Team:Virginia/Results
Loading
#ffd6a5
Results
Index:
Overview
Bacterial microcompartments (
BMCs
Bacterial microcompartments (BMCs) are organelle-like structures, consisting of a protein shell that encloses enzymes and other proteins. BMCs are typically about 40–200 nanometers in diameter and are entirely made of proteins. The shell functions like a membrane, as it is selectively permeable.
) were produced and extracted. The BMC shell size distribution was calculated and growth curves were determined for the cells transformed with the BMC part. Protein assays and western blot were performed to assess BMC protein concentration and identity.
Bacterial microcompartments (BMCs) are organelle-like structures, consisting of a protein shell that encloses enzymes and other proteins. BMCs are typically about 40–200 nanometers in diameter and are entirely made of proteins. The shell functions like a membrane, as it is selectively permeable.
The r_oligo flipper was successfully tested via RFP colony selection and used to create DNA scaffolds containing HTBS and zinc finger binding sequences. DNA scaffolds were produced via cotransformation of r_oligo and reverse transcriptase parts, extracted, and analyzed on agarose and polyacrylamide gels.
In order to demonstrate the functionality of the Manifold system, the BMC part, the DNA scaffold parts, and parts for various enzymes in the resveratrol biosynthetic pathway will be cotransformed in different combinations to test different aspects of the system.
Results
BMC Production
Our team was gifted the “pduABJKNU” and “pduABJKNUT” sequences for production of bacterial microcompartment (BMC) shells by the Warren Lab at the University of Kent in Canterbury, England. For our experiment, we opted to use the pduABJKNUT plasmid due to its enhanced results from prior studies and strong modeling evidence. Initially, production and extraction of both BMC plasmids was performed. BMCs were produced in order to verify that the plasmids provided by the Warren lab functioned as intended, to verify that we could successfully follow current BMC production methods for production and isolation without issues, and to obtain information regarding BMC size distribution under the culture conditions utilized. These plasmids were individually transformed into E. coli and BMC production was induced. BMCs were extracted and visualized via transmission electron microscopy (TEM). The images of extracted BMCs in Figure 1.1 and TEM images in Figure 1.2 highlight the successful extraction and visualization of the BMC plasmids, respectively.
Fig 1.1: Bacterial microcompartments were extracted from E. coli. The left image shows extracted pduABJKNUT BMCs and the right image shows extracted pduABJKNU BMCs. Further clarification of successful extraction was provided after the TEM images were taken.
Fig 1.2: TEM images of empty pdu BMCs. a.,c. pdu BMCs were composed of proteins pduABJKNU b.,d. pdu BMCs were composed of proteins pduABJKNUT. Images 4a and 4c were taken from the same sample, as were 4b and 4d. Scale bars of 0.5 μm were used in Figure 4a, 4b and 4d, while a scale bar of 1 μm was used in 4c.
According to the TEM images, the pduABJKNU BMCs were produced and isolated in a higher quantity than pduABJKNUT BMCs. In Figure 1.2c, 386 BMCs are present, resulting in a concentration of around 12.83 BMCs/μm2. In Figure 1.2d, 353 BMCs are present, resulting in a concentration of 7.40 BMCs/μm2. This observation was also qualitatively supported by the relative sizes of the protein pellets during BMC extraction (Figure 1.1). The images in Figure 1.2 were used to determine the distribution of BMC sizes in the two samples. These results are shown in Figure 1.3.
Fig 1.3: Distribution of empty BMC shell sizes for pduABJKNU and pduABJKNUT samples. There are 386 data points for the pduABJKNU BMCs and 353 data points for the pduABJKNUT BMCs.
In addition to the greater production of pduABJKNU BMCs, its protein concentration proved to be higher as well. The protein assays shown in Figure 1.4 were run for the two different BMCs to determine protein concentration. After the assays were completed and protein concentration was measured, the pduABJKNU Old BMCs relayed a higher protein concentration than pduABJKNUT BMCs, in which concentrations are 1.56 μg/mL and .27 μg/mL respectively. Results were run on a western blot in duplicates, and due to the difference in protein concentrations between the BMCs, only results from the pduABJKNU Old BMC appeared. The sizes of proteins are as expected for pduABKJNU BMCs. The western blot is shown in Figure 1.5.
