of 7
RESEARCHARTICLE
ENGINEERING
Biological matrix composites from cultured plant cells
EleftheriaRoumeli
a,b
,RodindeHendrickx
a
,LucaBonanomi
a
,AniruddhVashisth
c
,KatherineRinaldi
d
,andChiaraDaraio
a,1
EditedbyJoannaAizenberg,HarvardUniversity,Cambridge,MA;receivedNovember6,2021;acceptedFebruary3,2022
We present an approach to fabricate biological matrix composites made entirely from
cultured plant cells. We utilize the cell’s innate ability to synthesize nanofibrillar cell
walls, which serve as the composite’s fundamental building blocks. Following a con-
trolled compression/dehydration process, the cells arrange into lamellar structures with
hierarchical features. We demonstrate that the native cell wall nanofibrils tether adjacent
cells together through fibrillar interlocking and intermolecular hydrogen bonding. These
interactions facilitate intercellular adhesion and eliminate the need for other binders.
Our fabrication process utilizes the entire plant cell, grown in an in vitro culture; requires
no harsh chemical treatments or waste-generating extraction or selection processes; and
leads to bulk biocomposites that can be produced in situ and biodegrade in soil. The
final mechanical properties are comparable to commodity plastics and can be further
modulated by introducing filler particles.
sustainablematerials
|
biocomposites
|
biopolymers
Biocomposites have the potential to meet global material demand with renewable re-
sources and with reduced life cycle environmental impacts, compared with common
petroleum-based materials (1, 2). However, today’s green composites, which primarily
derive from crops and plant fibers, still rely on matrices or binders derived from petroleum
or on harsh thermal, mechanical, and chemical treatment of mature plants (3, 4). Most
recently, self-growing biocomposites have been proposed as a new class of multifunctional
materials, which capitalize on the ability of living matter to self-fabricate and replicate
(5–7). Bacteria often serve as a fabrication platform for pure nanocellulose (6, 8) and
polyesters (9). Some of the strongest and stiffest cellulose materials are made from bacteria
cultures (10, 11). More complex eukaryotic organisms, such as fungi and fermenting yeast,
have also been used in biocomposites. These approaches minimize waste as no part of the
cell is discarded, reduce production and harvest time, and avoid environmentally harmful
postprocessing steps (12–14). Remarkably, mycelium materials have already reached the
market for protective packaging, insulation, and acoustic panels (12, 15–17). However, the
main drawback of existing nonchemically processed eukaryotic biocomposites is that they
have low mechanical performance [modulus
<
1GPaandstrength
<
10 MPa (12, 18)],
which renders them unsuitable for engineering and structural applications. Plant materials
demonstrate an impressive range of mechanical properties. Their stiffness and strength,
for example, can vary over three orders of magnitude (19). This remarkable range depends
on the plant’s composition and hierarchical organization with nanoscale arrangement of
biopolymers within the cell wall and the distinct microscale patterning at the cell level
(
SI Appendix
) (19). Wood products, like medium-density fiberboard (MDF), combine
wood shavings with additives or binders, to expand their potential uses. Recently, chemical
and/or thermomechanical postprocessing of natural wood has been adopted to create
high-performance materials, with properties comparable to steel, ceramics, and insulating
foams (20–24). However, these materials require chemical treatment such as processing
in NaOH/Na
2
SO
3
or H
2
O
2
solutions for removal of most noncellulose components of
wood (24), which are energy-intensive treatments and produce waste.
Here we describe an approach to fabricate biocomposite materials based on the gradual
dehydration and compression of cultured plant cells. Our approach produces no waste
and reduces processing energy requirements, using only low-pressure filtration and oven
drying at mild temperatures (60
C). We use undifferentiated tobacco cells as a model
system. These cells multiply rapidly (a factor of approximately 80 to 100 every 7 d)
(25), can be used to produce materials in situ, are independent of seasonality and local
climate at production site, require no extraction processes, and can be cold pressed
in molds of different shapes and sizes. Our materials retain the native plant cell wall
composition and nanofibrillar structure naturally secreted by growing plant cells and
obtain a lamellar microstructure induced by the compression process. The obtained
densified bulk biocomposites achieve Young’s modulus and strength comparable to struc-
tural and engineered woods and commodity plastics. We characterize the microstructure,
composition, and mechanical properties of the produced materials and demonstrate that
Significance
Thedevelopmentofnovel
degradablebiocompositescan
contributetoansweringthe
increasingglobaldemandfor
sustainablematerials.Wepresent
amethodtoobtainself-bonded
biocompositematerialsfrom
culturedplantcells.Subjecting
cellstoacold-compression
moldingprocesscreates
hierarchicalbiocompositesthat
havestiffnessandstrength
comparabletocommodity
plastics,whilebeing100%
biodegradableinsoil.Introducing
fillersexpandstheattainable
functionalities,demonstratingthe
versatilityoftheproposed
platform.Theuseoffast-growing
plantcellsoffersthebenefitsof
shortharvesttime,zerobiomass
wasteduringprocessing,insitu
manufacturing,andnoarable
landrequirement.Theapproach
allowsthepossibilityoffurther
tuningthefinalmaterial
propertiesbygenetically
engineeringtheprocessedcells.
Author contributions: E.R., L.B., and C.D. designed re-
search;E.R.,R.H.,L.B.,A.V.,andK.R.performedresearch;
E.R.andA.V.analyzeddata;andE.R.andC.D.wrotethe
paper.
Theauthorsdeclarenocompetinginterest.
ThisarticleisaPNASDirectSubmission.
Copyright © 2022 the Author(s). Published by PNAS.
This article is distributed under
CreativeCommons
Attribution-NonCommercial-NoDerivatives License 4.0
(CCBY-NC-ND)
.
