Trilobatin as a Bitterness Blocker: What the Evidence Actually Shows
Search “natural bitterness blocker” and you will find a dozen supplier pages telling you that trilobatin removes the bitter aftertaste of stevia. None of them cite a sensory study. The reason is that the study they are describing does not exist.
What does exist is more useful than the marketing claim, and considerably more specific. A flavour house ran trilobatin against several bitter systems and published the results in a patent. It worked in some and failed in one. It worked at 5 ppm and started damaging the product at 20 ppm. And it did not remove bitterness so much as remove a particular part of it, leaving another part intact.
That is a usable ingredient profile. “Removes bitterness” is not.
Before going further, one thing should be stated plainly, because it shapes how much weight everything below can carry:
Most of the specific sensory data in this article comes from a single patent family, tested by panels of five to eight people.
That is screening-scale, not conclusion-scale. It is the best published data available on this molecule, it comes from a serious flavour laboratory, and it is still one source with small panels. Treat it as a well-informed starting point for your own trial, not as a settled result.
Each claim below is labelled by evidence type: patent sensory example, peer-reviewed receptor study, or inference. Those are not the same thing, and the distinction matters when you are deciding whether to fund a trial.
01 The compound
What Trilobatin Is
Trilobatin is a dihydrochalcone glycoside, CAS 4192-90-9. It also travels under p-phlorizin, phloretin 4′-glucoside and prunin dihydrochalcone. Worth knowing, because a specification sheet may not use the name you searched for.
Its commercial source is the leaf of Chinese sweet tea, Lithocarpus polystachyus, also published as L. litseifolius, a broadleaf evergreen from southern China where the leaves have been brewed as a sweet tea for centuries. It also occurs in the apple species Malus trilobata, which is where the name came from. It was first synthesised in 1942, and a 1963 US patent described its use as a sweetening agent.
Sweetness is usually quoted at around 300 times sucrose, alongside phlorizin, the other major dihydrochalcone in the same leaf. Both are described in the plant science literature as stable sweeteners and bitterness blockers. That dual character is what makes the molecule interesting, and also why so much writing about it is confused. Commercial grades and mogroside-free specifications are set out on our sweet tea extract page.
02 Definitions
Blocker, Masker, Modulator: Three Different Things
These get used interchangeably in supplier copy. They are different mechanisms and they fail in different ways.
Masking covers bitterness with something else. Add enough sweetness, vanilla or acid and the bitter note is still there but harder to find. It works, and it costs you: the product now tastes of the masking agent.
Blocking reduces the perception of bitterness itself, ideally without touching the other basic tastes.
Sweetness modulation changes the shape of the sweetness curve — onset, peak, linger, mouthfeel — without necessarily doing anything about bitterness.
The published evidence on trilobatin spans all three, which is precisely why claims about it are so slippery. At 100 ppm it is a weak sweetener. At 5 ppm it appears to block. And in a separate body of patent work its chemical class is used to modulate sweetness curves. Same molecule, three jobs, three dose ranges.
A defensible summary:
Trilobatin behaves as a concentration-dependent taste modifier with demonstrated bitter-reduction activity in specific systems, not as a universal bitterness blocker.
03 Temporal selectivity
It Removes the Middle, Not the Front
This is the most useful finding in the literature and it appears in no supplier article we are aware of.
Givaudan holds a patent family — US 9,844,224, published as US 2011/0086138, priority 2008 — covering trilobatin and a related dihydrochalcone as bitter blockers for alkaloid bitterness. The examples are small-panel sensory comparisons, sample against control.
In unsweetened cocoa at 10% in water, 5 ppm trilobatin produced a sample panellists found less bitter while the cocoa aroma stayed intact. The detail is in how the bitterness changed: alkaloid bitterness in the middle and towards the end of the temporal profile was markedly reduced, with no bitter aftertaste, while the upfront warm, woody, rounded dark-chocolate bitterness remained.
The energy drink example repeats the pattern. Against two commercial caffeinated drinks, 5 ppm made every panellist report less bitterness, with bitterness in the middle of the temporal pattern very effectively blocked. A slight bitter aftertaste remained, and the patent says plainly that this residue needs a partner ingredient working on the late phase.
So trilobatin is not a bitterness eraser. On this evidence it acts on a specific window of the temporal bitterness curve.
