How Is Reb M Made? Leaf Extraction vs Enzymatic Bioconversion vs Fermentation

How Is Reb M Made? Three Production Routes

The distinction is more precise than describing one route as “natural” and another as “fermented.” The manufacturing route is defined by where the process begins and how the target glycoside is formed. EU regulation increasingly reflects those differences explicitly. EUR-Lex

Route 1: Direct Extraction and Purification From Stevia Leaves

Reb M occurs naturally in Stevia rebaudiana. It is therefore technically possible to isolate the molecule directly from the leaf without converting another steviol glycoside first.

A frequently cited 2014 study demonstrated this at high purity. Researchers started with 2 kg of stevia leaves, extracted them with 40 L of water at 40°C for two hours, then used calcium-oxide clarification, ion-exchange treatment, resin fractionation, aqueous-ethanol elution, activated carbon treatment and repeated recrystallization.

The process ultimately yielded about 1.1 g of Reb M at greater than 98% HPLC purity. PMC

The published process can be simplified as:

Stevia leaves → water extraction → clarification → ion exchange → resin fractionation → ethanol elution → decolorization → recrystallization → high-purity Reb M

That result corrects an important misconception.

Direct leaf extraction is not incapable of producing high-purity Reb M. The practical problem is the amount of separation required before that purity is reached.

Purity Is Not the Same as Natural Abundance or Recovery

Reb M is a minor glycoside in conventional stevia leaf material. A 2024 review reports approximately 0.4–0.5% Reb M in dry stevia leaves, while also emphasizing the commercial difficulty of extracting and purifying such a low-abundance component. Institute of Food Technologists

Three numbers therefore need to be kept separate:

A finished fraction can assay above 98% Reb M even though Reb M represented only a small fraction of the original leaf.

High final purity therefore does not prove high natural abundance, high recovery, or favorable production economics.

Likewise, the 2014 result of approximately 1.1 g from 2 kg of leaf should not simply be converted into a “Reb M recovery rate.” The publication reports the isolated amount but does not provide the complete starting-leaf Reb M mass balance required for that calculation. PMC

For commercial sourcing, the real issue is therefore not whether high-purity leaf-isolated Reb M is technically possible. It is how much raw material and fractionation are required to isolate a naturally minor glycoside at scale.

Route 2: Enzymatic Bioconversion

Enzymatic production solves the low-abundance problem differently.

Instead of attempting to recover only the Reb M already present in the leaf, the process starts with existing steviol glycosides and uses glycosyltransferase enzymes to add the glucose units needed to form more highly glycosylated compounds such as Reb D and Reb M.

A simplified route is:

Leaf-derived steviol glycosides → precursor glycoside → selective enzymatic glycosylation → Reb M → enzyme removal → purification → crystallization

One EU-defined example, E 960c(ii), starts with highly purified Reb A extract obtained from Stevia rebaudiana. UDP-glucosyltransferase and sucrose synthase enzymes facilitate conversion, after which the enzymes are removed and the Reb M is concentrated and recrystallized. The defined finished material contains not less than 95% Reb M. EUR-Lex

Why Reb A Does Not Simply “Turn Into Reb M” in One Generic Step

Commercial descriptions sometimes reduce the process to:

Reb A → Reb M

That is useful shorthand, but it hides the difficult part of the process.

Reb M formation requires specific glycosylation reactions. Glucose has to be transferred to the appropriate positions while competing reactions and unwanted glycoside formation are controlled.

One practical pathway is:

Reb A → Reb D → Reb M

A 2026 Biotechnology Journal study used a stepwise enzyme cascade based on exactly this concept. One engineered glycosyltransferase converted Reb A to Reb D, and another converted Reb D to Reb M. At 3 L scale, the system converted 30 mmol of Reb A into 24.1 mmol of Reb M, giving a reported 80.34% overall yield under the study conditions. PubMed

The more important sourcing lesson is not the laboratory percentage itself:

Processes can differ in starting glycosides, enzymes, reaction sequence and resulting secondary glycoside profile.