1
To whom correspondence may be addressed. Email:
daraio@caltech.edu.
This article contains supporting information online at
https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.
2119523119/-/DCSupplemental
.
PublishedApril4,2022.
PNAS
2022 Vol.119 No.15 e2119523119
https://doi.org/10.1073/pnas.2119523119
1of7
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.
the incorporation of filler additives allows tuning of the material’s
performance and expands their functionalities, for example, creat-
ing magnetic and electrically conductive biocomposite materials.
Results and Discussion
We harvest plant cells from a suspension culture and compress
them in a permeable mold, to achieve a densified dehydrated
structure (Fig. 1
A
and
Materials and Methods
). During com-
pression, water diffuses through the plant cell wall, and the cell
volume is gradually reduced. When the cells reach a dry state,
corresponding to an approximate 98% weight loss, the resulting
bulk materials (Fig. 1
B
) consist of a hierarchical lamellar stack of
compacted cell walls. Cross-section scanning electron microscopy
(SEM) images of the resulting material (Fig. 1
C
and
D
) illustrate
the obtained anisotropic microstructure.
We characterize the cell morphology with light and laser scan-
ning confocal microscopy (Fig. 2
B
D
), which shows that the
plant cells are elongated, with a mean length of 170
±
60
μ
manda
mean width of 45
±
10
μ
m, and are surrounded by a thin primary
cell wall. By staining the cells we confirm that the cell walls
contain cellulose, pectin, and phenolic compounds (Fig. 2
B
D
)
as expected from this cell type (26, 27) (
Materials and Methods
).
Compositional analysis of the dry biocomposite material confirms
that it is composed of 15% cellulose, 20% hemicelluloses, 6.8%
pectins, and 6.3% phenolic compounds. Literature suggests that
in tobacco plant cells the remaining components are lipids, nucleic
acids, proteins, and inorganics (ash content), which together
account for about 45% of the dry mass (28). Residual water within
our samples is gravimetrically determined to be 7
±
3wt%.
Thus, the process results in a biocomposite material, composed
of a heterogeneous mixture of the natural cell wall biopolymers.
SEM and transmission electron microscopy (TEM) obser-
vations of the biocomposite materials reveal their hierarchical,
anisotropic, and lamellar microstructure composed of compacted
plant cells (Fig. 2
E
H
). TEM demonstrates that the nanofibrillar
structure of the primary cell walls is preserved during cell compres-
sion and dehydration (Fig. 2
F
and
G
). Accepted models suggest
that the primary cell wall is a multilayered structure consisting
of cellulose nanofibrils, arranged in various orientations within
each plane (from entirely isotropic to helically aligned, depending
on cell type and developmental stage), bound in a matrix of
hemicelluloses, pectins, and proteins (29). Even in the case of
randomly distributed cellulose nanofibrils in the plane of the
wall, the structure is considered highly anisotropic across thickness
(30). TEM images of our biocomposites show an average dehy-
drated cell wall thickness of 185
±
57 nm and cellulose crystalline
nanofibril bundles with diameters 1 to 30 nm being conformed
across the consecutive parallel planes (Fig. 2
G
and
H
). We observe
a hierarchical structure: at the supracellular level (micrometer
scale), a lamellar microstructure consisting of compacted cells
(Fig. 2
E
), and at the subcellular level (nanoscale), an anisotropic,
multilayered structure, derived from the natural organization of
the cell wall components (Fig. 2
G
and
H
). High-resolution TEM
(HRTEM) images show that the outer nanofibrils of the cell walls
tether to the adjacent walls (Fig. 2
H
, dangling fibrils pointed by
arrows).
Fourier-transform infrared (FTIR) spectroscopy of hydrated
and oven-dried cells and of the compacted biocomposite (Fig.
2
I
) reveals the predominant vibrations of carbohydrates (cellu-
lose, hemicelluloses, and pectin), proteins, and phenolic com-
pounds (26, 27, 31) in all samples (see
SI Appendix
, Table S1
,
for detailed bond assignment). Comparing the spectra reveals
that after processing, the biocomposite maintains all the native
cell carbohydrate components and retains the same degree of
pectin esterification [indicated by the ratio of the 1,735/1,414
peak intensities (26)] but has a slightly lower amount of protein
compared to the living cells [lower intensity of the 1,650 cm
1
peak (26)]. Moreover, the red-shifted hydrogen bonding band
at 3,000 to 3,700 cm
1
, in the dried biocomposite, compared
to the hydrated cells, reflects the strong intermolecular hydrogen
bonding between the cell wall biopolymers (
SI Appendix
)(32,33).
Based on HRTEM (Fig. 2
H
) and FTIR (Fig. 2
I
) analyses,
we postulate that the cell wall adhesion in our biocomposites
is provided by two mechanisms: 1) fibrillar interlocking and 2)
intermolecular interactions of the polymer chains in the adjacent
parts of the cell walls. Further, considering the role of pectins
in the structural adhesion of cells in plants (34), we hypothesize
Fig.1.
(
A
)Schematicofthefabricationmethod.Plantcellsarecultured,harvested,andsubjectedtoacontrolledcompressionanddehydration,resultingin
a
lamellardensifiedarchitecturewhendried.(
B
)Photographofthebiocomposite.(
C
and
D
)SEMtopandcross-sectionalviewsoftheanisotropicmicrostructure.
2of7
https://doi.org/10.1073/pnas.2119523119
pnas.org
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.
Fig.2.
(
A
)Photographofthecellculture.Microscopyimagesofthecellsstainedfor(
B
)pectins,(
C
)cellulose,and(
D
)phenolics.(Scalebarsin
B
D
,20
μ
m.)
