In our project, we aim to establish a standardized interface for protein modification on silicon dioxide (SiO2) surfaces. To achieve this, we intend to modify the LPG adapter protein. We have chosen a silica-binding peptide for binding to SiO2 and will utilize the avidin-biotin system, a powerful affinity-based coupling system, as the bridge between the interface and the target protein. We chose streptavidin from the avidin family.

1. SBP affinity selection:

We conducted a literature review on the SBP sequence and found that by adding a polyhistidine tag at the C-terminus of SBP, with isoleucine residues interspersed, the polyhistidine side chains will form an imidazole plane that interacts more strongly with silicon dioxide. Among the SBP variants we searched, SBP5 was ultimately selected due to its high affinity and relative stability in the intended environment.

2. Transformation of SBP oligomerization:

The affinity of SBP is influenced by steric hindrance between the SBP sequence and the avidin protein. Excessive steric hindrance can greatly reduce the binding strength. To enable the SBP to connect with avidin, it is necessary to introduce a linker sequence between SBP and avidin. In our study, we selected streptavidin, a tetrameric protein with a size of 66 kDa, from the avidin family. We chose a flexible linker sequence (GGGGS)*12 to provide multidirectional extension in space, thereby improving the efficiency of streptavidin and biotin binding.

Based on the design, we combined different components and performed codon optimization, resulting in the construction of a plasmid consisting of SBP，linker and streptavidin in the desired format.

We conducted dry lab validation and utilized the Alpha Fold 2 software for protein modeling to obtain the protein model of BS (as shown in the Figure 3). It was observed that the SBP sequence, to some extent, becomes entangled with streptavidin, resulting in the inability of the streptavidin domain to be exposed and the SBP unable to be displayed.

Due to the flexible nature of the linker we chose, it could potentially fold and wrap around the streptavidin, causing cross-interference among protein domains and preventing effective exposure, hence compromising the intended functionality. To maintain the relative positions of the structural domains, it is recommended to replace the flexible linker with a rigid one. This modification can help alleviate the issue and enhance the overall performance of the construct.

We conducted dry lab validation and utilized the Alpha Fold 2 software for protein modeling to obtain the protein model of BS (as shown in the Figure 3). It was observed that the SBP sequence, to some extent, becomes entangled with streptavidin, resulting in the inability of the streptavidin domain to be exposed and the SBP unable to be displayed.

Based on the design, we combined different components and performed codon optimization, resulting in the construction of a plasmid consisting of SBP，rigid linker and streptavidin in the desired format.

After modifying to a rigid linker, we are now able to clearly visualize the different domains of Basic Silinker. The rigid linker has successfully separated the mSA and SBP fragments.

However, we have observed that the target bands in the supernatant appear to be faint, indicating a low content of Basic Silinker. Due to misfolding, a significant amount of BS protein aggregates and forms inclusion bodies in the bacterial sediment. This poses a significant challenge for our protein expression, as it implies the need to scale up the culture in order to obtain sufficient amounts of BS protein. This is clearly unwelcome news for production purposes.

The second lane is supernatant, and the fourth lane is bacterial preposition, and it can be seen that there are a large number of inclusion bodies in the bacterial preposition.

After reviewing the literature, we found that mSA is difficult to express in E. coli [1], because mSA expressed in E. coli was insoluble which lead to form plenty of inclusion bodies[2], and it needed 7-10 days to be completely purified. This process may increase the total production cost and result in limited total yield, which made it impossible to purify sufficient protein linkers.

After reviewing the literature, it has been found that the addition of expression tags can enhance the soluble expression of mSA and increase protein expression. We found four solubilization tags that have been tested in previous studies and ultimately decided to add the thioredoxin (trx) sequence to promote the soluble expression of mSA[3].

We added the trx tag upstream of the pathway and synthesized the plasmid. After conducting gene sequencing at company, we successfully synthesized the plasmid and successfully transferred it into our engineered strain, E.coli BL21(DE3).

