Paper ID: 403066

µSynth: A Versatile Microfluidic Device for Automating Synthetic Biology

The synthetic biology process of specification, design, build, and test follows an iterative process that often requires extensive manual intervention. This process is used to engineer new microbes that contain the necessary genetic circuit and metabolic pathways to produce the required outputs for a wide range of applications such as bio-based chemicals and biofuels.1 While current available tools are useful in improving the synthetic biology process, further improvements in physical automation would help lower the barrier of entry into this field. For example, to build and test, there are currently automation tools to aid these processes2-4 but are still relatively underserved in terms of physical automation
technologies to build and to test DNA assembly.
Here, in response to this challenge, we introduce a new automated and versatile microfluidic
device that can assemble DNA plasmids using three assembly methods (Golden Gate, Gibson, and TAR cloning) with on-chip transformation. Our microfluidic device takes advantage of two droplet microfluidic platforms5: droplet-in-flow and digital microfluidics to integrate the various molecular biology steps. Figure 1 shows the microfluidic device comprising of a digital microfluidic chip for dispensing droplets and mixing droplets on-demand. The second part of the device consists of a channel region with integrated microvalves that will be used to incubate/store droplets and to conduct on-chip transformation. Figure 2 shows the device operation of the µSynth device (from assembly-to- transformation). To demonstrate the utility of our device, we used our device to assemble two sets of 16
plasmids: one set of plasmids for Golden Gate and Gibson assembly which contains a p15A origin of replication gene and kanamycin selection marker for bacteria and one set of plasmids for TAR cloning which contains both a 2-micron origin of replication gene with a tryptophan selection marker for yeast and a F1 origin of replication gene with an ampicillin selection marker for bacteria. Both sets of plasmids have the same DNA inserts: four promoter (Prom) variants (1, 2, 9, 11) and four bicistronic design (BCD) variants (1, 2, 20, 21) coupled with a gfp gene. Figure 3 shows the purification of the parts on a 1% agarose gel. Each assembled DNA plasmid is transformed into E.coli or S.cerevisiae using on-chip integrated electroporation. We also further evaluated our plasmids by Sanger sequencing. Figure 4 shows the results of our sequencing data for Golden Gate assembly (Gibson and TAR cloning not shown) using our microfluidic method. Excluding the beginning and ends of Sanger sequencing reactions, we obtained a high percentage of perfect sequence clones (95%) for the region spanning the BCD, insulator, promoter, and vector backbone.
We created the first automated microfluidic platform that integrates three DNA assembly methods with on-chip transformation and with minimal reagent use. As a proof-of-principle we showed a platform that can assemble two sets of 16 plasmid combinations, which we hope that we (or others) can expand on this microfluidic platform to assemble more combinations with integrated culture and screening.

1. P. P. Peralta-Yahya and J. D. Keasling, Biotechnology journal, 2010, 5, 147-162.

2. G. Linshiz, N. Stawski, G. Goyal, C. Bi, S. Poust, M. Sharma, V. Mutalik, J. D. Keasling and N. J. Hillson, ACS

synthetic biology, 2014, 3, 515-524.

3. C. N. Takahashi, A. W. Miller, F. Ekness, M. J. Dunham and E. Klavins, ACS synthetic biology, 2015, 4, 32-38.

4. E. Appleton, J. Tao, T. Haddock and D. Densmore, Nature methods, 2014, 11, 657-662.

5. S. C. C. Shih, P. C. Gach, J. Sustarich, B. A. Simmons, P. D. Adams, S. Singh and A. K. Singh, Lab on a chip,

2015, 15, 225-236.

Figure 1. Microfluidic device (µ Synth) for synthetic biology. (a) Schematic of the device, comprising of a bottom plate with patterned electrodes (shown in yellow) and a channel (shown in orange) to incubate droplets and to electroporate cells with the assembled plasmid. Top- plate for DMF and top PDMS layer for the channel is not shown for clarity. (b) Side view of the device showing the digital-to-droplet interface, not to scale

Figure 3. Gel electrophoresis (1%) images of the DNA fragments used for assembly. For (a) Golden Gate and (b) Gibson assembly, DNA parts were BsaI digested and purified. BCD-gfp (1, 2, 20, 21) show a band ~ 800 bp, promoter (1, 2, 9, 11) show a band ~ 100 bp, and the vector backbone showed a band ~ 2100 bp. For (c) yeast assembly, parts were DpnI digested at

37 °C for 1 hour and gel purified. BCD-gfp and
promoters show similar bands with a vector backbone having a band ~5600 bp. Lane abbreviations: L, 1 kb DNA ladder (Fisher Scientific) for Golden Gate; L, 1 kb plus DNA ladder (Fisher Scientific) for Gibson and yeast assembly.

Figure 2. DNA assembly and electroporation. Frames from a movie (left) with a corresponding schematic showing a simplified DNA assembly mechanism (right). First, (a) droplets containing DNA fragments (i.e. vector backbone, promoter, BCD-gfp) are dispensed and mixed and actuated to the channel. DNA ligase/polymerase is also added to the mixture for Golden Gate and Gibson assembly (not shown). (b) The droplets are incubated in the channel for X min at Y (X = 10 â?? Golden Gate and yeast, 15 â?? Gibson; Y = 25 °C â?? Golden Gate and yeast, 50 °C - Gibson). After incubation, (c) the assembled plasmid is mixed with cells shown as the red outline) and are electroporated (shown in d) by sending DC pulses to one microelectrode while grounding the other microelectrode. Scale bars, (a) 1 mm, (b, c) 1.5 mm, (d) 1 mm.

Figure 4. Sequencing results for Golden Gate DNA assembly. A library of 16 DNA plasmids were assembled using µSynth. Five colonies (that displayed green) were picked from LB agar plates and sequenced. The green portions of the sequencing arcs show regions where the sequencing result matches the expected sequence and the red portions of the sequencing arcs show regions where the sequencing result does not matches the expected sequence. The ligation region spanning the BCD, promoter (Prom) and vector backbone regions showed high sequence matching (~
95%).