SS-31 (Elamipretide): Cardiolipin Binding and Mitochondrial Membrane Research

SS-31 is a four-residue peptide in which cationic and aromatic side chains alternate: D-arginine, 2',6'-dimethyltyrosine, lysine and phenylalanine, capped as a C-terminal primary amide. PubChem records the molecular formula as C32H49N9O5 with a molecular weight of 639.8 g/mol and a monoisotopic mass of 639.3857 Da under CID 11764719.

Two design choices separate SS-31 from a conventional all-L peptide. The arginine at position 1 is in the D configuration, a substitution the Szeto-Schiller series was built around because non-natural stereochemistry at the N-terminus resists aminopeptidase cleavage. Position 2 carries 2',6'-dimethyltyrosine rather than tyrosine, adding steric bulk and aromatic character to the phenol ring.

Charge distribution follows from that architecture. The free N-terminal amine, the guanidinium of D-arginine and the lysine side chain each carry a positive charge at neutral pH, while the C-terminal amide removes the anionic carboxylate that a free acid would contribute. Mitchell and colleagues describe the resulting molecule as carrying a formal charge of +3 at neutral pH. PubChem computes an XLogP of 0 and a topological polar surface area of 267 Ų, values consistent with a strongly polar, non-lipophilic compound that nevertheless partitions onto membranes.

SS-31 binds anionic lipid bilayers, with a marked preference for cardiolipin. In model membranes containing 20 mol% tetraoleoyl-cardiolipin, isothermal titration calorimetry returned dissociation constants of 2.0-2.9 µM, compared with 6.0-13 µM for bilayers containing the monoanionic lipid phosphatidylglycerol at the same mole fraction.

Partition coefficients from the same 2020 study separate the two lipid classes further: roughly 5.07 × 10⁴ for cardiolipin-containing vesicles and 5.90 × 10⁴ for monolyso-cardiolipin, against 1.28 × 10⁴ for phosphatidylglycerol. Enthalpy changes were small (about −4.2 kJ mol⁻¹ for cardiolipin), while the entropic term dominated at roughly −22.7 kJ mol⁻¹, indicating an association driven mainly by entropy rather than by strong enthalpic contacts.

The consequence measured at the bilayer surface is electrostatic. Zeta potential measurements showed a saturable reduction in the magnitude of surface charge as peptide was added, plateauing near −30 mV, so the membrane remained net negative even at binding saturation. Molecular dynamics and NMR placed the basic side chains at the interfacial region tethered to headgroup phosphates, with aromatic rings only partially buried, and small-angle X-ray scattering detected no major topological alteration of the bilayer even at the highest peptide-to-lipid ratios tested.

Cardiolipin makes this selectivity biologically specific. The same source notes that cardiolipin comprises roughly 10-20 mol% of total lipid in the inner mitochondrial membrane and is largely confined to that compartment. Exogenously added SS peptides were reported to cross the plasma membrane by a nonsaturable route that does not require metabolic input, then to accumulate 1000-5000 fold at the inner mitochondrial membrane.

SS-31 has been examined for effects on cristae architecture, the folded inner-membrane geometry that houses respiratory complexes. In a rat cardiac ischemia-reperfusion model, serial block-face scanning electron microscopy with 3D reconstruction showed that the fragmentation of cristae networks seen after reperfusion was reduced when elamipretide was present, while respirometry in permeabilized ventricular fibers showed alleviation of the decrements in complex I, II and IV activity.

That structural result came with an informative negative. Mass spectrometry in the same work showed elamipretide did not restore cardiolipin concentration after ischemia-reperfusion, which argues the peptide acts on how cardiolipin is organized rather than on how much of it is present. In biomimetic membranes the peptide aggregated cardiolipin, consistent with a clustering effect at the lipid level.

Earlier mechanistic work identified a second consequence of the peptide-lipid complex. Cytochrome c bound to cardiolipin gains peroxidase activity, which oxidizes cardiolipin and initiates release of cytochrome c from the membrane. Birk and colleagues reported that the SS-31 and cardiolipin complex inhibited that peroxidase activity while leaving the electron-carrier function of cytochrome c intact, and that ATP-dependent cellular functions returned faster after ischemic challenge in the same system.

Genetic models of cardiolipin remodeling have been used the same way. In tafazzin-knockdown mice, cardiac mitochondria showed abnormal ultrastructural membrane morphology, vacuole accumulation, pro-fission conditions and defective mitophagy, and SS-31 exposure was reported to restore mitochondrial structure and mitophagy markers in the 2024 Scientific Reports study.

SS-31 interactors were mapped directly by Chavez and colleagues using chemical cross-linking coupled to mass spectrometry, published in PNAS in 2020. Every protein identified as an SS-31 binder was an established cardiolipin binder, and the set clustered into two functional groups: components of the oxidative phosphorylation pathway, and enzymes of 2-oxoglutarate metabolism and signaling.

Cross-linked residues mapped to regions frequently adjacent to known cardiolipin-protein interfaces, which links the proteomic result back to the lipid-binding result. The picture that emerges from the cross-linking data is of a peptide that concentrates at cardiolipin-rich patches and therefore ends up in proximity to the protein machinery those patches organize.

