Scanning X-ray Microdiffraction Studies of the Molecular Architecture of
Biological Tissues
A Dissertation Presented
by
Jiliang Liu
to
The Department of Electrical and Computer Engineering
In partial fulfillment of the requirements for the degree of
Doctor of Philosophy
in
Electrical and Computer Engineering
Northeastern University
Boston, Massachusetts
Aug 2015
ii
To my Family
The road to be a doctor is hard and lonely to me, especially when I try to achieve this
at foreign country. The spiritual support from my family is such essential to me for Ph.D.
period. My parents, my mother, always have strong faith on me, even at my most desperate
time. I could not image my success without their support. Thank you, mama and dad. I very
much appreciate my wife as well. She came to US and worked very hard to graduate as a
Master, she overcame all this for that we could get together. The toughest and happiest
moments of my Ph.D., we went through, will be deeply kept within my heart forever.
iii
Contents
List of Figures........................................................................................................................... vi
List of tables ............................................................................................................................. ix
List of Acronyms ....................................................................................................................... x
Acknowledgments .................................................................................................................... xi
Abstract of the Dissertation ...................................................................................................... xi
Chapter 1 Introduction ........................................................................................................... 1
1.1 Scanning X-ray Microdiffraction(SXMD) ...................................................................... 1
1.2 Myelin Sheath .................................................................................................................. 2
1.3 Arabidopsis ...................................................................................................................... 5
1.4 Brain tissue of from Alzheimer’s Disease (AD) Subjects ............................................... 6
Chapter 2 Experimental Procedures ...................................................................................... 7
2.1 SXMD .............................................................................................................................. 7
2.2 Neuron fibril for study of myelin sheath ......................................................................... 8
2.3 The longitudinal stem of Arabidopsis.............................................................................. 8
2.4 Brain tissues from AD patient ......................................................................................... 9
Chapter 3 Study of myelin sheath by SXMD ...................................................................... 10
3.1 Method for SXMD to study myelin ............................................................................... 10
3.1.1 Circular averaging of SXMD diffraction patterns .................................................. 10
3.1.2 Determination of periodicity of myelin sheath ....................................................... 11
3.1.3 The determination of orientation of myelin for membranes ................................... 11
3.1.4 The radiation damage of SXMD for myelin sheath ................................................ 11
3.2 Results ........................................................................................................................... 16
3.3 Discussion ...................................................................................................................... 20
3.3.1 Comparison with previous x-ray diffraction studies of nerve myelin. ................... 20
3.3.2 Deformation of myelin lamellar structure ............................................................... 21
iv
3.3.3 Study of structural variation of myelin sheath for fixed neuron fibers ................... 22
3.3.4 Repetition of Myelin Sheath for 1 µm beam and comparison SXMD to the TEM 22
Chapter 4 Study of molecular architecture of the plant cell wall in wild type Arabidopsis
................................................................................................................................................. 24
4.1 Method for studying on the plant cell of Arabidopsis by SXMD .................................. 26
4.1.1 X-ray Fluorescence Microscopy (XFM) ................................................................. 26
4.1.2 Fiber content calculation(Two Component Model) ................................................ 26
4.2 Results for wild type Arabidopsis.................................................................................. 29
4.2.1Anatomy of the Longitudinal Section ...................................................................... 29
4.2.2 Microfibril angle and cellulose orientation ............................................................. 33
4.2.3 Constituents of the Two Component Model ........................................................... 36
4.2.4 Ratio of Amorphous to Fibrillar cellulose .............................................................. 37
4.2.5 Wax diffraction pattern from the epidermis............................................................ 44
4.2.6 Mineral inclusions ................................................................................................... 47
4.3 Discussion ...................................................................................................................... 51
Chapter 5 Study of molecular architecture of plant cell walls in Arabidopsis harboring
mutations in lignin biosynthesis .............................................................................................. 55
5.1 Method for studying on the plant cell of Arabidopsis by SXMD .................................. 56
5.1.1 Separating oriented diffraction from patterns of SXMD and new fiber content
calculation ........................................................................................................................ 56
5.1.2 Determination of Microfibril Angle........................................................................ 59
5.1.3 Calculation of coherence length of crystallite ........................................................ 59
5.1.4 Analysis of small angle reflection........................................................................... 59
5.2 Results from lignin mutated Arabidopsis ...................................................................... 63
5.2.1 Studies of molecular architecture of 16 samples by SXMD ................................... 63
5.2.2 Oriented Fiber Content ........................................................................................... 71
5.2.3 Crystalline order in the cellulose fibrils .................................................................. 72
5.2.4 Microfibril Angle .................................................................................................... 73
v
5.2.5 Axial Coherence Length ......................................................................................... 74
5.2.6 Packing of cellulose fibrils ..................................................................................... 75
5.3 Discussion ...................................................................................................................... 84
5.3.1 Mutations in the lignin biosynthetic pathway affect deposition and degradation of
cellulose ........................................................................................................................... 84
5.3.2 Mutated lignin changes the orientation of cellulose fibrils ..................................... 85
5.3.3 The regularity of cellulose fibrils varied with lignin mutants ................................. 86
5.3.4 Packing of cellulose fibrils ..................................................................................... 87
5.3.5 Molecular architecture in the stem of lignin mutants ............................................. 87
Chapter 6 Study of structural variation of Amyloid (Aβ) fibril within brain tissue of AD
subjects .................................................................................................................................... 89
6.1 Method for studying on the brain tissue by SXMD ....................................................... 89
6.1.1 Background subtraction for patterns of SXMD ...................................................... 89
6.1.2 Identification accumulation of Aβ fibril by SXMD ................................................ 90
6.1.3 The Krakty plot of small angle diffraction ............................................................. 91
6.2 Results ........................................................................................................................... 96
6.3 Discussion .................................................................................................................... 101
Chapter 7 Future of SXMD ............................................................................................... 103
Bibliography .......................................................................................................................... 105
vi
List of Figures
Figure 1 The beamline 23IDB (GM/CA) at the Advanced Photon Source (APS) at Argonne
National Laboratory................................................................................................................... 4
Figure 2 the calculation of intensity of circular averaging ...................................................... 13
Figure 3 The periodicity of diffraction pattern ........................................................................ 13
Figure 4 The orientation of myelin sheath stacks is normal to the corresponding reflections.14
Figure 5 The multiple stacks of myelin in the scattering volume give rise to multiple sets of
lamellar reflections. ................................................................................................................. 14
Figure 6 Test for radiation damage ......................................................................................... 15
Figure 7 Microdiffraction from single, teased fibers of myelinated axons. ............................ 18
Figure 8 Mapping of Spacing of the most intense x-ray reflection in each pattern ................ 18
Figure 9 Mapping of the principal orientations of lamellar membranes in the region of a node
of Ranvier ................................................................................................................................ 19
Figure 10 Two component model intensity fitting for each microdiffraction pattern. ............ 28
Figure 11 R-factor - a measure of the goodness of fit for each diffraction pattern. ................ 28
Figure 12 Images of a 100 μm thick longitudinal section through the Arabidopsis stem: ...... 31
Figure 13 Different parts of the stem exhibits diffraction patterns with distinct features . ..... 32
Figure 14 MFA is determined by the anlge between splitted (2 0 0) reflection of cellulose
fibrils ....................................................................................................................................... 34
Figure 15 Trend of variation of MFA across the stem of Arabidopsis .................................... 35
Figure 16 Determination of amorphous component models. .................................................. 39
Figure 17 Two component model intensity fitting for each microdiffraction pattern. ............ 40
Figure 18 Comparison of calculated and observed intensities for seven diffractions typical to
various regions of the stem. ..................................................................................................... 41
Figure 19 Distribution of amorphous and fibrillar fractions of the two component model as
determined by fitting of equatorial data as shown in Figure 18. ............................................. 42
Figure 20 Comparison of integral intensity of (0 0 4) reflection and elemental fibrils. .......... 43
Figure 21 Diffraction pattern from the epidermis exhibits sharp reflection at small angle part .
................................................................................................................................................. 45
Figure 22 Wax secretions could be determined by the power of scattering at the edge of the
stem. ........................................................................................................................................ 46
Figure 23 Diffraction patterns from Column 27-30 show mineral diffraction peak seen as
bright points at the xylem and phloem. ................................................................................... 48
Figure 24 X-ray Fluorescence micrographs of the distribution of metals in the thin section of
Arabidopsis stem. .................................................................................................................... 49
vii
Figure 25 Each microdiffraction pattern shown in grid was separated to oriented and
disoriented pattern ................................................................................................................... 57
Figure 26 Integral intensity of oriented component and intensity of disoriented component . 57
Figure 27 Calculation of oriented fiber content....................................................................... 58
Figure 28 Determining the MFA by identify split scattering from cellulose crystallite. ........ 61
Figure 29 Calculation of axial coherent length for (0 0 4) reflection. ..................................... 61
Figure 30 Interference calculation and model and interference between two cylinders. ........ 62
Figure 31 Fitting of small intensity distribution of mutated sample. ...................................... 62
Figure 32 Wild type sample dried 3 and 30 days after harvest. .............................................. 63
Figure 33 Aldehyde sample dried 3 and 30 days after harvest. ............................................... 64
Figure 34 Aldehyde in G sample dried 3 and 30 days after harvest. ....................................... 65
Figure 35 High G sample dried 3 and 30 days after harvest. .................................................. 66
Figure 36 High S sample dried 3 and 30 days after harvest. ................................................... 67
Figure 37 High H sample dried 3 and 30 days after harvest. ................................................. 68
Figure 38 Ferulate incorporated lignin sample dried 3 and 30 days after harvest. .................. 69
Figure 39 Low lignin sample dried 3 and 30 days after harvest.............................................. 70
Figure 40 A longitudinal section of Arabidopsis stem was studied by SXMD. ...................... 77
Figure 41 Diffraction pattern was separated into circularly symmetric and oriented fractions.
................................................................................................................................................. 77
Figure 42 Example from high S shows separation of circularly symmetric and oriented
intensities. ................................................................................................................................ 78
Figure 43 Comparison of the highest oriented fiber content observed in each sample. .......... 79
Figure 44 8 Arabidopsis variants show intensities of the strongest equatorial reflections of
cellulose I. ............................................................................................................................. 79
Figure 45 Depiction of cellulose fibrils helically wrapped around a plant cell. ...................... 80
Figure 46 Increase of microfibril angle indicates alignment of cellulose fibrils tend to
transverse from the periphery of the stem to the center .......................................................... 80
Figure 47 Samples dried fresh and after 30 days in water display different histogram of axial
coherence length. ..................................................................................................................... 81
Figure 48 Comparison of maximum coherence lengths for each of the 16 samples. .............. 82
Figure 49 Modulation of small angle scattering by interference due to packing of cellulose
fibrils was observed to be strongest in the region near the pith. ............................................. 83
Figure 50 Eliminating the scatterings from mica window ...................................................... 92
Figure 51 Scattering from brain tissue. ................................................................................... 92
Figure 52 Extraction of intensity of tissues from pattern including strong background. ........ 93
Figure 53 Intensity distributions without background for SXMD on control ......................... 93
viii
Figure 54 Image of scanning features from section of Alzheimer Disease sample. ............... 94
Figure 55 The Krakty plot of small angle reflection for plaques accumulated region. ........... 95
Figure 56 Background subtracted diffraction pattern. ............................................................. 98
Figure 57 Montage of integral intensity for small angle and reflection at 4.7Å. .................... 98
Figure 58 Integral intensity maps correspond to difference brain sample. ............................. 99
Figure 59 Special small angle reflections. ............................................................................. 100
Figure 60 SAXS intensity for three amyloid fibrils models. ................................................. 102
ix
List of tables
Table 1 Coherence Length of Wax Crystals ............................................................................ 50
Table 2 Comparison of d-spacings of Bragg peaks observed in microscanning patterns with
those expected for specific mineral inclusions. ....................................................................... 50
x
List of Acronyms
AD Alzheimer’s Disease
CAA Cerebral Amyloid Angiopathy
Aβ Amyloid beta
MFA Microfibril angle
SXMD Scanning X-ray Micordiffraction
TEM Transmission Electron Microscopy
WT wild type
xi
Acknowledgments
I am very much appreciated for Professor Makowski’s supervising and directing of
my Ph. D. program. His enthusiasm on science and kindness to students deeply influence me
on the aspects of my research and life attitude. Cooperation with Professor Makowski is
always productive and very enjoyable. I would thank to all my colleagues from Makowksi’s
lab as well, especially, Dr. Inouye supplies much help on understanding theory of X-ray
diffraction.
I would like to thank Professor Kirschner, for his very useful suggestion on study of
myelin. Dr. Kirschner’s lab prepares the nerve fiber for SXMD study on myelin sheath. My
first SXMD analysis is basing on wonderful data from Dr. Riekel’s Lab at ESRF—ID13.
We collect significant amount of data from GM/CA at APS. Dr. Fischetti and Dr.
Venugopalan give us incredible help during data collection. Without their contribution, I
could not obtain results from SXMD. I thank very much to them and all other colleague at
GM/CA.
Arabidopsis samples were provided by Professor Chapple’s group from Purdue
University. Both wild type and lignin biosynthesis mutants exhibit affinity molecular
structural information. Their samples preparations are essential for our SXMD studies.
Therefore, I am sending my greeting to Professor Chapple and colleagues from Chapple’s
group.
I thank to Isabel Costantino for preparing brain section from Alzheimer’s subjects at
MGH.
In the end, my research was supported by the Center for Direct Catalytic Conversion
of Biomass to Biofuel (C3Bio), an Energy Frontier Research Center funded by the U.S.
department of Energy, Office of science, Basic Energy Science under Award #DE-
SC0000997.
xii
Abstract of the Dissertation
Scanning X-ray Micordiffraction Studies of the Molecular Architecture of
Biological Tissues
by
Jiliang Liu
Doctor of Philosophy in Electrical and Computer Engineering
Northeastern University, Aug 2015
Principle Adviser name, Dr. Lee Makowski
X-ray scattering is an important method to study atomic and molecular structures.
Here, I apply scanning x-ray micro-diffraction, a new advanced synchrotron technology, to
study the molecular structure of three tissues: 1. Myelin within the peripheral nervous system
(PNS); 2. Plant cell walls in Arabidopsis stems; 3. Protein aggregation in human brain
sections from Alzheimer's patients. A suite of custom software was developed to overcome
the challenge of processing a large amount of data collected by scanning micro diffraction
and to extract complex features from the scattering patterns of these different tissues. These
improvements in software have greatly expanded the utility scanning microdiffraction
technology for analysis of detailed information about the molecular architecture of myelin in
the nodal, paranodal, and juxtaparanodal regions; the structural heterogeneities within the
Arabidopsis stem; and pathological molecular structures that arise in Alzheimer's disease.
We anticipate significant expansion of the use of this method for studies of the molecular
architecture of intact tissues and the alteration of these structures due to wounds, specific
mutations or pathological conditions.
Keywords: Scanning Micro diffraction, Feature extraction, Data processing, Myelin sheath,
Arabidopsis, Cellulose, Amyloid, Fibril, Molecular structure
1
Chapter 1
Introduction
The development of scanning x-ray microdiffraction at synchrotron facilities has
opened the possibility of investigating structural variations within the tissues at μm to sub-μm
level resolutions. Compared to traditional x-ray scattering methods that use millimeter-sized
x-ray beams and average scattering from all material within samples, scanning x-ray
microdiffraction provides extensive information on the order and characteristics of
crystallized and non-crystallized material. Fratzl et al., 1997 (Fratzl, 1997); Riekel et al.,
1997 (Riekel, 1997) and Lichtenegger et al., 1999 (Lichtenegger, 1999) studied cellulose
structure in hard wood cell wall and human bone at the European Synchrotron Radiation
Facility (ESRF). The hardware used in our studies was developed for use by protein
crystallographers to identify well-ordered regions of poorly diffracting crystals. The
structure of G-protein-coupled receptors (GPCRs), contributing to the award of a Nobel prize,
was solved through the use of micro-diffraction applied to micro-crystals at sector 23
(GM/CA CAT) at the Advanced Photon Source (APS) (Rasmussen, 2007) where our data
collection was carried out.
Scanning biological tissues with micro-beams will produce thousands of times the
data with rich biological molecular structure information, compared with traditional methods.
It is not unusual for a scanning microdiffraction data set to exceed 10 GB, or, depending on
the sample and the specific questions asked, over 100 GB. To generate this information
requires automatic data capture and data analysis methods. The development of computation
methods combines traditional X-ray methodology and modern image processing algorithms.
Therefore, my Ph.D. project focuses on developing the unique custom software to
automatically extract information about molecular features at the nanoscale from scanning
microdiffraction data.
1.1 Scanning X-ray Microdiffraction(SXMD)
Many biological macro molecules, such as GPCRs, crystallize very poorly and well-
ordered domains of these crystals are often limited to no more than 5-10 μm in size. To study
the structure of these micro-crystals, the development of micro beams was required.
2
Currently, the most advanced synchrotron facilities, such as APS (Chicago, U.S.), ESRF
(Grenoble, France), and Spring-8 (Hyogo Prefecture, Japan) have beam lines capable of
producing micro-beams from 0.5-5 μm in diameter. The diffraction signals from micro-
crystals are highly enhanced by the micro-beam with the reduction of background scattering
from solvent and sample support. (Smith, 2012). Interests in membrane proteins such as
GPCRs have driven the development of microbeam facilities, greatly enhancing their study.
Use of microbeams for scanning x-ray microdiffraction has benefited greatly from these
advances.
Furthermore, the micro-beam opens the possibility of study of molecular structure
within real tissues. Conventional X-ray diffraction provides insight only into the averaged
molecular structure over scattering volumes hundreds or thousands of cubic μm in size. The
averaging implicit in the use of large scattering volumes in heterogeneous materials precludes
collection of data on crucial detailed architecture relevant to function or dysfunction of these
molecules. However, by scanning the real tissues with a micro beam, we revealed spatially-
dependent variations. For instance, a 1 μm beam, originated from beam line ID13 at ESRF,
enabled the study of structural variations of myelin sheath at specialized regions known as
the internode, juxtaparanode and node of Ranvier. Through data collection utilizing a 5 m
beam at GM/CA 23IDB at the APS, we uncovered tissue-specific molecular architecture in
the stem of Arabidopsis and studied the impact of mutations in lignin biosynthesis on the
structures of these tissues. Finally, we used the 5 μm beam scanning to carry out preliminary
studies of structural variations of fibrils of Aβ within plaques in tissue from subjects with
Alzheimer's Disease (AD). Figure 1 shows GM/CA at APS.
1.2 Myelin Sheath
Transmission of electrical signals along a nerve axon is greatly accelerated by the
presence of an insulating myelin sheath that is interrupted periodically with gaps for
electrical transmission. The structure of myelin sheath provides important clues to the
molecular basis of electrical transmission by the axon. The structure of myelin sheath in the
intermodal region had been studied for decades. Caspar and Kirschner, (Caspar and
Kirschner, 1971) determined the electron density map of the double-membrane structure of
the myelin sheath, calculated corresponding structure factor and interpreted their results in
3
terms of the known molecular constituents of myelin. Inouye et al., (1999) solved the P0
protein assembly in myelin membrane by X-ray diffraction.
However, the structure of the conducting gaps, known as the nodes of Ranvier,
remained poorly characterized (Poccia et al., 2013). Because of their small 4-5 m size,
conventional X-ray techniques were unable to address their structure. The electrochemical
properties of the system has not been completely elucidated in the absence of knowledge of
structural changes in the region of the nodes. Here we used a 1 μm beam, raster-scanned
across a single nerve fiber to extract detailed information about the structure of the nodes of
Ranvier and that of myelin in the vicinity of the nodes. Data from scanning microdiffraction
was used to determine the orientation of the lamellar membrane stack of the myelin assembly
and the membrane periodicity of myelin as it varied spatially in the nodal, paranodal, and
juxtaparanodal regions.
4
Figure 1 The beamline 23IDB (GM/CA) at the Advanced Photon Source (APS) at Argonne National
Laboratory.
The left figure is an aerial view of the synchrotron facilities (APS.ANL.GOV). The upper-right figure
shows the goniometer and sample holding arrangements at beamline 23IDB. The lower right figure shows
the enjoyable time of collecting data at GM/CA.
