Space exploration demands materials that can withstand extreme conditions while remaining lightweight, strong, and adaptable. Current options lack self-assembling, adhesive, and hydrogel-forming capabilities. To address this gap, we engineered a fusion protein combining hagfish intermediate filament proteins with mussel foot proteins.
The fusion protein can be effectively heterogeneously produced in E. coli BL21 (DE3) under the control of the conventional T7-LacO operon. This biomaterial forms threads and hydrogels with high water content, strong adhesion, and mechanical resilience.
We assessed the protein yield rate, and tested its hydrogel formation, adhesive strength, and structural stability, demonstrating its potential as a novel, multifunctional material for future space applications.
Hagfish intermediate filament (HIF) proteins can rapidly form hydrogels under seawater conditions[1, 2]. Two primary variants, α (alpha) and γ (gamma), have been successfully produced as recombinant proteins in E. coli [3] and are referred to as rHIFα and rHIFγ. Both share a common structure consisting of an N-terminus, a central rod domain (CRD), and a C-terminus[4]. The N- and C-terminal regions are unstructured, while the CRD forms an α-helical structure that contributes to the protein’s exceptional mechanical strength through coiled-coil with other threads [5].
To enhance adhesive properties, we incorporated segments from mussel foot proteins (MFPs), specifically mfp-3b and mfp-5, which are known for their robust adhesion in wet and harsh environments due to the DOPA groups. [6]
A flexible 10-amino-acid linker (GGGGSGGGS) was inserted between the HIF protein and the MFP segment to prevent structural interference and preserve hydrogel-forming capability. At the N-terminus, a 10×His tag was added to facilitate purification via Ni-IMAC, followed by an enterokinase cleavage site that allows removal of the tag if it adversely affects protein behavior. The entire construct was placed under control of a T7 promoter and terminator, with a standard RBS and LacO, and cloned into the commercial pET-28a vector for expression in E. coli BL21(DE3).
The following constructs were designed as illustrated in Fig. 1:
Figure 1. Gene circuit design and fusion protein design.
Figure 2. Expression vector for the fusion proteins, showing only rHIFα-mfp3b as an example.
The 3D structural modeling with AlphaFold has allowed us to visualize the designed structures. We found that the introduction of MFP segments does not adversely affect the overall structure and thus presumably retains the functionality of HIF proteins.
Figure 3. Protein Structures predicted by AlphaFold.
Past works in iGEM In 2014, the UCC_Ireland team expressed rHIFα (Part: BBa_K1518001) and rHIFγ (Part: BBa_K1518000) proteins, demonstrating that these recombinant proteins could be extracted from inclusion bodies and subsequently assembled into hydrogels [7]. Later, in 2020, the HKUST team sought to engineer a sustainable textile system leveraging the mechanical properties of HIF proteins and reported new signal peptide design and structural analyses [8]. To our knowledge, no other iGEM teams have worked specifically with HIF proteins.
Several iGEM teams have explored mussel foot proteins (MFPs). Among them, the 2024 NAU-China team’s work is most inspiring to us on self-healing adhesive materials for underwater soft robots. Their design involved fusing mfp-5 to a tandem repeat polypeptide derived from squid ring teeth, resulting in enhanced underwater adhesion and durability [9].
The designed construct was communicated to the commercial company Suzhou Antatares Biotech company. The DNA design was screened for safety and feasibility, and then synthesized as oligonucleotides. Successfully synthesized constructs were cloned onto pet28a vector and sent to us. We then verified the sequences by Sanger sequencing, as shown in Figure 4. The sequencing results confirmed that the plasmids were correctly assembled, aligning perfectly with our intended designs.
Figure 4. Success in constructions verified by Sanger sequencing. To avoid redundancy, only the result for rHIFα-mfp3b is shown. Other constructs are similar.
Following the successful construction of our plasmids, we transformed the engineered plasmid into E. coli BL21 (DE3) and induced the expression of our fusion proteins. The major experimental method is as follows:
The cell pellets were resuspended with 20 mM Tris-HCl buffer; the resuspended bacterial pellets were subjected to ultrasonication to disrupt the cells and release the proteins; centrifugation was performed to separate the supernatant; the pellet, potentially containing inclusion bodies, was resuspended in 20 mM Tris-HCl buffer to prepare for analysis; 50 μL of each sample was combined with SDS-Page loading buffer, then heated at 100°C for 5 minutes to denature the proteins. The samples were analyzed on SDS-PAGE gel.
