Restriction ligation and Gibson assembly

Restriction ligation cuts at known sites and seals compatible ends with a ligase. Gibson assembly chews back overlapping ends, fills the gaps and seals them in one isothermal reaction. They fail differently, and colony screening does not fully rescue a bad design. This note is how to choose and how to debug the join. The path from that join to an archived colony is in plasmid cloning from insert to colony. It is a research explainer, not a copied kit insert.

Enzyme classes live in the molecular biology catalogue. The map belongs in the quote request if you are asking for the enzymes or a synthetic fragment.

Restriction ligation, in physical terms

Type II restriction enzymes recognise a short DNA sequence and cut within or beside it, leaving blunt ends or short single-stranded overhangs. Two ends ligate efficiently when the overhangs match and a 5-prime phosphate is present for the ligase to join to a 3-prime hydroxyl. T4 DNA ligase is the usual enzyme class. It uses ATP. A ligation buffer that has been warmed and frozen until the ATP is dead is a silent reason for an empty plate.

Sticky ends find each other more readily than blunt ends. Directional cloning uses two different enzymes so the insert can enter in one orientation, provided both cut and the spacer between the sites was removed. Sites that sit too close together sometimes cut poorly in a double digest. If the gel still shows mostly single-cut vector, separate the digests or change the plan.

Star activity is cutting at sites that are merely similar to the true site. It shows up when the enzyme, the glycerol, the buffer or the incubation is outside the conditions that enzyme was characterised for. Extra bands after a "single" digest are the clue. Those extra ends will ligate in ways the map did not draw.

A phosphatase on the vector ends stops the vector from religating, because ligase needs the phosphate. The insert must still bring phosphates, which a normal restriction cut or a phosphorylated primer provides. Dephosphorylate the vector, not the insert. Clean the enzyme away afterwards so it does not nibble the insert in the ligation tube.

Incomplete cleanup also fails later. Active restriction enzyme carried into the ligation recuts the product as it forms. A column or a gel slice between digest and ligase is there for that reason. Confirm sizes on an agarose gel before the ligation so you are not joining the wrong band.

Gibson assembly, in physical terms

Gibson assembly uses three activities together. A 5-prime exonuclease exposes a single-stranded overlap, complementary overlaps anneal, a polymerase fills the gaps and a ligase seals the nicks. The classical reaction is isothermal, often near 50 Celsius in the original formulation. Use the time and temperature on the sheet that belongs to the tube you opened.

The overlaps are designed, usually into the primers that amplify each fragment. They need to be long enough to anneal after the chew-back and unique enough that fragment A does not prefer fragment C. Repeated promoters, identical tags on both ends, or a poly-A stretch make scrambled circles that still transform. A quick check is to read the overlap sequences next to each other before you order the primers. If two overlaps are the same sentence, redesign.

A PCR product that is mostly primer-dimer will assemble the dimer. Gel-purify when the PCR is not already a single band. A huge excess of one piece drives partial assemblies that skip a neighbour.

Gibson reactions do not enjoy being redesigned in the pipette. Adding "a little more enzyme" to a mix that sat warm during a shipping delay or a freezer failure is not a controlled experiment. Run a known assembly beside the new one if you suspect the mix. The NEB product pages illustrate how enzyme classes are described by their makers. They are not a protocol to copy volumes from.

Where each method fails

Colony screening starts after these controls, not instead of them. A vector-only reaction that already yields a carpet of colonies means the real plate will be mostly empties, and colony PCR will mostly disappoint you. Fix the join. If the vector-only plate is quiet and the real plate has colonies, screen a handful by PCR across a junction or by digest. Then sequence. The Sanger sequencing enquiry reference is a way to specify that read. It is an independent method reference. Ask whether a quotation is possible.

A positive colony PCR can still hide a deletion between the primers. A band that is only "close enough" on a low-percentage gel is how a missing domain gets archived. Run a percentage that can see the difference, or sequence earlier.

Design habits that prevent the rerun

For restriction maps, pick enzymes whose buffers can be made to agree, or accept a sequential digest. Heat-inactivate only the enzymes that actually die at that temperature. Many do not, and purification is then required. Keep glycerol low enough to stay out of the star-activity regime for that enzyme.

For Gibson maps, put the overlap on a sequence that is unique in the construct. Avoid the same restriction-site scar, the same primer tail, or the same epitope tag as the homology on both sides of a fragment. Check the assembled file by aligning the overlaps back to a GenBank record or to your own reference so a one-base frameshift is visible on paper before it is visible in a failed protein gel six weeks later.

If the insert is long or repetitive, a synthetic clone may be the calmer start. The custom gene synthesis enquiry reference is a place to specify the sequence and the backbone. It does not mean a synthesis run is already happening. You would still verify the junctions of whatever returns.

Safety and the bench climate

Restriction enzymes and assembly mixes are laboratory reagents, not household chemicals. Follow the storage on the tube. Several are in glycerol and do not enjoy being left in a warm pipette. In a hot room, a "cold block" that has been on the bench since morning is an ambient block. Chill it again, or the ligase works while you are still adding fragments and then dies in a way you cannot interpret.

Waste agar with antibiotic-resistant bacteria is a biosafety waste, even when the insert is a routine tag. Follow institutional rules. This page is not an approval for a hazardous construct.

Power cuts matter when a freezer holding enzyme aliquots warms and someone refreezes it without a note. Mark the aliquot and prove it with a control assembly before it touches a precious insert. Humidity does less to the join than heat does, but condensation in a tube cap will change a small-volume assembly if you do not spin the tube down before you open it.

What to send when asking for enzymes or fragments

Name the method, the number of fragments, the overhangs or the overlap length class, the vector marker, and how you will screen. Ask for the enzyme class or the assembly mix class you have already designed around. Do not ask a catalogue to choose between ligation and Gibson without the map. The choice is a property of the ends.