The "Meselson-Stahl Experiment"

The results of the first critical test of Watson and Crick's proposal that DNA replicates semiconservatively
were published in 1958 by M. S. Meselson and F. W. Stahl dent Their resúlts showed that the
chromosome (now known to contain a single Watson-Crick double helix of DNA) of the common colon
bacillus Escherichia colt replicated semiconservatively

Meselson and Stahl grew E coli cells for many generations in a medium in which the heavy isotope of
nitrogen, 15N, had been substituted for the normal light isotope, WN. The purine and pyrimidine bases
in DNA contain nitrogen; thus, the DNA of cells grown on medium containing N will have a greater
density (weight per unit volurne) than the DNA of cells grown on medium containing N. Since molecules
of differ- ent densities can be separated by a procedure called equilibrium density gradient
centrifugation, Mesel son and Stahl were ablę to distinguish between the three possible modes of DNA
replication by following the changes in the density of DNA of cells grown on 15N medium and then
transferred to 14N medium for various periods of time (so-called density transfer experiments)

The density of most DNAs is about the same as the density of concentrated solutions of heavy salts such
as cesium chloride (CsCI). For example, the density of 6 M CsCl is about 1.7 g/cm. Escherichia coli DNA
con- tsol taining 1"N has a density of 1.710 g/cm3. Substitution of the 5N for "N increases the density of
E. coli DNA to 1.724 cul ire g/cm3.

When a heavy salt solution such as 6 M CSCl is s the centrifuged at very high speeds (30,000-50,000 revo
the lutions per minute) for 48-72 hours, an equilibriurm density gradient is formed (Fig. 5.14). The
centrifugal force caused by spinning the solution at high speeds sediments the salt toward the bottom of
the tube Diffusio 1, on the other hand, results in movement of salt molecules back towaid ie (oP LIOW Sa
Coic tration) of the tube. After a sufficient period of high speed centrifugation, an equilibrium between
sedi i mentation and diffusion is reached, at which time a d linear gradient of increasing density exists
from the co top of the tube to the bottom of the tube (Fig. 5.14). If a DNA is present in such a gradient, it
will move to a P position where the density of the salt solution is equal n to its own density. Thus, if a
mixture of E. coli DNAse containing 1SN ("heavy" DNA) and E. coli DNA contain li ing "N C"light" DNA) is
subjected to CsCl equilibrium W density-gradient centrifugation, the DNA molecules will separate into
two "bands," one containing "heavy" i DNA and one containing "light" DNA (Fig. 5.14).

Meselson and Stahl took cells that had been grow ing in medium containing 15N for several generations
D (and thus contained "heavy" DNA), washed them to a remove the 15N-containing medium, and
transferred " them to medium containing 14N. After allowing the a cells to grow in the presence of 1 4N
for varying periods di of time, the DNA was extracted and analyzed in CsCl sh equilibrium-density
gradients. The results of their ex g periment (Fig. 5.15) are only consistent with semicon e servative
replication, excluding both conservative and M dispersive models of DNA synthesis. All the DNA iso-
lated from cells after one generation of growth in medium containing #N the densities of "heavy" DNA
and "light" DNA. This intermediate density is usually referred to as "hybrid density. After two generations
of growth in medium e contain.ng 4N, half of the DNA was of "hybrid" density f and half was "light."
These resu'ts are precisely those a predicted by the Watson and Crick semiconservative l mode of
replication (Fig. 5.15). One generation of A semiconservative replication of a parental double he- lix con
aining N in medium containing only MN n would produce two progeny double helices both of s which
had 15N in one strand (the "old" strand) and N in the other strand (the "new" strand). Such molecules
would be of "hybrid" density

Conservative replication would not produce any S DNA molecules with "hybrid" density; after one gener-
ation of conservative replication of "heavy" DNA in "light" medium, half of the DNA would still be
"heavy" and the other half would be "light." If replication were s dispersive, Meselson and Stahl would
have observed a I shift of the DNA from "heavy" toward "light" in each generation (i.e, "half heavy" or
"hybrid" after one gen- eration, "quarter heavy" after two generations, etc.). Meselson and Stahl's resuls
are clearly inconsistent with either of these possibilities.

Subsequent studies have verified Meselson and Stahl's conclusion that DNA replication is semiconser us
vative and have extended it to many other organisms, be including higher plants and animals.

Autoradiography of Replicating

Bacterial Chromosomes

The visualization of replicating chromosomes was first ce accomplished by J. Cairns in 1963 using the
technique (l called autoradiography. Autoradiography is a method ca for detecting and localizing
radioactive isotopes in fil cytological preparations or macromolecules by expo- w sure to a photographic
emulsion that is sensitive to el low-energy radiation. Autoradiography is particularly useful in studying
DNA metabolism because DNA can s, be specifically labeled by growing cells on Hthy midine, the
tritiated deoxyribonucleoside of thymine. Thymidine is incorporated exclusively into DNA; it is not
present in any other major component of the cell.

