Seamless Cloning Technologies

For years, DNA cloning has largely been synonymous with restriction enzymes—specifically, type II restriction enzymes like EcoRI and HinDIII, which bind and cut within short, four-to-eight nucleotide palindromic recognition sequences. Given two pieces of DNA, an insert and vector, cut with the same enzyme along with some DNA ligase to seal the nicks, researchers can link the two pieces into a recombinant plasmid, restoring the restriction sites at the junctions.

That approach works exceptionally well for many applications. But not always. For one thing, it requires that the piece of DNA you want to clone doesn’t contain within its sequence the enzyme-recognition site you hope to use. It’s particularly difficult to apply to complex cloning projects involving multiple fragments, which increasingly are the norm in, for instance, synthetic biology. And, it produces a “seam” or “scar”—short sequences at the cloning junctions that can potentially influence expression from the construct.

Researchers have developed multiple options to circumvent some or all of these problems. One is recombinase-based cloning, such as Thermo Fisher Scientific’s Gateway cloning system. Another is TA-cloning, commonly used to clone PCR fragments.

Today, researchers have another option: seamless cloning.

Golden Gate cloning

According to Ana Egana, production group leader for the cloning technologies group at New England Biolabs (NEB), seamless cloning strategies can be useful for any application. But where they really shine is in multifragment cloning projects. Traditionally, she explains, researchers could only clone one or perhaps a few inserts per cloning round. As the desired designs get more sophisticated, the number of rounds goes up, as does the experimental complexity and hands-on time.

But seamless methods are typically one-pot affairs in which vectors, inserts and enzymes are combined in a single tube, incubated and transformed. “There’s a great advantage in time,” she says. “It can be done in one day.”

Another advantage, says Sylvestre Marillonnett, a group leader at the Leibniz-Institut für Pflanzenbiochemie in Halle, Germany, is in experimental planning. As cloning projects get ever more complex—for instance, developing combinatorial libraries of promoters, untranslated regions, terminators and gene sequences for synthetic biology applications—it becomes more difficult to develop a working experimental scheme. “Sooner or later, you don’t have any free restriction enzymes you can use,” he says.

Seamless strategies enable researchers to clone whatever sequences they want, in whatever order, simply by ordering the proper series of oligonucleotides or using the proper adaptors.

In 2008, Marillonnett described one new strategy called Golden Gate. Available from NEB, Thermo Fisher Scientific and Addgene, among others, Golden Gate cloning relies on so-called class IIs restriction enzymes, like BsaI. BsaI recognizes the 5’-GGTCTC sequence. But it cuts several bases outside that sequence, producing a four-base overhang. Because that overhang is not part of the recognition sequence, researchers can use any four-base sequence they want, thus driving the system to assemble fragments in any desired order—provided the required pieces are flanked by type IIs sites at the outset. (Those sites are lost during assembly, thus producing seamless linkages.)

In one example, Marillonnett’s team—then at Icon Genetics in Germany—used Golden Gate to assemble nine pieces of the GFP gene in the proper order in a single-step reaction, with greater than 90% efficiency; in another case, they built a combinatorial library of gene fragments to build dozens of variations on the trypsinogen gene containing pieces from three mammalian homologs.

More recently, Marillonnett developed a variant of the system called “MoClo,” which hierarchically builds ever-larger designs by alternating between two type IIs enzymes, BsaI and BpiI—a strategy his team used to assemble 27 separate transcription units in a single 86-kb construct.

Stephen Ekker, professor of biochemistry and molecular biology at the Mayo Clinic, uses Golden Gate to assemble TALEN reagents for genome-editing applications.

Traditionally, he says, researchers assemble TALE (transcription activator-like effector) arrays one element (representing one nucleotide) at a time. Thus, to build an enzyme capable of binding a 15-mer sequence required combining at least 15 inserts—too many for a one-pot reaction.

Ekker’s team built five libraries, each containing 64 sets of TALE-array trimers—that is, recognizing the three-base sequences AAA, AAC, AAG and so on. These libraries differ in the tetranucleotide overhangs recognized by the type IIs enzymes. Thus, by picking one member of each library, researchers can, in one step, build a TALEN to recognize any 15-base sequence they choose.

Ekker has made this plasmid collection available as the FusX Automated TALEN Kit through a start-up company called Lifengine Technologies.

