Production of D-Allulose from D-Fructose Using Permeabilized Recombinant Cells of Corynebacterium glutamicum

The host matters as much as the enzyme

D-allulose sells because it tastes like sugar without the glucose spike. Making it at scale, however, forces a choice. Most enzymes convert D-fructose into D-allulose well in a test tube. When you build a real process, the production host matters just as much as the enzyme.

This PLOS ONE study tackles that choice head-on. The authors built recombinant Corynebacterium glutamicum cells that express a D-allulose 3-epimerase from Flavonifractor plautii. Then they permeabilized those cells and used them as the catalyst.

Why E. coli creates a bottleneck

D-allulose 3-epimerase flips fructose into allulose. Researchers have expressed that enzyme in E. coli for years. E. coli gives high activity and easy genetic tools. But ingredient buyers should care about one problem: E. coli is not considered suitable for food additive production.

C. glutamicum is different. Regulators recognize it as GRAS. It already produces amino acids and other food ingredients at industrial scale. Choosing this host removes a compliance hurdle before catalysis even starts.

The enzyme source also matters. It comes from Flavonifractor plautii, a gut bacterium. The team cloned it into C. glutamicum, then treated the cells to make them permeable. Permeabilized cells let fructose enter and allulose leave without full enzyme purification.

What the study did

The researchers did not stop at expression. They pushed the process toward industrial conditions.

They started with a high substrate load: 750 g/L D-fructose. That concentration mimics what a manufacturer would need for viable economics. The catalyst load was 10 g/L permeabilized cells. The reaction ran for just 40 minutes.

The system delivered 235 g/L D-allulose. Conversion reached 31%, and volumetric productivity hit 353 g/L/h.

Metric	Result	Why it matters
Catalyst loading	10 g/L permeabilized cells	Avoids enzyme purification step
Substrate	750 g/L D-fructose	Industrial-level sugar load
Reaction time	40 minutes	Fast reactor turnover
Product titer	235 g/L D-allulose	High concentration eases downstream recovery
Conversion	31%	Useful benchmark for process design
Volumetric productivity	353 g/L/h	Strong output per reactor volume per hour

That productivity number deserves attention. A rate of 353 grams per liter per hour means a modest reactor can generate meaningful output in under an hour.

What it means for manufacturers

Formulators often ask whether an allulose supplier can actually produce the sweetener without regulatory friction. This study points to a route that starts from a GRAS organism and uses a straightforward whole-cell catalyst.

Permeabilized cells cut steps. You avoid lysing cells and purifying the enzyme. You still get concentrated product. The shorter reaction time also reduces contamination risk and reactor fouling.

Buyers should ask two questions of any commercial allulose source. First, what production host do you use? Second, what form does the catalyst take? A GRAS whole-cell platform fits food production far better than an E. coli-based process.

Frequently asked questions

Does using C. glutamicum make the final allulose automatically food-grade?

No. The host is GRAS, but the final product still needs to meet purity, allergen, and residue standards. This study removes the host risk, not the entire regulatory burden.

Why use permeabilized cells instead of purified enzyme?

Permeabilized cells keep the enzyme inside the cell but let substrate and product cross the membrane. That saves time and cost in downstream processing.

Can I scale these numbers into a commercial reactor?

Carefully. The data came from a controlled study. Scale-up must account for mixing, temperature control, and enzyme stability over repeated use.

The process data you need

Allulose production has moved beyond the enzyme-discovery stage. This work shows a practical way to biocatalyze fructose into allulose in a GRAS host, in 40 minutes, at an industrial substrate concentration. That is exactly the kind of process data ingredient buyers need.