We conducted large-scale induction expression of the target protein using 1mM IPTG and achieved a significant yield after 16 hours of induction at 16°C. In the process of plasmid construction, we successfully incorporated a His tag into the target protein and subsequently purified it using nickel affinity chromatography based on the specific binding of the His tag. As shown in the figure below, a significant band of the target protein was successfully eluted using 400mM imidazole. Although there is still the formation of inclusion bodies, it still meets the requirements for subsequent experiments.

In order to confirm whether Basic Silinker can successfully function in the modification of functional proteins onto the surface of silica dioxide, we used a mutant variant of green fluorescent protein (eGFP) as a simulation for functional protein modification, providing a visual representation of the connection results. We deposited both the Basic Silinker-modified eGFP (BS-eGFP) and the unmodified eGFP onto a microscope slide.

The Basic Silinker only serves to attach proteins to the silica surface, but in practice and especially in the main application of mesoporous silica materials （e.g. in vivo drug release） , proteins on the surface of mesoporous silica need to be released in the presence of some specific enzymes to accomplish the process[4].We attempted to further design a Silinker that could release the target proteins, the Cut Silinker. We envisioned a protease slice segment in the middle of the SBP and streptavidin of the Basic Silinker, where a specific protease could cleave the site to disassociate the binding protein from the Silinker and complete the release process.

For this design, we found a short cleavable peptide (PLGVR) as an enzymatic cleavage site, which can be specifically cleaved by matrix metalloproteinase 2 (MMP2) to construct a plasmid in the form of SBP+PLGVR+mSA.

In the further design process, through literature review and rational deduction, we found that different proteases have different or similar site specificity[5]. While there are too many types of proteases, designing a corresponding Silinker for each protease based on its specificity site would lead to a lack of standardization in the design of Silinkers, which is contrary to our original intention.

We referenced the iGEM project of Pittsburgh 2019 to screen two sets of trans intronic peptides and use intronic peptide self-assembly to achieve modular substitutions at different enzymatic cut sites.

Initially, we believed that we could release surface-bound proteins by using specific proteases to cut the Cut Silinker at its corresponding enzymatic cleavage site. This means that for different enzymes, we need to redesign different fragments within Silinker, which requires the redesign of different Silinker proteins and the exploration of protein expression conditions in order to achieve the desired outcome. Undoubtedly, this is a significant undertaking. However, this method did not effectively achieve our goal of a unified protein junction. To address this, we tried to unify and modularize the Cut Silinker by incorporating an intronic peptide at both ends of the sequence for post-translational modification.

We devised a gene sequence structure consisting of the "N-terminal splice region of intein + extein sequence + C-terminal splice region of intein" During translation, we removed the matching N-terminal and C-terminal inteins from precursor proteins and linked the two extein sequences to form a mature protein [6]. This design separates the enzyme cleavage site in the Silinker as a distinct element. It allows for flexible connection of any desired enzyme cleavage site between SBP and streptavidin, while the self-scissoring connection of the two pairs of inteins further enables standardization and modularization.

For the specific selection of inteins, we finally selected gp41-1, NrdJ-1 (BBa_K3308007). The naturally fragmented gp41-1 intein is the result of macrogenome sequencing. It is one of the smallest reported fragmentation inteins with very fast trans-splicing activity, and it consists of an 88-residue N-terminal (IntN) and 37-residue C-terminal fragment. Its small size and strong protein splicing activity make it an attractive template for protein engineering[7].

Based on the design, we combined different components and performed codon optimization, resulting in three segragated parts of original pathway.

We performed secretory expression experiments of three small proteins with the following results:

In view of the difficulty of purification, we predicted the structure of the three parts of the cuttable silinker protein by alphafold2 and analyzed it.

For the structure of CS1, mSA was connected to the N-terminus of the intein GP41 through the flexible linker of (GSSSS)12. It can be seen that the structure of the mSA in the front part is relatively stable, showing a barrel-like structure, and there exists a certain distance from the intein, which is not easy to have an effect on the purification; however, the key His tag of the purified part and the trxA of the auxiliary purification are stacked on top of each other, and the overall structure may make it not easy to be better adhered when passing through the Ni-EDTA column. However, from a structural point of view, the relatively stable site of the intronic peptide and the relatively high variability may be better for the attachment reaction of the intein.

