Results

Overview

Our results consist of mainly 4 parts:

  • Firstly, developing and determining the right reporter system for evaluation.
  • Secondly, evaluating the efficiency of WT-T7RNAP against derived ancestral sequences.
  • Thirdly, evaluating the efficiecny of WT-T7RNAP against literature-reviewed variants.
  • Lastly, evaluating the efficiency of WT-T7RNAP with addition of psuedouridine.

    • Developing the Right Reporter System

    The very first part of our experiment was to design a reporter system to evaluate the efficiency of the wild-type T7RNAP, as well as any variants generated in the future. Established protocols that have previously characterized T7RNAP have utilized antibiotic resistance as a readout1, with better-performing variants capable of surviving at higher concentrations of antibiotic conditions2.

    To prevent manipulating the fitness phenotype of the organism, our design primarily revolves around utilizing the expression of fluorescence as a reporter to evaluate the transcriptional activity of different T7RNAP variants. Our initial design was a simple reporter cassette, a cyan fluorescent protein encoded downstream of a T7 promoter. This reporter cassette was then ligated to a high-copy number (pUC) Kanamycin resistance plasmid.

    The plasmid was constructed via ligation of different parts derived from the iGEM registry to construct the full reporter cassette and plasmid.

  • BBa_J428353 - pJUMP28-1A(sfGFP)
  • BBa_K3457003 - T7 Promoter
  • BBa_J428032 - Ribosomal Binding Site (RBS)
  • BBa_E0020 - eCFP
  • BBa_J428091 - T7Te

    • Fig 1: Reporter cassette used for evaluating T7RNAP

    Fig 2: Reporter plasmid - plasmid C1 (pUC; high-copy number plasmid)

    The reporter plasmid was then transformed into BL21 (DE3) cells for testing. Colony PCR was first performed to confirm successful plasmid transformation before a single colony was inoculated and cultured overnight.The overnight culture was then sub-cultured at a 1:50 dilution for 3 hours until the OD600 reached approximately 0.400, after which, it was induced with 0.5 mM IPTG for 2 hours. Fluorescence was observed only in the IPTG-induced BL21 (DE3) cells, confirming that the reporter plasmid was functional.



    Fig 3: Initial screening of our reporter plasmid C1. [TOP] Colony PCR confirmed successful transformation of plasmid C1 into BL21 (DE3) cells for all 8 colonies. Only 1 colony was inoculated and further cultured overnight. [LEFT] Observation of fluorescence under physical light. [RIGHT] Observation of fluorescence under blue-light.

    Unfortunately, upon further sub-culturing of the bacterial culture (-ve control) followed by repeated IPTG induction, no fluorescence was observed. We initially hypothesized the loss of fluorescence due to metabolic burden, caused by the expression of our protein of interest from a high-copy plasmid, resulting in a change in sequence within the reporter cassette. However, the plasmids isolated from the un-induced cultures showed no mutations within the T7 promoter sequence, nor within the coding sequence of the fluorescent protein.

    To further investigate the loss of fluorescence previously observed in reporter C1 (BL21 (DE3) ells transformed with plasmid C1), Sanger sequencing was employed to identify for possible loss-of-function mutations in the coding regions of the T7RNAP found within the genome. The T7RNAP coding sequence nested within the genome of BL21(DE3) cells was first amplified using PCR, then purified and sent for Sanger sequencing. Sanger sequencing results revealed a mutation in the active site of the T7RNAP3 - I810T. It was possible that this change in amino acid residue in the active site of T7RNAP resulted in the loss of function of the enzyme, thereby resulting in the loss of fluorescence observed subsequently.

    Fig 4: Isolated plasmids from two un-induced samples and amplified T7RNAP sequence from the genome of BL21 (DE3) cells were sent for Sanger sequencing. [TOP] Sequencing results showed that the reporter cassette of plasmid C1 remains unchanged, with the features still intact in the same position. Sequence alignment was also performed to verify that there were no major changes in the nucleotide sequence (not shown here). [BOTTOM] Sanger sequencing identified a mutation within the active site of T7RNAP, I810T. This could have resulted in the loss of function of function of the enzyme, thereby resulting in the loss of fluorescence observed in subsquent sub-cultures.

