This project focuses on developing a biosensor for rapid detection and sterilization of Escherichia coli (E. coli), Salmonella, and Staphylococcus aureus (S. aureus). We selected three phage-derived bacterial-binding proteins (PBPs) — T3 phage tail fiber protein (TFP), Salmonella phage CKT1 tailspike protein (TSP), and S. aureus phage plyV12 cell wall binding domain (CBD). These proteins were synthesized using synthetic biology and conjugated to gold nanoparticles (AuNPs) via electrostatic adsorption, resulting in AuNPs@TFP, AuNPs@TSP, and AuNPs@CBD. AuNPs@TFP and AuNPs@TSP showed strong binding affinity and induced a color change upon bacterial detection. The biosensor can rapidly detect these bacteria within 30 min using a smartphone app to analyze RGB values for real-time results. Additionally, under 808 nm NIR light for 10 min, the biosensor exhibited potent antibacterial activity.

Goal: To express and purify three distinct proteins: T3 phage tail fiber protein (TFP), the tailspike protein (TSP), and the cell wall binding domain (CBD) of the endolysin plyV12, we aim to construct three expression plasmids: pET28a-gp17, pET28a-Tailspike and pET28a-plyV12. These plasmids will be designed to facilitate the efficient expression of the respective proteins in BL21(DE3) bacterial, thereby enabling subsequent purification and functional studies.
1) Gene and Linearized Vector Amplification
To construct the pET28a-Tailspike, pET28a-gp17, and pET28a-plyV12 plasmids, we first amplified the target genes and linearized pET28a vector fragments using Polymerase Chain Reaction (PCR). The PCR products were then analyzed by agarose gel electrophoresis to verify the amplification and size of the DNA fragments.
The agarose gel electrophoresis results demonstrated successful amplification of the target genes and the linearized vector fragments. Distinct bands corresponding to the expected sizes of the tailspike (2067 bp), gp17(1689 bp), and plyV12(525 bp) genes were observed, confirming the presence of the desired PCR products (Figure 1). Additionally, the linearized pET28a vector fragment (5854 bp) was also clearly visible, indicating successful linearization.
These results indicate that the PCR amplification of the target genes and the linearized pET28a vector fragments were successful, providing the necessary components for subsequent cloning steps.
Figure 1. Agarose Gel Electrophoresis of PCR Amplified Products for Plasmid Construction
Agarose gel electrophoresis (1% agarose) of PCR amplified products used for the construction of plasmids. PCR product of the (A) gp17 gene and corresponding linearized pET28a vector. (B) gp17 gene and corresponding linearized pET28a vector. (C) Tailspike gene and corresponding linearized pET28a vector. Note: M indicated DNA ladder.
2) Gibson Assembly and Bacterial Transformation
Following the successful amplification of the target genes and linearized vector fragments by PCR, we proceeded with the Gibson Assembly to recombine the DNA fragments into the desired plasmids: pET28a-Tailspike, pET28a-gp17, and pET28a-plyV12. The Gibson Assembly reaction mixtures were then used to transform DH5α competent cells.
After transformation, the DH5α cells were plated on LB agar plates containing kanamycin, for selection. Numerous single colonies were observed after the plates were incubated overnight at 37°C (Figure 2), indicating successful transformation.

Figure 2: Bacterial Colony Formation on LB Agar Plates Post-Transformation
LB agar plates containing kanamycin selection were used to assess the success of the Gibson Assembly and transformation into DH5α cells. The plates were incubated overnight at 37°C. Plate showing colonies resulting from the transformation with (A) the pET28a-gp17 plasmid. (B) the pET28a-plyV12 plasmid. (C) the pET28a-Tailspike plasmid.
3)Colony PCR Screening
To identify potential positive clones, colony PCR was performed on several bacterial colonies from each transformation plate. The colony PCR products were analyzed via agarose gel electrophoresis. The gel images revealed distinct bands corresponding to the expected sizes of the Tailspike, gp17, and plyV12 gene inserts (Figure 3A), indicating the presence of the desired recombinant plasmids in several colonies.
These colonies will be further verified by sequencing. The sequencing results confirmed that the gene sequences in the plasmids were correct and matched the expected sequences for gp17, plyV12, and Tailspike (Figure 3B). This confirmed that the plasmid constructs pET28a-gp17, pET28a-plyV12, and pET28a-tailspike were successfully constructed without any mutations.