5
1.3 Arabidopsis
Plant cells have diameters of 10-40 m and can have lengths of the order of a
millimeter. The molecular architectures of plant cell walls are highly heterogeneous at
multiple scale levels. The structural inhomogeneity of plant cell walls are caused by the
variations in composition of plant cells. For decades, studies of plant cells revealed that three
main component of plant cell walls are high molecular weight polysaccharides, highly
glycosylated proteins and lignin. (Fraser, 2011) Cellulose fibrils, one of the most important
polysaccharide components, are essential elements for the scaffold of the cell wall. The
average structure of individual cellulose fibrils has been well studied by traditional X-ray
fiber diffraction methods. However, the relationship of function-structure of plant cell wall
remains obscure without detailed molecular information, especially about the diversity of
organization of cellulose fibrils. SXMD provides the possibility of re-investigating the
structural variation of cellulose within tissues at μm-level resolution. Riekel et al., (Riekel,
2001) applied this technique to wood in order to characterize the helical organization of
cellulose fibrils around plant cells. Here, we expand this approach for analysis of the
structural heterogeneities within the Arabidopsis stem. A raster made of 5 columns including
38 images each was collected from an Arabidopsis stem. The raster of diffraction patterns
distinguishes the five specific tissues arranged radially from the pith in the center, to the
epidermis at the periphery. The vascular tissues, xylem and phloem and cortex are set
between the pith and epidermis. X-ray fluorescence and scanning micro-beam also enable us
to study composites of mineral and wax present in the stem.
Furthermore, we reveal the interdependence among the synthesis and assembly of
difference polymeric species by comparison of plant cell wall of wild type (WT) and mutated
Arabidopsis. The analysis of SXMD of wild type and mutated Arabidopsis exhibits
significant structural changes of crystallized and non-crystallized material. The results also
suggest that some variants undergo an accelerated, time-dependent degradation of cellulosic
organization only modestly present in WT plants.
6
1.4 Brain tissue of from Alzheimer’s Disease (AD) Subjects
Finally, we carried out preliminary investigation on neural lesions associated with
Alzheimer’s Disease by SXMD. Alzheimer’s Disease (AD) is the only cause of death in the
top 10 in America that cannot be prevented, cured or slowed. It currently costs roughly $226
billion for treatments, and in an aging society, the costs are expected to increase dramatically
in the coming decades. It draws huge social attention (Alzheimer's Association, 2015) . We
applied micro-diffraction technology to scan pathological protein deposits in human brain
from anatomy material of subjects with Alzheimer's Disease. The goal is to elucidate the
molecular architecture of these deposits.
Amyloid plaque structure has been studied by X-ray scattering for decades. Eanes
and Glenner (Eanes and Glenner, 1968) studied the Aβ fibrils by X-ray diffraction and
indicated a cross-β conformation in fiber direction on the basis of strong reflections at
spacings of 4.7 Å and 10 Å Kirschner et al., (1986) reported purified samples from
Alziheimer’s subjects exhibit a sharp reflection at 4.7 Å and a diffuse one at about 10Å.
Recently, Tycko et al 2005 suggested that a single nucleation site may provide the seed for
formation of amyloid fibrils that propagate throughout the brain as the disease progresses.
Understanding the structural variation of amyloid fibrils in different parts of the brain will
provide evidence relevant to this hypothesis and important clues as to the mechanism of
aggregation of Aβ (Lu et al., 2013) .
With SXMD, we studied x-ray scattering from 18 μm thick sections of brain tissues
using a 5 μm beam and scanning with a 5 μm step size. The results of brain experiment
indicate the organization of pathological protein deposits at the margins of gray matter and
around the vasculature in Cerebral Amyloid Angiopathy (CAA). Krakty plots of small angle
region of SXMD indicates the presence of variations in the fibril structures of Aβ for both
healthy elderly people and AD subjects.
7
Chapter 2
Experimental Procedures
We applied SXMD to study molecular structure of three complex tissues: 1.
Peripheral Nervous System (PNS) myelin; 2. Plant cell walls from the stems of Arabidopsis;
3. Sections of human brain tissue from AD subjects and 'healthy' controls. By identifying the
‘fingerprint’ reflections of known constituents, like myelin membranes; cellulose fibrils and
amyloid fibrils, we observed the molecular structural variation within μm scale of these
tissues.
We have collaborated with scientists at two of the most advanced beamlines around
the world, ESRF-ID13 and 23IDB at APS. ESRF-ID13 provides 1μm beam enabling us to
scan single myelinated nerve axons. 23IDB at APS can deliver a 5 μm diameter x-ray beam,
that we have used to scan stems from Arabidopsis and human brain tissue sections. Details
of the experimental setup of SXMD and sample preparation of biological tissues are
discussed below.
2.1 SXMD
SXMD on myelin was performed at the ESRF-ID13 beamline. A monochromatic
beam was focused to a spot 1 µm full-width at half-maximum. The diffraction pattern was
recorded using a CCD detector. A single image frame contained 1024x1024 pixels with a
158 µm pixel size. The specimen-to-detector distance was 203.9 mm which was calibrated by
the Bragg reflections of silver behenate powder indexed by the fundamental period of 53.38
Å.
Micro-scanning experiments for Arabidopsis and brain sections were conducted at
23IDB beamline at the APS. The beam size was 5 μm, the wavelength of beam is 1.033 Å.
For scattering from wild type Arabidopsis we collected diffraction pattern images with
2048x2048 pixels and pixel size of 72 Å. The specimen-to-detector distance was 300 mm.
For scanning of lignin-mutated Arabidopsis and brain sections we binned pixel data to
produce diffraction pattern images with 1024x1024 pixels with a 144 µm pixel size. Features
of the diffraction patterns were not sufficiently sharp to warrant the smaller pixel size - no
information was lost in the binning. The specimen –to-detector distance was 416 mm.
8
2.2 Neuron fibril for study of myelin sheath
Myelin samples were prepared by Dr. Kirschner lab. Briefly, sciatic nerves were
dissected from mice that had been euthanized using CO2 inhalation followed by cervical
dislocation. Under a dissecting microscope, single fibers from mouse sciatic nerves that had
been fixed for 10-30 min in 2% paraformaladehyde-2.5 % glutaraldehyde (in 0.12 M
phosphate-buffered saline, at pH 7.4) were teased apart with very fine forceps after removal
of the perineurium using a 26-gauge, stainless-steel needle. The fibers with a small volume of
adhering solution were then aspirated into 0.7 mm-diameter x-ray capillaries, and sealed with
wax and fingernail polish enamel.
x-ray microdiffraction from myelin sheath was performed at the ESRF-ID13
beamline by Kirschner, Riekel and Burghammer. Inouye and Liu analyzed the collected data.
2.3 The longitudinal stem of Arabidopsis
Arabidopsis thaliana (Col-0 ecotype) plants were grown in a potting mix (Redi-
Earth; Scotts, http://www.scottspro.com) by the Chapple laboratory at Purdue University.
Primary inflorescence stems were harvested from six-week-old plants. Transverse and
longitudinal sections of the tissue were dried and cut into 100 µm thick sections using a
microtome. These sections were shipped to Northeastern University for mounting for x-ray
scattering experiments.
We further studied stems from seven variants of Arabidopsis thaliana with
engineered defects in lignin biosynthesis. These mutations were High G (C4H:F5H fah1),
High S (C4H:F5H), low lignin (ref3-2), ferulate incorporate lignin (ccr1), High H
(ref8ref4rfr1), Aldehyde (cad-c cad-d) and Aldehyde of G (cad-c cad-d fah1). These
mutations will affect the biosynthesis of lignin by suppression or enhancement of the
expression of G, S and H lignin units.(Meyer et al 1996, Turner et al 1997, Meyer et al 1998,
Goujon et al 2003, Sibout et al 2005, Schilmiller et al. and 2009, Bonawitz et al 2014)
Arabidopsis samples were cultured by Dr. Joanne C Cusumano and Dr. Jeong Im
Kim from Professor Clint Chapple’s lab. x-ray Microdiffraction was performed at GM/CA
CAT, APS-23IDB by Jiliang Liu and Lee Makowski with the assistance of Dr. Nargarajan
Venugopalan and Dr. Robert F. Fischetti.
9
2.4 Brain tissues from AD patient
Human brain harboring pathological protein deposits was identified and sectioned by
Isabel Constantino of Massachusetts General Hospital (MGH). The brain was fixed in
formaldehyde and embedded within paraffin for cutting into 18 μm thick sections by
microtome. The sections were heated to melt and remove paraffin and then washed by xylene,
ethanol and water to remove any residual paraffin. X-ray Microdiffraction was performed
by Jiliang Liu and Lee Makowski with the assistance of Dr. Nargarajan Venugopalan and Dr.
Robert F. Fischetti at GM/CA CAT, APS-ID23.
10
Chapter 3
Study of myelin sheath by SXMD
Myelin sheaths wrapping up axon gave rise to stacks of membrane with one
dimensional lamella structure. Caspar and Kirschner (Caspar and Kirschner, 1971)
determined the electron density map of the double membrane structure of myelin sheath,
calculated corresponding structure factors and interpreted their results in terms of the known
molecular constituents of myelin. Because of large x-ray beam size used in conventional x-
ray studies, the molecular architecture of the node of Ranvier with diameter of 4-5 μm had
not been studied using x-rays. This prevents complete elucidation of the electrochemical
properties of the system. Here we used a μm-sized beam, raster-scanned across a single nerve
fiber to collect diffraction data from the various regions around the nodes. From these
patterns, we extracted the essential structural information about the organization of
membranes in the nodal, paranodal, and juxtaparanodal regions including orientation and
membrane periodicity and electron density profiles.
3.1 Method for SXMD to study myelin
SXMD produces thousands of diffraction patterns that contain abundant structural
information. The size of SXMD data sets demands development of software to automatically
extract the essential structure information, such as periodicity and orientation of myelin
sheath. This section describes the custom algorithm that was developed and designed to
extract these features from the SXMD diffraction patterns.
3.1.1 Circular averaging of SXMD diffraction patterns
To calculate the intensity distribution of circular averaged patterns, we transform
index of pixels of patterns into polar coordinate systems with center information provided by
beamline. Then every pixel could be assigned to a coordinate including a radius to center and
an azimuthal angle to horizontal. We accumulate the intensities of pixels with same radius.
Then the circularly averaged intensity is determined by summing the intensities and dividing
by the number of pixels with corresponded radius. Figure 2 shows the center of pattern and
the determination of radius to calculate the circular averaged intensity.
11
D-spacing is the inverse of reciprocal spacing, 1/d. The calculation of reciprocal
coordinate is that: , θ is the Bragg angle, λ is the wavelength of X-ray.
(Guinier, 1994)
3.1.2 Determination of periodicity of myelin sheath
The repeat period of myelin sheath corresponds to the center-to-center distance of
lamellar membrane-pairs in the myelin sheath. Periodicity of myelin was determined by the
D-spacing of lamellar reflections at small angles. As shown in Figure 3, the reflections of
2nd-4th Bragg order for myelin could be well fit by three Gaussian functions after background
subtraction. D-spacing of each order is the D-spacing of the best position of its fitting
Gaussian curve. The D-spacing of each order should be the periodicity divided by the order.
The optimum estimate of repeat period was taken as the mean of the periodicity calculated
from the three observed reflections.
3.1.3 The determination of orientation of myelin for membranes
The stack of myelin membranes produces a series of Friedel reflections in the small
angle region. Figure 4 shows that the orientation of myelin lamella is normal to these
reflections. Therefore, we could determine the orientation of myelin lamella by identifying
the azimuthal angle of reflections. The calculation of orientation was complex at the
paranode and node region, as shown in Figure 5, due to the presence of multiple pairs of
Friedel reflections, indicating multiple stacks of myelin lamella in these regions. The
intensity distribution as a function of azimuthal angle was calculated by averaging the
intensity over a range of scattering angles (radii) ΔR. The positions of peaks in the intensity
distribution correspond to the azimuthal angle of reflections. The orientation of myelin
lamella was calculated as the angle half-way between the two corresponding peak positions.
3.1.4 The radiation damage of SXMD for myelin sheath
Synchrotron facilities provide high energy X-rays, capable of inducing significant
radiation damage to biological tissue samples. For our experiments on fixed myelin sheath at
ID13 at ESRF we used an exposure time of 500 msec. This was chosen after experiments to
determine the exposure time at which radiation resulted in negligible damage to tissue. To
determine the radiation damage to real sample, a single position in the sample was exposed
12
for a series of 200 msec exposure. Intensity of the 2nd-4th order myelin reflections were
monitored and no reduction in intensity was detected for exposures of less than 500 msec
(Figure 6). Even after 2 sec of exposure, little or no reduction of intensity was detected in
the 2nd-4th order of reflections.
13
Figure 2 Calculation of intensity of circular averaging
The left figure is a typical x-ray pattern from scanning microdiffraction, the right figure shows the intensity
distribution after calculation of circular averaging.
Figure 3 Periodicity of diffraction pattern
From the plot of circularly averaged intensity vs scattering angle the positions of the strong 2nd-4th
order lamellar reflections were used to calculate the periodicity of myelin lamellar. The peaks of plot
are related to 2nd
(blue), 3rd
(green) and 4th
(red) reflections. The lower-right image is the
enlargement of the lower-left image for x-axis range from 0 -75 pixel.
r
0 100 200 300 400 500 6000
50
100
150
200
250
Radius
inte
nsity
10 20 30 40 50 60 70
0
5
10
15
20
25
30
35
inte
nsity
Radius
14
Figure 4 Orientation of myelin sheath stacks is normal to the corresponding reflections.
Figure 5 Multiple stacks of myelin in the scattering volume give rise to multiple sets of lamellar
reflections.
The left figure indicates the two different reflection sets. The right figure shows azimuthal distribution of
intensity. The red and yellow arrows correspond to the reflections within red and yellow pi shapes.
0 50 100 150 200 250 300 350 40010
20
30
40
50
60
70
80
90
100
110
angle(degree)
inte
nsity
image mesh4 0480(2rd order)
15
Figure 6 Test for radiation damage (A) The small-angle region recorded for 200 msec per pattern at a single position along the nerve. (B)
Radial-averaged intensity for the sequentially-recorded patterns. The total time for the 10 patterns was 2.0
sec. The spectra show small variations of intensity, but no overall decrease in intensity, which indicates
little or no structure degradation due to radiation damage.
A
B
16
3.2 Results
Inouye et al, 2014 reported the structural variation at nodal, paranodal, and
juxtaparanodal regions as determined by SXMD. Figure 7A provides a mapping of lamellar
orientations in region scanned. Figure 7B is an optical micrograph of the same region and
Figure 7C is a larger region containing the area scanned.
The mapping of D-spacings enabled us to identify the variation of periodicity of
myelin sheaths across the grid. The D-spacing of the reflections exhibiting the highest
intensity was mapped onto a grid that correlated to the scanning positions for SXMD (Figure
8). The clusters of the D-spacing for patterns of SXMD was shown in histograms of
integrated intensities of identified Bragg reflections (Figure 8). As Figure 8B, C shows,
four principal clusters of D-spacing were identified. They were exhibited by four different
gray levels corresponding to d > 80 Å (black), 60 < d < 80 Å (dark gray), 45 < d < 60 Å (light
gray), and d < 45 Å (white) within the mapped grid (Figure 8A). These four clusters
corresponded to morphologically distinguishable regions of the nerve fibers. The paranodal-
nodal region was dominated by patterns with maxima in the range 45<d<60 Å, corresponding
to the 4th order of x-ray scatter from the myelin lamellae. Internodal regions were dominated
by scattering that had the greatest intensity in the range 60<d<80 Å, typically the 3rd order
lamellar reflection, which suggests that the molecular organization of the multilamellar
myelin in the internodal regions was different from that of the paranodal-nodal region.
Finally, regions that had a maximum intensity at d>80 Å were most often in the central
portion of the fiber (corresponding to face-on scattering, detailed below). This scattering is
most likely associated with the lateral organization of proteins in the plane of the membrane,
but could also be associated with the 2nd-order reflection of lamellar scattering from stacks of
myelin membranes.
The distribution of the orientation of the principal lamellar reflections for two
different raster scans, one from a single fiber and the other from the overlapping fibers (left
and right panels, respectively), mapped onto the grids as shown in Figure 9. The single lines
of Figure 9 correspond to the orientation normal to the Friedel pair reflections of each
diffraction pattern. These lines are parallel to the membrane planes and show how the
membrane wrapping shifts from being directed around the long axis of the fiber in the
internodal region to being perpendicular to the fiber at the node of Ranvier. Moreover, as
discussed below, some diffraction patterns exhibited multiple sets of lamellar reflections at
17
different angles about the center of the diffraction patterns, indicating the presence of two (or
more) lamellar stacks of membranes rotated relative to one another.
18
Figure 7 Microdiffraction from single, teased fibers of myelinated axons.
(A) Montage of the small-angle portions of diffraction patterns from a raster scan of a pair of teased nerve
fibers. The vertically-oriented nerve has a node of Ranvier slightly above its crossing with the horizontally
oriented nerve. Individual frames in (A), circled and numbered, were chosen for detailed analyses described
below. (B) Optical micrograph of the same field of view as the montage. (C) A larger field of view of the
nerve fibers.
Figure 8 Mapping of Spacing of the most intense x-ray reflection in each pattern
Mapping of spacing d (A-C) for the x-ray reflection having the greatest intensity in each diffraction pattern
of the raster-scan. (A) Map of positions with greatest intensity in three different ranges of scattering angle
as indicated by gray level. (B) Scatter plot of the different spacing as a function of sequential image number
indicating the most common scattering angles of intensity maxima. (C) Histogram of the data shown in (B).
19
Figure 9 Mapping of the principal orientations of lamellar membranes in the region of a node of
Ranvier
Mapping of the principal orientations of lamellar membranes in the region of a node of Ranvier (arrows)
for a single myelinated nerve (left) and a pair overlapping fibers (right). The vast majority of reflections in
the internode correspond to membranes oriented parallel to the long axis of the fiber. In the vicinity of the
node of Ranvier the lamellar stacks curl around and become more nearly perpendicular to the axis of the
nerve fiber (arrows). In the diagram on the right, where two distinct orientations were apparent in the
diffraction pattern, the orientations of both sets of lamellar reflections are shown by long red (stronger) and
short blue lines (weaker), respectively.
20
3.3 Discussion
3.3.1 Comparison with previous x-ray diffraction studies of nerve myelin.
The nerve myelin sheath has been studied for decades. Solving molecular architecture
of nerve myelin sheath by X-ray scattering has led to an elucidation of the distribution of
protein and lipid in the membranes and the role of those components in myelin assembly,
stability, and function (Caspar and Kirschner, 1971; Kirschner, 1992) However, because the
size of a single nerve fiber is typically 10–30 µm in the PNS, and the axial sizes of the node,
paranode, and juxtaparanode regions are 1 µm, 5 µm, and 10 µm, respectively (Arroyo,
1999), studies using conventional x-ray beams that are 100 m in diameter resulted is
observed diffraction that was dominated by the abundant, compact myelin arrays in the
internodal segments, but was in fact, a sum of scattering from different regions of the myelin
(Blaurock, 1966; Kirschner, 1984). Recently, the improvement of synchrotron facilities have
produced X-ray beams with µm, and even nm dimensions. In order to study the heterogeneity
of myelin structure, we applied 1 µm beam size produced by ESRF ID13 for study of the
molecular architecture of the myelin sheath in and bordering the internode.
Myelin (from sciatic nerve (PNS) and brain (CNS)) both freeze-dried (Ducic, 2011)
and fixed in formalin (De Felici, 2008; Yagi, 2009) has been studied previously by
synchrotron diffraction. Dehydration for freeze-drying of samples separates the membranes
into lipid-rich and protein-rich domains having different periodicities (Hollingshead et al,
1981), precluding the study of structural heterogeneities. Therefore, to obtain the regional
diversity of molecular architecture within the myelinated nerves we prepared glutaraldehyde-
fixed samples which cross-links molecules that are in proximity to one another, preventing
phase separation (Hirano, 1982).
Glutaraldehyde-fixed mouse sciatic nerves are slightly altered in structure due to
fixation. A slightly larger period of 178 Å compared to the native 176 Å was reported in
previous study by conventional X-ray ((Kirschner and Hollingshead, 1980). The electron
density distribution shows that the cytoplasmic membrane separation for the glutaraldehyde
treated myelin is smaller than that for the control, while the extracellular separation is larger.
