Articles

Gibson Assembly: Primer Design and Troubleshooting

Enzyme mechanisms, overlap calculations, multi-fragment limits, and avoiding common assembly failures

Molecular Cloning

Gibson Assembly: Primer Design and Troubleshooting

Gibson Assembly joins several DNA fragments in a single reaction at one temperature. Daniel Gibson and colleagues published it in 2009, and it removed the two things that made traditional cloning slow: the need for convenient restriction sites, and the scars those sites leave behind.

The protocol looks simple, which is part of the problem. The reaction is only as good as your overlaps and your DNA purity, and both are settled before you open the master mix. Most Gibson failures originate at the design stage or during a gel extraction.

This guide covers what the three enzymes do, how to design the overlaps, what actually happens as you add fragments, and the failures that look like enzyme problems but are not.

At a glance

  • Mechanism: Three enzymes at 50 °C. One digests the ends, one fills the gaps, one seals the nicks.
  • Design: 20 to 40 bp overlaps, balanced GC content, and each overlap matching only its intended partner.
  • Watch for: Guanidine salt from gel extraction, uncut vector, and efficiency falling as the fragment count rises.

How Gibson Assembly works

The master mix contains three enzyme activities working together at 50 °C:

  1. T5 exonuclease digests the 5' ends of each fragment, exposing single-stranded 3' overhangs. Because you design the fragment ends to share sequence, these overhangs find each other and anneal.
  2. High-fidelity DNA polymerase (typically Phusion) fills the remaining gaps. The 50 °C incubation does two jobs here: it suits the polymerase, and it inactivates the heat-labile T5 exonuclease early in the reaction (reported to be within the first 10 to 15 minutes), stopping ends from being digested indefinitely.
  3. Taq DNA ligase seals the nicks. Taq ligase rather than T4 is used because T4 denatures quickly at 50 °C. Taq ligase also uses NAD+ as its cofactor rather than ATP, and NAD+ degrades with repeated freeze-thaw cycles or poor storage. Keeping the master mix in single-use aliquots protects the ligation step, which is the step you do not notice failing until the plate is empty.

What comes out of the tube is a covalently closed plasmid, ready to transform.

Primer design

Gibson depends entirely on matching sequence between adjacent fragments, and you build this in by PCR.

A Gibson primer has two sections. The 3' end matches your template and does the priming. The 5' tail carries the 20 to 40 bases that match the adjacent fragment or the vector.

Aim for an overlap of 20 to 40 bp. Below about 15 bp, the overlap is usually too unstable to anneal reliably at 50 °C. Much longer than 40 bp and you are paying for primer length and increasing the chance of hairpins or primer dimers during the initial amplification, with no gain in the assembly.

Two things are worth checking before you order:

  • The melting temperature of the overlap: This needs to be high enough for the annealed overlap to hold at the reaction temperature. Published guidance differs between mixes, so take the figure from the protocol for the mix you use rather than from a general rule.
  • The uniqueness of each overlap: In a multi-fragment assembly, two overlaps that resemble each other will produce rearranged products, and they will do it efficiently. This is the error that gives you plenty of colonies, all of them wrong.

When you set up the PCR, calculate the melting temperature from the template-binding section of the primer alone. The 5' tail does not anneal until it has been copied into the product, so a Tm taken across the whole primer puts the annealing temperature far above the point at which the primer can bind, and the first cycles give you nothing to assemble (see PCR Fundamentals for Cloning).

What happens as you add fragments

Commercial literature often shows five or six fragments assembling into a vector in one reaction. This does work, but less often than the diagram suggests.

Every extra fragment adds a junction that has to find the right partner, and another opportunity for one piece to drop out. Efficiency falls as the fragment count rises, and it falls faster than most people expect.

For routine subcloning, two or three fragments including the vector is reliable. For five or six, build it in stages. Assemble the first group into one intermediate vector and the second group into another, verify both by sequence, then join the two verified halves in a final reaction. You spend one extra round of screening and you stop guessing which junction failed.

Cost is worth a thought too. Many labs make up their own Gibson mixes from bulk enzymes for routine two-fragment assemblies and keep the commercial kit for the difficult constructs.

Plan it first

Multi-primer design is hard to get right by eye, and one miscalculated overlap or shifted reading frame can cost you a week.

In Photo51, you can lay out the vector and inserts, map the overlap sequences onto the fragment ends, and inspect the finished construct before you order any oligos (or watch our step-by-step tutorial: Gibson Assembly Cloning in Photo51). Check each junction, read the frame across it, and look at whether each overlap matches the fragment you intended and nothing else in the set.

Troubleshooting

No colonies, and your overlaps are AT-rich

Symptom: Your design looked correct and your plates are empty.

Cause: An AT-rich overlap. A and T pair through two hydrogen bonds rather than three, so an AT-rich overlap is less stable and may not hold together at 50 °C.

Fix: Extend AT-rich overlaps to 30 or 40 bp to compensate, or move the junction a little along the sequence to a more balanced region. Moving the junction is usually the better answer, because it costs nothing and does not lengthen the primer.

Nothing at all, and your overlaps are very GC-rich

Symptom: Your plates are empty, and the overlaps in your design are above roughly 80% GC.

Cause: Very GC-rich single strands fold into stable hairpins. The reaction runs at a constant 50 °C with no denaturation step, so there is nothing to melt them, and the ends are never available to anneal.

Fix: Redesign the overlaps to a balanced GC content, around 40% to 60%. This is a design problem, so there is no additive or incubation change that will rescue it.

High empty-vector background

Symptom: You have dozens of colonies, and every colony PCR shows an empty vector.

Cause: Uncut vector carried through to the transformation, or a digested backbone whose two ends match each other and can rejoin.

Fix: Gel-extract the linearized vector to separate it from uncut supercoiled plasmid. If you amplified the backbone from a plasmid template, digest with DpnI to remove the methylated template. Cutting with two different enzymes leaves incompatible ends, which stops the backbone re-circularizing.

Poor purity readings after gel extraction

Symptom: Your gel-extracted fragments give a 260/230 ratio below about 1.5, and the assembly gives nothing.

Cause: Guanidine salt carried over from the gel-dissolving buffer. In Gibson this costs you twice: the salts inhibit all three enzymes, and the raised ionic strength stabilizes exactly the hairpins and mispaired overlaps described above, which a constant 50 °C reaction has no denaturation step to melt.

Fix: Where your PCR gives a clean band of the right size, use a column cleanup and skip the gel entirely. When you do have to gel-extract, the washing technique is what removes the salt, as set out in The Plasmid Cloning Workflow. Re-quantify after cleanup, because absorbance readings on contaminated DNA overstate how much you have, and you will add less than you think.