were compared with the data from the higher plant celluloses, based on
69
 70

Table 13. X-ray data for cotton (CF1) and ramie celluloses.

X-ray peaks: 101 101 002

Peak widths
Cotton 1.43 1.32 1.18
Ramie 1.42 1.48 1.40
Peak positions
Cotton 14.77 16.39 22.58
Ramie 14.83 16.44 22.66
Avicel 82 15.00 16.56 22.88
CF182 14.94 16.50 22.81
Ramie 82 15.06 16.50 22.75
Peak intensities
Cotton 729 0.90*Il01 6.30*101
Ramie 2044 0.92*I10
1
6 .16*I0
1

The data from reference 82 are calculated with intensity

functions for each peak. The other data are estimated
from Figures 15 and 16.

information in Tables 8, 9, and 10 (pp. 51-53) and Table 13. Comparisons of

the peak widths showed that, for all additives, their presence during

incubation reduced the degree of crystallinity and the crystallite size below
the magnitudes of both the control sample and the reference plant celluloses.
The positions of (101) and (101) peaks were shifted due to the influences of
the additives, xylan and xyloglucan in particular, to approach the ranges of
the cotton and ramie peak positions. The presence of xylan and xyloglucan
shifted the (002) peak positions to lower values. This change is interpreted
less readily because the information available on the (002) peak positions of
cotton and ramie (Table 13) is inconsistent. Comparisons of the peak

intensities revealed that the presence of the additives increased the (101)
and (101) intensity ratios toward the magnitudes found in cotton and ramie
 72
celluloses. The ratio of 0.60 for xyloglucan (Table 10) is lower than
expected and may be caused by a less accurate fit of this x-ray diffractogram
because the intensity ratios of celluloses produced with xylan and xyloglucan
after the 0.01 N NaOH and 0.1 N NaOH extractions are similar, which can be
seen in Figure 28 in Appendix B.

In the case of all additives, their presence during incubation of

bacterial cellulose has an effect on the peak widths, positions, and intensity
ratios. The effects of xylan and xyloglucan are generally stronger than the

effects of mannan and CMC, resulting in x-ray diffraction data that approach
the data from x-ray diffractograms of ramie and cotton celluloses. This
similarity can also be seen by comparing, in terms of the appearances, the
x-ray diffractograms of the cotton and ramie (Figures 15 and 16) with the
diffractograms of the celluloses modified by the presence of xylan or
xyloglucan (Figure 2). The celluloses produced in the presence of xylan or
xyloglucan no longer have the x-ray diffraction pattern of the typical
bacterial cellulose but more closely resemble the pattern of higher plant
celluloses.

Comparison of the Raman spectroscopy results

The Raman spectra of bacterial celluloses produced in the presence of an

additive (Figures 4-8) were compared with Raman spectra of ramie fibers
reported in the literature 85. The low frequency regions in the spectra of

both ramie and xylan-cellulose have bands which are wider and less resolved
than in a spectrum of control bacterial cellulose. The wider bands in the low

frequency regions of the spectra of ramie cellulose and xylan-cellulose

indicate that the cellulose crystallites are smaller in these celluloses than
 73

in the control bacterial cellulose. 83 Therefore, the results of Raman

spectroscopy support the results of x-ray diffraction that indicate celluloses
produced in the presence of additives have decreased crystallite dimensions.

Comparison of the electron microscopy results

The fibril widths of bacterial cellulose decreased when additives were

present in the growth medium (Table 12). The average width of fibrils in

control cellulose was 5.9 nm. The presences of CMC, xyloglucan, xylan, and
mannan decreased the average fibril width to 5.5, 4.7, 4.1, and 3.2 nm,
respectively. This trend agrees with the lower values reported in the
literature of fibril widths in higher plants, compared with widths of
bacterial cellulose fibrils. Boylston 84 reported widths for bacterial
cellulose fibrils of 5.3-5.5 nm, for cotton fibrils of 2.2-4.0 nm, and for
ramie fibrils of 2.5-4.4 nm. The widths of fibrils in a conifer (Pinus

densiflora) and a hardwood (Populus euramericana) are 2-5 nm as reported by

Harada and Goto. 86

In summary, the results from x-ray diffraction, Raman spectroscopy, and

electron microscopy all indicate that polysaccharides can have an impact on
the pattern of aggregation of cellulose when they are present during cellulose

production. The polysaccharides in this study which were most similar to

hemicelluloses in plant cell walls had the largest impact on the structure of

the cellulose. Furthermore, the changes were such that the bacterial
celluloses were modified to resemble the aggregation patterns found in
celluloses from higher plants. It follows, therefore, that hemicelluloses

present in higher-plant cell walls may form part of the mechanisms controlling

the patterns of aggregation among higher plant celluloses.

 74
RATIO OF CELLULOSE I- TO I, TERTIARY STRUCTURES

It is tempting to draw the conclusion that the presence of the

hemicellulose-like additives during the aggregation of bacterial cellulose
changed the ratio of cellulose Ia to Ib because the x-ray diffractogram

patterns and Raman spectra changed from a bacterial-type cellulose to a wood-

type cellulose. (See Background-section on cellulose Ia and Ib, and
hierarchical organization of cellulose structures). Cellulose Ia and Ib are
tertiary structures that have been defined from analysis with solid state 13C

NMR. The forms Ia and I can also be distinguished in the O-H region of Raman
spectra recorded primarily in the microprobe mode.

The magnetic resonance spectra of I -rich celluloses correlate well with

a wood-type x-ray diffraction pattern having decreased lateral dimensions and

lower degree of orientation. This correlation is true for all specimens
investigated except for one particular alga, Laminaria japonica. This alga
has an x-ray diffractogram pattern and a Raman spectrum of a wood-type
cellulose, but the 31C 10.
NMR spectrum shows a distinct cellulose I.. 87 Because
of this exception to the otherwise consistent division between Ia-rich

bacterial-type cellulose on one hand, and I,-rich wood-type cellulose on the

other hand, it is not possible to conclude from x-ray diffraction and Raman
spectroscopy anything about the ratio of Ia to I in the bacterial celluloses

produced in the presence of the additives. However, the changes in the x-ray
diffraction patterns and Raman spectra that did occur upon introduction of the
additives suggest that a change in the Ia-to-Ib ratio may have taken place.

