Preorganization: How a Rigid Cavity Pays the Entropy Bill Before Binding

What preorganization actually means

Preorganization is the degree to which a host molecule already possesses, in its unbound resting state, the conformation and electronic arrangement it will use to bind its guest. Cram stated the principle sharply in his 1986 Angewandte Chemie Nobel review: "the more highly hosts and guests are organized for binding and low solvation prior to their complexation, the more stable will be their complexes." The key word is prior. A perfectly complementary host that has to twist, fold, and shed solvent to reach its bound geometry pays for all of that during binding; a preorganized host has already paid.

It is essential to separate preorganization from complementarity, its sister concept. Complementarity is a static, geometric-and-electronic match: donor atoms of the right kind, spaced at the right distances, with the right charge complement to the guest (a cation wants convergent lone pairs; an anion wants convergent N–H or C–H⁺δ donors). Complementarity tells you whether a host can bind a given guest at all. Preorganization tells you what fraction of the binding energy survives after you subtract the cost of getting the host into its complementary shape. Two hosts with identical bound structures — identical complementarity — can differ by 10¹⁰ in affinity purely because one is rigid and the other flexible.

The classic ladder that makes this concrete is podand → coronand (crown) → cryptand → spherand. An open-chain polyether (podand) has every donor free to rotate away from the cavity. Closing it into a ring (Pedersen's crown ethers, 1967) restricts those rotations — the macrocyclic effect. Bridging the ring into a three-dimensional cage (Lehn's cryptands, 1969) restricts them further — the macrobicyclic cryptate effect. Rigidifying the whole framework so the donors cannot turn away, whether or not a guest is present (Cram's spherands, 1979), is the endpoint: total preorganization.

The thermodynamics: paying the entropy bill in advance

Binding affinity is set by the standard free energy of complexation, ΔG°bind = −RT ln Kₐ = ΔH°bind − TΔS°bind. Preorganization operates mainly through the entropic term, and understanding why requires bookkeeping the internal degrees of freedom. A flexible host in solution samples a broad conformational ensemble — dozens of low-energy rotamers about its single bonds. On binding, it collapses to essentially one conformation. Freezing each internal rotor costs roughly 4–6 J·mol⁻¹·K⁻¹ of conformational entropy; at 298 K that is on the order of +1.2 to +1.8 kJ·mol⁻¹ of unfavorable −TΔS° per frozen rotor — a factor of about 2–4 in Kₐ each. A host that freezes eight rotors during binding therefore concedes ~10–15 kJ·mol⁻¹ relative to one whose rotors were already frozen — several powers of ten in Kₐ.

A preorganized host has already paid this in its synthesis and resting state. Its rotors were locked when the ring or cage was closed, so they contribute little or no additional −TΔS° at the moment of binding. The entropy cost did not vanish; it was front-loaded into the covalent framework. This is the sense in which the cavity "pays the entropy bill before binding." There is also an enthalpic face: a flexible host must adopt torsional and van der Waals strain to wrap the guest (positive reorganization enthalpy), while a rigid host that already holds its lone pairs in a mutually repulsive, guest-ready arrangement carries that strain at rest — it is destabilized unbound, which paradoxically raises binding affinity because there is no extra enthalpic penalty to pay on complexation.

Do not neglect solvation, which Cram lists in the same breath as organization. A flexible host and its donor atoms are heavily solvated; those solvent molecules must be stripped (an enthalpy/entropy cost, though desolvation of the host often releases ordered solvent, a favorable entropy). A preorganized host with donors buried in a shielded, low-dielectric cavity is poorly solvated to begin with — Cram called this "low solvation" — so it forfeits little solvation stabilization on binding. The full accounting is ΔG°bind = ΔG°(complementarity) + ΔG°(reorganization) + ΔG°(desolvation); preorganization minimizes the last two penalties.