Figure 3: Colony PCR and Sequencing Results for the Identification of Positive Clones
(A) Agarose gel electrophoresis showing colony PCR results for the identification of positive clones containing the recombinant plasmids. Distinct bands at the expected sizes for the gp17, Tailspike and plyV12 inserts confirm the presence of the desired recombinant plasmids.
(B) Sequencing chromatograms of representative positive clones. The chromatograms show high-quality sequences with no mutations or insertions.

To express and purify the tail fiber protein (TFP) from bacteriophage T3, the tailspike protein (TSP) from Salmonella phage CKT1, and the cell wall binding domain (CBD) of the endolysin plyV12 from Staphylococcus phage, the constructed the corresponding expression plasmids: pET28a-gp17, pET28a-Tailspike, and pET28a-plyV12 were transformed into E. coli BL21 (DE3) cells for protein expression.
Protein expression was induced by adding 1 mM IPTG to the culture medium. The SDS-PAGE gel shows distinct bands at the expected molecular weights for plyV12, gp17 and Tailspike protein (Figure 4A), confirming successful IPTG-induced expression of the target proteins. After 24 hours of induction, the cells were harvested, lysed, and the His-tagged proteins were purified using nickel-affinity chromatography. The purified proteins were analyzed by SDS-PAGE to assess their purity and molecular weight (Figure 4B, C, D). The results demonstrated successful expression and purification of the target proteins.

Figure 4. SDS-PAGE Analysis of Expression and Purification of target Protein
SDS-PAGE Analysis of Protein Expression and Purification. (A) Total cell lysate of BL21(DE3) cells carrying pET28a-gp17, pET28a-Tailspike, and pET28a-plyV12 plasmid before/after IPTG induction were used as sample. SDS-PAGE Analysis of Purified Proteins, Different purified fraction of (B) purified TFP from BL21(DE3) cells carrying pET28a-Tailspike, (C) purified CBD from BL21(DE3) cells carrying pET28a-plyV12, (D) purified TSP from BL21(DE3) cells carrying pET28a-gp17.S indicated the supernatant of cell lysis. P indicated the pellet of cell lysis. FT indicated the Flow-through fraction. Wash indicated the wash solution fraction. Elution indicated the elution buffer of purified protein.

