Inhomogeneity of Cellulose Microfibril Assembly in Plant Cell Walls

Subscriber access provided by Kaohsiung Medical University

B: Biophysical Chemistry and Biomolecules

Inhomogeneity of Cellulose Microfibril Assembly in Plant Cell Walls Revealed with Sum Frequency Generation Microscopy Shixin Huang, Mohamadamin Makarem, Sarah Nelson Kiemle, Hossein Hamedi, Moujhuri Sau, Daniel J. Cosgrove, and Seong H. Kim J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.8b01537 • Publication Date (Web): 26 Apr 2018 Downloaded from http://pubs.acs.org on April 27, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Inhomogeneity of Cellulose Microfibril Assembly in Plant Cell Walls Revealed with Sum Frequency Generation Microscopy Shixin Huang,1+ Mohamadamin Makarem,1+ Sarah N. Kiemle,2 Hossein Hamedi,1 Moujhuri Sau,1 Daniel J. Cosgrove,2 and Seong H. Kim1* 1

Department of Chemical Engineering and Materials Research Institute, Pennsylvania State

University, University Park, PA 16802 2

Department of Biology, Pennsylvania State University, University Park, PA 16802

+ Equal contributions from these two authors *

Corresponding author: [email protected]

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT: Sum frequency generation (SFG) vibrational spectroscopy can selectively detect and analyze noncentrosymmetric components interspersed in amorphous matrices; this principle has been used for studies of nano-scale structure and meso-scale assembly of cellulose in plant cell walls. However, the spectral information averaged over a large area or volume cannot provide regio-specific or tissue-specific information of different cells in plants. This study demonstrates spatially-resolved SFG analysis and imaging by combining a broadband SFG spectroscopy system with an optical microscope. The system was designed to irradiate both narrow-band 800 nm and broad-band tunable IR beams through a single reflective objective lens, but from opposite sides of the surface normal direction of the sample. The developed technique was used to reveal inhomogeneous distributions of cellulose microfibrils within single cell walls such as cotton fibers and onion epidermis as well as among different tissues in Arabidopsis inflorescence stems and bamboo culms. SFG microscopy can be used for vibrational spectroscopic imaging of other biological systems in complement to conventional FTIR and confocal Raman microscopy.


Page 2 of 47

Page 3 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION Nonlinear optical spectroscopy such as sum frequency generation (SFG) and second harmonic generation (SHG) can reveal molecular or structural information that linear spectroscopy cannot find. Examples of such nonlinear optical studies include selective detection of interfacial species buried in random or amorphous bulk phases containing molecules that are chemically identical to the interfacial species, orientations of molecules or crystalline phases without inversion symmetry with respect to a reference axis, etc.1-7 The noncentrosymmetry and phase matching requirements of SFG and SHG allows selective detection of certain functional groups or molecules arranged without inversion symmetry in space, which cannot be done with linear spectroscopy due to interferences from random or centrosymmetric medium.4,

8

The same

principles have been used to study crystalline cellulose microfibrils in plant cell walls or lignocellulose biomass.9-14 The use of SFG for detection of cellulose in plant cell walls and biomass circumvents the problem of spectral interferences from amorphous hemicellulose and lignin.9-11 Taking advantage of this, SFG has been used for study of cellulose structures in native samples with minimal treatments (such as simple drying without any chemical staining or separation).9-11, 14-31 Initially, interpretation of SFG spectral features was difficult due to the lack of detailed knowledge on how SFG peaks of cellulose can be correlated to the nanoscale crystalline structure of cellulose itself and the mesoscale structural organizations among crystalline cellulose domains interspersed in amorphous matrices. Recently, significant understanding of SFG spectral interpretations has been made through control studies using wellcharacterized model systems and computational calculations.11-12, 32-35 These studies showed that at the nanoscale, the chain packing order in the crystal unit cell can be determined from the SFG



