Jeffrey Perkel has been a scientific writer and editor since 2000. He holds a PhD in Cell and Molecular Biology from the University of Pennsylvania, and did postdoctoral work at the University of Pennsylvania and at Harvard Medical School.
It’s been more than four decades since researchers launched the molecular-biology revolution with the invention of DNA cloning. In that time, the tools of the trade have changed relatively little, and for many researchers, DNA cloning still means restriction enzyme digestion and DNA ligation.
But that strategy has several significant limitations, according to Lisa Stillwell, a product development manager at Thermo Fisher Scientific who previously managed the company’s alternative cloning strategy lines. The first is the possibility (and indeed, with more complex projects, the likelihood) that key restriction sites will be present within the sequences to be assembled, complicating cloning strategies.
Restriction enzyme-based strategies also are not particularly amenable to assembling more than two fragments at a time. And even assembling two fragments is time-consuming and laborious, often requiring multiple steps between digestion and ligation, such as dephosphorylation and gel purification.
Finally, restriction enzyme-based cloning generates a “scar” or “seam” at the position where two fragments join—a molecular blemish that has the potential to subtly alter the behavior of the linked pieces of DNA. This is especially problematic for synthetic biologists, who often aim to turn DNA elements such as promoters and terminators into stand-alone components with predictable behaviors, regardless of genetic context.
Today, thanks largely to the needs and creativity of the synthetic-biology community, alternative, “seamless” cloning strategies have been developed. Here are some of the more popular options.
Gibson Assembly
One of the most popular of the new cloning approaches is Gibson Assembly®. Developed by Daniel Gibson of the J. Craig Venter Institute and used to build the genome of Venter’s heralded “synthetic bacterial cell,” Mycoplasma mycoides JCVI-syn1.0, Gibson Assembly is an isothermal, single-step process. All that’s required is a master mix comprising three enzymes and about an hour of incubation.
First, the DNA fragments to be cloned (including the vector) are PCR-amplified such that pieces that will be adjacent in the final assembly contain short stretches (15 bp to 30 bp) of overlap. These fragments are placed directly into the cloning reaction at 50oC, where a DNA exonuclease first chews back from the fragments’ 5’ ends to create a 3’ overhang. That creates single-stranded regions of homology between adjacent fragments, which then anneal. A DNA polymerase closes any remaining gaps, and DNA ligase seals the nicks, creating a contiguous piece of DNA ready for cloning. (The process is cleverly explained in song in this YouTube video.)
According to Cathy Shea, product manager for the Gibson cloning line at New England Biolabs (NEB, which licensed the technology from Synthetic Genomics), the reaction is optimized for assembly of up to six fragments. The smallest piece that can be assembled this way is about 120 bases, but fragments as long as 12 kb can be used. Justin Bingham, director of business development at SGI-DNA, a service provider and subsidiary of Synthetic Genomics that uses Gibson Assembly to build large synthetic DNAs, says the process can be applied to assemblies greater than 80 kb.
“Gibson [Assembly] has the advantage that it is quite rapid,” Shea says. PCR fragments can be used directly, for instance, so no cleanup is required. And Gibson Assembly can handle large numbers of fragments. But there are downsides, too. Because the process does not generate restriction sites, assembled fragments cannot be moved easily from vector to vector. As a result, Shea recommends planning ahead and engineering restriction sites into Gibson primers if you anticipate needing to move the cloned fragment later.
NEB has designed an online design tool called NEBuilder™ to assist in Gibson Assembly experiment planning.
GeneArt Seamless Cloning and Assembly
Life Technologies’ GeneArt® Seamless Cloning and Assembly Kits are similar to Gibson Assembly, but without the ligation step. DNA fragments containing sequence overlaps are added to the assembly reaction with the vector, digested with an exonuclease, annealed and filled in. This assembly is then transformed into cells, closing up the DNA circle via homologous recombination.
According to Stillwell, the GeneArt Seamless PLUS Cloning and Assembly Kit can assemble up to five pieces of DNA (four inserts and a vector)—up to 40 kb total—in a single 30-minute reaction at room temperature. The final ligation step is performed in E. coli. The GeneArt High-Order Genetic Assembly System can assemble up to 10 fragments and 110-kbp assemblies in S. cerevisiae, but that process requires three days.