(
E
)SEMimageofacross-section,demonstratingthelamellarmicrostructure.(
F
)TEMand(
G
and
H
)HRTEMimagesofcross-sections(arrowspointingtodangling
fibrilstetheringadjacentcellwalls,markedasCW1andCW2).(
I
)FTIRspectraof1)hydratedcells,2)driedcells,and3)drybiocomposite.Colorcodingfor
biopolymerpeakassignment:gray,allcarbohydrates;blue,cellulose;green,pectin;brown,phenolics;andyellow,proteins.(
J
)MDsimulationofafractionof
twoneighboringcellwalls(
Inset
showspectinandcellulosechainsdiffusedintheadjacentcellwallspace).Carbonatomsarecoloredgrayincellulose,orange
inphenolics,yellowinhemicellulose,andblueinpectinchainsforclarity.Oxygenatomsarecoloredred,andhydrogenatomsarewhite.(
K
)MDresultsfrom
tensileloadingofthecompactedcellwallssystem:1)totalenergyand2)hydrogenbondingenergy.
Inset
showschainsbeingunentangledandpulledoutfrom
thespacebetweenthetwocellwallswhensubjectedtotensileforces.(
L
)SEMofafracturedsurfaceofatensile-testedbiocomposite.(
M
)XRDpatternofthe
biocompositewithmarkedcontributionsfromcellulosepolymorphs
I
α
,
I
β
,II,andIII.
that esterified pectins may facilitate this intercellular bonding. We
conduct reactive molecular dynamic (MD) simulations (ReaxFF)
to verify our hypothesis and study the molecular interactions
between adjacent cell walls. Following the fabrication process, we
simulate the compression of the outer parts of two neighboring
cell walls (Fig. 2
J
and
K
and
SI Appendix,
Figs. S1–S3
). Cell walls
are modeled as mixtures of cellulose, hemicellulose, pectin, and
phenolics at the ratios identified from chemical analysis. The mod-
eled cell wall polymers encompass all functional groups (-OH,
-CO
2
CH
3
,-CH
2
OH, and -COOH) available for intermolecular
interactions within the cell walls. The compacted and equilibrated
system has a volume of 3.6
×
2.7
×
2.5 nm
3
and shows that all
polymer chains at the outer layers of neighboring cell walls interact
and diffuse in each other’s structure upon compressing. Thus,
the compression process leads to molecular interlocking between
adjacent cell walls. Then, we subject the system of compacted cell
walls to tensile testing, which shows that the total energy has a
substantial hydrogen bonding energy component, in agreement
with literature on self-adhesive cellulose materials (35) (Fig. 2
K
and
SI Appendix
,Fig.S3
). The fibrillar interlocking leads to chain
unfolding (unentanglement) and cascading hydrogen bond break-
ing and reformation events upon tensile loading (Fig. 2
K
,
Inset
)
(35, 36). These results match SEM observations of the tensile
tested fracture surfaces (Fig. 2
L
), which show fibrils bridging
neighboring parts of the matrix.
X-ray diffraction (XRD) patterns of the biocomposite reveal
multiple polymorphs of semicrystalline cellulose (I, II, and III,
marked in Fig. 2
G
) (23). Native cellulose from plant species
crystallizes in the type I polymorphs (
I
α
,
I
β
). Regeneration and
mercerization, ball-milling in presence of water and other meth-
ods (36) lead to the more thermodynamically stable cellulose
polymorph II, while ammonia treatment followed by thermal and
pressure treatments is known to convert either cellulose I or II to
III (37–39). In the densified biocomposites, cellulose microfibrils
partially undergo phase transformations to form crystal structures
II and III, likely in response to the pressure applied during
PNAS
2022 Vol.119 No.15 e2119523119
https://doi.org/10.1073/pnas.2119523119
3of7
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.
dehydration and the changing chemical environment during cell
dissociation. We postulate that upon compaction and cell death,
protein and nucleic acid losses, as suggested by FTIR, lead to
the diffusion of amine-rich compounds from the protoplasm to
the extracellular space through the cell walls, thus facilitating
the phase transformations, together with the extended period of
compression.
We perform tensile and three-point bending tests to character-
ize the mechanical performance of the dehydrated biocomposites.
Since our process results in bulk biopolymer structures, which
can serve as a standalone structural material as well as a polymer
matrix to be reinforced with fillers, we compare their mechanical
properties to 1) bulk synthetic polymers of similar density, which
also serve as both a host matrix and unfilled bulk material, and 2)
natural and engineered wood, which share similar composition,
have a hierarchical structure, and are also used as structural
bulk materials. We choose different softwoods (pine), hardwoods
(poplar, oak, and walnut), commercial plywood and MDF, and
synthetic plastics of similar density (polystyrene [PS], polypropy-
lene [PP], and low-density polyethylene [LDPE]) (Fig. 3
A
C
and
SI Appendix
,Fig.S4
). Stress–strain plots obtained from the
biocomposites (
SI Appendix
,Fig.S5
) show an initial linear elastic
response upon loading, both under tension and bending, followed
by a brittle failure at small strains (1
±
0.3%). The Young’s
modulus, calculated from the initial linear elastic part of the
tension experiments, is 2.5
±
0.4 GPa, and the ultimate strength
is 21.2
±
3 MPa. The flexural modulus is 4.2
±
0.4 GPa, and
the modulus of rupture is 49.3
±
3.2 MPa. Testing the flexural
properties of the biocomposite on the two perpendicular planes
(see schematic in
SI Appendix
,Fig.S6
) reveals that stiffness varies
by a factor of ca. 1.75 in the two directions, while strength remains
unaffected by orientation. The measured difference in stiffness
is due to the anisotropic microstructure of the biocomposite, as
discussed above. Tension tests show that our biocomposites are
stiffer than the other tested materials (Fig. 3
A
). However, natural
woods have higher strength (Fig. 3
B
), which can be explained by
their different cellular architectures, cell wall compositions, and
component arrangements within the secondary cell walls. The cells
used in our biocomposites originate from the herbaceous plant
Nicotiana tabacum
, and they naturally develop a thin, unlignified
primary cell wall (we confirm a low phenolics amount of 6.2 wt
%). These cells do not form secondary cell walls and cannot self-
organize in a hierarchical microstructure in our in vitro cultures.
