pe
pep-04425 v1 CC-BY-SA-4.0

Gramicidin D: natural antibiotic for skin and eye infections (Neosporin ingredient)

A natural antibiotic from soil bacteria that kills gram-positive bacteria by punching holes in their membranes; FDA-approved for topical use in products like Neosporin since 1955, not safe to inject or swallow.

statusbioassayed targetANTIMICROBIAL length15 aa refs3
fda-approved
status 5 / 5 · 2 contributors
prediction metrics boltz-2 2.2.1
ipTM0.000
pTM0.235
avg pLDDT94.6
ranking score0.804
STRUCTURE · PEP-04425 × ANTIMICROBIAL
ranking0.804
?
RECEPTOR UNKNOWN
peptide conformation only · no target structure
target interface 4.5Å peptide drag rotate · ctrl+scroll zoom · right-click pan
sequence15 aa
151015
VGALAVVVWLWLWLW
in the news 6 articles
overview readme

What this is

Gramicidin D is a naturally occurring linear antibiotic peptide isolated from the soil bacterium Brevibacillus brevis (formerly Bacillus brevis). It is not a single molecular species but a mixture of three closely related 15-amino acid linear peptides — gramicidin A (~85%), gramicidin B (~5%), and gramicidin C (~10%) — that differ only at position 11 (Trp in A; Phe in B; Tyr in C). Each variant inserts into bacterial lipid bilayers and creates a single-file cation channel that collapses gram-positive bacterial membrane potential. Gramicidin D is classified as a membrane-active antimicrobial peptide (AMP) with FDA-approved topical use; it has no approved systemic indication. It is distinct from the cyclic decapeptide gramicidin S, a separate compound from the same organism with a different mechanism.

The stored sequence VGALAVVVWLWLWLW (15 aa) represents the gramicidin A (valine) variant, the dominant (~85%) component of the mixture. The actual molecule contains D-amino acids at alternating positions — positions 2, 4, 6, 8, 10, 12, and 14 are D-form residues essential for the β6.3-helical channel geometry — and the N-terminus is formylated while the C-terminus carries ethanolamine; neither the D-amino acid stereochemistry nor these terminal modifications are captured in the standard one-letter sequence. Gramicidin B and C differ from A only at position 11 (Phe or Tyr instead of Trp).


History

The gramicidin story begins with soil-bacterium-derived antibiotic discovery. In 1939, René Dubos at Rockefeller University demonstrated that Bacillus brevis produced a bactericidal substance active against gram-positive pathogens — among the first demonstrations that soil microorganisms could be exploited as antibiotic sources (Dubos 1939). Dubos named the crude fraction "tyrothricin," which Rollin Hotchkiss and Dubos subsequently separated into two components: tyrocidin (a cyclic peptide) and gramicidin (the linear fraction). The linear fraction — later designated gramicidin D to distinguish it from the independently named gramicidin S — was notable because it was the first antibiotic peptide whose full amino-acid sequence was determined, an achievement by Sarges and Witkop in 1965.

Structural insights followed over the next three decades. Daniel Urry (1971) proposed the head-to-head β6.3-helical (πLD) model for gramicidin A channel formation — a dimer with alternating L and D residues winding a right-handed helix just wide enough for single-file water and ion passage. This model was confirmed by solid-state NMR and X-ray crystallography in the 1990s and became a cornerstone model system for studying membrane protein channel biophysics.

Clinically, gramicidin D was licensed by the FDA in 1955 as a component of Neosporin® (with neomycin and polymyxin B), one of the earliest multi-antibiotic topical formulations for infected conjunctiva and skin wounds (Alzain 2025). Neosporin and similar formulations remain among the most widely used over-the-counter antibiotics globally. Gramicidin D's clinical trajectory has been limited by its toxicity profile at systemic doses, which renders it unsuitable for anything beyond topical application; research interest now focuses primarily on gramicidin analogs with improved selectivity windows and reduced hemolytic activity.


What it does

Gramicidin D's mechanism is unusually well characterized at the atomic level relative to most antimicrobial peptides.

