The Nirenberg Experiment: Cracking the Genetic Code

Every protein in your body was built according to a molecular instruction set that uses a four-letter alphabet — the nucleotides A, U, G, C — to specify the sequence of twenty different amino acids. For decades after the double helix was described in 1953, how that translation worked was unknown. On the night of May 27, 1961, in a small laboratory at the National Institutes of Health, a 34-year-old biochemist and his German postdoctoral fellow ran a single experiment that produced a radioactive spike on a Geiger counter, breaking the genetic code open for the first time.

The Problem: Bridging DNA and Protein

By 1953, two pillars of molecular biology were in place. Oswald Avery's 1944 experiment had shown that genes are made of DNA. Watson and Crick's double helix model showed how DNA could be copied and how its sequence could carry information. But neither discovery explained the next step: how does a sequence of DNA bases become a sequence of amino acids in a protein?

The theoretical bridge was RNA. DNA in the nucleus is transcribed into messenger RNA, which moves to ribosomes in the cytoplasm where proteins are assembled. This much was known in outline by the late 1950s, largely from work by François Jacob and Jacques Monod on gene regulation in bacteria. But the code — the specific correspondence between RNA sequences and amino acids — was entirely unknown.

The problem was formidable. There are four RNA nucleotides: adenine (A), uracil (U), guanine (G), and cytosine (C). There are 20 amino acids. A one-letter code (four bases) can specify only four amino acids. A two-letter code (4² = 16 combinations) is still insufficient. A three-letter code (4³ = 64 combinations) gives more than enough: 64 possible triplets for 20 amino acids, with room for redundancy or stop signals.

Francis Crick had theorized in 1961 that the code was indeed triplet — each codon consisted of three nucleotides — based on genetic frameshift experiments. But theorizing that codons are triplets is not the same as knowing which triplet codes for which amino acid.

Source: Crick, F. H. C., Barnett, L., Brenner, S. & Watts-Tobin, R. J. (1961), "General nature of the genetic code for proteins," Nature, 192, 1227–1232 — the paper, published the same year as the Nirenberg experiment, establishing the triplet nature of the code from frameshift analysis.

The Theoretical Landscape Before Nirenberg

The genetic code problem attracted some of the best theoretical minds in biology and physics during the mid-1950s. George Gamow, a cosmologist who had helped develop Big Bang nucleosynthesis theory, proposed in 1954 that the code was a "diamond code" in which each amino acid was specified by a triplet whose identity was determined by the physical shape of a hole in the double helix. This idea was elegant and completely wrong. By 1956, combinatorial analysis had ruled out all direct template models in which the DNA or RNA physically templates the amino acid.

Crick proposed the "adaptor hypothesis" in 1955 — before it was published, in an internal MRC circulated note — arguing that small RNA adaptor molecules must mediate between codons and amino acids. He was right; the adaptors turned out to be transfer RNAs (tRNAs), discovered by Paul Zamecnik. But even knowing the machinery existed, nobody knew the correspondence.

Several groups had developed cell-free protein synthesis systems by 1960 — extracts from broken bacterial or wheat germ cells that could still synthesize proteins when provided with messenger RNA and the necessary molecular machinery. Nirenberg saw these systems as a potential tool for the code problem.

Source: Zamecnik, P. C. (1960), "Historical and current aspects of the problem of protein synthesis," Harvey Lectures, 54, 256–281 — review of cell-free translation systems, the essential technology that Nirenberg would exploit.

Marshall Nirenberg: The Newcomer Who Broke the Code

Marshall Nirenberg arrived at NIH in 1957 as a postdoctoral fellow at the age of 30, working in the laboratory of DeWitt Stetten. He had trained as a biochemist and had limited background in the cutting-edge molecular biology being developed at the MRC Laboratory of Molecular Biology in Cambridge and the Pasteur Institute in Paris.

This outsider status may have been an advantage. Nirenberg was not embedded in the social networks and theoretical preoccupations of the leading molecular biology groups. He approached the code problem from a biochemist's angle: build a cell-free translation system, add defined RNA sequences, and see what proteins come out.

