Research note 044 — 2026-09-04
Who's to Blame for Sake's Bitterness? — The Peptide Account and the Amino-Acid Account
Sake has a distinctive bitterness. At least two explanations for what carries this bitterness stand side by side in the literature. One attributes it to oligopeptides born when the steamed rice's protein — the rice's glutelin — is cut up by koji's enzymes. The other attributes it to free amino acids, arginine among them. This note does not try to settle which is correct; it lays out both explanations and their evidence side by side, along with how the processes of the rice polishing ratio, activated-carbon treatment, enzyme treatment, and storage move that bitterness.
On this page (9 sections)
- What this page establishes
- Bitter peptides are not unique to sake
- But pinning the blame on a single peptide is hard
- In sake, a rice-glutelin-derived peptide is a candidate
- Free amino acids — another explanation
- The word zatsumi, and the phrase tanrei-karakuchi
- Honest notes
- Frequently asked
- What is not known
What this page establishes
- Two explanations for what carries sake's bitterness stand side by side in the sources: rice-glutelin-derived peptides, and free amino acids such as arginine.
- The tendency for bitter peptides to decrease when the rice polishing ratio is lowered has been confirmed in only one variety (Nihonbare); the same tendency has not been seen in other varieties.
- That adding a high concentration of arginine to synthetic sake makes bitterness and sweetness more likely to be pointed out has been shown in one discrimination test.
- No experiment directly comparing whether the peptide account or the amino-acid account explains more of sake's bitterness has yet been found.
Bitter peptides are not unique to sake
That peptides can carry bitterness is not, in itself, a phenomenon unique to sake. A review on cheese positions bitterness as an inevitable consequence of the proteolysis that proceeds during ripening, and lays out the relationship between how far that breakdown has progressed and the accumulation of bitter peptides.
Scope: A review on cheese bitterness (1996). It covers isolation and detection methods for bitter peptides, their physicochemical properties, and control methods.
Show data
compound-class-bitter-peptides —is_produced_through→ "Cheese bitterness is described as an inevitable consequence of the proteolysis that proceeds during ripening, and the relationship between proteolytic activity and the accumulation of bitter peptides is organized in a review."There are also clues, from a separate experimental system, about how such peptides make bitterness felt. The bitter dipeptides Gly-Phe and Gly-Leu (25 mM) strongly activated hTAS2R1, one of the human bitter receptors, while the tasteless Gly-Gly activated no receptor even when tested up to 100 mM. Bitter tripeptides showed even stronger activation, and the effective concentration of the strongest, Phe-Phe-Phe, was in the micromolar range.
Scope: Heterologous-expression cell calcium assays: HEK293 (Maehashi 2008) and C6 glioma (Upadhyaya 2010)
Note: 2 independent lines of cell assay point to the same receptor. The contribution in vivo is not established.
Show data
compound-class-bitter-peptides —activates→ receptor-tas2r1The order of magnitude of the threshold has also been reported for other peptides. 12 purified and synthesized cheese-derived peptides showed clear bitterness, with recognition thresholds ranging from 0.05 to 6.0 mmol/L. The lowest was the decapeptide YPFPGPIHNS, whose bitter quality resembles caffeine's, at 0.05 mmol/L.
Scope: Peptides purified from aged Gouda cheese and then synthesized.
Note: The abstract gives neither the threshold-measurement method nor the panel size.
Show data
compound-class-bitter-peptides —has_recognition_thresholds_of→ "0.05-6.0 mmol/L (12 synthesized and purified peptides showed clear bitterness). The lowest threshold was 0.05 mmol/L, for the decapeptide YPFPGPIHNS, with a caffeine-like bitterness quality."But pinning the blame on a single peptide is hard
In a mixture where a whole protein has been digested, multiple receptors respond at the same time. In a bitter fraction from casein digested with trypsin, cells expressing hTAS2R1, hTAS2R4, hTAS2R14, and hTAS2R16 all responded, with hTAS2R1 and hTAS2R4 responding especially strongly. However, the authors who ran the experiment themselves state explicitly that this fraction is a mixture of many structurally different peptides and is likely to include nonbitter peptides as well.
Scope: HEK293 heterologous expression, calcium imaging
Note: The fraction is a mixture of numerous peptides; the correspondence between individual peptides and receptors cannot be identified from this experiment alone (see limits).
Show data
compound-class-bitter-peptides —activates_multiple_members_of→ receptor-tas2r-familyLimits have also been reported for using one specific peptide as a marker of bitterness. The concentration of β-CN f193-209, a peptide that has been linked to cheese's bitterness, correlated with sensory bitterness intensity at a correlation coefficient of 0.803 at 180 days of ripening, but that fell to 0.554 at 270 days. The assessment is that this peptide is not the sole cause but remains a partial marker.
