DNA Is the Genetic Material

In 1928 a mixture of living harmless bacteria and dead deadly ones killed mice, and the living bacteria that came out of those mice were the dead strain’s type.

Something in the dead cells had changed the living ones, and the change was inherited. It took about 25 years, three kinds of experiment and one model to name that something and to see how it could be copied. This lesson follows that evidence, keeping each result inside the limits its authors gave it. For nucleotide structure, see nucleotide metabolism.

The idea: a genetic material must store, copy and act

Three demands on any genetic material follow from what you met in chromosomes and meiosis. It must store a great deal of information that differs between individuals and species. It must be copied faithfully every time a cell divides, so both daughters get the same set. And it must act, by directing what the cell makes.

Chromosomes are made of protein and DNA. Protein looked like the better candidate: it is built from about 20 kinds of amino acid, while DNA has four kinds of base and was widely thought to be a simple repeating chain. The experiments below overturned that guess one step at a time.

Griffith to Avery: a heritable change you can purify

Pneumococcus (Streptococcus pneumoniae) comes in serological types, labelled II, III and so on, set by the polysaccharide capsule each strain makes. Smooth (S) strains have a capsule and can kill a mouse. Rough (R) variants lack it and are generally harmless.

Frederick Griffith, in a paper published in January 1928, injected mice with a small dose of living R bacteria derived from a type II strain together with heat-killed type III S bacteria. In one control series he showed the heated culture was sterile. Some mice died, and from them he recovered living encapsulated type III bacteria, which bred true. Neither ingredient alone gave this. Griffith did not identify the cause. He suggested that a specific protein structure in the dead cells let the living ones make the new capsule. What he left behind was a test: a trait you could switch on in recipient cells and score.

Think of it as a recipe card slipped into a kitchen. The R cells already had the kitchen; the dead cells supplied a card for a new capsule, and the change persisted in their descendants.

Later work at the Rockefeller Institute showed the change could be made in culture and with extracts free of cells. Oswald Avery, Colin MacLeod and Maclyn McCarty then purified the active fraction from type III cells and tested it on type II-derived R cells in culture. Their 1944 paper made the case in several independent ways:

1. Chemistry. The active material’s ratio of nitrogen to phosphorus, which ran from 1.58 to 1.75 across four preparations, and its absorption of ultraviolet light matched nucleic acid.
2. Enzymes. Crystalline trypsin and chymotrypsin did not abolish activity, and ribonuclease (RNase) left it intact. Enzyme preparations that broke down known DNA destroyed it.
3. Heredity. The change was specific to the donor’s type and passed on through successive generations of recipients.

The authors concluded that the active fraction was “principally, if not solely” highly polymerized DNA, within the limits of their methods. That hedge marks the objection they could not fully close: a protein present below the detection limit might be the real agent, with DNA as a carrier. The paper made DNA the leading candidate without ending the debate.

Hershey and Chase: labels that separate protein from DNA

Phage T2, a virus that infects Escherichia coli (see bacterial and phage genetics), is made of little more than a protein coat around DNA. After infection the cell makes new phage. So one part, or both, must carry the instructions in.

Alfred Hershey and Martha Chase used the chemistry of the two parts. Protein contains sulfur, in the amino acids cysteine and methionine, and DNA has none. DNA is rich in phosphorus, in its backbone, while protein has little. So they grew one batch of phage with radioactive sulfur-35, labelling protein, and another with phosphorus-32, labelling mostly DNA.

They let labelled phage attach to E. coli bacteria, sheared off the attached material by agitation in a blender, and centrifuged the cells into a pellet. In their summarized result about 75% of the sulfur came off into the fluid, but only about 15% of the phosphorus did. The stripped cells still produced phage. Progeny carried less than 1% of the parental sulfur and 30% or more of the parental phosphorus.

Keep the words “most” and “about” when you retell this. A quarter of the sulfur stayed with the cells, and Hershey and Chase wrote that whether sulfur-free material other than DNA entered had not been determined. They concluded that the coat probably had no function in growth inside the cell and assigned it a role in delivering the DNA. The result did not prove that no protein entered. It showed, in a second system with a different method, that what enters and is passed on travels with DNA.

