Bioisosterism and Scaffold Hopping

Bioisosterism and scaffold hopping are core structural replacement strategies used during hit-to-lead and lead optimization to modify a biologically active molecule so that it retains specific target recognition while improving ADMET properties, target selectivity, synthetic accessibility, or intellectual property freedom-to-operate. Bioisosteres are chemical substituents or functional groups with comparable steric, electronic, or physicochemical profiles that elicit similar biological responses when swapped on a drug molecule. Scaffold hopping, by contrast, replaces the central core framework (the Bemis–Murcko scaffold) of a ligand with a topologically distinct structure while holding the three-dimensional exit vectors of the peripheral pharmacophore elements fixed in space.

The conceptual foundation of bioisosterism originated in physical chemistry with Irving Langmuir (1919), who coined the term isosterism to describe molecules or ions possessing identical numbers and arrangements of valence electrons—such as N₂ and CO (both with 14 total and 10 valence electrons), or N₂O and CO₂ (22 electrons)—which display nearly identical crystal structures and boiling points. In 1925, Hans Gabriel Grimm formulated the hydride displacement law, showing that adding a hydrogen atom (a proton plus an electron) to an atom in the periodic table creates a "pseudoatom" with the valence behavior of the next group to the right (for example, –CH2– mimics –NH–, which mimics –O–, which mimics –F). Hans Erlenmeyer extended these rules to aromatic rings in the 1930s by showing that a vinylene group (–CH=CH–) in benzene is isosteric with a divalent sulfur atom (–S–) in thiophene. In 1951, Harris Friedman introduced the term bioisosterism to encompass all functional replacements—regardless of electron count—that produce broadly similar biological activity, a concept later divided by Alfred Burger (1970) and refined by Nicholas Meanwell into classical and non-classical bioisosteres.

Classical Bioisosteres and Grimm's Hydride Displacement

Classical bioisosteres obey Langmuir's and Grimm's valence-electron and steric rules, grouped by valency:

1. Monovalent Atoms and Groups: Hydrogen (–H) and fluorine (–F, van der Waals radii 1.20 Å vs. 1.47 Å); hydroxyl (–OH), amino (–NH2), and thiol (–SH); or chloro (–Cl), bromo (–Br), trifluoromethyl (–CF3), and cyano (–CN). Swapping –H for –F preserves steric fit while introducing strong inductive electron withdrawal (Hammett σ_p = +0.06, σ_m = +0.34) that blocks cytochrome P450 oxidation and modulates the p*K*_a of nearby amines by 1 to 2 log units.
2. Divalent Tethers: Methylene (–CH2–), ether (–O–), thioether (–S–), and secondary amine (–NH–) linkers, which share tetrahedral or near-tetrahedral bond angles (~105°–112°) while altering local polarity and hydrogen-bonding capacity.
3. Trivalent Atoms and Ring Equivalents: Replacing an aromatic methine (–CH=) with a ring nitrogen (–N=) converts benzene into pyridine, pyrimidine, or pyrazine, lowering Log*P* by ~0.6 to 1.2 units per nitrogen atom and eliminating electron-rich ortho/para sites prone to bioactivation into reactive quinone methides or epoxides. Similarly, replacing a phenyl ring (–CH=CH–) with thiophene (–S–), furan (–O–), or pyrrole (–NH–) preserves aromatic planarity and π-stacking interactions.

Non-Classical Bioisosteres of Carboxylic Acids, Amides, and Phenols

Non-classical bioisosteres do not share the same number of atoms or valence electrons as the parent functional group, nor do they follow simple periodic table hydride shifts; instead, they mimic spatial volume, acidity (p*K*_a), electrostatic potential surface, dipole moment, or hydrogen-bond geometry.

