Synthetic Genomics: Repurposing Biological Systems for Applications in Engineering Biology

Abstract

Substantial improvements in DNA sequencing and synthesis technologies and increased understanding of genome biology have empowered the development of synthetic genomics. The ability to design and construct engineered living cells boosted up by synthetic chromosomes provides opportunities to tackle enormous current and future challenges faced by humanity and the planet. Here we review the progresses, considerations, challenges, and future direction of the “design–build–test–learn” cycle used in synthetic genomics. We also discuss future applications enabled by synthetic genomics as this emerging field shapes and revolutionizes biomanufacturing and biomedicine.

The ability to read (by sequencing), write (by synthesis), and edit (by genome editing) DNA has rendered biology programmable. As an emerging discipline in synthetic biology, synthetic genomics combines the synthesis of DNA, computational design, and genomic modification of biological systems, which enables us to learn and better understand complex biological systems from the bottom-up. We are then in a better position to formulate a strategy that addresses serious problems facing human health and our environment. In December 2023, the UK government announced their national vision for engineering biology with a two-billion-pounds investment plan to accelerate biological engineering efforts. Earlier in 2023, the US, China, and Australia also released their own roadmaps and strategies that will allow them to expand their capabilities in this emerging field.¹ As a gateway to engineered biology, synthetic genomics may create new opportunities revolutionizing methods of food production, development of new medicines, production of alternative and environmentally friendly materials, new tools to protect our ecosystem, and preparation for solving problems in future scenarios that were barely imagined.

Milestones of Synthetic Genomics

Attempts to construct genomes of model organisms was first done successfully in the early 2000s. A groundbreaking achievement was the first synthesis of a synthetic viral genome, the poliovirus cDNA (∼7.9 kb). The viral genome was synthesized by the polymerase chain assembly (PCA) method using oligonucleotides followed by conventional cloning, demonstrating the feasibility of building a natural DNA template from scratch.² Single-step assembly by employing a pool of gel purified oligonucleotides greatly improved the assembly efficiency, leading to the construction of bacteriophage φX174 (∼5 kb) in two weeks.³ Construction of a chimeric T7 phage with genomically refactored regions (e.g., decoupling of genetic elements and introduction of unique restriction sites for manipulation) introduced the aspects of the “design–build–test–learn” concept and demonstrated the feasibility of studying the fundamentals of life using systematic engineering principles.⁴ Another landmark for synthetic genomics, reported in 2010, was the de novo synthesized ∼1.1 Mb Mycoplasma mycoides genome called JCVI-syn1.0, which was transplanted into a Mycoplasma capricolum recipient cell. This resulted in the generation of the first bacterium with a synthetic genome, also known as “Synthia”.⁵ Subsequent efforts including multiple rounds of transposon mutagenesis were made to obtain the genomically minimal synthetic cell called JCVI-syn3.0.⁶ A following study found seven genes are required together for the normal morphology in JCVI-syn3.0.⁷ The partial and complete constructions of synthetic Escherichia coli genomes with genome-wide substitutions of synonymous codons, which are designated as rE. coli-57 and syn61, demonstrate the plasticity of genetic code and deepen our understanding of the role played by distinct synonymous codons.^8,9 In parallel with the successful efforts of synthetic viral and bacterial genomes, the Sc2.0 yeast genome project was first initiated by Jef Boeke and then gradually became an international collaboration among research institutes across six countries through three continents starting in 2011.¹⁰ Thousands of designer changes, including deletions, insertions, substitutions, and relocations, were introduced into the Saccharomyces cerevisiae 2.0 yeast (Sc2.0) genome by following the design principles (viability, stability, and flexibility). The most unique and powerful design of Sc2.0 yeast is the Synthetic Chromosome Recombination and Modification by LoxP-mediated Evolution (SCRaMbLE). The SCRaMbLE system enables inducible rearrangements of chromosome segments by Cre-mediated recombination between distinct symmetrical loxP sites.¹¹ All of the aforementioned advances have paved the way for further research and diverse industrial applications. In the following, we will discuss the advances and future directions of the four key components (design, build, test, and apply) of synthetic genomics (Figure 1).

Figure 1.

Diagram showing the “design–build–test–learn” cycle of synthetic genomics, which lists key considerations, tools, and strategies involved in the design, construction, and functional evaluation of synthetic chromosomes. The iterative evolution of this cycle enables the creation of synthetic organisms that would be applied in agriculture, medicine, and manufacturing.