peak shape of the OH groups involved in hydrogen bonding interactions between chains.32-33 It was also discovered that the assembly patterns of nanocrystalline domains in the mesoscale dimension can substantially alter the SFG spectral features.11, 36 The polarity of crystallites within the SFG coherence length influences the relative intensities of CH (2840 – 2980 cm-1) and OH (3240 – 3450 cm-1) stretch modes differently, which allows independent determination of the overall packing polarity in the mesoscale.34, 36 For the uniaxially aligned cellulose sample, the preferential orientation direction can be determined from the SFG dependence on the probe beam polarization and the azimuth angle with respect to the laser incidence plane.18, 34 The CH/OH relative intensities in SFG spectra of uniaxially-aligned cellulose can be used to determine the inter-crystallite distance in the bulk sample.12 By detecting SFG signals at non-phase-matching scattering angles, SFG can distinguish the functional groups at the surface of the crystalline domain and those inside the domain.35 These newly discovered insights can make SFG a good analytical tool for study of nanoand mesoscale structural organizations of cellulose in plant cell walls, biomass, and cellulosecontaining engineering composites.

Dealing with biological samples, there is another level of complexity that must be considered. Biological samples are often inhomogeneous and sometimes hierarchically structured depending on biological functions – plant cell walls are not exceptions.37 Depending on the biological and mechanical functions of cells in plants, the nano- and meso-scale structural orders of cellulose microfibrils vary substantially.38-44 So, the area- or volume-averaged spectroscopic analysis may not be able to unveil cell-specific structural information. For that reason, it is necessary to conduct SFG analysis of cellulose in plant cell walls with sufficiently high spatial resolutions. There are many SFG microscopy systems developed and reported in the literature.45-51 The system reported in this paper is unique in that it uses broad-band (BB) SFG system for hyperspectral imaging and designed to work for analysis of rough and inhomegeneous 4 ACS Paragon Plus Environment

Page 4 of 47

Page 5 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


bulk samples such as plant cell walls in both dehydrated and hydrated states. The BB-SFG system collects the data of a broad spectral region of interest (determined by the bandwidth of input IR pulses), instead of recording the intensity of one wavenumber at a time. Thus, it allows collecting a full spectrum at each pixel. Since the areal power density of the beam focused with an objective lens with a high numerical aperture (NA) can be extremely high and the photon collection efficiency is also high, the spectral signal-to-noise ratio of SFG signals scattered from a small area could be even better than that of the spectra collected over a large area with normal lenses with large focal distances. In our system, broad-band spectra of plant cell walls are collected in 0.1 ~ 1 second per pixel, depending on signal intensity of the sample. This makes it possible to collect one hyperspectral image of a 300 µm × 300 µm area with a 5 µm/pixel resolution in an hour or less at one bandwidth setting of IR. Also, two input laser beams are focused from the opposite sides of the surface normal axis of the sample using one objective lens; this allows to study the polarization dependence of SFG signals as in typical table-top experiments.36 This uniquely-configured BB-SFG microscope system was used to reveal the inhomogeneity of cellulose microfibril structure and assembly inside a single cell walls as well as among cell walls of different tissues.

EXPERIMENTAL DETAILS The SFG system was constructed with a Ti:sapphire laser/amplifier system (Coherent, Libra), delivering 800 nm pulses with a bandwidth (FWHM) of 12 nm, a pulse duration of 85 fs, and a pulse energy of ~2.4 mJ at a 2 kHz repetition rate, and an optical parametric generation and amplification system (Coherent, OPerA Solo), producing tunable broad-band IR pulses in the 2.5‒10 µm range (corresponding to 4000–1000 cm-1) with a bandwidth of about 150‒200 cm-1 5 ACS Paragon Plus Environment


depending on the phase matching at the non-collinear difference frequency generation crystal.50 For spectral narrowing of the 800 nm pulse, a pair of angle-tuned Fabry-Pérot etalons (TecOptics; finesse = 75.5 and free spectral range = 483.3 cm-1 at 800 nm) were used. The pulse-width and spectral-width of the narrowed 800 nm beam were 2.1-2.4 ps and 0.78 nm, respectively. The broad-band IR and narrow-band 800 nm pulses were aligned to travel to the entrance port of a microscope (Olympus, BX51W1) in parallel, but spatially separated (approximately 1.5~3 mm between the nearest edges). The details of this system can be found in our previous publication.50

Figure 1. Schematic view of the BB-SFG set-up using a Schwarzschild-Cassegrain reflective objective lens. The narrow-band 800nm pulse and broad-band IR pulse are collinearly 6 ACS Paragon Plus Environment

