Science Wiki

DNA polymerase (EC 2.7.7.7) is a ubiquitous class of enzymes that catalyze the template-directed synthesis of DNA from deoxynucleoside triphosphates (dNTPs). Functioning as the central catalytic engines of DNA replication, DNA polymerases read an intact single-stranded DNA (or RNA) template in the 3′ → 5′ direction and covalently append complementary nucleotides onto a pre-existing primer 3′-hydroxyl terminus in the 5′ → 3′ direction according to Watson-Crick base pairing, enabling the genome to be duplicated prior to cell division. Because the term encompasses seven structurally and evolutionarily distinct protein families (Families A, B, C, D, X, Y, and RT), DNA polymerases represent a classic paradigm of convergent evolution toward a shared two-metal-ion phosphoryl transfer active site.

Beyond chromosomal duplication, specialized DNA polymerases operate across bacteria, archaea, eukaryotes, and viruses to execute the maturation of Okazaki fragments, gap filling in DNA repair (including base excision repair, nucleotide excision repair, and homologous recombination), damage-tolerant translesion synthesis, the extension of telomeres by the reverse transcriptase telomerase, and the somatic diversification of immunoglobulin genes in the vertebrate adaptive immune system.

The Right-Hand Fold (Palm, Fingers, and Thumb)

Most DNA polymerases have a shared overall shape that resembles a half-open right hand. The structure forms a cleft that binds the primer-template duplex and is divided into three subdomains. The best-studied polymerases, in families A, B, X, Y, and RT, all have this fold, although some polymerases (families C and D) differ in the structure of their catalytic core, as described below.

• Palm. The palm forms the base of the cleft and contains the catalytic carboxylate residues (usually two or three aspartates or glutamates) that bind the catalytic divalent metal ions. It also checks the geometry of the newly formed base pair by contacts in the minor groove.
• Fingers. The fingers contact the incoming dNTP and the templating base. In many polymerases, binding of a correct dNTP causes the fingers to rotate inward by tens of degrees. This converts an open complex, which has no catalytic activity, into a closed ternary complex that is competent for catalysis (enzyme·DNA·dNTP). In some polymerases, this change is a movement of an individual helix, and not a rigid movement of the whole subdomain.
• Thumb. The thumb wraps around the minor groove of the newly synthesized duplex. It positions the primer-template junction and contributes to processivity. After each nucleotide addition, the DNA moves by one base pair (about 3.4 Å) relative to the active site, which is called translocation.

Convergent Evolution of the Catalytic Core Folds

Comparison of the three-dimensional structures shows that the catalytic palm domain has more than one evolutionary origin, so the polymerases are a case of convergent evolution of a two-metal-ion active site. The principal folds are:

1. The RNA recognition motif (RRM)-like palm of families A, B, Y, and RT (and of many RNA-dependent RNA polymerases). It consists of a *β*-sheet flanked by *α*-helices. The Klenow fragment of *Escherichia coli* DNA polymerase I is the classical example.
2. The nucleotidyltransferase (Pol *β*-like) palm of family X (including eukaryotic Pol *β*) and family C (the main bacterial replicative polymerase, the *α* subunit of Pol III). The order of strands in the *β*-sheet differs from that of the RRM-like palm. The replicative polymerases of bacteria (family C) and eukaryotes (family B) are therefore unrelated in structure, and they likely arose independently.
3. The double-psi *β*-barrel (DPBB) fold of family D (Pol D, the main replicative polymerase of most archaea). The catalytic core of family D resembles that of cellular multi-subunit RNA polymerases, and it does not have the right-hand architecture.

Processivity and Sliding Clamps (β-Clamp and PCNA)

Processivity is the average number of nucleotides that an enzyme adds in one binding event. The catalytic core of a polymerase on its own is only moderately processive. Many add tens to a few hundred nucleotides before they dissociate from the DNA. Replicative polymerases are made highly processive by sliding clamps, which are ring-shaped proteins that encircle the DNA and bind the polymerase through short clamp-binding motifs (PIP boxes in eukaryotes):

• In bacteria, the clamp is the homodimeric *β* clamp (encoded by *dnaN*). Each monomer has three domains, so the ring has six domains.
• In archaea and eukaryotes, the clamp is the homotrimeric proliferating cell nuclear antigen (PCNA). Each monomer has two domains, so the ring also has six domains.