His collaborator for the crucial experiment was J. Heinrich Matthaei, a German postdoctoral scientist who joined Nirenberg's group in early 1961. Matthaei was responsible for much of the technical optimization of the cell-free system.

Source: Nirenberg, M. & Matthaei, J. H. (1961), "The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides," Proceedings of the National Academy of Sciences, 47(10), 1588–1602 — the foundational paper.

Cell-free translation system — an extract prepared from broken bacterial cells (in this case, E. coli) that retains the ribosomes, transfer RNAs, enzymes, and energy molecules required to synthesize proteins. When messenger RNA is added, the system translates it into protein.

The Poly-U Experiment: What Happened at 3 a.m.

By the spring of 1961, Matthaei had optimized the cell-free E. coli extract. The system could translate natural bacterial mRNAs into detectable proteins. The question was whether it would also translate synthetic, artificial RNA sequences.

The synthetic RNA available was poly-U — polyuridylic acid, a chain composed entirely of the nucleotide uracil, repeating indefinitely. Poly-U had been synthesized by Severo Ochoa's laboratory using the enzyme polynucleotide phosphorylase, which could link nucleotides together in vitro. Poly-U was a simple but artificial molecule: not a naturally occurring mRNA, just an unbroken string of U's.

The experiment Matthaei ran on the night of May 27, 1961 was systematic. He prepared 20 separate test tubes, each containing the cell-free extract plus all 20 amino acids, plus the necessary energy molecules. In each tube, a different one of the 20 amino acids was radioactively labeled with carbon-14 (a radioactive isotope), making it detectable if it became incorporated into a new protein chain.

To each tube, he added poly-U.

If poly-U were translated, it would produce a protein. If that protein incorporated a radioactively labeled amino acid, the tube with that amino acid labeled would show elevated radioactivity. All other tubes would show only background levels.

After one hour, he measured radioactivity in each tube. The result was unambiguous.

The control tube — no poly-U added — showed approximately 70 counts per milligram of protein.

The tube with labeled phenylalanine and poly-U added showed approximately 38,000 counts per milligram of protein.

The other 19 tubes showed background levels.

Poly-U had been translated, and it had produced polyphenylalanine — a chain of phenylalanine amino acids. A messenger RNA reading UUUUUUUUUUUU... produced a protein reading Phe-Phe-Phe-Phe... Since the code was a triplet, UUU was the codon for phenylalanine.

Source: Nirenberg, M. & Matthaei, J. H. (1961), "The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides," PNAS, 47(10), 1588–1602 — the original data, including the 38,000 vs. 70 counts comparison.

The signal-to-noise ratio was more than 500-to-one. There was no ambiguity.

The Moscow Announcement That Changed the Race

Nirenberg and Matthaei ran their experiment in May 1961. Nirenberg submitted a brief communication to the Biochemical and Biophysical Research Communications journal in August 1961. In late August, he traveled to the Fifth International Congress of Biochemistry in Moscow to present his data.

He was not a well-known figure. He was assigned a minor session. His abstract had attracted little attention before the meeting.

Matthew Meselson — who had conducted his own landmark DNA replication experiment in 1958 and was well connected in the molecular biology community — heard Nirenberg's talk and recognized what it meant immediately. Meselson arranged for Nirenberg to present again to a larger audience, one that included Francis Crick, Jacques Monod, and other leading molecular biologists.

The presentation was received as the breakthrough it was. Within days, the race to crack the remaining 63 codons was underway.

Source: Kay, L. E. (2000), Who Wrote the Book of Life? A History of the Genetic Code (Stanford University Press) — historical account of the Moscow announcement and its reception.

Crick wrote after the meeting: "Nirenberg's beautiful experiment with poly-U shook us all up. It was just the sort of direct biochemical approach to the problem that I had not foreseen and that cut right through the theoretical discussion."