Page 6 of 47

Page 7 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


propagating, but spatially separated, before they enter the reflective objective lens. When they come out of the lens, they are focused to the sample from the opposite sides of the surface normal axis (see Figure S1 in Supporting Information). Thus, the incidence of plane can be defined even though the objective lens is at the surface normal direction. Figure 1 schematically illustrates the microscope design. This configuration was chosen after considering various options [for details, see the Supporting Information, SI]. The broadband IR and narrow-band 800 nm pulses were reflected from the opposite sides of the center convex lens and then focused via the peripheral concave lens onto the sample from the opposite sides (see Figures S1 and S2 in SI). The incidence angle of the beams on the sample was ~17.8 ± 5.6o with respect to the surface normal when a 15× reflective lens (NA = 0.4; Newport Model #50105-02) was used and ~22.5 ± 7.5o when a 36× reflective lens (NA = 0.52; Newport Model #50102-02) was used. Note that the variance of the incidence angle shown here is the widest range calculated from the dimension of the lens (Figure S1). It can be reduced by making the width of the incidence beam smaller; this could increase the spatial resolution at the expense of the SFG signal intensity. For translucent samples, the SFG signal can be collected in the transmission mode using a condenser lens under the sample. For thick or opaque samples, the SFG signal can be collected in the reflection mode through the reflective lens and a dichroic mirror.50 The broad-band SFG signal was spectrally resolved using a volume-phase holographic (VPH) grating (Andor, Holospec) and recorded with a charge-coupled device camera (Andor, DU420A-BEX2-DD). The sample was step scanned using a nano-positioner with a translation range of 300 µm × 300 µm (MadCityLab, Nano-Bio300). In order to estimate the spatial resolution of the developed BB-SFG microscope, we employed a step-scan method using a sharp edge of a reference sample. The reference used for 7 ACS Paragon Plus Environment


this purpose was a thin film of uniaxially-aligned cellulose Iβ crystals. Its optical profilometry image is shown in Figure S3 in SI. By fitting the SFG intensity as a function of the reference sample position, the volume of the overlapped beam spot was determined to be ~7.9 µm × ~5.4 µm wide and ~26 µm deep for the 15× objective lens and ~4.1 µm × ~2.4 µm wide and ~15 µm deep for the 36× objective lens when the 800 nm beam diameter entering the reflective objective was ~7 mm. More details of the beam spot size analysis are provided in Figure S4 in SI. This BB-SFG microscope system was used to analyze individual fibers of cotton (Gossypium hirsutum), walls of individual abaxial epidermis cells of onion (Allium cepa), cell walls in the ultramicrotomed sections of inflorescence stems of 6-week and 8-week old Arabidopsis thaliana, and a stem of fully-grown bamboo (Bambusa multiplex). The data collected with the 15× objective are presented here. Details of the cotton fiber sample were described in our previous publication.26 The sample preparation of onion epidermis walls was described in detail elsewhere.11,

52

The bamboo sample was collected in a local garden.

Arabidopsis thaliana (ecotype Col-0) plants were grown in a chamber programmed for 12-h light / 12-h dark photoperiod (22°C /16 °C) at 120-150 µmol m−2 s−1 light intensity.14, 53 The stems were harvested at 6-weeks or 8-weeks after germination. Small segments (~1 cm long) were excised from the stems, flash-frozen in ShandonTM CryomatrixTM (Thermo Scientific) matrix, and then cryo-sectioned into 20-µm-thick transverse sections using a Leica CM1950 cryostat. For SFG analysis of samples in the fully-hydrated state, the sectioned samples were rinsed with D2O and then sandwiched with two glass cover slips with D2O in the gap. Using D2O circumvented the IR attenuation problem by water when the OH stretch modes were probed with SFG. Although this may prevent studying the OH groups at the cellulose surface accessible by D2O (because they will be exchanged to OD), it still allows the detection of the stretch modes of 8 ACS Paragon Plus Environment

Page 8 of 47

Page 9 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


OH groups inside the crystalline domain which will not be exchanged with OD.10 The birefringence image of crystalline cellulose domains were obtained with a cross-polarized microscope (Olympus BX63 Motorized microscope, Japan).

