Engineering

In order to immobilize sRAGE onto collagen-containing patches, we aimed to create a fusion protein that contains both sRAGE and CBD. In addition, we decide to keep the flexibility of this fusion protein in other application. Therefore, we designed the gene blocks containing sRAGE-EGFP-6xHis-Npu-DnaE^N (sRAGE-Int^N for short), Npu-DnaE^C-CWE-6xHis-CBD (Int^C-6xHis-CBD for short) and 6xHis-Npu-DnaE^C-CWE-CBD (6xHis-Int^C-CBD for short), because intein splicing not only enables the formation of the fusion protein but also provides greater flexibility for the use of CBD. At the early stage of the project, we planned to express the target proteins in large quantities using E. coli. However, after team experiments, we found that the expressed proteins are not soluble and folding properly. We then shifted to mammalian HEK293T cells. It was known that sRAGE is N-glycosylated in mammalian cells, and deletion of N-glycan increases the sRAGE affinity to AGE. Therefore, we ultimately decided to express the mutated sRAGE (N25Q and N81Q) -Int^N and CBD-Int^C in mammalian 293T cells.

To express our target protein, we designed four gene blocks and used the pET15b to construct our expressing plasmids. Each construct was then transformed into BL21 E. coli for protein expression.

Building constructs for E. coli

• Design & Build – Design and build constructs:
  The four gene blocks include sRAGE-Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD (Figure 1). The gene block Int^C-6xHis-CBD differs from 6xHis-Int^C-CBD in the way of order of intein and 6xHis tag, in case the 6xHis affects the intein splicing. The gene block EGFP-CBD serve as reporter for Dry Lab to detect the CBD anchorage on collagen layer.

  

  

  

  

• Test:
  After successfully assembling the constructs and amplifying the plasmids in DH5α, we transformed plasmids sRAGE-Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD into BL21 E. coli for protein expression. Initially, we induced the target proteins at 37°C with 1 mM IPTG for 2, 4, and 6 hours, but the target proteins were all found in the insoluble pellet. As shown in following figure 6xHis-Int^C-CBD proteins were all found in the insoluble pellet (Figure 2).

  

• Learn – The PBS based lysis buffer and sonication may not efficiently extract target protein:
  
  The decreases of induction temperature and IPTG concentration may increase the solubility of target protein by slowing down the protein production rate. Therefore, we decided to try different lysis buffers, and adjust the induction temperature and IPTG concentration.

Replaced lysis buffer & induced condition

• Design & Build – Prepare three different lysis buffers, lower the induction temperature, and reduce the concentration of IPTG:
  To ensure complete cell lysis, we used three different lysis buffers: PBS (1 mM PMSF, 0.1% Triton), PBS (1 mM PMSF, 0.5% Triton), and B-PER (1 mM EDTA) to lyse the cells. Additionally, we lowered the induction temperature to 20°C and reduced the IPTG concentration to ensure proper protein folding and normal expression.
• Test:
  The three different lysis buffers did not increase the solubility of induced protein. As shown in the Figure, sRAGE-Int^N was still found in the pellet (Figure 3-4).Decreasing the induction temperature to 20°C also did not enhance the solubility of target protein in the supernatant (Figure 5). Finally, we tested the protein induction by 0.5 mM, 0.25 mM, and 0.125 mM IPTG. As shown in figure, Int^C-6xHis-CBD remained in the pellet (Figure 6).

  

  

  

  

• Learn – The protein is trapped in the inclusion body:
  
  Based on attempts with different lysis buffers, we can rule out the possibility of incomplete cell lysis. Regardless of whether we lowered the induction temperature or reduced the IPTG concentration, the target protein was still found in the inclusion bodies. Therefore, we decided to refer to the relevant literature on inteins to explore other methods that might allow the target protein to be released from the inclusion bodies into the supernatant during lysis.

Urea denatured protein

• Design & Build – Denature proteins using 8M urea:
  To solubilize the target protein, we referred to the experimental procedures from the literature (Zettler, Schütz, & Mootz, 2009) and used the same lysis buffer as in the paper, buffer A-urea (50 mM Tris/HCl at pH 8.0, 300 mM NaCl, 8 M urea, 20 mM imidazole).
• Test:
  Showing sRAGE-Int^N as an example, 8M urea buffer could solubilize target proteins (Figure 7).