Similarly, Cut Silinker protein 2 presents an enveloped structural state, where the two parts of the inteins overlap with each other wrapping around the possible enzyme cleavage site, and also possibly interfering with the ligand reaction, which is taken into account during the modeling process (see the modeling section for more details), but the His tag used for purification is relatively outwardly oriented, which should theoretically be advantageous for the purification, but it still has a some difficulty and may be influenced by other factors.

For Cut Silinker protein 3, the structure prediction is relatively low confidence, which is due to the fact that the SBP we chose belongs to the flexible structure itself, and its properties are subjected to a relatively large concentration of solution ions. In the actual experimental process, unfortunately we were not able to develop a solution favorable for its purification, which has a lot to do with our failure in purifying the process. However, we have taken into account the variability of this factor in the modeling process, so that the variability of its interaction in the linkage reaction and contact with SiO2 has been simulated.

In cellular processes, protein dimerization plays a crucial role in substance transformation, energy conversion, and signal transduction. Therefore, we aim to design Pairing Silinker proteins that can efficiently bind to silicon dioxide nanoparticles and achieve diverse functions through dimerization. Each Pairing Silinker will have specific ion-binding sites, allowing control of the distance between two proteins through changes in substance concentration.

Taking into account various application scenarios and the widespread presence of calcium ions in biological systems, which serve multiple functions as intracellular signaling ions, we have chosen calcium ion-binding peptides as functional components[8]. We will construct a structure that resembles "P1 ⋯ Ca2+ ⋯ P2" (non-covalent binding of calcium ions to peptides) to achieve the primary function of Pairing Silinker through dimerization.

We have incorporated a calcium ion-binding peptide sequence into the linker region of Silinker's peptide chain. This sequence will undergo dimerization in the presence of calcium ions. We intend to express this protein in the BL21(DE3) strain.

We will validate the feasibility of our design by expressing the protein in the BL21(DE3) strain and observing dimerization in the presence of calcium ions. This experimental verification will help determine the effectiveness of the design and pave the way for further optimization and application of Pairing Silinker to achieve various functions.

The experimental results indicate that the soluble expression of PS in BL21(DE3) is very low, and even under non-induced conditions, inclusion bodies are formed.

BL21(DE3) exhibits significant issues with leaky expression, resulting in the predominant presence of proteins in the form of inclusion bodies. We believe that reducing leaky expression will enhance the soluble expression of the protein.

The BL21(DE3)pLysS strain carries the pLysS plasmid, which allows for constitutive expression of T7 lysozyme, thereby reducing leaky expression caused by T7 RNA polymerase. We intend to use this strain to express the PS protein.

We attempted to transfer the previously constructed pETDuet-1-PS plasmid from the previous loop into the BL21(DE3)pLysS strain and express the PS protein in this strain.

On top of the Basic Silinker and simple Cut Silinker, we want to achieve the intelligence of Silinker. It can be changed by sensing the external environment, so that our linker can not only be controlled by human beings, but also can still occur in the structure after entering the working environment which is beyond the control of human beings.

In order to achieve this result, we actively searched for substances that can respond to changes in the outside world. We found **Calcium binding peptide (**CBP), It undergoes dimerization in the presence of calcium ions. Such dimerization property can help us to achieve intramolecular binding of CBP, a set of structural folds that pulls the attached target protein closer to the silica carrier.

Based on the design, we combined different components, resulting in the construction of a plasmid consisting of TrxA, His-tag, thrombin site, streptavidin, CBP, flexible linker, CBP and SBP.

We conducted dry lab validation and utilized the Alpha Fold 2 software for protein modeling to obtain the protein model of Twisted Silinker (as shown in the figure 2). Observe that the flexible linker is unable to provide sufficient support, causing the silica-binding peptide and streptavidin to be spatially close to each other, which may result in cross-interference and hinder normal functionality.

