Difficulty: Intermediate | Prerequisites: Part 1 of these notes (Watson and Crick's double helix, Chargaff's rules, complementary base pairing)
Once Watson and Crick solved the double helix in 1953, the obvious next question was: how does DNA copy itself? The complementary structure hinted at an answer, each strand could serve as a template for its partner, but that had to be demonstrated experimentally. This section covers how Kornberg discovered DNA polymerase and how Meselson and Stahl proved that replication is semiconservative. Together, these experiments confirmed the first half of the central dogma: DNA → DNA.
Kornberg isolated DNA polymerase I and showed it synthesises DNA in the 5' to 3' direction using dNTPs, Mg²⁺, and a template strand. Meselson and Stahl used density-gradient centrifugation with heavy nitrogen (¹⁵N) to demonstrate that replication is semiconservative: each daughter molecule keeps one original strand and synthesises one new one.
DNA polymerase I (Pol I)
The first DNA-synthesising enzyme isolated (by Kornberg, 1956). It requires a DNA template, all four deoxyribonucleoside triphosphates (dATP, dCTP, dGTP, dTTP), and Mg²⁺. Think of it as the molecular machine that reads one strand and builds the complementary strand, one nucleotide at a time.
Semiconservative replication
The mode of DNA replication in which each daughter double helix contains one parental (old) strand and one newly synthesised strand. In simple terms, the original double helix "unzips" and each half is used as a blueprint to build a new partner.
Conservative replication (disproved)
A hypothetical model in which the original double helix stays intact and an entirely new double helix is made. Both original strands end up in one daughter molecule, both new strands in the other.
Dispersive replication (disproved)
A hypothetical model in which parental strands are broken into fragments and interspersed with new DNA in both daughter molecules. After several generations, all DNA would appear predominantly light.
Density gradient centrifugation (CsCl gradient)
A technique in which DNA is spun at high speed in a caesium chloride solution. DNA migrates to the position in the tube where its density matches that of the surrounding salt. Heavier DNA (e.g. ¹⁵N-labelled) settles further toward the bottom.
Complementary strands
The two strands of a DNA double helix, whose sequences are determined by base pairing (A with T, G with C). Knowing one strand's sequence automatically defines the other.
Watson and Crick noted in their 1953 Nature paper that complementary base pairing suggested a copying mechanism: separate the two strands, and each one can template a new partner.
This was an idea, not evidence. It still had to be shown that enzymes exist to carry out this synthesis and that the mechanism works the way the structure implies.
Arthur Kornberg used cell-free extracts of bacteria and calf thymus DNA as a template.
He isolated DNA polymerase I and showed it requires:
All four dNTPs (dATP, dCTP, dGTP, dTTP) as energy-rich substrates
Mg²⁺ as a cofactor
A DNA template to direct the sequence of the new strand
To determine the direction of synthesis, he started with unlabelled nucleotides, added ³²P-cytosine, then used a 3' phosphodiesterase to show the radiolabel was incorporated at the 3' end.
Conclusion: DNA polymerase links the 3' OH of the growing strand to the 5' phosphate of the incoming nucleotide. Synthesis proceeds 5' → 3', while the template is read 3' → 5'.
When Kornberg used DNA templates with varying A/T and G/C ratios from different species, the enzymatically synthesised product always matched the base ratios of its template.
This confirmed that the enzyme copies the template faithfully rather than producing random sequences.
Before Meselson and Stahl's experiment, three models were plausible:
Conservative: the parental double helix stays together; a completely new double helix is made alongside it. After one round, you get one heavy (H-H) molecule and one light (L-L) molecule.
Semiconservative: each daughter molecule gets one old strand and one new strand. After one round, you get two hybrid (H-L) molecules.
Dispersive: parental DNA is fragmented and mixed with new DNA in both daughters. After several rounds, DNA converges toward light density.
Grew E. coli in medium containing ¹⁵N (heavy nitrogen) so all DNA became uniformly heavy.
Transferred bacteria to ¹⁴N (light nitrogen) medium and allowed replication to continue.
At each generation, extracted DNA and centrifuged it on a CsCl density gradient.
Results:
Generation 0: single heavy band (¹⁵N-¹⁵N)
Generation 1: single intermediate/hybrid band (¹⁵N-¹⁴N). This rules out conservative replication, which predicts one heavy and one light band.
Generation 2: two bands, one hybrid (¹⁵N-¹⁴N) and one light (¹⁴N-¹⁴N). This rules out dispersive replication, which predicts a gradual shift rather than distinct bands.
Subsequent generations: hybrid band persists but becomes a smaller fraction; light band grows.
The pattern H-H → H-L → H-L + L-L is exactly what semiconservative replication predicts.
Even single-stranded RNA or DNA viruses must pass through a double-stranded replicative form to copy themselves.
