होम Analytical Chemistry Raman-activated Droplet Sorting (RADS) for Label-free High-throughput Screening of Microalgal...

Raman-activated Droplet Sorting (RADS) for Label-free High-throughput Screening of Microalgal Single-cells

, , , , , , , , , , , ,
यह पुस्तक आपको कितनी अच्छी लगी?
फ़ाइल की गुणवत्ता क्या है?
पुस्तक की गुणवत्ता का मूल्यांकन करने के लिए यह पुस्तक डाउनलोड करें
डाउनलोड की गई फ़ाइलों की गुणवत्ता क्या है?
भाषा:
english
पत्रिका:
Analytical Chemistry
DOI:
10.1021/acs.analchem.7b03884
Date:
November, 2017
फ़ाइल:
PDF, 1.13 MB
0 comments
 

To post a review, please sign in or sign up
आप पुस्तक समीक्षा लिख सकते हैं और अपना अनुभव साझा कर सकते हैं. पढ़ूी हुई पुस्तकों के बारे में आपकी राय जानने में अन्य पाठकों को दिलचस्पी होगी. भले ही आपको किताब पसंद हो या न हो, अगर आप इसके बारे में ईमानदारी से और विस्तार से बताएँगे, तो लोग अपने लिए नई रुचिकर पुस्तकें खोज पाएँगे.
Subscriber access provided by READING UNIV

Article

Raman-activated Droplet Sorting (RADS) for Label-free
High-throughput Screening of Microalgal Single-cells
Xixian Wang, Lihui Ren, Yetian Su, Yuetong Ji, Yaoping Liu, Chunyu Li,
Xunrong Li, Yi Zhang, Wei Wang, Qiang Hu, Danxiang Han, Jian Xu, and Bo Ma
Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03884 • Publication Date (Web): 03 Nov 2017
Downloaded from http://pubs.acs.org on November 4, 2017

Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth
Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim ; is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Analytical Chemistry

Raman-activated Droplet Sorting (RADS) for Label-free Highthroughput Screening of Microalgal Single-cells
Xixian Wang,a,c,‡ Lihui Ren,a,c,‡ Yetian Su,a Yuetong Ji,a,c Yaoping Liu,d Chunyu Li, a,c Xunrong Li,a,c Yi
Zhang,a Wei Wang,d Qiang Hu,b Danxiang Han,b Jian Xu,a,c,* and Bo Maa,c,*
a

Single-Cell Center, CAS Key Laboratory of Biofuels and Shandong Key Laboratory of Energy Genetics, Qingdao Institute
of Bioenergy and Bioprocess Technology, Chinese Academy of Sciences, Qingdao, Shandong 266101, China
b
Center for Microalgal Biofuels and Biotechnology, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, Hubei 430072, China
c
University of Chinese Academy of Sciences, Beijing, 100049, China
d
National Key Laboratory of Science and Technology on Micro/Nano Fabrication, Institute of Microelectronics, Peking
University, Beijing, 100871, China
ABSTRACT: Raman-activated cell sorting (RACS) has attracted increasing interest, yet throughput remains one major factor limiting its broader application. Here we present an integrated Raman-activated droplet sorting (RADS) microfluidic system for functional screening of live cells in a label-free and high-throughput manner, by employing AXT-synthetic industrial microalga Haematococcus pluvialis (H. pluvialis) as a model. Raman microspectroscopy analysis of individual cells is carried out prior to their microdroplet encapsulation, which is then directly coupled to DEP-based droplet sorting. To validate the system, H. pluvialis cells
containing different levels of AXT were mixed and underwent RADS. Those AXT-hyperproducing cells were sorted with an accuracy of 98.3%, an enrichment ratio of eight folds and a throughput of ~260 cells/min. Of the RADS-sorted cells, 92.7% remained
alive and able to proliferate, which is equivalent to the unsorted cells. Thus the RADS achieves a much higher throughput than existing RACS systems, preserves the vitality of cells, and facilitates seamless coupling with downstream manipulations such as single-cell sequencing and cultivation.

A single cell is the basic unit of function and evolution for
cellular lives on Earth. Thus accurate and high-throughput
sorting of single-cells are key tools for mechanistic dissection
of functional heterogeneity among cells and probing yet-to-becultured microbes in nature.1-3 A Single-cell Raman Spectrum
(SCRS) provides the intrinsic biochemical profile of a cell at a
given state, thus can be considered as a function-based instant
portrait-photo of the cell.4 Due to its label-free and noninvasive features, Raman-activated cell sorting (RACS), which
sorts cells based on their SCRS, has attracted increasing interest.
Recently, a series of core technologies and devices for
RACS have emerged.5-7 By coupling Raman microspectroscopy with optical tweezers (OT) and laser ejection, respectively,
static versions of RACS called Raman tweezers8,9 and Ramanactivated cell ejection (RACE)10,11 were developed. Although
these systems were simple and practical, the relatively low
sorting throughput greatly hampers their broader applications.
To achieve higher throughput, flow-mode RACS named Raman-activated microfluidic sorting (RAMS) were introduced
based on a trap-and-release strategy.12-16 Among these optical
tweezer based RAMS systems, cells were immobilized for
Raman measurement and then targeted cells were dragged to a
designated location for collection.12-16 However, the very low
throughput (~3 min per cell5) of optical tweezers was still an
unsolved problem; moreover, photo-damage to the target cells

can be significant when laser tweezers of visible wavelengths
were used.17
To tackle these limitations which are inherent to optical
tweezers, we have recently developed a high-speed RACS
system that combines positive dielectrophoresis (pDEP) for
the efficient trap-and-release of cells with a solenoid-valvesuction-based switch18 for cell separation.19 Compared to the
optical tweezers based RACS, throughput of our RACS system, at approximately one cell per second, represents a significant increase. However it was difficult to further increase the
throughput, due to the response delay of the solenoid and the
time delay required to resume stability of the fluids after suction. On the other hand, McIlvenna et al. proposed a pressuredividers-based switch to sort the cells with high accuracy, yet
the low throughput (0.5 Hz, i.e., one cell every two seconds)
was still recognized as one critical issue that remains to be
resolved.20
Droplet-based microfluidics has shown substantial promises
in the past few years.21 In fluorescence-activated cell sorting
(FACS), coupling compartmentalization of individual cells in
microdroplets and DEP-based cell sorting can achieve ultrahigh throughput, and moreover can be readily coupled to
downstream droplet-based DNA/RNA extraction or cell culture,22-24 e.g., isolation of cells in droplets can improve recovery rate of slow-growing species.25 However, droplet-based
microfluidic RACS prototypes have not been demonstrated,
which may be due to several reasons: (i) the convex/concave

