A microfluidic chip capable of generating and trapping emulsion droplets for digital loop-mediated isothermal amplification analysis

Yu-Dong Ma a, Kang Luo a, Wen-Hsin Chang a and Gwo-Bin Lee *abc
aDepartment of Power Mechanical Engineering, National Tsing Hua University, Hsinchu, 30013 Taiwan. E-mail: gwobin@pme.nthu.edu.tw; Tel: +886 3 5715131 Ext. 33765
bInstitute of Biomedical Engineering, National Tsing Hua University, Hsinchu, 30013 Taiwan
cInstitute of Nano-Engineering and Microsystems, National Tsing Hua University, Hsinchu, 30013 Taiwan

Received 18th September 2017 , Accepted 23rd November 2017

First published on 24th November 2017

Loop-mediated isothermal amplification (LAMP) is a nucleic acid amplification technique that rapidly amplifies specific DNA molecules at high yield. In this study, a microfluidic droplet array chip was designed to execute the digital LAMP process. The novel device was capable of 1) creating emulsion droplets, 2) sorting them into a 30 × 8 droplet array, and 3) executing LAMP across the 240 trapped and separated droplets (with a volume of 0.22 nL) after only 40 min of reaction at 56 °C. Nucleic acids were accurately quantified across a dynamic range of 50 to 2.5 × 103 DNA copies per μL, and the limit of detection was a single DNA molecule. This is the first time that an arrayed emulsion droplet microfluidic device has been used for digital LAMP analysis. When compared to microwell digital nucleic acid amplification assays, this droplet array-based digital LAMP assay eliminates the constraint on the size of the digitized target, which was determined by the dimension of the microwells for its counterparts. Moreover, the capacity for hydrodynamic droplet trapping allows the chip to operate in a one-droplet-to-one-trap manner. This microfluidic chip may therefore become a promising device for digital LAMP-based diagnostics in the near future.


Current nucleic acid technologies (NAT) permit researchers to detect and quantify DNA or RNA molecules in order to, for instance, diagnose cancers or verify the presence of 1) genetic abnormalities or 2) microbes.2–5 However, the quantity of the target nucleic acid in a sample is typically too little to be detected directly. Therefore, the first step in most nucleic acid detection protocols is the amplification of the target nucleic acid molecule to a detectable level. Several nucleic acid amplification techniques currently exist.6 Amongst them, polymerase chain reaction (PCR)7 is certainly the most commonly used for DNA or RNA (via a complementary DNA intermediate) amplification. In this process, nucleic acid molecules are amplified exponentially via reaction with primers at 2–3 different temperatures; as such, precise temperature control is required for such “thermocycling.” To eliminate the need for thermocycling, isothermal amplification methods, such as loop-mediated isothermal amplification (LAMP),8 were developed; not only do these methods reduce the complexity of the temperature control module, but they also consume less energy9 and are more amenable to microfluidics-based applications.10–13 Furthermore, when compared to traditional PCR, LAMP requires more primers (4 to 6, instead of 2 for PCR), which makes it more specific.14,15 Furthermore, only one hour or even less is needed to complete the isothermal amplification process due to the high efficiency of the Bacillus stearothermophilus (Bst) DNA polymerase.16 These advantages make it a promising tool for nucleic acid detection.

There are numerous biomedical applications that are reliant on quantification of nucleic acids. For instance, clinicians use NAT to quantify viral DNA from patient serum or biopsies in order to adjust drug dosages. This is routinely achieved with real-time PCR.17 However, real-time PCR-based quantification relies on a reference gene or production of a standard curve, both of which are labor-intensive and increase reagent demands. More importantly, real-time PCR cannot distinguish small target molecule copy number differences between samples.

In an effort to overcome the disadvantages associated with traditional and real-time PCR, digital DNA amplification techniques, such as digital PCR, have been developed.18–20 Compared to the former approaches, digital DNA amplification is based on the partitioning of a sample into many individual compartments (such as small droplets), in which the target molecule is either present (“positive”) or not (“negative”).21,22 After amplification, the ratio of positive to negative compartments is used to calculate the sample's initial template concentration. It is worth noting that the accuracy and dynamic range of detection scale with the number of compartments (i.e., the number of droplets).23

