Team:Leiden/Results

<!DOCTYPE html> DOPLLOCK iGEM Leiden

Description Proof-of-Concept Engineering Results Model Contribution Measurement Kit-Search
Parts Notebook Safety Experiments
Integrated Human Practices Public Engagement Education Entrepreneurship Implementation Survey
Team Members Attribution Collaborations Sponsors
Featured Newsletter Crowdfunding
☰
DOPL LOCK

Results

After we applied the bottom-up design to build up DOPL LOCK, we performed experiments to characterize components of our system, including origin of replications, promoters, toxins and antitoxin. Then, we focused on assembling the most compatible components to the DOPL LOCK system. On this page, detailed experimental results that led to the proof-of-concept are shown.

Introduction

Preventing horizontal gene transfer (HGT) and physically constraining GMOs are the most challenging issues to broaden the application of synthetic biology outside the labs. We designed DOPL LOCK to both minimize the chance of HGT and prevent the physical spread of the GMOs. In our system, the toxin-antitoxin (TA) expression needs to be balanced at a certain level. We applied the bottom-up design to build up DOPL LOCK in three steps. First, we started with the characterization of each genetic part that is influential to this balance, namely the origin of replication (Ori) of the plasmid vector, promoters upstream of the toxin and antitoxin genes, and the toxin-antitoxin genes themselves. All parts characterized in this project are shown in Figure 1. Secondly, we chose the best compatible parts and assembled them into a single TA system to find the balanced expression level. Finally, two verified TA systems can be assembled into the final version of DOPL LOCK on two different plasmids. During our lab time, we successfully characterized each part and found the best combinations of genetic parts, with the exception of the antitoxins. However, we failed to assemble a TA system in a single plasmid due to limited laboratory time. The results of our experiments are summarized as follows.

Figure 1 overview of DOPL LOCK system

Figure 1: overview of DOPL LOCK system

In the following part, we summarized the results of our experiment. More detailed information is shown in the Results section.

Characterization of parts

Origin of replication

Aim: Determine which combination of Oris maintained plasmid copy number in the cells at similar levels.

✔ Successfully construct pJUMP plasmid collection with the RFP cassette as reporter.

✔ Single transformation showed that the plasmid copy number had minor influence on the fluorescent protein production after overnight incubation.

✔ p15A and pBR322/ROB exhibited the best compatibility.

✔ pSC101 and pUC have low expression in RFP when co-transformed.

❔ Substitute mRFP1 with mCherry to resolve FRET with sfGFP.

(✔ and ❔ show "successful" or "remains a challenge")

Oris significantly influenced the plasmid copy number (Figure 2) [1]. To determine which Oris pair can maintain the plasmid copy number in the cell at similar levels, we used the JUMP plasmid collection that includes four types of Oris, two antibiotics selection markers and sfGFP reporter cassettes [2]. The sfGFP cassettes of the plasmids were switched to a RFP cassette () accordingly (all constructs are shown in the Parts page). Both vectors carrying sfGFP or RFP were co-transformed into the E. coli TOP10 strain and their fluorescence was measured by the plate reader.

Figure 2 Selection of Oris

Figure 2: Selection of Oris

The result showed that the Oris p15A and pBR322/ROB were the most compatible groups. Their co-transformed strain showed high levels of both sfGFP and RFP fluorescence signals (Figure 3). With this Ori combination, we could regulate the TA gene expression in two plasmids easily, since these two plasmid backbones have a more similar expression of the foreign genes they carried.

Figure 3 Fluorescence of co-transformed p15A and pBR322/ROB (in the middle) with their single-transformed control group (at the sides) with error bars. Green bars represent the sfGFP fluorescent signal while red bars represent RFP signals.

Figure 3: Fluorescence of co-transformed p15A and pBR322/ROB (in the middle) with their single-transformed control group (at the sides) with error bars. Green bars represent the sfGFP fluorescent signal while red bars represent RFP signals.

We also designed experiments using the fluorescent protein pair sfGFP-mCherry to avoid Förster resonance energy transfer between sfGFP-RFP [2]. However, the cloning didn't succeed as there were no correct plasmid constructs in the gel.



Promoters

Aim: Determine the arabinose inducibility of promoter pBAD and the relative transcription ability of pBAD (BBa_I0500) to constitutive promoter p2547 (BBa_J23100) .

✔ Successfully ligated the pBAD and p2547 promoter to mCherry (BBa_K3962340 BBa_K3962339).

✔ Tested the inducibility of L-arabinose on pBAD

✔ Calibration of pBAD with constitutive promoter p2547.

❔ Calibration of other constitutive promoters in BBa_J23100 promoter family.

The combination of a high-expression promoter of antitoxin and a relatively low-expression promoter of toxin is ideal for balancing a single TA system [4]. This approach prevents creating unnecessary constructs of constitutive promoters ligated with toxins. To simplify the experiment further, we first used mCherry as a reporter protein to characterize the inducibility of pBAD (BBa_I0500). Afterwards, we compared the transcriptional activity of pBAD with constitutive promoter p2547 (BBa_J23100) (Figure 4).

Figure 4 Selection of the promoters

Figure 4: Selection of the promoters

The results showed pBAD was tightly regulated by L-arabinose from concentration 0.00032% to 1% (w/v). Then, under the same condition, we relate mCherry expression of pBAD to constant expression of constitutive promoter p2547 (Figure 5). Their correlation was fit into a linear regression with R2 = 0.9586. The characterization of this part provides us with crucial information to select promoters based on the predicted toxin/antitoxin ratio in equilibrium states. The selected promoters can be used to assemble a single TA system to balance their expression.

