Massively Parallel Single-Molecule and Single-Cell Emulsion Reverse

Mar 27, 2012 - For a more comprehensive list of citations to this article, users are encouraged ... Single-Cell Forensic Short Tandem Repeat Typing wi...
3 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Massively Parallel Single-Molecule and Single-Cell Emulsion Reverse Transcription Polymerase Chain Reaction Using Agarose Droplet Microfluidics Huifa Zhang,† Gareth Jenkins,†,‡ Yuan Zou,† Zhi Zhu,† and Chaoyong James Yang*,† †

State Key Laboratory of Physical Chemistry of Solid Surfaces, The Key Laboratory for Chemical Biology of Fujian Province, Key Laboratory of Analytical Science and Department of Chemical Biology, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, China ‡ Institute of Biomedical Engineering, Imperial College London, South Kensington, London SW7 2AZ, United Kingdom S Supporting Information *

ABSTRACT: A microfluidic device for performing single copy, emulsion Reverse Transcription Polymerase Chain Reaction (RT-PCR) within agarose droplets is presented. A two-aqueous-inlet emulsion droplet generator was designed and fabricated to produce highly uniform monodisperse picoliter agarose emulsion droplets with RT-PCR reagents in carrier oil. Template RNA or cells were delivered from one inlet with RT-PCR reagents/cell lysis buffer delivered separately from the other. Efficient RNA/cell encapsulation and RT-PCR at the single copy level was achieved in agarose-in-oil droplets, which, after amplification, can be solidified into agarose beads for further analysis. A simple and efficient method to graft primer to the polymer matrix using 5′acrydite primer was developed to ensure highly efficient trapping of RT-PCR products in agarose. High-throughput single RNA molecule/cell RT-PCR was demonstrated in stochastically diluted solutions. Our results indicate that single-molecule RT-PCR can be efficiently carried out in agarose matrix. Single-cell RT-PCR was successfully performed which showed a clear difference in gene expression level of EpCAM, a cancer biomarker gene, at the single-cell level between different types of cancer cells. This work clearly demonstrates for the first time, single-copy RT-PCR in agarose droplets. We believe this will open up new possibilities for viral RNA detection and single-cell transcription analysis.

D

immobilized onto microbeads and other reagents.1,8,13 The use of microbeads is essential to maintain monoclonality of the droplets during downstream processing.1,8,13 Ideally, each aqueous droplet should encapsulate one microbead together with similar volumes/quantities of template and reagents to ensure reliable and efficient results. However, the low monodispersity of current techniques for emulsion generation leads to wide droplet to droplet variations in PCR reagent volumes and the probability of encapsulating one microbead in each droplet is low. This leads to low efficiencies and reduces

iseases often begin with cellular abnormalities in a small minority of cells1−4 within an organism. Therefore, the development of single-molecule and single-cell assays has great potential for a multitude of applications in oncology, including early cancer diagnosis, disease prognosis, and treatment evaluation.5 The use of water-in-oil droplets for the effective compartmentalization of reagents during emulsion PCR (ePCR) has become a well-established and effective method of performing single-cell or single-molecule analyses due to the benefits of being able to perform millions of independent reactions in parallel.6−9 Such advances have led to the ability to identify and quantify rare mutant genes within large populations8,10−12 and have also enabled the new generation of high-throughput sequencing systems.13,14 Current methods for performing ePCR rely upon the strong agitation of aqueous and oil phases to generate droplets together with primers © 2012 American Chemical Society

Received: December 13, 2011 Accepted: March 27, 2012 Published: March 27, 2012 3599

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

we believe our approach will lead to new possibilities for applications, such as quantitative viral RNA detection, rare RNA mutation analysis, and single-cell gene expression studies.

