Lipinski's Rule of Five

Lipinski's Rule of Five (Ro5, also called the Pfizer Rule of Five) is an empirical physicochemical heuristic in medicinal chemistry, chemoinformatics, and ADMET profiling that evaluates whether a chemical compound possesses the molecular weight, lipophilicity, and hydrogen-bonding polarity required to achieve adequate aqueous solubility and passive intestinal membrane permeability for oral bioavailability in humans. Formulated in 1997 by Christopher A. Lipinski, Franco Lombardo, Beryl W. Dominy, and Paul J. Feeney at Pfizer Central Research, the rule takes its name from the observation that all four of its numerical parameter cutoffs are multiples of the integer 5 (500, 5, 5, and 10).

During the early 1990s, the adoption of combinatorial chemistry and high-throughput screening (HTS) shifted pharmaceutical screening libraries toward high-molecular-weight, hydrophobic hits that displayed potent in vitro receptor binding yet failed in preclinical animal dosing due to negligible oral absorption. By analyzing the 90th-percentile property distributions of 2,245 candidate drugs from the World Drug Index (WDI) that had assigned United States Adopted Names (USAN) or International Nonproprietary Names (INN) upon entering Phase II human clinical trials, Lipinski and colleagues demonstrated that poor oral absorption or membrane permeation is statistically probable whenever a small molecule violates two or more of four thresholds.

Definition of the Four Rule-of-Five Parameters

Under Lipinski's formulation, an orally active small molecule generally satisfies at least three of the following four conditions:

1. Molecular Mass (MW) ≤ 500 Daltons (g/mol): Fewer than ~35–38 non-hydrogen heavy atoms.
2. Calculated Octanol–Water Partition Coefficient (cLog*P* or MLog*P*) ≤ 5.0: Where the thermodynamic partition coefficient is defined as the equilibrium concentration ratio of neutral solute between 1-octanol (a surrogate for the amphiphilic phospholipid bilayer interior) and water:

P = [Solute]_1-octanol / [Solute]_water , log P = log₁₀ P (Lipinski originally allowed Moriguchi log P [MLog*P*] ≤ 4.15 or Pomona MedChem cLog*P* ≤ 5.0).

1. Hydrogen-Bond Donors (HBD) ≤ 5: Defined operationally by Lipinski as the total count of nitrogen–hydrogen (N–H) and oxygen–hydrogen (O–H) covalent bonds in the molecule (note that a primary amine –NH2 counts as 2 HBDs).
2. Hydrogen-Bond Acceptors (HBA) ≤ 10: Defined operationally by Lipinski as the total count of nitrogen (N) and oxygen (O) atoms in the molecule, regardless of formal hybridization or protonation state.

Biophysical Derivation from Overton's Rule and Bilayer Diffusion

The four Ro5 boundaries arise directly from the physical chemistry of passive transcellular diffusion across the apical lipid bilayer of human small-intestinal enterocytes. By Fick's First Law and Overton's Permeability Rule, the apparent transmembrane permeability coefficient P_app (in cm/s) across a hydrophobic hydrocarbon slab of thickness δ (~30 Å for a biological membrane tail core) equals:

P_app = (K_mem/w · D_mem) / δ

where K_mem/w is the membrane–water partition coefficient (proportional to P_{octanol/water}) and D_mem is the intramembrane diffusion coefficient:

• Why HBD ≤ 5 and HBA ≤ 10: Before a solute can enter the low-dielectric hydrocarbon core of a lipid bilayer (relative permittivity ε_r ≈ 2.0) from bulk water (ε_r ≈ 78.4), it must shed its hydration shell. Breaking a single water–solute hydrogen bond incurs an enthalpic desolvation penalty of Δ*H*_desolv ≈ +4.0 to +6.0 kcal/mol (+16 to +25 kJ/mol), with O–H and N–H donors carrying a steeper penalty than acceptors. A molecule bearing >5 HBDs or >10 HBAs faces an activation free-energy barrier (Δ*G*^‡_perm) too high for passive bilayer crossing at 310 K.
• Why MW ≤ 500 Da: By the Stokes–Einstein and free-volume diffusion theories inside ordered acyl chains, D_mem decreases steeply with molecular volume (V_m), while large molecules also form higher-melting crystal lattices that depress aqueous solubility.
• Why log P ≤ 5: While increasing log P between 0 and 3 raises K_mem/w and accelerates membrane permeation, compounds with log P > 5 suffer extreme aqueous insolubility (by the General Solubility Equation of Yalkowsky and Valvani: log S_w = 0.5 − 0.01(MP − 25) − log P), sequester inside cell membranes, and undergo rapid oxidative clearance by hepatic cytochrome P450 enzymes.

Veber's Rules, Lead-Likeness, and Fragment Rule of Three

In 2002, Daniel F. Veber and colleagues at GlaxoSmithKline analyzed oral bioavailability measurements in rats across 1,100 drug candidates and showed that two structural molecular descriptors predict oral absorption even when MW > 500 Da: (1) 10 or fewer rotatable bonds (N_rot ≤ 10), because conformational flexibility lowers the fraction of conformers matched to membrane permeation, and (2) Topological Polar Surface Area (TPSA) ≤ 140 Ų (computed in microseconds via Peter Ertl's 2D fragment-contribution method).

