How the Model Guided Our Project

Our project revolves around a model that captures the structural behaviour of RNA and its role in gene regulation. The model describes how two sequences can fold into functional structures, and how these structures can be dynamically altered by external signals. By translating these structural changes into measurable effects on gene expression, the model provided us with a framework to design regulatory RNA switches, which we were able to build and test successfully in the lab.

This framework was not an add-on but central to our entire design–build–test–learn cycle. It allowed us to build a software following the model guidelines, with which we propose switchable RNA parts, predict their behaviour, and then guide the construction of new parts that we later successfully tested in the lab.

The experimental results obtained from testing the individual elements (SECIS and Aptamer) that made up our final switch design were fed back into the modelling of our SKIPPIT switch. This helped us refine both our understanding of RNA folding dynamics and the specific parameters of the system we were building. In this way, the model directly connected abstract structural predictions with concrete laboratory outcomes.

The approach is intentionally general: while we focused on one system for our proof of concept, the same modelling principles could inspire future iGEM teams to design other RNA-based regulatory elements. To make the model scalable for real design, we were transparent about key assumptions. Structural predictions were generated using RNAfold, which estimates the most stable folding pattern and its free energy. We assumed that these predicted minimum free energy (MFE) structures represent useful approximations of real conformations, even though RNA molecules in cells can occupy ensembles and are influenced by proteins and co-transcriptional folding. This simplification let us move quickly from sequence to predicted function, while keeping in mind that predictions would need experimental confirmation.

By grounding our software tool and lab work in this model, we ensured that modelling was not just theoretical but an active design driver. The insights shaped our constructs, explained unexpected results, and ultimately demonstrated that RNA switches can be engineered as modular, tunable elements for synthetic biology.

Software Development

The development of the TADPOLE software was a direct extension of our model: the evaluations, parameter choices, and everything else was based on this model. Through this software, we were able to determine the most promising designs to test in the lab, based on sound thermodynamic and structural criteria.

Model-to-Software Translation

Thermodynamic Validation: All energy calculations (Δ, G_OFF, G_ON) implemented directly from model equations

Structural Checks: Automated validation of OFF disruption, ON recovery, and CRE conformation

Pairing Analysis: Systematic counting and optimization of FRE-CRE interactions

Design Space Exploration: Systematic screening of linker sequences based on model criteria

Wet Lab Guidance

Our project focused on the SECIS + Theophylline Aptamer system. The model allowed us to identify and pre-select linker designs that were energetically more stable in the OFF state (inactive, without theophylline) and which, upon simulating ligand binding, would efficiently shift to the ON state (active).

This in silico validation allowed us to create a specific, optimized panel of candidates for synthesis and subsequent validation in the lab. In this way, the model transformed our strategy from an empirical "needle in a haystack" search to a targeted selection grounded in bioinformatics.

Before Model Guidance

Approach: Empirical trial-and-error

Efficiency: Low success rate

Cost: High synthesis and testing costs

Timeline: Extended experimental cycles

After Model Guidance

Approach: Targeted, model-driven selection

Efficiency: High-confidence candidates

Cost: Reduced synthesis burden

Timeline: Accelerated validation

Construct to Test in the Lab: SECIS + Theophylline System

SECIS Element Overview

During translation, the ribosome decodes an mRNA transcript, using tRNAs to add an aminoacid for each corresponding codon.

SECIS (Selenocysteine Insertion Sequence) is a short RNA element found in the 3' untranslated region (3' UTR) of eukaryotic mRNAs that enables the incorporation of the amino acid selenocysteine at UGA codons, which usually indicate a stop.

Functionally, SECIS performs a task comparable to canonical SCR elements: allowing translation to continue beyond a stop codon. However, unlike SCR elements, which are generally located immediately downstream of the stop codon, SECIS elements are positioned distally, often hundreds of nucleotides away.

Why SECIS is Ideal for Our System

Well-characterized structure: Known functional requirements and critical nucleotides
Clear functional readout: Stop codon readthrough can be easily measured
Structural flexibility: Can be disrupted and recovered through conformational changes
Eukaryotic compatibility: Functions in human cell systems

Selection of the SECIS Element

To select a suitable SECIS element for our constructs, we referred to the work by Touat-Hamici et al. (2018), which provides a comparative analysis of the 26 SECIS elements encoded in the human genome. Their study characterizes the activity and hierarchy of these elements in different cellular contexts, offering insight into which sequences are most robustly recognized by the eukaryotic translation machinery.

Based on this comparative dataset, we selected the well-performing SECIS DIO2, with documented efficiency, compatibility in human cells and a good understanding of its functionality. This ensures better reliability when adapting the system to include synthetic regulatory modules such as aptamers.

