Anita Ramanathan outlines four synthetic biology approaches that are improving and accelerating drug discovery.
The brainchild of multidisciplinary experts in computer science, physics, engineering, and biology, the field of synthetic biology has not only rapidly progressed in the last few decades but it’s also becoming more accessible and cost-effective. It has made its way into almost every aspect of drug discovery, whether it’s exploring basic biology, modelling diseases, identifying and validating targets, or performing drug screens.
Known for its ability to offer precision and flexibility while expanding the experimental scale, synthetic biology tools can reduce development timelines and costs, and lend a competitive advantage over traditional methods.
Here we outline four synthetic biology approaches that are improving and accelerating drug discovery:
Rapid and accurate DNA synthesis
Conventional DNA synthesis, typically used for making sequences such as primers, employs phosphoramidite chemistry, a method that was developed in the 1980s and is still considered the gold standard for DNA synthesis.
Although simple in its concept, the method itself is highly involved and requires the use of hazardous, unstable chemicals, making the whole ordeal labour-intensive and cost-prohibitive for individual laboratories to synthesise DNA sequences on their own. As such, the synthesis of DNA oligonucleotides for routine molecular biology applications is often outsourced to service provider companies that operate DNA synthesisers to meet these demands.
While chemical DNA synthesis methods are reliable for shorter sequences, such as PCR primers, as DNA sequences become longer, errors start becoming more pronounced. Issues with DNA fidelity, therefore, limit the maximum sequence length to ~300 base pairs. The ultimate goal for biologists is to be able to synthesize longer DNA fragments, at scale, and with high fidelity.
In recent times, to address the shortcomings of chemical-based DNA synthesis, several enzyme-based techniques have emerged that capitalise on the very enzyme that is most accomplished in synthesising DNA – the DNA polymerase. Developed with the idea of engineering an enzyme called terminal deoxynucleotidyl transferase (TdT), a type of DNA polymerase, these enzymatic DNA synthesis methods are now the focus of dozens of companies. As a result, several high-fidelity DNA synthesis methods are being developed, with some approaches directly modifying the TdT enzyme (at companies like Molecular Assemblies, Nuclera, DNA Script), another using a proprietary enzyme combination (Camena Bioscience), or a combination of enzymatic or chemical methods with in silico approaches (Evonetix).
Regardless of the underlying method being used, the way in which longer sequences can be generated at scale is by the parallel synthesis of several shorter DNA sequences and then re-assembling them into longer fragments.
“We facilitate parallel DNA synthesis by coupling our chip technology with our re-engineered chemical method that can thermally control DNA synthesis at every single action site on the microchip,” says Dr Michael Daniels, Head of Product Management at Evonetix.
In this approach, DNA synthesis is parallelised by using a silicon chip with thousands of tiny heaters that thermally control the DNA synthesis reactions occurring on its surface. Thousands of short fragments are synthesised on these reaction sites, which are then assembled into longer double-stranded DNA by selective release. Errors or mismatches are detected and removed by thermal control as well, which means the DNA being parallel-synthesised is not only longer, but also more accurate.
“Because we’re using thermal control, we can deal with some of the sequences that cause problems with existing DNA synthesis methods, such as long repeats or polyAs,” explains Dr Daniels. “We use temperature to make sure these are melted, and hairpin loops never form during the synthesis process. We’re also working towards using temperature to remove hetero duplexes during DNA assembly. If there’s a mismatch between two strands, the DNA will melt away at a lower temperature. This way, we can separate erroneous sequences from accurate ones. Improving the quality of the DNA is as important as improving the speed of synthesis.”
Even as methods in DNA synthesis continue to advance, scientists’ reliance on third-party companies for synthesising DNA creates a bottleneck, impeding research timelines.