Further, the electron density level at the cytoplasmic space is higher than at the extracellular
surface, suggesting that the glutaraldehyde crosslinks the abundant lysine and arginine
residues (Salem, 2010) in myelin basic protein (MBP). Our analysis of SXMD data indicated
21
a repeat period of 198 Å to 202 Å, larger by 22 to 26 Å, than that for fresh myelin. As shown
in electron density profile, while a cytoplasmic separation (34 Å) for extracellular of myelin
was similar to that of fresh myelin (32 Å), fixed samples exhibit significant larger
cytoplasmic separation,~48 Å, and intermembrane separation, ~41 Å. The changes in the
widths of the cytoplasmic appositions strongly indicate that packing of membranes varied in
the internodal region.
The study of SXMD shows that glutaraldehyde fixed samples have a larger
extracellular separation than fresh myelin (i.e., 60 Å for #11 and #18, and 53 Å for #42,
(positions marked in Figure 7A), compared to 48 Å for fresh myelin). Previous study of
glutaraldehyde-fixed myelin (Kirschner and Hollingshead, 1980) reported swelling at the
extracellular apposition as well, but not to the extent observed here. The absence in the teased
single fibers of the mechanical constraint provided by collagen (Rand, 1979) may give rise to
the greater extracellular swellings.
3.3.2 Deformation of myelin lamellar structure
Within the internodal region we commonly observed two or more distinct lamellar
domains with different orientations. The geometric features of wrapping of membrane stacks
around the axon gave rise to the most intense set of lamellar reflections. The weaker sets of
lamellar reflections correspond to a smaller population of membranes stacked at an angle—
usually close to 90o—to the dominant membrane packing direction. These weaker reflections,
which were observed in many of the en face diffraction patterns, may come from lamellar
layers deformed at Schmidt-Lanterman incisures (Mugnaini, 1977; MacKenzie, 1984).
Interestingly, we observed different intensity distribution for these two sets of lamellar
reflections in same diffraction patterns. This implied that the packing of layers at the
Schmidt-Lanterman incisures maybe different to the principal membrane within internodal
region. A highly localized tilting of the membrane stacking causes local deformation as well,
the deformations lead to the unequal intensity of centrosymmetrically-related reflections
(Friedel’s pairs), as shown in the vertically-oriented patterns #42, 449, 481, 635, and 637 in
Figure 7A.
In the center of the nerve fibers, many patterns contain two orthogonal sets of
reflections. Because the membrane planes of center region is normal to the incident beam,
lamellar stacks of membranes scatter from unexpected angles as described above, which
22
produces weaker reflections. The dominant scattering in these patterns can be accounted for
by membrane proteins arranged within the plane of the membrane. Electron microscopy
reveals that the continuous, small pocket of cytoplasm that defines the spiraling Schmidt-
Lanterman incisures results in a swelling deformation of the neighboring cytoplasmic space
(see review (Trapp and Kidd, 2004)); however, the SXMD observations suggests a more
complex geometry of the deformed myelin layers. Analyzing the orientation and positioning
of the reflections recorded in the current study using an even smaller x-ray beam—e.g., 0.5
m—may enable a more accurate three-dimensional reconstruction of incisures.
3.3.3 Study of structural variation of myelin sheath for fixed neuron fibers
The richness of information provided by scanning x-ray microdiffraction and the
details of molecular organization that this information illuminates suggests that similar
approaches using fresh nerve tissue will provide powerful new strategies for understanding
the underlying molecular foundation of a broad spectrum of myelinopathies. The studies of
myelin sheath membrane by SXMD indicate orientation of the lamellar membrane stacks and
membrane periodicity varied spatially. In the juxtaparanode-internode, 198-202 Å-period
membrane arrays oriented normal to the nerve fiber axis predominated, whereas in the
paranode-node, 205-208 Å-period arrays oriented along the fiber direction predominated. In
parts of the sheath distal to the node, multiple sets of lamellar reflections were observed at
angles to one another, suggesting that the myelin multilayers are deformed at the Schmidt-
Lanterman incisures. The significant increase in lateral resolution in mesh scans will allow a
more precise differentiation and identification of neighboring domain structures, not only in
myelinated fibers but also in other biological tissues and non-biological materials (Schroer,
2008)
3.3.4 Repetition of Myelin Sheath for 1 µm beam and comparison SXMD to the TEM
Because the periodicity of myelin sheath is around 200 Å, a 1 µm x-ray beam could
involve ~50 myelin sheath for scattering. Scanning nerve fibers with a 1 µm beam could
generate structural information with enough repetition of myelin sheath. Electron
microscopy (EM) studies show that the length of nodal and paranodel region are 3 µm and 13
µm. A 1 µm beam enables us to studiy the structural variation at these places. EM image of
myelin sheath indicated that membrane curls over at node of Ranvier region and is
consistence with our observation of orientation changes of myelin sheath at the nodal region.
23
The swolling of myelin ic quantitated by the increases of periodicity at node and paranodal
region observed here.
24
Chapter 4
Study of molecular architecture of the plant cell wall in wild
type Arabidopsis
There are more than 200 tons of biomass produced annually in the US, providing the
potential to prtially addressenergy needs by converting biomass to biofuel. However, the
digestibility of the plant cell wall, which is essential for biofuel production, is hindered by
rigid molecular structures. In order to break down this structure, we need to fully understand
its molecular architecture. SXMD was applied to study of the stems of Arabidopsis, a
commonly studied plant with complete genomic sequence available, to understand the
structural variation in real tissues and the interdependence of biosynthesis and assembly of
distinct constituents.
Here we use SXMD with μm size beam studing the five tissues of Arabidopsis stem
individually - pith, xylem, phloem, cortex and epidermis. These distinct tissues contain
corresponding cell types. As the stem matures, cellulose fibrils within the cell walls are
broken down to form amorphous components. These structural changes lead to hollowing of
stem and the expansion of cell wall in pith. The dense xylem cells constitute vascular tissues
for water and nutrient transport to support plant growth and survival. A solid layer of
relatively small cells encompass the phloem from the cortex, which is essential for stiffness
of the stem. The epidermis is coated by a wax-rich surface layer for protection of the plant.
We studied the physical variation of cell wall caused by cellulose organization or orientation
within these tissues at the sub-cellular level by SXMD.
Longitudinal sections of wild type Arabidopsis were prepared for the position-
resolved measurement of the local fibril orientation and organization by SXMD. X-ray
diffraction patterns of a 100 μm thick section were measured every 5 μm to map a region of
190 x 25 μm, spanning from the center of a stem to its periphery. We observed distinctly
different diffraction patterns in the interior, central and exterior regions of the stem. Based
on microbeam techniques, we identified the variation in abundance of crystalline and
amorphous cellulose; the number of cellulose chains per fibril; and the alternation of local
fibril orientation across cells. Diffraction analysis of the location and nature of mineral
inclusions was supplemented with x-ray fluorescence microscopy. The main structures of
plant wax in the epidermis were determined by analysis of the small- and wide-angle
25
scattering in the outer layer of the stem. From this work we constructed a comprehensive
analysis of the molecular architecture within the plant cell walls of the constituent tissues of
the Arabidopsis stem.
26
4.1 Method for studying on the plant cell of Arabidopsis by SXMD
4.1.1 X-ray Fluorescence Microscopy (XFM)
X-ray fluorescence is a very sensitive probe of elemental content of many metals, and
x-ray fluorescence microscopy (XFM) can provide information on deposition of mineral
elements within cell wall. We are able to identify and estimate the possible mineral by
combining the analysis of XFM and SXMD. Scanning x-ray absorption microscopy and x-
ray fluorescence microscopy (Paunesku T., 2006) were carried out on the samples to obtain
images of the subcellular architecture of the plant cell walls to provide corroborating
information on the distribution of minerals in the samples. Facilities at beamline 2ID-E at
the Advanced Photon Source (APS) were used to collect images in both stepping and fly scan
mode. The undulator x-ray beam of the APS was monochromated by using a double crystal
monochromator and a zone plate objective focused the x-rays onto the specimen. An order
sorting aperture was used to reject unfocused x-rays to reduce background. The sample was
raster scanned through the focal spot, and at each scan position, illuminated with x-rays.
Incident x-rays excite photo electrons resulting in vacancies in the inner shells of atoms that
are filled by outer shell electrons, either through an Auger or fluorescence process where the
excess energy is carried away by a photon. The fluorescence derived photons are detected by
an energy dispersive detector. By measuring the amount of electron hole pairs generated by
each photon, the chemical element from which it originated can be deduced. Since the
number of detected fluorescence photons is linear with the quantity of material present in the
illuminated spot, the amount can be quantified through use of elemental standards. Typically,
10 or more elements were quantitated simultaneously at each scan position. Images were
analyzed using the software package — MAPS (Vogt, 2003).
4.1.2 Fiber content calculation (Two Component Model)
The structure of cellulose Iβ had been solved by Nishiyama et al. 2002, and is widely
believed as the primary component of cellulose crystallite within the real plant cell wall
(Nishiyama Y., 2002). We evaluated a series of models of fibrils and sub-fibrillar aggregates
of cellulose that they were constructed using 1 to 36 Iβ cellulose chains. Equatorial scattering
for these aggregates was calculated using the cylindrical analog of the Debye formula
(Guinier A., 1994);
27
<I(1/d)> = jk fj(1/d)fk(1/d) Jo(2rjk1/d) 1
where <I(1/d)> is the cylindrically averaged equatorial intensity, 1/d is the radial component
of the reciprocal cylindrical coordinates, fj(1/d), the scattering factor of the jth atom, J0
corresponds to the zero order Bessel function and rjk the axially projected distance between
the jth and kth atoms. The sums are over all atoms in the axial repeat of the fibril. Following
other investigators (Fernandes, 2011; Thomas, 2013) we adjusted the a-axis of cellulose Iβ
crystal to fit the observed data. In this case, the changes of lattice of Iβ increase the lattice
constant along the a-axis from 7.78 Å to 8.09 Å.
The equatorial scatter was modeled as a two component system consisting of
amorphous and fibrillar forms of cellulose. Intensity was calculated as:
IModel = A·Ielementary fibrils + B · Iamorphous components 2
The intensities of two component — amorphous and fibrillar forms of cellulose were
determined by extensive modeling to identify those model components that most accurately
fit the observed scattering from specific tissues containing both amorphous and fibrillar
cellulose. A specific molecular model for the amorphous constituents was chosen on the
basis of comparison with scattering from pith where the amorphous scattering dominated;
fibrillar components were chosen to fit crystalline reflections observed in scattering from
vascular tissues where scattering from crystalline components was strongest. The amorphous
material is highly heterogeneous and includes amorphous cellulose; hemicellulose and other
non-cellulosic materials. The constituents giving rise to the amorphous scattering should not
be ignored. However, for convenience of comprehensive analysis at this study, we applied a
surrogate for this complex mixture — a single molecular model that closely fit the
amorphous scattering from the pith region. This, in turn, provided a basis for estimating the
ratio of amorphous to crystalline material within the various parts of the stem without bias as
to the identity of the amorphous materials. For each diffraction pattern, the constants of
proportionality, A and B were determined by a least-squares fit to the data. The sum of A and
B is equal to 1, subject to the linear combination feature of the fitting method.
28
Figure 10 Two component model intensity fitting for each microdiffraction pattern. (a) Intensity
calculated from elementry fibrils (blue) and amorphous component, which was simulated by a six
cellulose chains model (green). These curves have been normalized to indicate scattering from the
same number of electrons (by dividing the fibril scatter by 6). (b) Region of the equator over which
model fitting was carried out corresponding to the inset in (a).
Figure 11 R-factor - a measure of the goodness of fit for each diffraction pattern
(comparing the observed data with that calculated from a linear combination of amorphous and
fibrillar components).Data for all 5 rows of diffraction patterns in the montage are shown — each
row being rendered in a different color. R-factor = [(Iobserved - Icalculated)2]
1/2]/[ Iobserved
2]
1/2.
5 10 15 20 25 30 35 400
0.05
0.1
0.15
0.2
0.25
COLUMN
RF
AC
TO
R
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2x 10
6
1/d(1/ANGSTROM)
INT
EN
SIT
Y
Debye Calculation of Two Phase
0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.3 0.32-0.5
0
0.5
1
1.5
2
2.5
3
3.5x 10
4
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.3 0.320
20
40
60
80
100
120
1/d(1/ANGSTROM)
INT
EN
SIT
Y
Two Phase Model Calculation
Observed Intensity
Calculated Intensity
Intensity from Amorphous
Intensity from Fibrils
a b
29
4.2 Results for wild type Arabidopsis
4.2.1Anatomy of the Longitudinal Section
Figure 12a is an x-ray fluorescence image of the distribution of calcium in a
longitudinal section of Arabidopsis stem. Calcium is apparently distributed uniformly
through the plant cell walls. As a primary ionic elements of plant cell, the XFM of calcium
provides good contrast for visualizing the cell walls. Here we scanned only half of the stem,
the outside of the stem is at the right in optical image (Figure 12d); the center of the pith at
the left-hand edge of the field of view. The elongated shapes of plant cells are reflected in
the positions of cell walls. As Figure 12d shows, an optical microscope image taken of the
field of view indicates the location of the region in which we carried out scanning
microdiffraction. The quality of the image is compromised by the presence of a hole drilled
down the optical axis of the microscope to allow passage of the x-ray beam. The blue
rectangle indicates the location of the 190 μm x 25 μm (38 columns x 5 rows) grid on which
x-ray patterns were collected. Figure 12b is a montage of the central region of the 190
diffraction patterns in this grid and Figure 12c is a pseudo colored image of the grid coded to
indicate total scattering intensity. The distinctive characteristics of scattering from each
tissue enable us to estimate the approximate positions of pith, xylem, phloem, cortex and
epidermis. Little detail can be seen in the patterns in Figure 12c, and four representative
scattering patterns — from the pith, xylem, phloem and epidermis — are reproduced in
Figure 13.
We propose a prototypical fibrillar cellulose diffraction pattern based on the pattern
from the phloem in Figure 13c. The pattern exhibits relatively well oriented scattering that is
dominated by a pair of strong fingerprint reflections typical of cellulose crystallites. There are
two broad reflections on each side of the equator and a sharp reflection at wider angle on the
meridian of the pattern. The lower radius equatorial peak is a superposition of the (1 1 0) and
(1 -1 0) reflections from cellulose Iβ; and the (2 0 0) is the larger radius equatorial reflection.
The sharp meridional reflection is the (0 0 4). In contrast, the pattern from the pith (Figure
13a) is more typical of amorphous cellulose and is used as a model for scattering from
amorphous cellulose. The scattering in the pattern from pith is completely disoriented at
angles that span those of the strong equatorial peaks seen in Xylem (Figure 13b). In
scattering from pith, the (0 0 4) peak is particularly weak, indicating low cellulose content.
30
Figure 13b, from xylem, is very similar to Figure 13c except for the presence of numerous
sharp, punctate reflections due to crystalline mineral inclusions. Scattering from the cortex is
not readily distinguishable from that of the phloem. Figure 13d shows that scattering from
cellulose is nearly absent in the epidermis. Instead of cellulose scattering, we observed a
sharp peak at a spacing of 4.13 Å and a series of relatively sharp, small angle peaks due to
the lamella structure of waxy surface coating of the stem. Figure 13e shows the evolution of
the equatorial intensity from pith to epidermis; the transition from amorphous cellulose
scatter to fibrillar cellulose and concluding with the sharp small angle peaks from wax and
the almost the complete absence of scattering from cellulose in the penultimate layer. The
background intensity was calculated from the first and last patterns of each row in the series,
these diffraction patterns scatter at place just outside the sample.
31
Figure 12 Images of a 100 μm thick longitudinal section through the Arabidopsis stem: (a) X-ray
fluorescence microscope image of the distribution of calcium. The distribution of Calcium is
relatively even in the cell walls making Ca-distribution a high-contrast visualization of the location of
cell walls in the sample. Scale bar is 100 μm. (b) A montage of the 190 micro diffraction patterns
taken at the positions marked in (d). (c) A false color rendering of the total intensity of scattering in
each pattern of the montage. (d) An optical microscope image of the stem showing the grid that
corresponds to the positions at which microdiffraction patterns were taken. Scale bar is 20 μm.
a.
b.
1 5 10 15 20 25 30 35
1 3
5
d.
c.
32
Figure 13 Different parts of the stem exhibits diffraction patterns with distinct features. (a)
Diffraction from the pith region in the interior of the stem exhibits a single broad, disoriented
reflection. (b) Diffraction from the xylem region that the intensity distribution contains typical
reflections from cellulose fibrils. Random distributed sharp peaks appears to be scattering from
mineral inclusions; (c) diffraction from the phloem region exhibits prototypical cellulose fibril
scattering; (d) sharp small angle and wide-angle reflections typical of wax were observed in
diffraction from the epidermis (inset is an enlargement of the center of the pattern). (e) Equatorial
intensity as a function of units from the center of the diffraction pattern for 38 patterns spanning
from pith to epidermis.
a b
c d e
33
4.2.2 Microfibril angle and cellulose orientation
In many regions of the stem, we observed scattering that appeared to be a
superposition of two fiber patterns rotated at an angle to one another. This observation is
caused by the presence of two populations of cellulose microfibrils rotated relative to one
another. These two populations arise from the fact that cellulose microfibrils appear to wind
around plant cells in a helical path (Cave, 1966). Because of the geometry of helical
wrapping, microfibrils tilt relative to the principal axis of the elongated cells. The diffraction
patterns contain the scattering from both the front and back of the cell. Therefore, the
superposition could be separated with the two patterns rotated by an angle equal to twice the
angle between the microfibrils and the principal axis of the cell. The variation of this angle
has been studied at millimeter resolutions in wood (Cave, 1966) and more recently at higher
spatial resolutions by microdiffraction (Lichtenegger, 1999; Muller, 1998.). Here, we
measured the angular separation between the two superimposed patterns by fitting Gaussian
angular profiles to the data as shown in Figure 14. Measurement of the microfibril angle
(MFA), the angle between the cellulose fibers and the longitudinal axis (half the angle
between diffraction peaks), was possible using data from column 12 to column 35 of the
montage (Figure 15) . These areas correspond to pith, xylem, phloem and cortex and the
inner part of the epidermis. In the pith (columns 1-11) cellulose exhibits little azimuthal
orientation and from outer epidermis (columns 36-38) little scattering from cellulose is
discernible.
Figure 15a shows the variation in azimuthal positions of the two peaks (in the right
and left halves of the patterns) and 14b the corresponding MFA (half difference of angle
between the two peaks) as a function of column position in two rows of the montage. As
discussed above, x-axis of Figure 15 corresponds to the column 12 — 15 of scanning grid of
microdiffraction. The microfibril angle decreases from about 30o in the pith to about 10o in
the vessel tissue region. It remains at about 10o through the xylem and phloem regions and
decreases to near zero in the cortical region. The punctate approach to zero microfibril angle
in the pith/xylem region occurs at cell wall boundaries where the tilt of the cellulose fibrils is
either directly towards or away from the x-ray source.
34
Figure 14 MFA is determined by the angle between split (2 0 0) reflection of cellulose fibrils. (a)
Measurement of the azimuthal positions of the (2 0 0 ) for a diffraction pattern from column 23. The
angular separation was microfibril angle. The ring in black represents annulus over which the
intensity was measured. Inset shows the small angle area of the diffraction pattern indicating two
orientations of the microfibrils (MF). Diffraction peaks of microfibrils split at small angle and (2 0 0)
reflections (shown with arrow). (b) Azimuthal average intensity plot of ring indicated in (a). The red
lines correspond to background. The sharp low intensity feature is the shadow of the beam stop
holder. The straight red line is the minimum of plot outside the shadow . Two Gaussian functions
(green) are fit to the background-subtracted data and correspond to the azimuthal positions of the (2
0 0). The microfibril angle difference determined by distance between peak positions of the two
Gaussian functions.
a b
50 100 150 200 250 300 3500
10
20
30
40
50
60
70
80
90
100
long dry 5 2.0023
AZIMUTH(DEGREE)
INT
EN
SIT
Y
35
Figure 15 Trend of variation of MFA across the stem of Arabidopsis. (a) is the measured azimuthal
positions for peaks in two of the rows of the montage. (b) is MFA derived from the data in (a).
Horizontal axes represents the column position from outer pith to inner epidermis. x-axis of Figure 15
corresponds to the column 12 — 15 of scanning grid of microdiffraction.
