Approaches for Assembling Synthetic, Genome-Sized DNA
Synthetic Biology is the de novo design and construction of biological devices and systems for useful purposes . However, the size and sequence of potential constructs have been constrained by the limitations of the major techniques on which synthetic biology is based: 1) generation of recombinant DNA, and 2) amplification of cloned DNA copies using PCR and E. coli. The future of molecular biology and genetics research will be shaped by methods that move us past those limitations.
Dr Dan Gibson has spent most of the last decade working at the J. Craig Venter Institute, and more recently at Synthetic Genomics Inc. (SGI) and its subsidiary, SGI-DNA, to develop, establish, and commercialize methods that are already taking us far beyond those confines. In this article, Dr Gibson explains what some of the most important advances are, and how they are changing the field of synthetic biology.
Beginning of something big
Back in 2004, when Dr Gibson joined the J. Craig Venter Institute, the group was focused on producing a DNA construct of 583 kb, in the form of the bacterial genome of Mycoplasma genitalium. The M. genitalium genome was 18X larger than anything else that had been published before, mainly because extant methods did not allow for assemblies on this scale. Therefore, to be successful, Dr Gibson and his team had to invent entirely new methods to assemble DNA. When they did assemble this synthetic genome, they were the first to synthesize the complete set of genetic instructions for a bacterial cell. They published this groundbreaking work in 2008 .
In 2010, Dr Gibson and his colleagues assembled the larger, 1.1 Mb Mycoplasma mycoides genome, which, at this date, still holds the record for the largest synthetic DNA sequence ever built . This genome was then activated to produce the first synthetic cell, making headlines around the world. None of this would have been possible with restriction/ligation based cloning methods. So how did they do it?
Free from restriction
One major limiting factor in older cloning methods is the requirement for unique restriction enzyme recognition sequences in the working DNA. In short constructs of <1 kb, it is relatively easy to just add such a sequence that is compatible with plasmid and insert design. However, as the sequences for cloning become longer, the list of unique restriction sites that do not already occur and that will not fragment the target sequence grows significantly shorter.
An additional downside is that adding non-endogenous restriction sites to the construct creates a synthetic sequence that differs from the wild type sequence, and these may have unintended functional consequences. An ideal construct would not leave these "scar sequences" in the final assembly.
In 2009, Dr Gibson and his colleagues published their paper describing an isothermal assembly method that is now known as the Gibson Assembly™ Method . The method does not rely on specific sequences, but uses a unique enzyme mix to assemble double-stranded or single-stranded DNA elements that have overlapping ends of 20–80 bases. An overview comparing key differences between the Gibson Assembly Method and restriction/ligation approaches is shown in Figure 1.
Figure 1. Comparison of Restriction Cloning and Gibson Assembly™ Workflows. Traditional restriction cloning using compatible restriction endonucleases requires 1–2 days of preparative steps to generate cloning ends on insert and plasmid. Typically, only one insert can be ligated into the plasmid at a time. Generating longer inserts usually requires multiple rounds of restriction and ligation. Gibson Assembly is compatible with dsDNA and ssDNA inserts. It requires 20–80 bp sequence overlaps at the ends of the DNA elements to be assembled, and a linearized vector. The sequence overlaps are intrinsic to the desired construct and plasmid, eliminating the need for specific restriction sites. The Gibson Assembly Method allows for several inserts to be simultaneously assembled in a single isothermal reaction that takes only 1 hour, allowing for very large constructs to be rapidly generated. In principle, because it does not rely on restriction sites, the upper size limit for assembled sequences using this method is >1Mb.
The group also described a method for combining up to 25 large pieces of DNA with overlapping ends by transforming them into yeast. The yeast cells take them up and assemble these large DNA pieces into constructs of >1Mb. The resulting large, synthetic constructs are then available in that host organism for producing additional copies .
When asked if his team still uses restriction enzymes for any of their DNA assembly work, Dr Gibson responded, “No. There are no benefits to using restriction methods for gene assembly.” In addition, the Gibson Assembly Method is faster than traditional cloning approaches, requiring only 1 hour for an entire assembly reaction.
Reliance on small cells
Bacterial cells and yeast are still important tools for current generation gene construction. These cells are very useful for highly accurate copying of DNA constructs. However, E. coli and yeast both have an upper limit to the size of DNA that these cells can take in and copy—for E. coli this is approximately 300 kb and for yeast it is ~2 Mb.
In addition, some sequences can be toxic to host cells and, in some instances, in response to foreign DNA elements, E. coli will recombine or mutate the synthetic construct. In other cases, if a foreign gene is toxic to the host cells, few to no colonies will be obtained when cultures are plated.
Dr Gibson explains, “The issues posed by these host cells represent some of the biggest challenges for gene assembly because what happens within these cells is out of our control.” To address the host cell issues, Dr Gibson and his colleagues are currently developing in vitro methods to copy large constructs without the need for host cells that will hopefully circumvent many of these limitations.
Removing steps and controlling errors
The workflow for constructing large DNA assemblies used to require cloning multiple small pieces; weeding out errors by sequencing the pieces; then assembling small, error-free DNA segments into the final, large construct; and sequencing one more time. This is how Gibson and colleagues generated the 16 kb mouse mitochondrion genome . This method is time consuming, and the multiple rounds of cloning and sequencing add significant cost.
