Gibson Assembly vs. Golden Gate vs. Restriction Cloning: Choosing a Strategy
6 min read · Updated July 10, 2026
Picking a cloning method is often decided by habit rather than by what the construct actually needs. Restriction/ligation cloning, Gibson Assembly, and Golden Gate Assembly all join DNA fragments in a defined order, but they differ in scar sequence, fragment-number ceiling, and upfront design cost. This guide covers how each method works at the molecular level, when each one actually wins, and how to check an assembly design before committing bench time to it.
How restriction and ligation cloning works
Restriction and ligation cloning cuts vector and insert with restriction enzymes chosen so the resulting ends are compatible — matching sticky overhangs, or blunt ends — then joins them with T4 DNA ligase. It is the cheapest and most direct option when a construct needs one or two fragments and the parts already carry unique sites that do not cut anywhere else you need intact. The catch is that the whole strategy depends on site availability: as soon as a site you want to use also appears inside the insert or vector backbone, or two fragments do not share compatible ends, you are into partial digests, added linkers, or a different plan. Ligation efficiency and background also drop off past two or three fragments, and the digest can leave the restriction site itself behind as a scar at the junction — usually harmless, but it is fixed sequence you did not choose.
How Gibson Assembly works
Gibson Assembly replaces restriction sites with designed homology. Fragments are amplified, typically by PCR, with 15-40 bp of sequence at each end that matches the neighboring fragment, then combined in a single tube with a three-enzyme mix and incubated isothermally — commonly around 50°C for well under an hour. A 5' exonuclease chews back the 5' end of each fragment to expose single-stranded 3' overhangs; complementary overhangs from adjacent fragments anneal; a DNA polymerase fills in the remaining single-stranded gaps; and a DNA ligase seals the nicks to give a continuous, covalently closed product. Because the junction sequence is whatever you put in the primer tails, the joint carries no leftover recognition site, and the method does not depend on finding a compatible restriction site anywhere in the parts. The practical ceiling is fragment number: five or six fragments in one pot is routine, and beyond that junction mis-pairing and efficiency become more of a concern.
How Golden Gate Assembly works
Golden Gate Assembly is still restriction-based, but it uses Type IIS enzymes — BsaI, BsmBI, and BbsI are the common ones — that cut outside their own recognition sequence rather than within it. That has two consequences. First, the 4 bp overhang left behind is sequence you choose, not sequence dictated by the enzyme, so a whole set of parts can be designed with defined, ordered overhangs that assemble correctly, and only in the intended order, in one pot. Second, because the recognition site sits outside the fragment that ends up in the final construct, the assembled product is scarless in the same sense Gibson's is, even though the mechanism here is enzymatic digestion and ligation rather than exonuclease-polymerase chemistry. Golden Gate comfortably handles more fragments in one reaction than restriction/ligation cloning, and often more than a typical Gibson reaction, which is why it is the default for combinatorial or library-style work. The tradeoff is upfront design cost: every part needs to be checked for extra internal copies of the chosen Type IIS site — a step usually called domestication — before it can be reused reliably across builds.
Choosing between the three
Which method to reach for is mostly a function of fragment count and how reusable the parts need to be:
- Restriction/ligation cloning — one or two fragments, good unique sites already exist in the parts, no need to reuse the same fragments elsewhere.
- Gibson Assembly — several fragments that need to join at arbitrary junction sequence with no dependence on a specific restriction site, for a scarless, one-off construct.
- Golden Gate Assembly — many fragments, especially when the same parts get reassembled into different combinations repeatedly, and you can invest the design time to domesticate parts once.
Checking the design before you build it
Whichever method you pick, the failure mode is the same: an assembly that looks fine on paper but breaks at a junction. A homology arm one base short of matching, a Type IIS overhang that does not pair with its neighbor, or a restriction site buried inside an insert you did not screen can all turn into a failed transformation or, worse, a construct that grows but does not express correctly. SeqBench's Cloning Simulator assembles your fragments in silico for Gibson, Golden Gate, or restriction cloning and designs the junction primers needed to build them, so you can see the assembled sequence and junction layout before ordering anything. Pair it with the Restriction Sites tool to check for existing sites in your parts or leftover Type IIS sites that still need domesticating, and the Construct QC Linter to check the assembled sequence for correct frame across each junction and for unwanted restriction sites or premature stops introduced at a joint.
Frequently asked questions
Is Golden Gate assembly scarless?
Yes in the sense that matters: the Type IIS recognition site sits outside the fragment that ends up in the final construct, so the joined sequence itself carries no leftover site, only the 4 bp overhang you designed.
How many fragments can Gibson assembly do in one reaction?
Five to six fragments in a single pot is routine; beyond that, mis-pairing between homology arms and lower efficiency become more likely, and Golden Gate or a sequential build is usually more reliable.
What is domestication in Golden Gate cloning?
Domestication is removing or silently mutating extra internal copies of the chosen Type IIS recognition site (BsaI, BsmBI, BbsI, etc.) from a part so the enzyme only cuts at the two intended flanking sites.
When should I use restriction cloning instead of Gibson assembly?
When you only need one or two fragments and the parts already carry unique, compatible restriction sites — it is cheaper and skips designing homology-arm primers, but it stops scaling well past two or three fragments.
Related references
Common restriction enzymes: recognition sites, cut positions, NEB buffer activity, star activity and an interactive double-digest buffer finder.
Nucleotide ambiguity codes and their complements.
Selection markers, mechanisms and working concentrations for cloning.
Related tools
Assemble fragments and design junction primers for Gibson, Golden Gate or restriction cloning.
Scan a coding sequence for premature stops, cryptic RBS/polyA signals, unwanted restriction sites, GC extremes and repeats.
Find recognition and cut sites for common restriction enzymes.