Fig 1.4: Protein assays of pduABJKNU and pduABJKNUT BMCs, both Old and New. The concentration of protein in each of these samples (1.56 μg/mL for pduABJKNU and .27 μg/mL for pduABJKNUT) were verified by the western blot shown in Figure 1.5.
Fig 1.5: Western blot run in duplicates of pduABJKNUT Old, pduABJKNU Old, pduABJKNUT New, and pduABJKNU New.
In order to obtain useful data about the growth rates of transformed BL21(DE3) cells in Ampicillin, growth curves were measured for each of the cell lines. The resulting curves are shown in Figure 1.6. Logistic regressions were then fitted to each of these curves to allow for the doubling times of the cell lines and carrying capacities in the LB solutions to be calculated. These findings are reported in Table 1.
Fig 1.6: Growth curves for various BL21(DE3) transformants plotted against logistic regressions. The error bars assume a constant error of 5 minutes due to variation in sampling, and an error of 5% in the OD600 measurements due to observed imprecision in the spectrometer utilized. All samples were grown in LB medium containing 50 μg/mL Ampicillin except for “BL21(DE3)” (shown in blue) to which no Ampicillin was added. Plots labeled “New” were made from plates that had been stored at 4°C for one day, whereas plots labeled “Old” were made from plates that had been stored at 4°C for 6 days. The plot labeled “BL21(DE3)” was also made from a plate that had been stored at 4°C for one day.
Table 1: Calculated parameters and statistical analysis of values obtained from the logistic models shown in Figure 1.6. The samples here correspond to the same cell lines in Figure 1.6. The parameters a, k, and L in the “Values” column refer to the regression parameters as shown in Equation 1. The doubling values were calculated as 60*ln(2)/k. The carrying capacities simply refer to the value of L. And the shifts from control refer to the difference between the doubling time of sample and the doubling time of sample “BL21(DE3).” All errors were calculated according to 95% confidence intervals under a Student’s T Distribution, and error was propagated using the standard error formulas for normally distributed variables which are appropriate for use, as the t-distributions will be approximately normal under the given conditions.
Equation 1: Three-parameter logistic equation used to compute growth curve regression models. Parameters are L, a, and k. The dependent variable is OD600 and the independent variable is t.
The growth curves shown in Figure 1.6, and the corresponding doubling values shown in Table 1 imply that there is a slight difference in the doubling time between the transformed BL21(DE3) cells in ampicillin and the untransformed BL21(DE3) cells. The transformed cells appeared to grow slightly faster on average when compared to the untransformed cells. As of now, there is no proposed mechanism for why this should be true, but it is possible that it is due to random variation in the cell lines. Additionally, there was no statistically significant difference observed between the new and old doubling times of the pduABJKNUT plasmid containing cells. Furthermore, all of the observed doubling times align with the expected doubling times for E. coli in LB medium between 20 and 30 minutes. It should also be noted that no growth was seen in the “pduABJKNU New” culture which is most likely due to the inoculation of a dead colony. The observed carrying capacities also seem to align with values found in the existing literature for most for E. coli strains when cultured in LB medium.
DNA Scaffold Production
We successfully tested the functionality of the “empty” r_oligo flipper via RFP colony selection. In Figure 2.1, the fluorescent colonies are those with the r_oligo flipper containing RFP instead of the sequence for a particular DNA scaffold, which in the case of Figure 2.1 is the CoA 4x part. The colonies without fluorescence are those with the scaffold sequence successfully inserted into the r_oligo flipper via a single Golden Gate Assembly reaction. Transformed E. coli were grown on ampicillin plates as the r_oligo flipper has a backbone containing ampicillin resistance.
