Colony PCR for screening clones

Colony PCR asks whether a colony from a cloning plate carries a plasmid of the size you expect, before you spend a miniprep and a sequencing reaction on every colony on the dish. It uses the same endpoint chemistry as any other PCR, with a colony lysate as the template. A band is a screen. It is not a sequence, not proof of orientation, and not proof that the open reading frame is intact. Where this screen sits in the cloning path is described in plasmid cloning from insert to colony. The cycle itself is in how polymerase chain reaction works.

What the plate has already told you

Colonies on antibiotic agar show that a resistance marker was present. They do not show that an insert is present. Empty vector grows on the same drug. Ampicillin plates add a further trap: beta-lactamase can leak into the agar and let nearby cells that never took up a plasmid form small satellite colonies. Those satellites are not clones. How drug selection and blue-white colour should be read is set out in antibiotic selection and blue-white screening. Pick a well-isolated large colony, not a pinprick in a cleared halo, and not a colony from a lawn.

Blue or white colour, on a system built for it, only chooses which colonies are worth this PCR. A white colony can be a damaged lacZ fragment with no insert. A blue colony can hide a small in-frame insert. Screen the DNA either way if the claim needs the insert.

Flanking primers and insert primers answer different questions

Flanking primers bind the vector on either side of the cloning site. On an empty plasmid they copy only the short stretch between them. On a plasmid that took an insert they copy that stretch plus the insert. You see both outcomes as different sizes, which is what you want when the question is empty versus occupied. You do not learn orientation unless the size somehow depended on direction, which it usually does not. You do not learn whether the junctions are clean.

Insert primers bind inside the fragment you ligated. A true empty vector gives no product. A colony whose plasmid contains that sequence gives the internal amplicon. This is a sharp screen for "is this sequence somewhere in the well," and a weak screen for "is it in my vector." A contaminating fragment in the mix will amplify. An insert of a different gene will not, which is useful, and a truncated insert that lost the primer site will look empty, which is a false negative. Pair insert primers with the empty-vector control so silence means something.

One primer in the vector and one in the insert asks about orientation and presence together. The wrong orientation does not make a product. The size should match the map from the vector primer to the insert primer. Run the opposite orientation as a separate pair if you need to see the other direction rather than infer it from a blank. A blank can also be a failed lysis.

Write the expected sizes down before you load the gel. Include the bases between flanking primers on the empty map, not the insert length alone. A 500-base-pair insert with flanking primers 150 bases apart on the empty vector is a 650-base-pair product beside a 150-base-pair empty product. Those two must be separable on the percentage you pour.

Lysis is a class, not a recipe to invent

The polymerase has to see plasmid DNA, and a whole colony is mostly cells, agar and inhibitors. Use a lysis class the mix tolerates. Common classes are a brief heat step on a small suspension of cells in water or buffer, or a colony-PCR mix sold to accept intact cells. The decision-level rule is a barely visible touch of colony, not a glob. Too much biomass inhibits the reaction and turns real clones into false blanks.

Patch first. Stab or streak the same colony onto a numbered selective plate, then suspend what remains on the tip. If the PCR looks promising you still have the patch to inoculate. If you suspended everything, a positive band has no cells behind it.

Do not pool colonies to save reactions. A band from a pool does not name the colony, and a mixed pick already looks like two plasmids in one chromatogram later. One colony, one tube, one patch.

Controls that make the screen interpretable

Run an empty-vector colony through the same lysis and the same primers. With flanking primers it must produce the short backbone size. If it produces the long size, either it was not empty or the long band is contamination. With insert primers the empty colony must be blank. If it lights up, the insert primers are amplifying something that does not depend on your ligation.

Run a no-template tube with no cells at all. It must be blank at every size you intend to score. Colony PCR is famous for false positives because the ligation fragment and yesterday's amplicon are often on the same bench as the master mix. A no-template band at the insert size ends the session. Remake the mix away from post-PCR work, and do not score the colonies beside it.

A failed empty-vector reaction, with a blank where the backbone band should be, means the lysis or the flanking primers failed. Blank sample lanes then mean nothing. Fix the lysis or the primers before you call twenty colonies negative.

From the gel to a clone you can keep

Score sizes against the map and a marker that brackets both the empty and the insert products. A colony that shows both sizes may be a mixed pick. Restreak the patch to single colonies and screen again. A colony that shows an unexpected third size may be a deletion or a double insert. Do not average it into the "close enough" pile.

Take one or more size-matched patches to a miniprep. A diagnostic digest can confirm that extra DNA of the right length is in the plasmid. Sanger sequencing across the junctions is what shows the bases. One read does not cover a long insert. Mixed peaks mean the template was mixed. The Sanger sequencing enquiry reference is a way to frame that check. It is an independent method reference, not a statement that a sequencing bench is already running. Keep plasmid DNA as a backup separate from the plate you screened.

If every colony matches empty vector, the ligation or assembly did not deliver insert into the cells you plated. More colony PCR will keep reporting that. Return to the join, the molar ratio class, and the empty-vector control on the original plate. If no colony amplifies, including the empty-vector control, the screen is broken. Check lysis and primers. If the no-template tube is the only tube with a band, you do not have twenty recombinants. You have a contaminated mix.

Safety, warm plates and satellites

These are laboratory cloning strains, still living organisms, and the insert may change the institutional biosafety decision. Inactivate plates and cultures by the method your institution specifies. Do not tip antibiotic waste down a drain because the PCR already finished. This screen is research practice, not identification of a clinical isolate.

Plates that sit on a warm bench grow satellites around true ampicillin colonies and, in humid weather, sweat until neighbouring colonies merge. A pick then becomes a mixture that colony PCR will sometimes show as two bands and sometimes hide inside one. Screen colonies the day you need the answer, from a plate that has not been left out, and keep the patch plate from becoming the next warm incubation you forgot. A power cut to an incubator means you do not actually know how long selection ran. Look at the colonies before you lyse fifty of them from a lawn.

What an enquiry should include

Say you are screening bacterial colonies, the vector and the expected insert length, whether you need flanking-primer interpretation or only an enzyme for colony PCR, and that positives will go on to sequencing. Enzyme families are in the molecular biology catalogue. Send the lengths with the quote request. A polymerase that cannot tolerate colony debris is the wrong convenience. The nucleic acid analysis pathway is the context if the insert itself was amplified from a difficult extract. The multiplex PCR enquiry reference is relevant only if one colony reaction must report several junctions at once. Most screens are clearer as one pair and a size you can tell from the empty backbone.