  

• Learn – The effect of 8M urea is significant; the proteins are all in the supernatant, allowing for protein purification:
  
  Buffer A-urea dissolves the proteins in the inclusion bodies into the supernatant, including the target protein, allowing the experiment to proceed toward protein purification.

Protein purification

• Design & Build – Using Ni-IDA resin for small-scale protein purification:
  Since all the gene blocks are designed with a His tag, the target protein can be easily purified using immobilized nickel ions. Therefore, we first performed a small-scale protein purification using Ni-IDA resin to confirm that the target protein could be purified.
• Test:
  The SDS-PAGE analysis result indicated that the target protein was successfully captured by Ni²⁺ resin. However, only EGFP-CBD protein could be eluted from resin (Figure 8). The sRAGE-Int^N, Int^C-6xHis-CBD, and 6xHis-Int^C-CBD protein shows no or slightly elution (Figure 9).

  

  

• Learn – Elution buffer doesn’t have enough strength to elute the target proteins:
  
  Except for EGFP-CBD, the elution results of the other proteins were very poor. We believe that the current elution buffer (600 mM imidazole) does not have enough strength to elute the target proteins.

Trying different elution buffers for protein purification

• Design & Build – Increase the concentration of imidazole in the elution buffer and lower the pH:
  Since the elution of sRAGE-Int^N, Int^C-6xHis-CBD, and 6xHis-Int^C-CBD protein from resin is poor, we increased the imidazole concentration in the elution buffer to 1M and 2M, and in addition, we also try an acid elution buffer at pH 2.5.
• Test:
  sRAGE-Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD protein were eluted using three different elution buffers. Both 1M and 2M imidazole slightly increase the elution of Ni-IDA resin bound protein (Figure 10). The acid elution buffer did not work (Figure 11).

  

  

• Learn – The concentration of the eluted protein is not sufficient:
  
  Increasing the concentration of imidazole in elution is feasible, but the eluted protein concentration is not very high. Therefore, we reduced the volume of the elution buffer.

Increase the concentration of eluted protein

• Design & Build – Reduce the volume of the elution buffer to 200 µl:
  To increase the concentration of the eluted protein, we used an elution buffer with 1M imidazole and reduced the volume to 200 µl.
• Test:
  According to SDS-PAGE, the protein concentration in the elution increased significantly, but most target protein was still trapped on the Ni-IDA resin (Figure 12).

  

• Learn – The presence of inteins may affect protein purification:
  
  Although the protein concentration in the elution has increased, not all of the target proteins were eluted. In addition to continuing to search for the optimal protein purification conditions, we also suspect that the presence of intein may affect protein purification. Because in addition to EGFP-CBD, we also prepared a positive control (EGFP-4A-C7 from 2023 CCU-iGEM) (Figure 13.) to confirm that the target protein without the intein can be successfully purified.

  

Intein splicing

• Design & Build – Verify whether correct splicing can occur under unpurified conditions and evaluate whether the position of the 6xHis tag affects intein splicing:
  Although the purification of the sRAGE-Int^N, Int^C-6xHis-CBD, and 6xHis-Int^C-CBD protein was not completely successful, the reference suggested that intein splicing can still be performed in 8M urea buffer. We conducted two sets of intein splicing reactions under the conditions of 37°C for 0, 0.5, and 1 hours. One set is sRAGE-Int^N and Int^C-6xHis-CBD and the other one is sRAGE-Int^N and 6xHis-Int^C-CBD.
• Test:
  Both set of combination have been successfully spliced at 37°C for 0, 0.5, and 1 hour (Figure 14-15). However, after splicing, the protein bands on the SDS-PAGE are not very clear.

  

  

  

  

• Learn – Obtain more discriminative results using Western blot:
  
  Since the protein bands after splicing are too faint to distinguish clearly, we believe that using Western blot can help confirm the splicing results.

Western blot analysis

• Design & Build – Obtained the results of intein splicing using Western blot:
  To more accurately determine whether the intein splicing of sRAGE-6xHis-CBD was successful, we performed a Western blot.
• Test:
  According to the Western blot results, splicing between Int^C-6xHis-CBD and sRAGE-Int^N was successful at 37°C for 0 hr, 0.5 hr, and 1 hr. However, the amount of successfully spliced protein between 6xHis-Int^C-CBD and sRAGE-Int^N appeared to be low, and no successful splicing was observed at 0 hr (Figure 16).

