Team:Greatbay SCIE/Results
Results
Functional Characterization of all recombinant protein in the toolbox
Our project, SticKit, intends to create an underwater adhesive toolbox using various adhesive proteins that reveals impressive physical, chemical and biological properties, especially in aqueous environments. They are strong and difficult to obtain. Inspired by this and previous research papers, we had an idea in mind: to turn a microbial assembly line into a toolbox. Combining different parts from various organisms to obtain recombinant adhesive proteins, we hoped to achieve better performance. Our primary task was to produce these recombinant proteins and characterize their adhesion properties. We believed that this knowledge in different properties of bio-adhesive proteins will provide necessary information for future research and application. We also hoped that this project will give other iGEMers, who use adhesive proteins in their projects, a valid approach.
The backbones of “SticKit” are Mfp5-related recombinant proteins. Mfp5 (Mfp stands for Mussel foot protein) is a very adaptable adhesion contributor due to its strong adhesion and relatively small size (77 amino acids, 9kDa). Microbially synthesized repeats of mussel foot protein also display enhanced underwater adhesion (Kim, et al. 2018). Based on this evidence, we combined Mfp5, Mfp5-mfp5 and Mfp5-mfp3 to other platform proteins, such as CsgA – a fiberous protein in E.coli. Platform proteins can be seen as standardized and modularized parts used to design new functional recombinant proteins, just like building blocks. We also tried to standardize other promising proteins, like rBalcp19k and Fp1, to make this toolbox more adaptable to various environments and conditions.
Last but not the least, posttranslational modification played an important role in Mfp5 adhesion. Since different modification would change the properties of proteins and enhance their performances, this opened a new dimension for us: protein modification. Therefore, we integrated tyrosinase mTyr-CNK hydroxylation system into our toolbox.
E-coli as a cell factory to produce recombinant proteins
We selected the T7 expression system in order to achieve high levels of protein expression. E.coli BL21(DE3) Rosetta was used as expression hosts to compensate the issue of rare codons. All proteins were added with a 7xHis-tag on carboxyl terminal for purification. Linker 5’-GGGGSGGGGS-3’, contributing to the correct folding and corresponding biological activities of the different domains of the fusion proteins, was added between different proteins, making sure that they would not jam each other (Figure 1). csgA gene without secretion signals was amplified from E.coli MG1655 genome. Other parts like mfp5, mfp3, fp1 and rBalcp19k were obtained by synthesis. Gibson and Golden Gate assembly were used to construct all recombinant plasmids without scars. All genes of recombinant proteins were cloned into pET28b expression vectors, verified by gel electrophoresis (Figure 2) and sequencing.
Protein size and Isoelectric point were calculated by using software Geneious version 11.1.3., predicting between 18 and 33 kDa (Table 1). All recombinant adhesive proteins was considered basic if they had a pI (isoelectric point) greater than 7.
Recombinant plasmids shown in Figure 2 were transformed into BL21(DE3) Rosetta and grown to OD600 ~0.6 in LB broth containing 50 mg/mL kanamycin and 30 mg/mL chloramphenicol at 37°C. Protein expression was induced with 0.5 mM IPTG at 37°C for 5 h.
CsgA-Mfp5 design was based on Professor Zhong Chao’s work (Zhong, et al. 2014). He combined two important properties of glue-adhesion and cohesion to create Mfp5-CsgA. We developed his work further by designing a new recombinant protein based on Mfp5-CsgA. As shown in Figure 3, CsgA-mfp5, CsgA-mfp5-mfp5 and CsgA-mfp5-mfp3 proteins were purified under “Denatured Conditions” (see methods), as we cannot acquire proteins under native conditions. Weak bands of interest were presented on the lane E2/E1 for recombinant proteins. These proteins were verified by their correct positions on SDS-PAGE. However, due to our SDS-PAGE gels (12%)’s low resolution, the differences between their sizes were not obvious.
