DNA synthesis technologies to close the gene writing gap

Abstract

Synthetic DNA is of increasing demand across many sectors of research and commercial activities. Engineering biology, therapy, data storage and nanotechnology are set for rapid developments if DNA can be provided at scale and low cost. Stimulated by successes in next generation sequencing and gene editing technologies, DNA synthesis is already a burgeoning industry. However, the synthesis of >200 bp sequences remains unaffordable. To overcome these limitations and start writing DNA as effectively as it is read, alternative technologies have been developed including molecular assembly and cloning methods, template-independent enzymatic synthesis, microarray and rolling circle amplification techniques. Here, we review the progress in developing and commercializing these technologies, which are exemplified by innovations from leading companies. We discuss pros and cons of each technology, the need for oversight and regulatory policies for DNA synthesis as a whole and give an overview of DNA synthesis business models.

Keywords: DNA; Synthetic biology.

Fig. 1. The state of the art in DNA synthesis.

a, Productivity of DNA reading and DNA writing (synthesis) estimated in the number of nucleotides per person per day. The grey arrow denotes the current gap in productivity between reading DNA and writing DNA. The dashed oval outline highlights the time frame within which the DNA synthesis industry achieved the majority of important milestones to close the gap. DNA synthesis data (red line) are available only for column-based synthesis instruments. The number of transistors per chip (Moore’s law) is shown for comparison. The graph uses the data available in the literature. b, Timeline of milestones in DNA synthesis technologies discussed in the report^,,,,,,,,. For simplicity not all milestones are shown. NTP, nucleoside 5ʹ-triphosphate; PCA, polymerase cycling assembly; TdT, terminal deoxynucleotidyl transferase; TiEOS, template-independent enzymatic oligonucleotide synthesis. Copyright Wiley-VCH GmbH. Reproduced with permission from ref. .

Fig. 2. Mechanisms for solid-phase oligonucleotide synthesis and depurination.

A, Schematic representation of phosphoramidite solid-phase oligonucleotide synthesis. The method sequentially adds 5ʹ-dimethoxytrityl-protected nucleoside phosphoramidites 4 upon activation by tetrazole to ensure sequence-specific strand elongation^–. The steps of the process include a synthesis cycle comprising: protonation (step a), detritylation (step b), tetrazole activation and coupling (step c), capping of unreacted nucleotides on the resin (step d), oxidation (step e), detritylation (step f), which is repeated n times (step g), and followed by cleavage from the support (step h) and deprotection (step i) to yield a desired DNA 10. B, Mechanism of acid-catalysed depurination as a common side reaction in chemical synthesis^–,, comprising protonation (step j), depurination (step k), hydrolysis (step l) and elimination (step m). Resulting apurinic site 14 owing to the loss of a purine base (for example, adenine 13) is readily hydrolysed (steps c and d) during the basic work-up steps required for removing base protecting group (PG).

Fig. 3. Mechanisms for TiEOS.

A, Schematic representation of 3ʹ-protected nucleoside 5ʹ-triphosphate (NTP) approach. Resin beads are pre-loaded with an initiator DNA (iDNA) 19 to provide a template for binding of terminal deoxynucleotidyl transferase (TdT) and as a post-synthesis cleavage site^,,,,. Oligonucleotide synthesis then proceeds in a stepwise fashion in the 5ʹ-to-3ʹ direction. TdT ligates NTP 20 to the 3ʹ terminus of the growing oligonucleotide chain with each NTP protected at 3ʹ-OH with a protecting group (PG) 24–26 (refs. ^,,). The resin is washed to remove surplus reagents and the pyrophosphate by-product of the ligation. After deblocking or deprotection of the 3ʹ-PG (step b), the resin-bound 3ʹ-OH nucleophile of 22 becomes available for the next synthesis cycle (step c). The complete sequence is assembled by repeating the cycle of TdT-catalysed NTP(PG) coupling (step a) and deblocking (step b). On completion, the synthesized oligonucleotide 4 is cleaved from the solid support (step d) by uracil DNA glycosylase. B, Examples of NTP(PG)s used in the method — 3ʹ-azidomethyl-protected NTPs 24 by Nuclera Nucleics, Molecular Assemblies, 3ʹ-ONH₂-protected NTPs 25 by DNA Script and 3ʹ-O-2-nitrobenzyl 26 by Camena Bioscience^,,,,. C, Schematic representation for alternative (tethered) protecting strategies. 3ʹ-Unprotected NTPs (cytidine) are supplied pre-immobilized within the TdT-active site 28, via a short and labile linker^,. TdT then catalyses the incorporation of this NTP into the growing DNA strand 30 (step a) and sterically prevents the uncontrolled polymerization of the NTP until the linker is cleaved (step b), releasing the oligonucleotide 32. The cycle is repeated (step c) until the desired oligonucleotide 33 is completed (step d). Asp, aspartic acid; DTT, dithiothreitol; TCEP, tris-carboxyethylphosphine; TiEOS, template-independent enzymatic oligonucleotide synthesis.

Fig. 4. Schematic representations of assembly methods for DNA synthesis.

