Programmed magnetic manipulation of vesicles into spatially coded prototissue architectures arrays

Abstract

In nature, cells self-assemble into spatially coded tissular configurations to execute higher-order biological functions as a collective. This mechanism has stimulated the recent trend in synthetic biology to construct tissue-like assemblies from protocell entities, with the aim to understand the evolution mechanism of multicellular mechanisms, create smart materials or devices, and engineer tissue-like biomedical implant. However, the formation of spatially coded and communicating micro-architectures from large quantity of protocell entities, especially for lipid vesicle-based systems that mostly resemble cells, is still challenging. Herein, we magnetically assemble giant unilamellar vesicles (GUVs) or cells into various microstructures with spatially coded configurations and spatialized cascade biochemical reactions using a stainless steel mesh. GUVs in these tissue-like aggregates exhibit uncustomary osmotic stability that cannot be achieved by individual GUVs suspensions. This work provides a versatile and cost-effective strategy to form robust tissue-mimics and indicates a possible superiority of protocell colonies to individual protocells. To execute higher-order functions, cells self-assemble into spatially coded tissue configurations. Here the authors magnetically assembly giant unilamellar vesicles into three dimensional tissue-mimic structures with collective osmotic stability.

Full text

ARTICLE
Programmed magnetic manipulation of vesicles
into spatially coded prototissue architectures
arrays
Qingchuan Li1,2, Shubin Li1,2, Xiangxiang Zhang1, Weili Xu1& Xiaojun Han1*
In nature, cells self-assemble into spatially coded tissular configurations to execute higher-
order biological functions as a collective. This mechanism has stimulated the recent trend in
synthetic biology to construct tissue-like assemblies from protocell entities, with the aim to
understand the evolution mechanism of multicellular mechanisms, create smart materials or
devices, and engineer tissue-like biomedical implant. However, the formation of spatially
coded and communicating micro-architectures from large quantity of protocell entities,
especially for lipid vesicle-based systems that mostly resemble cells, is still challenging.
Herein, we magnetically assemble giant unilamellar vesicles (GUVs) or cells into various
microstructures with spatially coded configurations and spatialized cascade biochemical
reactions using a stainless steel mesh. GUVs in these tissue-like aggregates exhibit uncus-
tomary osmotic stability that cannot be achieved by individual GUVs suspensions. This work
provides a versatile and cost-effective strategy to form robust tissue-mimics and indicates a
possible superiority of protocell colonies to individual protocells.
https://doi.org/10.1038/s41467-019-14141-x OPEN
1State Key Laboratory of Urban Water Resource and Environment, School of Chemistry and Chemical Engineering, Harbin Institute of Technology, 92 West
Da-Zhi Street, Harbin 150001, China.
2
These authors contributed equally: Qingchuan Li, Shubin Li *email: hanxiaojun@hit.edu.cn
NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications 1
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved

During the evolution of life, one of the major transitions is
the appearance of multicellular systems with spatially coded
cell types1,2, which communicate and cooperate to exhibit
higher-order collective behaviors in the form of tissues or organs.
Mimicking these systems via the controlled assembly of synthetic
cell-like entities is expected to result in important implications for
the fabrication of artificial living systems and promising applica-
tions in the field of tissue engineering3.Sofar,variouskindsof
protocell entities, such as liposomes4–6, polymersomes7,proteino-
somes8, and water-in-oil emulsion droplets9–11 have been inte-
grated into rudimentary tissue-like assemblies that exhibit higher-
order behaviors as a collective including communication, defor-
mation, signaling, and differentiation. However, except the series of
breakthrough studies based on water-in-oil emulsion networks,
most of the current tissue-like assemblies are amorphous aggregates
of some protocell entities, especially for the lipid vesicle-based
systems that most closely resemble cells12. In a recent breakthrough,
Ces and coworkers13 sculpted small group of vesicles into defined
spatial organization using optical tweezers. The assembly of large
quantity of vesicle-based protocell entities into spatially coded and
communicating micro-architectures to mimick the existence form
of natural tissues remains a considerable challenge.
As a versatile, noninvasive, and cost-effective strategy, mag-
netic manipulation has been increasingly exploited for the scal-
able assembly of magnetic and nonmagnetic objects into two-
dimensional (2D) or three-dimensional (3D) metastructures
based on their responses to inhomogeneous magnetic field in two
mechanisms: positive magnetophoresis (moving to areas with
maximized field intensity) for magnetic objects14,15, and negative
magnetophoresis (moving to areas with minimized field intensity)
for objects with lower magnetic susceptibility than that of sus-
pension media16,17. Compared with the well-established study
and wide application of the first mechanism, the investigation of
the second mechanism is still in its infancy, but attracts intense
attentions in recent years because of its universal applicability for
different kinds of inanimate and living materials18,19, and unique
manipulation behaviors for the objects going beyond Earnshaw’s
theorem20. Early study in this area often required quite high
magnetic field intensity. However, a recent revolution by intro-
ducing paramagnetic dispersing environment brought up a
magneto-Archimedes effect21, enabling the manipulation of
nonmagnetic objects under weak magnetic field. Based on this
effect, 2D colloidal particles lattice was assembled on Ni grid with
different morphologies in paramagnetic Ho(NO
3
)
3
solution22.
Simple 3D structures, for example, the spheroidal tissue-like
models, have also been obtained via magnetic levitation devices in
a more biocompatible gadolinium-based nonionic paramagnetic
solution23. However, the application of this technique for the
formation of more complicated 3D aggregates, for instance, the
spatially coded tissue-like giant unilamellar vesicles (GUVs)
assemblies, has rarely been reported.
Herein, we describe the scalable magnetic assembly of cell-
mimic (GUVs) colonies with tissue-like complex 3D organiza-
tions using a stainless steel (SS) mesh with patterned microwells
in a paramagnetic solution media. The independent and colla-
borative influences of the microwell parameters (morphology and
arrangement) and directions of external magnetic field on GUVs
colonies formation with different spatial organizations are
investigated. Cascade enzyme reactions among these spatially
organized structures are engineered. Our work provides a method
to form higher-order tissue-like structures for synthetic biology,
tissue engineering, and the study of spatially compartmented
chemical reactions, and exhibits a further step for the controlled
magnetic manipulation.