  

• Learn – The position of 6xHis affected intein splicing:
  
  Under the same conditions, the splicing efficiency of Int^C-6xHis-CBD and sRAGE-Int^N is better than that of 6xHis-Int^C-CBD and sRAGE-Int^N. We speculate that the position of the 6xHis tag affects intein splicing.

E. coli can successfully express sRAGE-Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD, but soluble protein can only be obtained through urea denaturation. Although proteins denatured with urea can still be purified and undergo intein splicing, we are still concerned that, aside from the His tag used for purification and intein, most of the other proteins may be misfolded due to urea denaturation, resulting in loss of protein functionality. We can verify our concerns based on the results of EGFP-CBD [See the result of EGFP-CBD]. Therefore, to ensure that we obtain functional proteins, we have decided to switch the expression system to HEK293T cells.

To overcome the protein expression issues in E. coli, we shift to mammalian cells to expression our proteins.
We then designed five new constructs based on pcDNA3.1 for protein expression in HEK293T cells. sRAGE-Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD are adapted from our original E. coli-expressed constructs, while the fifth was an sRAGE N25Q and N81Q mutant (mut-sRAGE-Int^N), which harbors a higher AGE-binding affinity than sRAGE-Int^N.

Building constructs for HEK293T cells

• Design & Build – Designed and constructed five new pcDNA3.1-based constructs for HEK293T expression:
  Gene blocks of sRAGE-Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD were adapted from our original E. coli-expressed version (Figure 17). In addition, a mutant construct (mut-sRAGE–Int^N) carrying N25Q and N81Q substitutions was generated to compare the AGE-binding affinity with wild-type sRAGE–Int^N (Figure 17).

  

  

  

  

  

• Test:
  To assemble these constructs, we first cloned sRAGE -Int^N and mut-sRAGE -Int^N into pcDNA3.1 backbone using Gibson Assembly but failed. We then use BamHI and EcoRI restriction enzyme to clone insert into pcDNA3.1. The sequencing result showed an additional sequence located upstream of sRAGE (Figure 18).
  
  

  

• Learn – Investigating the impact of unexpected 6xHis–T7 insertion on protein expression:
  
  Based on the sequencing result, we knew that the unexpected 6xHis–T7 sequence in the vector eliminated the homologous overlap required for Gibson Assembly, leading to assembly failure. Importantly, the unintended upstream insertion did not affect the reading frame but likely disrupted the signal peptide.

Protein analysis in HEK293T cells expressed from an incorrect plasmid construct

• Design & Build – Examining the effects of unexpected 6×His–T7 Tag in sRAGE-Int^N secretion:
  We transfected the plasmid into HEK293T cells and examined expression using fluorescence microscopy and immunocytochemistry (ICC). To further assess whether the protein was secreted, we use Western blot to performed on both cell lysates and culture supernatants.
• Test:
  Fluorescence microscopy showed normal expression of the target protein in cells with green fluorescence (Figure 19), and ICC further showed the cytoplasmic location in HEK293T cells (Figure 20). For the localization, Western blot confirmed that the protein was at the expected size and there wasn’t any target protein secret out of cells (Figure 21).

  

  

  

• Learn – The extra peptide at N-termal of signal peptide blocks secretion without affecting expression or folding.
  
  The signal peptide of sRAGE appears highly sensitive to upstream extensions, and any addition before it can block secretion. Based on this finding, we proceeded to redesign the construct by removing the unintended T7–6xHis segment.

Remove unexpected sequence and check

• Design & Build – Vector reconstruction with HindIII/EcoRI:
  To restore proper secretion, we reconstructed the vector by removing the unwanted upstream sequence. A new primer containing a HindIII site complementary to the insert was designed to replace the sequence via overlap extension PCR. Then, we used HindIII and EcoRI digestion to clone sRAGE -Int^N, mut-sRAGE -Int^N, Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD into pcDNA3.1, and sent the constructs for sequencing.
• Test:
  Sequencing confirmed that the unwanted upstream sequence had been successfully removed and that the gene blocks were correctly inserted into the pcDNA3.1 vector. The cloning results for sRAGE–Int^N are presented in Figure 22.
  
