Thermostable DNA polymerases are DNA polymerases that originate from thermophiles, usually bacterial or archaeal species, and are therefore thermostable. They are used for the polymerase chain reaction and related methods for the amplification and modification of DNA. Thermostable DNA polymerases of natural origin are found in thermophilic bacteria, archaea and their pathogens.

Properties Several DNA polymerases have been described with distinct properties that define their specific utilisation in a PCR, in real-time PCR or in an isothermal amplification. Being DNA polymerases, the thermostable DNA polymerases all have a 5'→3' polymerase activity, and either a 5'→3' or a 3'→5' exonuclease activity. + Properties of thermostable DNA polymerasesPolymeraseTaqTth (Tfl)Bst Klenow fragment (BF)Tli (Vent)Deep VentPfx (KOD)Pfu (Pwo)OrganismThermus aquaticusThermus thermophilusGeobacillus stearothermophilusThermococcus litoralisPyrococcus sp. GB-DThermococcus kodakarensisPyrococcus furiosusOriginbacterialbacterialbacterialarchaealarchaealarchaealarchaealMolecular weight 80kDaPromega: Properties of Thermostable DNA Polymerases (PDF; 208 kB). Accessed september 27, 2012. 94kDa 67kDa 90kDa 90kDa 90kDa 92kDaExtension Temperature 74 °C 74 °C 65 °C 74 °C 75 °C 75 °C 75 °C5→3 Exonuclease Activity Yes Yes No No No3→5 Exonuclease Activity No No No Yes Yes Yes YesReverse Transcriptase Activity Weak Yes Weak No N/APCR Ends 3-A 3-A 3-A 70% Blunt; 30% Single-base Blunt Blunt BluntFidelity (errors per base and doubling)8 × 10−61.5 × 10−43-5.6 × 10−51.5 × 10−5Hong Guo Fan, Zhai Feng, Huang Wei-Hua: EP0810288 (A2) — 1997-12-03: A new DNA polymerase with proof-reading 3'-5' exonuclease activity2.8 × 10−62.7 × 10−64.0 × 10−63.5 × 10−61.2 × 10−57.6 × 10−61.3 × 10−6H.H. Hogrefe, M. Borns, High fidelity PCR enzymes, in: C.W. Dieffenbach, G.S. Dveksler (Eds.), PCR Primer: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2003.5.1 × 10−62.8 × 10−6Synthesis rate (bases/sec.)21–4761Agilent: High-Fidelity PCR Enzymes: Properties and Error Rate Determinations.19123120European Patent 1752534A1: Hochgeschwindigkeits-PCR unter Verwendung von Hochgeschwindigkeits-DNA-Polymerase, 2005-05-12 / 2007-02-14 by Toyo Boseki (applicant) & Masaya Segawa et al. (inventors).106–1389.3–25Processivity (bases)10–42<20>3006.4–20

Speed & processivity The baseline synthesis rates (speed, productivity) of various polymerases have been compared. The synthesis rate of Taq polymerase is around 60 base pairs per second. Among the unmodified thermostable DNA polymerases, only the synthesis rate of KOD polymerase is above 100 base pairs per second (approx. 120 bp/s). Among the modified thermostable DNA polymerases, various mutations have been described that increase the synthesis rate.US Patent 2013034879A1: DNA Polymerases, 2012-08-02 / 2007-02-14, Fermentas UAB et Al (applicant), Remigijus Skirgaila et al. (inventors).US Patent 2009280539A1: DNA Polymerases and related methods, 2009-04-16 / 2009-11-12, Roche Molecular Systems Inc (applicant), Keith A. Bauer (inventor). KOD polymerase and some modified thermostable DNA polymerases (iProof/Phusion, Pfu Ultra, Velocity or Z-Taq) are used as a PCR variant with shorter amplification cycles (fast PCR, high-speed PCR) due to their high synthesis rate. Processivity describes the average number of base pairs before a polymerase falls off the DNA template. The processivity of the polymerase limits the maximum distance between the primer and the probe in some forms of real-time quantitative PCR (qPCR).