the number of parallel reactions which can be successfully performed.15 Microfluidic techniques for droplet generation allow a high degree of droplet control with particularly good monodispersity.1,16,17 Hence, such techniques are well suited to ePCR. Microfluidic droplet techniques have been increasingly applied to PCR due to the inherent advantages of compartmentalizing reactions into discrete and well-defined volumes.18 The generation and manipulation of large combinatorial libraries is possible without cross-contamination between droplets and the common problem of amplification bias can be avoided.19 In addition, rapid thermocycling of small droplet volumes is possible enabling fast amplification and the high monodispersity allows for more quantitative detection compared with conventional emulsions. Microfluidics can also enable more sophisticated analyses such as multiplexing using novel chip geometries20,21 and may allow a more integrated approach to library preparation for high-throughput sequencing.22 Recently, we have developed an agarose droplet microfluidic approach for highly parallel and efficient single-copy DNA ePCR. Our method capitalizes on the unique thermoresponsive sol−gel switching property of agarose to eliminate the need to encapsulate microbeads.15 Entrapping the reaction components and amplicons within a homogeneous agarose solution matrix avoids the negative effects of steric hindrance and charge repulsion associated with microbeads. More importantly, using the sol−gel switching property of agarose to form stable beads after amplification ensures effective maintenance of monoclonality of each droplet and affords flexible schemes for downstream processing and applications such as sequencing, FACS analysis, and long-term storage. With this method, a highly efficient and cost-effective method for aptamer screening from a complex single-stranded DNA library was developed,23 demonstrating the flexibility of agarose droplet amplification for downstream applications. However, only DNA amplification has been demonstrated in previous work. To develop a method for RNA sequencing, single-cell gene expression studies,24−29 and for rapid, digital, quantitative detection of low-abundance (e.g., rare mutation) RNA and viral loading studies,30 we seek to apply the advantages of agarose droplet microfluidics to high-throughput, single-copy, reverse transcription PCR (RT-PCR). In this work, a two-aqueous-inlet microfluidic device for emulsion droplet generation was designed and fabricated to produce highly uniform and monodisperse picoliter agarose emulsion droplets containing well-defined amounts of sample and RT-PCR reagents. With this device, cells were injected from one inlet and cell lysis buffer from the other immediately prior to droplet generation. As a result, the cells remained intact until after encapsulation into the droplets. This method avoids premixing mRNAs from single cells and ensures uniform delivery of the agarose matrix together with RT-PCR components. Efficient RNA/cell encapsulation and RT-PCR at the single-molecule and singlecell level was thus achieved in agarose-in-oil droplets which, after amplification, could be solidified into beads for downstream processing and analysis. In this work, we first demonstrated high-throughput single RNA molecule RT-PCR using stochastically diluted solutions. We further presented results of single-cell RT-PCR using our method to show a clear differentiation in the level of gene expression of the EpCAM cancer biomarker gene between different types of cancer cells. For the first time, we demonstrated single-copy RT-PCR in agarose droplets and

■

EXPERIMENTAL SECTION Materials and Reagents. Glass (B270 glass, 150 μm thick) for microfluidic chip fabrication was purchased from Shaoguang Company Co., Ltd. (Changsha, China). 749 FLUID and 5225C FORMULATION AID were purchased from Dow Corning Company (Midland, MI). Silicone oil AR 20 (10836-500 ML), octadecyltrichlorosilane (OTS, 104817-25G), and agarose (A2576-5G, ultra-low gelling temperature) were purchased from Sigma-Aldrich (St. Louis, MO). One Step RT-PCR Kit was supplied by Tiangen Biotech (Beijing, China). Easy Taq DNA Polymerase was purchased from TransGen Biotech (Beijing, China). SYBR Green (20 × 500 μL, Biovision) was used to stain RT-PCR product in agarose beads. All the oligonucleotides used in the work were synthesized on a 12Column DNA Synthesizer (PolyGen GmbH) and HPLC purified (Agilent 1100) in house. N,N,N′,N′-Tetramethylethylenediamine (TEMED), ammonium persulfate (APS), acrylamide, and other reagents used in this work were purchased from Sigma (Shanghai, China). Kato III cell and MDA-MB-231 cell were gifts from Professor Wei Duan at the Deakin University, Australia. Device Microfabrication. Crossed-channel patterns were drawn using AutoCAD software and then printed onto a darkfield mask film. The microchannels on the mask were photolithographically transferred onto a glass surface and etched to produce a cross-channel with 170 μm (w) × 80 μm (d) for the aqueous channels and 230 μm (w) × 80 μm (d) for the oil channels, in a well-stirred bath containing diluted HF/ HNO3 solution. Five 1.8 mm diameter holes were drilled through the etched glass at channel terminals using a mechanical drill (TBM115, Proxxon). The etched glass was thermally bonded at 580 °C for 2 h to an identical thickness, featureless glass substrate in a Muffle furnace (Ney Centurion Q50, Dentsply Neytech, Burlington, NJ). After bonding, the microchannels were rinsed with isopropyl alcohol, acetone, piranha solution (H2SO4/H2O2, 3:1) and deionized water sequentially, and then dried with nitrogen gas. Following this, glass channels were hydrophobically treated with a 0.1% solution of octadecyltrichlorosilane in dry toluene for 5 min. The treated channels were rinsed using dry toluene, isopropyl alcohol, and deionized water sequentially. Primer Conjugation. To graft DNA primer onto the branch linear polyacrylamide (LPA), acrylic phosphoramidite was first synthesized. Briefly, 6-amino-1-hexanol (1 g, 8.53 mmol) and triethylamine (2.36 mL, 17 mmol) were cooled to 0 °C. Methacryloyl chloride (2.67 g, 2.55 mmol) was added dropwise, and the mixture was stirred at 0 °C for 2 h. After evaporation of all solvents, the residue was dissolved in 10 mL of ethanol and 20% sodium hydroxide (4 mL) was added into the solution. The solvent was evaporated, and 6-hydroxyhexyl methacrylamide was chromatographed on a column of silica gel G using ethyl acetate. N,N′-Diisopropylethylamine (DIPEA) (0.98 g, 7.50 mmol) was added slowly to a solution containing 6-hydroxyhexyl methacrylamide (0.50 g, 2.70 mmol) in anhydrous CH2Cl2 (10 mL) at 0 °C. Then, 2-cyanoethyl diisopropylchlorophosphoramidite (0.87 mL, 3.25 mmol) was added dropwise, and the reaction mixture was stirred at 0 °C for 2 h. After removing the solvent, the residue was purified by column chromatography (ethyl acetate/hexane/triethylamine 3600

