Label free screening of enzyme inhibitors at femtomole scale using segmented flow electrospray ionization mass spectrometry

Article abstract of DOI:10.1021/ac3011389

Droplet-based microfluidics is an attractive platform for screening and optimizing chemical reactions. Using this approach, it is possible to reliably manipulate nanoliter volume samples and perform operations such as reagent addition with high precision, automation, and throughput. Most studies using droplet microfluidics have relied on optical techniques to detect the reaction; however, this requires engineering color or fluorescence change into the reaction being studied. In this work, we couple electrospray ionization mass spectrometry (ESI-MS) to nanoliter scale segmented flow reactions to enable direct (label-free) analysis of reaction products. The system is applied to a screen of inhibitors for cathepsin B. In this approach, solutions of test compounds (including three known inhibitors) are arranged as an array of nanoliter droplets in a tube segmented by perfluorodecalin. The samples are pumped through a series of tees to add enzyme, substrate (peptides), and quenchant. The resulting reaction mixtures are then infused into a metal-coated, fused silica ESI emitter for MS analysis. The system has potential for high-throughput as reagent addition steps are performed at 0.7 s per sample and ESI-MS at up to 1.2 s per sample. Carryover is inconsequential in the ESI emitter and between 2 and 9% per reagent addition depending on the tee utilized. The assay was reliable with a Z-factor of ～0.8. The method required 0.8 pmol of test compound, 1.6 pmol of substrate, and 5 fmol of enzyme per reaction. Segmented flow ESI-MS allows direct, label free screening of reactions at good throughput and ultralow sample consumption.

Full text of DOI:10.1021/ac3011389

Article
pubs.acs.org/ac
Label Free Screening of Enzyme Inhibitors at Femtomole Scale Using
Segmented Flow Electrospray Ionization Mass Spectrometry
Shuwen Sun,† Thomas R. Slaney,† and Robert T. Kennedy*,†,‡
†University of Michigan, Department of Chemistry, Ann Arbor, Michigan, 48109, United States
‡University of Michigan, Department of Pharmacology, Ann Arbor, Michigan, 48109, United States
S
* Supporting Information
ABSTRACT: Droplet-based microfluidics is an attractive
platform for screening and optimizing chemical reactions.
Using this approach, it is possible to reliably manipulate
nanoliter volume samples and perform operations such as
reagent addition with high precision, automation, and
throughput. Most studies using droplet microfluidics have
relied on optical techniques to detect the reaction; however,
this requires engineering color or fluorescence change into the reaction being studied. In this work, we couple electrospray
ionization mass spectrometry (ESI-MS) to nanoliter scale segmented flow reactions to enable direct (label-free) analysis of
reaction products. The system is applied to a screen of inhibitors for cathepsin B. In this approach, solutions of test compounds
(including three known inhibitors) are arranged as an array of nanoliter droplets in a tube segmented by perfluorodecalin. The
samples are pumped through a series of tees to add enzyme, substrate (peptides), and quenchant. The resulting reaction mixtures
are then infused into a metal-coated, fused silica ESI emitter for MS analysis. The system has potential for high-throughput as
reagent addition steps are performed at 0.7 s per sample and ESI-MS at up to 1.2 s per sample. Carryover is inconsequential in
the ESI emitter and between 2 and 9% per reagent addition depending on the tee utilized. The assay was reliable with a Z-factor
of ∼0.8. The method required 0.8 pmol of test compound, 1.6 pmol of substrate, and 5 fmol of enzyme per reaction. Segmented
flow ESI-MS allows direct, label free screening of reactions at good throughput and ultralow sample consumption.
roplet-based microfluidics has emerged as a powerful tool
While such methods are powerful and rapid (allowing analysis
at up to kilohertz rates23), they are restricted to single
component detection unless wavelength resolution is used.
