Hydrogel-Immobilized Coacervate Droplets as Modular Microreactor Assemblies

Article abstract of DOI:10.1002/anie.201916481

Immobilization of compartmentalized microscale objects in 3D hydrogels provides a step towards the modular assembly of soft functional materials with tunable architectures and distributed functionalities. Herein, we report the use of a combination of micro-compartmentalization, immobilization, and modularization to fabricate and assemble hydrogel-based microreactor assemblies comprising millions of functionalized polysaccharide–polynucleotide coacervate droplets. The heterogeneous hydrogels can be structurally fused by interfacial crosslinking and coupled as input and output modules to implement a UV-induced photocatalytic/peroxidation nanoparticle/DNAzyme reaction cascade that generates a spatiotemporal fluorescence read-out depending on the droplet number density, intensity of photoenergization, and chemical flux. Our approach offers a route to heterogeneous hydrogels with endogenous reactivity and reconfigurable architecture, and provides a step towards the development of soft modular materials with programmable functionality.

Full text of DOI:10.1002/anie.201916481

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Hydrogel-immobilized coacervate droplets as modular micro-
reactor assemblies
Authors: Jianbo Liu, Liangfei Tian, Yan Qiao, Shaohong Zhou, Avinash
Patil, Kemin Wang, Mei Li, and Stephen Mann
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201916481
Angew. Chem. 10.1002/ange.201916481
Link to VoR: http://dx.doi.org/10.1002/anie.201916481
http://dx.doi.org/10.1002/ange.201916481

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
Hydrogel-immobilized coacervate droplets as modular micro-
reactor assemblies
Jianbo Liu,* [a,b] Liangfei Tian,[b] Yan Qiao,[b] Shaohong Zhou,[a] Avinash J. Patil,[b] Kemin Wang,[a] Mei
Li*[b] and Stephen Mann*[b]
Abstract: Immobilization of compartmentalized microscale objects in
three-dimensional hydrogels provides a step towards the modular
assembly of soft functional materials with tunable architectures and
distributed functionalities. Here we use a combination of micro-
compartmentalization, immobilization and modularization to fabricate
and assemble hydrogel-based micro-reactor assemblies comprising
millions of functionalized polysaccharide-polynucleotide coacervate
droplets. The heterogeneous hydrogels can be structurally fused by
interfacial crosslinking and coupled as input and output modules to
implement a UV-induced photocatalytic/ peroxidation nanoparticle/
functionally encoded hydrogel building blocks.[1a,5] Such
procedures have been primarily exploited for the printing of cell-
laden hydrogels from components such as hyaluronic acid[6] and
silk fibroin.[7]
Here we describe an approach to the fabrication and
assembly of soft matter micro-reactor modules based on the
hydrogel-mediated 3D immobilization of dispersed populations
of discrete molecularly crowded coacervate micro-droplets[8]
containing sequestered photocatalytic nanoparticles, enzymes
and DNAzymes. Aqueous suspensions of coacervate droplets
have been recently employed as membrane-free micro-reactors
capable of in vitro gene expression,[9] enhanced protein and
RNA catalysis,[10] electric field-induced energization,[11] enzyme-
based chemical signaling[12] and artificial predation.[13] Given this
wide range of biomimetic activities, we sought to exploit the
coacervate droplets as an alternative type of microscale agent
for developing multi-functional hydrogel-based modular micro-
reactor assemblies. We show that entrapment of the coacervate
droplets places diffusional and spatial constraints on their
localized reactivity such that fluxes of matter can be established
within the hydrogel-based modules. The soft modules can be
physically shaped and structurally fused by interfacial
crosslinking to produce interconnected multi-functional materials.
As proof-of-concept, we prepare a binary arrangement of input
and output modules that operate in sequence to implement a
DNAzyme reaction cascade that generates
a spatiotemporal
fluorescence read-out depending on the droplet number density,
intensity of photo-energization and chemical flux. Our approach
offers
a route to heterogeneous hydrogels with endogenous
reactivity and reconfigurable architecture, and provides a step
towards the development of soft modular materials with
programmable functionality.
Introduction
Immobilizing micro-compartmentalized synthetic objects such as
porous microbeads, liposomes, polymersomes and liquid
droplets in three-dimensional (3D) hydrogels offers diverse
opportunities for drug delivery and tissue engineering,[1] the
assembly of cell-like tissues[2] and organization of spatially
distributed microreactors for enhanced mechanical performance,
durability and reusability.[3] For example, supramolecular nested
microbeads were developed as building blocks to construct
macroscopic self-healing scaffolds,[1f] and enzyme-containing
hybrid microcapsules hierarchically organized in agarose
hydrogels to facilitate recycling of an immobilized biocatalyst.[3b]
As individual hydrogels can be shaped and assembled into
higher-order arrangements by 3D printing and controlled
UV-energized
nanoparticle/DNAzyme
photocatalytic/peroxidation cascade reaction under flow
conditions. Continuous photo-energization of the input module
transmits a hydrogen peroxide (H2O2) flux downstream, which in
turn activates the spatiotemporal production of a fluorescence
readout specifically in the coacervate droplets immobilized along
the length of the output module. We show that dynamical
changes between the activation and deactivation of fluorescence
as the H2O2 reaction-diffusion gradient advances along the
output module gives rise to a migrating fluorescence band
depending on the droplet number density, intensity of the photo-
activation and chemical flux originating from the droplet micro-
reactors.
