Golden Gate Cloning

Schematic workflow for generating complex combinatorial DNA libraries

Modular Cloning, or MoClo, is an assembly method introduced in 2011 by Ernst Weber et al., whereby using Type IIS restriction sites, the user can ligate at least six DNA parts together into a backbone in a one-pot reaction. It is a method based on Golden Gate Assembly, where Type IIS restriction enzymes cleave outside of their recognition site to one side, allowing for removal of those restriction sites from the design. This helps eliminate excess base pairs, or scars, from forming between DNA Parts. However, in order to ligate together properly, MoClo utilizes a set of 4-base pair fusion sites, which remain between parts after ligation, forming 4-base pair scars between DNA parts in the final DNA sequence following ligation of two or more parts.[9]

MoClo utilizes a parallel approach, where all constructs from tier-one(level 0 modules) have restriction sites for BpiI on both sides of the inserts. The vector(also known as "destination vector"), where genes will be added, has an outward-facing BsaI restriction site with a drop-out screening cassette.[8] LacZ is a common screening cassette, where it is replaced by the multigene construct on the destination vector.[8] Each tier-one construct and the vector have different overhangs on them yet complementary to the overhang of the next segment, and this determines the layout of the final multigene construct.[8] Golden Gate Cloning usually starts with level 0 modules.[5] However, if the level 0 module is too large, cloning will start from level -1 fragments, which have to be sequenced, to help cloning the large construct.[5] If starting from level -1 fragments, the level 0 modules do not need to be sequenced again, whereas if starting from level 0 modules, the modules must be sequenced.[5]

Level 0 modules are the base for MoClo system, where they contain genetic elements like a promoter, a 5' untranslated region (UTR), a coding sequence, and a terminator.[5] For the purpose of Golden Gate Cloning, the internal sequences of level 0 modules should not contain type IIS restriction enzymes sites for BsaI, BpiI, and Esp3I while surrounded by two BsaI restriction sites in inverted orientation.[5] Level 0 modules without type IIS restriction sites flanking can add the BsaI sites during the process of Golden Gate Cloning.[5]

If the level 0 modules contains any unwanted restriction site, they can be mutated in silico by removing one nucleotide from the Type IIS restriction site.[5] In this process, one needs to make sure that the introduced mutation will not affect the genetic function encoded by the sequence of interest.[5] A silent mutation in the coding sequence is preferred, for it neither changes the protein sequence nor the function of the gene of interest.[5]

Level -1 fragments are used to help cloning large level 0 modules.[5] To clone level -1 fragments, blunt-end cloning with restriction ligation can be used.[5] The vector used in cloning level -1 fragments cannot contain Type IIS restriction site BpiI that is used for the following assembly step.[5] Moreover, the vector should also have a different selection marker from the destination vector in next assembly step, for example, if spectinomycin resistance is used in level 0 modules, level -1 fragments should have another antibiotic resistance like ampicillin.[5] 

The level 1 destination vector determines the position and orientation of each gene in the final construct.[10] There are fourteen available level 1 vectors, which differ only by the sequence of the flanking fusion sites while being identical in the internal fusion sites.[10] Hence, all vectors can assemble the same level 0 parts.[10]

As all level 1 vectors are binary plasmids, they are used for Agrobacterium mediated temporary expression in plants.[10]

Level 2 vectors have two inverted BpiI sites from the insertion of level 1 modules.[10] The upstream fusion site is compatible to a gene cloned in level 1 vector while the downstream fusion site has a universal sequence.[10] Each cloning allows 2-6 genes to be inserted in the same vector.[10]

Adding more genes in one cloning step is not recommended, for this would result in incorrect constructs.[10] On one hand, this can induce more restriction sites in the construct, where this open construct allows additional genes be added.[10] On the other hand, this can also eliminate restriction sites, where this close construct stop the further addition of genes.[10]

Therefore, constructs of more than six genes need successive cloning steps, which requires end-linkers containing BsaI or BsmBI internal restriction sites and blue or purple markers.[10] Each cloning step needs to alternate the restriction site and the marker.[10] Furthermore, two restriction enzymes are needed, where BpiI is used for releasing level 1 modules from level 1 constructs and BsaI/BsmBI is for digesting and opening the recipient level 2-n plasmid.[10] When screening, the correct colonies should alternate from blue to purple every cloning step, but if a "closed" end-linker is used, the colonies will be white.[10] 

Level M vectors are similar to level 2 vectors, but have a BsaI site located upstream of the two inverted BpiI sites.[11] When one or several genes are cloned in a level M vector, a second BsaI is added at the end of the construct via a Level M end-linker (ref). This allows a fragment containing all assembled genes to be excised from the vector and subcloned in a next level of cloning (Level P).

Level P vectors are similar to level M constructs except that the BpiI sites are replaced by BsaI sites and the BsaI sites are replaced by BpiI sites. Several level M constructs with compatible fusion sites can be subcloned into a level P vector in one step. Theoretically, as many as 36 genes can be assembled in one construct using 6 parallel level M reactions (each required for assembly of 6 genes per level M construct) followed by one final level P reaction. In practice, fewer genes are usually assembled as most cloning projects do not require so many genes. The structure of level M and P vectors is designed in a such as way that genes cloned in level P constructs can be further assembled in level M vectors. Repeated cloning in level M and P vectors forms a loop that can be repeated indefinitely to assemble progressively large constructs.

In standard Golden Gate Cloning, the restriction sites from the previous tier construct cannot be reused.[12] To add more genes to the construct, restriction sites of a different Type IIS restriction enzyme need to be added to the destination vector.[12] This can be done using either level 2, or M and P. A variant version of level M and P is also provided by GoldenBraid.

GoldenBraid overcomes the problem of designing numerous destination vectors by having a double loop, which is the "braid," to allow binary assembly of multiple constructs.[12] There are two levels of destination plasmids, level α and level Ω.[12] Each level of plasmids can be used as entry plasmids for the other level of plasmids for multiple times because both levels of plasmids have different Type IIS restriction sites that are in inverted orientation.[12] For counterselection, the two levels of plasmids differ in their antibiotic resistance markers.[12]

The Golden Gate Cloning principle can also be applied to perform mutagenesis termed Golden Mutagenesis. The technology is easy to implement as a web tool is available for primer design (https://msbi.ipb-halle.de/GoldenMutagenesisWeb/) and the vectors are deposited at addgene (http://www.addgene.org/browse/article/28196591/).[13]