Figure 5. SDS-Page results of the designed proteins, for rHIFα, rHIFα–mfp3b, rHIFα–mfp5, rHIFγ, and rHIFγ–mfp5. (A) comparison between uninduced bacteria culture (-iPTG) and induced culture (+) (B) comparison between the soluble protein expressed in the supernatant and the inclusion body protein in the pellet.
1) SDS-Page Analysis
2) Experiment Technique
Due to our still-maturing electrophoresis techniques, there was a noticeable distortion between the marker lane and the sample lanes, which we aim to improve in future experiments.
Action
In Cycle 2, we first optimized the conditions for protein expression, especially with 1L shaking flask, to maximize yield and solubility. Induction strength (IPTG working concentration), induction temperature, and induction duration were investigated.
To find the optimal IPTG inducer concentration, seed cultures were prepared in TB medium (Terrific Broth) and inoculated into 1L Erlenmeyer flasks with three baffles containing 120 mL of TB medium with 100mg/ml Kanamycin (later we will refer to this as TB+Kan). Once the OD600 reached the mid-log phase (in our experiment, this corresponds to OD=4), different concentrations of IPTG (0.1 mM, 0.3 mM, 0.5 mM, 0.6 mM, and 0.8 mM) were added. After a 4-hour induction period with IPTG, the bacterial cultures were harvested and analyzed by SDS-Page. Then, we used the optimal IPTG concentration to explore the temperature conditions, and subsequently, using the optimal temperature condition to explore the induction durations.
Figure 6. SDS-Page results of the various fermentation conditions, using rHIFα–mfp3b, with different (A) IPTG induction concentration, (C) temperature of bacteria culturing after induction, and (E) culturing time after induction. (B), (D), and (F) shows the grey scale of the target band in the supernatant (sup) sample for each of the investigated conditions. The optimal fermentation conditions are determined to be 0.5 mM IPTG, 30°C, and 4 hours.
From the SDS-Page results analysis, the optimal fermentation conditions are determined to be 0.5 mM IPTG, 30°C, and 4 hours after seed inoculation. We then induced expression of all designed proteins in the 1L Erlenmeyer flask filling 120mL culturing medium. Upon finish fermentation, the bacteria were harvested by centrifugation and lysed by ultrasonication. The supernatant of the lysate after centrifugation were passed through Ni-TED IMAC columns driven by gravity to obtained purified proteins.【Immobilized-metal affinity chromatography (IMAC) Columns: TA-Ni FF (TED) metal chelating chromatography medium from Truking Micro-Sphere (25mL, Y5394) is used. Flow is driven by gravity.】
However, both the SDS-Page run and the Bradford assay of all the elution samples for all variants showed almost no protein extracted.
After careful discussion with Professor Yin from Beijing Normal University and Mr. Wang from LINK-Spider, we identified a potential cause. We noticed that most of the proteins remained in the flow-through even after passing the column 3 times. The hagfish intermediate filament proteins have unstructured N- and C- termini, potentially enclosing the 10×His-tag and leading to limited access to Ni-His interactions. We plan to use a mild urea buffer (4M Urea with 20mM Tris-HCl) to treat the protein, mildly denature the protein to expose the His tags.
Figure 7. Purification with 4M Urea (20 mM Tris-HCl, pH=8.0) lead to successful purification for the proteins. (A) Our team 3D printed a hollow rack to conveniently install and adjust gravity driven nickle column running. (B) SDS-Page of the samples during purification of rHIFγ-mfp5: supernatant (sup), and pellet from after the ultrasonication; sample after twice (2) flow-through (FT2), and with each elution (E1 – E6), we collected 1.5mL protein solution.
We used 20mM Tris-HCl + 4M Urea and adjusted the pH=8.0 to be compatible with all the isoelectric points for all designed proteins. For the washing buffer, we used 20mM Tris-HCl + 4M Urea + 20mM Imidazole; and for the elution buffer, we used 20mM Tris-HCl + 4M Urea + 500mM Imidazole.
We were able to successfully extract large quantities of the rHIFγ-mfp5 and rHIFα-mfp3b, but since we were running out of lab time, so we only obtained very limited amount of rHIFα, rHIFγ, and rHIFα-mfp5. Later, we will focus on testing the properties with rHIFγ-mfp5 and rHIFα-mfp3b.