Results
The setup for GUVs assembly. The assembly of diamagnetic
GUVs was carried out on a SS mesh (thickness ≈100 μm) with
patterned microwells (Supplementary Fig. 1) by mixing GUVs
mother dispersion electroformed in 400 mM sucrose solution
(Supplementary Fig. 2) with isotonic paramagnetic MnCl
2
or
Gadobutrol solution. The inner volume of GUVs was 400 mM
sucrose solution, and the outside solution was an isotonic mixture
of sucrose and paramagnetic compounds with relative lower
density. The GUVs encapsulated with sucrose solution were
heavier than their surroundings, so when no magnetic field was
applied, GUVs homogeneously deposited on the microwells and
grids under gravity (Supplementary Fig. 3). After the magnetic
field was applied from the bottom, the SS mesh exhibited a
paramagnetic response, i.e., magnetic moments parallel to the
external magnetic field were generated in the SS mesh (Supple-
mentary Fig. 4a), which resulted in the formation of magnetic
field gradient microenvironments around the SS mesh because of
the interplay between magnetized magnetic field from SS mesh
and external magnetic field from the magnets. These micro-
environments drove the aggregation of GUVs that deposited
around the SS mesh under gravity to form tissue-like colonies in
the paramagnetic media containing MnCl
2
. For a GUV in mag-
netic field with radius of Rat position r, the magnetostatic
potential energy U(r) was given by22
UðrÞ¼2πR3μ0
χGχS
χGþ2χSþ3HrðÞ
jj
2
;ð1Þ
where μ
0
is the magnetic permeability of vacuum, χ
G
and χ
S
are
the magnetic susceptibilities of GUVs and paramagnetic solution,
respectively, and H(r) is the magnetic field at the position of the
GUVs. For diamagnetic GUVs (χ
G
<χ
S
), the potential energy U(r)
was strictly positive as implied by Eq. (1), so GUVs tended to
move towards regions with minimum magnetic field in the gra-
dient microenvironments for lower U(r). Therefore, it can be
expected that the spatial organization of GUVs colonies can be
determined by the distribution of magnetic field around the
microwells. In this work, magnetic fields with three directions
versus the SS mesh surface, i.e., vertical (Fig. 1a), horizontal
(Fig. 2a), and inclined (Fig. 2d) magnetic fields, were adjusted to
modulate the magnetic field microenvironment around the SS
mesh for the assembly of GUVs in paramagnetic media. The
assembly device contained a cover slip, the SS mesh clinging to
the cover slip, and a Teflon cell enclosing the SS mesh (Supple-
mentary Fig. 5).
GUVs assembly under vertical magnetic field. The vertical
external magnetic field versus the SS mesh surface was provided
by putting the device on the top center of a permanent magnet
(Fig. 1a). As indicated by the horizontal (top in Fig. 1b) and
vertical (bottom in Fig. 1b) central section of the simulated
magnetic field distribution across the microwells, local micro-
environments with lower magnetic field strength than sur-
rounding space were generated in the patterned microwells. So
diamagnetic GUVs tended to aggregate in the microwells under
gravity and negative magnetophoresis. Moreover, the strength of
the magnetic field in the microwells is also inhomogeneous,
radially decreased from the center to the well walls, which caused
the preferential localization of GUVs around the microwells to
form toroidal microstructures. This phenomenon is more
superficially similar to the traditional manipulation behavior for
magnetic objects rather than diamagnetic entities. In previous
studies related to magnetic levitation or colloidal assembly at 2D
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x
2NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved

surfaces, diamagnetic objects tended to run away from the
magnets or the magnetized Ni grids, while magnetic materials
tended to be attracted and collected by them22. The interplay of
the magnetic field from the magnet and the magnetized SS mesh
resulted in the paramagnetic manipulation-like collection of
GUVs on the microwell walls. With the increase of time and
amount of GUVs, the microwells were gradually occupied by
GUVs from the well walls to the center (Fig. 1c and Supple-
mentary Fig. 6). Complete occupation of the microwells to form
columnar GUVs colonies arrays (Fig. 1d) was realized with
incubation time of 2 h in 0.04 mg mL−1GUVs solution (volume
=300 μL, V
GUVs mother solution
/VMnCl2
solution
=1/4). The top view
and side view from the laser confocal microscope indicated the
close packing of GUVs in the colony (Fig. 1e). The columnar 3D
structure of the tissue-like aggregates can be recognized from a
3D image constructed from serial sections of images in Z-stacks
(Fig. 1f). There existed the adhesive van der Waals force and
different kinds of repulsive forces, including undulation, hydra-
tion, and electrostatic forces, among GUVs. The balance of these
forces determined whether the GUVs were repulsive or adhesive.
However, this had no influence on the close packing of GUVs.
The adhesive GUVs in assembly solution containing MnCl
2
, and
the repulsive GUVs in 400 mM nonionic Gadobutrol solution can
both form closely packed GUVs colonies (Supplementary Fig. 7).
The driving forces for GUVs assembly and close packing were the
magnetic force and the gravity. The gravity facilitated the
deposition of GUVs around the mesh and the magnetic force
promoted their aggregation at region with lower magnetic field
rather than deform them (Supplementary Fig. 8).
