Bioactive Food Peptides: What Milk, Egg and Fish Proteins Release

Peptide science

Bioactive food peptides are the hidden cargo of ordinary dietary protein: short sequences that do nothing while they sit locked inside casein, ovalbumin or fish muscle, and that acquire measurable biological activity the moment a protease cuts them free. Food chemists have suspected this since the 1950s, but the field became commercially serious in 1995, when Japanese researchers fermenting milk with Lactobacillus helveticus purified two tripeptides — Val-Pro-Pro and Ile-Pro-Pro — that inhibited angiotensin-converting enzyme with IC₅₀ values of 9 and 5 µM. Three decades, dozens of randomised trials and one formal European rejection later, those two tripeptides remain the best test case we have for what food-derived peptides can and cannot do.

What “bioactive food peptides” actually means

The term describes sequences that are encrypted inside a larger dietary protein. In the intact protein they are inert. Cut the chain at the right two points and the fragment that falls out can bind a receptor, inhibit an enzyme, or chelate a metal. This is not exotic chemistry — it is the same logic the body uses deliberately when it carves active hormones out of larger precursors, a process covered in our guide to prohormone processing. The difference is that in food it happens by accident, or by industrial design.

Three routes release them. Ordinary gastric and pancreatic digestion does it uncontrolled, in the gut. Microbial fermentation does it with bacterial proteases — the 1995 sour-milk route. Industrial hydrolysis does it with a chosen enzyme under chosen conditions, which is how commercial ingredients are made. The parent protein is fixed; the peptide output is not. That single fact does most of the work in this article. (If the underlying vocabulary is new, start with what a peptide is.)

The best-characterised examples

• IPP and VPP (casein). The lactotripeptides. Ile-Pro-Pro inhibits ACE with an IC₅₀ of 5 µM, Val-Pro-Pro at 9 µM. They are not only a manufactured ingredient: they occur naturally in ripened cheese. Across 101 samples of ten Swiss varieties, combined VPP + IPP ranged from 1.6 to 424.5 mg/kg, with varietal means from 19.1 mg/kg to 182.2 mg/kg.
• Casokinins. The collective term for casein-derived ACE-inhibitory peptides, of which IPP and VPP are the famous two.
• Beta-casomorphin-7. Tyr-Pro-Phe-Pro-Gly-Pro-Ile, released from beta-casein residues 60 to 66, with opioid-like activity in laboratory assays — a food peptide that touches the same receptor family as the body’s own endogenous opioid peptides. It sits at the centre of the A1/A2 milk debate (see below).
• Lactoferricin. A pepsin fragment of bovine lactoferrin spanning residues 17 to 41, held in shape by an internal disulfide bond, and one of the most-studied dietary antimicrobial peptides.
• Egg peptides. Ovokinin, from ovalbumin residues 358 to 365, and the ovotransferrin tripeptide IRW. IRW is mechanistically interesting because its blood-pressure effect in rodents runs substantially through ACE2 upregulation and the Ang(1-7)/Mas axis rather than plain ACE inhibition — two different pharmacological targets producing a similar readout.
• Val-Tyr (sardine) and lunasin (soy). Val-Tyr is a dipeptide from sardine muscle and the only sequence quantified in an actual EU legal specification. Lunasin is a 43-residue soy peptide carrying an RGD cell-adhesion motif and a poly-aspartate tail.

What the human trials show — and how the effect behaves

The lactotripeptides have been tested more thoroughly than any other food peptide, which makes them a rare chance to watch an effect size behave under scrutiny. The most neutral synthesis is Fekete and colleagues (Nutrients, 2015), pooling 30 randomised controlled trials across 33 datasets.

• Pooled systolic effect −2.95 mmHg (95% CI −4.17 to −1.73); diastolic −1.51 mmHg (−2.21 to −0.80). Both statistically significant.
• Heterogeneity was high (I² = 77% for systolic), and the split was geographic: −5.54 mmHg in Japanese trials versus −1.36 mmHg in European ones, between-group p = 0.002.
• Small trials outperformed large ones — −4.93 mmHg below 50 participants against −1.99 mmHg at 50 or more (p = 0.011) — and the authors identified publication bias.
• Removing 14 low-quality trials shrank the systolic effect to −1.58 mmHg and the diastolic to −0.82 mmHg.
• 24-hour ambulatory blood pressure showed no significant change at all (p = 0.101 systolic).
• Doses of 10 mg/day or less outperformed higher ones. A non-dose-responsive pattern is exactly what a clean pharmacological effect should not look like.

A Japanese-only meta-analysis (Chanson-Rolle and colleagues, PLOS ONE 2015; 18 trials, 1,194 subjects) found a larger and more consistent effect: −5.63 mmHg systolic, I² = 44.4%, rising to −8.35 mmHg in already-hypertensive subgroups. A well-conducted Dutch trial (Engberink and colleagues, Hypertension 2008; 135 subjects, 14 mg/day for eight weeks) found nothing. Both are real results.