Because medicinal chemistry optimization from an initial screening hit to a clinical drug typically adds Δ*MW* ≈ +80 to +120 Da and ΔcLog*P* ≈ +1.0 to +1.5 ("molecular obesity"), Tudor Oprea (2001) introduced stricter lead-likeness cutoffs (MW ≤ 350 Da, cLog*P* ≤ 3.5), while Miles Congreve and colleagues at Astex (2003) formulated the Rule of Three (Ro3) (MW ≤ 300 Da, cLog*P* ≤ 3, HBD ≤ 3, HBA ≤ 3) for Fragment-Based Drug Discovery (FBDD). To ensure each heavy atom contributes productively to binding Gibbs free energy, Andrew L. Hopkins (2004) defined Ligand Efficiency (LE):

LE = −Δ*G*°_bind / N_heavy ≈ (1.4 · pIC₅₀) / N_heavy [kcal · mol⁻¹ · heavy atom⁻¹]

with ideal drug-like leads maintaining LE ≥ 0.30 kcal/mol per heavy atom. In 2012, G. Richard Bickerton and Hopkins replaced binary pass/fail cutoffs with the continuous Quantitative Estimate of Drug-likeness (QED ∈ [0, 1]), computed as the geometric mean of eight desirability functions over MW, ALOGP, HBA, HBD, TPSA, N_rot, aromatic rings, and structural alerts.

Lipinski explicitly noted that four major classes of orally active therapeutics are exempt from the Rule of Five: natural products, antibiotics (such as erythromycin and rifampicin), antifungals, and vitamins/cardiac glycosides. Many of these compounds are substrates for active transmembrane influx transporters (such as the peptide transporter PEPT1, organic anion transporting polypeptides OATPs, or nucleoside transporters) rather than relying on passive lipid diffusion.

Furthermore, modern drug discovery targeting flat protein–protein interactions and intracellular protein degradation operates extensively in Beyond Rule of Five (bRo5) space (MW = 500–1,200 Da):

• Chameleonic Macrocycles: Cyclic peptides such as cyclosporine A (MW = 1,202.6 Da, 5 HBDs, 12 HBAs) achieve ~30% oral bioavailability in humans by acting as "molecular chameleons"—exposing their amide N–H and C=O groups to water in aqueous solution, then folding inward to form internal transannular hydrogen bonds that mask their polarity inside the low-dielectric lipid membrane.
• PROTACs (Proteolysis-Targeting Chimeras): Heterobifunctional degraders (MW = 700–1,100 Da) linking a target warhead to an E3 ubiquitin ligase ligand achieve catalytic intracellular target degradation at sub-nanomolar concentrations despite minimal passive permeability.

Does Lipinski's Rule of Five predict whether a molecule will be pharmacologically active?

No. Lipinski's Rule of Five does not predict whether a molecule binds to a biological target or exhibits pharmacological activity; rather, it predicts whether a compound has the physicochemical properties required for oral absorption (aqueous dissolution and passive intestinal permeability) in humans. Drugs administered parenterally (via intravenous, intramuscular, or subcutaneous injection, or inhalation)—such as monoclonal antibodies, insulin, and glycopeptide antibiotics like vancomycin (MW = 1,449 Da)—do not need to cross the gastrointestinal epithelium and are completely unconstrained by the Rule of Five.

Why are all the cutoffs in Lipinski's rule multiples of 5?

When Christopher Lipinski and colleagues examined the 90th-percentile distributions of the 2,245 Phase II clinical candidates in the World Drug Index, the exact statistical 90th percentiles were approximately MW ≈ 480–500 Da, cLog*P* ≈ 4.9, OH+NH count ≈ 4.8, and N+O count ≈ 9.8. Lipinski deliberately rounded these empirical 90th-percentile values to the clean multiples of 5 (500, 5, 5, and 10) to create a memorable mnemonic rule of thumb that synthetic bench chemists could calculate mentally before synthesizing a molecule.

How can cyclosporine A be taken orally as a pill when its molecular weight is 1,202 Da?

Cyclosporine A (MW = 1,202.6 Da, 11 amino acids) violates three of the four Rule-of-Five parameters yet achieves ~30% human oral bioavailability because of two structural modifications evolved by its fungal producer (Tolypocladium inflatum): seven of its eleven backbone amide nitrogens are N-methylated (reducing its hydrogen-bond donor count from 11 to just 4), and its macrocyclic ring undergoes a solvent-dependent conformational switch (molecular chameleon behavior). Upon entering the non-polar interior of a cell membrane, the remaining four N–H donors fold inward to form intramolecular hydrogen bonds with backbone carbonyl oxygens, dropping the effective polar surface area below the threshold required for passive diffusion.

1. Lipinski, C. A., Lombardo, F., Dominy, B. W., & Feeney, P. J. (1997). Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews, 23(1–3), 3–25. https://doi.org/10.1016/S0169-409X(96)00423-100423-1)
2. Veber, D. F., Johnson, S. R., Cheng, H. Y., Smith, B. R., Ward, K. W., & Kopple, K. D. (2002). Molecular properties that influence the oral bioavailability of drug candidates. Journal of Medicinal Chemistry, 45(12), 2615–2623. https://doi.org/10.1021/jm020017n
3. Bickerton, G. R., Paolini, G. V., Besnard, J., Muresan, S., & Hopkins, A. L. (2012). Quantifying the chemical beauty of drugs. Nature Chemistry, 4(2), 90–98. https://doi.org/10.1038/nchem.1243
4. Doak, B. C., Over, B., Giordanetto, F., & Kihlberg, J. (2014). Oral druggable space beyond the rule of 5: Insights from drugs and clinical candidates. Chemistry & Biology, 21(9), 1115–1142. https://doi.org/10.1016/j.chembiol.2014.08.013