Cycle 0

Initial Plan

Fig.1 Conceptual diagram of quorum sensing & lysis

We designed a circuit using quorum sensing and lysis [1][2] to maintain a low E. coli population density. To compare the performance, we decided to prepare two systems with different styles of lysis. First, we planned to clone the lytic gene and confirm its function, then to confirm that the QS component was functional, and finally, to prove that E. coli exhibited the desired behavior by connecting the lytic gene system to the QS.

Cycle 1

Design 1

To test whether lysis occurs with the lytic gene, we designed two plasmids encoding the lytic genes strictly controlled by the araBAD promoter. We chose a low-copy plasmid as a vector to minimize leakage. Each plasmid had a different type of lytic gene: one had the lysis cassette from lambda phages [3][4] and one had ColE7 [5].

Build 1

The vectors were treated with restriction enzymes, and the insert was artificially synthesized with IDT. These were assembled using XE cloning [6]. Sanger sequencing confirmed that a plasmid with the expected sequence was produced.

Test 1

The plasmids were transformed into a ΔaraBAD E. coli strain and the growth curves with and without arabinose induction were compared.

Fig.2 Comparison of growth curves of lysis cassette and ColE7 with and without arabinose induction

After 3h, the increase of OD600 stopped only for those induced lysis with arabinose. However, after 6h, OD600 began to increase again. This phenomenon was observed repeatedly.

Learn 1

Different growth curves were obtained with and without arabinose induction: arabinose induction resulted in a reduced increase in OD600. It was also found that arabinose induction takes about 3 h. This is in agreement with Deborah & James (1997) [7]. Furthermore, we found that after 6h, OD600 began to increase even with arabinose induction. This is thought to be due to the strong negative selection pressure on the lytic genes, either by dropping the plasmid or by the emergence of mutants with mutations in genes required for lysis [8][9]. Therefore, we realized that we needed to find a way to delay the fixation of the mutant. This was considered in later experimental designs.

Cycle 2

Design 2

Next, to confirm that the LuxR and lux promoter required for quorum sensing are functional, we designed a plasmid in which sfGFP is expressed when the lux promoter is activated.

Build 2

Vectors were obtained by restriction enzyme processing and PCR, and inserts were artificially synthesized by IDT. These were assembled using XE cloning [6]. Sanger sequencing confirmed that a plasmid with the expected sequence was completed.

Test 2

OD600 and fluorescence were measured with and without activation of the QS pathway by the autoinducer AHL. The results are shown in Fig.3.

Fig.3 Fluorescence * / OD600 with and without AHL induction

Fluorescence * = Fluorescence measured - (Fluorescence of the blank)

Learn 2

AHL-induced activation of the lux promoter occurred as expected. This confirmed that the LuxR and lux promoter required for quorum sensing were functional.

Cycle 3

Design 3

To create E. coli that would lyse in response to quorum sensing, we designed a plasmid that encoded a lysis gene behind a lux promoter.

Build 3

The vector was treated with restriction enzymes and the insert was artificially synthesized with IDT. These were assembled using XE cloning [6].

Test 3

At the stage of liquid culture of E. coli for miniprep, many of the E. coli were lysed, resulting in a highly viscous culture.

Learn 3

We believe that the leakage of the lux promoter caused the unexpected expression of the lytic gene without AHL induction. We thought that if this was unaddressed, unintended selection pressure would cause the mutants to be fixed earlier. Therefore, we realized that we needed to use a less leaky lux promoter.

Cycle 4

Design 4

As a lux promoter with less leakage, we focused on the mutated versions of Mut5 and Mut9 lux promoters in Tsinghua 2018's characterization in Part: BBa_R0062. Tsinghua 2018 reports that Mut5 is a mutation that reduces leakage to some extent and maintains maximal activity; Mut9 is a mutation that almost eliminates leakage but also loses significant maximal activity. To check the leakage and activity of these two mutated versions of the lux promoter, we designed two plasmids that express sfGFP when the lux promoter is activated.

Fig.4 The mutations in the Lux promoter.

Mut5 has a G to T change at the -36 site, and Mut9 has a T to G change at the -37 site.

Build 4

The plasmids were created using PCR so that the point mutation would be inserted into the plasmid created in Cycle 2. The expected mutations were confirmed by Sanger sequencing.

Test 4

For each promoter, we measured OD600 and fluorescence without AHL induction and obtained the graphs shown in Fig.5 (a). The same measurements were also performed for the case with AHL induction, and the graph shown in Fig.5 (b) was obtained. For the samples without AHL induction, fluorescence at 0h was omitted because the low OD600 resulted in large absolute values.

