Articles

In-Fusion Cloning: Primer Design, Seamless Assembly and Troubleshooting

Takara's 15 bp homology system, in vivo nick repair, competent cell sensitivity, and troubleshooting

Molecular Cloning

In-Fusion Cloning: Primer Design, Seamless Assembly and Troubleshooting

In-Fusion is Takara's seamless cloning system. Like Gibson Assembly and NEBuilder HiFi, it joins DNA fragments through matching sequences at their ends rather than through restriction sites, so you are not limited to the sites your vector happens to contain.

Three things set it apart: the homology it needs is shorter (15 bp rather than 20 to 40), the in vitro incubation is about 15 minutes, and the final bonds are not made in the tube at all. The reaction gives you an annealed, nicked circle, and the bacteria finish the job after transformation.

This last point is easy to overlook, and it changes how the method behaves. A nicked plasmid is more fragile than a closed one, so In-Fusion can depend more on the health of your competent cells than Gibson or Golden Gate do.

At a glance

  • Mechanism: An exonuclease activity trims back fragment ends to expose single-stranded overhangs, which anneal through matching sequence. E. coli repairs the nicks after transformation.
  • Design: 15 bp of homology for a single insert, 20 bp for three or more fragments, placed at the very end of each fragment.
  • Watch for: EDTA in your elution buffer, uncut vector carried into the transformation, and tired competent cells.

How In-Fusion works

The reaction runs at 50 °C for about 15 minutes, in one tube, with no temperature cycling.

The enzyme mix contains an exonuclease activity that trims back one strand at each end of your double-stranded fragments. This exposes single-stranded overhangs. Because you design the fragment ends to share sequence with their neighbors, the overhangs find each other and anneal.

Takara does not publish the composition of the mix, so we cannot say what stops the trimming. What you need in practice is where it stops: once two complementary overhangs have annealed into double-stranded DNA, the ends are no longer available to be trimmed further, and the structure remains.

What comes out of the tube is a circular molecule held together by base pairing, with nicks still in the backbone. When you transform this into E. coli, the host's own repair machinery fills the remaining gaps and seals the nicks.

This is the practical difference from Gibson and Golden Gate, which both give you a covalently closed plasmid before you ever touch a tube of cells. With In-Fusion, the molecule that enters the cell is vulnerable, and cells that are stressed or old repair it slowly. A batch of competent cells that gives acceptable results with a closed plasmid can give you nothing at all here.

Primer design and the 15 base pair rule

You build the overlaps during PCR, so the design work happens before you amplify anything.

An In-Fusion primer has two sections: the 3' end anneals to your template and does the priming, and the 5' tail does not bind the template at first; it carries the sequence that matches the end of the adjacent fragment or the linearized vector.

Takara's recommendations depend on how many pieces you are joining:

  • One insert into a linearized vector: 15 bp of homology.
  • Three or more fragments: 20 bp of homology.

Two points about the length are worth noting. First, the homology has to sit at the absolute end of the fragment, not a few bases in. An overlap set back from the end will not be exposed properly by the trimming step. Second, longer is not better here. Shorter overlaps mean shorter primers, which cost less and are less likely to form hairpins or dimers during the amplification. A 15-nucleotide single strand also has less opportunity to fold back on itself than a 40-nucleotide one, which may be why In-Fusion performs reasonably well on sequences that give longer-overlap methods trouble.

Calculate the melting temperature (Tm) from the template-binding section of the primer alone (see PCR Fundamentals for Cloning). The 5' tail only starts to contribute once it has been copied into the product, so a Tm measured across the whole primer sets your annealing temperature too high for the first few cycles and leaves you with no product to assemble. This is a common reason a long cloning primer gives an empty gel, and it has nothing to do with the assembly.

Cost and when to use it

The In-Fusion mix is proprietary and costs considerably more per reaction than restriction enzymes and T4 ligase, or than a Gibson mix made up from bulk enzymes.

Most labs that use it do not use it for everything. It earns its cost on the constructs where the alternatives struggle: high-GC targets, very long PCR products, and libraries where you cannot afford a low-efficiency reaction. For routine two-part subcloning, a cheaper method will usually do the same job.

The 15-minute incubation is real, but it is worth being clear about what it saves. The reaction is a small part of the week. DNA preparation, transformation, overnight growth, colony screening and sequencing take as long as they do with any other method. In-Fusion is fast in the tube, but it is not a fast experiment.

Plan it first

Fifteen base pairs is short enough to design by hand and short enough to get subtly wrong by hand. A tail copied from the wrong strand, an overlap placed a few bases inside the fragment end, or a junction that shifts the reading frame by one will all still give you colonies, because the plasmid closes and carries the resistance marker whether or not the junction is what you wanted. Your plate looks like success; the mistake appears when you sequence, a round of screening later.

In Photo51, you can lay out the linearized vector and your inserts, map the 15 bp overlaps onto the fragment ends, and inspect the finished construct before you order any oligos (or watch our tutorial: InFusion Cloning in Photo51). Check each junction, read the frame across it, and confirm that each overlap matches the fragment you intended and nowhere else.

Troubleshooting

You get no colonies, and your DNA was eluted in TE buffer

Symptom: You quantified your DNA, calculated the amounts, ran the 15-minute incubation, and your plates are bare.

Cause: EDTA. The exonuclease activity in the mix needs magnesium, and TE buffer contains EDTA, which binds magnesium and takes it out of the reaction.

Fix: Never elute or resuspend DNA in TE for an In-Fusion reaction. Use nuclease-free water or an EDTA-free buffer such as Tris-HCl. If the DNA is already in TE, clean it up on a column and elute again before you set up the reaction.

Plenty of colonies, and every one is empty vector

Symptom: You have dozens of colonies, and colony PCR shows the backbone with no insert.

Cause: Uncut plasmid surviving into the transformation, or a digested backbone whose two ends can join to each other. Both carry the same resistance marker as your intended product, so selection does not remove them.

Fix: Cut the vector with two enzymes that leave incompatible ends, so it cannot close on itself. Gel-extract the linearized backbone to separate it from uncut supercoiled plasmid. If you amplified the backbone by PCR from a plasmid template, digest the product with DpnI to remove the methylated template.

A few small colonies, or none, with a vector and insert you have checked

Symptom: Your design is right, the DNA is clean and in the correct buffer, and you still get almost nothing.

Cause: Your cells. Because In-Fusion delivers a nicked circle, the construct has to survive long enough inside the cell for the host to repair it, and stressed or old competent cells do this slowly.

Fix: Use a fresh batch of high-efficiency competent cells, kept at the correct temperature and thawed on ice. Run the kit's positive control in the same transformation, so that a bare plate tells you whether the problem is your assembly or your cells. If you prepare your own competent cells, this is a method worth buying commercial cells for.

The clone is correct but the protein is wrong, or absent

Symptom: Sequencing confirms the insert is present and intact, but the protein does not express, or runs at the wrong size.

Cause: The reading frame shifted at a junction. Seamless means the method adds nothing of its own. It does not mean your junction is in frame.

Fix: You cannot fix this at the bench, so catch it in the design. Translate the construct across each junction before you order primers, and where your insert continues an upstream coding sequence, such as a tag, check that the bases between them come to a multiple of three.