Why UGT and Sucrose Synthase Are Often Used Together

Another common point of confusion is why glycosyltransferase and sucrose synthase repeatedly appear together in Reb M production literature.

The glycosyltransferase transfers glucose onto the steviol glycoside, but it requires an activated sugar donor such as UDP-glucose.

Continuously supplying purified UDP-glucose would be unattractive for a larger-scale process. Sucrose synthase can help regenerate the activated sugar donor using sucrose.

In practical terms:

UGT performs the target glycosylation. SuSy helps keep the glucose-donor system supplied.

A 2026 Food Chemistry study combined enzyme engineering, cascade biocatalysis and substrate feeding to produce Reb M, reporting a final titer of 28.38 g/L and 91.6% conversion in its optimized experimental system. PubMed

Those numbers should not be interpreted as a standard commercial manufacturing specification. They demonstrate how current process development is addressing enzyme efficiency and donor regeneration.

Recent Research Is Moving Beyond One Purified Precursor

A further development is the effort to reduce dependence on one highly purified starting glycoside.

Conventional bioconversion can require significant upstream separation to produce the desired precursor before conversion even begins.

A 2025 Green Chemistry study instead developed a two-enzyme cascade capable of using a naturally derived steviol-glycoside mixture. The optimized system produced 38.8 g/L Reb M with an 85.5% molar yield. 皇家化学学会出版物

The important development is the shift from:

purify one precursor → convert it → purify Reb M

toward:

use more of the original glycoside mixture → engineer the enzyme system to direct the mixture toward Reb M

That approach may reduce precursor-isolation burden and improve utilization of the original stevia glycoside pool.

However, this remains an important distinction between published process development and actual commercial production. A successful research system does not prove that every Reb M manufacturer currently uses mixed-glycoside feedstock.

Route 3: Fermentation-Derived Reb M

Microbial fermentation begins further upstream.

Instead of starting with Reb A, Reb D or another extracted steviol glycoside, the process supplies a microorganism with a carbon source and uses an engineered biosynthetic pathway to produce steviol glycosides during fermentation.

The EU now has a specific regulatory example:

E 960b(i) — Rebaudioside M from fermentation produced by *****Yarrowia lipolytica*****.

Under the EU definition, a genetically modified Yarrowia lipolytica strain ferments a simple sugar source. Biomass is subsequently removed, the steviol glycosides are concentrated, and downstream processing includes ion-exchange chromatography and crystallization from ethanol. EUR-Lex

The route can therefore be summarized as:

Simple sugar → microbial fermentation → steviol glycoside biosynthesis → biomass removal → purification → crystallization

This is fundamentally different from enzymatic bioconversion.

In bioconversion, a steviol glycoside already exists before the key conversion step begins.

In fermentation, the production microorganism carries the pathway responsible for biosynthesizing the steviol glycosides.

There is also an important compositional distinction. Under the current E 960b(i) definition, Reb M is the main component, but the final product must contain not less than 95% combined Reb M, Reb D, Reb A and Reb B. EUR-Lex

That is not the same statement as:

Reb M ≥95%.

If the Enzymes Are Made by Fermentation, Is the Reb M “Fermented”?

Not necessarily.

Fermentation can appear at more than one point in a manufacturing chain.

A microorganism can be fermented simply to manufacture an enzyme. That enzyme can then be separated and used in another process to convert leaf-derived steviol glycosides into Reb M.

The Reb M production step is still enzymatic bioconversion.

E 960c(ii) is a useful example. Defined genetically modified E. coli strains are used to produce the glycosyltransferase and sucrose synthase enzymes, but the commercial Reb M is subsequently formed through enzymatic conversion of highly purified, leaf-derived Reb A. EUR-Lex

By contrast, in E 960b(i), the Yarrowia lipolytica production organism itself biosynthesizes the steviol glycosides during fermentation from a simple sugar feedstock. EUR-Lex

This distinction is particularly useful when supplier descriptions use terms such as “fermentation technology” without explaining where fermentation actually occurs in the process.