Fig 2.1: Red E. coli colonies are those expressing the original r_oligo flipper, which contains RFP instead of a DNA scaffold insert. Non-fluorescent colonies are those expressing the r_oligo flipper containing a scaffold gene insert (CoA 4x) instead of RFP. Cells were grown on ampicillin plates.
The r_oligo flipper was used to create two versions of the scaffold part via several successive Golden Gate Assembly reactions. One contains two copies of the ZF-binding domains for resveratrol enzymes and six copies of those for Malonyl CoA-recycling enzymes (Resv 2x, CoA 6x). Another contains four copies of these domains for both pathways (Resv 4x, CoA 4x). The gel in Figure 2.2 shows successful creation of both of these parts.
Fig 2.2: Diagnostic digest of the (Resv 4x, CoA 4x) in lanes 2 and 3, and of the (Resv 2x, CoA 6x) part in lane 4. The expected results for all three parts are two fragments of about 1400 and 2100bp each.
An RT part was needed to allow for reverse transcription of r_oligo genes with HIV reverse transcriptase (HIV-RT) and Murine leukemia reverse transcriptase (ML-RT) to produce DNA scaffolds. While only HIV-RT is needed to produce the scaffolds, it has been documented that the co-expression of ML-RT leads to an increase in DNA production. Thus, we assembled the HIV-RT part containing p66 and p51 subunits with the ML-RT part using a three-insert Golden Gate assembly.
We validated the success of our Golden Gate assembly through the lack of fluorescence of colonies and a diagnostic gel. We assessed the functionality of our assemblies by co-transforming E. coli with r_oligo plasmids and reverse transcriptase plasmids, including the full parts, composites and forward sections.
Fig 2.3: The RT Full Part digested with XbaI and SpeI (expected sizes 2196bp and 5645bp). Lanes 5, 6, and 7 each have a different sample of the same RT Full Part plasmid.
Fig 2.4: The RT full part digested with only Xbal (expected size 7841bp). Lanes 5, 6, and 7 correspond to different samples of the RT Full Part plasmid, and are taken from the same samples, respectively, as lanes 5, 6, and 7 in Figure 2.3.
Co-transformation of DNA scaffold (ZF-binding domain and HTBS) parts and the full RT part containing MLRT and HIVRT in E. coli allowed production of DNA scaffolds in vivo. We grew and extracted these scaffolds via the technique described by Elbaz et al.: using an isopropyl-β-d-thiogalactoside-inducible (IPTG) promoter to form scaffolds in the presence of IPTG. DH10β E. coli cells were co-transformed with the Resv 2x, Resv 4x, CoA 4x, or CoA 6x scaffold part and RT part with each group grown with IPTG or without IPTG present. Each sample was compared with the PCR-purified dsDNA coding for its respective scaffold part. DNA scaffolds were extracted and then visualized on an agarose gel. Figure 2.5 shows the successful creation of the Resv 4x scaffold.
Fig 2.5: Agarose gel with dsDNA scaffolds extracted with phenol:chloroform:isoamyl. Lane 1: 20bp ladder. Lane 2: Resv 2x (not induced to make dsDNA scaffolds). Lane 3: Resv 2x (induced). Lane 4: PCR product of Resv 2x gene. Lane 5: Resv 4x (not induced to make dsDNA scaffolds). Lane 6: Resv 4x (induced). Lane 7: PCR product of Resv 4x gene. Lane 8: 1kb ladder. The ~100 bp band present in lane 6 shows successful formation of the Resv 4x scaffold.
In comparison to Genetic encoding of DNA nanostructures and their self-assembly in living bacteria by Elbaz et al
[1]Elbaz, J., Yin, P. & Voigt, C. Genetic encoding of DNA nanostructures and their self-assembly in living bacteria. Nat Commun 7, 11179 (2016).
, which pioneered the r_oligo structure, the bands on the polyacrylamide gels in the experimental groups with IPTG present show successful DNA scaffold formation and extraction (Figure 2.6).
