Tunable protein expression strength with toehold exchange riboregulators

Overview¶

Toehold exchange (THE) riboregulators are a class of RNA molecules that us toehold-mediated strand exchange (TMSE -- Zhang & Winfree (2009)) to regulate protein translation (Piorino et al. (2025), Schaffter et al. (2025)). In TMSE reactions an input RNA strand binds to an RNA duplex (gate) via a short toehold domain and, through sequence complementarity, initiates a process of branch migration that ultimately can displace an output RNA strand from the double stranded gate duplex. In the case of a THE riboregulator, the release of the output RNA results in exposure of a ribosome binding sequence (rbs) that connects RNA strand exchange to protein expression. THE riboregulators are a subset of a broader class of circuits, termed ctRSD circuits (Schaffter & Strychalski (2022), Schaffter et al. (2023), Bae et al. (2020)), that use self-cleaving ribozymes to cotranscriptionally produce the double-strand RNA gates needed for TMSE reactions. ctRSD circuits can be programmed to execute logic, signal amplification, and signaling cascades.

The mechanism of TMSE offers many design parameters for precisely tuning the strength of protein expression. In Figure 1 we emphasize methods that rely on design changes to the short input RNA rather than changes to the gate that contains the long protein coding sequence. In addition to transcriptional control of expression strength by changing DNA template concentrations (Figure 1, left), protein expression strength can be controlled at the translational level by varying the length of an unstructured domain at the 5’ end of the input strand (d domain). This unstructured domain serves as a standby site for the ribosome to bind to facilitate translation initiation after strand exchange (Karolchik et al. (2013)), with increasing lengths increasing translation strength (Figure 1, right). Lengths beyond 20 bases are not predicted to provide additional benefit (Karolchik et al. (2013)). Because the toehold exchange mechanism is reversible (Zhang & Winfree (2009)), protein expression can also be controlled by modulating the strength of the forward or reverse strand exchange reactions. Increasing the degree of complementarity between the input toehold of the input and gate (s’ domain) increases the rate of production of gate in the ON state. Conversely, increasing the degree of complementarity between the output toehold of the gate and input (e domain) decreases the rate of production of gate in the ON state. Both mechanisms thus tune protein expression strength at the post-transcriptional level (Figure 1, middle).

Together these results illustrate three distinct mechanisms for tuning protein expression with THE riboregulators:

1. Modulating the concentration of reactants (transcriptional control)

2. Modulating the rate and extent of reaction to produce products (post-transcriptional control)

3. Modulating the translation rate of the products (translational control)

Here we present the sequences and protocols for translational control experiments (modulating the length of the d domain on the input strand).

Design¶

For a detailed description of THE riboregulator design see Supplementary Section 2 of Schaffter et al. (2025). Supplementary Section S2 of Schaffter & Strychalski (2022) describes the design choices for the broader class of ctRSD gates.

Below we summarize a few design considerations of particular note:

1. THE riboregulators are currently designed with an 18-base linker sequence (n domain) upstream of the protein coding sequence (CDS) that adds 6 amino acids N terminus of the protein.

2. There is a start codon at beginning of the n linker domain for translation initiation; the start codon in native protein sequence is changed to a leucine codon (TTA) to minimize the chance for spurious initiation from gates in the OFF state.

3. Not all design trends in PURE produce the same results in cells or in cell lysate, notably the post-transcriptional control mechanisms (see Piorino et al. (2025) and Schaffter et al. (2025)).

To design alternative sequences, see Resources below for sequence compiling software to create new combinations of domains.

Annotated plasmid maps can be found on Benchling: Link

Materials and Equipment¶

Key Materials¶

DNA Components¶

Protocol¶

Preparation of HBC620 Pepper dye stock solution¶

The HBC620 dye is not necessary to assess THE riboregulator performance, but it does provide a measurement of transcription rate in an experiment. It should only be included if the RNA component you are testing encodes for the Pepper RNA aptamer, otherwise no signal will be produced.

1. Centrifuge the tube containing lyophilized HBC620 dye to ensure all the dye is at the bottom of the tube.

2. In a chemical fume hood, prepare a 10 mM stock solution by adding 100 % DMSO to the HBC620 dye tube (273.6 µL DMSO per 1 mg of HBC620 dye).