As shown by Fig. 1e and Supplementary Fig. 9a, the
electroformed polydisperse GUVs displayed heterogeneous size
distribution in the microwells. From the area close to the
microwell wall to that close to the center, the average diameter of
GUVs gradually decreased. This phenomenon was more
pronounced for the GUVs colony that partially occupied the
microwells via using fewer GUVs (Supplementary Fig. 9b). This is
because larger GUVs settled more quickly under gravity than
smaller ones24. Based on this size-dependent assembly
Magnet
100
µm
h
i
0.240.220.20
0.180.16
c
d
g
GUVs
ef
0.05
0.1
0.15
0.2
>0.25
Top view
SS mesh on magnet
Side view
Magnetic field strength (T)
Magnetic field strength (T)
b
95% occupation
85% occupation
70% occupation
a
Top view Side view
Fig. 1 Assembly of giant unilamellar vesicles (GUVs) on stainless steel (SS) mesh under vertical magnetic field. a Schematic illustration of the device
for GUVs assembly: a SS mesh placed on the top center of a magnet. bHorizontal (top) and vertical (bottom) central section of the simulated magnetic
field distribution across the microwells. The white dash box and the black dash circle respectively indicated the unit of the microwell array and the
approximating unit for decentralized microwells. The black dash arrow indicated the preferential localization of GUVs around microwell wall.
cFluorescence images of the GUVs colonies with different extent of occupation of the microwells. The white dash circle indicated microwell wall.
dFluorescence image of GUV colony arrays formed in SS mesh with microwell diameter of 250 μm. eTop view and side view along the yellow dash section
line of the GUVs colony taken by a laser confocal microscope. fA 3D image of GUVs colony obtained from serial sections of images in the Z-stacks taken
by a laser confocal microscope. gFluorescence images of GUVs colonies with different morphologies: from left to right, triangular, square, striped, and HIT-
like assemblies. The dash triangle, rectangle, and line illustrated rough outline of GUVs colonies. hThe schematic and simulated magnetic field distribution
of SS mesh with densely packed microwells. The white dash box presented the unit of the microwell array. The black arrows indicated the corners.
iFluorescence images of the Chinese ancient coin-like round GUVs colonies with square holes formed using the SS mesh with densely packed microwells.
The bottom is the enlarged image in the dash box of top image as indicated by the yellow dash arrow. The white dash box in bottom image indicated the
square hole in GUVs colony. The scale bars are 100 μm.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x ARTICLE
NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved

phenomenon, we programmed the formation of GUVs colony
with alternating GUVs layers of different average sizes (Supple-
mentary Fig. 9c) via the successive adding of GUVs for two times.
The GUVs in the colony were mutually isolated and retained
the integrity of their compartmentalized interior. No fusion/
hemi-fusion events and leakage of compartmentalized fluorescent
molecules were observed in the experiments (Supplementary
Movie 1 and Supplementary Figs. 10, 11). However, these GUVs
displayed no individual transitional motion (Supplementary
Movie 1 and Supplementary Fig. 10), but behaved as a jammed
and consolidated aggregate with collective stability due to their
close packing. When magnetic field was removed and the SS
mesh was inverted to provide a harmful gravity field that
promoted the GUVs colony disassembly, the contact forces25–27
among the closely packed GUVs kept the GUVs colony stable,
regardless of whether the GUVs were mutually repulsive or
adhesive (Supplementary Fig. 7).
The morphology of the obtained GUVs colonies was directly
related to the shape parameters of the microwells. The size of the
columnar GUVs colonies can be modulated by the diameter of
microwell templates (Supplementary Fig. 12). Columnar GUVs
colonies with aspect ratios (height/diameter) of 0.5 (Fig. 1d), 1.0,
and 1.5 (Supplementary Fig. 13) can be obtained via the variation
of the microwell aspect ratios. Moreover, through varying the
design of the structure of the microwells, GUVs colonies with
different morphologies, including triangular, square, striped, and
HIT-shape assemblies, were obtained (Fig. 1g).
From abovementioned results, we have demonstrated that the
magnetic field distribution in the microwells was the result of the
interplay of the magnetized magnetic field from SS mesh and the
external magnetic field from the magnets. In microwells, the
magnetized magnetic field was in opposite direction to the
external magnetic field, which weakened the magnetic field
strength. Therefore, the magnetic field distribution inside each
well can be adjusted by the SS mesh layout. For SS mesh with
decentralized microwells (Fig. 1b), the square unit of the
microwell arrays (white dash box) can be approximately
considered as a circular region (black dash circle). The magnetic
field was radially decreased from the center to the well walls
(Fig. 1b), resulting in axisymmetric columnar GUVs colonies
(Fig. 1c, d). However, when the microwells were quite hugging
(Fig. 1h), the square corners (indicated by the black arrows)
generated more magnetized magnetic field, which resulted in the
weakening of magnetic field strength in greater degree at the part
of microwells adjacent to the corner (bottom image in Fig. 1h).
The magnetic field microenvironments in the microwells then
guided the assembly of GUVs into Chinese ancient coin-like
round GUVs colonies with square holes (Fig. 1i). GUV colonies
with other morphologies, such as round colonies with oval,
elliptical, heart-shaped, half-round, or hexagonal holes, and
striped colonies with waved edges, can be predicted to generate
by varying the spatial organization of the microwells according to
the simulated results (Supplementary Fig. 14). Taken together,
from abovementioned results, GUVs colonies with various
morphologies were formed by the variation of microwell
morphologies and spatial organization under vertical
magnetic field.
GUVs assembly under horizontal and inclined magnetic field.
The above text addressed the influence of the morphology and
spatial organization of microwells on GUVs colonies formation
under external vertical magnetic field. The following text will
discuss the influences of external magnetic field with other
directions, i.e., horizontal magnetic field and inclined magnetic
field, on GUVs colonies formation. Under horizontal external
To the magnet center
0.05
0.1
0.15
0.2
0.25
Magnetic field strength (T)
Magnetic field strength (T)
B
N
S
N
S
Side view
Top view
B
0.08
0.1
0.12
0.14
0.16
0.18
0.2
Top view
ab
c
de
f
B
B
To magnet center (1 1)
(4 4)
Fig. 2 Assembly of GUVs on the SS mesh under horizontal and inclined magnetic field. a The schematic of the device for horizontal magnetic field.
bSimulated magnetic field distribution on the top surface of the SS mesh under horizontal magnetic field. cFluorescence images of the GUVs colonies
formed under horizontal magnetic field. The right image is the enlarged image in the yellow dash box of the left image. dThe schematic of the device for
inclined magnetic field (with directions between the vertical and horizontal magnetic fields) by putting the SS mesh on one side of the top of the magnet.