• Carboxylic Acid (–COOH, p*K*_a ≈ 4.0–4.5) Bioisosteres: Although a carboxylate anion forms strong salt bridges (Δ*G*° ≈ −3 to −5 kcal·mol⁻¹) with cationic arginine or lysine residues, its high polarity and negative charge at physiological pH 7.4 restrict passive membrane permeability and blood–brain barrier penetration. Moreover, carboxylic acids undergo phase II conjugation to form reactive acyl glucuronides that can covalently modify host proteins. Replacing –COOH with a planar 1*H*-tetrazole ring preserves acidity (p*K*_a ≈ 4.5–4.9) while delocalizing the negative charge across four nitrogen atoms; because the tetrazolate anion is roughly 10-fold more lipophilic than the corresponding carboxylate and resists acyl-glucuronidation and mitochondrial β-oxidation, this swap enabled the discovery of the orally bioavailable angiotensin II receptor blocker losartan. Another major carboxylic acid bioisostere is the acylsulfonamide (–CONHSO2R, p*K*_a ≈ 4.0–4.8), whose sulfonyl oxygens and acidic NH replicate carboxylate geometry while allowing extra hydrophobic contacts via the R-group, a design used in hepatitis C virus NS3/4A protease inhibitors (grazoprevir and paritaprevir).
• Amide Bond (–CONH–) Bioisosteres: Native amides are susceptible to enzymatic hydrolysis by amidases and proteases and carry a high desolvation penalty associated with the NH hydrogen-bond donor. Heteroaromatic 1,2,4-oxadiazoles, 1,3,4-oxadiazoles, and 1,4-disubstituted 1,2,3-triazoles (dipole moment μ ≈ 4.5 D vs. ~3.8 D for an amide) mimic the planar trans-amide geometry and hydrogen-bond acceptor vectors while eliminating proteolytic cleavage. Alternatively, trifluoroethylamine (–CH(CF3)NH–) and fluoroalkene (–CF=CH–) groups reproduce the C–N bond angle and dipole vector without a carbonyl oxygen.
• Phenol (–ArOH, p*K*_a ≈ 9.5–10.0) Bioisosteres: Phenols frequently form critical hydrogen-bond donor/acceptor contacts in receptor pockets (such as estrogen receptors and adrenergic receptors) but suffer rapid first-pass intestinal and hepatic metabolism via UDP-glucuronosyltransferases (UGTs) and sulfotransferases (SULTs), often resulting in oral bioavailability below 5%–10%. Replacing the phenolic hydroxyl group and an adjacent ring carbon with a bicyclic indazole, benzimidazole, benzimidazolone, or oxindole ring preserves the NH hydrogen-bond donor at the same vector while evading phase II glucuronidation.

Bicyclo[1.1.1]pentanes (BCP) and Cubanes

In 2009, Frank Lovering, Jack Bikker, and Christine Humblet demonstrated in their landmark analysis "Escape from Flatland" that the clinical success rate of drug candidates correlates positively with the fraction of sp³-hybridized carbons (Fsp³ = number of sp³ carbons / total carbons): average Fsp³ rises from 0.36 in discovery compounds to 0.47 in approved drugs. Over-reliance on Pd-catalyzed Suzuki–Miyaura cross-couplings historically enriched medicinal chemistry libraries with flat, aromatic biaryl rings that stack tightly in the solid-state crystal lattice (high melting point T_m), lowering aqueous solubility according to Yalkowsky's general solubility equation.

To replace flat para-disubstituted benzene rings (inter-bridgehead distance ≈ 2.82 Å) with saturated three-dimensional cages that maintain collinear 180° exit vectors, medicinal chemists substitute bicyclo[1.1.1]pentane (BCP) (bridgehead C1–C3 distance ≈ 1.85 Å), bicyclo[2.2.2]octane (BCO) (distance ≈ 2.58 Å), or cubane (diagonal distance ≈ 2.72 Å). For example, replacing the internal fluorophenyl ring of the γ-secretase inhibitor avagacestat with a BCP motif (Stepan et al., 2012) disrupted crystal packing, increased kinetic and thermodynamic aqueous solubility by >10-fold, enhanced Caco-2 passive permeability, and maintained nanomolar potency.

