Team:Rice/Results
Results
5. Conclusion
Strain Characterization
Growth Rates
To identify growth conditions suitable for all strains and measure growth rates for parameter estimation in modeling , we performed OD time course measurement for all strains used in our project (Figure 1). Cells were grown in LB at 30°C for 15 hours. The results demonstrate that all strains are able to grow under those conditions, with E. coli strains and B. subtilis exhibiting the highest growth rates.
Inhibitory Concentrations of Antibiotics
In order to identify the antibiotic concentrations to be used for selection with our strains, we performed titrations with neomycin and kanamycin by spotting cultures on LB agar plates with concentrations ranging from 0 to 200 mg/L. The results show that most strains can be inhibited with < 50 mg/L of kanamycin (Figure 2). The exceptions are S. meliloti which is resistant to kanamycin but can be inhibited with neomycin and B. subtilis, whose growth can be inhibited at 200 mg/L kanamycin or 100 ng/mL kanamycin. We conclude that KanR aph(3′)-II can be used as a broad host range antibiotic resistance cassette in our constructs since it is known to confer resistance to both neomycin and kanamycin.
Broad Host Range Parts
The goal of our project is to create a broad host range system that can be transferred across species without the need for genetic manipulation. To achieve this goal, we had to identify origins of replication and promoters that perform well in a wide variety of strains. We chose three origins of replication pBBR1, RSF1010, and pWV01 known from literature to have broad host range. To test the performance of origins in our bacteria, we transformed a prototypical vector for each origin of replication containing KanR aph(3′)-II antibiotic resistance cassette into all strains. Growth of bacteria on inhibitory concentration of the appropriate antibiotic indicated successful plasmid replication. The results of the experiment are summarized in Figure 3. The results of the experiments for the strains not shown in the table were inconclusive. Based on these results, we chose to use RSF1010 for our broad host range constructs.
Origins of Replication
Regulatory Elements
In addition to the origins of replication, we characterized five broad host range regulatory elements from Johns et al.1, which include a promoter and a ribosome binding site. Johns et al.1 created a library of broad host range regulatory elements and characterized them in E. coli, B. subtilis, C. glutamicum, P. aeruginosa, and V. natriegens. In attempt to obtain the elements which result in the most consistent reporter expression levels across strains, we parsed through the data on transcription levels reported in Johns et al. for 240 sequences using MATLAB script, normalized the values based on maximum and expression levels for each strain and extracted 10 sequences with the standard deviation < 0.1 which spanned a range of expression levels across the bacteria tested. We constructed cassettes involving mKate2 fluorescent protein under the control of five different regulatory elements (Figure 4). Elements were ranked based on transcription rates documented in Johns et al.
We transformed the constructs into four E. coli strains (MG1655, BL21, Nissle 1917, and DH10B), P. putida , and S. oneidensis. Fluorescence levels measured after 24 hours of incubation at 30°C demonstrate that broad host range regulatory elements resulted in the same order of magnitude of mKate2 expression across strains, with Strength 8 element demonstrating the lowest variability (Figure 5). We submitted broad host range regulatory elements as BioBricks.
Figure 5: mKate2 expression with broad host range regulatory elements across strains. mKate2 expression constructs with regulatory elements were transformed into four E. coli strains (MG1655, BL21, Nissle 1917, DH10B), P. putida , and S. oneidensis . Cells were grown in LB and fluorescence measured after 24 hours of incubation. Error bars show +/- 1 s.d. across three replicates.
Orthogonal Transcription
After successfully assembling modified and original UBER2 on plasmids containing RFS1010 origin, we co-transformed E. coli DH10B cells with either original or modified UBER (BBa_K2540011 ) and an expression cassette containing mKate2 controlled by T7 promoter (Figure 6) and collected time course fluorescence data over 15 hours at 37°C in LB.
Figure 6: constructs assembled to compare the performance of the original and modified UBER in E. coli mKate2 expression constructs with HW elements were transformed into four E. coli strains (MG1655, BL21, Nissle 1917, DH10B), P. putida , and S. oneidensis . Cells were grown in LB and fluorescence measured after 24 hours of incubation.
The results demonstrate that modifications introduced do not interfere with UBER function and result in higher fluorescence/OD compared to the original system. (Figure 7). The absence of fluorescence for negative control samples which did not have T7 RNA polymerase demonstrate that fluorescence levels are due to orthogonal transcription rather than transcription by host polymerase.