10 15 20 25 30 35 4060
80
100
120
COLUMN
AZ
IMU
TH
AL P
OS
ITIO
N
10 15 20 25 30 35 40220
240
260
280
300
COLUMN
10 15 20 25 30 35
0
10
20
30
COLUMN
AZ
IMU
TH
AL S
EP
ER
AT
ION
10 15 20 25 30 35
0
10
20
30
COLUMN
a b
36
4.2.3 Constituents of the Two Component Model
The pith contains abundant amorphous material including cellulose (about 30%) and
non-cellulosic components including lignin, hemicellulose, pectin and other minor
components. The heterogeneous mixture of constituents provides very distinctive pattern
with complex structural information. (Turner and Somerville, 1997). The lignin fraction
appears to be relatively small in pith (Zhong, 2000). Agarwal (Agarwal, 2013) reported that
amorphous cellulose, lignin and xylan (hemicellulose) exhibit broad reflections at scattering
angles similar to the (1 1 0)/(1 -1 0) and (2 0 0) reflections from cellulose Iβ. Although the
scattering from lignin and xylan exhibit peak positions slightly shifted from the observed data,
the nature of scattering from all of these components do not appear significant different from
what we observed in scattering from pith. This makes it exceedingly difficult to separate
scattering from individual components in this region. For qualitative analysis, we introduce
the two-component model to estimate the relative amounts of crystalline and amorphous
materials in the various tissues of the stem. Then, we need only to formulate a two-
component model that will account for all the scattering in the observed patterns and provide
a set of relative weights for amorphous and crystalline scattering at each observation point.
As a surrogate for the complex mixture of amorphous materials in the pith, we utilized a
model for amorphous cellulose that accounted for all the data and provided a metric for
estimating the relative amorphous content in each diffraction pattern.
Comparison of equatorial scattering from pith with that calculated from a wide range
of cellulosic aggregates (Figure 16) indicates that a unit of 4-6 chains formed by two layer
can account for virtually all the equatorial scattering observed from the pith in the region of
the (1 1 0)/(1 -1 0) and (2 0 0) reflections. On this basis, we used the scattering from a two-
layer, 6-chain aggregate of cellulose as a surrogate for estimating the relative scattering of
amorphous materials across the stem.
Diffraction patterns from vessel tissues of the stem - xylem and phloem, include two
equatorial reflections distinct to cellulose; one corresponding to the superposition of the (1 1
0) and (1 -1 0) reflections and the other to the (2 0 0) reflection of cellulose Iβ. Crystalline
cellulose-microfibrils give rise to those reflections. We studied various microfibril models,
including a 24 chain model (Fernandes, 2011; Thomas, 2013) and a 36 chain model (Ding
and Himmel, 2006). Extensive modeling carried out here indicated that the observations were
37
best fit using a 36 chain model with an elliptical/hexagonal cross section as suggested by
Ding and Himmel (Ding and Himmel, 2006).
4.2.4 Ratio of Amorphous to Fibrillar cellulose
Diffraction patterns involve scattering from a mixture of amorphous material and
cellulose microfibrils, the fraction for amorphous and microfibrils varied with tissues
specification. Therefore, we interpreted the equatorial scattering in terms of a linear
combination of scattering predicted from an elliptical, 36 chain model for the crystalline fibril
plus a two layer surrogate model for the amorphous components as shown in Figure 16.
Linear combinations of these two components proved adequate to fit the scattering from all
regions of the stem (Figure 18). A and B in Error! Reference source not found. are the
constants of proportionality of amorphous and microfibrils for each patterns. This enables us
to qualitatively estimate the relative amounts of fibrillar cellulose and amorphous
components in the different regions (Figure 19 and Figure 20).
Cellulose fingerprint reflections were generally not observed in pith, but rather
scattering in this region exhibited a single broad peak, as shown in Figure 18a. The diffuse
scattering extended from near the expected position of the (1 1 0)/(1 -1 0) reflection to just
beyond the (2 0 0) reflection with a profile similar to that predicted to arise from aggregates
of cellulose with 4-6 chains or hemicellulose. The scattering in pith was azimuthal isotropic
indicating little orientation for the amorphous material. The weak (0 0 4) reflection in this
region (Figure 13a) is suggestive of relatively low cellulosic content, disorder of fibrils in
the axial direction and the presence of non-cellulosic components. The weak (0 0 4)
reflection also suggests that what amorphous cellulose may be present may be either short
crystalline fragments, randomly oriented (Fink, 1987) or highly bent and twisted fibrillar
forms (Paakkari and Sermaa, 1989). The quality of intensity fitting from pith region to
epidermis region indicates that two phase model provides adequate estimation, as Figure 18
shown, where blue curves in each figure correspond to observed intensity. The relative
proportions of crystalline cellulose and amorphous components change dramatically with
position as shown in Figure 19a and b. The fibrillar content, Figure 19c, is defined as the
proportion of cellulose in the form of elementary, 36-chain fibrils. The residual, R-factor,
Figure 19d, a measure of the discrepancy between calculated and observed, is less than 15%
for nearly all patterns. This analysis allows the percentage of cellulose chains in elementary
38
fibrils ('apparent fibrillar content') to be calculated as a function of distance across the stem.
Because of the complexity and heterogeneity of the molecular architecture of cellulosic and
non-cellulosic material, we introduce 'apparent' fibrillar content. It should not be ignored that
individual cellulose molecules or single layers of molecules give rise to only very modest
scattering in the region of the equator analyzed. Conversely, the presence of non-cellulosic
material, such as hemicellulose, lignin and pectin, that may contribute to scattering in this
region should be included in estimates of amorphous content, leading to an underestimate of
the fibril content. Finally, disorder in the crystalline lattice of the fibrils (Fernandes, 2011)
may lead to an underestimate of fibrillar content. On this basis, we only provides relative,
not absolute, estimates for fibril content as a function of distance across the stem. The
highest apparent fibrillar content comes from diffraction patterns in column 29 or 32 of each
row (shown in Figure 19) , with the highest value observed being ~33%. This region may
correspond to cambium between xylem and phloem, where growth of cell wall is most active.
Gradual degradation of cell wall observed during maturation is suggestive of the decrease of
fiber content from xylem to pith region. In the pith region, lignin reduction causes that
microfibrils lose protection and tend to break down. Our data indicate the products of that
cellulosic material degradation may include many relatively small cellulose aggregates
composed of two molecular layers of cellulose, the presumptive products of this breakdown
process.
39
Figure 16 Determination of amorphous component models. Comparison of six small cellulose
crystallites are shown. A double layer model exhibits the best fit to observations.
0.1 0.15 0.2 0.25 0.3 0.350
2
4
6
8
10
12
1/d(ANGSTROM)
INT
EN
SIT
Y
long dry 5 2.0005
Observed Intensity
A6 Chains
2 Chains
4 Chains
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
f.XYLEM and PHLOEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
2
4
6
8
10
12
1/d(ANGSTROM)
INT
EN
SIT
Y
long dry 5 2.0005
Observed Intensity
A6 Chains
6 Chains
3 Chains
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
f.XYLEM and PHLOEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
40
Figure 17 Two component model intensity fitting for each microdiffraction pattern. (a) Scattering
from amorphous material as obseved at pith (blue) and scattering from air (green). Background is
based on the intensity from the first column, just outside the sample tissue. (b) Intensity of
diffraction from the most crystalline region (blue) and corresponding background from air (green).
(c) Intensities were modeled as a linear sum of scattering from two phases - amorphous and
crystalline. Scattering calculated from elementry fibrils (blue) and amorphous component, which
was simulated by a six cellulose chains model (green). (d) Region of the equator over which model
fitting was carried out corresponding to the inset in (c).
a b
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.450
10
20
30
40
50
60
70
1/d(1/ANGSTROM)
INT
EN
SIT
Y
Observed Intensity
Intensity of Background
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.450
20
40
60
80
100
120
140
160
180
1/d(ANGSTROM)
INT
EN
SIT
Y
Observed Intensity
Intensity of Background
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2x 10
6
1/d(1/ANGSTROM)
INT
EN
SIT
Y
Debye Calculation of Two Phase
0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.3 0.32-0.5
0
0.5
1
1.5
2
2.5
3
3.5x 10
4
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.3 0.320
20
40
60
80
100
120
1/d(1/ANGSTROM)
INT
EN
SIT
Y
Two Phase Model Calculation
Observed Intensity
Calculated Intensity
Intensity from Fibrils
Intensity from Amorphous
c d
41
Figure 18 Comparison of calculated and observed intensities for seven diffractions typical to various regions of
the stem.
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
a.PITH
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
b.PITH
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
c.PITH and XYLEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
d.PITH and XYLEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
e.XYLEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
f.XYLEM and PHLOEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
g.PHLOEM and CORTEX
1/d(1/ANGSTROM)
INT
EN
SIT
Y
42
Figure 19 Distribution of amorphous and fibrillar fractions of the two component model as
determined by fitting of equatorial data as shown in Figure 18. (a) Integral intensity of elementary
fibrils (b) Integral intensity of amorphous fraction. (c) apparent fibrillar content. (d) R-factor - a
measure of the goodness of fit for each diffraction pattern. Five different color curves correlate to
data for all 5 rows of diffraction patterns in the montage. Calculation for images were shown as: (a)
Integral Intensity of elementary fibrils = A·Ielementary fibrils; (b) Integral intensity of amorphous component
= B·Iamorphous component; (c) Apparent FIbrillar Content = [A·Ielementary fibrils]/[A·Ielementary fibrils +B·Iamorphous
component ] ; (d) R-factor = [(Iobserved - Icalculated)2]
1/2]/[ Iobserved
2]
1/2.
a b
c d
5 10 15 20 25 30 35 400
2000
4000
6000
8000
10000
12000
COLUMN
INT
EG
RA
L IN
TE
NS
ITY
OF
FIB
RIL
S
5 10 15 20 25 30 35 400
0.05
0.1
0.15
0.2
0.25
0.3
0.35
COLUMN
AP
PE
RA
NT
FIB
RIL
LA
R C
ON
TE
NT
0 5 10 15 20 25 30 35 400
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5x 10
4
COLUMN
INT
EG
RA
L IN
TE
NS
ITY
OF
A
MO
RP
HO
US
FR
AC
TIO
N
5 10 15 20 25 30 35 400
0.05
0.1
0.15
0.2
0.25
COLUMN
RF
AC
TO
R
43
Figure 20 Comparison of integral intensity of (0 0 4) reflection and elemental fibrils. The trend of
integral intensity of reflection (0 0 4), figure on the left, is consistent with the trend of integral
intensity of elemental fibrils, figure on the right.
5 10 15 20 25 30 35 400
100
200
300
400
500
600
COLUMN
INT
EG
RA
L IN
TE
NS
ITY
OF
(0
0 4
) R
EF
LE
CT
ION
5 10 15 20 25 30 35 400
2000
4000
6000
8000
10000
12000
COLUMN
INT
EG
RA
L IN
TE
NS
ITY
OF
FIB
RIL
S
44
4.2.5 Wax diffraction pattern from the epidermis
A series of sharp small angle reflections combined with a wide-angle reflection at
approximately 4.13 Å spacing appear in the last one or two patterns in each row of the
montage. The outermost layer of the stem corresponds to the epidermis, which contains an
abundance of plant wax. The waxy crystallite gives rise to the small and wide reflections at
the peripheral region of stem. Spacings of the small-angle reflections suggested the presence
of two co-existing phases of wax; one exhibiting a 78.6 periodicity with strong peaks at the
second, fourth and sixth order; and a second phase with a periodicity of 81.7Å with strong
third and fifth order peaks. The chemical composition of the Arabidopsis stem and leaf
waxes have been investigated (Rashotte A. M., 2001; Rashotte A.M., 2004; Jenks M., 1995;
Hannoufa A., 1993). The predominant wax components of Arabidopsis stems are C29 alkane,
ketone, and secondary alcohol, together with smaller quantities of C28 primary alcohols and
C30 aldehyde. The wax load was higher at stems than other parts of Arabidopsis. (Suh M.C.,
2005). The small angle reflection of wax for our observations are consistent with the
presence of C29 alkane with a 78.6 periodicity and C29 ketone with an 81.7Å periodicity.
C29 alkane exhibits diffraction with strong even order peaks whereas C29 ketone exhibits
strong odd order peaks (Ensikat, 2006). Therefore wax compound of epidermis of stem
appear to contain two components - C29 alkane and C29 ketone. The average coherence
length of the two phases was determined by the breadths of the peaks, and used as a measure
of the size of the lamellar crystals giving rise to these reflections. The 78.6 Å phase
exhibited a coherence length of 1290 + 320 Å, and the 81.7 Å phase, a coherence length of
420 + 30 Å as estimated from data in column 37 of all five rows of the montage. (Coherence
length calculation see section 5.2.5, standard deviation for coherence length was determined
by 40 patterns from SXMD. )
The (1 1 0) is Scattering from the epidermis of Arabidopsis stem exhibits a strong
reflection at 4.13 Å, that is consistent with the (1 1 0) reflection – the most significant
reflection from the wax crystallites (Ensikat, 2006) in the wide-angle regime. As shown in
Figure 22, as the micro beam approaches the edge of the stem, the coherence length of the
wax crystal does not change appreciably. However, the integral intensity increases. That
indicates that the size of wax crystallite is independent of the wax deposition.
45
Figure 21 The diffraction pattern from the epidermis exhibits sharp reflection in the small angle
regime. (a) Wax crystallites give rise to strong diffraction peaks at small angle region in patterns
from epidermis. (b) Distribution of intensity along the meridional direction. Positions of five peaks
correspond to those expected in a diffraction pattern from C29 alkane (C29H60 ) or C29 ketone
(C29H58O). Green line represents estimated background, red curve corresponds to scattering
expected for C29 alkane, pink curve corresponds to that for ketone.
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.090
50
100
150
200
250
300
350
400
1/d(ANGSTROM)
INT
EN
SIT
Y
Wax Reflection at Small Angle
Observed
Background
Without Background
Ketone
Alkane
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
f.XYLEM and PHLOEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
a b
46
Figure 22 Wax secretions could be determined by the power of scattering at the edge of the stem. (a)
diffraction pattern exhibiting the strong reflection at 4.13 Å. (b) intensities along the equator of this
pattern including observed (blue), background (brown), background subtracted intensities (red),
modeled scattering from cellulose (cyan) and remaining scattering assigned to scattering from wax
(green). The green curve is background, red curve represents amorphous cellulose and the (110)/(-1 1
0) and (200) reflections from cellulose Iβ. (c) Coherence length of wax diffraction pattern as a
function of position across the epidermis. (d) Integral intensity of wax diffraction across the
epidermis.
a b
c d
0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.280
50
100
150
200
250
300
1/d(ANGSTROM)
INT
EN
SIT
Y
Wax Reflection of (1 1 0)
Observed
Background
Without Background
Cellulose
Wax
wax diffraction at 4.13 A
0.1 0.15 0.2 0.25 0.3 0.350
20
40
60
80
100
120
f.XYLEM and PHLOEM
1/d(1/ANGSTROM)
INT
EN
SIT
Y
30 32 34 36 380
100
200
300
400
Pattern Position
Cohere
nt Length
of R
eflection (
1 1
0)
30 32 34 36 380
100
200
300
400
Pattern Position
30 32 34 36 380
100
200
300
400
Pattern Position
30 32 34 36 380
100
200
300
400
Pattern Position
Cohere
nt Length
of R
eflection (
1 1
0)
Row1 Row2
Row4Row3
30 32 34 36 380
50
100
150
200
250
Pattern Position
30 32 34 36 380
50
100
150
200
250
Pattern Position
30 32 34 36 380
50
100
150
200
250
Pattern Position
Inte
gra
l In
tensity o
f R
eflection (
1 1
0)
30 32 34 36 380
50
100
150
200
250
Pattern Position
Inte
gra
l In
tensity o
f R
eflection (
1 1
0)
Row3
Row1 Row2
Row4
47
4.2.6 Mineral inclusions
There are many very sharp peaks superimposed on the broad reflections from
cellulose at xylem and phloem of stem. They are due to small crystalline inclusions of
minerals in plants. Figure 23 gives three representative patterns that contain scattering from
minerals. X-ray fluorescence microscopy of the sample was carried out to aid in the
identification of the minerals giving rise to this scattering. As shown in Figure 24, XFM
suggests the presence of inclusions rich in silicon and iron and apparent co-localization of
potassium and chlorine. Calcium and zinc appear uniformly distributed throughout the
Arabidopsis cell wall and do not appear involved in formation of mineral inclusions. These
uniform distributions may correspond to calcium and zinc forming complexes with pectin.
Demarty et al., 1984 (Demarty M., 1984) has shown that most tissue calcium is associated
with the middle lamella of the cell wall. Absence of co-localization of calcium with the
silicon appears to preclude the possibility that calcium silicates are present.
The observed mineral diffraction patterns were compared with patterns from mineral
powders. (webmineral.com; rruff.info) The comparison indicated that no one mineral could
account for all of the sharp reflections observed and the presence of no particular mineral
could be unambiguously confirmed on the basis of the data collected (Table 2). In addition
to the minerals listed in the Table 2, KCl might account for the presence of reflections at
3.13 Å since sylvite (KCl) exhibits a 3.15 Å peak. Of the possible silicon minerals, quartz
would give rise to a peak at 3.34 Å that is not observed. The presence of the strong reflection
at 6.4 Å appears difficult to account for on the basis of any minerals commonly observed in
plants. Although there seems to be evidence for the presence of several of these minerals in
the Arabidopsis stem, unambiguous determination would require additional data.
48
Figure 23 Diffraction patterns from Column 27-30 show mineral diffraction peak seen as bright
points at the xylem and phloem. D-spacing of each reflection is shown in text. Column 27-30
correspond to xylem and phloem region as shown in Figure 12b
4.02
4.32
4.95
3.82
2.95
3.12
4.99
3.55
a c b
49
Figure 24 X-ray Fluorescence micrographs of the distribution of metals in the thin section of
Arabidopsis stem. Zinc and calcium are evenly distributed in the cell walls of the stem. Silicon-rich
inclusions are abundant in the xylem and phloem regions of the stem. Scale bar is 20μm.
50
Table 1 Coherence Length of Wax Crystals
Diffraction
pattern of C29
Alkane
D-
spacing
(Å)
Coherence
length of C29
alkane(Å)
Diffraction
pattern for
C29 Ketone
D-spacing
(Å)
Coherence
length of C29
Ketone(Å)
Second order 39.40 1498 Second order 40.74 458
Fourth order 19.64 1287 Third order 27.14 435
Sixth order 13.09 1217 Fifth order 16.43 411
Table 2 Comparison of d-spacings of Bragg peaks observed in microscanning patterns with those
expected for specific mineral inclusions.
(http://rruff.info/) The deviations from the observed spacing were shown in parenthesis
obs (Å)
Tridymite
SiO2 Cristobalite SiO2
Nitrocaclcite Ca(NO3)2· 4(H2O)
Hexahydrite MgSO4· 6(H2O)
Calcite CaCO3
Whewellite
Ca(C2O4) · (H2O)
Weddellite
Ca(C2O4) · 2 (H2O)
Humboldtine Fe(C2O4) · 3(H2O)
6.41 6.18(0.23)
5.84 5.64(0.2) 5.93(0.09)
5.25 5.41(0.16) 5.44(0.19)
4.95 5.13(0.18) 5.09(0.14) 4.8 (0.15)
4.34 4.33(0.01) 4.36(0.02) 4.39(0.05) 4.42(0.08) 4.7(0.36)
3.99 4.09(0.10) 4.04(0.05) 4.03(0.04) 3.88(0.11)
3.81 3.81 3.88(0.07) 3.86(0.05)
3.55 3.58(0.03) 3.45(0.1) 3.65(0.10) 3.60(0.05)
3.13 3.13 3.12(0.01) 3.19(0.06) 3.03(0.10)
2.93 2.97(0.04) 2.84(0.11) 2.93 2.97(0.04) 3.00(0.07)
51
4.3 Discussion
The study of SXMD on longitudinal section has revealed substantial architectural
heterogeneity in the plant cell walls of different tissues of the Arabidopsis stem. Within a
field of view smaller than the size of a conventional x-ray beam we collected 190 diffraction
patterns and carried out detailed analyses on each pattern. That diffraction patterns involve
the scattering scan across the stem with μm size step provided information of structural
variations by identifying the changes of the orientation and size of microfibrils and the
relative abundance of amorphous components and crystalline (fibrillar) cellulose. In the
epidermis, we observed dominant scattering from wax instead of cellulosic material. In some
tissues, scattering from mineral inclusions was superimposed on the cellulose scattering.