We conducted Bradford assay (Beyotime, P0006) to quantitatively measure the concentration of each eluted protein sample. Given that each sample’s volume is fixed at 1.5 mL, and given that we have recorded the volume of the bacteria culture when harvesting, we could calculate the weight (mg) of the total extracted proteins, and in turn assess the protein yield rate.
Compared to the previous research works, the protein yield is satisfying. Firstly, a previous master thesis research work [12] optimized all fermentation conditions in producing rHIFγ and rHIFα, and collected all proteins in the inclusion body to reach a yield of about 360 mg/L. In our case, we only extracted the soluble proteins in the supernatant. Meanwhile, a considerable portion of the protein still remain in the flow-through. The soluble protein production in our design could reach very high level, once the purification process can be further optimized.
Based on expert advice, we plan to conduct a series of property, function, and performance tests that are crucial for understanding the behavior of our engineered proteins in various environments and applications.
You can see our Hardware page 2025.igem.org/shsbnu-china/hardware for details of these testing builds.
We prepared samples by mixing 1mg/mL solution of rHIFγ-mfp5 and rHIFα-mfp3b at a 1:1 volume ratio. After 12000x centrifugation, 10min, we found white precipitations in the protein mixture solution (supernatant was mostly discarded so as not to disturb the pellets). This evidence partly proves the formation of hydrogel when the two protein co-assemble. However, further experiments and verifications are needed. See our Learn section for detailed discussions and experiment plan.
Figure 8. Hydrogel is formed as white precipitates with 1:1 rHIFγ-mfp5 and rHIFα-mfp3b.
To assess the adhesive properties of our proteins, we applied them as a glue on various material surfaces.
Figure 9. rHIFγ-mfp5, rHIFα-mfp3b, and rHIFα-mfp5 all exhibited adhesive power between different pairs of materials with varying strength: (A), (B) plastics – glass (strong) (C) plastics – metal (weak) (D) glass – glass (very strong) (E) glass – metal (strong).
rHIFγ-mfp5, rHIFα-mfp3b, and rHIFα-mfp5 all exhibited adhesive power between different pairs of materials with varying strength. In the case of Fig. 9 (C), the plastic metal pair are glued when picked up vertically, but broke upon a horizontal push. Meanwhile, rHIFα and rHIFγ don’t show adhesive strength and the tested pairs break immediately when picked up vertically.
Figure 10. Measurement of maximum adhesive force. (A) the controlled pulling device with a force measurement sensor using high-school physics lab class devices. The proteins were used as glues between two glass plates, and the surface area is 625mm². The two plates were pulled by the LEGO motor very slowly and the force at which the plates were separated is recorded as the max force. (B) rHIFγ-mfp5 and rHIFα-mfp3b at 0.5 mg/mL showed maximum adhesive force near 7000 N/m², close to normal student glues. Their mixture showed a statistically significant increase in the max adhesive force to almost 10,000 N/m². (***: p<0.001, **: p<0.01)
Figure 11. Maximum adhesive force of different concentrations, with rHIFα-mfp3b. The measurement setup is exactly the same as in the above measurement. In the range of 0 – 2.0 mg/ml concentration, we observed increasing max adhesive force when the concentration increases. We do not expect the trend to continue when the concentration is large enough.
Figure 12. Maximum adhesive force at different temperatures, with rHIFα-mfp3b. The max force quickly drops to non-adhesive when the temperature reached below -20°C. The max force increase to near 20,000 N/m² when heated up to 150°C. The two low temperature tests were achieved by placing the dried glued glass plates into -20°C freezer and -72°C freezer in our synthetic biology lab. The high temperature tests were achieved by placing the dried glued plates into the drying oven in our lab, and baked at the respective temperatures for 3 hours. 150°C was reported in the literature to be the temperature on the surface of the space suit when astronauts perform space walking missions.
Due to limited lab time and resources, large quantity purification was achieved only for rHIFα–mfp3b, and we focused our experimental testing on this protein. Other constructs will be characterized in future work when resources are available.
Building upon previous iGEM studies and literature reports, we observed hydrogel formation from rHIFα–mfp3b + rHIFγ–mfp5 mixture solutions. The low efficiency might come from low protein concentration. The research reported strong hydrogel assembly through ultrafiltration [12]. With additional time and resources, we plan to increase protein production, redissolve the protein, and apply ultrafiltration to test co-assembly conditions, and therefore, more hydrogel properties in a systematical way.