  

  

• Learn – Correction of construct errors by PCR enabled successful protein expression:
  
  This result demonstrates that incorrect sequences in a construct can be corrected through PCR using designed primers. Therefore, with the corrected constructs, we can proceed to transfect cells and express the intended protein.

Protein expression and purification of sRAGE -Int^N and mut-sRAGE -Int^N

• Design & Build – The proteins secreted in cuture medium were purified with Ni-IDA resin and verified by Western blot:
  We transfected the sRAGE -Int^N and mut-sRAGE -Int^N plasmids into HEK293T cells and confirmed protein expression by fluorescence microscopy. The cell culture medium containing secreted proteins was stored at –80 °C to avoid degradation. For each purification, we used Ni-IDA resin same protocol as for E. coli to express proteins and use Western blotting to assess purification success.
• Test:
  Fluorescence microscopy confirmed normal expression of the target protein in Figure 23 and Western blot results verified that the protein was secreted into the culture medium at the expected size. However, the protein failed to bind to Ni²⁺ resin and was detected entirely in the flow-through (Figure 24).

  

  

• Learn – Insufficient binding time likely reduced binding efficiency:
  
  Western blot analysis confirmed that the transfected protein exhibited the expected size, but its binding efficiency to the resin was relatively low. Therefore, we guessed that either the binding time was insufficient or that imidazole in the buffer was interfering with His-tag interaction.

Optimization of binding and elution conditions

• Design & Build – Binding the protein overnight in PBS without imidazole, then washing and eluting with high-imidazole buffers:
  To improve purification efficiency, the storage buffer was replaced with PBS lacking imidazole, and the binding step was extended from 1 hour to overnight. Additionally, because previous attempts failed to efficiently elute the protein, the elution buffer was replaced with PBS containing a higher imidazole concentration (1 M).
• Test:
  The Western blot analysis revealed that the target protein remained bound to the Ni-IDA resin and could not be eluted, even under high-imidazole conditions (Figure 25).

  

• Learn – Intein-containing proteins from E. coli and HEK293T cells were difficult to elute:
  
  We suspected that the failure to elute was either due to the intein domain obstructing His-tag accessibility or to suboptimal buffer composition. Since the target protein was not eluted, in addition to continuing to search for the optimal protein purification conditions, we also suspect that the presence of the intein may affect the functionality of the 6xHis tag. Therefore, we used a positive control, EGFP-4A-C7 provided by CCU-iGEM 2023, to verify that our hypothesis is plausible (Figure 26).

  

Expressing Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD protein in HEK293T

• Design & Build – Expression of Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD in HEK293T:
  Following the successful expression of sRAGE-Int^N and mut-sRAGE-Int^N, we proceeded to express Int^C-6xHis-CBD, 6xHis-Int^C-CBD, and EGFP-CBD in HEK293T. Then we used fluorescence microscopy to examine expression and performed Western blot to check the size of our protein.
• Test:
  Fluorescence microscopy confirmed normal expression of the EGFP-CBD protein in cells, as indicated by green fluorescence (Figure 27). The Western blot analysis further showed that EGFP-CBD was detected at the expected size. However, no detectable bands corresponding to Int^C-6xHis-CBD or 6xHis-Int^C-CBD were observed (Figure 28).

  

  

• Learn – Intein-containing constructs may have failed to express
  
  The absence of signal for Int^C–6xHis–CBD and 6xHis–Int^C–CBD suggests that transfection or expression of these constructs was unsuccessful. Further optimization such as adjusting transfection conditions will be required to determine whether the lack of expression is due to transfection inefficiency, instability of the construct, or rapid degradation of the protein.

We redesigned and rebuilt multiple pcDNA3.1 constructs for expression in HEK293T cells to overcome the limitations observed in E. coli expression. Initial cloning attempts failed due to primer mismatches and the introduction of an unwanted sequence, but vector reconstruction using HindIII/EcoRI and overlap extension PCR successfully resolved these issues, as confirmed by sequencing. The verified constructs were transfected into HEK293T cells, and protein expression and secretion were confirmed by fluorescence microscopy and Western blotting. However, purification using nickel beads revealed inefficient binding and elution. Optimization with overnight binding and higher imidazole concentrations improved binding but did not resolve elution, particularly for intein-containing proteins. These results suggest that either the intein domain or buffer conditions hindered elution, and further testing with proteins previously successfully purified will help clarify the underlying cause.