Figure 5The fluorescence intensity of strains carrying our construct under different arabinose concentration. The measurement was carried out for 20 h. The x-axis shows the time of each measurement. The y-axis shows the fluorescence of mCherry in the arbitrary unit (AU). The blue data points represent the fluorescent intensity of p2547::mCherry strain, red data points represent the pBAD::mCherry strain. The arabinose concentration is shown at the top of each graph.

Figure 5: The fluorescence intensity of strains carrying our construct under different arabinose concentration. The measurement was carried out for 20 h. The x-axis shows the time of each measurement. The y-axis shows the fluorescence of mCherry in the arbitrary unit (AU). The blue data points represent the fluorescent intensity of p2547::mCherry strain, red data points represent the pBAD::mCherry strain. The arabinose concentration is shown at the top of each graph.

To improve the reliability of our results, we intended to use another constitutive promoter to calibrate pBAD. However, we didn't succeed to ligate it with mCherry by 3A assembly.



Toxins

Aim: Determine which toxins are most effective in inhibiting cell growth.

✔ Inducible promoter pBAD was successfully ligated to each toxin gene (see the Parts Page).

✔ Toxin CcdB and HOK exhibited inhibitory activity when adding 1% arabinose as inducers.

✔ The MazF and RelE showed no influence on E. coli TOP10 strain in liquid medium.

❔ Toxicity assay based on LB-agar plates may be another method to assess toxin lethality.

❔ Toxin may work better in minimal medium which resemble natural environment.

The toxicity of toxins is crucial for one function of the DOPL LOCK system: preventing HGT. We evaluated the inhibitory effects of four toxins on bacterial growth based on the constructs of pBAD::toxin. By adding arabinose inducers, we observed distinct inhibitory effects of the CcdB and HOK toxins. The toxins MazF and RelE didn't show influence on the growth rate of bacteria. We propose using minimal medium or LB-agar plates as alternative assessment approaches in the future.

Figure 6The overnight growth curve of pBAD::CcdB (left) and pBAD::HOK (right) TOP10 strain under different arabinose concentrations. The high arabinose concentration was highlighted as red while the lower concentration was blue. It can be observed that the inhibitory effects increased as the concentration of arabinose increased.

Figure 6: The overnight growth curve of pBAD::CcdB (left) and pBAD::HOK (right) TOP10 strain under different arabinose concentrations. The high arabinose concentration was highlighted as red while the lower concentration was blue. It can be observed that the inhibitory effects increased as the concentration of arabinose increased.



Antitoxins

Aim: Determine the expression level required for antitoxin to neutralize toxicity.

✔ Constitutive promoter p162 (BBa_J23117) was successfully ligated to antitoxin CcdA and RelB(see the Parts Page).

❔ Toxicity assay with antitoxin cassettes.

The characterization of antitoxins could be done by co-transforming plasmids containing pBAD::toxin cassettes and constitutive promoter::antitoxin cassettes. The neutralizing effects of antitoxin could be assessed by gradually induced toxin gene expression. However, due to time restrictions, we only achieved ligation of p162 with CcdA and RelB by PCR-based cloning.

Part assembly

Creating the DOPL LOCK system

In the final stage of our lab time, we decided to refrain from balancing single TA systems due to limited lab time. Instead, We focused on creating the full DOPL LOCK system with constitutive promoters regulating both toxin and antitoxin genes. We successfully cloned the constructs of antitoxins p162::CcdA, p162::RelB, and of toxins p21::CcdB, p21::RelE by PCR-based cloning. However, 3A assembly of these constructs to the desired plasmid backbones didn't succeed.

Discussion

In summary, we characterized the components of the DOPL LOCK system to achieve the balance of TA gene expression levels of TA systems. The characterization result provided us with information to select the most compatible parts for the full DOPL LOCK system. We successfully characterized all parts except the antitoxins in our experiments.

For the choice of Oris, our results showed that p15A and pBR322/ROB were the most compatible Ori pair, as they showed high levels of fluorescence when co-transformed. However, pSC101 and pUC have a low expression in RFP when co-transformed. The first version of the DOPL LOCK system was based on plasmids with the Oris p15A and pBR322/ROB.

For promoters, the inducibility of pBAD was first analyzed by ligating to the mCherry fluorescent protein cassette. Then it was calibrated with the constitutive promoter p2547 successfully. The results could be applied to predict appropriate inducer concentrations to balance TA systems.

For the selection of the toxin, the toxicity assay showed that the toxins CcdB and HOK exhibited significant inhibitory activity on E.coli when induced by 1% arabinose. However, the toxins MazF and RelE showed no effect on the bacterial growth. Toxins tested in our experiment did not show instant kill effects, while previous research used them as kill switches [5]. There are several aspects that could explain these unexpected results. For instance, there is a possibility that the antitoxin gene pre-existed in our strains and neutralized the toxic effects as the four toxins we tested originated from E. coli [6]. Another possible explanation is that the strains already gained resistance to toxin before the expression of the toxin was induced. Future experiments could include colony counting units as an outcome measurement and/or use minimal medium to evaluate the toxic effect in alternative approaches. Furthermore, our model shows the highest concentration of MazF reached is about 3 × 10-4 mM when 1% of arabinose is treated to bacteria, which is not effective based on the results from CCU Taiwan in 2017. Using constitutive promoters instead of inducible promoters to regulate the MazF gene may ensure its toxicity on the cells.

