DSIP (Delta Sleep-Inducing Peptide): Nonapeptide Chemistry and Research Findings

DSIP was characterized in 1977 by Schoenenberger and Monnier, who reported a nonapeptide recovered from rabbit cerebral venous blood that enhanced delta and spindle activity in the electroencephalogram after intraventricular infusion. Nine synthetic peptides were compared across 58 rabbits under double-blind conditions, and only the complete nonapeptide produced the effect.

The 1977 report tested five plausible breakdown fragments spanning residues 1-8, 2-9, 2-8, 1-4 and 5-9, two nonapeptide analogues with two residues exchanged, and the tripeptide Trp-Ser-Glu. None reproduced the electroencephalographic signature. Requiring all nine residues was, at the time, the strongest argument that a discrete molecular entity rather than a degradation artefact had been isolated.

A 2025 historical review by Banks places the isolation inside a wider argument then unfolding about whether peptides could reach brain tissue at all. The same review notes that DSIP was explicitly shown not to be Pappenheimer's Factor S, the separate somnogenic material recovered from goat cerebrospinal fluid, and that a radioimmunoassay for DSIP requiring eight of its nine residues was developed to track it.

DSIP carries the formula C35H48N10O15, an average mass of 848.8 g/mol and a monoisotopic mass of 848.33 Da, indexed in PubChem as CID 68816 under CAS 62568-57-4 and the nonproprietary name emideltide. Three of the nine residues are glycine, one is tryptophan, and none is cysteine.

The residue composition sets the practical analytical handles. A single tryptophan at position 1 carries the 280 nm chromophore, so ultraviolet quantitation is straightforward but sensitive to oxidation of that one residue. Two acidic side chains plus the C-terminal glutamate carboxylate against one free amine leave the molecule anionic at neutral pH, and the absence of lysine, arginine and histidine means no strong cation-exchange handle exists for purification.

Solution-state spectroscopy published in 1994 described DSIP as sitting in a dynamic equilibrium between unordered and folded states, with beta-turn conformations occupying roughly 40 percent of the population. Three glycines make that flexibility unsurprising. Phosphorylation is the one well-documented covalent modification: casein kinase II phosphorylates the peptide in cell-free assays, and serine 7 is the site carried by the analogue reported in the literature as P-DSIP.

DSIP has no identified receptor. A 2006 review in the Journal of Neurochemistry titled the peptide a still unresolved riddle precisely because neither a gene encoding a precursor nor a binding site was ever found, leaving open whether the reported activity belongs to the nonapeptide itself or to something co-purified with it in the original preparations.

Transport across the blood-brain barrier is better documented than receptor pharmacology. Passage of synthetic DSIP into rabbit brain was reported in 1977, a vascularly perfused guinea pig preparation in 1989 described uptake consistent with a high-affinity saturable carrier rather than free diffusion, and Banks records in 2025 that DSIP and its analogues crossed from blood into brain tissue and cerebrospinal fluid in intact form in rats and dogs, attributing the passage to lipid solubility.

Proposed downstream mechanisms remain descriptive. A 2026 review in the International Journal of Molecular Sciences summarizes the peptide as probably acting at multiple sites with influence over GABAergic and NMDA glutamatergic signalling. Separate rodent work on the phosphorylated analogue under simulated high altitude reported that naloxone abolished the observed changes, implicating opioid signalling. None of these accounts amounts to a defined molecular target.

DSIP electroencephalography in rodents produced its clearest numbers with the phosphorylated analogue. A 1989 study in unrestrained rats delivered a ten-hour nocturnal intracerebroventricular infusion and recorded slow-wave activity 22 percent above baseline and paradoxical states 81 percent above baseline at 0.5 nmol, driven mainly by a greater number of episodes rather than longer ones.

The same study reported a narrow effective window. Quantities were tested from 0.025 to 25 nmol, and amounts both larger and smaller than 0.5 nmol were inactive. Potency of the phosphorylated analogue was put at five times that of the unmodified nonapeptide under matched conditions. An inverted-U response of that shape is a recurring feature of the DSIP literature and complicates comparison between studies that sampled only one level.