For most categories that is fine or irrelevant. For coffee, cocoa, dark chocolate and anything where some bitterness is the point, it is close to ideal — you want the opening bitterness that signals the product, and you want the long dry finish gone. An ingredient that removed bitterness uniformly would wreck those products.
It also tells you how to build the scoresheet. If your panel records a single bitterness score, you will miss the effect or badly under-read it. Score onset, mid-palate and finish separately, or the trial is wasted.
Figure 1. Schematic illustration of the temporal effect described in the Givaudan patent examples: alkaloid bitterness reduced in the middle and towards the end of the profile, while the upfront bitterness remains. Not reproduced from published curve data — the patent describes this effect in words and does not publish time–intensity measurements.
Patent sensory examples, five to eight panellists, applicant-generated, assignee Givaudan. The cocoa, energy drink, coffee, tea, caffeine and tannin results throughout this article all come from this same patent and the same panel series. They are not six independent confirmations.
04 Specificity
Not All Bitterness Is the Same
Here is where most writing about trilobatin falls apart, and where the same patent provides an answer.
Against caffeine at 200 ppm in water, trilobatin reduced bitterness from 5 ppm up to 20 ppm. At 1.25 ppm there was no effect. That gives a genuine lower bound rather than a marketing floor.
Against tannic acid at 500 ppm in water, trilobatin at 1.25 to 20 ppm did nothing. Samples and control were found similarly bitter. The patent states the conclusion directly: trilobatin does not mask tannin-derived bitterness.
Against black tea, 10 ppm produced an equivocal result. Most panellists found the sample slightly less bitter. Some detected no difference or were unsure.
| Bitter Source | Result | What It Means |
|---|---|---|
| Caffeine, 200 ppm in water | Reduced at 5–20 ppm, nothing at 1.25 ppm | Clean positive with a defined floor |
| Instant coffee | Less bitter at 10 ppm, no bitter aftertaste | Positive, but see the ceiling in the next section |
| Cocoa, 10% unsweetened | Less bitter at 5 ppm, aroma retained | Positive, with temporal selectivity |
| Commercial energy drinks | Less bitter at 5 ppm, slight aftertaste remains | Positive, may need a late-phase partner |
| Brewed black tea | Most found slightly less bitter, some found none | Weak and inconsistent |
| Tannic acid, 500 ppm in water | No reduction at 1.25–20 ppm | Negative |
Bitterness is not one sensation. Humans carry roughly 25 bitter taste receptors, and a compound that interferes with an alkaloid has no obligation to do anything about a polyphenol. Caffeine is an alkaloid. Tannins are polyphenols, and they contribute astringency as well as bitterness — a partly tactile sensation a taste-receptor intervention may not reach at all.
Two cautions on reading that negative result. It is one model system: tannic acid in water, at one concentration, over one dose range. It does not establish that trilobatin fails against every polyphenol in every matrix. What it does establish is this:
A positive result against alkaloid bitterness gives you no entitlement to assume a positive result against polyphenol bitterness.
If your bitterness comes from tea, grape, cranberry, cocoa polyphenols or a botanical active, that needs its own screen before you commit to a full trial. If it comes from caffeine, theobromine, theophylline or a botanical carrying them, you have direct sensory evidence to work from.
05 Dosage
The Working Window: Three Numbers That Are Not the Same Number
Most supplier pages give one recommended dosage. The data does not support one, and conflating three different thresholds is the commonest error in writing about this ingredient.
5 to 20 ppm — the demonstrated bitter-blocking range
From the caffeine model. Below 5 ppm, nothing. Screen here first.
100 to 200 ppm — the sweetness detection threshold for an average consumer
The patent's isointensity work put 100 ppm isosweet to 0.5% sucrose and 200 ppm isosweet to 1.0% sucrose. Critically, every panel used sweet-sensitive subjects able to detect 0.5% sucrose or less, and the patent states explicitly that the concentration detectable by an average consumer is higher — around 100 to 200 ppm.
Put those two together and the case for trilobatin as a modifier gets stronger, not weaker. You are working an order of magnitude below the point at which an ordinary consumer would notice added sweetness. The ingredient does its job invisibly.
20 ppm and up — where it starts costing you flavour
This is the number that gets left out. In the coffee example, bitterness masking at higher concentrations arrives together with suppression of the coffee flavour itself. The patent puts the onset of that effect at about 20 ppm, says 30 to 40 ppm still gives acceptable results, and states that at about 50 ppm or more the coffee flavour is significantly affected.