Regardless, the macroscopic mechanical performance of the bio-
composites is comparable to that of commercial engineered woods
and commodity plastics. They surpass literature-reported values
for bulk, three-dimensional composite biomaterials manufactured
through bottom-up methods by eukaryotic organisms that in-
clude plant cells, mycelium, or yeast matrixes. All these biocom-
posites have not been treated to remove native components, and
the microorganism cells serve in fact as the building blocks of those
materials (12, 13, 18, 20, 40, 43, 44) (Fig. 3
D
). We note, however,
that in particular, mycelium-based composites have lower density
than our biocomposites (18). Recently, microbial cultured cells
(from
Escherichia coli
) were used to create cell-based rigid films
through a simple casting and ambient drying method (45). The
localized mechanical properties of the microbial films were mea-
sured through nanoindentation, revealing a reduced modulus of
5 to 42 GPa and strength values of 60 to 800 MPa (calculated from
the measured hardness values). While those properties cannot be
directly compared to the plant cell bulk properties we report in
our work as they address different length scales, they demonstrate
another promising direction for cell-based materials.
A key factor in the design of sustainable materials is their end-
of-life fate. The realization of biological matrix materials, such as
those described here, offers an environmentally friendly alternative
to nondegradable materials, which typically survive in landfills. To
assess the biodegradability of our plant-based biocomposites, we
perform agricultural soil incubation tests (
Materials and Methods
),
comparing their mass loss with that of natural wood (46). Results
show an initial mass gain corresponding to water uptake from
the soil, in both natural wood and biocomposites (Fig. 3
E
). The
detectable mass loss due to biodegradation of the biocomposites
begins 3 wk after incubation, while for natural wood it begins
about 7 wk later. This can be associated to the presence of
lignin in natural wood, which is known to provide resistance to
Fig.3.
(
A
)Young’smodulusand(
B
)tensilestrengthofthebiocompositeandreferencematerials.(
C
)Materialdensity.(
D
)Comparisonofmechanicalproperties
of this work (blue circles are results from bending tests, and blue squares are results from tensile tests) and literature-reported biocomposites in
which
microorganismcellsserveasamatrix,andtheircompositionandnanostructurehavenotbeenaltered:yeast-basedcompositesrepresentedwithrhom
buses
andmycelium-basedmaterialswithcircles(12,13,18,40,41,42).(
E
)Biodegradationofthebiocompositeandnaturalpine.Samplenotation:BC,pure(without
fillers)biocomposite;1,pine;2,poplar;3,oak;4,walnut;5,plywood;6,MDF;7,PS;8,PP;9,LDPE.(ErrorbarsindicateSEfor
n
>
5measurements.)
I
α
,
I
β
,II,
andIII.CNF,cellulosenanofibrils;CNT,carbonnanotubes.
4of7
https://doi.org/10.1073/pnas.2119523119
pnas.org
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.
pathogen attacks on cell walls (47) and the water sensitivity of
our material. We observe an almost complete biodegradation of
the biocomposite 14 wk after initial incubation. A drawback of
rapid biodegradation is water sensitivity. We perform water uptake
and thickness swelling tests (
SI Appendix
) and find that our bio-
composites respond similarly to mycelium/wood biocomposites
(18). After more than 24 h of incubation our materials completely
disperse in water. However, when stored in ambient conditions,
they do not exhibit any swelling or fouling. In applications, the
water uptake could be mitigated with surface treatments or water-
resistant coatings.
The use of cell cultures for materials fabrication allows on-
demand property tuning by interfacing with additive particles
(48, 49). We demonstrate the ability to tune and introduce
properties in the composites, by incorporating filler additives. The
addition of different amounts of natural nanoclay (NC) platelets,
for example, changes the biocomposites’ compressive modulus
and strength (Fig. 4
A
). Upon the introduction of 0.15 wt %
NC the Young’s modulus and strength increase by 36 and 87%,
respectively. At 0.5 wt % NC the improvements are less signifi-
cant, 10 and 5%, respectively. At higher NC concentrations both
properties are reduced below the values of the unfilled material,
which is often observed in polymer nanocomposites because of
fillers’ aggregation acting as a stress concentrator (52). Different
filler particles expand further the biocomposites’ property space
(Fig. 4
B
). We plot the Young’s modulus as a function of density
of different plant cell-based biocomposites: pure cell matrix (BC),
biocomposites containing various amounts of carbon fibers (CF),
halloysite and montmorillonite NC, and graphene (G). Their
properties lie at the intersection of natural cellular materials,
including wood-based materials, and commercial plastics (Fig.
4
B
), presenting Young’s moduli spanning over one order of magni-
tude. We note that our biocomposites are outperformed by pure
cellulose materials and densified wood products. This is because
our approach preserves the entire natural cell wall composition
through mild processing. Filler additives also endow functional-
ities, such as electrical conductivity or magnetic properties. The
electrical conductivity of plant cell/CF composites, for example,
can be tuned varying the CF content (Fig. 4
C
). Similarly, the
addition of 13.5 wt % iron oxide nanoparticles (IN) in the plant
cell matrix conveys ferromagnetic properties, which allow the
biocomposite to support more than five times its weight when
attracted by a magnet (Fig. 4
D
). We foresee possible use of our
densified plant cell biocomposites in panels for packaging or non–
load-bearing automotive applications or space manufacturing.