Since large-scale systems consume large quantities of (typically costly) reagents, microfluidic-based devices have been widely adopted for digital DNA amplification techniques. Although such systems can produce large quantities of the requisite, small droplets, there are nevertheless some practical limitations. For instance, droplet-to-droplet coalescence may occur if they come into contact with each other24 during the heating process;25,26 this may compromise the accuracy of the quantification. This issue can be minimized by using compatible surfactants.27 However, in order to eliminate this issue completely, a spatially defined droplet array may be useful. Moreover, the stable, spatially defined droplet array technology was of great benefit in monitoring, identifying and indexing individual droplets24,28,29 for single cell analysis if a sensor was placed underneath.30 Several microfluidic devices capable of digital LAMP detection based on droplet array technology, including micro-well-based assays31,32 and SlipChip-based assays,33,34 have been developed. Each of these methods has certain advantages, though all have some limitations as well. For instance, the size of the digitized samples for microwell-based assays is dependent on interfacial tension and the viscosity of the oil phase. Therefore, the size distribution is not uniform, and such variation could affect the quantification accuracy. Alternatively, the SlipChip assays are efficient and effective in terms of droplet formation, though they are difficult to fabricate and operate.

To overcome the limitations of the aforementioned methods, a new microfluidic device capable of arraying emulsion droplets and carrying out digital LAMP was developed herein. It was capable of generating uniformly sized droplets (variation <3%), and the droplets could be hydrodynamically trapped in an arrayed structure (“droplet array”) with a well-designed order such that they could be immobilized at a fixed location for the subsequent detection over a period of time. As a proof of concept, the LAMP assay for the detection of vancomycin-resistant Enterococcus (VRE) bacteria was successfully implemented in water-in-oil droplets, and it was more sensitive and specific than conventional PCR.15 This is the first time that an integrated microfluidic chip has been demonstrated for this application, and this device may serve as a promising tool for digital DNA amplification techniques. Furthermore, it could be extended to applications that may involve DNA segments that may be prepared in order (such as DNA sequencing).


Materials and reagents

Mineral oil, Span 80, calcein, and manganese chloride were purchased from Sigma-Aldrich (USA). Bst DNA polymerase and the associated 10× ThermoPol® buffer were purchased from New England Biolabs (USA), and the deoxyribonucleotide triphosphate (dNTP) mix was purchased from Protech (Taiwan). The DNA template used in this study was isolated from VRE (clinical strain 102) and quantified with a NanoDrop ultraviolet (UV) spectrophotometer (DU530 UV/vis, Beckman, USA). The calculation of the copy number of DNA from concentration was based on a previous work.35 The primers for the LAMP reaction were designed for the vancomycin-resistant gene (vanA) with Primer Explorer V4 software (Eiken Chemicals Corporation, Tokyo, Japan). The primer sequences are as follows:


For on-chip and bench-top experiments, 30 μL of the reaction volume consisted of 10.5 μL of deionized-distilled water (ddH2O), 4 μL of betaine (0.8 M final concentration; Sigma-Aldrich, USA), 3 μL of 10× ThermoPol® buffer, 1 μL of 10 mM dNTP mix (2.5 mM of each of the four dNTPs), 1 μL of 10 μM B3 primers, 1 μL of 10 μM F3 primers, 3 μL of 20 μM BIP primers, 3 μL of 20 μM FIP primers (all primers were from Sigma-Aldrich, USA), 1 μL of 8000 U mL−1Bst DNA polymerase, 1 μL of the DNA template, and 1.5 μL of fluorescent dye (0.5 μM calcein and 10 μM manganese chloride premixed solution).

Chip design and fabrication

To fabricate the microfluidic chip, master molds were first formed via an SU-8 (MicroChem, USA) standard photolithography process.36 Briefly, after the cleaning process of a silicon wafer, a 50 μm-thick layer of SU-8 3035 was coated onto the silicon wafer with a spin-coater (M&R Nano Technology, Taiwan), and the composite unit was subsequently patterned with an UV exposure dosage of 250 mJ cm−2 through a photolithography mask. Then, a two-step, post-exposure bake process (PEB; 65 °C for 1 min and 95 °C for 5 min) was performed, and the fabrication of the master mold was accomplished after washing out the residual photoresist with acetone and isopropanol. Afterwards, a soft lithography process was used to replicate the inverse microfluidic structures on the master mold.37 Briefly, a polydimethylsiloxane (PDMS, Sylgard 184A/B, Dow Corning, USA) elastomer precursor and a curing agent were mixed together at a weight ratio of 10[thin space (1/6-em)]:[thin space (1/6-em)]1 and poured into the SU-8-patterned mold after degassing with a vacuum desiccator. Then, the molds were heated at 80 °C for 15 min (incompletely cured PDMS) and then poured another layer of oil-containing PDMS (i.e., PDMS that had been previously mixed with 5% w/w mineral oil and degassed). The two PDMS layers were heated overnight at 80 °C to cure and bond to each other, and the inverse microstructures were formed by a mechanical demolding process. Afterwards, two inlet ports and one outlet port were created with a biopsy punch (Harris Uni-Core™, Ted Pella, USA). Finally, a glass cover slip (0.2 mm, Electron Microscopy Sciences, USA) was bonded to the top of the microfluidic structure layer via oxygen plasma treatment, and the base of the microfluidic structure layer was bonded to a flat PDMS-coated glass substrate (0.7 mm, G-Tech Optoelectronics Corp, Taiwan) (Fig. 3b), such that the microfluidic chip capable of generating and trapping emulsion droplets was completely assembled.

