Bacterial Microcolonies in Gel Beads for High-throughput Screening

[Abstract] High-throughput screening of a DNA library expressed in a bacterial population for identifying potentially rare members displaying a property of interest is a crucial step for success in many experiments such as directed evolution of proteins and synthetic circuits and deep mutational scanning to identify gainor loss-of-function mutants. Here, I describe a protocol for high-throughput screening of bacterial (E. coli) microcolonies in gel beads. Single cells are encapsulated into monodisperse water-in-oil emulsion droplets produced with a microfluidic device. The aqueous solution also contains agarose that gelates upon cooling on ice, so that solid gel beads form inside the droplets. During incubation of the emulsion, the cells grow into monoclonal microcolonies inside the beads. After isolation of the gel beads from the emulsion and their sorting by fluorescence activated cell sorting (FACS), the bacteria are recovered from the gel beads and are then ready for a further round of sorting, mutagenesis or analysis. In order to sort by FACS, this protocol requires a fluorescent readout, such as the expression of a fluorescent reporter protein. Measuring the average fluorescent signals of microcolonies reduces the influence of high phenotypic cell-to-cell variability and increases the sensitivity compared to the sorting of single cells. We applied this method to sort a pBAD promoter library at ON and OFF states (Duarte et al., 2017).

multiple sequential rounds of this protocol. With this method, we sorted cells expressing a fluorescent reporter protein, while it could also be amended to screen for other readouts. If combined with a strategy to maintain fluorescent reaction products in the bead (Fischlechner et al., 2014), it can be used to screen for enzyme or pathway activities. Another option is to co-encapsulate sensor cells that fluoresce upon the production of a compound of interest (Meyer et al., 2015). It is also possible to assay cell growth by relying on the light scatter of the microcolonies or by staining them with a fluorescent biomass indicator dye (e.g., staining nucleic acids or proteins) (Weaver et al., 1991). When encapsulating multiple cells per bead, cell-cell interactions could also be screened for. Thus, the described protocol is broadly applicable in biology.  (Devenish et al., 2013). The design file of the device used in this study (40 µm flow-focusing channel) is available from the author's website: (Figure 1).   and incubate them at 37 °C for 12-18 h in a shaking incubator. Alternatively, the protocol can also be started from glycerol stocks (stored at -80 °C), for example, 15% (v/v) glycerol stocks prepared by adding 420 μl of the overnight culture to 180 μl of 50% (v/v) glycerol.

Materials and Reagents
2. Connect the high-speed camera to an inverted microscope (via C-mount).
3. Adjust the settings of the high-speed camera. Provide enough light so that the exposure time can be adjusted to the minimum (~50 µsec) and the frame rate to the maximum. Reducing the picture dimension also allows increasing the frame rate.
Note: To observe the droplet formation well, the frequency of the frames should be higher than the frequency of droplet formation. We are typically working at ≥ 600 frames per second. 4. Place the syringe pumps close to the microscope. In order to keep the tubing short and avoid gelation of the agarose in the tubing, the pump for the syringe with the aqueous solution should be as close as possible to the microscope stage. We used a lab jack to place it at the right height  8. Cut three pieces of tubing (best with a tubing cutter): used for the oil inlet, the aqueous solution inlet, and the outlet respectively. The tubes for the inlets must be long enough to connect the syringes with the device. The one for the oil can be quite long (typically 40 cm in our case). The one for the aqueous solution should be as short as possible to avoid gelation of the agarose in the tube. In our case, it is ≤ 5 cm. The exact length will depend on your microscope setup. The outlet tube is about 9 cm.
Note: As the used plasmids carry an antibiotic resistance gene, we perform this protocol without sterile technique. It is, however, to autoclave the tubing and the steel couplers, and to disinfect the syringes and the device with 70% ethanol.
9. Connect each tubing to a bent steel coupler by hand (wearing gloves). Note: With these flow rates we typically have a frequency of around 340 Hz, i.e., we produce 1.2 x 10 6 droplets/h. Seventy-four percent of them will not contain any cell. Some beads will also be lost during the process.
5. If you want multiple aliquots of the same sample (e.g., to test different incubation times), it is easier and more accurate to change the collection tube after a specific amount of time (e.g., 15 min for each aliquot) than later having to pipette aliquots of the emulsion. I suggest keeping one aliquot as a negative control without incubation.
6. If you want to make a second sample, let the device run while preparing the second sample in a second 100 μl syringe (+ a fresh aqueous inlet and outlet tubing). 13. After incubation, cool the emulsion on ice for at least 15 min. This is to solidify the gel beads.
C. Recover gel beads 1. If you would like to look at your droplets using (fluorescence) microscopy, you can pipette them onto a microscope slide.  7. If you need to stain the cells (e.g., with Syto 9), you can do that now. If they are already fluorescent, you can proceed to the next step. www.bio-protocol.org/e2911 c. For E. coli, a rough estimate is that OD600 of 1 corresponds to 5 x 10 8 -1 x 10 9 cells/ml.

Calculate your cell dilutions
Example: a. We have droplets with a 50 µm diameter. b. We want an average of 0.3 cell/droplet. c. We calibrated that the OD600 of 1 corresponds to 5 x 10 8 cells/ml. d. We measured the OD600 of our culture to be 0.2. e. Using equation (1) we calculate the volume of our droplets to be 6.5 x 10 -11 L, this means, we have 1.5 x 10 10 droplets/L and 1.5 x 10 7 droplets/ml of aqueous solution.
f. For an average of 0.3 cell we thus need 0.3 x 1.5 x 10 7 cells/ml = 4.5 x 10 6 cells/ml. g. With an OD600 of 1 corresponding to 5 x 10 8 cells/ml, 4.5 x 10 6 cells/ml correspond to an OD600 of 0.009. This means the final OD600 of our sample should be 0.009. As we measured an OD600 of 0.2 we will need to dilute the cells overall 22-fold. A 2-fold dilution is achieved by mixing the cells with the agarose, hence we prepare a 11-fold dilution.
h. These calculations will give you a first indication for your cell density. If the flow cytometry analysis ( Figure 4) indicates that too many or too few beads contain cells, the cell density should be accordingly adjusted in the next experiment.