For laboratory research use only - not for human consumption

Hub ยท Peptide Science

Cell-penetrating peptides: classes, cellular entry pathways, prediction and delivery limits

Cell-penetrating peptides (CPPs) are short natural or synthetic chains of 4 to 40 amino acids studied for their ability to cross biological membranes and carry molecular cargos into cells. Published reviews group CPPs into three primary physicochemical classes (cationic, amphipathic and hydrophobic) and describe two main entry routes: energy-independent direct translocation and energy-dependent endocytosis.

Published 2026-11-19Updated 2026-11-19Written by Form Laboratories research team11 min read

A biological cell wraps itself in a phospholipid barrier that keeps large polar molecules out while letting small nutrients and signals through. For laboratory researchers studying intracellular pathways, that selective wall blocks proteins, nucleic acids and imaging probes from reaching the cytosol on their own. Cell-penetrating peptides drew attention over thirty years ago because short sequences derived from viral and developmental proteins crossed that barrier without mechanical disruption.

This page is part of the peptide science hub. It summarises a December 2024 peer-reviewed review by Moreno-Vargas and Prada-Gracia covering what cell-penetrating peptides are, their cationic, amphipathic and hydrophobic classes, direct translocation versus endocytosis, computational prediction methods, and the documented limits of cargo delivery. For a concrete example of an HIV-1 Tat entry construct in laboratory literature, see the TAT-DATNT page. This is research reference material, for research use only, with no preparation steps or storage rules.

  • Guidethe peptide science hubPeptide science questions answered from published studies and regulator records: what peptides are, which are approved in Australia and by the FDA, and where research peptides sit.
  • Guidethe TAT-DATNT pageThe meaning of TAT-DATNT, its link to dopamine transporter research, a published sequence mismatch, and the limits of laboratory evidence.

Key facts

Key facts for Cell-penetrating peptides: classes, cellular entry pathways, prediction and delivery limits
Typical length4 to 40 amino acids, natural or synthetic (Moreno-Vargas and Prada-Gracia, 2024)
Historical namesProtein transduction domains (PTDs), membrane translocation sequences (MTSs) and Trojan peptides
Three primary physicochemical classesCationic (arginine/lysine-rich), amphipathic (around 40% of catalogued CPPs) and hydrophobic (around 15%), alongside cyclic variants
Studied cargo typesSmall molecules, short peptides, full-length proteins, plasmid DNA, siRNA, liposomes, nanoparticles, viral particles, radioisotopes and fluorescent probes
Main cellular entry routesEnergy-independent direct translocation (inverted micelle, pore formation, carpet-like models) and energy-dependent endocytosis (macropinocytosis, CME, CvME, coat-independent)
Key cargo-delivery limits citedEndosomal entrapment, low cell selectivity, proteolytic degradation, rapid renal clearance, physiological barriers (BBB, intestinal epithelium) and immunogenicity
FDA approval status citedNo CPP or CPP-cargo complex approved by the FDA as of December 2024 (Moreno-Vargas and Prada-Gracia, 2024)
SourceMoreno-Vargas LM, Prada-Gracia D, International Journal of Molecular Sciences, 2024;26(1):59 (PMID 39795918)
Results shown for Form Laboratories productsNone on this page
UseResearch reference material. Not for human or animal use
International Journal of Molecular Sciences 2024

What are cell-penetrating peptides?

Cell-penetrating peptides (CPPs) are short natural or synthetic peptides, typically 4 to 40 amino acids long, that can traverse cellular membranes while preserving the functional integrity of attached cargos. Early examples discovered more than three decades ago were identified inside larger natural proteins and termed protein transduction domains (PTDs), membrane translocation sequences (MTSs) or Trojan peptides. In laboratory research, CPPs are studied both as delivery vectors linked covalently or non-covalently to macromolecular payloads and as sequences carrying intrinsic functional motifs.

The first protein-derived examples established the field: the Tat peptide from the HIV-1 trans-activator of transcription protein (residues 49 to 57, RKKRRQRRR) and Penetratin (residues 43 to 58, RQIKIWFQNRRMKWKK) from the third alpha-helix of the Drosophila Antennapedia homeodomain. Chimeric designs later combined segments from separate natural or synthetic sequences, while the peptide versus protein guide sets out how short chains differ by size and convention from full-length proteins.

