Engineering Success

This page shows the iterations of our DBTL engineering cycle.

Engineering Success

Abstract

Engineering applies systematic principles to efficiently design and build functional systems. Unlike basic research, engineering focuses on creating solutions that meet specific goals. In synthetic biology, this involves modular biological parts, predictive DNA circuit design, measurement of system performance, and iterative optimization based on experimental data.

Our LEGO system (Light-induced Expression of GLP-1 and Insulin) illustrates this approach. It comprises a blue light-responsive transcriptional activator and a dual-hormone expression vector. The activator integrates a light-sensitive iLID/sspB switch, transcriptional activation domains, and the DNA-binding domain of Gal4, while the expression vector contains a 5×UAS-driven hsp70 promoter and a dual-hormone expression cassette, enabling precise and reversible control of insulin and GLP-1 expression and secretion.

Like other engineering fields, engineering a biological system is a complex, iterative process described by the Design-Build-Test-Learn (DBTL) Cycle:

Engineering Cycle Overview

Wet Lab
Cycles 1, 2, 3
Dry Lab
Cycles 4, 5

Cycle 1: Protein Engineering (iLID sspB, Gal4 VP64)

This DBTL cycle is about the conceptual design of the light-responsive transcriptional activation system and the initial engineering of optogenetic switches. The goal was to establish a reversibly activatable modular foundation for dual-hormone gene expression.

Design 1:

Build 1:

Generated DNA sequences for NLS-iLID-mCherry, NLS-miRFPnano3-sspB, iLID-mCherry, NLYN-miRFPnano3-sspB, NLS-iLID-VP64 and NLS-Gal4DBD-miRFPnano3-sspB.

The fusion protein constructs were cloned and expressed in HEK-293T cells.

Test 1:

We used confocal microscopy to verify two key points:

The imaging results clearly demonstrated correct nuclear targeting, successful dimerization under blue light, and reversibility, confirming the functionality of the optogenetic switch system.

Learn 1:

Before: Literature review indicates that the iLID-sspB system functions as a compact blue-light-inducible dimerization tool characterized by rapid response kinetics and reversibility (Guntas et al, 2015). While it has been widely implemented in mammalian cells and model organisms (Countryman et al., 2025), its application in light-gated transcriptional regulation remains unexplored. Studies indicate that when employing the CRY2-CIB1 optogenetic dimerization system for light-inducible transcriptional activation in mammalian cells, abnormal nuclear clearance of CRY2 occurs after 18 hours of blue light exposure (Pathak et al., 2017). This observation prompted our team to design parallel experiments to validate the sustained nuclear localization capability of the iLID-sspB system in mammalian cells.

After: From these experiments, we learned that the nuclear localization of iLID-sspB in mammalian cells remains unaffected by blue light stimulation. The iLID-sspB system responds to blue light reliably, rapidly, and reversibly in live cells. This provides a strong foundation for the next cycle, where we will use endogenous bioluminescence to trigger the system.

Recap of Cycle 1

In this first engineering cycle, we successfully designed, built, and experimentally validated the light-sensitive protein system iLID-sspB. The confocal microscopy experiments confirmed nuclear localization and light-controlled dimerization, and the system exhibited reversible control.

Cycle 2: Bioluminescence-Triggered Activation

This cycle introduces the luciferase NanoLuc (Nluc) as a photoswitch trigger, where NanoLuc interacts with its substrate FFz to generate localized blue light in situ, thereby testing the corresponding iLID-sspB response induced by bioluminescence.

Design 2:

The NLuc is fused within the optogenetic activator to produce localized blue light. The iLID-sspB dimerization system responds to this internal light source, thereby enabling non-invasive control.

Build 2:

Generated DNA sequences for Nluc-iLID-mCherry and NLS-Nluc-iLID-VP64.

The fusion protein constructs were cloned and expressed in HEK-293T cells.

Test 2:

FFz + NanoLuc generates sufficient blue light to trigger iLID-sspB dimerization.
The system demonstrated a rapid response to endogenously generated bioluminescence upon FFz addition.

Learn 2:

Before: Our initial design relied on external blue light delivery for system activation. However, through early-stage expert consultation with Professor Xu from Zhejiang University, we were alerted to the significant challenge of limited tissue penetration depth of external blue light, which would severely restrict the potential for future in vivo therapeutic applications. This critical feedback prompted us to explore alternative activation strategies. Literature research indicates that NanoLuc is a blue-light-emitting luciferase characterized by exceptionally high luminescence efficiency and intensity (England et al., 2016). It has been successfully applied to drive various optogenetic systems and is suitable for transcriptional activation control (Li et al., 2021; Yang et al., 2016). We hypothesized that replacing external light with this intrinsic bioluminescence source could enable non-invasive, in situ control of our iLID-sspB system.

After: The experimental results from Test2 robustly validated our hypothesis. The bioluminescence generated by the Nluc-FFz pair was indeed sufficient to trigger iLID-sspB dimerization effectively. More importantly, this shift from an external to an internal light source represents a pivotal learning point in our engineering journey. It directly addresses the penetration limitation issue raised during the expert interview, confirming the feasibility of non-invasive, chemically-induced control of our optogenetic system. Bioluminescence effectively triggers iLID-sspB, validating non-invasive, in situ control. This supports further integration with hormone expression systems and indicates potential for in vivo applications, guiding our efforts toward a more clinically relevant implementation of the LEGO system.