    Due to the unstable expression of the fluorescence observed, two additional reporter plasmids were cloned using Gibson assembly. The reporter cassette was cloned into a medium-copy number plasmid (BBa_J428341) and a low-copy number plasmid (BBa_J428350) to provide a more stable reporter system.

  • Plasmid C2 – pBR322_T7Prom_eCFP – Medium-copy plasmid (15 – 20 copies per cell).
  • Plasmid C3 - p15A_T7Prom_eCFP – Low-copy plasmid (10 copies per cell)
    • Fig 5: Plasmid map for plasmid C1 and C2. [LEFT] Medium-copy plasmid (BBa_J428341; 15 – 20 copies per cell). [RIGHT] Low-copy plasmid (BBa_J428350; 10 copies per cell)

    Similarly, the reporter plasmids (plasmid C2 and C3) were transformed into BL21 (DE3) bacterial cells for testing. Colony PCR was performed (results not shown) to confirm successful transformation before a single colony was inoculated for further culturing. Following the culturing, all three plasmids (C1, C2 and C3) were evaluated for their functionality as a reporter plasmid in BL21 (DE3) E. coli cells. Overnight cultures of BL21 (DE3) E. coli cells containing the three plasmids respectively were first sub-cultured (at 1:25 dilution) for 2-3 hours before IPTG induction. Two different IPTG-induction protocols were followed:

  • Fast induction: Incubate at 37 °C for 2 hours after IPTG induction.
  • Slow induction: Incubate at 18 °C for overnight after IPTG induction.

    • Qualitatively, across both IPTG-induction methods, the observations were consistent where fluorescence was only observed in BL21 (DE3) cells containing either reporter plasmids C2 or C3, and when they were induced with either 0.5mM or 1.0mM IPTG.

    Fig 6: Fluorescence (eCFP) was observed only for bacterial cells harbouring either plasmids C2 or C3, and induced with IPTG. Within each frame, from left to right, -ve control, 0.5mM IPTG, 1.0mM IPTG. [LEFT] Triplicates from fast induction. [RIGHT] Triplicates from slow induction.

    Following the qualitative analysis of our reporter systems, a quantitative analysis of our reporter systems was performed to further corroborate their functionallity. A similar IPTG induction protocol was followed, where bacterial cultures were induced with IPTG and incubated at 37°C for 2 hours. Our quantitative results corroborate our qualitative analysis, where only fluorescence was observed in IPTG-induced BL21 (DE3) cells harbouring plasmid C2 or C3.

    Fig 7: Quantification of relative fluorescence intensity for all three reporter systems. From left to right, C1(-IPTG), C1(+IPTG), C2(-IPTG), C2(+IPTG), C3(-IPTG) and C3(+IPTG) [LEFT] Replicate 1; n=8 technical replicates. [CENTER] Replicate 2; n=6 technical replicates. [RIGHT] Replicate 3; n=6 technical replicates.

    Having robustly evaluating our reporter systems through both qualitative and quantitative methods, we eventually decided on plasmid C3 as our final reporter plasmid. This was mainly based on two reasons:

  • By having our reporter cassette encoded on a low-copy number plasmid, we are able to increase the sensitivity of our assay, thereby allowing any difference in the fluorescence intensity to be easily detected. This aids in identifying better-performing variants in our downstream experiments.
  • The utilisation of a plasmid with a p15A origin of replication also allows for the compatibility with our T7-expressing plasmid, which has a different origin of replication (see Engineering Success). As such, by using plasmids from different incompatibility groups for our experiments, we can prevent a competition of cell machinery and resources between the two plasmids.