3. Vortex thoroughly until dye is dissolved.

4. Wrap tube in aluminum foil to prevent photobleaching.

5. Store at -20 °C when not in use.

Preparation of the molecular beacon (optional)¶

The molecular beacon is not strictly necessary to assess THE riboregulator performance, but it does provide a complementary measurement to fluorescent protein expression for strand exchange and can provide mechanistic insight. It is useful to include in experiments in which translation is absent or inhibited, such as experiments in PURE lacking ribosomes or elongation factors, as this enables measurement of strand exchange without fluorescent protein production.

1. Centrifuge tubes containing lyophilized DNA to ensure all the DNA is at the bottom of the tube.

2. Using nuclease free water, resuspend lyophilized DNA to 100 µM concentration (10 µL of water per 1 nmol of DNA).

3. Vortex thoroughly until DNA is dissolved.

4. In a 200 µL PCR tube, prepare 20 µM molecular beacon solution for thermal annealing (Table 6).

5. In a thermocycler, run the following thermal annealing protocol (Table 7).

6. Store annealed molecular beacon at -20 °C when not in use.

Preparation of linear DNA templates¶

Linear PCR product templates are recommended as they are simpler to work with and less prone to variability from preparation-to-preparation compared to plasmid extractions. We also find that linear PCR product templates produce higher fluorescent signals in PURE than plasmid templates.

Gene fragments containing full transcription units (spanning a promoter sequence to terminator sequence) can be ordered (Table 3 and Table 4) and PCR amplified as shown in Table 8. Alternatively, for THE riboregulator gates, a single gene fragment spanning the N-terminal linker sequence to the terminator sequence of the gate can be ordered and the 5’ UTR and promoter of the gate can be ordered separately and appended with overlap PCR (Table 5). This reduces cost and increases flexibility as multiple short 5’ UTR sequences can be ordered and appended to the single protein coding sequence. Table 9 shows how to set up the overlap extension PCR.

1. Prepare DNA templates:

   a. If ordering DNA as gBlock gene fragments, resuspend DNA to 10 ng/µL in nuclease free water.

   b. If ordering DNA as eBlock gene fragments, the DNA will come at 10 ng/µL in IDTE buffer and can be used directly.

2. Prepare DNA primers:

   a. Centrifuge tubes containing lyophilized DNA to ensure all the DNA is at the bottom of the tube.

   b. Using nuclease free water, resuspend lyophilized DNA to 100 µM concentration (10 µL of water per 1 nmol of DNA).

   c. Vortex thoroughly until DNA is dissolved.

3. In a 200 µL PCR tube prepare PCR solution (Table 8 or Table 9).

4. In a thermocycler, run the following protocol (Table 10). The same protocol can be used for single template PCR or overlap extension PCR.

5. Purify the PCR products from the reactions with the PureLink PCR cleanup kit.

   a. Add 300 µL of B2 buffer to PureLink column.

   b. Add all 75 µL of PCR reaction the PureLink column and pipette mix with B2 buffer ~10 times.

   c. Centrifuge PureLink column at 13000 rpm for 1 min, discard flow through.

   d. For overlap PCR products, a 300 µL wash with B3 buffer is recommended to remove short secondary products. DO NOT DO THIS FOR SHORT TEMPLATES THAT ENCODE INPUT RNAS or a lot of DNA will be lost.

   i. Add 300 µL of B3 buffer and repeat step c.

   e. Add 650 µL Wash buffer to PureLink column and repeat step c.

   f. Repeat step c. one more time to remove any residual Wash buffer.

   g. Add 35 µL of PureLink Elution buffer and let the column sit at room temperature for ~2 min.

   h. Move PureLink column to a fresh collection tube and centrifuge at 13000 rpm for 1 min.

   i. Save collection tube with eluted DNA.

6. Measure the concentration of the purified PCR product using A260 absorbance measurements (Nanodrop) or fluorogenic measurements (Qubit).

   a. For PCR products we find these two measurements are fairly similar. Qubit measurements were used in our published work.

7. Purified PCR products can be stored at -20 °C when not in use.

Preparation of plasmid templates¶

Methods for plasmid extraction can be found in Zhang & Winfree (2009). Additional details on best practices for DNA plasmid preparation for cell-free expression systems can be found in Romantseva et al. (2022).

To obtain similar results to linear templates with plasmid templates, we have identified three crucial points:

1. After plasmid extraction from cells a secondary PCR clean up should be performed on the plasmids (we use the PureLink kit specified in materials)

2. We found high copy number plasmids produced results most comparable to linear templates. Low copy number plasmids (~10 copies) deviated from linear template results even when using nominally the same concentrations.