The most recent animal work uses engineered constructs rather than the bare peptide. A 2024 Frontiers in Pharmacology study fused DSIP to a cell-penetrating carrier sequence, expressed it in Pichia pastoris, and tested it in Kun-Ming mice made insomniac with para-chlorophenylalanine. Daily wakefulness ran at 480 plus or minus 30 minutes in controls, 720 plus or minus 45 minutes in the model group, and 500 plus or minus 32 minutes with the fusion construct.

DSIP was carried into small controlled human studies during the 1980s and early 1990s, and the results were mixed. A double-blind trial published in Neuropsychobiology in 1992 enrolled 16 adults with chronic insomnia across five laboratory nights at 25 nmol/kg, measured higher efficiency and shorter latency against placebo, and concluded that the statistically significant effects were weak.

A 1987 study in European Neurology took a different design, following fourteen middle-aged adults with chronic insomnia across seven successive nights under double-blind, placebo-controlled conditions. That report described night rest and daytime function reaching the levels of normal controls, with alertness and daytime task scores rising. The contrast between a seven-night schedule and a three-night one is the most obvious variable separating the two outcomes.

Endocrine results did not transfer from rodents to humans. A 1995 Psychoneuroendocrinology study measured corticotropin and cortisol responses during and after infusion and found them almost identical to placebo, with the authors stating explicitly that their data do not support an inhibitory role for DSIP on corticotropin and cortisol secretion in man.

DSIP research extends well beyond electroencephalography, and most of it sits in rodent stress models. A 1985 Neuroendocrinology study reported that corticosterone rises driven by corticotropin-releasing factor fell significantly at 5 to 30 micrograms per kilogram, while corticosterone released by direct corticotropin stimulation was unchanged, placing the effect at the pituitary rather than the adrenal cortex.

Mitochondrial work from 2003 examined isolated rat brain mitochondria by polarographic oxygen measurement. DSIP raised state 3 respiration, the respiratory control ratio and the rate of ADP phosphorylation while uncoupled respiration stayed flat, and at 120 micrograms per kilogram before a hypoxic challenge it blocked the fall in mitochondrial respiratory activity that hypoxia otherwise caused. Rat hepatocyte studies published in 2016 reported catalase and superoxide dismutase changes that varied non-monotonically across 40, 120 and 360 micrograms per kilogram.

A 2021 focal stroke study in Sprague-Dawley rats is the most rigorous recent rodent report and is notable for what it did not find. Intranasal delivery at 120 micrograms per kilogram left infarct volume at day 21 statistically indistinguishable from vehicle, 20.9 plus or minus 6.9 percent against 24.1 plus or minus 4.6 percent, while rotarod coordination in the DSIP group returned by the seventh test day. A 2026 orthopaedic review reaches the same broad conclusion for the compound class, noting that preclinical reports outnumber clinical trials by a wide margin.

DSIP is supplied as a lyophilized solid and is characterized by the same two orthogonal checks used for any small synthetic peptide: reversed-phase HPLC for purity, commonly specified at 98 percent or better in reference-standard documentation, and mass spectrometry for identity against an expected protonated ion near m/z 849.34 monoisotopic or 849.8 on an average-mass readout.

Concentration is easy to verify because the molecule carries exactly one tryptophan and no tyrosine or cysteine. Absorbance at 280 nm therefore maps to a molar absorptivity close to 5,500 M-1 cm-1 by standard residue additivity, and any drift in that figure across a stored batch points at oxidation of the single indole side chain rather than at bulk loss of peptide.

Stability in biological matrices is poor and well documented. A 1987 Peptides study following labelled material in human and rat blood found rapid breakdown with release of a species eluting at the retention time of free tryptophan, and a 2026 review reports peptidase-driven clearance from cerebrospinal fluid on a timescale of minutes. Chemical degradation in storage runs on a slower clock: the aspartyl-alanine motif at positions 5 and 6 is a classic isoaspartate-forming site, and isoaspartyl DSIP is used in the enzymology literature as a defined substrate for protein L-isoaspartyl methyltransferase, with a reported Michaelis constant of 33.36 micromolar for the yeast enzyme.

Storage practice for material of this class follows the chemistry rather than any published protocol specific to the nonapeptide: sealed and desiccated at minus 20 degrees Celsius or colder for the dry solid, protected from light because of the tryptophan chromophore, and worked up in aqueous buffer near neutral pH where the net negative charge keeps the peptide soluble. Material described here is offered for laboratory research use only. It is not a medicine, not a food, and not intended for diagnostic use or for use in humans or animals.