“If you want to order gene-length sequences, rather than short oligos for PCR, the waiting times are around two or three weeks at least. The longer the DNA gets, the longer it will take for labs to receive it,” notes Dr Daniels. “If a pharmaceutical company is racing to identify a new vaccine, for instance, then waiting for three weeks to receive DNA sequences each time can slow down the development process compared to competitors who might have faster timelines.”
To enable scientists to better manage project timelines and no longer rely on vendor-imposed waiting times, technology providers are now making DNA synthesis more accessible, simple and affordable. Technologies such as benchtop DNA synthesizers (Evonetix) and desktop DNA printers (DNA Script) make it possible for scientists to generate DNA sequences on demand and no longer limit the scope of projects due to external constraints.
“Scientists tend to limit the ambition of their experiments because they know they can’t get the long DNA they want or can’t get it fast enough to complete the project within the period of funding,” says Dr Daniels. “But if we can democratise DNA synthesis with benchtop systems, it gives labs the opportunity to take on bigger projects.”
Applications: The ability to synthesise longer, accurate DNA sequences opens up a vast array of application possibilities: generating CRISPR libraries, in situ hybridisation probes, antibody libraries, vaccines, and so on, along with the ability to engineer metabolic pathways, microbes, and plants.
Synthetic libraries for antibody drug discovery
Developing therapeutic antibodies, one of the most successful approaches in drug development, involves identifying and characterising antibodies to be used as treatment against different types of cancers as well as autoimmune and cardiovascular diseases. Finding promising antibodies against specific molecular targets, however, is no easy feat. Once a specific target is identified, therapeutic antibody candidates are generated. Using phase display systems and single B-cell sequencing, hundreds, if not thousands, of high-affinity antibody candidates can be identified. The bottleneck in the process appears when these promising candidates need to be expressed as full-length antibodies for further characterisation.
Large amounts of antibodies need to be produced, multiple times, to study and examine functionality, and check for feasibility and efficiency of the antibody leads. This process can be drawn-out, expensive, and time and resource-intensive, often slowing down the momentum of a research study.
Modern advances in DNA synthesis, however, have made it possible to leverage gene-to- antibody production platforms where input candidate DNA sequences can be turned into purified antibodies. The recently launched high-throughput antibody production platform by Twist Bioscience is one such example. “We started by leveraging our ability to make DNA at scale to create high-value synthetic antibody libraries, and we’ve leaned on artificial intelligence to design these libraries,” Dr Emily Leproust, Founder and CEO of Twist Biopharma told DDW earlier this year1.
Gene-to-antibody production begins with scientists entering antibody sequences as starting material, after which bioinformatics tools filter out potentially non-viable sequences. Then, using DNA synthesis capabilities, oligo pools encoding for specific domains of the antibody are generated as libraries and cloned into a vector. “The libraries contain only the sequences that occur in the human repertoire, rather than randomly generated sequences,” adds Dr Leproust.
Parallel processing of DNA synthesis means tens of thousands of unique antibodies can be produced in a short amount of time and made available to the end user as easy-to-use cell lines.
Dr Leproust continues: “More and more therapeutic antibodies are being discovered from antibody libraries. Synthesising DNA at scale allows more libraries to be built, including target class-specific libraries.”
To accelerate drug development timelines, biotechs and pharma companies may choose to outsource some or all parts of antibody discovery and optimisation to DNA synthesis technology partners equipped with the tools and techniques to get reliable results faster. “We are even seeing some companies going completely ‘lab-free’, outsourcing the entire workflow,” says Dr Leproust. “By partnering with us, a pharmaceutical company can screen and identify candidate antibody hits in as little as eight weeks – something in the past that would take years.”
Cell-free systems for drug design and manufacturing
Cell-free systems help address the many limitations imposed by traditional cell-based methods. Although significant milestones in drug discovery have been achieved using cell-based approaches, including engineering CAR-T cells, the inherent reliance on a cellular host can add a layer of complexity and often confine the scope of an experiment.
