Getting Started with Golden Gate Assembly

Golden Gate Assembly is a multi-fragment, one-step cloning method used often for modular, hierarchical assembly of DNA into large constructs. But what does that mean? Is it useful for routine cloning? Why might a typical molecular biology lab choose Golden Gate Assembly over NEBuilder^® or traditional restriction enzyme cloning? This article aims to demystify Golden Gate, explains when and how to use it, and gives you everything you need to add this valuable method to your molecular biology toolbox.

 

What is Golden Gate Assembly?

Golden Gate Assembly (also known as Golden Gate Cloning) methods were invented by Marillonnet and coworkers in 2008 [1-3]. Like Gibson Assembly^®, it permits the seamless joining of multiple DNA fragments into a destination vector in a single reaction step. Golden Gate Cloning is often used for multi-fragment cloning as a cousin of Gibson cloning, NEBuilder, In-Fusion^®, and other homology-directed cloning approaches [4, 5]. These methods use regions of sequence homology at the ends of each fragment (15-30 nt), allowing them to overlap with matching regions on the vector or adjoining fragments in the assembly reaction. Complementary overhangs are generated dynamically within the homologous regions by an exonuclease and a polymerase, followed by the subsequent sealing of nicks by a ligase. Like Gibson Assembly, Golden Gate Assembly allows the precise assembly of multiple fragments in a defined order, without the need for multi-step protocols, and creates “scarless” constructs free from restriction site sequences at fusion sites.

 

Golden Gate Assembly is a Restriction Enzyme Cloning Method

Unlike the homology-directed cloning approaches described above, Golden Gate Assembly relies on restriction enzymes to generate complimentary overhangs for joining DNA fragments. It shares many similarities with classic restriction-based cut-and-paste cloning methods in terms of insert and vector design requirements (Figure 1: Golden Gate Assembly vs. Traditional Restriction Cloning). In both methods, a destination vector contains a cloning site incorporating the restriction enzyme recognition sequence for the enzymes you are using, and the vector must be free of other instances of the same recognition site(s) or it will be cut at these positions during the reaction. Inserts are typically amplified using PCR primers that add the desired flanking restriction recognition sites, and like the vector, the internal sequences must also be free of the chosen restriction enzyme recognition site.

Figure 1: Golden Gate Assembly vs. Traditional Cloning

 

In traditional restriction enzyme cloning, the vector is digested and purified resulting in a linearized backbone. The target insert is excised from another vector or amplified by PCR to incorporate the needed restriction sites, then digested and purified. The vector and insert are mixed in a ratio designed to optimize the yield of a circular product containing a single copy of the insert upon ligation, followed by transformation to yield single (clonal) assemblies. The most common restriction enzymes used in cloning are Type IIP enzymes, such as EcoRI, HindIII, BamHI, and many more, which recognize and cut palindromic sequences, generating short, self-complimentary overhangs or blunt ends. This feature can lead to vectors and inserts self-ligating, forming oligomers and other higher-order structures. To prevent unwanted ligation events, multiple restriction enzymes must be employed, the vector-to-insert ratio must be optimized to maximize the yield of the single-insert product, and vectors can be dephosphorylated to prevent antibiotic-resistant self-ligation products.

  

Type IIS Restriction Enzymes Enable the Unique Features of Golden Gate Assembly

If both traditional restriction enzyme cloning and Golden Gate Assembly use a restriction enzyme and ligase, why does traditional cloning require multiple steps, while Golden Gate can achieve this in a single-step reaction, joining multiple fragments in a pre-determined order? The key difference is in the type of restriction enzymes used. Golden Gate cloning uses Type IIS enzymes, such as BsaI-HF^®v2 (NEB #R3733), BsmBI-v2 (NEB #R0739), and PaqCI^® (NEB #R0745), which recognize non-palindromic sequences and, importantly, cut outside these recognition sites in a sequence-independent manner (Figure 2: Type IIP vs. Type IIS Restriction Enzymes). The most commonly used Type IIS enzymes leave 3 or 4 nucleotide 5′-overhangs, but unlike Type IIP enzymes, these can be any sequence and are independent of the binding site sequence specificity. Thus, you can select non-palindromic overhangs and prevent insert/vector self-ligation – no possibility of oligomers! By carefully selecting the junctions on each fragment, multiple inserts can be joined in a user-defined order, in a single step (see NEB’s NEBridge^® Ligase Fidelity Tools for more on how to ensure high accuracy assembly of multiple inserts).

