Team:Paris Bettencourt/Project/Foot Odor

Figure 1. B. subtilis is gram positive bacterium which is commonly found in the human foot flora, which degrades the leucine aminoacid present in the sweat to produce isovaleric acid. Our strategy is to perturb the leucine degradation pathway by knocking out the vital enzymes in the pathway. The leucine dehydrogenase knock out mutants lacks the capability to produce leucine and prevents the conversion of leucine to 4-methyl 2-oxopentanoate. The isovaleryl-coA dehydrogenase α and β subunit knock mutants lacks the capability of conversion of 4-methyl 2-oxopentanoate to isovaleryl-coA.

Aims and Achievement

Foot odor is generally perceived as socially awkward and negative. Although there are a many commercially available solutions for this problem, current products indiscriminately target bacteria on the foot skin microbiome. These type of products can have negative effects on skin health and microbiome dynamics. We aim to develop a targeted approach to prevent foot odor, by selectively killing microbes responsible for the biosynthesis of volatile compounds which compose the characteristic stinky feet smell, without destroying the beneficial microbes.

Introduction

Isovaleric acid is the characteristic odorant of foot odor. It is produced primarily from leucine, found in the sweat, via the leucine degradation pathway (Fig. 1). Of the many bacterial species found on the foot, only Bacillus frequency correlates with foot odor intensity (Ara et al., 2006). We therefore chose to focus our project on the leucine degradation pathway of the model microbe B. subtilis.

We hypothesize that a leucine degradation mutant of B. subtilis will produce less odor, yet still survive and replicate on the foot. A large population of mutants on the foot would exclude the wild-type strain and therefore attenuate odor. If the mutant strain can persist in the foot ecosystem, bacterial foot deodorants could provide a cheap alternative to traditional chemical deodorizers. .

Figure 2. Diffusion model of isovaleric acid through air. This animation shows the concentration of isovaleric acid diffusing through air from a foot over a time period of 0 to 1000 hours. From the graph on the left of the animation, you can see that the odor threshold is reached very early in the model, given that the concentration of isovaleric acid at the skin surface is kept at a constant concentration of 1.1*10^-6 M.

Figure 3. The B. subtilis wild type and mutant strains were grown overnight in M9 minimal medium. A double blind smell test was performed. The subjects were asked to rank the tubes in the scale of 0 to 5. The data from the smell test was analyzed and the error bars plotted are designated with 95% confidence interval. Though there is variation observed between the wild type strain and the bcd knock out strain, it is not statistically significant. In the case of bkdaα knock out strain and the wild type strain there is a statistically significant variation observed. This confirms that the bkdaα knock out strain produces less malodor compare to the wild type.


Figure 4. The B. subtilis wild type and mutant strains were grown over night in M9 minimal medium supplemented with 0%, 0.1% and 0.2% leucine. A double blind smell test was performed. The subjects were asked to smell the tubes and rank them in the scale of 0 to 5. The data from the smell test was analyzed and the error bars plotted are designated with 95% confidence interval. Though there is variation observed between the 0%, 0.1% and 0.2% leucine supplemented medium in bkdaβ knock out strain, the difference is not statistically significant. But in the case of the wild type, bcd knockout and bkdaα knockout strains supplemented with 0%, 0.1% and 0.2% leucine, there is a statistically significant variation observed. This confirms that the leucine concentration influence the malodor production.

Results

We found B. subtilis mutants ΔBKDAα knock out and ΔBKDAβ knockout to produce less odor than the wild type. We obtained 3 deletion strains from the Bacillus Genetic Stock Center (BGSC): ΔBCD, ΔBKDAα knockout and ΔBKDAβ knockout. Each of these mutants blocks the metabolic pathway connecting leucine to Isovaleryl-CoA. Double-blind smell tests with human smellers revealed that cultures of ΔBKDAα knockout and ΔBKDAβ knockout had significantly less intense odor than wild type. No significant difference was detected for the ΔBCD strain (Fig. 3).

Because our smell tests used human noses, we built a model to estimate the sensitivity of our instruments (Fig. X). Using COMSOL Multiphysics, we simulated odor diffusion as a system of differential equations. Isovaleric acid can be smelled at a concentration of 80 parts per trillion (Nagata, 1993). After diffusion equilibrium, this corresponds to a liquid phase concentration of 10 mg/L at a distance of 5 cm, comparable to the leucine concentration in human sweat (Callewaert et al., 2014). Our model confirms that human noses are able to detect isovaleric acid at physiologically relevant concentrations.

We next investigated the leucine-dependence of odor production in the B. subtilis and mutant strains. Strains were grown in M9 minimal media supplemented with 0%, 0.1% or 0.2% leucine. All strains produced detectable odor in all conditions, indicating that B. subtilis produces some odorant molecules independent of the leucine pathway. Odor production in the wild-type strain was strongly leucine dependant, suggesting leucine metabolism contributes significantly to odor. The odor of the BKDAβ knockout strain showed no leucine dependence, consistent with the successful inactivation of an odor pathway.