The geometrical features of helical wrapping of cellulose microfibrils cause a double-
orientation for cellulose reflection, that coming from opposite sides of a plant cell wall. We
separated the superimposed diffraction patterns to calculate the MFA, which is a measure of
the tilt of microfibrils to longitudinal axis. The overall pattern observed here – large angles in
the pith transitioning gradually to smaller angles in the xylem and phloem –appears similar to
those studied in other plants at substantially lower resolutions (typically mm). The trends in
fibrillar orientation, that could not be demonstrated previously are revealed by scanning
tissues with x-ray beams of substantially smaller dimensions. The thickness of plant cell wall
is similar to microbeam dimension — about 5 μm. When micro beam is incident at the
boundaries between cells, the cell wall is parallel to the x-ray beam and the tilt of fibrils is
either directly towards or away from the beam (as opposed to being orthogonal to the beam).
The encounter of boundaries of cell wall account for observation of punctuate regions of
essentially zero fibrillar angle. Microfibril angle variation may correlate with distribution of
turgor of plant stem. Turgor increases with stem radius inducing a higher stress on cellulose
fibers and potentially providing evolutionary pressure for fibrillar organization best suited to
tolerating that stress.
Equatorial reflections for most scattering patterns exhibited the highest intensity at
spacings from 0.15-0.3 Å-1. In the xylem and phloem, featured reflections from cellulosic
fibrils were dominated in scattering patterns. In the pith the observed scattering was rather
atypical, devoid of the signature reflections of cellulose Iβ including the (1 1 0)/(1 -1 0), (2 0
0) and exhibiting little orientation. This scatter, also contained a single broad reflection, that
52
extended in the spacing range of 0.15 - 0.3 Å-1. The single broad peak is due to a
heterogeneous mixture of amorphous components. A two-component model, consisting of a
crystalline, 36-chain fibril and a surrogate amorphous component fraction proved adequate to
reconstruct the observed scattering from all patterns except from the epidermis (which was
dominated by scattering from wax). The composition of the cell walls is undoubtedly
heterogeneous. However to identify the specific nature of the amorphous component required
a combination of chemical analysis and SXMD data. Here we proved that the cell walls
contain a mixture of amorphous components and fibrillar cellulose and the relative amount of
cellulose in fibrillar form rises from essentially zero in the pith to a maximum in the
cambium region between xylem and phloem.
The pith region appears to contain only primary cell walls constituted by a large
amount of amorphous materials including amorphous cellulose. The materials are largely
disoriented, with low fibrillar content and almost no mineral inclusions. Arabidopsis stems
may have in planta pathways for fibrillar degradation that lead to breakdown of fibrillar
architecture that was in place in the pith earlier in the plant life cycle. Because amorphous
material is the dominant component for cell wall in pith, pith are unlikely to provide the same
level of mechanical support present in cell walls with larger fibrillar content. Therefore, pith,
as a tissue without strict mechanical requirements may also represent an architecture that is
relatively easy to break down and reform.
Since the xylem appears to be a transition tissue from vascular tissues to pith, the
proportion of cellulose in fibrillar form increases from near zero at the inner part of the xylem
to near 33% at the outermost part. However, this variation of fiber content could also be a
reflection of a superposition of both pith and xylem in the relatively thick (100 μm) sections
used in this study. In either case, this appears to track according to the mechanical strain
associated with larger distance from the center of the stem. The xylem appears to have
transport properties for a substantial drifting of mineral and nutrient from root to apex, these
properties cause important functional constraints on the nanoscale architecture of the cell
wall. The cell walls of the phloem, by contrast, appear to have the maximum fibril content,
comparable to the outside of the xylem, and more or less constant across the breadth of the
tissue. This may be in response to the mechanical strain put on the phloem by virtue of its
position near the surface of the stem, or by the transport requirements of the tissue. Cortex
53
has lower fiber content than other vascular tissues but higher robustness. This is due to the
higher concentration of calcium in the cortex than xylem and phloem, as observed by XFM.
Van Buren (Van Buren, 1991) indicated that calcium ion and pectin have essential impact on
the plant robustness, accumulation of calcium ion and pectin in the cortex region will
enhance plant cell wall firmness. Wardrop (Wardrop, 1976) found unlignified cell walls
contain much more calcium than ligninified cell walls. It follows that lower fiber content of
the cortex may nonetheless be associated with higher physical robustness than the phloem.
The estimation of fibril size from x-ray scattering could be described with respect to
several aspects, such as homogeneity and crystallinity of samples (Nishiyama, 2009). As
reported previously (Fernandes, 2011; Thomas, 2013) the predicted scattering from ideal
crystallites of cellulose Iβ does not correspond well to the observed intensity for models with
any number of cellulose chains or any shape of fibril. In particular, the observed intensity
between the peak composed of the superposition of the (1 1 0) and (1 -1 0) reflections and
that of the (2 0 0) reflection indicates a larger lattice index for all crystallite models.
Fernandes et al. (2011) and Thomas et al. (2013) try to account for structural variation of
cellulose crystallite by introducing an asymmetric disorder term into their paracrystalline
model. Although this approach could explain the scattering we observe from the outer
regions of the stem (if we were to hypothesize a large degree of asymmetric disorder) it
cannot account for the scattering from the inner regions of the stem. In those regions the
observed intensity exhibits a broad intensity peak between the expected Bragg peak positions
of cellulose crystallite. This kind of distribution can be interpreted by an additional
amorphous component that may include cellulose, lignin, hemicellulose and pectin.
Consequently, we chose to construct a two-component model consisting of an amorphous
component and fibrillar crystallites. This approach provided a unified explanation of the
diversity of diffraction patterns observed across the Arabidopsis stem.
Scattering from epidermis exhibits a single wide-angle reflection at approximately
4.13 Å spacing corresponding to the spacing between alkane chains of wax. It further exhibits
a series of small-angle peaks that provide information about the nature of the layered
crystalline form of wax on the surface. We observe alkane and ketone forms co-existing in
the epidermis, giving rise to two sets of Bragg peaks with periodicities of 78 Å and 81 Å.
54
Coherent length results show that the size of the lamellar structures differ for the two waxes
observed and suggest that ketone and alkane assemble differently at the stem surface.
Mineral inclusions are common in plants and are assigned a number of functions for
plant maturity and growth. Results of SXMD show that the inclusions were largely confined
to vessel tissue, such as the xylem and phloem. Nitrogen, calcium, sulfur magnesium and
silicon are essential macronutrients for plants and are commonly found in vessel tissues. In
principle, identification of particular minerals involved can be made by correlating the
scattering angles of the reflections observed to those known to be characteristic of all
common minerals. In practice, however, mineral content can be heterogeneous and
observations of scattering peaks may be sparse with but a single peak observed from any
given inclusion. This may preclude an unambiguous identification of mineral inclusions
from the spacings of scattering peaks alone. Addition of data from XFM greatly simplifies
determination of the nature of the inclusions, providing a list of elements involved (or not
involved) in the inclusions. In this case, XFM demonstrated the presence of silicon in the
inclusions. We concluded from that data that some form of silicon oxide was involved, and
the some of the diffraction data was consistent with presence of cristobalite (SiO2), which has
a diffraction peak at 4.04Å. XFM also indicates high levels of rather uniformly distributed
calcium are present, consistent with earlier observations (Demarty M., 1984). It has been
suggested (Van Buren, 1991) that calcium ions frequently complex with cell wall pectin.
Studies of SXMD exhibit that the architecture of cell walls in the pith (and inner
xylem) have a fundamental difference from that in the fibril-rich walls of the transport tissues
and cortex. This structural distinction indicates that the fibrillar content of cell walls can be
adjusted to the requirements of the tissue of which they are a part. This reflects in vivo
control of fibril content within the cell wall. Study of differences in the synthetic apparatus
and deconstruction pathways as expressed in the pith and in the exterior regions of the stem
may provide clues to the elements involved in this control as a first step towards design of
plants that may exhibit far less recalcitrance to the physical and chemical processing required
for utilization of biomass as feedstock for chemicals and fuels.
55
Chapter 5
Study of molecular architecture of plant cell walls in
Arabidopsis harboring mutations in lignin biosynthesis
That SXMD study on the plant cell walls of wild type Arabidopsis reflected the well
known facts that plant cell walls are complex structures composed largely of high molecular
weight poly-saccharides, highly glycosylated proteins, and lignin (Somerville, 2004). The
degree of co-mingling and cross-linking intrinsic to the cell wall has essential impact on the
synthesis and assembly of different polymeric species. But the details of the co-dependencies
among the molecular architecture of cell wall and the synthesis of polymer within cell wall
are completely unknown.
Lignin is a complex non-cellulosic polymer, which is deposited within the plant cell
wall of vascular tissues of stem. Lignin is composed by three basic monomer units — p-
hydroxyphenyl (H unit), guaiacyl (G unit) and syringyl (S unit). G and S units constitute 90 %
of monomer of lignin. The understanding of lignin biosynthetic pathways (Fraser, 2011)
opens the possibility to construct numerous Arabidopsis mutants with altered lignin
compositions and disrupted microstructure. Physical changes of Arabidopsis stem, like
dwarfism, breakdown of vascular tissues, alteration of the plant cell wall microstructure, have
been observed for Arabidopsis mutants with lignin deficiency (Turner and Somerville, 1997).
We have collected SXMD data on wild type [WT] Arabidopsis and 7 variants, each identified
by a specific defect in some aspect of lignin biosynthesis. Variants studied include plants
with high levels of G lignin [high G], S lignin [high S], or H lignin [H lignin], high levels of
aldehyde [aldehyde], and of aldehyde with G lignin [aldehyde in G], high levels of ferulated
lignin [ferulated] and low lignin levels [low lignin]. Here we investigate the organization of
cellulose within thin sections of stem of these variants to determine if the alterations in lignin
biosynthesis lead to changes in the organization, orientation or order of cellulose fibrils in
these plants.
56
5.1 Method for studying on the plant cell of Arabidopsis by SXMD
5.1.1 Separating oriented diffraction from patterns of SXMD and new fiber content
calculation
Both WT and Arabidopsis mutants were cultured for 6 weeks, harvested and the
bottom of stems of WT and lignin biosynthetic mutants of Arabidopsis were cut into 100 µm
thick longitudinal sections. This work was carried out by the Chapple laboratory at Purdue
University. One sample set was immersed in water at room temperature for 3 days and
subsequently dehydrated in air at room temperature for 24 hours before data collection. A
second set was immersed in water at room temperature for 30 days prior to being dehydrated.
Background scattering was estimated from the first and last diffraction patterns of
each row of the montage. These patterns were collected at positions outside of the sample
and constitute diffraction from air and mica window. We subtracted these backgrounds from
all diffraction patterns prior to all other data analysis.
Instead of simply modeling the fibrillar and amorphous component of plant cell wall
structure with cellulose microfibril and amorphous cellulose (as carried out for WT and
described above), we separated the diffraction pattern into reflections from oriented and
disoriented material. The linear combination of the intensities of oriented material and
disoriented material, as shown in Figure 27, was used to determine the fiber content:
IFibrillar Content = [ Ioriented material]/[ Ioriented material+ Idisoriented material] 3
IFibrillar Content is a measure of oriented fibrillar content, Ioriented material is the normalized
integral intensity of elementary fibrils, Idisoriented material is the normalized integral intensity of
amorphous component.
57
Figure 25 Each microdiffraction pattern shown in grid was separated to oriented and disoriented pattern. a.
depolarized diffraction pattern. b. Azimuthal distribution at radius corresponding to ring insert in a.
c. diffraction pattern of disoriented material. d. diffraction pattern of oriented material.
Figure 26 Integral intensity of oriented component (blue) and intensity of disoriented component
(green)
a
d c
b
58
Figure 27 Calculation of oriented fiber content. Oriented and circularly symmetric intensities are
separated as in Figure 25. The Intensity of oriented material (left) and unoriented material (right) are then
integrated over a range of spacings corresponding to the positions of the (1 -1 0) and (2 0 0) reflections
from cellulose (indicated in orange in the figures and spanning 0.1 < 1/d < 0.3 A-1). The proportion of
oriented fibrillar material is then calculated as indicated in the text.
59
5.1.2 Determination of Microfibril Angle
The thin section of samples with thickness about 100 μm contains multiple cell walls
from different cell and tissues. Therefore, scattering of real tissue may overlap reflection
from mixed material. Lichtenegger et al., 1999 and Liu et al., 2013 reported MFA could be
used to determine the fiber orientation. Therefore, we developed an algorithm to determine
and separate the fiber orientation by azimuthal intensity distributions of diffraction patterns.
Microfibril angles were estimated from the angular variation of intensities in a polar
coordinate system (Figure 28). Azimuthal positions of peaks were determined by fitting of
Gaussians to the intensity as a function of at a radius corresponding to the (2 0 0) reflection
of the cellulose Istructure
5.1.3 Calculation of coherence length of crystallite
Axial periodicity of cellulose fibril corresponds to the (0 0 4) reflections. The breadth
of the (0 0 4) reflections calculated from the Scherrer equation enable us to estimate the axial
coherence length. Figure 29 shows Gaussian fitting of (0 0 4) reflection of cellulose fibrils.
Then coherent length could be calculated as below:
. 4
The green curve shows a Gaussian function fitting the reflection. The θ can be
determined by the peak position of the Gaussian and β is the half width of Gaussian function.
K is constant 0.9. is wavelength of incident x-ray.
5.1.4 Analysis of small angle reflection
The scanning microdiffraction technology on the lignin mutated sample reveals an
interesting observation that intensity distribution changed in small angle region for both WT
and lignin-mutated samples. A broad band is observed in xylem region for specific mutated
sample. Therefore, we try to interpret the intensity distribution by the solid cylinder model
and interference calculation between two cylinders to estimate the average size and distance
between fibers.
According to Debye calculation, the cylindrical average intensity could be calculated
as:
60
∫
5
is electron density of cylinder . R is the radius of cylinder. For convenience, we
let . Then
∫
We assume the cellulose fibril size of WT and mutated sample is constant, the broad
band only comes from interference. The interference can be calculated as:
∫ ( )
For quantitative analysis, a constant, 12 Å, was assigned to r, d is distance between
two solid cylinder. Figure 30 shows the intensity distribution of solid cylinder (blue curve)
and interference of two cylinders (red curve) at small angle region.
Since real tissues were constituted by complexes, a combination solid cylinder, as
Figure 31 shows, calculated from wild type and interference of solid cylinder will improve
the fitting of the intensity of real data. Disorganization of amorphous material may lead to a
reduction in the observed interference.
61
Figure 28 determining the MFA by identify split scattering from cellulose crystallite. (a) Diffraction
pattern exhibits separation of azimuthal positions of the (2 0 0 ) (microfibril angle). Inset shows the
small angle area of the diffraction pattern indicating two orientations of the microfibrils (MF).
Diffraction peaks of microfibrils split at small angle and (2 0 0) reflections (shown with arrow). (b)
Azimuthal average intensity plot of ring indicated in (a). Background was shown as red line through
1-360⁰. At 1- 180⁰ and 181-360⁰, two Gaussians functions (green) are fit to the background-
subtracted intensity to the azimuthal positions of the (2 0 0). Difference between peak positions of the
two Gaussian determine microfibril angle difference.
Figure 29 Calculation of axial coherent length for (0 0 4) reflection. A trace of the oriented intensity
including the (0 0 4) reflection of cellulose fibrils is shown in left. As shown in the enlargement of the
(0 0 4) on the right, the background subtracted reflection and (blue curve) and a Gaussian function
(green curve) fit to intensity distribution. Red line corresponds the maximum position of Gaussian
curve.
a b
0 0.1 0.2 0.3 0.4 0.50
0.5
1
1.5
2
2.5
3x 10
4
1/d
Inte
nsity
(0 0 4)
62
Figure 30 Interference calculation and model and interference between two cylinders. The left plot
shows scattering from two independent cylinders (blue) and the interference that would arise if they
were positioned a distance 'd' apart, causing interference at small angles. The right image is the
model two cylinders apart from each other with d.
Figure 31 Fitting of small intensity distribution of mutated sample. Interference must be added to the
calculation of scattering from a solid cylinder to obtain an adequate fit to observed intensity.
d
63
5.2 Results from lignin mutated Arabidopsis
5.2.1 Studies of molecular architecture of 16 samples by SXMD
5.2.1.1 Wild Type sample (WT)
The fiber content of wild type approached as high as ~ 25% in the xylem region (Figure 32).
The microfibril angle slowly increased towards the center of the stem, a trend opposite that of
fiber content. The (0 0 4) reflections are strong within the oriented part of the patterns and
reflects a relatively constant coherence length of about ~ 210Å at xylem. After storage in
water for 30 days, fiber content, represented by scattering from predominatly oriented
material, is significantly reduced. Poorly oriented patterns frequently preclude measurement
of microfibril angle and may have sufficiently weak (0 0 4) reflections to make measurement
of coherence length impractical. Extended storage in water results in lowered fiber content
and much wider variation of microfibril angle.
Figure 32 wild type sample dried 3 (left) and 30 (right) days after harvest. Optical micrographs
indicate the location of the microdiffraction scans. Fiber content, microfibril angle and axial
coherence length are plotted as a function of position across the stem.
64
5.2.1.2 Aldehyde sample (cad-c cad-d)
The fiber content of aldehyde mutation in xylem could approach as high as ~ 25% (Figure
33). microfibril angle is observed across the stem with the opposite trend to the fiber content.
The axial coherence length is, in general, smaller that wild type, about ~ 196Å at xylem.
Storage in water results in a significant reduction of fiber content, to less than 10%. For these
samples, even the best oriented patterns are usually too weak to make possible measurement
of microfibril angle and coherence length. The overall impression is that cellulose fibrils are
highly disordered and disoriented in all tissues within the stem.
Figure 33 Aldehyde sample dried 3 (left) and 30 (right) days after harvest. Optical micrographs
indicate the location of the microdiffraction scans. Fiber content, microfibril angle and axial
coherence length are plotted as a function of position across the stem.
65
5.2.1.3 Aldehyde in G sample(cad-c cad-d fah1)
The aldehyde in G exhibits properties similar to those of the aldehyde sample. The highest
fiber content of aldehyde in G was ~ 25% (Figure 34). Microfibril angle is observed to vary
modestly across the stem, generally exhibiting a trend opposite that of fiber content. The
coherence length is, like the aldehyde samples, smaller than wild type at about ~ 196Å in the
xylem. After 30 days storage in water a significant reduction of fiber content is observed,
down to less than 10%. The oriented patterns are again too weak to make possible
measurement of microfibril angle or coherence length. The cellulose fibrils are highly
disordered for the 30-day old sample. The thin sections of stem appeared collapsed in the
optical micrographs (Figure 34) reflecting the dramatic impact of this chaotic arrangement of
components on overall morphology.
Figure 34 Aldehyde in G sample dried 3 (left) and 30 (right) days after harvest.
Optical micrographs indicate the location of the x-ray microdiffraction scans. Fiber content, microfibril
angle and axial coherence length are plotted as a function of position across the stem.
66
5.2.1.4 High G sample (fah1)
Samples with high G lignin exhibit oriented fibril content comparable to wild type at the
xylem. As for wild type samples, microfibril angle can be observed across the stem and
exhibits the opposite trend of oriented fiber content. The (0 0 4) reflections are strong,
allowing measurement of coherence length within oriented patterns, which is equal to ~ 220Å
within the xylem. After 30 days storage, the fiber content decreased modestly, suggesting that
G lignin may be efficient in protecting cellulose from degradative processes that led to severe
disorder and disorientation in other samples.
Figure 35 High G sample dried 3 (left) and 30 (right) days after harvest. Optical micrographs
indicate the location of the x-ray microdiffraction scans. Fiber content, microfibril angle and axial
coherence length are plotted as a function of position across the stem.
67
5.2.1.5 High S sample(C4H:F5H:fah1)
The highest oriented fiber content observed in high S samples was only ~ 22% (Figure 36).
Microfibril angle is observed clearly across the stem, again exhibiting the opposite trend of
oriented fiber content. The (0 0 4) reflections are strong and well defined and reflect a
coherence length of ~ 210Å at xylem, about the same as wild type. After 30 days storage, the
fiber content decreases modestly, indicating that S-lignin, like G-lignin, may act to provide
some resilience of cellulose fibrils to degradative processes.