We introduce a method to create natural biocomposite mate-
rials based on cultured plant cells. The method capitalizes on the
plant cell’s ability to synthesize intricate multilayered structures
of cellulose, hemicellulose, phenolics, and pectin in their cell
walls, which we preserve as nanostructured building blocks. In
the future, the use of different cell cultures and/or genetically
modified species (53), as well as the modulation of processing
conditions to modify the cell arrangement at the microscale,
may allow the fabrication of materials with significantly altered
properties. Similar fabrication approaches can be envisioned for
many other biological systems (e.g., algae and fungi) that can
provide complex nanostructured elements as building blocks for
advanced composite biomaterials.
Materials and Methods
CellCultures.
Nicotiana tabacum Bright Yellow 2 (BY-2) cells were purchased
from DSMZ. Cells were kept in Linsmaier & Skoog medium with vitamins
(HIMEDIA-PT040) with 3% (wt/vol) sucrose at a pH of 5.8. The following
supplements were added: 1
μ
M 2,4-dichlorophenoxyacetic acid (2,4-D), 1
μ
M
a-naphthaleneacetic acid, and 1.46 mM KH
2
PO
4
. The cells were grown in 50 to
Fig.4.
(
A
) Compressive modulus and strength of biocomposites with NC platelets. (
B
) Young’s modulus versus density for various materials and our
biocomposites.Bluegroupscorrespondtobendingexperiments,andredgroupscorrespondtocompression.Thecellulose*areacorrespondstopurece
llulose
fibers,papers,andnanocellulose-basedproducts,includingbacterialcellulose(37,50,51).Densifiedwooddataarefromref.20.(
C
)IVcurvesforbiocomposites
with1and20wt%CF.(
D
)BiocompositewithINexhibitingmagneticproperties.
PNAS
2022 Vol.119 No.15 e2119523119
https://doi.org/10.1073/pnas.2119523119
5of7
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.
300 mL aliquots in 100 mL to 1 L flasks on a rotary shaker (130 rpm) at room
temperature and were subcultured biweekly at 1:10 to 1:60 dilutions.
Fabrication Method.
Living cells were harvested and vacuum filtered in an
Erlenmeyer flask, to remove water and growth media. The collected solid cell
clusters were placed in porous aluminum custom designed molds (Metapor
Aluminum) and subjected to a controlled compression. Pressure was gradually
increased at a 0.1
±
0.05 MPa/d rate to a maximum of 0.8 MPa, to compact
the dehydrating cells until they reached a 2
±
1 wt % solid residual mass.
All samples were subsequently dried for 48 h at 60
C in a benchtop oven
(Heratherm, Thermo Scientific). The residual water after the fabrication process
was 7
±
3 wt % as determined from thermogravimetric analysis measure-
ments of the dried samples. Each sample density was calculated from the ra-
tio of their mass (analytical balance XS205, Mettler Toledo) divided by their
volume.
Materials.
The natural wood materials tested were red oak (
Quercus rubra
),
black walnut (
Juglans nigra
), yellow poplar (
Liriodendron tulipifera
), and sugar
pine (
Pinus lambertiana
) and were kindly provided by California Institute of
Technology (Caltech) resources. The engineered wood samples were hardwood
plywood and medium-density untempered hardboard (MDF) provided by
Caltech resources. The commercial plastics were LDPE (King Plastic Corp.),
SIS-030E high-impact PS (Certene), and Densetec Copolymer PP (Polymer
Industries).
Cell Staining.
Cells were stained for cellulose in a 1% solution of alcian blue
in 3% acetic acid (MilliporeSigma). The staining for pectin was performed using
a 0.01% (vol/vol) ruthenium red solution in water, supplemented with 0.1%
(vol/vol) ammonia (MilliporeSigma).We used safranin O in a 1% solution to stain
phenolic compounds (MilliporeSigma).
LightMicroscopy.
AZeiss Axio Scope A1 (Zeiss) was used for optical imaging of
the untreated and stained cells. Two-photon analysis of the safranin-stained cells
was performed in a Zeiss LSM 710 confocal laser scanning microscope (Zeiss).
Image acquisition was implemented with a LC C-Apochromat 40
×
/1.1 W Korr
M27 objective, at an excitation wavelength of 488 nm and emission wavelength
606 nm.
Electron Microscopy.
SEM images were obtained using an FEI Nova 200
NanoLab Dualbeam Focused-Ion Beam/SEM (FEI), operating at 2 to 30 kV and
10 to 50 pA. For TEM, small sample pieces (1 to 2 mm
3
)werefixedfor1hwith
2%OsO
4
indH
2
O.Thepieceswererinsed3
×
withdH
2
O,dehydratedintoacetone
over 48 h,then infiltrated with Epon-Araldite resin (Electron Microscopy Sciences)
for 48 h. Then, 400 nm serial sections were cut with a UC6
μ
Ltramicrotome
(Leica Microsystems) using a diamond knife (Diatome US), placed on Formvar-
coated copper-rhodium grids (Electron Microscopy Sciences) and stained with
3% aqueous uranyl acetate and lead citrate. Colloidal gold beads (10 nm) placed
on both sides served as fiducial markers for subsequent image alignment. Grids
were placed in a dual-axis tomography holder (Model 2040, E.A. Fischione In-
struments, Inc.) and imaged with a TF-30ST electron microscope (ThermoFisher
Scientific) at 300 keV. Images were recorded with US1000 camera (Gatan, Inc.)
using the SerialEM software package (54), and the analysis was conducted using
the IMOD software package (55).
FTIR Spectroscopy.
FTIR spectra were collected with a ThermoNicolet iS10
spectrometer (Thermo Scientific) equipped with an attenuated total reflection
crystal.Wecollectedspectraof cellssuspendedingrowthmedia,driedintheoven
overnightat60
C,andof thedrybiocomposite.Thespectrawerecollectedacross
the range of 400 to 4,000 cm
1
with a resolution of 2 cm
1
and accumulated
64 scans for each spectrum.
