Second Harmonic Generation Mediated by Aligned Water in Starch Granules.

Subscriber access provided by Brought to you by ST ANDREWS UNIVERSITY LIBRARY

Article

Second Harmonic Generation Mediated by Aligned Water in Starch Granules Richard Cisek, Danielle Tokarz, Serguei Krouglov, Martin Steup, Michael J. Emes, Ian J. Tetlow, and Virginijus Barzda J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/jp508751s • Publication Date (Web): 26 Nov 2014 Downloaded from http://pubs.acs.org on November 28, 2014

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 free 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 accessible to all readers and 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.

The Journal of Physical Chemistry B 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 44

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

Second Harmonic Generation Mediated by Aligned Water in Starch Granules Richard Cisek,† Danielle Tokarz,† Serguei Krouglov,† Martin Steup,§ Michael J. Emes,‡ Ian J. Tetlow,‡ and Virginijus Barzda*,† †Department of Chemical and Physical Sciences, Department of Physics, and Institute for Optical Sciences, University of Toronto, 3359 Mississauga Road North, Mississauga, ON, Canada L5L 1C6 ‡Institute of Biochemistry and Biology, Department of Plant Physiology, University of Potsdam, Karl-Liebknecht-Str. 24-25 Building 20, 14476 Potsdam, Germany §Department of Molecular and Cellular Biology, College of Biological Sciences, University of Guelph, 50 Stone Rd. E., Guelph, Ontario, Canada N1G 2W1 Abstract:

The origin of second harmonic generation (SHG) in starch granules was investigated using ab initio quantum mechanical modeling and experimentally examined using polarization-in, polarization-out (PIPO) second harmonic generation microscopy. Ab initio calculations revealed that the largest contribution to the SHG signal from A- and B-type starch crystalline allomorphs originates from the anisotropic organization of hydroxide and hydrogen bonds mediated by

1 ACS Paragon Plus Environment


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

Page 2 of 44

aligned water found within the crystals. The hypothesis was experimentally tested by imaging maize starch granules under various hydration and heat treatment conditions that alter the hydrogen bond network. The highest SHG intensity was found in fully hydrated starch granules and heat treatment diminished the SHG intensity. The SHG PIPO imaging showed that dried starch granules have a much higher nonlinear optical susceptibility component ratio than fully hydrated granules. In contrast, deuterated starch granules showed a smaller susceptibility component ratio demonstrating that SHG is highly sensitive to the organization of hydroxyl and hydrogen bond network.

The polarization SHG imaging results of potato starch granules,

representing crystalline starch allomorph B, were compared to maize starch granules representing allomorph A. The results showed that the amount of aligned water was higher in the maize granules. Nonlinear microscopy of starch granules provides evidence that varying hydration conditions leads to significant changes in the nonlinear susceptibility ratio as well as the SHG intensity, supporting the hypothesis from ab initio calculations that the dominant contribution to SHG is the ordered hydroxide and hydrogen bond network.

Introduction: Living in a variable environment, essentially all organisms are capable of forming degrading reduced carbon compounds, which in most cases are starch or glycogen. Native starch (i.e. starch granules) is a hydro-insoluble carbon store almost ubiquitously occurring in plants whereas most heterotrophic organisms use another chemically similar yet hydro-soluble polyglucan, glycogen1. In addition, starch is an important starting material for various industrial applications including biotechnology, and biofuel production2,3. For many uses of starch, granules are disintegrated by enzymatic and/or hydrothermal treatments to obtain starch-derived materials with selective


Page 3 of 44

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


properties, such as a distinct thermostability and crystalline structure that are required to obtain specific textures in food (consistency, softness, or crispness) and to design biodegradability of compounds having suitable mechanical properties4. Often susceptibility to enzymatic hydrolysis is optimized to alter nutritional quality, biodegradability, or fermentability of starch granules. In both pure and applied science, new visualization techniques of starch granules are needed to characterize various starches and also, to optimize industrial procedures. Much work has been put forth on starch over the past 5 decades showing that the properties of starch granules and their derivatives are dependent on their water content4. For instance, the dehydration of starch B granules composed of the B-type crystalline allomorph (crystal-B), which exhibit a hexagonal C6v type symmetry, induces a change to starch A granules with the more tightly packed A-type crystal structure (crystal-A), which has monoclinic B2 type symmetry5. Therefore, this demonstrates that water is a necessary component of the structure, playing a crucial role in the organization of starch granules. In starch processing, hydration is strictly controlled to obtain desired properties of starch derivatives. Therefore, a measurement technique that can determine not only the crystalline structure of starch, but also the hydration state, is highly beneficial. The unique crystalline ultrastructure of starch granules renders intense second harmonic generation (SHG) signal in a nonlinear optical microscope6. Nonlinear optical microscopy is quickly becoming a standard analytical tool for structural investigations of biological samples. Although the resolution of SHG microscopy does not match the resolution of electron or atomic force microscopy, polarization-dependent nonlinear optical microscopy can deduce molecular organization at the nanoscale, as shown in several animal and plant tissue studies including collagen7-12, muscle13-15, carrots16 and starch17-20. Nonlinear optical microscopy allows



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

Page 4 of 44

convenient sample investigation without pre-treatment or the addition of labels, permitting a quick measurement of samples in their native state, and therefore avoiding artifacts. Indeed, the technique is well suited for investigating structures in vivo, since biological samples up to 1 mm thick can be studied without mechanical sectioning21. Therefore, nonlinear optical microscopy lends itself as a convenient research tool for structural analysis of individual starch granules. In this paper, a detailed investigation of the microscopic nonlinear optical properties of starch granules is presented. Modeling using ab initio quantum mechanical calculations is performed to determine the relationship between the molecular structures of two crystalline allomorphs, crystal-A and crystal-B, and their first hyperpolarizability tensors (β), which determine their SHG properties (intensity and polarization), showing that bound water within the crystals mediates SHG. Nonlinear optical imaging of starch granules is performed demonstrating that different hydration conditions affect the SHG signal from starch. Based on both theoretical and experimental results, the origin of SHG signal from starch is attributed predominantly to the ordered hydroxide and hydrogen bonds facilitated by aligned water

bound within starch

granules. Experimental Methods: Sample preparation: Starch granules from potato and maize were studied. A small piece of fresh potato from the local market was ground using a mortar and pestle. The resulting suspension containing starch granules was diluted with distilled water (1:3 [v/v]) and was then immobilized in polyacrylamide gel. In order to analyze dehydrated potato starch, the suspension of starch granules in water was centrifuged at 5000 g for 3 min. The pellet was washed with acetone and centrifuged again at


Page 5 of 44

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


5000 g for 3 min. This step was repeated once and the pellet was air dried. The dried granules were placed between microscope cover slips for imaging. For experiments with maize, dry maize granules (wild-type CG102) were placed in distilled water and allowed to stir overnight. The granule suspension was centrifuged for 5 min at 8000 g, and the pellet was diluted in distilled water (1:3 [v/v]). The same procedure was repeated with deuterated water as well as ethanol. Dry maize granules were prepared similar to potato starch granules. Immobilization of starch granules was achieved by use of polyacrylamide gel containing 0.68 M acrylamide (Sigma Aldrich), 8.3 mM bisacrylamide (Sigma Aldrich), 0.18 mM ammonium persulfate (Sigma Aldrich), and 26 mM tetramethylethylenediamine (TEMED; Sigma Aldrich). Distilled and deuterated water (Sigma Aldrich) were used in the preparation of the gel for the maize samples in distilled and deuterated water, respectively. The suspension of the sample and polyacrylamide gel was placed in between microscope coverslips and allowed to polymerize for at least 30 minutes and less than 4 hours before performing the measurements. In another experiment, starch was soaked in ethanol or water for 4 hours and 100 µL of the suspension was placed on glass and immediately imaged without immobilization. A heating experiment was also performed where three potato starch samples immobilized in polyacrylamide gel between coverslips were placed into 65ºC water for 15, 30 and 45 s and subsequently imaged. Similarly, hydrated maize starch immobilized in polyacrylamide gel was placed into 75ºC water for 5 minutes and immediately imaged. Laser and nonlinear optical microscope system: A home-built Yb:KGW laser operating at 1028 nm wavelength with 14.3 MHz repetition rate and ~450 fs duration pulses22 was used for SHG measurements. The pulse energy was below 2