So in a flavour-forward system the usable window is narrow and has a hard ceiling. More trilobatin does not mean less bitterness. Past a point it means a duller product.
For dairy, dairy-derived and dairy-alternative products the patent gives a different range — roughly 10 to 500 ppm, in some embodiments up to 750 ppm — because protein raises the sweetness detection threshold. A concentration validated in a clear beverage should not be carried into a milk or soy system.
Figure 2. The three thresholds are not one number. Ranges as reported in the Givaudan patent examples; the flavour-suppression band is from the coffee model and will sit differently in other matrices. Dairy and dairy-alternative systems tolerate considerably higher levels.
06 Two problems
Stevia and Monk Fruit Do Not Have the Same Problem
The category is routinely discussed as though stevia and monk fruit share one defect called aftertaste. They do not, and treating them as one problem produces the wrong response.
Steviol glycoside bitterness has been traced to specific receptors
Hellfritsch and colleagues at TU Munich and DIfE Potsdam combined human sensory testing with cell-based functional receptor assays across nine steviol glycosides, and identified two receptors, hTAS2R4 and hTAS2R14, as mediating the bitter off-taste. They also found a structure–activity pattern: glycosides carrying more glycosyl residues are sweeter and less bitter, which is the molecular reason Reb M behaves differently from stevioside. Evidence: peer-reviewed, Journal of Agricultural and Food Chemistry, 2012.
Mogroside behaviour is contested at the receptor level
This disagreement is worth knowing rather than resolving prematurely. A 2024 cell-based functional assay found mogroside V activated at least two TAS2R bitter receptors. A 2026 molecular docking and dynamics study of twelve mogrosides reached a different conclusion: sweet mogrosides did not bind the bitter receptors, and only two minor bitter mogrosides did.
The methods are not equivalent. A cell assay measures actual receptor activation; docking predicts binding and is prone to false negatives. But the computational study carries a commercially interesting implication. If the slight bitterness of commercial monk fruit traces to a few minor mogrosides rather than to mogroside V itself, then bitterness in monk fruit is partly a purity problem rather than a molecule problem — higher mogroside V content means fewer of the bitter minor species. Grades and mogroside profiles are set out on our monk fruit extract page.
That is a different fix from the stevia fix. For stevia you are managing a property of the main sweetener. For monk fruit you may be managing a contaminant.
| Attribute | Stevia, Reb A-rich | Monk Fruit |
|---|---|---|
| Primary complaint | Bitterness | Lingering sweetness |
| Bitterness origin | The glycosides themselves, via hTAS2R4 and hTAS2R14 | Disputed; possibly minor bitter mogrosides rather than mogroside V |
| Other off-notes | Licorice, herbal, astringency | Fruity or herbal character, varies with extract purity |
| Rational response | Grade selection plus taste modification | Purity first, then modification |
07 Evidence limits
What the Evidence Covers for Stevia and Monk Fruit, and What It Does Not
This section exists because the honest answer is narrower than the popular claim.
There is direct patent evidence that dihydrochalcones as a class improve the taste of steviol glycoside and mogroside sweeteners at sub-threshold concentrations. The Coca-Cola Company holds a patent family on exactly this — WO 2018/200663, published in the US as US 2020/0054058 — covering consumables containing steviol glycoside and/or mogroside sweeteners together with a dihydrochalcone present at or below its sweetness recognition threshold. Embodiments name rebaudioside M, rebaudioside A and siamenoside I.
Here is the qualification. The dihydrochalcones named in that patent's compound claims are phloretin, hesperetin dihydrochalcone, neohesperidin dihydrochalcone and hesperetin dihydrochalcone-4′-β-D-glucoside. Trilobatin appears elsewhere in the same document, in a list of additional sweeteners. Whether trilobatin falls inside the patent's general Formula I on a substituent-by-substituent reading is a question for a patent attorney, not a formulator, and we have not made that comparison.
So the position is:
- Trilobatin against alkaloid bitterness — direct sensory evidence, Givaudan examples
- Dihydrochalcones against steviol glycoside and mogroside off-notes at sub-threshold levels — direct patent evidence, Coca-Cola
- Trilobatin specifically against stevia bitterness — inference from class membership and shared working principle, not a documented result
That inference is reasonable. Trilobatin is a dihydrochalcone, it shares the sub-threshold mechanism, and it has demonstrated bitter-reduction activity in a related context. It is still an inference, and anyone deciding whether to fund a trial should know which part of the argument rests on data and which on chemistry.