eSimulated magnetic field distribution at the bottom surface of the SS mesh under inclined magnetic field. fFluorescence images of the GUVs colonies
under inclined magnetic field. The right image is the enlarged image in the yellow dash box of the left image. The dash circles in b,c,e, and findicate the
microwells. The scale bars in cand fare 200 μm.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x
4NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved

magnetic field by locating the SS mesh between two face-to-face
magnets (Fig. 2a), the region with minimum magnetic field
strength located on the frame of the SS mesh (Fig. 2b). So
stripped GUVs colonies perpendicular to external magnetic field
were formed on the frames rather than inside the microwells
(white dash circles) (Fig. 2c). Inclined external magnetic field
containing both horizontal and vertical components was provided
by putting the SS mesh on one side of magnet top surface
(Fig. 2d), which resulted in Janus distribution of magnetic field
strength at the bottom surface of the microwells. The part of
individual microwell away from the magnet center was in low
magnetic field strength (blue), while the opposite part was in high
magnetic field strength (red) (Fig. 2e). Therefore, as presented by
the fluorescence images, GUVs aggregated at the part of indivi-
dual microwell away from the magnet center rather than the
other side (Fig. 2f). Moreover, a gradient occupancy of the
microwells by GUVs was observed in the experiments (Fig. 2f and
Supplementary Fig. 15). From the microwells near the magnet
center to those close to the magnet edge, the amount of GUVs in
microwells gradually decreased. For example, more than half area
of microwell (1 1) in Fig. 2f was filled by GUVs, while microwell
(4 4) away from the magnet center was only occupied for about
one third of the area. We ascribe the gradient location of GUVs to
the non-uniform spatial distribution of the inclined magnetic
field. As presented by the simulated magnetic field distribution
around the magnet (Supplementary Fig. 16), the ratio of hor-
izontal component to the vertical component increased from the
magnet center to the edge, which resulted in decreased areas with
low magnet flux density in Fig. 2e for GUVs to occupy.
Formation of hybrid GUVs colonies. Biological tissues are
composed of coded microstructures with different cell types. To
mimic this structural complexity, we magnetically manipulated
the assembly of two kinds of GUVs, i.e., 1,2-dioleoyl-sn-glycero-
3-phosphoethanolamine-N-(7-nitro-2-1,3-benzoxadiazol-4-yl)
(NBD PE) labeled GUVs with green fluorescence (gGUVs) and
1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine with
triethylammonium salt (TR DHPE) labeled GUVs with red
fluorescence (rGUVs) into configurations with different spatial
organizations. Parallel coding with fully mixed gGUVs and
rGUVs was obtained for the assembly event of pre-mixed GUVs
under vertical magnetic field (Fig. 3a, Supplementary Fig. 17).
However, higher-order structures with serially coded GUVs
colonies can be obtained via the alternate addition of different
GUVs or the application of different magnetic fields. Firstly,
patterned layer-by-layer gGUVs and rGUVs colonies were
observed via the alternate addition of these two kinds of GUVs
under vertical magnetic field (Fig. 3b, Supplementary Figs. 18 and
19). The ratio of these two GUVs colonies in the microwells can
be modulated by varying the added amount of different GUVs
(Supplementary Fig. 20) and the structure of the microwells
sculpted morphologies of the microarchitecture (Supplementary
Fig. 21). The top and side view of the microstructures under laser
confocal microscope confirmed their coaxially coded configura-
tion (Supplementary Fig. 22). Secondly, asymmetrically config-
ured microstructures with two different GUVs colonies were
obtained via the alternate application of two inclined magnetic
fields (Case 1, Fig. 3c and Supplementary Fig. 23) or one inclined
one and another vertical one (Case 2, Fig. 3d and Supplementary
Fig. 23) for the chronological trapping of two GUV types. In case
1, the two inclined magnetic fields had different directions, with
one of them provided by putting the SS mesh on one side of the
top surface of the magnet and the other one provided by putting
it at the opposite side. In case 2, except for the rGUVs colonies
that were located opposite to the pre-trapped gGUVs in the
microwells, an additional rGUVs layer was formed near the pre-
trapped GUVs layer as indicated by the white arrows. This can be
attributed to the shaping effect of the pre-trapped GUVs colonies
on the spatial distribution of magnetic field in the microwells,
resulting in local weak magnetic field around them for GUVs
aggregation (Supplementary Fig. 4b). Thirdly, grid-like aggregates
composed of orthogonal gGUVs and rGUVs stripes (Fig. 3e,
Supplementary Fig. 24) can be generated by the successive
application of two horizontal magnetic field perpendicular to each
other for the respectively trapping of gGUVs and rGUVs. Finally,
through the successive application of one vertical magnetic field
for gGUVs and two perpendicular horizontal magnetic fields for
rGUVs, we obtained more complex structures containing
columnar gGUVs colonies in the microwells and meshed rGUVs
on the grid (Fig. 3f, Supplementary Fig. 24).
Versatility of the magnetic assembly method. Taken together,
via the modulation of SS mesh parameters (microwell morphol-
ogies and organizations) and external magnetic field, we proposed
a robust and cost-effective method to manipulate large quantity
of GUVs into tissue-mimic microstructures with different
morphologies and spatial organizations. Moreover, this method
can also be applied for the assembly of GUVs on Ni wires to form
1D colonies (Supplementary Fig. 25), on Ni foam to form 3D
networks (Supplementary Fig. 26), and even along the scratch on
a stainless sheet to generate 1D colonies arrays (Supplementary
Fig. 27). Other materials, including gel particles, emulsions,
bubbles and even cells, can all be trapped around the support to
form defined structures (Supplementary Fig. 28). The versatility,
universality, simplicity, and scalability of this method endowed it
with potential boom of applications in synthetic biology, tissue
engineering, photic and electronic devices fabrication, the engi-
neering of electrode surfaces, and the quality control of iron
materials, etc. In this paper, we mainly focus on their potential as
proto-tissues to provide a stable environment for individual cell-
mimic GUVs, and to spatialize biochemical reactions.
Osmotic stability of the GUVs colony. As cell mimics, individual
GUVs suspensions are very fragile. For example, an imbalanced
osmotic (hypotonic or hypertonic) condition can easily cause
their deformation or rupture28–30. This impeded their application
in advanced synthetic cells development, cell biology, and bio-
sensing, etc. However, in this work, when GUVs were magneti-
cally gathered to form GUVs colonies, they displayed
uncustomary stability. For colonies composed of GUVs encap-
sulating 400 mM sucrose, no morphology change for individual
GUVs was observed in a hypotonic condition provided by pure
water, and hypertonic conditions provided by 1000 mM glucose,
500 mM NaCl and even 333 mM CaCl
2
(Supplementary Fig. 29).