ACS Paragon Plus Environment

Analytical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

shape of the droplet surface creates a lensing effect that distorts the focus and reduces the spatial resolution, making it
difficult to accurately acquire the Raman signals of dropletencapsulated cells; (ii) the strong Raman background of oil
medium can adversely affect the system’s ability to accurately
distinguish the Raman signals of droplet-encapsulated cells;
(iii) the difficulty in integrating SCRS acquisition and analysis,
cell encapsulation and sorting within an automated system.
In this work, we established a droplet-based microfluidic
RACS named Raman-activated droplet sorting (RADS). Key
features of the RADS system include: (i) cells were interrogated for SCRS prior to droplet generation, which eliminates the
interference from droplet surface and oil; (ii) single-cell droplet encapsulation was directly coupled to DEP-based droplet
sorting, which simplifies the system design and increases sorting throughput and accuracy. Consequently, all single-cells
were encapsulated into droplets after Raman detection, while
only those droplets that contain the target cells would trigger
the DEP after a specific delay and thus be sorted into the collecting channel. Furthermore, a homemade software, QSpec,
was developed to automatically integrate and synchronize the
RADS system.
As a proof of concept for RADS, we employed microalgal
strain selection for astaxanthin (AXT) as a model. AXT is one
of the most potent antioxidants known has found great commercial potential in aquaculture, pharmaceutical and food industries. AXT is naturally produced by Haematococcus pluvialis (H. pluvialis),26 a unicellular freshwater green microalga,
thus the ability to rapidly screen AXT-hyperproducing H. pluvialis cells is pivotal to large-scale production of AXT. Traditionally, the screening process is based on cell culture (e.g.,
plate-based colony cultivation followed by liquid culture),
which however has been time-consuming and low-throughput
due to the low growth rate and susceptibility to contamination
of H. pluvialis cells.27 Via an improved FACS technique for
AXT, Ukibe et al were able to screen AXT-hyperproducing
mutants of the yeast Xanthophyllomyces dendrorhous.28 However they failed to screen AXT-hyperproducing H. pluvialis
cells, because fluorescence intensity of the large amount of
chlorophyll in microalgal cells masks that of AXT. In addition,
infrared (IR) spectroscopy was used for screening AXThyperproducing H. pluvialis cells.29,30 However, either AXT
extraction from cells with organic solvents or cell drying on
BaF2 substrates at 40 °C was required by IR, which makes the
procedure very tedious and in vivo screening of live cells not
practical. Unlike IR, Raman spectroscopy is friendly to aqueous samples and thus ideal for analyzing live cells, as water
molecules do not exhibit strong Raman scattering features. As
a result, Raman spectroscopy has been employed for quantitative profiling distribution, structural change and in vivo kinetics of AXT in H. pluvialis cells in a non-invasive and labelfree manner.30-34
To validate the RADS system, H. pluvialis cells containing
different levels of AXT were mixed and underwent RADS.
Those AXT-hyperproducing cells were sorted with an accuracy of 98.3%, an enrichment ratio of eight folds and a throughput of ~260 cells/min. Of the RADS-sorted cells, 92.7% remained alive and able to proliferate, which is equivalent to the
unsorted cells. Thus the RADS achieves a much higher
throughput than existing RACS systems, preserves the vitality
of cells, and facilitates seamless coupling with downstream
manipulations such as single-cell sequencing and cultivation.

Page 2 of 11

EXPERIMENTAL SECTION
Materials and Reagents.
Chemicals such as ethanol, isopropanol, Span 80 were purchased from Sinopharm Chemical Reagent (Shanghai, China).
Mineral oil, L-asparagine, yeast extract, propidium iodide (PI)
PF 127 and Na2EDTA·2H2O were purchased form SigmaAldrich (St. Louis, MO, USA). SU-8™ (GM 3025) was purchased from MicroChem (Massachusetts, USA). Polydimethylsiloxane (PDMS Sylgard 184) and curing agent (Sylgard
184) were purchased from Dow Corning (Midland, MI, USA).
All regents used in our experiment were analytical grade unless otherwise stated. All solutions were prepared with deionized water and were sterilized by filtration (0.22 µm microporous membrane filtration) or via an autoclave (121 °C,
30 min) prior to use.
Cell Culture and Preparation.
H. pluvialis (NIES-144) was purchased from the National
Institute for Environmental Studies (NIES, Japan). Single colonies from basal medium plates35 were inoculated into a liquid
basal medium at 22 °C under continuous low light illumination (20 µmol photons m-2 s-1) with shaking manually once per
day. To induce AXT accumulation, exponentially growing
cells were re-suspended into modified BBM medium (without
nitrogen source and containing 10 mM sodium acetate) in
triplicate, and exposed to continuous illumination of 150 µmol
photons m-2 sec-1 for various days (to generate a gradient of
averaged AXT content in the cells). Batches of cells, with
various average AXT contents, were harvested and filtered
with a cell strainer (40 µm microporous membrane filtration)
to remove debris and cell clusters, and then were washed with
sterile deionized water under 1000 g. Cell concentration was
measured using a cell count plate and was adjusted to ~8.02 ×
106 cells mL-1. Then cells from batches with different averaged
AXT contents were mixed to achieve a specific cell ratio. The
actual cell ratio was later determined by sampling >100 cells
using the Raman Points Mapping mode.
Chip Design and Fabrication.
We designed a microfluidic chip with PDMS as the structural material. The channel geometry (schematic illustrated in
Fig. 1a and b) was designed using AutoCAD 2013 (Autodesk,
USA). Width of the channels for loading the cells was designed as 50 µm. The microfluidic chip was fabricated via soft
lithography and rapid prototyping techniques.36 Briefly, a SU8™ mold with 50 µm in height was fabricated on a silicon
wafer. The PDMS layer was produced by pouring a mixture of
PDMS and curing agent in a mass ratio of 10 : 1 onto the SU8™ mold. After curing at 70 ºC for 2 h in an oven, the PDMS
sheet with channel networks was cut and peeled off from the
mold. The holes of inlets and outlets were punched using a
0.75 mm-diameter Harris Uni-Core biopsy punch (Electron
Microscopy Sciences). After oxygen plasma treatment
(PLASMA-PREEN II-862, Plasmatic Systems, Inc., United
States), the PDMS sheet was bonded with a PDMS-coated
glass substrate. After fabrication, the device was heated to 100
ºC on the hot plate and a low melting point In-Sn solder was
filled into the electrode channels. Small pieces of copper wire
were inserted to make the electrical connection with the solder
electrodes.
System Setup.
System setup was shown in Fig. 1a. The PEEK tubing (O.D.
= 0.03 inch, I.D. = 0.012 inch; Cole-Parmer, USA) was used
to connect the microfluidic device, syringe equipped on the
pumps (LSP01-2A, Longer Pumps, China) and tube for cell