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

Page 6 of 44

nJ at the sample. The laser beam was coupled to a home-built nonlinear optical microscope, which has been described previously23,24. Briefly, galvanometric scanning mirrors (VM1000, Cambridge Technology) were used to raster scan the beam through a 0.75 numerical aperture (NA) air objective (Carl Zeiss) at speeds up to 10 frames per second. SHG was collected in the forward direction by a home-built 0.85 NA collection objective. The SHG signals were filtered using FF01-515/25 (Semrock) and BG39 (CVI Laser Optics) filters and detected with photomultiplier tubes (H5783P Hamamatsu) set for single photon counting mode. A computer counting detection card (PCI-6602, National Instruments) was used for signal acquisition. Polarization measurements with the nonlinear optical microscope: For polarization measurements of SHG signals in the nonlinear optical microscope, the polarization-in, polarization-out (PIPO) technique was used as previously described10. Briefly, the nonlinear optical microscope was modified by addition of a linear polarizer (IR 1100 BC4, Laser Components) followed by a half-wave plate (532GR-42, Comar Optics) placed immediately before the excitation objective for rotation of the incident laser polarization. To determine the polarization of the SHG signal, a linear polarizer (10LP-VIS-B, Newport) was placed after the collection objective and before the detector. A typical PIPO measurement consisted of taking 100 images at 10 emission polarization angles for each of 10 half-wave plate angles, resulting in the sampled SHG polarization orientation range of 180° and the excitation polarization orientation range of 180°. Computational calculation of the first hyperpolarizability tensor: Ab initio calculations were performed using the computational chemistry package GAMESSUS25 (Gordon Group, Purdue University) to determine β for starch granules. β calculations were performed at a wavelength of 1030 nm, using the 6-311++G**D basis set and the time6 ACS Paragon Plus Environment

Page 7 of 44

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


dependent coupled perturbed Hartree-Fock (TDHF) method26,27 on the SciNet GPC Consortium at the University of Toronto, with the following convergence parameters28: ICUT=20, ITOL=30, IN-TTYP=HONDO, NCONV=7, ATOL=1.0D-0.7, BTOL=1.0D-0.7. The molecular coordinates from published crystal structures of starches A and B29,30, respectively, were used as an input. For bond additivity modeling, it was assumed that a hydroxide bond can be described by a single β tensor component along the bond orientation, and the β of the hydroxide bond was fit to the ab initio calculated parameters for the β of water by setting βzzz ≠ 0 . The fit was in agreement with the TDHF calculation of water, validating this assumption. The importance of hydrogen positions in SHG calculations of the starch crystalline allomorphs was deduced from β calculations using different hydrogen positions. Hydrogen positions were inferred from listed hydrogen bonds in crystallographic papers29,30 as well as, by performing geometrical optimization of hydrogen positions at the B3LYP level using the 6-311++G**D basis set. Calculations of β for the glucans from crystal-A and -B was performed by bond additivity followed by full TDHF quantum mechanical calculations. While TDHF is a rigorous method for estimating molecular properties, relating alterations of molecular structure to the SHG polarization response is non-intuitive. Therefore, the β of glucans was first calculated by a simpler bond additivity method, which describes each chemical bond by a single β tensor component aligned along the direction of the bond. Bond additivity has the additional advantage that the influence of each bond, based on the bond type and orientation geometry within the molecule, can be related directly to SHG polarization response. TDHF was used for all subsequent calculations using the interacting segment method31,32 where each unique glucan from the crystals (glucan) was segmented and a hydrogen atom was added as detailed below.



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

Page 8 of 44

Crystals-A and -B were segmented into polarizable units typically containing 1-3 glucans. Susceptibility tensor values were calculated by coherent summation of the segments, accounting for relative orientations. For the segment shown in Figure 1a (glucan from a helix), the segmentation was optimally achieved by cleaving half way through each glycosidic bond, at oxygens O1 and O4, since it introduced the least amount of error compared to keeping both glycosidic bonds, or cleaving one or both. Hydrogen atoms were added to the terminal oxygens, and were oriented in the same direction as the severed O-C bonds. Further improvement of the approximation was achieved with the introduction of an interaction factor32, which reduced error in the test cases studied. The interaction factor accounts for the effect on the β due to the interaction between electronic transition densities of the segmented polarizable units in close proximity. Interaction between nearest neighbors was found to be adequate for determination of the interaction factor. Notably the interaction factor accounts for interactions between two bonded glucans of the same helix (glucan pairs), while taking care not to count twice the contributions, similar to the method performed by Tuer et al.10. The interaction factors of β were introduced to the coherent summation to approximate the full β of the investigated structures. Additionally, due to substantial differences in the glucan configurations of unique glucans, the individual polarizable units were calculated separately, as well as their correction factors. Visualization of the first hyperpolarizability tensor: Visualization of β was achieved using the unit sphere representation33. Briefly, an effective SHG dipole is defined as: r

t

β eff = β : Eˆ (θ , φ ) Eˆ (θ , φ )

(1)


Page 9 of 44

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


t where β is the first hyperpolarizability tensor, and Eˆ (θ , φ ) is the unit vector incident electric

field with polarization defined in spherical coordinates, (θ , φ ) . A unit sphere is mapped out via sampling all possible incident polarizations by (θ , φ ) . At each point on the unit sphere defined by

r (θ , φ ) the corresponding β eff glyph is plotted. Once β was obtained, the effective SHG dipoles were calculated in MATLAB (The MathWorks) and visualized in ParaView (Kitware). Theoretical Methods: The second-order nonlinear susceptibility ratio from PIPO SHG microscopy: The second-order nonlinear optical susceptibility tensor component ratio is determined from PIPO SHG measurements10,12. The laboratory Cartesian coordinate system, xyz, is defined with respect to the principal propagation direction of the scanning laser, y, where xz is the image plane. Several assumptions are made for deducing the second-order nonlinear optical susceptibility tensor component ratio, χ(2)zzz/χ(2)zxx, including: (i) cylindrical symmetry along z applies to crystal-A and -B allomorphs34, (ii) crystal-A and -B in starch are oriented with the main axis within the scan image plane, which is valid when an equatorial optical section is analyzed due to the radial 3D arangement of the starch granule, and (iii) Kleinman symmetry, holds since starch does not absorb at 1030 nm or 515 nm. These assumptions result in two unique nonzero tensor elements, χ(2)zzz and χ(2)zxx. The SHG intensity can be written as a function of the laser polarization (θ) and the orientation of the analyzer (φ ) as12:

I 2ω

 χ ( 2)  ∝  zzz cos φ cos 2 θ + sin φ sin 2θ + cos φ sin 2 θ  ( 2)  χ zxx 

2

(2)

Birefringence was neglected in this equation since it was verified that birefringence is neglegable in starch granules less than 20 µm in diameter20. 9 ACS Paragon Plus Environment