The ring has a central channel that is wider (about 35 Å) than B-form DNA (about 20 Å). The clamp slides along the duplex, with water in the gap between it and the DNA. Clamps are closed rings and do not assemble spontaneously around DNA. A clamp loader, which is an AAA+ ATPase complex, opens the ring, recognizes the 3′ end of a primer-template junction, and uses ATP hydrolysis to close the ring around the duplex. The clamp loader is the *γ* complex in bacteria and replication factor C (RFC) in archaea and eukaryotes. With a clamp, replicative polymerases can add 10⁴ to more than 10⁵ nucleotides per binding event, and the speed of replication forks is about 500 to 1,000 nucleotides per second in *E. coli* and about 20 to 100 nucleotides per second in eukaryotes.

Reaction Thermodynamics and Pyrophosphate Hydrolysis

Polymerization is a template-directed phosphoryl transfer. The 3′-hydroxyl group at the end of the primer attacks the innermost (*α*) phosphorus of the incoming dNTP. A phosphodiester bond is formed, and pyrophosphate (PP_i) is released:

DNA_*n* + dNTP ⇌ DNA_*n*+1 + PP_i

The step is only modestly favorable. The standard free energy change is of the order of −15 to −20 kJ/mol, so the apparent equilibrium constant *K*′_eq = exp(−Δ*G*°′/*RT*) is on the order of 10³. For example, at 25 °C (*RT* ≈ 2.48 kJ/mol), a Δ*G*°′ of −17 kJ/mol gives *K*′_eq ≈ exp(6.9), or about 10³. The reverse reaction, called pyrophosphorolysis, can occur at a high PP_i concentration.

In cells, synthesis is made effectively irreversible by inorganic pyrophosphatase, which hydrolyzes PP_i to two molecules of orthophosphate (P_i):

PP_i + H₂O → 2 P_i (Δ*G*°′ ≈ −19 to −33 kJ/mol, depending on conditions and ions)

Because free energy changes add up for coupled reactions, the sum of the two steps is strongly negative, and a low concentration of PP_i pulls the polymerization to completion. The overall free energy of incorporation of a nucleotide also depends on base stacking and on the loss of conformational freedom of the template, and not only on the cleavage of the phosphoanhydride bond.

The Two-Metal-Ion Mechanism and Steric Gate Discrimination

Polymerases of all families use a common mechanism, proposed by Thomas A. Steitz and colleagues, which is called the two-metal-ion mechanism. Conserved carboxylates in the palm bind two divalent cations, which in cells are Mg²⁺ ions, separated by about 3.6 to 4 Å. Mn²⁺ can substitute in vitro with reduced fidelity.

• Metal A (the catalytic metal) coordinates the primer 3′-oxygen and the *α*-phosphate. As a Lewis acid, it lowers the p*K*_a of the 3′-hydroxyl group and promotes formation of the 3′-oxygen nucleophile, which attacks the *α*-phosphorus in line with the leaving group.
• Metal B (the nucleotide-binding metal) coordinates oxygens of the *α*-, *β*-, and *γ*-phosphates, together with basic residues of the fingers. It orients the triphosphate, neutralizes its charge, stabilizes the pentacoordinate transition state, and helps the pyrophosphate to leave.
• A third metal (C) has been seen in time-resolved crystallographic studies of human Pol *η* and Pol *β* and others. It appears transiently as the product forms and may help stabilize the products and lower the barrier for catalysis. It may not be a feature of every polymerase, and its role is under investigation.

The reaction is a concerted, S_N2-like in-line displacement and inverts the configuration at the *α*-phosphorus. To prevent the incorporation of ribonucleotides, which are present in cells at concentrations roughly 10 to 1,000 times higher than those of dNTPs (the ratio varies with the nucleotide and the cell type), polymerases use a "steric gate" residue that is positioned against the 2′-hydroxyl group of a ribose. It is a bulky residue such as glutamate (in many family A polymerases) or tyrosine or phenylalanine (in families B and Y and in reverse transcriptases). In spite of this, ribonucleotides are the most common non-canonical nucleotides in genomic DNA, and they are removed by ribonucleotide excision repair initiated by RNase H2.