Cairns grew E. coli cells in medium containing [3H]thymidine for varying periods of time, lysed the st cells
very gently so as not to break the chromosomes e (long DNA molecules are very shear sensitive), and d
carefully collected the chromosomes on membrane in filters. These filters were affixed to glass slides,
coated with emulsion sensitive to B-particles (the low-energy to electrons emitted during decay of
tritium), and stored rly in the dark for a period of time to allow sufficlent radioactive decays. The
autoradiographs observed( when the films were developed (Fig. 5.16) showed that se the chromosomes
of E coll are circular structures that exist as 0-shaped intermediates during replication. These
autoradiographs further indicated that the unwinding of the two complementary parental strands (which
is necessary for their separation) and their at semiconservative replication occur simultaneously or are
closely coupled. Since the parental double helix must rotate 360° to unwind each gyre of the helix, this
necessitates the existence of some kind of “swivel" in the chromosome. Present evidence suggests that
a transient single-strand break (cleavage of one phosdiester bond in one strand of the double helix) the
provides an axis of rotation to allow unwinding.

Cairns' interpretation of the autoradiographs was on that semiconservative replication started at a site
on tio the chromosome, which he caled the "origin," and proceeded sequentially and unidirectionally
around ) the circular structure (Fig. 5.16). Subsequent evidence has shown his original interpretation to
be incorrect s on one point: replication actually proceeds bidirec- n tionally, not unidirectionally. Each Y-
shaped structure d is a replication fork, and the two replication forks move in opposite directions
sequentially around the circular chromosome (Fig. 5.16).

Unique Origins and Bidirectional Replication

Cairns' results provided no information as to whether the origin of replication (site at which replication is
a initiated) is a unique site or occurs at random on the c chromosome. Moreover, his results did not
allow him to differentiate between uni- and bidirectional replica-h tion. We now have direct evidence
showing that replication in E. coli and several other organisms proceeds bidirectionally from a unique
origin. These features of DNA replication can be illustrated most simply and convincingly by experiments
with some of the small he bacterial viruses.

Bacteriophage lambda (phage X) is, like phage T2, is avirus that grows in E. coli. It has a small
chromosome the consisting of a single linear molecule of DNA only 17.5 nim um lorg. The phage A
chromosome is unique in that it has a single-stranded region, 12 nucleotides long, at epli te 5' end of
each complementary strand (Fig. 5.17).

These single-stranded ends, called 'cohesive" or is "sticky" ends, are complementary to each other. The i
cohesive ends of a lambda chromosoms can rhus base-pair to form a hydrogen-bonded ircular struc pe
ture. One of the first events to occur afte: a lambda chromosome is injected into a host cell is its conver-
st sion to a covalently closed circular molecule (Fig. br 5.17). This conversion from the hydrogen-bonded
circular form to the covalently closed circular form is to catalyzed by polynucleotide ligase, a very
importanthr enzyme that seals single-strand breaks in DNA double m helices. (Polynucleotide ligase is
required in most, f pe not all, organisms for DNA replication, DNA repair, and ch recombination between
DNA molecules.) The lambda un chromosome replicates in its circular form via de 0-shaped
intermediates (Fig. 5.18), as does the E. coli chromosome.

The feature of the lambda chromosome that facilitated the demonstration of bidirectional replication ma
was its differentiation into regions containing high m concentrations of adenine and thymine ("ATrich"re
regions) and regions with large amounts of guanine of and cytosine ("G-C rich" regions). In particular, it
th contains a few segments with very high A-T content nu ("A-T rich clusters"). These A-T rich clusters
were used an as physical markers by M. Schnös and R. B. Inman to re demonstrate, using a technique
called "denaturation ar mapping," that replication of the lambda chromosome is initiated at a unique
origin and proceeds bidirec- e tionally rather than unidirectionally.

When DNA molecules are exposed to high tem- c perature (100°C) or high pH (11.4), the hydrogen and la
hydrophobic bonds that hold the complementary r ds together in the double-helix configuration are g
broken, and the two strands separate. This process is ir called denaturation. Because A-T basepairs are
held is together by only two hydrogen bonds, compared with nt three hydrogen bonds in G-C base-pairs,
A-T rich molecules denature more easily (at lower pH or tem- f perature) than G-C rich molecules. When
lambda chromosomes are exposed to pH 11.05 for 10 minutes da under the appropriate conditions, the
A-T rich clusters ia enature to form "denaturation bubbles," which are detectable by electron
microscopy, whereas the G-C rich regions remain in the duplex state (Fig. 5.19) il These denaturation
bubbles can be used as physical on markers whether the lambda chromosome is in its gh mature linear
form, its circular form, or its 0-shaped replicative intermediates. By examining the positions ine of the
branch points (Y-shaped structures) relative to it the positions of the denaturation bubbles in a large ent
number of 0-shaped replicative intermediates, Schnös ed and Inman demonstrated that both branch
points are to replication forks that move in opposite directions on around the circular chromosome. A
summary of their me results is shown in Fig. 5.20. A schematic illustrating the rationale of this procedure
is shown in Fig. 5.2.

Bidirectional replication from a fixed origin has also been demonstrated for several organisms with
chromosomes that replicate as linear structures. Rep- lication of the chromosome of phage T7, another
small coliphage, begins at a unique site near one end to form a so-called "eye" structure (Fig. 5.22a) and
then pro- ceeds bidirectionally until one fork reaches the nearest end. Replication of the Y-shaped"
structure (Fig 5.22b) continues until the second fork reaches the other end of the molecule, producing
two progeny chromosomes.