The osmotic stability of the GUVs colony originated from the
resistance of the closely packed GUVs in the colony as a collective
to the external osmotic shock. In hypotonic condition provided
by pure water, the reinforced mechanical stability of individual
GUVs from the crowded GUVs surroundings impeded their
rupture and maintained their morphologies. In hypertonic con-
ditions, GUVs in unbalanced osmotic condition experienced a net
force from the high concentration region to the low concentration
region (Supplementary Fig. 30a). In previous studies, this unba-
lanced osmotic condition drove the motion of cancer cells and
GUVs to the region with low osmolyte concentration following
the osmotic engine model31,32. In our case, the forces generated
from the unbalanced osmotic condition compressed the GUVs
colony. This compression was better presented for the colony that
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x ARTICLE
NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved

partially occupied the microwell. Under isotonic condition, the
newly assembly GUVs colony contained some protrusions (yel-
low dash line in the left image of Supplementary Fig. 31). When
external solution was replaced with 1 M glucose solution to
introduce a hypertonic osmotic stress, the GUVs colony was
compressed and the osmotic stress smoothed the interface
between the GUVs colony and external solution (middle image in
Supplementary Fig. 31). Protrusions reappeared when the
hypertonic external solution was replaced with hypotonic pure
water (right image in Supplementary Fig. 31). The compression
from the external hypertonic osmotic stress sealed the voids
among GUVs near the external solution, resulting in the failure of
osmolyte and fluorescent dyes to penetrate into the GUVs colony
(Supplementary Fig. 30b) as confirmed by Supplementary
Fig. 30c–f. In isosmotic assembly solution containing MnCl
2
,
resorufin molecules in external solution gradually diffused into
the voids of GUVs colonies, as evidenced by the gradually
enhanced red fluorescence intensity of resorufin with time in
Supplementary Fig. 30c. In hypertonic glucose, NaCl, or CaCl
2
solutions, the red fluorescence of resorufin was not observed
(Supplementary Fig. 30d–f), indicating its failure to penetrate into
the GUVs colony. The block of the GUVs colony to small
molecules made it behave as an elastic collective under hypertonic
osmotic stress. The osmotic compression increased the elasticity
energy of the GUVs colony, resulting in a negative hydrostatic
energy in the GUVs colony that promoted the entering of water
into the colony. The negative hydrostatic energy (promoting
water in and elasticity energy release) balanced the osmotic stress
(promoting water out) to stabilize the GUVs colony under
hypertonic conditions. This osmotic stability of GUVs colonies
endowed them as robust models for widespread applications.
Moreover, it may also provide a tentative clue for the formation
+++
+++
+
+
Parallel coding
Serial coding
Vertical
magnetic field
Horizontal
magnetic field
Inclined
magnetic field
cdef
rGUVs
gGUVs
b
a
Fig. 3 Coding of spatially controlled GUVs colonies. a Schematic and fluorescence images of GUVs colonies formed via the parallel coding of giant
unilamellar vesicles with green fluorescence (gGUVs) and giant unilamellar vesicles with red fluorescence (rGUVs). Schematic and fluorescence images of
the serially coded GUVs colonies via application of vertical magnetic field for the alternative assembly of gGUVs and rGUVs (b), successive application of
two inclined magnetic fields with different directions respectively for gGUVs assembly (putting the SS mesh on one side of the magnet) and rGUVs
(putting the SS mesh on the other side of the magnet) (c), successive application of inclined magnetic field for gGUVs assembly and vertical magnetic
field for rGUVs assembly (d), successive application of two perpendicular horizontal magnetic fields for the chronological assembly of gGUVs and rGUVs
(e), and successive application of vertical magnetic field for gGUVs assembly and two perpendicular horizontal magnetic fields for rGUVs assembly
(f). The dash circles indicated the microwell wall. The scale bars are 100 μm.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x
6NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved

of proto-tissues from protocells on the early earth, since the
membrane-based protocells aggregates exhibit a mechanical sta-
bility in face of external shock.
Spatialized cascade reaction in the colonies. One unique prop-
erty of cells is their ability to spatialize biochemical reactions
among organelles or tissues for efficient biosynthesis or precise
signaling. In the following part, we evaluated the ability of the
spatially coded GUVs aggregates to mimic the spatialized bio-
chemical process in a simplified model. Two kinds of GUVs, i.e.,
gGUVs with melittin pores encapsulated with glucose oxidase
(GOD) and non-labeled GUVs encapsulated with horseradish
peroxidase (HRP), were magnetically assembled into coaxially
coded colonies in the microwells under vertical magnetic field
(Fig. 4a, b). With the addition of glucose and Amplex Red in the
external solution, glucose entered gGUVs through melittin pores,
and the uncharged Amplex Red passively diffused across the lipid
bilayers33. The GOD in gGUVs catalyzed the oxidation of glucose
to generate H
2
O
2
, which diffused into the non-labeled GUVs,
where they reacted with Amplex Red under the catalysis of HRP
to generate the product of resorufin with red fluorescence
(Fig. 4a). The non-labeled GUVs contained no protein pores, and
H
2
O
2
would also not oxidize the membranes to generate mem-
brane defects34,35 (Supplementary Fig. 32), so the charged
resorufin molecules were trapped in the GUVs. With the increase
of time, more resorufin molecules were formed, so the fluores-
cence intensity increased, until a plateau was achieved above 30
min (Fig. 4c, d). The red fluorescence of resorufin was mainly
observed in the non-labeled GUVs (Fig. 4e), which indicated the
ability of the magnetically coded GUVs aggregates to compart-
mentalize and spatialize biochemical reactions mimicking natural
tissues. Except for the small molecules mediated chemical com-
munication between artificial tissue-like assemblies, this techni-
que can also be utilized to study the chemical process between
GUVs colonies and cell colonies (Fig. 4f, g, and Supplementary
Gluocose
GOD
H2O2
H2O2
HRP
Amplex red
Resorufin
gGUVs with melittin Non-labelled GUVs
Gluocose
GOD
0 min 5 min
10 min 15 min
Cell death
a
b
c
f
g
Cell colony
GUVs colony
Cell colony
e
Fluorescence
intensity (a.u.)