Understanding that the flexible linker is too soft and random to be predictable, and may entangle with CBP to prevent the structure from being properly stretched, we decided to change the sequence of the linker and replace it with a rigid one to ensure the normal function of the structural domain.

We changed the original flexible linker to a rigid linker (EAAAK sequence, rigid structure), in an attempt to minimize interferences between linker and CBP, so that CBP can be effectively dimerized to achieve targeted results.

Based on the design, we combined different components, resulting in the construction of a plasmid consisting of TrxA, His-tag, thrombin site, streptavidin, CBP, rigid linker, CBP and SBP.

But then in the process of reviewing the literature, we found a problem that was difficult to solve, so we skipped the part of the wet experiment and looked for a solution by reviewing the literature.

After reading more literature related to CBP, one of the drawbacks that we found to be detrimental to our project was that in the presence of widespread CBP, it is highly likely that CBP will undergo intermolecular dimerization rather than undergo intramolecular binding. This forced us to review the literature extensively and iterate our design in time to change to a more efficient calmodulin and calmodulin binding peptide

After reviewing the literature, we finally chose to change to a more efficient combination of calmodulin and calmodulin-binding peptide, which happens in the presence of calcium ions. Unlike calcium-binding peptides, calmodulin-binding peptides tend to bind intramolecularly and are more likely to achieve the target function.

Furthermore, through extensive reading, we have discovered the existence of optimized monomeric forms of avidin. We were pleasantly surprised to find that this monomeric streptavidin possesses several advantages such as high solubility, enhanced metabolic stability, reduced immunogenicity, and toxicity. These characteristics make it highly suitable for applications in drug delivery and other related fields. Consequently, in our overall design, we have replaced avidin with monomeric avidin.

Based on the design, we combined different components, resulting in the construction of a plasmid consisting of TrxA, His-tag, thrombin site, Monomeric Streptavidin, CBP, linker, calmodulin and SBP.

We conducted dry lab validation and utilized the Alpha Fold 2 software for protein modeling to obtain the protein model of Twisted Silinker (as shown in the figure 5).

However, in the subsequent experimental design for functional validation, we found that Twisted Silinker requires more complex procedures to confirm proper folding. In order to obtain concise and definitive evidence that Twisted Silinker is capable of exhibiting the desired functionality, we urgently need a method that directly indicates the correctly folded structure of Twisted Silinker.

In addtion to that, we conducted large-scale induction expression and successfully extracted and purified Twisted Silinked under the guidance of Basic Silinker’s experience. As shown in the figure below, a band of the target protein was successfully eluted using lysine solution.

We chose to return to the Parts Registry, which is filled with creativity and ingenious ideas, to find inspiration. When we came across the part from iGEM21_ShanghaiTech_China, we knew it was exactly what we were looking for. It has two EF-hands motifs, each EF-hands can catch two calciums. When calmodulin combines with four calciums, calmodulin will change into the activated state. Then, activated calmodulin will be able to attach to calmodulin-binding peptide. This process will make the conformation of cpGFP tighter, which lead to an obvious increase of emitted light at about 510nm. So cpGFP plays the role of reporter. As intracellular calcium concentration increases, more GCaMP6ms can combine with calciums. So the cell which has a higher calcium concentration in its cytoplasm will be brighter under a fluorescence microscope.

We brought in GCaMP6ms and designed Twisted Silinker - GFP, which could make it easier for us to verify the function of Twisted Silinker. In our expectation, Twisted Silinker - GFP can emit green fluorescence in calcium solution of a certain concentration, indicating our Twisted Silinker can fold properly and work well. At the same time, we chose to replace streptavidin with monomeric streptavidin in an attempt to avoid the interference that occurs with structural entanglement of streptavidin and calmodulin. According to the literature, monomeric streptavidin has the highest affinity for biotin among current monovalent streptavidins. We believed that this affnity can can effectively correct the resistance and interference encountered by streptavidin during the folding process of Twisted Silinker.