crosslinking,[1a,4] they provide
fabrication of soft modular materials based on the integration of
a potential pathway to the
[a]
Dr J. Liu, S. Zhou, Prof. Dr K. Wang
State Key Laboratory of Chemo/Biosensing and Chemometrics,
College of Chemistry and Chemical Engineering,
Key Laboratory for Bio-Nanotechnology and Molecular Engineering
of Hunan Province,
Hunan University, Changsha 410082, P. R. China.
Email: liujianbo@hnu.edu.cn
Dr J. Liu, Dr L. Tian, Dr Y. Qiao, Dr A. J. Patil, Dr M. Li, Prof. Dr S.
Mann
Centre for Protolife Research and Centre for Organized Matter
Chemistry,
Results and Discussion
Preparation of Hydrogel-Based Droplet Micro-Reactor
Modules
Hydrogel-based droplet micro-reactor modules were produced
by in situ immobilization of millions of intact polysaccharide-
[b]
polynucleotide
coacervate
micro-droplets
containing
sequestered photoactive and catalytic components. Coacervate
micro-droplets were pre-prepared at pH 8.0 and room
temperature by mixing aqueous solutions of positively charged
diethylaminoethyldextran chloride (DEAE-dextran, Mw = 500
kDa) and double-stranded (ds) DNA at a DEAE-dextran : DNA
ratio of 2.4 : 1 (w/w) (Supporting Information, Figure S1). The
School of Chemistry, University of Bristol, Bristol BS8 1TS, UK.
Email: s.mann@bristol.ac.uk, Mei.Li@bristol.ac.uk
Supporting information for this article is given via a link at the end of
the document.
1
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
molecularly crowded droplets were near-neutral in surface
charge (+4.1 mV), polydisperse in size, less than 20 μm in
diameter (mean size = 5.3 μm) and exhibited homogeneous
green fluorescence after sequestration of the DNA-binding dye
SYBR Green (Supporting Information, Figure S2). Capture and
immobilization of the DEAE-dextran/dsDNA coacervate droplets
at typical loadings of ca. 106 droplets per mL was achieved by
slowly cooling a hot aqueous agarose solution containing the
micro-droplets from 60C to room temperature (Figures 1a-c).
Optical microscopy images showed discrete spherical droplets
within the hydrogel that remained unchanged in shape and size
for at least 30 days after immobilization (Figure 1d). Increasing
the loading to ca. 107 droplets mL-1 had negligible effect on
(elastic) modulus (G/) of ca. 4.3 ×103 Pa with G/ greater than the
loss (viscous) modulus (G//) for shear strain values less than
4.0% (Supporting Information, Figure S5).
Hydrogel-based
modules
comprising
immobilized
populations of functional coacervate droplets were prepared by
partitioning payloads into the molecularly crowded interior of the
coacervate micro-droplets prior to hydrogelation. A wide range
of functional components were spontaneously sequestered into
the DEAE-dextran/dsDNA droplets with high partition
coefficients (K). For example: organic dyes (calcein, K = 55;
rhodamine 6G, K = 90); enzymes (fluorescein isothiocyanate-
labelled glucose oxidase, FITC-GOx, K = 175; horseradish
peroxidase, HRP, K = 210); enzyme substrates and products
(Amplex red, K = 170; resorufin, K = 160); single-stranded (ss)
oligonucleotides (cyanine-5 (Cy5)-ssDNA, K = 560), a G-
quadruplex (G4-hemin DNAzyme, K = 145); and inorganic
nanoparticles (N-acetylcysteine-capped cadmium telluride
quantum dots, NAC-CdTe, K = 340; photocatalytic TiO2/Ag
hybrid clusters, K = 770). In each case, the functional payloads
were spatially localized at the micrometre scale within the
agarose matrix by compartmentalization within the entrapped
DEAE-dextran/dsDNA droplets (Figures 2a-f and Supporting
Information, Figure S6). Subjecting the hydrogel modules to gel
electrophoresis showed minimal movement of the guest
components out of the agarose matrix (Supporting Information,
Figure S7), consistent with their high binding affinity to the
coacervate phase.
hydrogel formation and produced
a densely crowded 3D
distribution of the embedded droplets (Figure 1e). Confocal laser
scanning microscopy images indicated that the entrapped
droplets were homogeneously distributed throughout the
hydrogel matrix (Figure 1f). In contrast, coacervate droplets
prepared in the absence of a hydrogel sedimented onto a glass
slide within ca. 30 min (Figure 1g). These observations were
consistent with SEM images of dried samples, which showed
intact or collapsed DEAE-dextran/dsDNA droplets in the
presence or absence of an agarose hydrogel, respectively
(Supporting Information, Figure S3).