The Raman spectra of the bacterial cellulose samples were compared in

order to base any conclusion of the Ia-to-Ib ratio on the hydrogen bonding
 75

patterns. The high frequency regions (3000-3700 cm-1) of the Raman spectra of
the control bacterial cellulose and of celluloses prepared in the presence of
additives were expanded. It was not possible to distinguish any features in
these spectra attributable to either of the cellulose Ia or I crystalline

forms. Reasons could be that the Raman spectra were not recorded in the
microprobe mode, or, for such detailed analysis, the amounts of the samples

were too small or the spectra quality was inferior.

CELLULOSE AGGREGATION ALTERATIONS RELATED TO THE ADDITIVE STRUCTURES

The results from the x-ray diffraction and Raman spectroscopy could be
divided into three groups: 1. control cellulose, 2. CMC- and mannan-influenced
celluloses, and 3. xylan- and xyloglucan-influenced celluloses. This suggests

that the additives interacted with the bacterial cellulose in two different
ways. The xyloglucan did not have functional groups, and the xylan had a low

degree of substitution of uronic acids as functional groups (DS=0.12).7 1

Hence, the conformations of xyloglucan and xylan were similar to the cellulose

conformation, and it was possible for xyloglucan and xylan to co-crystallize

with the cellulose. One mechanism could be by molecular replacement in the

crystalline lattice when xylan or xyloglucan were present during the

aggregation of nascent cellulose. This would create a defect structure within

the cellulose lattice.

The additives CMC and mannan were less similar to cellulose than xylan
and xyloglucan. CMC had extended functional groups in the form of charged

carboxymethyl side-chains which would prevent any closer interaction with the
cellulose aggregation. In terms of conformational dissimilarities of the

sugar moieties, the mannose unit in mannan was the least similar of the sugars
 76
in the additive backbones to the glucose unit in cellulose. This is due to
the axial hydroxyl group at C2 in mannose instead of an equatorial position as
in glucose and xylose. Another possible reason for the weak interaction
between mannan and cellulose was the relatively low molecular-weight of the
mannan fraction used in this study. The molecular weight of the mannan was
the lowest of the additives studied.

As a result of their conformation, CMC and mannan lowered the degree of

crystallinity of the bacterial cellulose and decreased the crystallite sizes.
Xylan and xyloglucan decreased the degree of crystallinity and, more
important, probably blended intimately with the cellulose during its

aggregation and changed the cellulose lattice structure.

The presence of mannan during cellulose aggregation decreased the

crystallite sizes and the fibril widths by approximately the same amounts as
the presence of xylan or xyloglucan. However, the aggregation pattern was not
changed as much by the presence of mannan as by the presence of xylan or

xyloglucan. The mannan molecules may have been capable of interacting with
the cellulose molecules during the aggregation and thereby decreased the

crystallite and fibril sizes. The dissimilarities of the cellulose and mannan
backbones may have prevented further association with additional cellulose
molecules and limited the changes of the cellulose crystalline lattice.

The effects of the different additives were similar but not identical.
The additives resemble different hemicelluloses that are associated with

different layers of the cell walls. It is likely that the hemicelluloses in

cell walls of higher plants interact, as selective moderators in a manner

similar to what has been shown in the results of this study, with the
 77
aggregation of cellulose as it is deposited in the cell walls.

POLYMER CHEMISTRY OF ADDITIVE SOLUTIONS

The sorption and reaction behavior of the additives could be a function

of the polymer chemistry of the solutions. One interpretation of the results

could be that the additives were merely precipitating on the cellulose and
were not incorporated in the cellulose lattice. The solubility of
B-(1-4)-linked polysaccharides depends on factors such as the molecular
weight, carbohydrate composition, and substitution pattern. Lower molecular-
weight polysaccharides are more readily dissolved than higher molecular-
weight fractions. Polysaccharide backbones consisting of hexoses, rather than
pentoses, are stiffer and, hence, less soluble. The polysaccharide solubility
is increased considerably if the backbone is substituted compared with an
unsubstituted polysaccharide such as cellulose.

The relative importance of these factors on the actual solubility of a

polysaccharide is difficult to quantify, which makes the relative solubilities

of the additives used in this thesis hard to predict because the additives
differ in several aspects. The mannan has a low molecular weight (higher
solubility) but is unsubstituted (lower solubility); the xylan backbone is
less rigid (higher solubility) and this additive has uronic acid substituents
(higher solubility); the xyloglucan has a very high molecular weight (lower

solubility) but also a high degree of side chains (higher solubility); the CMC
has a relatively high molecular weight (lower solubility) and charged

functional groups (higher solubility).

All additives were water-soluble and were dissolved in concentrations up

 78
to 0.8% in the incubation solutions for electron microscopy specimens. The
dissolution rates of the additives were quite different. By empirical
observations, xylan appeared to have the highest rate of dissolution, followed
by CMC, mannan, and xyloglucan. The xyloglucan never showed any signs of
precipitation but required storage at +2-4'C for approximately 12 hr to
dissolve. The ease of dissolution did not seem to correlate with the
approximate order of disrupting power on the cellulose lattice. Xylan and
xyloglucan were the most effective, followed by mannan and then CMC. This
observation in addition to the fact that the extracted cellulose samples were
prepared from incubation solutions containing 0.1% additive ruled out the

possibility that the additives were precipitating.

A dissolved molecule of the additives can take on different

conformations, such as a sphere or rod. The difference in conformation of the
additives in solution may have influenced their interaction with cellulose.

The higher molecular-weight additives are more likely to be in a coiled

spherical or rod-shaped form, which is less likely to be able to interact with
a cellulose molecule. However, in this case, xyloglucan had the highest

molecular weight, in addition to a strong effect on the cellulose aggregation

pattern. This result strongly suggests that the conformation of the additives
in solution did not influence their interactions with cellulose.

EFFECT OF ISOLATION PROCEDURES ON THE CELLULOSE AGGREGATION PATTERNS

As the bacterial celluloses produced in the presence of additives were

extracted with sodium hydroxide solutions of increasing strength, the

appearance of the x-ray diffractograms and Raman spectra changed. These

changes were more pronounced in cellulose produced in the presence of xylan
 79
and xyloglucan than in the control sample of cellulose and celluloses produced
in the presence of CMC and mannan. The extraction series of celluloses
produced in the presence of xylan is represented by the x-ray diffractograms,
Figure 17, and Raman spectra, Figure 18. Comparable results from the other
additives can be found in Appendix B and C.