Figure 36 High S sample dried 3 (left) and 30 (right) days after harvest. Optical micrographs indicate
the location of the x-ray microdiffraction scans. Fiber content, microfibril angle and axial coherence
length are plotted as a function of position across the stem.
68
5.2.1.6 High H lignin sample(ref8ref4rfr1)
The fiber content of High H at xylem approaches ~ 25% (Figure 37) similar to WT. There is
a distribution of microfibril angle across the stem with the opposite trend exhibited by
oriented fiber content. The (0 0 4) reflection gives rise to a mean coherence length of about ~
200Å, somewhat smaller than WT. After 30 days, structural degradation leads to significant
reduction of oriented fiber content. The weaker oriented patterns in general preclude
measurement of microfibril angle or coherent length. Qualitatively, the mutation gives rise to
degradation of cellulose organization similar to that observed in the aldehyde mutants.
Figure 37 High H sample dried 3 (left) and 30 (right) days after harvest. Optical micrographs
indicate the location of the x-ray microdiffraction scans. Fiber content, microfibril angle and axial
coherence length are plotted as a function of position across the stem.
69
5.2.1.7 Ferulic acid incorporated lignin sample (ccr1)
The fiber content of ccr1 at xylem never exceeded ~ 19% (Figure 38), significantly lower
than wild type. Microfibril angle could be observed in almost all positions across the stem.
The axial coherence length was ~ 210Å at xylem, about the same as wild type. After 30 days
storage, oriented fiber content is reduced. The resulting weaker oriented patterns make
measurement of microfibril angle difficult in many cases and decrease the mean coherence
length, indications of significant decrease in order and orientation of the cellulose in these
samples.
Figure 38 Ferulate incorporated lignin sample dried 3 (left) and 30 (right) days after harvest. Optical
micrographs indicate the location of the x-ray microdiffraction scans. Fiber content, microfibril
angle and axial coherence length are plotted as a function of position across the stem.
70
5.2.1.8 Low lignin sample (ref3-2)
The ref3-2 has much smaller fiber content compared to other mutants. Almost nowhere is
orientation adequate to make possible measurement of microfibril angle. The coherence
length is also significantly less than other samples, about ~ 160Å at xylem. After 30 days
storage, there is a further reduction of fiber content, but in some samples the orientation was
good enough to make possible measurement of microfibril angle. But most patterns were
sufficiently weak as to preclude measurement of microfibril angle or coherence length.
Figure 39 low lignin sample dried 3 (left) and 30 (right) days after harvest. Optical micrographs
indicate the location of the x-ray microdiffraction scans. Fiber content, microfibril angle and axial
coherence length are plotted as a function of position across the stem.
71
5.2.2 Oriented Fiber Content
Tissue specifications give rise to a wide diversity in the degree of orientation with
many patterns. As discussed in 5.1.1, scattering was separated into oriented and disoriented
components as described in Methods. For instance, Figure 41 shows that significant oriented
pattern was separated from high S sample by isolating the anisotropic part of the pattern
(green on the right) from the isotropic part (red and white on the right). The intensity as a
function of scattering angle could then be calculated for both of these fractions resulting in
the intensity distributions in Figure 42. The significance of the (1 1 0)/(1 -1 0), (2 0 0) and
(0 0 4) reflections of cellulose indicated that scattering from cellulose microfibril was
dominate the oriented part of the pattern. Scattering in the disoriented part of the pattern
exhibits an intensity distribution rather different from that observed in the oriented part of the
pattern, with broader peaks and substantial increase in features not normally associated with
scattering from cellulose. Amorphous material exhibited that the un-oriented feature is a
combination of scattering from poorly ordered cellulose and a preponderance of non-
cellulosic materials.
We have observed that mutations on lignin biosynthesis lead to variations of oriented
fraction within plant cell wall for different samples (as well as among different tissues within
individual stems). By introducing Fiber Content, a measure of the proportion of total
material made up of oriented cellulose fibrils, we provide the quantitative studies of fibrillar
material within cell wall. Figure 41 shows the separated oriented and un-oriented pattern
from single diffraction, both of whom exhibit featured reflections from cellulosic material.
But the primary scattering from oriented material can be attributed to cellulose fibrils.
Consequently, the ratio of oriented intensity to total intensity, including scattering from
oriented and disoriented components, provides a relative measure of the extent of structural
organization of cellulose fibrils within the scattering volume. This measurement enables the
estimation of a relative fraction of the oriented fiber content for plant cell wall. Several
similar measures appear in the literature. Although other measurements, such as crystallinity,
are based on different empirical methods leading to somewhat different values and
corresponding to somewhat different properties, most exhibit similar trends for similar tissues.
As seen in Figure 40b, the cell wall within vascular regions contain largest oriented fiber
72
content, particularly the xylem, and essentially zero in the pith and epidermis where no
scattering attributable to oriented, fibrillar cellulose is observed.
To identify the co-dependency among the oriented fiber content and mutations on
lignin biosynthesis, for the different variants, we compare the proportion of oriented fibrils of
the parenchyma, the region between xylem and phloem which exhibits the highest content of
oriented fibrils (Liu J., 2013). Figure 43 is a bar graph comparing the maximum values of
oriented fiber content observed in each of the 16 samples. The figure shows that for freshly
harvested plants (blue bars), significant reductions of the highest fiber content only happen to
ccr and ref3-2. ref3-2 has very low lignin content suggesting that assembly of cellulose
fibrils into oriented structures parallel to the stem relies on the presence of normal levels of
lignin. ccr has high levels of ferulated lignin and is far more digestible than WT. The
relatively high digestibility of ccr would suggest that its cellulose may also be susceptible to
degradation on storage in water, but our results indicate this is not the case.
We observed that after 30 days in water aldehyde, aldehyde in G and high H display
significant disordering of cellulose, but they show similar levels of oriented fiber content to
WT in fresh samples. These results were replicated on two sets of samples grown, harvested
and analyzed independently at different times. That the samples maintain physical integrity
after 30 days in water, indicates no evidence of microbial contamination. These observations
suggest that lignin organization may be important for protecting cellulose from hydrolytic
degradation that may occur in aqueous environments over time. Other variants showed much
less change on storage. Interestingly, ccr, which is known to be more digestible, does not
show a decrease in oriented fiber content after storage. No variant exhibited a significant
increase in oriented fibril content over wild type.
5.2.3 Crystalline order in the cellulose fibrils
The dimensions of the lattice of cellulose crystallite determine the positions of the (1
1 0)/(1 -1 0) and (2 0 0) reflections corresponding to the directions perpendicular to the fibril
axis. The widths of these reflections are affected by three aspects: crystallite breadth
(number of cellulose chains in the fibril), degree of order in the packing of cellulose chains,
and heterogeneity of crystal lattice constants. The difficulty in separating out these three
variables is a key reason for the continued debate over the number of cellulose molecules
73
making up an elementary fibril. Nevertheless, the homogeneity of crystalline order within
the cellulose fibrils averaged over the scattering volume could be measured by the sharpness
of these peaks. Figure 44 shows traces of the (1 1 0)/(1 -1 0) and (2 0 0) reflections in
scattering from the different lignin mutants. The pattern corresponding to the highest fiber
content from each sample is compared here for the 8 samples analyzed 3 days after
harvesting. The (2 0 0) reflection is broadest for ccr and ref3-2, the samples with the lowest
lignin content. Whether this is due to intrinsic disorder within individual fibrils or to a
structurally heterogeneous population of fibrils requires further experimental data to
determine.
5.2.4 Microfibril Angle
Microfibril angle (MFA) was calculated by transforming all intensity onto a polar
coordinate system and identifying angles of maximum intensity as detailed in Methods.
Because of the symmetry of scattering from cellulose, two independent measures of
microfibril angle are obtained for each pattern and these were averaged. In many cases, the
diffraction was dominated by circularly symmetric scattering and exhibits essentially no
oriented diffraction, making microfibril angle undefined. Figure 40b shows the distribution
of microfibril angle for the WT sample in Figure 40a. The helical wrapping of cellulose
fibrils around plant cells (Emons and Mulder, 1998), as diagrammed in Figure 45, give rise
to two sets of reflections from cellulose fibrils, the half of angle between them is the MFA.
The microfibril angle in the cortex and initial part of vascular tissue is relatively constant at
about 15⁰, but increasing to about 30o immediately adjacent to the pith. As seen in Figure 46,
the helical winding of cellulose about cells in wild type gradually increases going from cortex
to pith leading to variation of microfibril angle. Orientation was seldom observed in the pith,
precluding an estimate microfibril angle. For some of the mutants with sever reduction in
lignin content, orientation is so poor as to preclude any measurement of microfibril angle. For
instance, the dual reflections required for measurement of microfibril angle could only be
observed in xylem region of the ref3-2 mutants. Where measurable, the mean microfibril
angle within the xylem of ref3-2, is about 25⁰, greater than wild type and high S.
The microfibril angle of lignin mutants after 30 days in water was observed to vary
from 15ᵒ to 30ᵒ as well. Except High S mutants exhibiting a microfibril angle distribution
74
similar to fresh samples, the measurements of microfibril angle for other mutations of lignin
are precluded.
The microfibril angle exhibits an inverse correlation to oriented fiber content for
individual samples, but that correlation does not appear to hold between samples. In a
comparison of wild type and variants there is no well-defined correlation between microfibril
angle and fiber content. Measurement of microfibril angle is in some cases precluded for
mutants with severe reduction of lignin content. For instance, only small portions of ref3-2
mutant exhibits diffraction patterns with the split reflections required to measure microfibril
angle. The mean microfibril angles in xylem of ref3-2 is about 25⁰, which is greater than that
for WT and high S and suggests that the transverse assembly of cellulose is altered in ref3-2.
Patten et al. (2010) observed that within the interfascicular or vascular tissues of
Arabidopsis stem, the deposition of lignin decreased approaching the pith. Decreasing lignin
concentration correlates with decreased fiber content and increased microfibril angle, but to
what extent there are causal relationships among these three variables is unclear.
5.2.5 Axial Coherence Length
The axial coherence length is a key measurement for the crystallinity of cellulose.
The cellulose fibril within cell wall of real tissues is complex and imperfect crystalline,
because of its extreme length, very large surface area, curvature and tendency to twist, the
crystallinity is imperfect. In the axial direction, the coherence length as estimated by the
breadth of the (0 0 4) reflection in fiber diffraction patterns provides a useful measure of the
degree of imperfection of crystalline. Because of local stretching, twisting and curvature, the
periodicity of a cellulose fibril varies along its length. This leads to a phase difference in the
molecular repeating structure that increases progressively along the length of the fibril. The
coherence length is the distance along the fibril beyond which there is no ordered phase
relationship.
As discussed above, cellulose fibril is imperfectly crystalline within cell wall of
tissues due to the physical constraints placed on them by other cell wall constituents,
interactions with cross-linking polymers and greater degree of curvature of the fibrils. In a
comparison to isolated, purified cellulose crystallites, less well ordered cellulose fibrils
within real tissues have shorter coherence lengths. Therefore the coherence length may
75
provide insight into the interactions of cellulose fibrils with other cell wall constituents, and a
change in coherence length may reflect a disruption in those interactions, a perturbation in
the ordered process by which cellulose fibrils are assembled into the cell wall or enhanced
distortion of the cellulose fibrils caused by increased physical constraints due to interactions
with other cell wall constituents. Figure 32Figure 39 in Section 5.2.1 details variation of
coherence length among the tissues of the stem from different samples. The histograms of
the distribution axial coherence lengths for each of the lignin mutants were shown in Figure
47. Figure 48 provides comparison of the maximum (of histogram of) coherence lengths
observed for each of the samples. We observed that in fresh samples only ref3-2 appears to
have a significantly lower coherence length, but high G and high S appear to have slightly
larger average coherence lengths than wild type. This is perhaps suggestive of lignin
deficiency leading to fewer cross-linking constraints on cellulose fibril structure. Coherence
length decreased for all samples exposed to water for 30 days except for ref3-2. Coherence
length of ref3-2 is very low in fresh sample but appears to increase on storage in water —
perhaps through the relaxation of cross-linking constraints on it by other polymeric structures.
Interestingly, there is no direct correlation between the coherence length and fiber content for
the lignin mutants studied here.
In practice, the (0 0 4) reflections from the un-oriented fraction of cellulose are weak
and broad, making accurate measurement of coherence length difficult. Although we could
also in principle calculate the coherence length using curves similar to those in Figure 42, we
prefer to give a qualitative analysis. From a qualitative aspect, the curves suggest that the
coherence length is significantly less than for the oriented fraction of cellulose. In all
likelihood, this reflects the greater curvature expected in the unoriented fraction.
5.2.6 Packing of cellulose fibrils
Structural information about features ranging from 25 Å to 100 Å in size could be
extracted from the small angle region of the diffraction patterns. The intensity distribution in
this region corresponds to the scattering from individual cellulose fibrils and their relative
positions. When the fibrils are arranged in an organized fashion, regularly spaced side-to-
side, the intensity is modulated by an 'interference function' that provides information on the
spacing of fibrils in the material (Inouye, 2014; Thomas, 2013; Kennedy, 2007). Figure 49
contains an enlargement of the small angle region of exposure 61 and equatorial traces for
76
exposures 15, 30, 45 and 61, the intensity distribution in the small angle region varied from
the exterior to the interior of stem of a WT sample. A specific modulation of the small angle
scattering intensity with peak at 1/d ~ 0.017 Å-1 was observed from the exposure taken from
a diffraction pattern of a region immediately adjacent to the pith. This distinct modulation
suggests that the fibrils are spaced with a nearest neighbor distance of approximately 60 Å.
The observation of this interference near the pith is unexpected because this is the region of
the stem with the lowest (observable) oriented fibril content. The low fiber content indicates
that the region must be highly heterogeneous, with the small fraction of oriented fibrils well-
ordered in spatially confined regions.
77
Figure 40 A longitudinal section of Arabidopsis stem was studied by SXMD. a. optical image contains
a blue rectangle corresponding to the region scanned within stem. The grid includes three rows and
160 columns, each grid point being 5 μm square, making the grid 800x15 μm. A complete
microdiffraction pattern was taken at each grid point. b. Fiber content, microfibril angle and axial
coherence length as calculated from the diffraction patterns collected by scanning the rectangle in (a)
and plotted as a function of position across the stem. Data from all three rows are plotted. c.
Enlargement of microdiffraction patterns from selected regions of the scan. Selected regions
correspond roughly to the cortex, xylem and close to pith.
Figure 41 Diffraction patterns were separated into circularly symmetric (red and white in diagram to
right) and oriented fractions (green in right) using intensity at the scattering angle with greatest
intensity (the position of the (2 0 0) reflection marked as a circle in the scattering pattern to the left).
c
… … … … Positio Position
8 13 18 23 56 62
Position 1 Position 160
a b
78
Figure 42 Example from high S shows separation of circularly symmetric and oriented intensities. A
pattern from the High S sample demonstrates how oriented (top, middle) and unoriented (bottom,
middle) fractions were separated. The intensities were then plotted as a function of spacing 1/d
(roughly proportional to scattering angle) for the oriented (top right) and unoriented (bottom right)
fractions. The oriented portion includes well defined (1 1 0)/(1 -1 0), (2 0 0) and (0 0 4) reflections
characteristic of scattering from cellulose 1. The unoriented fraction had an intensity distribution
rather different from that expected for cellulose and probably represents a sum of scattering from all
constituents of the tissue minus that from the oriented cellulose fibrils.
High S (pattern
18)
0 0.1 0.2 0.3 0.40
1
2
3
4
5x 10
4
1/D
Inte
nsity
0 0.1 0.2 0.3 0.40
1
2
3
4
5x 10
4
1/D
Inte
nsity
79
Figure 43 Comparison of the highest oriented fiber content observed in each sample. The samples
dried 3 days after harvest are shown in blue bars. The red bars corresponds to samples dehydrated
after storage in water for 30 days.
Figure 44 Eight Arabidopsis variants show intensities of the strongest equatorial reflections of cellulose
I. In each case, the trace of intensity corresponds to that position in the sample that exhibited the
highest proportion of oriented cellulose fiber content. Intensities were normalized over the range 0.1-
0.3 Å-1
for comparison.
0
5
10
15
20
25
30
hig
he
st F
iber
co
nte
nt(
%) 3 day
30 day
80
Figure 45 Depiction of cellulose fibrils helically wrapped around a plant cell (left). Projection of
cellulose fibrils from front and back produce a double-orientation (center) that expresses itself in diffraction
patterns as a split or double diffraction pattern (right).
Figure 46 Increase of microfibril angle indicates alignment of cellulose fibrils tend to transverse from
the periphery of the stem to the center. The microfibril angle increases from the periphery of the
stem to the center, reflecting a change in the way cellulose fibrils coil around cells. Largest
microfibril angle is observed in regions of lowest oriented fibril content as can be seen in the plots in
the upper left. The grid within the optical image shows the region scanned and identifies positions
with fibril orientation corresponding approximately to the diagrams at the bottom of the figure.
2θ
2θ
Position 1 Position 160
81
Figure 47 Samples dried fresh and after 30 days in water display different histogram of axial
coherence length. (a) Histogram of axial coherence length for samples dried fresh and after 30 days
in water. The axial coherence length differs across the stem, tending to be largest in the vascular
tissues. (b) Gaussian curves fit to each of the histograms in (a) suppress the effect of random
variations and facilitates visual comparisons among the samples.
82
Figure 48 Comparison of maximum coherence lengths for each of the 16 samples. This is a
compilation of the peak positions of the smoothed curves shown in Figure 47b.
0
50
100
150
200
250
300
Axi
al C
oh
eren
ce L
engt
h(Å
-1)
3 day
30 day
83
Figure 49 Modulation of small angle scattering by interference due to packing of cellulose fibrils was
unexpectedly observed to be strongest in the region near the pith. (a) optical image of stem, the grid
containing 3x160 points corresponds the position of the microdiffraction scan. (b) the distribution of
fiber content and microfibril angle across the stem. (c) Diffraction pattern collected at position 61
close to the pit exhibits a strong modulation of small angle scattering (inset) attributed to partially
ordered packing of cellulose fibrils. (d) Comparison of intensity at small angle region for four
positions from epidermis to pith, and exhibiting the modulation at ~ 0.017 Å-1
due to interference.
15 30 45 61
a
c d
b
84
5.3 Discussion
The assembly of cellulose and lignin having a complex interdependency appears to be
essential to the intricate nanoscale architecture required for cell wall integrity, strength and
resiliency. Nevertheless, the possibility of indirect effects and individual variations obscure a
demonstration of this interdependency. By querying all stem tissues and multiple individual
plants with ~16,000 diffraction patterns we sought to minimize the potential impact of
individual variation and tissue-specific effects. The studies of Arabidopsis variants by
SXMD exhibits the systematic and reproducible differences in cellulose structure with altered
lignin metabolism and provide compelling evidence of the dependence of cellulose
architecture on lignin composition.
The vegetative apparatus of Arabidopsis is a ground rosette that develops a lignified
flowering stem (Dharmawardhana, 1992). Besides the xylem vessels, which are lignified, the
interfascicular parenchyma of the flowering stem differentiates into highly lignified fibers
with growth. Under certain growth conditions, lignins in Arabidopsis can represent up to 18%
of the dry weight of the extractive-free mature stem and are typical of angiosperms since they
contain both S and G units (Dharmawardhana, 1992). Phenotype could be significantly
changed by altering the lignin composition of the stem (Meyer, 1996; Meyer, 1998; Turner
and Somerville, 1997; Goujon, 2003; Schilmiller, 2009; Bonawitz, 2014; Ruegger, 1999;
Chapple, 1992). These phenotypic alterations are obvious in the whole plant and at the level
of electron microscopy which reveals in some variants wholesale disruption of the cell wall
structure and collapse of the vascular tissues. Finer-scale disruptions have not been fully
unexplored. How do the molecular changes that enforce these phenotypic changes associate
with normal lignin synthesis? By using scanning x-ray microdiffraction of lignin variants we
sought to reveal the changes in cellulose architecture triggered by alterations in lignin
composition and assembly.