Worked example: spherand vs. podand, and the price of a covalent bond

Cram's cleanest demonstration is the spherand–podand pair (Cram, Kaneda, Helgeson, Lein, J. Am. Chem. Soc. 1979 and follow-up quantitative work through the mid-1980s). Spherand-1 is a macrocycle of six directly (aryl–aryl) linked, meta-substituted anisyl (methoxybenzene) units whose methyl groups and aryl rings force all six oxygen lone pairs to converge on a single octahedral cavity ~1.6 Å in diameter (radius ~0.8 Å) — a near-perfect fit for Li⁺ (ionic radius 0.76 Å with its inner coordination geometry) and Na⁺. Crucially, X-ray structures show the free spherand already has this cavity: the empty host and the Li⁺ complex are almost superimposable.

The measured association constants in CDCl₃ saturated with D₂O are enormous — Kₐ for Li⁺ and Na⁺ exceeds 10¹⁴–10¹⁶ M⁻¹, so large that Cram had to measure them by picrate-extraction competition rather than direct titration. The matched flexible podand — the same six anisyl oxygens but with the ring cut open so the arms can rotate — binds the same cations with Kₐ smaller by factors of roughly 10¹⁰. Because ΔΔG° = −RT ln(Kₐ,spherand / Kₐ,podand), a 10¹⁰ ratio corresponds to about 57 kJ·mol⁻¹ at 298 K (RT ln10 ≈ 5.7 kJ·mol⁻¹ per decade). That ~57 kJ·mol⁻¹ is, essentially, the free-energy value of preorganization for this system.

Where does it come from? The bound geometries are nearly identical, so complementarity and host–guest contacts contribute similarly. The difference is almost entirely the reorganization and conformational-entropy penalty the podand pays and the spherand does not. Cram summarized it with an economic metaphor: the spherand's binding energy is "paid for" during synthesis when the ring is closed against strain, so the host "is like a spring already compressed" — release comes on binding. Selectivity follows the same logic: because the cavity is fixed, spherand-1 discriminates Li⁺/Na⁺ from K⁺ by >10⁴, since K⁺ (radius 1.38 Å) simply cannot enter — a hallmark of rigid, size-selective preorganization versus the softer, induced-fit selectivity of flexible hosts.

Kinetics: the cost of rigidity is slow exchange

Preorganization is a thermodynamic blessing and a kinetic curse. The very rigidity that removes the reorganization penalty also removes the host's ability to open a gate for the guest to enter and leave. Complexation and decomplexation both require the guest to squeeze past the framework, and a rigid framework does not flex to help. The consequence is dramatically slowed guest exchange: cryptands and spherands show decomplexation rate constants k_off orders of magnitude smaller than crowns. Lehn's [2.2.2]cryptand binds K⁺ with a k_off in the range of seconds⁻¹ or slower, whereas 18-crown-6 releases K⁺ on the microsecond timescale — a kinetic macrobicyclic effect of many orders of magnitude.

This kinetic penalty is why nature does not always maximize preorganization. An enzyme or a transport protein that bound its substrate irreversibly tight would never turn over. Biological ionophores such as valinomycin are only partially preorganized: valinomycin's cyclododecadepsipeptide bracelet is fairly ordered but flexible enough to wrap and release K⁺ fast enough for membrane transport (turnover ~10³–10⁴ ions·s⁻¹) while still achieving K⁺/Na⁺ selectivity near 10⁴. The design target is not maximum affinity but the right point on the affinity-versus-exchange-rate trade-off.

There is a subtle corollary for anion and neutral-molecule receptors, and for catalysis. A catalyst wants to bind the transition state tightly but the ground-state substrate and product only loosely — so extreme ground-state preorganization toward the substrate can be counterproductive if it also slows product release. The best-designed molecular receptors and catalysts therefore preorganize for the intended recognized species (often the transition state) while retaining just enough conformational freedom for acceptable kinetics.

Limits, subtleties, and enthalpy–entropy compensation

The clean "preorganization = entropy" story is a useful first approximation, not the whole truth. Careful calorimetry (van 't Hoff and direct ITC studies through the 1980s–2000s, notably by Angyal, Izatt, Christensen, and later Schmidtchen and Gibb) shows that enthalpy–entropy compensation muddies the ledger: measured ΔH° and ΔS° for macrocyclic binding often move in opposite directions and partially cancel, so the observed ΔG° enhancement can be smaller than a naive rotor-counting estimate predicts, and its partition between ΔH° and ΔS° varies with solvent and guest. The macrocyclic effect itself has been argued to be enthalpy-dominated in some systems (favorable ΔH° from better donor convergence and less unfavorable host desolvation) rather than purely entropic. The point survives — preorganized hosts bind more strongly — but attributing the gain cleanly to "frozen entropy" is often an oversimplification.

