how-to · 5 min read
Restriction enzyme digests: sites, cuts and fragment sizes
A restriction digest is one of the most useful checks in molecular biology: cut DNA at known sequences, separate the fragments on a gel, and compare the pattern with what the sequence predicts. The arithmetic behind the pattern is simple, but the difference between linear and circular DNA, and between blunt and staggered cuts, decides what the fragments actually are.
This guide explains recognition sites, cut offsets and fragment sizing, then works through a short sequence cut by EcoRI and BamHI in both linear and circular form. It closes with the practical reasons a real gel may not match the prediction.
Recognition sites and palindromes
Type II restriction enzymes bind a specific short sequence and cut DNA at or near it. Most of the common enzymes recognize palindromic sites, meaning the sequence on the top strand read 5′ to 3′ is identical to the sequence on the bottom strand read 5′ to 3′. EcoRI recognizes GAATTC, whose reverse complement is also GAATTC, so it can bind the site in either orientation.
Because of that symmetry, searching the entered top strand for the site is enough for a palindromic enzyme: a site on the bottom strand appears as the same text on the top strand. The reverse-complement search still matters for non-palindromic enzymes and for tools that may add them later, which is why a digest calculator should check both orientations rather than assume every site is symmetric.
Where the cut falls: offsets and overhangs
The recognition site says where the enzyme binds, but the cut offset says where it cuts. EcoRI cuts between the G and the first A of GAATTC on the top strand, leaving a four-base 5′ overhang of AATT. SmaI cuts in the middle of CCCGGG and leaves blunt ends, while PstI cuts CTGCA^G and leaves a four-base 3′ overhang. NotI recognizes an eight-base site and cuts after the second base, leaving a long overhang.
The overhang matters for downstream work. Compatible sticky ends let a fragment ligate into a vector cut by the same or a compatible enzyme, while blunt ends can ligate to any blunt end but usually with lower efficiency. When a calculator reports a cut position, it is really reporting the position of the break on the top strand, and the partner break on the bottom strand is offset from it.
Linear versus circular DNA
The topology changes the number of fragments. A linear molecule with n cut sites is divided into n + 1 fragments, because the two ends act as boundaries in addition to the cuts. A circular molecule with n cut sites gives n fragments, because there are no ends and the fragment from the last cut to the first wraps around the origin; a circle with no sites stays intact.
For example, a linear DNA with two EcoRI sites yields three fragments, while the same sequence as a circular plasmid with two sites yields two. A single cut in a circular plasmid linearizes it and produces one fragment equal to the full length, which is why a good digest of a plasmid often shows one band when only one enzyme cuts it.
Worked example: EcoRI and BamHI together
Take the 20 bp sequence CCCGAATTCAAAGGATCCCC. EcoRI recognizes GAATTC starting at the fourth base, and with its cut offset of one the top-strand cut falls after base 4. BamHI recognizes GGATCC starting at base 13 (counting from one), and with the same offset its cut falls after base 13. The linear molecule therefore has two cuts and three fragments: 4, 9 and 7 bp, which add up to 20 bp.
Treating the same sequence as a circle changes the answer. The two cuts at positions 4 and 13 leave 9 bp between them and 11 bp around the other side, so a circular molecule gives two fragments of 9 and 11 bp. Digesting with each enzyme alone gives different patterns again: EcoRI alone produces 4 + 16 bp on the linear molecule, and BamHI alone produces 13 + 7 bp. Comparing single-enzyme and double-enzyme patterns is how a cloning check confirms that both sites are present.
What the calculator does and does not model
A digest calculator searches the sequence for each selected enzyme, records the cut positions with the enzyme's offset, and computes fragment sizes for the chosen topology. It also reports ambiguity codes, because a site spanning an unknown base cannot be detected, and it rejects RNA letters rather than silently treating uracil as thymine and shifting every position.
The ideal prediction assumes complete digestion at every detected site. Real reactions can be affected by methylation, star activity at high enzyme concentration, partial digestion, enzyme kinetics and the way a supercoiled plasmid migrates on a gel. Fragment sizes are also calculated on the entered top strand, so staggered cuts leave overhangs and the two strands differ by a few bases at each end. Use the prediction as a design and checking tool, then confirm the real pattern against a properly digested sample.