Replication of DNA molecules in the chromo- somes of eukaryotes is also bidirectional (Chapter 6)
However, bidirection replication is not universal. The chromosome of coliphage P2, which like the
lambda chromosome is circular during replication, replicates unidirectionally from a unique origin.
Whether origins of replication in other organisms are always unique sites is also unknown. It is clear that
secondary, nor- mally inactive origins exist in some organisms. If the primary origin of replication of
phage T7 is deleted (physically removed from the chromosome), a second unique site on the T7
chromosome takes over this function. Why this secondary origin is inactive in the presence of the
primary origin is not known.

Figure 5.21 The use of electron microscope denaturation mapping to distinguish between (a)
unidirectional and (b) bidirectional modes of chromosome replication. Treatment of chromosomes at pH
11.05 for 10 minutes at 25°C under appropriate conditions will denature only regions contain- ing high
concentrations of A-T base-pairs. These "denatur- ation bubbles" can be used as physical markers for
specific sites on the chromosome. By examining the positions of the replication forks relative to these
markers in a population of replicating chromosomes, one can distinguish between uni- and bidirectional
replication.

DNA Polymerases and In Vitro DNA Synthesis Much has been learned about the molecular mecha- nisms
involved in biological processes by fractionating P cells into their various organelles, macromolecules and
other components, and then reconstituting sys- tems in the test tube, so-called in vitro systems, that are
capable of carrying out particular metabolic events. Such in vitro systems can be dissected biochemically
much more easily than in vito systems. Clearly, the information obtained from studies on in titro systems
has been invaluable. One should never assume, how ever, that a phenomenon demonstrated in vitro
occurs in vivo. Such an extrapolation should be made only when independent evidence from in vivo
studies vali- dates the in vitro studies The in vitro synthesis of DNA was first accom- plished by Arthur
Kornberg and his coworkers in 1957. Kornberg, who received the Nobel Prize in 1959 for this work,
isolated an enzyme from E. coli (initially called DNA polymerase or "Kornberg enzyme" now known as
DNA polymerase 1) that catalyzes the covalent addition of nucleotides to preexisting DNA chains. r The
enzyme requires the 5'-triphosphates of each of e the four deoxyribonucleosides: deoxyadenosine tri- d
phosphate (dATP), deoxythymidine triphosphate (dTTP d often written as TTP in the past because "ribo"
deriva- is tives of thymine had not been identified), deoxygua- e nosine triphosphate (dGTP), and
deoxycytidine triphosphate (dCTP). The enzyme is active only in the presence of Mg2 ions and
preexisting DNA This DNA must provide two essential components, one serving a primer function and
the other a template function a (Fig. 5.23). The overall reaction catalyzed by DNA ng polymerase I is
shown in Fig. 5.24; a diagram of the e functional sites of the enzyme is shown in Fig. 5.25. ire 1. Priner
DNA. DNA polymerase I cannot initiate the ts. lly he ms synthesis of DNA chains de novo. It has an
absolute requirement for a free 3' bydroxyl on a preexisting DNA chain. DNA polymerase I catalyzes the
formation of a phosphodiester bridge berween the 3' OH at the end of the primer DNA chain and the 5'-
phosphste of the incoming deoxyribonucleoxide. The direction of synthesis is thus always 5-3 (Fig 5.26).
urs nly 2. Template DNA. DNA polymerase I does not contain ali- sequence specificity. That is, the
enzyme requires a DNA template chain (Fig. 5.23) whose base sequence dic- tates, via the DNA base-
pairing requirements, the syn- thesis of a complementary base sequence in the strand being synthesized
57. for all ow ova Since Kornberg's discovery and extensive pio neering work with DNA polymerase I of E.
coli, a large Figure 5.26 Covalent extension of DNA chains in de 53' direction as catalyzed by all known
DNA poly merases. The existing chain terminates at the 3' end with C the nucleotide deoxyguanylate (or
deoxyguanosine 5-phosn phate). The diagram shows the DNA polymerase-catalgzed addition of
deoxythymidine monophosphate (from the pre cursor deoxythymidine triphosphate, diTP) to the 3' end
of the chain with the release of pyrophosphate (P,O,) iti of pr m. number of DNA polymerases have been
isolated and ba characterized from many different organisms. Three the different DNA polymerases (I, II,
and III) have been the identified and studied in E coli and B. subtilis. Simi- pri larly, three DNA
polymerases, called a, B, and y, have be been identified in several different eukaryotes, and of more
recently a fourth DNA polymerase, named 8, nas ex been isolated from calf thymus and rabbit bone mar
ex row. Thus, there a:e at least four different DNA poly po rat The precise functions of some of the
polymerases ac are still not clear. However, in E. coli and B. subtilis, Th DNA polymerase III, rather than
DNA polymerase Ias D me ases in eukaryotes. first believed, is the major replicative enzyme. Some of me
the strongest evidence for this has come from studies of mutants, so-called polA mutants, that are
deficient in rea DNA polymerase I. These polA mutants of E coli replicate their capacity to repair damage
to DNA caused, for example, by ultraviolet irradiation. This and other evidence suggest that a major
function of DNA poly merase I is DNA repair. Still other evidence indicates that DNA polymerase I is
responsible for the excision of the RNA primers used in the initiation of DNA synthesis (see pp. 118-119).