If mutant B. subtilis are to exclude the wild type on the human foot, they must be able to grow on sweat. We found that growth of all 3 mutant strains was moderately impaired on LB media, and severely impaired on synthetic sweat, relative to wild type (Figure X).

However, leucine-degradation mutant strains of B. subtilis showed robust survival on socks. We dipped cotton socks in dense M9 cultures of ΔBCD, ΔBKDAα knockout, ΔBKDAβ knockout or wild-type strains. Socks were hung to dry and sampled daily for the density of B. subtilis CFUs. After 5 days, we detected no significant decline in population viability. This suggests that, once introduced to socks, a metabolic mutant strain could persist and modulate odor production for an extended time.

Methods

Synthetic sweat: The chemical constituents present in the human sweat were analyzed with the aid of gas chromatography and mass spectroscopy. Synthetic sweat was prepared by diluting the exact concentration of all the amino acids and salts found in human sweat, the pH of the synthetic sweat was adjusted to 6.5 with the aid of a pH meter. The synthetic sweat was then autoclaved and stored at 4°C to avoid bacterial contamination.

M9 minimal media: The minimal media was prepared by diluting 1 ml of 50% glucose 1.1 ml of mgso4.3H20 2 ml of 1% w/v casmino acids 2 ml of 10 mg/ml of tryptophan in 100 ml of distilled water The media is then filter sterilized under a laminar air flow cabinet.

Strain: The leucine dehydrogenase, isovaleryl coA alpha subunit and isovaleryl coA beta subunit. The knockout strains were obtained from Bacillus Genetic Stock Centre, where the knockout library of the B. subtilis was generated by replacing each gene with the erythromycin cassette. The mutant strains are trp- so they require supplementation of tryptophan in the growth media. The knockout strains were validated with the PCR reaction.

Growth curve: The growth kinetics of the wild type and the mutant strains of B. subtilis were studied with the aid of micro plate reader to have insights about the fitness advantage and disadvantage within the mutant and wild type strain.

Smell test: The presence of isovaleric acid in the bacterial culture was sensed with double blind smell test, where neither the subject nor the experimenter is aware of the tube labels. In order to optimize this smell test we grew the B. subtilis in odorless M9 minimal media. The media was supplemented with 0.1% of leucine to study the influence of leucine in production of foot odor.

Sock experiment: The bacterial cell culture is diluted in synthetic sweat till it reaches 0.1 OD. The sock was soaked in the synthetic sweat and hanged till it dry. After this step 2 cm2 of sock was cut and soaked in 5 ml of 1% PBS. The tubes were stirred gently and supernatant is diluted in PBS to prepare 1X and 10X concentration of B. subtilis. The bacteria were then plated on LB agar.

Diffusion model of isovaleric acid: The model is based on a 2D axisymmetric geometry, with the air modeled as the space 5 cm around the skin surface. Here, the skin surface is kept at a constant concentration, C0 = 1.1E-6 M. This value refers to the approximate amount of isovaleric acid in sweat (Harvey, 2010). The diffusivity of isovaleric acid through the air was calculated using the Chapman-Enskog theory of gas diffusivity given by the following equation:
D = [1.858*10^-3* T^-3/2*√(1/M₁+1/M₂)] / [p * s₁₂² * w)]

- D is the diffusivity of the isovaleric acid.
- T is the room temperature (298 K).
- p is the air pressure (1 atm).
- M₁ is the mass of the isovaleric acid (102.13 g/mol).
- M₂ is the mass of the air (29 g/mol).
- s₁₂ is the average collision diameter, which is around 340 Angstroms for gas molecules in air.
- w is the dimensionless temperature-dependent collision integral, usually on the order of 1.

Thus, the diffusion coefficient of isovaleric acid was tabulated to be 1.7*10^-9 m² / s.

GC-MS analysis: For the GC-MS analysis, 1µl of the sample was injected into the instrument in split mode. A Rtx5MS- 30m column with 0.25-mm ID and 0.25µm df were used. The GC-MS run uses the following standardized parameters:

• Injection temperature: 300°C
• Interface temperature: 300°C
• Ion Source: 250°C
• Carrier Gas: Helium (flow rate of 1 ml min-1)

The temperature program used was 3 min of isothermal heating at 50⁰C followed by heating at 350⁰C for 10 mins. Mass spectra were recorded at 2 scan sec-1 with a scanning range of 40 to 850 m/z.
X Caliber software is used to process and analyze the data from a GC-MS run. Quantify each component based on peak areas and normalization based on the internal standard.