The pattern of aggregation of the molecules in a cellulosic material

provides a basis for the physical properties of a material. Properties such
as extensibility and tensile strength are related to the nature and degree of
crystallinity and the crystallite orientation of a cellulose fiber. 88 89
Therefore, if the aggregation pattern, or tertiary structure, of a cellulose
changes, it is expected that the physical properties of this material also
will change. The differences between the diffractograms (Figure 17) and the

spectra (Figure 18) indicate that the properties of the celluloses would
indeed be different depending on what procedure was used to prepare the
cellulose. If we go back to the quotation by Preston 7 referred to in the
Introduction, which stated that removal of noncellulosics would only sharpen
the x-ray arcs, the point could be made that photographs, instead of
diffractograms, of x-ray diffractions such as those in Figure 17 would

probably show the same effect described by Preston.

The discussion presented here of the influence of isolation procedures

on the cellulose structure is far from complete but does show evidence of an

influence. If the underlying phenomena that determine the tertiary structures

of cellulose can be understood, the properties of cellulosic materials may be

intentionally altered.
80
Figure 18. Raman spectra of control cellulose after the last extraction
and xylan-celluloses after each extraction step.
 82
ADDITIVES USED IN THIS STUDY AND INTERACTIONS IN CELL WALLS

There are obviously many differences between the additives used in this
study and the polysaccharides that can be found in the plant cell walls. One
difference is the origins of the mannan and xyloglucan, which were ivory nuts
and tamarind seeds, respectively, instead of plant cell walls. Mannan in
ivory nuts and xyloglucan in tamarind seeds function as storage reserve
polysaccharides rather than being involved in structural reinforcement of the
plants. There is a striking difference between the difficulty with which
primary wall xyloglucan can be isolated compared with the ease of isolating
xyloglucan from tamarind seeds. The xyloglucan in primary cell walls is
probably strongly associated with the cellulose there, whereas cellulose is
not present in the seeds.

The polysaccharides from seeds and primary walls are structurally

different. Mannan from ivory nuts is a homopolymer. In plant cell walls,
mannan occurs as a copolymer in the form of galactoglucomannan or glucomannan.
The structural differences between xyloglucans from seeds and primary walls
are small relative to the differences of the mannans. Xyloglucan in seeds is

a B-(1-4)-glucan with xylose side chains, either single or with a terminal

galactose. Primary cell-wall xyloglucan has the same type of backbone as seed
xyloglucan. The side chains often have terminal fucose units in addition to
the terminal galactose units. The ratios of sugar units are slightly
different in xyloglucans from seeds and primary walls, but essentially the

same.

Another difference is that the mannan and xylan used in this study was
of lower molecular weight than would be present in plant cell walls; the
 83
difference was especially large in the case of the mannan. The reason is that
the additives had to be water-soluble to be used with the resting cell method
for producing bacterial cellulose. However, the presence of this low
molecular weight xylan had a strong effect on the aggregation pattern of

cellulose and the presence of the mannan a more moderate effect. These
effects are expected to be stronger if a xylan or mannan of higher DP is used.

Clayton and Phelps 29 compared the rates of sorption of a glucomannan

(from white spruce) and a xylan (from white birch) and found that the sorption
rate of the glucomannan was about twice as high as the sorption rate of the
xylan. Xylan was expected to have a higher sorption than the glucomannan
because of the conformational similarities of the positions of the hydroxyl

groups with glucose in cellulose. The xylan had a lower degree of

polymerization than the glucomannan, which also led to the expectation of
higher sorption.

The authors suggested that the reason for the faster sorption of the
glucomannan is that it may not contain uronic acid groups, overriding such

factors as lower DP and conformational similarity. Stephen 61 , on the other

hand, suggested that, whereas the unsubstituted portions of the xylan molecule
can associate with cellulose chains, the absence of a primary alcohol group
greatly diminishes the capacity of xylan chains to join with each other or
with other polysaccharides. Mora et al. 3 found that the xylan-xylan

interactions were stronger than the cellulose-xylan interactions. The reason

for this discrepancy may be that the xylans referred to are of different

sources and, therefore, have slightly different structure.

The results of the present study showed that the ivory nut mannan
 84
fraction used here had a lower degree of influence on the cellulose structure
than the birch xylan. As before, the mannan backbone instead of a glucomannan
backbone or the low molecular weight may explain the different results.

In comparing the interactions of xylan and xyloglucan with cellulose,

Molinarolo2 7 found the maximum specific sorption of xyloglucan to be less than
half of the sorption of a glucuronoarabinoxylan. On the other hand, Hayashi
et al.25 found that xyloglucan sorption was decreased to half with a xylan
also present in the system but in a 10-fold excess. In this thesis, the
influences of xylan and xyloglucan were similar, which could indicate similar
magnitudes of interactions with cellulose. One needs to be careful in making
these comparisons because most studies in the literature are based on sorption
by mature cellulose whereas the influence on the aggregation of nascent
cellulose was studied in this thesis.

PROFILE FITTING OF X-RAY DIFFRACTOGRAMS

The profile fitting program was developed to fit x-ray diffractograms of

inorganic material or simpler diffractograms of polymers, and is less suitable

for complex patterns such as those of cellulose or of samples containing

material with lower crystallinity. The limitation in decreasing the fitting

step size gave rather crude peak profiles. The x-ray diffractograms of

bacterial celluloses before the alkaline extractions had substantial

background of amorphous material, and the profile fitting of these
diffractograms gave less reproducible results than the diffractograms recorded
after later extractions.

The goodness of an individual peak fit was given by an estimated

 85
standard deviation (a). Values of 1.0-1.5 for a were considered acceptable by
the software producer. In this thesis, the peak fits had a-values ranging
from 0.467 to 0.692 which indicated satisfactory fittings. The overall
fitting goodness depended primarily on that the number of fitted peaks
represented the physical situation. No statistical test of the overall
fitting was performed by the profile fitting program, but the goodness could
be estimated from the match of the fitted total profile with the experimental
diffractogram and from the magnitude of the residual data. The residual data
were calculated from the raw data minus the sum of all fitted profiles and the
background.