More importantly, they require engineering an optical change
into the reaction being screened by use of either labels or
coupled reactions. Incorporation of optical properties can slow
investigation of a new target and in some cases may not be
feasible. Even when feasible, labels can affect the reaction being
studied, generate false signals,26 and add expense to the screen
because of increased reagent cost. For these reasons, label-free
screening is desirable.26−28
A potentially powerful method for label-free screening at
droplet scale is mass spectrometry (MS). In one study,
segmented samples were deposited onto a plate for matrix-
assisted laser desorption ionization (MALDI)-MS analysis.29
This approach allowed optimization of a deacetylation reaction
using just 20 μg of substrate illustrating the feasibility of
screening at reduced scale. Electrospray ionization (ESI) would
be a valuable complement to MALDI methods. Several
techniques for interfacing droplets to ESI-MS have been
reported.30−32,34 Methods involve either extraction of droplets
into a stream that flows to the ESI source or directly passing the
segmented stream into the ESI source. While either approach
D
for analysis of low volume samples.1,2 In this approach,
samples are compartmentalized within an immiscible fluid that
is manipulated within a microfluidic channel or tube. This
technology is descended from continuous flow analysis,3
commonly used for automated analysis prior to multiwell
plates (MWP) and flow injection analysis. When droplets are
large enough to span the walls of the channel, they form plugs
and the resulting system is referred to as segmented flow.
Droplets can be moved through a microfluidic system by
pressure-induced flow where operations such as reagent
addition,4−6 dilution,7 splitting,8,9 or sorting10,11 may be
performed. A variety of chemical measurements including
enzymatic activity,12−14 enzyme kinetics,15−18 and cell-based
assays19−21 have been developed for droplet format.
Salient features of droplet systems are their small sample
volumes (picoliter to nanoliter volumes are typical) and high
throughput (e.g., reagents can be added to segmented samples
at rates as high as 10 Hz22). These properties make droplet
microfluidics attractive for screening applications. The small
scale can greatly reduce cost of reagents compared to screens
performed at microliter volumes in MWP while the high speed
enables the necessary throughput. Droplet microfluidics have
been used for screening for engineered proteins,23 protein
crystallization,24 and reaction catalysts.25
Received: April 30, 2012
Accepted: May 31, 2012
Most chemical measurements in droplets have relied on
optical detection such as fluorescence or colorimetric changes.
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 1. (A) Diagram of oil-segmented droplet generation from a MWP. The XYZ-positioner causes the Teflon tube to dwell in sample or oil for a
predetermined time and then move to another well. The syringe pump operates in refill mode at a constant flow rate. (B) Diagram of ESI-MS
analysis of sample droplets. The syringe drives droplets into the ESI source through a Pt-coated emitter. +1.7 kV is applied on the emitter.
may be used, we favor the latter because it is simple and allows
samples to remain encapsulated up to the point of detection to
avoid dilution and reduce potential for carry-over. Segmented
flow ESI-MS has been used to monitor chemical reactions32
and for analysis of fractions collected from capillary LC
columns.33 It has also been used to screen an enzymatic
reaction that did not involve a color change (acetylcholinester-
ase catalyzed hydrolysis of acetylcholine13). In that study,
however, the reaction was performed in MWP-scale and then
reformatted for segmented flow; therefore, miniaturization of
the reaction was not realized.
In this work, we seek to demonstrate the potential for label-
free screening of reactions at nanoliter scale using segmented
flow ESI-MS. The system is tested using cathepsin B catalyzed
proteolysis as a model reaction. Inhibitors of this protein are of
interest for treatment of Alzheimer’s disease,35 various types of
cancer,36 arthritic disease,37 and trematode infection.38 This
enzyme is presently screened using a fluorogenic substrate;
nevertheless, it provides a useful test case for segmented flow
ESI-MS.
on a PHD 200 programmable syringe pump (Harvard
Apparatus, Holliston, MA). The syringe and Teflon tube
were prefilled with perfluorodecalin (PFD, Acros Organics,
NJ). To fill the tube with oil-segmented droplets, a computer-
controlled XYZ-positioner (built in-house from XSlide
assemblies, Velmex Inc., Bloomfield, NY) was used to move
the inlet of the Teflon tube from sample to sample on the
MWP while the syringe was withdrawing at 150−180 nL/min.
The samples were covered with PFD to prevent aspiration of
air as the tube moving from sample to sample. (The edges of
the MWP were built-up to 3 mm height using epoxy to hold
PFD over the wells.) Droplets with 5 to 15 nL volume
separated by 10−20 nL oil segments were produced by
controlling the time that the tube dwelled in sample or oil. The
relative standard deviation (RSD) of droplet size was <5%.
MS Analysis. Teflon tubes containing segmented samples
were connected to a 75 μm i.d. Pt-coated fused-silica
electrospray emitter (FS-360-75-30-CE, New Objective,
Woburn, MA) pulled to 30 μm i.d. at the tip (Figure 1B).
