Inspiration

The escalating health crisis associated with excessive sugar consumption has become one of the most pressing global challenges of the 21st century. Diseases like obesity and type 2 diabetes are increasing at alarming rates thereby straining healthcare systems and reducing quality of life worldwide (Imperial College London Diabetes & Endocrine Centre, n.d.). These issues have sparked growing consumer and industry demand for natural, low-calorie alternatives to sugar that are both safe and sustainable.

Sweet proteins represent one of the most promising solutions. Among them, monellin which is derived from the West African fruit Dioscoreophyllum cumminsii stands out for its exceptional potency. It is nearly 3,000 times sweeter than sucrose with its caloric profile being negligible (Farag et al., 2022). However, despite this remarkable property, monellin’s lack of thermal and pH stability has severely limited its practical use in food and beverage products. The protein unfolds irreversibly when heated above 50 °C causing it to lose its sweetness and functional integrity (Liu et al., 2016).

Our team was inspired by both the potential health impact and the scientific challenge of addressing this limitation. The possibility of engineering a stable, scalable, and consumer ready protein sweetener that could replace sugar at industrial levels was both exciting and motivating.

Introduction and Background

Sweet proteins are a remarkable class of naturally occurring proteins that elicit an intense sensation of sweetness when consumed. These proteins, which include thaumatin from Thaumatococcus daniellii, brazzein from Pentadiplandra brazzeana, mabinlin from Capparis masaikai, and monellin from Dioscoreophyllum cumminsii, are all non-caloric and orders of magnitude sweeter than sucrose on a molar basis (Kant, 2005). Unlike conventional sugars which activate sweet taste receptors through direct ligand–receptor binding, sweet proteins bind at exosites. They bind to the heterodimeric T1R2/T1R3 sweet taste receptors on the human tongue thereby stabilizing the receptor in an active conformation that triggers the perception of sweetness (Perez-Aguilar et al., 2019).

Among these proteins, monellin has attracted our attention due to its extreme potency since it is 3000 times sweeter than sucrose (Farag et al., 2022). Native monellin is a heterodimer composed of two non-covalently associated polypeptide chains, designated as chains A and B. These were resolved through X-ray crystallography and deposited in the Protein Data Bank (PDB ID: 3MON) (Ogata et al., 1987). Chain A is a 44-residue polypeptide and chain B is a 50-residue polypeptide. Together they form a compact, globular structure stabilized by hydrophobic interactions and hydrogen bonds. This heterodimeric form is intrinsically unstable because the non-covalent association of the two chains dissociates irreversibly under physiological or industrial conditions. Since the active conformation of monellin’s sweetness depends on the precise assembly of both chains, dissociation results in a complete loss of sweetness perception (Leone et al., 2016). This instability has been a critical barrier to its industrial application in food and beverage products as it fails under standard pasteurization and baking regimes.

To overcome this limitation, researchers engineered a single-chain variant known as MNEI, monellin single-chain, in which chains A and B were covalently linked via a short peptide linker. This design effectively “locked” the two chains into a continuous polypeptide thereby eliminating the problem of dissociation under heat. The structure of MNEI was determined and deposited in the Protein Data Bank under PDB ID: 2O9U (Bank, n.d.). This design addressed one of monellin’s primary weaknesses, chain dissociation, and produced a protein that was easier to express recombinantly in microbial hosts. However, while MNEI exhibited improved folding robustness, it remained sensitive to denaturation at elevated temperatures and at neutral pH (Leone et al., 2016). This meant that MNEI could not withstand common industrial processes such as pasteurization and baking thus limiting its practical applications.

Previous studies have shown that monellin can be engineered for greater stability and sweetness, for example through rational mutagenesis guided by electrostatics or stabilizing hydrophobic packing mutations (Leone et al., 2016). These efforts demonstrated feasibility but remained limited in scope as they typically explored only a handful of targeted substitutions. These approaches were unable to capture the full mutational landscape of MNEI thereby leaving open the possibility of more stable and effective designs.

Our project builds directly on this foundation by applying a systematic, physics-based computational workflow. Using the structural data available for native monellin (PDB: 3MON) and the engineered MNEI variant (PDB: 2O9U), we first conducted an alanine scan to identify residues most critical to thermostability. Ten hotspot sites were identified and validated against existing literature. We then performed an exhaustive Rosetta mutation scan, testing every possible single and double substitution across these sites. To quantify the energetic effects of each mutation, we applied the Flex ddG protocol, which estimates the change in folding free energy (ΔΔG) thus allowing us to quantitatively predict how the mutations affected the protein’s stability. This allowed us to systematically rank thousands of variants and pinpoint the most stabilizing mutations.

By combining alanine scanning, Rosetta mutation analysis, and Flex ddG scoring, we created a comprehensive map of monellin’s stability landscape. From this, we selected a shortlist of top-ranked variants for experimental validation.

Methodology

Our methodology integrates a computational pipeline with a recombinant protein expression system to design and express our thermostable variants of monellin. The first stage of our workflow was conducted entirely in silico by leveraging structural data for monellin. Upon identifying 4 potentially thermostable variants we then proceeded onto the laboratory stage.

Firstly, the computationally selected variants were cloned into the expression plasmid pET-28a(+), which carries a T7 promoter for strong inducible expression and an N-terminal His-tag to enable purification (Li et al., 2022). Constructs were first transformed into Escherichia coli DH5-α, which was used exclusively for plasmid amplification. Amplified plasmids were subsequently introduced into E. coli NiCo21 (DE3), the host strain chosen for recombinant protein expression.