The function(s) carried out by DNA polyinerase II is uncertain, although it can function in DNA repair in
the absence of DNA poly merases I and IIL DNA polymerase IlI, however, plays an esserntial role in DNA
replication, because in mutant strains growing under conditions where no functional polymerase III is
synthesized, DNA synthesis stops. Most of the prokaryotic DNA polymerases studied so far not only
exhibit the 5 3' polymerase activity discussed earlier, but also have a 3'5' exonu clease activity. (An
exonuclease is an enzyne that degrades nucleic acids from the ends, as opposed to an endonuclease,
which degrades nucleic acids by mak ing internal cuts.) Both activities (polymerase and exonuclcase) are
present in the same protein macro- molecule. The 3 -5' exonuclease activity catalyzes the removal of
nucleotides, one by one, from the 3' ends of polynucleoxide chains. Some polymerases in Such as DNA
polymerase I of E. coli, also have 5 3 A ly exonuclase activity. When present, 5 3' exonu- with clease
activity is found at the site on the protein -phos molecule distinct from the active side catalyzing the zed
53' polymerase reaction and the 3-5' exonu he preclease reaction Both of these polymerase-associated
exonuclease activities play important roles in DNA metabolism The 35' exonuclease activity of DNA poly
merases carries out a critical proofreading" or ed- tting Junction that is necessary for the bigb degree of
fidelity characteristic of DNA replication When presented with a template-primer DNA that has a ter.
minal mismatch (an unpaired or incorrectly paired d and base or sequence of bases at the 3' end of the
primer), Three the 35 exonuclease of the polymerase clips off been the unpaired base or bases (Fg.5.27).
When an appro- Simi priately base-paired terminus results, the polymerase have begins resynthesis by
adding nucleotides to the 3' end s, and of the primer strand and continues until the template is 8, nas
exhausted. This proofreading function of the 3-5" mar exonuclease, built into DNA polymerases, is very
im- poly portant, for DNA replication must be extremely accu- rate. A tolerable mistake level could
probably not be erases achieved without such a proofreading mechanism btilis, Thus, it is somewhat
surprising that of the eukaryotic se I as DNA polymerases studied to date, only DNA poly. me of merase 8
possesses the 35' exonuclease proof tudies ent in reading activity The 5-3' exonuclease activity of many
colprokaryotic DNA polymerases is also very important. It Figure 5.27 "Proofreading" by the
3'5'cxonuclease exonucle activity of DNA polymerases during DNA r polymerase is presented with a
template and primer contain (b). The ing a 3' primer terminal mismatch (a), it will not catalyze terminus
covalent extension ("polymerization"). Instead, the 3extensio IF DNA functions in the removal of
segments of DNA Gamaged by ultraviolet light irradiation and other agents (see Chapter 11). The 53'
exonuclease of poly merases, such as E. coli DNA polymerasc I, alsc finc- tuons in the removal of RNA
primers from DNA (see pp. onstrat discon Au that the each re 118-119). 1his 53' exonuclease activity has
not direct been found in any of the eukaryotic DNA polyirerases. compl In eukaryotes, the functions of
the 53' exoru oppos clease activity associated with DNA prilymerases in tende tion (B synthe prokaryotes
must be carried out by a separate en.zsne. strand The "Growing-Point Paradox" and Discontinuous DNA
Synthesis (Fig. strand Studies of replicating DNA molecules by autoradiogr tion. phy and electron
microscopy indicate that the two 3'- progeny strands being synthesized at each replicaring segm fork are
being extended in the same overall di ection.joinin at least on the macromolecular level. Since the com
ligas plementary strands of a double helix have opposite DNA polarity, this means that synthesis is
occurring at the5' medi end of one strand (or 3'5) and the 3' end of the by gr other strand (53). As
discussed in the preceding medi section, however, all known polymerases have an ab- Whe solute
requirement for a free 3'-hydroxyl; they orly fore carry out 53' synthesis. This paradox existed for 1000
many years during which biochemists searchied in vain segn for new polymerases that could carry out 35
after synthesis. No such polymerase has yet been found. bou Instead, strong evidence has accunulated
indicating long that all synthesis occurs in the 5'3' direction. labe The resolution of the paradox resulted
from the dem exonuclease activity, an integral part of many DNA poly. DNAmerases, will cleave off the
terminal mismatched nucleotide ntain (b). Then, presented with a correctly base-paired primen talyze
terminus, DNA polymerase will catalyze 5 3' covalent 'extension of the primer strand (c) aged onstration
that the synthesis of one DNA strand is (s2e poly- unc e pp not ses discontinuous Autoradiography and
electron microscopy show that the two nascent DNA strands being synthesized at each replicating fork
are being extended in the same direction at the macromolecular level. Since the complementary strands
of a DNA double helix have oru opposite chemical polarity, one strand is being ex s in tended in an
overall 5'3' direction and the other ne. strand is being extended in an overall 35' direc tion (Fig. 5,28,
top). At the molecular level, however synthesis is actually occurring in opposite directions (Fig. 5.28,
bottom). At the molecular level both new strands are being synthesized in the 53' direc- tion. The strands
being extended in the overall two 3' 5' direction grow by the synthesis of short aring segments