Experimental setup

The experimental process of the arrayed emulsion droplet-based digital LAMP analysis is schematically illustrated in Fig. 1. First, an emulsion droplet formation device was integrated with a hydrodynamic trapping array. DNA target molecules, LAMP reagents, and mineral oil pre-mixed with a surfactant (2% (w/v) Span 80) were prepared freshly first before being injected into the microfluidic chip with a pressure-driven flow control system (MFCS™-EZ, Fluigent, France). The volume of the LAMP mixture for the digital LAMP assay was 30 μL as described above and the concentration of the VRE DNA template was tested with a series of dilutions (from 1/20 to 1/1000) of the stock DNA concentration (e.g., DNA stock concentration: 1.5 × 105 copies per μL; final concentration in 30 μL LAMP mixture: 5 × 104 copies per μL). Afterwards, DNA target molecules and LAMP reagents were then digitized on-chip into water-in-oil droplets using a droplet formation process known as flow-focusing.38 Then, emulsion droplets were hydrodynamically sorted into a pre-designed, arrayed trapping structure (i.e., the droplet array). After heating the chip with a Peltier device at 56 °C for 40 min, DNA targets were amplified by the LAMP amplification process. The fluorescence emerging from positive droplets (described in more detail below) was detected directly on the device using a fluorescence microscope. Note that a glass coverslip and oil-containing PDMS layer were used to reduce reagent evaporation during the LAMP process.39
image file: c7lc01004d-f1.tif
Fig. 1 Schematic illustration of the experimental process for emulsion droplet array-based digital LAMP analysis. The LAMP reaction components, template DNA, and oil were first injected into the microfluidic chip to form emulsion droplets. Then, the droplets were immobilized within a droplet array. After performing LAMP for 40 min, a fluorescence microscope was used to record the results.

The experimental setup comprised an optical microscope (BX43, Olympus, Japan), a cooled charge-coupled device (CCD) camera (Evolution™ VF Color Cooled, Olympus, Canada), and a microfluidic flow control system (MFCS™-EZ, Fluigent, France; Fig. 2). Reagents (carrier oil and samples) were first delivered to the microfluidic chip using the microfluidic flow control system. The formation of the microdroplets was observed under the microscope and recorded with the CCD camera. Note that the Peltier heater was placed under the LAMP reaction region of the chip to maintain the unit at a reaction temperature of 56 °C. After the reaction, the recorded images were analyzed with ImageJ software (National Institutes of Health, USA). Briefly, the images were adjusted using the “white” feature display mode and converted into “hue, saturation, and brightness (HSB)” color space. The “IsoData” threshold method was then used; this automatically divides the image into “object” and “background.”40 Finally, a “find edge” algorithm was used to highlight the full individual droplets such that the fluorescent droplets could be clearly identified after employing a noise reduction algorithm.41 All experiments were repeated at least twice to ensure the reproducibility (CPD = 0.05, n = 3; CPD = 0.01, n = 4; others, n = 2).

image file: c7lc01004d-f2.tif
Fig. 2 Schematic illustration of the experimental setup for the emulsion droplet array-based digital LAMP analysis.

Digital LAMP optimization

Betaine was one of the components in a typical LAMP reaction,8 which was used to decrease the melting temperature of double-stranded DNA during the amplification process.42 Previous studies revealed that betaine may improve the sensitivity,43 specificity43 and amplification efficiency in a LAMP assay.44,45 However, some studies have reported adverse effects of betaine.46,47 For instance, it was reported that betaine may inhibit the nucleic acid amplification and therefore it was not necessary to be added in the LAMP reaction. The inconsistency may be due to the amplification of different primers and DNA sequences.48 Therefore, two concentrations of betaine (0 M and 0.8 M) were used to test the effect of betaine in this study with the real-time LAMP assay prior to the digital LAMP analysis (described in more detail below). The reaction volume (30 μL) was the same as that described above, while the volume of ddH2O was altered accordingly when a reaction mixture without betaine was used.