Recombinant hybrid MAP Fp-151 was designed by a Korean research team (Hwang, et al. 2007). It was a novel fusion protein comprising six fp-1 decapeptide repeats at each fp-5 terminus. The predicted size of Fp151 is 24.83 kDa and the isoelectric point is 10.58. Recombinant hybrid fp-151 fused with histidine affinity ligand was successfully expressed in E. coli BL21(DE3) Rosetta. The team studied Fp151 for a decade and used acetic acid solution to extract it; however, we failed at replicating this method even when 25% acetic acid solution was used, it was impossible to eliminate the extra bands and obtain a single band of interest (Figure 4AB). The protein was successfully purified with the same denatured protein purification methods used for purifying CsgA-mfp5. Clear bands with correct sizes presented on SDS-PAGE gels after purification (Figure 4C, Lane E2 and E6).
Barnacle cement proteins are very promising in making biomedical glues. rBalcp19K from Balanus albicostatus had the properties of both self-assembly and adhesion. It could function in more basic condition than Mfps(Liang, et al. 2015). Thus, we designed a novel recombinant protein by combining it with Mfp5. We expected rBalcp19k-Mfp5 to be more adhesive than rBalcp19K, which solidified the modularization and standardized ability of Mfp5. We tried to express both rBalcp19k and rBalcp19k-mfp5 in E.coli BL21(DE3) Rosetta. They were successfully expressed and purified under “Native Conditions” (see methods), as bands of rBalcp19K appeared between 15kDa and 25kDa on 12% SDS-PAGE gel (Figure 5A) and bands of rBalcp19K-mfp5 appeared between 25kDa and 35kDa (Figure 5B). In SDS-PAGE of rBalcp19k, there were various unexpected band; using higher concentration imidazole as washing buffers still could not eliminate them. We hypothesized that they were polymers.
DOPA modification to make Mfp sticker
DOPA (Dihydroxyphenylalanine) was a crucial unnatural amino acid contributed to Mfp5/Mfp3 adhesion. Mfp5 was a core adhesion contributor protein in our toolbox. DOPA from natural Mfps could be posttranslational modified (PTM), but E.coli could not. We had two options, either manually modify it with tyrosinase in vitro or co-express tyrosinase along with Mfps in vivo. Tyrosinase could oxidize tyrosine into DOPA, however, DOPA will also be oxidized by tyrosinase and become Dopamine quinone (Figure 6). As a result, we selected mTyr-CNK, a tyrosinase from marine archaeon Candidatus Nitrosopumilus koreensis, for both in vivo and in vitro DOPA modification,. This tyrosinase exhibited a distinguished monophenolase/diphenolase activity ratio (Vmax mono/ Vmax di = 3.83), enhancing catalytic efficiency against L-tyrosine. In theory, modifying Mfp5 using tyrosinase should result in more DOPA instead of Dopaquine.
In vitro DOPA modification
We have produced mTyr-CNK individually and tried to modify it for in vitro DOPA modification. A clear induced band of mTyr-CNK was observed after expression with 500uM IPTG at 25℃ for 20 hours. The protein was verified by SDS-PAGE after it was purified under native condition. Interestingly, after overnight expression, the cell pellets of bacteria showed rose-like color (Figure 7A), a possible result of the interference of mTyr-CNK in pigment pathway of E.coli BL21 (DE3) Rosetta. However, the exact mechanism remains unknown due to the lack of research on tyrosinase mTyr-CNK. The protein yielded very high, approximately 7mg/L (Figure 7B), higher than any other recombinant protein in our toolbox. We used mTyr-CNK to modify other recombinant proteins; 10 ul 0.35mg/ml mTyr-CNK (in PBS, 0.02mM CuSO4) was added into 90ul protein solution of concentration 0.5m/ml (pH=6.0 PBS) for 3 hours in room temperature. DOPA modifications were verified by “NBT staining” (see methods). Since Dopa-containing proteins can catalyze redox-cycling reactions in an alkaline solution (pH 9), they can be be specifically stained by nitroblue tetrazolium (NBT) and glycinate solutions. The NBT assay was thus used to confirm the successful post-translational modification of tyrosine into Dopa in modified proteins. All in vitro modified recombinant protein produced positive result (turned purple), suggesting that tyrosines were successfully modified into DOPA, BSA protein was used as a negative control (Figure 7C).