Simulations.
Reactive molecular dynamics simulations (ReaxFF) within the
framework of Software for Chemistry and Materials (56) were used to simulate
the interactions between hemicellulose,cellulose,pectin,and phenols within the
cell walls of adjacent cells using the force field defined in ref.57.The equilibrated
system was composed of two compacted cell walls, following the compression
step of the fabrication process. The final system was virtually tested in tension at
strain rate of 1
×
10
8
s
1
pulling the two walls apart. The simulation setup is
further documented in
SI Appendix
.
XRD.
XRD patterns were collected using PANalytical X’Pert Pro (operating volt-
ageat40kV,currentat40mA,CuK
α
,
λ
= 0.1541 nm). An angular range of
2
θ
=10to60
with a step size of 0.1
C and a scanning speed of 0.008
Cs
1
was used for the measurements (Panalytical B.V.).
Chemical Analyses.
Compositional analysis was carried out using the
anthrone-sulfuric acid colorimetric method for cellulose, acidic hydrolysis for
hemicellulose, the carbazole colorimetric method for pectin, and the klason
method for phenolics.
MechanicalPropertiesCharacterization.
Three-point bending and tension
tests were performed in an eXpert 8612 axial-torsion tester (Admet), an Instron
5500, and an Instron E3000. The biocomposite samples were tested with 250 N
(for tensile tests) and 500 N load cells (for bending tests). For the flexural tests,
a minimum of five samples of each material were tested at a constant strain rate
of 0.004
±
0.001 s
1
until failure. Samples of
40
×
5
×
5 mm (length
×
width
×
thickness) dimensions were tested in two perpendicular directions, as
shownin
SI Appendix
, Fig. S6
.Naturalwoodsamplescutin35
×
7
×
4mmstrips
and plastic samples cut at 75
×
7
×
4 mm strips were tested at the same strain
rate as the biocomposites. For the tensile tests a minimum of five samples were
tested at a constant strain rate of 0.0025
±
0.0001 s
1
until failure. We applied
medium-density fiberboard end-tabs with dimensions 10
×
10
×
3mminall40
×
5
×
5 mm specimens, using a thin layer of polyvinyl acetate adhesive (Gorilla
Wood Glue). The natural wood tension samples, with dimensions
100
×
15
×
5 mm, and the plastic samples in a dog-bone configuration of 115
×
1.5
×
6 mm (American Society for Testing and Materials [ASTM] D638, type IV) were
tested at the same strain rate as the biocomposites. For the compression tests
of the biocomposite samples,
9
×
9
×
2 mm (length
×
width
×
thickness)
samples were tested under compression at a constant rate of 0.001 s
1
to a 10%
target strain.The compressive modulus was calculated from the linear part of the
unloading stress–strain curve, while the maximum stress value during compres-
sion, corresponding to the 10% strain, was referred to as compressive strength.
Biodegradation Tests.
We characterized the biodegradability of the biocom-
posites and pine by incubating rectangular shaped pieces of 0.05
±
0.01 g in
agricultural soil (FoxFarm Ocean Forest Potting Soil). Each pot was stored in an
outdoorslocationfor14wk.Thebiocompositesandnaturalwoodcontrolsamples
were recovered every 2 wk to measure their residual mass. Following literature
reported process (58), the samples were recovered, cleaned, and dried at room
temperature and subsequently weighed.
ElectricalProperties.
Copper tape was used for the electrodes, connected to a
Keithley 2636B source (Tektronix, Inc.). Voltage scans between
2and2V,with
a 0.1 V/s step were recorded.The slope of the obtained IVcurves was converted to
conductivity when multiplied with sample height and divided by cross-sectional
area.
Water Uptake.
Water uptake and thickness swelling tests were performed
according to ASTM D1037 with appropriate modifications (18). Ten dry samples
of
4
×
4 mm were immersed in 5 mL of distilled water,and their relative mass
and thickness increases were measured after 2 h.
DataAvailability.
All study data are included in the article and/or
SI Appendix
.
ACKNOWLEDGMENTS.
We thank Mr. M. Ladinsky, Dr. S. Amanatidis, Dr. A. M.
Jimenez,Dr.Y.Wei,Dr.M.Mello,Mr.A.Carim,Ms.S.Antilla,and Mr.D.A.Nguyen
for support in experiments.We thank Dr.R.Di Giacomo for useful discussions and
Prof. N. Lewis for providing access to the Raman facilities. We thank Caltech Kavli
Nanoscience Institute,Gordon and Betty Moore,and the Beckman Foundation for
support of electron microscopy facilities and the Caltech Beckman Institute and
the Arnold and Mabel Beckman Foundation for supporting the laser scanning
imaging facilities. This work was supported in part by the Resnick Sustainability
Institute at Caltech.
Author affiliations:
a
Division of Engineering and Applied Science, California Institute of
Technology, Pasadena, CA 91125;
b
Department of Materials Science and Engineering,
University of Washington, Seattle, WA 98195;
c
Department of Mechanical Engineering,
University of Washington, Seattle, WA 98195; and
d
Division of Chemistry and Chemical
Engineering,CaliforniaInstituteofTechnology,Pasadena,CA91125
6of7
https://doi.org/10.1073/pnas.2119523119
pnas.org
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.
1. A. K. Mohanty, M. Misra, G. Hinrichsen, Biofibres, biodegradable polymers and biocomposites: An
overview.
Macromol. Mater. Eng.
276–277
, 1–24 (2000).
2. A. K. Mohanty, S. Vivekanandhan, J. M. Pin, M. Misra, Composites from renewable and sustainable
resources: Challenges and innovations.
