A FLUID ARRAY DEVICE FOR
HIGH-THROUGHPUT PROTEIN SYNTHESIS
Z. Hugh Fan1,2, Ruba Khnouf2, Qian Mei1, Shouguang Jin3
1
Department of Mechanical & Aerospace Engineering, 2Department of Biomedical Engineering, 3Department of
Molecular Genetics and Microbiology, University of Florida, PO Box 116250, Gainesville, FL 32611, USA
ABSTRACT
We describe a miniaturized fluid array device for high-throughput cell-free protein synthesis (CFPS). The device
consists of 96 units and each unit is for expression of one protein; thus up to 96 proteins can be produced
simultaneously. The function of the fluid array was demonstrated by expression of a variety of proteins, with more than
2 orders of magnitude reduction in reagent consumption compared to a commercially available CFPS instrument. The
protein expression yield in the device was up to 87 times higher than that in a conventional microplate.
KEYWORDS: Cell-free protein synthesis, miniaturization, high-throughput, proteomics
INTRODUCTION
Biological protein synthesis (expression) is often required for studying a gene’s function since the corresponding
protein is needed to characterize the structure and biological properties. As more and more new genes are being
identified, there is a considerable need to have high-throughput methods to produce a large number of proteins in
parallel, matching the throughput and scale of gene discovery. Protein synthesis is primarily implemented using overexpression in E. coli cells, but cell-based methods can be difficult and cost-prohibitive to implement in a highthroughput format.
An alternative approach is cell-free protein synthesis (CFPS), which has been developed to address the limitations
encountered in E. coli-based protein production, including cytotoxicity and susceptibility to proteolysis [1, 2].
However, current CFPS still requires a large volume of reagents, which could be cost-prohibitive when used for highthroughput assays.
Miniaturization of CFPS is the logical step to meet the need of cost-effective, high-throughput protein production.
Miniaturization could also reduce the space required for implementing a large number of protein expression units
simultaneously [3]. However, current efforts in using microfabricated wells for CFPS have not taken advantage of
microtechnology for fluid manipulation [4, 5]. As a result, nutrients cannot be replenished and inhibitory byproducts
cannot be removed, thus significantly reducing protein expression yield. Published experimental results suggest that
high-yield protein expression can be obtained using continuous flow of a feeding solution, but not under static
conditions in a fixed reaction volume [1]. Efforts have also been made in applying microfluidics to CFPS [6-8], but
they have not addressed the need for high-throughput parallel protein production.
In this work, we report a fluid array device for high-throughput cell-free protein synthesis (CFPS), which has been
developed to address the limitations encountered in E. coli-based protein production [1, 2]. The array device is capable
of producing 96 proteins simultaneously. Production of six proteins with a combination of five enzymatic assays was
demonstrated in the fluid array.
EXPERIMENTAL
Device Fabrication. As shown in Figure 1, the fluid array device is made of 3 layers, comprising 96 units for
simultaneous production of the same number of proteins. Each unit is composed of three access holes and one reaction
chamber in the top layer, a dialysis membrane as the middle layer, and a feeding chamber in the bottom layer. Protein
synthesis takes place in the reaction chamber while the feeding chamber functions as the nutrient reservoir. A
microfluidic channel connecting them provides a means to supply nutrients continuously, ensure proper mixing, and
remove the reaction byproducts. The function of the membrane is to (1) allow small-molecule reactants/nutrients to
transfer; (2) retain proteins produced and large-molecule synthesis machinery; and (3) dilute the byproducts of the
reactions and reduce their effects on the reaction equilibrium.
(a)
(b)
access holes
reaction chamber
access holes
reaction
chamber
dialysis
membrane
feeding
chamber
feeding microchannel membrane
chamber
Figure 1: (a) Schematic of a fluid array consisting of 3 layers and 96 units. (b) Cross-sectional view of one
unit showing the relative position among components and between reagents. The drawing is not to scale.
978-0-9798064-3-8/µTAS 2010/$20©2010 CBMS
761
14th International Conference on
Miniaturized Systems for Chemistry and Life Sciences
3 - 7 October 2010, Groningen, The Netherlands
Protein Expression. Luciferase, -glucoronidase (GUS), alkaline phosphatase (AP), -lactamase (-lac), and galactosidase (lacZ gene product) were expressed using RTS 100 wheat germ kit (Roche). Green fluorescent protein
(GFP) was expressed in the RTS 500 E. coli kit. The reaction solution for protein expression and the feeding solution of
nutrients were prepared according to the manufacturer’s instructions. DNA vectors of GUS and GFP were obtained
from Roche while luciferase and lacZ were from Promega and Clontech, respectively. Both -lac and AP vectors were
constructed in house. To carry out protein expression in the fluid array, 10 L of the reaction solution was pipeted to the
reaction chamber and 200 L of the feeding solution was added to the feeding chamber through one of the access holes.
