palindromic recognition site

A word like 'level' or 'radar' reads the same forwards and backwards — that is a palindrome. DNA can be palindromic too, in a special double-stranded way, and restriction enzymes love to land on exactly these spots. The short stretch of DNA an enzyme recognizes and binds is called its recognition site.

Because DNA has two antiparallel strands and A pairs with T, G with C, a DNA palindrome means: reading one strand 5'-to-3' gives the same sequence as reading the partner strand 5'-to-3'. Take EcoRI's site: the top strand reads 5'-GAATTC-3', and the bottom strand also reads 5'-GAATTC-3' when you read it in its own 5'-to-3' direction. This twofold symmetry matters because most restriction enzymes work as a pair of identical protein halves (a dimer); each half grips one strand, and a symmetric site fits a symmetric enzyme perfectly. The site is usually 4 to 8 base pairs long — short enough to occur by chance every few hundred to few thousand bases, which conveniently chops a genome into manageable fragments.

Knowing exactly where an enzyme's recognition sites fall lets you predict and control how DNA will be cut — the foundation of building plasmids, mapping genes, and inserting one piece of DNA into another. A 'restriction map' is simply a diagram of where each enzyme's recognition sites lie along a DNA molecule. When you design a cloning experiment, you pick enzymes whose sites appear where you want the cuts and, crucially, not inside the gene you are trying to keep intact.