Calcein was used as the fluorescent dye for the detection of LAMP products, as this molecule's fluorescence is sensitive to changes in a solution's metallic ion concentrations.49 Divalent metallic ions such as Mg2+ enhance the fluorescence signal of calcein while transition metal ions such as Mn2+ quench it.50 In the LAMP reaction, pyrophosphate ions are produced from the polymerization of nucleotides. Initially, Mn2+ reacts with calcein, and minimal fluorescence is released. When pyrophosphate ions are produced, Mn2+ reacts instead with the pyrophosphate ions and forms insoluble salts, resulting in the emission of fluorescence signals. Since Mn2+ is no longer bound to calcein, the Mg2+ ions in the reaction mixture are capable of binding it, thereby enhancing the fluorescence signal. The concentration of Mn2+ in the reaction mixture is therefore a critical factor for this fluorescence detection process. Prior to the digital LAMP analysis, real-time LAMP was performed to optimize the reaction conditions and reagent volumes and concentrations on a real-time PCR system (StepOne™, Applied Biosystems, USA). The same reaction mixtures (30 μL) as those described above were prepared in 200 μL microtubes, which contained concentrations of manganese chloride ranging from 5 to 100 mM. The reaction program was set to 20 s at 56 °C for 180 cycles; the fluorescence signal was recorded after each cycle and analyzed using StepOne Software v2.3 (Applied Biosystems, USA). The assays were repeated three times to ensure the reproducibility. Finally, the real-time LAMP products were stained with ethidium bromide, and analyzed by an agarose gel (2%) electrophoresis assay with an UV imaging system (UVP BioDoc-It bioimage systems, Canada).

Results and discussion

The 4 cm × 2 cm (length × width) digital LAMP chip for emulsion droplet formation and arrayed trapping (Fig. 3(a)) consisted of five layers (Fig. 3(b)), as described in the Experimental section. Note that in digital LAMP, the droplets should be water-in-oil emulsions such that samples and reagents are both encapsulated in oil. In other words, the droplet formation channel should be hydrophobic. Therefore, a thin PDMS blank layer was spin-coated on the glass substrate to maintain the hydrophobicity of the channel.
image file: c7lc01004d-f3.tif
Fig. 3 (a) An image of the microfluidic chip. (b) An exploded view of the chip. The chip consisted of five layers, including a glass substrate, a PDMS blank layer, a structural layer featuring the microfluidic components, a 5% mineral oil-enriched PDMS cover layer, and a glass coverslip equipped with a PDMS inlet and a PDMS outlet. Note that the oil-enriched PDMS cover layer and coverslip were used to reduce evaporation during the LAMP reaction. (c) Detailed design of the microfluidic chip, including the (I) inlet for the continuous-phase liquid (oil), (II) inlet for the dispersed-phase liquid (samples and reagents), (III) flow-focusing area for droplet formation, (IV) fluidic resistor, and (V) arrayed droplet trapping structures.

The chip featured several key components (Fig. 3(c)), including an inlet for the continuous phase (oil), an inlet for aqueous phases (water), a flow-focusing area for droplet formation, a fluidic resistor for stabilizing the droplet formation, and an arrayed trapping structure for droplet trapping. The fluidic resistor was designed to reduce fluctuations arising from the mechanical instability of the pressure pump and the elasticity of the PDMS devices. The hydrodynamic arrayed trapping structure, which consisted of a serpentine-shaped main channel and hydrodynamic trapping sites, was first presented by Tan and Takeuchi1 for arraying microbeads. The ratio of the volumetric flow rate through the straight channel to the loop channel was designed to be 2.61[thin space (1/6-em)]:[thin space (1/6-em)]1, which means that about 72% of the fluid flowed through the straight channel and the remainder through the loop channel. It is worth noting that only solid microbeads were trapped in this design. To the best of our knowledge, no liquid droplets have been demonstrated in the literature.

After droplet formation, the emulsion droplets were delivered to the subsequent arrayed-trapping area in a well-ordered fashion. As a result, emulsion droplets were carried by the main flow into the traps in order and then immobilized in the traps. Note that a total of 240 trapping structures were designed in this work. Two design types were initially tested to trap the droplets (Fig. 4(a)). The original trapping structures did not have a narrow neck at the entrance. Even though the droplets could be transferred efficiently to the array with this design, backflow was an issue which may push out the droplets from the trapping structures during the LAMP process. The second design was therefore equipped with a narrow neck at the entrance, and the process of hydrodynamic, arrayed droplet trapping with this design is shown in Fig. 4(b). This design was capable of trapping droplets in arrays without inadvertently detaching them from the device during the LAMP process. Since the width of the narrow neck was less than the diameter of the droplets, the droplets were deformed as they were driven by the mainstream through the narrow neck and into the trapping structure. During the LAMP process, the backflow was not strong enough to deform the droplets to where they exited their chambers. Therefore, this design was found to be preferable. The emulsion droplets were used as reaction compartments for the digital LAMP assay. As a result, the diameter of the droplet is of great importance. The design objective of the emulsion droplet formation device was to generate emulation droplets of uniform sizes encapsulating DNA targets and LAMP reagents. A flow-focusing device was used to generate emulsion droplets by introducing the dispersed-phase liquid (samples and LAMP reagents) into the central channel and continuous-phase liquid (oil) into the neighboring sheath channels. With this approach, the emulsion droplets could be formed by the applied shearing force, and the size of the formed droplets could be fine-tuned based on the flow rate ratio between the central flow and the sheath flows.

