Design: We derived our backbone from the pET28:GFP plasmid via inverse PCR [1]. pET28:GFP contains a prokaryotic origin of replication, kanamycin antibiotic resistance, green fluorescent protein (GFP), a T7 promoter and a T7 terminator. The E. coli codon optimized Ex-4 sequence was inserted into the space occupied by GFP on the pET28:GFP plasmid. Our Ex-4 gene insert includes: BsaI restriction enzyme site, Shine-Dalgarno sequence, His-Tag, linker sequence, enterokinase cut site, and the Ex-4 codon optimized sequence. To learn more about each of the components of our Ex-4 gene insert, check out our Contributions page!
Build: Golden Gate Assembly (GGA) was used to insert Ex-4 into the plasmid backbone. GGA is a modular assembly method by which multiple DNA fragments can be simultaneously integrated into a plasmid backbone [2]. GGA exploits type IIS restriction enzymes, which cleave DNA surrounding the recognition site, so the final recombinant product has no indication of integration. All type II restriction enzymes leave sticky ends on the backbone and the gene insert. These sticky ends bind to one another since they are designed to have complementary nucleotide sequences. Specifically, type IIS restriction enzymes cut downstream of the binding site by a certain number of nonspecific nucleotides. This number varies based on the type IIS restriction enzyme used. This means the sticky ends can be engineered to seamlessly integrate DNA without inserting residual nucleotides from the restriction enzyme binding site. This is deemed “scarless integration”, and prevents mutagenesis during transcription of the amino acid sequences. The deletion of restriction enzyme sites also means digestion and ligation can be accomplished simultaneously.
Test: Once Ex-4 containing plasmids were constructed via a one step digestion and ligation GGA, they were transformed into DH10B E. coli. No colonies grew after this first GGA attempt.
Design & Build: Consequently, we decided to try a longer thermocycler protocol that cycles through digestion and ligation steps to increase the probability of successful plasmid construction. For this next attempt using GGA, we pursued both the shorter and longer protocols.
Test & Learn: These plasmids were again transformed into DH10B E. coli and incubated overnight. Colonies grew after this transformation, implying the longer thermocycler protocol may have contributed to a higher transformation efficiency. After testing both settings, we decided to use the longer thermocycler setting during GGA for constructing our pET28:Ex-4 plasmid. After we performed colony PCR on the DH10B transformants and performed gel electrophoresis, we sent samples with bands at the expected size to sequencing. Of the seven samples sent to sequencing, only one contained a mutation. Plasmid samples with our Ex-4 inserts displaying no mutations in sequencing results were then transformed into the protein producing E. coli BL21 and C41.
Design: For the construction of our S. cerevisiae genomic insert, we used a GAL1 promoter and a PGK1 terminator. The GAL1 promoter and PGK1 terminator sequences are derived from the yeast Modular Cloning (MoClo) library [3]. Protein production is initiated by the GAL1 promoter when S. cerevisiae is grown up in a galactose based liquid media.
We used the plasmid backbones mg-int-leu2-kan-1420 (hereafter referred to as pLEU) and mg-int-trp1-hyg-1432-4a (hereafter referred to as pTRP) for eventual genomic integration of our Ex-4 gene insert. pLEU has an origin of replication for E. coli, a yeast G418 resistance as a selection marker and LEU2 homology sites for integration into the host genome. pTRP is identical to the previous plasmid, except that it has hygromycin resistance in place of G418 resistance and TRP1 homology sites in place of LEU2. This means that this plasmid is integrated at a different spot inside the genome. Our insert was placed in between the two BsaI cut sites [4] already present on both the plasmid backbones. Because the promoter and terminator were cut out by our restriction enzyme digest, we added these parts as well during GGA.
The S. cerevisiae Ex-4 sequence was optimized through IDT [5]. An elastin-based polypeptide half-life extender sequence was added to our insert on the C-terminus. The half-life extender is composed of a ten-fold repeat of the VPGVG amino acid sequence, and increases the half-life of Ex-4 by 3.7 times [6]. We hypothesize this may allow for increased bioavailability of orally administered Ex-4. To learn more about each of the components of the S. cerevisiae genomic insert, check out our Contributions page!
Build: Our genomic insert for S. cerevisiae was assembled using the yeast MoClo Library [3]. This pivotal research in yeast enables rapid design of plasmids for protein expression. The MoClo library provides resources for easy, one-step constructions of yeast plasmids through GGA. Predesigned plasmid parts can be interchanged quickly, allowing for easy testing of multiple plasmid constructs.