Heterocycle Replacements, Pseudocycles, and 3D Shape Hopping

Coined by Gisbert Schneider and colleagues in 1999, scaffold hopping searches for isofunctional molecular structures with different core backbones. Hans-Joachim Böhm et al. (2004) classified scaffold hops into four structural categories according to the degree of topological divergence:

1. 1°-Hops (Heterocycle Replacements): Swapping core ring heteroatoms (such as pyrrolopyrimidine to pyrazolopyrimidine or quinazoline to quinoline in epidermal growth factor receptor [EGFR] and Janus kinase [JAK] inhibitors) to alter hinge-binding electrostatics and establish new composition-of-matter patents.
2. 2°-Hops (Ring Opening and Ring Closure / Pseudocycles): Breaking a covalent ring bond to increase conformational flexibility, or closing an acyclic linker into a macrocycle or fused ring to pre-organize the bioactive conformation (reducing the −*T*Δ*S*° entropy penalty of binding). A refined subclass is the intramolecular hydrogen-bonded pseudocycle, where an ortho-alkoxy benzamide or aminopyrimidine forms an internal 6-membered hydrogen-bonded ring in solution that mimics the planarity and geometry of a covalently fused quinazoline without the high crystal-lattice energy.
3. 3°-Hops (Peptidomimetic Scaffolds): Replacing a peptide α-helix or β-turn backbone with a small-molecule non-peptidic core (such as the benzodiazepine or cis-imidazoline Nutlin scaffolds) that projects amino acid side-chain mimics along identical vectors.
4. 4°-Hops (Topology-Based and Computational 3D Shape/Field Hops): Using virtual screening algorithms such as ROCS (Rapid Overlay of Chemical Structures) (Gaussian shape similarity combined with electrostatic field overlap) or topological pharmacophore fingerprint descriptors (CATS) to identify entirely unrelated chemical frameworks that superimpose in three-dimensional receptor space.

What is the difference between a classical and a non-classical bioisostere?

Classical bioisosteres share the same valence electron configuration, atom periodicity, or ring size as the group they replace—such as swapping hydrogen for fluorine (–H to –F), or a benzene ring (–CH=CH–) for a thiophene ring (–S–). Non-classical bioisosteres differ in atom count and electronic connectivity but mimic key biophysical properties such as p*K*_a, electrostatic potential, dipole vector, or spatial geometry—such as replacing a carboxylic acid (–COOH) with a 5-membered 1*H*-tetrazole ring.

Why is 1*H*-tetrazole the most widely used bioisostere for a carboxylic acid?

A 1*H*-tetrazole ring has an acidic NH proton with a p*K*_a of 4.5 to 4.9, matching the p*K*_a of a carboxylic acid (4.0 to 4.5) so both exist predominantly as anions at physiological pH 7.4. However, the negative charge of the tetrazolate anion is delocalized across four ring nitrogens over a larger surface area, making it roughly 10 times more lipophilic than a carboxylate—improving passive intestinal absorption—while eliminating toxic phase II acyl-glucuronide metabolite formation.

How does replacing a benzene ring with bicyclo[1.1.1]pentane (BCP) improve drug properties?

Flat aromatic benzene rings engage in intermolecular π–π stacking that stabilizes the solid-state crystal lattice, raising the melting point and severely limiting aqueous solubility. Bicyclo[1.1.1]pentane (BCP) is a rigid, three-dimensional, sp³-hybridized cage that maintains the linear 180° exit geometry of a para-substituted phenyl ring while breaking crystal planarity, eliminating aromatic CYP oxidation sites, and increasing aqueous solubility and metabolic stability.

1. Meanwell, N. A. (2011). "Synopsis of some recent tactical application of bioisosteres in drug design." Journal of Medicinal Chemistry, 54(8), 2529–2591. https://doi.org/10.1021/jm1013693
2. Lovering, F., Bikker, J., & Humblet, C. (2009). "Escape from flatland: increasing saturation as an approach to improving clinical success." Journal of Medicinal Chemistry, 52(21), 6752–6756. https://doi.org/10.1021/jm901241e
3. Stepan, A. F., Subramanyam, C., Efremov, I. V., et al. (2012). "Application of the bicyclo[1.1.1]pentane motif as a nonclassical phenyl ring bioisostere in the design of a potent and orally active γ-secretase inhibitor." Journal of Medicinal Chemistry, 55(7), 3414–3424. https://doi.org/10.1021/jm300094u
4. Böhm, H.-J., Flohr, A., & Stahl, M. (2004). "Scaffold hopping." Drug Discovery Today: Technologies, 1(3), 217–224. https://doi.org/10.1016/j.ddtec.2004.10.009