The emitter was mounted onto the nanospray source of the
mass spectrometer. A syringe pump (Fusion 400, Chemyx,
Stafford, TX) operated from 0.75 to 1.8 μL/min drove sample
droplets through the emitter. PFD emerging from the tip was
siphoned away using a Teflon tube placed near the tip as
described before.33 ESI voltage of +1.7 kV was directly applied
to the emitter. MS analysis was performed with a LTQ XL
linear ion trap MS (Thermo Fisher Scientific, Waltham, MA)
operated in positive mode, scanning from 350 to 690 m/z in
0.1 s. Extracted ion currents (XICs) of assay components were
used for analysis. Peak detection was performed using Qual
Browser (Version 2.0, Thermo Electron Co.).
EXPERIMENTAL SECTION
■
Chemicals and Materials. Unless otherwise specified, all
solvents were purchased from Honeywell Burdick & Jackson
(Muskegon, MI) and were certified ACS grade or better. All
reagents were purchased from Sigma Aldrich (St. Louis, MO).
Droplet Generation from MWP. Oil-segmented sample
droplets were generated by sampling from a MWP using a
protocol described previously (Figure 1A).34 Samples were
pulled into a 150 μm i.d. × 360 μm o.d. Teflon tube (IDEX
Health and Science, Oak Harbor, WA) connected to a 100 μL
Hamilton syringe (Fisher Scientific, Pittsburgh, PA) mounted
B
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 2. (A) Cathepsin B catalyzed proteolysis of ZRR-AMC and ESI mass spectrum of resulting reaction mixture. (B) Cathepsin B catalyzed
proteolysis of Ac-GFGFVGG-NH2 and ESI mass spectrum of reaction mixture.
In-Well Cathepsin B Inhibitor Screening. Cathepsin B
assays were initially tested by performing the reactions in vials
or wells and then formatting for segmented flow. These assays
used carboxybenzyl-Arg-Arg-7-amido-4-methylcoumarin (ZRR-
AMC) as substrate and were modified from previous reports.39
Twenty four test compounds, including three known cathepsin
B inhibitors: L-trans-3-carboxyoxiran-2-carbonyl-L-leucylagma-
tine (E-64), N-acetyl-Leu-Leu-Arg-al (leupeptin), and N-(Nα-
carbonyl-Arg-Val-Arg-al)-Phe (antipain) were dissolved in
water at 200 μM. (See Supporting Information for other test
compounds.) Fifty μL of each test compound solution was
mixed with 30 μL of 27 μg/mL bovine spleen cathepsin B and
200 μM 1,3-dithioerythritol (DTE) in 20 mM ammonium
formate buffer at pH 6.7 and incubated for 5 min at room
temperature. Eighteen μL of substrate 1 mM ZRR-AMC in
water was then added to start turnover. After 10 min of
incubation in a 40 °C water bath, 20 μL of a quenchant
consisting of 98% methanol and 2% acetic acid (v/v) was
added. Eighty μL of each quenched reaction mixture was
pipetted into a 384-well plate (Axygen Scientific, Union City,
Ca). Ninety six sample droplets (4 from each reaction mixture)
of 11 nL each were generated using the procedure described
above and pumped into MS at 0.75 μL/min through a Pt-
coated emitter. XICs of the product carboxybenzyl-Arg-Arg
(ZRR-OH) were used for analysis.
In experiments using unlabeled peptide Ac-GFGFVGG-NH2
(American Peptide Company Inc., Sunnyvale, CA) as substrate,
the procedures were the same except test compounds were
prepared at 40 μM and the enzyme was at 54 μg/mL. Each
reaction mixture was made into 14 nL droplets in quintuplicate.
XICs of the product FVGG-NH2 (m/z 378.5) were used for
analysis.
All-Droplet Cathepsin B Inhibitor Screening. To
perform the assay reaction in droplets, test compounds were
prepared as a 100-droplet array (8 nL each) consisting of 4
droplets for each of 25 test compounds at 100 μM. Droplets
were prepared with a 16 nL oil plug between them and loaded
into a 150 μm i.d. × 360 μm o.d. Teflon tube. Cathepsin B (54
μg/mL), ZRR-AMC (600 μM), and quenchant were added to
the test compound droplets in sequence using PDMS tees. (See
Supporting Information for test compounds.) Droplets were
pumped into the inlet channel of the tee at 2 μL/min while
solution to be added was pumped in an orthogonal channel at
200 nL/min. The resulting droplets exited the outlet channel
and were collected into a Teflon tube. Cathepsin B solution was
injected into the droplets first and required 1 min for 100
sample droplets. Substrate was then added, and the resulting
tube of droplets was sealed with Sticky Wax (KerrLab, Orange,
CA) and placed into a 40 °C water bath for 15 min. After
incubation, the ends of the tube were trimmed off and
quenchant was added. The final droplets containing quenched
reaction solutions were infused into the MS at 1.5 μL/min.