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

Figure 1. Schematic overview of the operating principle described in this work. A microfluidic droplet generator mixes the aqueous solution I (RTPCR and cell lysis reagents in 2% agarose solution) and aqueous solution II (RNA or cells in 2% agarose solution) with precise control of reagent and template concentrations, into highly monodisperse and size tunable droplets. After cell lysis, reverse transcription reaction, and PCR are performed on the resulting droplets, the droplets are cooled to form agarose beads. The oil is removed by washing, after which the beads are stained with SYBR Green in order to allow downstream analysis by flow cytometry or fluorescence microscopy.

sequence emulsion RT-PCR. RT-PCR and cell lysis cocktail solution (Solution I) was prepared with the following final concentrations: 2% agarose, 2 × RT-PCR buffer, 0.8 mM of dNTP Mixture, 2.4 units of RNasin, 0.75 units of Hotmaster Taq polymerase, 2.4 units of Quant RTase, and 1% Triton X100 for effective cell lysis,31,32 1 μM of forward primer (5′-GCT CAG GAA GAA TGT GTC TGT GA-3′), 2 μM reverse primer (5′-Acrydited-GCA GCC AGC TTT GAG CAA ATG ACA-3′) grafted on a linear polyacrylamide. To avoid amplification of genomic DNA, both primers were designed to span two adjacent exons of EpCAM RNA with an amplicon size of 125 bp. Cells were added in another 2% agarose solution (Solution II) and injected into the droplet generator from the inlet separately from the RT-PCR and cell lysis cocktail solution (Solution I). Agarose Droplet Generation and Single-Copy Emulsion RT-PCR. The flow-focusing microfluidic device consisted of a cross junction with four inputs and one outlet channel through which agarose droplets flowed out of the device. The microfluidic device was mounted on an inverted fluorescence microscope (Nikon Eclipse Ti−U, Japan). The sample inlets were connected with syringe pumps (PHD infusion pump, Harvard Apparatus) through plastic tubing. In the condition of 0.1/1.8 mL/h (agarose/oil) flow rate, the generated droplets had diameters of 48 ± 5 μm and were collected in a tube until 100 μL of the agarose solution was collected. The conditions for RT-PCR for collected droplets were as follows: 50 °C for 30 min (reverse transcription), 94 °C for 3 min (initial denaturation), 29 cycles of 94 °C for 30 s, 54 °C for 30 s,

40:60:3) and dried to yield the title compound as a colorless oil. Acrylic phosphoramidite obtained from the last step was used to couple to the reverse primer produced using a DNA synthesizer to obtain acrydited reverse primer. After successfully synthesizing the acrydited reverse primer, DNA was grafted onto linear polyacrylamide (LPA) by copolymerization with acrylamide. A total of 10 μL of 1 mM acrydited reverse primer was mixed with 10 μL of 40% acrylamide in 80 μL of DI water. Then, 1.4 μL of 10% APS and 1.4 μL of 10% TEMED were added in the solution. The solution was immediately mixed well and kept in a vacuum to remove air and the polymerization reaction took about 12 min to complete. Under this condition, it is estimated that there is one DNA chain in every 5633 repeat units. Preparation for Single RNA Sequence Emulsion RTPCR. The droplet cocktail for single RNA sequence RT-PCR was prepared according to the following final concentrations: 2% agarose, 2× RT-PCR buffer, 1 μM of forward primer (5′AGC GAC TCT GAG GAG GAA CAA G-3′), 2 μM of reverse primer (5′-Acrydited-GCC AGG AGC CTG CCT CTT T-3′, grafted on a linear polyacrylamide), 0.8 mM of dNTP Mixture, 2.4 units of RNasin, 0.75 units of Hotmaster Taq polymerase, 2.4 units of Quant RTase. A 78 bp synthetic cMyc RNA was used as the target sequence in the experiment. RNA template was dissolved in another 2% agarose solution and injected into the droplet generator from the inlet separately from the cocktail solution. Preparation for Single-Cell Emulsion RT-PCR. For single-cell RT-PCR, the preparation is similar to single RNA 3601

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

Figure 2. (A) Synthesis scheme of 5′-acrydite primer grafting to LPA; (B) MALDI-TOF characterization of acrydited reverse primer for single-copy RT-PCR. The calculated molecular weight is 6008.8. (C) Electrophoresis image of the grafted primer and RT-PCR products carried out using the grafted primer and RNA template. Lane 1, free reverse primer; lane 2, grafted reverse primer; lanes 3 and 7, marker; lanes 4 and 8, RT-PCR negative control (no RNA template); lanes 5 and 9, RT-PCR with free reverse primer; lanes 6 and 10, RT-PCR with grafted reverse primer.