Anyone telling you studies prove trilobatin eliminates stevia bitterness is describing a study that has not been published.
08 Regulatory
Regulatory Status: It Is a Flavouring, and That Is Not a Technicality
United States
FDA's Substances Added to Food inventory lists trilobatin with the technical effect flavoring agent or adjuvant, FEMA number 4674, FEMA GRAS publication 25, JECFA flavour number 2171. The entry was current as of July 2026.
That status lines up with the use case. Sub-threshold bitter modification at 5 to 20 ppm is flavouring use. Bulk sweetening at hundreds of ppm is a different regulatory question, and flavouring-agent status should not be read as clearance to use trilobatin as a sweetener at any concentration.
China
In 2017 Lithocarpus litseifolius was added to the novel food raw material catalogue by the then National Health and Family Planning Commission. The approved edible part is specified as the young leaf or bud. That is a real limit rather than boilerplate: material derived from mature leaves sits outside the approved scope, and a supplier should be able to say which they used.
European Union
Trilobatin is not an authorised sweetener. The EU sweetener list under Regulation 1333/2008 is a closed positive list and trilobatin carries no E number. Whether it appears on the Union list of flavouring substances under Regulation 1334/2008 is a separate and, as far as we have been able to establish, unresolved question — it should be verified directly against the current Annex I entry by CAS number before any EU launch.
Supplier statements on EU flavouring status, including ours, are not a substitute for that check. If an EU launch depends on it, obtain a written regulatory opinion.
09 Grades
Your Grade Determines Your Regulatory Object
This is the point most likely to cause a problem in a registration file, and a specification sheet will not tell you.
Sweet tea extract is sold across a wide purity range — commonly 30%, 50%, 80% and 98% trilobatin. These are not four strengths of one product. They can be two different regulatory objects.
| Grade | What It Is Regulatorily |
|---|---|
| 98% trilobatin | Corresponds to trilobatin the substance — the thing carrying CAS 4192-90-9 and FEMA 4674 |
| 80% trilobatin | Borderline against typical named-substance purity expectations; composition normally has to be declared |
| 50% / 30% trilobatin | Not trilobatin. A botanical extract containing trilobatin, carrying whatever else the leaf and the process left behind |
The EU flavouring framework makes the principle explicit in its own way: named flavouring substances are expected to be at least 95% pure, and below that the composition must be declared. Different jurisdictions draw the line differently, but the logic travels.
Compliance documentation prepared for a 98% grade does not automatically cover a 30% grade.
A customer who drops grade to save cost may have changed their regulatory position without noticing. Ask before the purchase order, not after.
10 Specification
The Specification Problem: Trilobatin Versus Phlorizin
Trilobatin and phlorizin co-occur in Lithocarpus leaf. They are structurally close — close enough that separating them takes real chromatographic capability, and suppliers without it ship mixed dihydrochalcone material with a ratio that moves between batches.
This matters because “sweet tea extract” is not a specification. Two lots with the same total dihydrochalcone number can behave differently if the trilobatin-to-phlorizin split has shifted, and all the sensory work described above is on trilobatin.
What to require on the COA:
- Trilobatin content by HPLC, reported on its own line
- Phlorizin content, reported separately — not folded into a total dihydrochalcone figure
- The analytical method and reference standard used
- Ratio data across batches, not a single sample
A supplier who can only give you one combined number is telling you something about their separation capability.
One further variable is worth putting to a supplier, though the evidence for it is agronomic rather than commercial: dihydrochalcone content in sweet tea leaf changes markedly with developmental stage, with young leaves measured well above 20% trilobatin and levels falling in later stages. That is a finding about raw material, not a demonstrated cause of finished-extract variability. But it does mean harvest window is a real variable in the upstream supply chain, and a supplier who cannot describe theirs is not controlling it.