The fittings of the outer peaks in the diffractograms (101 and 002) were
less variable than the fittings of the middle peaks in the diffractograms (101
and amorphous). A possible reason for this is that the fitting of the middle

peaks has error contributions from two sides. Different sensitivities of the
peaks to chemical treatments can also contribute to this phenomenon. The
small sample sizes in this study resulted in diffractograms with low signal-
to-noise ratios which also may have contributed to a less accurate fitting.

RAMAN BANDS AT 1152 AND 1506 CM-1

The Raman spectra of the bacterial celluloses after the last extraction

with 1.0 N NaOH had increased intensities at 1152 and 1506 cm-1 (1152 and 1510
cm-1 in the spectrum of xylan-cellulose). The 1152-band, but not the 1506-
band, is normally present at a much lower intensity in the Raman spectrum of

cellulose I. The intensities of the 1506-band were of increasing magnitude in

the spectra of xylan-, xyloglucan-, mannan-, CMC-, and control cellulose. The
intensity of the 1152-band correlated with the intensity of the 1506-band.
 86

The unknown bands may have originated in the bacterial cells which were
not removed completely from the cellulose samples. The alkaline extractions
in this study were performed at room temperature. However, to remove proteins
from wood celluloses, boiling in alkali is required. Boiling a bacterial
cellulose pellicle in 1% NaOH removed the cell material and did not result in
increased intensities at 1152 and 1506 cm-1. It is possible that the
extraction with 1.0 N NaOH degraded the bacterial cells, but did not fully
remove the degradation products.

The weights of the cellulose samples after the last extraction were
about one mg and increased in the following order: control-, mannan-, CMC-,
xylan-, and xyloglucan-cellulose - approximately opposite to the order of
increasing intensities of the unknown bands. The smaller samples were more

difficult to collect after washing, and less likely to be thoroughly washed.

This would suggest that the unknown bands were from some compounds that were

not washed off the cellulose products.

The sharpness of the unknown bands suggests they originate from an

isolated functional group in an organic molecule, or from an inorganic source.

If not from the bacterial cells, the unknown bands may be from a contaminant
present in the sodium hydroxide or on the glassware used.
 CONCLUSIONS

Bacterial cellulose was produced by nondividing cells in a system such

that polysaccharides added to the growth medium interacted with the cellulose
during its aggregation. The cellulose produced in the absence of and in the
presence of polysaccharides consisted of cellulose I as the predominant
polymorph, with subtle differences in the aggregation patterns. These
differences were found in the widths, intensity ratios, and positions of the
x-ray diffractogram peaks, and in the widths and resolution of the Raman
bands. These differences indicate that the presence of additives decreased

the degrees of crystallinity and orientation, and changed the unit cell
dimensions of the cellulose. The cellulose produced in the presence of
additives had narrower fibril widths than the controls, revealed via
transmission electron microscopy. The narrower fibrils agreed with the widths

of cellulose fibrils reported for higher plants.

The additives influenced the cellulose aggregation in different ways by

their presence. The presence of CMC slightly decreased the crystallite sizes

and fibril widths. The presence of xylan or xyloglucan caused a larger

decrease of the crystallite sizes and fibril widths, and also caused a change

in the aggregation patterns. The presence of mannan decreased the crystallite

sizes approximately as much as the presence of xylan or xyloglucan. However,

the cellulose aggregation patterns were not as strongly altered by the

presence of mannan as by the presence of xylan or xyloglucan. It was
concluded that xylan and xyloglucan were capable of incorporating into the

cellulose crystalline lattice because their backbones were relatively similar

to the cellulose backbone. Mannan were capable of associating with the

87
 88
aggregating cellulose but the greater dissimilarities between the mannan and
cellulose prevented further association with subsequently aggregating
cellulose, thereby causing decreased crystallite and fibril lateral dimensions
but limiting the changes of the cellulose crystalline lattice.

The presence of xylan or xyloglucan changed the aggregation pattern of

the bacterial cellulose to resemble the type of aggregation pattern found in
celluloses from higher plants. As a result of these observations,

polysaccharides in the plant cell walls are proposed to participate, in ways

similar to the effects found in this study, in determining the aggregation
patterns of celluloses in higher plants. Results from earlier studies 62 63.
indicate that different types of hemicelluloses are present in different
proportions in each cell wall layer. This study indicates that hemicelluloses

in the cell walls of higher plants may participate as selective moderators

which determine different modes of aggregation of cellulose during the
biogenesis of the separate plant cell-wall layers.

The procedure for isolation of a cellulose from hemicellulose, lignin,

and other wood components, is one factor which determines the pattern of
aggregation of the particular cellulose, and, thus, the nature and properties
of a cellulosic material.
 RECOMMENDATIONS

The presence of polysaccharide additives during aggregation of cellulose

may have an impact on the ratio of the cellulose crystalline.forms Ia and Ip,
as determined by solid state 13C NMR spectroscopy. A positive conclusion
about this impact could not be drawn on the basis of the x-ray diffraction and
Raman spectroscopy results in this study. A further investigation of the
ratio of cellulose Ia to I of the bacterial celluloses produced in the
presence of hemicelluloses would be valuable. The procedure for cellulose
production from resting cells would have to be modified or multiplied to
produce enough material for analysis with solid state 13C NMR.

The hydrogen bonding pattern of the samples could reveal the ratio of I.
to Ib if the Raman spectra would be recorded on oriented samples. This could
be accomplished by drying the celluloses under tension to get oriented
samples, and then recording Raman spectra in the microprobe mode with the
electric vector of the light parallel (0 ) to the fibril axis.

The synergistic effects of a mixture of hemicelluloses on the cellulose

aggregation pattern is another interesting area to investigate further. This

would more truly represent the situation in a plant cell wall where the

hemicelluloses are not present one at a time with the cellulose. It is

possible that the interactions with cellulose are stronger when mixtures of

hemicelluloses are used.