The synthesis of gold nanoparticles was achieved through the reduction of chloroauric acid (HAuCl₄) using sodium citrate as the reducing agent. Initially, the chloroauric acid solution was colorless (Figure 5A). Upon addition of trisodium citrate and subsequent heating, a distinct color change was observed. The solution gradually turned purple-red, indicating the successful formation of gold nanoparticles (Figure 5B). This color change is attributed to the specific surface plasmon resonance (SPR) of the gold nanoparticles, which confirms their synthesis.
Figure 5. Color Change Observed During the Synthesis of Gold Nanoparticles
(A) Initial colorless chloroauric acid (HAuCl₄) solution before the addition of sodium citrate. (B) The solution after the addition of trisodium citrate and heating, showing a purple-red color.
Following the synthesis of gold nanoparticles (AuNPs), we aimed to conjugate PBP proteins (TFP, TSP, and CBD) to AuNPs via electrostatic physical adsorption. This process was conducted in a mildly alkaline pH environment. Specifically, 100 μg of each protein TFP, TSP, and CBD was added to 1 mL of AuNPs solution. Additionally, pET28a empty vector protein was used as a control. The AuNPs solution was initially purple and exhibited an absorption peak at 530 nm. Upon successful protein conjugation, the maximum absorption peak was expected to shift from 530 nm to 533 nm.
After adding 100 μg pET28a empty vector protein to 1 mL of AuNPs, the absorption peak shifted from 530 nm to 533 nm. This shift in the absorption peak indicates successful conjugation (Figure 6A). Similar to pET28a empty vector protein, after adding 100 μg of TFP and TSP proteins to 1 mL of AuNPs, the absorption peak shifted from 530 nm to 533 nm. This shift in the absorption peak indicates successful conjugation of both TFP and TSP proteins to the AuNPs (Figure 6B，C). During the conjugation process of CBD protein to AuNPs, the solution changed from purple-red to colorless upon the addition of 100 μg CBD to 1 mL of AuNPs. Additionally, no significant 530 nm or 533 nm absorption peak was observed in the spectrum. These observations suggest that the conjugation of CBD protein to AuNPs was failed (Figure 6D).
We consulted the literature and found that the pH of the solution affects the degree of protonation/deprotonation of the chemical groups (i.e. amines and carboxyl groups) that confer electrostatic charge on proteins and AuNPs [1]. The pH of the buffer is critical because it must be above the isoelectric point of the protein so that the protein can displace the negatively charged citrate groups adsorbed on the surface of the AuNPs to form the AuNP@PBP [2]. Thus, we used an online tool (https://web.expasy.org/protparam/) to predict the theoretical isoelectric points (pI) of the three proteins. We found that the theoretical pI of TFP (gp17) and TSP (Tailspike) proteins are 5.74 and 5.0, respectively, which are smaller than pH 8.0, and thus can bind well to AuNP. On the other hand, the theoretical pI of CBD (plyV12) was 9.84, which was larger than pH 8.0, which might lead to the failure of AuNP@CBD preparation as CBD could not bind well to AuNP.
Figure 6. UV-Vis Absorption Spectra of AuNPs Conjugated with PBP
UV-Vis absorption spectrum of synthesized AuNPs and AuNPs after conjugation with (A) pET28a empty vector protein (B) TFP protein (C) TFP protein showing a shift in the absorption peak from 530 nm to 533 nm, indicating successful conjugation. (D) UV-Vis absorption spectrum of AuNPs after attempted conjugation with plyV12 protein, no significant absorption peak was observed, indicating unsuccessful conjugation.

To evaluate the ability of AuNPs conjugated with PBP proteins (AuNP@PBP) to detect specific bacteria, we incubated the AuNP@PBP with their respective target bacteria at room temperature for 30 min. The specific conjugates and corresponding bacteria were as follows: AuNP@TFP for E. coli detection, AuNP@TSP for Salmonella detection, and AuNP@CBD for S. aureus detection. After incubation, the supernatant was collected and analyzed using smartphone photography.
When AuNP@pET28a (empty vector control) was incubated with all three types of bacteria, the supernatant remained purple-red, indicating that AuNP@pET28a did not bind to the bacteria (Figure 7A). For AuNP@TFP and AuNP@TSP, incubation with varying concentrations of E. coli and Salmonella, respectively, as the bacterial concentration increased from (10^3) to (10^9) CFU/mL, the supernatant color changed from purple-red to colorless (Figure 7B, C). This indicates that AuNP@TFP and AuNP@TSP successfully bound to their respective target bacteria. Since the preparation of AuNP@CBD was unsuccessful, incubation with S. aureus did not result in a significant color change in the supernatant, which remained almost unchanged (Figure 7D).

Figure 7. Colorimetric Analysis of Bacteria Detection Using AuNP@PBP Conjugates
(A) Smartphone images of the supernatant after incubation with AuNP@pET28a (control) and E. coli, Salmonella, and S. aureus. (B) Smartphone images of the supernatant after incubation with AuNP@TFP and varying concentrations of E. coli ((10^3) to (10^9) CFU/mL). (C) Smartphone images of the supernatant after incubation with AuNP@TSP and varying concentrations of Salmonella ((10^3) to (10^9) CFU/mL). (D) Smartphone images of the supernatant after incubation with AuNP@CBD and S. aureus.
Subsequently, we performed an RGB colour analysis of the results of AuNPs@TFP and AuNPs@TSP, and found a correspondence between the increase in bacterial concentration and the corresponding B/R values. As a result, we obtained a standard curve between the logarithmic values of bacterial concentration and B/R. (Figure 8) For AuNP@TFP, the standard curve was y=0.0125x+1.0404 (R2=0.9583), and for AuNP@TSP, the standard curve was y=0.00098x+1.0684 (R2=0.9804).