The Race to Crack All 64 Codons

After the poly-U result, several competing approaches were used to determine the remaining codons. The race was primarily between Nirenberg's group at NIH and Severo Ochoa's group at New York University — both well-funded, technically sophisticated, and moving fast.

Random copolymers. Ochoa's group, which possessed expertise in polynucleotide synthesis, used polynucleotide phosphorylase to create RNA polymers of defined base composition but random sequence — for example, a polymer made of 75% U and 25% A. By calculating the statistical probabilities of different triplet combinations in such a polymer and correlating them with which amino acids were incorporated, they could infer codon compositions (which bases were in a codon) if not the precise order of the bases.

This approach assigned many codons to amino acids but could not determine base order within a triplet. For example, it could say "an amino acid is specified by a codon containing two U's and one A" but not whether that was UUA, UAU, or AUU.

Trinucleotide binding assay (Nirenberg-Leder technique). In 1964, Nirenberg and Philip Leder developed a more direct method. They showed that a specific trinucleotide (a three-nucleotide sequence, the minimum codon) could bind to a ribosome and recruit the transfer RNA carrying the corresponding amino acid — the ribosome would stall on the trinucleotide, and the correct aminoacyl-tRNA would bind. By testing all 64 possible trinucleotides against all 20 radiolabeled amino acids, they could directly read off which tRNA (and therefore which amino acid) was recruited by each codon.

Source: Nirenberg, M. & Leder, P. (1964), "RNA codewords and protein synthesis," Science, 145(3639), 1399–1407 — the trinucleotide binding assay, a key methodological advance enabling rapid systematic code assignment.

Gobind Khorana's chemical synthesis. H. Gobind Khorana at the University of Wisconsin developed methods to chemically synthesize RNA polymers with defined, repeating sequences rather than random ones. A polymer that read UCUCUCUCUC... alternating between U and C would produce only two possible codons in the reading frame: UCU and CUC. By testing what amino acids were incorporated, the codon assignments for both could be determined simultaneously. Khorana's approach provided unambiguous codon assignments for sequences that the statistical approach had left uncertain.

Source: Khorana, H. G. et al. (1966), "Polynucleotide synthesis and the genetic code," Cold Spring Harbor Symposia on Quantitative Biology, 31, 39–49 — comprehensive summary of Khorana's synthetic approach to codon assignments.

By 1966, the complete genetic code had been deciphered. All 64 codons were assigned: 61 coded for amino acids (with most amino acids specified by multiple redundant codons — the code is degenerate) and three (UAA, UAG, UGA) served as stop codons, signaling the end of a protein chain.

The Complete Genetic Code: Key Features

Feature	Details
Codon length	3 nucleotides (triplet)
Total codons	64 (4³)
Amino acid-coding codons	61
Stop codons	3 (UAA, UAG, UGA)
Amino acids encoded	20
Start codon	AUG (also encodes methionine)
Most amino acids	Specified by 2–6 different codons (degeneracy)
Universal?	Nearly: same code in bacteria, fungi, plants, and animals with minor exceptions
First codon cracked	UUU = phenylalanine (Nirenberg-Matthaei, 1961)
Code fully cracked	1966, collaborative effort (Nirenberg, Ochoa, Khorana)

The degeneracy of the code — multiple codons for the same amino acid — is not a redundancy in the sense of wasted capacity. It provides robustness against mutation: a change in the third nucleotide of a codon (which is most variable, since many amino acids differ only in their third codon position) often does not change the amino acid specified. This is called "wobble" at the third position and was described by Crick in 1966.

Source: Crick, F. H. C. (1966), "Codon-anticodon pairing: the wobble hypothesis," Journal of Molecular Biology, 19(2), 548–555 — description of the degeneracy pattern in the code.

Why the Code Is Nearly Universal — and What the Exceptions Tell Us

When the genetic code was first fully described, there was an open question: was it universal — the same in all organisms — or did different organisms use different codes? The answer, which emerged through the 1970s and 1980s as more genomes became available for study, was: nearly universal, with instructive exceptions.

