Loading Progress
0%

Results REE

Key Achievements

Backbone design and optimization for REE binding proteins cloning


To produce REE-binding proteins in E. coli, we built a genetic vector enabling controlled, high-level expression. The binding protein sequences were codon-optimized for E. coli BL21 (DE3), fused to DogCatcher for later attachment to the curli scaffold, and placed under a LacI-inducible T7 promoter system for precise control of expression.

The construct was cloned into a pET28a backbone carrying a kanamycin resistance gene and an origin of replication (ori) functional in both E. coli DH5α (cloning) and BL21 (protein production). The plasmid also contains a LacI nder constitutive expression of a LacI promoter, , a lac operator downstream of a T7 promoter for IPTG-inducible control of the synthetic gene, an optimized ribosome binding site (RBS), a T7 terminator, a copy number regulator that stabilizes replication (rop), and a mobilization sequence for plasmid transfer (bom).

Below is a representation of the synthetic gene assembled in the backbone:


pet28a-dogcatcher-8dq2-schema
Figure 1. Schematic of the pET28a-DogCatcher-8DQ2 Lanthanum binding protein expression construct. Kanᴿ (kanamycin resistance), ori (origin of replication), LacI, lac operator (IPTG-inducible control), T7 promoter (drives gene expression), RBS (ribosome binding site), REE-binding protein–DogCatcher gene (codon-optimized synthetic gene), T7 terminator (transcription termination), rop (plasmid copy number regulator), bom (mobilization sequence).

To clone the binding protein sequences together with the backbone vector, we followed a defined workflow, and the step-by-step process is presented after this summary:

First, each binding protein sequence was amplified by PCR using specific primers, and the corresponding segment of the pET28a backbone was amplified from an existing plasmid to provide compatible ends for Gibson assembly. The binding protein fragments and backbone were then assembled via Gibson assembly and transformed into E. coli DH5α, a strain optimized for efficient cloning. Positive clones were verified by colony PCR and sequencing, and the confirmed constructs were subsequently transformed into E. coli BL21 (DE3) to enable high-level, IPTG-inducible expression of the binding proteins.

Cloning of REE binding protein constructs with backbone in DH5α and BL21(DE3) strains


We performed PCR amplifications of the binding proteins and the backbone plasmid with the primers given in Table 1.

Table 1: Comprehensive list of all designed primers used for PCR amplification of binding proteins and backbone, including sequences, target regions, primer pairs, melting temperatures (Tm), and expected product lengths. N.B. : REE_11 and REE_12 are primers that were designed specifically to amplify the pET28a backbone to contain also the DogCatcher DNA sequence, for optimisation of Pr-binding protein cloning.

View Table PDF

We used following primer pairs for each target:

Gel picture PCR 1
Figure 2. PCR amplification of the backbone and REE binding protein sequences. Agarose gel electrophoresis was performed to verify PCR amplification of the backbone and all REE binding protein sequences. All reactions were successful, with the backbone appearing at the expected size of ~5.3 kb. The binding proteins showed bands corresponding to their predicted sizes: 8DQ2 at ~750 bp, GLAM at ~1.7 kb, 8FNS at ~750 bp, and 8FNR at ~750 bp. For the Praseodymium binding protein (6ZCW), two bands were observed: a minor unexpected band at ~1.5 kb and the expected product at ~2.2 kb. 5kb plus DNA ladder (100–5000 bp range) was used as a size reference.

These results confirm the successful amplification of the backbone and the majority of binding protein sequences, with the exception of a secondary product in Pr binding protein PCR that may represent a nonspecific amplification.

We purified the PCR products and used them for Gibson assembly. We then transformed the assembled constructs into E. coli DH5α. The following day, we observed colonies for the different REE binding proteins. To confirm correct insertion, we performed colony PCR on selected colonies.

Colony PCR gel picture 1
Figure 3. Colony PCR verification of metal binding protein constructs in E. coli DH5α. Colony PCR was performed on DH5α colonies transformed with the REE-binding protein plasmids using the primers PR_VV_054 (fw) and RAV159-pET28a-LacI (rev) (see Table 1). Three separate gels are shown, each displaying eight colonies per metal binding protein. Colonies 3 and 4 for each construct were selected for plasmid clean-up, as they exhibited the expected band sizes: 2183 bp for Pr, 2188 for Gd and 1267 bp for all other metal binding proteins. 5kb plus DNA ladder (100–5000 bp range) was used as a size reference.