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

Page 10 of 44

Effective SHG ratio measurements near the center of nucleation in starch granules: When the laser imaging plane is oriented parallel to the equator of a starch granule, the measured χ(2)zzz/χ(2)zxx will deviate from the actual χ(2)zzz/χ(2)zxx of a single crystal because the axial point spread function will determine the range of the cylindrical axis orientations of the crystals that are tilted off the image plane as much as α = tan −1 ( y D ) where y is the position along the laser propagation direction (where y = 0 is at the intersection of the xz plane) and D is the distance from the radial origin of the granule to the investigated pixel (focal volume in the image plane). The tilted crystals will contribute to the observed second-order nonlinear susceptibility component ratio, χ(2)zzz’/χ(2)zxx’, which is related to the intrinsic ratio, χ(2)zzz/χ(2)zxx, and the effective tilt angle, α, via12: ( 2) ( 2)  2 χ zzz '  χ zzz = − 3   cos α + 3 ( 2) ( 2) '  χ zxx χ zxx 

(3)

According to equation (3), χ(2)zzz’/χ(2)zxx’ is expected to be lower near the radial origin of all starch granules, and therefore, regions near the radial origin have modified values compared to the intrinsic susceptibility ratio. The nucleation region of a starch granule (hilum), is expected to be at the origin of a radially organized structure. SHG from circularly polarized excitation: Circularly polarized laser can be readily used for SHG imaging, having the advantage that all radial orientations are sampled equally in the structure. The SHG intensity, ICP, when circular polarization of fundamental radiation is used, is a function of the tilt off the plane angle, α, between the cylindrical fibers and the image plane20, and also depends on the susceptibility component values:


Page 11 of 44

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


( ) + ( 2χ ( ) sin α + cos α ( χ ( ) − χ ( ) )) 

 ( 2) ICP ( 2ω ) ∝ E 4 cos2 α  4 χ xxz 

2

2 xxz

2

2

2 zzz

2 zxx

2

(4)

Circularly polarized laser radiation is commonly used for SHG microscopy to obtain images that are independent of linear or elliptical polarization orientation with respect to the structural orientation of the sample within the image plane. The circularly polarized fundamental radiation can also be conveniently used for the orientation independent SHG intensity comparison between different samples, as will be presented in Figure 5. Results and Discussion: The calculated first hyperpolarizabilities of individual glucan molecules: Ab initio calculations of the first hyperpolarizabilities of crystal-A and -B allomorphs were performed in order to investigate the molecular origin of SHG from starch granules. The investigation included studying the first hyperpolarizabilities of glucans from published crystal structures using: (i) the bond additivity theory to gain an intuitive understanding of nonlinear optical contributions from individual bonds of the glucan, (ii) quantum mechanical calculations using TDHF theory to determine hydroxide bond orientation effects and (iii) TDHF theory to study the glucans from crystal-A and -B allomorphs in their helical configurations and also coupling effects between the glucans. The first hyperpolarizability of a glucan: The β of a generic glucan is calculated via additivity of the first hyperpolarizabilities of individual bonds (bond additivity method), previously shown to be accurate for simple organic molecules10,35. The β values used for C-H and C-O bonds (-33.2 and 30.0 au, respectively) were deduced from TDHF calculations of water, methane, and methanol molecules, while C-C bonds were neglected due to their central symmetry. In this approximation, the hydroxide bonds were 11 ACS Paragon Plus Environment


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

Page 12 of 44

neglected since they are not determined in the crystal structures. The investigated glucan with labelled atoms is visualized in Figure 1a, while the calculated β is conveniently shown via the unit sphere representation33 in Figure 1b, where the helical axis of the fibril from which the glucans are taken is oriented along the z direction indicated also in Figure 1a.

Figure 1. The first hyperpolarizability β of a generic glucan from crystal-A and -B allomorphs calculated via bond additivity: (a) the axes and nomenclature of the atoms in a glucan, (b) the unit sphere representation of β of a generic glucan at an orientation in (xz) plane, (c) the variation of the dominant β between all the unique glucans, for crystal-A (magenta) and crystal-B (cyan). The helical axis orientation is along z and the glucan is oriented in its position within the helix. Positions of atoms were obtained from references29,30. The projections of the first hyperpolarizability unit spheres visualize one dominant β component oriented almost parallel to the x-axis (Figure 1c). This dominant β of a glucan does not contribute to the β of a helix since the imposed cylindrical symmetry around the helical axis cancels any first hyperpolarizabilities oriented perpendicular to the helix. The dominant β along the x-axis results from two unbalanced and nearly parallel bonds, H7-C6 and C3-O3 (Figure 1a), and according to the bond additivity model, they constitute 97% of the total β of a glucan. The remaining 3% of the glucan β originates from the remaining bonds due to imperfect antiparallel arrangements.


Page 13 of 44

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


Crystal-A and -B allomorphs have three and two unique glucans, respectively, and their orientation slightly varies with respect to the helical axis. Variations in the dominant hyperpolarizability orientation of all glucans are revealed in Figure 1c where magenta arrows represent the effective single β vector of each unique glucan from crystal-A, and cyan arrows represent the β vector of glucans from crystal-B, calculated analogous to Tuer et al.33. The observed variation in dominant β occurs due to a slight twisting of two bonds, C3-O3 and H7C6, from the molecular plane, while the variation in alignment within the helix between glucans in crystal-A and -B is insignificant (less than 1°). Therefore, the small structural and alignment variations between the glucans are not expected to induce a significant change in the β between crystal-A and -B. Bond additivity calculations were performed on each unique glucan of crystal-A and -B, accounting for all tensor components, followed by a cylindrically symmetric rotational averaging, in order to reveal average β parameters per glucan representing the crystals. After cylindrical averaging, two unique nonzero tensor components remain: βzzz = -2.4 au and βzxx = 2.8 au for crystal-A, and βzzz = -4.9 au and βzxx = 2.1 au for crystal-B. This means that if glucans are the only contributors to the β of starch then the measured second-order nonlinear optical susceptibility component ratio, χ(2)zzz/χ(2)zxx, for starch granules should be in the range between 2.5 and -1. However, experimental measurements found in literature have concluded that χ(2)zzz/χ(2)zxx in starch A is positive and larger than 3, performed by fitting SHG intensity data obtained via modulating the polarization of the laser19,34. Therefore, this result indicates that additional factors such as the influence of hydroxide bonds and the coupling between glucans plays a major role in the determination of the β of crystal-A and -B. TDHF calculation of the role of hydroxide bonds on the first hyperpolarizability of a glucan



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

Page 14 of 44

The orientation of hydroxide bonds is not provided in the published crystal structure of crystal-A and -B, therefore hydrogen positions were predicted and the structure was optimized via TDHF (see experimental methods). Figure 2 illustrates that the rotation of a single hydroxide bond (O2) in a glucan (compare Figure 2, a and d) changes the β values (Figure 2, b and e) including the tensor components that survive when cylindrical symmetry is applied (Figure 2, c and f). The large effect of hydroxide bonds on the β of glucans calculated by TDHF can be clearly understood from the bond additivity model, which showed that a hydroxide bond has a β magnitude of 26.6 au, as deduced from calculations of water molecules with TDHF, and therefore dominates the SHG response of a glucan. When the hydroxide bond is oriented in the direction of the helical axis then the contribution to the cylindrically averaged β of glucan can be dominant. Since each glucan has three hydroxide bonds, even partial alignment of the bonds along the helical axis will provide a high χ(2)zzz/χ(2)zxx as long as the bonds are not in a centrally symmetric orientation with each other.