Science
362
, 536–542 (2018).
3. A. Mohanty, M. Misra, L. Drzal,
Natural Fibers, Biopolymers, and Biocomposites
(CRC Press, 2005).
4. O. Faruk, A. K. Bledzki, H. P. Fink, M. Sain, Biocomposites reinforced with natural fibers: 2000–2010.
Prog. Polym. Sci.
37
, 1552–1596 (2012).
5. P. Q. Nguyen, N. D. Courchesne, A. Duraj-Thatte, P. Praveschotinunt, N. S. Joshi, Engineered living
materials: Prospects and challenges for using biological systems to direct the assembly of smart
materials.
Adv. Mater.
30
, e1704847 (2018).
6. M. Schaffner, P. A. R
̈
uhs, F. Coulter, S. Kilcher, A. R. Studart, 3D printing of bacteria into functional
complex materials.
Sci. Adv.
3
, eaao6804 (2017).
7. C. Gilbert
et al.
, Living materials with programmable functionalities grown from engineered microbial
co-cultures.
Nat. Mater.
20
, 691–700 (2021).
8. P. A. R
̈
uhs,F.Storz,Y.A.L.G
́
omez, M. Haug, P. Fischer, 3d bacterial cellulose biofilms formed by foam
templating.
npj Biofilms Microbiomes
4
, 21 (2018).
9. R. A. Verlinden, D. J. Hill, M. A. Kenward, C. D. Williams, I. Radecka, Bacterial synthesis of
biodegradable polyhydroxyalkanoates.
J. Appl. Microbiol.
102
, 1437–1449 (2007).
10. S. Wang
et al.
, Super-strong, super-stiff macrofibers with aligned, long bacterial cellulose nanofibers.
Adv. Mater.
29
, 1702498 (2017).
11. S. Wang
et al.
, Transparent, anisotropic biofilm with aligned bacterial cellulose nanofibers.
Adv. Funct.
Mater.
28
, 1707491 (2018).
12. M. Haneef
et al.
, Advanced materials from fungal mycelium: Fabrication and tuning of physical
properties.
Sci. Rep.
7
, 41292 (2017).
13. L. Valentini
et al.
, Fermentation based carbon nanotube multifunctional bionic composites.
Sci.Rep.
6
,
27031 (2016).
14. N. M. Pugno, L. Valentini, Bionicomposites.
Nanoscale
11
, 3102–3111 (2019).
15. M. Jones
et al.
, Thermal degradation and fire properties of fungal mycelium and mycelium-biomass
composite materials.
Sci. Rep.
8
, 17583 (2018).
16. E. Karana, D. Blauwhoff, E. J. Hultink, S. Camere, When the material grows: A case study on designing
(with) mycelium-based materials.
Int. J. Des.
12
, 119–136 (2018).
17. K. Cerimi, K. C. Akkaya, C. Pohl, B. Schmidt, P. Neubauer, Fungi as source for new bio-based materials:
A patent review.
Fungal Biol. Biotechnol.
6
, 17 (2019).
18. W. Sun, M. Tajvidi, C. G. Hunt, G. McIntyre, D. J. Gardner, Fully bio-based hybrid composites made of
wood, fungal mycelium and cellulose nanofibrils.
Sci. Rep.
9
, 3766 (2019).
19. L. J. Gibson, The hierarchical structure and mechanics of plant materials.
J. R. Soc. Interface
9
,
2749–2766 (2012).
20. J. Song
et al.
, Processing bulk natural wood into a high-performance structural material.
Nature
554
,
224–228 (2018).
21. L. A. Berglund, I. Burgert, Bioinspired wood nanotechnology for functional materials.
Adv. Mater.
30
,
e1704285 (2018).
22. F. Jiang
et al.
, Wood-based nanotechnologies toward sustainability.
Adv. Mater.
30
, 1703453 (2018).
23. T. Li
et al.
, Anisotropic, lightweight, strong, and super thermally insulating nanowood with naturally
aligned nanocellulose.
Sci. Adv.
4
, eaar3724 (2018).
24. C. Chen
et al.
, Structure–property–function relationships of natural and engineered wood.
Nat. Rev.
Mater.
5
, 642–666 (2020).
25. T. Nagata, Y. Nemoto, S. Hasezawa, “Tobacco by-2 cell line as the “hela” cell in the cell biology of
higher plants” in
International Review of Cytology
, K. W. Jeon, M. Friedlander, Eds. (Academic Press,
1992),
vol. 132
, pp. 1–30.
26. M. C. McCann, K. Roberts, Changes in cell wall architecture during cell elongation.
J. Exp. Bot.
45
,
1683–1691 (1994).
27. M. C. McCann, J. Shi, K. Roberts, N. C. Carpita, Changes in pectin structure and localization during the
growth of unadapted and NaCl-adapted tobacco cells.
Plant J.
5
, 773–785 (1994).
28. Y. Yuan
et al.
, Microwave-assisted hydrothermal extraction of non-structural carbohydrates and
hemicelluloses from tobacco biomass.
Carbohydr. Polym.
223
, 115043 (2019).
29. Y. Zhang
et al.
, Molecular insights into the complex mechanics of plant epidermal cell walls.
Science
372
, 706–711 (2021).
30. D. J. Cosgrove, Plant cell wall extensibility: Connecting plant cell growth with cell wall structure,
mechanics, and the action of wall-modifying enzymes.
J. Exp. Bot.
67
, 463–476 (2016).
31. M. C. McCann, N. J. Stacey, R. Wilson, K. Roberts, Orientation of macromolecules in the walls of
elongating carrot cells.
J. Cell Sci.
106
, 1347–1356 (1993).
32. Y. Hishikawa, E. Togawa, T. Kondo, Characterization of individual hydrogen bonds in crystalline
regenerated cellulose using resolved polarized FTIR spectra.