In other experiments, native substrate was used with a Teflon
tee.40 Five nL of 80 μM test compounds were loaded into the
Teflon tube with 10 nL plugs of oil between them. Enzyme
(108 μg/mL), substrate (600 μM), and quenchant were added
using the same flow rates. Resulting droplets were pumped into
the MS at 0.75 μL/min.
PDMS Tee Fabrication. Polydimethylsiloxane (PDMS)
reagent addition tees were fabricated using soft lithography.9,41
The droplet channel was rendered hydrophobic by silanization
immediately after plasma bonding by pumping 1:100 (v/v)
trichloro(1H,1H,2H,2H-perfluorooctyl) silane solution in
anhydrous hexadecane through it. A 3 cm length of presilanized
50 μm i.d. × 150 μm o.d. fused silica capillary (Polymicro
Technologies, Phoenix, AZ) was inserted from the side of the
tee into the droplet inlet channel, and a 6 cm length of
nonderivatized 20 μm i.d. × 150 μm o.d. capillary was inserted
into the reagent channel. A 3 cm length of 150 μm i.d. × 360
μm o.d. Teflon tube was inserted into the outlet channel.
Capillaries and the Teflon tube were glued over the ends with 5
min Epoxy (Devcon, Danvers, MA).
RESULTS AND DISCUSSION
■
In-Well Cathepsin B Inhibitor Screening. Cathepsin B is
commonly screened using the fluorogenic substrate ZRR-AMC
by the reaction shown in Figure 2A. By performing the reaction
in a low concentration buffer (6 mM ammonium formate and
60 μM DTE) and quenching the reaction with 98% methanol
and 2% acetic acid, both the substrate and product can be
readily detected by ESI-MS (Figure 2A). These results suggest
that the buffer conditions are both suitable for retaining
enzymatic activity and are MS friendly. Although a fluorescent
substrate for cathepsin B has been devised, the MS can also
C
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
detect native substrates. To demonstrate this, we synthesized
the nonfluorescent heptapeptide Ac-GFGFVGG-NH2 that
should serve as a good substrate according to previous studies
on the selectivity of this enzyme.42,43 The reaction and mass
spectrum of the reaction mixture illustrating detection of
substrate and product (FVGG-NH2) are shown in Figure 2B.
These results illustrate the potential of ESI-MS for detecting
different substrate−product pairs, reactions of unlabeled
substrates, and substrate and product simultaneously as long
as reaction conditions are enzyme and ESI-MS compatible.
We then tested both assays in a screening format. Reaction
solutions were reformatted to segmented flow from a MWP
and then analyzed by ESI-MS. The result of screening 24 test
compounds (one negative control, three positive controls, and
20 randomly chosen small molecules as potential inhibitors) in
quadruplicate using the substrate ZRR-AMC is illustrated in
Figure 3A. The result of a similar experiment using 12 samples
The assay has good reproducibility with RSD < 3% in droplet
peak height for each test compound set. The MS detection is
also rapid with the throughput of 0.7 Hz.
The suitability of the assay for screening was evaluated using
Z-factor.44 Z-factor is a statistic defined as Z = 1 − 3× (σpos
+
σneg)/(μpos − μneg), where σpos and σneg represent the SD of the
response in positive and negative control, and μpos and μneg are
the mean response of each. Z-factor exceeding 0.5 is considered
suitable for HTS. In our experiment, Z-factors were 0.92 (E-
64), 091 (leupeptin), and 0.91(antipain) for the fluorescent
substrate and 0.93 (E-64), 0.93 (leupeptin), and 0.70 (antipain)
for the unlabeled substrate.
The assays can also provide quantitative characterization of
the reaction. A linear relationship between sample concen-
tration and relative signal intensity was obtainable from 10
to180 μM (R2 = 0.991). IC50s of three inhibitors determined
from dose response analysis under our experiment conditions
generally agreed with reported values (see Supporting
Information). These results illustrate that each reaction result
can be rapidly analyzed at nanoliter scale by segmented flow
ESI-MS and that the results are suitable for screening and
quantitative assay of inhibitors.