ACS Paragon Plus Environment

Page 3 of 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Analytical Chemistry

collection. The tubing and syringe for cell loading were treated
with hydrophilic regent “5% PF127” for 10 min and washed
with sterile deionized water for 1 min prior to operation. Mineral oil containing 2.5% (w/w) Span 80 surfactant was used
for droplet generation. A high-voltage amplifier (PC 2000,
Tianjin Dongwen High Voltage Power Supply Co. Ltd., China), controlled by a digital I/O unit (DIO-1616LX-USB,
CONTEC) connected with the computer, was employed to
generate DEP to trigger the target droplet sorting.
Raman microscopy was carried out on a customized
LabRam HR micro-Raman setup,10,19 which is equipped with a
Nd:YAG 532 nm laser emitter (Ventus, Laser Quantum Ltd.,
United Kingdom) as the excitation light source, an electronmultiplying charge-coupled device (EMCCD) (Newton
DU970N-BV, Andor, United Kingdom) for collecting the
Raman signals, a high-speed CCD camera (Pike F-032, Allied
Vision Technologies, China) for monitoring the cell and droplet flow and a 60 × water objective (NA = 1.0, Olympus, United Kingdom) to focus the laser beam on the sample. A 600
lines/mm grating was used for the measurements, resulting in
a spectral resolution of ∼1 cm-1 with 1581 data points ranging
from 600 to 2100 cm-1. A 660 nm LED array was used as the
light source for monitoring the sorting process.
A homemade software, QSpec, was developed to control the
electronics (EMCCD and high-voltage amplifier, etc.) and
adjust the system parameters (e.g. acquisition time and DEP
duration). With the QSpec, all the RADS units including microfluidic device, Raman system and DEP system were integrated and can be operated in an automatic mode.
Operation Procedure for RADS.
There are five steps for the cell sorting, as shown in Fig. 2
and Video S1 in the Supporting Information: Step 1, cell suspension was loaded into the chip; Step 2, cells were focused
into single-cell flow by two pinch flow; Step 3, cells were
passed through the laser point and analyzed by Raman; Step 4,
single-cells were encapsulated into single-droplets; Step 5,
after a specific delay, target droplets were sorted by DEP and
flowed into collecting channel, while non-target droplets
would flow into waste channel.
Cell Recovery.
The sorted droplets containing target cells were collected
and filtered using Parylene C film featuring pores with 12 µm
in diameter (smaller than the cell size which is > 20 µm in
diameter). During the filtering, oil would pass through the film
easily with a filter paper placed underneath the film, while the
cells would be trapped on the film. After wash with culture
media, the Parylene C film containing target cells was transferred into the flask containing 15 ml basal medium and shaken gently to rinse the cells off from the film. Finally, the
Parylene C film was removed and the cells were cultured as
described above.
Validation and Evaluation Assays.
For the assay that validates sorting accuracy, the cells were
reinjected into the chip, and the SCRS was acquired without
sorting, and analyzed to verify whether the sorting result met
the sorting criteria. Two assays were employed to evaluate the
vitality of post-sorting cells. (i) To determine whether the
post-sorting cells are alive or dead, the cells were stained with
PI (5 µmol L-1) for 10 min to measure the survival rate. (ii) To
assess the robustness of cellular proliferation, the recovered
cells were cultured in 15 ml basal medium and the concentration was monitored for 18 days using cell counting plate. For
these two assays, the sorting criteria was set as 1516 cm-1 -

1800 cm-1 > 200 (the chip background is < 60) to sort the cells,
and the unsorted cells were used as the control group.