Across published studies summarised by Moreno-Vargas and Prada-Gracia, CPPs have been coupled to small molecules, short peptides, full-length proteins, DNA, siRNA, liposomes, nanoparticles, viral particles, fluorescent probes and radioisotopes. Because CPPs lack a single shared amino-acid motif, their uptake behaviour depends jointly on sequence composition, three-dimensional conformation, membrane lipid environment and whether cargo is attached covalently or through non-covalent electrostatic or hydrophobic complexes.

  • GuidepeptidesWhat peptides are: short chains of amino acids made by the body, some approved as medicines. Regulator and study evidence, safety records and Australian legal status. Research material is not a medicine.
  • Guidethe peptide versus protein guidePeptide vs protein, hormone, steroid and collagen compared: size conventions, natural vs synthetic, research vs compounded, and what published trials and regulators record. Research material is not a medicine.
International Journal of Molecular Sciences 2024

How are cell-penetrating peptides classified?

Moreno-Vargas and Prada-Gracia group CPPs into three primary physicochemical classes based on charge, hydrophobicity and residue distribution: cationic, amphipathic and hydrophobic, alongside cyclic CPPs as a structurally constrained category. A majority of catalogued CPPs carry a net positive charge at physiological pH. Amphipathic CPPs account for roughly 40% of identified sequences, whereas hydrophobic CPPs make up about 15%; later discoveries have also added neutral and anionic examples.

Cationic CPPs (cCPPs) contain a continuous stretch of basic residues, predominantly arginine and lysine, without folding into an amphipathic helix (typically adopting disordered or random-coil conformations in solution). Arginine's guanidinium side chain forms stable bidentate hydrogen bonds with negatively charged phosphate, sulfate and carboxylate groups on cell-surface glycosaminoglycans such as heparan sulfate; as a result, polyarginines (with hepta-arginine R7 as a minimum threshold and R8 to R10 widely studied) show higher uptake than polylysines of the same length. Prototypes include HIV-1 Tat(49-57) (examined in a dopamine-transporter research construct on the TAT-DATNT page) and Penetratin.

Amphipathic CPPs (aCPPs) pair polar or charged regions with non-polar stretches rich in alanine, valine, leucine and isoleucine, frequently folding into an amphipathic alpha-helix with one hydrophobic face and one cationic, anionic or polar face. The review divides them into four subgroups: primary amphipathic CPPs (typically 20 or more amino acids with sequential hydrophobic and hydrophilic domains, such as pVEC, MPG and Pep-1), secondary amphipathic CPPs (shorter sequences that adopt helical structure upon contacting negatively charged membranes or polyanions, such as MAP, M918, anionic GALA, cationic KALA and bacterial p28), beta-sheet amphipathic peptides (such as EAK16 and RAD16) and proline-rich amphipathic peptides (such as SAP and Bac-7).

Hydrophobic CPPs (hCPPs) consist solely of non-polar residues or carry a low net charge (under 20% of the sequence) alongside a hydrophobic motif or lipid group critical for entry; many natural examples originate from signal peptide sequences. They split into linear hydrophobic peptides built from natural amino acids (such as Bax-inhibiting pentapeptides like VPMLK and VPTLK, or phage-display sequences Pep-7, SG3 and FGF) and chemically modified designs, including hydrocarbon-stapled alpha-helical peptides (cross-linked at i/i+3, i/i+4, i/i+7 or i/i+11 positions), prenylated peptides bearing farnesyl (C15) or geranylgeranyl (C20) chains, and lipid-tethered pepducins such as palmitoylated P1pal-7. Cyclic CPPs (cyCPPs), built via solid-phase peptide synthesis and closed through disulfide, lactam or backbone bridges, add conformational rigidity and greater resistance to enzymatic proteolysis.

  • Guidethe TAT-DATNT pageThe meaning of TAT-DATNT, its link to dopamine transporter research, a published sequence mismatch, and the limits of laboratory evidence.
  • Guidesignal peptide sequencesWhat a signal peptide sequence means, how cells recognise a protein's address, and why protein targeting differs from cosmetic signal-peptide language.
  • Guidesolid-phase peptide synthesisHow solid-phase peptide synthesis builds chains on an insoluble resin, why deletion and truncated by-products form, and what that means for analytical testing. Research reference only.
International Journal of Molecular Sciences 2024

How do cell-penetrating peptides enter cells?