SECIS DIO2 Selection Criteria

High Activity: Ranked among the top performers in Touat-Hamici et al. comparative study

Human Compatibility: Proven function in human cellular systems

Well-Characterized: Extensive literature on structure and function

Robust Recognition: Reliable interaction with eukaryotic translation machinery

SECIS Structure Prediction and Design Strategy

The structure of the selected SECIS element was predicted using the SECISearch3 web server. This method is homology-based, and aligns candidate sequences to a built-in covariance model that represents known SECIS elements.

As output, SECISearch3 also offers a colored representation that highlights the key nucleotides for function in green. We focused our design precisely to disrupt these nucleotides in the OFF state 💡.

SECIS DIO2 Structure Analysis

The structure of the chosen SECIS element (DIO2) was predicted by SECISearch3. The prediction highlights in green the most important nucleotides for its function, with a non-canonical pair marked in pink.

Critical SECIS Structural Features

Green-marked nucleotides: Essential for SECIS function - these are the primary targets for disruption in OFF state

Non-canonical pairs: Include C-A pair and A-G, G-A, U-U pairs that are crucial for function

Structural integrity: These features must be preserved in the ON state for proper function

RNAfold vs SECISearch3: Validation Challenges

An important feature of the SECIS element is its non-canonical base pairs, some unsupported by RNAfold, which include the C-A pair, but also the A-G, G-A, and U-U from the first green-marked section.

The fact that RNAfold does not predict these pairings raises some concerns. For instance, it needs to be checked if the SECIS structure in the bound (ON) state is the one predicted by SECISearch3. An alternative approach was needed to check the bound (ON) state's feasibility.

Structural Prediction Comparison

Key differences observed between RNAfold and SECISearch3:

Yellow highlights: Mismatches that RNAfold is unable to predict (non-canonical pairs)
Purple highlights: Extra pairs predicted by RNAfold but not by SECISearch3
Overall assessment: Despite differences, RNAfold structure fits well with homology-based SECISearch3 prediction

Resolution: Taking into account these inaccuracies, RNAfold could be used to explore the effect of sequence variants on the SECIS + aptamer structure. We took the structure predicted by RNAfold as the desired structure that the isolated SECIS element has to adopt in the bound (ON) state.

Riboswitches: Natural RNA Switches

Riboswitches are systems in which a structured RNA sequence, called an aptamer, changes conformation when it binds to specific small molecules. This change in conformation can shift how a gene is expressed by affecting transcription, translation initiation, or RNA stability.

In a functional, designed aptamer–RNA construct, the sequence, including the aptamer and other structural RNA elements, folds in the absence of the ligand in a specific way. When the ligand is present and binds to the aptamer, this structure is disrupted, resulting in a new conformation.

Riboswitch Mechanism

Unbound state: Aptamer (red) and functional RNA elements (other colors) adopt one conformation

Bound state: Ligand (small brown molecule) binding changes the configuration, disrupting the original structure

Functional outcome: Conformational change affects gene expression through various mechanisms

Key Factors for Successful Riboswitch Design

After reviewing extensive literature on aptamers, we found some key factors that can improve the likelihood of ligand binding to the aptamer:

Energy Balance

The energy difference between the ligand-bound and unbound conformations must be relatively small. This allows the system to change between conformations without a high energetic barrier.

Interaction Optimization

The formation of multiple strong interactions, especially ones that the ligand would need to disrupt in order to bind to the aptamer, should be avoided.

These factors were explored further for the development of our system and directly informed our model's energy balance requirements.

The Theophylline Aptamer: An Ideal CRE

For this project, the theophylline aptamer💡, a synthetic RNA of about 30 nucleotides, was selected as our Conformational RNA Element (CRE).

Reasons for Selection

High Specificity

It binds with high specificity to its ligand, theophylline. The binding affinity is 10,000-fold greater than for caffeine, a closely related molecule.

This exceptional specificity ensures that the switch will only respond to theophylline and not to other similar compounds that might be present in the cellular environment.

Synthetic Ligand

Theophylline is not naturally found in mammalian cells, eliminating background interference. It has also been used safely on HEK cells, which are highly relevant as these are the cells used for the experimental part of this project.

This synthetic nature provides clean experimental conditions and precise control over switch activation.

Cross-System Functionality

It has been shown to work in both prokaryotic and eukaryotic systems, providing versatility for different experimental contexts and potential applications.

Defined Structure

Its secondary structure is well understood, simplifying its use in a structure-based design. This structural knowledge was crucial for our model-guided approach .

Design Flexibility

Multiple versions of the aptamer are available, allowing for optimization based on specific design requirements and constraints.