Aside from modified methods, including more TA systems is also an important future aspect for the DOPL LOCK system. Toxins that act as nucleases, like HOK, are preferred as they prevent HGT further by degrading any DNA that could be taken in by environmental bacteria.

References

  1. https://blog.addgene.org/plasmid-101-origin-of-replication
  2. All natural. (2007). Nature Chemical Biology, 3(7), 351-351. https://doi.org/10.1038/nchembio0707-351
  3. Bajar, B. T., Wang, E. S., Zhang, S., Lin, M. Z., & Chu, J. (2016). A Guide to Fluorescent Protein FRET Pairs. Sensors (Basel, Switzerland), 16(9), 1488. https://doi.org/10.3390/s16091488
  4. Chan, C. T. Y., Lee, J. W., Cameron, D. E., Bashor, C. J., & Collins, J. J. (2016). 'Deadman' and 'Passcode' microbial kill switches for bacterial containment. Nature Chemical Biology, 12(2), 82-86. https://doi.org/10.1038/nchembio.1979
  5. Valenzuela-Ortega, M., & French, C. (2021). Joint universal modular plasmids (JUMP): a flexible vector platform for synthetic biology. Synthetic Biology, 6(1). https://doi.org/10.1093/synbio/ysab003
  6. Xie, Z., Qi, F., & Merritt, J. (2013). Development of a Tunable Wide-Range Gene Induction System Useful for the Study of Streptococcal Toxin-Antitoxin Systems. Applied and Environmental Microbiology, 79(20), 6375-6384. https://doi.org/doi:10.1128/AEM.02320-13
  7. Yamaguchi, Y., & Inouye, M. (2011). Regulation of growth and death in Escherichia coli by toxin–antitoxin systems. Nature Reviews Microbiology, 9(11), 779-790. https://doi.org/10.1038/nrmicro2651

Introduction

In our design, two TA systems are required to balance the expression of antitoxin and toxin. There are four critical components that determine the balance in this system, namely plasmid Ori, promoter in front of both toxin and antitoxin genes, toxin and antitoxin. To build up the system, we first collected multiple parts for each component and characterized them in a standardized protocol to determine which combination of all four components was the best option.

Characterization of origin of replication

RFP switch

To determine the compatibility of different Oris, we first obtained JUMP plasmids [1] from Addgene. This is a collection of plasmids designed by Valenzuela-Ortega et al., containing different Oris and sfGFP as the reporter [1].

We switched the already present sfGFP cassette of the JUMP plasmids to a mRFP1 cassette ( BBa_J04450). The products were transformed into E. coli DH5α and plated on agar plates containing kanamycin and spectinomycin. We isolated the plasmids from the successful transformants. Then, we restricted it with EcoRI & PstI and ran the gel electrophoresis (Supplementary Figure 1). The result showed that we successfully constructed four pJUMP plasmids with different Oris and the same RFP cassette.

Short description, in 3 words

Supplementary Figure 1: The gel result of RFP switches. The first lane was the DNA ladder. The four lanes on the left were original pJUMP-4x plasmids with sfGFP as control (pJUMP46: 2311bp, 902bp; pJUMP47: 3039bp, 902bp; pJUMP48: 2508bp, 902bp; pJUMP49: 3039bp, 902bp). The other four lanes are pJUMP-4x plasmids with RFP cassette after switching (pJUMP46: 2311bp, 1110bp; pJUMP47: 3039bp, 1110bp; pJUMP48: 2508bp, 1110bp; pJUMP49: 3039bp, 1110bp). Some contamination was observed in pJUMP47-RFP.

sfGFP-RFP Cotransformation

Vectors containing sfGFP or RFP (listed in Table 1) were co-transformed into E. coli TOP10 in 16 combinations. Sixteen co-transformed strains and eight single-transformed strains (used as control) were cultured overnight for the measurement.

First, they were diluted to OD600 = 0.5 and transferred to 96-well plates for fluorescence measurement via plate reader (Spark, Tecan Group Ltd, Zürich, Switzerland).

Table 1: The plasmids and their constructs used in the co-transformation

NAME RESISTANCE ORIGIN OF REPLICATION INSERT INCOMPATIBILITY GROUP COPY NUMBER
pJUMP26 Kanamycin p15A sfGFP B Medium/Low copy number
pJUMP27 Kanamycin pSC101 sfGFP C Low copy number
pJUMP28 Kanamycin pUC sfGFP A High copy number
pJUMP29 Kanamycin pBR322/ROB sfGFP A Medium/High Copy number
pJUMP46 Spectinomycin p15A RFP B Medium/Low copy number
pJUMP47 Spectinomycin pSC101 RFP C Low copy number
pJUMP48 Spectinomycin pUC RFP A High copy number
pJUMP49 Spectinomycin pBR322/ROB RFP A Medium/High Copy number


For the control groups with only one plasmid transformed, we surprisingly observed that the plasmid copy number only had minor impacts on the fluorescence (Figure 7). Minor differences caused by the plasmid copy number of Ori could be observed regarding RFP fluorescence (pUC > pBR322/ROB > pSC101 ≈ p15A). The sfGFP fluorescence of pSC101 and pUC showed much lower signals. Therefore, we focused on the RFP expression in the following data analysis.

Figure 7 Fluorescence of single transformants with different Ori measured by plate reader. The left side shows sfGFP fluorescence from original JUMP plasmids.The right side shows RFP fluorescence of different pJUMP backbones inserted with RFP cassette.