All-Droplet Cathepsin B Inhibitor Screening. With the
assay established, we sought to miniaturize the system to take
advantage of the small quantities used for actual detection of
segmented flow ESI-MS. To do so, we performed the entire
Cathepsin B inhibitor screen in an all droplet format (Figure
4A). In this approach, 25 test compounds were first formatted
as droplets (8 nL each, in quadruplicate) in a tube. The
droplets were then pumped through a series of reagent addition
tees where 2−3 nL of enzyme, substrate, and quenchant could
be added for each step. The final droplet size was about 16 nL.
Fixing the droplet flow rate at 2 μL/min, one reagent addition
was completed at 1.4 Hz. 100 droplets were analyzed in 2.2 min
(0.8 Hz), as illustrated in Figure 4B.
XICs of the assay product ZRR-OH allowed easy detection
of inhibitors and noninhibitors. Some carry-over was observed
in that signal of the first droplet of each inhibitor set was higher
than the rest. Also, signal for the first droplet of noninhibitor
was lower than its replicates if following an inhibitor, as pointed
out in Figure 4B. To avoid the impact of carry-over, RSD and
Z-factors were calculated using the signal intensity of the 2nd−
4th droplet of each set. The RSD was 11% on average, which
was larger than in-well screening due to the higher throughput
which compromised the number of data points for each droplet
and the accumulation of the slight variation in each reagent
addition. The Z-factors were 0.82, 0.85, and 0.76 for E-64,
leupeptin, and antipain, respectively. For each reaction, 1.6
picomoles of ZRR-AMC, 0.8 picomoles of test compound, and
5 femtomoles of Cathepsin B were consumed. This represents a
1000-fold reduction compared to common MWP platforms.
Importantly, for the droplet screening, the small amounts used
can be achieved with no waste or requirements for larger
reservoirs of compounds.
For these experiments, we used the reagent addition tee
shown in Figure 4A. The geometry of the tee was designed to
avoid formation of reagent-only droplets between sample
droplets. As reagent continuously flows out of the reagent
capillary, the interfacial tension between water and PFD
prevents it from being detached from the end of the channel.
When an aqueous droplet arrives at the junction, it merges with
the accumulated reagent and flows away from the tee. However,
if oil gaps between droplets are too long or the reagent flow
Figure 3. (A) Result of in-well screening of Cathepsin B inhibitors
using ZRR-AMC as the substrate. XICs show signal for ZRR-OH and
the known inhibitors E-64, leupeptin, and antipain. Each of 24 reaction
mixtures was formatted in quadruplicate. Droplets with low product
signal are inhibitor sets. (B) Result of screening using Ac-GFGFVGG-
NH2 as the substrate. XICs are product FVGG-NH2 and the substrate.
Twelve reaction solutions were formatted in quintuplicate (see
Supporting Information for all test compounds).
in quintuplicate using unlabeled peptide substrate is shown in
Figure 3B. Each droplet is detected as a burst of current
separated by low signal that corresponds to oil exiting the
emitter nozzle. As shown previously, PFD does not spray or
generate signal under these conditions.33 Each sample current
burst consists of a series of mass spectra (∼10 spectra for each
droplet) so that different m/z channels can be simultaneously
monitored allowing detection of the substrate, product, and
inhibitors. The impact of inhibitors on the enzyme is evident in
low signal intensity for product when inhibitors are present.
D
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
addition.46 The broad outlet of the tee slows down the reagent
that prevents reagent-only droplets from forming, but it also
slows down the passing droplet that leads to longer diffusive
mixing of sample and reagent at the junction. Commonly,
carry-over is circumvented by inserting a “rinse” droplet
between samples.29 We achieved a similar effect here by
analyzing samples in quadruplicate and discarding the first
sample of each set. The carry-over of the second blank droplets
was <6%, even after 3 additions, which did not affect the Z-
factor.
Using replicate droplets is effective and easy to implement. It
also adds little time to a small scale screen. However, for large
screens, it would be desirable to avoid carry-over completely to
minimize the time spent rinsing between samples. We therefore
evaluated a Teflon tee.40 This tee yielded 2−5% carry-over per
reagent addition (Supporting Information, Figure S-3B.) The
raw trace for assay product FVGG-NH2 from a small-scale
screen with the nonfluorescent substrate is shown in Figure 5B.