RESULTS AND DISCUSSION Polarization and Azimuth Angle Dependences of SFG Spectral Features of Uniaxially-Aligned Cellulose Crystals. Even though only one reflective objective lens is used at the surface normal axis, two incidence beams are off the surface normal by the same degree in the opposite sides (see Figure S2 in SI). Thus, the plane of incidence can be defined and polarization dependent measurements can be performed.36 In SFG experiments, the polarization combination of probe beams are expressed in a combination of three polarization symbols (s or p) in the order of SFG signal, 800 nm input beam, and IR input beam. If the sample has a preferential alignment direction, azimuth angle (φ) dependent measurements can also be performed.36 Figure 2 displays the SFG spectra of the uniaxially-aligned cellulose Iβ film at two polarization combinations (pps and sps) and at two azimuth angles (φ = 0o and 90o with respect to the laser incidence plane). The raw data and relative intensity ratios are shown in Figure S5 and S6 in SI. ()

The second-order susceptibility tensor, , responsible for SFG is proportional to the product of the derivatives of polarizability and dipole with respect to the SFG-active normal ()

mode.34, 54 The experimental data of the SFG signal intensity, ∝ , as a function of the probe beam polarization and azimuth angle have been measured for a uniaxially-aligned control sample using a table top system and compared with theoretically-computed SFG 9 ACS Paragon Plus Environment


spectra.36 This previously reported data set can be used as a qualitative guidance for interpretation of the data obtained with the microscope system although the probe beam incidence and detection angles are different. When the pps polarization is used (Figure 2a), the OH signal intensity of cellulose is relatively strong when the cellulose chains are perpendicular to the laser incidence plane (φ=90o); in contrast, it is small when the chains are parallel to the laser incidence plane (φ=0o) (Figure S6b). Similar to the ssp polarization measurement with the table-top set-up,36 the OH/CH intensity ratio in the pps polarization is larger when the cellulose chain axis is parallel to the IR electric field. When the sps polarization is used (Figure 2b), the OH/CH intensity ratio does not vary significantly with the azimuth angle (Figure S6b), although the OH peak shape varies with the angle. These polarization and azimuth angle dependences are consistent with the trend observed for the area-averaged SFG spectra in the table-top experiment.36 Also, changes in the CH stretch region are noted when the azimuth angle is changed.55 The spectral shape and relative intensity changes with the azimuth angle in the pps polarization combination can be useful to distinguish the mesoscale packing of cellulose in plant cell walls.


Page 10 of 47

Page 11 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. SFG spectra of a uniaxially-aligned cellulose Iβ film at azimuth angles of 0o and 90o at polarization combinations of pps and sps. The p and s represents the electric vector of the light parallel and perpendicular to the laser incident plane, respectively. The spectra are collected in the transmission mode and plotted as a function of wavenumber of the incident IR beam. The spectra shown here are averaged and normalized; the raw spectra are shown in Figure S5 and the relative intensities are sown in Figure S6. It should be noted that due to the noncentrosymmetric symmetry of the crystal structure, cellulose has an intrinsic birefringence56 and rotates the polarization of light. This property is widely used to detect the presence of cellulose in plant cell walls (as shown in Figure 3b). The intrinsic birefringence of cellulose may alter the polarization state of the IR and 800 nm beam as well as the SFG beam propagating into the sample. So, it is difficult to analyze the observed polarization dependence of SFG signal theoretically; nonetheless, the data shown in Figure 2 can



