Nucleic acid synthesis is the process of building DNA or RNA chains, whether by the molecular machinery inside your cells or by chemists and engineers working in a lab. Every time a cell divides, it copies roughly three billion DNA base pairs with remarkable speed and accuracy. Every time a cell needs a protein, it transcribes a stretch of DNA into messenger RNA. And increasingly, researchers synthesize nucleic acids from scratch for purposes ranging from mRNA vaccines to storing digital data. The term covers a surprisingly wide landscape, from the ancient chemistry that may have sparked life on Earth to cutting-edge machines that print custom DNA sequences on silicon chips.
How Cells Copy Their Own DNA
When a cell prepares to divide, specialized enzymes called DNA polymerases read one strand of the double helix and assemble a matching strand nucleotide by nucleotide. The process is fast, but it is also astonishingly precise. The polymerase itself selects the correct base most of the time, and a built-in proofreading function catches most remaining mistakes. Proofreading works by shifting a mismatched nucleotide from the polymerase’s building site over to a separate editing site, snipping off the wrong base, and then sliding the strand back for another attempt. This activity alone improves accuracy by roughly a hundred- to a thousand-fold.1PubMed Central. Fidelity of DNA replication-a matter of proofreading Recent structural work on the human mitochondrial DNA polymerase has revealed that the enzyme accomplishes this through a “bolt-action” mechanism, cycling back and forth between its two catalytic sites without ever letting go of the DNA strand.2Nature Communications. Structural basis for DNA proofreading
One complication of replication is that the two strands of the double helix run in opposite directions. The polymerase can only build in one direction, so one strand (the “leading” strand) gets copied smoothly while the other (the “lagging” strand) has to be assembled in short pieces called Okazaki fragments. Each fragment starts with a small RNA primer that the cell later removes and replaces with DNA. In both bacteria and more complex organisms, an enzyme displaces the RNA primer into a flap-like structure, which is then cut away and the resulting gap sealed by a ligase.3PubMed Central. Okazaki fragment metabolism Work in yeast has confirmed two parallel routes for this removal: a flap-cleavage pathway and a direct exonuclease pathway, both operating in living cells.4Journal of Biological Chemistry. Direct Visualization of RNA-DNA Primer Removal from Okazaki Fragments Provides Support for Flap Cleavage and Exonucleolytic Pathways in Eukaryotic Cells
Even the final joining step matters for fidelity. DNA ligase, the enzyme that seals Okazaki fragments together, turns out to have its own quality-control role. When researchers studied a yeast strain carrying a low-fidelity version of the ligase, they found it sealed fragments even when the polymerase had slipped in an extra base, leading to insertion mutations. Normal ligase refuses to join such imperfect substrates, acting as an additional checkpoint.5Nature Communications. High-fidelity DNA ligation enforces accurate Okazaki fragment maturation during DNA replication
Transcription and RNA Synthesis in the Cell
When a cell needs to express a gene, it does not copy the entire genome. Instead, an RNA polymerase reads one gene’s worth of DNA and builds a single-stranded RNA copy. The basic chemistry is similar to replication: nucleotides are added one at a time, each selected by base-pairing with the template strand. But the details differ in ways that affect accuracy and regulation.
Structural studies of viral RNA polymerases have illuminated the earliest moments of this process. During initiation, the polymerase must bring together two nucleotide substrates and catalyze the first bond between them. Crystal structures of a bacteriophage RNA polymerase show that only the correct base pairing at the second template position triggers a conformational shift in the enzyme that completes the active site and allows the reaction to proceed.6PubMed Central. X-ray crystal structures elucidate the nucleotidyl transfer reaction of transcript initiation using two nucleotides Kinetic work on a related phage system suggests that the rate-limiting step during initiation is either the formation of the first bond or a conformational change just before it, a bottleneck that likely helps the enzyme reject wrong starting nucleotides and maintain the fidelity of RNA synthesis.7PubMed. Kinetic mechanism of transcription initiation by bacteriophage T7 RNA polymerase
Building DNA and RNA from Scratch in the Lab
Chemical synthesis of nucleic acids dates back decades, but the core method used today, phosphoramidite chemistry, remains dominant. In this approach, nucleotides are added one at a time to a growing chain that is tethered to a solid support, typically controlled-pore glass beads. Each addition cycle involves removing a protective group from the chain’s end, coupling the next nucleotide, capping any chains that failed to extend, and then oxidizing the new linkage to stabilize it. The same basic process works for both DNA and RNA, with adjustments for different protecting groups, and it can accommodate non-natural modifications at various positions on the nucleotide.8PubMed Central. Protocol for the Solid-phase Synthesis of Oligomers of RNA Containing a 2′-O-thiophenylmethyl Modification and Characterization via Circular Dichroism
One persistent headache with chemical synthesis is depurination, a side reaction in which the acid used to remove protective groups also damages purine bases (A and G). Researchers have measured how quickly this happens with different acids and concentrations, finding that the choice between dichloroacetic acid and trichloroacetic acid, and their concentrations, dramatically affects the trade-off between complete deprotection and unwanted base loss.9PubMed Central. Kinetic studies on depurination and detritylation of CPG-bound intermediates during oligonucleotide synthesis This balancing act is why synthesizing very long oligonucleotides chemically remains challenging. Even a small per-step failure rate compounds over hundreds of cycles, so chemical synthesis is generally practical for sequences up to about 200 bases.
