The Thermal Photocopier

Kary Mullis, Yellowstone Extremophiles, and the Polymerase Chain Reaction

Driving his silver Honda Civic north along California Highway 128 through the moonlit redwood groves of Mendocino County on an evening in April 1983, the Cetus Corporation chemist Kary Mullis turned over a persistent signal-to-noise bottleneck in clinical genetics. At Cetus's biotechnology laboratories in Emeryville, Mullis's group synthesized short single-stranded chains of DNA—oligonucleotides fifteen to twenty bases long—for colleagues trying to diagnose sickle-cell anemia from a drop of fetal blood by spotting a single point mutation inside the human beta-globin gene. On paper, a labeled synthetic probe should have bound cleanly to its matching target sequence. In the test tube, the experiment failed repeatedly because the human genome contains three billion base pairs: a single-copy gene represents less than one-millionth of the total DNA mass, causing the true signal from a few thousand target molecules to vanish beneath random background sticking on the rest of the chromosome.

Up to kilometer post forty-six near Cloverdale, every molecular biologist in the world treated that scarcity as a detector problem, trying to brew hotter radioactive phosphorus labels or wash nitrocellulose filters more carefully. Halfway up the mountain grade toward his cabin in Anderson Valley, Mullis realized he could invert the ratio of signal to background before touching a detector at all. Instead of searching for a single twenty-base needle inside a three-billion-base haystack, he could use two opposing synthetic DNA primers oriented $5' \to 3'$ toward each other on opposite strands to bracket the exact target segment and force a DNA polymerase enzyme to double the target segment in an iterative chain reaction.

Every single cycle of Mullis's Polymerase Chain Reaction (PCR) executed a three-step thermodynamic ratchet inside a tiny plastic microcentrifuge tube containing target DNA, a vast excess of the two flanking twenty-base primers, free deoxynucleotide triphosphates (dATP, dCTP, dGTP, dTTP), and magnesium ions:

1. 1. Thermal Denaturation ($94^\circ\text{C}\text{ to }96^\circ\text{C}$): Heating the aqueous solution near boiling breaks the weak hydrogen bonds holding the two complementary strands of the long genomic double helix together, melting it into two separate single strands without breaking the covalent sugar-phosphate backbone.
2. 2. Primer Annealing ($50^\circ\text{C}\text{ to }60^\circ\text{C}$): Cooling the tube rapidly allows the short synthetic primers—present in a millionfold molar excess—to find and hydrogen-bond to their complementary flanking target sites before the massive genomic strands can re-zip with each other.
3. 3. Polymerase Extension ($72^\circ\text{C}$): A DNA polymerase enzyme latches onto the $3'$ hydroxyl end of each annealed primer and walks down the template strand, snapping free nucleotide building blocks into place to synthesize two brand-new double-stranded copies.

In the first cycle, the newly synthesized strands run on past the opposite primer site like long-tailed comets. From the second and third cycles onward, however, strands copied from previous copies terminate precisely at the far primer boundary. With each subsequent temperature cycle $n$, the number of exact-length double-stranded target amplicons $N(n)$ generated from $N_0$ initial template molecules grows exponentially according to the amplification law (with per-cycle efficiency $E \approx 1$):

After twenty cycles ($2^{20}$), a single target segment is multiplied one millionfold; after thirty cycles ($2^{30}$), it is multiplied into 1.07 billion identical copies, while the unbracketed three billion base pairs of background DNA remain unamplified.

When Mullis and research associate Fred Faloona ran the first successful PCR experiments in late 1983 and Randall Saiki and Henry Erlich applied it to the human beta-globin gene in 1985, the mechanical execution was grueling. They used the Klenow fragment of Escherichia coli DNA polymerase I, an enzyme evolved for the $37^\circ\text{C}$ human gut. Every time the test tube was heated to $95^\circ\text{C}$ to melt the DNA strands apart, the heat permanently denatured and destroyed the E. coli enzyme. Lab technicians had to sit beside three water baths with a stopwatch for four hours, moving tubes by hand every few minutes and uncapping each tube after every single $95^\circ\text{C}$ boiling step to pipette in fresh micro-drops of Klenow enzyme.

To close the loop into a sealed, hands-free instrument, Mullis, Saiki, Susanne Stoffel, and David Gelfand turned to an extremophile bacterium discovered in September 1966 by Indiana University microbiologist Thomas Brock inside the $70^\circ\text{C}$ outflow channels of Mushroom Pool, a geothermal hot spring in the Lower Geyser Basin of Yellowstone National Park. Named Thermus aquaticus, the yellow-pigmented microbe thrived in near-boiling water because its replication enzyme, Taq DNA polymerase, was cross-linked by dense hydrophobic cores and salt bridges that survived $95^\circ\text{C}$ denaturation intact while synthesizing DNA at 150 bases per second at $72^\circ\text{C}$. By 1988, automated Taq thermal cyclers could take a single molecule of DNA from a crime-scene bloodstain or a forty-thousand-year-old bone shard and amplify it into more than a billion identical copies before lunch.