Overview

Aiming at developing a customizable meal replacement in terms of flavour and function, we constructed a system consisting of several modules.

The core module is a genetic circuit which can be induced by light and temperature simultaneously to express embedded structural genes selectively.

The flavour module contains enzymes that catalyze the transformation from basic biomolecules such as glucose and amino acids to special scent molecules. These enzymes are reconstructed into the core circuit to make the flavor of our meal replacement alterable. Available choices of flavours are vanilla, nut, and sweet rose.

The function module is a collection of three proteins – Amuc_1100, ovalbumin, and RuBisCO. They are expected to provide weight-loss, muscle-building, and stress-relief functions respectively. These proteins are embedded into the core circuit together with flavour-producing enzymes to achieve on-demand and easy-to-use substance production.

Disclaimer: Our project only promotes the idea of customizable meal replacement. Although we have conducted in vitro experiments to validate our design, our project does NOT support food-level validation, i.e., testing our product on any animals, including humans.

Chassis selection

To find the best chassis bacterium, we ask ourselves: what is the ultimate form of our proposed product? Obviously, we cannot simply use engineered E. coli that are grown on normal LB medium and feed it to our customers. The best combination of our chassis bacterium and its culture medium should be close to ordinary food. Hence, we chose engineered Lactobacillus delbroeckii subsp. bulgaricus (L. bulgaricus) and milk to produce a customizable meal replacement milkshake. Milk is a perfect medium for L. bulgaricus to grow and produce our desired substances, and the whole culture can be drunk – just like ordinary yogurt!

But that is not the whole story – many kinds of bacteria can ferment in milk; how did we choose the best one? Several factors are considered:

Safety. Current studies show that L. bulgaricus does not produce pathogenic substances, and its cell wall contains no lipopolysaccharide which can be a detrimental endotoxin in the human body. Moreover, lactobacillus-derived plasmid vectors have good species specificity and are therefore not susceptible to horizontal gene transfer.

Fermentation ability. L. bulgaricus is capable of producing a wide range of nutrients and has well-established food-grade culture technology. Since our sight extends towards entrepreneurship, L. bulgaricus can be a top-notch option for the chassis and fermentation bacteria. As we investigate deeper about L. bulgaricus, we learnt that it exclusively produces L-lactic acid from glucose via homofermentative pathways at an easily achieved 45~62℃.

The difficulty of genetic engineering. We have found a precedent for plasmid electro-transformation in L. bulgaricus (Zink et al., 1991). Moreover, the CRISPR-Cas system in L. bulgaricus has been studied by a few research groups.

Regretfully, due to the delivery difficulty caused by COVID-19 and the hardship of experimentation techniques, we failed to complete our lab work on L. bulgaricus. For replacement, we used E. coli to test our systems and managed to complete several full cycles of experimenting and engineering. In conclusion, our proposed chassis is L. bulgaricus and the experimental chassis is E. coli.

Core module: 2×2 transfer switch

The 2×2 transfer switch is derived from a bistable switch which functions by expressing two different genes in the presence and absence of blue light. We modified it to form two genetic circuits that work exclusively at high and low temperatures respectively. Together, these two circuits make up a transfer switch which intakes a combination of light and temperature and outputs an expected substance out of four options.

Blue light/Temperature	High	Low
Present	1	2
Absent	3	4

The photoactive bistable switch

The bistable switch utilizes LOV-bound lacI (lacILOV) to achieve blue light-induced state transition. When blue light is absent, the activity of lacI is maintained, which inhibits the expression of lac operon. As blue light is shed over the bacteria, the alteration of LOV conformation deteriotes genes in the lac operon are able to express. One gene in the lac operon is the desired production protein or enzyme (here, we use mCherry to denote the expression), and the other is cI:LVA which encodes λ-repressor that eliminates the activity of cI:LVA promoter in another direction. On the contrary, the absence of blue light leads to the recovery of cI:LVA promoter activity, which enables downstream gene to be expressed.

Specifically speaking, LOV is a photo-sensitive domain which deforms when cast light of 465nm. CI:LVA expresses a λ-repressor (the cI part) and a degradation tag (the LVA part and subsequent ssrA tag), and the degradation effect raises its cycling rate, in order to make the bistable switch more effective. CI binds to cI:LVA promoter and prohibits its downstream transcription. lacILOV is expressed using a strong and broad-spectrum promoter J23119.