Published literature describes two overarching cellular entry routes for CPPs: energy-independent direct translocation across the plasma membrane and energy-dependent endocytosis inside membrane-bound vesicles. Many CPPs use more than one pathway depending on peptide concentration, charge, length, secondary structure, cargo size, cell type, temperature, pH and local membrane lipid composition. Electrostatic binding to negatively charged cell-surface glycosaminoglycans, particularly heparan sulfate proteoglycans, is widely viewed as an initial trigger that clusters surface components and initiates either route.

Direct translocation is a single-step, energy-independent process first inferred when CPP entry persisted at low temperatures (such as 4 degrees Celsius) or in the presence of metabolic and endocytic inhibitors. It requires electrostatic contact with negatively charged membrane lipids and temporary or permanent membrane destabilisation, and is more likely at higher peptide concentrations (above 10 micromolar generally, or above 5 micromolar for arginine-rich CPPs entering via membrane nucleation zones) and for primary amphipathic peptides such as MPG and transportan analogues. Moreno-Vargas and Prada-Gracia summarise three main physical models: inverted micelle formation (membrane invagination capturing hydrophilic peptide-cargo assemblies and releasing them cytoplasmically), pore formation (barrel-stave helical bundles, toroidal lipid-bending pores, or transient water pores observed in molecular dynamics simulations of Tat), and the carpet-like or membrane-thinning model (parallel surface accumulation reducing local surface tension until peptides intercalate and the bilayer reseals).

Endocytosis is an active, energy-dependent two-step route (vesicular uptake followed by endosomal escape) and is widely accepted as the primary entry pathway at lower concentrations (below 5 micromolar for arginine-rich CPPs) and especially when CPPs carry macromolecular cargos. Within pinocytosis, the review outlines four mechanisms: macropinocytosis (rapid, receptor-independent, lipid-raft and actin-driven membrane ruffling that forms large macropinosomes, strongly implicated as a major CPP route), clathrin-mediated endocytosis (CME, a receptor-, clathrin- and dynamin-dependent pathway progressing from coated pits to early endosomes, acidic late endosomes and lysosomes), caveolae-mediated endocytosis (CvME, slower flask-shaped invaginations taking over 20 minutes and organised by caveolins and cavins in cholesterol- and sphingolipid-rich domains), and clathrin- and caveolae-independent endocytosis mediated by lipid rafts, flotillins or ARF pathways.

International Journal of Molecular Sciences 2024

How are cell-penetrating peptides predicted computationally?

Because experimentally catalogued CPP datasets have grown steadily, researchers use sequence descriptors, classical machine learning and deep learning models to predict whether an amino-acid sequence can cross cell membranes. Early statistical models from Hallbrink and colleagues (2005) and Hansen and colleagues (2008) mapped peptides onto 26 dimensionless z-descriptors capturing physicochemical traits such as molecular weight, lipophilicity (logP), steric bulk and accessible surface area. In 2010, Dobchev and colleagues introduced the first machine learning CPP predictor, an artificial neural network trained on 52 CPPs and 10 non-CPPs that achieved 83% accuracy on its validation set.

Later tools combined larger feature sets (from 61 descriptors and 400 pseudo-amino-acid or dipeptide features up to 15,537 descriptors in CellPPDMod) with support vector machines, random forests, gradient-boosted trees (XGBoost, LightGBM) and stacked ensembles such as CellPPD, MLCPP, MLCPP 2.0, SkipCPP-Pred and StackCPPred. Deep learning architectures including ITP-Pred (CNN-BiLSTM), SiameseCPP, DeepTPpred and PractiCPP learn sequence patterns directly without manual feature selection; across both classical and deep models, positively charged residues (arginine, lysine, histidine), aromatic and hydrophobic residues (tryptophan, leucine, alanine, isoleucine), charge distribution and lipophilicity recur as informative predictors.

Moreno-Vargas and Prada-Gracia highlight clear boundaries on current prediction tools. Many models act as binary classifiers (CPP versus non-CPP) trained on short sequences from limited datasets, with only a subset (such as CellPPD, MLCPP 2.0 and DeepCPPred) attempting to separate high from low uptake efficiency, and specialised models such as Multi_CycGT targeting cyclic sequences. Predicting specific entry pathways or cargo-specific delivery remains largely unexplored, and the field still lacks a unified benchmark database and standard evaluation metrics for comparing predictors across laboratories.