RESULTS AND DISCUSSION
RADS Chip Design and System Development.
In the present implementation of FADS, fluorescence signal
was measured after single-cell droplet encapsulation.22-24
However, owing to the lensing effect of the convex/concave
shape of the droplet surface and the strong inherent Raman
interference from oil, it was difficult to accurately acquire the
SCRS of cells encapsulated in droplets. To tackle this challenge we proposed to interrogate SCRS of the cells prior to
droplet encapsulation. Accordingly, we fabricated a RADS
(Fig. 1b), in which the droplet generation unit (ii) designed to
encapsulate single-cells was set after the pinch flow focusing
unit (i) that converges cells into the center line of the channel
for efficient SCRS acquisition. Based on this design, the efficiency of SCRS acquisition was significantly improved and
the effect of droplet on SCRS acquisition completely eliminated. Moreover, the DEP-based droplet sorting unit (iii) designed to isolate the droplets containing the target cells was set
directly after the droplet encapsulation unit, so as to realize
simultaneous droplet encapsulation and sorting, which increases sorting accuracy and simplifies system design. To bias
all droplets flow into the “waste” channel in the absence of
DEP, which was a central prerequisite for accurate sorting,
resistance of the “collect” channel was designed to be 1.5
times higher than that of the “waste” channel. Furthermore, to
balance the resistance changes caused by the bias of droplet
flow, the sorting junction was implemented as several small
islets. These designs endow the RADS chip the ability to
screen AXT-hyperproducing H. pluvialis cells with highthroughput and high-accuracy.
In our Raman system, the shortest acquisition time for AXT
is ~30 ms under laser power of 100 mW with 1000-fold degeneration (the laser power on sample is ~135 µW) (Fig. S1a).
Based on the fact that the cell size used in our experiments is
~30 µm in diameter, the optimized cell velocity in the channel
should be 1 mm s-1. Based on the relationship between flow
velocity and cell velocity (Fig. S1b), the flow velocity for
loading the cells should be 7 µL h-1 (4 µL h-1 for cell sample
and 3 µL h-1 for pinch flow). Under these conditions, the efficiency of SCRS acquisition was increased significantly, to ~98%
(Fig. S2 and Video S2).
In the RADS system, DEP-based droplet sorting was applied to isolate those containing target droplets cells. As previously reported, droplet spacing plays an important role in successful sorting.22,37 When being physically too close to each
other, the droplets would congest at the sorting junction, causing them to flow into the collecting channel in the absence of
DEP, and might also induce droplets coalescence in the present of DEP. Increasing the flow rate of mineral oil would
increase the droplet spacing. However, extremely high flow
rate of the mineral oil would also decrease the droplet size and
increase the droplet speed, which was adverse to efficiency of
cell encapsulation and droplet sorting. In our experiments, the
optimal flow rate for mineral oil was 120 µL h-1, with which
the droplets were generated with 50 µm in diameter. Under
these conditions, application of 3 ms of an 800 Vp-p (-400 V
to +400 V) voltage was sufficient for isolation of the target
droplets.
To increase the portion of droplets containing only one cell
while reduce that of those harboring multiple cells, samples of

ACS Paragon Plus Environment

Analytical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

~8.02 × 106 cells mL-1 (the final concentration was ~4.58 × 106
cells mL-1 due to the loading of pinch flow) were loaded to
encapsulate theoretically λ = 0.3 cells in droplets with ~50 µm
diameter. As estimated by the Poisson statistics, this setting of
cell density would result in 74.08% of the droplets containing
no cells, 22.22% containing a single cell, 3.3% containing two
cells and 0.38% containing over two cells.24 During sorting,
there would be a specific time delay for the flowing cells to
reach the sorting junction after SCRS acquisition. Under the
above optimized parameters the time delay was calculated to
be 145 ms by recording the real-time cell flow using a highspeed CCD. Hence, the high-voltage amplifier controlled by a
digital I/O unit would be triggered to sort the target droplets
after 145 ms when the Raman system acquires a signal that
meets the sorting criteria. A 660 nm LED array, which was
found to exert no interference on SCRS acquisition, was employed to monitor the entire sorting process in real-time. Finally, to automate the function of each unit and integrate Raman
microscopy and DEP-based droplet sorting, the QSpec software was developed.
Quantitative Raman Detection of AXT in the PDMS-glass
Hybrid Chip.
Previous studies proved that Raman signal can model concentration and distribution of AXT in H. pluvialis.30-34 The
Raman spectrum of AXT showed two most prominent characteristic bands at 1516 cm-1 and 1157 cm-1 which were assigned
to C=C and C-C stretching vibrations of the chain bonds, respectively. Here, the cells after stress-induced AXT accumulation for various days, which featured distinct averaged AXT
concentrations at each of the days, were interrogated for SCRS
with laser of 100 mW and filter of 1000-fold degeneration on
a calcium fluoride (CaF2) slide (Fig. 3a i). These SCRS were
then treated as the benchmark curve of AXT concentration in
H. pluvialis. Our RADS process was carried out on the
PDMS-glass hybrid chip. To test whether background interference from PDMS or glass affects SCRS-based measurement of
AXT in H. pluvialis, the cells were also analyzed in RADS
chip and compared to those analyzed on CaF2 slide under identical conditions (Fig. 3a ii). Pre-processing of all the raw spectra was performed with LabSpec 5 (Horiba Scientific, Orsay,
France), including background subtraction and baseline correction via a polynomial algorithm with a degree of seven. As
shown in Fig. 3a and b, although the Raman intensity measured in the RADS chip decreased significantly at both 1157
cm-1 and 1516 cm-1 as compared to that measured on the CaF2
slide, the tendency of AXT accumulation over time was identical between RADS chip and CaF2 slide. Furthermore, a good
linear relationship between the Raman intensity obtained in
RADS chip and on CaF2 slide at both 1157 cm-1 and 1516 cm-1
was found, as illustrated in Fig. 3c. Collectively these data
indicated that the Raman background interference from the
RADS chip would not compromise the quantitative detection
of AXT in H. pluvialis.
Quantitative Characterization of AXT in Real-time with
Raw Data in Flow State.
During real-time sorting, Raman signal was acquired in
flow state without pre-processing. Thus to ensure that such
raw data can quantitatively detect AXT, the cells induced for
different days were further interrogated for SCRS in flow state.
To increase signal intensity in RADS chip, laser of 100 mW
and filter of 100-fold degeneration were applied for SCRS
acquisition. The raw spectrum illustrated in Fig. 4a revealed
that the Raman intensity of the two characteristic bands at