Synthetic Genome Design and Principles

Establishing “best practice” design principles plays a fundamental role in the successful construction of synthetic chromosomes and in reliably achieving desired phenotypes. It is important to keep in mind that the ultimate objective of synthetic genomics is not only to create new versions of organisms with enhanced or novel properties but also to prevent or minimize fitness impairment. Achieving success on the genome scale will require both sufficient understanding of the target biological system and the optimal balance among different considerations including introduced function, genome stability, biological compatibility, and technological feasibility. Appropriate computer-aided design tools and relevant data sets are particularly important to the latter. For instance, software-guided segmentation¹² of large DNA fragments helps to overcome some incompatibility between the designed genome sequence and constraints from DNA synthesis and assembly steps.

The current genome design of biological systems often relies on a top-down strategy, where the wild-type genome of a target organism serves as the reference. In contrast, the bottom-up design strategy refers to the selection of functional bioparts followed by the gradual assembly of them to form modules and eventually a synthetic genome. Due to the higher predictability and operability, the majority of the synthetic genomics projects employ the top-down design strategy, which can be categorized into four categories:¹⁰ (i) distinguishing the designer sequence from the native version; (ii) genome refactoring that enables decoupling or relocation of genes into functional modules; (iii) reduction of genome redundancy and genome minimization; and (iv) introduction of enabling design features for enhanced or novel functions. Please note that these designs are not mutually exclusive and are often correlated. For instance, the genome-wide changes themselves could be used to distinguish the synthetic and native sequences. In addition, the systematic introduction of newly designed elements might favor or require the removal of genomic and functional redundancies.

Synthetic Genome Refactoring and Construction

Genome construction consists of three major steps: oligonucleotide synthesis, stepwise genome assembly, and delivery of the synthetic chromosome into cells. As one of the underlying technologies in genome writing, oligonucleotide synthesis has made enormous progress in the past two decades in terms of the scale and cost, which has fueled growth of the DNA synthesis industry and has made genome synthesis possible.¹³ Despite relatively few improvements in phosphoramidite chemistry, miniaturization and automation, which are driven by recent developments in microfluidics and semiconductor technologies, account for most of the cost reduction. It is worth pointing out that the current chemistry-based method is challenging to directly synthesize >200 nt oligos because the synthesis efficiency for each step is often between 98.5% and 99.5%. Therefore, the maximum final yield of desired 200 nt products would only account for 36% of the synthesized product. In contrast, the engineered terminal deoxynucleotidyl transferase (TdT) offers great potential to produce single-stranded DNA (ssDNA) with high efficiency and fidelity. Currently, many research institutes and startup companies are developing enzyme mutants that exhibit improved polymerization performance and other desired features (e.g., high specificity and stability). It is self-evident that our capability to directly synthesize single-strand DNA with several kilonucleotide lengths would fundamentally transform our ways of genome synthesis.

The synthesized oligos are not ready-to-use materials for genome construction; instead, the double-stranded linear DNA constructs assembled from multiple oligos are used for building a gene and eventually a chromosome. Methods based on the T4 polynucleotide ligase and DNA polymerase have been developed to assemble shorter oligos into gene-sized DNA fragments. The PCA method represents one of the most commonly used methods due to its simplicity and versatility. However, it should also be noted that some sequence features (e.g., repetitive DNA sequence and high-GC content) would have huge impacts on the amenability, production time, and commercial price of this step. Various strategies may be utilized to reduce the cost of gene assembly by PCA. First, error reduction and error correction can be achieved using certain nucleases and/or high-fidelity DNA polymerase. In addition, automation is highly suitable for the PCA assembly because all steps in the workflow can occur in a cell-free system. Finally, miniaturization of the assembly reaction reduces the consumption of expensive reagents such as polymerase. To facilitate the construction of megabase-sized chromosomes, larger DNA constructs (from 10 to 100 kb) are often needed. Both in vivo and in vitro assembly strategies using distinct principles have been developed, as summarized elsewhere.¹⁰ Gibson assembly and Golden Gate/MoClo assembly become the most commonly used in vitro methods. Due to the intrinsic capability of budding yeast S. cerevisiae to carry out efficient homologous recombination, S. cerevisiae has been employed as a preferred platform to assemble large DNA constructs, even at the chromosome scale.¹⁴ Besides, Bacillus subtilis is another ideal host for large in vivo DNA assembly.¹⁵ Although the versatility of the B. subtilis platform is limited due to fact that large DNA assemblies need to be integrated into the genome, it allows shuttling of assembled DNA between organisms by conjugation in contrast to S. cerevisiae. Amplification of the successfully assembled DNA fragments is also important to prepare enough materials for genome synthesis. Despite the fact E. coli is the widely used host for amplification of assembled DNA constructs, the cell-free technology serves as a promising solution to amplify large DNA molecules.¹⁶

