In our research, we aim to investigate the potential of the lab-engineered protein cage Mi3 ^[1] as a versatile diagnostic tool for rapid detection of various disease biomarkers, inflammatory factors, and environmental pollutants. By fusing SpyCatcher-Mi3 subunits with conjugated tags, they can easily assemble with cargo proteins that carry corresponding tags. During the expression process, the 60 SpyCatcher-Mi3 subunits spontaneously self-assemble into a dodecahedral SpyCatcher-Mi3 protein cage. Utilizing the 60 sites on this protein cage, we incorporate a large number of SpyTag-eGFP fluorescent reporter molecules to illuminate the cage. At the same time, by introducing a specific ratio of specific antibody molecules, we utilize immunorecognition to capture biomarkers, enabling efficient detection of specific biological molecules.

However, the choice of which bioconjugation system to use, the selection of the right 'adapters' to achieve the 'plug-and-play' product positioning, and the adjustment of the assembly ratio between SpyTag-eGFP and 'adapter' proteins all remain critical questions that require urgent solutions. We are, therefore, committed to an engineering cycle of "research-design-build-test-learn" to discover the best conditions for building a modular detection system.

1.1 RESEARCH

To identify a biologically conjugation system that is both easily preparable and structurally stable for building the entire device, we conducted a thorough literature review and tested three different conjugation systems: SpyCatcher-SpyTag ^[2], SnoopCatcher-SnoopTag^[3], and DogCatcher-DogTag^[4]. Our goal was to determine the most appropriate conjugation tag for merging with Mi3, which would enable the efficient loading of cargo proteins.

1.2 DESIGN

Informed of the amino acid sequences of Mi3, SpyCatcher, SnoopCatcher, and DogCatcher, we initiated the design and construction of cloning strategies for three recombinant plasmids. Following a thorough review of the literature, we decided to utilize the widely-used commercial plasmid, pET28a, for creating the recombinant plasmids: pET28a-SpyCatcher-Mi3, pET28a-SnoopCatcher-Mi3, and pET28a-DogCatcher-Mi3. The plasmid design is illustrated in Figure 2.

1.3 BUILD

As shown in Figure 3a, we successfully obtained the target fragment of plasmid pET28a-SnoopCatcher-Mi3, which measures 386 bp, through PCR amplification using the upstream and downstream primers. In Figure 3b, we employed PCR to amplify the target fragment containing DogCatcher from the plasmid pET28a-DogCatcher-Mi3, and it measured 367 bp. In Figure 3c, through PCR, we amplified the target fragment containing SpyCatcher-Mi3 from the plasmid pET28a-SpyCatcher-Mi3, and it measured 1011 bp. These outcomes confirm the successful construction of the recombinant protein plasmids.

1.4 TEST

To test the success of cloning, we transfected the plasmids into Escherichia coli BL21 (DE3) receptor cells. Subsequently, the transformed cells were seeded onto LB agar plate containing kanamycin and incubated at 37°C for 12-16 hours. On the following day, uniform bacterial colonies were observed on the agar plate. Colonies were selected and subjected to colony PCR, followed by bacterial strain preservation.

We extracted the plasmid and sent it to Sangon Biotech for sequencing, obtaining the expected sequencing results. After activating the preserved bacterial strains and enlarging the culture, we induced them at 22°C using 1 mM IPTG for 12-14 hours. We then centrifuged the bacterial culture at 10,000 rpm, 4°C for 5 minutes to obtain more cells. After high-pressure homogenization and further centrifugation at 10,000 rpm, 4°C for 30 minutes, we obtained the supernatant of cell lysis. The supernatant, cell lysate pellet, and whole-cell samples from strains A, B, and C were individually subjected to SDS-PAGE. The electrophoresis results depicted in Figures 4, 5, and 6 reveal that only SpyCatcher-Mi3 exhibited soluble expression, at an approximate concentration of 1 mg/mL. In contrast, SnoopCatcher-Mi3 forms a significant amount of inclusion bodies during the expression process, while DogCatcher-Mi3 is expressed insolubly. Both situations pose significant challenges for protein purification.