How does Golden Gate enable restriction and ligation in a single reaction? In traditional cloning, the ligase joins the generated self-complementary overhangs to any other copy with the same overhang – whether it’s the destination vector, another insert copy, or the fragment you just cut from the vector. If a Type IIP restriction site is present, the regenerated cut site becomes a substrate for cleavage, resulting in a cycle of cutting and re-ligation that hampers assembly. However, Type IIS enzymes cut away from the recognition site and the entirety of the recognition site is in one product of the restriction reaction. If you place the recognition sites such that the cut insert fragments and vector backbone no longer retain the recognition sequences, ligation of the desired fragments form seamless junctions without restriction sites. This design ensures the desired product cannot be cleaved again, while any reformed empty vector or insert fragments can be digested. As the reaction cycles through cutting, melting, annealing, and ligation, the desired product accumulates without the need for additional steps, purifications, or phosphatase treatments required in traditional cloning. In Golden Gate Assembly, a vector that fails to join with the insert is re-linearized by the restriction enzymes, reducing the background of empty vectors in the final product.

Figure 2: Type IIP vs. Type IIS Restriction Enzymes

 

In summary, Golden Gate Assembly conceptually requires much of the same design logic as classic cloning:

• A cloning site on the vector containing the restriction specificity to be used
• Flanking sites on the insert with the restriction sites
• No other sites for the same enzyme elsewhere in the vector or insert sequences

 

But Golden Gate Assembly has several advantages:

• Simplified protocol
• Low hands-on time
• No purification step between restriction and ligation
• Scarless ligation
• Low background
• Potential for modular design [2, 6, 7]
• Easily scalable to multiple inserts and combinatorial libraries [3, 8]

One limitation of Golden Gate Assembly is that fewer Type IIS enzymes have been validated for this method compared to the available enzymes for traditional cloning – there are only about a half dozen commonly used Type IIS enzymes, most with 6-base recognition sites. In traditional cloning, it is usually easy to find enzyme pairs that do not cut within your insert sequence, and many standard vectors have multiple cloning sites (MCS) compatible with these enzymes. In contrast, Golden Gate has fewer standard vectors and may require modification of the insert DNA to remove internal recognition sites. In the next section, we will cover in detail how to obtain and prepare suitable vectors and inserts for your Golden Gate Assembly reaction.

 

Design and Preparation of Vector and Insert DNA for Golden Gate Assembly-Based Cloning

Obtaining a Golden Gate Assembly-compatible vector

A common challenge to implementing Golden Gate Assembly is finding a suitable vector. Vectors for Golden Gate need to include all the standard features required for your application, such as a replication origin, a selectable marker (e.g., antibiotic resistance gene), and promoters for expression. They must also be free of extraneous Type IIS recognition sites for the enzymes you are using, and they must contain a Golden Gate cloning site for inserting your fragment (Figure 3: pGGAselect Vector Map). The Golden Gate cloning site looks somewhat different from a classic multiple-cloning site for restriction enzyme cloning. It will have two Type IIS recognition sites for each enzyme, arranged so that these sites will point away from each other such that when the sequences are cut, the region containing the binding sites will be excised from the vector. Some variants of a Golden Gate cloning site will also include a counterselection marker, such as the Superfolder GFP (sfGFP) gene to help distinguish uncut vectors from those that have incorporated your desired insert.