Figure 7: Fluorescence of single transformants with different Ori measured by plate reader. The left side shows sfGFP fluorescence from original JUMP plasmids.The right side shows RFP fluorescence of different pJUMP backbones inserted with RFP cassette.

The co-transformation result showed p15A was highly compatible with pBR322/ROB. When compared with other co-transformation with pUC and pSC101 (Figure 8 and 9), they exhibited both high sfGFP and RFP fluorescence. We also discovered that plasmid combinations among incompatibility groups (shown in Table 1) may also generate poor signals. For instance, pSC101 showed poor RFP signals with all the other plasmids, despite it being the only plasmid from incompatibility C (Figure 10). Strains co-transformed with plasmids containing Ori pUC did not show any patterns related to incompatibility groups or plasmid copy numbers. (Figure 11).

Figure 8 The fluorescence of single-transformed strains of pJUMP46-sfGFP and its co-transformants with other JUMP plasmids with RFP cassette.

Figure 8: The fluorescence of single-transformed strains of pJUMP46-sfGFP and its co-transformants with other JUMP plasmids with RFP cassette.

Figure 9 The fluorescence of single-transformed strains of pJUMP49-RFP and its co-transformants with other JUMP plasmids with sfGFP.

Figure 9: The fluorescence of single-transformed strains of pJUMP49-RFP and its co-transformants with other JUMP plasmids with sfGFP.

Figure 10 The fluorescence of single-transformed strains of pJUMP47-RFP and its co-transformants with other JUMP plasmids with sfGFP cassette.

Figure 10: The fluorescence of single-transformed strains of pJUMP47-RFP and its co-transformants with other JUMP plasmids with sfGFP cassette.

Figure 11 The fluorescence of single-transformed strains of pJUMP48-RFP and its co-transformants with other JUMP plasmids with sfGFP cassette.

Figure 11: The fluorescence of single-transformed strains of pJUMP48-RFP and its co-transformants with other JUMP plasmids with sfGFP cassette.

In summary, the Ori pair of p15A and pBR322/ROB showed excellent compatibility (Figure 12). Both sfGFP and RFP fluorescence in co-transformed cells were even larger than the control. pJUMP26 and pJUMP49 were selected as the backbone of the DOPL LOCK. Whether these two plasmids both stably maintained in bacteria was verified by confocal microscopy (Figure 13). The microscopy imaging showed low heterogeneity of the culture. Most bacteria expressed both sfGFP and RFP.

Figure 12 Fluorescence of co-transformed p15A and pBR322/ROB (in the middle) with their single-transformed control group (at the sides) with error bars. Green bars represent the sfGFP fluorescent signal while Red bars represent RFP signals.

Figure 12: Fluorescence of co-transformed p15A and pBR322/ROB (in the middle) with their single-transformed control group (at the sides) with error bars. Green bars represent the sfGFP fluorescent signal while Red bars represent RFP signals.


Figure 13 The confocal microscopy image of the <em>E. coli</em> strain co-transformed with pJUMP49-RFP and pJUMP49-sfGFP. (A) The green fluorescence emission of the strain at 510 nm. (B) The red fluorescence emission of the strain at 608 nm. (C) The strain under normal light. (D) The merged RGB image of the green and red fluorescence.

Figure 13: The confocal microscopy image of the E. coli strain co-transformed with pJUMP49-RFP and pJUMP49-sfGFP. (A) The green fluorescence emission of the strain at 510 nm. (B) The red fluorescence emission of the strain at 608 nm. (C) The strain under normal light. (D) The merged RGB image of the green and red fluorescence.


sfGFP-mCherry Co-transformation

Even though we found the best combination of Oris in the sfGFP-RFP co-transformation assay, we also noticed the Förster resonance energy transfer (FRET) existed between two fluorescent proteins [2]. To avoid that, we performed experiments to substitute RFP with mCherry. However, the plate reader showed that the plasmid heterogeneity was present in strains transformed with plasmids containing pSC101 and pBR322/ROB (Figure 14). Due to time restriction, we stuck to previous results from the sfGFP-RFP co-transformation for the following experiment.

Figure 14 Fluorescence of single transformants with JUMP plasmids with different Oris. The fluorescence was measured by the plate reader.

Figure 14: Fluorescence of single transformants with JUMP plasmids with different Oris. The fluorescence was measured by the plate reader.

Characterization of promoters

The major goal of our initial experiments is to balance the amount of toxin and antitoxin in bacterial cells. Ideally, all the toxin and antitoxin genes in our system should be regulated by constitutive promoters to prolong the state of balance. To do this, we calibrated inducible promoter pBAD (BBa_I0500) with constitutive promoter BBa_J23100 (referred to as p2547 in the following). pJUMP49 plasmids containing pBAD::mCherry (BBa_K3962340) and p2547::mCherry (BBa_K3962339) were transformed into E. coli TOP10 strain. Their fluorescence under different L-arabinose concentrations was measured.

Short description, in 3 words

Supplementary Figure 2: The gel result for plasmids involved in the inducibility test restricted with EcoRI and PstI. (A) The first lane shows the plasmid construct of pJUMP49-pBAD::mCherry (2964bp, 2096bp) and the third lane is the control plasmid pJUMP49-sfGFP (2964bp, 902bp). (B) The second lane shows the plasmid construct of pJUMP49-p2547::mCherry (2964bp, 922bp)

Calibration of pBAD with p2547

The result showed constant expression from p2547::mCherry and the expression from pBAD::mCherry increased gradually as the concentration of arabinose increased (Figure 15). The lowest inducibility was observed at the concentration of 0.00032% (w/v). We also concluded that pBAD is tightly regulated as it showed low leaky expression when no arabinose was added.