(syr.thesized 53'), and the subsequent tion. joining of these short segments by polynucleotide ligase. The
evidence for this discontinuous mode of osite DNA replication has come from studies in which inter he 5'
mediates in DNA synthesis were radioactively labeled f the by growth of cells for very short periods of
time in ding medium containing PHthymidine ( pulse-labeling"). ab When L coli cells were pulse-labeled
for 15 seconds for example, all the label was fourid in small pieces, for 1000-2000 nucleotides long.
These small pieces or vain segments of DNA, often called "Okazaki fragments" 5 after R Okazaki, who
first iden ified them, are smaller, about 100-200 nucleotides long, in eukaryotes. When longer pulse-
labeling periods are used, more of the ion. label is recovered in large DNA molecules-probably the size of
molecules containing all the DNA present in Figure 5.28 Discontinuous DNA synthesis. (1) Relatively C
low resolution techniques such as autoradiography and elec- tron microscopy show that both nascent
DNA chains are extended in the same overall direction at each replication fork. Since the two chains have
opposite polarity, the overall or macromolecular direction of extension must be 53' on one chain and 3 5'
on the other chain. Both are being w extended from left to right, for example, at the fram replication
fork. (2) Higher-resolution biochemical tech- niques such as pulse-labeling and density-gradient analysis
show that replication is actually discontinuous for the chain being extended in the overall 3' 5' direction.
Short tiv fragments are synthesized in the 5-3' direction and w subsequently are joined by polynucleotide
ligase (C"DNA x ligase"). periods the radioactivity present in short DNA "fragments" becomes
incorporated in chromosome-sized DNA molecules during subsequent growth of the cells on medium
containing nonradioactive thymidine. This is important because it indicates that the "Okazaki frag-
ments" are true intermediates in DNA synthesis rather than some kind of metabolic by product.
Extensive evidence has shown that DNA synthesis is continuous or the strand growing in the overall 5'3'
direc tion (sometimes called the "leading" strand) and is discontinuous for the strand growing in the
overall 35 direction (sometimes called the "lagging intict chromosomes. In short pulse-labeline strand) as
is shown in Fig, 5.28. Initiation and the "Primer Problem As has been emphasized earlier, all known DNA
poly merases have an absolute requirement for a free 3' OH on a DNA primer plus an appropriate DNA
template strand for activity. Thus, no known DNA polymerase cun initiate the syntesis of a neu strand of
DNA. Since the synthesis of each "Okazaki fragment" requires an initiation event, an efficient mechanism
of chain initiation is essential for ongoing DNA replication. RNA polymerase, a complex enzyme that
catalyzes the synthesis of RNA molecules from DNA templates, has long been known to be capab'e of
initiating the synthesis of new RNA chains at specific sites on the DNA. When this occurs, an RNA-DNA
hybrid is formed in which the nascent RNA is hydro gen-bonded to the DNA template. Since DNA poly
merases are capable of extending polynucleoti elativelychains containing an RNA primer with a free 3'-
OH l scientists in several laboratories began testing the idea lilcation that DNA synthesis is initiated by
RNA primers. There is overal now definitive evidence supporting the proposal that 3 DNA syntbesis is
"primed by short segments of RNA, e being which are later removed by a 53' exonuclease framed and
replaced by DNA prior to covalent sealing by anasis Polynucleotide ligase (Fig. 5.29), In E. coll, the RVA
echain primers are excised by the 5-3 exonuclease ac- Short tivity of DNA polymerase I. This occurs
simultaneously on and with the synthesis of new DNA strands (replacing the CDNexcised RNA primer
strands) by the3'poly- nd elec ins are l tech merase activity of this enzyme (Fig 5.29) The synthesis of the
RNA primers is catalyzed by enzymes called primases, which have properties quite distinct from those of
the RNA polymerases. The E. coli primase is the product of the dnaG gene. In prokary otes, the RNA
primers are 10-60 nucleotides in length. In eukaryotes they are quite shorn, about 10 nucle. otides long.
The use of RNA primers is almost certainly the most common mechanism used to initiate DNA synthesis.
Nevertheless, certain viruses appear to have Figure 5.29 Schematic illustration of the initation of DNA
synthesis via RNA primers. A short RNA strand is synthesized to provide a 3'-OH primer for DNA
synthesis. This RNA primer is subsequently removed and replaced with DNA by the dual 5'3' exonuclease
and 53polymerase volved quite different mechanisms for the intiation of DNA synthesis (see Kornberg,
1980). The Complete "Replication Apparatus" Is Complex When Watson and Crick worked out the
double-helix structure of DNA, they immediately recognized that the complementary nature of the two
strands prcvided a simple basis for the faithful duplication of genetic material. Meselson and Stahl's
demonstration of the semiconservative replication of the E. coli chromo- some solidified the concept
that the two strands of the double helix unwind and serve as templates for the activities built into DNA
polymerase . DNA ligase then covalently closes the nascent DNA chain, catalyzing the formation of
phosphodiester linkages between adjacent 3'-hydroxyls and 5'-phosphates. synthesis of complementary
strands. Thus, a parental double helix directs the synthesis of two identical progeny double helices.
Kornberg's isolation of an enzyme, DNA polymerase I, capable of synthesizing DNA in vitro appeared to