Page 4 of 11

1516 cm-1 and 1157 cm-1 was still positively correlated with
AXT concentration. The Raman intensity of 1157 cm-1 and
1516 cm-1 was further characterized using the absolute Raman
intensity (defined as the actually detected value) shown in Fig.
S3a and the relative Raman intensity (defined as the Intensity1157 cm-1 - Intensity1800 cm-1 or Intensity1516 cm-1 - Intensity1800 cm-1;
1800 cm-1 was used due to the absence of biological molecular
vibrations from 1800 cm-1 to 1840 cm-1. 10) shown in Fig. 4b,
indicating an identical tendency compared with those measured on CaF2 slide. As shown in Fig. S3b and Fig. 4c, both the
absolute and relative Raman intensity exhibited a good liner
relationship (R2 > 0.9) with the benchmark intensity acquired
on CaF2 slide, indicating that the raw spectrum acquired in
flow state can also be employed for quantitative characterization of AXT. In fact, compared with the absolute intensity (R2
< 0.97), the relative Raman intensity (R2 > 0.97) exhibited an
even higher, i.e., nearly linear, correlation with the Raman
intensity on CaF2 slide. Based on these results, the relative
Raman intensity at 1516 cm-1 was chosen as the sorting criterion in the following sorting experiments.
Integration for Automated and Programmable Cell Sorting.
To achieve automated and reliable operation of the RADS
system with high-throughput and high-accuracy, SCRS acquisition and DEP-based droplet sorting were synchronized via
software. The work flow of whole sorting process was shown
in Fig. S4. Basically, except the loading of cells, pinch buffer
and mineral oil, which was controlled by the pump, the whole
process was run automatically based on predefined operational
parameters, and the sorting criteria can be set freely according
to the specific sorting requirement. During the sorting, the
Raman spectrum was acquired continuously, resulting in enhancement of spectra analysis speed and sorting throughput.
Furthermore, the Raman spectra acquired in real-time were
saved automatically, so as to facilitate off-line spectra analysis.
It should be noted that the SCRS acquisition and DEP-based
droplet sorting were controlled by multithreading, i.e., SCRS
acquisition ran simultaneously during the droplet sorting. This
design further increased the sorting efficiency.
Validating the Sorting Efficiency of RADS.
Traditionally, centrifugation or demulsification reagents
such as Pico-Break™ were used to recover the cells from sorted droplets.38,39 These recovery strategies can suffer from
drawbacks such as harmful effects on cell and low efficiency.
In our experiments, an ultra-high porosity Parylene C film
featuring pores with 12 µm in diameter (Fig. 5a) was used to
recover the sorted cells. The fabrication procedure of the
Parylene C film was described in our previous works. 40,41 As
shown in Fig. 5b, the cells were trapped on the film with high
efficiency and could also be released easily with almost no
cells attached on the film after washing (Fig. 5c). With this
recovery strategy, a high targeted cell separation yield of ~96%
was achieved.40,41
To validate the sorting efficiency, cells induced for 0 day
and 3 days respectively (i.e., in the absence or presence of
AXT; Fig. 4) were mixed at specific ratios and underwent the
RADS, where the sorting criterion was set as 1516 cm-1 - 1800
cm-1 > 1200. As shown in Fig. 6a and b, among the over 60
post-RADS cells that were randomly selected and then verified for AXT content, only one cell failed to meet the preset
sorting criterion, as indicated by blue spectra. On average, as
shown in Fig. 6c, the ratio of AXT-hyperproducing cells was
elevated to 98%, (from original samples of 12%; three inde-

ACS Paragon Plus Environment

Page 5 of 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Analytical Chemistry

pendent experiments), representing around eight folds enrichment.
Furthermore, to test the performance of our RADS system
under a more demanding situation, the cells induced for 1 day
and 2 days respectively (i.e., the two kinds of cells to be separated both produce AXT, but at different levels of abundance;
Fig. 4) were mixed at different ratios and then served as starting material for RADS. Correspondingly, the sorting criterion
was set as 1516 cm-1 - 1800 cm-1 > 2500. As shown in Fig.
6d and e, among the over 60 post-RADS cells randomly selected and verified, only 15 cells failed to meet the sorting
criteria, as indicated by blue spectra. On average, as demonstrated in Fig. 6f, the ratio of AXT-hyperproducing cells was
elevated to 78%, (from original samples of 11%, three independent experiments), which represented around seven folds
enrichment. The sorting efficiency is not 100%, which might
be due to several reasons: (i) there was a certain degree of
system error in the device setup; (ii) cells were detected in
their flow state yet the raw data from the slide were used for
setting sorting criteria; (iii) distribution of AXT in H. pluvialis
cells was not absolutely uniform; (iv) according to Poisson
statistics, there are very low ratio of droplets containing over
one cell.
Theoretically, the sorting throughput under our experimental settings can reach > 500 cells min-1, which is mainly
determined by the speed of Raman signal acquisition and single-cell encapsulation. In comparison, our actual sorting
throughput is ~260 cells min-1. The lowered performance is
mainly due to the sedimentation of the H. pluvialis cells (due
to their averaged large size of ~30 µm in diameter), which
however can be further increased by improving the efficiency
of cell loading. Due to the much weaker signal of spontaneous
SCRS than that of fluorescence probes and the resulted much
longer SCRS acquisition time, throughput of our RADS system is still much lower than present droplet-based FACS systems (hundreds to thousands cells per seconds).22-24 However,
our RADS system is non-labelling required and features the
highest sorting throughput among existing RACS techniques,
such as Raman tweezers (minutes per cell)8,9, RACE (minutes
per cell)10,11, optical tweezer based RAMS (minutes per cell)1216
, trap-free RAMS (~30 cells per minute)20 and DEP based
RAMS (~60 cells per minute)19. Collectively, these results
suggest robust performance of the RADS system in screening
AXT-hyperproducing H. pluvialis cells.
Assessing the Effect of RADS on Cell Vitality.
Considering that cells experienced laser exposure and DEPbased sorting during the screening process, which might introduce a negative effect on activity or vitality of the sorted cells,
assays for survival rate and proliferation were conducted. Few
dead cells were present, as indicated by PI staining, in both
unsorted cells (Fig. 7a) and sorted cells (Fig. 7b), and no significant difference was observed in averaged survival rate
(91.9% for the unsorted cells and 93.7% for the sorted cells;
Fig. 7c) between them. Moreover, cellular proliferation was
monitored by detecting cell concentration for every two days.
As shown in Fig. 7d, proliferation rate of the sorted cells was
nearly identical to that of the unsorted cells, at least within 18
days post-RADS. The high survival rate of RADS sorted cells
was due to several potential reasons: (i) the laser power on the
sample was far lower than 1 mW; (ii) the SCRS acquisition
time was quite short (30 ms); (iii) SCRS acquisition was carried out in flow state in liquid; (iv) all the single-cells were
encapsulated into droplets, which protect the cells during

DEP-based sorting; (v) Parylene C film was used to recover
the sorted cells, which was gentle and harmless to the cells.
Collectively, all these non-invasive procedures have resulted
in the high survival rate of RADS-sorted cells.

