November 1, 2012 (Vol. 32, No. 19)

Assembly, Editing, and Horizontal Interspecies Transfer of Large Genetic Constructs

One of the major goals of the emerging area of synthetic biology is to design and build engineered biological systems in a reliable way, as engineers today design integrated circuits based on the known physical properties of their materials. It is foreseeable, thereby, that the last two steps in such a workflow would be to convert a large digitized sequence stored in a computer into a real DNA molecule (as large as a chromosome) and then to install it in the right contextual host.

Three fundamental shortcomings to these processes are 1) gene synthesis, editing, and assembly reactions are less successful with larger constructs, 2) long DNA molecules are sensitive to in vitro manipulation, and 3) a variety of organisms are refractory to transformation. To help overcome these difficulties, we have developed a set of tools based on well-characterized biological properties such as homologous recombination and DNA conjugative transfer.

Figure 1. Seamless assembly of multiple DNA fragments: DNA fragments of the designated size were either PCR-amplified or released from precloned constructs by restriction digests as indicated and purified by standard silica spin columns. Reactions were set up by combining 1 or 4 fragments as indicated with the corresponding linearized vectors, and incubated at room temperature with the GeneArt® enzyme mix. Assemblies totaling up to 10 kbp were cloned into a pUC19 derivative vector, whereas larger constructs were cloned into pYES7L (see text). Reaction aliquots were transformed into One Shot® DH10B™ T1R SA cells. The asterisk indicates that the PCR-amplified fragment was not purified prior to assembly. Bars indicate the cloning efficiency (CE) and squares indicate the number of colonies per plate (CFU/plate). Experiments were performed in triplicate.

Assembling the Constructs

The DNA assembly strategy uses a proprietary enzyme mix based on the “chew-back and repair” principle exploited by homologous recombination. Up to four DNA fragments can be simultaneously recombined into any linearized vector, typically in 30 minutes, without extra DNA sequences, restriction endonucleases, or ligation reactions. Adjacent molecules of sizes ranging from 50 bp to 10 kbp must share unique common end sequences that can be incorporated by PCR amplification. To facilitate the design, an intuitive webtool assists in delineating the constructs, intermediate molecules, and required oligonucleotides.

A number of cloning configurations were successfully tested using fragments of different sizes resulting in final constructs of up to 40 kbp (Figure 1). Results showed that between 100 to 10,000 colonies per plate are obtained with cloning efficiencies ranging from 72% to 98% depending on the assembly’s complexity.

Figure 2. Multisite-directed mutagenesis: A) Methodology of the approach. (See text for details.) Dotted lines represent methylated DNA. Small arrows correspond to the oligonucleotides and large arrows represent the PCR-amplified fragments. Numbers indicate the mutation sites. B) Typical mutagenesis performance. We generated 5-, 10-, and 14-kbp plasmids containing a lacZa fusion with 3 mutations, where each individual mutation (spanning either 1 or 3 bp) inactivated its alpha complementation capacity. The multisite mutagenesis efficiency was determined by calculating the percentage of blue bacterial colonies that appeared on X-gal LB agar plates. Where indicated, the distances between 2 sites are shown. The asterisk indicates that the two mutations are covered by a single pair of oligonucleotides. Experiments were performed in triplicate.

Editing the Constructs

In order to further modify an existing construct without having to recreate the whole assembly de novo, we have developed a multisite-directed mutagenesis approach based on homologous recombination that results in high efficiencies even when applied to relatively large episomes. The strategy is summarized in Figure 2A.

Plasmids of up to 14 kbp are subjected to a 12−20-min methylation reaction followed by 12 to 18 cycles of either a single multiplex or three independent PCR reactions. The mutation site in each oligonucleotide must be flanked with at least 10 unchanged nucleotides at both sides. After amplification, an aliquot of the reaction is subjected to a 15-min pulse of recombination activity, followed by transformation into One Shot® MAX Efficiency® DH5α™-T1R cells, where the methylated template is degraded by the mcrBC restriction system.

The high mutagenesis efficiencies observed (between 70% and 95%) significantly reduce the number of clones needed to be validated compared to other methodologies (Figure 2B). The approach also works with sites in close proximity (Figure 2B) and with oligonucleotides harboring up to 12 degenerated nucleotides for single-site and 3 degenerated nucleotides for multisite mutagenesis (not shown).

Transferring the Constructs

Once the molecules are assembled and edited, ex vivo manipulation becomes particularly challenging as very large episomes are extremely susceptible to physical damage and not easily transformed into the final host. To overcome this limitation we elaborated a simple in vivo transfer method based on an established triparental interspecies plasmid conjugation. A diagram of the approach is shown in Figure 3A.

First, the inserts are assembled into a high-capacity, broad-host range, and mobilizable plasmid, pYES7L. Then, cells harboring the construct are mixed with helper and recipient cells, incubated for a few hours, and plated onto the appropriate selection agar medium. Assemblies of up to 25 to 40 kbp in size could be efficiently transferred to different Gram (-) bacteria and yeast (Figure 3B). The capacity and host range of the plasmid are dictated by the nature of the DNA replication factors of the host.

Figure 3. Horizontal DNA transfer: A) Diagram of the approach. (See text for details.) Ori T, origin of transfer; tra, transfer genes necessary for conjugation; pRK2013, helper plasmid; pYES7L, assembly vector. The plasmid pYES7L harbors an RK2 replication origin functional in several Gram (-) bacteria and an autonomous replicating sequence and centromere for replication in yeast. B) Typical transfer performance: Tri-parental matings were setup as shown in (A) and in the text. Conjugal performance corresponds to a 25 kbp plasmid. Up to 40 kbp size plasmids were successfully transferred with somewhat lower efficiencies. Transconjugants were selected in agar medium containing the appropriate selection reagents (LB agar plus 100 µg/mL spectinomycin, 100 µg/mL rifampicin for bacteria, and CSM-tryptophan agar plates for yeast).


By the redesign and reprogramming of existing biological systems, synthetic biology and other fields have underscored the need for robust systems to generate and transfer large DNA assemblies. The approaches presented here enable researchers to readily design, construct, edit, and mobilize plasmids with enough capacity to encode complex genetic pathways. Size and complexity may no longer be limitations to build and transfer DNA. Rather, a better understanding of the cell’s physiology is needed in order to fully exploit these and other DNA assembly and transfer technologies.

Xiquan Liang, Ph.D., Lansha Peng, and Chang-Ho Baek, Ph.D., are staff scientists; Ke Li, is a scientist III; Todd Peterson, Ph.D., is head of R&D; and Federico Katzen, Ph.D. ([email protected]), is a senior staff scientist, all at Life Technologies. Reagents referenced in this article (GeneArt® Seamless PLUS Cloning and Assembly, GeneArt® Seamless Cloning and Assembly Enzyme Mix, and GeneArt® Site-Directed Mutagenesis PLUS System) are for Research Use Only. Not for human or animal theraeputic or diagnostic use.

Previous articleInspiration Enters Chapter 11, Seeks Buyer
Next articleScientists Release 1,000 Genomes Map of Genetic Variation