Example for an ssRNA virus: the viral genome (sense strand) templates a complementary strand (antisense), forming a dsRNA intermediate. The antisense strand then templates new copies of the sense strand.
Double-stranded DNA has a built-in advantage: each strand is already a backup copy of the other.
Direction of synthesis: 5' → 3' (template read 3' → 5')
Kornberg's requirements: dATP + dCTP + dGTP + dTTP + Mg²⁺ + DNA template
Pyrophosphate (PPi) is released with each nucleotide addition, then split by pyrophosphatase (drives the reaction forward)
Kornberg's discovery of DNA polymerase is the conceptual foundation for PCR (polymerase chain reaction), which uses a heat-stable DNA polymerase to amplify DNA in a test tube. Every COVID test, forensic DNA analysis, and cloning experiment depends on the same enzymatic logic Kornberg characterised in 1956. The semiconservative model also underpins how we think about mutation inheritance: because each daughter cell gets one original strand, mutations on one strand become fixed only after a second round of replication copies them into both strands.
Students sometimes say "DNA polymerase synthesises DNA in the 3' to 5' direction." It does not. Synthesis is always 5' to 3'. The template is read 3' to 5', which may be the source of confusion.
The Meselson-Stahl experiment did not directly observe strands separating. It measured the density of whole DNA molecules. The strand-level interpretation follows logically from the density shift pattern.
Conservative replication is not ruled out by generation 0 or by generation 2 alone. The critical result is generation 1: a single hybrid band, not two separate bands.
"Dispersive" does not mean the DNA is randomly shattered. The model proposed that strands break every ~10 bp and serve as primers for new synthesis, producing mosaic strands.
⚠️ Be able to draw or describe the predicted banding pattern for all three replication models at generations 0, 1, and 2.
⚠️ Know the direction of synthesis (5' → 3') and that the template is read 3' → 5'. This distinction comes up repeatedly.
⚠️ Understand why pyrophosphate release and subsequent cleavage by pyrophosphatase drives nucleotide incorporation forward thermodynamically.
⚠️ Kornberg's requirements (four dNTPs, Mg²⁺, template) are a common short-answer or fill-in-the-blank target.
True or False: DNA polymerase I synthesises DNA in the 3' to 5' direction.
Fill in the blank: In the Meselson-Stahl experiment, after one round of replication in ¹⁴N medium, the DNA formed a single ______ density band.
True or False: Conservative replication predicts two hybrid-density bands after one generation.
Fill in the blank: DNA polymerase requires all four ______, Mg²⁺, and a DNA template.
True or False: Single-stranded RNA viruses can replicate without ever forming a double-stranded intermediate.
Answers: 1. False (5' to 3'). 2. Intermediate/hybrid. 3. False (it predicts one heavy and one light band). 4. Deoxyribonucleoside triphosphates (dNTPs). 5. False (they pass through a double-stranded replicative form).
Q: What evidence from Kornberg's experiment demonstrated that DNA synthesis proceeds 5' to 3'?
A: Kornberg added ³²P-labelled cytosine to an otherwise unlabelled reaction. A 3' phosphodiesterase cleaved the radiolabel from the product, showing that the labelled nucleotide had been incorporated at the 3' end of the growing chain. This means new nucleotides are added to the 3' hydroxyl group, so the chain extends in the 5' to 3' direction.
Q: How does the Meselson-Stahl generation-1 result distinguish semiconservative from conservative replication?
A: After one round of replication in ¹⁴N medium, semiconservative replication predicts every molecule is hybrid (one ¹⁵N strand, one ¹⁴N strand), producing a single band at intermediate density. Conservative replication predicts the parental molecule stays fully heavy and the new molecule is fully light, producing two distinct bands. Meselson and Stahl observed a single intermediate band, ruling out conservative replication.
Q: Why is the dispersive model ruled out at generation 2 rather than generation 1?
A: At generation 1, both semiconservative and dispersive models predict intermediate-density DNA (for different reasons). At generation 2, semiconservative replication predicts two distinct bands (hybrid and light), while dispersive replication predicts a single band that has shifted further toward light density. The appearance of two discrete bands at generation 2 rules out dispersive replication.
This material follows directly from Part 1 (the double helix structure). It also sets the stage for Part 3 (transcription), because once you understand how DNA copies itself, the next question is how its information gets read out into RNA. In later coursework, Kornberg's DNA polymerase connects to the replication fork machinery (helicase, primase, ligase, leading/lagging strands), which builds on the 5' to 3' directionality established here.
DNA replication, semiconservative replication, Meselson and Stahl experiment, CsCl density gradient, Kornberg DNA polymerase I, conservative replication, dispersive replication, 5 prime to 3 prime synthesis, dNTPs, template strand, phosphodiester bond, pyrophosphate, replicative form, central dogma DNA to DNA, chapter 2 molecular biology