Overview¶

The pH-sensing developer cells detect acidic conditions (pH 6.0–6.5) using a pH-responsive ssDNA system Hwang et al. (2025). Acidification releases a trigger ssDNA, which activates a toehold switch RNA and induces expression of the protein of interest. To generate a visible colorimetric output, a two-vesicle population system is used. Upon sensing acidic pH, the sensor cells express phospholipase A1 (PLA1), which disrupts neighboring CPRG-loaded vesicles and releases CPRG into the surrounding solution. External β-galactosidase then converts the yellow CPRG substrate into purple CPR, producing a visible color change.

Schematic

Data

Reagents¶

Constructs¶

Annealing was performed in a thermocycler using the following protocol: 5 min at 95 °C, followed by stepwise cooling at 2 min per degree to the melting temperature of the construct (Tm = 52 °C), 30 min per degree from 52 °C to Tm − 10 °C (42 °C), and finally 2 min per degree until reaching 25 °C. The annealed construct was stored at −20 °C.

Protocols¶

CPRG-LUV preparation¶

Lipids for SUVs and LUVs were added from chloroform stocks to the bottom of a 1 mL glass vial and dried under argon/N₂ gas. They were then dried under vacuum for 20 min, and the inner solution was added directly to the lipid film. The process is the same as for GUVs up to the film-drying step. After that, it differs: for GUVs, mineral oil is added, whereas for SUVs and LUVs the inner solution is added directly to the film.

1. Add the inner solution directly to the dried lipid film.

2. Vortex until the film is no longer visible.

3. Sonicate the suspension for 10 min.

4. Run 5 freeze-thaw cycles. One cycle is three operations:

   • Freeze in liquid nitrogen.

   • Thaw in a 35 °C water bath.

   • Vortex for 30 s.

5. Combine each of 12 tubes holding 50 µL LUVs with 200 µL outer solution.

6. Centrifuge at 10,000 g for 10–20 min to pellet the vesicles.

7. Pool the pellets into two tubes. Take 15 µL from each of six tubes, 90 µL in total.

8. Add 160 µL outer solution to each tube, for a final volume of 250 µL.

9. Wash 10 times. One wash is three operations:

   • Centrifuge at 4,000 g for 10 min.

   • Remove 200 µL supernatant without disturbing the pellet.

   • Add 200 µL fresh outer solution.

   Alternate the tube orientation with each wash, so the pellet moves to the opposite side. This releases unencapsulated or trapped solution held between vesicles.

10. After the final wash, leave the tubes on ice for at least 30 min.

11. Tap the tubes gently to resuspend the pellet. Do not pipette.

CPRG-SUV preparation¶

Used from Full Demo Attempt 2 onward. The LUV route above was used for pH Sensor Validation in Solution and Full Demo Attempt 1. The lipid film is prepared the same way as for CPRG-LUV preparation above.

1. Add the inner solution (1 mL) directly to the dried lipid film.

2. Vortex until the film is no longer visible.

3. Sonicate the suspension for 10 min.

4. Run 5 freeze-thaw cycles. One cycle is three operations:

   • Freeze in liquid nitrogen.

   • Thaw in a 35 °C water bath.

   • Vortex for 30 s.

5. Assemble a liposome extruder (Avanti A81411) with a 400 nm pore size membrane filter, and pass the lipid suspension through it 13 times to make SUVs.

6. Centrifuge at 15,000 g for 10–20 min to pellet the vesicles.

7. Pool the pellets into two tubes. Take 15 µL from each of six tubes, 75 µL in total.

8. Add 175 µL outer solution to each tube, for a final volume of 250 µL.

9. Wash 10 times. One wash is three operations:

   • Centrifuge at 10,000 g for 10 min.

   • Remove 200 µL supernatant without disturbing the pellet.

   • Add 200 µL fresh outer solution.

   Alternate the tube orientation with each wash, so the pellet moves to the opposite side. This releases an unencapsulated or trapped solution held between vesicles.

10. After the final wash, leave the tubes on ice for at least 30 min.

11. Tap the tubes gently to resuspend the pellet. Do not pipette.

PLA1-GUV preparation¶

The inverted emulsion method.

1. Combine three lipid stocks in a 20 mL glass vial. All three stocks are in chloroform.

   • 41.0 µL of 25 mg/mL POPC.

   • 1.16 µL of 50 mg/mL cholesterol.

   • 1.95 µL of 1 mg/mL rhodamine-PE.

