Engineering-Project

Overview of our engineering strategy

This page presents the Design-Build-Test-Learn (DBTL) cycles of our project PFAway. Here, we describe the rationale behind our design choices, the strategies we used to build and test our system, and the lessons we learned along the way. You will also find how we managed our experiments, overcame challenges, and improved our approach step by step. Throughout PFAway, we have completed three engineering cycles, focusing on three modules: the construction of a strain capable of orthogonal replication, the development of a screening technology for DeHa2, and the enhancement of P. putida’s fluoride resistance.

For each module, we performed a series of tests, analyzed the results, and adapted our strategy to improve outcomes. To visualize this process:

• DBTL cycles (left boxes) are supported by experimental data (right boxes).
• Down arrows indicate the iterations, highlighting the thoughts and new strategies that emerged from previous results.
• Colors highlight outcomes of the cycles: red for unsuccessful attempts, green for successful results.

Building an E. coli strain capable of orthogonal replication

Our primary objective was to introduce the replication operon (see part BBa_25HKSDJN) into E. coli DH10B in order to enable accelerated mutagenesis of the linear replicon (see part BBa_25794BPU) for the directed evolution of DeHa2. Inspired by Tian et al., 2024, we first attempted chromosomal integration of the operon, before shifting to a strategy whereby the operon is carried on a low-copy plasmid.

We used the pTarget plasmid to deliver guide RNAs targeting the SS9 genomic region for Cas9, and the pCas plasmid to supply the molecular components necessary for both λ-Red recombination and Cas9 machinery. Together, these plasmids enable the insertion of the replication operon at the SS9 site within the bacterial chromosome (see Design page).

Transformation of E. coli DH10B with pCas, pTarget, and, the replication operon.

Colony PCR of our strains after chromosomal insertion trials
Figure 1: Colony PCR of our strains after chromosomal insertion trials.
Well Mq shows the 1 kb Plus ladder from invitrogen, while well 1 to 16 corresponds to PCR on different colonies using primers flanking the chromosomal insertion. Bands at 1302 bp represent the SS9 site without insertion, whereas bands at 4982 bp correspond to the SS9 site with the chromosomal insertion.

WGS (data not shown) and colony PCR (Figure 1) indicate that the replication operon was not integrated at the SS9 site. Expected bands for successful insertion (4982 bp) were not observed; only bands corresponding to the empty SS9 site (1302 bp) are present.

Colony PCR screening of transformants and Whole Genome Sequencing (WGS).

Despite several attempts, no successful integration was detected, neither from the chromosome sequencing data nor in colony PCR using primers flanking the SS9 site. These results suggest that the chromosomal insertion at SS9 was not efficient in our hands.

We chose a one-copy plasmid (OCP). Using Gibson assembly, we attempted to insert the replication operon into the plasmid (see Design page).

Gibson assembly of gBlocks A–B–C into the pUD1387 OCP backbone to construct pUD1387-operon.

Alignment of the expected one-copy plasmid and the plasmid extracted from transformants.
Figure 2: Alignment of the expected one-copy plasmid and the plasmid extracted from transformants.
The upper band corresponds to the expected one-copy plasmid containing the full synthetic operon, while the lower band represents the sequence of the plasmid extracted from the transformants. A mismatch or missing region is shown in red.

In Figure 2, alignment of the expected one-copy plasmid with the plasmid extracted from transformants shows that only part of the synthetic operon (gBlocks B and C) was successfully integrated. These results indicate that the current assembly approach needs optimization. Further experiments are ongoing, and results are pending.

Transformation of pUD1387-operon and sequencing.

Transformation was partially successful: only gBlocks B and C were integrated into the plasmid, whereas gBlock A was missing. This indicates incomplete assembly or recombination in E. coli.

Development of a screening and selection platform for DeHa2 evolution

Our objective was to implement a fluorescence-based biosensor for the detection of fluoride ions produced during the defluorination of short-chain PFAS by DeHa. The fluorescence intensity would report on the activity of the DeHa2 variants generated by orthogonal replication. The most active enzyme variants would then be selected. We first attempted to use a riboswitch-sfGFP construct that was described by the iGEM team SDU Danemark 2023. Unfortunately, its sensitivity was insufficient for the low fluoride concentrations expected during the first evolution cycles. Inspired by recent work on RNA aptamers, we then turned to the FluorMango aptamer reported by Husser et al., 2023. This RNA aptamer fluoresces upon fluoride binding and is expected to provide a specific signal. However, its extracellular nature made the aptamer difficult to use for bulk screening. To overcome this limitation and enable high-throughput applications, we developed a microfluidic-compatible screening protocol, see details in the Design page.

               

Inspired by SDU Danemark 2023, we constructed a pUC19 plasmid carrying a constitutive promoter that produces an RNA containing a riboswitch. In the presence of fluoride, this riboswitch enables the translation of sfGFP, resulting in fluorescence (see Design page).