3. For plasmids Qubit concentration measurements are necessary to get close correlation to linear template results. Nanodrop measurements indicate much higher plasmid concentrations than Qubit measurements; this difference is not as great for linear templates.

Kinetic plate reader experiments¶

This protocol is for experiments demonstrating how to use toehold exchange riboregulators to obtain translational control of protein expression. Translation strength is controlled by varying the length of the d domain on the input strand. We will test 4 different d domain lengths: 1-base (low expression), 5-base (medium expression), 10-base (high expression), and 20-base (maximum expression). Additionally, we include a control that uses a decoy input (Io) that is not complementary to the toehold exchange riboregulator. This serves as the OFF state for the system. Io is included so that all 5 samples have the same total concentration of DNA template. This keeps the transcriptional load across all the samples the same to enable direct comparison of translational differences due to differences in d domain length. For experiments that vary the concentration of complementary inputs, we recommend adding Io to each sample such that the total DNA template concentration across all the samples is the same.

The protocol below is for testing the 5 samples described above with 3 technical replicates for each sample. The volumes of DNA templates to add to the reactions are based on stock concentrations of 100 nM.

1. Thaw cell-free expression reagents from -80 °C on ice.

   a. PURExpress Solution A and Solution B or

   b. Nucleus SMix, PMix, tRNA, and ribosomes

2. Thaw DNA templates (purified PCR products or plasmids), annealed molecular beacon, and HBC620 dye stock solution from -20 °C at room temperature.

3. While solutions are thawing, set up the plate reader protocol

   a. Incubate samples at 37 °C

   b. Set the protocol to conduct fluorescence measurements every 5 min, reading from the bottom of the plate

   c. Set wavelengths for fluorescence measurements as described in Table 11

4. Add the DNA templates to 0.5 mL DNA LoBind tubes as specified in the first 4 columns of Table 12, then cap tubes.

   a. Note the volumes in Table 12 assume 3 technical replicates with 10 % excess volume

5. Prepare Master Mix in Table 13 (PURExpress) or Table 14 (Nucleus Cytosol) in a 1.7 mL or 2 mL microfuge tube.

6. Pipette mix the Master Mix with a pipette volume set to 50 µL.

7. Add the volume of Master Mix specified in the 5th column of Table 12 to each of the DNA LoBind tubes.

8. Pipette mix each of the samples upon addition of the Master Mix.

9. Take 10 µL at a time from each sample in DNA LoBind tubes and transfer to a well in a 384-well plate; there should be 3 wells for each sample, corresponding to 3 technical replicates.

10. Check for bubbles and apply a clear plastic cover sheet to the plate.

11. Put the plate in plate reader and run the protocol.

12. Let reaction run for at least 4 hours.

13. Stop or pause the experiment in the plate reader software and remove the plate.

14. Add 3 µL of 10 µmol/L MB trigger RNA to each reaction well and pipette mix.

    a. This step is only needed when using the molecular beacon (MB)

    b. Addition of excess MB trigger RNA will cause all of the MB molecules in each well to react and release a fluorescent strand, allowing for an internal maximum fluorescence value of MB normalization in each well

15. Cover plate with a clear sheet, put back in the plate reader, and continue the measurements.

16. Let the reaction proceed for another hour or leave overnight.

17. Stop the experiment and export the data for analysis.

Results¶

PURExpress

Nucleus Cytosol

Resources¶

Analysis Notebooks¶

Two analysis notebooks accompany this DevNote:

• PURExpress Analysis (experiments/20250311-PURExpress-THE/analysis.ipynb)

• Nucleus Cytosol Analysis (experiments/20260512-Cytosol-THE/analysis.ipynb)

External Tools and Sequences¶

• ctRSD and THE riboregulator sequence compiler

  • Documentation: ctrsd-simulator.readthedocs.io

  • CoLab: sequence compiler CoLab

• Annotated sequence / plasmid maps

  • Benchling: toehold-exchange-riboregulator-published-sequences

Credits¶

Developers¶

• NIST Cellular Engineering Group Link

  • Samuel Schaffter, Fernanda Piorino, Eugenia Romantseva

License¶

The text and figures in this DevNote are made available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. The content of this DevNote is not subject to the CERN Open Hardware Licence – Permissive (CERN-OHL-P). The corresponding author, Samuel W. Schaffter, is an inventor on two patent applications pertaining to ctRSD circuits (PCT/US2022/053229) and THE riboregulators (PCT/US25/13610).