89
 ACKNOWLEDGEMENTS

I sincerelythank Dr. Rajai Atalla, Dr. Norm Thompson, Dr. Terry

Conners, and Dr. Morris Johnson, the members of the thesis advisory committee,
for their guidance, contributions, and encouragement throughout the course of
this thesis.

The financial support of this research by The Gunnar and Lillian

Nicholson Graduate Fellowship and Faculty Exchange Fund, and the Institute of
Paper Science and Technology and its member companies is gratefully

acknowledged.

Many of the staff, faculty, and students at the Institute have

contributed to this investigation with their advice and assistance for which I
am very grateful. Special thanks are in order to Becky Whitmore and Clark
Woitkovich for their help with the Raman spectrometer and x-ray
diffractometer, and to Dr. Sue Molinarolo who generously provided the
xyloglucan.

I thank Dr. Candace Haigler and Dr. Moshe Benziman for valuable

discussions regarding preparation of bacterial cellulose. The gift of

bacterial culture from Dr. Haigler is highly appreciated.

While working on this thesis, my life was made easier by the help from
many wonderful people. Fellow students at the Institute have supplied plenty
of encouragement, great company, and challenging discussions. Lots of thanks

to Beth, Bob, and Pat, who helped me through the last year in Appleton. I
also thank the U.S.D.A. Forest Products Laboratory for letting me use the

90
 91
facilities in Madison, and John and John for their help.

I thank my family back home for their encouragement which kept coming so
regularly in a stamped envelope. Finally, I thank my husband Allan for his
support and patience during the completion of this thesis.
 LITERATURE CITED

1. Ben-Hayyim, G. and Ohad, I., J. Cell Biol. 25:191-207 (1965).

2. Haigler, C.H.; Brown, R.M., Jr.; Benziman, M., Science 210:903-6 (1980).
3. Brown, R.M., Jr.; Haigler, C.H.; Cooper, K., Science, 218:1141-42 (1982).
4. Haigler, C.H.; White, A.R.; Brown, R.M., Jr.; Cooper, K.M., J. Cell
Biology 94:64-69 (1982).
5. Kai, A.; Koseki, T., Chemistry Letters, The Chemistry Society of Japan,
pp. 607-10 (1985).
6. Mullis, R.H.; Thompson, N.S.; Parham, R.A., Planta (Berl.) 132:241-48
(1976).
7. Preston, R.D. The Physical Biology of Plant Cell Walls, Chapman and Hall,
London, 1974, p. 170.
8. Atalla, R.H., in "Cellulose: Structural and Functional Aspects", (J.F.
Kennedy, G.O. Phillips and P.A. Williams, eds.) Ellis Horwood, in press.
9. Atalla, R.H., J. Appl. Polym. Sci.: Appl. Polym. Symp. 37:295-301 (1983).

10. Atalla, R.H.; VanderHart, D.L., in "Cellulose and Wood, Chemistry and
Technology", Proc. 10th Cellulose Conf. (C. Schuerch, ed.) Wiley Interscience,
John Wiley & Sons, New York, 1989, pp. 169-88.
11. Asai, T, Acetic Acid Bacteria, Classification and Biochemical Activities.
University Park Press, Baltimore 1968.

12. Pasteur, L., Etudes sur le vinaigre, Paris, 1868.

13. Brown, A.J., J. Chem. Soc. 49:432-39 (1886), 51:643 (1887).
14. Zaar, K., Cytobiologie 16:1-15 (1977).
15. Brown, R.M., Jr.; Willison, J.H.M.; Richardson, C.L., Proc. Natl. Acad.
Sci. USA 73(12):4565-69 (1976).
16. Zaar, K., J. Cell Biology 80:773-77 (March, 1979).

17. Schramm, M.; Hestrin, S., J. Gen. Microbiol. 11:123-9 (1954).

18. Williams, W.S.; Cannon, R.E., Appl. Environm. Microbiol. 55(10):2448-52
(1989).
19. Haigler, C.H., Dept. Biol. Sci., Texas Tech. Univ., Lubbock, TX, Personal
communication, 1988.

92
 93
20. Allen, R.L.M., Colour Chemistry, Thomas Nelson and Sons Ltd., London,
1971. p. 278-79.
21. Venkataraman, K., The Analytical Chemistry of Synthetic Dyes, Wiley-
Interscience Publication, New York, 1977. p. 7.
22. Haigler, C.H.; Chanzy, H. J. Ultrastructure and Molecular Structure
Research 98:299-311 (1988).
23. Kai, A., Makromol. Chem., Rapid Commun. 5(6):307-10 (1984).

24. Haigler, C.H.; Benziman, M., in "Cellulose and Other Natural Polymer
Systems: Structure, Biogenesis and Degradation" (Brown, R.M., Jr., ed.)
Plenum, New York, 1982. pp. 273-297.
25. Hayashi, T.; Marsden, M.P.F.; Delmer, D.P., Plant Physiol. 83:384-89
(1987).
26. Valent, B.S; Albersheim, P., Plant Physiol. 54:105-8 (1974).
27. Molinarolo, S.L. Sorption of Xyloglucan onto Cellulose Fibers. Doctoral
Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1989.
28. Walker, E.F., Tappi 48(5):298-303 (1965).
29. Clayton, D.W.; Phelps, G.R., J. Polym. Sci.: Part C 11:197-220 (1965).
30. Mora, F.; Ruel, K.; Comtat, J.; Joseleau, J.-P., Holzforschung 40:85-91
(1986).
31. Most, D.S., Tappi 40(9):705-712 (1957).
32. Laffend, K.B.; Swenson, H.A., Tappi 51(3):118-122 (1968).

33. Russo, V.A., Thode, E.F., Tappi 43(3):209-18 (1960).

34. Dugal, H.S.; Swanson, J.W., Tappi 55(9):1362-1367 (1972).

35. Blackwell, J; Kurz, D.; Su, M.-Y.; Lee, D.M., in "Structures of Cellulose"
(R.H. Atalla, ed.) ACS Symp. Ser. No. 340, Am. Chem. Soc., Washington, DC,
1987, pp. 199-213.
36. Simon, I.; Scheraga, H.A.; Manley, R.St.J., Macromolecules 21:983-990
(1988).
37. Simon, I.; Glasser, L.; Scheraga, H.A.; Manley, R.St.J., Macromolecules
21:990-998 (1988).
38. Sarko, A., in "Cellulose Structure, Modification and Hydrolysis" (R.A.
Young and R.M. Rowell, eds.) John Wiley & Sons, New York, 1986, pp. 29-49.

