DNA polymerase is the enzyme that actually builds new DNA strands during PCR. It reads the original template strand and assembles a complementary copy one nucleotide at a time, turning a tiny sample of DNA into billions of identical copies over the course of 25 to 35 heating and cooling cycles. Without it, none of the other PCR components, the primers, the free nucleotides, the buffer, would do anything useful.
How Polymerase Builds New DNA
DNA polymerase works by grabbing individual nucleotides (the A, T, G, and C building blocks floating freely in the reaction mix) and snapping each one into place opposite its matching partner on the template strand. It can only add nucleotides in one direction, extending the chain from what chemists call the 3′ end. This is a hard constraint: the enzyme needs a free 3′-OH group, a specific chemical handle, to attach the next nucleotide to. It cannot start building a strand from scratch.
That limitation is exactly why PCR requires primers. Primers are short, pre-made pieces of single-stranded DNA designed to match the beginning of the target sequence. They bind to the template and provide that essential 3′-OH starting point. Once a primer is in place, the polymerase latches on and begins extending it, reading the template and adding complementary nucleotides one by one.
The enzyme also requires magnesium ions as a cofactor. Magnesium activates the polymerase and influences how DNA strands separate and re-bind during the reaction. Most PCR protocols use a magnesium chloride concentration between 1.5 and 3.0 mM, and getting this concentration right matters: too little magnesium and the polymerase barely works, too much and the reaction becomes sloppy.
What Happens During Each PCR Cycle
A single PCR cycle has three temperature steps, and the polymerase is active in the third one.
First, the reaction is heated to around 94–98°C to separate (denature) the double-stranded DNA into two single strands. Then the temperature drops to roughly 50–65°C so the primers can bind (anneal) to their matching sequences on those single strands. Finally, the temperature rises to 72–80°C for the extension step. This is where the polymerase does its job, moving along each primed template and synthesizing a new complementary strand.
The standard polymerase used in PCR, Taq polymerase (originally isolated from a heat-loving bacterium), adds nucleotides at a rate of about 10 to 45 per second at its optimal temperature, though some engineered versions reach 155 nucleotides per second. A typical extension step lasts anywhere from 15 seconds to a few minutes depending on how long the target sequence is.
After one cycle, every original double-stranded DNA molecule has become two copies. After 30 cycles, a single molecule has theoretically become over a billion.
Why Taq Polymerase Survives the Heat
The denaturation step that pulls DNA apart at 94–98°C would destroy most enzymes. Taq polymerase tolerates it because it evolved in Thermus aquaticus, a bacterium that lives in hot springs. This heat stability was the breakthrough that made modern PCR practical. Before Taq, researchers had to manually add fresh enzyme after every denaturation step, making the process tedious and error-prone.
Accuracy and Error Rates
Every time a polymerase copies DNA, there is a small chance it inserts the wrong nucleotide. Taq polymerase makes roughly 1 error per 20,000 to 50,000 nucleotides copied. That sounds rare, but over 30 cycles of amplification, errors accumulate. Early-cycle mistakes get copied into every subsequent generation, so a single wrong base in cycle 3 shows up in a large fraction of the final product.
Taq is relatively error-prone because it lacks a built-in proofreading ability. Some polymerases have what’s called 3′-to-5′ exonuclease activity, essentially a backspace key that lets them detect a mismatched nucleotide, remove it, and try again. Taq doesn’t have this.
When accuracy matters, such as cloning a gene or detecting mutations, researchers use high-fidelity polymerases like Pfu, Pwo, or Phusion. These enzymes have proofreading capability and produce roughly 10 times fewer errors than Taq, with error rates around 2 to 3 per ten million nucleotides copied per cycle. The tradeoff is that proofreading polymerases tend to be slower, so extension times may need to be longer.
Processivity: How Far Before It Falls Off
Processivity describes how many nucleotides a polymerase adds in a single stretch before it detaches from the template. A highly processive enzyme stays locked on and copies long stretches efficiently. A low-processivity enzyme falls off frequently, requiring it to rebind before it can continue.
Taq polymerase on its own has moderate processivity. For most PCR targets under a few thousand bases, this is perfectly adequate because the enzyme has the full duration of the extension step to complete the job, re-binding as needed. For very long targets (over 5–10 kilobases), specialized long-range polymerase blends are available that combine high processivity with proofreading.
Hot-Start Polymerases Reduce Errors
One practical problem with standard Taq is that it has some activity even at room temperature. While you’re setting up the reaction and the tube is sitting on the bench, the polymerase can start extending primers that have bound nonspecifically, producing unwanted byproducts. These artifacts compete with the real target during amplification.
Hot-start polymerases solve this by keeping the enzyme inactive until the first high-temperature denaturation step. The most common approach uses an antibody that physically blocks the enzyme’s active site. At room temperature, the antibody stays bound and the polymerase can’t work. Once the reaction hits 94–95°C, the antibody denatures and falls off, releasing fully active polymerase. Other approaches use chemical modifications or physical barriers to achieve the same result. The payoff is cleaner, more specific amplification with fewer background bands.
Choosing the Right Polymerase
The choice of polymerase shapes the entire PCR experiment. For routine genotyping or diagnostic detection where speed matters more than perfection, Taq is the standard. It’s fast, robust, and inexpensive. For any application where the amplified DNA will be sequenced, cloned, or used in downstream experiments that depend on an accurate sequence, a high-fidelity enzyme like Pfu or Phusion is worth the added cost and slower extension time.
Some applications use blends that mix a fast, non-proofreading polymerase with a small amount of a proofreading one, balancing speed with improved accuracy. Others use engineered polymerases designed for difficult templates, such as GC-rich sequences that form stubborn secondary structures. In every case, the polymerase remains the core engine of the reaction: everything else in the tube exists to help it do its job.