Enzymatic DNA Synthesis Without a Template
An emerging alternative to chemical synthesis uses enzymes instead of chemical reagents to add nucleotides one at a time. The workhorse enzyme for this approach is terminal deoxynucleotidyl transferase, or TdT, which naturally adds random nucleotides to DNA ends in immune cells. By feeding TdT nucleotides carrying a removable blocking group on their sugar, researchers force the enzyme to add exactly one base per cycle and then stop. After the block is removed, the next cycle begins.
The main challenge has been making TdT fast and reliable enough for practical use. One group reported engineering a TdT through 32 rounds of directed evolution, changing about a fifth of the protein’s sequence. The resulting enzyme raised incorporation efficiency by roughly 200-fold to above 99% for blocked nucleotides, and extension times dropped by more than 600-fold to about 90 seconds per cycle.10Nucleic Acids Research. Evolving a terminal deoxynucleotidyl transferase for commercial enzymatic DNA synthesis A separate effort using structure-guided engineering with AlphaFold3-predicted models produced a variant achieving over 98% stepwise efficiency for short sequences.11Synthetic and Systems Biotechnology. Semi-rational engineering of terminal deoxynucleotidyl transferase for high-efficiency enzymatic de novo DNA synthesis Enzymatic synthesis avoids the harsh chemicals and organic solvents of phosphoramidite chemistry, and proponents hope it will eventually produce longer sequences with fewer errors.
Amplifying Nucleic Acids
Synthesizing a nucleic acid sequence is one thing; getting enough of it to use is another. The polymerase chain reaction, or PCR, remains the standard method for copying a specific DNA segment millions of times over. It works by cycling between high temperatures (to separate the two strands) and lower temperatures (to let primers bind and a heat-stable polymerase extend them). The classic enzyme is Taq polymerase, but newer alternatives offer advantages. A proofreading polymerase from the deep-sea archaeon Pyrococcus abyssi, for instance, showed not only better tolerance for high denaturation temperatures than Taq but also greater resistance to common inhibitors, plus the ability to extend primers carrying up to two mismatches at their ends.12PubMed Central. PCR performance of a thermostable heterodimeric archaeal DNA polymerase
Not every application requires a thermal cycler. Isothermal amplification methods run at a single temperature, which makes them attractive for point-of-care diagnostics and field testing. Strand displacement amplification relies on a restriction enzyme that nicks one strand of a recognition site and a polymerase that extends from the nick while pushing the downstream strand out of the way.13Nucleic Acids Research. Strand displacement amplification—an isothermal, in vitro DNA amplification technique Loop-mediated isothermal amplification (LAMP) uses a polymerase with strong strand-displacing activity and a set of specially designed primers, two inner and two outer, that generate looped DNA structures enabling continuous amplification without temperature changes.14Nucleic Acids Research. Loop-mediated isothermal amplification of DNA LAMP gained widespread attention during the COVID-19 pandemic as a faster, simpler alternative to PCR for field diagnostics.