39. Hayashi, J.; Kon, H.; Takai, M.; Hatano, M.; Nozawa, T., in "Structures of
Cellulose" (R.H. Atalla, ed.) ACS Symp. Ser. No. 340, Am. Chem. Soc.,
 94
Washington, DC, 1987, pp. 135-50.
40. Atalla, R.H., in "Structures of Cellulose" (R.H. Atalla, ed.) ACS Symp.
Ser. No. 340, Am. Chem. Soc., Washington, DC, 1987, pp. 1-14.
41. French, A.D.; Roughead, W.A.; Miller, D.P., in "Structures of Cellulose"
(R.H. Atalla, ed.) ACS Symp. Ser. No. 340, Am. Chem. Soc., Washington, DC,
1987, pp. 15-37.
42. Miller, D.P., Tappi Int. Dissolving Pulps Conf. 1987, pp. 177-80.
43. Atalla, R.H., Appl. Polym. Symp. 28:659-69 (1976).
44. Isogai, A.; Ishizu, A.; Nakano, J.; Atalla, R.H., in "Structures of
Cellulose" (R.H. Atalla, ed.) ACS Symp. Ser. No. 340, Am. Chem. Soc.,
Washington, DC, 1987, pp. 292-301.
45. Atalla, R.H., in "Industrial Polysaccharides, Biomedical and
Biotechnological Advances", Proceedings of the 3rd International Workshop, (V.
Crescenzi, I.C.M. Dea, S. Paoletti, S.S. Stivala and I.W. Sutherland, eds.)
Gordon and Breach Science Publishers, in press.
46. Marrinan, H.J.; Mann, J., J. Polym. Sci. 21:301-11 (1956).

47. Atalla, R.H.; VanderHart, D.L., Science 223:283-85 (1984).

48. VanderHart, D.L.; Atalla, R.H., Macromolecules 17:1465-72 (1984).
49. Horii, F.; Yamamoto, H.; Kitamaru, R.; Tanahashi, M.; Higushi. T.,
Macromolecules 20:2946-49 (1987).
50. Hirai, A.; Horii, F.; Kitamaru, R., Macromolecules 20:1440-42 (1987).
51. Chanzy, H.; Henrissat, B.; Vincendon, M.; Tanner, S.F.; Belton, P.S.,
Carbohydrate Research, 160: 1-11 (1987).
52. Thompson, N.S., Appleton, WI, Personal communication, 1989.
53. Kooiman, P., Recueil Des Travaux Chimiques Des Pays-Bas 80:840-865 (1961).
54. Hayashi, T.; Maclachlan, G., in "Cellulose Structure, Modification and
Hydrolysis" R.A. Young and R.M. Rowell, eds.) John Wiley & Sons, New York,
1986, pp. 67-76.
55. Hayashi, T.; Maclachlan, G., Plant Physiol. 75:596-604 (1984).

56. Hayashi, T.; Delmer, D.P., Carbohydrate Res. 181:273-77 (1988).

57. Talmadge, K.W.; Keegstra, K.; Bauer, W.D.; Albersheim, P., Plant Physiol.
51:158-73 (1973).
58. Nealey, L.T. The Isolation, Characterization, and Biological Testing of
Xyloglucan from Suspension Cultured Loblolly Pine Cell Spent Medium. Doctoral
 95

Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1987.

59. Hayashi, T.; Wong, Y.-S., Maclachlan, G., Plant Physiol. 75:605-10 (1984).
60. Fry, S.C., Planta 169:443-453 (1986).
61. Stephen, A.M., in "The Polysaccharides" (G.O. Aspinall, ed.) Academic
Press, New York, 1983, Vol. 2, pp. 97-193.
62. Sjostrom, E., Wood Chemistry, Fundamentals and Applications, Academic
Press, New York, 1971.
63. Byers, E.M. An Autoradiographic Study of the Hemicellulose Distribution
in the Walls of Pinus resinosa Tracheids. Doctoral Dissertation. Appleton,
WI, The Institute of Paper Chemistry, 1989.
64. Aspinall, G.O.; Hirst, E.L.; Percival, E.G.V.; and Williamson, I.R., J.
Chem. Soc. pp. 3184-88 (1953).
65. Aspinall, G.O.; Rashbrook, R.B.; Kessler, G., J. Chem. Soc. pp. 215-21
(1958).
66. Timell, T.E., Can. J. Chem. 35:333-38 (1957).
67. Hercules Inc. Wilmington, Delaware, Cellulose gum, 1968.

68. Thiem, J.; Sievers, A.; Karl, H., J. Chromat. 130:305-13 (1977).
69. Henderson, M.E. Mechanisms of Alkaline Glycosidic Bond Cleavage in 1,5-
Anhydro-4-O-B-Mannopyranosyl-D-Mannitol. Doctoral Dissertation. Appleton,
WI, The Institute of Paper Chemistry, 1986. pp. 67-68.
70. Borchardt, L.G.; Piper, C.V. TAPPI 53(2):257-60 (1970).
71. Duckert, L.; Byers, E.; Thompson, N.S., Cellulose Chem. Technol. 22(1):29-
37 (1988).
72. Hestrin, S., in "Methods in Carbohydrate Chemistry" (Whistler, ed.) Vol.
III, Acad. Press, New York, 1963. pp. 4-9.

73. Haigler, C.H. Alteration of Cellulose Assembly in Acetobacter xylinum by

Fluorescent Brightening Agents, Direct Dyes, and Cellulose Derivatives.
Doctoral Dissertation. Chapel Hill, North Carolina, The University of North
Carolina at Chapel Hill, 1982.
74. Benziman, M., Dept. Biol. Chem., Inst. Life Sciences, The Hebrew
University of Jerusalem, Israel, Personal communication, 1986 and 1988.

75. Sandford, P.A.; Baird, J., in "The Polysaccharides" (G.O. Aspinall, ed.)
Academic Press, New York, 1983, Vol. 2, pp. 411-490.

