If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

Main content

Mode of DNA replication: Meselson-Stahl experiment

A key historical experiment that demonstrated the semi-conservative mechanism of DNA replication.

Key points:

  • There were three models for how organisms might replicate their DNA: semi-conservative, conservative, and dispersive.
  • The semi-conservative model, in which each strand of DNA serves as a template to make a new, complementary strand, seemed most likely based on DNA's structure.
  • The models were tested by Meselson and Stahl, who labeled the DNA of bacteria across generations using isotopes of nitrogen.
  • From the patterns of DNA labeling they saw, Meselson and Stahl confirmed that DNA is replicated semi-conservatively.

Mode of DNA replication

Imagine yourself in 1953, after the double helix structure of DNA has just been discovered1. What burning questions might be on your mind, and on the minds of other scientists?
One big question concerned DNA replication. The structure of the DNA double helix provided a tantalizing hint about how copying might take place1,2. It seemed likely that the two complementary strands of the helix might separate during replication, each serving as a template for the construction of a new, matching strand.
But was this actually the case? Spoiler alert: The answer is yes! In this article, we'll look at a famous experiment, sometimes called "the most beautiful experiment in biology," that established the basic mechanism of DNA replication as semi-conservative—that is, as producing DNA molecules containing one new and one old strand3.

The three models for DNA replication

There were three basic models for DNA replication that had been proposed by the scientific community after the discovery of DNA's structure. These models are illustrated in the diagram below:
Schematic representation of models of DNA replication.
  1. Conservative. Replication produces one helix made entirely of old DNA and one helix made entirely of new DNA.
  2. Semi-conservative. Replication produces two helices that contain one old and one new DNA strand.
  3. Dispersive. Replication produces two helices in which the individual strands are patchworks of old and new DNA.
_Image modified from "Basics of DNA replication: Figure 1," by OpenStax College, Biology (CC BY 3.0)._
  • Semi-conservative replication. In this model, the two strands of DNA unwind from each other, and each acts as a template for synthesis of a new, complementary strand. This results in two DNA molecules with one original strand and one new strand.
  • Conservative replication. In this model, DNA replication results in one molecule that consists of both original DNA strands (identical to the original DNA molecule) and another molecule that consists of two new strands (with exactly the same sequences as the original molecule).
  • Dispersive replication. In the dispersive model, DNA replication results in two DNA molecules that are mixtures, or “hybrids,” of parental and daughter DNA. In this model, each individual strand is a patchwork of original and new DNA.
Most biologists at the time would likely have put their money on the semi-conservative model. This model made a lot of sense given the structure of the DNA double helix, in which the two DNA strands are perfectly, predictably complementary to one another (where one has a T, the other has an A; where one has a G, the other has a C; and so forth)2,4. This relationship made it easy to imagine each strand acting as a template for the synthesis of a new partner.
However, biology is also full of examples in which the “obvious” solution turns out not to be the correct one. (Protein as the genetic material, anyone?). So, it was key to experimentally determine which model was actually used by cells when they replicated their DNA.

Meselson and Stahl cracked the puzzle

Matt Meselson and Franklin Stahl originally met in the summer of 1954, the year after Watson and Crick published their paper on the structure of DNA. Although the two researchers had different research interests, they became intrigued by the question of DNA replication and decided to team up and take a crack at determining the replication mechanism5.

The Meselson-Stahl experiment

Meselson and Stahl conducted their famous experiments on DNA replication using E. coli bacteria as a model system.
They began by growing E. coli in medium, or nutrient broth, containing a "heavy" isotope of nitrogen, 15N. (An isotope is just a version of an element that differs from other versions by the number of neutrons in its nucleus.) When grown on medium containing heavy 15N, the bacteria took up the nitrogen and used it to synthesize new biological molecules, including DNA.
After many generations growing in the 15N medium, the nitrogenous bases of the bacteria's DNA were all labeled with heavy 15N. Then, the bacteria were switched to medium containing a "light" 14N isotope and allowed to grow for several generations. DNA made after the switch would have to be made up of 14N, as this would have been the only nitrogen available for DNA synthesis.
Meselson and Stahl knew how often E. coli cells divided, so they were able to collect small samples in each generation and extract and purify the DNA. They then measured the density of the DNA (and, indirectly, its 15N and 14N content) using density gradient centrifugation.
This method separates molecules such as DNA into bands by spinning them at high speeds in the presence of another molecule, such as cesium chloride, that forms a density gradient from the top to the bottom of the spinning tube. Density gradient centrifugation allows very small differences—like those between 15N- and 14N-labeled DNA—to be detected.
Diagram of a test tube containing CsCl, nitrogen-14-labeled DNA, and nitrogen-15-labeled DNA following high-speed centrifugation. The density of the medium in the test tube is greatest at the bottom and least at the top, thanks to the formation of the CsCl gradient. The nitrogen-14-labeled DNA forms a band relatively close to the top of the test tube, while the nitrogen-15-labeled DNA forms a band closer to the bottom of the test tube. The positions of the bands reflect their relative densities.
_Image modified from "Meselson-Stahl experiment diagram en," by Mariana Ruiz Villareal (public domain)._