Figure 2. Effect of hydroxyl orientation on the first hyperpolarizability of a glucan. Results of TDHF calculations of the β of a single glucan with two distinct positions for hydroxide orientations (a, d) highlighted by a circle around the bond. The total β tensors are shown via the unit sphere representation in (b, e), while (c, f) shows the contribution to the helix, calculated via


Page 15 of 44

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


inducing C6v symmetry on the tensor along the helical (z) axis. Images (c, f) were scaled up by 5 times indicated at top right to clearly visualize the arrows. The role of coupling between glucans on the first hyperpolarizability: Calculation of the β due to coupling between glucans was also performed using TDHF. The coupling β between two glucans C(x1,x2), can be defined as: C ( x1 , x 2 ) = H ( x1 , x 2 ) − H ( x1 ) − H ( x 2 )

(5)

where H(x1,x2) is the β of glucan x1 and x2 together, while H(x1) or H(x2) is the β of glucan x1 or x2 alone. Directly bonded glucans (Figure 3a) in a helix had a total coupling contribution (Figure 3e) oriented along the coupling bond direction, however the angle to the helical axis was quite large resulting in a reduced contribution to the tensor components of the helix (Figure 3i) after performing cylindrical averaging. The effect of coupling glucans from adjacent helices within a double helix (Figure 3b) was also investigated, revealing the total β (Figure 3f) as well as the helical contribution (Figure 3j), which had a similar magnitude as the coupling due to direct bonding (Figure 3i). The coupling due to the presence of a hydrogen bond oriented with the OHO in the downward (Figure 3c) and upward (Figure 3d) directions had a magnitude 2-5 times greater than coupling due to direct bonding shown in Figure 3a. This observation is in agreement with previous ab initio investigations that showed the β magnitude enhancement between phenol and trimethylamine due to hydrogen bonds, as well as in hydrogen bonded complexes of nitropyridines with hydrogen fluoride36,37. The hydrogen bond in Figure 3, c and d, occurs between two helices of a double helix, and has an angle of ~60º to the helical axis, which corresponds to χ(2)zzz/χ(2)zxx = 0.7 if treated like a simple dipole38. However, many other hydrogen bonds exist in crystal-A and -B such as between glucans within a single helix and also between



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

Page 16 of 44

glucans of adjacent double helices, which are hydrogen bonded via water bridges. Therefore, it is likely that these hydrogen bonds dominate the β of crystals in starch. Further investigations with molecular dynamic simulations would need to be performed in order to properly account for the contributions of different hydrogen bond configurations.

Figure 3. First hyperpolarizability due to coupling between glucans. Three examples of the first hyperpolarizabilities due to coupling are shown: bonded glucans (a, e, i), glucans from two adjacent helices (d, f, j) and hydrogen bonded glucans at two different hydrogen bond orientations (c, g, k) and (d, h, l). (a, b, c, d) shows the considered molecules, (e, f, g, h) shows the total β due to coupling and (i, j, k, l) shows only the components of the β pertaining to cylindrical symmetry. The first hyperpolarizabilities due to interactions were calculated with equation (5). The first hyperpolarizabilities of crystal-A and -B allomorphs: The β of crystal-A and -B were approximated by performing TDHF calculations using the interacting segment method, taking into account the glucans, hydroxide and hydrogen bonding networks, and coupling between bonded glucans.


Page 17 of 44

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. Hydrogen bonding networks for crystal-A and -B allomorphs with optimized hydroxide orientations. The double helical structure of (a) crystal-A and (d) crystal-B are shown29,30. The highlighted regions in (a) were extracted and the initial hydrogen positions were altered and geometrically optimized giving a number of configurations where only 2 examples are shown in (b, c). Similarly, the highlighted regions in (d) were extracted and the initial hydrogen positions were altered and geometrically optimized seen in (e, f). The hydrogen bonding network is shown with dotted lines in (b, c, e, f). The structures are oriented with the helical axis pointing upwards. The orientations of hydroxide and hydrogen bonds were estimated by selecting large portions of the crystals encompassing a continuous hydrogen bonding network (Figure 4, a and d). Different hydroxide bond orientation combinations were used for two crystal allomorphs in order to calculate theoretical β for crystal-A (Figure 4, b and c) and crystal-B (Figure 4, e and f). The orientations of hydroxide bonds at O2, O3 and O6 of glucans (Figure 1a) were the primary concern since the influence of hydroxide bond orientations, as well as the hydrogen bond orientations were found to be the largest contributors to the β of glucans in TDHF calculations. Figure 4, b and c, demonstrates 2 examples of geometrically optimized sections of the hydrogen



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

Page 18 of 44

bonding network within the crystal-A double helical structure, which gives a positive secondorder nonlinear optical susceptibility tensor component ratio. Differences between Figure 4, b and c, are largely in the positions of the hydrogen atoms with respect to two nearby oxygen atoms (Figure 4, b and c, i-vi). Similarly, Figure 4, e and f, demonstrates 2 examples of geometrically optimized sections of the hydrogen bonding network between and within the crystal-B double helical structure. Differences between Figure 4, e and f are largely in the positioning of the hydrogen atoms with respect to two nearby oxygen atoms (Figure 4, e and f, iiv) and the 3D position of hydrogen atoms in space (Figure 4, e and f, v). Table 1 presents the calculated cylindrical χ(2) components of crystal-A and -B at the two different bond configurations (Figure 4, b, c, e and f). By applying cylindrical averaging to TDHF calculated β values, the χ(2) values representing the crystal-A and -B helices were obtained (Table 1). Table 1. Second-order nonlinear optical susceptibilities of crystal-A and -B allomorphs. χ(2) calculations were performed on crystal-A and -B glucans with two different hydrogen positions. For crystal-A, the rows A(b) and A(c) correspond to hydrogen bond positions in Figure 4 b, c. For crystal-B, the rows B(e) and B(f) correspond to hydrogen bond positions in Figure 4 e, f. Calculations were performed with TDHF using the interacting segment model and scaled per glucan. Values are expressed in atomic units (au). The summed susceptibility component ratio, χ(2)zzz/ χ(2)zxx, was calculated by summing the individual tensor components of glucans, hydrogen bonds and coupling. Starch A(b) A(c) B(e) B(f)

Glucans χ(2)zzz χ(2)zxx 10.4 7.7 -4.6 5.4 -7.5 -1.4 -20.3 -1.1

Hydrogen Bonds χ(2)zzz χ(2)zxx -28.6 -16.9 11.2 -3.5 -11.9 0.1 55.8 13.3

χ(2)zzz 3.4 -2.7 0.3 -8.0

Coupling χ(2)zxx -1.1 -1.4 0.3 -0.6

Totals χ(2)zzz/χ(2)zxx 1.4 9.0 21.3 2.4


Page 19 of 44

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


Table 1 shows results from TDHF calculations for crystal-A and -B. The calculations were performed using the interacting segment model, where each three and two bonded glucans, respectively for crystal-A and -B, was calculated accounting for hydrogen bonds and coupling as described in experimental methods, and were subsequently averaged with the cylindrical symmetry operation, and scaled per glucan. The ordering of hydroxide bonds via different connected molecular networks in crystal-A and -B allomorphs produces different χ(2) values for each combination of hydroxide positions. Table 1 shows that hydrogen bonds have the largest contribution to the χ(2) of crystal-A and -B helices. The fact that the hydrogen bond is formed from the hydrogen atom on the right glucan in Figure 4b, i, while the hydrogen bond is formed from the hydrogen atom on the left glucan in Figure 4c, i, causes each hydrogen bond in the hydrogen bond network to differ between Figure 4b, i-vi, and Figure 4c, i-vi. This resulted in an overall difference of 39.8 au for χ(2)zzz and 13.4 au for χ(2)zxx when only considering the effect of the hydrogen bond orientations (Table 1). Similarly, for crystal-B, the hydrogen bond is formed from the hydrogen atom on the left glucan in Figure 4e, i, while the hydrogen bond is formed from the hydrogen atom on the right glucan in Figure 4f, i. This causes each hydrogen bond in the hydrogen bond network to differ between Figure 4e, i-v, and Figure 4f, i-v. The different orientations of the hydrogen bond network resulted in an overall difference of 67.7 au for χ(2)zzz and 13.2 au for χ(2)zxx when considering only the effect of the hydrogen bonds (Table 1). Since the positioning of a single hydrogen bond causes an overall change in the orientation of hydroxide bonds and the positioning of other hydrogen bonds, it is difficult to predict how χ(2)zzz and χ(2)zxx change with slight differences amongst a single hydrogen bond orientation. Other possible arrangements of the hydrogen bonding network could be determined or calculated from molecular dynamic simulations, leading to different χ(2)zzz/χ(2)zxx including negative values,