76. Philips Electronic Instruments, Inc. Instructions for Profile Fitting.

Publication no. 4835.015.13900, Philips, Mahwah, NJ, June 1987.
 96
77. Woitkovich, C.P., Institute of Paper Science and Technology, Atlanta, GA,
Personal communication, 1989.
78. Hayat, M.A. Basic Techniques for Transmission Electron Microscopy,
Academic Press, Orlando, 1986. pp. 266-70.
79. Frei, E.; Preston, R.D. Proc. Roy. Soc. B. 169:127-45 (1968).
80. Gast, J.C. A Nuclear Magnetic Resonance Study of the Glycosidic Linkage
of the Xylo- and Cello-Oligosaccharides. Doctoral Dissertation. Appleton,
WI, The Institute of Paper Chemistry, 1983.
81. Gorin, P.A.J., Can. J. Chem. 51:2375-83 (1973).
82. Isogai, A., Dept. Forest Products, Fac. Agriculture, Univ. Tokyo, Japan,
Unpublished work, 1989.
83. Wiley, J.H. Raman Spectra of Celluloses. Doctoral Dissertation.
Appleton, WI, The Institute of Paper Chemistry, 1986.
84. Boylston, E.K.; Hebert, J.J., J. Appl. Polym. Sci. 25(9):2105-7 (1980).
85. Wiley, J.H.; Atalla, R.H., Carbohydrate Research 160:113-29 (1987).
86. Harada, H.; Goto, T., in "Cellulose and Other Natural Polymer Systems:
Structure, Biogenesis and Degradation" (Brown, R.M., Jr., ed.) Plenum, New
York, 1982. pp. 383-401.

87. Hackney, J.M., U.S.D.A. Forest Products Laboratory, Madison, WI, Personal
communication, 1989.
88. Ward, K., Jr., Textile Res. J. 20(6):363-72 (1950).
89. Howsmon, J.A.; Sisson, W.A., in "Cellulose and Cellulose Derivatives" (E.
Ott, H.M. Spurlin, and M.W. Grafflin, eds.) Interscience Publishers, New York,
1954. Part I(IVB), pp. 231-347.
 APPENDIX A
LIQUID STATE 13 C NMR SPECTRA

Table 14. 13C NMR spectrum of mannan.

Position (ppm) Intensity Assignment 80 ,81

60.818 strong C6int

61.385
61.858 weak
62.429
67.174 med-str C4nred
68.743 weak
69.893 weak
70.101 weak
70.356 strong C2int
70.893
71.093 weak
71.608
71.786 strong C3int
73.086 weak
73.348 weak
73.535
73.966 weak
74.204 weak
75.425 strong C5int
76.652 strong C4int
76.921 med-str
77.966
81.819 weak
100.398 strong Clint

int = internal carbohydrate moiety

nred = nonreducing end group

97
 99

Table 15. 13C NMR spectrum of xylan.

Position (ppm) Intensity Assignment 80

59.586 medium a-C5red
63.033 strong C5int+P-C5
71.328 medium C4nred
72.067 med-str
72.885 strong C2int or C3int
74.305 strong C3int or C2int
75.854 strong C4int+9C4
82.368 medium
97.399 weak f-Clred
100.822 weak
101.910 strong Clint+Clnred
176.766 medium -COO-
red = reducing end group
int = internal carbohydrate moiety
nred = nonreducing end group
 APPENDIX B

X-RAY DIFFRACTOGRAMS AND PROFILE FITTING RESULTS

Table 16. X-ray diffractograms in this Appendix.

Diffractogram Sample
Figure 21. Control cellulose
ABX987.RD before alkaline extractions
AB1074.RD after 0.01 N NaOH extraction
AB1154.RD after 0.1 N NaOH extraction
AB1194.RD after 1.0 N NaOH extraction

Figure 22. Cellulose produced in CMC

ACM981.RD before alkaline extractions
CM1071.RD after 0.01 N NaOH extraction
CM1151.RD after 0.1 N NaOH extraction
CM1191.RD after 1.0 N NaOH extraction

Figure 23. Cellulose produced in mannan

NMA989.RD before alkaline extractions
MA1075.RD after 0.01 N NaOH extraction
MA1155.RD after 0.1 N NaOH extraction
MA1195.RD after 1.0 N NaOH extraction

Figure 24. Cellulose produced in xylan

AXY985.RD before alkaline extractions
XY1073.RD after 0.01 N NaOH extraction
XY1153.RD after 0.1 N NaOH extraction
XY1193.RD after 1.0 N NaOH extraction

Figure 25. Cellulose produced in xyloglucan

AXG983.RD before alkaline extractions
XG1072.RD after 0.01 N NaOH extraction
XG1152.RD after 0.1 N NaOH extraction
XG1192.RD after 1.0 N NaOH extraction

101
102
106
109
 APPENDIX C

RAMAN SPECTRA

Table 17. Raman spectra of the following samples in this Appendix.

Figure 29. Control cellulose

3875-98-7 before alkaline extractions
3875-107-4 after 0.01 N NaOH extraction
3875-115-4 after 0.1 N NaOH extraction
3875-119-4 after 1.0 N NaOH extraction
Figure 30. Cellulose produced in..CMC
3875-98-1 before alkaline extractions
3875-107-1 after 0.01 N NaOH extraction
3875-115-1 after 0.1. N NaOH.extraction
3875-119-1 after 1.0 N NaOH extraction
Figure 31. Cellulose produced in mannan
3875-98-9 before alkaline extractions
3875-107-5 after 0.01 N NaOH extraction
3875-115-5 after 0.1 N NaOH extraction
3875-119-5 after 1.0 N NaOH extraction

Figure 32. Cellulose produced in xylan

3875-98-5 before alkaline extractions
3875-107-3 after 0.01 N.NaOH extraction
3875-115-3 after 0.1 N-NaOH extraction
3875-119-3 after 1.0 N NaOH extraction
Figure 33. Cellulose produced in xyloglucan
3875-98-3 before alkaline extractions
3875-107-2 after 0.01 N NaOH extraction
3875-115-2 after 0.1 N NaOH extraction
3875-119-2 after 1.0 N NaOH extraction
 APPENDIX D
CONFORMATIONAL COMPONENT RESOLUTION OF RAMAN SPECTRA