The last step of synthesizing a genome is to deliver it into a living cell. There are two main strategies available: (i) one-step delivery and (ii) stepwise replacement of the native chromosome by the synthetic fragments. Although direct transplantation of a synthetic genome is possible and seems straightforward, its success has only been demonstrated in a few bacterial species. In addition, isolation and manipulation of genomes are challenging as Mb-scale DNA is extremely fragile. In contrast, these problems can be avoided by iteratively substituting wild-type genomes with their synthetic counterparts. This strategy was utilized and validated in the Sc2.0 and syn61 projects.⁹ It should be noted that the delivery of the synthetic chromosome and the testing step are often interconnected and indivisible. Another advantage of stepwise substitution delivery is the convenience of mapping the potential sequence feature introduced by design, which is also often designated as the design bug and is responsible for the observed fitness defect during the replacement process. Taken together, the key point is that there might not be a single best method that meets all needs in the process of building a synthetic genome from scratch, and combinations of distinct approaches and flexible adjustments would provide the best outcome.

Viability and Functional Testing

After obtaining the synthetic DNA construct in designer cells, the next key step is to test and validate their function. The purpose of “testing” is involved in three major aspects: (i) determining if the synthetic DNA ranging from chromosome to genome size matches the designed sequence in vivo; (ii) determining if the phenotype of synthetic cells matches expectations; and (iii) identifying and repairing discrepancies. The first two parts are straightforward because we are now able to sequence DNA in a matter of hours and check resulting phenotypes using respective assays under various conditions within a few days. In fact, the major challenge in building a fully functional, artificially designed genome is to screen for design bugs and carry out the costly and laborious task of repairing (or debugging). As our understanding about complex biological systems is still incomplete and massive changes are introduced into DNA sequences on the genome level, the design bugs are often inevitable. As the clue of design flaw, fitness defects under specific conditions can be measured and quantified by using a variety of phenotypic assays (e.g., direct growth measurement and indirect phenotypic reporters). As mentioned above, stepwise substitution of the host genome by synthetic fragments could generate a pool of intermediate strains, which serve as valuable material for understanding the biology behind the observed phenotype. By coupling the genotype and phenotype of the intermediate strains, the design deficiency can be mapped to a specific chromosome region. To further facilitate this process, additional debugging strategies have been developed by using the CRISPR/Cas9 system and the meiotic recombination process of budding yeast.¹⁷ With the progress of different genome synthesis projects, many interesting bugs and the underlying mechanisms have been described, which greatly expands our understanding of genome biology.¹⁰ Apart from the phenotypic measurement, extensive characterization of the synthetic organism via trans-omics tests also enables us to discover subtle differences between the synthetic cell and their wild-type counterparts as shown in the case of Sc2.0 yeast chromosome II.¹⁸ The debugging process would be a far more daunting task for building a genome via a one-step genome delivery strategy because the inability to boot up the synthetic genomes may be due to the failure of genome transfer or the synthetic construct resulting in nonviable cells. For instance, considerable efforts were made to identify quasi-essential genes, which are not strictly essential but are nevertheless needed for long-term fitness, during the construction of a minimized bacterial genome.¹⁹ Once the design bugs have been revealed, they could be rapidly corrected by genome editing tools such as CRISPR/Cas9 as well as methods based on homologous recombination. For situations where the source of a growth defect is unknown, adaptive evolution could serve as a powerful way to repair the design flaw.

Despite the different strategies and methods available to verify, characterize, and learn the synthetic genomes, improving the testing step would require more powerful platforms that could simultaneously test fitness and perform trans-omics measurements in a highly parallel and high-throughput manner. By doing this, huge biological data sets would be generated. We envision that the wealth of this information would grow rapidly with increasing numbers of genome writing projects, open collaborations, and data sharing. By applying the artificial intelligence (AI) and machine learning (ML) techniques, these data sets would serve as valuable training data to build computational models that might be used to reduce the bugs in future designs and perhaps be used to create automated design tools.