Theophylline Aptamer in Our Model

Binding Energy: Literature value of ~9.5 kcal·mol⁻¹ provides strong switching capability

Target Energy Gap: Δ ≈ 4.5 kcal·mol⁻¹ (approximately half of binding energy)

Structural Change: Well-characterized conformational change upon theophylline binding

Model Integration: Perfect fit for our FRE-CRE framework as a reliable, well-understood CRE

Aptamer Variant Selection

We reviewed several ^{[27,28,29,30,31,18,35]} theophylline-responsive designs from the literature to select the best for our needs. Our criteria included:

Confirmed activity in eukaryotic cells
Absence of self-cleaving ribozymes
Simple and stable secondary structures
Experimental validation

Variant	Pros	Cons	Key Features	Decision
Gallivan	High activation, widely studied, optimized design	Limited testing in eukaryotes	Spacer between aptamer and RBS improves function	Possible
Anzalone	Used in a similar system, well documented	Many sequences tested, might be hard to reproduce	Good documentation	✅ Best option
Smolke (2005)	Designed for eukaryotes, acts in trans	Needs extra components	Ligand triggers aptamer–mRNA interaction	Possible
Smolke (2012)	Mechanism matches design goals	Includes self-cleaving ribozymes	Not suitable for our system	❌ Rejected
Ogawa (2008)	Mechanism matches design goals	Self-cleaving motifs present	Translation blocked by hybridizing with RBS	❌ Rejected
Ogawa (2011)	Uses IRES for translation bypass	Limited in vivo data	Similar functioning as in prokaryotes	Possible
Hartig (2012)	Used in HEK cells	Self-cleaving	Derived from Ogawa variants	❌ Rejected
Hartig (2013)	Designed for viral control	Self-cleaving	Derived from Ogawa variants	❌ Rejected

Selected Aptamer: Theo-ON-5 from Anzalone et al. (2016)

The theophylline aptamer chosen had to have proven functionality. In a previous work, Anzalone et al. (2016) developed ligand-responsive RNA switches capable of regulating frameshift in eukaryotic cells.

Their system is conceptually similar to the SECIS+Aptamer, as both join an RNA structural element with an aptamer to regulate gene expression. From the aptamers tested, Theo-ON-5 was the most effective, showing the largest difference in frameshift percentage between 0 mM and 30 mM theophylline.

Given the robust experimental validation and modularity of their aptamer designs, it was decided to adopt the theophylline aptamer from their study as the regulatory core of this thesis's synthetic switch.

Structural Analysis of Selected Aptamer

The tested sequences differ from each other in the number of nucleotides at the bottom part of the sequence, but they conserve the core functional part responsible for the binding of the ligand. The core functional region remains conserved. Initial studies highlighted the importance of the core binding pocket. However, later findings pointed out that nearby regions (such as the so-called "entry site") are also important for ligand recognition and structural control.

3D Structure Validation

PDB Structure: 3D structure of the aptamer (PDB code: 8D28) and its predicted 2D structure by RNAfold. Key nucleotides for binding with theophylline are highlighted.

Computational Validation: When it comes to structure prediction, the structure predicted with RNAFold, agrees with NMR data. This means that the current version of RNAfold can be used with a certain degree of confidence.

Experimental Design: Dual Luciferase System

In order to explain the design of the plasmids, using SnapGene software (www.snapgene.com), this section describes the essential features required for a minimal understanding of the experimental system.

Model-Guided Design Strategy

Step 1: Identify critical SECIS nucleotides using SECISearch3 predictions (green-marked regions)

Step 2: Design linkers that pair with these critical regions in OFF state

Step 3: Account for RNAfold limitations with non-canonical pairs

Step 4: Validate that theophylline binding can release these pairings

Step 5: Confirm SECIS structure recovery in ON state using RNAfold as reference

Step 6: Select optimal candidates based on energy balance (Δ ≈ 4.5 kcal·mol⁻¹)

Theophylline Aptamer Integration

Binding Energy: Literature value of ~9.5 kcal·mol⁻¹ provides strong switching capability

Target Energy Gap: Δ ≈ 4.5 kcal·mol⁻¹ (approximately half of binding energy)

Structural Compatibility: Theophylline aptamer conformation changes can propagate through linker to SECIS

Experimental Validation: Well-established protocols for theophylline-responsive systems

🧬 SECIS DIO2

Function: Stop codon readthrough

Source: Human genome (top performer)

Validation: SECISearch3 structure prediction

🔗 Optimized Linkers

Design: Model-guided selection

Target: 6-10 nucleotides

Function: Disrupt SECIS in OFF, release in ON

☕ Theophylline Aptamer

Binding: ~9.5 kcal·mol⁻¹

Response: Conformational change upon binding

Validation: Well-established experimental protocols

Dual Luciferase Mechanism

The system is based on dual luciferase. It consists of two proteins that emit light, with a stop codon placed in the middle. If no readthrough occurs, only the first protein is expressed. However, if readthrough occurs, the second protein will also be expressed.