CONCLUSIONS
In this study, we established a RADS system that automatically integrates SCRS acquisition and DEP-based droplet sorting in a mode of continuous flow. In our RADS, the droplet
technique was adopted to encapsulate and sort single-cells,
which combines many of the advantages of plate screening
and existing RACS systems: (i) dielectrophoresis (for sorting),
which can be difficult to apply directly to cells in suspension,
can be used to sort droplets with good stability and high
throughput; (ii) vitality of cells can be protected during sorting,
i.e., recovery rate of viable cells post-sorting is improved; (iii)
sorted cells are still compartmentalized in droplets, which
facilitates seamless integration with downstream extraction of
DNA, RNA, proteins and metabolites from the cells for further
analysis (e.g., sequencing); (iv) coupling with high sensitivity
Raman techniques (e.g. Coherent Anti-Stokes Raman Spectroscopy and Stimulated Raman Scattering) is facilitated,
which might lead to ultra-high throughput RADS system. By
acquiring Raman signals of individual cells prior to droplet
encapsulation, the system successfully tackles the difficulties
associated with Raman interrogation of droplet-encapsulated
cells: primarily the lensing effect of the convex/concave shape
of droplet surface and the strong inherent Raman interference
from oil medium. Moreover, droplet encapsulation was directly coupled to DEP-based droplet sorting, which simplifies
system design and operation. Finally, based on our RADS
system, AXT-hyperproducing H. pluvialis cells were sorted
with high-throughput (~260 cells min-1) and high-accuracy
(~98%), achieving enrichment ratio of eight folds on average.
Our recent work suggested that SCRS, via the ramanome
approach, can quantitatively distinguish bacterial species,10
measure the general metabolic activity of cells or probe the
catabolic activity targeting a specific substrate,9 model the
intracellular levels of triacylglycerols (TAG)42 and starch,43
measure drug sensitivity,44 distinguish cellular drug responses
based on cytotoxicity mechanism45 and trace cross-species
metabolite exchange.46 Hence, by coupling with the ramanome
approach, the RADS can enable the sorting for a broad range
of (and theoretically unlimited) cellular functions. On the other hand, the “trap-free” SCRS acquisition strategy adopted in
this RADS system might have certain limitations, for example,
(i) longer SCRS acquisition time might be required, e.g., when
the signal associated with the targeted phenotype in an SCRS
is weak; (ii) small-size cells such as E. coli might require additional optimization, due to the reduced Raman scattering
cross-section. As a result, the RADS at its current form is the
most suitable for sorting cells that feature high Raman intensity and large scattering cross-section. To tackle these limitations, one potential solution is a RADS chip in which pDEPbased single-cell trap-and-release for SCRS acquisition was
performed before droplet encapsulation and sorting. As pDEPbased single-cell trap-and-release has been established as a
generic strategy to interrogate the intrinsic SCRS of cells,19 the
pDEP-RADS should be feasible and can tackle those cells
with weak SCRS signal, by extending the exposure time of
cells during SCRS acquisition. We envision that RADS may
become a generally applicable, versatile and high-throughput

ACS Paragon Plus Environment

Analytical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

RACS platform for label-free functional screening of cells and
subsequent omics profiling at single-cell resolution.

ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website.
Additional supporting figures (PDF)
Additional supporting videos (WMV)

AUTHOR INFORMATION
Corresponding Author
*E-mail: mabo@qibebt.ac.cn. Phone: +86-0532-80662657.
*E-mail: xujian@qibebt.ac.cn. Phone: +86-0532-80662653.

Author Contributions
‡

These authors contributed equally.

Notes
The authors declare no competing financial interest.
Project funded by China Postdoctoral Science Foundation.

ACKNOWLEDGMENT
The authors gratefully acknowledge financial support from National Natural Science Foundation of China (31600076, 31470220,
31670101, 31327001, 31601084, 21775155 and 31425002), Chinese Academy of Sciences (KFJ-SW-STS-165), China Postdoctoral Science Foundation Grant (2017M612364 and 2017T100522)
and Shandong Province Natural Science Foundation
(ZR2014CQ005). We thank the partial support from State Development & Investment Corporation (SDIC).

REFERENCES
(1) Navin, N.; Kendall, J.; Troge, J.; Andrews, P.; Rodgers, L.;
McIndoo, J.; Cook, K.; Stepansky, A.; Levy, D.; Esposito, D.; Muthuswamy, L.; Krasnitz, A.; McCombie, W. R.; Hicks, J.; Wigler, M.
Nature 2011, 472, 90-94.
(2) Rinke, C.; Schwientek, P.; Sczyrba, A.; Ivanova, N. N.; Anderson, I. J.; Cheng, J. F.; Darling, A.; Malfatti, S.; Swan, B. K.; Gies, E.
A.; Dodsworth, J. A.; Hedlund, B. P.; Tsiamis, G.; Sievert, S. M.; Liu,
W. T.; Eisen, J. A.; Hallam, S. J.; Kyrpides, N. C.; Stepanauskas, R.;
Rubin, E. M.; Hugenholtz, P.; Woyke, T. Nature 2013, 499, 431-437.
(3) Rosenfeld, N.; Young, J. W.; Alon, U.; Swain, P. S.; Elowitz,
M. B. Science 2005, 307, 1962-1965.
(4) Xu, J.; Ma, B.; Su, X. Q.; Huang, S.; Xu, X.; Zhou, X. D.;
Huang, W. E.; Knight, R. Engineering 2017, 3, 66-70.
(5) Song, Y.; Yin, H.; Huang, W. E. Curr. Opin. Chem. Biol. 2016,
33, 1-8.
(6) Zhang, Q.; Zhang, P.; Gou, H.; Mou, C.; Huang, W. E.; Yang,
M.; Xu, J.; Ma, B. Analyst 2015, 140, 6163-6174.
(7) Chrimes, A. F.; Khoshmanesh, K.; Stoddart, P. R.; Mitchell, A.;
Kalantar-Zadeh, K. Chem. Soc. Rev. 2013, 42, 5880-5906.
(8) Huang, W. E.; Ward, A. D.; Whiteley, A. S. Environ. Microbiol. Rep. 2009, 1, 44-49.
(9) Berry, D.; Mader, E.; Lee, T. K.; Woebken, D.; Wang, Y.; Zhu,
D.; Palatinszky, M.; Schintlmeister, A.; Schmid, M. C.; Hanson, B.
T.; Shterzer, N.; Mizrahi, I.; Rauch, I.; Decker, T.; Bocklitz, T.; Popp,
J.; Gibson, C. M.; Fowler, P. W.; Huang, W. E.; Wagner, M. Proc.
Natl. Acad. Sci. U.S.A. 2015, 112, E194-203.
(10) Wang, Y.; Ji, Y.; Wharfe, E. S.; Meadows, R. S.; March, P.;
Goodacre, R.; Xu, J.; Huang, W. E. Anal. Chem. 2013, 85, 1069710701.