11 Bench work
A Bench Protocol That Will Produce an Answer
Given the narrow window and the temporal selectivity, a single-concentration trial is close to useless. Screen a series.
| Prototype | Trilobatin |
|---|---|
| Control | 0 ppm |
| A | 5 ppm |
| B | 10 ppm |
| C | 20 ppm |
| D | 40 ppm |
Score these separately rather than as one bitterness number:
- Sweetness onset
- Peak sweetness
- Mid-palate bitterness
- Lingering bitterness
- Astringency and dryness
- Sweet linger
- Base flavour suppression — the one most panels forget, and the one that defines your ceiling
Hold everything else constant, and run it in the actual matrix at actual pH and serving temperature. The patent notes these blockers may work under low pH conditions. That is the word used, and a claim that soft should be tested rather than assumed.
If the bitterness in your product comes from polyphenols, run a small screen on that question before committing to a full trial.
12 Sourcing
What to Ask a Supplier
| Weak Question | Better Question |
|---|---|
| What's the trilobatin content? | Is that trilobatin specifically or total dihydrochalcones, and what is the phlorizin figure? |
| Is it food grade? | Which regulatory route does this grade correspond to in my destination market, and does the documentation match that route? |
| Do you have a COA? | Does it carry trilobatin and phlorizin as separate HPLC lines, with method and reference standard named? |
| How much should I use? | At what concentration, in which matrix, at what pH, and against which bitter compound was that recommendation established? |
| Can it block stevia bitterness? | What supports that — published sensory data, patent examples, or inference from compound class? |
| Is the sample representative? | Same plant part, harvest window, cultivation region and process route as bulk supply? |
That last one has a specific edge for this ingredient. China's novel food listing covers the young leaf and bud. A supplier who cannot tell you which plant part went into the extract cannot tell you whether it sits inside that listing.
13 Bottom line
The Short Version
Trilobatin has real bitter-reduction data behind it, from one flavour-house patent family with small panels, and that data is more specific than the marketing. It worked against alkaloid bitterness — caffeine, cocoa, energy drinks — at 5 to 20 ppm. It did not work against tannin bitterness in the one direct test published. It was equivocal in tea. It acts on the middle and end of the bitterness curve rather than the beginning, which is an advantage where some bitterness is wanted. It has a ceiling near 20 ppm in flavour-forward systems, past which it starts flattening the product. Its effect on stevia specifically is a reasonable inference from its chemical class rather than a published result.
For a formulator: worth screening, on a concentration series, in your own matrix, with a scoresheet that separates the phases of the bitterness curve.
For a buyer, three questions:
Which compound am I actually buying. Which regulatory object does this grade correspond to. Can the supplier show me trilobatin and phlorizin separately.
If you are evaluating trilobatin for a specific bitterness problem, send us the bitter source, the matrix, the finished pH and your destination market. We will tell you whether the published evidence covers your case or whether you are running the first trial — and supply the grade, COA and sample for it.
14 Questions
Frequently Asked Questions
Does trilobatin remove the bitter aftertaste of stevia?
There is no published sensory study showing that trilobatin alone removes steviol glycoside bitterness. What exists is direct patent evidence that dihydrochalcones as a class improve the taste of steviol glycoside and mogroside sweeteners at sub-threshold concentrations, and separate direct sensory evidence that trilobatin reduces alkaloid bitterness. Applying the first to trilobatin is a reasonable inference from chemical class, not a documented result. Treat it as worth screening rather than as established.
Is trilobatin a sweetener or a bitterness blocker?
Both, at different concentrations, which is why the category is confusing. At 100 to 200 ppm an average consumer perceives it as weakly sweet — roughly equivalent to 0.5 to 1.0% sucrose. At 5 to 20 ppm it sits well below that threshold and has demonstrated bitter-reduction activity against alkaloids. In the United States its regulatory status reflects the second use: FDA lists it as a flavoring agent or adjuvant, not as a sweetener.
How much trilobatin should I use?
There is no single recommended dosage, and quoting one is the commonest error in supplier literature. The demonstrated bitter-blocking range against caffeine is 5 to 20 ppm, with no effect at 1.25 ppm. In flavour-forward systems such as coffee, base flavour suppression begins at about 20 ppm, remains acceptable to 30 to 40 ppm, and becomes significant at about 50 ppm. Dairy and dairy-alternative systems tolerate considerably more because protein raises the sweetness threshold. Screen a series in your own matrix rather than adopting a number.
Does trilobatin mask all types of bitterness?
