Introduction
Throughout the course of our project, we followed the engineering design cycle of designing, building, testing, and learning in iterative fashion. A few examples of this are highlighted below.
Cycle 1
Design
To test the efficiency of the AMP + encapsulin expression system, we planned a growth assay comparing the growth rates of bacteria with our experimental construct to relevant controls. This resulted in the following bacterial strains:
- Vehicle Control: pETDuet vector
- Negative Control: T4GALA encapsulin+pETDuet vector
- Positive Control: AMP+pETDuet vector
- Experimental Construct: T4GALA encapsulin+AMP+pETDuet vector
We first designed our constructs for cloning via Gibson assembly, with each segment containing 20 base pair overlaps on either end. We ensured that this included overhangs compatible with the sites created by the restriction enzymes we planned to use for linearizing our plasmid vector. We also codon optimized our sequences for our expression strain of BL21 E. coli.
Build
We used PCR to extract the encapsulin sequence from a plasmid. We ordered our AMP sequence from IDT. We linearized our pET Duet vector using two restriction enzymes, NdeI and PacI. We then assembled all of these sequences into plasmids using Gibson assembly. Transformation was then conducted with DH5ɑ E. coli. Then ampicillin resistant cells were miniprepped in preparation for sequencing.
Test
The positive control AMP plasmid as well as the encapsulin + AMP plasmid were submitted to Eurofins Scientific for Sanger sequencing to confirm Gibson assembly success.
Learn
The location-specific insertion of our sequences confirmed Gibson assembly success, although there were mismatches throughout despite generally accurate sequencing. We attributed this difficulty to technical issues in sequencing. In the future, we could repeat sequencing with different primers to confirm our result, sequence the pure AMP insert gBlock to ensure accuracy, or repeat Gibson assembly and subsequent transformation to produce the purest production of the plasmid if mutations were found. Additionally, we learned the importance of sequencing primer choice, as when we initially sequenced the plasmid with generic T7 primers, we observed nonspecific primer binding. Ultimately, we believe we were successful in engineering our designed plasmids.
Cycle 2
Design
After cloning, we transformed these plasmids into BL21 E. coli, a gram-negative bacteria known to be susceptible to our AMP [1]. We hypothesized that, after induction, the experimental construct would improve bacterial growth relative to the positive control, which would demonstrate inhibited growth or cell death as the AMP produced its toxic effects. We planned to induce protein expression and take OD600 readings to measure bacterial growth over a period of several hours.
Build
In order to build the bacterial strains that we would use in our growth assay, each plasmid was successfully transformed into BL21 E. coli bacteria, as confirmed by antibiotic selection with ampicillin. These 4 strains were used to inoculate 8 samples of LB broth treated with ampicillin. They were then either grown in the presence or absence of IPTG for several hours.
Test
OD600 readings were taken manually approximately every half hour for 7 hours. We used these results to find the growth rates of each of the bacterial colonies. The growth rate was determined for the logarithmic phase of growth for each sample. Then the average rate and variance of the rates were calculated.
Learn
This growth assay was run manually, and it was logistically challenging to take 8 accurate OD600 measurements every half hour for 7 hours. This led to noisy data, seen below, which were hard to draw conclusions from. Similarly, the manual sampling made it difficult to run replicates for accurate estimations of effect size. We determined that this experiment was inconclusive, and began the planning phase over again to conduct a more precise experiment.
Figure 1: Growth of bacteria under induced expression of indicated protein. Note high noise caused by manual measurement.
Cycle 3
Design
To address the logistical challenges of our manual experiment, we designed a protocol for an automated growth assay. Rather than growing cultures in flasks, we grew our cultures in wells of a 96 well plate. We chose an automated method because it allowed us to eliminate noise from human error from our measurements while simultaneously increasing the number of replicates.
Build
We obtained access to a plate reader capable of running an automated growth curve experiment and built an assay that measured bacterial concentration for 3 technical replicates of 3 biological replicates with or without induction by IPTG for each of our bacterial strains.
Test
Our proposed plate was assembled and OD600 measurements were collected for each well every 5 minutes over the course of 12 hours.
Learn
This assay produced high quality data that our team was able to analyze for a meaningful assessment of the efficacy of our project. We learned definitively that the effect on bacterial growth rate caused by the encapsulin mitigating AMP toxicity was smaller than we had expected, and this informed our analysis of the results of our project.
Figure 2: Example curves from automated growth assay. Note the high quality of the curves.
References
- Lee, T. H., Carpenter, T. S., D'haeseleer, P., Savage, D. F., & Yung, M. C. (2020). Encapsulin carrier proteins for enhanced expression of antimicrobial peptides. Biotechnology and bioengineering, 117(3), 603–613. https://doi.org/10.1002/bit.27222