Results of the experiment

When DNA from the first four generations of E. coli was analyzed, it produced the pattern of bands shown in the figure below:
The steps of the Meselson-Stahl experiment and the banding pattern produced in the experiment. In the diagram there is a fluid-filled flask labeled N-15, and an enlarged image showing a picture of DNA and E.coli and the picture has the title All DNA is initially N15 labeled. There are 2 arrows pointing from the flask. One arrow points towards another fluid filled flask labeled N14 and next to the arrow it says add N15 labeled cells to N14 medium and grow for four generations. The second arrow from the N15 labeled flask points to Column 0 in the data table. The data table shows the percentage of the N14 and N15 labeled DNA in each generation. On the left side of the data table there is a test tube with a positive symbol at the bottom of the test tube and a negative symbol at the top of the test tube and the word Density written in the test tube. Above the test tube is the label Cesium chloride. Next to the test tube are 5 columns. The columns are labeled 0, 1, 2, 3 and 4 generation at the top of the columns. At the bottom, the columns are labeled 0,20,40,60,80 Time in minutes. In generation zero at time 0 minutes 100% of the DNA is N15 labeled. In generation 1 at time 20 minutes 100% of the DNA is N15 labeled. In generation 2 at 40 minutes, 50% of the DNA is N14 labeled and 50% of the DNA is N15 labeled. In generation 3 at 60 minutes 75% of the DNA is N14 labeled and 25% of the DNA is N15 labeled. In generation 4, at 80 minutes 88% of the DNA is N14 labeled and 12% of the DNA is N15 labeled.
_Image modified from "Basics of DNA replication: Figure 2," by OpenStax College, Biology (CC BY 3.0). Original artwork from "Meselson-Stahl experiment diagram en," by Mariana Ruiz Villareal (public domain)._
What did this result tell Meselson and Stahl? Let's walk through the first few generations, which provide the key information.

Generation 0

DNA isolated from cells at the start of the experiment (“generation 0,” just before the switch to 14N medium) produced a single band after centrifugation. This result made sense because the DNA should have contained only heavy 15N at that time.

Generation 1

DNA isolated after one generation (one round of DNA replication) also produced a single band when centrifuged. However, this band was higher, intermediate in density between the heavy 15N DNA and the light 14N DNA.
The intermediate band told Meselson and Stahl that the DNA molecules made in the first round of replication was a hybrid of light and heavy DNA. This result fit with the dispersive and semi-conservative models, but not with the conservative model.
The conservative model would have predicted two distinct bands in this generation (a band for the heavy original molecule and a band for the light, newly made molecule).

Generation 2

Information from the second generation let Meselson and Stahl determine which of the remaining models (semi-conservative or dispersive) was actually correct.
When second-generation DNA was centrifuged, it produced two bands. One was in the same position as the intermediate band from the first generation, while the second was higher (appeared to be labeled only with 14N).
This result told Meselson and Stahl that the DNA was being replicated semi-conservatively. The pattern of two distinct bands—one at the position of a hybrid molecule and one at the position of a light molecule—is just what we'd expect for semi-conservative replication (as illustrated in the diagram below). In contrast, in dispersive replication, all the molecules should have bits of old and new DNA, making it impossible to get a "purely light" molecule.
Diagram of the Meselson-Stahl experiment. All DNA is initially nitrogen-15-labeled. A DNA sample is taken prior to adding the bacteria to nitrogen-14 medium, and when centrifuged in a CsCl gradient, the DNA forms a single band low in the tube (indicating DNA labeled entirely with nitrogen-15). This is labeled as "generation 0."
The bacteria are then added to nitrogen-14 medium and grown for four generations. At each generation (which takes about 20 minutes to grow), a DNA sample is taken and analyzed by centrifugation in a CsCl gradient.
  • Generation 0 (see above). 100% of DNA in nitrogen-15 band.
  • Generation 1. 100% of DNA in a band intermediate in position between nitrogen-14 and nitrogen-15 bands.
  • Generation 2. 50% of DNA in a band intermediate in position between nitrogen-14 and nitrogen-15 bands. 50% of DNA in nitrogen-14 band.
  • Generation 3. 25% of DNA in a band intermediate in position between nitrogen-14 and nitrogen-15 bands. 75% of DNA in nitrogen-14 band.
  • Generation 4. 12% of DNA in a band intermediate in position between nitrogen-14 and nitrogen-15 bands. 88% of DNA in nitrogen-14 band.
The right panel of the figure is a cartoon illustrating how these results can be explained by the semiconservative model. The starting double helix is fully labeled by nitrogen-15 (generation 0). Replication of this helix produces two helices that each contain one nitrogen-15 (old) and one nitrogen-14 (new) strand (generation 1). Replication of these two helices produces four helices, two which are also nitrogen-15/nitrogen-14 hybrids and two which are purely made of nitrogen-14 (generation 2). Replication of the generation 2 helices produces eight helices, two of which are nitrogen-15/nitrogen-14 hybids and six of which are purely made of nitrogen-14 (generation 3). Replication of the generation 3 helices produces sixteen helices, two of which are nitrogen-15/nitrogen-14 hybrids and fourteen of which are purely made of nitrogen-14 (generation 4).
_Image modified from "Basics of DNA replication: Figure 2," by OpenStax College, Biology (CC BY 3.0). Original artwork from "Meselson-Stahl experiment diagram en," by Mariana Ruiz Villareal (public domain)._

Generations 3 and 4

In the semi-conservative model, each hybrid DNA molecule from the second generation would be expected to give rise to a hybrid molecule and a light molecule in the third generation, while each light DNA molecule would only yield more light molecules.
Thus, over the third and fourth generations, we'd expect the hybrid band to become progressively fainter (because it would represent a smaller fraction of the total DNA) and the light band to become progressively stronger (because it would represent a larger fraction).
As we can see in the figure, Meselson and Stahl saw just this pattern in their results, confirming a semi-conservative replication model.

Conclusion

The experiment done by Meselson and Stahl demonstrated that DNA replicated semi-conservatively, meaning that each strand in a DNA molecule serves as a template for synthesis of a new, complementary strand.
Although Meselson and Stahl did their experiments in the bacterium E. coli, we know today that semi-conservative DNA replication is a universal mechanism shared by all organisms on planet Earth. Some of your cells are replicating their DNA semi-conservatively right now!

Want to join the conversation?