The RNA thermometers

The RNA thermometers are in essence two temperature-sensitive RBSs. Below 37℃, RBS NoChill-06 forms a secondary structure that inhibits translation, which means it only works when the temperature is above 37℃. In contrast, RBS F2 works exclusively under 27℃. We integrated NoChill-06 into the bistable switch to make it work only under high temperatures, and F2 for low temperatures.

NoChill-06 is able to form intramolecular hydrogen bonds, and consequently, it transforms to a condensed conformation which blocks translation. The Gibbs free energy of such a secondary structure rises with temperature, making it more and more unstable (Zadeh et al., 2011). Above 37℃, NoChill-06 can no longer hold the folded conformation and regains its linear structure, permitting the downstream gene to be translated.

Temperature/℃	Gibbs free energy/(kcal/mol)
25	-19.93
30	-18.16
37	-15.70

F2 utilizes endogenous RNase E to achieve temperature-controlled activity. At higher temperature (above 37°C), the RC segment (RNase E cleavage site) of the element is truncated by RNase E, allowing the element to terminate translation. At lower temperatures (below 27°C), the ARC segment (Anti-RNase E cleavage site) is complementarily paired with the RC segment to form a stem-loop structure that cannot be sheared by RNase E, leaving the element intact and ready for downstream gene expression. Without the ARC segment, the element would be permanently shut down.

Flavor and functions

Biosynthesis of scent molecules

Vanilla

Vanilla flavor is mellow, elegant, delightful, and akin to a velvety marshmallow. The origin of this special flavor is vanillin, i.e., 4-hydroxy-3-methoxybenzaldehyde, a cymbidium-derived perfume that is commonly used. Our synthesis method is inspired by the method of Ni et al. (2015), which mimics the natural pathway for the de novo synthesis of vanillin.

The synthesis pathway initiates from intrinsic molecules sources like tyrosine (Try). Three genetic circuits expressing multiple enzymes are designed to finish the whole reaction. The first circuit transforms tyrosine into ferulic acid, using three enzymes in a polycistron. The second circuit utilizes two enzymes to produce vanillin out of ferulic acid. Additionaly, the third circuit is to upregulate the amount of tyrosine in the bacterium cell, by expressing four enzymes which sequentially turn glucose into tyrosine.

Sweet rose

Sweet rose has a sugary, balmy, and comforting flavor. We utilize endogenous phenylalanine (Phe) and three exogeneous enzymes to produce 2-phenylethanol which possesses a special fragrance of sweet rose (Dunkel et al., 2014).

The synthesis pathway involves three enzymes. AT is an aminotransferase that is encoded by bcaT, PDC stands for phenylpyruvate decarboxylase encoded by aro10, and ADH can carry out the redox between 2-phenylethanol and its ketone form. For experimental convenience, we attached a GFP reporter to the 3’-end of the ADH gene, so as to validate and quantify the expression of the proteins in the polycistron.

Nut

Nut flavor is woody, toasty, and savory; you can always feel its lingering aftertaste between the lips and teeth. We constructed a circuit that transforms endogenous leucine (Leu) to 3-methylbutylaldehyde, the molecule that gives out the special nut scent (Dunkel et al., 2014).

Production of functional proteins

Amuc_1100: Weight-loss

The Amuc_1100 protein is a thermostable outer-membrane protein of Akkermansia muciniphila. A. muciniphila is a gram-negative bacterium which colonizes the intestines of humans and rodents. Studies have revealed the relevance of A. muciniphila abundance decline with various digestive system diseases and metabolic disorders. Specifically, Anhê and Marette (2017) demonstrated that Amuc_1100 is positively correlated with mice obesity alleviation and gut-barrier integrity protection. Therefore, we choose Amuc_1100 as a weight-loss ingredient, hoping to promote the idea of using this protein in dietary schemes.

Ovalbumin: Muscle-building

Ovalbumin, derived from Gallus gallus (jungle fowl), is the main component of the protein in egg whites (approximately 55%). In fowl eggs, egg white proteins provide significant, abundant and comprehensive sources of amino acids for embryonic growth. As people always consume eggs and their products as daily food, we choose ovalbumin as a safe, nutritional and easy-to-synthesize source of amino acids for muscle building.