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

Page 20 of 44

however experimental measurements showed only positive values of the ratio. In general, this data suggests that the cylindrically averaged χ(2) component values of both crystal-A and -B allomorphs are dominated by hydroxide bonds at specific orientations and therefore, the origin of SHG signal from crystals in starch is attributed predominantly to anisotropically oriented bound water present in crystal-A and -B. In order to determine whether this conclusion is valid, SHG intensity and χ(2)zzz/χ(2)zxx was measured experimentally in starch granules at different hydration conditions. SHG intensity from starch granules at different hydration conditions: SHG of starch granules from maize and potato, containing crystal-A and -B allomorphs, respectively, were measured at two different hydration conditions (see Figure 5). For SHG intensity comparisons, circularly polarized excitation was used to obtain signal from the entire granule (see experimental methods). Equation (4) shows the relation between SHG intensity and the tensor values, χ(2)zzz and χ(2)zxx. The dried granules of both maize and potato starch (Figure 5b and d, respectively), are about half as intense as their hydrated counterparts (Figure 5a and c, respectively). The role of scattering of the dry starch granule in air as compared to the hydrated starch granule in water was estimated via a Fresnel calculation of the reflection of a Gaussian beam near the edge of a 20 µm diameter starch granule, resulting in maximum of 20% SHG intensity difference due to scattering. The role of scattering was experimentally verified by measuring the SHG intensity for dry potato starch granules suspended in ethanol for 4 hours. Ethanol has a similar index of refraction as water (1.33 versus 1.36), therefore, scattering should be similar, while the hydrogen bond network created by ethanol is likely different than water. The SHG signal of starches in ethanol was 30% lower than in hydrated starch granules (not shown), while dry starches exhibited 50% less intensity than hydrated granules, providing


Page 21 of 44

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


evidence that the difference in SHG intensity between hydrated and dry starch granules cannot be explained by scattering alone. The crystalline order of starch granules is increased during hydration39, which allows the formation of more hydroxide and hydrogen bonds corroborating the ab initio results. In addition, the hydrated maize (Figure 5a) had a slightly higher SHG intensity (20% more) than potato starch granules (Figure 5c). Since the crystallinity of maize and potato granules is similar40 according to acid hydrolysis, X-ray diffraction and

13

C-NMR, it

suggests that slightly more aligned water is contained in the crystalline structure of the A allomorph. The crystallographic content of water is markedly higher in the B-type allomorph, indicating that only part of the water content in crystal-B is aligned that gives rise to the SHG signal. More detailed analysis of the SHG signal in starch under different hydration conditions is conducted with polarization-dependent SHG microscopy.

Figure 5. SHG intensity analysis of starch granules. The SHG intensities of maize (a, b) and potato starch granules (c, d) were studied using different hydration conditions including hydrated in water (a, c) and air dried (b, d). The circular laser polarization is indicated at the top righthand corner of a SHG image, while the scale bars on the lower right corner represent 50 µm. The 21 ACS Paragon Plus Environment


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

Page 22 of 44

intensity scale bar represents the photon counts in an individual pixel and has a maximum of 500 for all images. Second-order nonlinear susceptibilities of starch granules at various hydration conditions: PIPO SHG microscopy was performed on wild-type (WT) maize starch granules at different hydration conditions and compared to potato starch. Figure 6 shows PIPO analysis of typical WT maize granules which have been hydrated (a-d), deuterated (e-h) and dried (i-l). The starch granules are visualized with SHG contrast by adding images of all the different polarizations (Figure 6, a, e and i).

Figure 6. Polarization analysis of starch granules. Maize starch granules were studied under different hydration conditions including hydrated in (a-d) water, (e-h) treated in deuterated water and (i-l) air dried. Maize starch was also compared to a fresh potato starch granule (m-p). (a, e, i, m) SHG intensity images of the starch granules obtained by a sum of the images recorded at various polarizations where the number on the top left corner of the SHG images represents the typical SHG intensity of the starch granules in photon counts per pixel, and the scale bars represent 10 µm. (b, f, j, n) Fitted χ(2)zzz/χ(2)zxx values for each pixel in (a, e, i, m) represented with


Page 23 of 44

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


color where blue depicts a χ(2)zzz/χ(2)zxx value of 3 and red represents a χ(2)zzz/χ(2)zxx value of 8. (c, g, k, o) The χ(2)zzz/χ(2)zxx value occurrence histograms of the fitted pixels in (b, f, j, n). The fitted orientation of the crystal axis for each sixth pixel is shown in (d, h, l, p). Fitting of starch granules polarization data was performed using equation (2), assuming that within the laser focal volume the different crystalline regions emit SHG which coherently sum19 rather than undergoing incoherent summation due to hyper-Rayleigh scattering. The fitting of polarization data for each pixel had a high goodness of fit parameter (R2>0.9) and therefore, the coherent scattering model was deemed valid. The fits of each pixel of the hydrated starch granule image (Figure 6a) revealed χ(2)zzz/χ(2)zxx values which are visualized with the corresponding color in Figure 6b. This shows that the starch granule inner core has a lower χ(2)zzz/χ(2)zxx value near 3, while the periphery has a higher χ(2)zzz/χ(2)zxx value, which was also previously observed41. The origin of the variation in χ(2)zzz/χ(2)zxx is attributed to the larger spread of helix tilt out of the image plane due to the radial arrangement of the starch granule. Therefore, during analysis of the χ(2)zzz/χ(2)zxx parameter of starch granules, the regions located more than ~4 µm from the radial origin of the starch granule approach closer to the ratio values of the crystalline helices without tilt. Therefore, starch granules of similar size with diameters larger than ~12 µm were chosen for this analysis. Figure 6c shows the distribution histogram of χ(2)zzz/χ(2)zxx for the fitted pixels in Figure 6b. From 19 starch granules, the mean χ(2)zzz/χ(2)zxx was found to be 4.5 ± 0.1 with a width of 1.3 ± 0.1. The fitted orientation of the cylindrical axis in each pixel is shown in Figure 6d, visualizing the radial structure of the starch granule12. The experimental values for χ(2)zzz/χ(2)zxx of maize granules obtained with PIPO SHG microscopy falls into the range of possible values obtained theoretically by TDHF calculations for crystal-A. Similarly, the experimental χ(2)zzz/χ(2)zxx values