Therefore, in this project, we chose to utilize the SpyCatcher-SpyTag system to facilitate the conjugation of the target protein with Mi3. Furthermore, we combined SpyTag-eGFP with SpyTag, resulting in the recombinant plasmid, pET28a-SpyTag-eGFP, which is illustrated in Figure 7. Following this, we effectively purified the SpyTag-eGFP fusion protein, as shown in Figure 8.

Following that, we combined SpyCatcher-Mi3 and SpyTag-eGFP at various ratios. As depicted in Figure 9, we conducted grayscale analysis on the electrophoresis gel images using ImageJ. This analysis showed that the highest assembly efficiency, which reached 63.57%, was attained when the input ratio of SpyTag-eGFP to SpyCatcher-Mi3 was 3:2, as indicated in Table 1.

In conclusion, this project utilized the SpyCatcher-SpyTag bioconjugation system, known for enabling soluble expression without the formation of inclusion bodies, to connect SpyTag-eGFP with the protein cage Mi3. The highest assembly efficiency was attained when the input ratio between these two components was set at 3:2.

1.5 LEARN

Based on our research, in the initial stages of the project, we investigated three common bioconjugation systems: SpyCatcher-SpyTag, SnoopCatcher-SnoopTag, and DogCatcher-DogTag.However, we encountered issues when fusing the latter two with Mi3, leading to significant challenges in protein expression and purification. SnoopCatcher-Mi3 resulted in the formation of numerous inclusion bodies during expression, and DogCatcher-Mi3 exhibited insoluble expression. Only SpyCatcher-Mi3 demonstrated soluble expression without the formation of inclusion bodies. Consequently, we opted to utilize the SpyCatcher-SpyTag bioconjugation system for the entire system's construction. We generated the SpyTag-eGFP recombinant protein and assembled it with SpyCatcher-Mi3. Grayscale scanning analysis revealed that the highest assembly efficiency was achieved with a 3:2 input ratio, reaching 63.57%.

2.1 RESEARCH

In the second step, to attach the detection antibodies to the protein cage, we considered two strategies. The first is utilizing coiled coil ^[5] to connect the antibodies by wrapping them with small helical structures. The second is using Streptavidin ^[6] to connect them through interaction with biotin.

To determine which approach is more suitable, we conducted simulated experiments.

According to the RMSD results, we observed that the conformation of the Streptavidin component remained consistently stable, whereas the conformation of the coiled coil component fluctuated within a range of 0-0.35 nm for most of the time, indicating relative instability.

The analysis of RMSF data reveals that the atoms in the coiled coil component generally exhibit higher flexibility compared to the atoms in the Streptavidin component.Furthermore, it indicates that the coiled coil component displays excessive flexibility, which may lead to significant conformational changes before binding to another helix. We speculate that this could potentially reduce the affinity between the two helices.

The gyration radius also demonstrates that the Streptavidin component has a higher compactness compared to the coiled coil component, indicating greater stability. Therefore, we have chosen Spy Tag-Streptavidin as our connector element.

2.2 DESIGN

We selected pET28a as the vector plasmid, becausethe strong T7 promoter on the vector ensures the efficient expression of the SpyTag-Streptavidin target protein， while the 6xHis tags at both sides of the MCS region serve as protein purification tags.

To determine whether to attach SpyTag at the N-terminal or C-terminal, we used the complex module in Colabfold to simulate the assembly of SpyCatcher with SpyTag-Streptavidin in the condition of SpyTag in the N and C-terminal respectively. In both cases, we observed that they maintained favorable conformations. Consequently, we chose N-terminal and simulated the assembly in a water environment at 25°C for 50 ns. The final analysis of RMSD, RMSF, and Rg showed relatively small values, suggesting that the stability of this conformation meets our expectations. Subsequently, we proceeded with plasmid construction.