enables downstream analysis using fluorescence microscopy and flow cytometry to detect and quantify agarose beads containing DNA clones. During emulsion amplification, it is important to maintain the monoclonality of each droplet and prevent diffusion of amplicons into neighboring droplets. In conventional bead based ePCR, microbeads are used as solid supports for the attachment of primers in order to keep amplicons within the aqueous droplets. However, the use of a solid support can lead to steric hindrance and other negative effects, reducing amplification efficiency. Tethering one of the primers within a homogeneous agarose droplet should overcome such limitations. Previously, we have used a Schiff-base reaction to conjugate forward primer to the agarose matrix.15 However, the conjugation process is time-consuming with a less than 10% conjugation efficiency. In this work, we explored a new approach for primer conjugation to effectively prevent diffusion of RT-PCR reaction products. In this approach, the reverse primer was first tethered to a polymer chain by synthesizing 5′acrydite primer and was then grafted to a linear polyacrylamide (LPA) chain (Figure 2A).33 As Figure 2B shows, MALDI-TOF mass analysis indicated successful synthesis of the acryditedDNA. To confirm the success of the grafting, RT-PCR was carried out using the grafted primer and RNA template. The grafted primer and RT-PCR products were analyzed using 3% agarose gel electrophoresis. If the primer is successfully grafted to LPA, the grafted primer and RT-PCR products will be physically trapped in the 3% agarose gel pores and will not migrate during gel electrophoresis. As shown in Figure 2C, the grafted primer (lane 2) and amplification products (lanes 6 and 10) barely moved under electrophoresis, suggesting successful grafting of reverse primer onto the LPA. No free primer band was observed from lane 2, suggesting a nearly 100% conjugation efficiency. Such a physical trapping of amplicons inside agarose matrix prevents their diffusion out of the agarose, thus, permitting long-term processing of agarose beads without concern over the leakage of DNA products.

and 65 °C for 30 s, followed by a single final extension for 5 min at 65 °C in a Peltier thermal cycler (BIO-RAD, Richmond, CA). The amplified emulsions were solidified to agarose beads by cooling to 4 °C. After removing the oil phase by acetone, isopropyl alcohol, and deionized water sequentially, the agarose beads were stained by dispersing in 100 μL deionized water with 6 μL 20× SYBR Green for 1 h. Single-copy RT-PCR fluorescence images were observed with an inverted fluorescence microscope (Nikon Eclipse Ti− U, Japan). A total of 5 μL from each sample was exposed for 3 s at 10-fold magnification. . The fluorescence intensity of the stained beads was measured by a flow cytometry (FACSAria, BD Biosciences, Franklin Lakes, NJ) by counting 10 000 events.

■

RESULTS AND DISCUSSION The operating principle of our single-copy agarose droplet RTPCR approach is illustrated in Figure 1. A two-aqueous-inlet microfluidic droplet emulsion generator was used to produce highly uniform monodispersed picoliter agarose emulsion droplets containing RT-PCR reagents in carrier oil. Target solution is injected into the device through one inlet, while RTPCR and cell lysis reagent is injected from the other. This twoinlet design avoids premixing of the reaction components. Either cells or RNA targets are statistically diluted prior to injection such that there is no more than one cell or RNA molecule in each generated droplet. With this method, millions of agarose droplets can be generated in less than 1 h. After collection, droplets are thermocycled to perform RT-PCR with the resultant amplicons being trapped within the droplets. After thermocycling, the solution phase agarose droplets are solidified to solid gel phase by cooling the agarose droplets below the gelling point of the agarose and in this way producing stable agarose beads. Hence, the monoclonality of the RT-PCR product generated in each droplet can be maintained after removal of the oil phase. To detect and visualize agarose droplets containing DNA, SYBR Green is used to stain the DNA product trapped in the agarose beads. This staining label 3602

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

Figure 3. (A) Flow streams merge together to mix agarose solutions containing target and RT-PCR reagents (fluorescence used to aid visualization). (B) Image of droplet generation with a frequency of 300 Hz. (C) Fluorescence microscope image showing monodispersity of agarose droplets, 48 μm in diameter with a C.V. of 3.5%. (D) Agarose droplet size could be fine-tuned by changing the oil flow rate to meet different needs (agarose flow rate vs oil flow rate in mL/h is indicated).