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

Page 24 of 44

of potato granules (Figure 6o) was also found to fall into the range of theoretically calculated values for crystal-B. Several assumptions are used when relating the hyperpolarizability calculated by TDHF to the measured nonlinear susceptibility. The local birefringence of starch granules is neglected, however, it plays a role in the relation between the Lorentz local field affecting the hyperpolarizability and the Maxwell macroscopic field of the susceptibility42, resulting in a potential error of ~3.1% in χ(2)zzz/χ(2)zxx, estimated using the local refractive index value of 1.49 and a birefringence43 of 0.015 at both SHG and laser wavelengths. Additionally it is assumed that the molecules stay in their crystal registers, neglecting Brownian motion during the PIPO SHG measurement. Since TDHF calculations of the χ(2) of both crystal-A and -B allomorphs suggest that SHG signal from starch granules is influenced by bound water, the ratio χ(2)zzz/χ(2)zxx was studied for starch granules in varying hydration conditions. Starch granules treated with deuterated water exhibited a different morphology and different PIPO SHG properties compared to the starch granules hydrated in water. The exchange between hydrogens and deuterons in water and hydroxides of glucans in starch is exothermic resulting in efficient molecular exchange44,45,46. The typical SHG image of WT maize starch granules in D2O (Figure 6e) demonstrated an enlarged central region that does not exhibit SHG compared to granules in water, while the χ(2)zzz/χ(2)zxx values were found to be significantly smaller (Figure 6, f and g). Of the 23 granules investigated, the mean χ(2)zzz/χ(2)zxx was found to be 3.7 ± 0.1 with a width of 0.8 ± 0.1. The smaller SHG parameter is attributed to a reorganization of the hydrogen bonding network in starch granules which contributes largely to the SHG signal. Since the angle between deuterons in D2O is 103.0° as compared with hydrogens in water (101.4°)47, an overall decrease in long range alignment of D-O bonds with the crystal axis and in turn lowering of χ(2)zzz/χ(2)zxx is


Page 25 of 44

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


expected. In amylose gels, higher viscosity is observed upon deuteron exchange showing that hydrogen bonds after deuteration get stronger and affect the bonding structure46. Dried starch granules were found to have much larger χ(2)zzz/χ(2)zxx values than the deuterated and the hydrated granules. As well, the distribution of χ(2)zzz/χ(2)zxx was also found to be significantly broader (Figure 6, j and k). From 5 dried starch granules, the average χ(2)zzz/χ(2)zxx was found to be 5.7 ± 0.5 with a width of 4.5 ± 0.6. The increased average χ(2)zzz/χ(2)zxx for dried starch granules indicates that the loss of water selectively retains the hydroxide and hydrogen bonds oriented closer to the helical axis of the crystal domains, while higher mobility water that assumes larger orientation distribution from the cylindrical axis, is eliminated first during the drying process. The increased width of the χ(2)zzz/χ(2)zxx distribution in dried granules suggests increased variation of the crystalline structure (amount of hydration) between the different domains in the granule. In the preceding analysis, the polarization effects of scattering of dry starch granules was neglected which is justified since excitation with s and p polarization components have an estimated maximum transmission difference of only ~3.5% near the starch granule edge. The TDHF calculations of different hydroxide and hydrogen bond orientations in starch A (Figure 4 and Table 1) are in rough agreement with the experimental data showing that large variations of the χ(2)zzz/χ(2)zxx are indeed possible when hydration conditions are altered. The example networks shown in Figure 4c gives a χ(2)zzz/χ(2)zxx of 9 which roughly corresponds to the dried granules case. The experimental value of the hydrated maize granules is only achieved when averaging between the cases seen in Figure 4 a and c, and likely requires molecular dynamics simulations to account for the dynamic interactions between water molecules. Starch granules type A versus B:



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

Page 26 of 44

The difference in the χ(2)zzz/χ(2)zxx of hydrated maize starch (starch A, Figure 6, a-d) was compared to fresh potato starch (starch B, Figure 6, m-p) to investigate the difference in crystalline allomorphs A and B. The typical χ(2)zzz/χ(2)zxx for maize starch was found to be 4.7 ± 0.1, and 5.6 ± 0.2 for potato starch (results were stated as: average ± standard error, where 14 maize and 8 potato granules with diameters 12-20 µm were investigated). The difference in χ(2)zzz/χ(2)zxx values is statistically significant according to the t-test with a confidence level >99%, and the data is in agreement with the ab initio calculations which show that starch B can have higher χ(2)zzz/χ(2)zxx (see Table 1). While the variation in the helical structure of the allomorphs is small where the helical pitch only varies by 1.8º as obtained from the published crystal structures30,48, starch B contains substantially more water (27%, compared with 7% in the A allomorph of fully hydrated starches). However, the SHG intensity of hydrated maize and potato is similar (Figure 5, a, c), showing that large amount of water is not aligned within potato granules. In addition, the χ(2)zzz/χ(2)zxx in potato granules is higher, leading to the higher SHG intensity according to equation (4) (for the same concentration of aligned water), which is not the case for the demonstrated results in Figure 5. These two effects correspond to potato starch granules having much lower water alignment than maize granules. Thermally induced loss of SHG in starch granules: In order to further demonstrate that the SHG signal generated from starch granules is mediated by bound water, starch granules were subjected to heat treatment, which is known to break hydrogen bonds. The crystallinity of potato starch granules decreases during granule heating starting from 60ºC as measured by X-ray diffraction experiments49. Therefore, three samples of starch granules from the same potato were embedded in polyacrylamide gel between microscope


Page 27 of 44

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


coverslips, and immersed in 65ºC water for 15 s, 30 s and 45 s, respectively, followed by SHG imaging using circularly polarized light. The heat treatment resulted in significant SHG signal decrease with increased incubation time (Figure 7, a, b and c). The SHG signal and structure of starch granules immersed in hot water for 15 s did not significantly change (Figure 7a), however, 30 s immersion resulted in a decrease in SHG intensity accompanied by the emergence of channels or cracks (Figure 7b). Heat treatment of 45 s often completely destroyed the granules, but some remained with low SHG intensity (Figure 7c). Similar observations were seen with maize starch at 75ºC (data not shown).

Figure 7. SHG images of heated potato starch granules. Potato starch granules immobilized in polyacrylamide gel on a microscope cover slip were heat-treated in 65ºC water for 0 s, 30 s and 45 s, resulting in typical SHG images (a), (b) and (c), respectively. The circular laser polarization is demonstrated at the top right-hand corner of each SHG image, while the scale bars on the lower right corner represent 5 µm. The loss of SHG due to heating is explained using results from the ab initio calculations. According to the ab initio calculations, the hydroxide and hydrogen bonds dominate the SHG, at the phase transition temperature the heating breaks hydrogen bonds50, which in turn causes loss of SHG, and the accompanying loss of radial order. Conclusions: The ab initio calculations indicate that SHG from crystal-A and -B originates predominantly from the anisotropic organization of hydroxide and hydrogen bonds of the glucans and water 27 ACS Paragon Plus Environment


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

Page 28 of 44

molecules. Experimental evidence shows that starch granules treated under different hydration conditions exhibit significantly different SHG intensities and χ(2)zzz/χ(2)zxx values. Furthermore, a variation in χ(2)zzz/χ(2)zxx is observed in different allomorphs of starch granules. Also, heat treatment was found to affect the SHG intensity. The experimental data are in line with the ab initio TDHF calculations showing that ordered hydroxide and hydrogen bonds, which are mediated by water in the crystalline starch, are responsible for SHG emission in starch granules. Therefore, SHG microscopy can be utilized to measure hydration of starch granules. In addition, the investigation demonstrates that PIPO SHG microscopy is sensitive to the organization of water and therefore can be applicable to study other systems containing aligned water, providing a sensitive imaging technique.