Channel formation: Individual gramicidin monomers insert into one leaflet of the bacterial phospholipid bilayer in a β6.3-helical conformation. Two monomers from opposite bilayer leaflets associate at their formylated N-termini, forming a head-to-head dimer channel that fully spans the hydrophobic core of the membrane. The interior of the channel is approximately 4 Å in diameter — just wide enough for a single file of water molecules and unhydrated monovalent cations.

Cation conductance: The gramicidin channel is highly selective for monovalent cations (Na⁺, K⁺, Rb⁺, Cs⁺, H⁺) and essentially impermeable to divalent cations and anions. In bacteria that maintain Na⁺ and K⁺ gradients across their membranes for metabolic energy, gramicidin channel insertion collapses the transmembrane ion gradient, depolarizes the membrane potential, and disrupts ATP synthesis. The resulting energy failure and ionic dysregulation is bactericidal.

Selectivity for gram-positive bacteria: Gram-positive organisms (e.g., Staphylococcus aureus, Streptococcus pyogenes, Bacillus subtilis) have a single lipid bilayer accessible to hydrophobic channel-forming peptides. Gram-negative bacteria have an outer membrane with lipopolysaccharide that sterically and electrostatically excludes gramicidin D in most conditions, explaining the compound's predominantly gram-positive activity spectrum (Alzain 2025).

Mammalian toxicity: The same channel-forming mechanism that kills bacteria also permeabilizes red blood cell membranes at sufficiently high local concentrations, causing hemolysis. This hemolytic activity is the primary reason gramicidin D is confined to topical use; at the low concentrations achievable in topical formulations, local antimicrobial effect is achieved without systemic exposure sufficient to cause hemolysis (Yang 2025).

Anti-HIV activity: Gramicidin D has demonstrated activity against HIV in vitro as a component of certain topical microbicide formulations. The mechanism is proposed to involve disruption of the viral lipid envelope, though this application remains investigational.


Evidence

  • Human: Gramicidin D as a component of Neosporin® (with neomycin and polymyxin B) has demonstrated efficacy for superficial ocular infections (bacterial conjunctivitis, keratitis) and wound infections, with more than 70 years of post-licensure use since FDA approval in 1955. Approximately 9 registered studies involving gramicidin appear on ClinicalTrials.gov, reflecting investigation of new formulations and combination products rather than de novo efficacy demonstration for the established topical indication (Alzain 2025). Post-licensure surveillance has confirmed the absence of significant systemic absorption at topical doses.
  • In vitro: The β6.3 head-to-head helical dimer model, originally proposed by Urry (1971), was validated by solid-state ¹³C NMR and crystal structures of gramicidin A in lipid bilayers and organic solvents. The gramicidin channel has become one of the most thoroughly characterized model systems in membrane biophysics, with the key ion selectivity sequence (Cs⁺ > Rb⁺ > K⁺ > Na⁺ > Li⁺) arising from differential dehydration energies at the channel mouth (Dubos 1939). Multiple in vitro studies confirm that hemolytic activity at supraphysiological concentrations results from the same channel-forming mechanism; minimum inhibitory concentrations against S. aureus in topical conditions and hemolytic doses differ by orders of magnitude, establishing the therapeutic window that makes topical use safe (Yang 2025).

Myths and misconceptions

  • "Gramicidin D and gramicidin S are the same compound." They are not. Gramicidin D (the Dubos 1939 linear fraction) and gramicidin S (a cyclic decapeptide) are structurally and mechanistically distinct. Both are produced by B. brevis, but they were discovered independently and have different spectra and mechanisms. Gramicidin D forms cation channels; gramicidin S disrupts membranes via a different cyclic peptide mechanism. They are sometimes conflated in antimicrobial peptide reviews, but their pharmacology is not interchangeable.
  • "The stored 15-residue sequence is a single pure peptide." The stored sequence VGALAVVVWLWLWLW represents gramicidin A (the dominant ~85% component). Commercial gramicidin D is always a mixture of A, B, and C variants; all three are biologically active and contribute to the preparation's total antibacterial effect. Any synthetic version based solely on the gramicidin A sequence would not replicate the full pharmacology of the natural mixture.
  • "Resistance to gramicidin D is emerging as a clinical concern." Unlike broad-spectrum antibiotics, gramicidin D's channel-forming mechanism is tied to fundamental membrane physics rather than specific enzymatic targets, making resistance development much harder. No clinically significant resistance emergence has been documented over seven decades of widespread topical use, in contrast to resistance rates seen with systemic beta-lactams or fluoroquinolones.