First off, genetically encoding desired features into living cells requires designing an appropriate vector, successfully transfecting it across the cell membrane, integrating the new DNA into the host’s genome, and finally, a reporting marker to examine changes. The tedious and time-consuming nature of this approach restricts the number of design-build-test cycles possible, especially when time and resources are capped. In cell-free systems, the molecular machinery is extracted from the cells, essentially making them ‘open’ and easier to modify. They contain all the necessary enzymes for transcription and translation, making it possible to execute the central dogma outside a cell.
“The main idea of cell-free systems is to take biological processes, such as gene expression, outside the cell and into a test tube,” says Dr Nadanai Laohakunakorn, University of Edinburgh, who specialises in engineering cell-free systems. “This approach has been around for a while. Researchers used it to elucidate the genetic code and mechanism of protein synthesis back in the 1960s. Now, however, we have finer control over these processes and better capabilities in terms of DNA synthesis and assembly due to recent advances in synthetic biology.”
“It’s now possible to programme cell-free systems to make RNA and proteins that interact with each other,” Dr Laohakunakorn continues. “The goal is no longer just about making one protein but a handful of them, for instance, a group of transcription factors that form a network. This gives us the opportunity to construct complex genetic networks, study their properties and learn about how they function… all outside the living cell.”
Cell-based systems can be classified into two types, cell extract-based, where components of the whole cell are removed as a lysate, and purified enzyme-based, where individual protein elements are isolated, purified, and reconstituted into a defined solution. Cell extract-based systems are quick, more cost-effective, and commonly used for different applications. However, they are prone to easy degradation and may not offer precise control over the molecular machinery. Purified enzyme- based systems, on the other hand, are considered ‘cleaner’ and minimal. Although requiring technical prowess to perform from scratch, ready-to-use kits for these so-called PURE (protein synthesis using recombinant elements) cell-free systems are now commercially available.
Applications: Being more flexible to biological manipulation and faster to generate results, one direct benefit of using cell- free systems in drug discovery is the accelerated time to design prototypes or biotherapeutics.
“In our recent publication, we used cell-free systems to accelerate the design of therapeutics that work both prophylactically and therapeutically against SARS- CoV-2,” says Dr Michael Jewett of Northwestern University, an expert in cell-free synthetic biology who collaborated with a team of protein design experts to address the growing concerns around Covid-19 variants. “One of the challenges we face with SARS-CoV-2 is that the virus continually evolves, making RNA vaccines and antibody therapies less effective. So, we wondered if we could design a therapeutic that essentially works against all variants of concern,” shares Dr Jewett.
“With a team of researchers, we applied computation protein design to create a minibinder protein that forms a tripod-like structure that multivalently binds to the three receptor binding domains of the SARS-CoV-2 spike proteins. We also observed that this protein neutralised the variants of concern, including the omicron variant,” explains Dr Jewett. The results of the study appeared in Science Translational Medicine earlier this year2.
In this study, cell-free systems lent speed of testing and fast- paced screening of a novel therapeutic agent without having to design elaborate in vivo or cell-based studies from the get-go. Dr Jewett continues: “Our collaborators would design the proteins and we’d test them in cell-free systems. We then picked out the top candidates and performed neutralisation assays to demonstrate that they work – all without ever going into a cell.
In addition to accelerating design, cell-free systems support yet another dimension of drug development – local drug manufacturing. The coronavirus pandemic has made us aware of the many supply-chain complexities involved in vaccine production and, subsequently, making vaccines available on time and to the communities that need them. “Because vaccine production is largely centralised and requires cold-chain management, pharma companies build manufacturing plants that cost hundreds of millions to billions of dollars, and then dispatch vaccines from these centralised facilities to local communities,” explains Dr Jewett. “We’ve been trying to rethink this paradigm: What if cell-free systems could be used for local drug manufacturing?”