Temperature-dependent
rheological
measurements
showed negligible changes in the solid-like viscoelastic
properties of the agarose hydrogels with or without the
embedded coacervate micro-droplets. In both cases, thermal
hysteresis was observed for the gel-to-sol transition with gelling
and melting temperatures of approximately 33.0 ± 1.0 and 90.0 ±
1.0 C, respectively (Supporting Information, Figure S4). Plots of
the shear modulus against strain gave values for the storage
Functional Modules and Modular Assembly
We exploited the high loading capacity of the DEAE-
dextran/dsDNA coacervate droplets to prepare
a single
Figure 1. Preparation of a hydrogel-based droplet micro-reactor module. a) Scheme illustrating capture and immobilization of DEAE-dextran/dsDNA coacervate
micro-droplets in an idealized agarose hydrogel to produce a hydrogel-based droplet micro-reactor module. b,c) Photographs of a coacervate micro-droplet
aqueous suspension (DEAE-dextran (12.0 mg mL-1)/dsDNA (5.0 mg mL-1)) before (b) and after in situ hydrogelation (c). Hydrogelation was initiated by addition of
an equal volume of 1 wt% agarose at 60C followed by cooling to room temperature; scale bars: 20 mm. d,e) Optical microscopy images of micro-reactor modules
comprising ca. 106 (d) or 107 coacervate droplets mL-1 (e) embedded within an agarose hydrogel (0.5 wt%) after storage for 30 days at room temperature; scale
bars; 10 µm. f,g) Confocal fluorescence microscopy images of a population of DEAE-dextran/DNA coacervate droplets showing a homogeneous 3D distribution
when immobilized within the agarose hydrogel (f); in contrast, the droplets sediment onto a glass slide in the absence of a hydrogel (g). Coacervate droplets were
stained with the DNA binding dye SYBR-green; scale bars: 20 µm.
2
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
hydrogel-based droplet micro-reactor module capable of
processing a localized two-step enzyme cascade. For this, we
co-sequestered GOx, HRP and Amplex red in the droplets,
embedded the loaded droplets within the hydrogel, and then
added glucose to trigger GOx-mediated formation of hydrogen
peroxide (H2O2). Subsequent H2O2/HRP-mediated peroxidation
of Amplex red to resorufin was confirmed by fluorescence
microscopy images, which showed red fluorescence
predominantly within the immobilized coacervate micro-droplets
(Figure 2g). The kinetic profiles were fitted to a Michaelis–
Menten model with values for the Michaelis constant (Km) and
turnover number (kcat) of 120 mM, and 120 s-1, respectively
(Figure 2h and Supporting Information, Figure S8). These values
were different from those determined in an agarose hydrogel
matrix prepared in the absence of the coacervate droplets (Km,
82 mM; Kcat, 190 s-1), suggesting that the glucose binding
efficiency was reduced in the modules, possibly due to partial
deactivation of the enzymes in the presence of high
polyelectrolyte concentrations. However, direct comparison of
the kinetic profiles was ambiguous given that the effective
concentrations of enzymes and substrates partitioned within the
coacervate droplets were higher than those in the bulk phase,
and that diffusional limitations due to rapid substrate depletion in
the coacervate droplets could not be ruled out. Nevertheless, it
seemed clear that the large loading capacity of the coacervate
microdroplets for GOx and HRP (K = 175 and 210, respectively)
gave rise to enhanced levels of enzyme activity in the micro-
reactor module compared with hydrogels loaded with the same
enzyme concentrations but prepared in the absence of
coacervate micro-droplets (Supporting Information, Figure S9).
Given the above observations, we constructed self-
supporting
stacks
of
DEAE-dextran/dsDNA
hydrogels
comprising multiple zones of matrix-immobilized coacervate-
based droplet micro-reactors. For this, alginate (1 wt%;
agarose/alginate volume ratio = 1 : 1), was added to the agarose
hydrogels to enable sequential adhesion of individual reaction
modules via Ca2+-induced crosslinking (Supporting Information,
Figure S10). Three separate agarose/alginate hydrogels
containing embedded coacervate micro-droplets loaded with
FITC-GOx, NAC-CdTe or Cy5-DNA were prepared individually in
16 mm-diameter glass tubes comprising a detachable stopper.