Figure 4. (A) Diagram of all-droplet screening system coupled to ESI-
MS analysis. Test compounds arrayed as droplets were pumped into a
series of PDMS tees where enzyme, substrate, and quenchant were
added in. (For the actual experiment, the third Teflon tube was sealed
on both ends and incubated in a water bath prior to pumping droplets
through the quenchant addition tee.) The inset is a photomicrograph
of the PDMS tee. The droplet inlet is silanized 50 μm i.d. × 150 μm
o.d. capillary; the outlet is 150 μm i.d. × 360 μm o.d. Teflon tube and
the reagent channel is nonderivatized 20 μm i.d. × 150 μm o.d.
capillary. (B) Results of all-droplet screening using ZRR-AMC as the
substrate. XICs are ZRR-OH, antipain, E-64, and leupeptin. Each test
compound has 4 replicates (see Supporting Information for details).
Figure 5. (A) Photomicrograph of Teflon tee. The droplet channel is
150 μm i.d. × 360 μm o.d. Teflon tube. The reagent channel is 40 μm
i.d. × 190 μm o.d. nonderivatized capillary. (B) Trace resulting from
all-droplet screening with the Teflon tee using the nonfluorescent
substrate. XIC is the product FVGG-NH2 (see Supporting
Information for test compounds). Samples were run in quintuplicate.
rate is too high, reagent droplets may form and be swept off by
the oil. The results of having reagent-only droplets include
irreproducible addition ratio, nonuniform droplet train, and
uneven pressure, all of which are detrimental to multistep
reactions. As described elsewhere, a larger outlet effectively
prevents reagent-only droplet formation.45 Indeed, no un-
expected signal was detected between droplets (Figure 4B)
showing that this design prevented reagent-only droplets.
Carry-Over and Alternative Reagent Addition Device.
Although the all-droplet system performed well, we further
studied it to determine if improvements were possible. Our
carry-over measurement showed that <1% of carry-over was
caused by the emitter, in agreement with previous results (see
Supporting Information, Figure S-3A).34 One reagent addition
using the PDMS tee resulted in ∼9% of carry-over, which
increased to 16% with 2 additions and 25% with 3 additions
(Supporting Information, Figure S-3A). This carry-over appears
The total carry-over in the first droplet was 10−14% and
negligible in the following droplets. The low carry-over of the
Teflon tee can be attributed to (1) a narrow outlet that allowed
relatively high velocity through tee; (2) partial coverage of the
reagent outlet by Teflon that prevents droplets from sticking to
the capillary. Both tees were demonstrated as good reagent
addition devices for this application. Other tee designs have
been reported to yield low carry-over and high addition ratio
that may produce even better results.5
Throughput. The emphasis in this work is on miniaturizing
label-free enzyme assays. This miniaturization should be
valuable when protein target or reagents are difficult or
expensive to obtain. Miniaturization is also beneficial in scale-
up of a screen. The biggest expense of large scale screens is
often reagent costs,29 so reduction of volume can make screens
more affordable. Droplet assays also have potential utility in
́
to be related to the low Peclet number (ratio of advection rate
to diffusion rate) developed in the tee. Advection facilitates
merging while diffusion causes contamination during reagent
E
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
(5) Li, L.; Boedicker, J. Q.; Ismagilov, R. F. Anal. Chem. 2007, 79,
2756−2761.
high-throughput screening. For the all droplet-assay, the steps
to completing a screen are: (1) droplet generation, (2) all-
droplet reactions, and (3) MS analysis. In this experiment,
around 10 min is required to load 96 droplets using the XYZ-
positioner; however, faster positioners and parallel operation
can greatly reduce this time. Further, it may be feasible to
reformat libraries to droplets in advance, thus eliminating this
time for a screen. Reagent addition was performed at 1.4
samples/s, which was comparable with other studies, although
rates up to 10 Hz have been reported.22 The reactions
themselves require 15−20 min of incubation. This time cannot
be eliminated; however, if the fluidics is operated in a steady
stream, then after an initial lag time, samples would be
produced at a rate limited by the reagent addition rate.
Alternatively, tubes could be prepared in parallel, and reactions
could be done in batches. Finally, the ESI-MS rate was 0.5 to 1
Hz here, and faster rates may be possible with a faster scanning
MS.
(6) Song, H.; Chen, D. L.; Ismagilov, R. F. Angew. Chem., Int. Ed.
2006, 45, 7336−7356.
(7) Niu, X.; Gielen, F.; Edel, J. B.; deMello, A. J. Nat. Chem. 2011, 3,
437−442.
(8) Adamson, D. N.; Mustafi, D.; Zhang, J. X. J.; Zheng, B.;
Ismagilov, R. F. Lab Chip 2006, 6, 1178−1186.
