the phosphodiester bond
/ FOS-foh-dy-ESS-ter /
If the sugar-phosphate backbone is a string of beads, then the phosphodiester bond is the knot that ties each bead to the next. It is the single, specific chemical link that holds a strand of DNA or RNA together end to end.
The name breaks down neatly. A phosphate group can form ester bonds to two different alcohols (the OH groups on sugars); 'di' means two, 'ester' is that bond type — so a phosphodiester bond is one phosphate holding hands with two sugars at once. Concretely, the phosphate bridges the 3' oxygen of one nucleotide's sugar to the 5' oxygen of the next nucleotide's sugar, which is why chemists call it a 3'-5' phosphodiester bond. Each new link is forged by a condensation reaction (losing a small molecule) and costs energy, which is why the cell uses energy-rich nucleotide triphosphates to build a strand.
These bonds matter because their strength and direction define the molecule. They are covalent and strong, so the genetic message does not fall apart on its own — a great virtue for an archive. Their built-in directionality (always 3'-to-5' bridges) is what gives a strand a 5' end and a 3' end. And because they are specific bonds, enzymes can find and cut them exactly: nucleases sever phosphodiester bonds, and ligases re-seal them, which is the chemistry behind cutting and pasting DNA in the laboratory.
DNA ligase, the enzyme that seals the last gap in a newly copied strand and that biologists use to glue cloned fragments together, does one job: it forges a single phosphodiester bond.
One phosphate, two sugars, a directional 3'-to-5' link.
The phosphodiester bond joins a sugar to the next sugar, not a base to a base. The hydrogen bonds between paired bases (A-T, G-C) are a completely different, much weaker kind of contact, easily broken by heat while the backbone bonds hold.