RuBisCO: Stress-relief

RuBisCO (ribulose-1,5-bisphosphate carboxylase) is the key enzyme in photosynthesis that determines the rate of carbon assimilation and is also vital in plant photorespiration. Studies have shown that in the mammalian gut, RuBisCO can be digested into rALPs that have anti-anxiety activity, which has been proven by experiments on mice. Hence, we choose a subunit of RuBisCO which is capable of producing rALPs as a potential non-addictive and effective anti-anxiety product.

Biosafety: Oxygen-controlled Suicide

Introduction

The ccdA-ccdB toxin-antitoxin system, combined with an oxygen-sensitive regulating circuit, is used as a suicide switch. When engineered bacteria are exposed to atmospheric concentrations of oxygen, the switch is activated and the abundance of toxin proteins in the cell increases dramatically, resulting in effective cell death.

Mechanism

Most prokaryotic chromosomes contain toxin-antitoxin modules, which often consist of a pair of genes encoding two components: one of which is a stable toxin protein, while the other is an unstable homologous antitoxin. The ccdA-ccdB toxin-antitoxin system we use is a kill switch that is non-toxic to human and other mammalian cells. The mechanism of ccdB is to prevent DNA pro-cyclase from functioning, which leads to double-strand breaks in DNA, while ccdA prevents the toxicity of ccdB by forming a tight ccdA-ccdB complex (De Jonge et al., 2009).

FNR is a transcriptional regulator containing a [Fe-S] cluster of sensing proteins that senses oxygen concentration. HIP-1 is a promoter that binds to FNR. Under hypoxic conditions, the FNR protein forms a homodimer and cannot bind to the HIP-1 promoter, which turns on the expression of downstream genes; under normoxic conditions, the FNR dimer is depolymerised, the HIP-1 promoter is repressed and the expression of downstream genes is stopped. (Mengesha et al, 2006)

In our design of the suicide switch, the toxin ccdB coding region is placed below the T7 promoter and the antitoxin ccdA coding region is placed below the FNR-regulated HIP-1 promoter, whose activity is controlled by the oxygen concentration present in the environment. Under hypoxic conditions, the FNR protein forms a homodimer that induces the initiation of downstream ccdA expression; under normoxic conditions, the FNR dimer is depolymerised, the initiation of ccdA transcription is inhibited and the toxin ccdB acts to kill the engineered bacteria.

Visit our Parts page to get the Part names and sequences of our genes and circuits.

References

Anhê, F. F., & Marette, A. (2017). A microbial protein that alleviates metabolic syndrome. Nature medicine, 23(1), 11-12.

De Jonge, N., Garcia-Pino, A., Buts, L., Haesaerts, S., Charlier, D., Zangger, K., ... & Loris, R. (2009). Rejuvenation of CcdB-poisoned gyrase by an intrinsically disordered protein domain. Molecular cell, 35(2), 154-163.

Dunkel, A., Steinhaus, M., Kotthoff, M., Nowak, B., Krautwurst, D., Schieberle, P., & Hofmann, T. (2014). Nature’s chemical signatures in human olfaction: a foodborne perspective for future biotechnology. Angewandte Chemie International Edition, 53(28), 7124-7143.

Zink, A., Klein, J. R., & Plapp, R. (1991). Transformation of Lactobacillus delbrueckii ssp. lactis by electroporation and cloning of origins of replication by use of a positive selection vector. FEMS microbiology letters, 78(2-3), 207-212.

Mengesha, A., Dubois, L., Lambin, P., Landuyt, W., Chiu, R. K., Wouters, B. G., & Theys, J. (2006). Development of a flexible and potent hypoxia-inducible promoter for tumor-targeted gene expression in attenuated salmonella. Cancer biology & therapy, 5(9), 1120-1128.

Ni, J., Tao, F., Du, H., & Xu, P. (2015). Mimicking a natural pathway for de novo biosynthesis: natural vanillin production from accessible carbon sources. Scientific reports, 5, 13670.

Zadeh, J. N., Steenberg, C. D., Bois, J. S., Wolfe, B. R., Pierce, M. B., Khan, A. R., ... & Pierce, N. A. (2011). NUPACK: Analysis and design of nucleic acid systems. Journal of computational chemistry, 32(1), 170-173.