Based on the design, we combined different components, resulting in the construction of a plasmid consisting of TrxA, His-tag, thrombin site, Monomeric Streptavidin, GCaMP6m and SBP.

We conducted dry lab validation and utilized the Alpha Fold 2 software for protein modeling to obtain the protein model of Twisted Silinker - GFP (as shown in the figure 8).

We conducted large-scale induction expression and successfully extracted and purified Twisted Silinked - GFP. As shown in the figure below, a significant band of the target protein was successfully eluted using 400mM imidazole. Although there is still some impure protein and inclusion bodies, it still meets the requirements for subsequent experiments.

Firstly, we conducted pre-experiment to design appropriate gradients of both calcium chloride concentration and reaction time. Results of the preliminary experiment are shown below.

Based on the preliminary experimental results, the three different ratios of calcium ions to Twisted Silinker - GFP protein did not show significant and stable fluorescence signals.

In subsequent experiments using a larger system, we set the gradient of calcium ion concentrations to be 0.02M, 0.1M, and 0.5M, and the gradient of Twisted Silinker - GFP protein solution volumes to be 100μL, 150μL, and 200μL. The fluorescence emission of Twisted Silinker - GFP protein samples were incubated with calcium ion solution for 20s, 1min, and 15min was tested.

During the incubation, the values of the experimental group and the control group were similar, and the values of the experimental group did not show significant changes during testing. We considered variables such as reaction time, temperature, calcium ion concentration, and buffer solution in multiple controlled experiments. But unfortunately, we were unable to obtain the desired results. Therefore, it is regrettable that we cannot conclude that Twisted Silinker - GFP protein can fold in the presence of calcium ions.

Through the experiment, we found that Twisted Silinker - GFP did not folded correctly and then fluoresce in a solution with a certain concentration of calcium ions, as we expected. Considering the reaction time, temperature, calcium ion concentration, buffer and other variables, we carried out several control variable experiments, but failed to detect the ideal value.

We speculated that because the prokaryotic system is difficult to express calmodin with correct spatial structure, the originally extracted Twisted Silinker - GFP cannot fluoresce normally in calcium ion solution as expected. So in the future, we hope to replace the chassis organism, such as using yeast to express Twisted Silinker, and measure its fluorescence emission in solutions with different calcium ion concentrations.

During the development of CS, PS, and TS, we realized that our approach to enhancing Silinker with multiple functions conflicted with our initial goal of modularizing and standardizing protein linkers. Additionally, the continuous design, expression, and validation of new proteins proved to be a lengthy and costly process, limiting the industrial and commercial potential of Silinker. While Cutting Silinker suggested that internal peptides could be a useful tool to address this issue, the low programmability of proteins and their challenging direct synthesis remained Silinker's pain points.

Except for BS, none of the functionalities of PS, CS, and TS Silinkers met our expectations. This highlights that the design of protein linkers for signal response is influenced by numerous factors, many of which are largely unknown.

In contrast, nucleic acids, both DNA and RNA, can be rapidly and massively synthesized, and their strong programmability arises from complementary base-pairing. We believe that nucleic acids will serve as substitutes for proteins in the Silinker family, functioning in signal recognition and response.

As a result, building upon BS, we have developed a new generation of Silinker called Nucleotide Silinker, which relies on nucleic acids as functional linkers. By creating a fusion protein of mSA and TrwC enzyme, we aim to develop a Silinker capable of linking to any nucleic acid adapter.[9]

The amount of NS expression was very low, and we were not able to obtain sufficient and high purity NS protein. We guessed that NS might be toxic to the engineered bacteria based on the low cell concentration in culture.

Based on our experience with expressing PS, we attempted to express NS in the BL21 (DE3) pLysS strain. This strain carries the pLysS plasmid, which can produce T7 lysozyme, reducing leaky expression caused by T7 RNA polymerase and minimizing the potential toxicity of NS to the engineered bacteria.

We speculate that TrwC may suppress the resistance of the pLysS plasmid, leading to the death of transformants under antibiotic selection conditions.