Figure 15The fluorescence intensity of strains carrying our construct under different arabinose concentration. The measurement was carried out for 20h. Blue lines and dots were fluorescent of p2547 while red ones were from pBAD. The arabinose concentration was shown at the bottom of each graph.

Figure 15: The fluorescence intensity of strains carrying our construct under different arabinose concentration. The measurement was carried out for 20h. Blue lines and dots were fluorescent of p2547 while red ones were from pBAD. The arabinose concentration was shown at the bottom of each graph.

The relation between the arabinose concentration (w/v, log5) and the relative fluorescence of pBAD fitted nicely into a linear regression model with R2=0.9586 (Figure 16). From the regression, it could be calculated that at the arabinose concentration of 0.14% the transcriptional activity of pBAD is equal to p2547.

Figure 16 Linear regression of arabinose concentration relative fluorescence of pBAD compared with p2547. The orange dash lines showed how p2547 related to pBAD transcriptional activity at the arabinose concentration of 0.14% (log<sub>5</sub>-1.22).

Figure 16: Linear regression of arabinose concentration relative fluorescence of pBAD compared with p2547. The orange dash lines showed how p2547 related to pBAD transcriptional activity at the arabinose concentration of 0.14% (log5-1.22).

Detailed information about this part can be found on the Measurement Page.

Characterization of toxins

In this project, we chose three type II TA systems (CcdA/CcdB, RelB/RelE, MazE/MazF) and one type I TA system (HOK/SOK) as candidates for DOPL LOCK. In type II TA systems, both toxin and antitoxin are proteins and the antitoxins act as antagonists to the toxin activities [3]. While in the type I TA system, the antitoxin is a small antisense RNA that base-pairs with the toxin encoding mRNA [4]. All antitoxins are degraded more rapidly than toxins.

Ligation

To validate the toxicity, the toxin genes were first ligated to an inducible promoter pBAD and transformed into the TOP10 strain independently (all constructs are shown in the Parts page). The gel electrophoresis showed correct constructs of pBAD::toxin (Figure 17).

Figure 17 The gel result for plasmids containing constructs of pBAD::toxin. These plasmids were restricted with EcoRI and PstI. (A) The second lane shows the plasmid construct of pJUMP27-pBAD::MazF (2947bp, 1718bp). The third lane shows the plasmid construct of pJUMP27-pBAD::HOK (2947bp, 1809bp). The fourth lane shows the plasmid construct of pJUMP27-pBAD::RelE (2947bp, 1676bp). (B) The second lane shows the plasmid construct of pJUMP27-pBAD::CcdB (2947bp, 1688bp).

Figure 17: The gel result for plasmids containing constructs of pBAD::toxin. These plasmids were restricted with EcoRI and PstI. (A) The second lane shows the plasmid construct of pJUMP27-pBAD::MazF (2947bp, 1718bp). The third lane shows the plasmid construct of pJUMP27-pBAD::HOK (2947bp, 1809bp). The fourth lane shows the plasmid construct of pJUMP27-pBAD::RelE (2947bp, 1676bp). (B) The second lane shows the plasmid construct of pJUMP27-pBAD::CcdB (2947bp, 1688bp).

OD assay

We first tested the overexpression of the toxins in DH5α strain. However, the DH5α strain metabolizes arabinose, which was used as the inducer of pBAD in our experiments. The induction of arabinose was not enough to induce bacterial death. Therefore, we switched to the E. coli strain TOP10 for the toxicity assay.

All strains containing the pBAD::toxin construct first grew overnight in LB medium, and were treated with different concentrations of arabinose at the exponential growth phase (OD = 0.5). Their ODs were measured by the plate reader overnight in intervals of 30 min. On the same plate, LB medium was used as blank control. Bacteria with no plasmids and bacteria containing the pBAD::mCherry construct were used as negative control. The bacteria treated with antibiotics were used as a positive control. Both strains were treated at the same arabinose concentration as toxin-producing strains.

From the plate reader results, the TOP10 control strain showed that the arabinose concentration used had no influence on cell growth. pBAD::mCherry strains showed cells treated with 1% arabinose had a lower OD value after 9 h (Supplementary Figure 3). This indicated that the high-level expression of proteins induced a burden to the bacterial growth. The result of toxin-producing strains showed that different toxins have different effects on bacterial growth. CcdB was the most promising toxin as it is the only one that inhibited bacterial growth significantly at the arabinose concentration of 0.5% (Figure 18). HOK also showed inhibitory effects but was less toxic than CcdB (Figure 19). Both toxins had higher inhibitory effects than pBAD::mCherry strain. Surprisingly, MazF and RelE showed no influence on bacterial growth (Figure 20 and 21).

Short description, in 3 words

Supplementary Figure 3: The overnight growth curve of pBAD::RelE TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.
Figure 18 The overnight growth curve of pBAD::CcdB TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 18: The overnight growth curve of pBAD::CcdB TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 19 The overnight growth curve of pBAD::HOK TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 19: The overnight growth curve of pBAD::HOK TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 20 The overnight growth curve of pBAD::MazF TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 20: The overnight growth curve of pBAD::MazF TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 21 The overnight growth curve of pBAD::RelE TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

Figure 21: The overnight growth curve of pBAD::RelE TOP10 strain under different arabinose concentrations. The high arabinose concentration is highlighted as red while the lower concentration is blue.