Scaling Up and Assembling Genes
Synthesizing short oligonucleotides is routine, but most applications in synthetic biology require much longer sequences, from entire genes to whole metabolic pathways. The solution is to synthesize short overlapping fragments and then assemble them. Microarray-based synthesis can produce enormous numbers of unique oligonucleotide sequences in parallel on a silicon chip or glass slide, offering huge throughput advantages and lower reagent costs per sequence.15PubMed Central. Automated high-throughput DNA synthesis and assembly A recent system using a microchip-based approach with an iterative “identification, sorting, synthesis, recycling” mechanism increased DNA product concentrations by four to six orders of magnitude over previous chip methods, simplifying downstream assembly of full-length genes.16PubMed Central. Scaling DNA synthesis with a microchip-based massively parallel synthesis system
Once short pieces are in hand, they need to be stitched together. Gibson Assembly, an isothermal method developed by Daniel Gibson, combines three enzymes in a single tube: an exonuclease that chews back the ends of overlapping fragments to create single-stranded overhangs, a polymerase that fills gaps, and a ligase that seals the joints. This method can seamlessly join multiple fragments into constructs spanning hundreds of kilobases.17PubMed Central. Enzymatic assembly of DNA molecules up to several hundred kilobases Robust methods for assembling hundreds of DNA fragments into constructs over 20 kilobases have made the optimization of large genetic elements for metabolic engineering increasingly routine.18PubMed. Methods and applications for assembling large DNA constructs
Catching Errors After Synthesis
Chemical synthesis introduces errors, typically substitutions, deletions, and insertions, at rates that make error correction essential for any construct longer than a few hundred bases. A systematic comparison of error-correction enzymes tested several candidates head to head. ErrASE proved most effective, dropping the error frequency from an intentionally elevated rate of about 45 errors per kilobase down to roughly 8 after two rounds of treatment. T7 Endonuclease I came in second at about 9 errors per kilobase, followed by MutS at about 11.19PubMed Central. A systematic comparison of error correction enzymes by next-generation sequencing The same study found that adding ligase to the T7 Endonuclease I treatment actually made things worse, and that T7 Endonuclease I’s performance varied widely depending on protocol details and enzyme concentration.
Newer approaches aim for higher throughput and automation compatibility. One team engineered a version of E. coli’s MutS mismatch-recognition protein and combined it with a biotin-streptavidin magnetic bead system, creating a platform that physically pulls error-containing duplexes out of a mixture. This approach avoids the cleavage-based strategies that can fragment the desired product and integrates more easily into automated synthesis pipelines.20PubMed Central. Engineered eMutS based magnetic separation system for high-fidelity DNA synthesis
Nucleic Acids as Drugs
Synthetic nucleic acids are not just research tools. A growing number of approved drugs are oligonucleotides, short strands of modified DNA or RNA designed to silence disease-related genes or correct faulty RNA processing. The key to making these work as medicines has been chemical modification. Unmodified RNA or DNA would be chewed up by enzymes in the blood within minutes and would struggle to enter cells. Modifications such as phosphorothioate backbones, 2′-fluoro, 2′-O-methyl, and 2′-O-methoxyethyl (MOE) sugar substitutions dramatically improve stability, binding affinity, and pharmacokinetic properties.21Nucleic Acids Research. Chemistry, structure and function of approved oligonucleotide therapeutics22PubMed. The Medicinal Chemistry of Therapeutic Oligonucleotides Despite decades of progress, getting oligonucleotides past cell membranes and out of endosomes remains a significant barrier.23PubMed Central. Therapeutic Oligonucleotides: An Outlook on Chemical Strategies to Improve Endosomal Trafficking
The mRNA vaccines against COVID-19 brought a different branch of nucleic acid therapeutics into the spotlight. These rely on in vitro transcription using T7 RNA polymerase to produce full-length mRNA from a DNA template. One challenge is that the polymerase sometimes generates truncated or aberrant RNA products alongside the desired full-length strand. Engineering the polymerase itself can help: mutating a single amino acid (K389A) in T7 RNA polymerase raised the proportion of intact, full-length mRNA from lower baselines to about 88%, and combining mutations pushed integrity above 91%.24PubMed Central. Effective Synthesis of High-Integrity mRNA Using In Vitro Transcription
Getting Synthetic Nucleic Acids into Cells
Making a synthetic nucleic acid is only half the battle. You also need to deliver it to the right cells and get it working once inside. For mRNA therapeutics and many oligonucleotide drugs, lipid nanoparticles have become the dominant delivery vehicle. These tiny fat droplets protect the cargo from degradation and ferry it across the cell membrane through processes like macropinocytosis and clathrin-mediated endocytosis. The real bottleneck is what happens next: the nanoparticles tend to get trapped in endosomes, acidic compartments inside the cell. Only a fraction escape into the cytoplasm where the mRNA can be translated into protein. Positively charged lipid components help by interacting with the negatively charged endosomal membrane, promoting fusion and leakage of the cargo into the cell’s interior.25Nature Reviews Materials. Lipid nanoparticles for mRNA delivery The COVID-19 vaccines and the siRNA drug Onpattro demonstrated the real-world viability of this delivery strategy.26PubMed. The impact of lipid compositions on siRNA and mRNA lipid nanoparticle performance for pulmonary delivery
Xeno Nucleic Acids and Expanding the Genetic Alphabet
Natural nucleic acids use a deoxyribose sugar (in DNA) or a ribose sugar (in RNA) as their backbone. But researchers have engineered polymerases capable of reading and writing “xeno” nucleic acids, or XNAs, which swap in completely different sugar or backbone chemistries. One example is HNA, which uses a six-membered anhydrohexitol sugar instead of the usual five-membered ring. A team engineered a variant of an archaeal DNA polymerase, introducing 18 mutations, to give it the ability to reverse-transcribe HNA back into DNA, enabling the information stored in these artificial polymers to be read and copied.27Nucleic Acids Research. Structural insights into a DNA polymerase reading the xeno nucleic acid HNA XNAs are interesting for several reasons. They resist natural nucleases, making them potentially more stable for therapeutic use. They could also serve as an orthogonal information system that does not cross-talk with natural biology, a useful property for biocontainment in synthetic organisms.