E 960a, E 960b and E 960c: The EU Manufacturing Distinction

For buyers supplying the European market, including Poland, the E-number can reveal important information about the production category.

In 2021, the EU renamed the previous E 960 designation as E 960a — Steviol glycosides from Stevia and introduced E 960c — Enzymatically produced steviol glycosides. EUR-Lex

Fermentation-derived Reb M was subsequently given a specific definition under E 960b(i). EUR-Lex

All of these share a single ADI of 4 mg per kg body weight per day, expressed as steviol equivalents. The ADI was reconfirmed by the EFSA FAF Panel in September 2024 after a full re-evaluation of exposure data across all E 960 subcategories.

E 960a Does Not Mean “Reb M 95%”

This distinction is especially useful for procurement.

E 960a tells the buyer the regulatory family. It does not tell the buyer that Reb M is the dominant glycoside.

The designation therefore should not be treated as equivalent to:

If a specification only identifies the ingredient as E 960a, the buyer still needs to review the actual steviol-glycoside profile and assay.

E 960c Is Also More Than One Process

“Enzymatic Reb M” is not one single EU manufacturing definition.

E 960c(i) covers Reb M produced by enzyme modification of purified steviol glycosides from stevia, while E 960c(ii) specifically describes Reb M produced via enzymatic conversion of highly purified Reb A extracts. EUR-Lex

This is why two suppliers can both describe a material as:

while still using different defined starting materials or process specifications.

2026 Update: Enzymatic Production Is Still Evolving

The manufacturing framework continues to develop.

On 17 September 2026, EFSA published a scientific opinion evaluating another modified enzymatic process for producing Reb M from purified steviol-glycoside leaf extracts.

The assessed system uses UDP-glucosyltransferase, UTP-glucose-1-phosphate uridylyltransferase and sucrose synthase enzymes produced by a newly developed genetically modified E. coli K-12 DSM 35119 strain. EFSA concluded that there was no safety concern for the assessed Reb M under the described conditions. EFSA

EFSA also considered that the proposed process differed sufficiently at the manufacturing-definition level that a new specification entry would be more appropriate than simply treating it as identical to the existing entries. EFSA

For buyers, the important point is not the microorganism strain number. It is that:

07 Natural positioning

Can Enzymatic Reb M Be Called "Natural"?

This is the question behind the question. When a European brand asks about production routes, the regulatory classification is usually the stated concern. The unstated concern is whether the ingredient can sit inside a "natural sweetener" or "clean label" positioning without creating a problem later.

The short answer is that there is no single authority that resolves this, because the EU has no legal definition of "natural" for food ingredients in the way it defines "organic". What exists instead is a patchwork of guidance, case law and industry convention that varies by claim type.

Leaf extraction (E960a)

The simplest case. The product is a water or ethanol extract of a plant leaf. It can be described as a stevia leaf extract without qualification. No enzyme, no fermentation organism, no GMO-adjacent questions. If your marketing requires "from the stevia leaf" to be literally true at every stage of production, this is the only route that delivers it.

Enzymatic conversion (E960c)

The starting material is still a stevia leaf extract — specifically, high-purity Reb A obtained from the leaf. The conversion enzymes are processing aids that are removed before the product is finished. The end product is a molecule that occurs naturally in the stevia leaf.

Whether that qualifies as "natural" depends on who is asking. Under EU Regulation 1334/2008 on flavourings, an enzymatic process applied to a material of plant origin can qualify as a "natural" production method for flavouring purposes. But steviol glycosides are regulated as food additives under Regulation 1333/2008, not as flavourings, and the flavouring regulation's "natural" definition does not automatically transfer. In practice, many brands describe enzymatic Reb M as "from stevia" or "stevia-derived" and avoid the word "natural" in the regulatory sense while using it in the marketing sense. Whether that distinction holds depends on the enforcement environment in the destination market.