2.3 BUILD

We have commissioned GenScript to construct the pET28a-Streptavidin-SpyTag plasmid and subsequently transformed it into Escherichia coli BL21 (DE3) host cells. After confirming the correctness of the sequenced plasmid, we induced protein expression in E. coli by adding 1mM IPTG at 22°C for 10 hours.

2.4 TEST

The structure of SpyTag-Streptavidin is highly stable and shows minimal sensitivity to denaturation by SDS. It requires boiling at 100°C for more than 20 minutes to achieve complete denaturation. Therefore, when we conducted tests on our samples without boiling, we observed bands at approximately 88 kDa. While test in denaturing gel electrophoresis, in addition to bands around 88 kDa, bands at around 22 kDa also appeared, and these bands intensified with prolonged boiling time. This electrophoretic pattern is consistent with the tendency of SpyTag-Streptavidin to tetramerize. Under the influence of high temperature and denaturing agents, the tetramers partially dissociate into monomers, and the degree of dissociation increases with longer boiling times.

Based on these observations, we can preliminarily confirm that the purified protein is SpyTag-Streptavidin.

To further validate the biological activity of SpyTag-Streptavidin, we took advantage of its specific binding to biotin and conducted fluorescence characterization by adding biotin-PEG-FITC to the protein samples.

Analyzing the data above, SpyTag-Streptavidin, in the presence of Biotin-PEG-FITC, exhibits a significant fluorescence signal. In comparison to its fluorescence intensity values, the fluorescence from SpyTag-Streptavidin and blank control can be omitted.

These two aspects proved that the protein is indeed SpyTag-Streptavidin and demonstrates its biological function of binding to Biotin.

To verify the assembly efficiency, we mixed SpyCatcher-Mi3 and SpyTag-Streptavidin in a 1:1 ratio, added Biotin-PEG-FITC, and assessed FITC fluorescence using NFC as a representation.

Examining the NFC data, the assembled sample exhibits a significantly higher ratio of fluorescence labeling compared to the unassembled one, indicating a successful assembly.

Validation of Expression Levels: We observed a higher abundance of inclusion bodies upon induction of expression at 22°C.

Based on the grayscale analysis data, it is evident that nearly 80% of the protein samples are present in the sediment. This shows the challenge of substantial inclusion bodies formation during the purification of SpyTag-Strepavidin.

2.5 LEARN

To analyze the issue of SpyTag-Streptavidin forming inclusion bodies during the purification process, we hypothesize that inducing at 22°C might result in the protein folding too rapidly, leading to misfolding and the formation of inclusion bodies. As a result, we have reduced the expression temperature to 16°C.

By comparing and analyzing the average grayscale values under induction conditions at 16°C and 22°C, we have observed a significant reduction in the formation of inclusion bodies when the induction temperature was lowered. However, complete prevention of inclusion body formation was not achieved, which to some extent, still affected the final protein yield.

3.1 RESEARCH

To attain the utmost signal amplification ratio, it is imperative to enable the protein cage to accommodate a maximal quantity of SpyTag-eGFP. Simultaneously, a proportionate presence of SpyTag-Streptavidin is essential to ensure the overall cage's proficiency in antigen capture. Hence, we optimized of the assembly ratio of these two proteins, aiming for the best detection efficacy.

3.2 DESIGN & BUILD

We initially conducted assembly experiments with SpyCatcher-Mi3 and SpyTag-eGFP to explore which assembly ratio would yield the highest fluorescence signal. Based on the preliminary investigation and the fluorescence standard curve generated, we identified the assembly ratio with the highest fluorescence value and proceeded with further detailed exploration of three protein assembly ratios in that vicinity.

3.3 TEST

We characterize the assembly of these three proteins in our samples using SDS-PAGE and a nanoscale flow cytometer.