provide the final link in what was thought to be an elegantly simple mechanism for the replication of the
genetic material-but such was not the case. Twenty years later, scientists are still trying to work out the
details of the mechanism of DNA replication DNA replication is complex. It is carried out by a
multienzyme complex, often called the replicatton apparatus or the replisome. In eukaryotes, the com-
ponents of the replication machir.ery are just begin Figure 5.30 Complexity of the E. coli replication
appara tus. Only those proteins that have been purified (or partially Str purified) and studied in vitro are
shown. Other gene pro ucts, such as the products of genes dnaj, dnak, dnal, dnaPco and dnaT, are known
to be required for replication. Howma ever, these gene-products have not yet been identified. (After A.
Kornberg, DNA Replication, Freeman, San FrancisCO, 1980.) pol tor 5.3 as pro Ver ma acc ning to be
identified. Even in prokaryotes, DNA repli hel of how some of these proteins function in DNA repli- mai
cation requires many different proteins, and the details gyr cation are still being investigated today. For
example, DNA replication in E coli requires at least two dozep top different gene-products. Many of these
gene-products of have been purified and their roles in DNA replication (se studied in vitro. Figure 5.30
shows the involvement of lyze some of these E. coli proteins in DNA replication; it is intended to illustrate
the complexity of the replication process rather than to illustrate the specific roles of the individual gene-
products. Pri pri call prir First, the two complementary strands of the paren des tal double helix have to
be unwound and separated so of that each can serve as a template for the synthesis of a new daughter
strand. Unwinding and movement of the replication fork occur processively with the strands being
transiently unwound ahead of the fork as it moves along the chromosome. Three different types of
proteins appear to contribute to unwinding the strands to car str all syn str of double helices. (1) DNA
unwinding proteins or Fig DNA helicases are directly involved in catalyzing the the unwinding of the
double helices. In E. coli, two differ ent helicases are involved. One helicase, the product of the rep gene,
binds to and stimuliles sp the strand that has 3' to 5' polarity in the direction of replication fork
movement. The other helicase (exact identity still uncertain) binds to and assists unwinding of the strand
that has 5' to 3' polarity in the direction that the fork is moving. (2) DNA single strand bind ing proteins
(SSBPs) bind tightly to single-stranded regions of DNA produced by the action of the helicases and help
sabilize the extended single stranded tem plates needed for polymerization. The SSBPs bind to DNA as
tetramers, and their binding exhibits cooper tivity (ie, the binding of one tetramer stimulates the binding
of additional tetramers to adjacent segments of single-stranded DNA). The binding of SSBP to single
stranded DNA tends to hold that DNA in an extended configuration and prevents it from folding back orn
itself. Single-stranded DNA that is saturated with bound SSBP replicates over 100 times faster than
uncom plexed single-stranded DNA in vitro. Presumably, un complexed single strands of DNA form
secondary structures that interfere with the movement of DNA polymerases or other components of the
replication dnaP complex along the molecule in the normal processive How manner. (3) Finally, DNA
gyrases, which catalyze the ormarion of regative supercoils in DNA (see Fig 536), are essential for
replication and are believed to play a key role in the unwinding process. Supercoiling has been proposed
to help "drive the unwinding process, however, we still do not know how this works Very recently, it has
been suggested that DNA gyrase may function by removing positive supercoils that accumulate in front
of the replication forks as the repli helicases unwind the double helices. In any case, DNA details gyrases
are essential for DNA replication and somehow repli maintain pre- and postreplicative DNAs in the
proper ng nd apparas e p artiall (After ncisco, ample, topological structures. dozep Nascent DNA strands
are then initiated by the use of RNA primers by the mechanisın discussed earlier ication (see Fig. 5.29)
Synthesis of the RNA primers is cata ent of lyzed by a special class of enzymes called primases n; it is
Primase activity requires the formation of a complex of cation primase and at least six other proteins;
this complex is of the called the primosome. In addition to primase, the rimosome contains prepriming
proteins tentatively paren designated proteins i, n, n' and n" plus the products so of genes dnaB and
dnaC (Table 5.4). The primosome is of a carries out the initial priming reaction for the leading of the
strand (the strand extended continuously in the over all 5' to 3' direction) and the repeating priming of
the as it synthesis of "Okazaki fragments" for the lagging pes of strand (synthesized discontinuously in
the overall 3' trands to 5' direction-but 5' to 3' at the molecular level, see oducts ted trands ins or Fig
5.28). The functions of the individual proteins in the primosome are still uncertain g the differ- The
covalent extension (see Fig. 526) of the prmed DNA chains during chromosome replication in E. coli is
carried out by DNA polymerase III. Unlike i DNA polymerase I of E. coli (which is a single poly peptide; see
Fig. 5.25), DNA polymerase III is a com plex enzyme containing seven different polypep tides (Fig. 5.31),
and all of these polypertides inust be s present for proper replicative function. The 5' to 3'p polymerase
