A major problem with cancer therapies is the low specificity of many treatment drugs. Conventional therapies are often administered in a systemic fashion, leading to numerous unwanted off-target effects. To mitigate such issues, there is thus a need for targeted drug delivery systems for anticancer drugs.

Bacteria have long been considered promising candidates for drug delivery. Some strains of bacteria are highly anaerobic, and can only survive in hypoxic environments. As the tumor microenvironment is one of the few few hypoxic sites in the human body. Anaerobic bacteria will therefore thrive in hypoxic tumour cores and die in oxygenated regions, allowing greater specificity in targeting. Clostridium novyi, an anaerobic bacterium, was found to successfully reduce tumour sizes in phase I clinical trials. However, Clostridium is relatively more difficult to manipulate than other strains of bacteria, making it an unideal choice of a delivery vector.

Besides strictly anaerobic microbes, there exist other bacteria strains which are facultative anaerobes, the most well-known being Escherichia coli. These strains can survive in both aerobic and anaerobic environments by changing their gene expression programmes accordingly. Thus, we can utilise the natural ability of E. coli to express certain genes under anaerobic conditions to express therapeutic genes/drugs only in the hypoxic core of the tumour.

However, the tumour core is not the only hypoxic location in the body. For example, the bone marrow and gut are also hypoxic environments in which the anaerobic expression programme may be activated. As a result, there is a need for more specific control of regulation of the therapeutic drugs or genes.

Quorum sensing (QS) is a system by which bacteria communicate with each other according to the density of the population. Through this system, bacteria are able to coordinate changes in gene expression and thereby alter their phenotype to better adapt to the environment. Our group aims to harness this system to regulate gene expression in our modified bacteria. As more bacteria gather in the hypoxic tumour core, they sense the increase in population density, and then express the gene of interest. Therefore, by placing the gene of interest under both hypoxic and QS regulation, we can precisely regulate the expression of the therapeutic drug or gene.

The esaR binding site however, would be our repressor site, where esaR will bind to when there’s no quorum sensing. esaR is a homologue of luxR. Unlike luxR which is activated upon binding to AHL, esaR loses its ability to bind on the binding site after forming a complex with AHL. Generally, esaR is paired with another protein - esaI. esaI synthesises AHL, which interferes the binding capabilities of esaR.

High metabolic activity and dysfunctional vasculature often result in oxygen deficiency in solid tumour cores (Smallbone, Gatenby & Maini, 2005). This “hypoxic core” is an anaerobic environment, which contrasts with the aerobic environment in most other regions of the body. Therefore, we propose to use an anaerobic switch so that the target genes will specifically be transcribed at the tumour region in response to the anaerobic environment.

In the facultative anaerobe Escherichia coli K-12, the transcription factor Fumarate and Nitrate reductase Regulatory (FNR) protein is encoded by the fnr gene (Spiro & Guest, 1987). FNR activates various fnr-dependent promoters in the anaerobic pathway of E. coli (Spiro & Guest, 1987), such as the nirB promoter (pnirB), which drives the expression of NADH-dependent nitrite reductase (Jayaraman et al., 1987). pnirB has partial activity when induced under anaerobic conditions and optimal activity in the presence of nitrite in addition to anaerobic growth conditions (Griffiths & Cole, 1987; Jayaraman et al., 1987). We have selected pnirB as the switch to induce the expression of target genes under anaerobic conditions. As the non-pathogenic E. coli K-12 strain BL21 (Life Technologies) that we used expresses the fnr gene endogenously, pnirB will be regulated by FNR protein in this system.

High metabolic activity and dysfunctional vasculature often result in oxygen deficiency in solid tumour cores (Smallbone, Gatenby & Maini, 2005). This “hypoxic core” is an anaerobic environment, which contrasts with the aerobic environment in most other regions of the body. Therefore, we propose to use an anaerobic switch so that the target genes will specifically be transcribed at the tumour region in response to the anaerobic environment.