*Mail: DV 4059A 3359 Mississauga Rd. N., Mississauga, ON, Canada L5L 1C6. Tel.: +1 905 828 3821 Fax: +1 905 828 5425 E-mail: [email protected] This research was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). Computations were performed on the GPC supercomputer at the SciNet HPC Consortium. SciNet is funded by: the Canada Foundation for Innovation under the auspices of Compute Canada; the Government of Ontario; Ontario Research Fund - Research Excellence; and the University of Toronto. SHG, second harmonic generation; PIPO, polarization-in, polarization-out; 3D, threedimensional. References:


Page 29 of 44

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


(1)

Fettke, J.; Hejazi, M.; Smirnova, J.; Hochel, E.; Stage, M.; Steup, M. Eukaryotic Starch

Degradation: Integration of Plastidial and Cytosolic Pathways. J. Exp. Bot., 2009, 60, 2907-2922. (2)

Jane, J.-l. 2006. Current Understanding on Starch Granule Structures. J. Appl. Glycosci.,

2006, 53, 205-213. (3)

Farris, P. L. Economic Growth and Organization of the US Corn Starch Industry. In

Starch Chemistry and Technology; Whistler, R. L.; BeMiller, J. N.; Paschall, E. F., Eds.; Academic Press: Burlington, 2009; Vol. 1, pp 11-21. (4)

Buléon, A.; Veronese, G.; J. L. Putaux, J. L. 2007. Self-Association and Crystallization

of Amylose. Aust. J. Chem., 2007, 60, 706-718. (5)

Vermeylen, R.; Derycke, V.; Delcour, J. A.; Goderis, B.; Reynaers, H.; Koch, M. H. J.

Structural Transformations During Gelatinization of Starches in Limited Water: Combined Wide- and Small-Angle X-Ray Scattering Study. Biomacromolecules, 2006, 7, 1231-1238. (6)

Mizutani, G.; Sonoda, Y.; Sano, H.; Sakamoto, M.; Takahashi, T.; Ushioda, S. Detection

of Starch Granules in a Living Plant by Optical Second Harmonic Microscopy. J. Lumin., 2000, 87-9, 824-826. (7)

Freund, I.; Deutsch, M.; Sprecher, A. Connective-Tissue Polarity - Optical 2nd-Harmonic

Microscopy, Crossed-beam Summation, and Small-angle Scattering in Rat-Tail Tendon. Biophys. J., 1986, 50, 693-712. (8)

Stoller, P.; Reiser, K. M.; Celliers, P. M.; Rubenchik, A. M. Polarization-Modulated

Second Harmonic Generation in Collagen. Biophys. J., 2002, 82, 3330-3342.



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

(9)

Page 30 of 44

Campagnola, P. J.; Millard, A. C.; Terasaki, M.; Hoppe, P. E.; Malone, C. J.; Mohler, W.

A. Three-Dimensional High-Resolution Second-harmonic Generation Imaging of Endogenous Structural Proteins in Biological Tissues. Biophys. J., 2002, 82, 493-508. (10) Tuer, A. E.; Krouglov, S.; Prent, N.; Cisek, R.; Sandkuijl, D.; Yasufuku, K.; Wilson, B. C.; Barzda, V. Nonlinear Optical Properties of Type I Collagen Fibers Studied by Polarization Dependent Second Harmonic Generation Microscopy. J. Phys. Chem. B, 2011, 115, 1275912769. (11) Chen, X. Y.; Nadiarynkh, O.; Plotnikov, S.; Campagnola, P. J. Second Harmonic Generation Microscopy for Quantitative Analysis of Collagen Fibrillar Structure. Nat. Protoc., 2012, 7, 654-669. (12) Tuer, A. E.; Akens, M. K.; Krouglov, S.; Sandkuijl, D.; Wilson, B. C.; Whyne, C. M.; Barzda, V. Hierarchical Model of Fibrillar Collagen Organization for Interpreting the Secondorder Susceptibility Tensors in Biological Tissue. Biophys. J., 2012, 103, 2093-2105. (13) Chu, S. W.; Chen, S. Y.; Chern, G. W.; Tsai, T. H.; Chen, Y. C.; Lin, B. L.; Sun, C. K. Studies of χ(2)/χ(3) Tensors in Submicron-scaled Bio-tissues by Polarization Harmonics Optical Microscopy. Biophys. J., 2004, 86, 3914-3922. (14) Plotnikov, S. V.; Millard, A. C.; Campagnola, P. J.; Mohler, W. A. Characterization of the Myosin-based Source for Second-harmonic Generation from Muscle Sarcomeres. Biophys. J., 2006, 90, 693-703. (15) Samim, M.; Prent, N.; Dicenzo, D.; Stewart, B.; Barzda, V. Second Harmonic Generation Polarization Properties of Myofilaments. J. Biomed. Opt., 2014, 19, 0560051-0560056.


Page 31 of 44

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


(16) Tokarz, D.; Cisek, R.; Krouglov, S.; Kontenis, L.; Fekl, U.; Barzda, V. Molecular Organization of Crystalline Beta-Carotene in Carrots Determined with Polarization-dependent Second and Third Harmonic Generation Microscopy. J. Phys. Chem. B, 2014, 118, 3814-3822. (17) Chu, S. W.; Chen, I. H.; Liu, T. M.; Chen, P. C.; Sun, C. K.; Lin, B. L. Multimodal Nonlinear Spectral Microscopy Based on a Femtosecond Cr:Forsterite Laser. Opt. Lett., 2001, 26, 1909-1911. (18) Cisek, R.; Spencer, L.; Prent, N.; Zigmantas, D.; Espie, G.; Barzda, V. Optical Microscopy in Photosynthesis. Photosynth. Res., 2009, 102, 111-141. (19) Psilodimitrakopoulos, S., Amat-Roldan, I.; Loza-Alvarez, P.; Artigas, D. Estimating the Helical Pitch Angle of Amylopectin in Starch Using Polarization Second Harmonic Generation Microscopy. J. Opt., 2010, 12, 0840071-0840076. (20) Cisek, R. Structural Investigation of Biological and Semiconductor Nanostructures with Nonlinear Multicontrast Microscopy. Ph.D. Thesis, University of Toronto, December 2013. (21) Hsieh, C. S.; Chen, S. U.; Lee, Y. W.; Yang, Y. S.; Sun, C. K. Higher Harmonic Generation Microscopy of In Vitro Cultured Mammal Oocytes and Embryos. Opt. Express, 2008, 16, 11574-11588. (22) Major, A.; Cisek, R.; Sandkuijl, D.; Barzda, V. Femtosecond Yb:KGd(WO4)(2) Laser with > 100 nJ of Pulse Energy. Laser Phys. Lett., 2009, 6, 272-274. (23) Greenhalgh, C.; Prent, N.; Green, C.; Cisek, R.; Major, A.; Stewart, B.; Barzda, V. Influence of Semicrystalline Order on the Second-Harmonic Generation Efficiency in the Anisotropic Bands of Myocytes. Appl. Optics, 2007, 46, 1852-1859. 31 ACS Paragon Plus Environment


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

Page 32 of 44

(24) Carriles, R.; Schafer, D. N.; Sheetz, K. E.; Field, J. J.; Cisek, R.; Barzda, V.; Sylvester, A. W.; Squier, J. A. Imaging Techniques for Harmonic and Multiphoton Absorption Fluorescence Microscopy. Rev. Sci. Instrum., 2009, 80, 10364-10371. (25) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Kosecki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J. et al. General Atomic and Molecular Electronic Structure System. J. Comput. Chem., 1993, 14, 1347-1363. (26) Davidson, E. R.; Feller, D. Basis Set Selection for Molecular Calculations. Chem. Rev., 1986, 86, 681-696. (27) Karna, S. P.; Dupuis, M. Frequency-dependent Nonlinear Optical-properties of Molecules - Formulation and Implementation in the HONDO Program. J. Comput. Chem., 1991, 12, 487-504. (28) Loken, C.; Gruner, D.; Groer, L.; Peltier, R.; Bunn, N.; Craig, M.; Henriques, T.; Dempsey, J.; Yu, C-H.; Chen, J. et al. SciNet: Lessons Learned from Building a Power-Efficient Top-20 System and Data Centre. J. Phys.: Conf. Ser., 2010, 256, 0120261-01202635. (29) Imberty, A.; Perez, S. A Revisit to the 3-Dimensional Structure of B-Type Starch. Biopolymers, 1988, 27, 1205-1221. (30) Popov, D.; Buleon, A.; Burghammer, M.; Chanzy, H.; Montesanti, N.; Putaux, J. L.; Potocki-Veronese, G.; Riekel, C. Crystal Structure of A-Amylose: A Revisit from Synchrotron Microdiffraction Analysis of Single Crystals. Macromolecules, 2009, 42, 1167-1174.