Fig 2.6: Extracted DNA scaffolds from Elbaz et al.
A 3A assembly was used in order to ligate the full DNA scaffold part (2x,6x or 4x,4x and the full RT part containing MLRT and HIVRT into the same pSB1A3 plasmid. Future wetlab tasks will involve verification of this ligation via restriction enzyme digest analysis and cotransformation of this plasmid along with the full BMC part and resveratrol and malonyl CoA enzyme parts in order to test the Manifold system.
Enzyme Production
To complete the Manifold system, we needed to modify enzymes in the resveratrol pathway so that they could bind to our DNA scaffolds. This involved adding a zinc finger scaffold gene to a standard GFP protein gene. After assembly and transformation, we saw GFP presence in our cells under UV light (Figure 3.1).
Fig 3.1: GFP-zinc finger fusion proteins fluorescing in E.coli.
We created resveratrol pathway enzymes with zinc finger fusions and used a Golden Gate Assembly to form the full parts. Verification of the transformed and extracted assemblies is still underway.
Demonstrating Manifold
Fig 4.1 This dot chart details the final experimentation of the Manifold system. Extracted resveratrol and other measurements, if applicable, will be compared between various experimental groups to assess the functionality and benefits of a full Manifold system.
Comparison of groups 4 and 5 will determine whether differences in resveratrol production in (2x6x) and (4x4x) systems show that Manifold resolves flux imbalances in the resveratrol biochemical pathway. Comparison of groups 7 and 4 will show whether the CoA scaffold increases resveratrol yield. Comparison of groups 5 and 6 will show whether the RT part is needed to see resveratrol yield increase. Comparison of groups 2 and 3 will show whether BMCs alone do not increase resveratrol yield. Comparison of groups 5 and 8 will show whether the full Manifold system increases resveratrol yield beyond CoA and resveratrol scaffolds alone. Comparisons of groups 3 and 4, and 3 and 5 will both determine whether the full system increases resveratrol yields beyond BMCs, resveratrol enzymes and CoA enzymes alone.
As of October 21st, 2021, we are in the process of cotransforming, inducing, and gathering data on the Manifold system.
To test the affinity of our zinc finger fusion proteins to bind to our DNA scaffolds, we will run a “pseudo-southwestern blot” in which we will transfer a western blot of our scaffolds onto a membrane and wash it with purified, extracted GFP-zinc finger protein. If the DNA scaffold bands on the membrane turn green, we will know that our scaffolds can bind to our zinc finger fusion enzymes.
Finally, we also intend to do a mobility shift assay to assess the binding of our zinc finger fusion-resveratrol pathway enzymes to our DNA scaffolds.
Improving Manifold
After verification of resveratrol production, the next step in Manifold's journey, following the engineering design principle, is to test Manifold with different products. Our current experimental design presents a proof of concept by creating resveratrol, one drug of the many that Manifold could potentially produce. The next goal of Manifold is to test it at larger scales in order to determine whether it would be useful in the chemical industry. We hope to accomplish this by efficiently engineering Manifold such that the least amount of input and costs yield the greatest amount of product. Streamlining this process will help us expand to a larger-scale manufacturing of drugs and chemicals as needed by the industry.
Another key way for us to improve Manifold is to continue modeling its molecular dynamics in order to better understand the specifics and possible restraints of the system. This comes through understanding how the BMCs are being created, the assembly of our scaffolds, and the attaching of enzymes. Using modeling as a tool to visualize this data, we hope to improve our scaffold and enzyme design so that Manifold can truly be a platform technology.
Another future goal of Manifold focuses on outreach through presenting our system to other companies and labs as a technology that can be utilized to improve their own drug manufacturing.
Home
Project
Parts
Outreach
Team
Awards
Resources
Chemistry Building
485 McCormick Rd
CHARLOTTESVILLE, VA 22903
485 McCormick Rd
CHARLOTTESVILLE, VA 22903