No, and there is a direct negative result on this. In the same patent series that showed bitterness reduction against caffeine, cocoa and energy drinks, trilobatin at 1.25 to 20 ppm produced no reduction in the bitterness of 500 ppm tannic acid in water. Caffeine is an alkaloid; tannins are polyphenols and also contribute astringency. A positive result against alkaloid bitterness does not transfer to polyphenol bitterness by default, and a polyphenol-driven product needs its own screen.
Does trilobatin work in coffee and tea?
Differently in each. In instant coffee, 10 ppm reduced bitterness with no bitter aftertaste, and panellists noted that the mid-to-late bitterness fell while the characteristic upfront bitterness remained. In brewed black tea at the same concentration the result was equivocal: most panellists found the sample slightly less bitter, some detected no difference. Tea bitterness carries a substantial polyphenol component, which is consistent with the weaker outcome.
Is trilobatin approved for use in food and beverages?
It depends on the market and on what you are using it for. In the United States it appears in FDA's Substances Added to Food inventory as a flavoring agent or adjuvant, with FEMA number 4674 and JECFA flavour number 2171. In China the source plant was added to the novel food raw material catalogue in 2017, with the edible part specified as the young leaf or bud. In the European Union it is not an authorised sweetener, and its status on the Union list of flavouring substances should be verified against the current Annex I entry before launch. Flavouring status in any market should not be read as clearance for bulk sweetener use.
Should I buy the 30% or the 98% grade?
They are not the same product scaled differently. A 98% grade corresponds to trilobatin the substance, the identity that carries the CAS and FEMA numbers. A 30% or 50% material is a botanical extract containing trilobatin, which is a different regulatory object and carries other leaf constituents. Compliance documentation prepared for one does not automatically cover the other. Decide on regulatory route and dosing accuracy first; price per kilogram is the wrong starting point for this particular decision.
What is the difference between trilobatin and phlorizin?
Both are dihydrochalcones and both occur in the same sweet tea leaf, which is why lower-purity extracts contain a mixture of the two. They are structurally close enough that separating them requires real chromatographic capability. This matters commercially because the published sensory work is on trilobatin, so a lot specified only by total dihydrochalcone content can behave inconsistently if the ratio moves between batches. Ask for the two figures as separate HPLC lines.
15 Evidence base
References
Sources are grouped by subject. Where evidence comes from patent examples rather than peer review, that is stated in the entry.
Sensory and taste-modifier evidence
[1] Jia Z, Gray K, Potineni RV (assignee Givaudan SA). Bitter alkaloid containing consumables comprising bitter blockers. US 2011/0086138 A1; granted as US 9,844,224 B2. Priority 23 May 2008. Source of the cocoa, energy drink, coffee, black tea, caffeine and tannic acid sensory examples, the 3–200 ppm working range, the isointensity determinations and the coffee flavour-suppression thresholds. Evidence note: applicant-generated sensory data, panels of five to eight trained subjects, not independent peer review. All six matrix results cited in this article come from this single patent family. Google Patents
[2] The Coca-Cola Company. Sweetness and taste improvement of steviol glycoside and mogroside sweeteners with dihydrochalcones. WO 2018/200663 A1; US 2020/0054058 A1. Priority US 62/489,845, 25 April 2017. Establishes the sub-threshold dihydrochalcone principle for steviol glycoside and mogroside systems. Evidence note: the dihydrochalcones named in the compound claims are phloretin, hesperetin dihydrochalcone, neohesperidin dihydrochalcone and hesperetin dihydrochalcone-4′-β-D-glucoside; trilobatin appears in the additional-sweetener list. Google Patents
Receptor-level studies
[3] Hellfritsch C, Brockhoff A, Stähler F, Meyerhof W, Hofmann T. Human psychometric and taste receptor responses to steviol glycosides. Journal of Agricultural and Food Chemistry. 2012;60(27):6782–6793. DOI: 10.1021/jf301297n. Combined human sensory testing and cell-based functional receptor assays across nine steviol glycosides; identifies hTAS2R4 and hTAS2R14 as mediating the bitter off-taste, and the glycosylation structure–activity pattern. PMID 22616809
[4] Belloir C, Jeannin M, Karolkowski A, Scott C, Briand L. A receptor-based assay to study the sweet and bitter tastes of sweeteners and binary sweet blends: the SWEET project. Chemical Senses. 2024;49:bjae041. DOI: 10.1093/chemse/bjae041. Open access. Cell-based TAS1R2/TAS1R3 and TAS2R screening; source of the finding that mogroside V activated at least two TAS2Rs. Conflict-of-interest note: one author is affiliated with Cargill, a commercial sweetener producer. Oxford Academic