“It’s huge. And the machinery inside that yeast cell is amazing,” Stillwell says.
Users can plan their experiments using Life Technologies’ GeneArt Primer and Construct Design Tool.
Type IIs cloning
An alternative strategy, sometimes called Golden Gate cloning, relies on so-called type IIs restriction enzymes.
Traditional “type II” restriction enzymes bind and cut within palindromic sequences (e.g., GGATCC) to create an overhang. Ligation of two such ends cut with the same enzyme will restore the restriction site. Type IIs enzymes bind asymmetric recognition elements and cut one or more bases outside them. For instance, the enzyme BsaI recognizes the sequence 5’-GGTCTCN^-3’, producing a four-base 5’ overhang. Joining those ends—which can contain any four-base tag the user desires—eliminates the restriction site, theoretically creating a seamless junction.
Life Technologies offers GeneArt Type IIs Assembly Kits for three different enzymes, AarI, BsaI and BbsI, and users can plan type IIs-based experiments with the company’s online tool. “This approach has been used to assemble multiple (up to eight) DNA fragments in any order, into any compatible vector, without scars,” Stillwell says.
Researchers at Icon Genetics in Germany, led by Sylvestre Marillonnet, now at the Leibniz Institute of Plant Biochemistry, developed a standardized, automatable form of this strategy, called MoClo. (An alternative strategy, called GoldenBraid, also has been developed.) The idea behind MoClo, Marillonnet explains, is to mimic the BioBrick strategy of flanking common components, such as promoters, with common restriction sites to simplify assembly and facilitate combinatorial designs.
“The method allows you to pipette in one tube, say, 10 different plasmids—a recipient vector and nine containing the pieces you want to clone. And in one step you can assemble all these in the vector,” he says. Marillonnet’s team at Icon Genetics used MoClo to assemble in three steps an 86-kb plasmid containing 27 transcription units, each comprising a promoter, coding sequence and terminator, with 81 fragments total.
Unlike Gibson Assembly and GeneArt Seamless Cloning, the method does not rely on homology between adjacent fragments and therefore can correctly assemble pieces with repeating elements. The key downside, Marillonnet says, is that, as with traditional restriction enzymes, the fragments to be cloned cannot contain internal recognition sequences for the type IIs enzymes.
Marillonnet has made the necessary vectors available at Addgene, but experiments must be designed manually, as no software tool is yet available.
A place for restriction enzymes
So, which strategy should you choose? Each has strengths and weaknesses, so it really depends on application. And many strategies, in fact, are complementary, Shea says. Gibson Assembly, for instance, is relatively easy to start using—no special vectors are required, for instance—and especially useful for direct cloning of PCR fragments. But after a fragment is cloned, it can only be excised via PCR unless flanking restriction sites also were engineered in. And it, like GeneArt Seamless Cloning, is trickier if the fragments to be joined are homologous to one another—if they contain repeating elements, for instance.
For combinatorial assemblies—for instance, to test multiple promoters on a single transcription unit—Golden Gate is an attractive option. But if the fragments to be assembled contain type IIs restriction sites, they first must be deleted.
Yet restriction enzymes still have a place in today’s molecular lab. For one thing, most labs have years’ worth of constructs made using restriction enzymes, and they remain the easiest way to excise and move those fragments around, not to mention checking clones for accuracy.
“They’re very basic tools for analyzing DNA,” Stillwell says, “but people are moving away from them for more demanding cloning projects.”
Of course, if you really want to build a large, complex DNA sequence, you can simply have it made for you. DNA synthesis companies now can build and quality-control massive sequences on demand—up to 2 Mbp in the case of SGI-DNA, nearly twice the size of Venter’s synthetic cell.
“There’s a lot of labor that goes into [building] a 75-kb construct,” says Bingham. You need to find an oligo source, order the DNA, assemble them into intermediates, clone them and sequence them to ensure the sequence is correct. Then, depending on the size of the desired construct, those pieces in turn must be assembled into larger constructs and sequenced again.
Some researchers may find the simplicity of ordering their precise sequence via the web and receiving it in the mail beats any manual process, no matter how fast and seamless it might be.