Because RT-PCR is performed in an agarose matrix, a realtime quantitative RT-PCR assay was performed (n = 3) using bulk solutions of agarose to evaluate if agarose would affect RTPCR. It was found that the efficiency of RT-PCR in 2% agarose solution was as high as the RT-PCR in 0% agarose solution (see Supporting Information). At this concentration, the agarose could still be thermally switched to form a solid gel phase suitable for further analysis and processing. Therefore, we chose 2% as the optimized concentration of agarose in droplets to achieve droplet RT-PCR. To generate agarose droplets, a cross-channel microfluidics device was used in this study. In this design, two aqueous inlet channels merge together into a single channel in order to bring together the separate streams of agarose solution prior to droplet formation. In one inlet, the agarose solution contains the target RNA or cell and in the other inlet, it contains all the necessary reagents for RT-PCR (i.e., reverse transcriptase, Taq polymerase, forward and reverse primers, and dNTPs) and cell lysis. Statistically diluted targets are encapsulated into uniform picoliter agarose-in-oil droplets. Figure 3A shows the two inlet streams being brought together with fluorescence used to aid visualization. Diffusional mixing is minimal across the microchannel and uniform amounts of each stream enter each droplet after which rapid mixing occurs immediately after droplet formation.34 This method ensures that the reverse transcription reaction does not occur before droplet formation and that mixing is rapid immediately afterward, once all reagents are encapsulated in the droplet. This design is of particular importance for the single-cell RT-PCR experiments where no cells are lysed until after encapsulation into droplets to maintain integrity of the transcriptome of each cell and avoid cross-contamination. The laminar inlet zone leads to a droplet generator which uses flow focusing of oil streams to generate droplets containing the RT-PCR mixture (Figure 3B). The droplets generated are highly monodisperse and their size can be tuned by simple adjustment of the flow rates of aqueous and oil phase flows (Figure 3D). For example, using a chip with an oil channel of 230 μm and aqueous channel of 170 μm in width, uniform agarose droplets in the range of 40−140 μm can be easily obtained. Such monodispersity is very important to ensure uniform distribution of target molecules and uniform delivery of RT-PCR amplicons for sequencing or other quantitative genetic analysis. This allows for accurate control

over the reaction volumes used for RT-PCR. Figure 3C shows highly monodisperse 48 μm diameter droplets (C.V. = 3.5%). To demonstrate the feasibility of our agarose droplet microfluidic approach for single-molecule emulsion RT-PCR, a 78bp cMyc RNA sequence was chosen as the template for single-molecule RT-PCR in agarose droplets. With the use of the microfluidic chip, large populations of 2% agarose droplets containing RT-PCR reagents were generated for RT-PCR experiments. Droplets were collected and RT-PCR was performed. After cooling for 1 h at 4 °C, stable agarose beads were formed. After solidification and removal of the oil phase, agarose beads were stained with SYBR Green (which binds to double-stranded nucleic acids to form a fluorescent complex allowing visualization and quantification) and analyzed using a microscope. Fluorescent beads were observed after amplification using 1.5 RNA copies per droplet, while no fluorescence was seen in the control sample containing 0 copies of template (Figure 4), suggesting the feasibility of running transcription reaction and PCR in agarose emulsion droplets. Statistical analysis of large populations of agarose beads resulting from stochastically diluted levels of RNA templates (0, 0.15, 0.5, 0.75, 1, 1.5 copies/droplet) using fluorescence flow cytometry was performed to further evaluate the feasibility of single RNA template RT-PCR in agarose droplets. Figure 5A

Figure 4. Optical (A and C) and fluorescence (B and D) microscope images of agarose after RT-PCR at 0 (A and B) and 1.5 (C and D) copies RNA/droplet input concentration. 3603

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

Figure 5. Fluorescence flow cytometry analysis of agarose beads after amplification at the single-molecule level (A). Theoretical percentage of microbeads carrying RT-PCR product, calculated according to Poisson distribution and experimental data obtained from flow cytometry analysis (B).

Figure 6. Fluorescence flow cytometry analysis of agarose beads resulting from single-cell RT-PCR. Control (A): No cells. MDA-MB-231 (B): cell line with low EpCAM expression level. Kato III (C): cell line with high EpCAM expression level.

shows fluorescence flow cytometry analysis results. Increasing populations of positively amplified beads (right peaks) were observed with increasing RNA copies per droplet. Theoretically, the probability of a droplet having n RNA templates can be calculated by Poisson statistics f(n, λ) = (e−λλn/n!), where λ is the average number of RNA molecules per droplet before amplification.35 The percentage of positive beads (n ≠ 0) observed over a range of template concentrations at the singlecopy level agreed closely with the probability predicted by Poisson statistics (Figure 5B), demonstrating the feasibility of performing high-throughput single-molecule RT-PCR in agarose droplets. Agarose droplet RT-PCR efficiency was further evaluated by quantifying the numbers of amplicons in each positive agarose beads with a real-time quantitative PCR assay (see Supporting Information). After 29 cycles of agarose droplet emulsion RTPCR from RNA concentration of 1.5 copies RNA/droplet, the average numbers of amplicons (N) in each microbead were about 3.89 × 107. Correspondingly, the RT-PCR efficiency (E = ((N/1.5)1/29 −1) × 100%) in agarose droplets with a template concentration of 1.5 copies/bead, as calculated was 80.1%. Singe-cell RT-PCR was then carried out with two different cell lines each expressing a different level of Epithelial cell adhesion molecule (EpCAM). EpCAM is overexpressed in most solid cancers and it has recently been identified as a cancer stem cell marker.36−39 In this study, two different cells (Kato III cell and MDA-MB-231 cell) which have different