The results indicated that most of the picked colonies contained the expected fragments, consistent with the sizes observed during the initial PCR verification of the binding protein sequences. Sequencing further confirmed successful cloning of the constructs encoding the La+Ce, Gd, and Dy binding proteins, but not the Nd and Pr binding proteins. Glycerol stocks were prepared from these verified colonies, and overnight cultures were grown to extract plasmids via miniprep.

We then transformed confirmed plasmids into E. coli BL21(DE3). Colonies grew successfully on selective plates, and we repeated the colony PCR to verify the presence of the correct inserts in the production strain.

Colony PCR gel picture 2
Figure 4. Colony PCR verification of REE binding protein constructs in E. coli BL21(DE3). Colony PCR was performed on BL21(DE3) colonies transformed with the REE-binding protein plasmids for La+Ce, Gd, and Dy using the primers PR_VV_054 (fw) and RAV159-pET28a-LacI (rev) (see Table 1). Two separate gels are shown, each displaying eight colonies per metal binding protein. Colonies 2 and 4 for each construct were selected for plasmid clean-up, as they exhibited bands of the expected sizes: 2188 bp for Gd and 1267 bp for all other metal binding proteins. 5kb plus DNA ladder (100–5000 bp range) was used as a size reference.

These results confirm successful transfer and stable maintenance of the binding protein constructs in the protein production strain (E. coli BL21 (DE3)). We prepared glycerol stocks of the verified strains.


Optimization and re-cloning of Nd and Pr binding protein constructs

▼

We performed additional trials for the Nd and Pr binding proteins, which we did not successfully clone in the first attempt. To address this, we re-amplified the PCR products and repeated the Gibson assembly using optimized fragment ratios and assembly conditions to improve construct formation efficiency.

Gel picture PCR 2
Figure 5. PCR verification of Nd and Pr binding protein sequences. Agarose gel electrophoresis shows the PCR products for the Nd (8FNS) and Pr (6ZCW) binding proteins. Nd PCR was performed with a 50 µL reaction, producing a single band at the expected size of ~750 bp. Pr PCR was performed with 25 µL and 50 µL reactions, and two bands were observed: an unexpected band at ~1.5 kb and the expected product at ~2 kb. A 5kb plus DNA ladder (100–5000 bp) was used as a size reference.

For the Pr binding protein, we attempted to isolate the correct ~2 kb PCR band and performed Gibson assembly; however, colonies either did not grow or, when a colony was obtained, sequencing results were incorrect. Despite repeated attempts and troubleshooting, we were ultimately unable to clone the Pr binding protein.

In contrast, the Nd (8FNS) PCR product was successfully purified and used for Gibson assembly. The assembled constructs were transformed into E. coli DH5α, and colonies grew reliably, allowing verification via colony PCR.

Colony PCR gel picture 3
Figure 6. Picture of the agarose gels of colony PCR in DH5α E. coli for Nd. Colony PCR (primers REE_6 and REE_3) of the transformed DH5α E. coli with Nd binding protein. We tested 8 colonies as shown on the gel. Every colony showed good band sizes and we selected 5, 6 and 7 to send for sequencing. We used a 5000bp DNA Marker, 100-5000bp.

Colony PCR, followed by sequencing, confirmed the correct assembly of the construct in DH5α. The Nd binding protein construct was then successfully transformed into and maintained within the protein production strain. Glycerol stocks of the verified strain were prepared, enabling subsequent protein expression and functional testing. Overall, optimization of the Gibson assembly conditions allowed successful cloning of the Nd binding protein sequence into E. coli BL21(DE3).

Assessment of metal ions capture and release using different Curli-Binding protein constructs


To demonstrate the capacity of our system to capture different REE, we established a rigorous protocol and carried out experiments to quantify its ability to bind, retain, and subsequently release metal ions into the supernatant for collection.


Curli Fiber Production and Assessment with Congo Red Staining

▼

Our host lab provided us two curli-producing strains:

  • csgA- strain (negative control): this strain has the csgABCDEFG operon deleted from the genome.
  • csgA+ DogTag strain: this strain of curli-producing E. coli, carries an optimised DogTag expected to efficiently interact with the DogCatcher.