Page 6 of 11

(11) Song, Y.; Kaster, A. K.; Vollmers, J.; Song, Y.; Davison, P.
A.; Frentrup, M.; Preston, G. M.; Thompson, I. P.; Murrell, J. C.; Yin,
H.; Hunter, C. N.; Huang, W. E. Microb. Biotechnol. 2017, 10, 125137.
(12) Dochow, S.; Krafft, C.; Neugebauer, U.; Bocklitz, T.; Henkel,
T.; Mayer, G.; Albert, J.; Popp, J. Lab Chip 2011, 11, 1484-1490.
(13) Dochow, S.; Beleites, C.; Henkel, T.; Mayer, G.; Albert, J.;
Clement, J.; Krafft, C.; Popp, J. Anal. Bioanal.Chem. 2013, 405,
2743-2746.
(14) Lau, A. Y.; Lee, L. P.; Chan, J. W. Lab Chip 2008, 8, 11161120.
(15) Casabella, S.; Scully, P.; Goddard, N.; Gardner, P. Analyst
2016, 141, 689-696.
(16) Xie, C.; Chen, D.; Li, Y. Q. Opt. Lett. 2005, 30, 1800-1802.
(17) Pilat, Z.; Jezek, J.; Sery, M.; Trtilek, M.; Nedbal, L.; Zemanek, P. J. Photochem. Photobiol., B 2013, 121, 27-31.
(18) Zhang, Q.; Zhang, P.; Su, Y.; Mou, C.; Zhou, T.; Yang, M.;
Xu, J.; Ma, B. Lab Chip 2014, 14, 4599-4603.
(19) Zhang, P.; Ren, L.; Zhang, X.; Shan, Y.; Wang, Y.; Ji, Y.;
Yin, H.; Huang, W. E.; Xu, J.; Ma, B. Anal. Chem. 2015, 87, 22822289.
(20) McIlvenna, D.; Huang, W. E.; Davison, P.; Glidle, A.;
Cooper, J.; Yin, H. Lab Chip 2016, 16, 1420-1429.
(21) Shang, L.; Cheng, Y.; Zhao, Y. Chem. Rev. 2017, 117, 79648040.
(22) Baret, J. C.; Miller, O. J.; Taly, V.; Ryckelynck, M.; ElHarrak, A.; Frenz, L.; Rick, C.; Samuels, M. L.; Hutchison, J. B.;
Agresti, J. J.; Link, D. R.; Weitz, D. A.; Griffiths, A. D. Lab Chip
2009, 9, 1850-1858.
(23) Agresti, J. J.; Antipov, E.; Abate, A. R.; Ahn, K.; Rowat, A.
C.; Baret, J. C.; Marquez, M.; Klibanov, A. M.; Griffiths, A. D.;
Weitz, D. A. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 4004-4009.
(24) Mazutis, L.; Gilbert, J.; Ung, W. L.; Weitz, D. A.; Griffiths,
A. D.; Heyman, J. A. Nat. Protoc. 2013, 8, 870-891.
(25) Liu, W.; Kim, H. J.; Lucchetta, E. M.; Du, W.; Ismagilov, R.
F. Lab Chip 2009, 9, 2153-2162.
(26) Boussiba, S. Physiol. Plant. 2000, 108, 111-117.
(27) Sandesh Kamath, B.; Vidhyavathi, R.; Sarada, R.; Ravishankar, G. A. Bioresour. Technol. 2008, 99, 8667-8673.
(28) Ukibe, K.; Katsuragi, T.; Tani, Y.; Takagi, H. FEMS Microbiol. Lett. 2008, 286, 241-248.
(29) Liu, J. H.; Song, L.; Huang, Q. Lett. Appl. Microbiol. 2016,
62, 185-191.
(30) Liu, J.; Huang, Q. Appl. Spectrosc. 2016, 70, 1639-1648.
(31) Kaczor, A.; Turnau, K.; Baranska, M. Analyst 2011, 136,
1109-1112.
(32) Kaczor, A.; Baranska, M. Anal. Chem. 2011, 83, 7763-7770.
(33) Collins, A. M.; Jones, H. D.; Han, D.; Hu, Q.; Beechem, T. E.;
Timlin, J. A. PloS One 2011, 6, e24302.
(34) Li, K.; Cheng, J.; Ye, Q.; He, Y.; Zhou, J.; Cen, K. Bioresour.
Technol. 2017, 244, 1439-1444.
(35) Kobayashi, M.; Kakizono, T.; Nagai, S. J. Ferment. Bioeng.
1991, 71, 335-339.
(36) Duffy, D. C.; McDonald, J. C.; Schueller, O. J.; Whitesides,
G. M. Anal. Chem. 1998, 70, 4974-4984.
(37) Fiedler, B. L.; Van Buskirk, S.; Carter, K. P.; Qin, Y.; Carpenter, M. C.; Palmer, A. E.; Jimenez, R. Anal. Chem. 2017, 89, 711-719.
(38) Liu, X.; Painter, R. E.; Enesa, K.; Holmes, D.; Whyte, G.;
Garlisi, C. G.; Monsma, F. J.; Rehak, M.; Craig, F. F.; Smith, C. A.
Lab Chip 2016, 16, 1636-1643.
(39) Zhang, Q.; Wang, T.; Zhou, Q.; Zhang, P.; Gong, Y.; Gou, H.;
Xu, J.; Ma, B. Sci. Rep. 2017, 7, 41192.
(40) Liu, Y.; Wang, W.; Wu, W.; Li, H.; Sun, Y. Highly precise
and efficient cell separation with Parylene C micropore arrayed filtration membrane. MicroTAS’2015, 389-391.
(41) Liu, Y.; Wang, W.; Wu, W.; Li, H.; Yang, F. Filtration membrane with ultra-high porosity and pore size controllability fabricated
by Parylene C molding technique for cell separation from bronchoalveolar lavage fluid (BALF). Transducer’2015, 1767-1769.
(42) Wang, T.; Ji, Y.; Wang, Y.; Jia, J.; Li, J.; Huang, S.; Han, D.;
Hu, Q.; Huang, W. E.; Xu, J. Biotechnol. Biofuels 2014, 7, 58.