Fidelity The error rates of various polymerases (fidelity) have been described. The error rate of Taq polymerase is 8 × 10−6 errors per base, that of Advantage HF 6.1 × 10−6 errors per base, that of Platinum Taq High Fidelity 5.8 × 10−6 errors per base and doubling, that of TaqPlus 4 × 10−6 errors per base and doubling, that of KOD polymerase 3.5 × 10−6 errors per base and doubling, that of Tli polymerase and Herculase 2.8 × 10−6 errors per base and doubling, that of Deep Vent 2.8 × 10−6 errors per base and doubling, that of Pfu, Phusion DNA Polymerase (identical with iProof DNA Polymerase) and Herculase II Fusion 1.3 × 10−6 errors per base and doubling and that of Pfu Ultra and Pfu Ultra II 4.3 × 10−7 errors per base and doubling. A newer analysis found slightly different error rates: Deep Vent (exo-) polymerase (5.0 × 10−4 errors per base and doubling), Taq polymerase (1.5 × 10−4 errors per base and doubling), Kapa HiFi HotStart ReadyMix (1.6 × 10−5 errors per base and doubling), KOD (1.2 × 10−5 errors per base and doubling), PrimeSTAR GXL (8.4 × 10−6 errors per base and doubling), Pfu (5.1 × 10−6 errors per base and doubling), Deep Vent DNA polymerase (4.0 × 10−6) errors per base and doubling, Phusion (3.9 × 10−6 errors per base and doubling), and Q5 DNA polymerase (5.3 × 10−7 errors per base and doubling). Yet another found error rates of 3–5.6 × 10−6 for Taq, 7.6 × 10−6 for KOD, 2.8 × 10−6 for Pfu, 2.6 × 10−6 for Phusion, and 2.4 × 10−6 for Pwo. To reduce the number of mutations in the PCR product (e.g. for molecular cloning), more template DNA and less cycles can be used in the PCR.

Yield Bacterial thermostable DNA polymerases generally produce higher product concentrations than archaeal, but with more copy errors. This is partly due to different degrees of exonuclease activity: higher activity leads to stronger proofreading but lower rate of synthesis, which leads to lower yields when used alone.

In the bacterial thermostable DNA polymerases, a Klenow fragment (Klen-Taq) or a Stoffel fragment can be generated by deleting the exonuclease domain in the course of protein design, analogous to the DNA polymerase from E. coli, which results in a higher product concentration. Two amino acids required for the exonuclease function of Taq polymerase were identified by mutagenesis as arginines at positions 25 and 74 (R25 and R74). A histidine to glutamic acid mutation at position 147 (short: H147E) in KOD polymerase lowers the relatively high exonuclease activity of KOD.

Structure DNA polymerases are roughly shaped like a hand with a thumb, palm and fingers. The thumb is involved in binding and moving double-stranded DNA. The palm carries the polymerase active site, whereas the fingers bind substrates (template DNA and nucleoside triphosphates). The exonuclease activity is in a separate protein domain. Mg2+ is a cofactor.

The polymerase active site in the palm catalyses the prolongation of DNA, starting from a primer bound to a template DNA single strand:

deoxynucleoside triphosphate + DNAn pyrophosphate + DNAn+1.

Bacterial polymerases Several homologs of DNA polymerase I from thermostable bacteria are used for PCR. Sources for these enzymes include various Thermus species (T. aquaticus = Taq, T. thermophilus = Tth), Thermatoga (Tt. neapolitana = Tne, Tt. maritima = Tma), and Geobacillus stearothermophilus (Bst).

The DNA polymerase I homologs are A-type DNA polymerases. In addition to 5'→3' polymerase activity, the A-type polymerases have 5'→3' exonuclease activity and generate an adenosine overhang (sticky ends) at the 3' end of the newly generated strand. The Klenow fragment of Bst (BF) has a strand displacement activity which allows for use in isothermal amplification without the necessity of denaturation of the DNA in a thermocycler, and its 5'→3' exonuclease activity is deleted for higher yield.