The code is the same in bacteria, fungi, plants, and animals. The codon UUU specifies phenylalanine in E. coli, in yeast, in corn, in fish, and in humans. This universality is strong evidence that all life on Earth descended from a common ancestor that used this specific code, because there is nothing intrinsically necessary about the UUU-phenylalanine assignment — it is a frozen accident of early evolutionary history.

Exceptions exist in mitochondrial genomes, where the reduced number of genes in the mitochondrial chromosome has allowed some codon assignments to drift. Human mitochondria use UGA as a tryptophan codon rather than a stop codon, for example. Some ciliates (single-celled organisms) use UAA and UAG as glutamine codons rather than stop codons.

Source: Barrell, B. G., Bankier, A. T. & Drouin, J. (1979), "A different genetic code in human mitochondria," Nature, 282, 189–194 — the first demonstration of a genuine exception to the universal genetic code.

These exceptions are informative precisely because they are rare. They represent evolutionary divergence from the standard code after the lineage separated from other life — the deviation requires dozens of simultaneous changes to tRNA molecules and synthetases, which explains why the code is nearly but not absolutely universal.

The Nobel Prize and the Question of Credit

In 1968, the Nobel Prize in Physiology or Medicine was awarded to Marshall Nirenberg, H. Gobind Khorana, and Robert Holley. Holley had determined the structure of alanine tRNA (the first tRNA to be fully sequenced), which was the physical structure of the adaptor molecule Crick had predicted. Nirenberg and Khorana were recognized for cracking the code itself.

Severo Ochoa did not receive the 1968 Nobel for this work (he had received a Nobel in 1959 for work on polynucleotide synthesis). The omission of Ochoa from the 1968 prize was noted, given his group's substantial contributions to the systematic codon assignments. Nobel Prizes are limited to three recipients per year, and the committee's choices inevitably involve judgment calls about relative contribution.

Heinrich Matthaei, who ran the poly-U experiment in the early hours of May 27, 1961, did not share the prize. He had returned to Germany by the time the prize was awarded, and the Nobel was given to Nirenberg as the senior researcher. The question of Matthaei's contribution relative to Nirenberg's — who conceived the strategy, and who executed the crucial experiment — has been a minor historical controversy.

Source: Strasser, B. J. (1999), "Sickle cell anemia, a molecular disease," Science, 286, 1488–1490 — on the politics of Nobel attribution in molecular biology.

What the Genetic Code Meant for Biology

The cracking of the genetic code was not simply an important experiment; it completed a conceptual framework. The double helix showed how genetic information could be stored and copied. The central dogma — DNA to RNA to protein — described how information flowed. The genetic code specified the translation rules: the exact dictionary by which an RNA sequence is read into a protein sequence.

Together, these three elements constituted a complete mechanistic account of heredity at the molecular level. You could, in principle, read a gene's nucleotide sequence and predict the exact amino acid sequence of the protein it would produce. The gap between genotype and phenotype had been bridged — not fully, since gene expression, protein folding, and protein-protein interactions add layers of complexity, but in terms of the fundamental information transfer, it was bridged.

Source: Nirenberg, M. (1963), "The genetic code: II," Scientific American, 208(3), 80–94 — Nirenberg's own popular account of the code-breaking research, written when the work was in progress.

The practical consequences followed over subsequent decades. Recombinant DNA technology, developed in the 1970s, depended on the ability to design synthetic genes. Protein engineering requires knowing which codons specify which amino acids. The polymerase chain reaction (PCR), which amplifies specific DNA sequences, depends on primers designed using genetic code logic. Modern mRNA vaccines for COVID-19 use sequence-optimized mRNA — choosing among synonymous codons to improve translation efficiency — a technology that would be impossible without the complete code.

What People Get Wrong About the Genetic Code Experiment

"Nirenberg and Crick cracked the code together." Crick did essential theoretical work on the code — the adaptor hypothesis, the triplet argument, the wobble model — but he was not involved in the actual deciphering experiments. The code was broken by biochemists using cell-free translation systems, not by structural biologists. Crick was a prominent observer and interpreter of the breakthrough, not its executioner.