5.3.1 Mutations in the lignin biosynthetic pathway affect deposition and degradation of
cellulose
The cell wall within the vascular tissue contains higher fiber content in the
Arabidopsis stem, perhaps the parenchyma, the region between xylem and phloem, exhibits
the largest fiber content. Variants with high G, high H and aldehyde exhibit maximum fiber
85
content comparable to that of WT. Modest reduction of maximum fiber content has been
observed in plants with abnormally high S lignin. However, the ccr and ref3-2 mutants
display a significant decrease of maximum fiber content (Figure 43). Turner et al. (Turner
and Somerville, 1997) and Goujon et al. (Goujon, 2003), reported a reduction of cellulose
content in the stem was induced by the ccr mutants, consistent with our results. Similarly, the
ref3-2 variant exhibits lower lignin and cellulose content in the secondary cell wall
(Schilmiller, 2009), as reflected here. These observations indicate that lignin is essential to
the deposition of cellulose fibril within the plant cell wall. The subnormal deposition of
cellulose fibrils in ccr and ref3-2 may be a factor in the observed dwarfism of these plants.
Aldehyde, aldehyde in G and high H mutants have lower lignin content but exhibit
the highest observed fiber content of these mutants comparable to wild type in fresh sample.
Nevertheless, after 30 days in water, a significantly lowered content of oriented cellulose
fibrils was observed for thin sections of these plants, indicating that the lowered lignin
content has increased their susceptibility to degradation processes. In contrast, plants with
high G or high S show no significant alteration in fibril content after 30 days in water. This
indicates that lignin expressing high levels of G or S subunits may prevent cellulose fibrils
from the relevant degradative processes.
5.3.2 Mutated lignin changes the orientation of cellulose fibrils
Micro x-ray technology has been used to study the variation of microfibril angle
( Lichtenegger et al 1999) . The microfibril angle is thought to reflect the helical arrangement
of cellulose microfibrils around the plant cell. Emons and Mulder (1998) and Lichtenegger et
al., (Lichtenegger, 1999) described how the helical geometry of the architecture of the plant
cell determines the microfibril angle. The micorfibril angle is measured as half the angle
between the split reflections of cellulose (see Figure 45&Figure 46). The microfibril angle
represents the tilt of cellulose fibril relative to long axis of the stem. The average tilt of
cellulose fibrils in the vascular tissues is never more than ~ 30⁰ in these variants. For
comparison, the maximum tilt of cellulose microfibrils in roots of Arabidopsis was reported
to be ~45⁰ (Burk, 2002; Anderson, 2010).
Turner and Somerville (1997) reported that deposition of lignin within the plant cell
wall for Arabidopsis stem decreases as one approaches the pith. Liu et al. (2013) noted
86
correlation among the reduction of lignin content in cell wall and a decrease in fiber content
which inversely correlates with changes of microfibril angle in Arabidopsis stem. The
correlation in wild type suggests that lignin may be essential for anchoring cellulose
microfibrils within the polysaccharide matrix at appropriate tilts relative to the stem axis.
The structural resilience of cell walls may be reduced by the disruption of lignin synthesis.
WT plants and those over-expressing S-lignin contain well-ordered cellulose
microfibrils exhibiting obvious oriented scattering for measurement of fibril reflections. By
contrast, in the mutants with severe reduction of lignin content, the degree of orientation was
inadequate for estimation of fibrillar angle. Interestingly, where measurable, the mean value
of microfibril angle in these mutants is considerably larger than others.
5.3.3 The regularity of cellulose fibrils varied with lignin mutants
Coherence length gives an important clue for understanding the integrity of cellulose
fibrils within plant cell walls. Figure 47 and Figure 48 exhibit the range of variation of
coherence lengths within these samples. The most obvious reduction in coherence length is
observed in ref3-2, a plant with severe lignin deficiency (Schilmiller et al, 2009). It is
possible that the curvature of the fibrils may give rise to the lowered coherence length. These
observations support the notion that lignin plays a key role in maintaining proper assembly of
cellulose fibrils within cell walls. Cellulose fibrils may not keep their integrity without cross-
linking by lignin. Disordered cellulose fibrils may be an underlying contributor to the
extreme dwarfism displayed by this plant. Another mutant with lignin deficiency, ccr,
exhibits relatively high coherent length compared to ref3-2, about 210Å, which is comparable
to WT. Lower oriented fiber content was observed in Both ref3-2 and ccr. The implication is
that lowered lignin content disrupts the natural order of cellulose fibrils within the cell wall.
The coherent lengths of high S and G plants appear to be similar to wild type. The
aldehyde mutations show lower coherent length to wild type, but greater than ref3-2. This
indicates that cellulose fibrils are disordered, and probably highly curved in aldehyde and
high H plants. This result may alter physical properties of the cell walls, lowering stiffness of
cell wall and leading to limp floral stem.
87
5.3.4 Packing of cellulose fibrils
The well-organized side-to-side packing of cellulose fibrils in WT samples may give
rise to interference at small angles, as shown Figure 49. The interference was observed in a
region adjacent to the pith. The scattering volume corresponding to the interference indicates
that cellulose fibrils packed with a center to center distance of ~ 60 Å. Langan et al. (Langan,
2014) reported that the result of molecular dynamic simulations suggest formation of larger
fibrils of diameter of 60-70 Å with the reduction of lignin. The interference reported here is
not consistent with the coalescence of fibrils and presumably reflects a distinct phenomenon.
It is interesting to notice that interference reflections are not observed in scattering from
lignin mutants overexpressing the S and G subunits.
Small angle reflections at a spacing of ~39 Å are observed in high H and ccr1
samples. This reflection is close to that expected for paraffin. It is possible that paraffin
deposition as well as cellulose may be affected by lignin mutations, but we have not explored
this possibility in detail. Complete interpretation of this small scattering requires more
experiment data and chemical analysis in future.
5.3.5 Molecular architecture in the stem of lignin mutants
In section 5.2.1, data on the oriented fiber content, microfibril angle and axial
coherence lengths are presented for 16 samples. The combinations of those charts enable us
to have a comprehensive overview of the tissue-specific variations in these properties among
the samples. Individual variations are inevitable in comparisons of these types, but replicate
experiments indicated that the overall trends reported here are representative. Oriented fiber
content appears to be comparable for wild type and variants. Disruption of lignin biosynthesis
resulted in either a decrease in the proportion of oriented cellulose fibrils, or no apparent
change. Plants with high G or high S behaved quite similar to wild type, with high S having,
perhaps, somewhat lower oriented cellulose fiber content. Interestingly, average axial
coherence of high G or high S are comparable or slightly higher than that of wild type.
Aldehyde-containing lignins and high H plants also had similar fiber content to WT, but
displayed significant sensitivity to degradation during storage in water. A broader distribution
of axial coherence lengths for aldehyde-containing lignins and high H plants indicate
inhomogeneity in the spatial constraints on cellulose organization. ccr and ref-2 had lower
88
oriented fiber content than wild type. ccr and ref3-2 in water for 30 days show little change in
oriented fiber content. However, their axial coherence lengths increased after storage,
especially that of ref3-2 which increased significantly. These observations indicate that the
axial periodicity of cellulose fibrils in these samples is maintained with more precision than
WT. One explanation for this observation would be that cross-links among cellulosic
structures in WT plants may disrupt the precise repetitive structure of cellulose; whereas in
ref3-2 the cross-links may be disrupted, allowing the cellulose to relax and take on a more
periodic structure.
The impression is that the disruption of lignin biosynthesis may have impact on either
the assembly of cellulosic structures within the plant cell walls, or resilience of these
structures to degradation.
89
Chapter 6
Study of structural variation of Amyloid (Aβ) fibril within
brain tissue of AD subjects
The molecular mechanisms of the nucleation and growth of Alzheimer’s disease(AD)
lesions and their spread throughout the brain during disease progression are poorly
understood. Recent studies indicate that Aβ amyloid is capable of self-propagation and that
multiple strains may exist (Petkova et al, 2005; Lu et al, 2013; Watts et al, 2014). This may a
key feature of the mechanisms leading to disease progression and tissue lesions. Here we
demonstrate the use of scanning X-ray microdiffraction (SXMD) to detect structural
differences among fibrils in intact amyloid plaques in histological sections of human brain
tissue from Alzheimer’s Diseases(AD) subjects.
6.1 Method for studying on the brain tissue by SXMD
6.1.1 Background subtraction for patterns of SXMD
Brain samples from AD subjects were fixed with formalin, embedded in paraffin,
sliced to 18 µm thickness using a microtome and then treated with xylene and ethanol to
remove the paraffin. This part of the sample preparation was carried out by Isabel Costantino
(MGH). These sections were subsequently attached to 20 µm thick mica windows; and
mounted on a sample holder compatible with the goiniometer at GM/CA (beam line 23IDB at
the Advanced Photon Source). To ensure no visible radiation damage to the samples, we
limited exposure time to 1 sec. These restrictive experimental conditions lead to diffraction
patterns that exhibited relatively low signal to noise ratios. Therefore, data analysis methods
focused minimizing the impact of the noise and accurately subtracting background scattering
from the diffraction patterns.
Background scattering was largely due to scattering from air (between sample and
detector) and the mica window. Mica is a highly crystalline mineral that occasionally gave
rise to strong Bragg reflections within patterns. But the relative absence of small angle
background or diffuse wide-angle rings makes mica an attractive substrate on which to mount
the samples. We developed and applied automatic image processing methods for spot
90
detection and elimination, in order to minimize the impact of mica scattering on the data used
for analysis of amyloid structure.
Figure 50 shows that the strong reflections from mica window could be extracted
from the diffraction patterns by use of an intensity threshold filter. We first masked the mica
scattering in the diffraction pattern to ensure complete elimination of the scattering from
mica, which is much stronger than the scattering from tissues. Once masking was complete,
the intensity from tissue at the position of the mica spots (that had been eliminated) were
estimated by a 2-dimensional polynomial fitting routine that utilized peripheral scattering
from tissues to interpolate the intensity of these regions. Since most all scattering from these
tissues was circularly symmetric, the one dimensional, circular average of each pattern was
calculated and then used for further analyses.
The intensity distribution from each pixel in the SXMD scan is made up of :
scattering from amyloid (the unknown of interest); scattering from tissue; scattering from
mica; and scattering from air and camera components. The intensity in each pattern was
estimated as a linear combination of scattering from regions of tissue and scattering from air
and camera components:
A+B=1
I is the circular averaged intensity of one diffraction patterns of SXMD. is the
normalized intensity for scattering of tissues (green curve in left image of Figure 52),
is the normalized intensity for scattering of air and mica (blue curve in left image
of Figure 52).
6.1.2 Identification of accumulation of Aβ fibril by SXMD
The possible aggregation of plaques within the brain sections could be determined by
the integral intensity map of small-angle scattering, as shown in Figure 54. The plaques of
amyloid will introduce a small-angle reflection which is not present in scattering from tissue
that is lacking plaques. Therefore, the map of integral intensity at 0.06 Å-1-0.1 Å-1 shows the
potential position of amyloid plaques.
91
6.1.3 The Krakty plot of small angle diffraction
The subtle changes of small angle reflections indicate the structure variation of
plaques (Figure 55). The Krakty plots provide an enhancement of relatively small
differences between the small angle reflections of different plaques. The Krakty plot exhibits
the distribution of (
)
vs
, where the 1/d corresponds to the reciprocal spacing, the
I(1/d) is the scattering intensity at 1/d spacing.
92
Figure 50 Eliminating the scatterings from mica window
Mica windows, that hold the sample for SXMD produces significant diffraction peaks overlapping the
diffraction patterns from tissues. We apply the image processing method described in the text to eliminate
the mica scattering.
Figure 51 Scattering from brain tissue.
Scattering from tissues (green box) was determined by subtracting the background which was chosen as the
scattering from the empty interior of void blood vessels (white box).
r θ
Pixel
93
Figure 52 Extraction of intensity of tissues from pattern including strong background. Right:
Comparison of scattering from background tissue (green) and tissue containing amyloid (blue and
red) Left: Scaled curves from background only and tissue only. The right image demonstrates that
the ratio of background for diffraction from this tissue could be estimated by the linear combination
approach described in the text.
Figure 53 The intensity distributions after background subtraction for SXMD on control tissue. The
left image exhibits the 2500 intensity distributions. The right images shows the mapping of integral
intensity at small angle indicating the potential plaque rich region within tissue.
Pixel
94
Figure 54 Image of scanning features from section of Alzheimer Disease sample. Optical image and
corresponding intensity mapping is shown at left. Center (color) image is a mapping of diffraction
from region marked by black square. The right image shows one example of the circularly averaged
intensity in which the small angle region used to generate the mapping is indicated in solid red.
95
Figure 55 Krakty plots of small angle reflection for plaques accumulated region. The intensity
mappings (top) show the distribution of small angle reflections (upper left) and 4.7Å scattering
(upper right), features expected from Aβ reflections. The intensities plotted in the bottom figure, are
colored to correspond to the positions marked by arrows in the mappings. The small angle spacing
was enlarged for the region highlighted in the inset. The Krakty plots (bottom right) are color coded
similarly.
Krakty Plot
Mapping of small angle reflection Mapping of 4.7 Å reflection
96
6.2 Results
Characterization of a 250x250 μm field of view in tissue samples utilized 50x50
(2500) diffraction patterns taken with a 5μm x-ray beam. We examined samples from
subjects with Alzheimer's Disease (AD), Alzheimer's disease with cerebral amyloid
angiopathy (AD_CAA), no symptoms of cognitive decline but heavy plaque burden
(mismatch), and healthy older patients who exhibited no cognitive decline at time of death
(control), collecting more than 40,000 diffraction patterns. Our results on human brain tissue
demonstrate that detailed characterization of the structure of Aβ fibrils in amyloid is possible
using x-ray scattering from thin sections of brain tissue.
As mention above, features in the x-ray patterns from SXMD can be used to map the
distribution of molecular constituents and their structural characteristics. Eanes and Glenner
(Eanes and Glenner, 1968)studied amyloid fibers from systemic amylidoses and determined
that the constituent proteins took on a cross-beta conformation. Kirschner and his colleagues
(1986) demonstrated that amyloid from subjects with Alzheimer Disease also exhibited a
cross-β conformation as indicated by scattering features from the purified sample that
ehxibited a sharp reflection at 4.7 Å and a diffuse one about 10 Å. We characterize these
scattering from cross-β conformation in fiber as the scattering 'fingerprint' of amyloid. These
characterizations support the capability of SXMD for identifying regions of a thin section
containing amyloid with no need for staining. But, instead of sharp reflections, two diffuse
bands at ~4.6Å and ~10Å have been observed. As Figure 56 shown, reflections at ~4.6Å are
stronger than one at ~10Å, but much diffuse compared to a purified sample. Scanning micro-
diffraction exhibits stronger small angle scattering at some regions of brain section,
distributed in such a way that it may relate to amyloid plaques. The intensity mapping shown
in Figure 58 identifies the positions of these material aggregations, which may or may not be
observable in optical images. Figure 59 also exhibits a distinctive reflections at ~62.5Å, the
reflection may correspond to lipid had been reported in Kirschner et al. 1986 (Kirschner,
1986). However, the treatment with xylene and ethanol may have removed all lipidic material
from the samples prior to x-ray exposure.
We observed that the regions of strong small angle scattering identified as potential
plaques of Aβ tend to accumulate around the blood vessels. The distinctive 4.7 Å and 10 Å,
scattering observed in these regions around vasculature are representative of the fingerprint
97
reflections of extracted Aβ plaques (Figure 56). We observe significant small angle
reflections that correlate to the characterized Aβ scattering. In some cases, the scattering from
small angle region exhibits distinctly different distributions. These differences may be related
to different strains of Aβ.
As shown in Krakty plot of Figure 55, the 250x250 μm scan from the amygdala of a
control subject exhibits several putative plaques according to the distribution of SAXS
intensity. These plaques appear to exhibit three distinct classes of structure as judged by their
SAXS intensities and Krakty Plot of SAXS. This indicates that brain tissue of an 86 year old
control subject with no history of dementia may contain distinct structures from proximal
amyloid plaques surrounding the blood vessel. Presence of amyloid plaques was confirmed
by neuropathological analysis and immuno-histochemical studies.
98
Figure 56 the left image shows the background subtracted diffraction pattern. The right image show
the circular averaged intensity, two reflection at ~10Å (yellow arrow) and ~4.6Å (green arrow).
Figure 57 The montage of integral intensity for small angle and reflection at 4.7Å. The regions studied is marked
by a black square in the optical micrograph in (a). A longitudinal section of a small blood vessel is apparent in
the upper left portion of the region of interest. (b) is the mapping of small angle scattering intensity from the
2500 diffraction patterns that were collected from that region. (c) is a mapping of the correlation coefficient
beween each scattering pattern and the average of several selected from the most intense region of small angle
scattering to the left of the blood vessel.
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
0
10
20
30
40
50
60
70
80
90
100
a b c
99
Figure 58 the integral intensity maps from different brain samples. The upper left is control sample.
the upper right is AD sample. the lower left is AD-CAA sample. the lower right is a sample from a
subject with Pick's disease.
10 20 30 40 50
5
10
15
20
25
30
35
40
45
50
10 20 30 40 50
5
10
15
20
25
30
35
40
45
50
100
Figure 59 Unusual small angle reflections. The upper left is an optical micrograph of the sample, the
up-right is the corresponding integral intensity map. The lower left is s diffraction pattern, the lower
right is an enlargement of the small angle region exhibiting the 62.5 Å spacing diffraction peak.
AD + CAA
INT at 62.5/6.66 A
5 10 15 20 25 30 35 40
5
10
15
20
25
30
35
40
101
6.3 Discussion
As shown in Figure 59, scanning microdiffraction of human brain exhibits relatively
high small-angle intensity in regions with rich amyloid plaque. A 4.7 Å reflection, a
prototypical reflection from amyloid fibrils, was observed, however, slightly weaker and
more diffuse than from extracted amyloid fibrils (e.g. Kirschner et al., 1986). At small angle
region, we occasionally observe a feature at ~ 62.5 Å as well, which may correspond to
lipofuscin (Kirschner et al., 1986). These results indicate the amyloid fibril structure within
real brain tissue may not be as rigid as amyloid extracted samples. Aβ amyloid is predicted to
be capable of forming multiple strains, each of which may be capable of self-propagation
(Malinchik et al, 1998; Petkova et al, 2005; Lu et al, 2013; Watts et al, 2014). Figure 60
exhibits ssNMR-based models of distinct amyloid strains (Petkova et al, 2002 and 2005)
calculated using CRYSOL (Svergun, 1995). Amyloid fibrils with structures as constructed
from PDB files 2LMP (blue); 2M4J (red); 2LMN (black) are predicted to give rise to small-
angle (SAXS) data with distinctive features that should be readily distinguished. These
differences may be examples of the kinds of differences apparent in the Krakty plots from
different plaques as described in Figure 55.
Scanning x-ray microdiffraction provides a very rich source of data on the details of
molecular organization. Extending the approach used here to fresh frozen tissue may provide
a powerful new strategy for understanding the underlying molecular foundation of
Alzheimer's disease. We have demonstrated that scanning x-ray microdiffraction (SXMD)
can detect structural differences among fibrils in intact amyloid plaques in histological
samples of human brain tissue from Alzheimer's Disease (AD) subjects. The results of these
experiments will provide a basis for estimating the frequency of distinct nucleation events in
amyloid formation and dissemination; the relationship between amyloid in different regions
of the brain; the correlation of amyloid strains among proximate plaques; the origin of
amyloid associated with vasculature in CAA; the relationship between amyloid at the core of
plaques and their margins; and the relationship between plaques and associated oligomers.
102
Figure 60 SAXS intensity for three amyloid fibrils models. The three models constructed on the basis
of subunit structures determined by ssNMR – pdb files 2LMP, 2M4J, 2LMN.
103
Chapter 7
Future of SXMD
Advanced synchrotron facilities provide much of the technology for scanning X-ray
microdiffraction, opening the possibility of comprehensive studies of complex materials
aimed at understanding the underlying molecular foundations of hierarchical materials.
Compared to traditional x-ray scattering methods that use millimeter-sized x-ray beams and
average scattering from all material within samples, scanning x-ray microdiffraction provides
extensive information on the order and characteristics of crystalline and non-crystalline
material.
For studies of real tissues, SXMD provides an opportunity to understand molecular
structural information on µm length scales. This extends research range of structural biology,
and introduces the opportunity to study the molecular anatomy of tissues, generating high
resolution information with sub-cellular point-to-point resolution. Another advantage of
SXMD is that sample preparation is easier than other methods for structural biology. In lieu
of strictly limitation on sample thickness for transmission electron microscopy (TEM), less
than 1 µm to avoid multiple scattering, SXMD can scan tissues with µm – mm thickness.