image file: c7lc01004d-f4.tif
Fig. 4 Two droplet trapping designs. (a) The design presented by Tan et al.1 (b) The design utilized herein, which features a narrow neck at the entrance of the trapping structure. (c) After the traps were filled, the subsequent droplet travelled.

In order to characterize the size distribution of the emulation droplets generated by the microfluidic chip, the sizes of about 300 droplets were measured with ImageJ. The average diameter was measured to be 74.63 μm with a coefficient of variation of 2.67% (Fig. 5); such uniformity in droplet size is superior to that of a previous study.32 Note that the variation in size of the droplets can bias the Poisson-based calculations of template quantification.

image file: c7lc01004d-f5.tif
Fig. 5 Size distribution of generated emulsion droplets. The average diameter was measured to be 74.63 μm with a standard variation of 2.67%.

To optimize the concentration of betaine and Mn2+ in the reaction mixture, real-time LAMP experiments were performed (Fig. 6). As a result, the LAMP products were detected earlier with the addition of 0.8 M betaine while the fluorescence signal was elevated by about 8% after the LAMP reaction when compared to the mixture without betaine (Fig. 6(a)). Therefore, on the basis of the above analysis, the LAMP reaction mixture was optimized with 0.8 M betaine. Fig. 6(b) shows real-time LAMP results with different concentrations of Mn2+. The reaction with a 5 mM Mn2+ concentration exhibited the highest fluorescence signal after the LAMP reaction. However, the initial background fluorescence (noise) was relatively strong. Upon increasing the concentration of Mn2+ to 10 mM, a better signal-to-noise ratio was achieved. It is worth noting that it takes longer to detect the rising fluorescence signals as the concentration of Mn2+ increases. When the concentration of Mn2+ exceeds 20 mM, the fluorescence signal was hardly detected, even when the reaction time was extended to more than one hour. It should be noted that Mn2+ did not interfere with the DNA amplification process and the ladder-like LAMP products were still clearly observed on gel electropherograms (Fig. 6(c)). Therefore, the optimal concentration of Mn2+ was determined to be 10 mM.

image file: c7lc01004d-f6.tif
Fig. 6 Optimization of the LAMP reaction (carried out in a real-time PCR thermocycler). (a) Effect of betaine in the LAMP reaction. With betaine, the fluorescence signal was enhanced earlier than the one without betaine. (b) Real-time LAMP reaction with different concentrations of Mn2+. As the concentration of Mn2+ increased, it took longer to observe calcein fluorescence signals above the background level. A reaction containing 10 mM Mn2+ was found to exhibit the highest signal-to-noise ratio and was used for the experiments as detailed in Fig. 7 and 8. RFU = relative fluorescence units. (c) The LAMP products could be clearly observed by gel electrophoresis; L: 100 bp DNA ladder; N: negative control.

Fig. 7 shows the fluorescence signals from the droplet LAMP experiments. The experiment was performed by utilizing an array-free chip to facilitate this testing process (Fig. S1). The average number of copies of the DNA template in each droplet was 10. Fig. 7(a) and (b) represent the fluorescence images before and after the LAMP reaction, respectively, and no fluorescence was emitted prior to increasing the temperature to 56 °C. It can be clearly seen that the droplets which encapsulated the LAMP reaction mixture did not present fluorescence signals before the LAMP amplification process. After the reaction, all droplets produced strong fluorescence signals, indicating that the LAMP reaction was successful and that positive reaction signals could be detected with microscopy in volumes of only 0.22 nL. It was worth noting that a large amount of “dots” appearing in the droplets were the precipitate of magnesium pyrophosphate, which was the by-product of the LAMP reaction.51

image file: c7lc01004d-f7.tif
Fig. 7 Fluorescence signals of the droplet array-based digital LAMP experiment. (a) Before the reaction. (b) After the reaction.