Several further caveats matter for rigorous work:

• Standard-state and concentration dependence. The translational/rotational entropy lost when two particles become one (the cratic or connection entropy, roughly −30 to −50 J·mol⁻¹·K⁻¹, i.e. up to ~+15 kJ·mol⁻¹ of −TΔS°) is paid by every bimolecular association, preorganized or not, and depends on the chosen 1 M standard state. It is not part of the preorganization advantage and must be bookkept separately.
• Over-rigidification. A host preorganized for the wrong geometry is worse than a flexible one, because it cannot adapt at all. Rigidity only pays if the fixed shape is the complementary shape.
• Guest strain. If binding forces the guest to desolvate or deform (e.g., partially dehydrating a metal ion), that cost offsets the host's preorganization advantage.

Modern statements therefore frame preorganization as one of three coupled design variables — complementarity, preorganization, and (de)solvation — rather than a standalone entropy trick. Jean-Marie Lehn folded all three into the broader program of molecular recognition that defined supramolecular chemistry.

History and reach: from Pedersen's crowns to designed catalysts

The lineage is well documented. Charles Pedersen discovered crown ethers serendipitously at DuPont in 1967 while trying to make a bis(phenol) ligand; a trace cyclic byproduct bound K⁺ so avidly that its UV spectrum shifted, and dibenzo-18-crown-6 was born. Jean-Marie Lehn generalized the idea into three-dimensional cryptands (1969) and articulated supramolecular chemistry and molecular recognition as a field. Donald Cram pushed rigidity to its logical limit with spherands and, in the process, named and quantified the principle of preorganization, showing it was the missing variable that explained why some complementary hosts bound tightly and others did not. The three shared the 1987 Nobel Prize in Chemistry "for their development and use of molecules with structure-specific interactions of high selectivity."

The principle's reach extends far beyond ion binding. In enzymology, the debate over how much of catalysis comes from a rigid, transition-state-complementary active site (preorganized electrostatics, championed by Arieh Warshel) versus induced-fit dynamics is a direct descendant: Warshel's central claim is that enzymes work largely by preorganizing their polar environment to stabilize the transition-state charge distribution, so no reorganization energy (in the Marcus sense) is paid during the chemical step. In drug design, macrocyclization and conformational restriction of lead compounds are standard tactics to raise affinity and selectivity by cutting the entropic penalty of binding — the medicinal-chemistry version of turning a podand into a crown.

Contemporary supramolecular chemistry — foldamers, self-assembled coordination cages (Fujita, Nitschke, Raymond), molecular machines (Sauvage, Stoddart, Feringa, Nobel 2016), and templated synthesis — all deploy preorganization deliberately: build the binding site into a rigid scaffold, and let the front-loaded organization buy affinity, selectivity, and control. Cram's compressed-spring metaphor, coined for a bowl of anisole rings, remains the mental model for a large fraction of designed molecular recognition.

Preorganized vs. flexible hosts: where the free energy goes when a guest binds

Property	Preorganized host (spherand/cryptand)	Flexible host (podand/glyme)
Binding conformation before binding	Already adopted; cavity pre-formed	Disordered; must fold around guest
Conformational entropy cost on binding	Small — few rotors freeze	Large — many single bonds are frozen (−TΔS° unfavorable)
Reorganization enthalpy paid on binding	Near zero (already 'strained' at rest)	Significant strain/torsion adopted on binding
Typical Kₐ for Li⁺/Na⁺ (CDCl₃)	10¹⁴–10¹⁶ M⁻¹ (spherand)	10⁴–10⁸ M⁻¹ (open-chain podand)
Complexation/decomplexation kinetics	Slow (rigid gate; k_off can be 10⁻⁴–10⁻⁶ s⁻¹)	Fast (freely adapting)
Selectivity origin	Fixed cavity size = size selectivity	Induced fit = weaker discrimination