The yield of rHIFα–mfp3b (~193 mg/L) is comparable to previous reports of recombinant HIF proteins (~360 mg/L) [12]. However, the reported yields included proteins retrieved from inclusion bodies, whereas our protocol focused on the soluble fraction, although extraction losses occurred due to incomplete binding on the Ni-TED affinity column. The presence of the mfp3b may partially enhanced solubility. Future optimization will involve consulting with protein biochemistry experts to improve binding efficiency and overall yield.
The surface temperature during spacewalk missions can reach as low as −173 °C and as high as +120 °C [11]. Our preliminary tests find that the adhesive force decreases at low temperatures. This reduction may result from decreased DOPA oxidation at low temperature conditions. In the future, we will try to change the terminal domains or fusion with motifs to address the low temperature performance issues.
In fact, in the next engineering cycle, we did construct rHIFα-mfp variants where the N- terminal region were deleted. Unfortunately, we do not have enough time to systematically produce and characterize these proteins. In the future, we will further explore the structure-property relations.
Based on the previous work [10], it has been suggested that the selective removal of either the N-terminal or C-terminal regions of rHIFα or rHIFγ proteins can enhance its mechanical strength when the proteins were spun into fibril threads and tested under tensile stretch. We have decided to engineer new protein constructs by removing one terminal region at a time, i.e., δN or δC, to assess the impact on the protein's strength and functionality.
Figure 13.1 Gibson Assembly strategy to remove the N- and C- terminal regions from the rHIF proteins.
Figure 13.2: Agarose gel electrophoresis image showing the plasmid successfully assembled using the Gibson assembly method.
After constructing the plasmids, we proceeded to sequence them to validate the correctness of the assembled sequences. The sequencing results indicated successful construction for the following plasmids:
Figure 14. Sequencing results verified the successful construction of three constructs.
We were able to test the expression of two proteins rHIFα-δN, and rHIFα-mfp5-δN. We ran out of time before wiki freeze to express γ: N-CRD. The expression was carried out under the previously determined optimal conditions.
Figure 15. SDS-Page results to verify the induced expression of two rHIF proteins in which the N- terminal regions are removed. For the rHIFα-mfp5-δN construct, the anticipated protein size is around 60 kDa, and clearly shown in the correct region in the page bands. For the rHIFα-δN construct, the anticipated protein size is around 45 kDa, and since there is a strong intrinsic band in the negative control, and also in all other expression, the results are inconclusive on the page results.
Following expression, we purified the proteins using gravity flow nickel column chromatography with 20mM Tris-HCl + 4M Urea as the buffer.
Figure 16. SDS-Page shows the purification sample of rHIFα-δN and rHIFα-mfp5-δN. Through nickle column purification by catching the 10×His tag, we can verify that both proteins are expressed successfully.
Figure 17. Simple tests verified rHIFα-mfp5-δN has adhesive power as expected from its mfp5 group. Due to time constraints, we will perform max force measurement in the future.
Due to time constraints, we have opted to focus on the production and extensive testing of the previous proteins, and haven’t obtained the chance to produce in large quantity of these structure-altered new proteins. Later when we have the time and resources, we will apply the optimal fermentation and purification conditions that we previously identified to manufacture these proteins, and subsequently use our designed physicochemical property measurement methods to successfully characterize them.
Unfortunately, many plasmids designs like rHIFα–mfp3–δN and rHIFγ–mfp5–δC were not successfully constructed according to the sequencing results. This outcome suggests possible challenges such as low assembly efficiency or limitations in the cloning design.
In our Gibson assembly strategy, the PCR-amplified vector fragment was relatively long, approaching 7000 bps, which increased the likelihood of false-positive colonies.
Future plan: We will redesign our cloning strategy to reduce the length of the vector and thus the chance of false positives, or optimizing primer design to improve assembly accuracy.
Due to limited time, we were unable to produce large quantities of recombinant protein for detailed quantitative analysis.
Future plan: With more time, we plan to repeat the protein expression and purification process, then systematically measure the properties of the new constructs (rHIF–δN and rHIF–δC). Specifically, we will examine whether removing the N- or C-terminal region affects protein yield, solubility, hydrogel formation, and adhesive strength. By conducting controlled experiments under the same conditions as in the previous cycle, we aim to compare how structural changes influence these key physical and functional properties.
Building on the learnings from this cycle, we have several plans for the next stages of our project: