PCR Steps: What Happens to Your DNA at Each Temperature

The PCR Thermal Cycling Process

Polymerase chain reaction (PCR) is a way of making the DNA sequence you care about double itself, over and over. The trick is that DNA is thermodynamically stable at some temperatures and unstable at others, so a thermal cycler can do the work simply by moving a fixed reaction tube through three temperatures in a loop. Each of the three PCR steps — denaturation, annealing, extension — exploits a different piece of the chemistry, and understanding the chemistry is what separates a protocol that "sort of works" from one you can debug.

Overview: The Three PCR Steps

The rest of this article goes through each step the way a PI would think about it: what the molecules are doing, why the numbers matter, and how to know you have it right. To keep things concrete we'll follow one real target throughout — a 131 bp fragment of human GAPDH amplified with the widely used OriGene qPCR pair HP205798 (forward GTCTCCTCTGACTTCAACAGCG, reverse ACCACCCTGTTGCTGTAGCCAA, melting temperatures 61.6°C and 65.1°C by the SantaLucia nearest-neighbour model). We use it in our validation benchmarks, so every number below is reproducible.

Step 1: Denaturation (94–98°C)

Denaturation is where the double helix comes apart. Two forces hold the strands together: hydrogen bonds between paired bases (two for each A–T pair, three for each G–C pair) and base stacking between neighbouring bases on the same strand. Neither is a covalent bond — they are individually weak and are constantly breaking and re-forming even at room temperature. Nuclear magnetic resonance studies measure the lifetime of a closed A–T base pair at roughly 1–30 ms and a G–C pair at 10–50 ms at physiological temperature: the helix is always "breathing," with bases flicking open and closed.

The purpose of heating to 94–98°C is to tip that equilibrium so far toward the open state that the two strands separate completely. Because strand separation is a cooperative melting transition, it happens fast. In a stop-flow study of PCR-like conditions, a 100 bp duplex denatured with a half-life of about 100 ms when the temperature was just 2.4°C above its melting temperature — and the rate climbs steeply as you go hotter. At 94–98°C, denaturation of a short amplicon is complete in well under a second, which is why cycle denaturation of 15–30 s is generous.

Why GC content matters: G–C pairs have three hydrogen bonds to A–T's two, so GC-rich regions are harder to melt. Templates above about 65% GC may need the higher end of the range (98°C) or slightly longer initial denaturation. Long templates (>5 kb) also melt more slowly and benefit from a 2–5 min initial denaturation that you don't repeat in the cycles.

Step 2: Annealing (50–65°C)

Once the template strands are separate, the cycler drops to the annealing temperature (Ta) so the primers can bind. A primer is a short single strand complementary to a stretch of your target. At Ta, the primer and its template reach a binding equilibrium. Here is the mental model: at the primer's melting temperature (Tm), about half of the primer molecules are bound to their template at any instant. Because you want most primers bound, you anneal a few degrees below the lower of the two primer Tm values — typically Ta = Tm − 3 to 5°C — which shifts the equilibrium toward the bound state.

For our GAPDH pair, the forward primer has a SantaLucia Tm of 61.6°C and the reverse 65.1°C. Using the lower Tm, a sensible Ta sits around 56–59°C; the nearest-neighbour calculations in our validation data suggest 58.3°C for this pair. Anneal too high (above ~62°C) and the primers barely bind — you get no product. Anneal too low (say 50°C) and the primers will bind to partially complementary sites elsewhere in the genome — you get extra bands and primer dimers. That is why annealing is the single most specific step in the reaction.

Setting the annealing temperature:

• Calculate primer Tm with the SantaLucia nearest-neighbour model (most accurate) or the Wallace rule (Tm = 2(A+T) + 4(G+C))
• Set Ta 3–5°C below the lower primer Tm
• Keep forward and reverse Tm within 2–5°C of each other (our GAPDH pair is 3.5°C apart — comfortable)
• Run a temperature gradient (55–65°C in 2°C steps) once, empirically, to find your Ta

Annealing quality ultimately comes from the primers themselves. Primers with strong self-complementarity or cross-dimer potential will produce non-specific bands even at an ideal Ta — which is why the fix is often in the design, not the cycler. See our primer design rules.

Step 3: Extension (68–72°C)

At extension temperature the polymerase takes over. The enzyme binds the primer-template junction and adds dNTPs one at a time to the 3' end of the primer, reading the template strand. The speed matters less than people think for small products, but it sets your cycling times:

Taq's ~1,000 bp/min figure is the practical guidance (Thermo Scientific's manual quotes roughly this rate at 72°C; QIAGEN's datasheet lists 2–4 kb/min under optimal buffer conditions). Our 131 bp GAPDH amplicon therefore needs less than 10 seconds of true synthesis — a 30 s extension is already a 3–4-fold safety margin. The rule "1 min per kilobase" is for amplicons larger than 1 kb; using more than necessary is harmless, using too little starves long products.

Walking the Whole Cycle: Our GAPDH Amplicon

Putting it together for a 35-cycle reaction on the GAPDH pair:

95°C 3 min (initial denaturation) 95°C 20 s ┐ 58°C 30 s ├── 35 cycles 72°C 30 s ┘ 72°C 5 min (final extension) 4°C hold

Cycle 1: the 95°C melt separates the two strands of the genomic template. Cycle 1 product is a single copy of the forward or reverse primer extended across part of the target. From cycle 3 onward, the amplification is between the two primers and the amplicon doubles each cycle, growing as 2ⁿ. After 35 cycles that is roughly 3.4 × 10¹⁰ copies of the 131 bp region — more than enough for gel detection, sequencing, or downstream analysis. This is why extension time can stay flat at 30 s across all 35 cycles: the product length (131 bp) never changes, only the number of copies.

Initial Denaturation and Final Extension

Most protocols add two bookend steps outside the cycle:

• Initial denaturation (2–5 min at 94–98°C): Fully melts long, structured genomic templates and activates hot-start polymerases, which are blocked by a chemical or antibody modification until heated.
• Final extension (2–10 min at 72°C): Finishes every partially extended product so the pool is uniformly full-length — important before cloning or sequencing.

Common PCR Cycling Issues and Solutions

How Cycle Number Affects PCR Outcome

During the exponential phase (roughly cycles 1–25), the amplicon doubles every cycle and the yield grows cleanly. After cycle 30–35 you enter the plateau phase: primers and dNTPs run low, polymerase activity drops, and the product stops accumulating. Thirty to thirty-five cycles is the sweet spot for most reactions. Too few cycles gives faint bands; too many adds background and primer dimers. For very low-abundance targets, nested PCR or up to 40 cycles can help — but include no-template controls so you can tell real amplification from artefacts.

Ramping Rates and Their Impact

The ramping rate — how fast the block heats and cools between steps — is a hidden dial. Fast ramping shortens the run but can hurt specificity: as the block cools from 95°C toward Ta, primers can start binding non-specifically in the window before Ta is reached. Slow ramping (1–2°C/s) gives cleaner annealing for difficult templates at the cost of a longer run. Some protocols insert an explicit slow ramp between denaturation and annealing for exactly this reason. Most modern cyclers let you set the rate (typically 1–5°C/s), so it is worth testing once whether yours is a "fast block" or "slow block" cycler.