Recap of Cycle 2

In this cycle, we successfully engineered and validated a bioluminescence-triggered activation mechanism for our optogenetic switch. By integrating NanoLuc, we overcame the limitation of external light penetration, establishing a foundation for non-invasive control of therapeutic hormone expression via oral administration of the small molecule FFz, laying the groundwork for in vivo bioluminescence-activated dual-hormone expression and future therapeutic applications.

Cycle 3: Dual Hormone Expression Design

This cycle involved the design, construction, and validation of a dual-hormone expression cassette containing insulin and GLP-1 sequences, enabling the secretion of both hormones in various mammalian cell types beyond pancreatic β-cells, and ensuring the “single activation, dual effect” principle.

Design 3:

The processing of native human insulin relies on pancreatic β-cell-specific prohormone convertases PC1/3 and PC2 for post-translational modification and precise C-peptide excision. To engineer a universally processable insulin sequence compatible with diverse mammalian cell types, we replaced the native PC1/3 and PC2 cleavage sites with a Furin cleavage motif (RRKR) while retaining the N-terminal signal peptide, thereby ensuring efficient secretion.

The active fragment of human GLP-1 (7-37) was engineered with an A8G substitution to enhance resistance to DPP-IV enzymatic hydrolysis, and fused with the secretion signal peptide of Exendin-4 (Parsons et al., 2007), thereby enabling efficient expression and secretion of bioactive GLP-1 in mammalian cells.

IRES sequence ensures both proteins are translated independently from a single mRNA transcript.

A mCherry fluorescent protein sequence, serving as a reporter gene, was inserted downstream of the dual-hormone expression cassette under the guidance of an IRES sequence.

The dual-hormone expression cassette is driven by a 5×UAS-Hsp70 promoter, where the 5×UAS sequence serves as the targeted binding motif for Gal4DBD, providing a binding scaffold for the transcriptional activator.

Build 3:

The engineered insulin construct consists of an insulin signal peptide followed by Chain B-Furin-Chain C-Furin-Chain A, with its interchain A-B disulfide bond sites predicted by Disulfide by Design 2.0 and confirmed to be identical to those in native human insulin (Craig et al., 2013): The engineered GLP-1 construct: Extendin-4EX leader sequence-Furin-GLP-1(7-37)(A8G). The binding targeting of the NLS-Gal4DBD-miRFPnano3-sspB module within the optogenetic activator to the 5×UAS motif was simulated and analyzed using AlphaFold 3 (Abramson et al., 2024). Constructs cloned into mammalian vectors and transfected into HEK-293T cells for light-responsive co-expression of insulin and GLP-1.

Test 3:

Confocal microscopy was used to quantify reporter fluorescence intensity in HEK-293T cells transfected with the LEGO system under blue-light stimulation or dark conditions, demonstrating blue-light-dependent activation of target gene expression.
Quantitative PCR was employed to assess the transcriptional levels of the two engineered hormone genes and the reporter gene in HEK-293T cells under blue-light stimulation or dark conditions. The results demonstrated significantly elevated transcriptional levels of the target genes in light-stimulated cells, validating light-inducible expression.
Live-cell confocal microscopy imaging of PI3K-Akt signaling pathway activation was performed to validate the bioactivity of insulin present in the supernatant of HEK-293T cells following blue-light stimulation.

Learn 3:

Before: We evaluated multiple strategies for reconstituting native insulin and bioactive GLP-1 in common mammalian cells such as HEK-293T, and ultimately selected the engineered approach described in this cycle.

After: The engineered insulin and GLP-1 were successfully co-expressed under optogenetic activation, and our designed universally applicable insulin for mammalian cells was demonstrated to be functionally active.

Recap of Cycle 3

Successful design, build, and validation of light-responsive insulin + GLP-1 co-expression with functional bioactivity, providing a strong foundation for light-triggered hormone therapy.

Cycle 4: Design and Iteration of the Software

This cycle focused on developing the computational core that translates clinical needs into executable commands for the LEGO system. The goal was to create a reliable, accurate and adaptable software model for calculating hormone doses and their corresponding light parameters.

Design 4:

We designed a software architecture to perform a two-step translation. First, it calculates therapeutic hormone doses (insulin or insulin/GLP-1) from patient-specific physiological parameters (e.g., body weight, blood glucose), using adapted pharmacological models. Second, it converts these biological doses into precise light parameters required to activate the LEGO system in vivo

Build 4:

We built a web-based application with a modular structure, implementing key modules including a user profile manager, a dose calculation engine, and a dose-to-light translator. The initial version provided a functional interface for data input and result display.

Test 4:

We verified the software's logical consistency and output accuracy by comparing software results with manual calculations across a wide range of simulated patient scenarios. We also ensured that the generated light parameters matched the effective ranges established in prior Wetlab experiments.