EpCAM expression levels40 were used as the targets for singlecell RT-PCR in agarose droplets. It has been estimated that there are about 1700 copies of EpCAM RNA in a MDA-MB231 cell, and 893 100 copies in a Kato III cell.41 To avoid amplification of genomic DNA, both primers were designed to span two adjacent exons of EpCAM RNA to produce a 125 bp amplicon. A real-time quantitative RT-PCR assay was performed (n = 3) to confirm the expression level difference between two cell lines. It was found that the cycle threshold number (Ct) of 10 000 Kato III cells was 19.8, while the Ct of 10 000 MDA-MB-231 cells was about 24.5, confirming that the RNA expression level in Kato III cells is significantly higher than the level in MDA-MB-231 cells (see Supporting Information). Fluorescence flow cytometry analysis of large populations of agarose beads resulting from stochastically diluted cells of 0.4 cells/droplet was performed to further evaluate the feasibility of single-cell RT-PCR in agarose droplets (Figure 6). No positive beads were found in the control sample (Figure 6A), while the percentages of positive beads of MDA-MB-231 (Figure 6B) and Kato III (Figure 6C) were 31.7% and 30.2%, respectively, which agree closely with the theoretically predicted Poisson distribution of 33% for the 0.4 cells/bead used. The larger shift of positively amplified beads (Figure 6 right peak) represents the higher level of EpCAM RNA expression. As shown in Figure 6, a significant shift of Kato III cells compared to MDA-MB-231 cells was observed, which is consistent with the result of the quantitative RT-PCR assay. Our results here establish the feasibility of 3604

