Science Wiki

DNA synthesis is the biochemical or chemical assembly of deoxyribonucleotide monomers into covalently linked DNA polymers. In living organisms, template-directed DNA synthesis is executed *in vivo* during chromosomal DNA replication and DNA repair, operating via semi-conservative 5′ → 3′ enzymatic polymerization to transmit genomes across cellular generations. In modern biotechnology and synthetic biology, artificial DNA synthesis encompasses both *in vitro* template-directed enzymatic amplification (such as the polymerase chain reaction and reverse transcription) and *de novo* template-independent synthesis (solid-phase phosphoramidite chemical synthesis and enzymatic TdT synthesis) used to manufacture custom oligonucleotides, synthetic genes, and entire designer chromosomes from digital sequence files.

3′-Hydroxyl Nucleophilic Attack and Pyrophosphate Cleavage

Whether executed inside a dividing cell or in an *in vitro* reaction vessel, template-directed enzymatic DNA synthesis proceeds via a conserved S_N2 phosphoryl transfer reaction. The reaction strictly requires a single-stranded template and an annealed primer strand presenting an unblocked 3′-hydroxyl (3′-OH) terminus.

During each catalytic cycle, the primer 3′-oxygen acts as a nucleophile, launching an inline attack on the innermost phosphorus atom (the α-phosphate) of a complementary incoming 2′-deoxynucleoside 5′-triphosphate (dNTP):

DNA_*n* + dNTP ⇌ DNA_*n*+1 + PP_i

While the isolated formation of the 3′–5′ phosphodiester bond is modestly exergonic (ΔG°′ ≈ −15 to −20 kJ/mol), the overall pathway is driven irreversibly forward in biological systems by the subsequent hydrolysis of the released inorganic pyrophosphate leaving group (PP_i + H₂O → 2 P_i, ΔG°′ ≈ −19 to −33 kJ/mol) catalyzed by inorganic pyrophosphatase.

Two-Metal-Ion Mg²⁺ Coordination and 5′→3′ Directionality

Within the palm subdomain of all biological DNA polymerases, the pentacoordinate transition state is stabilized by two conserved divalent cations—physiologically magnesium (Mg²⁺_A and Mg²⁺_B)—coordinated by invariant aspartate and glutamate carboxylate side chains in the active site. Metal Ion A acts as a Lewis acid that lowers the p*K*_a of the primer 3′-OH from ~13 to ~7 to generate a reactive 3′-O⁻ alkoxide nucleophile, while Metal Ion B chelates the α-, β-, and γ-phosphates of the incoming dNTP to neutralize negative electrostatic charge and facilitate departure of the MgPP_i²⁻ complex. Because the activating triphosphate group is carried on the 5′ carbon of the incoming monomer and added exclusively to the 3′-OH primer terminus, all template-directed enzymatic DNA synthesis proceeds strictly in the 5′ → 3′ direction.

Semi-Discontinuous Replication at the Replication Fork

Within living cells, chromosomal DNA synthesis is executed by a multi-megadalton molecular machine known as the replisome. Because the parental double helix consists of two antiparallel strands while DNA polymerases synthesize solely in the 5′ → 3′ direction, the replication fork synthesizes the two daughter strands via an asymmetric, semi-discontinuous mechanism:

• Leading Strand Synthesis: The parental template strand oriented 3′ → 5′ toward the advancing replication fork is copied continuously. Following synthesis of a single initial RNA primer by primase at the origin, the leading-strand polymerase (DNA polymerase III in bacteria; Pol ε in eukaryotes) tracks in the 5′ → 3′ direction directly behind the unwinding replicative helicase.
• Lagging Strand Synthesis: The opposing parental template strand (oriented 5′ → 3′ toward the fork) must be copied discontinuously in the direction away from fork movement as a series of Okazaki fragments (1,000–2,000 nt in bacteria; 100–200 nt in eukaryotes). Primase repeatedly synthesizes new RNA primers as template ssDNA is exposed. Once each Okazaki fragment is extended by Pol III or Pol δ, DNA polymerase I (in prokaryotes) or RNase H2 and FEN1 (in eukaryotes) excise the RNA primers and fill the gaps with DNA, enabling DNA ligase to seal the nicked sugar-phosphate backbone.

Thermal Cycling in the Polymerase Chain Reaction (PCR)

The ability to direct enzymatic DNA synthesis in a cell-free buffer transformed molecular biology and clinical diagnostics. The cornerstone of *in vitro* template-directed synthesis is the polymerase chain reaction (PCR), conceived by Kary Mullis in 1983.