ACS Paragon Plus Environment

Page 7 of 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Analytical Chemistry

(43) Ji, Y.; He, Y.; Cui, Y.; Wang, T.; Wang, Y.; Li, Y.; Huang,
W. E.; Xu, J. Biotechnol. J. 2014, 9, 1512-1518.
(44) Tao, Y.; Wang, Y.; Huang, S.; Zhu, P.; Huang, W. E.; Ling,
J.; Xu, J. Analytical chemistry 2017, 89, 4108-4115.

(45) Teng, L.; Wang, X.; Wang, X.; Gou, H.; Ren, L.; Wang, T.;
Wang, Y.; Ji, Y.; Huang, W. E.; Xu, J. Sci. Rep. 2016, 6, 34359.
(46) Wang, Y.; Song, Y.; Tao, Y.; Muhamadali, H.; Goodacre, R.;
Zhou, N. Y.; Preston, G. M.; Xu, J.; Huang, W. E. Anal. Chem. 2016,
88, 9443-9450.

ACS Paragon Plus Environment

Analytical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Page 8 of 11

Figure 1

Fig. 1 Overview of RADS system setup for high-throughput screening of AXT-hyperproducing H. pluvialis cells. (a) Schematic diagram
of RADS system setup. (b) Schematic illustration of the chip design. The red spheres represent target cells, and the green triangle indicates
the position of the 532 nm laser beam. The inserted photograph is a bright-field image of the main functional units: i, pinch flow focusing
unit; ii, cross-junction based droplet generation unit and iii, DEP-based droplet sorting unit. Scale bar represents 100 µm.

Figure 2

Fig. 2 Step-by-step operation process for DEP-based single-cell sorting. (1-5) Trajectory of the target cell, consequently sorted into the
collecting channel based on the DEP trigger. (6-9) Trajectory of the non-target cell, resulting in flowing into the waste channel without the
DEP trigger. Scale bar represents 200 µm.

ACS Paragon Plus Environment

Page 9 of 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Analytical Chemistry

Figure 3

Fig. 3 Quantitative Raman detection of AXT in live individual H. pluvialis cells. (a) Single-cell Raman spectra of H. pluvialis cells induced for different days on CaF2 (i) and in chip (ii). Colored lines and dark grey shading indicate the mean value and the SEM of 60 cells,
respectively. (b) Raman intensity of H. pluvialis cells induced for different days at 1157 cm-1 and 1516 cm-1 on CaF2 and in chip. Error bars
represent the SEM of 60 cells. (c) Linear correlation between the Raman intensity detected on CaF2 and in chip at 1157 cm-1 (i, R2 = 0.992)
and 1516 cm-1 (ii, R2 = 0.996). Error bars represent the SEM of 60 cells.

Figure 4

Fig. 4 Quantitative characterization of AXT in real-time with raw data in the flow state. (a) Single-cell Raman spectra of H. pluvialis cells
induced for different days in the RADS chip. Colored lines and dark grey shading indicate the mean value and the SEM of 60 cells, respectively. (b) Relative Raman intensity of H. pluvialis cells induced for different days at 1157 cm-1 and 1516 cm-1 on CaF2 and in chip. Error
bars represent the SEM of 60 cells. (c) The linear correlation between the Raman intensity detected on CaF2 (processed data) and in the
RADS chip (raw data) at 1157 cm-1 (i, R2 = 0.973) and 1516 cm-1 (ii, R2 = 0.983). Error bars represent the SEM of 60 cells.

ACS Paragon Plus Environment

Analytical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Page 10 of 11

Figure 5

Fig. 5 Cell recovery via an ultra-high porosity Parylene C film. The images demonstrate the ultra-high porosity Parylene C film before (a),
during (b) and after (c) filtration. The red arrow in (c) indicates the cell attached on the film after washing. Scale bars represent 50 µm.

Figure 6

Fig. 6 Validation of sorting efficiency of RADS. (a-c) RADS were performed by sorting AXT-hyperproducing cells from a mixture that
contained ~12% target cells (i.e., cells induced for 3 days, which produced AXT), which was generated by mixing the cells induced for 0
day and those induced for 3 days. (d-e) RADS were performed by sorting AXT-hyperproducing cells from a mixture that contained ~11%
target cells, which was generated by mixing the cells induced for 1 day and those induced for 2 days. (a and d) Raman spectra of > 60 sorted cells. Red spectra indicate the target cells and the blue spectra represent the non-target cells. (b and e) Distribution of the relative Raman
intensity of sorted cells at 1516 cm-1. The dotted line at 1200 in (b) and at 2500 in (e) indicate the sorting criteria. (c and f) Enrichment of
the target cells by the RADS system. “Sorted” indicates the positively-sorted cells and “Waste” indicates the negatively-sorted cells. Error
bars indicate the SEM of three independent experiments.

ACS Paragon Plus Environment

Page 11 of 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Analytical Chemistry

Figure 7

Fig. 7 Evaluation of the viability of post-RADS cells. (a and b) PI staining for detecting dead cells of the control group (a) and the experimental group (b). White circles indicate the dead cells. (c) Survival rate. n.s.: not significant. (d) Proliferation assay. Error bars indicate the
SEM of three independent experiments.

TOC: RADS for label-free high-throughput screening of microalgae

ACS Paragon Plus Environment