Scope: Reduced-fat Cheddar cheese (Broadbent 2002). The association with bitterness is cited by that paper from earlier literature.
Note: The association with bitterness itself is cited by Broadbent 2002 from earlier literature (that paper's refs 4 and 28); there is no threshold measurement within this paper itself.
Show data
compound-class-bitter-peptides —includes_chymosin_derived_member→ "β-CN f193-209: a chymosin-derived peptide that has been associated with bitterness in cheese. It is not hydrolyzed by lactocepin, and its concentration shows no difference among starters (0.3-0.9 mg/g, P>0.5). Combined with αs1-CN f1-9, an index for predicting the tendency toward bitterness has been proposed (r=0.84)."The same difficulty has been pointed out in another cheese study. The correlation between the concentration of individual peptides and the bitterness score was weak (correlation coefficient 0.56 to 0.58), and only rose to 0.84 with an index combining two peptides. That study's authors write that there must indeed be other peptides contributing to bitterness as well.
Scope: Aged Gouda and aged sharp Cheddar (water-soluble extracts).
Note: The explanatory power of any single peptide alone is limited: in Broadbent 2002, the correlation between a single peptide's concentration and the bitterness score stayed at r=0.56-0.58, versus r=0.84 for a combined 2-peptide index (see limits).
Show data
ingredient-cheese —derives_bitterness_in_part_from→ compound-class-bitter-peptidesEven so, there is an example of a reconstruction experiment that succeeded by adding multiple peptides together. 5 peptides analyzed as chiefly carrying the bitterness of aged, real sharp Cheddar were added, at the same concentrations, to a young mild Cheddar, and the resulting bitterness intensity was reproduced to a point with no significant difference from the real sharp Cheddar. Among them, GPVRGPFPIIV was analyzed as making the main contribution, and it had increased up to 28.7-fold during ripening.
Scope: Aged sharp Cheddar versus young mild Cheddar. A taste-reconstitution experiment.
Note: The sequences are as given in the abstract (residue numbers are not written because they are not given in the abstract).
Show data
compound-class-bitter-peptides —reconstitute_bitterness_of→ ingredient-cheeseIn sake, a rice-glutelin-derived peptide is a candidate
Returning to sake, there is research showing the possibility that oligopeptides — born when steamed-rice protein, the acidic subunit of rice's glutelin, is digested by koji's enzymes — are involved in sake's flavor. The details of their structure and their detection and recognition thresholds were covered in a separate research note, 「発酵は苦味を作ってから壊す」 (in translation: "Fermentation Creates Bitterness, Then Breaks It Down"). Here, we look at how the amount of that peptide moves across the brewing process.
How the amount of this high-molecular-weight peptide moves as more of the rice is polished away (as the rice polishing ratio falls) turned out to differ by variety. Nihonbare showed a tendency for the amount of high-molecular-weight peptide to decrease as the rice polishing ratio fell, but Yamada Nishiki and Gohyakumangoku showed no such tendency.
Scope: In a steamed-rice digestion test on 3 varieties prepared with rice polishing ratios stepped down from 90% to 40%, Nipponbare showed a tendency for the amount of high-molecular-weight peptide to decrease as the rice polishing ratio fell (i.e., as milling became more thorough), but Yamada Nishiki and Gohyakumangoku showed no such tendency.
Note: Hashizume (2008), Figure 4. Measured data showing that the relationship between rice polishing ratio and bitter-peptide generation depends on the raw-rice variety and is not uniform. Based on a digestion test with koji mold (strain RIB128) (pH 4.3, 37°C, left standing for 1–2 days).
Show data
compound-class-sake-glutelin-peptides —are_not_uniformly→ "Changes in the amount of high-molecular-weight peptide generated in steamed-rice protein digests, as a function of rice polishing ratio (90–40%)"Activated-carbon treatment after brewing also moves the amount of this peptide. As the amount of activated-carbon treatment increased — 200, 500, 1000 mg/l — the peptide concentration in the sake fell, and at the 1000 mg/l treatment most of it fell below the detection limit. In untreated daiginjo at a rice polishing ratio of 40%, one peak was 10.5 to 12.1 mmol/L, compared with a lower 0.8 to 3.5 mmol/L in commercial futsū-shu A-C.
Scope: The more the amount of activated-carbon treatment (200/500/1000 mg/l) was increased, the more the peptide concentration in sake (peaks 13,17,18,20) fell, and with 1000 mg/l treatment most fell below the detection limit. In untreated daiginjo (2001–2004 vintage, 40% rice polishing ratio), peak 13 was 10.5–12.1 mM, compared with a lower 0.8–3.5 mM in commercial futsū-shu (regular sake) A–C (consistent with the tendency for futsū-shu using more activated carbon to be lower).
Note: Hashizume (2008), Table 4. Mean ± standard deviation (3 measurements).
Show data
compound-class-sake-glutelin-peptides —is_reduced_by→ "Activated-carbon treatment"With enzyme-based treatment, it has been reported that the effect of reducing bitterness comes paired with another change. Adding to sake an enzyme preparation with high acid carboxypeptidase activity reduced bitter peptides while increasing a different short-chain peptide, (pGlu)-L-ethyl ester. Meanwhile, an enzyme preparation with different properties changed which kind of peptide increased. For enzymes showing carboxypeptidase-type activity, there is also a statement that 「苦味ペプチドの低減に寄与する一方で，呈味の優れないジペプチドの生成に関わる」 (in translation: "while it contributes to reducing bitter peptides, it is also involved in producing dipeptides with poor flavor").
Scope: According to an abstract by Hashizume et al. (publication year not confirmed, original not read in full) introduced in "Research Achievements on Alcoholic Beverages in Reiwa 4" (Journal of the Brewing Society of Japan, vol. 118, no. 4, p.220), adding a peptidase preparation with high ACP activity to sake reduced bitter peptides and increased (pGlu)L-ethyl. The source reports that, on the other hand, adding an enzyme preparation with high (pGlu)-peptide ethyl ester-forming activity and low ACP activity increased (pGlu)LF-ethyl and (pGlu)LFNP-ethyl in sake (without increasing (pGlu)L-ethyl), and it was suggested that this increase in oligopeptide ethyl esters gives a good effect on sake's taste. A separate abstract by Hashizume et al. in the same journal, p.233, states explicitly that an enzyme showing carboxypeptidase-type activity "contributes to the reduction of bitter peptides while also being involved in the generation of dipeptides with poor taste," suggesting that reducing bitterness can carry a quality trade-off as well.
Note: Original not read in full. A secondary citation from an abstracting journal (a summary by the editorial staff). Even for enzyme treatment of the same peptide lineage, the treatment "reduction of bitter peptides = desirable" coexists with the treatment "increase in the esterified derivative = a good effect on taste"; this is included as an example showing this two-sidedness in how bitterness is treated. The addition of the enzyme preparation with high ACP activity itself also carries a two-sided effect, reducing bitter peptides while increasing (pGlu)L-ethyl (see also the separate Hashizume et al. abstract on p.233 of the same journal, vol.118 no.4).
Show data
compound-class-sake-glutelin-peptides —has_bitterness_quality_reduced_by→ "Addition to sake of a peptidase preparation with high acid carboxypeptidase (ACP) activity."Storage — time itself — also moves bitterness. Sake that has been hi-ire (heat-pasteurized) is recorded as showing bitterness and a smooth pungency increasing with storage, while astringency and a rough pungency decrease. In the same table's entry for nama-zake (unpasteurized sake, before hi-ire), no increase or decrease in bitterness is recorded.
Scope: According to Table 4 of Utsunomiya et al. (2006) (evaluation terms describing flavor changes accompanying sake storage), for pasteurized sake, the tastes "bitterness, stimulant taste (smooth)" increase with storage while "astringency, stimulant taste (rough)" decrease. The same table's entry for unpasteurized sake (before pasteurization) does not record any increase or decrease in bitterness.
Note: Based on the storage flavor-change table (Table 4), excluding hineka and namahineka (stale off-flavors). Note that the behavior of bitterness during storage is written up differently for unpasteurized sake and pasteurized sake.
Show data
ingredient-sake —has_perceived_bitterness_affected_by→ "Increase: storage of pasteurized (heat-sterilized) sake."Not all peptides work only in the direction of bitterness, though. A report is also introduced stating that peptides in sake make a favorable sensory contribution. This statement's cited source has not been checked directly, and it is not clear whether the peptide it refers to is the same as the glutelin-derived bitter peptide.
Scope: Sake's peptides are reported to make a favorable sensory contribution to flavor (introducing a citation to an earlier study).
Show data
ingredient-sake —has_positive_sensory_attributes→ compound-class-bitter-peptidesFree amino acids — another explanation
There is also an explanation that attributes sake's bitterness not to peptides but to free amino acids. One review sums up sake's distinctive bitterness as originating from valine, isoleucine, leucine, proline, and arginine. This review presents no citation numbers or individual quantitative or sensory-test data.
Scope: A review statement that "sake's distinctive bitterness is attributed to valine, isoleucine, leucine, proline, and arginine."
Note: A statement on p.336 of the Ogawa (2015) review. It is presented as a general summary without an explicit citation number, and this source contains no individual quantitative or sensory data.
Show data
ingredient-sake —has_bitterness_attributed_to→ "Valine, isoleucine, leucine, proline, arginine."There is also an experiment that tested part of this explanation. Arginine was added, at 1500 ppm — 10 times the amino-acid concentration in commercial sake — to a synthetic sake formulated with ethanol, sugar, acid, salt, and aroma compounds, and a 3-point (triangle) discrimination test comparing it against the unadded version gave correct discrimination in 23 of 25 trials (significant at the 0.1% level). In a 3-point preference test, the arginine-added sample drew frequent comments of 「苦味，甘味」 (in translation: "bitterness, sweetness") and was judged, at the 0.1% significance level, to be an undesirable amino acid in terms of flavor. However, this 1500 ppm is a high condition, 10 times the concentration found in natural sake, and the panel was centered on a small, untrained group of students and faculty.
Scope: In a 3-point (triangle) discrimination test on pure synthetic sake (formulated with 20% ethanol, glucose, succinic acid, lactic acid, NaCl, KH2PO4, isoamyl alcohol, isoamyl acetate, and ethyl caproate) to which arginine was added at 1500 ppm, 10 times the amino-acid concentration in commercial sake, 23 of 25 answers were correct (significant at the 0.1% level). In a 3-point preference test, the arginine-added sample drew frequent comments of "bitterness, sweetness" and was judged, at the 0.1% significance level, to be a sensorially undesirable amino acid.
Note: Iwano et al. (2004). The panel was small and untrained: 3 faculty members and 2 undergraduate thesis students, 5 people x 5 repetitions = 25 judgments. This is an examination under a high-concentration condition, 10 times the concentration in natural sake, and does not show the size of the contribution at ordinary concentrations. The English abstract states explicitly, "Arginine was judged as the amino acid causing bitterness."
Show data
ingredient-sake —has_perceived_bitterness_associated_with→ compound-arginineArginine coming up as a candidate for bitterness is not a sake-only story either. In a sesame paste soaked for 2 hours, germinated for 36 hours, and roasted at about 130°C for 40 minutes, arginine content was reported at 90.82 mg per 100 g, with a TAV (a rough measure of contribution to flavor) of 1.82, and it was named as one of the candidate bitterness components. However, this report does not present a sensory-test method, and TAV does not establish a causal relationship within a food matrix.
Scope: Paste SG2-36, after 2 hours of soaking, 36 hours of germination, and roasting at about 130°C for 40 minutes
Note: Arg's TAV was above 1, but the present paper does not present a sensory-test design, and 『最も苦い』 (in translation: 'the most bitter') depends on the description in an earlier experiment.
Show data
seasoning-germinated-sesame-paste —has_candidate_bitterness_contributors→ compound-arginineThe word zatsumi, and the phrase tanrei-karakuchi
Among the words used to describe sake's aroma and flavor, some are composite terms that include bitterness without being bitterness itself. The definition of the word 「雑味」 (in translation: "zatsumi"), its treatment relative to standard terminology, and how it connects to bitter peptides were covered in a separate research note, 「発酵は苦味を作ってから壊す」 (in translation: "Fermentation Creates Bitterness, Then Breaks It Down"). Here we focus on the opposition between the peptide account and the amino-acid account.
This opposition also bears on industry trends concerning sake's raw rice. There is an explanation that the reason low-protein raw rice is called for lies in the trend toward lighter, more delicate sake quality and the increase in ginjo-type sake.
Scope: An explanation of the research motivation, stating 『低タンパク質質の原料米が求められる背景には，酒質の淡麗化や吟醸タイプ清酒の増加があると考えられる』 (in translation: "the reason low-protein raw rice is called for is thought to be the trend toward lighter, more delicate sake quality and the increase in ginjo-type sake").
Note: A statement from the introduction (p.574) of Hashizume (2008). This is the author's background explanation of an industry trend, and is treated as discursive framing (there is no numerical verification in the body text).
Show data
ingredient-sake —is_associated_with→ "Industry trend toward demanding low-protein raw rice (sake quality becoming lighter and more delicate, and an increase in ginjo-type sake)"The same review states that, from the recent preference for a highly aromatic, light taste — the so-called tanrei-karakuchi preference — attention to ginjo sake has been increasing, and it adds the words 「米白ければ酒にくせなし」 (in translation: "if the rice is white, the sake has no kuse"). However, this 「くせ」 (kuse) does not name bitterness or zatsumi directly; it is a general phrase referring to the lightness or purity of sake quality.
Scope: 『近年の香り高く味が淡い嗜好，いわゆる淡麗辛口嗜好からの吟醸酒の注目度が増している。まさに“米白ければ酒にくせなし”である。』 (in translation: "In recent years, attention to ginjo sake has been increasing, driven by the preference for a highly aromatic, light taste — the so-called tanrei-karakuchi preference. It is precisely 'if the rice is white, the sake has no kuse.'")
Note: Discursive framing from Ogawa (2015), review, p.331. Note that 『くせ』 (in translation: kuse) here is not naming bitterness or zatsumi directly; it is cited as an idiomatic phrase linking the rice polishing ratio to lightness of taste in general. Because this is not discourse that speaks of bitterness directly, it is kept as a weak, contextual connection.
Show data
ingredient-sake —is_associated_with→ "Preference for the tanrei-karakuchi (light and dry) style — a recent preference for sake that is highly aromatic and light in taste"Honest notes
- The statement about the enzyme preparation (acid carboxypeptidase) has not been checked directly against the original paper; it is a summary that carries over the abstract's own wording as is.
- The arginine test in synthetic sake was conducted under a high condition, 10 times the concentration found in natural sake, and the panel, too, was a small, untrained group.
- The rice-polishing-ratio experiment is a comparison of only 3 varieties, and it cannot be generalized to other varieties.
- The findings about receptor activation, thresholds, and the weakness of attribution to a single peptide for cheese-derived peptides are not direct measurements of sake's own peptides; they are laid out here as contextual information for comparison.
- No experiment directly comparing which of the peptide account or the amino-acid account explains more of sake's bitterness has been found within the scope of this inventory.
Frequently asked
Does raising the rice polishing ratio reduce sake's bitterness?
In an experiment that digested steamed-rice protein in stages, the variety Nihonbare showed a tendency for the amount of bitter peptide to decrease the more the rice polishing ratio was lowered (that is, the more thoroughly the rice was milled). However, in the same experiment, the two other varieties, Yamada Nishiki and Gohyakumangoku, showed no such tendency. In other words, the relationship between the rice polishing ratio and bitter peptides changes by variety. This experiment is limited to a comparison of a small number of varieties, and what happens in other varieties has not yet been confirmed. How the rice polishing ratio works on bitterness from the free-amino-acid side is also not included in this material.
Does sake become more bitter when stored?
A record of stored sake that had been hi-ire (heat-pasteurized) sets out that bitterness and a smooth pungency increase with storage, while astringency and a rough pungency decrease. Meanwhile, the same record's entry for nama-zake, before hi-ire, does not note any increase or decrease in bitterness itself. In other words, the answer may change depending on whether hi-ire was done, and the relationship between storing nama-zake and bitterness cannot be determined from this material. Whether the cause of increasing bitterness is a change in peptides or a change in amino acids also cannot be confirmed from this record alone.
What is not known
- When sake's bitterness increases with storage, whether that change is due to an increase or decrease in peptides or in free amino acids has not been confirmed from this record alone.
- How the operation of reducing bitter peptides through enzyme treatment moves bitterness on the free-amino-acid side has not yet been confirmed.
- How strong the receptor responses and bitterness thresholds reported for cheese-derived peptides are, compared with the actual concentrations of bitter peptides in sake, cannot be determined from this material.
- No experiment has been found in this inventory that individually confirms how much valine, isoleucine, leucine, and proline each contribute to bitterness in sake.
Related:sake ／sake bitter peptides from rice glutelin ／bitter peptides ／L-arginine ／zatsumi
Related topic hub: Bitterness, Astringency, and Egumi
Related research notes: Fermentation Creates Bitterness, Then Breaks It Down — Miso, Soy Sauce, Sake ／Standards Certify Bitterness, and the Shelf Chooses It ／The Map of "Bitter" — From One Word to East and West
Source ledger: Source list
How to cite this page
Aoki, M. (NIGAMI LAB) (2026). Sake Bitterness: Peptides or Amino Acids?. https://nigamilab.com/en/notes/sake-bitterness-debate/
Each claim card has a stable anchor (#clm-…). The dataset is CC BY 4.0; see the data page. Dataset DOI: 10.5281/zenodo.22739101