Team:Lethbridge/Results


Results



Overview: What We Accomplished


  • Tested multiple chassis for our system in order to provide the best chance for being able to pass the GI tract
  • Was able to determine a potential storage method of microalgae to increase the shelf life of our system
  • This season we were able to express, purify and run testing on three proteins
  • Were able to confirm binding of insulin variant SCi57 while still fused to the RFP protein


Microalgae Degradation Tests


As the stomach is a harsh environment, sitting at an average pH of ~1.5 (Beasley et al., ), we needed to ensure that our chassis of choice could handle the low pH over time. For that reason we ran degradation tests on our initial chassis of choice Arthrospira platensis aka spirulina and Chlamydomonas reinhadtii. Our results showed that in similar pH conditions to the stomach, Spirulina and C. reinhardtii are drastically degraded in these harsh conditions.





Figure 1: Arthrospira culture placed in growth media modified to different pH points. The OD750 was measured over time. The blue line represents the population in media at pH 2.5 and the orange line represents the media at pH 7.0. n=3





Figure 2: C. reinhardtii cultures placed in growth media modified to different pH points. The OD750 was measured over time. The blue line represents the population in media at pH 2.5 and the orange line represents the media at pH 7.0. n=3



As compared to the neutral pH media control, both species showed a drastic decrease in cell populations over the three hour time period in the acidic pH. This data was confirmed by student t-test analysis to determine significance. Due to this significant decrease in cell turbidity, these species of microalgae are therefore unable to survive in the stomach.

Therefore we decided to find a new chassis of choice. After our interview with Dr. Stephen Rader he gave us a culture of Cyanidioschyzon merolae as a gift and we were able to test this species with our previous degradation protocol






Figure 3: C. merolae cultures placed in growth media modified to different pH points. The OD750 was measured over time. The blue line represents the population in media at pH 2.5 and the orange line represents the media at pH 7.0. n=3



As expected, the turbidity of the C. merolae culture was stable and did not significantly change over time. The pH treatments of C. merolae confirmed that it would likely be stable in the stomach's pH conditions, making it a perfect chassis for our work and proof of concept for our project.



Lyophilization Testing


Another challenge of our project was the ability for us to compete for the storage times that purified insulin can currently provide (up to one year). Our first option was the use of a lyophilizer to freeze dry the microalgae cultures. Using small sample sizes we tested this. Our results showed that there is a potential for lyophilization to retain the insulin protein stability.





Figure 4: C. merolae samples lyophilized for 21 hours ran on a 10% SDS-PAGE to determine if protein degradation occurred. The three samples were compared to the control mean and quantified using ImageJ software.



Although the integrity of the proteins seems valid, where a degradation of the reference protein was not significantly changed, the attempt to deploy it in the acidic conditions showed less than optimal results. Once we retested the cells in low pH levels after lyophilized for the 21hour period, the results showed that the cell wall was ultimately compromised. Although we decided to do a quick qualitative test, the yellowing of the culture indicated that cell death had occurred. A blue colored supernatant was also observed which is hypothesized to be chlorophyll proteins entering the medium indicating that the cell wall was severely damaged after lyophilization and pH treatment.



Figure 5: Left: lyophilized samples resuspended in pH 2.5 allen media and incubated for 3 hours at 37 degrees Celsius. Center: Lyophilized samples resuspended in pH 8 allen media and incubated for 3 hours right: lyophilized samples resuspended in pH 7 medium and incubated for 3 hours.

Glycerol Storage


Our second storage idea was to determine if storage in glycerol and then held at low temperatures would affect the microalgae. It should be noted that the intake of glycerol by diabetics does not affect glucose levels and is not harmful unless in incredibly large dosages (Thornit et al., 2009). As like before, we then added the samples to pH changed media and recorded OD at 750nm over time.



Figure 6: C. merolae cultures under different pH treatments overtime. n=3



As seen compared to the control, the initial population had dropped, likely due to the flash freezing. Although sugar based cryogenics are suggested for microalgae, it does not prevent a substantial amount of cell death (Aujero & Oseni, 1979). However, there seems to be a stability retained over time. With statistical analysis, the slight decrease in acidic conditions was confirmed to be significant while variations in neutral pH and basic pH were not. We will have to continue testing to optimize the storage and determine if glycerol and freezing of the culture is supportive of delivering enough insulin into the body through the digestive system.



Insulin Purification and Activity


Although we were able to design a various amount of constructs that include expression in both E. coli and C. reinhardtti (https://2019.igem.org/Team:Lethbridge/Parts), this season we were only able to study three proteins in E. coli. Our bottleneck for our project was the biobrick cloning and the short time scale when attempting to clone and screen microalgae. It should be noted that although we have chosen the chassis C. merolae our microalgae constructs have been codon optimized for our original chassis C. reinhardtii.



We were able to purify three proteins this season: mRFP-SCI57 (BBa_K2148018), mRFP1-Proinsulin (BBa_K2148016) and the insulin receptor protein (BBa_K2148013).

