Choosing restriction sites that actually cut

A restriction site printed on a plasmid map is a string of bases. A cut is what happens when the enzyme, the buffer and this DNA molecule agree. Choosing sites that actually cut means refusing the string until you have checked methylation, star conditions and whether both enzymes of a double digest can work in one buffer. This explainer is for that choice. How the pieces then become a colony is in plasmid cloning from insert to colony. The gel that shows whether the cut happened is in agarose gel electrophoresis for DNA.

Enzyme classes live in the molecular biology catalogue. The card for the enzyme you will pipette outranks a remembered recipe.

Recognition is not a cut on this DNA

Type II restriction enzymes bind a short recognition sequence and hydrolyse the backbone, leaving blunt ends or short overhangs. The map search finds the sequence. It does not know whether an adenine or a cytosine in that sequence was methylated inside the host, whether the enzyme stock still has activity, or whether the salt in the tube matches the enzyme. Those are properties of a preparation.

Count sites on the molecule you will digest. An empty-vector map can show one site that the recombinant no longer has, or miss a second copy inside the insert. A site you chose because it was unique on a parent file can cut twice on the DNA in the tube. The band pattern is then correct and the plan is wrong.

PCR products and plasmid minipreps are different substrates. Ordinary E. coli methylation is an in vivo event. A polymerase chain reaction does not copy Dam or Dcm marks, so an amplicon can cut at a site that the same sequence, prepared from a standard cloning strain, refuses. After you propagate the clone, the block can return. Choose the site for the DNA you will cut that day, and write down which host will carry it next.

Methylation overlap, as the card describes it

Dam methylation marks the adenine in GATC. Dcm methylation marks a cytosine in the motif CCWGG. When that motif overlaps a restriction site, the enzyme may be blocked, impaired or indifferent. The difference is documented on the enzyme card. This page will not invent a list of blocked enzymes, because the list belongs to the card for the protein in your freezer, not to a blog paragraph that cannot see your isoschizomer.

The usual workaround, when the card says the overlap blocks cutting, is to prepare the plasmid from a dam/dcm-deficient strain, or to cut a PCR copy that never saw those methylases. Switching enzymes to an isoschizomer is a different workaround only when the replacement card says the methylation does not block it. Guessing from the enzyme's surname is how people order a twin that fails the same way.

Public digest notes, including the Addgene restriction digest protocol, are a place to see how laboratories record enzyme, buffer and DNA. Follow the card you have. Their volumes are not your volumes.

Star activity when the reaction is outside the card

Star activity is cutting at sequences that only resemble the true site. Extra bands after a digest you called single are the clue. The ends those extra cuts create will ligate in ways the map did not draw, and colony screening then finds deletions and scrambled junctions.

Three conditions show up again and again, and they are conditions, not a reason to abandon the enzyme. Glycerol comes in with the enzyme stock. A large enzyme volume in a small digest pushes glycerol up, and many cards warn once that share of the reaction is high, often discussed around a few percent. Salt that is wrong for that enzyme, including a buffer that is too low in salt for an enzyme that needs it, opens star cuts for some enzymes and simply stalls others. Too much enzyme, or a long incubation past the time the card ties to the DNA mass, does the same favour for star activity. Use the unit range and the clock the card gives, then stop the reaction. Heat kill only if that card says heat kills this enzyme.

If the gel shows extra bands, do not ligate the smear and hope the right junction wins. Lower the enzyme, correct the buffer, shorten the time, and rerun a sample next to uncut plasmid. Addgene's molecular biology reference is useful background for how these controls are talked about. It is not a star-activity table you should copy in place of the card.

Double digests and the buffer both enzymes tolerate

A directional clone needs both enzymes to have cut, and the small stump between two close sites to have left the vector. Each enzyme has buffers where it is happy and buffers where it is sluggish. A double digest is honest when you pick a buffer both cards accept at useful activity. If no such buffer exists, digest in sequence. Heat-inactivate the first enzyme when its card allows, change the salt, and add the second. When heat will not kill it, purify the DNA between the two reactions so the first enzyme does not keep working under the wrong conditions.

Incomplete double digests are a cloning problem disguised as a ligation problem. Single-cut vector has compatible ends and recircularises. Uncut supercoiled plasmid transforms very well. Both outcompete a rare proper insert. On the plate you see colonies. In colony screening you see parent plasmid. Insert orientation tests then have nothing real to orient. The fix is earlier: a gel of the digest, with uncut plasmid in the next lane, before ligase ever enters the tube.

Sites that sit a few bases apart can fail even in a good buffer, because the second enzyme has little duplex left to hold. If the gel still looks like a single cut, separate the digests or move one site. Do not add more of both enzymes and call it thorough.

How an incomplete cut becomes a failed clone

Walk the digest forward one step. Linear vector plus insert, both fully cut, can ligate. A tube that still contains mostly uncut parent will transform as parent. The antibiotic plate cannot tell those colonies apart, because the marker never depended on the insert. People then screen twelve colonies, find twelve empty backbones, and redo the ligation with the same dirty digest.

Branch there. If uncut DNA remains, purify the linear band or repeat the digest. If only one of two enzymes cut, the cloning is no longer directional, and insert orientation becomes a coin toss you will have to test later. If extra star bands are present, stop. Those molecules are a library of unintended ends.

A control substrate with a known unblocked site tells you whether the enzyme stock is alive. Without that control, a failed cut on a methylated plasmid and a dead enzyme look the same.

Safety, and the gel you use to judge the cut

The digest itself is a small aqueous reaction. The judgement step is often an agarose gel. Ultraviolet light used to see the DNA damages eyes, skin and the DNA you may want to clone. Follow the stain safety note on the bottle you actually have. Do not look into a bare UV box. Containment of the host strain remains an institutional decision. Nothing here is a clinical or forensic method.

When the incubation clock lies

A long digest in a water bath assumes the bath stayed at the card's temperature. In a building where power drops, the bath cools and the clock on the wall does not. An incomplete cut from a cooled incubation looks, the next morning, like a ligation that refused the insert. If the power failed, do not restart by faith. Put a sample on a gel beside uncut plasmid and decide from the bands whether the sites actually cut. Heat left in a closed room is a different problem: an enzyme aliquot that sat warm on the bench can lose activity before it ever meets the DNA. Return stocks to the temperature on their label between uses.

What to send when you enquire about an enzyme

Name the recognition sequence, the enzyme name on the card, the host strain and whether it is dam/dcm proficient, whether you need a double digest, and the fragment sizes on the map. Say if the substrate is a plasmid prep or a PCR product. A synthetic insert that removes a problem site can be discussed through the custom gene synthesis enquiry reference. Confirmation of a later clone can be framed with the Sanger sequencing enquiry reference. Those pages are enquiry references. They do not mean EVRINTH runs synthesis or sequencing. Put the scientific requirement in the quote request and ask whether a quotation is possible.