DNA Assembly

Golden Gate Assembly comprises a number of constituent elements: Type IIS Restriction Enzymes, Parts and Plasmids, and Assembly Standards. On this page we will explore these elements.

Golden Gate Assembly is based on the use of Type IIS restriction endonucleases.

Classic Type II restriction enzymes such as EcoRI ↗ or PstI ↗, used in the BioBrick ↗ assembly standard, bind to specific DNA sequences (recognition sites) and cut within those sites. Most sites are palindromic, meaning recognition sites have the same sequence on both strands of the DNA.

Type IIS restriction enzymes bind to their respective recognition sites, but then cleave the DNA at defined positions away from the recognition sites. After cleaving, single-stranded overhangs remain.

Type IIS recognition sites are also not palindromic, which means that the two strands of the DNA it comprises do not have the same sequence. This permits the flanking of a DNA part, ex. a promoter or gene, with the same recognition site in opposing directions such that the part can be cleaved from its backbone, removing the recognition sites but leaving behind those single-stranded overhangs.

Comparison of Type II (HindIII) and Type IIS (BsaI) enzymes. Notice how BsaI allows for cleaving of the part leaving 4bp overhangs flanking it.

DNA parts that have overhangs with complementary sequences can then anneal to one another. A single Type IIS restriction enzyme can cleave all of the selected parts from their respective plasmid backbones allowing for a one-pot assembly reaction. This offers efficiencies in process and time when compared to assembly methods that rely on Type II enzymes.

One can use Type IIS restriction enzymes to efficiently and scarlessly assemble small components into a much larger DNA sequence using unique native sequences as overhangs. For example, T7 bacteriophage ↗ and even small genomes have been assembled seamlessly from synthesized parts.

However, it is the development of assembly standards in which these overhangs have been standardized, where the real power of Golden Gate Assembly comes into play for DNA construction to engineer biology.

A Golden Gate part is a synthetic DNA sequence flanked by a pair of specific Type IIS restriction enzymes, whether you designed and built it yourself or found it in one of the collections or toolkits. These parts represent a genetic element of a certain “part type”.

In your iGEM distribution you will find a range of part types in the various collections and toolkits. Some part types are regulatory sequences such as promoters, ribosome binding sites (RBSs) for prokaryotes, or terminators whilst others are coding region sequences such as the coding sequences (CDSs) and fusion partners. You most likely will design and build one or more of your own parts to work with those in the distribution.

Some collections have more part types. For example, the E. coli Protein Expression Toolkit has many different signal sequences, purification tags, protease cleavage sites and flexible linkers to give a great many options for protein purification in E. coli. The Open Yeast Collection has other part types outside of the transcription unit including homology arms for genome integration, yeast origins of replication and assembly connectors for hierarchical assembly (more about this later).

Some parts types are composites of regulatory sequences and coding sequences creating complete transcription units. The various yeast selection markers found in the Open Yeast Collection are examples of these composites. They all have a yeast promoter, selection marker gene (CDS) and yeast terminator. These selection markers comprise a full transcription unit and are used for selection of assembled constructs in yeast.

As you look through a collection, keep in mind the part types within. This may differ from collection to collection, and may reflect different assembly standards and interoperability.

All parts that exist within a collection are cloned into a plasmid backbone. The sole purpose of the backbone is to propagate the plasmid in E. coli so you can extract the part for use in assembly. The backbone typically has a strong origin of replication for high yields when performing plasmid DNA extraction. The backbone also uses the same antibiotic selection for all of the parts in the collection to permit one-pot-assembly. This contrasts with the destination vector - the backbone that you assemble all your parts in - which has a different antibiotic selection marker.

As you look through a collection or library, keep in mind the plasmid backbone used for parts, as well as the destination vector, this may differ from collection to collection.

For example, the iGEM Engineering collections use a pSB1C3 based plasmid backbone: chloramphenicol resistant, high copy number. The FreeGenes collections use the pOpen-v3 (BBF10K_003498 ↗) backbone: ampicillin / carbenicillin resistant, high copy number.