To produce curli fibers, single colonies were inoculated overnight and then plated and incubated 37°C, and left in the incubator for 48 hours. We performed the Congo Red staining assay to confirm curli production of the provided strain.

Quantification of Curli Fiber Production in csgA+ strain Using Congo Red Staining:

Curli CsgA- vs CsgA+ visual comparison
Figure 7: Visual comparison between csgA- (control) and csgA+ strain, before and after the OD 1 dilution necessary for the test. Congo Red stain is expected to bind to curli fibers and when bound, the intensity of the red color is amplified, additionally, this dye emits fluorescence detectable (excitation at 497, emission at 614nm), allowing a precise quantification. Here even visually, we can already see the difference in red intensity between the strain with, and without production of curli fibers.

Because visual comparison alone is not informative enough, we quantified Congo Red fluorescence associated with curli fibers at 497/614 nm using a microplate reader. To minimize bias from cell density, both csgA- (control) and csgA+ cultures were first adjusted to an OD600 of 1, then measured again. Background signal was subtracted using a PBS blank, and fluorescence was normalized to OD600 to account for differences in cell number.

Table 2: Normalized Congo Red fluorescence per OD for csgA- and csgA+ strains (fluorescence per cell density).

Samples OD after normalization
csgA– 353.6
csgA+ 885.4

After correction, the csgA- strain displayed a normalized fluorescence of approximately 354 a.u., whereas the csgA+ strain reached approximately 885 a.u., representing an approximately 2.5-fold difference. This result confirms that the csgA+ strain produces a significant amount of curli fibers and is suitable for downstream experiments.


Assembly of the system

After synthesizing both components - the curli fibers and the REE-binding proteins - we assembled them into a functional unit designed for REE capture. Our protocol included module preparation, system assembly, exposure to metal solutions, controlled release of bound ions, and quantification of recovery. Full details of the procedures and their optimization process are provided in the Experiments and Engineering Success sections of the wiki.

To assemble the system, we combined the cells with the curli fibers (carrying a Dogtag) and the REE binding proteins fused to a DogCatcher in an Eppendorf tube containing a PBS buffer and incubated the mixture at room temperature for 3 hours in a shaking incubator. This incubation period allowed the DogTag-DogCatcher interaction to occur efficiently, ensuring that as many binding proteins as possible were covalently attached to the curli fiber scaffold.


Exposure of the system to REE ions

In electronic devices and e-waste, REE are typically present as oxides or alloys. Because the oxide form is the most common starting point in industry, we chose to work with REE oxides. To make these metals available for capture by our system, we first dissolved the oxide powders in an acidic solution, thereby releasing the metal ions into solution (the full dissolution protocol is described here).

For testing, we prepared REE solutions at known concentrations and combined them with our assembled system in the presence of HEPES buffer. HEPES was chosen because it does not interact with REE or interfere with the REE concentration measurement assay, while also stabilizing the proteins.


Measuring the REE capture efficiency of the system

One of the major challenges of the project was accurately quantifying how much REE our system could capture. To do this, we adapted an absorbance-based assay using the arsenazo dye developed by Hogendoorn et al. This method was ideal for our purposes: it is fast, sensitive, and can be performed on small volumes using a standard spectrophotometer, with a useful detection range of 0.1–10 µM. Although it was originally designed for La and Eu, we adapted it to other REE, also drawing inspiration from the iGEM Aachen 2023 team, who had modified the same protocol to suit their experiments.

The assay uses the arsenazo III dye, which forms differently colored complexes with divalent and trivalent metal ions under acidic conditions. Since REE readily hydrolyze at low pH, the assay buffer - 2.7-2.8 pH citric acid/phosphate - keeps the solution acidic and stable while ensuring the REE remain soluble. By mixing our REE-containing sample with the dye and buffer, we could observe a color shift and measure its absorbance. Higher absorbance corresponds to higher REE concentration.

To ensure accurate interpretation of our measurements, we generated calibration curves for each REE across a concentration range of 1-250 µM. An example is shown in Figure 8, which displays the calibration curve for La³⁺ in HEPES buffer.

Calibration curve La3+
Figure 8. Representative calibration curve in HEPES buffer for lanthanum (La³⁺) spanning concentrations from 1 to 250 µM.