How do you find a suitable destination vector that meets these requirements? Several options are available from commercial sources like Addgene. The pGGAselect vector (also included in NEBridge^® Golden Gate Assembly kits) has a cloning site compatible with BsaI, BsmBI, and BbsI (Figure 3: pGGAselect Vector Map). There are also numerous vectors designed to work with pre-existing Golden Gate “standards”, with sets of preselected fusion sites and a compatible vector system, such as the MoClo standard.

Figure 3: pGGAselect Vector Map

 

If you are comfortable with modifying vectors, you can make an existing vector Golden Gate-compatible. First, you need to remove any existing recognition sites for the Type IIS enzymes you plan to use, typically by site-directed mutagenesis (SDM). Care must be taken to avoid disrupting critical vector features; for example, silent mutations can be made within the coding sequence. Next, you can add a Golden Gate cloning site by copying one from an existing vector and inserting it into your vector via blunt cloning, SDM, or by amplifying the vector with primers that append the Type IIS site. A detailed look at how to make your own Golden Gate-compatible vector is beyond the scope of this document, but NEB Technical Support is available for assistance.

Generating insert DNA for Golden Gate Assembly

The most common way to generate insert DNA is to PCR amplify the gene of interest from your genomic or plasmid source. PCR primers can be designed to add the necessary Type IIS restriction sites during amplification, and there are several design tools to help you generate the appropriate primers, such as the NEBridge Golden Gate Assembly Tool. It is important to ensure that the insert does not contain internal recognition sites for the Type IIS enzyme being used, or the insert would be cut during the assembly. If working with genomic DNA, it is often easier to change to a different enzyme and add the appropriate site via PCR than to remove an internal site. When using plasmids, SDM can be used to introduce silent mutations, as with the vectors. Alternatively, with the low cost of modern DNA synthesis, fully synthetic DNA can be obtained (e.g., IDT gBlocks™ and Twist gene fragments). In these cases, “domestication” to remove native Type IIS sites can be easily done in silico when ordering the DNA.

 

Selecting your Golden Gate Assembly Reagents

Before finalizing your design, it’s important to pick the appropriate reagents.

Restriction Enzyme Selection

Which restriction enzyme should you use? BsaI is frequently used for Golden Gate Assembly, and in most cases it’s a good place to start. Specifically, we recommend BsaI-HFv2, which we have shown gives optimal results in Golden Gate Assembly. You can also use the NEBridge^® Golden Gate Assembly Kit (BsaI-HFv2) (NEB #E1601), which includes the enzyme and ligase in a master mix, along with the reaction buffer and pGGAselect destination vector. If your desired insert contains BsaI sites, changing to a different enzyme may be easier than modifying your DNA. Alternatives like BsmBI-v2, PaqCI, or BbsI-HF^® can be used without changing your insert sequence. PaqCI has the added advantage of requiring a longer recognition site, making it particularly useful for large DNA fragments or constructs, as the probability of an undesired recognition site occurring within the sequence is much lower.

Ligase Selection

For ligation, we always recommend a T4 DNA ligase product. Our NEBridge^® Golden Gate Assembly Kits include an optimized ligase mixed at the ideal ratio with the included restriction enzyme. Alternatively, the NEBridge^® Ligase Master Mix is a ligase/buffer premix that can be used with multiple NEB Type IIS restriction enzymes.

Competent Cells

Lastly, you will need high-quality competent cells for transformation. For most cloning applications, we recommend NEB^® 5-alpha Competent E. coli. For large inserts, NEB^® 10-beta Competent E. coli or its electrocompetent version is a better choice.

 

Before you Begin – Double Check Your Design!

As a final check before beginning—actually, before ordering your primers or DNA—it’s important to double check your design using a tool such as the NEBridge Golden Gate Assembly tool. Enter the full sequences for your vector and inserts to confirm that they assemble correctly in silico into the final construct you intend. The tool also ensures proper restriction site orientation and helps you avoid common pitfalls like overlooked internal restriction sites. When you’re new to Golden Gate Assembly, it’s easy to accidentally place Type IIS recognition sites in the wrong location or orientation, or to overlook an internal restriction site that would lead to assembly failure.