In vivo DOPA modification by co-expression
It is suggested that in vivo DOPA modifications could hydrolyze higher proportion of tyrosine to DOPA and simplify later processes (Choi, et al. 2012). This way, proteins could be extracted more easily and perform better. Thus, we created a prototype in vivo co-expression bi-plasmid system of recombinant proteins and mTyr-CNK. A pET vector can be used in combination with a vector system with p15A replicon, recombinant pACYDuet-1-mTyr-CNK plasmid was prepared from pACYCDuet-1. We transformed two-plasmid system into BL21 (DE3) to modify in-vivo (Figure 8A), using the same methods in expression as mTyr-CNK with purification tag and protein purification as their non co-expression versions, we obtained co-CsgA-mfp5, co-CsgA-mfp5-mfp5, co-Fp151, co-rBalcp19k-mfp5(Figure 8B). Their final yields were 0.4mg/L, 0.4mg/L, 4.75mg/L, 3.25mg/L respectively. NBT staining was used to verify the presence of DOPA; unfortunately, it was tested negative.
mTyr-CNK did not oxidize tyrosine into DOPA, which was likely due to the low expression of tyrosinase mTyr-CNK and recombinant proteins. As shown in Figure 9A and Figure 7B, when expressed alone, mTyr-CNK produced a high yield, but it did not express in co-expression strains under the same expression condition (Figure 9B). The expression level of recombinant proteins fell. Based on this observation, we modeled co-expression system to discuss the mechanisms behind it (see model). Modeling results should generate better gene circuits designs, improving the catalytic efficiency of the co-expression system.
Function analysis
Surface coation analysis
After obtaining a small amount of recombinant proteins, surface coating analysis (see methods), the qualitatively assessment of the surface adsorption ability of recombinant proteins, was performed on 2 of the most commonly used bio-related surfaces: hydrophilic glass slides and hydrophobic polystyrene tissue culture plates (Figure 10). As shown in Figure 11, Mfp5 related proteins (unmodified) exhibited higher surface absorption abilities than other recombinant proteins, whereas almost all absorbed BSAs were washed away. Basing on our current understandings of Mfp5, we were able to predict the adhesiveness of this protein, especially when it is combined with different proteins, making it more adaptable in many environments. Furthermore, DOPA modification significantly improved the surface absorption abilities of Mfp-related recombinant proteins, suggesting the positive contribution of DOPA in adhesive proteins’ performances. Future research on mechanisms of DOPA may reveal more applications of DOPA, possibly integrating it with other proteins, like fibrin, in order to enhance it.
Other adhesion analysis assay
We attached 2 pieces of plastics together (common 96-well plate and pipette tips) using small amounts of adhesive proteins. As shown in Video(left), negative control BSA protein showed no adhesion to the plastic, while CsgA-Mfp5 (left 1) and CsgA-Mfp5-Mfp5 (left 4) was strongly adhesive, though Fp151(left 2) and rBalcp19k (left 3) were slightly less adhesive.
Other than that, we tested the proteins’ adhesiveness using different materials. In these experiments, we mainly used Fp151 protein due to its higher yields. Results indicated that Fp151 proteins were more adhesive to plastic and glass, while showing no obvious adhesion to rubber band fragments and pieces of paper. Sodium hyaluronate, a widely used moisturizing material in facial masks, was added in a 1:2 mass ratio to recombinant proteins to enhance their performances. The low protein volumes greatly limited the characterization. Thus, we should aim to increase yields first.