Since the assay provides the highest precision at lower concentrations, initial measurements were taken at higher sample concentrations, followed by serial dilutions. This approach allowed us to obtain reliable absorbance values within the optimal range. Absorbance readings from the samples were then converted into REE concentrations based on the calibration curves established with known standards.

We performed each calibration curve measurement in technical triplicates to ensure that absorbance values are consistent and accurately transposed.


First measurement - REE depletion from the supernatant

We incubated the assembled system (REE binding protein displayed on curli fibers via Dog catcher - Dog tag interaction) with the REE solution for 30 minutes at room temperature in a shaking incubator. After incubation, we centrifuged the tubes to pellet the cell with the curli fibers and collected 470 µL of the supernatant to measure its absorbance using the arsenazo assay. This first measurement gave us an indirect estimate of how much metal the system had captured: the initial REE concentration we added minus the remaining concentration measured in the supernatant.

For instance, when starting with a concentration of 150 µM, the absorbance decreased from 0.500 (control without the Curli-BP system) to 0.300 (after incubation with the Curli-BP system). This 0.200 difference corresponds to the removal of ~50 µM of REE ions from the solution by the system.

Although this provided a useful first estimation, it did not distinguish between metal ions truly bound to the system and those simply lost during the washing and handling steps. To obtain a more accurate measurement, we performed a second test.


Second measurement - REE release from the pellet

While the first measurement quantifies the amount of REE left in the supernatant, in this measurement we released the REE bound to the pellet containing Curli-BP system and quantified it. After the first measurement, we washed the pellet twice with a HEPES buffer to remove any traces of unbound metal. We then resuspended the pellet in Milli-Q water adjusted to a low pH (2-2.4). It has been shown (e.g. for Lanmodulin proteins) that lowering the pH triggers the release of bound REE ions by inducing conformational changes or partial denaturation (Deblonde et al., 2020).

We incubated the system for 10 minutes under these conditions, then centrifuged it again and collected 470 µL of the supernatant. Measuring the REE in this supernatant with the arsenazo assay gave us the amount of REE ions that had been truly captured and then released by our system.

To illustrate the workflow, Figure 9 illustrates schematically the two complementary measurements used to quantify REE capture and release:

Two-step assay
Figure 9. Two-step assay for quantifying REE capture and release by the Curli-BP system. The first step measures metal depletion from the supernatant after incubation with the Curli-BP system, providing an indirect estimate of capture. The second step measures REE released from the pellet under low-pH conditions, confirming the amount of REE truly bound to the system.

Validation of REE capture and release using lanmodulin and other binding proteins

To validate our system and establish a proof of concept, we first tested it using Lanmodulin, a binding protein known to preferentially interact with Lanthanum (La). All three initial experiments were conducted with La metal, while the final experiment extended the protocol to Ce, Gd, and Dy to demonstrate broader applicability. Congo Red-stained curli were used across all experiments, with the exception of the last one, where unstained curli were applied.

1. The Curli-binding protein system reliably captures 50 µM La³⁺ across different ratios of Curli to binding proteins

▼

To evaluate the system’s ability to capture La³⁺ in this experiment, we tested a wide range of Curli-to-BP ratios in volumes containing either cells expressing the Curli fibers or cell lysates of cells overexpressing the binding proteins (1:1, 1:5, 1:10, 1:25, 1:50, 1:75, 1:100, 1:250, 1:500, 1:750 and 1:1000). We exposed each configuration to 50 µM La³⁺. The varying range of ratios here was to determine optimal ratio and potential saturation of the DogTag-DogCatcher interaction.

Lanthanum capture and recovery by the Curli-BP system (50 µM La³⁺, Curli:BP ratio 1:25).
Figure 10. Lanthanum capture and recovery by the Curli-BP system (50 µM La³⁺, Curli:BP ratio 1:25). Panel A: La³⁺ remaining in the supernatant after 30 minutes of incubation with the system. Conditions include REE only, Curli alone with and without REE, and Curli + BP 1:25 with and without REE, illustrating the fraction of REE not captured by the system. Panel B: La³⁺ released into the supernatant after exposure of the pelleted Curli-BP system to a low-pH buffer (pH 2.4), for the same ratio and initial REE concentration, illustrating recovery of bound REE ions. Error bars represent the mean ± standard deviation of three technical replicates, except for the negative control conditions, which was tested once.