Learn 4:

The initial model calculations were functionally correct. However, testing revealed the need for greater flexibility to accommodate clinical variability. This informed the critical design addition of an adaptive rule configurator, allowing user adjustment of key variables (e.g., insulin sensitivity factor) to enhance personalization and clinical applicability.

Recap of Cycle 4

Cycle 4 yielded a robust software engine capable of generating personalized light control parameters for the LEGO system. The key learning was the importance of user-adjustable parameters, which increased versatility. The final output of this cycle—a defined set of light parameters (intensity, duration, interval and repetition)—served as the primary input requirement for the Hardware design in Cycle 5.

Cycle 5: Design and Iteration of the Hardware

This cycle focused on building the physical device to reliably deliver the light regimens defined by the Software. The goal was to build a wearable, user-friendly, and precise blue-light delivery system based on the specifications generated in Cycle 4.

Design 5:

Informed by the software's outputs, we specified a hardware device with the following core specifications: a wearable form factor, precise PWM-controlled blue LEDs (465-470 nm), Wi-Fi connectivity for receiving commands, and the ability to execute complex, time-based light profiles ("animations") generated by the software's calculations.

Build 5:

We constructed a prototype centered on an ESP32 microcontroller, integrating an LED array, a Li-Po battery, and basic firmware. The initial firmware enabled manual adjustment of light intensity and duration via a simple built-in web server hosted on the device.

Test 5:

We measured the prototype's light output stability and spectrum using a spectrometer. We tested the reliability of the Wi-Fi connection and the accuracy of command execution during prolonged operation periods.

Learn 5:

This cycle produced a robust hardware device capable of faithfully executing the complex light profiles defined by the Software from Cycle 4. Importantly, the completion of this cycle marks the establishment of a fully integrated, end-to-end therapeutic pipeline: from patient data input in the Software, to precise light regimen execution by the Hardware, culminating in LEGO system activation for hormone production.

Recap of Cycle 5

Cycle 5 delivered a fully functional hardware controller and demonstrated its seamless integration with the software, thereby completing the operational backbone of our proposed therapy. The key learning is that while the integrated control system is functionally operational, its ultimate validation through biological testing is the essential next step.

References

Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J., Bambrick, J., Bodenstein, S. W., Evans, D. A., Hung, C. C., O'Neill, M., Reiman, D., Tunyasuvunakool, K., Wu, Z., Žemgulytė, A., Arvaniti, E., Beattie, C., … Jumper, J. M. (2024). Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature, 630(8016), 493–500. https://doi.org/10.1038/s41586-024-07487-w

Countryman, A. D., Doherty, C. A., Herrera-Perez, R. M., & Kasza, K. E. (2025). Endogenous OptoRhoGEFs reveal biophysical principles of epithelial tissue furrowing. Nature communications, 16(1), 7665. https://doi.org/10.1038/s41467-025-62483-6

Craig, D. B., & Dombkowski, A. A. (2013). Disulfide by Design 2.0: a web-based tool for disulfide engineering in proteins. BMC bioinformatics, 14, 346. https://doi.org/10.1186/1471-2105-14-346

England, C. G., Ehlerding, E. B., & Cai, W. (2016). NanoLuc: A Small Luciferase Is Brightening Up the Field of Bioluminescence. Bioconjugate chemistry, 27(5), 1175–1187. https://doi.org/10.1021/acs.bioconjchem.6b00112

Guntas, G., Hallett, R. A., Zimmerman, S. P., Williams, T., Yumerefendi, H., Bear, J. E., & Kuhlman, B. (2015). Engineering an improved light-induced dimer (iLID) for controlling the localization and activity of signaling proteins. Proceedings of the National Academy of Sciences of the United States of America, 112(1), 112–117. https://doi.org/10.1073/pnas.1417910112

Li, T., Chen, X., Qian, Y., Shao, J., Li, X., Liu, S., Zhu, L., Zhao, Y., Ye, H., & Yang, Y. (2021). A synthetic BRET-based optogenetic device for pulsatile transgene expression enabling glucose homeostasis in mice. Nature communications, 12(1), 615. https://doi.org/10.1038/s41467-021-20913-1

Parsons, G. B., Souza, D. W., Wu, H., Yu, D., Wadsworth, S. G., Gregory, R. J., & Armentano, D. (2007). Ectopic expression of glucagon-like peptide 1 for gene therapy of type II diabetes. Gene therapy, 14(1), 38–48. https://doi.org/10.1038/sj.gt.3302842

Pathak, G. P., Spiltoir, J. I., Höglund, C., Polstein, L. R., Heine-Koskinen, S., Gersbach, C. A., Rossi, J., & Tucker, C. L. (2017). Bidirectional approaches for optogenetic regulation of gene expression in mammalian cells using Arabidopsis cryptochrome 2. Nucleic acids research, 45(20), e167. https://doi.org/10.1093/nar/gkx260

Yang, J., Cumberbatch, D., Centanni, S., Shi, S. Q., Winder, D., Webb, D., & Johnson, C. H. (2016). Coupling optogenetic stimulation with NanoLuc-based luminescence (BRET) Ca++ sensing. Nature communications, 7, 13268. https://doi.org/10.1038/ncomms13268