The double helix and the evidence behind it

How can a molecule with four bases store and copy information? Three lines of evidence constrained the answer.

Base ratios. Around 1950 Erwin Chargaff’s laboratory at Columbia University measured base composition in DNA from many sources. Composition varied from species to species, but in each sample the amount of adenine was close to that of thymine, and guanine close to cytosine. These ratios hold for double-stranded DNA as a whole, not for each strand alone.

Diffraction. At King’s College London, Rosalind Franklin and her student Raymond Gosling, and separately Maurice Wilkins with Alexander Stokes and Herbert Wilson, recorded X-ray patterns from DNA fibres. Franklin and Gosling showed that fibres switch between A and B forms as water content changes. For the B form they measured a 3.4-angstrom spacing and a 34-angstrom repeat along the fibre, concluded the structure was probably helical with phosphate on the outside, and considered two chains probable. In 1952 Gosling, working under Franklin’s direction, recorded the B-form pattern later called Photograph 51. Wilkins showed it to James Watson in early 1953. Wilkins, Stokes and Wilson found similar patterns from DNA of several sources, including several biological sources, and noted that diffraction could not reveal the order of the bases.

Model building. At the Cavendish Laboratory in Cambridge, Watson and Francis Crick built a model that fitted these constraints. Two right-handed chains coil around one axis and run in opposite directions, which is what antiparallel means. Phosphates sit outside; the bases sit inside, joined across the chains by hydrogen bonds, adenine with thymine and guanine with cytosine. The model, Franklin and Gosling’s paper and the Wilkins paper appeared together in Nature in April 1953.

Two details are often garbled. In modern structures a G-C pair has three hydrogen bonds and an A-T pair two, but the helix is held together largely by the stacking of flat bases on one another (see water and weak bonds). And the 1953 model had 10 base pairs per turn, an assumption; later measurements of B-DNA in solution gave about 10.5.

Franklin died in 1958. The 1962 Nobel Prize in Physiology or Medicine went to Watson, Crick and Wilkins, and the prize is not awarded after death.

Worked problem: reading base composition and helix length

Assume perfect Watson-Crick pairing in any duplex, and percentages taken over all bases in the sample.

(a) A double-stranded DNA sample is 22% A. Pairing gives T = A = 22%. That leaves 100 - 44 = 56% for G plus C, and G = C, so each is 28%.

A = T, G = C, and A + T + G + C = 100% (double-stranded DNA only)

(b) A viral genome measures A 25%, T 33%, G 24%, C 18%. The bases total 100%. T is present and U is absent, so this is DNA. But A (25) does not equal T (33) and G (24) does not equal C (18), so the bases are not paired across two strands. The genome is single-stranded DNA.

(c) A 3,000 bp segment of B-form DNA, with a rise of 0.34 nm per base pair and about 10.5 bp per turn.

length = base pairs × rise = 3,000 × 0.34 nm = 1,020 nm, about 1.0 micrometre turns = base pairs ÷ bp per turn = 3,000 ÷ 10.5 = 285.7, about 286

With the 1953 model’s 10 bp per turn you would get 300 turns; 10.5 is the solution value. The rise is per base pair, so do not double it for two strands.

Pairing as a copying rule

Specific pairing means one strand fixes the sequence of the other. Watson and Crick saw the consequence. In a second Nature paper later in 1953, on the genetical implications of the structure, they proposed that the two chains separate and each serves as a template on which a new partner is built. A gene could be copied by the same rule that holds the helix together.

An analogy: each strand is a photographic negative of the other. Separate them, print from each, and you get two copies of the original pair.

That was a proposal, not a test. In 1958 Matthew Meselson and Franklin Stahl grew E. coli on heavy nitrogen, moved it to light nitrogen and separated DNA by density. After one generation all the DNA was of hybrid density; after two, hybrid and light DNA were present in equal amounts. That is what templating predicts; DNA replication takes it further.

Design problem

A purified extract from heat-killed S type III cells turns R cells (derived from type II) into type III. Activity remaining after treatment: none 100%, trypsin plus chymotrypsin 100%, RNase 100%, DNase 0%, heating to 65 C 90%. A critic says a protein below 0.1% of the preparation could be the agent. What do the data show, and what experiments would answer the critic?