As of 6 October 2026, four papers first published 24–29 September 2026 infer antimicrobial-peptide (AMP) selectivity from mismatched membrane and host-cell readouts. Microbial evidence in this set is in-vitro killing, permeabilization, biofilm work, or a prior antimicrobial claim paired with new bilayer imaging. Host-side evidence is hemolysis, unnamed cytotoxicity, or a cancer-versus-healthy contrast. The abstracts generally omit media, inoculum, and lipid composition, so the numbers cannot be ranked across studies.
Microbial killing is not a single membrane assay
The designed cationic peptide RK-14 (RIWKIWKKIWKIWK-NH₂) was reported as bactericidal against methicillin-resistant Staphylococcus aureus and ESBL-positive Escherichia coli, with MICs of 16–64 µM and MBCs of 32–128 µM (PubMed 42809173). Membrane disruption was supported by membrane-binding affinity, increased propidium iodide uptake, and fluorescence microscopy showing fewer cells and morphological changes. The same peptide showed concentration-dependent antibiofilm activity (80–99% inhibition) and 87–91% eradication of mature catheter-associated biofilms at 1 × MIC, with reductions in luxS, csgD, icaD, and spa transcript levels relative to untreated controls. Gene-expression changes sit downstream of growth and membrane stress; they do not by themselves prove a pore mechanism.
Peptide-enriched seed fractions from Capsicum frutescens were tested against Candida rather than bacteria (PubMed 42803916). At 200 µg mL⁻¹, fractions F2 and F3 inhibited C. albicans growth by 83% and 69% and C. parapsilosis by 71% and 57%. Both reduced viability and increased membrane permeabilization. Further analyses associated F2 with cell-wall integrity changes and F3 with mitochondrial dysfunction. Partial sequences resembled plant defense peptides (heveins, defensins, vicilins). Because the dose is given in mass concentration against yeasts, it is not numerically comparable to RK-14 micromolar MICs against MDR bacteria.
A review of gramicidin S (GS) analogues treats bacterial-membrane disruption as the primary mechanism, with rapid bactericidal activity and limited reported resistance, while noting that GS has a long clinical history (PubMed 42788235). That article is a review, not a new matched MIC series.
Mastoparan-1 (MP-1) from Polybia paulista is described as possessing antimicrobial properties, but the assays reported in this abstract are model membranes and mammalian cells (PubMed 42800250). Atomic force microscopy (AFM) showed that perturbation morphology can change substantially after single-point mutations. AFM morphology is a physical membrane readout; it is not interchangeable with MIC, MBC, or propidium iodide uptake.
Hemolysis, unnamed cytotoxicity, and PS-rich models answer different questions
RK-14 was described as showing low cytotoxicity at effective concentrations, with hemolysis increasing at higher concentrations, and a favorable preliminary selectivity profile. The abstract does not name the mammalian cell type, the hemolysis species, or the concentration at which hemolysis rose (PubMed 42809173).
The Capsicum fractions were assessed for hemotoxicity by in-vitro sheep erythrocyte lysis and were reported to have low hemolytic activity, interpreted as selectivity toward fungal cells (PubMed 42803916). Sheep red-cell lysis does not measure nucleated-cell membrane repair, serum binding, or tissue exposure.
The GS review identifies cytotoxicity as the constraint on broader application and summarizes analogue strategies—β-turn modification, non-proteinogenic amino acids, and machine learning–guided design—aimed at improving selectivity by tuning cationic charge, hydrophobicity, amphipathicity, and conformational flexibility (PubMed 42788235). No new matched host-cell dataset is reported in the abstract.
MP-1 is the only paper here that pairs model membranes with cell-based assays for a host-versus-host contrast: enhanced activity against a number of cancer cell lines relative to healthy cells, with the authors judging it likely that outer-leaflet phosphatidylserine (PS) contributes. Mutant screening identified derivatives with enhanced selectivity for cancer-like membrane models and breast cancer cell lines (PubMed 42800250). That is a cancer-versus-healthy comparison, not a microbial-versus-mammalian therapeutic index.
What media, inoculum, and membrane models leave open
None of these abstracts report inoculum density, growth-medium recipe, or bilayer mole fractions. RK-14 is said to retain activity under various temperature and ionic conditions, without listing those conditions. Hemolysis, unnamed cytotoxicity, PS-doped model bilayers, and AFM each weight different membrane properties. Planktonic MIC is a different exposure than catheter biofilm at 1 × MIC.
Selectivity-related readouts in four late-September 2026 papers (as of 6 October 2026)
| First publication date | System | Microbial or membrane readout | Host or model selectivity readout | Evidence stage |
|---|---|---|---|---|
| 29 Sep 2026 | RK-14 | MIC/MBC, PI uptake, biofilms | Cytotoxicity (cell type not named); hemolysis higher at higher concentrations | In-vitro peptide |
| 28 Sep 2026 | C. frutescens F2/F3 | Candida growth and permeabilization | Sheep erythrocyte lysis | In-vitro fractions |
| 25 Sep 2026 | Gramicidin S analogues | Review of membrane-active MDR activity | Cytotoxicity as a design constraint | Review |
| 24 Sep 2026 | Mastoparan-1 mutants | AFM perturbation morphology; model membranes | Cancer versus healthy cells; cancer-like (PS) models | Model membrane plus cell assays |
Checklist when reading an AMP selectivity claim:
- Is the host readout sheep (or other) hemolysis, a named nucleated cell line, or only a lipid bilayer?
- Are microbial data planktonic MIC/MBC, a dye-uptake assay, biofilm eradication, or AFM?
- Are concentrations in the same units and against the same organism class?
- Does the report give medium, inoculum, ionic conditions, and lipid composition?
- Is a review being treated as if it were a new head-to-head experiment?
A specific remaining question, given this as-of-6-October-2026 slice rather than a complete methods review, is whether a peptide judged selective by low sheep-erythrocyte lysis versus Candida growth at 200 µg mL⁻¹ would keep that ranking in the RK-14 format (micromolar MIC/MBC and PI uptake on MDR bacteria) and in the MP-1 format (defined PS content, AFM morphology, and healthy versus breast-cancer cell lines). Until those parameters are matched, selectivity is local to the assay, not a transferable peptide property.
Frequently Asked Questions
Can RK-14 MIC values be compared with the Capsicum fraction doses?
Not numerically. RK-14 MICs and MBCs are reported in micromolar units against MRSA and ESBL-positive E. coli, whereas the Capsicum F2 and F3 fractions were tested at 200 µg mL⁻¹ against Candida species with percent growth inhibition as the endpoint.
Does low hemolysis mean an AMP is selective for microbes over mammalian cells?
Not from these abstracts. The Capsicum fractions were screened by sheep erythrocyte lysis; RK-14 cytotoxicity was not tied to a named cell type, and hemolysis increased at higher concentrations. The MP-1 work compares cancer-like membranes and cancer versus healthy cells, which is a different contrast.
Explore Further
Browse our research peptide catalog and review third-party lab reports & COAs for published batches.
---
Research use only. The information above is provided for educational and laboratory research purposes only. The compounds discussed are not approved for human or veterinary use, diagnosis, treatment, or the prevention of any disease. Nothing here is medical advice.