    • Evaluating the efficiency of WT-T7RNAP against derived ancestral sequences

    Ancestral sequence reconstruction has a been a long-standing technique in evolutionary biology to infer sequences of ancient proteins based on existing sequences4. One of the key takeaways from our interviews (see Human Practices) was the difficulty behind developing novel protein sequences through protein engineering. As such, we set out to use ancestral reconstruction to generate functional distant homologs through in silico methods. A programme called Molecular Evolutionary Genetics Analysis version X (MEGAX)5 was utilized for our ancestral reconstruction workflow:

  • Data Procurement
    • Two datasets were generated by using the wild-type T7RNAP sequence as the query to search for homologous sequences using BLASTP (Basic Local Alignment Search Tool Protein). The first dataset generated ~100 sequences (denoted Blast100), all derived within the same taxonomy of Studiervirinae (tax id: 2731653). A larger dataset of ~250 (denoted Blast250) sequences was also derived with no limitations for a more diverse and possibly biologically relevant dataset.

  • Sequence Alignment
    • The protein sequences within each dataset were then aligned and distant sequences were manually removed. Sequences that resulted in large gaps for all other sequences were subsequently removed to "reduce” these misalignment between sequences.

  • Model Selection
    • An initial substitution model was ran to find the best model for ancestral reconstruction for each dataset. The LG model (G+I) was used for Blast100, while the LG model (G+I+F) was used for Blast250.

  • Phylogenetic Tree Construction and Ancestral State Reconstruction
    • A maximum likelihood tree (phylogenetic tree) was then generated based on the best-fit model of each dataset. MEGAX was then utilized to infer the ancestral sequence based on the constructed phylogenetic tree and selected outgroup.
    Fig 8: Inferred ancestral tree for both datasets. Box in red in wild-type T7RNAP sequence. [LEFT] Inferred ancestral tree generated from Blast100; boxed in blue are nodes of interests. Nodes 119 (denoted as RNAPAnc119) and 137 (denoted as RNAPAnc137) were selected for further investigation. [RIGHT] Inferred ancestral tree generated from Blast250; circled in red are nodes of interests. Node 302 (denoted as RNAPAnc302) was selected for further investigation.

  • Structure Prediction using Alphafold
    • Alphafold was then utilized to first predict the structure of all ancestral sequences of interest, providing insights into the protein's folding and functional domains. PyMOL was also employed to visualize the predicted structure, allowing for a detailed comparison with the wild-type and the RMSD value. After confirming the feasiblity of the structure through computational analyses, the ancestral sequences were then cloned into our stable T7-expressing plasmid, plasmid 1c (see Engineering Success for details), for downstream testing and analysis.


    Fig 9: Predicted structure for ancestral sequences of interest using Alphafold. [TOP] Alphafold predicted structures for all 3 ancestral sequences; and compared against the structure of wild-type T7RNAP (cyan). [BOTTOM LEFT] Sequence coverage of RNAPAnc119. [BOTTOM CENTER] Sequence coverage of RNAPAnc137. [BOTTOM RIGHT] Sequence coverage of RNAPAnc302.

    The ancestral sequences were subsequently cloned into plasmid 1c using Gibson assembly, replacing the wild-type T7RNAP sequence. Purified plasmids were then transformed into competent E. coli reporter cells to compare their efficiency against the wild-type for an initial screening. Cells were then sub-cultured, followed by IPTG induction. Comparing the different variables tested, it is evident that the wild-type T7RNAP expressed from a plasmid (Stbl3 - 1c/C3) has shown the highest relative fluorescence intensity among all, higher than that observed by BL21 (DE3) cells transformed with the reporter plasmid. Comparing both the ancestral sequences (RNAPAnc119 and RNAPAnc137), little to no fluorescence intensity was similar, showing that the derived ancestral sequences have much lower processivity than the wild-type. As such, more developmental effots to improve the processitivity of the ancestral sequences are required before a comparable efficiency is observed. The level of fluorescence intensity observed from that of the uninduced negative control at 120mins is also higher than that of some of the induced samples. As the sequence of the antisense oligonucleotide (ASO) was designed specifically for the wild-type T7RNAP, it is possbily that the ASO are not acting against the ancestral sequences. This results in the ancestral sequences to not be supressed by the ASO system. Hence, future works also include developing and refining the sequence of the ASO for different ancestral sequences.

    Fig 10: Comparison of fluorescence measured at different time point after IPTG induction (-ve(120min,10min, 30min, 60min, 120min). 6 technical measurements were taken for each sample. The negative control sample were measured after the 120 minutes timepoint.