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

(2) Asano, M.; Toda, M.; Sakaguchi, N.; Sakaguchi, S. J. Exp. Med. 1996, 184, 387−396. (3) Sugars, K. L.; Rubinsztein, D. C. Trends Genet. 2003, 19, 233− 238. (4) Hebbel, R. P.; Yamada, O.; Moldow, C. F.; Jacob, H. S.; White, J. G.; Eaton, J. W. J. Clin. Invest. 1980, 65, 154−160. (5) Spratlin, J. L.; Serkova, N. J.; Eckhardt, S. G. Clin. Cancer Res. 2009, 15, 431−440. (6) Leamon, J. H.; Link, D. R.; Egholm, M.; Rothberg, J. M. Nat. Methods 2006, 3, 541−543. (7) Kelly, B. T.; Baret, J. C.; Taly, V.; Griffiths, A. D. Chem. Commun. 2007, 1773−1788. (8) Dressman, D.; Yan, H.; Traverso, G.; Kinzler, K. W.; Vogelstein, B. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 8817−8822. (9) Guttenberg, Z.; Muller, H.; Habermuller, H.; Geisbauer, A.; Pipper, J.; Felbel, J.; Kielpinski, M.; Scriba, J.; Wixforth, A. Lab Chip 2005, 5, 308−317. (10) Zeng, Y.; Novak, R.; Shuga, J.; Smith, M. T.; Mathies, R. A. Anal. Chem. 2010, 82, 3183−3190. (11) Pekin, D.; Skhiri, Y.; Baret, J. C.; Le Corre, D.; Mazutis, L.; Ben Salem, C.; Millot, F.; El Harrak, A.; Hutchison, J. B.; Larson, J. W.; Link, D. R.; Laurent-Puig, P.; Griffiths, A. D.; Taly, V. Lab Chip 2011, 11, 2156−2166. (12) Hindson, B. J.; Ness, K. D.; Masquelier, D. A.; Belgrader, P.; Heredia, N. J.; Makarewicz, A. J.; Bright, I. J.; Lucero, M. Y.; Hiddessen, A. L.; Legler, T. C.; Kitano, T. K.; Hodel, M. R.; Petersen, J. F.; Wyatt, P. W.; Steenblock, E. R.; Shah, P. H.; Bousse, L. J.; Troup, C. B.; Mellen, J. C.; Wittmann, D. K.; Erndt, N. G.; Cauley, T. H.; Koehler, R. T.; So, A. P.; Dube, S.; Rose, K. A.; Montesclaros, L.; Wang, S. L.; Stumbo, D. P.; Hodges, S. P.; Romine, S.; Milanovich, F. P.; White, H. E.; Regan, J. F.; Karlin-Neumann, G. A.; Hindson, C. M.; Saxonov, S.; Colston, B. W. Anal. Chem. 2011, 83, 8604−8610. (13) Margulies, M.; Egholm, M.; Altman, W. E.; Attiya, S.; Bader, J. S.; Berka, J.; Braverman, M. S.; Chen, Y.-J.; Chen, Z.; Dewell, S. B.; de Winter, A.; Drake, J.; Du, L.; Fierro, J. M.; Forte, R.; Gomes, X. V.; Goodwin, B. C.; He, W.; Helgesen, S.; Ho, C. H.; Hutchinson, S.; Irzyk, G. P.; Jando, S. C.; Alenquer, M. L. I.; Jarvie, T. P.; Jirage, K. B.; Kim, J.-B.; Knight, J. R.; Lanza, J. R.; Leamon, J. H.; Lee, W. L.; Lefkowitz, S. M.; Lei, M.; Li, J.; Lohman, K. L.; Lu, H.; Makhijani, V. B.; McDade, K. E.; Myers, E. W.; Nickerson, E; Nobile, J. R.; Plant, R.; Puc, B. P.; Reifler, M.; Ronan, M. T.; Roth, G. T.; Sarkis, G. J.; Simons, J. F.; Srinivasan, M.; Tartaro, K. R.; Tomasz, A.; Vogt, K. A.; Volkmer, G. A.; Wang, S. H.; Wang, Y.; Weiner, M. P.; Willoughby, D. A.; Yu, P.; Begley, R. F.; Rothberg, J. M. Nature 2005, 437, 376−380. (14) Novak, R.; Zeng, Y.; Shuga, J.; Venugopalan, G.; Fletcher, D. A.; Smith, M. T.; Mathies, R. A. Angew. Chem., Int. Ed. 2011, 50, 390−395. (15) Leng, X. F.; Zhang, W. H.; Wang, C. M.; Cui, L. A.; Yang, C. J. Lab Chip 2010, 10, 2841−2843. (16) Link, D. R.; Grasland-Mongrain, E.; Duri, A.; Sarrazin, F.; Cheng, Z. D.; Cristobal, G.; Marquez, M.; Weitz, D. A. Angew. Chem., Int. Ed. 2006, 45, 2556−2560. (17) Tan, Y. C.; Cristini, V.; Lee, A. P. Sens. Actuators, B 2006, 114, 350−356. (18) Huebner, A.; Sharma, S.; Srisa-Art, M.; Hollfelder, F.; Edel, J. B.; Demello, A. J. Lab Chip 2008, 8, 1244−1254. (19) Schaerli, Y.; Wootton, R. C.; Robinson, T.; Stein, V.; Dunsby, C.; Neil, M. A. A.; French, P. M. W.; deMello, A. J.; Abell, C.; Hollfelder, F. Anal. Chem. 2009, 81, 302−306. (20) Shen, F.; Du, W. B.; Kreutz, J. E.; Fok, A.; Ismagilov, R. F. Lab Chip 2010, 10, 2666−2672. (21) Brouzes, E.; Medkova, M.; Savenelli, N.; Marran, D.; Twardowski, M.; Hutchison, J. B.; Rothberg, J. M.; Link, D. R.; Perrimon, N.; Samuels, M. L. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14195−14200. (22) Coupland, P. Lab Chip 2010, 10, 544−547. (23) Zhang, W. Y.; Zhang, W. H.; Liu, Z. Y.; Li, C.; Zhu, Z.; Yang, C. J. Anal. Chem. 2012, 84, 350−355. (24) Bontoux, N.; Dauphinot, L.; Vitalis, T.; Studer, V.; Chen, Y.; Rossier, J.; Potier, M. C. Lab Chip 2008, 8, 443−450.

running massively parallel single-cell RT-PCR in agarose droplets, which enables the comparison of gene expression level at the single-cell level over a large population of cells.

■

CONCLUSIONS In conclusion, we have demonstrated, for the first time, singlecopy RT-PCR using microfluidic generated agarose emulsion droplets. We have further proven the use of microfluidics to exercise precise control over the RT-PCR reaction volume and concentration, with a uniform distribution of reagents within the droplets. Reverse transcription and PCR in the one-step protocol used is well controlled with this method as reaction components are mixed quickly at the point of droplet formation. A simple and efficient method to graft primer to the polymer matrix using 5′-acrydite primer has been developed to ensure highly efficient trapping of RT-PCR products in agarose. Performing RT-PCR in a homogeneous agarose solution allows for higher RT-PCR efficiency because steric hindrance effect and charge repulsion problems are avoided as compared to running amplification on solid beads. Rapid sol− gel phase transition of agarose efficiently traps RT-PCR product within each agarose droplet to maintain monoclonality for downstream processing. The coupling of agarose droplet amplification to flow cytometry enables rapid and sensitive statistical analysis of large droplet populations. Finally, the ability to perform massively parallel RT-PCR in agarose droplets with input targets down to the single copy limit is of great significance to many fundamental research and clinical applications. We believe this technique will prove particularly useful for amplification of RNA libraries for RNA sequencing, and for performing high-throughput single-cell gene expression studies, rare RNA mutation analysis, and digital RNA detections.42

■

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.