Figure 10 panel A, shows the amount of La³⁺ remaining in the supernatant after 30 minutes of incubation with our system in a 50 µM La³⁺ solution, representing the fraction of REE not captured. In this assay, lower REE concentrations in the supernatant correspond to higher capture efficiency.

As anticipated, curli fibers alone were able to capture some La³⁺ (~20 µM), consistent with their known non-specific metal-binding ability (Tay et al., 2018). In contrast, the addition of binding proteins allowed the scaffold to sequester the full 50 µM La³⁺ at various Curli:BP ratios. While in this figure we only show a few ratios, all the tested ratios demonstrated complete REE capture. These results confirm that our dual-modular system effectively sequesters all available metal ions under the tested conditions, providing a robust platform for further optimization.

While panel A illustrates the system's efficiency in removing REE from solution, panel B quantifies the fraction of captured REE that can be subsequently released from the Curli-BP complex, enabling recovery of the bound ions.

In this second measurement experiment, which corresponds to the first of the release assays quantifying recovery of bound metal, the REE measured in the supernatant reflects the amount successfully released and recovered by the system. After exposure to a low-pH buffer (pH 2.4), the recovered metal ions were slightly lower than the initially captured amount (~50 µM captured vs. ~40 µM released), likely due to minor losses during wash steps or incomplete release under these conditions. Nevertheless, the results are encouraging. Curli fibers alone released only ~10 µM, indicating that unspecific binding contributes minimally to the system's overall performance. In the Curli + BP condition, ~30 µM of the recovered REE can be attributed to the binding protein, highlighting its key role. The consistent capture observed across all Curli-to-BP ratios demonstrates that the system performs robustly even with minimal binding protein, suggesting that further experiments with higher metal concentrations or optimized ratios could reveal its full potential.


2. Curli + BP system captures ~50 µM La³⁺ even in a more concentrated solution containing 200 µM La³⁺, with lower Curli:BP ratios capturing more metal

▼

This experiment directly followed the previous one, aiming to determine the limits of our system by increasing the REE concentration and assessing whether higher Curli:BP ratios improve REE capture and subsequent recovery after exposure to a low-pH buffer.

Lanthanum recovered into the supernatant from Curli-BP system after ion release at varying ratios (200 µM La³⁺, Curli:BP ratios 1:100-1:1000).
Figure 11. Lanthanum recovered into the supernatant from Curli-BP system after ion release at varying ratios (200 µM La³⁺, Curli:BP ratios 1:100-1:1000). Measurement of La³⁺ released into the supernatant after exposure of pelleted Curli-BP system to a low-pH buffer (pH 2.4). The figure includes multiple Curli:BP ratios (1:100, 1:500, 1:1000), illustrating the experimental setup used to assess recovery of bound metal ions at higher initial REE concentrations. Error bars represent the mean ± standard deviation of three technical replicates.

Figure 11 shows that at a high metal concentration (200 µM La³⁺), increasing the BP proportion results in a captured REE concentration similar to that observed in Figure 10, indicating that REE binding reaches a plateau under these conditions. Capture efficiency begins to plateau between 1:500 and 1:1000 ratios, suggesting the system approaches saturation, likely due to a limited amount of curli fibers available for binding protein attachment. If saturation occurred at the protein binding sites instead, increasing the BP proportion would provide additional sites for REE ion capture. To further scale the system’s REE sequestration capacity, more curli fibers would need to be introduced to provide additional binding sites. The 1:500 ratio is already efficient and well-suited for scale-up, allowing higher curli volumes while maintaining effective BP binding. Building on this, we tested the system at higher curli volumes and metal concentrations to evaluate its scalability and efficiency.

Moreover, metal capture begins to plateau between 1:500 and 1:1000 ratios, indicating the system approaches saturation. This apparent saturation is likely due to a limited amount of curli fibers available for the binding proteins to attach. If the saturation occurred at the protein binding sites instead, increasing the ratio of binding protein to curli would directly provide more sites to capture metal ions. To scale up the REE concentration that our system can sequester, we would need to increase the amount of curli fibers, thereby providing more binding sites for the proteins. 1:500 ratio is already efficient and more adapted for scale-up, allowing increased curli volume while maintaining effective BP binding. Building on this, we tested the system at higher curli volumes and metal concentrations to evaluate its scalability and efficiency.