International Journal of Molecular Sciences 2024

What are the known limits of CPP cargo delivery?

Moreno-Vargas and Prada-Gracia report that despite decades of preclinical work and multiple clinical trials, no CPP or CPP-cargo complex has yet received FDA approval. Five recurring scientific and technical barriers limit cargo delivery: endosomal entrapment, low cell selectivity with off-target uptake, rapid proteolytic degradation and poor pharmacokinetics, restricted passage across tight physiological barriers, and immunogenicity or membrane cytotoxicity.

Endosomal entrapment is a central bottleneck whenever CPPs enter by endocytosis: peptide-cargo complexes remain sequestered inside endosomes and are routed toward lysosomal degradation or recycling instead of reaching the cytosol or nucleus. Adding endosomolytic agents, pH-sensitive fusogenic peptides or membrane-disruptive motifs can raise cytosolic release, yet those modifications often increase cytotoxicity risk, requiring a narrow balance between escape and membrane damage. Similarly, strongly membrane-active amphipathic peptides such as MAP can cause membrane leakage and cell toxicity at concentrations as low as 1 micromolar, and certain hydrocarbon-stapled peptides have been reported to induce non-specific cell lysis.

Systemic stability and selectivity pose separate hurdles. Because CPPs are peptide chains, endogenous proteases in blood and gastrointestinal fluids degrade them rapidly, while small peptides face fast renal clearance, short plasma half-life and poor oral bioavailability. Chemical stabilisation strategies (such as incorporating D-amino acids, N-methylation or backbone cyclisation) improve protease resistance but can lower uptake efficiency or alter biological activity. Meanwhile, cationic CPPs readily bind negatively charged glycosaminoglycans across many cell types, causing off-target uptake unless paired with targeting ligands (such as antibodies, homing peptides or receptor ligands), while viral-derived or strongly cationic sequences can trigger immune responses. Tight physiological barriers including the blood-brain barrier and intestinal epithelium also remain difficult to cross reliably at scale.

International Journal of Molecular Sciences 2024

Evidence limits

What does this page not show?

This page shows no purity, identity, stability or biological test result for any Form Laboratories product. It gives no preparation steps, no reconstitution instructions, no formulation recipe and no storage rules. Moreno-Vargas and Prada-Gracia reviewed laboratory, animal and early clinical drug-delivery studies across diverse investigational sequences; those published findings do not establish approved human use, clinical efficacy or safety for any research compound. Form Laboratories supplies research material for research use only, not for human or animal use.

Exact molecular details of how individual CPPs bind membrane lipids, transition between secondary structures and escape endosomes also remain active areas of academic study rather than settled rules that generalise across every peptide-cargo pair.

Comparison

Physicochemical CPP classes and cyclic variants compared by sequence features, structural traits, named examples and cited characteristics from Moreno-Vargas and Prada-Gracia (International Journal of Molecular Sciences, 2024;26(1):59). Describes published literature only, not any Form Laboratories product.

Comparison: Cell-penetrating peptides: classes, cellular entry pathways, prediction and delivery limits
ClassSequence and charge featuresStructural traits and examplesKey uptake and limit notes cited
Cationic CPPs (cCPPs)Continuous stretch of basic residues (predominantly Arg and Lys); net positive charge at physiological pHLack an amphipathic helix (disordered or random coil); e.g. HIV-1 Tat(49-57), Penetratin, polyarginines (R8 to R10)Arg guanidinium binds surface glycosaminoglycans (heparan sulfate); endocytosis below 5 micromolar and direct translocation at higher levels; prone to endosomal sequestration and low cell selectivity
Amphipathic CPPs (aCPPs)Both polar/charged and non-polar regions (rich in Ala, Val, Leu, Ile); roughly 40% of catalogued CPPs; includes cationic, anionic and polar facesPrimary (e.g. pVEC, MPG, Pep-1), secondary alpha-helical (e.g. MAP, M918, GALA, KALA, p28), beta-sheet (EAK16, RAD16) and proline-rich (SAP, Bac-7)Conformation often shifts at the lipid interface; primary aCPPs frequently enter via direct translocation, while some membrane-active sequences (such as MAP at 1 micromolar) show marked cytotoxicity
Hydrophobic CPPs (hCPPs)Non-polar residues or low net charge (below 20% of sequence) plus a hydrophobic motif or lipid group; around 15% of catalogued CPPsLinear natural sequences (e.g. Bax-inhibiting pentapeptides VPMLK/VPTLK, Pep-7, SG3, FGF) and chemically modified stapled peptides, prenylated peptides (C15/C20) and pepducins (P1pal-7)Often derived from signal peptides; prenylated designs can enter via ATP-independent non-endocytic routes, while some stapled designs show reduced permeability or non-specific cell lysis
Cyclic CPPs (cyCPPs)Closed-loop chain formed by covalent disulfide, lactam or backbone cross-linking between termini or side chainsSynthesised via solid-phase peptide synthesis and cyclisation; conformational rigidity resists proteolysisDirect phospholipid interaction and endocytic uptake with vesicle-budding endosomal escape proposed in late endosomes; specialised predictors such as Multi_CycGT developed for cyclic designs