PCR replaces biological helicases and primases with physical thermal cycling and synthetic oligonucleotide primers. Each cycle alternates between (1) denaturation (~94–98 °C) to melt the target duplex into single strands, (2) annealing (~50–65 °C) to hybridize forward and reverse primers flanking the target locus, and (3) extension (~68–72 °C), during which a thermostable DNA polymerase isolated from hyperthermophilic extremophiles (such as *Taq* polymerase from *Thermus aquaticus* or high-fidelity proofreading *Pfu* polymerase from *Pyrococcus furiosus*) synthesizes new complementary strands, doubling the target amplicon copy number every cycle (2^*n* amplification). Isothermal enzymatic variants—such as Loop-Mediated Isothermal Amplification (LAMP) using *Bst* polymerase at 60–65 °C and Multiple Displacement Amplification (MDA) using bacteriophage Φ29 polymerase at 30 °C—couple strand-displacement synthesis with primer design to amplify DNA at a constant temperature without a thermal cycler.

RNA-Dependent DNA Synthesis (Reverse Transcription)

Furthermore, by employing retroviral reverse transcriptase (RT, an RNA-dependent DNA polymerase from Moloney murine leukemia virus or avian myeloblastosis virus), researchers synthesize complementary DNA (cDNA) directly from single-stranded RNA templates primed with oligo(dT) or random hexamers. Coupled with PCR (RT-PCR and RT-qPCR), reverse transcription enables quantitative profiling of cellular transcriptomes, cDNA cloning of intron-free eukaryotic coding sequences, and clinical detection of pathogenic RNA viruses.

3′→5′ Solid-Phase Synthesis on Controlled Pore Glass (CPG)

Unlike biological polymerases, which require a pre-existing template strand, synthetic biology relies on de novo chemical oligonucleotide synthesis to construct arbitrary user-designed sequences from scratch. Developed by Marvin Caruthers and Serge Beaucage in 1981 (building on earlier phosphodiester and phosphotriester chemistry by Har Gobind Khorana and Robert Letsinger), virtually all industrial oligonucleotide synthesis is executed via the automated solid-phase phosphoramidite method (on column synthesizers or high-density inkjet/electrochemical silicon microarrays).

In direct contrast to 5′→3′ enzymatic synthesis, chemical phosphoramidite synthesis proceeds in the 3′ → 5′ direction because the primary 5′-hydroxyl group is less sterically hindered and more nucleophilically reactive than the secondary 3′-hydroxyl group. The initial 3′-most nucleotide is covalently tethered via its 3′-OH to a solid support matrix—typically Controlled Pore Glass (CPG) beads or polystyrene—while reactive exocyclic amine groups on the nucleobases (A, C, and G) are permanently shielded by benzoyl or isobutyryl protecting groups to prevent branched side-reactions.

The Four-Step Iterative Cycle (Deprotection, Coupling, Capping, and Oxidation)

Addition of each successive nucleotide monomer occurs through a four-step anhydrous acetonitrile chemical cycle (~3–5 minutes per base):

1. Step 1 — Detritylation (Deprotection): The acid-labile 4,4′-dimethoxytrityl (DMT) protecting group blocking the 5′-carbon of the support-bound terminal nucleoside is cleaved using a mild organic acid (2–3% trichloroacetic acid or dichloroacetic acid in dichloromethane). This exposes a free, reactive 5′-hydroxyl (-OH) group and releases an orange-colored DMT⁺ cation whose absorbance at 498 nm is spectrophotometrically monitored to quantify per-step coupling efficiency.
2. Step 2 — Coupling: The next nucleoside monomer—prepared as a protected 5′-DMT-2′-deoxynucleoside-3′-O-(β-cyanoethyl-N,N-diisopropyl)phosphoramidite—is injected alongside an weak acidic azole activator (such as 1H-tetrazole or 5-ethylthio-1H-tetrazole). The tetrazole protonates the diisopropylamino group on the 3′-phosphoramidite phosphorus (P^III), converting it into a reactive leaving group that is immediately attacked by the exposed 5′-OH of the growing chain to form a dinucleoside phosphite triester (P^III) linkage (~99.0–99.5% yield).
3. Step 3 — Capping: Because ~0.5% to 1.0% of immobilized 5′-OH chains fail to react during coupling, the column is treated with acetic anhydride and *N*-methylimidazole (NMI) to permanently acetylate any unreacted 5′-OH termini. Capping prevents failed chains from extending in subsequent cycles, ensuring that incomplete deletion mutants (*n*−1, *n*−2) are truncated into short fragments easily removed during post-synthesis HPLC or PAGE purification.
4. Step 4 — Oxidation: Because the trivalent phosphite triester (P^III) linkage formed in Step 2 is chemically unstable to acid cleavage, it is treated with aqueous iodine (I₂) in water/pyridine/THF to oxidize the phosphorus atom to a stable pentavalent phosphotriester (P^V) intermediate (protected by a β-cyanoethyl group).