All known DNA polymerases add nucleotides to a free 3′-hydroxyl group, so DNA is synthesized only in the 5′ to 3′ direction. Two constraints follow from this and shape the organization of the replication fork.

Energetics of 5′→3′ Synthesis and Proofreading Compatibility

The 5′ to 3′ direction of synthesis is explained in part by the energetics of proofreading. In 5′ to 3′ synthesis, the activating triphosphate is carried by the incoming nucleotide. If a wrong nucleotide is removed from the primer end by an exonuclease, the new end is a 3′-hydroxyl group that can continue elongation. In a hypothetical 3′ to 5′ synthesis, the triphosphate would be at the end of the growing chain. Removal of a wrong nucleotide would leave a 5′-monophosphate, which cannot be extended without re-activation, so proofreading would end chain growth. This is a hypothesis consistent with the biochemistry, and it is not the only possible reason. A few exceptions to the general rule exist in other enzymes, such as the 3′ to 5′ addition of a nucleotide to the tRNA His 5′ end by the enzyme Thg1, which is not a replicative DNA polymerase.

Because the two strands of the double helix are antiparallel, synthesis of both strands in the 5′ to 3′ direction means that the replication fork works in an asymmetric way. As a helicase (bacterial DnaB, or the eukaryotic CMG complex) unwinds the parental duplex, the leading strand is synthesized continuously in the direction of fork movement. The lagging strand is made discontinuously in the opposite direction as a series of Okazaki fragments (about 1,000 to 2,000 nucleotides in bacteria and about 100 to 200 nucleotides in eukaryotes). Flap endonucleases (FEN1 in eukaryotes) and, in bacteria, the 5′ to 3′ exonuclease of Pol I remove the RNA primers, and DNA ligase seals the nicks.

Primer Initiation Mechanisms (Primase, Protein Priming, and PrimPol)

Replicative DNA polymerases cannot start synthesis on a bare single-stranded template. They need a primer base-paired to the template, which provides a 3′-hydroxyl group. The main ways that primers are provided are:

• RNA primers made by primase. In most cellular replication, a primase (DnaG in bacteria, the primase subunits of the Pol *α*-primase complex in eukaryotes, and archaeal primases) synthesizes a short RNA primer of about 8 to 12 nucleotides.
• Protein priming. Some linear viral genomes are replicated from a primer that is a terminal protein. This occurs in bacteriophage *Φ*29 and PRD1, and in adenoviruses. A serine, threonine, or tyrosine hydroxyl group on the terminal protein attacks the first dNTP and forms a covalent protein-nucleotide bond.
• tRNA and target-site priming. Retroviruses, including HIV, use a host tRNA to prime reverse transcription. Non-LTR retrotransposons, such as LINE-1, use a nick in the target DNA to prime target-primed reverse transcription.
• PrimPol. The human enzyme PrimPol, a member of the archaeo-eukaryotic primase superfamily, has both primase and polymerase activity and can start DNA synthesis *de novo* (reported in 2013). It re-primes DNA synthesis downstream of lesions that stall the replication fork, in the nucleus and in mitochondria.

The Multi-Step Fidelity Cascade

Cells copy DNA with an error rate of about 10⁻⁹ to 10⁻¹⁰ per base pair per replication. The accuracy is not due to one mechanism but to a product of several steps:

If a polymerase makes a mismatch with a probability *ε*₁ at the first step, and each later step removes a fraction of those errors, then the overall error rate is approximately the product:

*ε* ≈ *ε*₁ × *ε*₂ × *ε*₃

For example, with *ε*₁ = 10⁻⁴, *ε*₂ = 10⁻², and *ε*₃ = 10⁻², the overall error rate is 10⁻⁸. The real steps are not entirely independent, and error rates differ with the sequence context and with the polymerase.

Mutations in the proofreading domains of the human replicative polymerases Pol *ε* (the gene *POLE*) and Pol *δ* (*POLD1*) increase the mutation rate greatly and are found in some tumors, such as certain endometrial and colorectal cancers. These "ultramutated" tumors have a high mutational burden, which has relevance for immunotherapy.