Improve the yield
Overview of protein expression level
To compare the expression of each recombinant protein, using the same condition, we induced expressions in a small amount of proteins. The OD600nm of the culture mediums were measured to proofread sample volumes, making sure that similar numbers of cells were tested. The growth of E. coli with recombinant proteins was relatively poor when compared with E. coli with pET28b empty vector (data not shown), spotting no significant differences. Gray-level of the SDS-PAGE of whole cell lysates was quantitatively analyzed, results suggested that the expression of Fp151 had the highest expression level. Mfp5-Mfp3 and rBalcp19k also had relatively high expression. No obvious protein bands of interest were observed for CsgA related proteins and rBalcp19K-mfp5 on gels compared with lane pET28b empty vector. We purified some recombinant adhesive proteins and protein concentrations were measured by BCA assay, as shown in Table 2. Unsurprisingly, the yield, between 0.5mg/ml to 4mg/ml, was poor. CsgA-mfp5 and Csg-mfp5-mfp5 performed well in surface coating analysis. Though we changed the culturing conditions, such as induction times (4, 8, 20 hours) and incubating temperatures (37℃, 25℃), no bands of interest appeared. We shared these results with Prof. Zhong (Integrated Human Practice), he suggested that since CsgA is an amyloidogenic protein, it is a major subunit of adhesive curli fiber in E. coli. Expression of CsgA related recombinant proteins in a large quantity inside E. coli cells after induction by IPTG may lead to aggregate formation, resulting in growth deficiency.
Using sfGFP to characterize recombinant protein expression
We attempted to figure out the reasons behind the poor expressions, when fusing sfGFP on the C terminal of CsgA-mfp5. CsgA-sfGFP, Mfp5-sfGFP and sfGFP were set as controls (Figure 14). All 4 fragments were cloned into pET28b and transformed into E .coli BL21 (DE3).
25-ml cultures of E. coli BL21(DE3) bearing sfGFP fusion constructs grew in LB medium containing kanamycin (50 mg/ml) overnight. E. coli with fusion constructs, diluted at 1000-fold in 200-μL cultures, were grown to ~0.2/0.5/0.8 OD600 nm in a 96-well plate with cover and induced at 37℃ with 500μM IPTG for 22 hours. OD600nm and fluorescence were measured (488-nm excitation, 530-nm emission,10-nm band pass for GFP) with a Microplate Fluorescence Reader (THERMO Varioskan Flash). Fluorescence was normalized by dividing by the OD600 nm. We continuously monitored the OD600nm and fluorescence of these four strains, plotted their growth graph and induced fluorescence. We added IPTG to these strains in different times of their log phase, such as OD600nm = 0.2(early), 0.5(medium), 0.8(late).
Results were measured by ratio of fluorescence of OD600nm. As predicted, it showed that sfGFP had a much higher expression level than others. CsgA-sfGFP had the 2nd highest expression level, while Mfp5-sfGFP and CsgA-Mfp5-sfGFP expressed poorly (Figure 15ABC). This showed that the Mfp5 and related proteins had difficulties in expression, explaining why we failed at finding a band on SDS-PAGE after induction with IPTG.
Interestingly, adding inducer in an early stage of log phase (OD = 0.2) would delay the growth of all strains, except sfGFP (Figure 15D), causing them to re-enter lag phase. The more difficult the protein expressed, the longer the lag phase. During this second-lag-phase, fluorescence was growing continuously (Figure 15G), which meant that proteins were still accumulating in cells. However, when growth entered log phase, the normalized fluorescence was reduced to some extent, due to the rapid increase in OD600nm value. These phenomena were generally not observed when adding IPTG in medium (OD = 0.5) and late log phase(OD = 0.8) (Figure 15EF).
We were mainly concerned about the total yields, in other words, the absolute fluorescence of these cultures. Thus, we compared the total yields of all 4 constructs by adding inducers in different times of log phase (Figure 17). Results showed that adding IPTG in early log phase significantly reduced overall expression, fluorescence reached its peak in 5 hours. We concluded that an inducer should be added in late log phase in order to reach higher expression levels; 5 hours induction time was enough to obtain optimum yield.