Construction of pUC19-sfGFP plasmid with fluoride riboswitch upstream of sfGFP.

Endpoint OD600 measurement of E. coli carrying pUC19-sfGFP under increasing NaF concentrations.
Figure 3: Endpoint OD600 measurement of E. coli carrying pUC19-sfGFP under increasing NaF concentrations.
Measurements were taken after overnight culture in a 96-well microplate. pUC19 bearing strain was used as a negative control, and p15A-PihfbsfGFP bearing strain was used as a positive control expressing sfGFP under the control of a constitutive promoter.
Normalized fluorescence (fluorescence/OD600 nm) of E. coli carrying pUC19-PJ23119-FRSsfGFP under increasing NaF concentrations.
Figure 4: Normalized fluorescence (fluorescence/OD600 nm) of E. coli carrying pUC19-sfGFP under increasing NaF concentrations.
Fluorescence measured at 𝜆excitation = 485 nm and 𝜆emission = 516 nm.

In Figure 3, growth curves (OD600) of E. coli carrying pUC19-sfGFP show reduced viability as NaF concentration increases. Normalized fluorescence (fluorescence/OD600) across NaF concentrations confirms that the riboswitch-sfGFP construct lacks sufficient sensitivity to respond to the low fluoride levels expected during initial mutagenesis cycles. Even at 25 mM NaF, no significant fluorescence was observed, casting some doubts on the utility of this biosensor under our conditions.

Fluorescence was measured in 96-well microplates with increasing NaF concentrations.

The initial riboswitch design was not sensitive enough, making it incompatible with the small amounts of fluoride expected to be released by DeHa2 during early evolution cycles.

A double-stranded DNA coding the FluorMango aptamer served as a template for in vitro transcription with the T7 RNA polymerase. This allowed for the expression of the aptamer, which was subsequently tested in microplates (see Design page).

The aptamer was produced by in vitro transcription of the DNA template.

Specificity tests of the FluorMango aptamer after 1 hour of incubation.
Figure 5: Specificity tests of the FluorMango aptamer after 1 hour of incubation.
Fluorescence is shown as a function of the tested molecules: TFA, NaF, NaCl, and microfluidics oil (Fluo-oil 135) with or without surfactant (FluoroSurf O at 2%). n = 3 biological replicates with two technical replicates each. For TFA and oil: n = 1 biological replicate. Fluorescence was measured at 𝜆excitation = 510 nm and 𝜆emission = 550 nm.
Calibration of the FluorMango aptamer after 1 hour of incubation.
Figure 6: Calibration of the FluorMango aptamer after 1 hour of incubation
Relative fluorescence is shown as a function of increasing NaF concentrations, measured at 𝜆excitation = 510 nm and 𝜆emission = 550 nm. Each biological replicate (n = 3) included two technical replicates. Grey points represent individual measurements (n = 3), and black symbols indicate the mean values. The shaded area represents the confidence interval at 95%.

In Figure 5, specificity tests of FluorMango aptamer in the presence of various salts and small molecules confirm the findings of Husser et al., 2023 and demonstrate strong aptamer specificity. In Figure 6, calibration experiments across different fluoride concentrations show that the biosensor produces a clear, proportional, and highly fluoride-specific fluorescence response.

Fluorescence measurement in 384-well microplates in the presence of NaF or other compounds (negative controls).

The aptamer functions as a fluoride-specific biosensor. However, its extracellular nature (it does not permeate the cell wall) requires a compatible strategy to screen DeHa2 variants produced in the orthogonal replication strain.

Develop a droplet microfluidic screening platform to compartmentalize the FluorMango aptamer and a single cell of the orthogonal replication strain to quantify fluoride release by DeHa2 variants.

Water-in-oil emulsion droplets containing the FluorMango aptamer and its ligand TO1-biotin, with or without fluoride ions, were generated in a microfluidic device. Droplets were screened one by one for their fluorescence and were subjected to sorting based on intensity thresholding.                    
Confocal fluorescence microscopy images of emulsion droplets containing the FluorMango aptamer and TO1-biotin after 1 h of incubation.
Figure 7: Confocal fluorescence microscopy images of emulsion droplets containing the FluorMango aptamer and TO1-biotin after 1 h of incubation.
Droplets without fluoride (NaF –) show no detectable fluorescence, whereas droplets with fluoride (NaF +) display strong fluorescence measured at 𝜆excitation = 480/40 nm and 𝜆emission = 527/30 nm.
Microfluidic sorting of a mixed droplet population (20% NaF +, 80% NaF –).
Figure 8: Microfluidic sorting of a mixed droplet population (20% NaF +, 80% NaF –).
Fluorescence measured at 𝜆excitation = 480/40 nm and 𝜆emission = 527/30 nm. Sorting was realized at 𝜆excitation = 488 nm and 𝜆emission= 525 nm. Post-sorting analysis shows strong enrichment of fluorescent droplets, confirming the selection efficiency of the FluorMango–microfluidics platform.