Once the full-length sequence is assembled, concentrated ammonium hydroxide (NH₄OH) at 55 °C cleaves the oligonucleotide from the CPG support, removes all β-cyanoethyl phosphate protections, and deprotects the exocyclic base amines to yield native single-stranded DNA. Because cumulative yield decays exponentially with length (*Y* = *c*^*N*−1; at 99.3% step yield, a 100-mer yields 0.993⁹⁹ ≈ 50%, whereas a 300-mer yields only ~12%), chemical synthesis is practically capped at 150 to 300 nucleotides. To build multi-kilobase genes, plasmids, or synthetic genomes (such as J. Craig Venter's 1.08-Mb *Mycoplasma mycoides* JCVI-syn1.0 and the synthetic yeast *Sc2.0* project), overlapping 60–200 nt oligos are stitched together hierarchically using Polymerase Cycling Assembly (PCA), Golden Gate Type IIS assembly, Gibson isothermal assembly, and yeast homologous recombination.

Template-Independent Terminal Deoxynucleotidyl Transferase (TdT) Cycling

While phosphoramidite chemistry remains the industrial foundation of gene synthesis, its reliance on hazardous organic solvents (acetonitrile, dichloromethane, trichloroacetic acid) and acid-induced depurination during repeated detritylation steps impose environmental and length bottlenecks.

To overcome these chemical limits, structural synthetic biology platforms have commercialized template-independent enzymatic DNA synthesis utilizing engineered variants of Terminal Deoxynucleotidyl Transferase (TdT)—a Family X DNA polymerase naturally expressed in immature vertebrate B and T lymphocytes (where it adds random non-templated "N-nucleotides" to V(D)J junctions to diversify antibodies and T-cell receptors). In engineered two-step aqueous synthesis cycles, TdT is supplied with dNTPs carrying a reversible 3′-ONH₂ or 3′-O-azidomethyl blocking group (or TdT is covalently tethered to a single dNTP via a cleavable linker). TdT appends exactly one blocked nucleotide onto the 3′-OH of a support-bound initiator oligo at physiological pH in water, after which mild aqueous cleavage removes the 3′-blocking group (or tether) to regenerate a native 3′-OH for the next base addition—enabling green, aqueous synthesis of >300–1,000 nt strands and high-density DNA digital data storage.

Why does chemical phosphoramidite DNA synthesis run in the 3′→5′ direction while biological enzymes synthesize DNA in the 5′→3′ direction?

Biological DNA polymerases evolved to synthesize in the 5′ → 3′ direction because the incoming dNTP carries its own high-energy 5′-triphosphate group, allowing a 3′→5′ proofreading exonuclease to clip off a misincorporated 3′ nucleotide and still leave a reactive 3′-OH ready for the next incoming dNTP. Conversely, chemical phosphoramidite synthesis operates without proofreading in an anhydrous organic solvent and runs 3′ → 5′ because coupling relies on nucleophilic attack by the growing chain's hydroxyl group: the primary 5′-hydroxyl (-CH₂OH) is far less sterically crowded and significantly more reactive than the secondary 3′-hydroxyl group on the deoxyribose ring, giving >99.3% coupling yields.

Why is it difficult to chemically synthesize a single oligonucleotide longer than 200 to 300 nucleotides?

Two mathematical and chemical constraints limit phosphoramidite oligo length: first, the exponential decay of step-wise yield (*Y* = *E*^*N*−1), where even a 99.5% coupling efficiency per cycle reduces the fraction of full-length product to 0.995²⁰⁰ ≈ 36.7% at 200 bases and 13.5% at 400 bases; second, during every cycle's acidic detritylation step (Step 1), the growing oligonucleotide is exposed to trichloroacetic acid, which accumulates trace acid-catalyzed depurination (cleavage of adenine and guanine bases from the sugar ring) that causes strand scission during final ammonium hydroxide deprotection.

How are entire multi-kilobase genes or megabase bacterial genomes constructed if chemical synthesizers only make 200-base oligonucleotides?

Synthetic biologists use a hierarchical bottom-up assembly pipeline: first, a target gene sequence is computationally partitioned into overlapping 60- to 150-nt sense and antisense oligonucleotides synthesized on microarrays or columns; second, these oligos are pooled and stitched into 500- to 2,000-bp double-stranded fragments via Polymerase Cycling Assembly (PCA) or Golden Gate assembly and error-corrected using mismatch-cleaving endonucleases (such as MutS or Surveyor/T7 Endonuclease I); third, verified kilobase fragments are joined into multi-kilobase plasmids or megabase chromosomes using Gibson isothermal assembly (5′ exonuclease + Phusion polymerase + *Taq* ligase) and *in vivo* homologous recombination in *Saccharomyces cerevisiae*.