 

Protocol for single-insert Golden Gate Cloning

This section provides a sample protocol to help get you started with a single-insert cloning reaction using the NEBridge Golden Gate Assembly Kit (BsaI-HFv2). For protocols covering more complex assembly (up to 50 parts in one reaction!), our Golden Gate Assembly page.

 

Materials & Method

• NEBridge Golden Gate Assembly Kit (BsaI-HFv2) (NEB #E1601)
• NEBridge Golden Gate Enzyme Mix (BsaI-HFv2) (NEB #M2616)
• T4 DNA Ligase Reaction Buffer (NEB #B0202)
• pGGAselect DNA (destination vector), 75 ng/μl
• Prepared insert DNA flanked by BsaI sites (PCR amplicons should have an additional 6bp of 5′ extension beyond the recognition sites), 0.05 pmol
• Nuclease Free Water
• NEB 5-alpha Competent E. coli (High Efficiency) (NEB #C2987)
  
  1. Set up the reaction: 2 μl T4 DNA ligase reaction Buffer, 1 μl pGGAselect destination vector, 0.05 pmol insert fragment, and water to 19 μl total reaction volume. Mix gently by pipetting.
  2. Add 1 μl NEBridge Golden Gate Enzyme Mix (BsaI-HFv2) and mix gently by pipetting. Do not vortex!
  3. Place in a thermocycler and incubate 30 cycles at 37ºC for 1 minute for DNA cutting by BsaI-HFv2 and 16ºC for 1 minute for DNA ligation by T4 ligase. For a fast protocol, you can alternately incubate 15 minutes at 37ºC, but this will result in fewer colonies.
  4. Incubate at 60°C for 5 minutes before transformation. This step inactivates the ligase but not the restriction enzyme, ensuring that remaining vector is linearized, minimizing transformation background.
  5. Chill on ice.
  6. Use 2 μl of the reaction to transform 50 μl of competent cells. If the reaction will not be used immediately for transformation, store at -20°C.
  7. After plating, colonies can be checked for the desired insert by colony PCR or by purifying the plasmid and sequencing the insert.

 

Ready For More?

Find more information on Golden Gate Assembly products and protocols.

Learn about the MoClo Standard and hierarchical assembly. 

Publication: Learn how to design and implement high complexity assemblies of 12-36+ fragments.

Tutorial Video: Learn about our online design tools.

Publication: See how NEB scientists have applied Golden Gate Assembly to bottom-up genome assembly.

Publication: Learn how to use Golden Gate Assembly to make genes from pooled oligos.

Application Note: Accelerate DNA construction to protein expression with a rapid 1-day in vitro workflow using NEBridge Golden Gate Assembly.

Application Note: DNA shuffling using NEBridge Golden Gate Assembly for protein engineering.

 

References

1. Engler, C., R. Kandzia, and S. Marillonnet, 2008. PLOS ONE, 3(11): p. e3647.
2. Weber, E., et al., 2011. PLOS ONE, 6(2): p. e16765.
3. Marillonnet, S. and R. Grützner, 2020. Curr. Prot. Mol. Biol.,130(1): p. e115.
4. Gibson, D.G., et al., 2009. Nat Methods, 6(5): p. 343-345.
5. Chao, R., Y. Yuan, and H. Zhao, 2015. FEMS Yeast Res, 15(1): p. 1-9.
6. Marillonnet, S. and S. Werner, 2015. Methods Mol Biol, 1321: p. 269-84.
7. Bird, J.E., J. Marles-Wright, and A. Giachino, 2022. ACS Synthetic Biology, 11(11): p. 3551-3563.
8. Sikkema, A.P., et al., 2023. Curr Protoc, 3(9): p. e882.