Protein Assays. The synthesized proteins were measured using a commercial plate reader (Mithras LB 940,
Berthold Technologies). GFP was detected by measuring the amount of its native fluorescence. Luciferase was
determined by adding 30 L luciferase assay reagent (LAR, Promega) into the corresponding reaction chamber by an
injector in the plate reader and measuring luminescence. GUZ was quantified by injecting 30 L 100 M 4methylumbelliferyl -D-glucuronide (MUG, Marker Gene Technologies), incubating for 6 minutes, and then measuring
the fluorescence. The same approach was used to measure AP while the assay substrate was 30 L of 10 M 3phenylumbelliferone 7-O-phosphate hemipyridinium (PPH, Marker Gene Technologies). LacZ was determined by
injecting 30 l 200 M fluorescein mono--D-galactopyranoside (FMG, Marker Gene Technologies), incubating for 10
minutes, measuring the fluorescence. -lac was measured by adding 90 L 2 mM m-([(phenylacetyl)glycyl]oxy)benzoic
acid (PBA, Calbiochem). After incubation for 5 minutes, the sample was transferred into a cuvette and its absorbance
was measured at 314 nm.
Normalized Expression Yield
Expression Yield (normalized)
RESULTS AND DISCUSSION
1.2
High-yield synthesis. We compared the fluid array with
a standard microplate for protein synthesis. Their temporal
1
profiles of luciferase expression are shown in Figure 2. In
the standard microplate, all protein expression components
0.8
were mixed and then pipetted into a well. In the fluid array,
array device
however, the nutrients in the feeding chamber were
0.6
continuously supplied through the channel and membrane
0.4
into the reaction chamber. The protein synthesis stopped
after half an hour in the standard microplate whereas it
standard plate
0.2
continued for up to 6 hours in the fluid array. Continuous
supply of nutrients was due primarily to the concentration
0
gradient of each component between two chambers [9]. The
0
1
2
3
4
5
6
7
fluid array enabled the reaction to last 12 times longer, with a
Expression Time (h)
52 fold increase in luciferase expression yield.
We studied the expression of five other proteins in the Figure 2: Comparison between luciferase synthesis in
fluid array, including GFP, GUS, lacZ, -lac, and AP. For all the fluid array device (open circles) and in a standard
of these proteins, we obtained higher expression yields in the microplate (solid circles). The normalized synthesis
fluid array when compared to expression in a standard yield of luciferase is plotted as a function of the
microplate. Among these proteins, GUS, luciferase and GFP synthesis time. Each data point represents an average
had significant improvement and their expression yield obtained from three repeat experiments.
increased 87, 52, and 45 fold, respectively.
1.2
In addition to higher synthesis yield, one of the major
2 nm
advantages of the fluid array device is much lower reagent
consumption and significant cost-saving when it is compared to
the conventional cell-free protein synthesis apparatus. The
0.8
10 nm
volume of the reaction solution in fluid array is 10 μL, which is 2
orders of magnitude less than 1 mL of a reaction solution used in
15 nm
tube
RTS 500 (a kit for a commercially available CFPS instrument)
[10]. The cost-saving would be substantial when a large array is
0.4
required for high-throughput applications.
100 nm
Efforts of membrane. In the previous CFPS work [1, 2, 9],
dialysis membranes are employed for separating the reaction and
0.0
feeding solutions. We investigated the use of nanoporous
0
2
4
6
8
10
membranes for CFPS since it is possible to laminate a
Time (hr.)
nanoporous membrane thermally with other plastic layers as
practiced in industry. Dialysis membranes are required to be Figure 3: The effects of the pore size of nanoporous
stored in a solution, making it difficult to handle during the membranes on the protein synthesis yield. The
manufacturing process. Membranes with different pore sizes are expression yield is normalized against the 2-nm-pore
investigated. The effects of the pore size on the luciferase membrane. The temporal profile of the luciferase
synthesis yield are shown in Figure 3. When the 100-nm synthesis in a microcentrifuge tube is also provided
membrane is used, no significant difference exists between the for comparison.