A, Gibson assembly^,,. Two duplex DNA strands 34 and 35 are selected with a complementary terminal overlap region (black). Digestion with T5 exonuclease (step a) degrades each strand of the DNA duplexes in the 5ʹ-to-3ʹ direction, yielding the sticky ends of 36 and 37, followed by the annealing of complementary sticky ends between the two DNA duplexes (step b). Phusion polymerase and Taq ligase are then combined (step c) to ligate the two short DNA duplexes into a single, long DNA duplex construct 38 (refs. ^,). The process is repeated (step d) for gene assembly. B, Polymerase cycling assembly (PCA)^,. High-purity synthetic oligonucleotides 39 are designed, such that annealing of complementary overlaps generates the desired long duplex DNA construct (step a). The desired construct is then assembled in either a single step from 43 or two steps from 39 using a DNA polymerase (step b) to yield template 40. PCR amplification (steps c and d) of 41 amplifies the desired long duplex DNA construct 42.

Fig. 5. Schematic representation of thermally controlled DNA synthesis.

Terminally protected oligonucleotide strands 44 and 45 are immobilized on discrete reaction sites (sites 1 and 2). Thermal heating of a chosen site (site 2) cleaves terminal protecting groups (PGs) 47 (step a), enabling selective elongation of strands on this site^,,, (steps b and c) Elongation of a desired oligonucleotide is performed via TiEOS or the phosphoramidite method to selectively generate 49 (refs. ^,). Repeated steps a–c on other selected reaction sites (for example, site 1) sequentially produce bespoke oligonucleotides 50 (steps d and e). Thermally assisted reagent treatment of selected sites cleaves safety catch linkers and liberates oligonucleotides 53 from a chosen site^, (step f). The liberated oligonucleotides (step g) anneal to complementary chip-bound oligonucleotides 52 producing perfectly annealed double-stranded DNA 54, which has a higher denaturation temperature than that of DNA duplexes formed by oligonucleotides with mismatches 55. The heating of reaction site 1 allows the mismatched oligonucleotides 56 to be washed away (step h). Thermally assisted reagent treatment of site 1 cleaves safety catch linkers and releases the desired duplex DNA 57 from the chip^, (step i). This liberated DNA is annealed to a chip-bound complementary DNA duplex 58 to form the nicked construct 59 (step j). The process is repeated to elongate the double-stranded DNA until the desired gene is assembled.

Fig. 6. Gene synthesis strategies.

A, Gene synthesis from diverse oligonucleotide libraries. 5ʹ-Phosphorylated sense oligonucleotide strands 60 and 61 are annealed (step a) to complementary 5ʹ-phosphorylated antisense strands 62 and 63. The resulting DNA duplexes 64 and 65 have two 5ʹ-overhangs or sticky ends that are used to anneal (step b) the duplexes into an extended, ‘nicked’ duplex 66 (‘nick’ highlighted in magenta). T4 DNA ligase is then used (step c) to stitch the oligonucleotides at the nick site into an elongated, larger DNA duplex 67. Cycles of annealing and ligation are repeated until the desired gene is assembled^– (step d). B, DNA microarrays. A library of bespoke single-stranded oligonucleotides 68 is generated with 3ʹ-terminal and 5ʹ-terminal DNA ‘barcodes’ on a miniaturized chip^,,,,,, (step a). These sequences are cleaved (step b) from the microchip to yield a pool of template oligonucleotides with a range of DNA ‘barcodes’ (only two, black and brown, ‘barcodes’ are shown for clarity)^,. Primers selectively anneal to either the ‘brown’ 69 or ‘black’ 70 DNA barcodes and specifically amplify oligonucleotides via PCR (step c), according to the identity of the barcode at its 3ʹ and 5ʹ termini^–. The resulting duplex DNA constructs 71 and 72 still contain the DNA barcodes at their termini, which must be removed prior to gene assembly. DNA barcodes are cleaved (step d) from the duplex DNA 71 and 72 by type IIS restriction endonucleases (REN), giving rise to assembly pools of sequences 73 and 74 with sticky ends. Duplex DNA fragments are annealed (step e) via complementary sticky ends and assembled into desired genes 75 and 76 via Gibson assembly^,,. C, Rolling circle amplification (RCA). Template plasmid DNA 77 with a desired gene cassette (green) and protelomerase sites (magenta) is thermally denatured (step a) to create a single-stranded template 78 (ref. ). A complementary primer binds to the protelomerase sites 79 (step b) and the template is amplified via RCA (steps c–e), to produce double-stranded concatemeric DNA 82 with alternating copies of the desired cassette (green) and the unwanted plasmid backbone (black)^,–. Protelomerase then cuts (step f) the duplex at its recognition sites and ligates the cut ends generating covalently closed ‘doggybone’ DNA (dbDNA) 84 and a circular plasmid DNA 83 as a by-product. The circular backbone of the plasmid DNA is subsequently cut (step g) by REN and digested (step h) by exonucleases.

Fig. 7. User autonomy in DNA synthesis.

DNA synthesis technologies plotted versus complexity levels of the current offerings (left axis) by different companies who develop these methods and versus user autonomy (deskilling) that these companies offer (right axis), from expert involvement to complete autonomy for the end user. For example, template-independent enzymatic oligonucleotide synthesis (TiOES) is the basic technology for DNA Script who offer customized and multiple sequences, while integrating automation into their synthesis workflows, which results in the development of benchtop DNA printers that the end user can buy and use with a minimum expert involvement. Gibson, Gibson assembly approaches; POS, phosphoramidite oligonucleotide synthesis; RCA, rolling circle amplification.