What a hydrogen bond is
A hydrogen bond forms when a hydrogen already attached to an electronegative atom (the donor: N–H, O–H, sometimes S–H) lines up with a lone pair on another electronegative atom (the acceptor: typically O or N). The hydrogen sits in the middle and is shared, a bit like a handshake. In a drug, the groups that supply these are counted as hydrogen-bond donors (the N–H, O–H groups) and hydrogen-bond acceptors (the lone pairs on carbonyls, ethers, ring nitrogens, and so on).
What makes hydrogen bonds special among the binding forces is directionality. They are strongest when the donor, hydrogen, and acceptor are roughly in a straight line, at a distance of about 2.7–3.2 Å between the heavy atoms. Bend that angle or stretch that distance and the bond weakens fast. This pickiness is a gift: it lets a target distinguish a molecule that presents an acceptor at exactly the right spot from one that is a few tenths of an ångström off. Hydrogen bonds are the scalpel of recognition.
Why an extra hydrogen bond rarely doubles potency
Beginners often reason: "the crystal structure shows an unmet acceptor near my donor, so let me add a hydroxyl and gain a strong hydrogen bond." Sometimes that works — and sometimes the new compound is no better, or worse. The reason is water. Before binding, your donor was hydrogen-bonding to water, and the protein's acceptor was too. To form the new drug–protein bond, you must break both of those water bonds. You are not creating a bond out of nothing; you are swapping water bonds for a protein bond. The net gain is only the difference.
There is a flip side, and it is just as important. An unsatisfied polar group — a donor or acceptor that you bury in the pocket but that finds no partner — is a real penalty. You paid to strip its water and gave it nothing in return. So the practical rule is symmetric: form the bond when you can, but if you cannot, do not bury a naked polar group; either find it a partner or remove the group entirely.
Putting it to work
- Look at the structure and ask which polar atoms of your ligand actually face a complementary protein partner — and which are dangling into solvent or, worse, buried unmet.
- For a hoped-for new hydrogen bond, check geometry: can your group reach the partner at a near-linear angle and the right distance, without straining the rest of the molecule?
- If a donor is hurting potency or properties, consider swapping it for a bioisostere that keeps the geometry but changes the polarity (e.g., trading an N–H for a ring nitrogen).
- Use selective hydrogen bonds to win selectivity: a partner residue present in your target but absent in a look-alike off-target is a precious handle.
Naive view: add an OH -> + one H-bond -> much tighter binding
Real ledger: break ligand-OH ... water
break protein-O ... water
form ligand-OH ... protein-O
-----------------------------------------
net ~ (new bond) - (two water bonds) -> often small
Unsatisfied buried donor: strip its water, form NOTHING -> clear penalty