Engineering Success

Overview

In APOPTO-SENSE 2.0, we applied synthetic biology engineering principles to design a modular biosensing system for detecting chemotherapy-induced apoptosis in cancer cells. Following the Design-Build-Test-Learn (DBTL) cycle, we engineered a dual-cell platform using synthetic Notch (synNotch) receptors and orthogonal transcription factors to convert phosphatidylserine (PS) exposure—a key apoptosis marker—into quantifiable fluorescent signals. This addresses the need for rapid, personalized drug sensitivity testing in oncology.

Our engineering process emphasized modularity, orthogonality, and dose-dependency, drawing from established tools like synNotch (previously used in iGEM for antigen sensing) and Gal4-UAS systems. We iterated through multiple DBTL cycles to optimize specificity, linearity, and robustness, starting with a linear amplification MVP and planning expansions for multiplexed cell death detection. For project context, see Description; for protocols, see Experiments.

Design Principles

We designed APOPTO-SENSE 2.0 as a composable system with two core modules: a sensor for apoptosis detection and a reporter for signal amplification. Key principles included:

  • Modularity: Interchangeable domains in synNotch allow customization (e.g., swapping recognition elements for different cell death modes).
  • Orthogonality: Use of yeast-derived Gal4 to avoid interference with mammalian pathways, ensuring low background noise.
  • Dose-Dependency: Linear response circuits to correlate output with input (apoptotic cell quantity), critical for drug screening.
  • Scalability: PiggyBac integration for stable cell lines, enabling high-throughput applications.

Module 1: Apoptosis Sensor (SynNotch Receptor)

Initial Design: Fused Annexin V (extracellular domain for PS binding) to a Notch transmembrane core and Gal4-VP64 intracellular payload. This creates a "plug-and-play" receptor activated by PS exposure on apoptotic cells.

Rationale: Annexin V's high affinity (Kd ~1-10 nM) and calcium-dependency ensure specificity to apoptotic membranes. Gal4-VP64 provides orthogonal nuclear translocation upon cleavage.

Engineering Considerations: We selected a minimal Notch TMD to minimize steric hindrance and ensured the construct fits within PiggyBac vectors for stable integration.

Module 2: Signal Reporter (Linear Amplification Circuit)

Initial Design: Gal4-VP64 binds 5x UAS repeats upstream of a minimal promoter driving TagBFP expression. Constitutive PGK-mCherry serves as an internal reference.

Rationale: UAS multiplicity amplifies signal linearly (output proportional to input), avoiding bistable "all-or-nothing" responses from positive feedback. TagBFP/mCherry ratio normalizes for variables like cell number.

Engineering Considerations: Minimal promoter reduces basal expression (<5% of induced levels); spectral separation (TagBFP: 390/450 nm; mCherry: 587/610 nm) enables ratiometric quantification.

DBTL Cycle 1: Initial Prototype and Testing

Design: Assemble synNotch and reporter plasmids via Gibson cloning, targeting HEK293T cells for engineering.

Build: Synthesized genes, amplified in E. coli, and transfected using PiggyBac for stable lines (see Experiments for protocols).

Test: Induced apoptosis in HL-60 cells with raphasatin (10 µM) and co-cultured with sensor cells. Measured TagBFP/mCherry ratio via fluorescence microscopy.

Learn: Low activation observed (ratio ~0.3 vs. expected >1.0). Hypothesized: Insufficient UAS repeats or suboptimal transfection efficiency. No off-target activation in controls (specificity confirmed).

DBTL Cycle 2: Iteration for Improved Sensitivity

Design: Increased UAS repeats from 5x to 9x for stronger amplification; optimized transfection reagent ratio (2:1 DNA:lipid).

Build: Re-assembled plasmids; verified via sequencing and gel electrophoresis.

Test: Repeated co-culture assays with serial dilutions of apoptotic HL-60 (10^4-10^6 cells). Quantified linearity (R² >0.85) and LOD (~5% apoptotic fraction).

Learn: Improved ratio to >1.5 in apoptotic samples, but variability in expression (CV ~20%). Identified issue: Transient transfection instability; switched to stable lines for consistency.

DBTL Cycle 3: Refinement for Specificity and Robustness

Design: Added EDTA controls to block Ca2+-dependent Annexin V binding, confirming PS specificity. Incorporated constitutive mCherry for normalization.

Build: Generated polyclonal stable HEK293T lines with dual selection (puromycin/blasticidin).

Test: Co-cultured with non-apoptotic cells or PS-masked apoptotic cells. Assessed crosstalk via flow cytometry (no activation in negatives).

Learn: Achieved high specificity (>95% signal reduction in controls) and precision (CV <10%). However, extended co-culture (>24h) led to sensor cell stress; optimized to 15h.

Key Iterations and Optimizations

  • From Linear to Potential Feedback: MVP focused on linear circuits for dose-dependency; future cycles could add positive feedback (e.g., Gal4 auto-activation) for ultrasensitivity, but we avoided due to stability risks.
  • Failure Analysis: Early low expression traced to promoter leakiness—mitigated by minimal promoter use.
  • Modular Expansions: Designed for orthogonality; e.g., integrate TetR for multiplexing necrosis detection.
  • Quantitative Metrics: Used ImageJ for fluorescence quantification; validated linearity with Hill coefficient ~1 (indicating non-cooperative response).

For full results, see Results.

Proof of Concept and Validation

Engineering success was validated through:

  • Specific activation only in apoptotic co-cultures (see figure legend in Results).
  • Dose-dependent outputs correlating with apoptosis rates (confirmed by Annexin V flow cytometry).

Future Engineering Directions

  • Multiplexing: Engineer parallel synNotch variants for detecting ferroptosis (e.g., via lipid peroxidation sensors) or pyroptosis (caspase-1 cleavage).
  • High-Throughput Adaptation: Integrate into microfluidic chips for automated drug screening.
  • Clinical Translation: Optimize for patient-derived organoids, incorporating biosafety features (e.g., kill switches).

This iterative engineering process demonstrates how DBTL cycles transformed a conceptual design into a functional biosensor, advancing synthetic biology in oncology.