Homology-based cloning

Other seamless cloning strategies are based on the annealing of insert and linearized vector DNA via homologous sequence overhangs in the absence of type IIs enzyme. Commercialized by NEB, Takara Bio USA and Thermo Fisher Scientific, among others, these methods differ in the particulars, but all involve the same basic concept.

The insert or inserts to be cloned are PCR-amplified or -digested with restriction enzymes such that the ends of each fragment contain a short (15-base pair or longer) homologous overlap with either the ends of the linearized vector or another adjacent insert, depending on the cloning strategy. The vector is likewise linearized at the desired location. These DNA segments are then incubated together with an enzyme mix that chews back each end to produce single-strand overhangs. This enables the inserts to spontaneously anneal to the linearized vector and/or adjacent insert at the exposed homologous overlaps, after which the annealed molecule is directly transformed into competent E. coli, where any gaps are filled and repaired in vivo, as with In-Fusion HD Cloning from Takara. Alternatively, a DNA polymerase fills in the gaps, and a DNA ligase seals the nicks in vitro prior to transformation, as in NEB’s Gibson Assembly® and NEBuilder® HiFi DNA Assembly Kits.

These strategies can be incredibly efficient, says Malathi Raman, European product manager for cloning at Takara Bio Europe. Cloning efficiencies for the In-Fusion® kit  exceed 95% for inserts ranging from 20-bp oligos to 15-kb fragments. “In most cases, it is sufficient to screen just two colonies per construct,” she says.

Indeed, there’s really no reason to use traditional cloning methods, Raman says; she advises researchers to use seamless methods full-time. After all, she says, cloning is not the point of an experiment but its beginning:

“If you spend less time in cloning, then you have more time to spend on the actual research and to get results faster.”

But which method to choose? According to Axel Trefzer, a research and development director in the synthetic biology business at Thermo Fisher Scientific, the decision often comes down to the specific sequence and available reagents. Golden Gate strategies, for instance, require the introduction of type IIs recognition sequences into vectors and insert libraries, and the thoughtful selection of overhangs to make those pieces fit together. And as with other restriction enzyme-based approaches, researchers must remove unwanted recognition sequences to avoid cutting within their inserts (or choose another type IIs enzyme).

Homology-based approaches like Gibson can be incompatible with repetitive sequences, as they can cause misassembly errors. Homology-based clones should also be sequenced, as the amplification step can introduce unwanted mutations—something that is not true of type IIs approaches.

All these strategies vary somewhat in the number of elements that can be assembled in one step and the maximum size of the constructs that be built, but according to Trefzer, that’s rarely a significant concern. And in any event, online design tools are available (from both NEB and Thermo Fisher Scientific) to help researchers in planning their experiments.

Open-source plasmid distributor Addgene tracks the cloning methods researchers use to create the plasmids they submit, says executive director Joanne Kamens. She estimates the company’s collection includes thousands of reagents created using Gateway, as it has been around the longest. But other technologies are gaining ground. “I guess we have hundreds of plasmids created using Gibson Assembly,” she says.

That’s not to say researchers need to throw their restriction enzymes away. According to Penny Devoe, associate director of portfolio management for DNA cloning at NEB, though seamless strategies are on the rise, many customers continue to use restriction enzyme-based cloning, either for simplicity or historical reasons, not to mention in pre-cloning fragment preparation, post-cloning verification and many other applications wholly unrelated to cloning.

And a new strategy also is emerging, notes Ekker. With the cost of high-quality, long-length DNA synthesis falling, some labs—including Ekker’s—opt to synthesize DNA from scratch rather than assemble it from existing pieces.

If nothing else, Ekker says, gene synthesis—for instance, using gBlocks® gene fragments from Integrated DNA Technologies or GeneArt™ gene synthesis from Thermo Fisher Scientific—provides an easy alternative to begging other labs for reagents, especially if a material-transfer agreement is required. “It’s sort of the get-out-of-jail-free card for tech transfer,” he says.

But more to the point, this approach is incredibly powerful. “I start feeling like a real genome engineer,” Ekker says, “because you can really customize that DNA. Remove repeats, remove CpGs, codon optimize for the cell type, optimize the Kozak sequence—all these things that used to be a pain, you can do at once, and you can make variants at the same time.”

And hey, who couldn’t use a little extra space in their enzyme box?