Time (min)
d
010203040
gGUVs colony Colony of non-labelled
GUVs
Fig. 4 Spatialized biochemical reactions in tissue-like GUVs aggregates. a Schematic illustration for the chemical communication between the colony of
gGUVs with melittin and the non-labeled GUVs colony. bFluorescence image and bright field image of the GUVs aggregates with two coaxial GUVs
colonies: fluorescence image of the colonies of gGUVs with melittin (left), bright field image of the two kind of colonies (middle), merged image (right).
cFluorescence images of the GUVs aggregates against time after the biochemical reaction was initiated by the addition of glucose and Amplex Red.
dVariation of the fluorescence intensity of the product against time. The error bar represents the standard error of mean (SEM), n=3 independent
experiments. eFluorescence images of the tissue-like GUVs aggregates with fluorescent resorufin product: fluorescent image of gGUVs colony (left),
fluorescent image of the product of resorufin (middle), merged image (right). fSchematic illustration of the cell death caused by H
2
O
2
that is generated by
the GUVs colonies. gImages for the H
2
O
2
caused cell death: bright field image of GUVs colonies and cell colonies (the first one), fluorescence image of live
cells (the second one), fluorescence image of dead cells (the third one), and merged image (the last one). GOD in aand frepresents glucose oxidase. HRP
in arepresents horseradish peroxidase. The scale bars were 100 μm. Source data are provided as a Source Data file.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x ARTICLE
NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Fig. 33). The GUVs encapsulated with GOD generated H
2
O
2
,
which diffused to the cell colony and caused cell death. The live
cells were stained with fluorescein diacetate (FDA) with green
fluorescence, and dead cells were labeled by propidium iodide
(PI) with red fluorescence. After 6 h of incubation in 400 mM
glucose solution, almost all the cells died as indicated by the
negligible green fluorescence in the second image of Fig. 4g and
the evident red fluorescence in the third image of Fig. 4g. As a
control, GUVs colonies contained no cell death because of the
absence of red fluorescence (Supplementary Fig. 34). According
to the above experimental result, this technique holds great
potential in the investigation of more complicated biological
processes for the study of cell biology and the development of
tissue models with higher-order collective behaviors.
In summary, we obtained tissue-mimicking GUVs aggregates
arrays with different morphologies and spatially coded config-
urations using a SS mesh under magnetic field. GUVs in these
aggregates exhibited uncustomary stability in hypotonic or
hypertonic conditions in comparison with individual GUVs
suspensions, which made them robust models for application in
synthetic biology and cell biology, and suggests possible clues for
the evolution of multicellular cells on early earth. Via the spatial
coding of GUVs or cells, designated GUVs were illumined and
cell death was triggered by enzyme reactions, proving the ability
of the model to mimic the spatialized biochemical processes in
natural tissues. This work paved the way for the study of higher-
order tissue behaviors via the groundbreaking manipulation of
diamagnetic objects into defined 3D structures.
Methods
Materials. 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2-dipalmi-
toyl-sn-glycero-3-phosphocholine (DPPC), and 1,2-dioleoyl-sn-glycero-3-phos-
pho-L-serine (sodium salt) (DOPS) were purchased from Avanti Polar Lipids
(USA). 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-1,3-benzox-
adiazol-4-yl) (NBD PE) and Texas red labeled 1,2-dihexadecanoyl-sn-glycero-3-
phosphoethanolamine, triethylammonium salt (TR-DHPE) were obtained from
Invitrogen (China). Sucrose, glucose, HRP, GOD, Amplex Red, melittin, FDA, PI,
manganese (II) chloride, and Gadobutrol were purchased from Sigma (China).
Cylindrical NdFeb magnets (1T, diameter =3 cm, thickness =1 cm) were bought
from Gates Qiangci Company (Shanghai, China). The square SS mesh (1 cm ×1
cm) with thickness of 100 μm was custom-made by RGRS Company (Shenzhen,
China). Indium tin oxide (ITO) electrode was purchased from Hangzhou Yuhong
technology Co. Ltd (China). Millipore Milli-Q water with a resistivity of 18.2 MΩ
cm was used in the experiments.
Giant unilamellar vesicles (GUVs) formation. Two kinds of mother GUVs
samples, i.e., DMPC/NBD PE (w/w, 95/5) GUVs and DMPC/DOPS/TR-DHPE
(w/w/w, 95/4.5/0.5) GUVs, were formed using electroformation method in 400 mM
sucrose solution using two face-to-face electrode layout of ITO electrodes. Lipid
thin films were formed in the following procedure: 20 μL of lipid solution (5.0 mg
mL−1) was deposited on ITO electrode, and spread using a needle, followed by
drying in a vacuum desiccator for 2 h. The two slides of ITO electrodes were then
assembled with a 2 mm thick Teflon spacer with a 2 cm × 1 cm hole, as reported
elsewhere28,36,37. To form GUVs, the electroformation instrument was placed on a
hot plate with temperat ure of 45 °C, and an AC electric field with amplitude of 5 V
and frequency of 10 Hz was applied for 2 h. The GUVs were observed under the
fluorescence microscope.
GUVs colonies formation. GUVs colonies were formed under magnetic field in a
home-made device (Supplementary Fig. 5). The device was assembled by adhering
a square Teflon cell with opening size of ~1.1 cm × 1.1 cm to a cover slip using
vacuum grease, followed by putting the SS mesh on the top of the cover slip in the
cell. Before using it for GUVs assembly, the device was firstly treated in below
procedure to avoid GUVs rupture during GUVs entrapment experiments. A total
of 200 μL of DPPC ethanol-water solution with ethanol volume percentage of 40%
and DPPC concentration of 0.10 mg mL−1(similar composition used for bicelles
formation38–40 by us) were added in the cell. The device was then heated at 50 °C
for 5 min, and washed using 133 mM MnCl
2
solution for at least three times,
resulting in the formation of supported DPPC membranes on the cover slip and SS
mesh. To form GUVs colony arrays, GUVs mother dispersion in 400 mM sucrose
solution was mixed with isotonic MnCl
2
(133 mM) solution to obtain a mixture
with volume of 300 μL. The mixture was added in the cell, and then the device was
put in magnetic field generated by NdFeB magnets. The GUVs concentration was
controlled by varying the volume ratio of GUVs sucrose solution and 133 mM
MnCl
2
solution with fixed final solution volume of 300 μL. Magnetic fields with
three different directions were used in the GUVs colonies formation experiments.