COMPUTER PROGRAM USED FOR RESOLUTION OF RAMAN SPECTRA (From ref. 77)
5 REM Analysis of Raman spectrum for the composition of kI, kII, and kO.
6 REM Filename: ANA3COMP.BAS
10 DIM DTA(7500)
20 K= 0
30 L= 0
40 M= 0
50 N= 0
60 = 0
70 P= 0
80 Q= 0
90 Q1= 0
100 Q2 = 0
110 REM
120 OPEN "I",#1,"kI.prn"
140 FOR I = 0 TO 500
160 INPUT #1,DTA(I)
170 NEXT I
180 CLOSE #1
200 OPEN "I",#2,"kII.prn"
210 FOR I = 2048 TO 2548
220 INPUT #2,DTA(I)
230 NEXT I
240 CLOSE #2
300 OPEN "I",#3,"kO.prn"
310 FOR I = 4096 TO 4596
320 INPUT #3,DTA(I)
330 NEXT I
340 CLOSE #3
400 OPEN "I",#4,"unknown.prn"
410 FOR I = 6144 TO 6644
420 INPUT #4,DTA(I)
430 NEXT I
440 CLOSE #4
500 REM
510 REM
1110 FOR I = 0 TO 300
1120 A = DTA(I)
1130 B = DTA(I+2048)
1140 C = DTA(I+4096)
1150 Y = DTA(I+6144)
1160 D = A*A
1170 E = B*B
1180 F = C*C
1190 G = A*B
1200 H = A*C
1210 J = B*C
1220 J1 = A*Y
116
 117
1230 J2 = B*Y
1240 J3 = C*Y
1250 K = K+D
1260 L = L+E
1270 M = M+ F
1280 N= N+ G
M=M+F
1290 N=N+G
0 = 0 + H
O=O+H
1300 P= P+J
1310 Q = Q + J1
1320 Q1 = Q1 + J2
1330 Q2 = Q2 + J3
1340 NEXT I
1350 Al = (L*M)-P*P
1360 A2 = (K*M)-O*O
1370 A3 = (K*L)-N*N
1380 A4 = (O*P)-(M*N)
1390 A5 = (N*P)-(L*O)
1400 A6 = (N*O)-(K*P)
1410 R = (K*L*M)+(2*N*O*P)-(P*P)*K-(O*O)*L-(N*N)*M
1420 S1 = (A1*Q)+(A4*Q1)+(A5*Q2)
1430 Xl = S1/R
1440 S2 = (A4*Q)+(A2*Q1)+(A2Q(A6*Q2)
1450 X2 = S2/R
1460 S3 = (A5*Q)+(A6*Q1)+(A3*Q2)
1470 X3 = S3/R
1480 U= 0
1490 FOR I = 0 TO 300
1500 A = DTA(I)
1510 B = DTA(I+2048)
1520 C = DTA(I+4096)
1530 Y = DTA(I+6144)
1540 T = (Y-(X1*A)-(X2*B)-(X3*C))
1550 Tl = T*T
1560 U = U + Tl
1570 NEXT I
1580 Xl = INT ((Xl*1000)+.5)/1000
1590 X2 = INT ((X2*1000)+.5)/1000
1600 X3 = INT ((X3*1000)+.5)/1000
1610 V = (U/298)^.5
1620 W = ((A1/R)A.5)*V*1.96
1630 W1 = ((A2/R)A.5)*V*1.96
1635 W2 = ((A3/R)A.5)*V*1.96
1640 W = INT ((W*1000)+.5)/1000
1650 W1 = INT ((Wl*1000)+.5)/1000
1660 W2 = INT ((W2*1000)+.5)/1000
1670 PRINT "KI = ";X1;"+-";W
1680 PRINT "KII = ";X2;"+-";W1
1690 PRINT "KO = ";X3;"+-";W2
1693 TOTAL = XI + X2 + X3
1695 PRINT "TOTAL: ";TOTAL
1697 PRINT "SUM OF SQUARED ERRORS: ";U
1700 END
 118

Table 18. Conformational components of bacterial celluloses.

CON CMC MAN XYL XGL

k, 0.01 N 0.730.05 0.380.04 0.590.04 0.380.04 0.440.04

kI 0.1 N 0.770.03 0.690.04 0.670.02 0.510.04 0.510.03
kI 1.0 N 0.730.02 0.580.03 0.560.02 0.500.02 0.580.02
ki1 0.01 N -.070.16 0.060.15 -.23+0.14 0.110.15 -.100.12
k1i 0.1 N 0.040.11 -.120.12 -.060.08 0.160.13 0.030.09
kii 1.0 N -.050.08 -.010.09 -.020.08 -.060.08 0.090.06

ko 0.01 N 0.310.15 0.530.14 0.620.13 0.430.15 0.620.12

k o 0.1 N 0.170.10 0.410.12 0.370.08 0.300.12 0.450.09
k o 1.0 N 0.320.07 0.430.08 0.430.07 0.550.07 0.330.06

CON: Control cellulose without additive; CMC: Cellulose produced with

CMC; MAN: Cellulose produced with mannan; XYL: Cellulose produced with
xylan; XGL: Cellulose produced with xyloglucan.
 APPENDIX E
PHOTOMICROGRAPHS OF BACTERIAL CELLULOSE RIBBONS

I
d,..

1.
.i- -

4)

Figure 34. Photomicrographs of control cellulose; no. 2412, 2535, and

2538; scale bar: 0.3Am, X48,000.

119
 120

K c, . '-
In
..--
'
' ' i-
:
a ,
r A
I.'M
.

CMC,
no. 2446: 0.4% CMC; scale bar: 0.3Am, X48,000.
 121

no. 2492: 0.4% mannan, and no. 2494: 0.1% mannan; scale bar: 0.3gm,
X72,000.
 122

-M*-- -
Figure 37. Photomicrographs of xylan-cellulose; no. 2452: 0.8% xylan,
scale bar: 0.4 #im,X36,000; no. 2457: 0.4% xylan, and no. 2465: 0.1%
xylan, scale bar: 0.3 jm, X48,000.
 123

Figure 38. Photomicrographs of xyloglucan-cellulose; no. 2507 and 2509:

0.8% xyloglucan, no. 2514 and 2524: 0.1% xyloglucan; scale bar: 0.3 im,
X48,000.