Common questions

Q: Why is gramicidin D only used topically if it's bactericidal? A: Gramicidin D's cation-channel mechanism is non-selective at higher concentrations — the same physical disruption that kills bacteria also lyses mammalian red blood cells (hemolysis). At typical topical concentrations in eye drops or ointments, the drug achieves effective local antimicrobial activity without significant systemic absorption or exposure sufficient to cause hemolysis. If administered systemically at doses needed to treat deeper infections, it would cause intravascular hemolysis. This is a fundamental property of membrane-active AMPs that form channels in eukaryotic membranes above a threshold concentration (Yang 2025).

Q: What is the difference between gramicidin D and tyrothricin? A: Tyrothricin is the crude antibiotic fraction isolated by Dubos from B. brevis culture — it contains both gramicidin (the linear peptide fraction, later called gramicidin D) and tyrocidin (a cyclic decapeptide with broader spectrum and higher hemolytic toxicity). Gramicidin D is the purified linear-peptide sub-fraction of tyrothricin. Most modern pharmaceutical preparations specify gramicidin D specifically (not tyrothricin) to ensure a defined composition (Dubos 1939).

Q: Does gramicidin D work against MRSA? A: Gramicidin D retains activity against methicillin-resistant Staphylococcus aureus (MRSA) in topical conditions, because methicillin resistance (mediated by the mecA gene encoding an altered penicillin-binding protein) does not affect membrane ion-channel vulnerability. The drug's activity is independent of beta-lactam targets. This makes it valuable in topical combinations for superficial MRSA skin and eye infections, though its hemolytic properties preclude systemic use.


Known effects

  • Bactericidal against gram-positive pathogens — FDA-approved (topical, since 1955; Neosporin® combination)
  • Ocular and wound infection treatment — established clinical use in topical formulations
  • Hemolysis at systemic concentrations — well-characterized in vitro and in vivo; limits use to topical routes
  • Membrane depolarization in bacteria — mechanistic; confirmed by electrophysiology

Safety signals

Gramicidin D has a well-established safety profile at approved topical doses, with more than 70 years of use in Neosporin® and comparable formulations. Hemolysis is the primary safety concern and is concentration-dependent: the compound is confined to topical application precisely because the systemic concentrations required for bactericidal effect against deep infections would cause intravascular hemolysis. At topical doses in ophthalmic or dermatological use, systemic absorption is not considered significant and hemolysis has not been a clinical concern at those exposure levels. Allergic contact dermatitis has been reported with multi-ingredient Neosporin formulations; attribution to individual components (gramicidin D vs. neomycin vs. polymyxin B) requires clinical evaluation. The compound is cytotoxic to human cells at high local concentrations, consistent with its non-selective membrane-channel mechanism (Alzain 2025; Yang 2025).


Regulatory status

  • US: FDA-approved as a component in topical antibiotic formulations (Neosporin®) since 1955. Available over-the-counter for skin and eye infections. Not approved for systemic use.
  • WADA: No prohibition listed for gramicidin D in its current topical-antibiotic indication.

