Engineering Overview

Our goal was to build a low-cost, yeast-based detection system that reports BPA with a visible output suitable for homes, schools, and community testing.

We engineered a yeast two-hybrid (Y2H) BPA sensor where BPA-induced dimerization of ERα (ESR1) reconstitutes a split transcription factor (LexA-BD/B42-AD) and turns on lacZ. A BPA-biased design (ESR1-300+ truncation + five LBD mutations at aa 519–535) produced a same-day visible response: 901 Miller Units at 3 h with 1 μM BPA vs 17.9 MU control (~50×), while E2 stayed at 15.6 MU (~control). A full-length ESR1 version worked after overnight growth (129 MU; ~13×) but was slower and less specific.

To ensure fair comparisons across tubes, color readouts were normalized by reaction time, assay volume, and cell density (OD₆₀₀) and reported as Miller Units; when the plate reader saturated (“over”), we measured a 10× dilution and corrected.

We followed the iGEM engineering cycle throughout our process: Design → Build → Test → Learn.

Design

We evaluated two ESR1 designs within the same Y2H architecture:

• (i) A truncated ESR1-300+ carrying five ligand-binding-domain substitutions (aa 519–535; Y, G, N, T, L; codons TAC, GGC, AAC, ACC, CTC) to bias for BPA and improve folding in yeast.
• (ii) A full-length ESR1 carrying the same five substitutions as an architectural control.

Both designs used identical bait/prey vectors and the same LexA-operator→lacZ reporter:

• Bait (pGilda): LexA-BD::ESR1-300+ (5×LBD) via EcoRI/BamHI
• Prey (pB42AD): B42-AD::ESR1-300+ (5×LBD) via EcoRI/XhoI
• Reporter: LexA operators → minP → lacZ → CYC1t
• Mutagenesis: Segmented site-directed mutagenesis to install five LBD mutations; assemble fragments.

*See process tab in our Wet Lab for more details.

Build

The bait plasmid (pGilda) encodes the LexA DNA-binding domain fused to ESR1, and the prey plasmid (pB42AD) encodes the B42 activation domain fused to ESR1. To bias the sensor toward BPA and improve folding in yeast, we engineered an ESR1-300+ truncation and installed five substitutions in the ligand-binding domain (aa 519–535; Y, G, N, T, L with codons TAC, GGC, AAC, ACC, CTC) using segmented site-directed mutagenesis.

Inserts were cloned into pGilda with EcoRI/BamHI and pB42AD with EcoRI/XhoI, maintaining reading frames at the BD/AD junctions. The reporter is a LexA operator promoter driving lacZ with a yeast terminator. All junctions were confirmed in-frame; promoters and terminators were correctly oriented; and no internal EcoRI/BamHI/XhoI conflicts remained post-assembly, as verified by diagnostic PCR, Sanger sequencing, and sequence alignment.

Test

Each assay was run under four conditions (25–30 °C):

• Blank: Water/buffer only (matrix control)
• Control (−ligand): Yeast + substrate, no BPA/E2 (leakiness)
• BPA 1 μM
• E2 1 μM (estradiol; related molecule)

The lacZ reporter produces a yellow product visible to the human eye. We timed each reaction, read OD₄₂₀, and converted to Miller Units (MU = 1000 × OD₄₂₀ / [time × volume × OD₆₀₀]). If the reader saturated (“over”), we measured a 10× dilution and used that value. The Blank remained clear, and the Control of Mutant-300 was low, confirming negligible background and leak.

Two Versions of the Sensor

1. A) Mutant-300 (ESR1-300+ with 5 LBD substitutions): Designed for BPA bias and improved yeast folding.
2. B) Full-length ESR1 (same 5 LBD substitutions): Architectural control; larger protein, typically slower in yeast.

For the full-length ESR1 build, we used the same vectors and restriction sites with the same five LBD substitutions, generating pGilda–LexA-BD::ESR1-Full (5×LBD) and pB42AD–B42-AD::ESR1-Full (5×LBD) analogs.