“One of the big advantages of cell-free systems is that they can be freeze-dried, stored, and shipped anywhere, and they still maintain their stability. This opens up a whole new world of ‘just add water’ type of applications,” notes Dr Jewett. “Cell-free systems can be programmed to make, say, a conjugate vaccine, and can then be used in smaller bioreactors in cities or states to locally manufacture drugs in a fairly higher throughput fashion. For instance, cell-free reactions can make enough medicine for about 40,000 people at 50 cents a dose in something the size of a 1L carton of milk.”
Engineering biosensors for specific functions
Biosensors have long been used as valuable tools for designing therapeutic strategies. They detect specific intracellular changes or the accumulation of small molecules of interest and then translate these into measurable outputs. However, these output signals that are closely linked to the underlying metabolism at or near the drug target site are not always precisely quantifiable. To warrant a biotherapeutic application, biosensing elements not only need to ‘sense’ a physiological change but will also need to produce quantitative signals that can trigger a dose-dependent response, such as delivering a therapeutic at the tumour site.
Recent efforts in synthetic biology are catered towards bringing more precision to engineering biosensors and making measurements more quantitative. “We want to transform bioengineering from a more qualitative approach to a quantitative one,” says Dr David Ross, a project leader in the Cellular Engineering Group at the National Institute of Standards and Technology (NIST). This research group at NIST focuses on designing living measurement systems and engineering them to sense and respond to stimuli in programmed ways. “We’re developing novel ways to better engineer biosensing proteins to sense new molecules or targets – or re-engineer existing biosensors to detect what they already do but in a more quantitatively controlled manner,” explains Dr Ross.
To engineer new protein biosensors or modify current ones, protein variants with different combinations of mutations are created. But rather than screening for ‘winner variants’, which is the conventional approach used in directed evolution, here, every single protein variant is examined. “We’ve developed measurements and machine learning methods where we can test a million different variants simultaneously for the desired function in a single experimental run. This is really important because it’s not possible to predict quantitatively how a change to a protein sequence is going to affect its function,” says Dr Ross.
Applications: Being able to engineer biosensors to perform specific functions can immediately open doors to improved drug discovery and faster design-build-test iterations. For instance, a particular human GPCR protein that is a promising drug target can be engineered into yeast cells, thereby providing millions of biosensors that can then be used to rapidly screen drug candidates and measure protein function without having to generate cell lines or transgenic mice.
Engineering biosensors also find an important place in the emerging field of live biotherapeutics such as the use of tumour-targeting bacterial species that have been carefully modified for controlled and sustained delivery of therapeutic agents into the tumour microenvironment.
By 2027, the synthetic biology market across all industries, drug discovery included, is expected to reach US $32.9 billion3, indicating the growing demand for synthetic biology tools. With so many start-ups bringing innovative techniques into the market, as well as increased academic funding to advance synthetic biology research, it might seem hard to predict what new method will emerge next. However, when observed collectively, synthetic biology tools most likely serve one of three pursuits in biology. “Measure, make, and model,” informs Dr Jewett, the past interim Editor in Chief of ACS Synthetic Biology. “Synthetic biology tools allow us to measure, make or model biological systems, and in doing so, they enhance our ability to reprogram the living world.”
Included in the DDW Synthetic Biology eBook, sponsored by Evonetix
- https://www.ddw-online.com/why- synthetic-biology-is-the-next-big- thing-for-biopharma-15377-202202/
- https://www.science.org/ doi/10.1126/scitranslmed.abn1252
- https://www.businesswire.com/ news/home/20220713005665/ en/32.9-Billion-Worldwide- Synthetic-Biology-Industry-to- 2027—Featuring-Amyris-Codexis- and-Danaher-Among-Others— ResearchAndMarkets.com
About the author
Anita Ramanathan is a science writer and award-winning speaker based in Bristol, UK. In her capacity as a science writer/editor at several digital publications, including NIH Research Matters, she has crafted dozens of stories buried under numbers and scientific findings. A story-teller at heart, Ramanathan also delivers science communication workshops.