be used as an empirical rule to estimate the relative orientation of cellulose microfibrils with respect to the laser incidence plane. The main peak at 2944 cm-1 is tentatively assigned to the CH2 asymmetric stretch mode of the exocyclic CH2OH side group coupled with the CH stretch modes in the axial positions of the 6-membered ring.34 The peaks at 2920-2925 cm-1 and 2964-2970 cm-1 could be ascribed to the CH2/CH coupled modes;34 their relative intensities compared to the 2944 cm-1 peak appear to be sensitive to the spatial arrangement of crystalline cellulose domains with respect to the laser incidence plane.11, 18, 55 The small peak at 2850-2870 cm-1 is presumed to have the contribution from the CH2 symmetric stretch mode coupled with the CH stretch modes.55 The broad peak centered at 3320~3330 cm-1 is the OH stretch modes that are delocalized throughout the entire crystalline domain.33-34 The peaks at higher than 3400 cm-1 are the weakly hydrogen bonded OH groups that may exist at the surface of crystalline cellulose domains.35 Inhomogeneity of Cellulose Microfibril Assembly in a Single Cell Wall. Fullymatured cotton fibers are single trichome cells with a thick secondary cell wall consisting of mostly cellulose.57-58 The cotton fiber has a kidney-shape cross-section due to collapsed lumen.17 Cellulose microfibrils are arranged helically around the fiber wall with periodic reversals in the deposition angle; fibers become twisted at these reversal regions.57-58 Although their tensile strength along the fiber axis is large, individual fibers can easily be bent or twisted since they are thin (10-20 µm in diameter). Figure 3 displays the optical and birefringence images as well as SFG hyperspectral images at 2944 cm-1 (the main peak in the CH stretch region) and 3320 cm-1 (the main peak in the OH stretch region) of a single cotton fiber. The fiber sample was naturally dried upon opening of the fully-matured cotton boll in the growth field, but it was immersed in D2O and analyzed in D2O to reduce the scattering loss of light at the fiber surface. 12 ACS Paragon Plus Environment

Page 12 of 47

Page 13 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In Figure 3, the variance in birefringence intensity could be caused by multiple factors. The first is the change in local thickness as the fiber with the kidney shape cross-section twists, which will change the degree of rotation of the polarized light traveling through the sample. The second is the change in the cellulose microfibril angle since it will also alter the rotation of the polarized light. The hyperspectral images of the 2944 cm-1 and 3320 cm-1 SFG intensities show that the maximum intensity locations are different in two spectral images; and they are also different from the brightness in the birefringence image. The overall SFG intensity at different locations might vary; because the thickness of the cotton fiber is comparable with the focal depth of the 15× objective lens, this may happen if the cotton fiber is not perfectly flat on the glass substrate. However, the relative intensity difference between the 2944 cm-1 and 3320 cm-1 peaks in individual pixels cannot be explained by the sample position with respect to the maximum focal plane since they are collected at the same location.



Figure 3. Optical image from bright-field microscopy, birefringence image from crosspolarization microscopy, and SFG hyperspectral images at 2944 cm-1 and 3320 cm-1. The SFG spectra were collected with the 15× objective lens. The polarization combination of probe beams was pps and the laser incidence plane is in the vertical direction of the SFG image. The SFG signals were collected in the transmission mode. The contour lines in the SFG spectral images are drawn in comparison with the optical image as a guide to eyes.


Page 14 of 47

Page 15 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. Averaged SFG spectrum of cellulose in a single cotton fiber and SFG spectra of different pixels marked as 1−8 in the 2944 cm-1 intensity map. Each spectrum is from the ~8µm × ~5.5µm elliptical-shape beam spot. The polarization combination of probe beams was pps and the laser incidence plane is in the vertical direction of the SFG image. Figure 4 displays the SFG spectrum of cellulose averaged over the entire fiber region imaged and the SFG spectra of cellulose at eight different locations. The averaged spectrum shows a single dominant peak at 2944 cm-1 and a weak peak at ~2860 cm-1 in the CH stretch region and a relatively weak peak at 3320cm-1 in the OH stretch region. Among 8 locations analyzed, locations 1, 2, and 6 give the spectral features similar to the averaged spectrum. In locations 3, 4, and 5 where the cotton fiber is severely bent, the SFG spectral features are drastically different from the averaged spectrum. Especially, the CH peaks at 2860-2870 cm-1, 2915-2920 cm-1, and 2964-2970 cm-1 are greatly enhanced. These peaks are relatively stronger when the cellulose chains are aligned parallel to the IR electric field at the pps polarization (see φ