The plasmid backbones, GAL1 promoter, Ex-4 insert, and PGK1 terminator were assembled simultaneously through a one tube GGA reaction. These customized plasmids were subsequently grown in DH10B E. coli post GGA and then introduced into our yeast strains. We transformed pLEU:Ex-4 into S. cerevisiae strain number ROY 6783. ROY 6783 has a wild type LEU2 gene in the chromosome, allowing integration and auxotrophic selection within the leucine production site of the genome. For the pTRP:Ex-4 plasmid, we utilized S. cerevisiae strain number ROY 4513. ROY 4513 has a wild type TRP1 gene in the chromosome, allowing integration and auxotrophic selection within the tryptophan production site of the genome.
Testing: We performed transformation of S. cerevisiae with the linearized integrative plasmids purified from our DH10B E. coli colonies. An additional linear plasmid was utilized in the pLEU:Ex4 transformation culture. This plasmid contained the CRISPR-Cas9 protein coding gene and a single guide RNA targeting the LEU2 integration sites. The CRISPR-Cas9 protein cleaves and creates a chromosome break on the integration sites, which promotes homologous repair and increases the efficiency of integration of the pLEU plasmid.
The genomic DNA samples from pLEU:Ex4 and pTRP:Ex4 transformants were used for PCR reactions to check for gene insert integration via amplification. For each template DNA, the forward primer on the GAL1 promoter and the reverse primer on the PGK1 terminator were used to amplify the complete sequence of Ex-4. Bands at the expected size confirmed the presence of the desired protein coding sequence. Another set of primer pairs designed around the LEU2 homology region and TRP1 homology region were run in separate PCR reaction mixtures to amplify the wildtype and integrated sequence of LEU2 and TRP1 integration sites. Lastly, the primer pair used previously to confirm GGA before the S. cerevisiae transformation step was used as a positive control.
PCR products of the Ex-4 amplicon from the S. cerevisiae genome were prepared and sent for single primer extension Sanger Sequencing. The same forward primer and reverse primer used for PCR reactions were used for separate sequencing on the forward and reverse strands of each sample. Lastly, we lysed our cells to extract Ex-4 and purified it through immobilized metal affinity chromatography (IMAC). We ran our purified protein on an SDS-PAGE to see if there was a His-tag protein at the expected size of our Ex-4 protein.
Learn: Our genomic DNA gel electrophoresis showed bands at the expected nucleotide size for all sets of primers at both the LEU2 and TRP1 sites. To confirm that these nucleotide segments matched what we digitally designed, we sent our PCR products for Sanger Sequencing. We observed a 100% match on the Ex-4 protein coding region from all four sequencing results. Our SDS-PAGE results further confirmed that there was a His-tag protein of the appropriate size being produced by our yeast. Our work with S. cerevisiae also taught our team how to successfully perform a genomic integration of DNA into >S. cerevisiae.
Design: A codon optimized sequence for the GLP-1 receptor (GLP-1 R) was designed for E. coli to test the binding affinity of our Ex-4 in both E. coli and S. cerevisiae. This sequence was inserted into the same pET28:GFP backbone as the Ex-4 insert for E. coli. The gene insert for the GLP-1 R includes: BsaI restriction enzyme site, Shine-Dalgarno sequence, His-tag, linker sequence, TEV cut site, and GLP-1 R codon optimized sequence. The insert includes a TEV cut site instead of an enterokinase cut site, but both have equivalent proteolytic functions. We inserted an alanine in between the TEV cut site and the GLP-1 R sequence because TEV protease will only cut before serine, glycine, alanine, methionine, cysteine, or histidine amino acids [7]. To learn more about each of the components of the GLP-1 R gene insert, check out our Contributions page!
Build: The GLP-1 R was inserted into the pET28:GFP backbone through GGA. Initial assembly protocol suggested a 1:1 ratio of insert to backbone. We transformed our GGA assembled plasmids into DH10B E. coli, and incubated overnight with the expectation of seeing colonies the next day. To confirm the colonies we grew contained successfully ligated GGA plasmids, we performed colony PCR and gel electrophoresis. The GLP-1 R insert is 505 base pairs long, so using a 10kb ladder we were expecting bands to show up near the bottom of the gel.
Test & Learn: Our gel electrophoresis had bands well below the expected position, indicating that our GGA was not successful in correctly assembling the GLP-1 R plasmid. The bands traveled far enough down the gel that they seemed to be 100 base pairs long. We hypothesized that our PCR primers ligated together, creating primer dimers, and the GLP-1 R sequence was not successfully inserted into the pET28:GFP backbone.
Design: To test this hypothesis, we performed GGA with a 2:1 ratio of insert to backbone with the hope that increasing the number of GLP-1 R inserts would increase the probability of successful plasmid construction.
Build & Test: We transformed these GGA reactions into DH10B E. coli as we had done previously, and let them grow overnight. We performed colony PCR and gel electrophoresis to determine if the new GGA reactions were successful.