Mechanism

Gramicidin D adopts a β6.3-helical (πLD) conformation in lipid bilayers, with alternating L and D residues producing a right-handed helix. The tryptophan residues at the C-terminus (positions 11, 13, 15 in gramicidin A) anchor the molecule at the bilayer–water interface via cation–π and hydrogen-bonding interactions, orienting the helix perpendicular to the membrane plane. Two monomers from opposing leaflets dimerize via formyl-group N-terminal contacts to form a 26-residue transmembrane pore. The single-file channel (internal diameter ~4 Å) conducts monovalent cations with selectivity Cs⁺ > Rb⁺ > K⁺ > Na⁺ > Li⁺ — reflecting differential dehydration enthalpies at the channel entry — while excluding divalent cations and anions. In gram-positive bacteria, insertion of gramicidin channels into the cytoplasmic membrane collapses transmembrane ion gradients, depolarizes membrane potential (as confirmed by DiSC3(5) fluorescence assays used in experimental comparisons), and terminates ATP synthesis. The D-amino acids at alternating positions are not merely structural curiosities: they are required for the geometry that allows head-to-head dimerization across the bilayer, and any L-for-D substitution abolishes channel-forming activity (Alzain 2025; Yang 2025).


Related peptides

  • Polymyxin B — co-formulated with gramicidin D in Neosporin®; targets gram-negative outer membranes by a distinct lipopolysaccharide-disruption mechanism, complementing gramicidin D's gram-positive spectrum
  • Gramicidin S — a cyclic decapeptide produced by the same organism (B. brevis), discovered independently; disrupts membranes via amphipathic β-sheet mechanism rather than channel formation; historically grouped with gramicidin D by name despite unrelated structure and pharmacology
Hypotheses4 directions▾ collapse

Research directions for this peptide, selected from the current sources — hypotheses you can explore and model. None of it is proven yet; tap any one to see the full thinking.

openupdated 2026-06-05

Could a pill form of gramicidin A kill the dangerous gut bug C. difficile without getting absorbed into the body and causing harm?

C. difficile causes severe, recurring hospital-acquired diarrhea and is growing harder to treat. If this approach works, it could give doctors a bactericidal option that stays inside the gut, where it is needed, while avoiding the blood toxicity that has always ruled gramicidin out as an internal medicine.

The hypothesis
Gramicidin A is active against Clostridioides difficile (a gram-positive, spore-forming anaerobe) at clinically achievable intraluminal concentrations when delivered as a non-absorbed oral formulation, because the colonic mucosa provides a natural absorption barrier that recreates the same concentration-confinement rationale that makes topical use safe.
Why it’s plausible
The readme explains gramicidin D's systemic exclusion is purely pharmacokinetic (hemolysis at high systemic exposure), not a target-based limitation. C. difficile is gram-positive with a single accessible lipid bilayer, matching gramicidin D's spectrum. The colon maintains a physical barrier analogous to skin/cornea - a luminal antibiotic formulation (e.g., poorly absorbed lipid suspension, similar to oral vancomycin for C. diff) would keep gramicidin concentrated at the mucosal surface while limiting absorption. No resistance has emerged over 70 years of topical use, a major advantage over fidaxomicin and vancomycin against which C. diff resistance is emerging.
Why it matters
C. difficile infection is a major nosocomial problem with few non-absorbable bactericidal options and growing resistance to existing treatments. A repurposing strategy exploiting gramicidin D's gram-positive spectrum and 70-year resistance-proof track record could address a high-unmet-need indication without requiring new resistance-mechanism studies.
Plausibility.53
Novelty.50
Impact.62
Basis · grounding1 paper · 2 computed/notes
[1]
noteNo clinically significant resistance emergence documented over seven decades of widespread topical use; gram-positive spectrum; bactericidal by membrane depolarization.
[2]
noteSystemic toxicity (hemolysis) is the only barrier to non-topical use; topical use safe because local antimicrobial effect achieved without systemic exposure.
[3]
paper
AMPs exhibit low MICs against bacterial strains with few treatment choices, making them a promising target for novel therapeutics.
doi: 10.2147/idr.s514825
openupdated 2026-06-05

When the mouth turns acidic from sugar, does gramicidin kill bacteria by letting their stored acid energy leak out, rather than by the salt disruption most scientists assumed?