FAQ

What are cell-penetrating peptides?

Cell-penetrating peptides (CPPs) are short natural or synthetic peptides, typically 4 to 40 amino acids long, studied for their ability to cross cellular membranes and transport attached cargos such as small molecules, proteins, nucleic acids or imaging probes while keeping cargo structure intact (Moreno-Vargas and Prada-Gracia, 2024).

What are the three main classes of cell-penetrating peptides?

Based on physicochemical properties, Moreno-Vargas and Prada-Gracia classify CPPs into three primary classes: cationic CPPs (arginine- and lysine-rich sequences such as Tat, Penetratin and polyarginines), amphipathic CPPs (combining polar and non-polar regions and accounting for roughly 40% of catalogued CPPs) and hydrophobic CPPs (non-polar or low-charge sequences and lipidated or stapled variants, around 15% of catalogued CPPs), alongside cyclic CPPs.

How do cell-penetrating peptides enter cells?

CPPs enter cells through two main routes: energy-independent direct translocation (proposed via inverted micelle formation, barrel-stave or toroidal pores, or carpet-like membrane thinning, more common at higher concentrations) and energy-dependent endocytosis (macropinocytosis, clathrin-mediated, caveolae-mediated and coat-independent pathways, predominant at lower concentrations and with large cargos), often initiated by electrostatic binding to cell-surface glycosaminoglycans (Moreno-Vargas and Prada-Gracia, 2024).

How are cell-penetrating peptides predicted?

Computational CPP predictors evolved from 26 physicochemical z-descriptors to classical machine learning models (neural networks, support vector machines, random forests and gradient-boosted ensembles such as CellPPD and MLCPP 2.0) and deep learning architectures (such as ITP-Pred, SiameseCPP and PractiCPP) that identify sequence patterns involving basic, aromatic and hydrophobic residues. Many models remain binary classifiers limited by small datasets and a lack of unified evaluation benchmarks (Moreno-Vargas and Prada-Gracia, 2024).

Are any cell-penetrating peptides approved by the FDA, and does this page show Form Laboratories product results?

As of December 2024, Moreno-Vargas and Prada-Gracia report that no CPP or CPP-cargo complex has received FDA approval, citing non-specificity, instability and suboptimal pharmacokinetics. This page is research reference material only: it shows no test result for any Form Laboratories product and gives no preparation steps, formulation recipe or storage rules. Form Laboratories supplies research material for research use only, not for human or animal use.

Citations

1 sources
  1. Moreno-Vargas LM, Prada-Gracia D. International Journal of Molecular Sciences. 2024;26(1):59.
    doi:10.3390/ijms26010059

Written by

Form Laboratories research team

Perth, Western Australia. We supply research-use-only reference material.

Written in-house
These articles are written and maintained by our own research team, not by a named clinician.
Built from the record
Every figure comes from published literature, listed in full above with a link to the original paper.
No use directions
We describe what the research measured. We give no use directions, preparation steps or advice.

Form Laboratories supplies research-use-only reference material. This article is an educational summary of published third-party literature. It is not medical advice, not a use instruction, and research material is not for human or animal use.

2026 Form Laboratories - Research use only - Perth WAPerth WA 6000 - dispatched Australia-wide via Australia Post or CouriersPlease