Absolute quantification of nucleic acids is of great importance in clinical applications. The accuracy of digital LAMP was analyzed by serially diluting genomic VRE DNA from 1/20 to 1/1000 of the stock sample concentration, which equated to 0.01–0.5 DNA copies per droplet (CPD) (stock DNA concentration = 10 CPD). Fig. 8(a) shows representative results of digital LAMP performed on the arrayed trapping device after incubation at 56 °C for 40 min. Note that a few droplets with weak fluorescence were observed after cessation of the LAMP reaction; this was likely due to non-specific amplification (e.g., primer-dimers). Future works would attempt to design better primers and optimize the LAMP reaction components and conditions to reduce this issue.52,53 It should be noted that these primer-dimer droplets were not counted as positive signals after the ImageJ “thresholding” analysis (Fig. 8(b)); therefore, the accuracy of the quantification was not compromised due to these primer-dimers. After the thresholding analysis, the positive droplets could be clearly identified and directly counted (Fig. 8(b)). As the DNA template was further diluted, the fraction of droplets with positive calcein fluorescence decreased (Fig. 7(c)), as expected. Random, uniform, and independent separation is crucial for using the Poisson statistics to accurately quantify DNA with digital LAMP. If more than one DNA molecule is entrapped in a droplet, the correspondence between the number of positive signals and the starting DNA concentration will fit a sigmoidal curve instead of a linear one. Therefore, the four lowest DNA concentrations were plotted against their respective number of fluorescent droplets (Fig. 8(d)), and a linear relationship was revealed. The lowest observed number of fluorescent droplets was 1, which corresponds to a DNA concentration of 50 copies per μL (i.e., 50 copies randomly distributed across ∼5000 droplets). The limit of detection of our developed system is, therefore, one molecule of DNA. These results demonstrated that the developed chip was capable of performing a precise and accurate quantification of DNA in an unknown concentration sample (R2 = 0.9986).

image file: c7lc01004d-f8.tif
Fig. 8 The quantification results and analysis of emulsion droplet-based digital LAMP on an arrayed trapping device. (a) Fluorescence images of digital LAMP products with serially diluted DNAs (concentrations ranging from 1/20 to 1/1000 of the stock sample concentration [5 × 104 copies per μL]). (b) The “thresholding” images produced using ImageJ (see the main text for details.). (c) A regression curve obtained by plotting the fraction of fluorescent (positive) droplets against the dilution factor according to the Poisson distribution. (d) A positive, linear, statistically significant correlation between the DNA concentration and the number of positive fluorescence droplets. (CPD = 0.05, n = 3; CPD = 0.01, n = 4; others, n = 2). CPD = copies per droplet. Error bars represent standard deviation.


We have developed an integrated microfluidic device capable of carrying out all stages of the digital LAMP process. The microfluidic chip featured emulsion droplet formation and hydrodynamic trapping modules that were used in tandem to create a droplet array; digital LAMP was then carried out in each droplet. This is the first time that emulsion droplets have been used for an array-based digital LAMP assay, and the droplets were near-uniform in size (mean = 74.63 μm). Only one droplet could enter a single trap structure (0.22 nL) in a well ordered manner, meaning that this system could be used for other emulsion-based applications, such as emulsion PCR preceding certain types of DNA sequencing (e.g., Roche 454) in the future. The emulsion droplet LAMP assay was further tested by quantifying the concentration of the vanA gene from VRE bacteria. Not only could it accurately quantify the concentration of this gene, but it could also detect as little as one copy of DNA in 40 min. Therefore, the novel microfluidic system developed herein may serve as a promising tool for digital LAMP-based diagnostics.

Conflicts of interest

There are no conflicts to declare.


The authors would like to acknowledge financial support from Taiwan's Ministry of Science and Technology (MOST 105-2119-M-007-009 and MOST 105-2221-E-007-007 to GBL). Partial financial support from the “Towards a World-Class University” Project (104N2751E1) is also greatly appreciated. The authors also thank Dr. Jiunn-Jong Wu for providing the VRE bacteria.