Therefore, SXMD appears to be able to fulfill the research gap for structural biology between
optical microscopy and TEM.
We have shown that scanning microdiffraction applied to myelinated axons, plant
stems and brain sections provides significant information on the heterogeneity of molecular
structure in these complex tissues. In the future, reduction of beam size to ~ 0.5 μm should
enable us obtain detailed molecular information at significantly higher resolution.
But scanning complex tissues with micro-beams produces thousands of times the data
collected by traditional methods; data rich in information about the molecular architecture of
the tissue. It is not unusual for a scanning microdiffraction data set to exceed 10 GB, or,
depending on the sample and the specific questions asked, over 100GB. To generate this data
requires automatic data capture and data analysis methods. To extract information from this
data requires automatic programs that can extract a wide variety of image features from each
pattern. Nevertheless, there is no well established program or algorithm designed to solve
very distinctive problems, that are commonly encountered due to artifacts from facility and
104
special sample preparation. For instance, camera elements count for beam stop, mica sheet —
sample holder for brain section, induce very strong scattering to overlap our real data.
Therefore, a specific program is necessary for SXMD analysis.
My Ph.D. project focused on developing the unique custom software to
automatically extract information about molecular features at the nanoscale from scanning
microdiffraction data. The development of computation methods combines traditional X-ray
methodology and modern image processing algorithms. The studies of myelin sheath,
Arabidopsis and AD subjects by SXMD demonstrate that my custom algorithms perform
well and efficiently for extraction of structural information from microdiffraction patterns.
Based on these engineering tools and crystallography methods, we should be able to extend
these approaches to a wide range of complex tissues.
105
Bibliography
Retrieved from Redi-Earth; Scotts, http://www.scottspro.com: http://www.scottspro.com
Alzheimer's Association. (2015, 6 4). Retrieved from alz.org:
http://www.alz.org/facts/overview.asp
Agarwal U. P., Reiner, R. R., and Ralph, S. A.. (2012). Estimation of Cellulose Crystallinity
of Lignocelluloses Use Near-IR FT-Raman Spectroscopy and Comparison of the
Raman and Segal-WAXS Methods. Journal of Agricultural and Food Chemistry,
61(1), 103-113.
Anderson, C. T., Carrol, A., Akhmetova, L., and Somerville, C. (2010). Real-time imaging of
cellulose reorientation during cell wall expansion in Arabidopsis roots. Plant
physiology, 152(2), 787-796.
APS.ANL.GOV. (n.d.). APS-Welcome. Retrieved from
http://www.aps.anl.gov/About/Welcome/
Arroyo, E. J., Xu, Y. T., Zhou, L., Messing, A., Peles, E., Chiu, S. Y., and Scherer, S. S.
(1999). Myelinating Schwann cells determine the internodal localization of Kv 1.1,
Kv 1.2, Kvbeta 2 and Caspr. Jounral of Neurocytology, 28(4/5)333-347.
Blaurock A. E., and Worthington, C. R. (1966). Treatment f low angle x-ray data from
planar nad concetric multilayered structures. Biophys Journal, 305-312.
Bonawitz N. D., Kim, J. I., Tobimatsu, Y., Ciesielski, P. N., Anderson, N. A., Ximenes, E.,
Maeda, J., Ralph, J., Donohoe, B. S., Ladish, M., and Chapple, C. (2014). Disruption
of Mediator rescues the stunted growth of a lignin-deficient Arabidopsis mutant.
Nature, 509(7500), 376-380.
Burk, D. H., and Ye, Z. H. (2002). Alteration of oriented deposition of cellulose microfibrils
by mutation of a katanin-like microtubule-severing protein. The Plant Cell, 14(9),
2145-2160.
Caspar, D. L. D. and Kirschner, D. (1971). Myelin Membrane Structure at 10 A Resolution.
Nature New biology, 231:46-52.
Cave, I. D. (1966). Theory of X-ray measurement of microfibril angle in wood. . Forest
Products Journal, 16, 37-42. .
Chapple C. C. S., Vogt, T., Ellis, B. E., and Somerville, C. R. (1992). An Arabidopsis mutant
defective in the general phenylpropanoid pathway. The Plant Cell, 4(11), 1413-1424.
106
De Felici M, Felici, R., Ferro, C., Tartari, A., Gambaccini, M., and Finet, S. (2008).
Structural characterization of the human cerebral myelin sheath by small angle x-ray
scattering. Physics in Medicine and Biology, 53(20):5675-5688.
Demarty M., Morvan, C. and Thellier, M. (1984). Calcium and the cell wall. Plant, Cell &
Environment, 7(6), 441-448.
Dharmawardhana, D. P., Ellis, B. E., and Carlson, J. E. (1992). Characterization of vascular
lignification in Arabidopsis thaliana. Canadian journal of botany, 70(11), 2238-2244.
Ding, S. Y. and Himmel, M. E. (2006). The maize primary cell wall microfibril: a new model
derived from direct visualization. Journal of Agricultural and Food Chemistry,
54,597-606.
Ducic T, Quintes, S., Nave, K. A., Susini, J., Rark, M., Tucoulou, R., Alvera, M., Guttmann,
P., and Salditt, T., (2011). Strucutre and composition of myelinated axons: a
multimodel synchrotron sepctro-microscopy study. Journal of Structural Biology,
173(2):5675-5688.
Eanes, E. D. and Glenner, G. G. (1968). X-ray diffraction studies on amyloid filaments.
Journal of Histochemistry & Cytochemistry, 16(11), 673-677.
Emons, A. M. C. and Mulder, B. M. (1998). The making of the architecture of the plant cell
wall: How cells exploit geometry. PNAS, 95(12),7215-7219.
Ensikat, H., Boese, M., Mader, W., Barthlott, W., and Koch, K. (2006). Crystallinity of plant
epicuticular waxes:electron and X-ray diffraction studies. Chemistry and Physics of
Lipids, 48,168-172.
Fernandes, A. N., Thomas, L. H., Altaner, C. M., Callow, P., Forsyth, V. T., Apperley, D. C.,
Kennedy, C. J., and Jarvis, M. C. (2011). Nanostrucutre of cellulose microfibrils in
spruce wood. Proccedings of the National Academy of Sciences, 108(47),1195-1203.
Fink H. P., Philpp, B., Paul, D., Serimaa, R., and Paakkari, T. (1987). The strucutre of
amorphous cellulose as revealed by wide-angle X-ray scattering. Polymer,
28(8),1265-1270.
Fraser C. M., and Chapple, C. (2011). The phenypropanoid pathway in Arabidopsis. In The
Arabidopsis Book. American Society of Plant Biologist.
Fratzl, P., Jakob, H. F., Rinnerthaler, S., Roschger, P, and Klaushofer, K. (1997). Position-
resolved small-angle X-ray scattering of complex biological material. Journal of
Applied Crystallography., 30, 765-769.
Goujon T., Ferret, V., Mila, I., Pollet, B., Ruel, K., Burlat, V., Joseleau, J., Barriere, Y.,
Lapierre, C., and Jouanin, L. (2003). Down-regulation of the AtCCR1 gene in
107
Arabidopsis thaliana: effects on phenotype, lignins and cell wall degradability. Planta,
217(2), 218-228.
Guinier A. (1994). X-ray diffraction in Crystals, imperfect crystals, and amorphous bodies.
Courier Corporation.
Hannoufa A., McNevin, J. and Lemieux, B. (1993). Epicuticular waxes of eceriferum
mutants of Arabidopsis thaliana. Phyochemistry, 33,851-855.
Hirano A, and Dembitzer, H. M. (1982). Further studies on the transverse band. Journal of
Neurocytology, 11(6):861-866.
Hollingshead C. J., Caspar, D. L. D., Melchior, V., and Kirschner, D. A. (1981). Compaction
and particle segregation in myelin membran arrays. The Journal of Cell Biology,
89(3):631-644.
Inouye H., Liu, J., Makowski,. L., Palmisano, M., Burghammer, M., Riekel, C. and Kirschner,
D. A. (2014). Myelin Organization in the Nodal, Paranodal, and Juxtaparanodal
Regions Revealed by Scanning X-Ray Microdiffraction. PLoSone, e100592.
Inouye H. , Tsuruta, H., Sedzil, J., Uyemura, K., and Kirschner, D. A. (1999). Tetrameric
assembly of full-sequence protein zero myelin glycoprotein by synchrotron x-ray
scattering. Biophysical Journal, 423-437.
Janet L Smith, Fischetti, R. F., and Yamamoto, M. (2012). Micro-crystallography comes of
ages. Current Opinion in Strucutural Biology, 22:602-612.
Jenks M. A., Tuttle, H. A., Eigenbrode, S. D., and Feldmann, K. A. (1995). Leaf epricuticular
waxes of the eceriferum mutans in Arabidopsis. Plant Physiology, 108(1),369-377.
Kennedy, C. J., Cameron, G. J., Sturcova, A., Apperley, D. C., Altaner, C., Wess, T. J., and
Jarvis, M. C. (2007). Microfibril diameter in celery collenchyma cellulose: X-ray
scattering and NMR evidence. Cellulose, 14(3), 235-246.
Kirschner D. A. (1992). in Myelin:Biology and Chemistry. In D. Martenson, Boca Ration (pp.
3-78). CRC Press, Inc.
Kirschner D. A and Hollingshead C. J. (1980). Processing for electron microscopy alters
membrane strucutre and packing in myelin. Journal of Ultrastructure research,
73(2):211-232.
Kirschner DA, Ganser, A. L., and Caspar, D. L. (1984). Diffraction studies of molecular
organization and interactions in myelin membranes. In Myelin (pp. 51-95). New York:
Plenum Press.
Kirschner, D. A., Abraham, C., and Selkoe, D. J. (1986). X-ray diffraction from intraneronal
paired helical filaments and extraneuronal amyloid fiber in Alzheimer disease disease
108
indicates cross-beta conformation. Proceeding of the National Academy of Science,
83(2),503-507.
Langan, P., Petridis, L., O'neill, H. M., Pingali, S. V., Foston, M., Nishiyama, Y., Schulz, R.,
Lindner B, Hanson, B., Harton, S., William T. H., Urban, V., Evans, B. R.,
Gnanakaran, S., Ragauskas, A. J., Smith, J. C. , and Brian H. D. (2014). Common
processes drive the thermochemical pretreatment of lignocellulosic biomass. . Green
Chemistry, 16(1), 63-68.
Lichtenegger, H. Muller, M., Paris, O., Riekel, C., and Fratzl, P. (1999). Imaging of the
helical arrangement of cellulose fibrils in wood by synchrotron X-ray
microdiffraction. J. Applied Crystallography., 32,1127-1133.
Liu, J., Inouye, H., Venugopalan, N., Fischetti, R. F., Gleber, S. C., Vogt, S., Cusumano, J.
C., Kim, J. I., Chapple, C., and Makowksi, L. (2013). Tissue specific specialization
of the nanoscale architecture of Arabidopsis. Journal of Structural Biology,
184(2),103-114.
Lu, J. X., Qiang, W., Yau, W. M., Schwieters, C. D., Meredith, S. C. and Tyco, R. (2013).
Molecular structure of β-amyloid fibrils in Alzheimer’s disease brain tissue. Cell,
154(6), 1257-1268.
MacKenzie, M. L., Ghabrel, M. N., and Allt, G. (1984). Nodes of Ranvier and Schmidt-
lanterman incisures: an in vivo lanthamum tracer study. journal of Neurocytology,
1043-1055.
Malinchik, S. B., Inouye, H., Szumowski, K. E., and Kirschner, D. A., (1998), Structure of
Analysis of Alzheimer's ß(1-40) Amyloid: Protofilament Assembly of Tubular Fibrils.
Biophysical Journal, Vol 74, 537-545.
Meyer, K. Cusumano, J. C., Somerville, C., and Chapple, C. (1996). Ferulate-5-hydroxylase
from Arabidopsis thaliana defines a new family of cytochrome P450-dependent
monooxygenases. Proceedings of the National Academy of Sciences, 93(14), 6869-
6874.
Meyer, K. SHirley, A. M., Cusumano, J. C., Bell-Lelong, D. A., and Chapple, C. (1998).
Lignin monomer composition is determined by the expression of a cytochrome P450-
dependent monooxygenase in Arabidopsis. Proceedings of the National Academy of
Sciences, , 95(12),6619-6623.
Mugnaini, E, Osen, K. K., Schnapp, B., and Friedrich, Jr, V. L. (1977). Distribution of
schwann cell cytoplasm and plasmalemmal vesicles in peripheral myelin sheaths. An
electron microscopic study with thin setions and freeze-fracturing. Journal of
Neurocytology, 647-668.
109
Muller, M. Czihak, C., Vogl, Fratzl, P. Schober, H., and Riekel, C. (1998.). Direct
Observation of Microfibril Arrangement in a Single Native Cellulose Fiber by
Microbeam Small-Angle X-ray Scattering, . Macromolecules, 31, 3953-3957.
Nishiyama Y., Langan P. and Chanzy, H. (2002). Crystal Structure and Hydrogen-Bonding
System in Cellulose Iβ from Synchrotron X-ray and Neutron Fiber Diffraction.
Journal of the American Chemical Society, 124, 9074-9082.
Nishiyama, Y. (2009). Structure and properties of the cellulose microfibril. Journal of Wood
Science, 55,241-249.
Paakkari T. and Sermaa R. (1989). Structure of Amorphous cellulose. Acta polymerica, 40,
731-734.
Patten A. M., Jourdes, M., Cardenas, C. L., Laskar, D. D., Nakazawa, Y., Chung, B.,
Franceschi, V. R., Davin, L. B., and Lewis, N. G. (2010). Probing native lignin
macromolecular configuration in Arabidopsis thaliana in specific cell wall types:
further insights into limited substrate degeneracy and assembly of the lignins of ref8,
fah1-2 and C4H:F5H lines. Molecular Biosystems, 6(3),499-515.
Paunesku T., Vogt, S., Maser, J., Lai, B., and Woloschak, G. (2006). X-ray fluorescence
microprobe imaging in biology and medicine. Journal of Cellular Biochemistry., 99,
1489-1502.
Petkova, A. T., Isshii, Y., Balbach, J. J., Antzukin, O. N., Leapman, R. D., Delaglio, F., and
Tyco, R. (2002). A structural model for Alzheimer’s beta-amyloid fibrils based on
experimental constraints from solid state NMR. Proceedings of the National
Academy of Science, 99(26): 16742–16747.
Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W. M., Mattson, M. P., and Tyco, R. (2005).
Self-Propagating, Molecular-Level Polymorphism in Alzheimer’s beta-Amyloid
Fibrils. Science, 307, 262-265.
Poccia, N., Di Cola, E., Caporale, A. S., Bianconi, A., Campi, G., Hawkins, T. A., and Ricci,
A. (2013). Mapping early stage of myelin degradation at nanoscale resolution.
Scientific Report, 5430.
Rand R. P., Fuller, N. L., and Lis, L. J. (1979). Myelin swelling and measurement of forces
between myelin membranes. Nature, 279:258-260.
Rashotte A. M., Jenks, M. A., and Feldmann, K. A. (2001). Cuticular waxes on eceriferum
mutants of Arabidopsis thaliana. Phytochemistry, 57(1),115-123.
Rashotte A. M., Jenks, M. A., Ross, A. S., and Feldmann, K. A. (2004). Novel eceriferum
mutants in Arabidopsis thaliana. Planta, 219(1),5-13.
110
Rasmussen, S. G., Choi, H. J. Rosenbaum, D. M., Kobilka, T. S., Thian, F. S., Edwards, P. C.,
Burghammer, M., Ratnala, V. R. P., Sanishvili, R., Fischetti, R., Schertler, G. F. X.,
Weis, W. I., and Kobila, B. K. (2007). crystal strucutre of the human beta2
adrenergic G-protein-coupled receptor. nature, 450,383-387.
Riekel C., Craig, C., Burghammer, M., and Muller, M. (2001). Microstructural homogeneity
of support silk spun by Eriophora fuliginea(C. L. Koch) determined by scanning X-
ray microdiffraction. Naturwissenschaften, 88,67-72.
Riekel, C., Cedola, A., Heidelbach, F., and Wagner, K. (1997). Microdiffraction experiments
on single polymeric fibers by synchrotron radiation Macromolecules.
Macromolecules, 30(4),1033-1037.
rruff.info. (n.d.). Retrieved from rruff.info: rruff.info
Ruegger M., Meyer, K., Cussumano, J. C., and Chapple, C. (1999). Regulation of ferulate-5-
hydroxylase expression in Arabidopsis in the context of sinapate ester biosynthesis.
Plant physiology, 119(1), 101-110.
Salem M, Mauguen, Y., and Prange, T. (2010). Revisiting glutaraldehyde cross-linking: the
case of the Arg-Lys intermolecular doublet. Acta Crystallogrographica Section F:
Structural Biology and Crystallization Communications, 66(3):225-228.
Schilmiller A. L., Stout, J., Weng, J. K., Humphreys, J., Ruegger, M. O., and Chapple, C.
(2009). Mutations in the cinnamate 4‐hydroxylase gene impact metabolism, growth
and development in Arabidopsis. The Plant Journal, 60(5), 771-782.
Schroer, C. G., Boye, P., Feldkamp, J . M., Patommel, J., Schropp, A., Schwab, A., Stephan,
S, Burghammer, M. , Schoder, S., and Riekel, C. (2008). Coherent x-ray diffraction
imaging with nanofocused illumination. Physical review letters, 101(9),090801.
Somerville, C. Bauer, S., Brininstool, G., Facette, M., Hamann, T., Milne, J., Osborne, E.,
Paredez, A., Persson, S., Raab, T., Vorwerk, S., and Youngs, H. (2004). Toward a
systems approach to understanding plant cell walls. Science, 306(5705), 2206-2211.
Suh, M . C. Samuels, A. L., Jetter, R. Kunst, L., Pollard, M., Ohlrogge, J. and Beisson, F.
(2005). Cuticular Lipid Coposition, Surface Structure and Gene Expression in
Arabidopsis Stem Epidermis. Plant Physiology, 139,1649-1665.
Svergun, D. B arberato, C., and Koch, M. H. J. (1995). CRYSOL - a Program to Evaluate X-
ray Solution Scattering of Biological Macromolecules from Atomic Coordinates.
Journal of applied crystallography, 28, 768-773.
Thomas, L. H., Forsyth, V. T., Sturcova, A., Kennedy, C. J., May, R. P., Altaner, C. m.,
Apperley, D. C., Wess, T. C., and Jarvis, M .C. (2013). Structure of cellulose
microfibrils in primary cellwalls from collenchyma. Plant Physiology, 161,465-476.
111
Trapp B. D. and Kidd G. J. (2004). Structure of the myelinated axon. In Myelin Biology and
Disorders (pp. Vol. 1:3-27). Amsterdam: Elsevier/Academic Press.
Turner S. R. and Somerville C. R. (1997). Collapsed xylem phenotype of Arabidopsis
identifies mutants deficient in cellulose deposition in the secondary cell wall. The
Plant Cell online, 9, 689-701.
Van Buren, J. P. (1991). The chemistry and technology of pectin. N.Y.: Academic Press, Inc.
Vogt, S. (2003). MAPS : A set of software tools for analysis and visualization of 3D X-ray
fluorescence data sets. In Journal de Physique IV( Proceedings), Vol104, 635-638.
Wardrop, A. B. (1976). Lignification of the plant cell wall. Appl. Polym. Symp. , 28, 121-144.
Watts, J. C., Condello, C., Stohr, J., Oehler, A., Lee, J., DeArmond, S. J., Lannfelt, L.,
Ingelsson, M., Giles, K., and Prusiner, S. B. (2014). Serial propagation of distinct
strains of Aβ prions from Alzheimer’s disease patients. Proc. Natl. Acad. Sci., 111,
10323–10.
webmineral.com. (n.d.). Retrieved from webmineral.com: webmineral.com/Mineral_
Definition.shtml
Yagi, N. (2009). A scanning SXAS/WAXS study of rat brain. journal of Physics: Conference
Series, 272:1-6.
Zhong R., Ripperger, A., and Ye, Z. H. (2000). Ectopic deposition of lignin in the pith of
stems of two Arabidopsis mutants. Plant Physiology, 123,59-70.