■

AUTHOR INFORMATION

Corresponding Author

*Mailing address: Key Laboratory of Analytical Science, Key Laboratory for Chemical Biology of Fujian Province, State Key Laboratory of Physical Chemistry of Solid Surfaces, and Department of Chemical Biology, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, China. Phone: (+)00865922187601. Fax: (+)00865922189959, E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The first two authors contributed equally to this work. We gratefully acknowledge the financial support from the NSFC (21075104), National Basic Research Program of China (2010CB732402), the Natural Science Foundation of Fujian Province for Distinguished Young Scholars (2010 J06004) and the Department of Business Innovation and Skills, U.K.

■

REFERENCES

(1) Kumaresan, P.; Yang, C. J.; Cronier, S. A.; Blazej, R. G.; Mathies, R. A. Anal. Chem. 2008, 80, 3522−3529. 3605

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606

Analytical Chemistry

Article

(25) Marcus, J. S.; Anderson, W. F.; Quake, S. R. Anal. Chem. 2006, 78, 3084−3089. (26) Marcus, J. S.; Anderson, W. F.; Quake, S. R. Anal. Chem. 2006, 78, 956−958. (27) Mary, P.; Dauphinot, L.; Bois, N.; Potier, M. C.; Studer, V.; Tabeling, P. Biomicrofluidics 2011, 5, 024109. (28) Toriello, N. M.; Douglas, E. S.; Thaitrong, N.; Hsiao, S. C.; Francis, M. B.; Bertozzi, C. R.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 20173−20178. (29) Zhong, J. F.; Chen, Y.; Marcus, J. S.; Scherer, A.; Quake, S. R.; Taylor, C. R.; Weiner, L. P. Lab Chip 2008, 8, 68−74. (30) Shen, F.; Sun, B.; Kreutz, J. E.; Davydova, E. K.; Du, W. B.; Reddy, P. L.; Joseph, L. J.; Ismagilov, R. F. J. Am. Chem. Soc. 2011, 133, 17705−17712. (31) Bukrinsky, M. I.; Sharova, N.; Mcdonald, T. L.; Pushkarskaya, T.; Tarpley, W. G.; Stevenson, M. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6125−6129. (32) Kuriyama, R.; Borisy, G. G. J. Cell Biol. 1981, 91, 822−826. (33) Zhu, Z.; Wu, C. C.; Liu, H. P.; Zou, Y.; Zhang, X. L.; Kang, H. Z.; Yang, C. J.; Tan, W. H. Angew. Chem., Int. Ed. 2010, 49, 1052− 1056. (34) Hosokawa, K.; Fujii, T.; Endo, I. Anal. Chem. 1999, 71, 4781− 4785. (35) Zhang, C.; Da, X. Chem. Rev. 2010, 110, 4910−4947. (36) Al-Hajj, M.; Wicha, M. S.; Benito-Hernandez, A.; Morrison, S. J.; Clarke, M. F. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3983−3988. (37) Li, C. W.; Heidt, D. G.; Dalerba, P.; Burant, C. F.; Zhang, L. J.; Adsay, V.; Wicha, M.; Clarke, M. F.; Simeone, D. M. Cancer Res. 2007, 67, 1030−1037. (38) Yamashita, T.; Ji, J. F.; Budhu, A.; Forgues, M.; Yang, W.; Wang, H. Y.; Jia, H. L.; Ye, Q. H.; Qin, L. X.; Wauthier, E.; Reid, L. M.; Minato, H.; Honda, M.; Kaneko, S.; Tang, Z. Y.; Wang, X. W. Gastroenterology 2009, 136, 1012−1024. (39) Dalerba, P.; Dylla, S. J.; Park, I. K.; Liu, R.; Wang, X. H.; Cho, R. W.; Hoey, T.; Gurney, A.; Huang, E. H.; Simeone, D. M.; Shelton, A. A.; Parmiani, G.; Castelli, C.; Clarke, M. F. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10158−10163. (40) Shigdar, S.; Lin, J.; Yu, Y.; Pastuovic, M.; Wei, M.; Duan, W. Cancer Sci. 2011, 102, 991−998. (41) Prang, N.; Preithner, S.; Brischwein, K.; Goster, P.; Woppel, A.; Muller, J.; Steiger, C.; Peters, M.; Baeuerle, P. A.; da Silva, A. J. Br. J. Cancer 2005, 92, 342−349. (42) Zhu, Z.; Jenkins, G.; Zhang, W.; Zhang, M.; Guan, Z.; Yang, C. J. Anal. Bioanal. Chem. 2012, DOI: 10.1007/s00216-012-5914-x.

■

NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on April 4, 2012. Reference 13 was corrected, and the updated version was reposted on April 6, 2012.

3606

dx.doi.org/10.1021/ac2033084 | Anal. Chem. 2012, 84, 3599−3606