    Evaluating the efficiecny of WT-T7RNAP against literature-reviewed variants

    Site-Directed Mutagenesis

    The project also aims to evaluate a series of mutations derived from literature, focusing on their functional implications to the T7RNAP. Below is a table summarizing the specific mutations along with their corresponding references. There are also future efforts to combine these mutations found across different literatures to developing better-performing variants.

    Code Mutation Reference
    A S43Y Wu, H., Wei, T., Yu, B., Cheng, R., Huang, F., Lu, X., Yan, Y., Wang, X., Liu, C., & Zhu, B. (2021). A single mutation attenuates both the transcription termination and RNA-dependent RNA polymerase activity of T7 RNA polymerase. RNA Biology, 18(sup1), 451–466. https://doi.org/10.1080/15476286.2021.1954808
    B G47A Dousis, A., Ravichandran, K., Hobert, E. M., Moore, M. J., & Rabideau, A. E. (2022). An engineered T7 RNA polymerase that produces mrna free of immunostimulatory byproducts. Nature Biotechnology, 41(4), 560–568. https://doi.org/10.1038/s41587-022-01525-6
    K +884G
    C P266L Guillerez, J., Lopez, P. J., Proux, F., Launay, H., & Dreyfus, M. (2005). A mutation in T7 RNA polymerase that facilitates promoter clearance. Proceedings of the National Academy of Sciences, 102(17), 5958–5963. https://doi.org/10.1073/pnas.0407141102
    D S430P Boulain, J.-C., Dassa, J., Mesta, L., Savatier, A., Costa, N., Muller, B. H., L’hostis, G., Stura, E. A., Troesch, A., & Ducancel, F. (2013). Mutants with higher stability and specific activity from a single thermosensitive variant of T7 RNA polymerase. Protein Engineering Design and Selection, 26(11), 725–734. https://doi.org/10.1093/protein/gzt040
    F S767G
    I F880Y
    M I810S
    G F849I
    L Q744R
    E E643G
    K T3M Jiang, K., Yan, Z., Bernardo, M. D., Sgrizzi, S. R., Villiger, L., Kayabolen, A., Kim, B., Carscadden, J. K., Hiraizumi, M., Nishimasu, H., Gootenberg, J. S., & Abudayyeh, O. O. (2024). Rapid Protein Evolution by Few-Shot Learning with a Protein Language Model. https://doi.org/10.1101/2024.07.17.604015
    H F880A Gardner, L. P., Mookhtiar, K. A., & Coleman, J. E. (1997). Initiation, elongation, and processivity of carboxyl-terminal mutants of T7 RNA polymerase. Biochemistry, 36(10), 2908–2918. https://doi.org/10.1021/bi962397i


    Several of the mutations were cloned into T7RNAP individually, and evaluated for their efficiency after IPTG-induction. The variation of the fluorescence measured between each construct is high. At timepoint 10min, the fluorescence intensity for most variants are negative, with fluorescence increasing with time (higher fluorescence intensity measured at later timepoints of 30min and 60min).

  • The fluorescence intenisty measured for constructs a - e, and g are relatively similar, although with slight fluctuations.
  • The fluorescence intenisty measured for construct h, i , k, and m are exceptionally low compared to the entire panel tested.
  • The fluorescence intenisty measured for construct J appear to be the highest amongst the variant constructs; ableit still lower than the wild-type T7RNAP.
  • Additional established variants will be added to the panel to further include more beneficial mutations.
    • Fig 11: Comparison of different constructs with site directed mutagenesis at different time point after IPTG-induction (10min, 30min, 60min, 120min). 6 technical replicates were taken for each variant. The negative control sample were measured after the 120 minutes timepoint.

    Random Mutagenesis

    Another method that can be utilized to identify beneficial mutations is through screening mutants after random mutageneis. Random mutagenesis is a widely utilized technique in molecular biology and directed evolution that generates genetic diversity by introducing random mutations into DNA sequences. The various methods to induce random mutagenesis include error-prone PCR (epPCR), chemical mutagenesis, transposon-based methods, saturation mutagenesis and mutator strains6.