Dr Gibson and his team are now working to find new solutions for building sequences that they can be confident are much closer to 100% pure—approaches without intervening cloning and sequencing steps, which will ultimately drive assembly costs down.
The current method, developed internally and commercialized by SGI-DNA, requires only 1 hour for an entire assembly and represents a significant market opportunity. After assembling and amplifying 60mer DNA oligos from IDT into 2–3 kb constructs, they use a proprietary enzyme mix developed for error correction that is able to effectively degrade the majority of incorrect sequences, leaving a pool of very high-quality, assembled DNA. Next, the scientists reamplify that population and, in a matter of hours, are left with a population enriched for the correct sequence that is much faster to screen. Using this approach, one can create sequences between 10–30 kb that are the size of an entire biological pathway in less than a week, and without ever having to clone into E. coli.
When asked about the types of DNA sequences they are able to synthetically assemble, Dr Gibson replied, “We have been able to make everything that we have attempted to so far.” This is not to say that everything is practical to make. Some repetitive motifs and/or domains that are particularly rich in A/T or G/C elements pose special challenges. Researchers at SGI-DNA have developed tools to deal with these demanding sequences.
Cost remains one of the major limitations. At this time, it still costs over $1 million to build a 1 Mb bacterial genome. “Not a lot of labs can afford that. If we can come up with ways to reduce overall costs by reducing the cost of DNA oligos, sequencing, and labor through increased automation, this technology will be more available to more labs and open the door to more research and even greater possibilities,” noted Dr Gibson.
Even longer DNA
Size is probably not much of a limitation for researchers anymore. “People are only starting to realize that we have the ability to build these really large constructs, which was previously a major bottleneck. The lag has now shifted back to DNA designs. If you can easily build constructs of 30–100 kb that can constitute entire biological pathways, all the way up to a whole bacterial genome, what people need to start thinking about is what would they build?”
For their part, Dr Gibson and other scientists at the Venter institute are currently using these methods to understand what a minimal cell might look like. Dr Gibson explained, “We are currently working on building "minimal genomes" that will include only the minimum elements required for a living cell. Building these types of constructs will help us know what these genomes and cells need to look like, and will give us a cell where we completely understand what every single gene is doing.”
Dr Gibson envisions that, “This project will really accelerate gene design and move the field forward because you will have a genomic framework, or 'chassis' to which you can start adding elements or modules. You can then start swapping in and out modules and testing these new modules for functionality.”
Current developments at SGI-DNA
Dr Gibson and his team are always looking for opportunities to refine and improve the DNA assembly methods that he and his colleagues have developed. They are also working on ways to improve large gene delivery to host cells, as well as methods that rely less on host organisms for copying their synthetic constructs.
In addition, as part of his work with SGI-DNA, Dr Gibson is working to make large DNA constructs more cost effective, and developing manufacturing methods to commercialize such constructs. Dr Gibson hopes to eventually be able to produce large assemblies for under a penny/bp by taking advantage of low cost oligonucleotides, improved automation, and reducing DNA sequencing costs.
Working with IDT
In February of this year, SGI-DNA and IDT announced an expanded collaboration. The agreement aims at developing various co-branded DNA products up to 2 Mb orderable from IDT.
Dr Gibson believes the partnership with IDT and SGI-DNA is a perfect match, “…because you have IDT, who makes the best oligos in the world, combined with SGI-DNA and their unique skills, assembling those oligos into very large and/or complex constructs fast and efficiently.”
Dr Gibson earned his PhD in molecular biology from the University of Southern California (Los Angeles, California, USA) in 2004, after which he moved to Maryland with his wife. Upon arriving in Maryland, he attended a job fair on synthetic life at the J. Craig Venter Institute, and was drawn in by their project “to develop approaches for building a synthetic cell from scratch”. Dr Gibson contacted Nobel Laureate, Dr Hamilton (Ham) Smith, who still leads the synthetic biology group at the Venter institute, and is co-founder of SGI, who granted him an interview. Dr Gibson started work at the JCVI a few days later.
Dr Gibson’s success has not kept him out of the lab. He told us, ”I still try to get into the lab and do at least one experiment every week. It helps me keep in touch with all that’s happening, and it is one of the things I enjoy most about my job—so I am not ready to give that up. I have been saying that for years, and so far so good.”
- Schmidt and Markus (2012). Synthetic Biology: Industrial and Environmental Applications (3rd ed.). Weinheim, Germany:Wiley–Blackwell. pp. 1–67. ISBN 3-527-33183-2.
- Gibson DG, Benders GA, et al. (2008) Complete chemical synthesis, assembly, and cloning of a Mycoplasma genitalium genome. Science, 319(5867):1215–1220.
- Gibson DG, Glass JI, et al. (2010) Creation of a bacterial cell controlled by a chemically synthesized genome. Science, 329(5987):52–6.
- Gibson DG, Young L, et al. (2009). Enzymatic assembly of DNA molecules up to several hundred kilobases. Nature Methods, 6(5):343–345.
- Gibson DG, Benders GA, et al. (2008) One-step assembly in yeast of 25 overlapping DNA fragments to form a complete synthetic Mycoplasma genitalium genome. Proc Natl Acad Sci, 105(51):20404–20409.
- Gibson DG, Smith HO, et al. (2010) Chemical synthesis of the mouse mitochondrial genome. Nat Methods, 7(11):901–903.