ACS Omega
2
, 1469–1476 (2017).
33. S. Cichosz, A. Masek, IR study on cellulose with the varied moisture contents: Insight into the
supramolecular structure.
Materials (Basel)
13
, 4573 (2020).
34. A. Peaucelle, R. Wightman, H. H
̈
ofte, The control of growth symmetry breaking in the Arabidopsis
hypocotyl.
Curr. Biol.
25
, 1746–1752 (2015).
35. H. Zhu
et al.
, Anomalous scaling law of strength and toughness of cellulose nanopaper.
Proc. Natl.
Acad.Sci.U.S.A.
112
, 8971–8976 (2015).
36. Q. Meng, T. J. Wang, Mechanics of strong and tough cellulose nanopaper.
Appl. Mech. Rev.
71
,
040801 (2019).
37. R. J. Moon, A. Martini, J. Nairn, J. Simonsen, J. Youngblood, Cellulose nanomaterials review:
Structure, properties and nanocomposites.
Chem.Soc.Rev.
40
, 3941–3994 (2011).
38. A. Mittal, R. Katahira, M. E. Himmel, D. K. Johnson, Effects of alkaline or liquid-ammonia treatment on
crystalline cellulose: Changes in crystalline structure and effects on enzymatic digestibility.
Biotechnol. Biofuels
4
, 41 (2011).
39. Y. Liao
et al.
, The role of pretreatment in the catalytic valorization of cellulose.
Mol.Catal.
487
, 110883
(2020).
40. L. Valentini, S. Bittolo Bon, S. Signetti, N. M. Pugno, Graphene-based bionic composites with
multifunctional and repairing properties.
ACS Appl. Mater. Interfaces
8
, 7607–7612 (2016).
41. L. Valentini, S. Bittolo Bon, N. M. Pugno, Graphene and carbon nanotube auxetic rubber bionic
composites with negative variation of the electrical resistance and comparison with their nonbionic
counterparts.
Adv. Funct. Mater.
27
, 1606526 (2017).
42. R. Di Giacomo
et al.
, Candida albicans/MWCNTs: A stable conductive bio-nanocomposite and its
temperature-sensing properties.
IEEE Trans. NanoTechnol.
12
, 111–114 (2013).
43. R. Di Giacomo, C. Daraio, B. Maresca, Plant nanobionic materials with a giant temperature response
mediated by pectin-Ca2+.
Proc.Natl.Acad.Sci.U.S.A.
112
, 4541–4545 (2015).
44. R. Di Giacomo
et al.
, Bio-nano-composite materials constructed with single cells and carbon
nanotubes: Mechanical, electrical, and optical properties.
IEEE Trans. NanoTechnol.
12
, 1026–1030
(2013).
45. A. Manjula-Basavanna, A. Duraj-Thatte, N. S. Joshi, Robust self-regeneratable stiff living materials
fabricated from microbial cells.
Adv. Funct. Mater
.
31
, 2010784 (2021).
46. K. G. Satyanarayana, G. G. Arizaga, F. Wypych, Biodegradable composites based on lignocellulosic
fibers—An overview.
Prog. Polym. Sci.
34
, 982–1021 (2009).
47. K.Iiyama,T.Lam,B.A.Stone,Covalentcross-linksinthecellwall.
Plant Physiol.
104
, 315–320 (1994).
48. F. Natalio
et al.
, Biological fabrication of cellulose fibers with tailored properties.
Science
357
,
1118–1122 (2017).
49. F. Natalio, Biological fabrication: Future perspectives on biological fabrication and material farming
(small methods 1/2019).
Small Methods
3
, 1800055 (2019).
50. A. J. Ben
́
ıtez, A. Walther, Cellulose nanofibril nanopapers and bioinspired nanocomposites: A review
to understand the mechanical property space.
J. Mater. Chem. A Mater. Energy Sustain.
5
,
16003–16024 (2017).
51. Y. C. Hsieh, H. Yano, M. Nogi, S. Eichhorn, An estimation of the young’s modulus of bacterial cellulose
filaments.
Cellulose
15
, 507–513 (2008).
52. E. Roumeli
et al.
, Factors controlling the enhanced mechanical and thermal properties of
nanodiamond-reinforced cross-linked high density polyethylene.
J. Phys. Chem. B
118
,
11341–11352 (2014).
53. E. Roumeli
et al.
, Structure and biomechanics during xylem vessel transdifferentiation in Arabidopsis
thaliana.
Plants
9
, 1715 (2020).
54. D. N. Mastronarde, Automated electron microscope tomography using robust prediction of specimen
movements.
J. Struct. Biol.
152
, 36–51 (2005).
55. D. N. Mastronarde, Correction for non-perpendicularity of beam and tilt axis in tomographic
reconstructions with the IMOD package.
J. Microsc.
230
, 212–217 (2008).
56. A. C. T. van Duin, S. Dasgupta, F. Lorant, W. A. Goddard, Reaxff: A reactive force field for hydrocarbons.
J. Phys. Chem. A
105
, 9396–9409 (2001).
57. A. Vashisth, C. Ashraf, W. Zhang, C. E. Bakis, A. C. T. van Duin, Accelerated ReaxFF simulations for
describing the reactive cross-linking of polymers.
J. Phys. Chem. A
122
, 6633–6642 (2018).
58. H. Wang, D. Wei, A. Zheng, H. Xiao, Soil burial biodegradation of antimicrobial biodegradable PBAT
films.
Polym. Degrad. Stabil.
116
, 14–22 (2015).
PNAS
2022 Vol.119 No.15 e2119523119
https://doi.org/10.1073/pnas.2119523119
7of7
Downloaded from https://www.pnas.org by California Institute of Technology on April 5, 2022 from IP address 131.215.225.169.