Learn: The gel electrophoresis showed the same bands as the previous iteration; the GGA primers formed primer dimers and the pET28:GFP backbone annealed to itself. We then decided to send our GLP-1 R insert and pET28:GFP backbone to sequencing to determine if the ends of the backbone or insert degraded, thus preventing proper plasmid assembly. The sequencing results for the backbone indicated that our backbone did not degrade, however, the sequencing results for the insert were inconclusive.
Design: We decided to try 2:1 GGA again, but this time, use fresh reagents and amplify our GLP-1 R insert to ensure the GLP-1 R insert's ends are correct.
Build: We performed PCR to amplify our GLP-1 R insert and used this amplified insert in the 2:1 GGA reaction. These GGA reactions were transformed into DH10B E. coli and incubated overnight. We performed colony PCR and gel electrophoresis as we had done in the last iteration.
Test: For the first time, we saw bands at the expected position of around 500 bases! We excitedly performed miniprep on those colonies to extract the plasmid and sent them to sequencing to further confirm that the plasmid was constructed correctly. Sequencing results concluded that we properly assembled the GLP-1 R insert into the pET28:GFP backbone. These iterations taught us that GGA is finicky and can fail in multiple ways. Our success came from improving upon multiple iterations of the engineering cycle. Our next steps were then transforming into C41 E. coli, and performing IMAC and SDS-PAGE.
Learn: The SDS-PAGE had a band in the pellet at 17kD, the expected weight of the GLP-1 R. This suggests that the receptor is insoluble in E. coli. Our next steps include refolding the receptor in E. coli, so it will bind correctly to our Ex-4.
Design: To test that our recombinant Ex-4 proteins function as expected, we planned an Octet binding affinity assay between Ex-4 and GLP-1 R. The Octet machine determines the dissociation constant (kd) of a protein-ligand interaction using antibodies and a fiber optic biosensor [8]. First, this test requires the removal of one of the binding partner’s His-tag using a protease. Next, Penta-His antibodies on the biosensor tip attach to our His-tagged protein. Then, the tip submerges into the non-tagged protein sample, allowing binding reactions to occur. We decided to remove the E. coli Ex-4 His-tag and attach our purchased GLP-1 R on the Octet biosensor for our first round of testing.
Build: Contaminating salt content was removed from the chosen His-tagged Ex-4 samples using diafiltration with 3kD spin filters. After our Ex-4 was resuspended in Tris-HCL buffer, we incubated the protein with enterokinase overnight to cleave Ex-4’s His-tag. Batch purification isolated tag-free Ex-4, which was then analyzed via spectrophotometry at A280 to determine protein concentration.
Test & Learn: Protein concentration of our sample was around 0.001mg/mL, and did not indicate successful isolation of Ex-4. The 3kD spin filters we used during diafiltration may not have collected our 6.4kD Ex-4, as indicated by the DuBois lab who had similar issues isolating small proteins.
Design: Since a 1kD spin filter would have been difficult for our team to obtain, we hypothesized the His-tags may be kept on both binding partners. Ensuring the Octet biosensor was fully saturated in His-tagged GLP-1 R, and the Ex-4’s His-tag was shielded using a His-tag antibody, we believed an accurate binding reaction would still occur.
Build & Test: We performed the binding affinity assay using an Octet machine under the conditions described on the previous page. Controls for this assay included an unloaded sensor submerged into Ex-4 and a GLP-1 R loaded sensor submerged into His antibodies.
Learn: The graph produced in real time from this assay indicated this experiment was not successful, suggesting binding did not occur. Our next step is to obtain a 1kD spin filter to properly capture Ex-4 so we can perform the binding affinity assay under the proper condition where only one binding partner has a His-tag.
Design: The first model we wrote predicted Ex-4's binding affinity in the presence of its endogenous competitor, GLP-1, aptly named the competitive binding model. A crucial step in this process was deciding which formula best described our system, where both Ex-4 and GLP-1 generate the desired pharmacological effect. After speaking with Dr. Michael Stone, a biochemistry professor at UC Santa Cruz, we chose a variation of the Cheng-Prusoff equation [7] for its history of representing competitive binding interactions. For all parameters involved (e.g. IC50, Ki), we found literature values by searching through online databases such as PubMed and Google Scholar.
Build and Test: We programmed the competitive binding model in Python 3 using Jupyter Notebook. While we initially attempted to use MATLAB, Python was ultimately the better choice due to our team's past experience with the language. We ran the code and adjusted our graph’s specifications for clarity of our model.
Learn: We learned this model does not appropriately represent our system because it does not account for the effect of GLP-1 and Ex-4 both producing insulin. The Cheng-Prusoff equation tells us the concentration at which Ex-4 inhibits 50% of the natural ligand binding. We changed our equation to fractional occupancy because it tells us how many ligands are bound compared to the total amount of receptors in a system. This way, we would be able to more accurately compare the competitive binding between Ex-4 and GLP-1.