8945d_003-012 6/18/03 11:04 AM Page 9 mac85 Mac 85:1st shift: 1268_tm:8945d:

Classic Experiment 4.3

PROVING THAT DNA REPLICATION IS

SEMICONSERVATIVE

he discovery that the structure of DNA is a double helix, containing two

T complementary strands of DNA, led to a number of hypotheses about how
DNA might be replicated. Although the possible replication mechanisms were rel-
atively easy to deduce, proving which occurs in vivo was a more difficult task. In
1958, Mathew Meselson and Franklin Stahl used the newly developed techniques
of density-gradient centrifugation, to show that DNA replication proceeds in a
semiconservative fashion.

exhibiting very small differences in density. The tools were

Background now available for a definitive test to determine whether
During the 1950s, scientists uncovered many of biological DNA replication occurs by a semiconservative or conserv-
facts we now take for granted, beginning with the discov- ative mechanism.
ery that genetic information is passed on through deoxyri-
bonucleic acid (DNA), and continuing through the eluci-
dation of DNA’s three-dimensional structure. As the The Experiment
decade neared a close, biologists were ready to study how
DNA passed on genetic information from the parental to Meselson and Stahl reasoned that if one could label the
the progeny generation. parental DNA in such a way that it could be distinguished
James Watson and Francis Crick had hypothesized, from the daughter DNA, the replication mechanisms could
based on their double-helical model of DNA, that replica- be distinguished. If DNA replication is semiconservative,
tion occurs in a semiconservative fashion. That is, the then after a single round of replication, all DNA molecules
double helix unwinds, the original parental DNA stands should be hybrids of parental and daughter DNA strands.
serve as templates to direct the synthesis of the progeny If replication is conservative, then after a single round of
strand, and each of the replicated DNA duplexes contains replication, half of the DNA molecules should be composed
one old (parental) strand, and one newly synthesized only of parental strands and half of daughter strands.
strand, often called the “daughter” strand. Another To differentiate parental DNA from daughter DNA,
hypothesis proposed at the time was conservative replica- Messelson and Stahl used “heavy” nitrogen (15N). This
tion, whereby after replication the parental strands isotope contains an extra neutron in its nucleus, giving it
formed one DNA duplex and the two daughter stands a higher atomic mass than the more abundant “light”
formed the second duplex. nitrogen (14N). Since nitrogen atoms make up part of the
When these hypotheses were first proposed, little purine and pyrimidine bases in DNA, it was easy to label
experimental evidence was available to support one over E. coli DNA with 15N by growing bacteria in a medium
another. In 1957, however, Messelson and Stahl, along containing 15N ammonium salts as the sole nitrogen
with Jerome Vinograd, developed density-gradient cen- source. After several generations of growth, the bacteria
trifugation, a technique that can separate macromolecules contained only 15N-labeled DNA. Now that the parental
8945d_003-012 6/18/03 11:04 AM Page 10 mac85 Mac 85:1st shift: 1268_tm:8945d:

DNA was labeled, Meselson and Stahl abruptly changed DNA produced during a second round of replication was
the medium to one containing 14N as the sole nitrogen analyzed, two distinct species were observed. One corre-
source. From this point on, all the DNA synthesized by the sponded to hybrid molecules; the other corresponded to
bacteria would incorporate 14N, rather than 15N, so that 14
N-labeled DNA. With each subsequent round of replica-
the daughter DNA strands would contain only 14N. As the tion the proportion of hybrid DNA decreased as the
bacteria continued to grow and replicate their DNA in the amount of 14N-labeled DNA increased. As the diagrams in
14
N-containing medium, samples were taken periodically the Figure show, the sedimentation patterns observed by
and the bacterial DNA was analyzed with the newly devel- Messelson and Stahl are consistent only with a semiconser-
oped technique of equilibrium density-gradient centrifuga- vative model of replication.
tion. In this type of analysis, a DNA sample is mixed with
a solution of cesium chloride (CsCl2). During long periods
of high-speed centrifugation the CsCl2 forms a gradient,
and the DNA migrates to the position where the density of Discussion
the DNA is equal to that of the CsCl2. If the DNA sample
contains molecules of different densities they will migrate For Meselson and Stahl to prove that DNA replication
to different positions in the gradient. Because 15N has a proceeds in a semiconservative manner, they not only
greater density than 14N, 15N-labeled DNA has a greater had to design a clear, easily interpretable experiment,
density than 14N-labeled DNA. The higher-density (15N) but also develop the technology to do it. The beauty of
DNA will sediment to a different position than the lower- this classic experiment is that each of the possible mod-
density (14N) DNA. Hybrid DNA molecules, containing els would produce distinctly different results, so that
both 15N and 14N, will sediment at an intermediate den- interpretation of the experimental data was unambigu-
sity, depending on the ratio of heavy nitrogen to light ous. This study remains a shining example of defining a
nitrogen. problem and employing the proper methodology to
The Figure illustrates the results obtained by Meselson solve it.
and Stahl. Before any DNA replication had occurred in the By demonstrating that DNA replication occurs in a
14
N-containing medium, all DNA sedimented as a single semi-conservative fashion, Meselson and Stahl opened up
species, corresponding to 15N-labeled DNA. As DNA repli- the field of DNA replication for in depth research. With
cation proceeded, the amount of (15N)-DNA decreased, and the correct model in hand, researchers could now turn to
a second DNA species, consisting of hybrid DNA molecules unraveling the precise mechanism of DNA replication. In
containing 15N- and 14N-labeled strands, appeared. DNA addition, equilibrium density-gradient centrifugation
collected after completion of the first round of replication became a widely used tool for the analysis of complex
was found to sediment with the second species. When the mixtures of DNA.
8945d_003-012 6/18/03 11:04 AM Page 11 mac85 Mac 85:1st shift: 1268_tm:8945d:

Predicted results Actual results

Conservative model Semiconservative model

Parent strands
synthesized
in 15N
Light Heavy
(14N) (15N)

One band One band:

H–H H–H

H–H
H H H H

First doubling
in 14N

Two bands: One band:

+ H–H + L–L + H–L (hybrid)

Old
strand
New H–L
strand
H H L L L H H L
New strands

Second doubling
in 14N

Two bands: Two bands:

H–H + L–L H–L + L–L

L–L H–L

H H L L L L L L H L L L H L L L

Experimental demonstration by Messelson and Stahl that DNA replication is semiconservative. After several generations of growth in a
medium containing “heavy” (15N) nitrogen, E. coli were transferred to a medium containing the normal “light” isotope (14N). Samples
were removed from the cultures periodically and analyzed by equilibrium density-gradient centrifugation in CsCl to separate heavy-heavy
(H-H), light-light (L-L), and heavy-light (H-L) duplexes into distinct bands. The actual banding patterns observed were consistent with the
semiconservative mechanism. [From H. Lodish et al., 1995, Molecular Cell Biology, 3rd ed. W. H. Freeman and Company. See M.
Messelson and W. F. Stahl, 1958, Proc. Nat’l. Acad. Sci. USA 44:671; photographs courtsey of M. Messelson.]