Figure 8. Standard curve between the logarithmic values of bacterial concentration and B/R
(A) AuNP@TFP detection of E. coli; (B) AuNP@TSP detection of Salmonella
In order to detect bacterial concentration by smartphone, we built the standard curve obtained from the experiment into the smartphone APP. After taking a picture to record the incubation supernatant of bacteria with AuNP@TFP or AuNP@TSP, we uploaded the picture to the APP, which will automatically read the RGB value of the picture and calculate the B/R value, and finally show the corresponding bacterial concentration according to the standard curve (Figure 9).

Figure 9. Bacteria detection results based on smartphone APP

AuNPs have significant near-infrared absorption capabilities that enable them to conduct light energy into heat and make them promising candidates for targeted photothermal therapies [3]. After the above experiments, we successfully prepared novel colorimetric biosensors, including AuNPs@TFP and AuNPs@TSP. After detecting the bacteria, we irradiated the sensors with near-infrared light at a wavelength of 808 nm with a power of 2 W.
The experimental results showed that the temperatures of AuNPs@TFP and AuNPs@TSP gradually increased (28℃-45℃) with the increase of irradiation time (0-10 min). The treated samples were subjected to plate-coating operation and after incubation overnight, we found that the number of bacteria decreased, indicating the effectiveness of photothermal sterilization (Figure 10).
However, after 10 min of NIR light irradiation, there were still a few colonies growing, and in the subsequent experiments, we could adjust the distance between the NIR light and the sample, which allowed the temperature to be further increased for better bactericidal effect.
Figure 10. Photothermal sterilisation results of (A)AuNP@TFP and (B)AuNP@TSP

In this project, we identified three phage-derived bacterial-binding proteins (PBPs) that specifically bind to E. coli, Salmonella, and S. aureus. We successfully constructed plasmids containing the corresponding protein genes and expressed and purified these proteins in BL21(DE3) cells. Concurrently, we synthesized gold nanoparticles (AuNPs) and conjugated the PBPs to the AuNPs through electrostatic adsorption. Analysis of the absorption peaks indicated that AuNPs@TFP and AuNPs@TSP were successfully prepared, while AuNPs@CBD was not. The successfully prepared AuNPs@TFP and AuNPs@TSP demonstrated strong binding affinity to E. coli and Salmonella, respectively, causing a color change in the detection system. This biosensor allows for rapid colorimetric detection of these bacteria within 30 minutes, providing a promising approach for bacterial detection in the food industry.

In this experiment, the preparation of AuNPs@CBD was unsuccessful, likely due to an inappropriate conjugation method. Due to time constraints, we were unable to explore alternative conjugation methods. Therefore, in future work, we will attempt various conjugation strategies to ensure efficient binding of the protein to AuNPs. Additionally, we will optimize the amount of AuNPs@PBP added to bacterial samples to enhance detection efficiency. We also plan to compare this method with other existing bacterial detection methods to confirm its accuracy and reliability. This will help in establishing a robust and efficient biosensor for bacterial detection in various applications, particularly in the food industry.

[1] Capocefalo A, Bizien T, Sennato S, Ghofraniha N, Bordi F, Brasili F. Responsivity of Fractal Nanoparticle Assemblies to Multiple Stimuli: Structural Insights on the Modulation of the Optical Properties. Nanomaterials (Basel). 2022;12(9):1529.
[2] Waldeisen JR, Wang T, Ross BM, Lee LP. Disassembly of a core-satellite nanoassembled substrate for colorimetric biomolecular detection. ACS Nano. 2011; 5(7):5383-9.
[3] Addae E, Dong X, McCoy E, Yang C, Chen W, Yang L. Investigation of antimicrobial activity of photothermal therapeutic gold/copper sulfide core/shell nanoparticles to bacterial spores and cells. J. Biol. Eng. 2014; 8: 11.