The design of RIOT sensor 10 (BBa_K1897028) and 11 (BBa_K1897029) have higher sensitivity to lactate and lower basal expression as compared to the natural promoter and its modified version, lldRO1-J23117-lldRO2, designed by team ETH_Zurich 2015 (BBa_K1847008). However, to evaluate whether which design is optimal for biomolecule delivery systems, further characterization is required. We replaced the reporter gene sfGFP by Alanine Racemase (ALR, BBa_K1172901) by PCR overlap using appropriate primers and generate a RIOT sensor collection (Fig 1).

Figure 1: Design of the RIOT sensor collection

This collection is a combination of three promoters and three ribosomal binding sites (RBS) of different strengths. The promoters consists of a modified lactate sensor p70 (BBa_K1847008), a shorter version of the promoter region for wild-type lldPRD operon (BBa_K1897037, derived from Part BBa_K822000) called p62 and a constitutively expressed promoter J23100 (BBa_K314100).

Alanine Racemase (ALR) is an enzyme that catalyzes the reaction of L-alanine ⇌ D-alanine. D-alanine is an essential amino acid that is required for bacterial cell survival (Walsh, 1989). One of our goals is to make the RIOT sensor activated in the presence of high lactate concentration, which enhance the survival of the bacteria in tumour environment. Detailed working mechanism of RIOT sensors can be found at the Introduction page. When being integrated into the whole system, RIOT sensors are expected to increase the successful rate of detection of cancer cells or delivery of biomolecules that kill cancer cells while leaving normal cells unharmed.

We transformed E. coli Nissle ∆alr, a D-alanine auxotrophic mutant bacteria (Hwang et al, 2016) with RIOT sensors 1 – 9. This is an engineered probiotic microbe which aims to eliminate and prevent pathogen infection in the mammalian gut. RIOT sensors, which express ALR in the presence of L-lactate, will be able to rescue the transformed bacteria grown in media without D-alanine supplement. Therefore, the performance of RIOT sensors were assessed by the growth rate of transformed bacteria over a range of lactate concentration in M9 minimal media via natural logarithm (LN) of OD600. We used linear regression analysis to fit the data of bacterial growth rate using four data points of LN (OD600) values from 90 to 180 min of each sensor. This is because the growth rate of each constructs appeared to deviate from each other at the 90-minute time point (Figure 3) . We postulated that induction of ALR by lactate might take about 90 minutes for these sensors. The slope of each fitted linear line represents the growth rate of bacteria (P-value < 0.01).

Since naturally occurring amino acids have an L-configuration in mammals and D-alanine was also shown to be absent from human plasma protein (Hoeprich, 1965; Nagata, Masui & Akino, 1992), bacteria transformed with a well-controlled sensor will not be able to grow in the absence of both D-alanine and lactate. The results shown in Figure 2 suggested that RIOT sensors 1 – 6 with strong and medium RBS exhibited high leakiness. The transformed bacteria were able to survive after 6 hours even without addition of lactate. Moreover, Table 1 showed that transformed bacteria with sensors 1 and 2, regardless of lactate concentration appeared to have similar initial growth rate compared to those were grown in the media containing D-alanine supplement (0.008 – 0.009 min-1, P-value < 0.01). This is probably because the amount of ALR expressed at the lowest lactate concentration is still sufficient to convert enough L-alanine into D-alanine, which help the bacteria to maintain their normal growth. The insignificant differences in growth rate of bacteria treated with various lactate concentrations indicated that RIOT sensor using lldRO1-J23117-lldRO2 promoter ligated with strong RBS designed by team ETH_Zurich 2015 was not sensitive enough to detect small changes in lactate concentration in the tumour microenvironment.

Figure 2: Comparison among RIOT sensors with strong and medium ribosomal binding site (RBS) for sensitivity to L-lactate

1. (A) RIOT sensor 1: p70-34-ALR-Terminator (BBa_K1897019)
2. (B) RIOT sensor 2: p62-34-ALR-Terminator (BBa_K1897022)
3. (C) RIOT sensor 3: J23100-34-ALR Terminator (BBa_K1897025)
4. (D) RIOT sensor 4: p70-32-ALR-Terminator (BBa_K1897020)
5. (E) RIOT sensor 5: p62-32-ALR-Terminator (BBa_K1897023)
6. (F) RIOT sensor 6: J23100-32-ALR Terminator (BBa_K1897026)

D-alanine auxotrophic mutant bacteria were transformed with plasmids containing ALR sequence under the control of RIOT sensors with strong or medium RBS. Their overnight culture were spun down and resuspended with M9 minimal media without D-alanine supplement. Then, they were subjected to different treatments: without and with addition of D-alanine supplement (50 μg/ml) and with addition of lactate of 10^-4 to 10^-2 M. Their OD600 readings were measured for 6 hours. LN of OD600 readings reflects the bacterial growth rate.