Each material was gently removed using a glass rod and the
modules attached at their ends in a stepwise fashion by
immersion in aqueous 50 mM CaCl2 solution for 4 h at room
temperature. The resulting cylindrical stack comprised discrete
zones of green, yellow and red fluorescence that were joined
0.2
0.1
0
0
100 200 300 400
Glucose (mM)
Figure 2. Functional modules and modular assembly. a,b) Fluorescence microscopy images of a bulk hydrogel containing an entrapped population of FITC-GOx-
loaded DEAE-dextran/dsDNA coacervate droplets ((a), module 1), and corresponding confocal image of the homogeneous 3D distribution of immobilized droplets
(b) (green fluorescence, FITC-GOx; 106 droplets mL-1); scale bars = 10 mm (a) and 20 µm (b). c-f) As for (a) and (b) but for hydrogel-based modules 2 (c,d) and 3
(e,f) functionalized with NAC-CdTe quantum dots (yellow fluorescence) or Cy5-DNA (red fluorescence), respectively; scale bars, 10 mm (c,e) and 20 µm (d,f). g)
Fluorescence microscopy image showing red fluorescence associated with resorufin production specifically within hydrogel-immobilized coacervate droplets
containing GOx, HRP and Amplex red after addition of glucose (10 mM); scale bar, 20 µm. h) Plots of initial rate of enzyme activity against glucose concentration
for GOx/HRP cascade reactions in a hydrogel matrix (0.5% wt agarose, pH = 7.4, green triangles) or within hydrogel-immobilized coacervate droplets (0.5% wt
agarose, ca. 106 coacervate droplets mL-1, pH=7.4, blue circles). (GOx, 33 mU mL-1 (added concentration, Co) = 486 mU mL-1 (sequestered concentration, Cs);
HRP, 1 U mL-1 (Co) and 15 U mL-1 (Cs); Amplex red, 16.7 µM (Co) and 250 µM (Cs). i-k) Three-level stack prepared from sequential Ca2+-mediated crosslinking of
modules 1, 2 and 3; graphic shows the idealized 3D arrangement of the stack (i), and optical and fluorescence microscopy images are shown in (j) and (k),
respectively; scale bars, 10 mm. l) Scanning confocal fluorescence microscopy image of the adhesion interface (dotted white line) between modules 1 and 2. The
imaged area is denoted by the small rectangular white box shown in (k); scale bar: 100 µm. m) Fluorescence microscopy images recorded 12 h after immersion of
a linear stack of modules 1, 2 and 3 in water in the dark showing no changes in fluorescence intensities; scale bar, 10 mm. n) As for (m) but in the absence of
entrapped coacervate droplets showing ca. 37% decrease in fluorescence intensity at the boundaries of the modules due to inter-zonal diffusion; scale bar, 10
mm.
3
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
continuously across their interfaces (Figures 2i-k). Scanning
confocal fluorescence microscopy images of the adhesion
interfaces showed a clear demarcation in the positioning of the
sequestered FITC-GOx, NAC-CdTe or Cy5-DNA components
(Figure 2l). Moreover, time-lapse images recorded over a period
of 12 h showed no significant changes in the fluorescence
intensities associated with the different modules, indicating that
the constituent coacervate micro-droplets and their functional
components remained immobilized within the stacked hydrogel
matrix (Figure 2m and Figure S11). In contrast, FITC-GOx, NAC-
CdTe and Cy5-DNA migrated across the hydrogel boundaries in
the absence of the entrapped coacervate micro-droplets (Figure
2n). We attribute these differences to the high sequestration
potential of the molecularly crowded DEAE-dextran/dsDNA
droplets. This was consistent with a series of experiments that
showed that release of the guest molecules from the
coacervate-containing hydrogel matrix into the bulk aqueous
phase was strongly inhibited when compared with leaching rates
indicated that the DNAzyme activity was reduced within the
coacervate micro-droplets whether in suspension (initial rate, 60
nM s-1) or immobilized within the hydrogel (11 nM s-1) compared
with the value determined in free solution (130 nM s-1) (Figure
3e).
Having established that the individual hydrogel-based
droplet micro-reactor modules were functionally active, we
prepared input and output modules with disk-like and
rectangular morphologies, respectively, and connected them in
series by interfacial crosslinking. In so doing, we produced an
arrangement
capable
of
implementing
a
photocatalytic/peroxidation cascade under reaction-diffusion
conditions (Supporting Information, Figure S15). Continuous
exposure of the upstream input module to UV irradiation resulted
in TiO2/Ag nanoparticle-mediated photocatalytic reduction of
dioxygen within the embedded coacervate droplets followed by
mass transport of H2O2 into the output module. Subsequent
oxidation of Amplex red to resorufin by G4-hemin-mediated
measured for coacervate-free hydrogels (Supporting Information, peroxidation within the immobilized coacervate droplets gave
Figure S12).
rise to a red fluorescence output specifically in the output
module. Moreover, because the resorufin fluorescence was
quenched at high concentrations of H2O2 by molecular
degradation to resazurin,[16] the high H2O2 diffusion gradient
generated in the input module produced a fluorescent band that
migrated along the reaction-diffusion field in the output domain
(Figure 3f). This effect was attributed to dynamical changes
between the activation and deactivation of fluorescence as the
H2O2 chemical gradient advanced along the output module
during which resorufin was produced by G4-hemin activity and
then degraded in situ as the H2O2 concentration reached a
critical value (typically above 500 µM; Supporting Information,
Figure S16).