"The first experiment decoded phenylalanine because it was the easiest." Poly-U was chosen for technical reasons, not because phenylalanine was expected. Poly-U was available from Ochoa's laboratory and was easy to synthesize. It happened to decode a single codon for a single amino acid — if the code had been more complex for that sequence, the experiment might have been less clean. The clarity of the result was partly luck.

"One experiment cracked the whole code." The poly-U result broke open the problem, but cracking all 64 codons required five more years of work by multiple groups using complementary methods. The full code was not complete until 1966.

"The genetic code is a universal constant of nature." The code is highly conserved but not universal. Mitochondrial genomes and some protists use variant codes. The universality of the standard code reflects common ancestry, not physical necessity — in principle, life could have evolved different codes, and in a few lineages, it has.

FAQs

Q: What was the poly-U experiment?
A: Nirenberg and Matthaei added synthetic poly-uridylic acid (a chain of repeated U nucleotides) to a cell-free bacterial extract containing all 20 amino acids, with each amino acid labeled in a separate test tube. Only the tube with labeled phenylalanine showed high radioactivity, proving that a sequence of U's produces a protein made only of phenylalanine — and therefore that the codon UUU specifies phenylalanine.

Q: What is a codon?
A: A codon is a sequence of three nucleotides in messenger RNA that specifies one amino acid (or a start or stop signal) during protein synthesis. The genetic code is the complete dictionary mapping all 64 possible three-nucleotide combinations to their corresponding amino acids or signals.

Q: Why is the genetic code called "degenerate"?
A: Degeneracy in this context means that most amino acids are specified by more than one codon. For example, the amino acid leucine is encoded by six different codons (UUA, UUG, CUU, CUC, CUA, CUG). This redundancy means many mutations in the third position of a codon do not change the amino acid — providing some protection against the effects of random mutation.

Q: Who else contributed to cracking the genetic code?
A: Severo Ochoa's group at NYU used random RNA copolymers to assign codon compositions. H. Gobind Khorana at Wisconsin synthesized defined repeating RNA sequences to make unambiguous assignments. Robert Holley at Cornell determined the structure of the first tRNA. Nirenberg and Philip Leder developed the trinucleotide binding assay. The full code required contributions from all these groups.

Q: Why did Nirenberg win the Nobel Prize over Ochoa?
A: Nirenberg's group made the crucial first breakthrough (poly-U) and developed the trinucleotide binding assay that systematically assigned individual codons. Ochoa had already received a Nobel Prize in 1959. The 1968 Nobel was shared by Nirenberg, Khorana, and Holley. Nobel Prizes can only go to three recipients per award, which creates inevitable omissions in work that involves many contributors.

Q: Is the genetic code the same in all organisms?
A: Nearly. The standard genetic code is universal across bacteria, archaea, and eukaryotes. Exceptions occur in mitochondrial genomes (which have evolved slightly different codes in various lineages) and in some single-celled organisms. These deviations are rare and require extensive molecular machinery changes, which explains why the code remained stable for billions of years in most lineages.

Q: What did Nirenberg's experiment mean for understanding life?
A: It completed the molecular-level description of how genetic information becomes biological function. Combined with the double helix and the central dogma, the genetic code meant that, in principle, you could read a DNA sequence and know exactly what protein it would produce. This was the foundation for genetic engineering, synthetic biology, and genomics as practiced today.

The Takeaway

A 34-year-old biochemist working at 3 a.m. with a cell-free bacterial extract and a synthetic RNA polymer produced a 500-fold spike in radioactivity that unlocked the instruction manual for every protein ever made by life on Earth. The genetic code that Nirenberg, Matthaei, Khorana, and their colleagues fully deciphered by 1966 is written in every living cell, has been stable for billions of years, and is nearly identical across all known life — a testament both to the efficiency of the solution and the shared ancestry of all biology.

---

Read next:

• The Meselson-Stahl Experiment
• The Hershey-Chase Experiment
• The Avery-MacLeod-McCarty Experiment