= 90o case in Figure 2a). At location 5 where the degree of fiber bending is the most severe, the relative intensities of 2920 cm-1, 2944 cm-1, and 2970 cm-1 are different from the φ = 90o spectrum in Figure 2a. It is speculated that it might be related to the elastic distortion or deformation of cellulose microfibrils associated with the high degree of bending. Further details could not be determined in this study and will be the subject of a separate study in the future. It is noted that the 2915-2920 cm-1 and 2964-2970 cm-1 components are barely noticed as shoulders in the averaged spectrum. The 3320 cm-1 / 2944 cm-1 intensity ratio is also larger at the bent regions (locations 3, 4, and 5) of the fiber than the relatively straight regions (locations 1, 2, 6, 7, and 8). This must be due to the change in the cellulose chain orientation with respect to the probe beam polarizations. Again, this ratio in the SFG spectrum of location 5 is much larger than that of the φ = 90o spectrum in Figure 2a (pps polarization), supporting the argument that it could not be fully explained using the microfibril orientation change at the highly bent region. At locations 7 and 8, the 3320 cm-1 / 2944 cm-1 intensity ratio is nearly zero, which is similar to the spectrum shown for the φ = 0o case in Figure 2a. Thus, it can be said that the cellulose microfibrils in locations 7 and 8 are relatively parallel to the laser incidence plane. Overall, the data presented in Figure 4 is congruent with the fact that the microfibril angle with respect to the fiber axis varies from locations to locations along the cotton fiber.59-60 Because the SFG coherence length is larger in the forward scattering direction than the backward scattering,61-62 the SFG intensity is generally stronger in the transmission-mode detection than the reflection-mode detection. Figure 5 compares the SFG hyperspectral images and full spectra of cotton fibers collected in the reflection- and transmission-mode detections. For this sample, the transmission signal is about 100 times stronger than the reflection signal. In


Page 16 of 47

Page 17 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


addition, the maximum intensity positions are different between two detection modes of the same sample (Figures 5d-5g). This implies the average packing or orientation of cellulose microfibrils detected in the top part of the sample with a shorter coherence length are different from those detected through the entire thickness of the sample with a longer coherence length. The broadband spectra of the cotton fibers collected during the image scan are averaged and plotted in Figure 5h. The reflection spectrum looks similar to that of position 5 in Figure 4, while the transmission spectrum resembles the averaged spectrum in Figure 4. This may imply variations in the mesoscale packing or orientation of cellulose microfibrils along the fiber cross-section. The difference in the OH/CH ratio between the two spectra in Figure 5h could be due to the difference in thickness dependence of the SFG signal scattering in different detection modes.50

Figure 5. Comparison of reflection- and transmission-mode detections. The optical image in (a) shows cotton fibers immersed in D2O. The SFG detection in two modes is schematically illustrated in (b) and (c). Hyperspectral images of (d,e) 2944 cm-1 and (f,g) 3310 cm-1 SFG peaks of the sample shown in (a). The reflection and transition SFG spectra obtained by averaging the



signals obtained during the imaging scans are shown in (h). The transmission spectrum intensity is reduced by a factor of 100 to plot in the same scale. The abaxial epidermis cell walls of onion bulb are a good model system for primary cell walls.63-68 The abaxial epidermis cell is easily split open along the sidewalls when the skin is peeled from each bulb scale (Figure 6a).11 The peeled skin is the outer wall of the epidermal cell and the cytoplasm is easily washed off. The peeled cell wall is only 5-8 µm thick in the fully hydrated state, depending on scales, and 2-4 µm thick when dried.11, 69 It can be laid flat on a glass substrate (see Figures S7 and S8 in SI); thus, the entire sample can be kept at the maximum focal plane (Figure S4 in SI).


Page 18 of 47

Page 19 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. Line scan across the single cell wall of the abaxial epidermis of onion (Allium cepa) bulb in (a,c,e) fully hydrated in D2O and (b,d,f) fully dried in air. (a,b) show the optical image; (c,d) display the SFG spectra at the CH and OH stretch regions collected with the 15× objective lens; (e,f) are the SFG spectra averaged along the scanned line. The polarization combination of the probe beams was pps and the laser incidence plane is in the vertical direction of the image in (c) and (d). The SFG signals were collected in the transmission mode. Figure 6c displays SFG spectra collected during the line scan over a 150 µm distance along the long axis of a single epidermis cell wall in the fully-hydrated state (in D2O) and Figure 6e is the averaged spectrum of the entire line. The averaged spectrum clearly shows a strong peak at 2944 cm-1, which is characteristic of cellulose microfibrils fully-extended and deposited with equal probabilities for the opposite polarities along a specific assembly direction (as in the cotton fiber case; see the averaged spectrum in Figure 4). The OH/CH intensity ratio is significantly larger in the onion epidermis, compared to the secondary cell wall cotton fiber case. One of the reasons for this could be due to the lower cellulose content in the onion epidermis.11, 67