A 2023 GeroScience study extended this to nucleotide transport. Elamipretide was reported to bind the adenine nucleotide translocator and ATP synthase directly, and in mitochondria isolated from old skeletal muscle it increased ADP sensitivity by increasing ADP uptake through the translocator. Abundance of ADP and ATP pathway proteins did not change; what changed was post-translational, with decreased protein S-glutathionylation including of the translocator itself.

SS-31 appears most often in recent cell work as a probe of stress-induced senescence and of mitochondrial function under insult, typically at a 1 µM working concentration. Two 2026 cardiomyoblast studies from overlapping groups report quantified senescence-marker changes in the H9C2 line, and a third 2026 study examines the peptide against α-synuclein membrane binding.

In the radiation study published in the Journal of Radiation Research, γ-irradiation at 2, 5 and 10 Gy inhibited H9C2 proliferation and induced senescence-associated β-galactosidase staining. SS-31 at 1 µM over 7 days lowered that staining from 67% to 38% in cells irradiated at 10 Gy, blunted increases in the senescence markers p16 and p21 in both H9C2 cells and human iPSC-derived cardiomyocytes, lowered the secretory markers TNF-α, IL-6 and IL-1β, and reversed a radiation-driven rise in mitochondrial respiration.

The doxorubicin study in Biology used a 3 hour exposure to 50 nM doxorubicin followed by 3 days of culture, a protocol that produced roughly 50% β-galactosidase staining and halted proliferation. SS-31 at 1 µM lowered staining from 51.4% to 35.8% and blunted the p16, p21 and secretory-phenotype increases, but did not reverse the cell-cycle arrest, a dissociation the authors report explicitly.

The 2026 Chemical Biology and Drug Design report moved the peptide into a protein-aggregation system. Fluorescence correlation spectroscopy and fluorescence anisotropy showed SS-31 displacing both wild-type and N-terminally acetylated α-synuclein from negatively charged small unilamellar vesicles in a concentration-dependent manner; thioflavin-T assays and transmission electron microscopy showed reduced membrane-induced aggregation and altered fibril morphology; and Seahorse Mito Stress Test measurements showed restoration of respiratory parameters in neuroblastoma cells exposed to α-synuclein oligomers.

Rodent work from 2025 supplies the matching in-vivo numbers. In embryonic rat spinal cord neuron cultures challenged with 125 nM rotenone, SS-31 lowered lactate dehydrogenase release from about 60% above control to 9% above control; with 100 µM glutamate, the caspase-3/7 signal fell from a 2.6-fold increase to a 0.16-fold increase, and neurite length loss of roughly 80% was limited to 4.82% at 100 µM peptide.

SS-31 reaches mitochondria by a different route than the lipophilic cations that dominated earlier mitochondria-targeted chemistry. Triphenylphosphonium-conjugated molecules accumulate inside the matrix in proportion to the negative-inside membrane potential, so their localization tracks the potential itself. SS-31 instead binds cardiolipin directly, and it does so in protein-free liposomes that have no membrane potential at all.

That distinction matters for interpreting experiments in depolarized or damaged mitochondria, where a potential-driven probe loses its targeting exactly when the system is most perturbed. The liposome measurements described above were made on pure lipid vesicles, and binding curves for yeast mitochondria were reported to closely resemble the vesicle curves, which supports charge-driven lipid association as the governing interaction.

The framing of the molecule in the literature has also shifted. The 2025 review in Biomedicine and Pharmacotherapy notes that initial descriptions of the mechanism emphasized reactive oxygen species scavenging, and that the last decade of work has redirected attention toward modulation of mitochondrial membrane electrostatic potentials and the assembly of cardiolipin-dependent protein complexes. Research reports frame the peptide as an interfacial membrane binder rather than as a bulk antioxidant.

SS-31 is supplied as a lyophilized powder, and its computed physicochemical profile matches that format: an XLogP of 0, a topological polar surface area of 267 Ų and a formal charge of +3 at neutral pH describe a highly polar salt-like solid rather than a lipophilic compound. Published methods sections prepare working solutions in aqueous buffer or culture medium.

Working concentrations in the literature span a narrow band by model type. Recent cardiomyoblast and iPSC-derived cardiomyocyte work used 1 µM; primary neuron cultures used up to 100 µM; biophysical binding assays operated in the 2-3 µM range set by the measured dissociation constants, in buffers containing large unilamellar vesicles at defined lipid mole fractions. Assays under calcium stress in the 2020 biophysics work went up to 5 mM calcium chloride.

Analytical confirmation follows standard peptide practice: identity by electrospray mass spectrometry against a monoisotopic mass of 639.3857 Da, giving a singly protonated ion near m/z 640.4, and purity by reversed-phase HPLC reported on the lot certificate of analysis. Lyophilized peptide material is normally kept frozen, protected from light and moisture, with aqueous stock solutions aliquoted so that repeated freeze-thaw cycles are avoided.

This material is offered for laboratory research use only. It is not a drug, not a food, and not intended for human or veterinary use.