One explanation for the low toxicity of MazF and RelE is that the bacteria proliferate rapidly in nutrient-rich medium and only essential metabolic pathways were triggered. Therefore, we tried to use M9 minimal medium for the toxicity assay. However, the optimization of the M9 minimal medium and the rest of this research line were abandoned due to time constraints. For further optimization of the experiments, this could be a focus for future experiments.

Characterization of antitoxins

To investigate the antagonistic ability of the antitoxins, plasmids that constitutively expressed an antitoxin were required. Then, they could be co-transformed with the pBAD::toxin plasmids which had been cloned and validated above (Figure 22). By inducing different levels of toxin expression, the appropriate expression levels of antitoxin required to neutralize the effect of toxicity could be identified.

First, we used restriction enzyme cloning to ligate the constitutive promoters with the antitoxin. However, it turned out to be really difficult as the length of promoters we used were less than 100 bp. We also tried to increase the molar ratio between the promoter gene and plasmid backbone to more than 200:1, but didn't succeed either. To overcome this issue, we changed the cloning method to PCR-based cloning, in which the constitutive promoters were parts of the primers. We designed cloning primers (listed in the Experiments page) and optimized the conditions of PCR including the primer concentrations and melting temperature. The gel electrophoresis showed successful ligation of p162::CcdA and p162::RelB (Figure 22) (see the Parts page). However, due to time limitations, we didn't perform toxicity assays with plasmids carrying antitoxin cassettes. The ligation products were directly used to create the full DOPL LOCK system.

Figure 22 Gel electrophoresis results of p162::CcdA (second lane, 420 bp), p21::CcdB (third lane, 506 bp), p21::RelE (fourth lane, 494bp), and p162::RelB (fifth lane, 441bp).

Figure 22: Gel electrophoresis results of p162::CcdA (second lane, 420 bp), p21::CcdB (third lane, 506 bp), p21::RelE (fourth lane, 494bp), and p162::RelB (fifth lane, 441bp).

Part assembly

In our design, the inducible promoters were designed to be ligated with antitoxin genes as the function of a kill switch, and the constitutive promoters were designed to be ligated with the toxins. This construct can also be the tool to balance the single TA expressions. However, we only ligated the pBAD promoter with MazE and RelB by restriction enzyme cloning. Ligations of other constructs didn't succeed.

In the final stage of our lab time, we decided to skip balancing single TA systems. We tried to create the full DOPL LOCK system with constitutive promoters regulating both toxin and antitoxin genes. The constitutive promoters p162 (BBa_J23117) and p21 (BBa_J23113) were selected]. We successfully created the constructs of antitoxins p162::CcdA and p162::RelB (Figure 22), and toxins p21::CcdB and p21::RelE (Figure 22) via PCR-based cloning (see the Parts page). 3A assembly was performed to create constructs of pJUMP26-p162::CcdA-p21::RelE and pJUMP49-p162::RelB-p21::CcdB. No colony could be observed in the plates. The gel electrophoresis result showed the 3A assembly did not succeed.

References

  1. Bajar, B. T., Wang, E. S., Zhang, S., Lin, M. Z., & Chu, J. (2016). A Guide to Fluorescent Protein FRET Pairs. Sensors (Basel, Switzerland), 16(9), 1488. https://doi.org/10.3390/s16091488
  2. Brantl, S. (2012). Bacterial type I toxin-antitoxin systems. RNA Biology, 9(12), 1488-1490. https://doi.org/10.4161/rna.23045
  3. Fraikin, N., Goormaghtigh, F., Melderen, L. V., & Margolin, W. (2020). Type II Toxin-Antitoxin Systems: Evolution and Revolutions. Journal of Bacteriology, 202(7), e00763-00719. https://doi.org/doi:10.1128/JB.00763-19
  4. Valenzuela-Ortega, M., & French, C. (2021). Joint universal modular plasmids (JUMP): a flexible vector platform for synthetic biology. Synthetic Biology, 6(1). https://doi.org/10.1093/synbio/ysab003
  5. Zhang, S.-P., Wang, Q., Quan, S.-W., Yu, X.-Q., Wang, Y., Guo, D.-D., . . . He, Y.-X. (2020). Type II toxin–antitoxin system in bacteria: activation, function, and mode of action. Biophysics Reports, 6(2), 68-79. https://doi.org/10.1007/s41048-020-00109-8

Discussion

For building up the DOPL LOCK system, the major difficulty lies in balancing the gene expression level of toxin-antitoxin (TA) systems. To achieve this, we first characterized the influential parts of TA gene expression, including the origin of replication (Ori) of the plasmids, promoters, toxins and antitoxins. The result of the characterization provided us with information to select the most compatible parts for the full system.

For the selection of the various Oris, our results show four Oris exhibited high-level expressions of GFP and RFP fluorescence when transformed independently (Figure 7). Regarding the plasmid with RFP, the fluorescent patterns were consistent with the previous characterization of JUMP plasmids, where Ori p15A, pSC101 showed lower signals and Ori pBR322/ROB showed moderate signals [1]. However, the expression level difference between these Oris was much greater in the previous research. This could be because the bacterial strain DH5α, instead of the JM109 strain, was used in our experiment.