In order to determine if our insulin was active after processing we decided to express and purify the insulin receptor protein. Our affinity purifications were done using either the AKTA START system or by gravity column. All of the size exclusion purifications were done using the AKTA PURE system



Figure 7: 16.5% Tricine gel of the Nickel affinity chromatography of the receptor protein run on a gravity column.



The monomeric receptor was observed at the bottom of the selected region but the band at 25kDa was larger, which may indicate that receptor protein might have interacted with itself, forming a tetramer.



Figure 8: Size Exclusion purification chromatogram of receptor in SEC column. The blue line indicates the A280 absorbances and the orange line indicates the A260 absorbance.



Figure 9: 16.5% tricine gel of Size Exclusion Chromatography of receptor protein.



The last elution of the SEC purification showed a thicker band at 25kDa which we suggested is the tetrameric receptor protein.



The first insulin derivative we chose as a proof of principle was the proinsulin with modified cut sites first proposed by Tsinhua 2014. With the insertion of the furin cut sites, our in vitro studies were easier as only one protease for insulin processing was needed and (as it is a human protease) ensures that the proinsulin would be processed (Hay and Dockerty, 2003). in vivo.The proinsulin construct was expressed in E. coli and purified using nickel affinity chromatography and size exclusion chromatography.



Nickel affinity chromatography utilizes electrostatic interaction between nickel ion and histidine residues of hexahistidine tag in recombinant proteins to purify the target proteins from non-specific proteins. The DNA sequence for our recombinant insulin and receptor proteins were designed to have the hexahistidine tag sequences at the N terminus. Hence, nickel affinity chromatography technique was used to check the translation of histidine tag from DNA sequences into amino acid sequences.





Figure 10: 10% SDS-PAGE of proinsulin gel. The gel was initially run at 100V for 10 minutes on 6% stacking gel and run at 200V for 50 minutes on 10% separating gel. Then, the gel was stained with coomassie.



Figure 11: Nickel affinity chromatogram of proinsulin. The chromatogram was created by ÄKTA Start. Binding buffer contained 10mM imidazole and elution buffer contained 250mM imidazole. At wash phase with 30mM imidazole in the solution (91.6% binding buffer and 8.4% elution buffer), insignificant amount of proinsulin and most of non-specific cellular proteins were eluted. At elution phase with 118mM imidazole in solution (55% binding buffer and 45% elution buffer), the insulin with histidine tag was eluted.



Figure 12: 10% SDS-PAGE of SCI gel. The gel was initially run at 100V for 10 minutes on 6% stacking gel and run at 200V for 50 minutes on 10% separating gel. Then, the gel was stained with coomassie.





Figure 13: The chromatogram was created by ÄKTA Start. Binding buffer contained 10mM imidazole and elution buffer contained 250mM imidazole. SCI was also eluted at 118mM imidazole.



The 10% SDS gel showed that histidine tag of our insulin proteins was successfully translated. However, the size of the protein was slightly smaller than expected size. We had two possible reasons for the truncation. One is the protease activity that caused the truncation of the protein, or the proteins were not stable at storage conditions. The optimized protein storage condition was not established due to insufficient amount of time.


The absorbance of A280 detects the presence of Trp and Tyr residue or disulfide bond of Cys-Cys residue of proteins. Proinsulin has 10 and 14 residues of Trp and Tyr (about 10% of the entire sequence) and SCI has 3 and 15 residues of Trp and Tyr (about 6% of the entire sequence), so the peaking of A280 reading provided information that the histidine tag of insulins worked as expected in nickel affinity chromatography.


For the proinsulin construct, the mRFP1 and three different insulin peptides are linked with furin cut sites. Hence, the three different insulin peptides were required to further analysis whether they form a complex or they remained as three individual peptides. Similar to the proinsulin construct there is a furin cut site between the mRFP1 and the insulin analog. SCI57 is a single chain analog (Hua et al., 2008) and therefore needed no more processing.


Because of multiple bands appeared on 10% SDS gel, the insulin proteins were further purified with SEC. However, the gel picture and chromatogram of the size exclusion of the insulin with RFP was not provi ded because the strong turbidity of the RFP protein sample caused the faulty absorbance reading at 280 and 260nm. We recovered the SEC fractions that had bright red color and ran another gel to confirm the identity of the protein, but about similar amount of impurities were eluted with our insulin proteins.


Since our objectives was not obtaining the homogenetically purified insulin, we attempted to process the proinsulin and SCI57 with the Furin protease (which both removes the his-mRFP1 and process the proinsulin into the bioactive form)





Figure 14: 16.5% Tricine gel of insulin constructs treated with furin. 1 unit of furin was added to the 500uL of insulin samples and incubated for 16 hours in 25oC



Figure 15: A 16.5% Tricine PAGE with mRFP1-Proinsulin and mRFP1-SCI57 treated with Furin Protease at 25 degrees Celsius for 16 hours. The gel was run at 100V for 10 minutes, then 180V for 60 minutes and stained with coomassie. Lane 1= 10-240kDA marker (Biobasic), lane 2= mRFP1-Proinsulin 10 units Furin, Lane 3= mRFP1-Proinsulin 20 units Furin, Lane 4= mRFP1-Proinsulin untreated, Lane 5= mRFP1-SCI57 10 units Furin, Lane 6= mRFP1-SCI57 20units Furin, Lane 7=mRFP1-SCi57 untreated.



