Water networks in docking: the waters you displace decide potency

Most docking pictures show a ligand dropping into an empty pocket. The real pocket is full of water, and a large fraction of the binding free energy comes not from the ligand touching the protein but from what happens to those water molecules when the ligand arrives. Get the water story wrong and your scoring function will confidently rank the wrong analog. Get it right and a single methyl group that displaces an unhappy water can buy you an order of magnitude in affinity.

Why a pocket is not a vacuum

Before a ligand binds, the pocket is solvated. Each water sits in some local environment: some are perfectly happy, making a full complement of hydrogen bonds; others are trapped in a greasy, enclosed spot where they cannot satisfy their hydrogen-bonding partners and pay an entropic penalty for being pinned in place. Binding is a swap. The ligand evicts some of those waters back into bulk solvent and leaves others in place, sometimes bridging the ligand to the protein.

The thermodynamic accounting is the whole game. Displacing a high-energy, frustrated water is strongly favorable: that water was unhappy in the pocket and is much happier back in bulk, so kicking it out releases free energy. Displacing a contented, well-satisfied water is the opposite — you pay to remove something that was already comfortable. This is why two ligands that make almost identical contacts with a protein can differ enormously in potency: they are displacing different waters.

The classic result that made this concrete

The framework most people cite traces to work from the Friesner and Berne groups in the late 2000s. By running molecular dynamics on the solvated apo pocket and clustering the water density into discrete hydration sites, they could assign each site an enthalpy and entropy relative to bulk — the approach commercialized as WaterMap. In their factor Xa analysis, the pattern of which hydration sites a ligand displaced tracked the measured binding affinities across a congeneric series, even where the direct protein-ligand contacts looked similar. The earlier “hydrophobic enclosure” work explained the underlying physics: waters in tightly enclosed, hydrophobic sub-pockets are the most frustrated, so displacing them is where the biggest potency gains hide.

Conserved waters: the ones you keep

The flip side matters just as much. Some pocket waters are so well coordinated that the smart move is to design around them rather than displace them. A conserved bridging water that donates and accepts hydrogen bonds between the ligand and the protein can be worth more than a direct contact, and a series of FDA-approved drugs are predicted to rely on a ligand-contacting water network. The trap is that displacing one water in a connected network can destabilize the rest, so an “obvious” modification that should pick up a contact instead collapses the network and loses potency. Treating these waters as part of the receptor, not as empty space to be filled, is the difference between a rational SAR step and a surprise.

What this means for docking

Standard docking treats the pocket as dry and bakes a crude desolvation term into the scoring function. That is fine for triage but it is exactly the approximation that breaks down on the analogs you care most about. A few practical habits help:

• Keep crystallographic waters that are clearly conserved.If a bridging water appears in multiple structures of the same target, dock with it present rather than deleting every water by reflex.
• Treat displaceable waters explicitly when you can.Methods that let docking add or remove discrete waters with a desolvation penalty — the hydrated-docking force fields built for this — recover poses that dry docking misses.
• Be suspicious of a potency jump your contacts cannot explain. If an analog gains affinity without any new protein contact, a displaced unhappy water is the usual culprit, and that insight generalizes to the next compound.

Try the docking yourself

Open Studio and dock a small congeneric series against your target, then compare the poses and interaction fingerprints between analogs that differ by a single substituent. When molecular docking gives two close analogs very different scores, ask which pocket waters each one is displacing before you trust the ranking. Liganx brings molecular docking online in the browser so you can run the comparison and inspect every pose without a local setup.

Primary sources

• Abel R, Young T, Farid R, Berne BJ, Friesner RA. Role of the Active-Site Solvent in the Thermodynamics of Factor Xa Ligand Binding. J Am Chem Soc 130, 2817-2831 (2008). doi:10.1021/ja0771033
• Young T, Abel R, Kim B, Berne BJ, Friesner RA. Motifs for molecular recognition exploiting hydrophobic enclosure in protein-ligand binding. Proc Natl Acad Sci USA 104, 808-813 (2007). doi:10.1073/pnas.0610202104
• Forli S, Olson AJ. A force field with discrete displaceable waters and desolvation entropy for hydrated ligand docking. J Med Chem 55, 623-638 (2012). doi:10.1021/jm2005145