Biosen-Driven Growth-Coupled Evolution of Epimerases for Improved D-Allulose Biosynthesis

D-Allulose Needs Better Enzymes

If you formulate with D-allulose, the enzyme is the bottleneck. D-allulose comes from D-fructose, and the conversion depends on ketose 3-epimerases (KEases). Those enzymes are not fast enough, stable enough, or cheap enough for many production goals. This paper offers a way to fix that. It also gives you a clear picture of what the next generation of D-allulose biocatalysts may look like.

The Problem with Current Epimerases

Enzyme engineers face a tedious loop. They mutate an enzyme, test each variant, and hope for an improvement. That works, but it is slow. KEases are particularly tricky because their activity and thermostability often move in opposite directions. A mutant that reacts faster may fall apart at 60 °C. The challenge is finding variants that do both.

This study, published in Advanced Science (DOI: 10.1002/advs.202306478), tackles that problem with a different screening logic. The researchers built a growth-coupled in vivo platform. A D-allulose-dependent biosensor called Biosen sits inside the host cell. When an enzyme variant produces D-allulose, the biosensor triggers cell growth. Better variants produce more D-allulose. More D-allulose means faster growth. The selection pressure does the screening for you.

What the Researchers Did

The team started with ADAE, a D-allulose 3-epimerase. They created a library of ADAE mutants and put those mutants into cells carrying the Biosen system. Then they applied evolutionary pressure. Cells with weak variants lagged behind. Cells with strong variants outgrew the rest.

After screening, they isolated a mutant called M42. This is not just a small tweak. M42 performed significantly better across multiple metrics that matter in an industrial setting.

What the Data Show

M42 improves the two things formulators care about most: speed and stability.

Metric	Reported result
Catalytic activity	6.28-fold increase over the starting enzyme
Thermostability at 60 °C	Half-life 2.5 times longer than the starting enzyme
D-Fructose to D-allulose conversion	32.2% in 3 hours

The catalytic activity gain is large. The thermostability gain is equally important. Many lab-generated mutants fall apart under heat. M42 holds up longer. That makes higher reaction temperatures feasible, and higher temperatures usually mean faster reactions and lower enzyme loading.

The conversion number matters too. A 32.2% yield from D-fructose within 3 hours gives process engineers a realistic starting point. That is not a theoretical trace yield. It is a practical batch result.

What This Means for Manufacturers

M42 is not yet a commercial product. You cannot order it today. But the implications are direct.

Enzyme cost is a major factor in D-allulose pricing. A variant with 6.28-fold higher activity can cut enzyme consumption per batch. A 2.5-fold longer half-life at 60 °C means the enzyme can stay active longer in the reactor. It may also allow multiple reuse cycles. Those numbers translate to lower production costs if the enzyme performs similarly in a scaled process.

For formulators, the bigger story is the screening platform itself. Growth-coupled evolution does not rely on testing thousands of colonies by hand. It lets the cells tell you which enzyme is best. That accelerates development of new biocatalysts, not just for D-allulose but potentially for other rare sugars made by KEases.

Questions You Might Ask

Is M42 ready for food production? Not yet. The paper stops at laboratory characterization. A commercial enzyme needs process development, regulatory review, and scale-up. But M42 gives enzyme suppliers a strong engineering target.

Does this work with other sugars? The platform was developed for KEases. Those enzymes are widespread in rare-sugar production. The same growth-coupled approach could likely be adapted to other epimerases with the right biosensor.

Is the 32.2% conversion commercially relevant? It is a strong lab result. The final yield in a plant will depend on substrate concentration, temperature, pH, enzyme loading, and downstream recovery. Still, hitting 32% in three hours is the kind of performance that makes process developers pay attention.

The Practical Outlook

The enzyme is the limit for cost-effective D-allulose. This paper shows a faster way to remove that limit. The M42 variant is proof that both catalytic speed and thermostability can be improved in a single round of growth-coupled evolution. That combination is exactly what the market needs. Expect more engineered epimerases to follow, and watch for the ones that come through biosensor-driven selection.