How Nucleic Acids May Have First Appeared
Before any cells existed, the building blocks of nucleic acids had to arise from simple chemistry. One influential line of research has shown that the precursors of ribonucleotides, amino acids, and lipids can all be derived from hydrogen cyanide and a handful of related compounds under conditions plausible for early Earth, suggesting that the raw materials for RNA, proteins, and cell membranes could have emerged together rather than sequentially.28PubMed Central. Common origins of RNA, protein and lipid precursors in a cyanosulfidic protometabolism
Getting from raw nucleotides to functional RNA strands without enzymes is a harder problem. Experiments have demonstrated that activated short RNA oligomers can serve as helpers, catalyzing template-directed copying of RNA with all four bases at an overall fidelity of about 98%. That level of accuracy would theoretically support genomes of around 50 nucleotides, long enough for simple functional ribozymes. The same research group showed that this nonenzymatic process could produce a significant portion of an active hammerhead ribozyme, a small RNA catalyst, from monomers and trimers alone.29PubMed Central. Nonenzymatic copying of RNA templates containing all four letters is catalyzed by activated oligonucleotides More recent work has shown that ribozymes themselves can mediate both sense and antisense RNA synthesis in conditions mimicking Hadean-era microenvironments, and that this copying can even occur in the absence of a seeding template when the ribozyme itself serves as the starting template.30Nature Communications. Ribozyme-mediated RNA synthesis and replication in a model Hadean microenvironment
Biosecurity and Screening of Synthetic Nucleic Acids
As DNA and RNA synthesis becomes cheaper and more accessible, the risk of misuse grows. Someone with bad intentions could, in principle, order synthetic genes encoding dangerous toxins or components of pathogens. For years, biosecurity guidance focused primarily on double-stranded DNA, because that was the main product of gene synthesis companies. But the landscape has changed. Updated screening frameworks now recommend that providers screen orders of all forms of synthetic nucleic acids, including single-stranded DNA, single-stranded RNA, and double-stranded RNA, not just the traditional double-stranded DNA. The guidance also extends to manufacturers of benchtop synthesis equipment, recognizing that as machines become small enough for individual labs, the screening burden can no longer rest solely on commercial gene-synthesis providers.31PubMed Central. Enhancing Gene Synthesis Security: An Updated Framework for Synthetic Nucleic Acid Screening and the Responsible Use of Synthetic Biological Materials
The practical challenge is that screening must keep pace with synthesis speed and cost. When ordering a gene took weeks and cost thousands of dollars, there was ample time to review sequences against databases of known pathogens. Now that turnaround times are measured in days and costs have dropped by orders of magnitude, automated screening systems need to be fast, accurate, and integrated into the ordering workflow without creating bottlenecks that push users toward unscreened alternatives.
From Khorana’s First Gene to Genome-Scale Writing
The idea that you could build a functional gene entirely from chemical parts was first realized by Har Gobind Khorana, who in 1970 synthesized the first fully artificial gene outside a living organism by linking nucleotides in a precise sequence. By 1979, he had demonstrated that this synthetic gene worked inside a bacterium.32PubMed Central. Unraveling the Genetic Code: The Legacy of Har Gobind Khorana (January 9, 1922, to November 9, 2011) At the time, each nucleotide coupling was painstaking and expensive. Today, microchip-based synthesis platforms can produce hundreds of thousands of unique oligonucleotide sequences simultaneously, and assembly methods can stitch them together into constructs spanning tens or even hundreds of kilobases. The trajectory from Khorana’s single gene to modern genome-writing projects illustrates how improvements in both synthesis chemistry and assembly logic have compounded over half a century, turning what was once a heroic one-off demonstration into an increasingly routine manufacturing process.