Figure 7: Performance of original vs. modified UBER in E. coli mKate2 expression driven by either original or modified UBER was measured over 15 hours. Modifications introduced do not interfere with UBER function and result in higher fluorescence/OD compared to the original system. Error bars show +/- 1 s.d. across three replicates.
Next, we assembled the orthogonal transcription system (UBER and mKate2 expression cassette) on a single plasmid to facilitate cross-species expression (Figure 8).
Figure 8: UBER (original and modified) and mKate2 expression cassette assembled on a single plasmid
E. coli DH10B cells containing original UBER and mKate2 had detectable fluorescence (Figure 9). However, no fluorescence over background was measured for the modified UBER. A similar trend was observed for S. oneidensis. The opposite was observed in P. putida: modified UBER resulted in high mKate2 fluorescence while no fluorescence was detected for the original UBER. We suggest that inconsistencies in the expression levels may be due to heterogeneity of cell populations: when cells containing the orthogonal transcription system are plated on LB agar plates, both fluorescence and non-fluorescent colonies are observed. Future work will involve investigating this problem (one potential cause may be recombination within our circuit).
Figure 9: performance of orthogonal transcription systems in P. putida, S. oneidensis, and E. coli. The original UBER results in high fluorescence in E. coli and S. oneidensis and low fluorescence in P. putida.The system containing modified UBER yields high fluorescence in P. putida while no fluorescence in detected for E. coli or S. oneidensis. Error bars show +/- 1 s.d. across three replicates.
Orthogonal Translation
To identify an orthogonal 16S rRNA sequence that is likely to work across most of the strains used in our project, our team applied our software to the genomes of E. coli, B. subtilis , L. lactis, P. putida, S. oneidensis, S. meliloti, and C. glutamicum. The ASD sequence “AUCAGCAGCAUA” was highly ranked for all bacteria except B. subtilis, so we decided to use it to create a “universal” orthogonal ribosome construct. The wild-type E. coli 16S rRNA was mutated to incorporate the altered anti Shine-Dalgarno sequence, and constructs shown in Figure 10 were built to test the function of orthogonal ribosomes in E. coli.
Figure 10: constructs used to test orthogonal translation in E. coli. Shine-Dalgarno sequence of the oRBS is unable to interact with host ribosome and no fluorescence is observed when o16S rRNA is not present (negative control). When plasmid containing o16S rRNA is introduced, orthogonal ribosomes are created which are able to translate mRNA containing oRBS, resulting in reporter expression.
When the construct containing mKate2 with oRBS (orthogonal ribosome binding site BBa_K2540011) was transformed into E. coli MG1655, no fluorescence over background was detected. However, when mKate2 plasmid was co-transformed with a plasmid expressing orthogonal 16S rRNA, a 100~ fold increase of fluorescence compared to negative control was observed (Figure 11). This demonstrates that only orthogonal ribosomes are able to initiate translation of the mKate2 mRNA.
Figure 11: characterization of orthogonal translation constructs in E. coli . IPTG-dependent mKate2 fluorescence is observed when plamids containing o16S rRNA controlled by Plac promoter and oRBS-mKate2 are co-transformed into E. coli (left). At 0.1 mM IPTG, ~100 fold fluorescence over negative control is observed, demonstrating that translation is orthogonal (right). Error bars show +/- 1 s.d. across three replicates.
Conclusion
In the course of our project, we performed characterization of broad host range parts to be used in PORTAL. We identified antibiotic resistance cassette, origin of replication, and regulatory elements which can make PORTAL easily transferable across bacteria. In addition, we separately constructed and characterized orthogonal transcription and orthogonal translation components of PORTAL. Our results demonstrate that orthogonal transcription part needs to be further improved to eliminate inconsistencies in expression across bacteria. Orthogonal translation successfully functions in E. coli, with host ribosome unable to initiate translation of mRNA containing the orthogonal RBS. We submitted modified UBER, broad host range regulatory elements, and oRBS as BioBricks. However, we were not able to submit o16S rRNA since it contains an internal EcoRI cut site.
References
[1] Chubiz, L. M., & Rao, C. V. (2008). Computational design of orthogonal ribosomes. Nucleic Acids Research, 36(12), 4038–4046.
[2] Kushwaha, M., & Salis, H. M. (2015). A portable expression resource for engineering cross-species genetic circuits and pathways. Nature Communications, 6(1), 7832.