Can Microbes Divide the Work of Making Fragrance Ingredients?

Compounds used in fragrances include a family called esters, such as linalyl acetate, which helps shape lavender’s aroma. When microbes are used to make them, is it best to entrust every step, from starting materials to finished product, to a single cell?

In 2026, Dianqi Yang and colleagues used genetically engineered Escherichia coli to investigate how to produce a variety of esters and how to selectively produce larger amounts of a desired compound. The central aim was not merely to increase microbial output, but to design enzyme substrate preferences and intracellular precursor supply together.

An enzyme that joins two starting materials

Broadly speaking, esters are compounds whose structures join an alcohol-derived part to an acid-derived part. The enzymes used in the study, called AATs, combine an alcohol with an acid-derived substrate in a form that reacts readily inside cells.

These enzymes do not necessarily select just one substrate with strict specificity. If they accept multiple alcohols and acid-derived substrates, they can produce diverse compounds. When a factory seeks a single product, however, that same flexibility can generate unwanted compounds as well.

Cells also contain unequal amounts of different substrates. The proportions of products therefore depend not only on the enzyme’s properties but also on what is supplied and in what quantities.

First, exploring the range of compounds that could be made

The researchers examined 14 AATs originating from bacteria, yeasts, and plants. They had E. coli produce these enzymes and supplied various combinations of alcohols and acids, including monoterpenes that can serve as aroma compounds. Monoterpenes are a group of compounds generally built on a ten-carbon skeleton.

Products were confirmed using methods including gas chromatography–mass spectrometry, which separates components and examines their molecular characteristics. This exploration yielded 71 esters across three structural groups. These included compounds with a monoterpene structure not only on the alcohol side but also on the acid side, or on both sides.

The 71 compounds here resulted from experiments in which enzymes combined externally supplied starting materials. This does not mean microbes produced all 71 from simple nutrients at large scale. The sequence was first to broaden the available reactions, then to refine how particular compounds were made.

Adjusting enzymes to select the desired substrates

The researchers next addressed selective production of geranyl butyrate from the alcohol geraniol. Because acetate-side precursors are also abundant inside cells, an unwanted consequence is production of geranyl acetate alongside geranyl butyrate.

Starting with a strawberry-derived enzyme, they computationally examined the three-dimensional structure of the substrate-binding region and replaced one amino acid in the enzyme. They tested modified enzymes both in cells and in isolated reactions, measuring products and reaction rates rather than relying only on computational predictions.

One mutation suppressed the reaction using the acetate-side substrate. Further adjustment of the proportions of supplied butyric acid and geraniol increased the share of geranyl butyrate among the two products. Both changing the enzyme’s shape and adjusting substrate allocation helped.

Improved selectivity does not necessarily mean a faster reaction, however. With the first enzyme mutation, the isolated enzyme’s catalytic efficiency toward the butyrate-side substrate decreased. The researchers therefore also used a related enzyme more likely to remain soluble inside cells, seeking to combine selectivity with productivity.

Related enzymes do not necessarily change by the same mechanism

The comparison showed that modifying the same position can be useful in related enzymes, but that one simple mechanism cannot explain why selectivity changes.

The authors’ computations and reaction measurements suggest that substrate orientation, the ease of adopting a reaction-ready arrangement, and enzyme solubility contribute differently in different enzymes. A single design rule such as “make the cavity larger” is insufficient: actual reactions need testing even for closely related enzymes.

Dividing the work between two strains rather than one cell

Following the exploration using externally supplied substrates, the researchers connected the production of precursors from nutrients such as glycerol with their conversion into esters. Under some conditions, they also incorporated a pathway that uses ethanol to help supply precursors.

They used coculture, growing together two E. coli strains with different roles. One produces alcohols such as linalool or geraniol; the other converts them into esters. Rather than combining different microbial species, they engineered the same bacterium for different jobs.

For linalyl acetate production in test tubes, the highest value with all steps in one strain was about 0.058 g/L. Coculture reached 3.22 g/L after adjustment of the strain mixing ratio and precursor-supply pathways. This compares conditions designed and optimized within the study; it does not show that the same fold increase would occur in every production system.

The authors suggest that division of labor can ease both the burden of expressing many genes in one cell and competition between pathways for the same starting material. Another advantage is the ability to improve and recombine the alcohol-producing and finishing strains separately.

What was achieved in a small fermenter?

The researchers also tested coculture in a fermenter with a capacity of 1 L. The starting culture medium was 400 mL, with an organic-solvent layer added to extract products and glycerol supplied during the process. A 1 L vessel capacity does not mean that the cultures were grown from the outset in 1 L of medium alone.

In these experiments, linalyl acetate reached 11.50 g/L after 96 hours, while geranyl butyrate reached 3.16 g/L after 83 hours under separate production conditions. Each condition was examined in three biological replicates. These are the production titers reported in the paper, not yields of purified fragrance ingredients ready for sale.

Geranyl butyrate accounted for 98.97% of the combined amount of geranyl butyrate and geranyl acetate. This figure describes selectivity between two esters. It does not mean that the entire extract, including organic solvent and other components, was 98.97% pure.

For linalyl acetate, analysis also confirmed selective production of one of two mirror-image molecular forms. The result illustrates the value of biological reactions not only for making a named compound but also for controlling its molecular form.

Coculture production was demonstrated for seven compounds in total, including linalyl acetate and geranyl butyrate. However, titers varied widely among the other compounds; using the same approach does not make every fragrance ingredient equally efficient to produce.

A technology for selective ingredient production, not scent replication

Linalyl acetate is an important constituent of lavender essential oil, but the study did not reproduce the entire oil using microbes. Nor did these production experiments assess perceived odor similarity, preference, or product safety.

The 1 L results demonstrate a step up from test-tube production. They do not establish stable operation at commercial scale, purification costs, or environmental impacts relative to conventional methods. Further validation is needed before practical implementation.

Even so, the sequence of examining the range of substrates accepted by enzymes, adjusting enzymes and supply ratios for a desired reaction, and then dividing the production process offers a design path that could extend to other esters. The authors see potential not only in fragrance and food ingredients, but also in discovering and improving enzymes that produce various useful plant compounds.

The research received public funding, including support from the National Natural Science Foundation of China. Three authors declared that they are inventors on patent applications related to the microbial production technology.

Summary

The study treated expanding the range of esters that microbes can make and selectively producing one desired compound as connected challenges. After obtaining 71 compounds in an exploration with externally supplied substrates, the researchers adjusted parts of enzymes and substrate ratios, then demonstrated production of seven compounds through coculture of two strains.

The important point is that finding one good enzyme is not the end of the task. Enzyme selectivity, the materials supplied by cells, and the burdens and competition among process steps need to be designed together. The authors show how this approach can provide a foundation for more flexible, selective production of fragrance ingredients.

References

Yang, D., Liang, H., Li, X., et al. (2026). A microbial platform for selective biosynthesis of monoterpene esters. Nature Communications, 17, 4851. DOI: 10.1038/s41467-026-71681-9. Full text. © The Author(s) 2026. The original article is published under CC BY-NC-ND 4.0.