The vertical magnetic field was provided by putting the SS mesh on the top center
of the magnet. The horizontal magnetic field was provided by putting the SS mesh
between two face-to-face magnets. An inclined magnetic field was provided by
putting the SS mesh on one side of the top of the magnet. To obtain colonies with
coded GUVs assemblies, GUVs solution containing different GUVs were
successively added, or magnetic fields with different directions were successively
applied. The most GUVs magnetic entrapment experiments were lasting more
than 2 h.
Spatialized chemical communications. The communication between two GUVs
populations or one GUVs population and one cell colony was investigated. For the
study of communication between two GUVs populations, gGUVs and non-labeled
GUVs were electroformed in 400 mM sucrose solution containing 12 μgmL
−1
GOD and 1.2 μgmL
−1HRP, respectively. gGUVs encapsulated with GOD was
firstly magnetically trapped in the microwells under vertical magnetic field, incu-
bated in solution with 12 μgmL
−1melittin for 2 h, and carefully washed with
excess MnCl
2
solution to remove the non-encapsulated GOD and free melittin.
Then the non-labeled GUVs encapsulated with HRP were added and trapped in the
microwells under vertical magnetic field followed by the remove of non-
encapsulated HRP via careful washing with MnCl
2
solution. To initiate the reac-
tions, the external solution was replaced with 400 mM glucose containing 50 μM
Amplex Red. The fluorescent product of the cascade reaction was monitored using
fluorescence microscope. For the investigation of the communication between
GUVs population and cell colony, non-labeled GUVs encapsulated with GOD were
trapped in microwells, incorporated with melittin, and washed with excess
Gadobutrol solution. Then HEPG2 cells were added to form coaxial aggregates of
GUVs population (encapsulated with GOD) and cell colony. For comparison,
coaxial aggregates of GUVs population with no GOD and cell colony were also
fabricated. The two microstructures were then incubated in 400 mM glucose
solution for 6 h. The communication between GUVs and cells was verified via the
check of cell viability. Live and dead cells were stained with FDA and PI,
respectively.
Characterization. The topology of the SS mesh was characterized by fluorescence
microscope (Olympus IX73, Japan) and scanning electron microscopy (Quanta 200
FEG, Netherlands). The fluorescence images of the GUVs colonies were obtained
by fluorescence microscope and laser confocal microscope (Olympus FV 3000,
Japan).
Simulation. The magnetic field distribution around the SS mesh was simulated
using COMSOL Multiphysics 4.3 software. The magnetic susceptibilities of GUVs,
SS mesh, and 133 mM MnCl
2
solution were −1.0 × 10−5, 2, and 0.03, respectively.
The external magnetic field generated by magnets magnetized the SS mesh for
GUVs assembly.
Reporting summary. Further information on research design is available in
the Nature Research Reporting Summary linked to this article.
Data availability
The source data underlying Fig. 4d and Supplementary Figs. 6b, 9a, b, c, 11b, 32b, and d
are provided as a Source Data file. The data that support the findings of this study are
available within the paper and its supplementary information. All other relevant data are
available from the authors upon reasonable request.
Received: 22 July 2019; Accepted: 11 December 2019;
References
1. West, S., Fisher, R. M., Gardner, A. & Kiers, E. T. Major evolutionary
transitions in individuality. P. Natl Acad. Sci. USA 112, 10112–10119 (2015).
2. Grosberg, R. K. & Strathmann, R. R. The evolution of multicellularity: a minor
major transition? Annu. Rev. Ecol. Evol. Syst. 38, 621–654 (2007).
3. Mantri, S. & Sapra, K. T. Evolving protocells to prototissues: rational design of
a missing link. Biochem. Soc. Trans. 41, 1159–1165 (2013).
4. Parolini, L. et al. Volume and porosity thermal regulation in lipid mesophases
by coupling mobile ligands to soft membranes. Nat. Commun. 6, 5948 (2015).
5. Parolini, L., Kotar, J., Di Michele, L. & Mognetti, B. M. Controlling self-
assembly kinetics of DNA-functionalized liposomes using toehold exchange
mechanism. ACS Nano. 10, 2392–2398 (2016).
6. Karlsson, A. et al. Molecular engineering: networks of nanotubes and
containers. Nature. 409, 150–152 (2001).
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x
8NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved

7. Jin, H. et al. Reversible and large-scale cytomimetic vesicle aggregation: light-
responsive host-guest interactions. Angew. Chem. Int. Ed. 50, 10352–10356
(2011).
8. Gobbo, P., Patil, A. J., Li, M. & Mann, S. Programmed assembly of synthetic
protocells into thermoresponsive prototissues. Nat. Mater. 17, 1145–1153
(2018).
9. Villar, G., Graham, A. D. & Bayley, H. A tissue-like printed material. Science
340,48–52 (2013).
10. Booth, M. J., Schild, V. R., Graham, A. D., Olof, S. N. & Bayley, H. Light-
activated communication in synthetic tissues. Sci. Adv. 2, e1600056 (2016).
11. Dupin, A. & Simmel, F. C. Signalling and differentiation in emulsion-based
multi-compartmentalized in vitro gene circuits. Nat. Chem. 11,32–39 (2019).
12. Rideau, E., Dimova, R., Schwille, P., Wurm, F. R. & Landfester, K. Liposomes
and polymersomes: a comparative review towards cell mimicking. Chem. Soc.
Rev. 47, 8572–8610 (2018).
13. Bolognesi, G. et al. Sculpting and fusing biomimetic vesicle networks using
optical tweezers. Nat. Commun. 9, 1882 (2018).