If proton leakage turns out to be the main killing mechanism in acidic conditions, it could explain why gramicidin keeps working where other antibiotics fail, and could open a path to using it in mouthwashes or wound gels targeting the organisms behind tooth decay and infected wounds.

The hypothesis
The proton (H+) conductance of the gramicidin A channel, rather than Na+/K+ collapse alone, is the primary driver of bactericidal activity against acid-tolerant gram-positive pathogens such as Streptococcus mutans and Lactobacillus species, because these organisms rely heavily on proton motive force maintenance even when Na+/K+ gradients are partially buffered.
Why it’s plausible
The readme notes the gramicidin channel conducts monovalent cations including H+, and that bactericidal activity arises from collapsing transmembrane ion gradients. For acid-tolerant gram-positive organisms that maintain tight proton motive force control, proton conductance through the channel would be thermodynamically the most damaging flux given the large inward H+ gradient in acidic oral/wound environments. This has not been disentangled from bulk cation collapse in the existing literature, which focuses on Na+/K+ gradients.
Why it matters
If proton conductance is the dominant lethal flux in specific pathogens, this would explain why gramicidin D remains active in acidic wound or oral environments and would open a rationale for deploying it against oral biofilm pathogens (dental caries organisms) where the pH gradient is large and other antibiotics fail.
Plausibility.52
Novelty.35
Impact.47
Basis · grounding1 paper · 1 computed/note
[1]
noteChannel is selective for monovalent cations including H+, Na+, K+; bactericidal effect attributed to collapse of transmembrane ion gradient and disruption of ATP synthesis.
[2]
paper
Gramicidin D is a linear peptide mixture active against gram-positive bacteria; mechanisms involve membrane disruption.
doi: 10.2147/idr.s514825
openupdated 2026-06-05

Could gramicidin slip into the power stations of cancer cells more easily than into normal red blood cells, giving it a window where it is lethal to tumors but not yet toxic to the body?

If cancer cell mitochondria are genuinely more vulnerable to gramicidin than red blood cells are, it would mean a dose range exists where the peptide disrupts tumors while leaving healthy cells largely intact. That selectivity window, if confirmed, could make gramicidin a starting point for a new class of mitochondria-targeting cancer therapies.

The hypothesis
The gramicidin A dimer channel, when inserted into mitochondrial inner membranes of cancer cells with elevated mitochondrial membrane potential, selectively collapses that potential at lower local peptide concentrations than required to lyse erythrocytes, because the mitochondrial inner membrane is enriched in cardiolipin (a negatively charged phospholipid absent in RBCs) that accelerates gramicidin insertion kinetics.
Why it’s plausible
Erythrocytes have no mitochondria and rely on their plasma membrane to resist gramicidin-induced hemolysis. Cancer cells with hyperpolarized mitochondria (a hallmark enabling selective targeting by mitochondria-targeting cations) have cardiolipin-rich inner membranes. Cardiolipin's negative charge would lower the electrostatic insertion barrier for gramicidin more than the RBC plasma membrane's predominantly zwitterionic composition. The same sequence (VGALAVVVWLWLWLW) with its tryptophan anchors and hydrophobic core would insert into any accessible bilayer; the selectivity question is which membrane is encountered and with what kinetics. Mitochondrial outer membrane permeabilization is a known apoptotic trigger in cancer cells.
Why it matters
If gramicidin A preferentially collapses mitochondrial potential in cancer cells over erythrocyte plasma membranes due to cardiolipin-driven insertion kinetics, it defines a mechanistic selectivity window for an anticancer application that the existing topical antimicrobial framing has overlooked, and connects to the broader field of mitochondria-targeting peptides.
Plausibility.42
Novelty.47
Impact.53
Basis · grounding3 computed/notes
[1]
sequenceVGALAVVVWLWLWLW - highly hydrophobic sequence with tryptophan anchors; will insert into any accessible bilayer with hydrophobic core.
[2]
noteChannel-forming mechanism is non-selective at sufficient concentrations and operates on any lipid bilayer accessible to the peptide; hemolysis occurs when RBC membranes are reached at high local concentrations.
[3]
noteGramicidin D mechanism: membrane depolarization via cation channel collapse of transmembrane ion gradient and ATP synthesis disruption.
openupdated 2026-06-05

Does gramicidin attack bacterial membranes first mainly because they carry a stronger negative charge, rather than because the peptide changes shape to fit?