The compatibility of these Oris was evaluated by co-transformation of two plasmids with different reporter genes and the fluorescence of the co-transformants were measured. The results showed that p15A and pBR322/ROB were the most compatible Ori pair, while pSC101 and pUC have low expression in RFP when co-transformed (Figure 8-11). The first version of the DOPL LOCK system was based on plasmids with the Oris p15A and pBR322/ROB. Their co-transformed strains showed high levels of both sfGFP and RFP fluorescence, even higher than the single-transformed strains.

Interestingly, incompatibility groups did not have a significant influence since we still applied antibiotic selection pressure. This phenomenon has also been observed by Velappan et al [2]. They showed that different plasmids containing identical Oris can be stably maintained in bacteria over 3 days without antibiotic selections [2].

To avoid the Förster resonance energy transfer (FRET) effect, we designed experiments to use the sfGFP-mCherry fluorescent protein pair instead of the sfGFP-RFP pair. However, we observed heterogeneity in plasmids with Oris pSC101 and pBR322/ROB from flow cytometry. A possible explanation is that mutations occurred in the fluorescent protein gene expression cassette. This mutation alleviates the expression burden, resulting in strains that are more competitive and will likely become the dominant phenotypes in the population.

It has been suggested that plasmid gene expression is not only related to elements of the plasmid backbone but also to the E. coli host strain and the cultivation conditions [3]. The plasmid copy number was not the only crucial factor of plasmid gene expression in our characterization experiment. Furthermore, the effect of different Oris on the plasmid copy number has not been well-investigated. For instance, the Ori pUC used in our experiment was generally regarded as a high-copy plasmid [4]. However, it was also reported to yield a lower copy number of up to 15 [3]. Therefore, characterization of Oris is crucial for plasmid gene regulation and should be focused on further experiments.

Additionally, the fluorescent proteins were very stable once formed [5]. In our experiments, they accumulated overnight and reached an equilibrium at around 9 h (Figure 15). This gene regulation of fluorescent proteins may not resemble antitoxins since most antitoxins were rapidly degraded. It would be interesting to use fluorescent proteins with degradation tags to simulate the antitoxin regulation. This would provide more information for selecting which parts to use. We would also like to test more plasmid backbones that are frequently used in Standard European Vector Architecture (SEVA) and Biobricks system to provide more flexibility to our system.

It has been suggested that the combination of high expression of pBAD and the relatively low expression of the promoter makes the pair suitable for studies of TA systems. We designed the same strategy to balance the TA systems. The pBAD (BBa_I0500) was calibrated with p2547 (BBa_J23100) for the selection of low expression promoters that regulate toxin genes. Our results show that when 0.14% L-arabinose is added, pBAD will express the same level of protein as p2547 (Figure 16). This result is in accordance with research of Santos et al., where the pBAD transcription activity is higher than p2547 at an arabinose concentration higher than 0.1% [6]. The calibration helped us predict the appropriate promoters that could be used to balance TA systems. Taking CcdB/CcdA as an example, the antitoxin–toxin ratio greater than one could prevent the harmful effect of CcdB in plasmid-containing bacteria [7]. We may predict that, in the TA system comprising p2547::CcdB and pBAD::CcdA, the antitoxins likely neutralize toxic effects when a concentration higher than 0.14% of arabinose is used.

For the selection of the toxins, the toxicity assay showed CcdB and HOK exhibited significant inhibitory activity on E.coli when induced by 1% arabinose (Figure 18-19). However, the MazF and RelE showed no effect on the bacterial growth (Figure 20-21). The toxins tested in our experiment did not show the instant kill effects as described in previous research, where they are used as kill switches. As the four toxins we tested originated from E. coli, there is a possibility that the antitoxin gene already existed in our strains and neutralized the toxic effects. It is also shown in literature that the TA system is tightly regulated in E. coli, as bacterial stress responses can trigger the transcription of chromosomally encoded TA systems without releasing active toxins [8]. The overexpression of toxins in our experiment may not exceed the regulation ability of the pre-existing TA system.

Another possible explanation for the low toxicity is that resistant strains had already existed in the culture before overexpression of the toxin was induced. The mutation for escaping toxins is developed in 3 - 12 h when the toxin is over-expressed, according to Garcia et al [9]. Here, they found that the mutations that inactivate the toxins were primarily in the −10 promoter region of the plasmids [10]. Other research showed that bacteria escaped biocontainment predominantly by inactivating mutations in the toxin genes [10]. As a single TA system could be inactivated easily by mutations, inclusion of additional TA pairs or multiple toxin genes should further reduce the plasmid escape rate.

Besides mutation, the growth environment is also an influential factor for toxicity. CFU cell viability assays are another frequently used method to evaluate how toxins influence bacterial biofilm formation. It has been shown by Kolodkin-Gal et al. that different TA systems affected bacterial cell death differently in the liquid medium and during biofilm formation [11]. In our human practices, Nathan Fraikin also suggested applying colony counting units (CFU) instead of optical density measured at 600 nm (OD600) to assess the toxic effects on biofilm formation in the future. It has been shown that induction of RelE or MazF transcription does not confer cell killing but instead induces a static condition in which the cells are still viable but unable to proliferate [12]. The bacteriostatic effects of these two toxins may only be observed via CFU cell viability assays. Therefore, the CFU cell viability assay is suggested as a future experiment in order to clarify the effect of the different toxins on cellular growth.