🔴 OFF State

Mechanism: The ribosome ends translation at the STOP codon and only the first protein is synthesised.

Result: Single luciferase signal

SECIS Status: Disrupted by CRE interaction

🟢 ON State

Mechanism: The ribosome bypasses the STOP codon and the two proteins are synthesised.

Result: Dual luciferase signal

SECIS Status: Functional, enables readthrough

Experimental Considerations

Design Specifications

3' UTR Placement: The SECIS element (and also the aptamer) is placed at the 3′ UTR.

Ribosome Protection: The ribosome covers about 30 nucleotides, even if it reads only three. Since the second protein and the SECIS element are consecutive, the SECIS element could influence the ribosome during the translation of the second protein.

Safety Distance: To ensure a safety distance between the end of the second protein and the SECIS element, 30 nucleotides of context were included. These correspond to the 30 nucleotides present in the SECIS DIO2 context in the human genome.

Model-Guided Linker Analysis

Typical linker length ranges from 4 to about 10 nucleotides. For this study, linkers of length 5, 7, 8 and 9 were chosen for laboratory testing. Linker 5 presented some complexities and will be discussed after the others for clarity.

Comprehensive Linker Evaluation

Linker 7

Energy Difference

~3.5 kcal/mol (OFF more stable than ON)

Theophylline Accessibility

Yellow nucleotides (binding sites) mostly free and accessible

SECIS Disruption

Some key nucleotides paired with aptamer in OFF state

SECIS Recovery

Clearest recovery in ON state, resembles independent form

Aptamer-SECIS Pairings

Strong pairings help disrupt SECIS but may hinder ON transition

Linker 8

Energy Difference

4.9 kcal/mol (strongest OFF preference)

Theophylline Accessibility

Binding sites less accessible than other linkers

SECIS Disruption

Key nucleotides paired within SECIS itself (weaker repression)

SECIS Recovery

Minor pairings with context, but function should remain intact

Aptamer-SECIS Pairings

Too few pairings, resulting in weaker repression

Linker 9 ⭐

Energy Difference

4.2 kcal/mol (good OFF stability)

Theophylline Accessibility

Binding sites less accessible but still functional

SECIS Disruption

Key nucleotides effectively sequestered in OFF state

SECIS Recovery

Minor pairings with context, function preserved

Aptamer-SECIS Pairings

Best balance: sequesters SECIS without blocking ON transition

Linker 5 (Modified)

Energy Difference

Only 0.9 kcal/mol (insufficient OFF stability)

SECIS Modifications

First 7 nucleotides changed: ACCAGUG → AAAAGGA

Design Rationale

Mutations distant from key nucleotides, complementary pairs preserved

Limitation

Low energy difference does not guarantee stable OFF state

Linker 5 Special Case

For shorter linkers (<7 nucleotides), no structures were found that satisfied ON/OFF requirements. To broaden possibilities, mutations were introduced into the SECIS element. Since mutations in structural elements can affect both function and folding, precautions were taken:

Limited scope: Only the first seven nucleotides of the SECIS element were mutated, as they are distant from key nucleotides
Paired mutations: Complementary nucleotides were mutated in pairs to preserve base-pairing
Result: A viable linker of length 5 was obtained, but the energy difference was only 0.9 kcal/mol, which does not guarantee stability of the OFF state

Summary of Linker Properties

Linker Length	Pros	Cons
7	Good energy difference (3.5 kcal/mol). Likely prevents readthrough. Aptamer sites accessible.	Many aptamer–SECIS pairings.
8	Strong energy difference (4.9 kcal/mol). Few aptamer–SECIS pairings.	OFF state may not fully repress SECIS. Aptamer sites less accessible.
9	Good energy difference (4.2 kcal/mol). OFF state strongly represses readthrough. Few aptamer–SECIS pairings.	Aptamer sites less accessible.
5	Few aptamer–SECIS pairings.	Low energy difference. OFF state may not fully repress SECIS. Aptamer sites less accessible.

From Theory to Practice: Model-Driven Experimental Design

Our model didn't just provide theoretical insights—it became the practical foundation for experimental design. By pre-validating designs computationally, we transformed wet lab work from exploratory to confirmatory, dramatically improving efficiency and success rates. The systematic linker analysis demonstrates how computational predictions can guide targeted experimental validation.

The Linker: The Dynamic Hinge of the RNA Switch

Simply placing a Conformational RNA Element (CRE) next to a Functional RNA Element (FRE) is not enough to create a reliable ON/OFF switch. The key component that orchestrates their interaction is the linker, a short RNA segment, typically 6–10 nucleotides long, that acts as a mechanical hinge connecting the sensor (CRE) to the effector (FRE).

