hydrogen bond
A hydrogen bond is like a tiny magnetic handshake. One molecule offers a hydrogen atom that sits next to a needy, electron-hungry atom such as nitrogen, oxygen or fluorine; another molecule offers a lone pair of electrons to welcome it. The two snap into a directional grip that is much weaker than a real chemical bond but strong enough to hold a drug snugly inside its target.
More precisely, a hydrogen bond forms between a donor (an X–H group where X is electronegative, so the hydrogen carries a partial positive charge) and an acceptor (an atom with a lone pair, carrying partial negative charge). It is mostly electrostatic, with some directional, partly covalent character, and is strongest when the donor, hydrogen and acceptor line up in a roughly straight line. Typical strengths range from about 1 to 7 kcal/mol.
In drug design hydrogen bonds give binding both strength and specificity: a well-placed donor or acceptor can dramatically improve potency and selectivity. But there is a catch — to bind its target, the drug and the pocket must each shed the hydrogen bonds they made with water. A hydrogen bond only pays off if the new contact is better than the water contact it replaces, which is why simply adding more polar groups does not reliably improve affinity.
In many kinase inhibitors, two or three hydrogen bonds to the conserved 'hinge' region of the ATP pocket anchor the drug and largely set its potency.
Hinge-binding hydrogen bonds in kinase inhibitors.
Counting potential hydrogen-bond donors and acceptors is part of Lipinski-style druglikeness rules, because too many of them hurt membrane permeability.