Photocatalytic/Peroxidation Serial Processing Under Flow
Conditions
Given that individual hydrogel-based droplet micro-reactor
modules could be structurally integrated by interfacial
crosslinking, we prepared a construct comprising a binary
combination of modules capable of serial processing a UV-
initiated
nanoparticle/DNAzyme
photocatalytic/peroxidation
cascade reaction under flow conditions (Figures 3a,b). The
upstream input module consisted of a population of hydrogel-
immobilized
containing high concentrations of sequestered positively
charged TiO2/Ag nanoparticles (K = 770) capable of UV-light
induced reduction of dioxygen to H2O2 (Supporting Information,
Figures S13 and S14).[14] We tested the viability of the input
module by determining the yield of H2O2 produced by photo-
excitation under UV irradiation (365 nm, 60 min) using an ABTS
DEAE-dextran/dsDNA
coacervate
droplets
We tested the coupling of the photocatalytic input and
peroxidase output modules under different conditions. In general,
optical and fluorescence monitoring of the interface between the
two micro-reactor modules showed a clear demarcation in the
formation of resorufin specifically at the output side of the
colorimetric assay in the absence of any auxiliary electron donor. boundary (Supporting Information, Figure S17). Continuous
We compared the results with yields obtained under the same
conditions but from aqueous dispersions of TiO2/Ag
nanoparticle-containing coacervate droplets or TiO2/Ag
nanoparticles alone (Figure 3c); in both cases, significant and
similar yields of H2O2 were observed (38 and 30 µM,
respectively), indicating that sequestration of the nanoparticles
within the coacervate droplets did not deactivate the
photocatalytic components. Embedding the TiO2/Ag-loaded
coacervate droplets in an agarose/alginate hydrogel slightly
reduced the photocatalytic yield to a value of 28 µM, possibly
due to scattering of the UV light by the hydrogel matrix.
exposure of the input module to UV-light produced a diffusive
downstream flow of H2O2 such that a time-dependent red
fluorescence readout migrated progressively along the length of
the output module. For example, under close to optimum
conditions (106 droplets mL-1; input module, 50 mW UV, TiO2/Ag
= 1 mg mL-1; output module, G4-hemin = 0.5 µM, Amplex red =
25 µM), the fluorescence intensity, which attained a steady state
value after 15 min, migrated along the output module at a linear
rate of ca. 2.0 µm s-1 (Figure 3f and Movie S1). Significantly, as
shown in Figure 3f, the tail of the fluorescence output was
progressively quenched with time such that the readout became
spatiotemporally organized to produce a moving front that
increased marginally in width along the chemical gradient
(Figure 3g). High resolution fluorescence profiles recorded at
different positions along the length of the output module
indicated that the fluorescence was activated locally after a
specific delay time, increased in intensity to a maximum value as
the localized Amplex red was consumed and resorufin produced,
and then decayed at a relatively slower rate over a period of
100-150 min as the H2O2 increased in concentration (Figure 3h).
As a consequence, the width of the migrating fluorescence band
remained essentially constant when the rates of fluorescence
activation and deactivation were approximately equal at the front
The downstream output module comprised
a similar
arrangement of hydrogel-immobilized coacervate droplets as
present in the input module but in place of the inorganic
nanoparticles the micro-droplets were loaded with Amplex red (K
= 170) and a single-stranded G4-hemin DNAzyme (K = 145)
capable of functioning as a peroxidase-mimic in the presence of
H2O2.[15] As a consequence, addition of H2O2 to the module gave
rise to red fluorescence (resorufin production) specifically within
the embedded DEAE-dextran/dsDNA coacervate droplets
(Figure 3d), indicating that the G4-hemin DNAzyme remained
active after sequestration and immobilization. Time-dependent
changes in the fluorescence intensity recorded at 585 nm
4
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
and tail of the band, respectively. In practice, both processes
were accompanied by peak broadening as the band moved
along the output module due to shear-induced dispersion of the
chemical gradient (Supporting Information, Figure S18).
fluorescence intensity was recovered in this time period for
resorufin-containing
consequence, UV-energization of an identical system fabricated
in the absence of the coacervate droplets produced
coacervate-free
hydrogels.