In the facultative anaerobe Escherichia coli K-12, the transcription factor Fumarate and Nitrate reductase Regulatory (FNR) protein is encoded by the fnr gene (Spiro & Guest, 1987). FNR activates various fnr-dependent promoters in the anaerobic pathway of E. coli (Spiro & Guest, 1987), such as the nirB promoter (pnirB), which drives the expression of NADH-dependent nitrite reductase (Jayaraman et al., 1987). pnirB has partial activity when induced under anaerobic conditions and optimal activity in the presence of nitrite in addition to anaerobic growth conditions (Griffiths & Cole, 1987; Jayaraman et al., 1987). We have selected pnirB as the switch to induce the expression of target genes under anaerobic conditions. As the non-pathogenic E. coli K-12 strain BL21 (Life Technologies) that we used expresses the fnr gene endogenously, pnirB will be regulated by FNR protein in this system.

Our aim was to first characterise the different parts used in the above two-switch regulatory system (Figure 1). As such, we designed the following four parts flanked by the BioBricks Prefix (BBPrefix) and BioBricks Suffix (BBSuffix):

1. pnirB-invasin (BBa_K1804002)
2. pnirB-invasin-listeriolysin
3. plac-green fluorescent protein (gfp) (BBa_K1804001)
4. plac-invasin

In pnirB-invasin (BBa_K1804002), the invasin (inv) gene is placed under the control of pnirB. Similarly, the invasin (inv) and listeriolysin O (LLO) genes are placed under the control of pnirB in pnirB-invasin-listeriolysin. As such, bacteria transformed with plasmids encoding pnirB-invasin can invade mammalian cells under anaerobic conditions and bacteria cells transformed with plasmids encoding pnirB-invasin-listeriolysin can invade mammalian cells, as well as escape the endosome.

The constitutive lacI promoter (plac) (Shong & Collins, 2013) driving the expression of invasin (plac-invasin) was chosen as a positive control for pnirB-invasin (BBa_K1804002) to show that invasin expressed from pnirB-invasin was functional through an invasion assay. In order to ensure that plac was functional, plac-gfp (BBa_K1804001) was constructed and functionally characterised by measuring fluorescence intensities.

Unfortunately, our team only managed to construct and functionally characterise plac-gfp (BBa_K1804001) and pnirB-invasin (BBa_K1804002).

plac-gfp was constructed by prepending the KpnI sequence and appending the XhoI sequence to the start and stop codons of gfp (BBa_E0040) respectively. This was done via extension polymerase chain reaction (PCR). Restriction enzyme digest with KpnI and XhoI was then carried out to replace esaR and esaI with the PCR product in the pAC-esaR-esaI plasmid (Figure 1). This placed gfp under the control of the plac promoter. The resultant pAC-plac-gfp plasmid encoded gfp under the control of plac.

The BioBricks Prefix and Suffix were then prepended and appended to the start of plac and end of gfp specific primers, respectively, to produce BBPrefix-plac-gfp-BBSuffix by PCR. To increase the percentage of cleavage during restriction digest, primers with short (referred to as junk) sequences prepended to BBPrefix (FP_BBP_Junk) and appended to BBSuffix (RP_BBS_Junk) were used to amplify BBPrefix-plac-gfp-BBSuffix by PCR. The resultant PCR product Junk-BBPrefix-plac-gfp-BBSuffix-Junk was then cloned into pSB1C3 by restriction digest with EcoRI and PstI followed by ligation and transformation into E. coli BL21 to produce plac-gfp in pSB1C3 (Figure 2). Figure 3 shows the gel electrophoresis of the PCR products formed after PCR of the correct pSB1C3-plac-gfp clone, PCR-purified Junk-BBPrefix-plac-gfp-BBSuffix-Junk (positive control), pAC-plac-gfp (negative control) and water (negative control) using the BBPrefix and BBSuffix primers. Colony PCR was performed with the BBPrefix and BBSuffix primers to confirm if clone was correct.

plac-invasin (Figure 3) was intended to be constructed by two rounds of extension PCR of the invasin gene from BBa_K299812. The first round of primers are FP1_placInv and RP2_invendBBsuffix. The second round of primers are FP2_BBprefixplac and RP2_invendBBsuffix. Unfortunately, due to time constraints, the construction of this part is incomplete.

pnirB-invasin (Figure 5) was constructed by extension PCR of the invasin gene from BBa_K299812 using BBPrefix-pnirB as the forward primer and a reverse primer for the end of invasin fused to BBSuffix (invendBBSuffixLonger) (Figure 6). BBPrefix-pnirB-invasin-BBSuffix was then amplified into Junk-BBPrefix-pnirB-invasin-BBSuffix-Junk using FP_BBP_Junk and RP_BBS_Junk (Figure 7). Junk-BBPrefix-pnirB-invasin-BBSuffix-Junk was then cloned into pSB1C3 by restriction digest with EcoRI and PstI followed by ligation and transformation into E. coli BL21. The clone was confirmed with colony PCR of Biobricks Prefix/Suffix primers and internal primers invlloF1/InvendBBsuffixlonger (Figure 8).