(1) Fluorescence Characterization: We established six distinct assembly ratios by gradually increasing the SpyTag-eGFP units. After assembly, we conducted extensive dialysis to eliminate any unbound proteins.

Note: Before analysis with the nanoscale flow cytometry, it is imperative to dilute the samples to ensure that the concentration of the assembled samples remains approximately uniform.

(2) Electrophoresis Characterization: SDS-PAGE characterization of the assembly of SpyCatcher-Mi3 and SpyTag-eGFP samples has been detailed in the first module.

Additionally, due to the large size of the SA-Mi3 subunit and its partial denaturation even under denaturing electrophoresis conditions, resulting in broader bands, electrophoresis is no longer suitable as a method for characterizing the assembly of the MI3-eGFP-SA triad of proteins.

(3) Nanoscale flow cytometry Characterization: We have learned that the assembly of SpyCatcher-Mi3 subunits with SpyTag-eGFP at a ratio of approximate 6:5 results in the highest fluorescence signal (as shown in Figure 21). To further investigate this, we set up three different ratios of 60 : 5, 60 : 10, and 60 : 15 (SpyCatcher-Mi3 : SpyTag-Streptavidin). Biotin-PEG-FITC was added to the system, followed by thorough dialysis, and nanoflow cytometry was performed to obtain the following results.

In contrast to the results from Control Loop 1, where the optimal assembly ratio for SpyCatcher-Mi3 subunits to SpyTag-eGFP was approximately 6:5 with the highest assembly efficiency but not the highest fluorescence intensity, it suggests that when optimizing assembly conditions, we need to consider a combination of various factors depending on the intended use.

As more SpyTag-Streptavidin-FITC is added to the system, the fluorescence intensity on the protein cage increases. This indicates that at this assembly ratio, SpyCatcher-Mi3 can successfully assemble with SpyTag-Streptavidin, and the fluorescence is positively correlated with the added SA concentration.

Following this, we referred to the previous results and set up two ratios: SpyCatcher-Mi3 subunit to SpyTag-eGFP to SpyTag-Streptavidin, which are 60:55:5 and 60:50:10, to explore which ratio yields the maximum fluorescence intensity.

3.4 LEARN

When SpyCatcher-Mi3, SpyTag-eGFP, and SpyTag-Streptavidin are assembled in a ratio of 60:50:10, the fluorescence intensity is the highest and relatively uniform. However, when the ratio is 60:55:5, the fluorescence intensity is slightly lower than the former, and the fluorescence is more dispersed.

However, we theorized that it might form a concatenate aggregate such as MI3-SA-MI3-SA-... To verify the existence of possible aggregates, we centrifuged the assembly samples at 14,000 rpm, 4°C for 10 minutes. We observed a clear green precipitate at the bottom of the centrifuge tube. By using the BCA method, we measured a protein aggregation rate of 38.7%. (To rule out aggregation due to sample instability, we also assembled samples containing 50 SpyTag-eGFP but without SA and Mi3. After assembly, we centrifuged them at 14,000 rpm, 4°C for 10 minutes and observed and measured the aggregation.

DLS measurements of the particle size of the assembled samples also provided data of around 1000nm (with a hydration layer), confirming our hypothesis of the presence of aggregates. We are not sure about the extent of this serial aggregation phenomenon, but it is certain that the assembled Mi3-eGFP-SA sample is a heterogeneous system, containing particles of different sizes assembled with SpyTag-eGFP and SA. The majority of the sample is assembled in the ratio we need.

This aggregation phenomenon may adversely affects the results of experiments such as microcalorimetry. Therefore, special attention was paid to the pretreatment of assembled samples in subsequent experiments related to affinity validation and measurement

4.1 RESEARCH

To validate our monitoring system, it's crucial to determine the capture efficiency of antibodies loaded on the protein cage for the target antigen. However, the concentrations of our samples are relatively low, and accurate measurement of protein samples demands a high-quality detection environment. Therefore, we chose microcalorimetryto assess the index parameter.