    By generating large libraries of mutants, we can screen for variants with improved properties or novel functionalities, such as enhanced enzyme activity of stability under different conditions.

    We thus employed epPCR to generate a library of mutant T7RNAPs, in which the wild-type T7RNAP sequence was amplified using a low-fidelity DNA polymerase that introduces point mutations during amplification. The resulting PCR products, which contain a library of mutations, were then cloned into plasmid 1c, replacing the wild-type sequence. Sanger sequencing results from few colonies show random point mutations along the T7RNAP sequence (not shown). The library of mutant plasmids was then transformed into our competent E. coli reporter cells and plated onto CmR + Kan LB Agar plates.

    Fig 12: Library of mutant T7RNAP plasmids transformed into reporter E. coli cells

    The cells were collected and diluted 100x, before aliquoting 125ul each into a 96 well plate for further sub-culture. However, wells along the 96-well plate were completely dried up, with partial drying of culture observed throughout the plates. As such, the identification of mutant variants was unable to proceed due to in-optimization of experimental protocols.

    Fig 13: Observed dried up cultures along the edges of the 96-well plates, resulting in auto-fluorescence when observed under blue light.

    Evaluating the efficiency of WT-T7RNAP with addition of psuedouridine

    One of the final experiments performed was based on one of the feedback questions received during our public engagement efforts (see Human Practices). One of the questions we received prompted us to determine if our current expression system (the combination of plasmid 1c with C3) was capable of incorporating pseudouridines into mRNA transcripts, or if the addition of pseudouridines would impair the catalytic function of the T7 RNA polymerase. Bacterial cultures containing the various T7-expression systems were induced with IPTG w/o pseudouridines. The lack of significant differences in the levels of fluorescence detected between the same expression system suggests that pseudouridine does not have a substantial effect on the experimental setup. This indicates that the presence of pseudouridine may not influence the reporter's sensitivity or the overall outcome of the experiment, leading to the conclusion that it might not be a critical factor in this particular context. Further investigation may be necessary to explore its role or to confirm these findings under different conditions.

    Fig 14: Samples were induced with IPTG w/o pseudouridines to determine the effect of pseudouridines on expression levels.

    Given the incorporation of pseudouridines in the production of mRNA vaccines, the ability to incorporate such uncanonical bases would prove to be useful for such applications. Future experiments such as using in-vitro transcription with modified bases could thus be performed to select for variants that are capable of incorporating these unnatural bases.

    References

    1 Ikeda, R. A., Ligman, C. M., & Warshamana, S. (1992). T7 promoter contacts essential for promoter activity in vivo. Nucleic acids research, 20(10), 2517–2524. https://doi.org/10.1093/nar/20.10.2517

    2 Boulain, J. C., Dassa, J., Mesta, L., Savatier, A., Costa, N., Muller, B. H., L'hostis, G., Stura, E. A., Troesch, A., & Ducancel, F. (2013). Mutants with higher stability and specific activity from a single thermosensitive variant of T7 RNA polymerase. Protein engineering, design & selection: PEDS, 26(11), 725–734. https://doi.org/10.1093/protein/gzt040

    3 Bonner, G., Patra, D., Lafer, E. M., & Sousa, R. (1992). Mutations in T7 RNA polymerase that support the proposal for a common polymerase active site structure. The EMBO Journal, 11(10), 3767–3775. https://doi.org/10.1002/j.1460-2075.1992.tb05462.x

    4 Joy, J. B., Liang, R. H., McCloskey, R. M., Nguyen, T., & Poon, A. F. Y. (2016). Ancestral Reconstruction. PLoS Computational Biology, 12(7), e1004763–e1004763. https://doi.org/10.1371/journal.pcbi.1004763

    5 Kumar, S., Stecher, G., Li, M., Knyaz, C., & Tamura, K. (2018). MEGA X: Molecular Evolutionary Genetics Analysis across Computing Platforms. Molecular Biology and Evolution, 35(6), 1547–1549. https://doi.org/10.1093/molbev/msy096

    6 Labrou N. E. (2010). Random mutagenesis methods for in vitro directed enzyme evolution. Current protein & peptide science, 11(1), 91–100. https://doi.org/10.2174/138920310790274617