Figure 7: shows confocal fluorescence microscopy images of emulsion droplets containing the FluorMango aptamer and its ligand TO1-biotin after 1 hour of incubation with (NaF +) or without (NaF –) fluoride. Droplets without fluoride exhibited no detectable fluorescence, while droplets with fluoride displayed strong fluorescence, confirming the system’s functionality. Figure 8 presents a mixture of 20% fluoride-containing droplets and 80% fluoride-free droplets subjected to microfluidic sorting. After sorting, fluorescence microscopy revealed a strong enrichment of fluorescent droplets, demonstrating the efficiency of the selection process. Together, these results validate the FluorMango–microfluidics strategy as a robust platform for high-throughput screening. Once coupled with DeHa2-mediated defluorination, this system could be applied to select the most efficient variants after successive mutagenesis rounds.

A mock library consisting of a mixed population of droplets either containing (fluorescent) or not containing (nonfluorescent) NaF was prepared at 1:5 ratio, i.e., 20% of active droplets. Sorting efficiency was determined by sorting the fluorescent droplets that were subsequently recovered and analyzed under fluorescence microscopy.

The sorted fraction of droplets contained 80% of fluorescent droplets, demonstrating successful enrichment. We conclude that FluorMango coupled with droplet-microfluidic is a viable strategy for high-throughput screening and selection of orthogonal replication–generated variants.

Enhancing P. putida’s fluoride resistance

Our objective was to evaluate the fluoride tolerance of Pseudomonas putida KT2440 as it must withstand the rise of fluoride ions released during the degradation of short-chain PFAS by the evolved DeHa enzyme. As a strategy to enhance fluoride resistance, we focused on overexpressing FluC, a fluoride-specific transporter capable of expulsing intracellular ions out of the cells. Initial experiments without FluC overexpression provided data to feed our computational model, which indicated that overexpression would be necessary for P. putida’s viability. By overexpressing FluC, we aimed to improve fluoride tolerance, ensuring sustained growth and efficient PFAS degradation under high intracellular fluoride stress.

Wild-type P. putida KT2440 was grown in different media to generate diverse data for feeding our computational model (see Model page).

P. putida KT2440 was grown in 96-well microplates in M9 minimal medium supplemented with acetate under increasing fluoride concentrations supplied in the form of NaF to assess its natural tolerance.

Growth curves of P. putida KT2440 exposed to increasing NaF concentrations.
Figure 9: Growth curves of P. putida KT2440 exposed to increasing NaF concentrations.
Growth was monitored over time in M9 minimal medium supplemented with 45 mM acetate, using a 96-well microplate at 30 °C under different NaF concentrations. Solid lines represent mean values, while dashed lines correspond to individual replicates (n = 3).

In Figure 9, growth curves of P. putida exposed to different NaF concentrations over time are shown. Cell growth decreases as fluoride concentration increases, reflecting stress at levels relevant for PFAS degradation. These data provide a baseline measurement of natural fluoride tolerance in the strain, from which a MIC of 75 mM was estimated.

OD600 growth curves were monitored in liquid culture supplemented with NaF, enabling inhibition constant calculation to feed the computational model for simulations of PFOS degradation with WT P. putida.

Native fluoride tolerance may be insufficient to support degradation of polyfluorinated compounds, suggesting that enhancing fluoride resistance could be required.

A pSEVA438 plasmid carrying a constitutive promoter was constructed to drive FluC transcription (see Design page).

The pSEVA438-fluC plasmid was assembled to overexpress FluC in P. putida.

Growth curves of P. putida KT2440-pSEVA438-fluC overexpressing FluC under increasing NaF concentrations.
Figure 10: Growth curves of P. putida KT2440-pSEVA438-fluC overexpressing FluC under increasing NaF concentrations.
Growth was monitored in M9 minimal medium supplemented with 45 mM acetate, using a 96-well microplate at 30 °C under different NaF concentrations. Solid lines represent mean values, while dashed lines correspond to individual replicates (n=3).

In Figure 10, growth curves of P. putida overexpressing FluC are shown under varying NaF concentrations. Compared to the baseline, cells exhibit improved growth at high fluoride concentrations, with MIC increasing from 75 mM to 100 mM. This demonstrates that constitutive FluC overexpression significantly enhances fluoride resistance. The data also confirm that the engineered strain maintains improved growth at fluoride concentrations relevant to PFAS degradation, highlighting FluC overexpression as a promising strategy to mitigate fluoride stress.

P. putida KT2440 was transformed with pSEVA438-fluC and growth was monitored under increasing NaF concentrations.

Constitutive FluC expression improves fluoride resistance, enabling robust growth under high fluoride stress that may be caused by efficient PFAS degradation.

Next, find out about our results here: