Aspartame

Aspartame is a dipeptide: L-aspartic acid joined to L-phenylalanine, with the carboxyl end of the phenylalanine capped as a methyl ester[1]. It is one of the most widely eaten peptides, and it shows two things that apply to peptides in general unusually clearly. Its taste depends entirely on the exact arrangement of its atoms, including the handedness of both amino acids. And once swallowed, it is taken apart like any other small peptide in food.

It is also one of the most argued-over food additives. This page covers its chemistry first, then what the regulators and the 2023 cancer review actually concluded.

Aspartame at a glance
StructureL-α-aspartyl-L-phenylalanine methyl ester (L-Asp-L-Phe-OMe)
FormulaC14H18N2O5
Molecular weight294.3 Da; calculated
SweetnessAbout 200 times sucrose
Broken down toAspartate, phenylalanine and methanol
Acceptable daily intake40 mg/kg body weight (JECFA); 50 mg/kg (US FDA)

Found in a peptide synthesis

Aspartame was not designed. Chemists at G. D. Searle were synthesising the last four residues of the hormone gastrin, Trp-Met-Asp-Phe-NH2, and one of them, James Schlatter, was crystallising an intermediate, aspartylphenylalanine methyl ester, when he “noticed that it was sweet”[2]. The authors' paper says no more about how, and the authors were frank that this was how sweeteners were usually found: “The discovery of sweet-tasting compounds has been completely accidental”[2].

Their first tests put it at 100 to 200 times sweeter than sucrose, depending on concentration and on what other flavours were present, with no unpleasant aftertaste[2]. The US regulator now gives a round figure of about 200 times[3]. A later account dates the discovery to 1965; the US Food and Drug Administration first approved the sweetener in 1974, then stayed that approval while objections were heard[4].

Why only one isomer is sweet

Aspartame drawn as three boxes: aspartic acid, phenylalanine and a methyl ester group, both amino acids marked L with the S configuration. Dashed lines mark the peptide bond and the ester bond as the places cut in the gut, giving aspartate, phenylalanine and methanol. Below, four boxes show the four stereoisomers; only L-Asp-L-Phe-OMe is sweet, the other three are bitter.
Aspartame is L-Asp-L-Phe-OMe; both amino acids are L, which by the Cahn–Ingold–Prelog rules is (S) for each. Of its four stereoisomers, only this one is sweet. Structure and the taste of the isomers from Mazur, Schlatter and Goldkamp (1969)[2]; breakdown products from Stegink (1987)[4]; mass calculated.

The sweetness is not borrowed from the parts. On their own, L-aspartic acid is tasteless and L-phenylalanine is bitter[2]. So the Searle chemists set about changing the molecule one piece at a time, tasting 1% solutions from a cotton swab, the approximate threshold at which untrained tasters notice sucrose[2].

The aspartic acid turned out to be untouchable. Replacing it with any of sixteen other amino acids gave esters that were all bitter. Moving the peptide bond to aspartic acid's side-chain carboxyl, or using glutamic acid with its one extra carbon, removed the sweetness. So did capping the free amino group[2].

The handedness mattered just as much. The chemists made every stereoisomer of aspartylphenylalanine methyl ester, and every one except the L,L form was bitter, including the two that keep one L-amino acid and change the other[2]. They concluded that the free amino group, one carboxyl group of aspartic acid, the distance between them and the configuration of the chiral carbon were “completely critical”, and suggested that these charged groups must bind directly to a receptor site in the taste buds[2].

The other end was more forgiving. Swapping phenylalanine for tyrosine or methionine kept most of the sweetness, though the methionine version developed a sulfurous taste on standing. The methyl ester was needed: the free acid and the simple amide were not sweet, and larger esters grew steadily less sweet. Their summary was that the aspartic acid part could not be changed, but there was “room for substantial manipulation of the phenylalanine portion”[2].

What the body does with it

Aspartame is not absorbed as aspartame. One route is for digestive enzymes in the gut to split it straight into aspartate, phenylalanine and methanol. The other is for the methyl ester to be removed first, leaving the dipeptide aspartyl-phenylalanine, which is taken up into the cells lining the intestine by the transporters that carry dietary di- and tripeptides and is then split inside those cells. Either way, aspartate, phenylalanine and methanol reach the portal blood, which Stegink compares with the way amino acids arrive from dietary protein and methanol from the polysaccharides in food[4]. The 2023 cancer review found this metabolism to be similar in humans and experimental animals[5].

The question for safety was therefore whether those three products ever reach harmful levels. Lewis Stegink's group at the University of Iowa tested this directly in volunteers, at the request of Searle, which asked them to help design and run clinical and animal studies[4].

Projections of the highest likely daily intake ranged from 22 to 34 mg per kilogram of body weight, the top figure being the 99th percentile of projected consumption. At 34 mg/kg, methanol in the blood stayed below the limit of detection, 0.4 mg/dL. Even at 200 mg/kg, blood formate, the toxic product of methanol, did not rise; only urinary formate increased for a few hours[4]. For comparison, a drink sweetened with aspartame was estimated to contain about 55 mg of methanol per litre, against an average of 140 mg per litre in fruit juice[4].

Stegink’s summary was that “blood formate concentrations as well as plasma aspartate and phenylalanine concentrations remained within the normal postprandial range after aspartame ingestion and were far below concentrations of these compounds associated with adverse effects”[4]. He was also clear about the limits of the work: these were studies of single large doses, and they did not test the effect of repeated doses[4].

The diketopiperazine

Aspartame can also close on itself into a ring, aspartylphenylalanine diketopiperazine. This happens over time, at extremes of pH and temperature, and especially in liquids[6]. The ring compound was tested for safety in its own right, and JECFA in 1980 and the European Scientific Committee for Food in 1985 set a separate acceptable daily intake for it, 7.5 mg/kg body weight[6].

The same reaction, and why some peptide sequences undergo it easily, is described on the peptide stability and storage page. In the United States, tabletop aspartame sweeteners must carry instructions not to use them in cooking or baking[1].

Phenylketonuria and the label

More than half of the aspartame molecule by weight is phenylalanine: 34 mg/kg of aspartame delivers up to 19 mg/kg of it[4]. Grossly raised blood phenylalanine, of the kind found in children with classic phenylketonuria, is associated with intellectual disability[4]. In the United States, every product containing aspartame must therefore carry the statement “PHENYLKETONURICS: CONTAINS PHENYLALANINE”[1].

The Iowa studies also looked at carriers of the condition, who have one working copy of the gene. They cleared phenylalanine from aspartame more slowly than other people, but their peak blood levels stayed well below those associated with harm, and the authors judged that aspartame posed no significant risk to them[4].

The 2023 cancer reviews

In June 2023 a working group of 25 scientists from 12 countries met at the International Agency for Research on Cancer (IARC) and classified aspartame as “possibly carcinogenic to humans”, Group 2B[5]. Every strand of evidence was rated only “limited”: in humans, in experimental animals, and in mechanistic studies[5].

The human evidence was limited to one cancer, hepatocellular carcinoma of the liver, and rested on three cohort studies that linked artificially sweetened drinks with liver cancer. The group judged these studies to be of high quality, but concluded that “chance, bias, or confounding could not be ruled out with reasonable confidence”. For other cancers the evidence was rated “inadequate”[5]. The animal studies were mixed: several negative studies predated modern laboratory standards, and the positive ones raised questions about design, such as not adjusting for litter effects. A minority of the group would have rated the animal evidence higher and placed aspartame in Group 2A[5].

The largest study measuring aspartame from all foods, the French NutriNet-Santé cohort, followed 102 865 adults, 78.5% of them women, for a median of 7.8 years. Those who consumed more aspartame than the median among consumers had a 15% higher rate of cancer overall than non-consumers (hazard ratio 1.15, 95% CI 1.03 to 1.28), with higher rates of breast cancer and of obesity-related cancers[7]. Every participant's intake was below the acceptable daily intake[7]. The authors listed selection bias, residual confounding and reverse causation as limitations[7], and the IARC group noted that these findings were not consistent across all studies and that this cohort had not looked at liver cancer[5].

IARC's classification identifies a hazard; it does not say how much is risky. That second question went immediately to the Joint FAO/WHO Expert Committee on Food Additives (JECFA), which found “no sufficient reason to change the previously established acceptable daily intake” of 40 mg/kg body weight. In WHO's illustration, a 70 kg adult would need to drink more than 9 to 14 cans of diet soft drink a day, at 200 to 300 mg of aspartame per can, to exceed it[8]. Both IARC and WHO called for more research, including studies with longer follow-up[8]. The US FDA, whose own acceptable intake is 50 mg/kg, stated that it “disagrees with IARC's conclusion that these studies support classifying aspartame as a possible carcinogen to humans”[3].

A note on sources. Much of the early safety work on aspartame was done by, or at the request of, its manufacturers: the 1969 discovery paper came from Searle[2], the Iowa studies were designed with Searle's involvement[4], and the most extensive safety review was written mainly by staff of The NutraSweet Company[6]. The NutriNet-Santé study was publicly funded[7].

In short

  • Aspartame is L-aspartyl-L-phenylalanine methyl ester, found by accident in 1965 during the synthesis of a gastrin peptide. It is about 200 times sweeter than sugar.
  • Only the L,L stereoisomer is sweet. Changing either chiral centre, replacing the aspartic acid, or removing the methyl ester makes it bitter or tasteless; the phenylalanine end tolerates more change.
  • In the gut it is cut into aspartate, phenylalanine and methanol, either directly or via the dipeptide Asp-Phe taken up by peptide transporters. At the highest projected intakes, blood methanol stayed undetectable in single-dose studies.
  • Over time, especially in liquids at extremes of pH and temperature, it closes into a ring, a diketopiperazine, which has its own separate safety limit.
  • In the United States, products containing it must warn people with phenylketonuria, because more than half of the molecule by weight is phenylalanine.
  • In 2023 IARC placed aspartame in Group 2B on limited evidence, mainly for liver cancer. JECFA kept the acceptable daily intake at 40 mg/kg, and the US FDA disagreed with the IARC classification.

References

  1. US Code of Federal Regulations. Title 21, §172.804 Aspartame. Electronic Code of Federal Regulations. ecfr.gov
  2. Mazur RH, Schlatter JM, Goldkamp AH (1969). Structure-taste relationships of some dipeptides. Journal of the American Chemical Society 91(10):2684–2691. DOI: 10.1021/ja01038a046
  3. U.S. Food and Drug Administration. Aspartame and other sweeteners in food. Content current as of 27 February 2025. fda.gov
  4. Stegink LD (1987). The aspartame story: a model for the clinical testing of a food additive. American Journal of Clinical Nutrition 46(1):204–215. DOI: 10.1093/ajcn/46.1.204
  5. Riboli E, Beland FA, Lachenmeier DW, Marques MM, Phillips DH, Schernhammer E, et al. (2023). Carcinogenicity of aspartame, methyleugenol, and isoeugenol. The Lancet Oncology 24(8):848–850. DOI: 10.1016/S1470-2045(23)00341-8
  6. Butchko HH, Stargel WW, Comer CP, Mayhew DA, Benninger C, Blackburn GL, et al. (2002). Aspartame: review of safety. Regulatory Toxicology and Pharmacology 35:S1–S93. DOI: 10.1006/rtph.2002.1542
  7. Debras C, Chazelas E, Srour B, Druesne-Pecollo N, Esseddik Y, Szabo de Edelenyi F, et al. (2022). Artificial sweeteners and cancer risk: results from the NutriNet-Santé population-based cohort study. PLOS Medicine 19(3):e1003950. DOI: 10.1371/journal.pmed.1003950
  8. World Health Organization (2023). Aspartame hazard and risk assessment results released. News release, 14 July 2023. who.int