   This gives 89.9 / 10 / 0.1 mol% POPC, cholesterol and Rhod-PE.

2. Remove the chloroform under a gentle argon flow.

3. Dry the vial in a vacuum desiccator for 30 min.

4. Rehydrate the lipids in 3 mL mineral oil, to 0.5 mM total lipid in oil.

5. Seal the vial. Disperse the lipids in four operations:

   • Bath sonicate for 20 min.

   • Incubate at 60 °C for 1 h.

   • Vortex for 2 min.

   • Bath sonicate again for 20 min.

6. Gently layer 300 µL of the lipid-in-oil dispersion over 400 µL of vesicle outer solution in a 1.5 mL tube.

7. Incubate the tube at room temperature for 10–20 min.

8. Prepare the inner encapsulation solution in a separate 1.5 mL tube during that incubation.

9. Add 600 µL of the lipid-in-oil dispersion to the inner solution.

10. Pipette the mixture up and down thoroughly for 3 min, to produce water-in-oil monolayer emulsion droplets.

11. Add the droplets on top of the oil-water interface formed in step 6.

12. Centrifuge at 2,500 g for 15 min at 15 °C.

13. Remove the oil phase carefully with a pipette.

14. Collect the vesicles into a 0.2 mL PCR tube with a fresh pipette tip.

15. Resuspend the vesicles gently.

LGA Gel Preparation¶

Materials

1. UltraPure™ DNase/RNase-Free Distilled Water, Invitrogen™ Cat number 10977015

2. Low-gelling agarose, Sigma, A9414-5G

Procedure:

1. Make 2.7% agarose stock dissolved in UPDI water (UltraPure™ DNase/RNase-Free Distilled Water)

   For 1500 uL stock preparation, weigh 40.5 mg directly in 2 mL epitube and add 1500 uL of UPDI water). Heat in 85-95 C heat block until fully dissolved. Vortex occasionally.

2. Synthetic cell (vesicle) embedment in 0.7% agarose

   Compositions:

   Component	Fraction	Volume	Notes
   energy solution (2x)	50 v/v%	30 uL
   pH buffer mix*	12 v/v%	7.2 uL
   2.7% agarose dissolved in UPDI	26 v/v%	15.6 uL
   GUVs, SUVs or substrates	12 v/v%	7.2 uL	in 46.67% of Tris1M-HEPES1.15M solution; osmolarity 1170-1180 mOsm
   Total	100 v/v%	60 uL

   1. First, cool the 2.7% agarose stock to 65C.

   2. Mix the agarose with Energy solution and pH buffer mix* first, and cool to 45C.

   3. Add CFE-containing vesicles.

   For 0.35% final agarose concentration, half of gel volume can be replaced with UPDI water.

3. pH buffer mix* preparation

   pH buffer mix was prepared so that resulting pH to be 1) pH 7.4 or 2) pH 6.5 (also to match osmolarity)

   1. For pH7.4: use Tris1M-HEPES1.15M buffer stock (~2520 mOsm)

   2. For pH6.5:

      Component	Fraction	Volume
      Tris1M-HEPES1.15M buffer stock	38.89 v/v%	466.67 uL
      2M HEPES stock	48.02 v/v%	576.19 uL
      37% HCl	6.67 v/v%	80 uL
      Ultra-pure distilled water	6.43 v/v%	77.14 uL
      Total	100 v/v%	1200 uL

Color development¶

The neutralization/basic buffer was prepared by mixing 1 mL of 1 M Tris base with 2.5 µL of 37% HCl. The resulting solution has a pH of 9.9 and a measured osmolarity of 1,165 mOsm/kg.

To reach a final β-galactosidase concentration of 1.2 U/mL, a 3x concentrated solution (3.6 U/mL β-galactosidase in pH 9.9 basic buffer) was prepared from a 1000 U/mL β-galactosidase stock. 30 µL of this solution was added to 60 µL of gel/solution, giving a total volume of 90 µL and a final concentration of 1.2 U/mL β-galactosidase.

Compositions¶

PLA1-expressing GUVs¶

CPRG-loaded LUVs¶

Outer solution¶

The energy solution is a 2× sub-mix of ten components per Sun et al. 2013, made up to 4000 µL. It is not reproduced here: experiments/energy-solution-composition.csv, and its vendor data is in Table 1.