Unfortunately, the Furin was unable to cut neither the mRFP1-Proinsulin or mRFP1-SCI57 protein. Proinsulin was expected to have bands at 3.43, 2.58, and 2.38kDa and SCi57 was expected to be seen at 6.508kDa. We hypothesized that perhaps there was some type of hinderance to furin binding due to the proteins structure. We were able to model this and found that this was not the case. Future work in determining the cutting issue will be necessary.



Since the Furin protease was unable to cleave both the mRFP1-Proinsulin and mRFP1-SCI57 insulin proteins we decided to move ahead and test SCI57. As this insulin is a single chain there is a chance that it is still active despite the mRFP1 still being fused.



In order to determine if binding was occurring we ran both a MicroScale Thermophoresis assay (MST) and an electrophoretic mobility shift assay (EMSA) to confirm binding.



MST analysis




Figure 16: A time scale graph showing the MicroScale Thermophoresis assay indicating protein-protein interactions between the mRFP1-SCI57 protein and the insulin receptor. The blue data indicates fluorescence reads with just the mRFP1-SCI57 construct at 500nM concentration. The green data represents the samples containing the mRFP1-SCI57 and insulin receptor at 500nM concentration and ~17,000nM receptor respectively. The increase in fluorescence reads with the mRFP1-SCi57 + insulin receptor protein indicates a binding event between the two proteins.



The time graph shows that there is a slight shift in the fluorescent reads indicating a binding event. This was also confirmed by the Monolith.NT equipment software. This experiment was done twice with the same results. However, we hope to retest the construct without RFP to ensure there are no binding of the receptor protein with the RFP as a control.



Electrophoretic Mobility Shift Assay


We wanted to complement our MST data with a protein-protein EMSA. However, as seen in Figure 17, our results were inconclusive. Due to time constraints we were unable to optimize our conditions to the shift clearly.



Figure 17: An Electrophoretic Mobility Shift Assay of insulin receptor and mRFP1-SCI57 run on a 12.5% Native PAGE at 60V. Samples were run from high to low concentrations of mRFP1-SCI57 protein and a constant amount of Receptor Protein. Lane 1= mRFP1-SCI57 protein, Lane 2=5000nM mRFP1-SCI57, Lane 3= 500nM mRFP1-SCI57, Lane 4= 50nM mRFP1-SCI57, Lane 5= 5nM mRFP1-SCI57, Lane 6= .5nM mRFP1-SCI57, Lane 7= .05nM mRFP1-SCI57, Lane 8= .005nM mRFP1-SCI57, Lane 9= .0005nM mRFP1-SCi57, Lane 10= Receptor protein only.



Overall, we have been able to express and purify two insulin proteins from E. coli and have attempted to test activity with the insulin activity with one protein. For algae, we have been able to find a suitable chassis for our work and have started optimizing storage conditions for microalgae for future work and for production storage that is safe for use and safe for diabetics.



Future Directions


  • Although were were able to express proteins in a pUC57 plasmid, we hope to take our finalized cloned constructs in pET58a(+) and reconfirm expression and purification of our proteins
  • determine why Furin protease is not cutting the proteins and optimize the cutting. As well, relating this to biological abundance of furin as a part of modelling final bioactivity and protein levels after delivery
  • Optimize binding to better inform the activity of our process proteins, especially the proinsulin variant we have been working with
  • Move towards expressing our insulin in microalgae and optimizing our microalgae system including the standardization of the insulin expression, storage, and dosage modelling


    • References


    Aurojero, E., & Oseni, M. (1979). Viability of frozen algae used as food for larval penaeids. Aquaculture Department Quarterly Research Report, 3, 11-16

    Beasley, D., Kilts, A., Lambert, J., Fierer, N., and Dunn, R. (2015) The Evolution of Stomach Acidity and Its Relevance to the Human Microbiome. PLOS ONE. DOI:10.1371/journal.pone.0134116

    Hay, C., and Docherty, K. (2003) Enhanced expression of a furin-cleavable proinsulin. Journal of molecular endocrinology. 3, 597-607.

    Hua, Q., Nakagawa, S., Jia, W., Huang, K., Philips, N., Hu, S., and Wiess, M. (2008) Design of an active ultrastable single-chain insulin analog: synthesis, structure, and therapeutic implications. The Journal of Biological Chemistry. 21, 1473-14716.

    Thornit, D., Sander, B., la Cour, M., & Lund-Anderson, H. (2009). The effects of peroral glycerol on plasma osmolarity in diabetic patients and healthy individuals. Basic & Clinical Pharmacology & Toxicity, 105, 289-293.