Translesion Synthesis (TLS) Across Damaged Templates

The active sites of the high-fidelity replicative polymerases (Pol III, Pol *δ*, and Pol *ε*) are tight and exclude bulky lesions such as ultraviolet-induced cyclobutane pyrimidine dimers (CPDs), (6-4) photoproducts, cisplatin adducts, and abasic sites. When one of these polymerases meets a lesion, the replicative polymerase stalls. The helicase may continue to unwind, which exposes single-stranded DNA coated with replication protein A (RPA). Stalled forks can collapse and produce double-strand breaks.

Cells have two main ways of coping: re-priming downstream of the lesion, which leaves a gap that is later filled by recombination or by template switching, and translesion synthesis (TLS). TLS is carried out mostly by family Y polymerases, together with Pol *ζ* (family B), which acts as an extender. In eukaryotes, the RAD6-RAD18 ubiquitin ligase monoubiquitinates PCNA at lysine 164, and the modified PCNA recruits TLS polymerases that have ubiquitin-binding domains and PIP boxes, which causes a polymerase switch.

Family Y polymerases have small thumb and fingers domains and an extra "little finger" (or polymerase-associated domain). Their active sites are larger and more open, and they have no proofreading exonuclease. They can insert nucleotides opposite distorted bases and bypass lesions that block replicative polymerases. On undamaged templates, their error rates are high, from about 10⁻¹ to 10⁻³ per base pair. Their access to DNA is therefore limited to sites of damage.

The Seven Structural Families (A, B, C, D, X, Y, and RT)

On the basis of sequence similarity and structure, DNA polymerases are grouped into families, which are denoted A, B, C, D, X, and Y, with the reverse transcriptases (RT) included as a seventh group.

Prokaryotic DNA Polymerases (Pol I through Pol V)

In *Escherichia coli*, five DNA polymerases are known:

• Pol I (family A, *polA*). It is a monomer of about 103 kDa with a 5′ to 3′ polymerase, a 3′ to 5′ proofreading exonuclease, and a 5′ to 3′ nuclease (flap endonuclease) domain. It removes RNA primers by nick translation and fills gaps in repair. The proteolytic Klenow fragment lacks the 5′ to 3′ nuclease domain.
• Pol II (family B, *polB*). It has proofreading activity, is induced in the SOS response, and takes part in replication restart and in the repair of some lesions.
• Pol III holoenzyme (family C). It is the main replicative enzyme, with about 10 to 20 copies per cell. It is an asymmetric assembly of about 900 kDa that copies both strands at about 750 to 1,000 nucleotides per second. Its parts are:
• Two or three catalytic cores, each containing *α* (DnaE, the polymerase), *ε* (DnaQ, the proofreading exonuclease), and *θ* (HolE, which stabilizes *ε*).
• Two *β* sliding clamps (each a dimer of DnaN).
• A clamp loader complex, which has the subunits *τ* (or a mix of *τ* and *γ*), *δ*, *δ′*, *χ*, and *ψ* (encoded by *dnaX*, *holA*, *holB*, *holC*, and *holD*). The *τ* subunits link the polymerase cores to each other and to the DnaB helicase, and *χ* binds the single-stranded DNA-binding protein (SSB).
• Pol IV (family Y, *dinB*) and Pol V (family Y, *umuDC*). They are error-prone translesion polymerases without exonuclease activity. They are induced in the SOS response, when RecA nucleoprotein filaments on single-stranded DNA promote autocleavage of the LexA repressor. Pol IV is present at about 250 copies per cell and rises about tenfold on induction. Pol V is active as UmuD′₂C (called Pol V Mut) after it assembles with RecA and ATP, and it bypasses UV photoproducts. Its activity is tightly regulated because of its mutagenicity.

Eukaryotic DNA Polymerases (Replicative and Specialized Isoforms)

Human cells have at least 15 to 16 DNA polymerases in the nucleus and in mitochondria. Nuclear chromosomal replication is divided among three family B enzymes:

• Pol *α*-primase. A complex of the catalytic subunit POLA1 (p180, without proofreading), the regulatory POLA2 (p70), and the primase PRIM1 (p49) and PRIM2 (p58, with an iron-sulfur cluster). The primase makes an RNA primer of about 8 to 10 nucleotides, and the complex hands it off to POLA1, which extends it by about 20 deoxynucleotides. The result is an RNA-DNA primer of about 30 nucleotides. RFC then displaces Pol *α* and loads PCNA. Pol *α* starts the leading strand once at each origin and primes every Okazaki fragment.
• Pol *ε*. A complex of POLE (the catalytic subunit, with polymerase and proofreading domains), POLE2, POLE3, and POLE4. It associates with the CMG helicase (Cdc45-MCM2-7-GINS) and in the current model it synthesizes most of the leading strand. A hoop-shaped P domain gives it processivity.
• Pol *δ*. A complex of POLD1 (the catalytic subunit, with proofreading), POLD2, POLD3, and POLD4. Pol *δ* acts with PCNA, and it is the main lagging-strand polymerase. It displaces the 5′ end of the downstream fragment to produce a flap that is cut by FEN1 and Dna2. It also fills gaps in nucleotide excision repair, mismatch repair, and recombination. Some analyses indicate that Pol *δ* also contributes to the leading strand in some conditions, and the division of labor is an area of active research.
• Pol *γ* (family A). It consists of a catalytic subunit (POLG, with polymerase, 3′ to 5′ exonuclease, and 5′-dRP lyase activities) and a dimeric accessory subunit (POLG2). It replicates and repairs mitochondrial DNA with the Twinkle helicase and mitochondrial SSB. Mutations in *POLG* cause mitochondrial diseases, such as Alpers syndrome, progressive external ophthalmoplegia, and ataxia-neuropathy syndromes.

Specialized polymerases in repair and TLS include:

• Pol *β* (family X). It is a 39 kDa monomer without proofreading that carries out short-patch base excision repair. Its 8 kDa N-terminal domain has 5′-deoxyribose phosphate (5′-dRP) lyase activity, in which a lysine forms a Schiff base intermediate and removes the sugar by *β*-elimination. The 31 kDa polymerase domain fills the one-nucleotide gap.
• Pol *λ*, Pol *μ*, and TdT (family X). They act in NHEJ. TdT is expressed in immature lymphocytes and acts without a template in V(D)J recombination, where it adds random nucleotides ("N nucleotides") at the ends of the coding joints and so increases the diversity of antigen receptors.
• Pol *η* (family Y, *POLH*). Its large active site can hold both thymines of a *cis-syn* thymine dimer and inserts two adenines opposite them, which is mostly error-free. Loss-of-function mutations cause the variant form of xeroderma pigmentosum (XP-V). Pol *η* also contributes to the A:T mutations of somatic hypermutation of immunoglobulin genes.
• Pol *ι*, Pol *κ*, Rev1, and Pol *ζ*. Pol *ι* reads template bases through Hoogsteen base pairing. Pol *κ* (family Y) can extend mismatched termini and bypass some bulky lesions. Rev1 (family Y) acts as a deoxycytidyl transferase, in which an arginine side chain, and not the template base, serves as the templating residue, and it is also a scaffold that binds other TLS polymerases. Pol *ζ* (family B, REV3L with REV7 and the accessory subunits POLD2 and POLD3) extends distorted primer ends produced by other TLS polymerases.
• Pol *θ* (family A, *POLQ*). It has an N-terminal helicase-like domain and a C-terminal low-fidelity polymerase domain. It mediates theta-mediated end joining (TMEJ, a type of microhomology-mediated end joining). Cells deficient in BRCA1 or BRCA2 depend on Pol *θ* for survival, which is why it is a target for cancer drug development, and inhibitors have been tested in clinical trials.

Non-Orthologous Replicative Polymerases Across Domains

DNA polymerases are among the best examples of how different structural folds converged on the same chemistry. The replicative polymerases of the three domains of life are not homologous in a simple way. Bacteria use family C, eukaryotes use family B (Pol *α*, *δ*, and *ε*), and archaea use family D (Euryarchaeota) or family B. The replication proteins, including helicases, primases, and clamp loaders, also show differences, and they are used to infer the nature of the replication apparatus of the last universal common ancestor and the origins of DNA viruses. DNA polymerases of viruses of all kinds have varied origins, and some were captured from their hosts.

The Viral Replication Hypothesis and LUCA

Phylogenetic analyses indicate that the family B polymerases and the RT family, and the nucleotidyltransferase fold, are very old. A hypothesis in which DNA replication evolved more than once, and was transferred to cells by viruses, has been proposed ("viral replication hypothesis"). It is controversial, and a clear answer is not yet available.

DNA polymerases with different properties underlie molecular biology, genomics, and molecular diagnostics.