 

The plac promoter is constitutively active in our strain of E. coli BL21. After construction of the pSB1C3-plac-gfp vector, the activity of the plac promoter was measured by quantification of fluorescence intensity. BL21 was used as the negative control and BL21 transformed with pAC-plac-gfp as the positive control. As seen in Figure 5, GFP is expressed in both strains of BL21 containing the plac-gfp vectors but not in the negative control. When fluorescence data was quantified in arbitrary units (A.U.), the intensity of plac-gfp in pSB1C3 was lower relative to that of plac-gfp in the pAC vector (Figure 9). Nonetheless, our results indicated that plac is indeed able to drive gene expression under normal growth conditions. Supplementary Table 1 shows the descriptive statistics of the fluorescence intensities determined.

The invasin protein from Yersinia facilitates bacterial invasion into mammalian cells via an integrin dependent mechanism (Leong et al., 1990). We induced the expression of invasin by growing B21 carrying pSB1C3-pnirB-invasin under anaerobic conditions. After 2 and 6 hours, the bacteria were allowed to invade a monolayer of HepG2 mammalian cells in an invasion assay, where results are represented in colony-forming units (CFU) of bacteria successful in invading mammalian cells (Figure 7). Our results show that BL21, and BL21 with pSB1C3-nirB-invasin grown under aerobic conditions have low invasion ability. BL21 has low invasion ability under both aerobic and anaerobic conditions . However, BL21 with pSB1C3-pnirB-invasin exhibits a 25-fold increase in invasion compared to the control grown at aerobic conditions, and this increase is statistically significant (p-value < 0.05) (Figure 11). It is notable that after 6 hours of growth in anaerobic conditions, the pnirB-invasin construct can no longer confer an increased invasive phenotype, which may be due to the high fitness cost of producing the transgenic invasin protein.

From our experiments, we have demonstrated that the lacI promoter is able to drive the expression of a target gene (in this case, gfp) in pAC and pSB1C3. It also suggested that pnirB is able to control the invasion phenotype conferred by invasin expression in BL21 to some degree. However, there are some limitations, as we have not fully characterised the optimal conditions for activating the pnirB promoter. It is also unknown if the expression of the invasin gene results in high fitness cost that might affect the ability of transformed bacteria to serve as a drug delivery vehicle. This invasion phenotype can be further characterised with the construction of the pSB1C3-plac-invasin vector.

Anderson, J. C., Clarke, E. J., Arkin, A. P., & Voigt, C. A. (2006). Environmentally controlled invasion of cancer cells by engineered bacteria.Journal of molecular biology, 355(4), 619-627.

Griffiths, L., & Cole, J. A. (1987). Lack of redox control of the anaerobically-induced nirB+ gene of Escherichia coli K-12. Archives of microbiology, 147(4), 364-369.

Jayaraman, P. S., Peakman, T. C., Busby, S. J. W., Quincey, R. V., & Cole, J. A. (1987). Location and sequence of the promoter of the gene for the NADH-dependent nitrite reductase of Escherichia coli and its regulation by oxygen, the Fnr protein and nitrite. Journal of molecular biology, 196(4), 781-788.

Jayaraman, P. S., Gaston, K. L., Cole, J. A., & Busby, S. J. W. (1988). The nirB promoter of Escherichia coli: location of nucleotide sequences essential for regulation by oxygen, the FNR protein and nitrite. Molecular microbiology, 2(4), 527-530.

Leong, J. M., Fournier, R. S., & Isberg, R. R. (1990). Identification of the integrin binding domain of the Yersinia pseudotuberculosis invasin protein. The EMBO journal, 9(6), 1979.

Shong, J., & Collins, C. H. (2013). Engineering the esaR promoter for tunable quorum sensing-dependent gene expression. ACS synthetic biology, 2(10), 568-575.

Spiro, S., & Guest, J. R. (1987). Regulation and over-expression of the fnr gene of Escherichia coli. Journal of general microbiology, 133(12), 3279-3288.