Additionally, when discussing our experiments with Prof. Dr. Dennis Claessen, he advised us to use minimal medium instead of LB-broth to test the toxin efficiency. This would resemble the natural environment much better, would the organism accidently escape from the biocontainment. It has been discovered that the ability of MazE to reverse the bactericidal effect of MazF was less effective in M9 minimal medium than in the rich LB medium [13]. However, additional nutrients need to be added to the minimal medium for different strains. Because of the limited lab time, we didn't succeed in optimizing the components of the minimal medium. Therefore, this could be an opportunity for further experiments to improve the toxin efficiency and characterization.

To gain more insights into the regulation of the toxin gene, we built a model to investigate how arabinose influences the expression of pBAD::MazF. The model shows the highest concentration of MazF reached is about 3 × 10-4 mM when 1% of arabinose is treated to bacteria. It is suggested that this concentration is not sufficient for toxicity, based on the results from CCU Taiwan in 2017. After reaching the peak, the concentration of MazF decreases over time due to the degradation of the arabinose. Using constitutive promoters instead of inducible promoters to regulate the MazF gene may ensure its toxicity on the cells.

All figures mentioned in the discussion panel can be found in the results panel.

References

  1. Afif, H., Allali, N., Couturier, M., & Van Melderen, L. (2001). The ratio between CcdA and CcdB modulates the transcriptional repression of the ccd poison–antidote system. Molecular Microbiology, 41(1), 73-82. https://doi.org/https://doi.org/10.1046/j.1365-2958.2001.02492.x
  2. Amitai, S., Yassin, Y., & Engelberg-Kulka, H. (2004). MazF-mediated cell death in Escherichia coli: a point of no return. Journal of Bacteriology, 186(24), 8295-8300. https://doi.org/10.1128/JB.186.24.8295-8300.2004
  3. Chan, C. T. Y., Lee, J. W., Cameron, D. E., Bashor, C. J., & Collins, J. J. (2016). 'Deadman' and 'Passcode' microbial kill switches for bacterial containment. Nature Chemical Biology, 12(2), 82-86. https://doi.org/10.1038/nchembio.1979
  4. Damalas, S. G., Batianis, C., Martin-Pascual, M., de Lorenzo, V., & Martins dos Santos, V. A. P. (2020). SEVA 3.1: enabling interoperability of DNA assembly among the SEVA, BioBricks and Type IIS restriction enzyme standards. Microbial Biotechnology, 13(6), 1793-1806. https://doi.org/https://doi.org/10.1111/1751-7915.13609
  5. Fernandez-Garcia, L., Kim, J.-S., Tomas, M., & Wood, T. K. (2019). Toxins of toxin/antitoxin systems are inactivated primarily through promoter mutations. Journal of Applied Microbiology, 127(6), 1859-1868. https://doi.org/https://doi.org/10.1111/jam.14414
  6. Jahn, M., Vorpahl, C., Hübschmann, T., Harms, H., & Müller, S. (2016). Copy number variability of expression plasmids determined by cell sorting and Droplet Digital PCR. Microbial Cell Factories, 15(1), 211. https://doi.org/10.1186/s12934-016-0610-8
  7. Kolodkin-Gal, I., Verdiger, R., Shlosberg-Fedida, A., & Engelberg-Kulka, H. (2009). A Differential Effect of E. coli Toxin-Antitoxin Systems on Cell Death in Liquid Media and Biofilm Formation. PloS one, 4(8), e6785. https://doi.org/10.1371/journal.pone.0006785
  8. Lee, C. L., Ow, D. S. W., & Oh, S. K. W. (2006). Quantitative real-time polymerase chain reaction for determination of plasmid copy number in bacteria. Journal of Microbiological Methods, 65(2), 258-267. https://doi.org/https://doi.org/10.1016/j.mimet.2005.07.019
  9. LeRoux, M., Culviner, P. H., Liu, Y. J., Littlehale, M. L., & Laub, M. T. (2020). Stress Can Induce Transcription of Toxin-Antitoxin Systems without Activating Toxin. Molecular Cell, 79(2), 280-292.e288. https://doi.org/https://doi.org/10.1016/j.molcel.2020.05.028
  10. Pedersen, K., Christensen, S. K., & Gerdes, K. (2002). Rapid induction and reversal of a bacteriostatic condition by controlled expression of toxins and antitoxins. Molecular Microbiology, 45(2), 501-510. https://doi.org/https://doi.org/10.1046/j.1365-2958.2002.03027.x
  11. Tombolini, R., Unge, A., Davey, M. E., de Bruijn, F. J., & Jansson, J. K. (1997). Flow cytometric and microscopic analysis of GFP-tagged Pseudomonas fluorescens bacteria. FEMS Microbiology Ecology, 22(1), 17-28. https://doi.org/10.1111/j.1574-6941.1997.tb00352.x
  12. Valenzuela-Ortega, M., & French, C. (2021). Joint universal modular plasmids (JUMP): a flexible vector platform for synthetic biology. Synthetic Biology, 6(1). https://doi.org/10.1093/synbio/ysab003
  13. Velappan, N., Sblattero, D., Chasteen, L., Pavlik, P., & Bradbury, A. R. M. (2007). Plasmid incompatibility: more compatible than previously thought? Protein Engineering, Design and Selection, 20(7), 309-313. https://doi.org/10.1093/protein/gzm005

Find Out More

©2021 All rights reserved.

Contact Us

  • Email
  • igem@science.leidenuniv.nl
  • Address
  • Sylviusweg 72, 2333BE Leiden
    the Netherlands