Notes and references

  1. W.-H. Tan and S. Takeuchi, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 1146–1151 CrossRef CAS PubMed .
  2. Y.-K. Tong and Y. D. Lo, Clin. Chim. Acta, 2006, 363, 187–196 CrossRef CAS PubMed .
  3. D. Sidransky, Science, 1997, 278, 1054–1058 CrossRef CAS PubMed .
  4. F. Barany, Proc. Natl. Acad. Sci. U. S. A., 1991, 88, 189–193 CrossRef CAS .
  5. M. Grompe, Nat. Genet., 1993, 5, 111–117 CrossRef CAS PubMed .
  6. R. F. Massung, Emerging Infect. Dis., 2005, 11, 357 CrossRef .
  7. K. B. Mullis and F. A. Faloona, Methods Enzymol., 1987, 155, 335–350 CAS .
  8. T. Notomi, H. Okayama, H. Masubuchi, T. Yonekawa, K. Watanabe, N. Amino and T. Hase, Nucleic Acids Res., 2000, 28, e63 CrossRef CAS PubMed .
  9. Y. Mori and T. Notomi, J. Infect. Chemother., 2009, 15, 62–69 CrossRef CAS PubMed .
  10. X. Fang, Y. Liu, J. Kong and X. Jiang, Anal. Chem., 2010, 82, 3002–3006 CrossRef CAS PubMed .
  11. X. Fang, H. Chen, S. Yu, X. Jiang and J. Kong, Anal. Chem., 2010, 83, 690–695 CrossRef PubMed .
  12. C.-H. Wang, K.-Y. Lien, J.-J. Wu and G.-B. Lee, Lab Chip, 2011, 11, 1521–1531 RSC .
  13. J. Luo, X. Fang, D. Ye, H. Li, H. Chen, S. Zhang and J. Kong, Biosens. Bioelectron., 2014, 60, 84–91 CrossRef CAS PubMed .
  14. K. Nagamine, T. Hase and T. Notomi, Mol. Cell. Probes, 2002, 16, 223–229 CrossRef CAS PubMed .
  15. M. Parida, S. Sannarangaiah, P. K. Dash, P. Rao and K. Morita, Rev. Med. Virol., 2008, 18, 407–421 CrossRef CAS PubMed .
  16. T. Notomi, H. Okayama, H. Masubuchi, T. Yonekawa, K. Watanabe, N. Amino and T. Hase, Nucleic Acids Res., 2000, 28, e63 CrossRef CAS PubMed .
  17. C. A. Heid, J. Stevens, K. J. Livak and P. M. Williams, Genome Res., 1996, 6, 986–994 CrossRef CAS PubMed .
  18. B. Vogelstein and K. W. Kinzler, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 9236–9241 CrossRef CAS .
  19. P. Sykes, S. Neoh, M. Brisco, E. Hughes, J. Condon and A. Morley, BioTechniques, 1992, 13, 444–449 CAS .
  20. M. Li, W.-D. Chen, N. Papadopoulos, S. N. Goodman, N. C. Bjerregaard, S. Laurberg, B. Levin, H. Juhl, N. Arber and H. Moinova, Nat. Biotechnol., 2009, 27, 858–863 CrossRef CAS PubMed .
  21. R. Sanders, J. F. Huggett, C. A. Bushell, S. Cowen, D. J. Scott and C. A. Foy, Anal. Chem., 2011, 83, 6474–6484 CrossRef CAS PubMed .
  22. B. J. Hindson, K. D. Ness, D. A. Masquelier, P. Belgrader, N. J. Heredia, A. J. Makarewicz, I. J. Bright, M. Y. Lucero, A. L. Hiddessen and T. C. Legler, Anal. Chem., 2011, 83, 8604–8610 CrossRef CAS PubMed .
  23. E. A. Ottesen, J. W. Hong, S. R. Quake and J. R. Leadbetter, Science, 2006, 314, 1464–1467 CrossRef CAS PubMed .
  24. R. R. Pompano, W. Liu, W. Du and R. F. Ismagilov, Annu. Rev. Anal. Chem., 2011, 4, 59–81 CrossRef CAS PubMed .
  25. B. Xu, N. T. Nguyen and T. N. Wong, Biomicrofluidics, 2012, 6, 12811–12818 CrossRef PubMed .
  26. Q. Zhong, S. Bhattacharya, S. Kotsopoulos, J. Olson, V. Taly, A. D. Griffiths, D. R. Link and J. W. Larson, Lab Chip, 2011, 11, 2167–2174 RSC .
  27. C. Holtze, A. C. Rowat, J. J. Agresti, J. B. Hutchison, F. E. Angile, C. H. J. Schmitz, S. Koster, H. Duan, K. J. Humphry, R. A. Scanga, J. S. Johnson, D. Pisignano and D. A. Weitz, Lab Chip, 2008, 8, 1632–1639 RSC .
  28. L. Labanieh, T. N. Nguyen, W. Zhao and D. K. Kang, Micromachines, 2015, 6, 1469–1482 CrossRef PubMed .