Smallbone, K., Gavaghan, D. J., Gatenby, R. A., & Maini, P. K. (2005). The role of acidity in solid tumour growth and invasion. Journal of Theoretical Biology,235(4), 476-484.

 

Extension PCR was carried out to amplify target parts with specific flanking sequences, followed by deoxyribonucleic acid (DNA) gel electrophoresis to check for the amplification of the target parts. After the required part was created, restriction enzyme digest, ligation and transformation into bacteria (E. coli BL21 or dh5α) were carried out. The bacteria were then plated onto LB agar with the appropriate antibiotic and grown overnight in 37 ℃ in a stationary incubator. Colony PCR or PCR of miniprepped plasmids were then conducted to select for positive clones. For minipreps, bacteria was inoculated into LB media containing the appropriate antibiotic and grown overnight at 37 ℃ with constant agitation at 250 revolutions per minute (rpm).

Green fluorescence protein (GFP) fluorescence intensity quantification was carried out after growth of transformed cells expressing GFP in LB media with the appropriate antibiotic. The optical density (OD) of the bacteria was then measured using a spectrophotometer (UV-1800 Shimadzu Spectrophotometer) at 600 nm and diluted to an OD of 0.5. The bacteria were then visualised under a confocal laser scanning microscope (Confocal Leica SPE2). The wavelength of the laser used was 488 nm, and the fluorescent signal detection range was between 500 and 540 nm, corresponding to the wavelength of the light emitted by GFP. Each experiment was carried in duplicates for the sample, positive control, and negative control.

The GFP fluorescence intensities were measured using ImageJ (Figure 12). For each group (sample, positive control and negative control), 15 bacteria cells were selected randomly from at least two images and their fluorescence intensities quantified, unless the fluorescence intensity was too low to be detected. Graphpad Prism 6.0 was used to analyse the fluorescence intensities.

Bacteria were grown in Luria Broth (LB) at 37 ℃ with shaking at 250 rpm, while mammalian cells were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with fetal bovine serum (FBS) in a humidified CO2 incubator. HepG2 cells were seeded in 24-well plates at a concentration of 2.5 x 105 cells/well and grown overnight. Stationary cultures of E. coli BL21 and BL21 + pSB1C3-pnirB-invasin were prepared by inoculating one colony in 3 mL LB and grown overnight. Bacteria were then diluted 10X in LB, and grown for 2 hours and 6 hours at 30 degrees with shaking at 250 rpm under anaerobic conditions in a BD GasPak EZ Incubation Container with activated Gazpak, and under aerobic conditions for 6 hours. The O.D. values of the cultures were then measured.

The cells in the 24-well plates were washed with 500 ul sterile phosphate buffered solution (PBS). Following which, 500 ul of fresh antibiotic-free DMEM with FBS was added into each well. The bacteria cells were then diluted in DMEM, and used to infect the cells with a multiplicity of infection of 100:1 for three hours in the CO2 incubator. Subsequently, the cells were washed with PBS, and DMEM with 1000 ug/mL kanamycin was added for a brief incubation of 1 hour in the CO2 incubator. Each well was then washed twice with PBS, and 100 ul of 0.1% Triton X-100 was subsequently added to lyse the cells. Serial dilutions of the lysate were carried out and plated on LB agar. After overnight growth at 37 ℃, the number of colonies on each plate was counted and colony forming units was calculated. Statistical tests were conducted using GraphPad Prism 6.0.

For further details on our experimental protocols and daily lab work, please refer to our Protocols page and Notebook page .

The gfp gene was obtained from basic BioBricks Part BBa_E0040.

The plac sequence 5’ TTTACACTTTATGCTTCCGGCTCGTATGTT 3’ was obtained from the pAC-EsaR-EsaI plasmid (Figure 1) (Shong & Collins, 2013), which was a gift from Cynthia Collins (Addgene plasmid # 47660).

The pnirB sequence, which is 5’ CATTAAGGAGTATAAAGGTGAATTTGATTTACATCA ATAAGCGGGGTTGCTGAATCGTTAAGGTAGGCGGTA 3’, was obtained from Jayaraman et al., 1988.

The sequences of the primers of our own design with start and end sites with respect to the submitted and theoretical parts used can be downloaded here.