The next stage of the workflow will focus on experimental characterisation of the expressed proteins. Thermal stability will be assessed using thermal shift assay, while circular dichroism (CD) spectroscopy will be employed to probe secondary structure and folding quality (Greenfield, 2006). These assays will allow direct comparison between wild-type MNEI and our computationally redesigned variants, providing validation of our physics-based computational predictions. In the future, we intend to collaborate with universities to complete the biophysical and functional characterisation of our variants to estimate melting temperatures and to assess secondary structure. With academic support, in the longer term, we envisage transferring production into alternative hosts such as yeast or other generally recognized as safe (GRAS) organisms to enable larger-scale production and reduce costs. This aligns with the broader sustainability goals of reducing the environmental footprint of sugar production.

Advantages and Applications

A thermostable monellin-inspired sweetener has the potential to deliver broad health, environmental, and industrial benefits. A protein-based sweetener offers a fundamentally different solution. It provides intense sweetness while being completely non-caloric, since it is digested into amino acids without contributing to blood glucose levels. Recent clinical trials with sweet proteins have demonstrated no effects on insulin levels even at sweetness levels equivalent to 75 g of sucrose, underlining their safety and suitability for diabetic and health-conscious populations (Samish et al., 2024b).

Thermostability is the key property that would enable monellin to move from laboratory curiosity to real-world application. A stable protein would withstand pasteurisation, baking, and long-term storage, making it compatible with the workflows of the food industry. Studies with related proteins have shown that designer sweeteners can reduce added sugar in food formulations by 40–70% without compromising taste, while remaining stable in complex matrices such as dairy, beverages, and confectionery (Samish et al., 2024). This opens the possibility of reformulating everyday products with substantially lower caloric content, without the bitterness often associated with synthetic sweeteners.

In addition to health benefits, protein sweeteners offer clear sustainability advantages. Sugarcane and beet production are resource-intensive, demanding extensive land, irrigation, and chemical inputs, while contributing significantly to global greenhouse gas emissions (Jamis, 2025). Recombinant protein production through microbial fermentation provides a lower-carbon, resource-efficient alternative. Unlike plant-derived monellin, which is naturally rare and difficult to source from Dioscoreophyllum cumminsii, our research would enable scalable, reliable supply without ecological impact.

Industrial Application

In terms of industrial applications, a thermostable monellin based sweetener offers a practical alternative to added sucrose. Conventional sugar contributes significantly to caloric load when used for formulation of beverages, dairy and baked goods whereas our sweetener has the potential to deliver high sweetener at micro doses. Because required dosage is small, cost is governed largely by expression titre and simple downstream operations enabling competitive prices for products using monellin. These features position a stable monellin variant as a straightforward and economic alternative for sugar in the various industries without sacrificing taste acceptability.

References

• Bank, R. P. D. (n.d.). RCSB PDB - 2O9U: Monellin (MNEI) at 1.15 resolution. https://www.rcsb.org/structure/2O9U
• Farag, M. A., Rezk, M. M., Elashal, M. H., El-Araby, M., Khalifa, S. a. M., & El-Seedi, H. R. (2022). An updated multifaceted overview of sweet proteins and dipeptides as sugar substitutes; the chemistry, health benefits, gut interactions, and safety. Food Research International, 162, 111853. https://doi.org/10.1016/j.foodres.2022.111853
• Greenfield, N. J. (2006). Using circular dichroism spectra to estimate protein secondary structure. Nature Protocols, 1(6), 2876–2890. https://doi.org/10.1038/nprot.2006.202
• Imperial College London Diabetes & Endocrine Centre. (n.d.). Diabetes trends. https://icldec.ae/diabetes-health-hub/about-diabetes/diabetes-trends/
• Jamis, J. (2025, July 1). Where do GHG emissions in sugarcane come from? Bonsucro. https://bonsucro.com/ghg-emissions-in-sugarcane/
• Kant, R. (2005). Sweet proteins – Potential replacement for artificial low calorie sweeteners. Nutrition Journal, 4(1). https://doi.org/10.1186/1475-2891-4-5
• Leone, S., Pica, A., Merlino, A., Sannino, F., Temussi, P. A., & Picone, D. (2016). Sweeter and stronger: enhancing sweetness and stability of the single chain monellin MNEI through molecular design. Scientific Reports, 6(1). https://doi.org/10.1038/srep34045
• Li, L., Li, H., Tian, Q., Ge, B., Xu, X., Chi, Y., Zhao, H., Liu, Y., Jia, N., Zhou, T., Zhu, Y., & Zhou, Y. (2022). Expression and purification of soluble recombinant β-lactamases using Escherichia coli as expression host and pET-28a as cloning vector. Microbial Cell Factories, 21(1). https://doi.org/10.1186/s12934-022-01972-5
• Liu, Q., Li, L., Yang, L., Liu, T., Cai, C., & Liu, B. (2016). Modification of the sweetness and stability of Sweet-Tasting protein Monellin by gene mutation and protein engineering. BioMed Research International, 2016, 1–7. https://doi.org/10.1155/2016/3647173
• Ogata, C., Hatada, M., Tomlinson, G., Shin, W., & Kim, S. (1987). Crystal structure of the intensely sweet protein monellin. Nature, 328(6132), 739–742. https://doi.org/10.1038/328739a0
• Perez-Aguilar, J. M., Kang, S., Zhang, L., & Zhou, R. (2019). Modeling and structural characterization of the sweet taste receptor heterodimer. ACS Chemical Neuroscience, 10(11), 4579–4592. https://doi.org/10.1021/acschemneuro.9b00438
• Samish, I., Lesmes, U., & Lifshitz, Y. (2024). 1837-LB: Designer protein Sweelin, a novel hypersweet, extremophile protein as a sugar substitute. Diabetes, 73(Supplement_1). https://doi.org/10.2337/db24-1837-lb