Fig.5 Graph of OD600 and fluorescence without AHL induction

(a) Fluorescence was measured when no AHL was added to examine the leakage of the original lux promoter and its mutants Mut5 and Mut9.
(b) To examine the maximal activity of the original lux promoter and its mutants Mut5 and Mut9, we induced them with 100 nM AHL and measured their fluorescence.
Fluorescence * = Fluorescence measured - (Fluorescence in the blank)

Learn 4

Compared to the lux promoter without the mutation, Mut5 and Mut9 showed a slight decrease in maximum activity, Mut5 showed no change in leakage, and Mut9 showed a decrease in leakage. This indicates that Mut9 is the better choice.

Cycle 5

Design 5

As an additional plasmid to the E. coli used in Cycle 4, we designed a plasmid encoding the luxI gene that produces AHL.

Build 5

The vector was amplified by PCR and the insert was artificially synthesized by IDT. These were assembled using XE cloning [6]. Sanger sequencing confirmed that the sequence was as expected.

Test 5

We confirmed by measuring OD600 that lysis occurs in response to the number of surrounding E. coli by having E. coli perform quorum sensing.

Fig.6 Graph of OD600 with and without induction of luxI controlled by araBAD promoter at arabinose concentration of 0.2%

(a) With the lambda phage lysis cassette, OD600 stopped increasing at about OD600=0.5 regardless of the mutation of the lux promoter.
(b) When ColE7 was used as a lysogenic gene, OD600 stopped increasing at about OD600=1, although there were some differences depending on the mutation of the lux promoter.

Learn 5

Both of the two genes successfully caused lysis by quorum sensing. We have registered parts of the sequence of the lytic genes controlled by the mutated version of the lux promoter used in this experiment (BBa_K4655015, BBa_K4655017). However, in both cases of the two types of lytic genes, the non-lysogenic mutants outnumbered the lysogenic strain after 10 hours, and the increase in OD600 could not be prevented. This indicates that it is necessary to prevent the fixture of the non-lysogenic mutant. In the case of the lambda phage lysis cassette, the OD600 decreased to 1/3 of the maximum value, and based on the observations made in previous experiments, it is thought that almost all bacteria are lysed at this time. However, based on our modeling results, we believe that this problem can be solved by appropriately reducing the concentration of arabinose used for induction (see Model page for details).

Finally, from these results and the sfGFP leakage experiment described in Protein Degradation page, we found that the lambda phage lysis cassette of the two lysis genes we used is more suitable for our project.

References

[1] Hsu, C. Y., Yu, T. C., Lin, L. J., Hu, R. H., Chen, B. S. (2014) Systematic approach to Escherichia coli cell population control using a genetic lysis circuit. BMC Systems Biology, 8(S7). https://doi.org/10.1186/1752-0509-8-S5-S7

[2] Cahill1, J., Young, R. (2019). Phage Lysis: Multiple Genes for Multiple Barriers, Advances in Virus Research, 103, 33-70. https://doi.org/10.1016/bs.aivir.2018.09.003

[3] Savva, C. G., Dewey, J. S., Moussa, S. H., To, K. H., Holzenburg, A., Young, R. (2014). Stable micron-scale holes are a general feature of canonical holins. Molecular Microbiology, 91 (1), 57-65. https://doi.org/10.1111/mmi.12439

[4] Abeysekera, G. S., Love, M. J., Manners, S. H., Billington, C., Dobson, R. C. J. (2022). Bacteriophage-encoded lethal membrane disruptors: Advances in understanding and potential applications, Frontiers of Microbiology 13. https://doi.org/10.3389/fmicb.2022.1044143

[5] Chen, Y. R., Yang, T. Y., Lei, G. S., Lin, L. J., Chak, K. F. (2011). Delineation of the translocation of colicin E7 across the inner membrane of Escherichia coli. Archives of Microbiology 193, 419-428. https://doi.org/10.1007/s00203-011-0688-7

[6] Liu, A.Y., Koga, H., Goya, C., Kitabatake, M. (2023). Quick and affordable DNA cloning by reconstitution of Seamless Ligation Cloning Extract using defined factors. Genes to Cells, 28(8), 553-562. https://doi.org/10.1111/gtc.13034

[7] Siegele, D. A., Hu, J. C. (1997) Gene expression from plasmids containing the araBAD promoter at subsaturating inducer concentrations represents mixed populations. PNAS, 94(15), 8168-8172. https://doi.org/10.1073/pnas.94.15.8168

[8] You, L., Cox, R. S., Weiss, R., & Arnold, F. H. (2004) Programmed population control by cell–cell communication and regulated killing. Nature, 428 (6985), 868-871. https://doi.org/10.1038/nature02491

[9] Liao, M.J., Din, M.O., Tsimring, L., Hasty, J. (2019) Rock-paper-scissors: Engineered population dynamics increase genetic stability. Science, 365 (6457). 1045-1049. https://doi.org/10.1126/science.aaw0542