Polymerase Chain Reaction (PCR) and Thermostable Enzymes

The polymerase chain reaction (PCR), developed by Kary Mullis in 1983, amplifies a target sequence exponentially by cycles of denaturation (about 94 to 98 °C), primer annealing (about 50 to 65 °C), and extension (about 68 to 72 °C). The number of copies after *n* cycles with an efficiency *E* (between 0 and 1) is:

*N*_*n* = *N*₀ (1 + *E*)^*n*

For perfect doubling (*E* = 1), 30 cycles give about 2³⁰ ≈ 10⁹ copies per starting molecule. An efficiency of 0.9 gives 1.9³⁰ ≈ 2.3 × 10⁸, so even a small loss of efficiency has a large effect.

Mesophilic enzymes such as the Klenow fragment of *E. coli* Pol I denature at each high-temperature step, so practical PCR depends on thermostable polymerases:

• *Taq* polymerase (Taq polymerase) is a family A polymerase from *Thermus aquaticus*. Its half-life is about 40 minutes at 95 °C, and its extension rate is about 60 nucleotides per second at 70 to 75 °C. It has a 5′ to 3′ exonuclease domain, which is used to cleave hydrolysis probes in TaqMan assays in quantitative PCR, and it has no 3′ to 5′ proofreading activity. The error rate is about 10⁻⁵ to 10⁻⁴ per base per duplication, and the enzyme adds an untemplated 3′ adenosine to products, which is used in TA cloning.
• Archaeal family B polymerases (*Pfu* from *Pyrococcus furiosus*, KOD from *Thermococcus kodakarensis*, and Vent) are more thermostable (half-lives of hours at 95 °C) and have proofreading activity, which gives blunt-ended products and a fidelity several-fold to about 50-fold higher than that of *Taq*. These enzymes have a pocket that binds uracil in the template, which causes them to stall at uracil resulting from cytosine deamination. This limits the use of templates that contain uracil, for example from damaged DNA.
• Fusion polymerases (such as Phusion and Q5) join a family B polymerase to a small DNA-binding domain such as Sso7d, from *Sulfolobus solfataricus*. They have high processivity, fast extension, and fidelity that is more than 50-fold higher than that of *Taq* by some measures, with values for different enzymes and different assays reported in the range of tens to hundreds of times.

Strand Displacement and Isothermal Amplification (Φ29 and Bst)

Some polymerases displace the downstream strand without a helicase. This permits amplification at one temperature.

• Bacteriophage *Φ*29 DNA polymerase (family B) has an insertion that forms a tunnel around the template and downstream DNA. It has high processivity (more than 70,000 nucleotides per binding event), proofreading activity, and strong strand displacement at 30 °C. It is used for multiple displacement amplification (MDA), for whole-genome amplification from single cells, for rolling circle amplification, and as a motor in some nanopore sequencing methods.
• Large fragment of *Bst* DNA polymerase (family A, from *Geobacillus stearothermophilus*) has strand displacement activity at 60 to 65 °C. It is used in loop-mediated isothermal amplification (LAMP), including point-of-care diagnostics.

DNA Sequencing Platforms (Sanger, Illumina SBS, and PacBio SMRT)

DNA sequencing technologies use engineered polymerases that can incorporate modified nucleotides:

1. Sanger sequencing. Early versions used the Klenow fragment and then phage T7 DNA polymerase. Wild-type polymerases discriminate against dideoxynucleotides (ddNTPs). A single substitution in the O helix of the fingers (F667Y in *Taq*, which corresponds to F762Y in Pol I and Y526 in T7 polymerase in a modified numbering) removes much of the discrimination. Sequenase is an engineered T7 polymerase with a deleted exonuclease, and the sequencing enzymes allow fluorescent terminators to be incorporated with similar efficiency.
2. Cyclic reversible termination (Illumina sequencing by synthesis). Engineered family B polymerases derived from the archaeon *Thermococcus* species 9°N incorporate nucleotides that have a cleavable fluorophore on the base and a reversible 3′-blocking group. Each cycle adds one nucleotide, and imaging is followed by removal of the fluorophore and blocking group to restore a 3′-hydroxyl group.
3. Single-molecule real-time (SMRT) sequencing. A single engineered *Φ*29 DNA polymerase is immobilized at the bottom of a zero-mode waveguide. The fluorophore is attached to the terminal phosphate of the nucleotide and not to the base, so it is released with the pyrophosphate when the nucleotide is incorporated. The product is native, unmodified DNA, which allows reads of tens of kilobases and the detection of base modifications from the kinetics of incorporation.

