Available via license: CC BY-SA 4.0
Content may be subject to copyright.
AN OPEN-CHANEL MICROFLUIDIC MEMBRANE DEVICE FOR IN SITU
HYPERSPECTRAL MAPPING OF ENZYMATIC CELLULOSE HYDROLYSIS
H.Y.N. Holman1*, W. Zhao1, J.D. Nill2, L. Chen1, S.R. Narayansamy1, and T. Jeoh2
1Molecular Biophysics and Integrated Bioimaging, Lawrence Berkeley National Laboratory, California,
2 Biological and Agricultural Engineering, University of California, Davis, California
ABSTRACT
Synchrotron infrared hyperspectral microscopy is a label-free and non-invasive technique well suited for imaging
of chemical events in situ. It can track the spatial and temporal distributions of molecules of interests in a specimen
in its native state by the molecule’s characteristic vibrational modes. Despite tremendous progress made in recent
years, IR hyperspectral imaging of chemical events in biomaterials in liquids remains challenging because of the
demanding requirements on environmental control and strong infrared absorption of water1-2. Here we report a
multi-phase capillary-driven membrane device for label-free and real-time investigation of enzymatic
deconstruction of algal cellulose purified from Cladophora aegagropila.
KEYWORDS: Cellulose hydrolysis, infrared spectral imaging, FTIR
INTRODUCTION
Commercialization of advanced biofuels hinges on efficient hydrolysis to release glucose from the cellulose
fraction of lignocellulosic plant biomass which is then fermented to bioethanol or other biofuels and bioproducts.
Cellulose functions as a structural polysaccharide in plant cell walls and resists enzymatic hydrolysis by packing
into highly ordered (crystalline) fibrils that limits enzyme access to the glycosidic bonds (C ̶ O ̶ C). A fundamental
understanding of the reaction mechanisms by which cellulase hydrolyzes cellulose remain unsolved3, due in part to
limitations of imaging capabilities to capture at micro-to-nano-scale structural and chemical information in real-
time and in situ. Herein, we develop a novel integrated open-channel platform for label-free tracking of the location
and kinetics of cellulose hydrolysis. Two key features include (1) a multi-layer multi-phase open-channel membrane
device that controls moisture, temperature, and enzyme-solution flow, (2) long-term continuous measurements of
C-O stretches of the glucose residues and the C-O-C of the cellulose. As a proof-of-concept for profiling cellulose
deconstruction at a molecular-structural level, we demonstrated the temporally- and spatially-resolved chemical
modification of cellulose imparted by a cellobiohydrolase enzyme Cel7A.
EXPERIMENTAL
The polydiemthylsilonxane (PDMS) device (Figure 1) composed of a gold-coated silicon nitride porous
membrane between a feeding channel and a humidity controlled viewing chamber. It allows cellulose to be attached
and maintained on the upper membrane surface in a micrometer-thick layer of enzyme solution while buffer media
is replenished from the feeding microchannels below. The humidity sensor is placed next to an observation window
to monitor humidity in the humidity control chamber that uses vapor-based capillary-driven microfluidic regulation.
For enhanced signal-to-noise detection, we use an infrared or atomic force microscope to focus an infrared beam,
resulting in studying hydrolysis with a micrometer-to-nanometer spatial resolution.
Figure 1: A. Microfluidic membrane device
schematic. B. Cellulose fibrils are maintained in
thin film of Cel7A-medium fluid that allows
probing with IR light. Soluble enzyme is
supplied from the media channel below the
membrane. C. Device. D, E. Results of
Computational Fluid Dynamic (CFD)
simulation of the temperature and relative
humidity fields for the device which are
consistent with values from the temperature and
humidity sensors.
RESULTS AND DISCUSSION
Figure 2. Imaging of cellulose during
enzymatic hydrolysis. Continuous live
measurements of hydrolysis were performed
using synchrotron infrared (SIR)2 illumination,
using a 8x8 grid with 2.5 m step size and 20
minute spacing for 14 hours. SIR spectra were
obtained using transflection measurement mode
between 4000 and 650 cm-1 at 4 cm-1 spectral
resolution, and averaged 16 scans.
Figure 3. Visualization of spatially resolved
deconstruction of cellulose during
hydrolysis. A. IR spectrum of C. aegagropila
cellulose and their diagnostic markers. B.
Intensity heat map of glycosidic bond shows
the cellulose “particle” shrinks and thins from
outer to inner regions (imaged region in Figure
2). C, D. Tracking the location-specific
kinetics of cellulose depletion and impact of
hydrolysis on molecular ordering in cellulose.
Cellulose hydrolysis study was performed (Figure 2) and the optimum flow rate ratio for the two fluid inputs
(Figure 1) was 1:2.5 (DI water in the upper channels: Cel7A-buffer medium in the lower channels) to maintain a
micrometer thick liquid film and pH during live infrared imaging of hydrolysis. Algal cellulose fibrils purified from
C. aegagropila were deposited onto the porous membrane and incubated overnight in the refrigerator for
attachment. Prior to use, unbound cellulose was rinsed off the surface with enzyme-free buffer. At the start of the
experiment, enzyme (Trichoderma reesei Cel7A (TrCel7A)) at a concentration of 3.5 g/mL in sodium acetate pH
5 was injected into the bottom media channel at a flow rate of 0.5 L/min. DI water was injected into the upper
channel to feed the capillaries in the humidity control chamber at a flow rate of 0.2 L/min to maintain a relative
atmospheric humidity of ~85% at 37 °C. As a polymer of glucose, cellulose features C-O bonds at five of the six
carbons of the glucose residues that contribute to absorption peaks (Figure 3A) prominently at 1034 cm-1, 1060 cm-
1, and 1160 cm-1, corresponding to C-O stretches at the sixth (C6) and third carbon (C3) of the glucose residues,
and the glycosidic bond (C-O-C) of the polysaccharide. We documented spatially heterogeneous and temporal
changes in the abundance (Figure 3B, 3C) and molecular ordering (Figure 3D) of cellulose during enzymatic
hydrolysis over 12 hours. We observed a decline in absorbance intensity of the 1160 cm-1 glycosidic bond peak,
and wavenumber shifts in the prominent 1034 cm-1, 1060 cm-1, and 1160 cm-1 peaks. In some regions of cellulose,
‘blue shift’ of these three peaks indicated that cellobiohydrolase action removed ‘less ordered’ cellulose, leaving
intact the intra-and intermolecular hydrogen bonding by C3 and C6 in cellulose. In regions of thick cellulose, ‘red
shift’ of these same peaks indicated disordering, likely due to active decrystallization of cellulose by enzyme complexation
and hydrolysis.
CONCLUSION
This work demonstrates for the first time the capability of in situ IR imaging for profiling the spatial and
temporal changes in cellulose molecular ordering and deconstruction during enzymatic hydrolysis by coupling
infrared microscopy with open-channel temperature and humidity controlled microfluidics membrane devices.
ACKNOWLEDGEMENTS
This work was supported by the U.S. DOE, Office of Biological and Environmental Research (BER),
REFERENCES
[1] Holman HY et al. Anal Chem, 2010, 82(21), 8757-8765
[2] Loutherback K et al. Protein Pept Lett, 2016, 23(3), 273-282.
[3] Jeoh, T et al. Biotechnol. Bioeng. 2017, 114 (7), 1363–1631. https://doi.org/10.1002/bit.26277.
CONTACT: H.Y.N. Holman; phone +1-510-486-5943; hyholman@lbl.gov; Website: https://bsisb.lbl.gov/