[5] Qin et al. Insights into the interaction mechanisms between mogrosides and sweet and bitter taste receptors obtained by integrating molecular docking and dynamics simulations. Journal of the Science of Food and Agriculture. 2026. DOI: 10.1002/jsfa.70879. Twelve mogrosides examined; reports that sweet mogrosides did not bind T2Rs and only two minor bitter mogrosides did. Evidence note: computational docking and dynamics, not a functional cell assay; conclusions differ from reference [4]. Wiley
Compound and source material
[6] Shang A, Liu H-Y, Luo M, et al. Sweet tea (Lithocarpus polystachyus Rehd.) as a new natural source of bioactive dihydrochalcones with multiple health benefits. Critical Reviews in Food Science and Nutrition. 2022;62(4):917–934. DOI: 10.1080/10408398.2020.1830363. Review of phlorizin, phloretin and trilobatin as the main dihydrochalcones of sweet tea. PMID 33030031
[7] Integrated metabolomic and transcriptomic analysis reveals the biosynthesis mechanism of dihydrochalcones in sweet tea (Lithocarpus litseifolius). Frontiers in Plant Science. 2025. DOI: 10.3389/fpls.2025.1629266. Source of the relative-sweetness figure of approximately 300× sucrose for trilobatin and phlorizin and their description as stable sweeteners and bitterness blockers. Article
[8] Temporal dynamics of bioactive compounds in sweet tea (Lithocarpus litseifolius (Hance) Chun): linking harvest stages to flavor and health benefits. Food Research International. 2025. Source of the developmental-stage variation in dihydrochalcone content, including young leaves measured above 23% trilobatin. Evidence note: an agronomic finding about raw material, not a demonstrated cause of finished-extract batch variability. PMID 40790702
Regulatory
[9] US Food and Drug Administration. Substances Added to Food inventory — TRILOBATIN. CAS 4192-90-9; technical effect: flavoring agent or adjuvant; FEMA No. 4674; FEMA GRAS publication No. 25; JECFA flavour No. 2171. Entry current as of 28 July 2026. FDA
[10] Regulation (EC) No 1333/2008 on food additives, and Regulation (EC) No 1334/2008 on flavourings, together with Commission Implementing Regulation (EU) No 872/2012 establishing the Union list of flavouring substances. Basis for the statements that the EU sweetener list is a closed positive list and that named flavouring substances are expected to be at least 95% pure. Evidence note: the current status of trilobatin on the Union list of flavouring substances was not established at the time of writing and should be verified against the live Annex I entry by CAS number. European Commission
[11] National Health and Family Planning Commission of the People's Republic of China. Novel food raw material listing for Lithocarpus litseifolius (Hance) Chun, 2017, with the edible part specified as the young leaf or bud.
Reviewed by
Xinjuan XuScientific Advisor, PhytoVale · Associate Professor, Stevia Agronomy & Crop Physiology Expert. Research focus: stevia cultivation, crop physiology, glycoside accumulation and raw material quality.
Written by
FloraContent Editor at PhytoVale. Product content and editorial specialist, focused on botanical ingredient specifications, sourcing information and B2B technical content.
This article is intended for ingredient sourcing, formulation and technical information only. It does not constitute regulatory, legal, medical, nutritional or finished-product formulation advice. PhytoVale supplies sweet tea extract, stevia extract and monk fruit extract commercially, and readers should weigh that interest when assessing the comparisons made here; where the evidence is thin, indirect or inferential, we have tried to say so in the text. Much of the sensory data cited originates in patent examples generated by the applicant using small panels, and should be treated as a starting point for the purchaser's own trials rather than as an independently validated result. Figure 1 is a schematic illustration of an effect described in words in the source patent and is not reproduced from published measurement data. Permitted uses, technical-effect classifications and maximum levels for trilobatin and sweet tea extract vary by market, food category, plant part and purity grade, and are subject to change; all regulatory statements should be verified against current legislation for the destination market before commercial launch.
Screening Trilobatin for a Bitterness Problem?
Tell us the bitter source, your matrix and finished pH. We’ll tell you whether the published evidence covers your case or whether you’d be running the first trial — and provide sample availability, COA and pricing.
