Polymerase Chain Reaction (PCR)

In the early 1980s, studying a specific stretch of DNA meant cloning it into a bacterial vector, growing up sufficient cells, and extracting the product through a laborious series of steps that took days to weeks and required substantial starting material. Detecting a pathogen's genetic sequence in a clinical sample, or identifying a single-nucleotide variant in a patient's genome, was a research exercise, not a routine clinical test. The sensitivity of available hybridization methods was limited by the amount of target DNA present in any given specimen.

Kary Mullis, a biochemist at Cetus Corporation in Emeryville, California, conceived the polymerase chain reaction one spring night while driving on Highway 128 toward his cabin in Mendocino County; he pulled over at mile marker 46.7 to check his arithmetic. His idea was that if primers flanking a target sequence were extended by a polymerase and the product was heated apart and primed again, the target would double with each cycle; 20 cycles give about a million copies, within a few hours. The method was first described in December 1985, in a Science paper from the Cetus group that used it to amplify beta-globin sequences 220,000-fold for prenatal diagnosis of sickle cell anemia in under a day. The initial protocols used a DNA polymerase that was destroyed by each heating step and had to be replenished manually each cycle. That limitation was solved in 1988 when Randall Saiki and colleagues at Cetus switched to a heat-stable polymerase from the bacterium Thermus aquaticus, which also improved the specificity, yield and length of the products.

Saiki was first author on both the 1985 and 1988 papers, and the heat-stable enzyme made it practical to run the whole cycle automatically in a machine. Once automation was practical, PCR spread quickly through molecular biology laboratories. By 1993 the Nobel committee noted its use in detecting HIV and faulty genes in hereditary diseases, and PCR was soon applied to tests such as direct detection of Mycobacterium tuberculosis in sputum and diagnosis of herpes simplex encephalitis from cerebrospinal fluid.

Mullis received half of the 1993 Nobel Prize in Chemistry, for his invention of the PCR method; the other half went to Michael Smith of the University of British Columbia for site-directed mutagenesis. The academy's announcement noted that simple equipment could now multiply a DNA segment millions of times in a few hours. Forensic applications followed in parallel with clinical ones: DNA profiling from trace samples became central to criminal investigation, and the announcement mentioned DNA recovered from fossil material of extinct animals.

The method became familiar to the general public during the COVID-19 pandemic, when RT-PCR for SARS-CoV-2 RNA was the reference test for diagnosis. Its origins were less smooth than its spread. By Mullis's own account in his Nobel lecture, his first PCR paper was rejected by Science, and the first researcher outside Cetus to use the method was Alan Wilson.