Function of the Linker

The linker is not a passive connector. Its sequence and length determine how the CRE and FRE fold and interact. A properly designed linker ensures that the system can reliably toggle between OFF and ON states.

Design Principles

Avoid excessive binding: The linker should allow enough pairing to maintain the OFF state, but not so much that the switch gets "stuck" and cannot activate when the signal arrives
Optimal length: Too short a linker may prevent the necessary OFF-state interaction; too long can introduce unwanted folding or too much flexibility, reducing switch precision. Typically, 6–10 nucleotides work well
Enable conformational change: The linker must be designed so that when the CRE changes shape, the energy released is sufficient to break the OFF-state pairings, allowing the FRE to fold into its functional ON-state

Key Aspects to Consider: Validation Criteria

Beyond the basic thermodynamic requirements, successful RNA switch design requires careful validation of structural and functional criteria. These checks ensure that your designed switch will behave as intended in both OFF and ON states.

OFF Disruption: Ensuring Functional Repression

In the OFF state, the FRE must be prevented from performing its function. This is achieved by disrupting the specific structural features that the FRE requires to work.

Validation Check: OFF Disruption

What to verify: In the predicted OFF fold, confirm that the FRE's key functional structure is sequestered or paired with other parts of the construct.

Why it matters: If the FRE can still adopt its functional structure in the OFF state, your switch will leak—it will be partially active even when it should be completely repressed.

Example: SECIS Element Disruption

For a SECIS element (which requires a specific hairpin structure), verify that in the OFF fold ^[18,35]:

The stem region of the SECIS hairpin is disrupted by pairing with CRE or linker nucleotides
Critical loop nucleotides are sequestered in alternative base pairs
The overall SECIS structure cannot form as predicted by the folder

ON Recovery: Confirming Functional Restoration

When the switch activates, the FRE must be able to recover its functional structure. This doesn't need to be perfect, but it should be within acceptable tolerance of the target structure.

Validation Check: ON Recovery

What to verify: In the predicted ON fold, confirm that the FRE adopts (or closely approximates) its target functional structure.

Tolerance: Small deviations are acceptable, but core functional elements must be preserved.

Structural Similarity Metrics

Base pair recovery: What percentage of the target FRE base pairs are recovered in the ON state?
Critical motif preservation: Are essential structural motifs (loops, bulges, specific base pairs) maintained?
Overall fold similarity: Does the predicted ON structure resemble the known functional structure?

CRE Reacted Conformation: Validating Sensor Response

The CRE must adopt the correct conformation when it responds to its signal. This ensures that the conformational change can properly propagate to the FRE. For known ligand-binding elements such as theophylline aptamers, structural validation can rely on experimental studies^{[27,28,29,30,31]}.

Binding-Based CREs (Ligand Sensors)

Validation: Check that the CRE prediction under ON constraints matches the known or expected bound conformation^{[27,28,29,30,31]}.

Method: Compare the predicted ON-state CRE structure with experimentally determined or literature-reported bound structures^{[27,28,29,30,31]}.

Example: For a theophylline-binding aptamer, verify that the predicted bound conformation matches the known theophylline-bound structure from crystallography or NMR studies^{[27,28,29,30,31]}.

Environment-Driven CREs (Temperature, pH Sensors)

Validation: Check the CRE predicted under activating environmental conditions^[27,28].

Method: Use structure prediction tools with appropriate environmental parameters (e.g., different temperatures for thermosensors)^[27,28].

Example: For a temperature-sensitive CRE, predict structures at both baseline (e.g., 30°C) and activating (e.g., 42°C) temperatures to confirm the expected conformational change occurs^[27,28].

FRE–CRE Pairing Balance: The Goldilocks Principle

The number of base pairs between the FRE and CRE in the OFF state must be "just right"—enough to ensure repression, but not so many that the switch cannot activate^[33,36].

Pairing Balance Guidelines

Too few pairings (< 4-6 bp): OFF state may not be stable enough to repress the FRE effectively. The switch may leak^[33,36].

Too many pairings (> 12-15 bp): Creates a large energetic barrier. The CRE signal may not provide enough energy to break all the pairs and activate the switch^[33,36].

Optimal range: Typically 6-12 base pairs, depending on the strength of your CRE signal (E_bind)^[33,36].

Counting FRE-CRE Interactions

In the predicted OFF structure, count:

Direct base pairs: FRE nucleotides paired with CRE nucleotides^[33,36]
Linker-mediated pairs: FRE paired with linker, linker paired with CRE^[33,36]
Stacking interactions: Adjacent unpaired bases that contribute to stability^[33,36]

Energy Balance: Thermodynamic Validation

The energy balance between OFF and ON states is the fundamental thermodynamic requirement that determines whether your switch will work. This builds on the basic energy assumptions but provides practical validation steps^[32,33,34].