If charge is the gating factor, formulators could tune how selectively gramicidin hits bacteria versus human cells by adjusting the acidity of a cream or gel, or by pairing it with agents that alter bacterial surface charge. That would give a rational, predictable way to make gramicidin-based products safer or more potent for specific uses.

The hypothesis
The high avg_plddt of 94.6 for the gramicidin A monomer reflects the intrinsic rigidity of the alternating D/L beta-helical backbone, and this structural pre-organization means the membrane insertion barrier is primarily electrostatic (not conformational), so that gramicidin A inserts faster into negatively charged bacterial membranes (high phosphatidylglycerol content) than into zwitterionic mammalian membranes (high phosphatidylcholine/sphingomyelin content), quantitatively explaining its therapeutic window.
Why it’s plausible
The boltz-2 monomer prediction gives avg_plddt of 94.6, indicating high structural confidence for the monomer fold. However, ptm is only 0.23, suggesting the terminal regions that mediate dimerization and membrane anchoring are less well-defined in solution. The high local confidence combined with the known rigid beta6.3-helical structure in membranes implies the peptide arrives at the membrane in a near-final conformation. The rate-limiting step for insertion would then be overcoming the electrostatic barrier of the headgroup layer, which is lower for negatively charged PG-rich bacterial membranes than for PC-rich mammalian membranes. This mechanistic framing is more precise than the existing explanation (which cites gram-negative outer membrane exclusion and hemolysis separately).
Why it matters
If the electrostatic insertion rate is the primary selectivity determinant, then surface-charge-modifying formulation strategies (e.g., acidic pH topical vehicles) or bacterial membrane-charge-altering combination partners could predictably tune gramicidin D's therapeutic window, providing a rational path to improving its selectivity index.
Plausibility.38
Novelty.33
Impact.48
Basis · grounding3 computed/notes
[1]
structureboltz-2 monomer: avg_plddt 94.6 (high backbone confidence), ptm 0.23 (lower terminal confidence), iptm null (no interface modeled, consistent with membrane-embedded dimer not captured in solution monomer prediction).
[2]
noteAlternating D-amino acids at positions 2,4,6,8,10,12,14 essential for beta6.3-helical channel geometry; N-terminus formylated, C-terminus ethanolamine - pre-organized structure.
[3]
noteMammalian toxicity: same channel-forming mechanism lyses RBC at sufficiently high local concentrations; MIC vs hemolytic doses differ by orders of magnitude.
details expand to inspect
full evidence table1 metrics
metricvaluetool
ranking score 0.8040136694908142 boltz-2
3-letter notation
Val-Gly-Ala-Leu-Ala-Val-Val-Val-Trp-Leu-Trp-Leu-Trp-Leu-Trp
recipeboltz-2 2.2.1
parametervalue
modelboltz-2 2.2.1
weights
hardwarevast_v100_32gb
mlx version
python
random seed1
msa strategynone_monomer
runtime
predicted by
predicted at2026-05-23
citationbibtex
peptidemodel (2026). Gramicidin D: natural antibiotic for skin and eye infections (Neosporin ingredient) (pep-04425, v1). PeptideModel. https://peptidemodel.com/card/pep-04425
@peptide{pep04425,
  sequence = {VGALAVVVWLWLWLW},
  target   = {antimicrobial},
  author   = {peptidemodel},
  year     = {2026},
  status   = {bioassayed}
}
clinical trials 8 on ct.gov · checked 2026-05-22
ct.gov trials 8
by phase
3phase 11phase 22phase 32no phase
by status
2completed2recruiting1withdrawn3unknown
references 3 papers
discussion no comments
sign in to comment
peptidemodel.com CC-BY-SA-4.0 research only · not for human use