Page 33 of 44

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


(31) Miller, C. K.; Ward, J. F. Measurements of Nonlinear Optical Polarizabilities for Some Halogenated Methanes: The Role of Bond-Bond Interactions. Phys. Rev. A, 1977, 16, 11791185. (32) Sundberg, K. R. A Group-Dipole Interaction Model of the Molecular Polarizability and the Molecular First and Second Hyperpolarizabilities. J. Chem. Phys., 1977, 66, 114-118. (33) Tuer, A.; Krouglov, S.; Cisek, R.; Tokarz, D.; Barzda, V. Three-Dimensional Visualization of the First Hyperpolarizability Tensor. J. Comput. Chem., 2010, 32, 1128-1134. (34) Psilodimitrakopoulos, S.; Amat-Roldan, I.; Artigas, D.; Loza-Alvarez, P. ThreeDimensional Polarization Second Harmonic Generation (3D-PSHG) Imaging: the Effect of the Tilted-off the Plane SHG Active Structures. Proc. SPIE, 2011, 7903, 79030H1-79030H8. (35) Levine, B. F.; Bethea, C. G. Second and Third Order Hyperpolarizabilities of Organic Molecules. J. Chem. Phys., 1975, 63, 2666-2682. (36) Dopieralski, P.; Panek, J.; Mierzwicki, K.; Latajka, Z.; Ratajczak, H.; Barnes, A. Theoretical Study on the Polarizability and Hyperpolarizability of Hydrogen Bonded Complexes of Nitropyridines with Hydrogen Fluoride. J. Mol. Struc-THEOCHEM, 2009, 916, 72-75. (37) Latajka, Z.; Gajewski, G.; Barnes, A. J.; Xue, D.; Ratajczak, H. Hyperpolarizabilities of Some Model Hydrogen-Bonded Complexes: PM3 and ab initio Studies. J. Mol. Struct., 2009, 28, 121-124. (38) Tiaho, F.; Recher, G.; Rouede, D. Estimation of Helical Angles of Myosin and Collagen by Second Harmonic Generation Imaging Microscopy. Opt. Express, 2007, 15, 12286-12295.



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

Page 34 of 44

(39) Cheetham, N.; Tao, L. Variation in Crystalline Type with Amylose Content in Maize Starch Granules: An X-ray Powder Diffraction Study. Carbohydr. Polym., 1998, 36, 277-284. (40) Buleon, A.; Colonna, P.; Planchot, V.; Ball, S. Starch Granules: Structure and Biosynthesis. Int. J. Biol. Macromol., 1998. 23. 85-112. (41) Psilodimitrakopoulos, S.; Amat-Roldan, I.; Loza-Alvarez, P.; Artigas, D. Effect of Molecular Organization on the Image Histograms of Polarization SHG Microscopy. Biomed. Opt. Express, 2012, 3, 2681-2693. (42) Boyd, R. W. Nonlinear Optics; Academic Press: Amsterdam, 2008. (43) Frey-Wyssling, A. On the Molecular Structure of Starch Granules. Am. J. Bot., 1969, 56, 696-701. (44) Taylor, N. W.; Zobel, H. F.; White, M.; Senti, F. R. Deuterium Exchange in Starches and Amylose. J. Phys. Chem., 1961, 65, 1816-1820. (45) Hennig, H.-J. Accessibilität der Hydroxylgruppen in Stärke für Deuteriumoxid. Starch/Stärke, 1977, 29, 289-290. (46) Tomasik, P.; Schilling, C. H. Chemical Modification of Starch. In Advances in Carbohydrate Chemistry and Biochemistry; Horton, D., Ed.; Academic Press: San Diego, 2004; Vol. 59, pp 175-403. (47) Soper, A. K.; Benmore, C. J. Quantum Differences Between Heavy and Light Water. Phys. Rev. Lett., 2008, 101, 0655021-0655024.


Page 35 of 44

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


(48) Imberty, A.; Chanzy, H.; Perez, S.; Buleon, A.; Tran, V. The Double-Helical Nature of the Crystalline Part of A-Starch. J. Mol. Biol., 1988, 201, 365-378. (49) Le Bail, P.; Bizot, H.; Ollivon, M,; Keller, G.; Bourgaux, C.; Buleon, A. Monitoring the Crystallization of Amylose-lipid Complexes During Maize Starch Melting by Synchrotron XRay Diffraction. Biopolymers, 1999, 50, 99-110. (50) Hennig, H. J.; Scholz, P. Nuclear Magnetic-Resonance Studies of Behavior of Adsorbed Water-Molecules in Native Starch. Z. Phys. Chem. Neue. Fol., 1976, 103, 31-44.



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

Page 36 of 44

Table of Contents Image:


Page 37 of 44

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


82x30mm (300 x 300 DPI)

ACS Paragon Plus Environment


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

82x59mm (300 x 300 DPI)


Page 38 of 44

Page 39 of 44

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


82x55mm (300 x 300 DPI)



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

82x65mm (300 x 300 DPI)


Page 40 of 44

Page 41 of 44

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


82x82mm (300 x 300 DPI)



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

82x69mm (300 x 300 DPI)


Page 42 of 44

Page 43 of 44

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


27x8mm (300 x 300 DPI)



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

50x50mm (300 x 300 DPI)


Page 44 of 44

Three-dimensional structural imaging of starch granules by second-harmonic generation circular dichroism.

NIR-triggered drug delivery by collagen-mediated second harmonic generation.

Second harmonic generation microscopy in zebrafish.

Investigating starch gelatinization through Stokes vector resolved second harmonic generation microscopy.

Corrigendum: Investigating starch gelatinization through Stokes vector resolved second harmonic generation microscopy.

Thermal characteristics of second harmonic generation by phase matched calorimetry.

Second-harmonic generation imaging of cancer.

Unified approach to Čerenkov second harmonic generation.

Monitoring membrane potential with second-harmonic generation.

THz Electric Field-Induced Second Harmonic Generation in Inorganic Ferroelectric.

Microscopic approach to second harmonic generation in quantum cascade lasers.

Giant Optical Second Harmonic Generation in Two-Dimensional Multiferroics.

Second-harmonic generation in AlGaAs microdisks in the telecom range.

Dynamic control of light beams in second harmonic generation.

Measuring microtubule polarity in spindles with second-harmonic generation.

Multipolar second harmonic generation in a symmetric nonlinear metamaterial.

Extraordinary Second Harmonic Generation in tungsten disulfide monolayers.

Second harmonic generation correlation spectroscopy for single molecule experiments.

Second harmonic generation on self-assembled tilted gold nanowires.

Surface plasmon-enhanced transverse magnetic second-harmonic generation.

Fiber-based dual-focus time-demultiplexed second harmonic generation microscopy.

Automated cardiac sarcomere analysis from second harmonic generation images.

Application of quantitative second-harmonic generation microscopy to dynamic conditions.

Precise Determination of the Crystallographic Orientations in Single ZnS Nanowires by Second-Harmonic Generation Microscopy.