Archaeal polymerases

B-family Archaea use a B-type DNA polymerase to copy their genome. The same enzyme is used for PCR. Sources for these enzymes include: Pyrococcus: P. furiosus = Pfu, P. woesei = Pwo, P. abyssi = Pab. Thermococcus: Vent/Tli from Tc. litoralis, KOD/Pfx from Tc. kodakarensis, Tc. aggregans = Tag, Tc. celer = Tce, Tc. gorgonarius = Tgo, TNA1 from Tc. onnurineus NA1, Tc. peptonophilus = Tpe, Tc. thioreducens = Tthi, Nanoarchaeum equitans = Neq. Unlike other archaeal B-family polymerases, Neq polymerase is not inhibited by uracil.

The B-type DNA polymerases produce blunt ends (the Tli polymerase produces an overhang in about 30% of the products). It displays 3'→5' exonuclease activity, unlike the bacterial 5'→3'. This activity plays a proofreading role. In archaeal polymerases, the error rate suffers when a Klenow fragment analogue is generated, as the correcting exonuclease activity is removed in the process.

Some B-family polymerases are known for their ability to handle DNA with modified bases. Specifically, this has been demonstrated for Phusion (commercial Pfu-Sso7d) and KOD.

Y-family The Y-family DNA polymerases are used in DNA repair, specifically translesion synthesis (see for an overview). They are found in all domains of life. Thermostable Y-family DNA polymerases such as Dpo4 from Sulfolobus solfataricus are able to bypass DNA lesions that would block regular A- and B-family DNA polymerases, making them suitable for amplifying ancient DNA.

Modified polymerases Mutagenesis Specific amino acid residues can be changed on a DNA polymerase to provide many features such as: disabling (or weakening) of the exonuclease domain, more frequent activation of ddNTP (useful in Sanger sequencing), hot starting.

A modified Pfu polymerase was also generated by protein design (Pfu Ultra).Holly H. Hogrefe, M. Borns: High fidelity PCR enzymes. In: C.W. Dieffenbach, G.p. Dveksler (Eds.): PCR Primer: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2003.

Chimera The individual protein domains (exo, finger-palm, thumb) from different thermostable DNA polymerases can be mixed and matched to produce functional enzymes with new properties.

Processivity enhancement by domain fusion The processivity of a DNA polymerase in PCR can be greatly enhanced by grafting it with a DNA-binding protein domain, creating a fusion protein. This is especially useful for DNA polymerases of Pyrococcus origin, which have high fidelity but low apparent speed due to low processivity. The thermostable DNA-binding protein SSo7d is fused with Pfu to produce an enzyme combining low error rate of archaeal and the high synthesis rate of bacterial thermostable DNA polymerases (Q5 polymerase). The product is also capable of long-range synthesis. Such fusion polymerases are sold under various tradenames (Phusion) under premium prices since 2014, but they can also be made and purified in a lab. (The commercial versions include additional mutations for performance.) A fusion protein of the PCNA homologue from Archaeoglobus fulgidus was also generated with archaeal thermostable DNA polymerases. Fusion proteins of thermostable DNA polymerases with the thermostable DNA-binding protein domain of a topoisomerase (type V, with helix-hairpin-helix motif, HhH) from Methanopyrus kandleri were generated (TopoTaq and PfuC2).

Fusion with a single-strand DNA binding domain has also been done.

Enzyme blends Enhanced results are also achieved with mixtures of thermostable DNA polymerases of both types with a mixing ratio of the enzyme activities of type A and B polymerases of 30 to 1, e.g. Herculase and TaqPlus as a commercial mixture of Taq and Pfu polymerase, Expand as a commercial mixture of Taq and Pwo,Bruno Frey, Bernhard Suppmann: Demonstration of the Expand TM PCR System's Greater Fidelity and Higher Yields with a lacI-based PCR Fidelity Assay. (2000). Expand High Fidelity as a commercial mixture of Taq and Tgo, Platinum Taq High Fidelity as a commercial mixture of Taq and Tli (Vent), and Advantage HF 2 as a commercial mixture of Titanium Taq and an unnamed proof-reading polymerase. These mixtures can be used for long-range PCR to synthesize products of up to 35kb length.

Mixtures do not need to followed the A+B formula. A mixture of KOD and KOD(exo-) is also suitable for long-range PCR.

PCR enhancers