Fermentation (E960b)

The hardest case. The starting material is glucose. No stevia leaf is involved at any stage. The product is biosynthesised by a yeast. The molecule at the end is identical to the one in the leaf, but the process has no physical connection to the plant. Describing fermentation-derived Reb M as "from stevia" would be factually incorrect. Some brands position it as "nature-identical" or "biofermented", but those terms do not have regulatory definitions either.

08 GMO status

GMO-Derived Enzymes, Non-GMO Product: How the EU Draws the Line

The enzymes used in the E960c routes — UDP-glucosyltransferase and sucrose synthase — are produced by genetically modified micro-organisms. In the E960c(i) route the host is K. phaffii. In the E960c(ii) route it is typically E. coli. The enzymes are used as processing aids and are removed from the final product by filtration and heat treatment.

Under EU Regulations 1829/2003 and 1830/2003, the GMO labelling obligation applies to products that contain or consist of GMOs. A product manufactured with the aid of a GMO-derived enzyme, where the final product does not contain viable cells or detectable recombinant DNA, is not subject to GMO labelling. The EU specifications in Regulation 231/2012 explicitly require that viable cells and DNA of the production organisms are not detectable in the finished additive.

This is where the residual DNA and residual protein tests on the COA become more than a quality control line item. They are the documentary evidence that the product has crossed the line from "made with a GMO tool" to "does not contain GMO material". If a supplier's COA does not report these tests, the buyer has no way to confirm that the product meets the EU specification on this point, and no documentation to show an auditor.

09 Organic certification

Organic Certification and Enzymatic Reb M: The Tension and How It Gets Resolved

EU organic legislation under Regulation 2018/848 prohibits the use of GMOs and products produced from or by GMOs in organic production. At first reading that appears to rule out enzymatic Reb M entirely, because the enzymes come from genetically modified yeasts or bacteria.

The resolution sits in the distinction between a processing aid and an ingredient. The enzymes are classified as processing aids. They are used during manufacture and removed before the final product. They are not present in the finished material and are not declared on the label. Under this interpretation, the final product is not "produced from" a GMO in the regulatory sense, because the GMO-derived substance does not persist in it.

Organic certification bodies have evaluated this distinction and issued certificates for enzymatically converted steviol glycosides, including Reb D and Reb M grades. The assessment is made by the certifier on the specific process, the specific documentation and the specific test results for that operator. It is not a blanket exemption for all enzymatic routes.

If your brand already carries an organic claim and you are considering switching from leaf-extracted Reb A to enzymatic Reb M, the question to ask is not "is enzymatic Reb M organic?" in the abstract. The question is whether your specific certifier accepts the specific processing aid documentation and residual testing from your specific supplier. That conversation should happen before the purchase order, not after the product has shipped.

10 Labelling and ADI

What Goes on the Label and How the ADI Limits Your Formulation

The E number is mandatory and route-specific

In the EU, the E number declared on a food label must correspond to the actual production route of the ingredient. A Reb M produced by enzymatic conversion from Reb A using E. coli-derived enzymes is E960c(ii). The same molecule produced by leaf extraction is E960a. The same molecule produced by fermentation is E960b(i). They are not interchangeable on a label, even though the molecule is identical, because the E number identifies the production method and its associated specification.

If your supplier cannot tell you which E960 subcategory applies to their product, your regulatory team cannot complete the label. That single question — "which E number?" — should be the first line of any enquiry for Reb M destined for the European market.

ADI and the steviol equivalence conversion

The shared ADI of 4 mg/kg bw/day is expressed as steviol equivalents, not as steviol glycosides. The conversion from glycoside weight to steviol equivalent weight uses a molecular weight ratio that differs by glycoside. For Reb M the conversion factor is approximately 0.25 — meaning 1 mg of Reb M corresponds to about 0.25 mg of steviol equivalents. For Reb A the factor is approximately 0.33.

This matters for formulation. A product sweetened with Reb M can carry a higher glycoside load within the same ADI budget than one sweetened with Reb A, because each milligram of Reb M contributes less steviol equivalent. That is a real formulation advantage, and it is one that the cost comparison between routes should account for.