Mutant-300 cultures reached OD₆₀₀ ≈ 1 by ~3 h (ready for same-day reading), whereas the full-length set typically required longer growth (overnight) before a clear readout.

A) Mutant-300 (ESR1-300+ 5×LBD)

Visual Results

Quantitative Results (Miller Units)

3 h — Mutant-300: BPA = 901 MU (~50× control 17.9); E2 = 15.6 MU (≈ control).

B) Full-length ESR1 (5×LBD)

Visual Results

Quantitative Results (Miller Units)

Overnight — Full-length: BPA = 129.3 MU (~13× control 9.67); E2 = 49.7 MU (~5× control).

Conclusion

The ESR1-300+ 5×LBD design meets our pre-set criteria (BPA ≫ control, BPA ≥ E2) and enables a same-day response. The full-length version is functional but slower and less specific, making Mutant-300 more suitable for field use. This matches our tube results: a single bright BPA tube at 3 h for Mutant-300, and broader yellowing at 24 h for full-length.

Learn

The ESR1-300+ (5×LBD) design delivers a same-day, BPA-biased response with low background (901 MU at 3 h vs 17.9 MU control ~50×), while E2 remained low (15.6 MU). The full-length version functions but needs longer growth and shows higher E2 activation (BPA 129 MU ~13×, E2 49.7 MU ~5×), reducing specificity.

• Mutant-300 is the preferred design for same-day detection.
• Standardize a 3-hour color assay at 25–30 °C with visual or smartphone thresholding for field use.

Limitations & Next Iterations

Our target users (homes, schools, community groups) need a sensor that works outside the lab—without incubators or cold-chain storage. Lyophilization (freeze-drying) turns yeast into a stable, shippable, room-temperature pellet that can be rehydrated on site.

Demonstrating that our Mutant-300 sensor retains performance after drying and storage is essential for a “fieldable, same-day BPA check.” Robustness testing answers four key questions:

• Shelf life & storage: Does the dried kit still work after weeks–months at room temp or 4 °C?
• Transport tolerance: Do pellets survive temperature swings and humidity?
• Usability: After rehydration with real samples, do we still get accurate results?
• Reproducibility: Are results consistent across batches and time?

Concentrations Iterations

To move beyond a single “worked at 1 µM” result, we hope to run a 3-hour dose–response with the Mutant-300 design to quantify sensitivity and dynamic range.

BPA will be tested across 0–10 µM (e.g., 0, 0.05, 0.1, 0.3, 1, 3, 10 µM) alongside our standard Blank and −ligand Control, holding temperature (25–30 °C), assay volume, and timing constant.

For each tube we’ll record start/stop times, read OD₄₂₀, correct any saturated reads with a 10× dilution, normalize by OD₆₀₀, and report Miller Units (mean ± SD, n≥3).

From these data we’ll extract the EC₅₀ (midpoint), LoD (mean blank + 3 SD), and fold-activation vs control to define a practical same-day threshold.

Success looks like a clear, by-eye call at ≤3 h with LoD ≤ 0.3 µM, EC₅₀ ≤ ~1 µM, and ≥10× fold at 1 µM, confirming that users can detect BPA without instruments while our quantitative metrics back the visual readout.

Robustness Iteration: Lyophilized vs Fresh

Goal

Prove the kit works from a dry pellet.

Change

Lyophilize cells with 5–10% trehalose (± skim milk, sorbitol). Trehalose replaces water and vitrifies membranes/proteins, preventing damage during drying/rehydration; skim milk/sorbitol cushion and osmo-protect cells to preserve the 3-hour BPA readout after storage.

Build

Prepare pellets, vacuum-seal, and store at both 4 °C and room temperature.

Test

Rehydrate in sample; measure MU after 3 hours. Repeat after 1, 4, and 8 weeks of storage.

Success Metric

Lyophilized signal ≥ 70% of fresh; baseline unchanged.

Decision

If passed, freeze-drying becomes the default kit format.