As
a
a
We used a fluorescence recovery after photobleaching
(FRAP) experiment to confirm that diffusion of the resorufin
signal along the output module was severely restricted in the
presence of the coacervate micro-droplets (Supporting
Information, Figure S19). No fluorescence recovery was
observed over a period of 250 s in the presence of the
immobilized coacervate droplets, consistent with strong
partitioning of the fluorescence read-out specifically within the
droplet micro-reactors. In contrast, ca. 77% of the red
continuous red fluorescence band along the entire output
module due to unrestricted diffusive mixing of Amplex red and
resorufin (Supporting Information, Figure S20). Similar
observations were recorded when the number of hydrogel-
immobilized droplets was reduced from 106 to 105 droplets mL-1
(Supporting Information, Figure S21), indicating that emergence
of the migrating resorufin front was dependent on a critical
number density of embedded micro-reactor droplets.
a
d
hv
H2O2 transport
16
3
2
X-axis
12
8
Filter paper
hv
1
0
G4-hemin
Amplex red
4
0
380 420 460 500
0
100 200 300
Time (s)
Wavelength (nm)
H2O2
g
5 min
Output module
Filter paper
Input module
Input module
10 min
15 min
25 min
40 min
55 min
75 min
95 min
120 min
150 min
185 min
Output module
0
3
6
9
12
X-axis (mm)
8
12
0
2
4
6
10
14
16
5 min
X-axis (mm)
10 min
15 min
25 min
40 min
55 min
75 min
95 min
120 min
150 min
185 min
h
i
X=5 mm
X=12 mm
(2)
(3)
12
8
(1)
4
(4)
0
0
0
70
140
Time (min)
210
280
30
60
Time (min)
90
120
Figure 3. Photocatalytic/peroxidation serial processing under flow conditions. a) Graphical representations: linear arrangement of two hydrogel-based droplet
micro-reactor modules operating collectively as a photocatalytic/peroxidation cascade (top) under flow conditions, and details of the upstream input (TiO2/Ag
nanoparticles, yellow image, bottom left) and downstream output (G4-hemin DNAzyme, red image, bottom right) modules. Exposure of the input module to UV
irradiation results in H2O2 production followed by downstream diffusive flow into the output module and subsequent formation of a red fluorescence readout
(resorufin). Unidirectional liquid flow along the x axis is facilitated by attaching a strip of filter paper downstream of the output module and by gradual addition of
water in front of the input zone. b) Photograph showing experimental set-up. Typically, the input and output modules were 16 mm and 20 mm in diameter and
length, respectively. c) UV-vis absorption spectra showing bands at 420 nm corresponding to the formation of oxidized [ABTS·-] due to photocatalytic H2O2
production in the presence of TiO2/Ag nanoparticles dispersed in bulk solution (blue), or sequestered within coacervate droplets dispersed in aqueous solution
(red), or immobilized within an agarose/alginate hydrogel (input module, green); a.u = arbitrary units. d) Fluorescence microscopy image (excitation, 561 nm;
emission, 585 nm) of the output module after addition of aqueous H2O2 (10 µL, 10 mM) showing red fluorescence (resorufin) in the coacervate droplets; scale bar,
20 µm. e) Time-dependent plots showing increases in resorufin concentration associated with DNAzyme activity in aqueous solution (blue), sequestered within
coacervate droplets dispersed in aqueous solution (red), or immobilized within an agarose/alginate hydrogel (output module, green). f) Time-lapse fluorescence
microscopy images recorded from the output module showing the downstream propagation of a resorufin red fluorescence band (x axis; x = 0 at input/output
interface). Experimental conditions: 106 droplets mL-1; input module (not imaged), 50 mW UV (continuous), TiO2/Ag = 1 mg mL-1 (added concentration, Co) and
15.7 mg mL-1 (sequestered concentration, Cs); output module, G4-hemin = 0.5 µM (Co) and 7.3 µM (Cs), Amplex red = 25 µM (Co) and 368 µM (Cs). g) Plots of red
fluorescence intensity along the output module (x axis) as shown in (f) recorded at various time intervals over a period of 185 min. h) Corresponding time-
dependent changes in red fluorescence at x = 5 mm (red) and x = 12 mm (blue) showing fluorescence activation and quenching. A period of ca. 55 min separates
the two resorufin readouts. i) Plots of fluorescence peak positions with time for resorufin bands produced under different reaction conditions. All experiments
performed with continuous UV irradiation, 106 droplets mL-1, TiO2/Ag = 1 mg mL-1 (Co) and 15.7 mg mL-1 Cs) and G4-hemin = 0.5 µM (Co) and 7.3 µM (Cs) with the
following specifications: 1 (red triangles), 50 mW UV, Amplex red = 25 µM (Co) and 368 µM (Cs); 2 (green squares), 100 mW UV, Amplex red = 25 µM (Co) and
368 µM (Cs); 3 (blue diamonds), 50 mW UV, Amplex red = 8 µM (Co) and 118 µM (Cs); 4 (yellow circles), 50 mW UV, Amplex red = 25 µM (Co) and 368 µM (Cs),
catalase = 10 pM (Co) and 147 pM (Cs). Constant migration speeds of 2.0, 2.9, 4.2 and 0.7 µm s-1 were observed, respectively, along the x axis.