Step-by-Step Energy Validation

1. Calculate the Energy Gap

Δ = G_ON - G_OFF

Where:

G_OFF = MFE of unconstrained fold (no external factor)
G_ON = MFE with CRE constrained to reacted conformation

2. Check Default OFF Preference

Δ > 0

This ensures OFF is more stable than ON in the absence of the trigger^[33,34].

3. Verify Switching Capability (Binding-based CREs)

Δ < E_bind

This ensures the binding energy can overcome the energy gap and favor the ON state.

Combined: 0 < Δ < E_bind

Practical Energy Validation Workflow

Step 1: Predict MFE for unconstrained sequence → G_OFF

Step 2: Predict MFE with CRE forced into bound/reacted conformation → G_ON^[32,33]

Step 3: Calculate Δ = G_ON - G_OFF

Step 4: Check that 0 < Δ < E_bind (for binding-based CREs)^[32,33,34]

Step 5: Optimize if needed by adjusting linker sequence or length

Complete Validation Example

System: SECIS element + theophylline aptamer

Energy Results:

G_OFF = -42.3 kcal·mol⁻¹ (unconstrained fold)
G_ON = -37.8 kcal·mol⁻¹ (theophylline-bound constraint) ^[32,33,34]
Δ = -37.8 - (-42.3) = +4.5 kcal·mol⁻¹
E_bind = 9.5 kcal·mol⁻¹ (theophylline) ^{[27,28,29,30,31]}

Validation: ✓ 0 < 4.5 < 9.5 → Energy balance is correct ^[32,33,34]

Structural checks:

✓ SECIS hairpin disrupted in OFF state ^[18,35]
✓ SECIS structure recovered in ON state ^[18,35]
✓ Theophylline aptamer adopts known bound conformation ^{[27,28,29,30,31]}
✓ 8 base pairs between SECIS and aptamer regions (optimal range)^[33,36]

🔍 Structural Validation

Verify OFF disruption, ON recovery, and CRE conformation through structure prediction and comparison

⚖️ Energetic Validation

Ensure proper energy balance: OFF preferred by default, but ON becomes favored after signal

🔗 Interaction Validation

Count and optimize FRE-CRE pairings to achieve the right balance between repression and switchability