(9) Nie, J.; Kennedy, R. T. Anal. Chem. 2010, 82, 7852−7856.
(10) Abate, A. R.; Agresti, J. J.; Weitz, D. A. Appl. Phys. Lett. 2010, 96,
203509.
(11) Ahn, B.; Lee, K.; Panchapakesan, R.; Oh, K. W. Biomicrofluidics
2011, 5, 024113.
(12) Cai, L. F.; Zhu, Y.; Du, G. S.; Fang, Q. Anal. Chem. 2012, 84,
446−452.
(13) Pei, J.; Li, Q.; Kennedy, R. T. J. Am. Soc. Mass Spectrom. 2010,
21, 1107−1113.
(14) Pei, J. A.; Nie, J.; Kennedy, R. T. Anal. Chem. 2010, 82, 9261−
9267.
(15) Ahn, K.; Agresti, J.; Chong, H.; Marquez, M.; Weitz, D. A. Appl.
Phys. Lett. 2006, 88, 264105.
CONCLUSION
■
(16) Mazutis, L.; Baret, J. C.; Treacy, P.; Skhiri, Y.; Araghi, A. F.;
Ryckelynck, M.; Taly, V.; Griffiths, A. D. Lab Chip 2009, 9, 2902−
2908.
A label-free all-droplet assay system was developed in this work.
Its robustness, ultralow reagent consumption, and high
throughput were demonstrated by the screening of Cathepsin
B inhibitors. Compared with MWP-based fluorescence assay
systems, this approach eliminates the need of labeling and
reduces the sample requirement 1000-fold. These results
suggest the potential for screening reactions for optimization,
chemical probe discovery, and drug discovery, especially when
reagent or protein target are expensive or difficult to obtain.
Further development is required for routine high-throughput
screening. One limitation of the system is the build-up of carry-
over in a multistep reaction. Modifications of the tee such as
changing the dimension of the channels will be explored in
order to reduce the carry-over. In pursuit of higher throughput,
parallel droplet generation, parallel droplet-based reactions, and
a faster mass spectrometer may be incorporated.
(17) Roach, L. S.; Song, H.; Ismagilov, R. F. Anal. Chem. 2005, 77,
785−796.
(18) Srisa-Art, M.; Dyson, E. C.; Demello, A. J.; Edel, J. B. Anal.
Chem. 2008, 80, 7063−7067.
(19) He, M. Y.; Edgar, J. S.; Jeffries, G. D. M.; Lorenz, R. M.; Shelby,
J. P.; Chiu, D. T. Anal. Chem. 2005, 77, 1539−1544.
(20) Clausell-Tormos, J.; Lieber, D.; Baret, J. C.; El-Harrak, A.;
Miller, O. J.; Frenz, L.; Blouwolff, J.; Humphry, K. J.; Koster, S.; Duan,
H.; Holtze, C.; Weitz, D. A.; Griffiths, A. D.; Merten, C. A. Chem. Biol.
2008, 15, 875−875.
(21) Shim, J. U.; Olguin, L. F.; Whyte, G.; Scott, D.; Babtie, A.; Abell,
C.; Huck, W. T. S.; Hollfelder, F. J. Am. Chem. Soc. 2009, 131, 15251−
15256.
(22) Trivedi, V.; Doshi, A.; Kurup, G. K.; Ereifej, E.; Vandevord, P. J.;
Basu, A. S. Lab Chip 2010, 10, 2433−2442.
(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) Zheng, B.; Roach, L. S.; Ismagilov, R. F. J. Am. Chem. Soc. 2003,
125, 11170−11171.
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
(25) Kreutz, J. E.; Shukhaev, A.; Du, W. B.; Druskin, S.; Daugulis, O.;
Ismagilov, R. F. J. Am. Chem. Soc. 2010, 132, 3128−3132.
(26) Lunn, C. A. Future Med. Chem. 2010, 2, 1703−1716.
(27) Cunningham, B. T.; Laing, L. G. Expert Opin. Drug Discovery
2008, 3, 891−901.
AUTHOR INFORMATION
Corresponding Author
*E-mail: rtkenn@umich.edu.
Notes
The authors declare the following competing financial
interest(s): We have a financial interest in a company that is
licensing some of the technology described here.
■
(28) Shiau, A. K.; Massari, M. E.; Ozbal, C. C. Comb. Chem. High
Throughput Screening 2008, 11, 231−237.