When the cellulose content is lower and thus the interfibrillar distance is larger, then the

symmetry cancellation of the OH dipoles among adjacent fibrils running in antiparallel directions is less; thus, the OH/CH intensity ratio will be larger.12 In Figure 6c, it is important to note that the measured intensities of the CH and OH stretch modes and their relative intensity (OH/CH) vary along the scan line. Based on the knowledge established from previous studies with model systems,12, 36 these variances must be due to local inhomogeneity in interfibrillar distance and/or orientation of cellulose microfibrils. Visualization of cellulose microfibrils in the same type of onion epidermis cell wall using atomic force microscopy clearly showed that cellulose microfibrils exposed at the cell wall surface are 19 ACS Paragon Plus Environment


not perfectly straight and aligned; their orientations and distances between them vary gradually.37 It is unlikely that the low intensity region is due to the lack of cellulose microfibrils in specific regions. The onion epidermis wall consists of multiple lamellae and cellulose microfibrils in one lamella are tilted by 62±16° from those in the adjacent lamella.37 Although there could be a region where cellulose microfibril density is lower than other regions of the same lamella, it is very unlikely that such regions of multiple lamellae are coincidentally positioned in the same location to give low SFG intensities. Figures 6b, 6d, and 6f show the data of an air-dried onion epidermis cell walls. Because cellulose microfibrils are laid within the plane parallel to the cell wall, the lateral dimension change of the cell wall upon drying is negligible (less than 1-2%). But, the change in thickness is significant (up to ~60%) upon drying in ambient air since the hydrated gel-like pectin matrix between lamellae is collapsed into a glassy state upon dehydration.70-71 It is remarkable that the SFG peak shape in the CH stretch mode is changed significantly upon simple air-drying; instead of the 2944 cm-1 peak, two peaks at 2920 cm-1 and 2968 cm-1 are dominant. And the OH/CH intensity ratio is much lower compared to the fully-hydrated state in D2O. Such changes are tentatively attributed to local strains imposed to the cellulose microfibrils during the collapse of the pectin matrix upon dehydration of the cell wall. Since cellulose microfibrils are interspersed in the hydrated gel-like pectin matrix, local physical strains could be induced to the cellulose microfibrils when the pectin matrix collapses into the hard glassy state. Such local strains would not be homogeneous because of the presence of load-bearing spots in the cellulose microfibril network that prevent the cell wall from creeping in the lateral dimension.72 The strained crystals of cellulose could make vibrational modes that are normally SFG-inactive in the fully-relaxed state to become SFG-active due to local strain


Page 20 of 47

Page 21 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


gradients.73 Further details in the CH mode change cannot be elucidated in this experimental study; it may require theoretical calculations to predict SFG spectra features of a strained unit cell using first-principles computational methods.34 Figures 6d and 6f show that the OH component at ≥3400 cm-1 is enhanced and becomes comparable with the 3320 cm-1 peak intensity upon drying. The higher wavenumber in the OH peaks means weaker hydrogen bonding interactions.38, 74-76 The weakly-hydrogen-bonded OH groups appearing at ≥3400 cm-1 could be assigned to the hydroxyl groups at the surface region of the cellulose crystal.35 An alternative interpretation could be the local strain of cellulose microfibrils upon dehydration of the cell wall. Such strains could result in distortion of the crystalline cellulose domains and an increase in the hydrogen bond distance between cellulose chains. This may also reduce the coherence of the OH vibration modes along the microfibril direction. These could cause the reduction of the overall intensity of the OH SFG peaks (making the OH/CH ratio small) and the relative enhancement of the weakly hydrogen-bonded OH peak at ≥3400 cm-1 compared to the strongly hydrogen-bonded peak at

Inhomogeneity of Cellulose Microfibril Assembly in Plant Cell Walls

Recommend Documents