Applications of Synthetic Genomics

Synthetic genomics allows us not only to study and learn fundamentals of life via the “bottom-up” strategy but also to design and build new organisms and biological systems that can be used to solve many of the grand challenges facing us today. Synthetic biology may allow us to cure human disease, tackle environmental problems, find alternative methods to produce industrial materials, and produce foods more sustainably. Synthetic genomes with designer features enable the creation of cells with enhanced biological functions and unique properties (e.g., metabolic diversity and extremophile tolerances). This is probably best exemplified by the inducible evolution system called SCRaMbLE that is embedded in the Sc2.0 yeast strains. By exploring the enormous amount of genomic diversity post-SCRaMbLE, researchers can rapidly acquire mutant strains with improved capability to produce bioproducts for industrial applications.²⁰⁻²³

The creation of cells with altered genetic codes by de novo genome synthesis opens the door to promising applications such as the biosynthesis of unnatural polymers consisting of noncanonical amino acids (ncAAs), biocontainment of genetically modified organisms, and the generation of virus-resistant cells used for producing therapeutic biologics. Inspired by examples of the deviations from the standard genetic code found in nature,²⁴ synthetic biologists now could use engineered aminoacyl-tRNA synthetase/tRNA pairs to site-specifically incorporate ncAAs into proteins of interest, a technology called genetic code expansion (GCE). As each of the 64 triplet codons are already utilized to guide the synthesis of proteins, one of the key challenges to performing GCE in living cells is to create a blank codon that is explicitly assigned to a designed ncAA. Genome-wide replacement of synonymous codons by multiplex editing methods such as MAGE technology²⁵ and high-fidelity genome synthesis represents a fundamental way to generate blank codons for genetically encoding ncAA.²⁶ The genomically recoded organism (GRO) serves as an ideal chassis for GCE because it enables the incorporation of multiple distinct ncAAs into proteins²⁸ and the significant improvement of GCE efficiency.²⁷ By introducing the orthogonal aminoacyl-tRNA synthetase/tRNA pairs into the GRO, robust biocontainment systems relying on ncAA have been developed, which could prohibit unwanted propagation of GRO in open environments. Recent studies also show the GRO with altered genetic code is resistant to viral infections and transfer of mobile genetic elements.^29,30

Despite these exciting achievements using E. coli as the model organism, much less progress has been made in the eukaryotic system, which is more suitable for expression of proteins that require post-translational modifications and folding. The genome-wide substitution of the TAG stop codon by TAA in the design of Sc2.0 yeast aims to free a blank codon for GCE; however, it is still challenging to completely repurpose the specificity of eRF1, the single omnipotent release factor required for cell viability. To promote future applications of GCE in budding yeast, a well-studied model eukaryote and a commonly used chassis in biomanufacturing, we are constructing a synthetic yeast strain with a compressed genetic code to release multiple blank codons including sense codons. The unexpected risk of the recoded yeast on natural ecosystems and human health could be minimized by using the ncAA-dependent biocontainment system.³¹ For other synthetic genomics studies in eukaryotic systems, different pilot projects were proposed in Genome Project–Write (GP-write). A notable plan is to create virus-resistant human cells that could be used to produce antibodies, growth factors, vaccines, and other valuable biologics without the risk of viral contamination.

Conclusion and Outlook

Synthetic genomics represents an emerging field and benefits from the decreased cost of DNA synthesis and rapid development of DNA assembly and manipulation approaches. If a great number of designed versions of large DNA fragments or even chromosomes could be constructed and tested in a short period of time, this field is going to grow explosively. Although we now know that genomes exhibit high plasticity by learning from different genome synthesis projects, very small genomic changes might cause a severe growth defect. Thus, our abilities to prevent or minimize design bugs and to achieve fast and effective debugging are crucial to increase the success rate of booting up of the synthetic genomes to generate living cells. Although many challenges still need to be overcome, we are confident that synthetic genomics serves as a gateway to engineering biology which will revolutionize humans’ way of manufacturing and therapies.

Author Contributions

X.F. and Y.S. both wrote the manuscript and read and approved the final version.

The authors declare no competing financial interest.