Other Molecular Biology Applications

Polymerases are also used for reverse transcription PCR, using reverse transcriptases or engineered thermostable polymerases with reverse transcriptase activity, for site-directed mutagenesis and cloning, for DNA labeling by nick translation, and in the construction of synthetic DNA. Terminal deoxynucleotidyl transferase is used to add tails to DNA ends and in enzymatic DNA synthesis methods. Directed evolution has created polymerases that accept xeno nucleic acids (XNAs) as substrates and templates, and that tolerate inhibitors in blood or soil samples.

Antiviral, Chemotherapeutic, and Genetic Syndromes

• Antivirals. Nucleoside analogs that terminate chains, such as aciclovir (acyclovir), target the DNA polymerases of herpesviruses after activation by a viral kinase, and nucleoside and non-nucleoside reverse transcriptase inhibitors are central in the treatment of HIV. Foscarnet mimics pyrophosphate and binds the polymerase directly. Polymerase inhibitors are also used for hepatitis B.
• Chemotherapy. Many nucleoside analogs, such as cytarabine, gemcitabine, and fludarabine, are incorporated into DNA by cellular polymerases and arrest DNA synthesis. Cisplatin adducts are bypassed by TLS polymerases, and TLS inhibition has been proposed to overcome resistance.
• Inherited disease. Mutations in *POLG* cause mitochondrial DNA depletion and deletions. Mutations in *POLH* cause XP-V. Mutations in the exonuclease domains of *POLE* and *POLD1* cause polymerase proofreading-associated polyposis and a predisposition to cancer.
• Antibiotics. Bacterial DNA polymerase III is an attractive target, and inhibitors of the *α* subunit have been studied, but few have entered clinical use.

Why can't DNA polymerases initiate synthesis de novo on a bare single-stranded template like RNA polymerases can?

High-fidelity DNA polymerases require a base-paired primer terminus with an accurately positioned 3′-hydroxyl group so that the Palm and Thumb subdomains can verify Watson-Crick minor-groove geometry before catalyzing bond formation. Initiating a dinucleotide *de novo* from two free monomers in solution is entropically disfavored and has a high error rate (~10⁻²). By delegating initiation to a specialized RNA primase (which synthesizes an 8–12 nt RNA primer that is later completely excised and replaced with proofread DNA), the cell eliminates initiation errors from the permanent genome.

Why do archaeal Family B PCR polymerases (*Pfu*, *KOD*, *Q5*) stall when amplifying bisulfite-treated or ancient DNA?

Hyperthermophilic archaea experience high rates of spontaneous cytosine deamination to uracil at 85–100 °C. To prevent C → T mutations, archaeal Family B polymerases evolved a specialized "uracil-binding pocket" in their N-terminal domain that scans the single-stranded template four bases ahead of the catalytic site and tightly binds any deoxyuridine residue, halting replication until repair enzymes excise the uracil. To amplify uracil-containing bisulfite-converted or ancient DNA, engineered point mutants (such as *Pfu* V93Q or Q5U) with an inactivated uracil-binding pocket must be used.

How does the antiviral drug acyclovir selectively inhibit herpesvirus DNA polymerase without poisoning human replicative polymerases?

Acyclovir is an acyclic guanosine analog prodrug that lacks a 2′ and 3′ carbon on its sugar ring. Uninfected human cells express cytosolic thymidine kinases that have negligible affinity for acyclovir, leaving the prodrug unphosphorylated and inactive. In cells infected by herpes simplex virus (HSV) or varicella-zoster virus, the viral thymidine kinase efficiently monophosphorylates acyclovir, which cellular kinases then convert to acyclovir triphosphate (ACV-TP); viral DNA polymerase incorporates ACV-TP with ~30-fold higher affinity than human polymerases, resulting in immediate obligate chain termination.

• DNA replication
• RNA polymerase
• Reverse transcriptase
• Polymerase chain reaction
• Translesion synthesis
• Replisome
• DNA repair