The osmolarity of the Tris1M-HEPES1.15M stock (1 M Tris, 1.15 M HEPES, pH 8), approximately 2,520 mOsm/kg, was measured, not calculated, using a Vapro® Vapor Pressure Osmometer (Model 5600). The stock was diluted (e.g., 1/4, 1/3) at least five times so that the readings fell within the calibrated range (100-1000 mOsm/kg). Each reading was multiplied by its dilution factor, and the reported value is the average of these measurements.

Gel¶

The gel composition is already given: LGA Gel Preparation’s step 2 table, under Protocols. The demo used a total volume of 7.2 µL, composed of: 3 µL of PLA1-expressing vesicles (prepared in 1175 mOsm/kg solution, pH 8), 3 µL of CPRG-loaded SUVs/LUVs (prepared in 1175 mOsm/kg solution, pH 8), and 1.2 µL of 1175 mOsm/kg solution, pH 8.

Results¶

pH Sensor Validation in Solution¶

Worked: An acid-responsive visible color change was observed in solution.

One field of view each, Rhodamine in yellow and Alexa Fluor 647 in red. Each panel is a 502 µm square at 0.465 µm per output pixel, taken at the same stage coordinates in every well, so no panel was positioned to favour a result. All four share one contrast stretch per channel, pooled across the wells, so brightness is comparable between panels: Rhodamine 2410 to 50690 and Alexa Fluor 647 1904 to 30863. Rendered from the OME-Zarr stores by general/render-microscopy-fovs.py.

D4

D5

E4

E5

Conditions¶

384-well square plate, nine wells, 60 µL each. The pH 7.6 arm is plate column 4, the pH 6.3 arm is plate column 5, and H5 is empty in the build sheet, which is why one arm has five wells and the other four. The vesicle slot is a constant 7.2 µL on every well except H4: outer solution plus PLA1-GUV plus CPRG-LUV sums to 7.2 on rows D, E, F and G.

These four wells are also shown as interactive viewers on Interactive viewers.

Full Demo Attempt 1¶

An acid-responsive visible color change was observed. However, the control group (CPRG-loaded LUVs with β-galactosidase but without PLA1-expressing GUVs) also developed color, so the visible difference was not very clear. The only difference between this attempt and pH Sensor Validation in Solution was the absence or presence of agarose gel. Therefore, gel embedment seems to cause CPRG leakage from the LUVs even without PLA1.

Full Demo Attempt 2¶

To decrease CPRG leakage from the LUVs, I extruded them (through 400 nm membranes) to break them down into smaller sizes, because smaller liposomes are known to be more stable than GUVs.

Results: The background signal was slightly lower than in Full Demo Attempt 1, but the difference was still not distinguishable by eye. In addition, the yields of SUVs and PLA1-expressing GUVs were lower than before, so the overall signal was low.

First experiment 7.4 then adjusted to 7.8 (mixture of Tris and Hepes 1M and 1.15M) for acidic condition add HEPES and HCL (wanted to minimize amount of HCl) (might consider using MES buffer for lower pH range).

Full Demo Attempt 3¶

To increase the SUV yield, I increased the starting volume. I also did more rounds of washing to reduce any further background signal. I also reduced the gel concentration from 0.7% to 0.35% to reduce any stress caused by gel embedment. In parallel, I ran a solution experiment to determine whether the issue really originated from the gel embedment.

Results: It worked. A visible color difference was observed both in gel and in solution.

Notes¶

In contrast to previous descriptions, no gramicidin is used. Gramicidin A was not included in the colorimetric demonstration because it caused a portion of the CPRG-loaded vesicles to rupture prematurely, leading to nonspecific color development. Although the absence of gramicidin A may reduce pH-sensing efficiency, PLA1 expression driven by proton diffusion into a subset of vesicles with relatively permeable membranes appeared to be sufficient to initiate the vesicle-lysis cascade.

What’s next¶

QR-code patterned color change in hydrogel¶

To explore a smartphone-readable output format, we tested the two-vesicle colorimetric system in a patterned hydrogel shaped like a QR code rather than in a bulk gel. The goal was to determine whether a mobile app could read the pattern and identify whether the sensor was in the ON or OFF state, indicating detection of the target condition. A visible color change was successfully observed under acidic conditions (pH 6.3), but the color developed outside the patterned region, likely due to diffusion of CPRG and/or CPR over time, preventing successful detection by Samuel Chen’s Hydrogel Sensor app. In contrast, the yellow CPRG substrate remained relatively confined within the pattern, possibly because CPRG encapsulated in SUVs diffused less readily. These results suggest that future improvements could include detecting color changes in the outer gel region or using a less diffusive substrate such as X-gal.