We call the 5’ BsaI site overhang the 5’ fusion site and the 3’ overhang the 3’ fusion site. All parts will have fusion sites flanking them. In an assembly standard the 3’ fusion site of a part has the same overhang sequence as the 5’ fusion site of the next part in the assembly. This is the same for the 5’ fusion site which is the 3’ fusion site of the previous part. The actual region of the part involved in assembly spans from the 5’ fusion site to the 3’ fusion site - without the Type IIS recognition sequences.

As you use any collection of Golden Gate / Type IIS parts, it is critical that you are aware of the fusion sites it uses, as defined by the assembly standard it follows.

Please also note that some assembly standards will use different terminology for these overhangs, with the 5’ fusion site and 3’ fusion site being referred to as the prefix and suffix, respectively.

An example of BBa_B0034, an RBS, in MoClo format. The part is flanked by BsaI sites in opposite orientation and the fusion sites for an RBS part.

As an example, here is BBa_B0034. BBa_B0034 is a commonly used ribosome binding site in iGEM, and variants of it be found in the E. coli Protein Expression Toolkit and iGEM Engineering Collections. This particular version is BBF10K_003410 ↗ and includes extra bases of padding.

As an RBS part that adheres to the MoClo assembly standard, it has a 5' fusion site of TACT and a 3' fusion site of AATG. When cleaved from its plasmid backbone, these overhangs will allow it to anneal to a promoter and CDS part.

When you have all of the parts you want to assemble, you will need to assemble them into a vector. That is the role of the destination vector, also called a receiving vector or assembly vector. Like the parts' plasmid backbone, the destination vector has an origin of replication as well as an antibiotic selection marker.

The destination vector must also have a different selection marker than that of all of the part backbones used in the assembly. For example, all of the FreeGenes parts are all in an ampicillin resistant backbone and the destination vectors are either kanamycin, chloramphenicol or spectinomycin.

An example of a destination vector for MoClo. Parts would be assembled into it with BsaI to build a transcription unit, removing the dropout / counter-selection marker. Correct assemblies could be selected against uncut ones (those where the dropout was not removed).

Unlike the parts vector, destination vectors also have a counter-selection marker in what is known as a dropout marker (dropout for short). Dropouts are transcription units that allow one to differentiate between assembled vectors and uncut destination vectors. Common reporter genes are LacZ𝛼 for blue/white selection, RFP for fluorescent selection and toxin/antitoxin selection such as ccdB.

In Golden Gate Assembly, the dropouts include Type IIS restriction enzyme recognition sites that are inverted relative to the backbone compared to all of the parts in the assembly. During an assembly, these inverted recognition sites drop out from the destination vector along with the counter-selection marker, leaving the appropriate fusion sites for the completed assembly.

Most destination vectors and their dropouts have overhangs that reflect the original MoClo transcription unit fusion sites - GGAG and CGCT.

Destination vectors may also have a second Type IIS restriction enzyme that differs from the one used to assemble into the vector. This Type IIS resriction enzyme will flank an assembled transcription unit, in order for that transcription unit to be assembled to others in the future. These, along with their respective fusion sites, will be defined by the assembly standard and its hierarchical assembly scheme.

A summary of a Golden Gate (MoClo) assembly: assembling four parts (promoter, rbs, cds, terminator) into a destination vector in a single reaction.

Once you have all of your parts, whether they are from an existing collection or made by you, the next step is to assemble everything into a final construct. This is known as One-pot Assembly since you put everything for the assembly - parts, destination vector, Type IIS restriction enzyme, T4 DNA Ligase and buffers - in a single PCR tube. When setting up an assembly it is critical that all of the parts are equimolar so there is no imbalance in the ratios of overhangs from part to part.

The assembly standard is what defines the rules for DNA part constructions, including where each part type falls within the assembly schema. More specifically it defines the fusion site overhangs of all parts in the schema (syntax).

The original Golden Gate syntax was defined in the Modular Cloning (MoClo) assembly standard paper ↗.