Microcalorimetry enables precise evaluation of binding interactions and capture efficiency between the antibodies and the target antigen, even when dealing with low-concentration samples. This highly sensitive technique measures the heat changes associated with biochemical reactions, making it well-suited for our purpose.

4.2 TEST

(1) Following high-speed centrifugation and treatment with Triton-20 (final concentration of 0.05%), the assembled samples displayed no signs of aggregation or adsorption, and they exhibited uniform fluorescence

The data presented above confirm the effectiveness of the previously established sample pre-processing method, which greatly aided in subsequent experiments. The substantial overlap of MST curves also indirectly suggests that the assembled samples to be tested exhibit uniform fluorescence, thus facilitating the acquisition of authentive data.

(2) The steps and precautions for microcalorimetry experiments Building upon the optimal assembly ratio confirmed in the third engineering cycle, we established a modular protein system with a ratio of SpyCatcher-Mi3 subunits to SpyTag-eGFP to SpyTag-StrepStreptavidin as 60:50:10. For the validation of the system's affinity, we selected the commonly used BSA antigen and antibody. To ensure the reliability of the experimental data, we established 16 points for affinity determination via gradient dilution.

4.3 LEARN

Through ITC verification, we have drawn the following two conclusions about the protein system we constructed:

(1) MST curves exhibit a strong degree of overlap, signifying uniform fluorescence in the assembled Target sample (Mi3-eGFP-SA-IgG).

(2) The affinity constant (Kd) between IgG on the assembled protein cage and BSA is 53.7 nM, slightly higher than the typical BSA and BSA antibody interaction. The uniformity of the fluorescence signal and the elevated affinity constant both suggested that the protein system we constructed efficiently captured the target antigen and generated a stable fluorescence signal strength. This provided data support for our subsequent applications in semi-quantitative paper-based testing.

1. W. Kang, X. Ma, D. Kakarla, H. Zhang, Y. Fang, B. Chen, K. Zhu, D. Zheng, Z. Wu, B. Li, C. Xue, Angew. Chem. Int. Ed. 2022, 61, e202214001; Angew. Chem. 2022, 134, e202214001
2. ZAKERI B, FIERER J O, CELIK E, et al. Peptide tag forming a rapid covalent bond to a protein, through engineering a bacterial adhesin [J]. P Natl Acad Sci USA, 2012, 109(12): E690-E7.
3. VEGGIANI G, NAKAMURA T, BRENNER M D, et al. Programmable polyproteams built using twin peptide superglues [J]. P Natl Acad Sci USA, 2016, 113(5): 1202-7.
4. Keeble AH, Yadav VK, Ferla MP, Bauer CC, Chuntharpursat-Bon E, Huang J, Bon RS, Howarth M. DogCatcher allows loop-friendly protein-protein ligation. Cell Chem Biol. 2022 Feb 17;29(2):339-350.e10. doi: 10.1016/j.chembiol.2021.07.005. Epub 2021 Jul 28. PMID: 34324879; PMCID: PMC8878318.
5. Woolfson DN. Understanding a protein fold: The physics, chemistry, and biology of α-helical coiled coils. J Biol Chem. 2023 Apr;299(4):104579. doi: 10.1016/j.jbc.2023.104579. Epub 2023 Mar 5. PMID: 36871758; PMCID: PMC10124910.
6. Fogen D, Wu SC, Ng KK, Wong SL. Engineering Streptavidin and a Streptavidin-Binding Peptide with Infinite Binding Affinity and Reversible Binding Capability: Purification of a Tagged Recombinant Protein to High Purity via Affinity-Driven Thiol Coupling. PLoS One. 2015 Sep 25;10(9):e0139137. doi: 10.1371/journal.pone.0139137. PMID: 26406477; PMCID: PMC4583386.