  29. A. Huebner, D. Bratton, G. Whyte, M. Yang, A. J. Demello, C. Abell and F. Hollfelder, Lab Chip, 2009, 9, 692–698 RSC .
  30. J. Q. Boedicker, M. E. Vincent and R. F. Ismagilov, Angew. Chem., Int. Ed., 2009, 48, 5908–5911 CrossRef CAS PubMed .
  31. Q. Zhu, Y. Gao, B. Yu, H. Ren, L. Qiu, S. Han, W. Jin, Q. Jin and Y. Mu, Lab Chip, 2012, 12, 4755–4763 RSC .
  32. A. Gansen, A. M. Herrick, I. K. Dimov, L. P. Lee and D. T. Chiu, Lab Chip, 2012, 12, 2247–2254 RSC .
  33. D. A. Selck, M. A. Karymov, B. Sun and R. F. Ismagilov, Anal. Chem., 2013, 85, 11129–11136 CrossRef CAS PubMed .
  34. B. Sun, F. Shen, S. E. McCalla, J. E. Kreutz, M. A. Karymov and R. F. Ismagilov, Anal. Chem., 2013, 85, 1540–1546 CrossRef CAS PubMed .
  35. Y. D. Ma, W. H. Chang, K. Luo, C. H. Wang, S. Y. Liu, W. H. Yen and G. B. Lee, Biosens. Bioelectron., 2018, 99, 547–554 CrossRef CAS PubMed .
  36. F. C. Huang, C. S. Liao and G. B. Lee, Electrophoresis, 2006, 27, 3297–3305 CrossRef CAS PubMed .
  37. M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and S. R. Quake, Science, 2000, 288, 113–116 CrossRef CAS PubMed .
  38. M. Joanicot and A. Ajdari, Science, 2005, 309, 887–888 CrossRef CAS PubMed .
  39. X. Bian, F. Jing, G. Li, X. Fan, C. Jia, H. Zhou, Q. Jin and J. Zhao, Biosens. Bioelectron., 2015, 74, 770–777 CrossRef CAS PubMed .
  40. T. W. Ridler and S. Calvard, IEEE Trans. Syst. Man. Cybern., 1978, 8, 630–632 CrossRef .
  41. L. B. Pinheiro, V. A. Coleman, C. M. Hindson, J. Herrmann, B. J. Hindson, S. Bhat and K. R. Emslie, Anal. Chem., 2012, 84, 1003–1011 CrossRef CAS PubMed .
  42. W. A. Rees, T. D. Yager, J. Korte and P. H. Vonhippel, Biochemistry, 1993, 32, 137–144 CrossRef CAS PubMed .
  43. E. Mok, E. Wee, Y. Wang and M. Trau, Sci. Rep., 2016, 6, 37837 CrossRef CAS PubMed .
  44. D. Zhou, J. Guo, L. Xu, S. Gao, Q. Lin, Q. Wu, L. Wu and Y. Que, Sci. Rep., 2014, 4, 4912 CrossRef CAS PubMed .
  45. H. B. Zhao, G. Y. Yin, G. P. Zhao, A. H. Huang, J. H. Wang, S. F. Yang, H. S. Gao and W. J. Kang, Indian J. Microbiol., 2014, 54, 80–86 CrossRef CAS PubMed .
  46. C. Ma, Y. Wang, P. Zhang and C. Shi, Anal. Biochem., 2017, 530, 1–4 CrossRef CAS PubMed .
  47. S. Y. Chen, F. Wang, J. C. Beaulieu, R. E. Stein and B. L. Ge, Appl. Environ. Microbiol., 2011, 77, 4008–4016 CrossRef CAS PubMed .
  48. W. Henke, K. Herdel, K. Jung, D. Schnorr and S. A. Loening, Nucleic Acids Res., 1997, 25, 3957–3958 CrossRef CAS PubMed .
  49. L. Saari and W. Seitz, Anal. Chem., 1984, 56, 810–813 CrossRef CAS .
  50. N. Tomita, Y. Mori, H. Kanda and T. Notomi, Nat. Protoc., 2008, 3, 877–882 CrossRef CAS PubMed .
  51. Y. Mori, K. Nagamine, N. Tomita and T. Notomi, Biochem. Biophys. Res. Commun., 2001, 289, 150–154 CrossRef CAS PubMed .
  52. D. J. Bacich, K. M. Sobek, J. L. Cummings, A. A. Atwood and D. S. O'Keefe, BMC Res. Notes, 2011, 4, 457 CrossRef CAS PubMed .
  53. Y. Kimura, M. J. de Hoon, S. Aoki, Y. Ishizu, Y. Kawai, Y. Kogo, C. O. Daub, A. Lezhava, E. Arner and Y. Hayashizaki, Nucleic Acids Res., 2011, 39, e59 CrossRef CAS PubMed .


Electronic supplementary information (ESI) available. See DOI: 10.1039/c7lc01004d
These authors contributed equally.

This journal is © The Royal Society of Chemistry 2018