References

Five decades of eukaryotic transcription. Nat Struct Mol Biol 26, 757 (2019). https://doi.org/10.1038/s41594-019-0303-1
Friedlander, T., Prizak, R., Guet, C. et al. (2016). Intrinsic limits to gene regulation by global crosstalk. Nat Commun 7, 12307. https://doi.org/10.1038/ncomms12307
Barroso, G. V., Puzovic, N., & Dutheil, J. Y. (2018). The Evolution of Gene-Specific Transcriptional Noise Is Driven by Selection at the Pathway Level. Genetics, 208(1), 173–189. https://doi.org/10.1534/genetics.117.300467
Cramer, P. (2019). Organization and regulation of gene transcription. Nature 573, 45–54. https://doi.org/10.1038/s41586-019-1517-4
Blazeck, J., & Alper, H. S. (2012). Promoter engineering: Recent advances in controlling transcription at the most fundamental level. Biotechnology Journal. https://analyticalsciencejournals.onlinelibrary.wiley.com/doi/abs/10.1002/biot.201200120
Khan, A., Nasim, N., Pudhuvai, B., Koul, B., Upadhyay, S. K., Sethi, L., & Dey, N. (2023). Plant Synthetic Promoters: Advancement and Prospective. Agriculture, 13(2), 298. https://doi.org/10.3390/agriculture13020298
Xu, N., Wei, L., & Liu, J. (2019). Recent advances in the applications of promoter engineering for the optimization of metabolite biosynthesis. World J Microbiol Biotechnol 35, 33. https://doi.org/10.1007/s11274-019-2606-0
De Winter, S., Konstantakos, V., & Aerts, S. (2025). Modelling and design of transcriptional enhancers. Nat Rev Bioeng 3, 374–389. https://doi.org/10.1038/s44222-025-00280-y
Sánchez, Á., & Kondev, J. (2008). Transcriptional control of noise in gene expression. Proc. Natl. Acad. Sci. U.S.A. 105(13), 5081–5086. https://doi.org/10.1073/pnas.0707904105
Splinter, E., & de Laat, W. (2011). The complex transcription regulatory landscape of our genome: control in three dimensions. The EMBO Journal 30, 4345–4355. https://doi.org/10.1038/emboj.2011.344
Duveau, F., Hodgins-Davis, A., Metzger, B. P. H., et al. (2018). Fitness effects of altering gene expression noise in Saccharomyces cerevisiae. eLife 7:e37272. https://doi.org/10.7554/eLife.37272
Zhang, Q., et al. (2024). Transcriptional bursting dynamics in gene expression. Frontiers in Genetics. https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2024.1451461/full
Davies, K., & Nowak, K. (2006). Molecular mechanisms of muscular dystrophies: old and new players. Nat Rev Mol Cell Biol 7, 762–773. https://doi.org/10.1038/nrm2024
Ogawa, A. (2011). Rational design of artificial riboswitches based on ligand-dependent modulation of internal ribosome entry. RNA, 17(3), 478–488. https://doi.org/10.1261/rna.2433111
Werstuck, G., & Green, M. R. (1998). Controlling gene expression in living cells through small molecule-RNA interactions. Science, 282(5387), 296–298. https://doi.org/10.1126/science.282.5387.296
Shanidze, N., et al. (2020). A Theophylline-Responsive Riboswitch Regulates Expression of Nuclear-Encoded Genes. Plant Physiol, 182(1), 123–135. https://doi.org/10.1104/pp.19.00625
Bayer, T. S., & Smolke, C. D. (2005). Programmable ligand-controlled riboregulators of eukaryotic gene expression. Nat Biotechnol, 23(3), 337–343. https://www.nature.com/articles/nbt1060
Mariotti, M., et al. (2016). Lokiarchaeota marks the transition between the archaeal and eukaryotic selenocysteine encoding systems. Molecular Biology and Evolution, 33(9), 2441–2453. https://academic.oup.com/mbe/article/33/9/2441/2925203
Chen, Y., Ho, J. M. L., Shis, D. L., et al. (2018). Tuning the dynamic range of bacterial promoters regulated by ligand-inducible transcription factors. Nature Communications 9(1), 64. https://www.nature.com/articles/s41467-017-02473-5
Haugen, S. P., Ross, W., & Gourse, R. L. (2008). Advances in bacterial promoter recognition and its control by factors that do not bind DNA. Nat Rev Microbiol 6(7), 507–519. https://doi.org/10.1038/nrmicro1912
Jiang, C., & Pugh, B. F. (2009). Nucleosome positioning and gene regulation: advances through genomics. Nat Rev Genet 10(3), 161–172. https://doi.org/10.1038/nrg2522
Jang, H. S., Shin, W. J., Lee, J. E., & Do, J. T. (2017). CpG and Non-CpG Methylation in Epigenetic Gene Regulation and Brain Function. Genes (Basel) 8(6), 148. https://doi.org/10.3390/genes8060148
Segert, J. A., Gisselbrecht, S. S., & Bulyk, M. L. (2021). Transcriptional Silencers: Driving Gene Expression with the Brakes On. Trends Genet 37(6), 514–527. https://doi.org/10.1016/j.tig.2021.02.002
Rodnina, M. V., et al. (2019). Translational recoding: Canonical translation mechanisms reinterpreted. Nucleic Acids Research 48(3), 1056–1067. https://doi.org/10.1093/nar/gkz783
Ge, H., & Marchisio, M. A. (2021). Aptamers, riboswitches, and ribozymes in S. cerevisiae synthetic biology. Life 11(3), 248. https://doi.org/10.3390/life11030248
Lynch, S. A., Desai, S. K., Sajja, H. K., & Gallivan, J. P. (2007). A high-throughput screen for synthetic riboswitches. Chemistry & Biology 14(2), 173–184. https://doi.org/10.1016/j.chembiol.2006.12.008
Akhter, S., Tang, Z., Wang, J., et al. (2024). Mechanism of ligand binding to theophylline RNA aptamer. J Chem Inf Model 64(3), 1017–1029. https://doi.org/10.1021/acs.jcim.3c01454
Wrist, A., Sun, W., & Summers, R. M. (2020). The theophylline aptamer: 25 years as an important tool in cellular engineering research. ACS Synthetic Biology 9(4), 682–697. https://doi.org/10.1021/acssynbio.9b00475
Suess, B., Fink, B., Berens, C., Stentz, R., & Hillen, W. (2004). A theophylline responsive riboswitch controls gene expression in vivo. Nucleic Acids Research 32(4), 1610–1614. https://doi.org/10.1093/nar/gkh321
Warfield, B. M., & Anderson, P. C. (2017). Molecular simulations and Markov state modeling reveal the structural diversity and dynamics of a theophylline-binding RNA aptamer. PLoS ONE 12(4), e0176229. https://doi.org/10.1371/journal.pone.0176229
Zimmermann, G. R., Jenison, R. D., Wick, C. L., Simorre, J.-P., & Pardi, A. (1997). Interlocking structural motifs mediate molecular discrimination by a theophylline-binding RNA. Nature Structural Biology 4(9), 644–649. https://doi.org/10.1038/nsb0897-644
Hausser, J., Mayo, A., Keren, L., & Alon, U. (2019). Central dogma rates and the tradeoff between precision and economy in gene expression. Nature Communications 10(1), 68. https://doi.org/10.1038/s41467-018-07391-8
Mishler, D. M., & Gallivan, J. P. (2014). A family of synthetic riboswitches adopts a kinetic trapping mechanism. Nucleic Acids Research 42(10), 6753–6761. https://doi.org/10.1093/nar/gku262
Gebauer, F., & Hentze, M. W. (2004). Molecular mechanisms of translational control. Nature Reviews Molecular Cell Biology 5(10), 827–835. https://doi.org/10.1038/nrm1488
Touat-Hamici, Z., Bulteau, A.-L., Bianga, J., Lobanov, A. V., & Gladyshev, V. N. (2018). Selenium-regulated hierarchy of human selenoproteome. BBA - General Subjects 1862(11), 2493–2505. https://doi.org/10.1016/j.bbagen.2018.04.012
Liang, J. C., Chang, A. L., Kennedy, A. B., & Smolke, C. D. (2012). A high-throughput, quantitative cell-based screen for efficient tailoring of RNA device activity. Nucleic Acids Research 40(20), e154. https://doi.org/10.1093/nar/gks636
Lorenz, R., Bernhart, S. H., Höner zu Siederdissen, C., Tafer, H., Flamm, C., Stadler, P. F., & Hofacker, I. L. (2011). ViennaRNA Package 2.0. Algorithms for Molecular Biology, 6, 26. https://doi.org/10.1186/1748-7188-6-26
Turner, D. H., & Mathews, D. H. (2009). NNDB: The nearest-neighbor parameter database for predicting stability of nucleic acid secondary structure. Nucleic Acids Research, 38(Database issue), D280–D282. https://doi.org/10.1093/nar/gkp892
Jenison, R. D., Gill, S. C., Pardi, A., & Polisky, B. (1994). High-resolution molecular discrimination by RNA. Science, 263(5152), 1425–1429. https://doi.org/10.1126/science.7510417
Mathews, D. H., Moss, W. N., & Turner, D. H. (2006). RNA secondary structure prediction. Current Opinion in Structural Biology, 16(3), 270–278. https://doi.org/10.1016/j.sbi.2006.04.008