The MoClo Assembly Standard fusion sites for building a complete transcription unit in a single assembly reaction.

The MoClo Golden Gate assembly standard defines the part types and their overhangs for the transcription unit. Originally it was focused on bacterial genetics but has been expanded to include a wide range of organisms.

Thankfully, many of the assembly standards that have expanded upon MoClo generally follow the same syntax for building a transcription unit. Where they differ might include, an extension of this syntax to allow for greater flexibility and complexity when building a transcription unit. Or they may suggest a different method and syntax for building multi-transcription unit constructs.

As an example, the iGEM Type IIS standard is based completely on Loop Assembly ↗, which in turn uses the MoClo syntax for creation of transcriptional units, and others that overlap with it like Phytobricks ↗.

Hierarchical assembly is a multilevel system that starts first with standardized basic parts (Level 0) that make up a collection or toolkit and your own parts. These parts act as building blocks for assembly to typically build a transcription unit (Level 1). Two or more transcription units can then, if needed, be assembled (Level 2) to create a multi-transcription unit construct.

The majority of the parts in the iGEM Distribution are Level 0 (L0) parts. There are also a selection of destination vectors to assemble L0 parts into a Level 1. For example, perhaps your goal is to express and purify the Moloney Murine Leukemia Virus (MMLV) Reverse Transcriptase (BBF10K_003286 ↗) from the E. coli. Protein Expression Toolkit. All of the L0 parts to create a Level 1 expression construct are available in the E. coli Protein Expression Toolkit and the Open Enzyme Collection in your distribution.

Often you may only need a single Level 1 transcription unit to create a working construct for your needs. But, what if you wanted to assemble two or more transcription units into one construct? Perhaps you are building a metabolic pathway or expressing a restriction enzyme and its methylases.

The JUMP collection in the iGEM Distribution, is a set of MoClo compatible destination vectors that also enables hierarchical assembly. As an example, with pJUMP29-1A / BBa_J428341 an engineer can assemble L0 parts into it using BsaI to make a trancription unit (GGAG - CGCT). This transcription unit will be flanked by BsmBI and a pair of fusion sites. Transcriptional units assembled into pJUMP29-1A, pJUMP29-1B, pJUMP29-1C, and pJUMP29-1D, can then be assembled into Level 2 vector, pJUMP49-2A / BBa_J428361, vector with BsmBI.

Once you have designed your construct and have your parts, you can proceed to an actual assembly.

First, prepare each part sample; in the case of the iGEM Distribution this would include transforming each sample, making a glycerol stock, miniprepping, and any quality control measures.

Once you have minipreps of all parts, you will need to quantify the DNA (typically in ng/µL). We recommend tracking your parts in a spreadsheet and your notebook, include the quantification, the total plasmid size in base pairs, and the part size (i.e the length from prefix to suffix). From this, calculate the molar concentration of each part with the following equation or use the online NEB calculator ↗.

moles dsDNA (mol) = mass of dsDNA (g)/((length of dsDNA (bp) x 617.96 g/mol/bp) + 36.04 g/mol)

It is essential to add equal molar amounts of each part in a Golden Gate assembly. An imbalance with one part can reduce the efficiency of an assembly reaction.

We recommend this very useful protocol ↗ developed and documented by Shyam Bhakta at the Bennet Lab, Rice University. The documentation covers a variety of Type IIS restriction enzymes.

iGEM Distribution

The iGEM 2023 distribution has a number of Type IIS libraries and toolkits such as the iGEM Engineering Committee Collections, Asimov’s Mammalian Parts Collection ↗, Open Bioeconomy Lab’s E. coli Protein Expression Toolkit (PET), the Open Enzyme ↗ and Open Reporters ↗ Collections as well as Open Science Network Societies Open Yeast Collection (OYC) ↗. With these libraries you can design and build expression systems for testing in the Design, Build, Test and Learn Cycle ↗. You can augment these collections by creating your own parts through PCR or gene synthesis.