Two caveats deserve to be said out loud. Every product tested also delivers calcium, potassium, magnesium and phosphorus, all independently associated with blood pressure, and not every placebo was mineral-matched. And of the three syntheses above, the one with the least industry involvement is the one that found the smallest effect.

Why European regulators refused the claim

In 2011 Valio Ltd applied under Article 13(5) of Regulation (EC) No 1924/2006 for permission to state that “Peptides IPP and VPP help to maintain normal blood pressure” (Question No EFSA-Q-2011-00121). EFSA delivered its opinion on 30 September 2011 (EFSA Journal 2011;9(9):2380) and concluded that, on the data presented, a cause-and-effect relationship had not been established. Commission Regulation (EU) No 1017/2013 then listed the claim formally among rejected health claims. It is not authorised in the EU today.

On beta-casomorphin-7, EFSA’s DATEX working group reviewed the evidence in 2009 and concluded that a cause-and-effect relationship between oral intake of BCM-7 and the aetiology or course of any suggested non-communicable disease could not be established, and that a formal risk assessment was not warranted. The A1/A2 biochemistry is real — histidine at position 67 leaves that bond cleavable, proline resists it — but the leap from that polymorphism to a human health difference is not established, and the marketing is well ahead of the evidence.

In the United States these ingredients sit under DSHEA. Structure/function claims are permitted with prior substantiation and must carry the statutory disclaimer that the statement has not been evaluated by the FDA; disease claims remain drug territory.

The proline shield, and the wall it does not get you over

Most peptides are destroyed long before they reach the bloodstream, which is why peptide drugs are generally not pills. IPP and VPP are partial exceptions for a specific structural reason: proline- and hydroxyproline-containing sequences resist digestive proteases, and a C-terminal Pro-Pro pair resists even proline-specific peptidases. Di- and tripeptides can also cross the intestinal wall intact through peptide transporters, which is why the survivors of this process tend to be exactly two or three residues long.

Survival is not the same as exposure, though. In volunteers given an enriched drink, IPP was detectable in circulation and VPP was not — and most commercial products contain more VPP than IPP. The proposed mechanisms (ACE inhibition, reduced bradykinin degradation, opioid-like signalling, endothelin and nitric-oxide effects) remain unsettled. Our guides to peptide pharmacokinetics and peptide degradation cover the same wall from the pharmaceutical side.

Why a food peptide is not a defined peptide

Here is the whole legal chemical definition of the EU-authorised sardine peptide ingredient: peptides under 2 kDa, at least 85 g per 100 g; Val-Tyr, 0.1 to 0.16 g per 100 g; ash no more than 10 g per 100 g; moisture no more than 8 g per 100 g. Read that again. Eighty-five percent of the product is specified only by a molecular-weight cutoff — an unresolved mixture of thousands of sequences — and exactly one sequence is quantified, at roughly a tenth of a percent.

A defined peptide is specified the opposite way: one sequence written out in full, purity of 98% or better, named and capped impurities, a stated counterion, and a net peptide content that distinguishes peptide from salt and water (see acetate versus TFA salt). Add the cheese finding — a 265-fold range within a single food category — and the conclusion writes itself: a peptide named on a food label is a description of a mixture, not a quantity of anything. Worth remembering if you have ever used a reconstitution calculator or a concentration converter and taken for granted that the number on the vial means one substance.

One more detail from the meta-analysis: fermentation-derived products pooled at −3.82 mmHg and enzymatically hydrolysed ones at −2.77 mmHg, but the difference between routes was not statistically significant (p = 0.433).

Frequently asked questions

Are bioactive food peptides the same thing as peptide drugs?

No, and the difference is one of definition rather than origin. A peptide drug is one sequence, quantified, with a specified impurity profile. A food peptide ingredient is usually a hydrolysate defined by a size range, with at most one or two sequences measured. Some individual food peptides are genuinely well characterised; the products built around them generally are not.

Does A2 milk avoid beta-casomorphin-7?

The biochemistry supports the mechanism: proline at position 67 in A2 beta-casein resists cleavage at that bond, so less BCM-7 is released than from A1. What is not established is that this produces a health difference in people. EFSA’s 2009 review found no demonstrable cause-and-effect link to any suggested disease, and some studies have failed to detect casomorphins where the hypothesis predicted them.

Should I read the blood-pressure numbers as a reason to change anything?

No. This article describes what published trials and regulators concluded; it is not health guidance. The EU has formally refused the blood-pressure claim for IPP and VPP, the pooled effect is small and heterogeneous, and it does not survive 24-hour ambulatory measurement. Any decision about blood pressure belongs with a qualified clinician.

Why do IPP and VPP survive digestion when most peptides do not?

Length and proline. At three residues they are short enough to cross the intestinal wall through peptide transporters rather than needing full hydrolysis, and their proline content — particularly the C-terminal proline pair — blocks the proteases that would otherwise clear them. Chemists exploit the same trick deliberately in drug design; see peptide half-life.

Informational only — not medical advice · 21+