Impact on our project

How the Model Guided Our Project

Software Development

Model-to-Software Translation

Wet Lab Guidance

Before Model Guidance

After Model Guidance

Construct to Test in the Lab: SECIS + Theophylline System

SECIS Element Overview

Why SECIS is Ideal for Our System

Selection of the SECIS Element

SECIS DIO2 Selection Criteria

SECIS Structure Prediction and Design Strategy

SECIS DIO2 Structure Analysis

Critical SECIS Structural Features

RNAfold vs SECISearch3: Validation Challenges

Riboswitches: Natural RNA Switches

Riboswitch Mechanism

Key Factors for Successful Riboswitch Design

Energy Balance

Interaction Optimization

The Theophylline Aptamer: An Ideal CRE

Reasons for Selection

Theophylline Aptamer in Our Model

Aptamer Variant Selection

Selected Aptamer: Theo-ON-5 from Anzalone et al. (2016)

Structural Analysis of Selected Aptamer

3D Structure Validation

Experimental Design: Dual Luciferase System

Model-Guided Design Strategy

🧬 SECIS DIO2

🔗 Optimized Linkers

☕ Theophylline Aptamer

Dual Luciferase Mechanism

🔴 OFF State

🟢 ON State

Experimental Considerations

Design Specifications

Model-Guided Linker Analysis

Comprehensive Linker Evaluation

Linker 7

Linker 8

Linker 9 ⭐

Linker 5 (Modified)

Linker 5 Special Case

Summary of Linker Properties

From Theory to Practice: Model-Driven Experimental Design

The Linker: The Dynamic Hinge of the RNA Switch

Function of the Linker

Design Principles

Key Aspects to Consider: Validation Criteria

OFF Disruption: Ensuring Functional Repression

Validation Check: OFF Disruption

Example: SECIS Element Disruption

ON Recovery: Confirming Functional Restoration

Validation Check: ON Recovery

Structural Similarity Metrics

CRE Reacted Conformation: Validating Sensor Response

FRE–CRE Pairing Balance: The Goldilocks Principle

Pairing Balance Guidelines

Counting FRE-CRE Interactions

Energy Balance: Thermodynamic Validation

Step-by-Step Energy Validation

1. Calculate the Energy Gap

2. Check Default OFF Preference

3. Verify Switching Capability (Binding-based CREs)

Practical Energy Validation Workflow

Complete Validation Example

🔍 Structural Validation

⚖️ Energetic Validation

🔗 Interaction Validation

References