5
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
Other experiments indicated that the reaction-diffusion
process was also influenced by time-dependent changes in the
localized H2O2, Amplex red and resorufin concentrations. As
such, coupling between the input and output modules could be
modified by altering the fluxes used to establish the
photocatalytic/peroxidase cascade. For example, increasing the
power of the continuous UV-light source from 50 to 100 mW
increased the migration rate of the fluorescence band from 2.0
to ca. 2.9 µm s-1 due to the increase in the H2O2 concentration
gradient generated in the input module (Supporting Information,
Figure S22). The rate of propagation was increased to ca. 4.2
µm s-1 when the concentration of sequestered Amplex red was
decreased to 8 µM (Supporting Information, Figure S23). We
attributed this to an increase in the H2O2 flux associated with the
reduced levels of H2O2 depletion from the advancing
concentration gradient at higher [H2O2]/[Amplex red] and
[H2O2]/[resorufin] molar ratios. In contrast, a decrease in the
speed of the fluorescence band to ca. 0.7 µm s-1 was obtained
by co-sequestration of catalase into the coacervate droplets
contained within the output module (Supporting Information,
Figure S24). Under these conditions, catalase activity in the
droplets provided a competitive enzyme-mediated pathway for
consumption of the H2O2 as it diffused along the chemical
gradient.
that the hydrogel-based modules can be internally connected by
the flow of molecular intermediates and selectively interfaced
with the surrounding environment using UV-irradiation. As a
consequence,
we
demonstrate
a
UV-induced
photocatalytic/peroxidation nanoparticle/DNAzyme cascade with
a spatiotemporal fluorescence read-out specifically along the
output module, which depends on the droplet number densities,
intensity of photo-energization in the input module and the
downstream flux of H2O2. These factors influence the dynamical
changes between the activation and deactivation of fluorescence
as the H2O2 reaction-diffusion gradient advances along the
output module to produce a read-out in the form of a migrating
fluorescence band. The system operates efficiently due to the
relatively high diffusion coefficient of H2O2 but is likely to become
less effective when the carrier is a larger molecule. Taken
together, our work offers
heterogeneous hydrogels
a
with
promising approach to
endogenous reactivity,
reconfigurable architecture and programmable functionality for
potential uses in areas such as bioengineering and flow
chemistry.
Acknowledgements
A simplified model of the spatiotemporal distribution of
resorufin along the output module was developed to simulate the
emergence and propagation of the fluorescent band. Using a
finite explicit approach with closed boundary conditions we
calculated the mass balance at different locations of the
unidirectional H2O2 chemical gradient assuming that the
production and depletion of resorufin occurred in two discrete
steps and that H2O2 in the input module was photocatalytically
generated at a continuous rate of 0.1 µM s-1 (Supporting
Information, Figure S25). The calculated 1D time-dependent
distributions of resorufin in the simulated hydrogel-based reactor
migrated progressively along the length of the output module,
consistent with the experimental observations and in agreement
with a dynamical process involving both fluorescence activation
and deactivation.
We thank the BBSRC (BB/P017320/1), the ERC Advanced
Grant Scheme (EC-2016-ADG 740235), and BrisSynBio, a
BBSRC/EPSRC
Synthetic
Biology
Research
Centre
(BB/L01386X/1), for financial support. The work was partly
supported by the National Natural Science Foundation of China
(21190040, 21175035), J. Liu is grateful for financial support
from the China Scholarship Council.
Keywords: hydrogel • coacervate • micro-reactor • cascade
reaction
[1]
a) M.Y. Chiang, Y.W. Hsu, H.Y. Hsieh, S.Y. Chen, S.K. Fan, Sci. Adv.
2016, 2, e1600964; b) F. Pati, J. Jang, D.H. Ha, S. Won Kim, J.W. Rhie,
J.H. Shim, D.H. Kim, D.W. Cho, Nat Commun. 2014, 5, 3935; c) J. L.
Guo, Y. S. Kim, V. Y. Xie, B. T. Smith, E. Watson, J. Lam, H. A. Pearce,
P. S. Engel, A. G. Mikos, Sci. Adv. 2019, 5, eaaw7396; d) M. Marguet,
C. Bonduelle, S. Lecommandoux, Chem. Soc. Rev. 2013, 42, 512-529;
e) I. Kusters, N. Mukherjee, M.R. de Jong, S. Tans, A. Koçer, A.J.M.
Driessen, PLOS One 2011, 6, e20435; f) Z. Yu, J. Liu, C.S.Y. Tan, O.A.
Scherman, C. Abell, Angew. Chem. Int. Ed. 2018, 57, 3079-3083.
a) D. K. Baxani, A. J. L. Morgan, W. D. Jamieson, C. J. Allender, D. A.
Barrow, O. K. Castell, Angew. Chem. Int. Ed. 2016, 55, 14240-14245;
b) M. Mariam Bayoumi, H. Bayley, G. Maglia, K. T. Sapra, Sci. Rep.
2017, 7, 45167.