762
LacZ
GUS
Luciferase
GFP
AP
Blac
Co
Neg
Expression Yield
device and a microcentrifuge tube. This result indicates that reagents can flow between two chambers and proteins
synthesized cannot be retained. When the pore size decreases from 100 nm to 15 nm, the protein synthesis yield improves
considerably. The synthesis yield further improves when the pore size decreases to 10 nm, and then to 2 nm. It is
understandable because the molecular cutoff of 2 nm membrane is 8 KDa and that of a 10 nm membrane is about 100 KDa
while the molecular weight of luciferase is 61 KDa. A fraction of luciferase molecules likely pass through the 10 nm
membrane, but very few luciferase molecules transport through the 2 nm membrane. The results indicate that the pore size
of the membrane is critical for the protein synthesis yield. The pores should be large enough to allow small molecules (e.g.,
amino acids and ATP) to pass, but small enough to
prevent large molecules (e.g., proteins and ribosomes)
from transporting.
1
High-throughput synthesis. High-throughput
0.8
protein synthesis is illustrated in Figures 4. Each unit
0.6
in the device was used for one condition for protein
synthesis and assay. Rows A to F were designated for
0.4
producing six different types of proteins. Row G was
0.2
used for co-expression of all of these six proteins.
0
Row H functioned as a negative control, which
1 2
A GFP
contained no DNA expression vectors. At the end of
B luciferase
3 4
C
5 6
GUS
D
protein synthesis, appropriate assay reagents were
E
7
lacZ
F
8
AP
added according to columns 1-12 and the signal in
9
G
10
H
E-lac
11
each unit was then measured. The adjacent two
12
co-expression
negative
columns were subjected to the same assay to enhance
the precision. The same assays were performed in a
standard microplate for comparison, showing higher
Figure 4: Simultaneous synthesis and detection of six proteins
protein synthesis yields in the fluid array.
in a fluid array. Proteins are indicated on the right of each
row, including green fluorescent protein (GFP), CONCLUSION
The miniaturized fluid array device has been glucoronidase (GUS), -galactosidase (lacZ), -lactamase (developed for high-throughput cell-free protein lac), and alkaline phosphatase (AP). At the end of protein
synthesis, having a potential to match the throughput synthesis, appropriate assay reagents were added according to
and scale of gene discovery. It has addressed the columns 1-12 and the signal in each unit was then measured.
cost (of setup and reagents) and other limitations of GFP was detecting by its natural fluorescence. The assay
E. coli cells-based production of recombinant reagents for detecting GUS, luciferase, lacZ, AP, and -lac
proteins for high-throughput applications. The array were 4-methylumbelliferyl -D-glucuronide (MUG), luciferase
device has been demonstrated for expression of a assay reagent (LAR), fluorescein mono--D-Galactopyranoside
variety of proteins, with more than 2 orders of (FMG), 3-phenylumbelliferone 7-O-phosphate hemipyridinium
magnitude reduction in reagent consumption. As a (PPM), and m-([(phenylacetyl)glycyl]oxy)benzoic acid (PBA),
result, there is a great potential of the array device in respectively.
proteomics applications.
ACKNOWLEDGEMENTS
This work was supported in part by Defense Advanced Research Projects Agency (DARPA) via Micro/Nano
Fluidics Fundamentals Focus Center at the University of California at Irvine, and the University of Florida via UF
Opportunity Fund.
REFERENCES
[1] Spirin, A. S., Baranov, V. I., Ryabova, L. A., Ovodov, S. Y., Alakhov, Y. B., Science 1988, 242, 1162-1164.
[2] Katzen, F., Chang, G., Kudlicki, W., Trends Biotechnol 2005, 23, 150-156.
[3] Gilbert, M., Albala, J. S., Curr Opin Chem Biol 2002, 6, 102-105.
[4] Angenendt, P., Nyarsik, L., Szaflarski, W., Glokler, J., et al., Anal. Chem. 2004, 76, 1844-1849.
[5] Mei, Q., Fredrickson, C. K., Jin, S., Fan, Z. H., Anal Chem 2005, 77, 5494-5500.
[6] Yamamoto, T., Fujii, T., Nojima, T., Lab Chip 2002, 2, 197-202.
[7] Tabuchi, M., Hino, M., Shinohara, Y., Baba, Y., Proteomics 2002, 2, 430-435.
[8] Asthana, A., Kim, K. O., Perumal, J., Kim, D. M., Kim, D. P., Lab Chip 2009, 9, 1138-1142.
[9] Mei, Q., Fredrickson, C. K., Simon, A., Khnouf, R., Fan, Z. H., Biotechnol Progr 2007, 23, 1305-1311.
[10] Betton, J. M., Curr Protein Pept Sci 2003, 4, 73-80.
CONTACT
*Z. Hugh Fan, tel: +1-352-8463021; [email protected]
763