(29) Hatakeyama, T.; Chen, D. L.; Ismagilov, R. F. J. Am. Chem. Soc.
2006, 128, 2518−2519.
(30) Fidalgo, L. M.; Whyte, G.; Ruotolo, B. T.; Benesch, J. L. P.;
Stengel, F.; Abell, C.; Robinson, C. V.; Huck, W. T. S. Angew. Chem.,
Int. Ed. 2009, 48, 3665−3668.
ACKNOWLEDGMENTS
This work was supported by NSF grant CHE-058903. We
thank New Objective Inc. for providing Pt-coated emitters.
■
(31) Kelly, R. T.; Page, J. S.; Marginean, I.; Tang, K. Q.; Smith, R. D.
Angew. Chem., Int. Ed. 2009, 48, 6832−6835.
(32) Zhu, Y.; Fang, Q. Anal. Chem. 2010, 82, 8361−8366.
(33) Li, Q.; Pei, J.; Song, P.; Kennedy, R. T. Anal. Chem. 2010, 82,
5260−5267.
REFERENCES
■
(34) Pei, J.; Li, Q.; Lee, M. S.; Valaskovic, G. A.; Kennedy, R. T. Anal.
Chem. 2009, 81, 6558−6561.
(1) Casadevall i Solvas, X.; deMello, A. Chem. Commun. 2011, 47,
1936−1942.
(2) Chiu, D. T.; Lorenz, R. M. Acc. Chem. Res. 2009, 42, 649−658.
(3) Furman, W. B. Continuous-Flow Analysis. Theory and Practice; M.
Dekker: New York, 1976.
(35) Hook, V.; Funkelstein, L.; Wegrzyn, J.; Bark, S.; Kindy, M.;
Hook, G. Biochim. Biophys. Acta, Proteins Proteomics 2012, 1824, 89−
104.
(36) Sevenich, L.; Schurigt, U.; Sachse, K.; Gajda, M.; Werner, F.;
(4) Abate, A. R.; Hung, T.; Mary, P.; Agresti, J. J.; Weitz, D. A. Proc.
Natl. Acad. Sci. U.S.A. 2010, 107, 19163−19166.
Muller, S.; Vasiljeva, O.; Schwinde, A.; Klemm, N.; Deussing, J.;
F
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Peters, C.; Reinheckel, T. Proc. Natl. Acad. Sci. U.S.A. 2010, 107,
2497−2502.
(37) Pozgan, U.; Caglic, D.; Rozman, B.; Nagase, H.; Turk, V.; Turk,
B. Biol. Chem. 2010, 391, 571−579.
(38) Jilkova, A.; Rezacova, P.; Lepsik, M.; Horn, M.; Vachova, J.;
Fanfrlik, J.; Brynda, J.; McKerrow, J. H.; Caffrey, C. R.; Mares, M. J.
Biol. Chem. 2011, 286, 35770−35781.
(39) de Boer, A. R.; Letzel, T.; van Elswijk, D. A.; Lingeman, H.;
Niessen, W. M. A.; Irth, H. Anal. Chem. 2004, 76, 3155−3161.
(40) Slaney, T. R.; Nie, J.; Hershey, N. D.; Thwar, P. K.; Linderman,
J.; Burns, M. A.; Kennedy, R. T. Anal. Chem. 2011, 83, 5207−5213.
(41) McDonald, J. C.; Whitesides, G. M. Acc. Chem. Res. 2002, 35,
491−499.
(42) Biniossek, M. L.; Nagler, D. K.; Becker-Pauly, C.; Schilling, O. J.
Proteome Res. 2011, 10, 5363−5373.
(43) Gosalia, D. N.; Salisbury, C. M.; Ellman, J. A.; Diamond, S. L.
Mol. Cell Proteomics 2005, 4, 626−636.
(44) Zhang, J. H.; Chung, T. D. Y.; Oldenburg, K. R. J. Biomol.
Screening 1999, 4, 67−73.
(45) Sivasamy, J.; Chim, Y. C.; Wong, T.-N.; Nguyen, N.-T.; Yobas,
L. Microfluid. Nanofluid. 2009, 8, 409−416.
(46) Song, H.; Li, H. W.; Munson, M. S.; Van Ha, T. G.; Ismagilov,
R. F. Anal. Chem. 2006, 78, 4839−4849.
G
dx.doi.org/10.1021/ac3011389 | Anal. Chem. XXXX, XXX, XXX−XXX