On this page
The genetic code table in your textbook looks like something that was always known. In 1959 nobody knew a single entry. Marshall Nirenberg filled in the first one with a cell extract, a radioactive amino acid and an artificial RNA made of only one base. This post explains how his experiments worked and walks through the probability reasoning his field used to read the code.
Who was Marshall Nirenberg?
His Nobel biography says he was born in New York City on 10 April 1927 and moved with his family to Orlando, Florida, in 1939. At the University of Florida he took a BSc in 1948 and an MSc in zoology in 1952, with a thesis on the ecology and taxonomy of caddis flies. He then turned to biochemistry and earned a PhD at the University of Michigan in 1957.
He did postdoctoral work at the National Institutes of Health (NIH) in Bethesda and in 1960 became a research biochemist there. In 1959 he began to study the steps that connect DNA, RNA and protein. His Nobel facts page records that he died on 15 January 2010.
What problem was he trying to solve?
In his Nobel lecture, Nirenberg explains that RNA was thought to be a template for protein, but direct biochemical evidence was lacking. He wanted a cell-free system: a soup of broken E. coli cells that could make protein in a test tube, where he could add one RNA at a time and see what came out.
The first obstacle was that the extract made protein without any added RNA, using the cell's own templates. The fix was to add DNAase and incubate the extract briefly. Protein synthesis stopped, and from then on it depended on whatever RNA template was added. A filter method for collecting radioactive protein cut the time per experiment about four-fold.
What did poly-U show?
Heinrich Matthaei joined Nirenberg in these studies. RNA from yeast, ribosomes and tobacco mosaic virus stimulated the incorporation of every amino acid tested. Poly-U, a synthetic RNA made only of uridylic acid, was different. It stimulated phenylalanine incorporation "rather specifically", and the product was polyphenylalanine. Their paper, The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides, appeared in PNAS in October 1961.
Nirenberg drew three conclusions in his lecture: RNA is a template for protein, U in poly-U corresponds to phenylalanine, and the RNA's shape matters, because double- or triple-stranded poly-U paired with poly-A did not work as a template. Later work showed poly-C gave proline and poly-A gave lysine, while poly-G gave no activity.
Worked example: what a mixed RNA can tell you
The next step used random RNAs with two bases, made by the enzyme polynucleotide phosphorylase. The base order is random, but the proportions are set. Here is the logic, with illustrative numbers.
Suppose an RNA contains U and G in a 5:1 ratio. The chance of U at any position is 5/6, and of G is 1/6.
UUU: (5/6)^3 = 125/216 = 0.579.
Each triplet with two U and one G (UUG, UGU, GUU): (5/6)^2 x (1/6) = 25/216 = 0.116 each.
Each triplet with one U and two G: (5/6) x (1/6)^2 = 5/216 = 0.023 each.
GGG: (1/6)^3 = 1/216 = 0.005.
Check: 125 + 3(25) + 3(5) + 1 = 216, so the probabilities add to 1.
If an amino acid is incorporated at about 0.116/0.579 = 20% of the phenylalanine level, its codon probably contains two U and one G. But the method cannot tell you the order: UUG, UGU and GUU are equally likely. Nirenberg's lecture says this first phase gave the base composition of codons, not their sequence.
Common mistakes:
Forgetting the exponent. A triplet needs three independent positions, so multiply three probabilities.
Counting one arrangement. There are three arrangements of two U and one G.
Reading order from composition. Mixed RNAs show which bases, not which sequence.
How did the triplet binding test settle the order?
Nirenberg and Philip Leder found a way to read order directly. In RNA codewords and protein synthesis (1964), they showed that a trinucleotide of known sequence on a ribosome makes the matching amino acid-carrying tRNA stick. Their abstract reports that pUpUpU, pApApA and pCpCpC, "but not dinucleotides", directed the binding of phenylalanine, lysine and proline tRNAs. The ribosome-bound tRNA was caught on a cellulose nitrate filter, and the rest washed through.
In his lecture, Nirenberg gives two results that every student can use. AAA bound lysine tRNA, while AA did not, "hence, 3 sequential bases in mRNA correspond to 1 amino acid in protein". And three triplets from a poly-(U,G) digest, GUU, UGU and UUG, turned out to code for valine, cysteine and leucine. The composition method could never have separated those three.
Making the triplets was the hard part. Nirenberg's group used enzymes. Har Gobind Khorana's group made all 64 trinucleotides by chemical synthesis, and the two approaches together assigned the codons. Our profile of Har Gobind Khorana covers that side, and our genetic code guide shows how to use the finished table.
The 1968 Nobel Prize in Physiology or Medicine went to Robert Holley, Khorana and Nirenberg "for their interpretation of the genetic code and its function in protein synthesis".
What can a student take from his way of working?
Remove the background first. Destroying the extract's own templates made a new signal visible.
Make experiments faster. A quicker assay meant more RNAs tested.
Know what your method can and cannot show. Mixed RNAs gave composition, so he invented a test for order.
Ask for help. His lecture thanks colleagues whose advice on nucleic acids he relied on throughout.
Practising codon problems
Probability and codon questions reward a clear set-up. The Fermi Problem Library has worked problems with a step-by-step solver that explains the reasoning rather than printing an answer, so you can check each multiplication.
Try it now: for an RNA with A and C in a 3:1 ratio, work out the probability of AAA and of each triplet with two A and one C.