14. Massana-Cid, H., Meng, F., Matsunaga, D., Golestanian, R. & Tierno, P.
Tunable self-healing of magnetically propelling colloidal carpets. Nat.
Commun. 10, 2444 (2019).
15. Bharti, B., Fameau, A. L., Rubinstein, M. & Velev, O. D. Nanocapillarity-
mediated magnetic assembly of nanoparticles into ultraflexible filaments and
reconfigurable networks. Nat. Mater. 14, 1104–1109 (2015).
16. Mirica, K. A., Ilievski, F., Ellerbee, A. K., Shevkoplyas, S. S. & Whitesides, G.
M. Using magnetic levitation for three dimensional self-assembly. Adv. Mater.
23, 4134–4140 (2011).
17. Erb, R. M., Son, H. S., Samanta, B., Rotello, V. M. & Yellen, B. B. Magnetic
assembly of colloidal superstructures with multipole symmetry. Nature 457,
999–1002 (2009).
18. Durmus, N. G. et al. Magnetic levitation of single cells. P. Natl Acad. Sci. USA
112, E3661–E3668 (2015).
19. Ge, S. & Whitesides, G. M. “Axial”magnetic levitation using ring magnets
enables simple density-based analysis, separation, and manipulation. Anal.
Chem. 90, 12239–12245 (2018).
20. Earnshaw, S. On the nature of the molecular forces which regulate the
constitution of the luminferous ether. Trans. Camb. Philos. Soc. 7,97–112
(1842).
21. Gao, Q. H. et al. Label-free manipulation via the magneto-Archimedes effect:
fundamentals, methodology and applications. Mater. Horiz. 6, 1359–1379
(2019).
22. Demirörs, A. F., Pilla, P. P., Kowalczyk, B. & Grzybowski, B. A. Colloidal
assembly directed by virtual magnetic moulds. Nature 503,99–103 (2013).
23. Tocchio, A. et al. Magnetically guided self-assembly and coding of 3D living
architectures. Adv. Mater. 30, 1705034 (2018).
24. Varnier, A. et al. A simple method for the reconstitution of membrane
proteins into giant unilamellar vesicles. J. Membr. Biol. 233,85–92 (2010).
25. Zhou, J., Long, S., Wang, Q. & Dinsmore, A. D. Measurement of forces inside
a three-dimensional pile of frictionless droplets. Science 312, 1631–1633
(2006).
26. Corwin, I. E., Jaeger, H. M. & Nagel, S. R. Structural signature of jamming in
granular media. Nature 435, 1075–1078 (2005).
27. Brujić, J. et al. 3D bulk measurements of the force distribution in a
compressed emulsion system. Faraday Discuss. 123, 207–220 (2003).
28. Zong, W. et al. A fissionable artificial eukaryote-like cell model. J. Am. Chem.
Soc. 139, 9955–9960 (2017).
29. Li, S., Wang, X., Mu, W. & Han, X. Chemical signal communication between
two protoorganelles in a lipid-based artificial cell. Anal. Chem. 91, 6859–6864
(2019).
30. Chabanon, M., Ho, J. C., Liedberg, B., Parikh, A. N. & Rangamani, P. Pulsatile
lipid vesicles under osmotic stress. Biophys. J. 112, 1682–1691 (2017).
31. Stroka, K. M. et al. Water permeation drives tumor cell migration in confined
microenvironments. Cell 157, 611–623 (2014).
32. Shoji, K. & Kawano, R. Osmotic-engine-driven liposomes in microfluidic
channels. Lab a Chip 19, 3472–3480 (2019).
33. Piwonski, H. M., Goomanovsky, M., Bensimon, D., Horovitz, A. & Haran, G.
Allosteric inhibition of individual enzyme molecules trapped in lipid vesicles.
P. Natl Acad. Sci. USA 109, E1437–E1443 (2012).
34. Yoshimoto, M. et al. Phosphatidylcholine vesicle-mediated decomposition of
hydrogen peroxide. Langmuir 23, 9416–9422 (2007).
35. Tai, W.-Y. et al. Interplay between structure and fluidity of model lipid
membranes under oxidative attack. J. Phys. Chem. B 114, 15642–15649 (2010).
36. Li, Q., Wang, X., Ma, S., Zhang, Y. & Han, X. Electroformation of giant
unilamellar vesicles in saline solution. Colloid Surf. B 147, 368–375
(2016).
37. Ghellab, S. E., Li, Q., Fuhs, T., Bi, H. & Han, X. Electroformation of double
vesicles using an amplitude modulated electric field. Colloid Surf. B 160,
697–703 (2017).
38. Li, Q. & Han, X. Self-assembled “breathing”grana-like cisternae stacks. Adv.
Mater. 30, 1707482 (2018).
39. Li, Q., Li, C., Mu, W. & Han, X. Topological defect-driven buckling of
phospholipid bicelles to cones for micromotors with modulated heading
pathways. ACS Nano. 13, 3573–3579 (2019).
40. Li, Q. & Han, X. Self-assembled rough endoplasmic reticulum-like proto-
organelles. iScience 8, 138–147 (2018).
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant
No. 21773050), the Natural Science Foundation of Heilongjiang Province for Dis-
tinguished Young Scholars (JC2018003).
Author contributions
X.J.H. supervised the research. X.J.H., Q.C.L., and S.B.L. conceived and designed the
experiments. Q.C.L., S. B. L., X. X. Z, and W. L. X. performed experiments. X.J.H., Q.C.L.,
S. B. L., X. X. Z, and W. L. X. analyzed the data. X.J.H., Q.C.L., and S. B. L wrote the
paper, and all authors commented on the paper. Q. C. L and S. B. L contributed equally
to this work.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
019-14141-x.
Correspondence and requests for materials should be addressed to X.H.
Peer review information Nature Communications thanks Atul Parikh and the other,
anonymous, reviewer(s) for their contribution to the peer review of this work. Peer
reviewer reports are available.
Reprints and permission information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2020
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-14141-x ARTICLE
NATURE COMMUNICATIONS | (2020) 11:232 | https://doi.org/10.1038/s41467-019-14141-x | www.nature.com/naturecommunications 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com