11 Other markets

FDA GRAS, JECFA and the Markets That Follow Neither EU nor US Rules

FDA GRAS: a different system

The US does not classify steviol glycosides by production route. There are no E960 subcategories. Instead, the system runs on GRAS (Generally Recognised as Safe) notifications, filed individually by each manufacturer for their own product.

A GRAS notification is not an FDA approval. It is a manufacturer's self-determination that their product is safe for its intended use, submitted to FDA for review. If FDA has "no questions", it means the agency did not find a basis to disagree. That is a weaker statement than "FDA approved this ingredient".

Between 2024 and mid-2026, FDA issued "no questions" letters for enzymatic Reb M from at least six different companies, including GRN 1184 (SIB), GRN 1203 (Layn, for Reb M2), GRN 1206 (Adorvia), GRN 1260 (PureCircle / Ingredion), GRN 1294 (Arzeda / Manus Bio) and GRN 1307 (Shanghai Shenghe). Each notification covers only the notifier's own product and process.

JECFA and Codex

JECFA sets the specifications that Codex Alimentarius member countries adopt. The current JECFA specification requires total steviol glycosides of not less than 95 percent on a dried basis. JECFA does not subdivide by production route. There is no JECFA equivalent of E960c(i) versus E960c(ii).

For markets that adopt Codex standards without additional national requirements — and that includes a significant share of the Middle East, Southeast Asia and parts of Latin America — the regulatory question is simpler: does the product meet the 95 percent total SG specification? The production route is not part of the regulatory identity.

If your customer base spans multiple jurisdictions, the safest approach is to carry a product that meets the JECFA specification and the EU specification simultaneously, with the E number documentation ready for EU and the GRAS documentation ready for the US. That single SKU then covers most of the global market without reformulation.

12 COA by route

COA by Production Route: What to Check and What Most COAs Leave Out

The items that belong on a Reb M COA depend on which route produced it. A generic "Reb M 95%" certificate that does not state the production route is not a usable purchasing document for any market that distinguishes between them.

The two lines that matter most for EU-facing procurement are the E number subcategory and the residual DNA test. If either is missing, the COA is not fit for purpose. The first tells your regulatory team what to put on the label. The second is the evidence that the product meets the EU specification for absence of GMO material.

For a detailed breakdown of how minor glycoside composition affects taste — particularly the Reb C and dulcoside A thresholds that predict bitterness — see our stevia aftertaste guide. Our stevia extract product page lists the specific grades we supply across the Reb A, Reb D and Reb M range, with the production route and assay basis stated for each.

Leaf Extraction vs Bioconversion vs Fermentation: What Actually Changes?

This is why the phrase “Reb M 95%” alone is not enough to establish what a buyer is actually purchasing.

Same Reb M Molecule Does Not Mean the Same Commercial Ingredient

Chemically, Rebaudioside M remains the same defined molecule.

A commercial ingredient, however, is more than the identity of its main molecule.

Reb M materials can differ in secondary glycoside profile, assay definition, manufacturing specification, process-related controls and regulatory documentation.

The EU provides an especially clear example.

Under E 960c(ii), the finished product is defined as containing not less than 95% Reb M. EUR-Lex

Under E 960b(i), Reb M is the main component, but Reb M + Reb D + Reb A + Reb B together must total at least 95%. EUR-Lex

Both specifications therefore contain the number 95%, but the number describes different things.

For procurement, one question resolves much of the ambiguity:

Why a Simple GMO / Non-GMO Question Can Be Incomplete

The role of a genetically modified production organism can also differ by process.

In an enzymatic route such as E 960c(ii), genetically modified microorganisms are used to produce the enzymes, while the Reb M is subsequently formed through a separate conversion of stevia-derived glycosides. The EU specification also requires that viable production cells and their DNA are not detected in the finished additive. EFSA

In E 960b(i), the engineered microorganism is instead the organism responsible for biosynthesizing the steviol glycosides during fermentation. EUR-Lex

For a technical buyer, a more informative question is therefore:

Any final labeling or certification claim should then be checked against the requirements of the destination market and the relevant certification scheme rather than inferred solely from the production-route name.