activity and the 5' to 3' exonuclease activ p ity are both present on the a polypeptide of DNA ex
polymerase III. The 3' to 5' proofreading activicy (see of Fig. 5.27) of polymerase II is present on the e pol
D peptide. The functions of the other subunits are stil! pr uncertain. Subsequent to DNA polymerase III
activity Ta at the replication fork, DNA polymerase I catalyzes the re removal of the RNA primers by the
concerted action of po its 5' to 3' exonuclease activity and its 5' to 3' poly ys merase activity, and DNA
ligase catalyzes covalent clo- Sti sure of the resulting single-stranded "nick" (Fig. 5.30) Several of the
components essential for DNA rep- ge lication have been identified genetically, that is, E. coli ou strains
carrying mutations (heritable changes in the fur genetic material, see Chapter 11) that result in the
inability to replicate DNA under certain conditions ly (usually high temperature) have been identified. m.
When these mutations were charatterized genetically p (see Chapters 7 and 8), they were found to
identify a be set of genes (designated dnas, dnaB, etc.) whose 3' products are required for DNA synthesis
in vivo. The tiv. products of some of these genes are known. For A example, dnaE, dnaN, dnax, and dnaZ
code for four ee of the seven subunits (polypeptides) of the complete DNA polymerase III enzyme, and
dnaG specifies the til! primase. The products and functions of others (see ity Table 5.4) are still unknown.
Other components of the the replication enzymes (e.g, some of the subunits of DNA of polymerase III)
were discovered by bioc emical anal- ly yses, and the genes that encode these proteins have lo- still not
been identified It is hoped that the exact functions of the many ep gene products involved in replication
will be worked out during the next few years. Attempts to isolate intact, the functional replisomes,
however, have been largely un the successful. Reconstitution of subconiplexes of replica Figure 532
(Right page) Three stages in the replication of te the single-stranded DNA of bacteriophage (PX174. (Top,
a-e) Stage I: conversion of the single-stranded chromosome (a) to a double-stranded pareptal replicative
(RF) form (e). (b) sy fr ad Synthesis of the complementary negative" ( strand is in initiated by the
synthesis of a short RNA primer. This reaction A is catalyzed by primase and requires the activity of a
complex a of at least six different priming proteins; this complex is tiv sometimes called the
"primosome." (c) DNA polymeraseIII cle next catalyzes the covalent addition of deoxvribonucleotides an
to the 3' end of the RNA primer. Synthesis of the comple. tiv mentary negative strand then takes place
discontinuousty stu (see Figs. 5.28, and 5.29) until the "positive" (7) surand na t:mplate is exhausted. The
primosome appears to travel a around the circular emplate strand, pausing uo iritiate the R synthesis of
each new,"Okazaki fragment." (d) Excision of P the RNA primers and gap filling appear to be catalyzed by
ur DNA polymerase I, as shown in Fig. 5.29. Polynucleotide (E ligase ("DNA ligase") then catalyzes the
formation of a c covalent linknge between the adjacent 3'-OH and 5'-PO4 R roups, to produce the closed,
double-stranded parental RF ne e). (Center e-1) Stage II: "rolling circle" replication of the s parental RF (e)
to produce a population of progeny RFs (I) fro (f) The positive strand of the parental RRis cut at the
originth by the site -specific endonuclease ("nickase) activity of the si px174 gene A protein. The gene A
protein nicks the parental ve RF only at the origin, it will not cut most other DNA mole sy cules at all. (f-h)
During the nicking event, the gene l protein becomes covalently attached to the 5' phosphor a group of
the positive strand. It remains linked to the 5' purified proein tion apparatuses from successful This is
undoubtedly a result of the fact thar the complexes are held together by relatively weak protein-proein
interactions, which are disrupted during the isolation procedures In addition, replication com 0 kD lexes
may be membrane-bound and require mem- e structures for their assembly. There is consider able
evidence that replication forks are associated with the cell membrane in prokaryotes and with the
nuclear envelope in eukaryotes For excellent, more detailed accounts of replica tion and the components
of the replication apparatus the reader is referred to Kornberg's DNA Replication and 1982 Supplement
to DNA Replication holoen subunits Phage x174 and "Rolling Circle" Replication Bacteriophage (ox174 is
representative of a group of small viruses, both bacterial and eukaryotic, that store their gc netic
information in a single-stranded, circular molecule of DNA. When these viruses infect a host cell, E colf in
the case of dx174, the single-stranded viral tion ofterminus until a complete p p,a-c) synthesized (g) The
5' end of the positive strand is displaced mc (a) from the negative strand and deoxyribonucléotides are c)
(b) added to the free 3'-OH as the circle (maintained by the rand is Intact negative strand) rotates about
its axis. (g h) The gene eaction A protein remains bound at the replication fork as it travels mplexaround
he circular negathe plex is tive strand origin has been symthesized, the gene A protein rase III cleaves the
nascent origin and simultaneously ligates the 3 oides and 5' termini to produce a covalently closed
circular posi y positive strand has been (h-I) Once a ner posi- ompl srand le tive strand (I) Synthesis of the
complementary negative uousth strand ther akes place discontinuousy as in saage l using the nascent