Using Pichia pastoris as Chassis
Previous experiments demonstrated strong negative impacts on growth and expression, as adhesive proteins accumulated. This suggested that adhesive proteins may disrupt the normal functioning of cells. In order to improve chassis, we turned to yeasts. Unlike E. coli, yeasts could easily secret proteins out of the cells, so adhesive proteins were less problematic until it became too concentrated. Thus, Pichia pastoris was chosen as our premier chassis, improving the yields. Pichia pastoris only secreted a small amount of its own proteins to culture mediums, which gave us a relatively pure solution of recombinant proteins, making purification processes simpler.
Molecular Cloning
We synthesized 3 plasmids of recombinant proteins (CsgA-mfp5-mfp5, Fp151, Cp19k-mfp5). Since AOX1 promoter was a strong promoter and the alpha-factor would secret the protein of interest into culture medium, pPIC9K vector was selected. CsgA-mfp5-mfp5 and rBalcp19K-mfp5 were successfully cloned to yeast Pichia pastoris (see methods) and 5 strains in total were verified by gel electrophoresis (Figure 19). The molecular cloning of rBalcp19k-mfp5 was 1 step closer to success. Colony PCR was positive but sequencing indicated that gene rBalcp19K was missing. We needed more time to repeat this experiment.
Protein expression and purification
Proteins were expressed in small scale induced by methanol (see methods). Unfortunately, no proteins of interest were found in culture medium of CsgA-Mfp5-Mfp5 after induced expression for 48 hours at 30℃. This suggested that the proteins were not expressed nor secreted as expected under this condition. However, 4 strains of rBalcp19k-mfp5 demonstrated clear bands around 25kDa on Lane S (culture medium) after expressing them for 48 hours. The predicted size of rBalcp19k-mfp5 is 28kDa, larger than shown in the gel. This may be due to the acidification of culture medium after long-time incubation.
Future outlook
Synthetic Biology, a shortcut to nanoengineering, allows us to synthesize materials efficiently and sustainably. Protein engineering offers us a glance to the bright and promising futures of advanced materials and processing technologies. They will be widely applicable in both laboratories and day-to-day lives.
Overall, the characterizations of various proteins opens up a new dimension for the Materials Science. Adhesive proteins have great potentials in biological compatibility and underwater adhesion, further research will standardize and modularize more parts.
For now, improving yields should be scientists’ primary goals before applying these new materials to daily lives. The recent progress of bioplastic well exemplifies this.
Yet, this is only the beginning of a long and fantastical journey. ‘Magics’, like living organisms producing glues on various living and non-living surfaces, is emerging from 21st century’s horizon. Bacteria and viruses are future pioneers of nano-robots and biofouling. Production of proteins from other organisms in the human body will be able to cure certain diseases, such as diabetes.
We sincerely hope that this project will inspire future biologists into researching recombinant proteins and provide them a new perspective. This will be our pathway to future Biology and Materials Science.
Reference
Choi, Y. S., et al. (2012). In vivo modification of tyrosine residues in recombinant mussel adhesive protein by tyrosinase co-expression in Escherichia coli. Microb Cell Fact 11:139.
Do, H., et al. (2017). A tyrosinase, mTyr-CNK, that is functionally available as a monophenol monooxygenase. Sci Rep 7(1):17267.
Hwang, D. S., et al. (2007). Practical recombinant hybrid mussel bioadhesive fp-151. Biomaterials 28(24):3560-8.
Kim, E., et al. (2018). Microbially Synthesized Repeats of Mussel Foot Protein Display Enhanced Underwater Adhesion. ACS Appl Mater Interfaces 10(49):43003-43012.
Liang, C., et al. (2015). Protein Aggregation Formed by Recombinant cp19k Homologue of Balanus albicostatus Combined with an 18 kDa N-Terminus Encoded by pET-32a(+) Plasmid Having Adhesion Strength Comparable to Several Commercial Glues. PLoS One 10(8):e0136493.
Zhong, C., et al. (2014). Strong underwater adhesives made by self-assembling multi-protein nanofibres. Nat Nanotechnol 9(10):858-66.