Since RIOT sensor 3 and 6 contain a constitutively expressed promoter, they served as positive controls for Alanine Racemase expression.
A-C: N=1, D-F: N = 3 and the error bars stand for SEM.

Table 1: Comparisons of bacteria growth rate among RIOT sensors with weak and strong ribosomal binding site.

RIOT sensors 7 and 8 contain a weak RBS (N = 3 ± SEM). RIOT sensors 1 and 2 contain a strong RBS (N = 1 ± SEM).

Doubling time = LN (2) / Growth rate.

In contrast, RIOT sensors 7, p70-33-ALR-Terminator (BBa_K1897021) and 8, p62-33-ALR-Terminator (BBa_K1897024) with weak RBS appeared to exhibit very minimal leakiness (Figure 3). The transformed bacteria were unable to grow in the absence or at low lactate concentration (10^-4 M). The addition of lactate at high concentration (10^-3 and 10^-2 M) rescued the transformed bacteria grown in media without D-alanine. And as expected, RIOT sensor 9, J23100-33-ALR Terminator (BBa_K1897027) which served as a positive control, was able to express ALR constitutively. According to Table 2, the growth rate of bacteria transformed with sensor 7 and 8 grown in lactate concentration of 10^-4 M was not significantly different from zero. Moreover, RIOT sensor 7 exhibited an increase in growth rate or a decrease in doubling time from 92.7 to 79.5 minutes when lactate concentration increased from 10^-3 M to 10^-2 M. However, the growth and doubling time of sensor 8 were almost similar when lactate concentration increased.

In conclusion, we have characterized the output of lldRO1-J23117-lldRO2 (BBa_K1847008 from ETH_Zurich 2015) in a new chassis which was E. coli Nissle ∆alr, a D-alanine auxotrophic mutant bacteria to detect lactate in tumour microenvironment. The characterization result of RIOT sensor 7 shows that we have successfully minimized the leakiness and enhance the sensitivity to lactate of sensor using the lldRO1-J23117-lldRO2 promoter from ETH_Zurich without adding LldP and LldR. Bacteria transformed with sensor 7 are expected to be able to survive in small numbers in mild lactate concentration of the blood and once reaching tumour site, their growth will be increased by the high lactate concentration there, which will enhance the detection signal or the delivery of biomolecules into tumour cells.

Figure 3: Comparison among three RIOT sensors with weak ribosomal binding site (RBS) for sensitivity to L-lactate

1. (A) RIOT sensor 7: p70-33-ALR-Terminator (BBa_K1897021)
2. (B) RIOT sensor 8: p62-33-ALR-Terminator (BBa_K1897024)
3. (C) RIOT sensor 9: J23100-33-ALR Terminator (BBa_K1897027) contains a constitutively expressed promote and served as a positive control for Alanine Racemase expression.

D-alanine auxotrophic mutant bacteria were transformed with plasmids containing ALR sequence under the control of three RIOT sensors with weak RBS. Their overnight culture were spun down and resuspended with M9 minimal media without D-alanine supplement. Then, they were subjected to different treatments: without and with addition of D-alanine supplement (50 μg/ml) and with addition of lactate of 10^-4 to 10^-2 M. Their OD600 readings were measured for 6 hours. LN of OD600 readings reflects the bacterial growth rate. Mean LN(OD600) was calculated with N=3 and the error bars stand for SEM.

• Hoeprich, P. D. (1965). Alanine: Cyeloserine Antagonism. VI. Demonstration of D-Alanine in the Serum of Guinea Pigs and Mice. Journal of Biological Chemistry, 240, 1654-60.
• In Young Hwang, Elvin Koh, Adison Wong, John C. March, William E. Bentley, Yung Seng Lee and Matthew Wook Chang (2016). Engineered probiotic microbes eliminate and prevent pathogen infection in the mammalian gut. Manuscript submitted.
• Nagata, Y., Masui, R., & Akino, T. (1992). The presence of free D-serine, D-alanine and D-proline in human plasma. Experientia, 48(10), 986-988.
• Walsh, C. T. (1989). Enzymes in the D-alanine branch of bacterial cell wall peptidoglycan assembly. Journal of biological chemistry, 264(5), 2393-2396.