positive strand as templare. The ragments travel are initiaied by RVA primers in this sage also, however,
these ate ie RNA primers are not sheorwn in the diagram. (H-I) The sion of parental RF continues to
replicate by the rolling circle mode zed by until a population of about 60 progeny RFs are produced
deotide (Bot.om, I-q) Stage III: synthesis of single-stranded progeny n of a chromosomes, (-p) Rolling
circle replication of progeny 5 PO RFs occurs just as for parental RFs in stage II, except that Instead, the
positive ntal RFnegative strand s are not synthesized. 1 of the strands are packaged in progeny virions.
(n-p) The switch positive strand syn- RFs (I). from RF synthesis (stage I1) to orlgin thesis (stage 11l) results
from the binding of newly synthe y of thesied viral coar proxeins to the nascent positive strand, pre
arentalventing it from serving as a template for negative strand molesynthesis.(q) Maturation of the
progeny virion completes the plage 0x174 life circle Approxímately 500 progeny virions sphor e produced
per infected cell. DNA (called the "positive" (+) strand) is converted to a double helical form (called the
"replicative form RF) by the synthesis of a complementary" negative cast () strand. This double-stranded
parental RF, then x replicates, which in turn asymmetrically to produce a large population of viral (+)
process results. The viral projection, "strands are then incorporated into protein coats to ecoplete the
reproductive cycle. The replication of the pX174 chromosome can be divided into three of stages: (1)
parental (+) RF parental strand, (2) n parental RF progeny RFs, and (3) progeny RFs progeny (+) strands
(Fig. 5.32). In the last two termediate in the replication of the DNA of bacteriophage px174. A single-
stranded tail is seen extending from a act en double-stranded, circular replication form (RF. (From at K.
Koths and D. Dressler, Proc. Natl. Acad. Sci. 75: 605, 1978.) th me ter to re ci stages, DNA synthesis occurs
by a different mechanism called "rolling circle" replication. po Most of the features of "rolling circle"
replication l are the same as those discussed earlier for replication 5 via the more common 6, "eye, and
Y-shaped struc. val tures. In this case, however, the replicative structure is St a circular DNA molecule
with a single-stranded tail (Fig 5.33) is or Rolling circle replication is initiated when the sequence-specific
endonuclease activity of the phage be X174 gene A protein cleaves the positive strand of the parental RF
at the origin of replication (Fig. 5.32) This so endonuclease activity is site-specific, it cuts the pX174 ria
chromosome at only one site, the origin of replicaex tion. It produces 3'-OH and 5'-phosphate termini at
rR the site of the cut in the () strand; the (-) strand Ch remains intact. The 5' end of the (+) strand is un-
wound and "peeled off" while the () strand rotates about its axis (thus the name rolling circle"). This P
yields the circle with its tail (Figs. 5.32 and 5.33). As initially proposed by W. Gilbert and D. Dressler, the
rolling circle model of DNA replication included a se specific enzyme, called a "transferase," which
attached of the 5' end of the (+) strand to a specific site on the cell ch membrane. Although most, if not
all, replicating chro- tra mosomes are attached to the mémbrane, little is known about the specific
nature of such attachments. tio In any case, membrane attachment is not an essential feature of rolling
circle replication. As the circle rotates and the 5' end is displaced, DNA polymerase catalyzes gra covalent
extension at the other (3'-OH) end. During parental RF to progeny RF replication, the nascent positive
strands are used as templates for the discontinuous synthesis of complementary negative strands. In
some cases, the synthesis of the comple- mentary strand may occur discontinuously on the sin- gle-
stranded tail before synthesis of the first strand has been completed. In such cases, a double-stranded tai
will be produced. The switch from double-stranded RF DNA synthesis to single-stranded viral (+) DNA
syn- thesis occurs when specific proteins of the viral coat are produced in the cell. Rolling circle
replication continues, but as the viral strand is displaced, these coat proteins bind to it and prevent the
synthesis of complementary () strands (Fig. 5.32) The phage x174 gene A protein is a key protein in px174
replication. It possesses a remarkable set of e activities. (1) Gene A protein possesses a site-specific
endonuclease activity that cleaves the positive strand om at the origin. (2) Gene A protein then
maintains 8.) the energ of the cleaved phosphodiester linkage by means of a covalent attachment of the
5'-phosphory terminus to itself. (3) Gene A protein remains bound to the 5'-terminus of the positive
strand and to the replication fork while the fork traverses the complete sm circular minus-strand
template. (4) When a complete positive strand has been synthesized, gene A protein on cleaves the new
origin, ligates the 3' hydroxyl and on 5'phosphoryl termini, and once again becomes co- uc valently linked
to the newly generated 5'-positive- is strand terminus. This cycle of gene A protein activities ail is
repeated until a population of progeny RFs (stage II) or progeny positive strands (stage III) is produced. To
date, evidence for rolling circle replication has ge been found for (1) single-stranded DNA viruses like he
px174, (2) the replication associated with chromo- his some transfer during "mating" (conjugation) in
bacte- 74 ria (see Chapter 8), and (3) the repiication of small a extrachromosomal DNA molecules
carrying clusters of at rRNA genes during oögenesis in amphibians (see he nd Chapter 15).