Conclusion
In conclusion, we describe an approach to the fabrication and
assembly of micro-reactor modules based on the hydrogel-
mediated immobilization of dispersed suspensions of
molecularly crowded DEAE-dextran/dsDNA coacervate micro-
droplets containing photocatalytic nanoparticles, enzymes and
DNAzymes. The ability of coacervates to sequester high
concentrations of guest components compared with
[2]
[3]
a) H.P. De Hoog, I.W. Arends, A.E. Rowan, J.J. Cornelissen, R.J. Nolte,
Nanoscale 2010, 2, 709–716; b) S. Zhang, Z. Jiang, J. Shi, X. Wang, P.
Han, W. Qian, ACS Appl. Mater. Interfaces 2016, 8, 25152−25161.
O.F. Khan, D.N. Voice, B.M. Leung, M.V. Sefton, Adv. Healthcare
Mater. 2015, 4, 113–120.
[4]
[5]
[6]
[7]
conventional
water-containing
droplets
prepared
by
a
emulsification or vesicularization for example offers
A. Nadernezhad, N. Khani, G.A. Skvortsov, B. Toprakhisar, E. Bakirci,
Y. Menceloglu, S. Unal, B. Koc, Sci. Rep. 2016, 6, 33178.
straightforward approach to droplet-based micro-reactors with
high cargo loading. Moreover, entrapment of high number
densities of the reactive coacervate droplets in agarose/alginate
K.M. Park, J.A. Yang, H. Jung, J. Yeom, J.S. Park, K.H. Park, A.S.
Hoffman, S.K. Hahn, K. Kim, ACS Nano 2012, 6, 2960-2968.
S.H. Kim, Y.K. Yeon, J.M. Lee, J.R. Chao, Y.J. Lee, Y.B. Seo, M.T.
Sultan, O.J. Lee, J.S. Lee, S.I. Yoon, I.S. Hong, G. Khang, S.J. Lee, J.J.
Yoo, C.H. Park, Nat Commun. 2018, 9, 1620.
polysaccharide hydrogels provides
a facile approach to
developing spatially segregated populations of micro-reactors
that can be assembled into input and output modules by
interfacial crosslinking. Immobilization of the droplets imposes
spatial and diffusional barriers on their localized reactivity such
[8]
a) S. Koga, D. S. Williams, A. W. Perriman, S. Mann, Nat. Chem. 2011,
3, 720-724; b) W. M. Aumiller, C. D. Keating, Nat. Chem. 2016, 8, 129-
6
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
137; c) K. K. Nakashima, J. F. Baaij, E. Spruijt, Soft Matter 2018, 14,
361-367.
[9]
D. T.-Y. Tang, D. van Swaay, A. deMello, J. L. Ross Anderson, S.
Mann, Chem. Commun. 2015, 51, 11429-11432.
[10] a) J. Crosby, T. Treadwell, M. Hammerton, K. Vasilakis, M. P. Crump, D.
S. Williams, S. Mann, Chem. Commun. 2012, 48, 11832-11834; b) B.
Drobot, J. M. Inglesias-Artola, K. Le Vay, V. Mayr, M. Kar, M. Kreysing,
H. Mutschler, T.-Y. D. Tang, Nat. Commun. 2018, 9, 3643; c) R. R.
Poudyal, R. M. Guth-Metzler, A. J. Veenis, E. A. Frankel, C. D Keating,
P. C. Bevilacqua, Nat. Commun. 2019, 10, 490.
[11] Y. Yin, L. Niu, X. Zhu, M. Zhao, Z. Zhang, S. Mann, D. Liang, Nat.
Commun. 2016, 7, 10658.
[12] L. Tian, N. Martin, P. G. Bassindale, A. J. Patil, M. Li, A. Barnes, B. W.
Drinkwater, S. Mann, Nat. Commun. 2016, 7, 13068.
[13] Y. Qiao, M. Li, R. Booth, S. Mann, Nat. Chem. 2017, 9, 110-119.
[14] D. Tsukamoto, A. Shiro, Y. Shiraishi, Y. Sugano, S. Ichikawa, S.
Tanaka, T. Hirai, ACS Catal. 2012, 2, 599–603.
[15] P. Travascio, A.J. Bennet, D.Y. Wang, D. Sen, Chem. Biol. 1999, 6,
779-787.
[16] V. Towne, M. Will, B. Oswald, Q. Zhao, Anal. Biochem. 2004, 334,
290–296.
7
This article is protected by copyright. All rights reserved.

10.1002/anie.201916481
Angewandte Chemie International Edition
RESEARCH ARTICLE
Entry for the Table of Contents
A binary arrangement of connected
input and output hydrogel-based
droplet micro-reactor modules is
constructed and used to implement a
UV-induced photocatalytic/
peroxidation nanoparticle/DNAzyme
cascade with a spatiotemporal
fluorescence read-out.
Jianbo Liu,* Liangfei Tian, Yan Qiao,
Shaohong Zhou, Avinash J. Patil,
Kemin Wang, Mei Li* and Stephen
Mann* Page No. – Page No.
Hydrogel-immobilized coacervate
droplets as modular micro-reactor
assemblies
8
This article is protected by copyright. All rights reserved.