3. Scaling up the experiment with increased Curli + BP volumes and higher metal concentrations resulted in more absolute REE ions captured

▼

To assess the scalability of our system, we increased the amount of curli biomass added to the reaction (16 µL, 32 µL, and 64 µL of curli mixture in PBS, each prepared at the same OD and under the same conditions as in all other experiments). For each curli condition, the BP lysate volume was adjusted according to our standard 1:500 curli-to-BP ratio (8 mL, 16 mL, and 32 mL, respectively). In parallel, we tested higher metal loading by doubling the La³⁺ concentration to 500 µM, 1000 µM, and 2000 µM. This setup allowed us to evaluate system performance across increasing biomass input and metal challenge.

Lanthanum recovered into the supernatant from the Curli-BP system after ion release at increasing La³⁺ concentrations, Curli:BP ratio 1:500, with scaled-up Curli and BP volumes.
Figure 12. Lanthanum recovered into the supernatant from the Curli-BP system after ion release at increasing La³⁺ concentrations, Curli:BP ratio 1:500, with scaled-up Curli and BP volumes. Measurement of La³⁺ released into the supernatant after exposure of pelleted Curli-BP system to a low-pH buffer (pH 2.4) under scaled-up conditions. The assay tests increasing volumes of curli-producing E. coli (16, 32, 64 µL) and proportionally increased binding protein, while maintaining a Curli:BP ratio of 1:500. Conditions include Curli alone and Curli + BP controls. Error bars represent the mean ± SD of three technical replicates, except for "Curli 64 µL + BP system + La³⁺ 2mM", which was performed in duplicate.

Figure 12 illustrates that released REE ions for conditions with Curli alone are very low (REE values around 3–7 µM), even as the curli volume increases (16, 32, 64 µL). This confirms that curli fibers by themselves capture only a small fraction of the REE. Curli + BP condition captures roughly ten times more metal than curli alone at the same concentration, showing that the majority of metal binding is due to the binding protein, not the curli fibers. Our system consistently achieved a maximum recovery of ~40 µM La³⁺, independent of the initial metal concentration, even under scaled-up conditions. While we could not increase this value, the absolute amount of La³⁺ captured increased, because we worked with bigger volumes.


4. Binding proteins for Ce, Dy, and Nd successfully captured and released their respective REE, demonstrating that our approach is promising across both different REE and binding proteins

▼

Using the same protocol, we tested two new binding proteins with three additional REE (with Ce being targeted by the same binding protein as La). All assays were carried out at a REE concentration of 150 µM.

Cerium, Gadolinium, and Dysprosium recovered from the Curli–BP system after ion release (150 µM REE, Curli:BP ratio 1:500)
Figure 13. Cerium, Gadolinium, and Dysprosium recovered from the Curli–BP system after ion release (150 µM REE, Curli:BP ratio 1:500). Measurement of REE ions released into the supernatant after exposure of pelleted Curli–BP system to a low-pH buffer (pH 2.4). Conditions include Curli alone and Curli + BP, with or without REE. The assay was conducted for multiple binding proteins and REE under standardized metal concentrations. Error bars represent the mean ± standard deviation of three technical replicates. Statistical comparison (two-tailed t-test) between Curli alone + REE and Curli + BP + REE showed significant differences for Ce (P = 0.0353) and Gd (P = 0.0418), while Dy was not significant (P = 0.2463).

Figure 13 shows that, for Ce and Gd, Curli + BP recovers more metal ions than Curli alone, while the difference for Dy is not significant. In this experiment, Curli alone exhibits a higher apparent capacity to capture ions compared to previous assays, roughly doubling the recovered REE concentration. This difference likely reflects a change in experimental conditions: unlike earlier assays, unstained Curli fibers were used here instead of Congo Red-stained fibers. While Congo Red is commonly used to quantify curli, it may partially occupy binding sites or influence molecular interactions, thereby reducing non-specific metal capture. Consequently, using stained curli appears to increase assay specificity by limiting unintended metal interactions and ensuring that measured binding reflects primarily BP-mediated capture.

This experiment confirms that our approach can be extended to other binding proteins and to capture other REE. It also highlights the importance of maintaining consistent experimental conditions to ensure reliable comparisons. As further steps, we aim to test the Nd-binding protein, test BP specificity, and further explore the effect of Congo Red staining on Curli fibers in this assay. We are also planning to further improve binding affinity and specificity of the BPs by in silico guided approaches.

References