Written and maintained by CASRAI Editorial Board
Last updated
Short answer
Recombinase polymerase amplification (RPA) is an isothermal method for copying a specific DNA sequence. Instead of heating and cooling to separate DNA strands, it uses a recombinase protein to pair primers with matching sites in double-stranded DNA, then a strand-displacing polymerase extends them. The original 2006 report ran at 37 °C and needed no thermal denaturation of the template.
Definition and scope
RPA was described by Piepenburg and colleagues in PLoS Biology (2006) as a way to amplify DNA at a constant, low temperature without thermal denaturation of the template. It sits alongside other isothermal methods such as rolling circle amplification, which also copies nucleic acid without temperature cycling but uses a different mechanism. Detection methods that rely on isothermal pre-amplification, such as the Cas13-based approaches, are covered on their own page.
This page is a laboratory reference on the concept. It does not give a reaction recipe or a stepwise protocol, and it is not a guide to diagnostic use.
How RPA works in practice
According to Piepenburg et al. (2006), the reaction combines three proteins from bacteriophage T4: the recombinase uvsX, its loading factor uvsY, and the single-stranded DNA-binding protein gp32. The polymerase is the large fragment of Bacillus subtilis Pol I (Bsu), which displaces strands as it extends. The mixture also contains a crowding agent (Carbowax20M) and an energy-regeneration system of ATP, phosphocreatine and creatine kinase, plus dNTPs.
The mechanism, as the authors describe it: recombinase-primer filaments scan double-stranded DNA and exchange strands at sites that match the primer. The displaced strand is held open by gp32, the polymerase extends the primer, and repeating this on both strands gives exponential amplification. The authors note that performance depends on a balance between recombinase filament formation and disassembly.
What the reaction contains compared with PCR
| Element | RPA as described by Piepenburg et al. (2006) |
|---|---|
| Template denaturation | Not required; the recombinase opens duplex DNA at primer-matching sites |
| Temperature regime | Constant; standard conditions 37 °C |
| Strand-opening proteins | T4 uvsX recombinase, uvsY loading factor, gp32 single-stranded DNA-binding protein |
| Polymerase | Strand-displacing Bsu (large fragment of B. subtilis Pol I) |
| Other components | Crowding agent (Carbowax20M), ATP, phosphocreatine, creatine kinase, dNTPs |
The practical consequence reported by Boyle et al. (mBio, 2013) is that RPA avoids “complex thermocycling equipment”; their lateral-flow readout still needed a basic heat source. Isothermal does not mean enzyme-light: the protein set above is larger than a single polymerase. Labs that run PCR can compare this with their thermal cycler specifications: the cited RPA work calls only for a constant-temperature heat source.
Reported conditions and sensitivity
Published and vendor sources give different headline numbers because they describe different assays. The table lays out what each source actually states, so the figures are not mixed together.
| Source | Temperature | Time | Sensitivity or other statement |
|---|---|---|---|
| Piepenburg et al., PLoS Biology 2006 | 37 °C standard | 60 min standard; complex genomic targets in under 30 min | Fewer than ten genomic copies detected; two copies inconsistent |
| Boyle et al., mBio 2013 (HIV-1 proviral DNA) | 39 °C used; 25 to 42 °C range described | 1 min preincubation plus 20 min | Reliable detection of 3 proviral copies; lateral flow positive at 1 copy in 6 of 8 replicates |
| Abbott (TwistDx) product page, checked 2026-10-29 | Not stated | Results “typically” within 3 to 10 minutes | Labelled for research use only, not for diagnostic procedures |
Comparing the rows shows why a single “RPA takes 20 minutes at 37 to 42 °C” statement hides detail: the published sensitivity figures are for particular targets and primer sets, and the 2006 paper reported single-digit rather than guaranteed single-copy detection.
Primer and probe design constraints
Boyle et al. (2013) report that RPA primers are ideally 30 to 35 nucleotides long, probes 35 to 52 nucleotides, and amplicons 100 to 200 base pairs. For PCR, see the separate primer selection guide; those design rules are a different matter and are not stated here for RPA. The same study found tolerance to mismatches: amplification continued with up to 9 changes across the primer and probe binding sites, the polymerase-gene primers amplified 98.6% of 72 variants, and the LTR primers 93% of 15 viruses. Those authors also state that the assay was tested only in their own laboratory, that comparison with a gold-standard PCR assay remained to be done, and that a residual component inhibitory to RPA was found in some plasmid preparations but not in genomic DNA stocks. As with PCR, template quality is a variable the lab controls.
Where it matters for labs and supply decisions
The Abbott (TwistDx) product page describes two formats: TwistAmp lyophilised (freeze-dried) kits, described as suited to field use, and liquid kits that supply the core RPA components for users to combine into their own master mixes. It states that kits are for research use only and not for diagnostic procedures. Teams comparing RPA reagents with a conventional PCR master mix should note that these are different product architectures: a lyophilised pellet versus bulk liquid components. This page makes no statement about price, availability or which kit is better.
Common misunderstandings
- “Isothermal means no enzymes beyond a polymerase.” The described reaction uses a recombinase, a loading factor, a single-stranded binding protein and a strand-displacing polymerase.
- “RPA always detects a single copy.” The 2006 paper reported fewer than ten copies, with two copies inconsistent; the 2013 study reported reliable detection of 3 copies.
- “A research-use kit is a diagnostic.” The vendor page labels its kits for research use only.
- “Any sequence mismatch ruins the reaction.” The 2013 study showed amplification across up to 9 changes in primer and probe sites, though recovery was below 100% for the variant panels tested.
FAQ
What temperature does RPA run at?
Piepenburg et al. used 37 °C as the standard condition. Boyle et al. describe a broad range of 25 to 42 °C and ran their assay at 39 °C.
How long does an RPA reaction take?
Reported times differ by assay: 60 minutes standard and under 30 minutes for complex genomic targets in 2006, about 20 minutes in the 2013 HIV-1 assay, and “typically” 3 to 10 minutes on the vendor page.
How sensitive is RPA?
Fewer than ten copies in the original report, and reliable detection of 3 proviral copies in the 2013 study, with the limits described above.
Does RPA need a thermal cycler?
No. The method is isothermal and does not require thermal denaturation of the template; a heat source holding a constant temperature is enough in the cited work.
Sources
- Piepenburg O et al. PLoS Biology 4(7):e204, 2006. journals.plos.org (checked 2026-10-29).
- Boyle DS et al. Rapid Detection of HIV-1 Proviral DNA for Early Infant Diagnosis Using Recombinase Polymerase Amplification. mBio, 2013. PMC3622927 (checked 2026-10-29).
- Abbott Global Point of Care, TwistDx page. globalpointofcare.abbott (checked 2026-10-29).