How to Verify a Supplier's Reb M Production Route

A supplier's production description should be supported by documentation, not just marketing terminology.

This is much more useful than relying on a quotation that simply says:

For example, “leaf-derived Reb M” could mean Reb M physically isolated from stevia leaves, or it could mean that leaf-derived steviol glycosides were later enzymatically converted into Reb M.

Likewise, “fermented Reb M” should distinguish a process in which the microorganism actually biosynthesizes the steviol glycosides from one in which fermentation was used only to manufacture the enzymes.

And “Reb M 95%” should always be accompanied by a definition of what the 95% represents.

The process statement, specification and batch COA should tell the same story.

Which Production Route Fits Your Project?

There is no single Reb M route that is automatically appropriate for every product.

The correct choice depends on the requirements of the finished formulation.

An EU project may need to confirm whether the intended material falls under E 960a, E 960b(i) or E 960c, while another project may place greater emphasis on individual Reb M assay, secondary glycoside profile, documentation, supply consistency or formulation performance.

The useful decision is therefore not:

It is:

For buyers comparing grades, that question should be answered before two Reb M quotations are treated as technically equivalent.

Frequently Asked Questions

Conclusion

Reb M can reach the market through three fundamentally different manufacturing routes.

Direct extraction isolates Reb M that Stevia rebaudiana has already produced.

Enzymatic bioconversion starts with existing steviol glycosides and uses selective glycosylation to form more Reb M.

Microbial fermentation uses a production organism to biosynthesize steviol glycosides from a carbon source.

The target molecule may be the same, but that does not make every commercial Reb M ingredient equivalent. Starting material, secondary glycoside profile, assay definition, regulatory classification and supporting documentation can differ.

For procurement, the most reliable comparison therefore goes beyond the product name. Confirm the starting material, production route, individual Reb M assay, glycoside profile, applicable E 960 specification and supporting documentation before treating two Reb M offers as equivalent.

References

Prakash I, Markosyan A, Bunders C. Development of Next Generation Stevia Sweetener: Rebaudioside M. Foods. 2014;3(1):162–175. DOI: 10.3390/foods3010162. PMC

Okonkwo CE, Adeyanju AA, Onyeaka H, et al. A review on rebaudioside M: The next generation steviol glycoside and noncaloric sweetener. Journal of Food Science. 2024;89:6946–6965. DOI: 10.1111/1750-3841.17401. Institute of Food Technologists

Luo Z, Guo X, Deng Z, et al. Designing an artificially concise biocatalytic cascade for the manufacture of rebaudioside M from a naturally derived steviol glycoside mixture. Green Chemistry. 2025;27:15694–15705. DOI: 10.1039/D5GC03693C. 皇家化学学会出版物

Luo X, Wang Y, Zhou M, et al. Natural sugar substitute Rebaudioside M: Efficient and cost-effective biosynthesis via host engineering, cascade biocatalysis, and substrate feeding. Food Chemistry. 2026;509:148545. DOI: 10.1016/j.foodchem.2026.148545. PubMed

Huang Y, Ge Y, Zhang M, Zhou W. Multi-Enzyme Cascade Reaction of Crude Enzyme Strategy for the Economical and Efficient Bioconversion of Rebaudioside A to Rebaudioside M. Biotechnology Journal. 2026;21(6):e70270. DOI: 10.1002/biot.70270. PubMed

Commission Regulation (EU) 2021/1156. E 960a Steviol glycosides from Stevia and E 960c enzymatically produced steviol glycosides. EUR-Lex

Commission Regulation (EU) 2022/1922. Specifications for enzymatically produced rebaudiosides, including E 960c(ii). EUR-Lex

Commission Regulation (EU) 2025/652. E 960b(i), Rebaudioside M from fermentation produced by Yarrowia lipolytica. EUR-Lex

EFSA FAF Panel. Safety evaluation of the food additive rebaudioside M produced via enzymatic conversion of purified steviol glycosides leaf extracts. EFSA Journal. 2026;24(9):e10232.