DNA Replication and Repair

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 533

13 DNA Replication and Repair 13.1 13.2 13.3

DNA Replication DNA Repair Between Replication and Repair

The Human Perspective:

The Consequences of DNA Repair Deficiencies

R

eproduction is a fundamental property of all living systems. The process of reproduction can be observed at several levels: organisms duplicate by asexual or sexual reproduction; cells duplicate by cellular division; and the genetic material duplicates by DNA replication. The machinery that replicates DNA is also called into action in another capacity: to repair the genetic material after it has sustained damage. These two processes—DNA replication and DNA repair— are the subjects addressed in this chapter. The capacity for self-duplication is presumed to have been one of the first critical properties to have appeared in the evolution of the earliest primitive life forms. Without the ability to propagate, any primitive assemblage of biological molecules would be destined for oblivion. The early carriers of genetic information were probably RNA molecules that were able to self-replicate. As evolution progressed and RNA molecules were replaced by DNA molecules as the genetic material, the process of replication became more complex, requiring a large number of auxiliary components. Thus, although a DNA molecule contains the information for its own duplication, it lacks the ability to perform the activity itself. As Richard Lewontin expressed it, “the common image of DNA as a self-replicating molecule is about as true as describing a letter as a self-replicating document. The letter needs a photocopier; the DNA needs a cell.” Let us see then how the cell carries out this activity. ■

Three-dimensional model of a DNA helicase encoded by the bacteriophage T7. The protein consists of a ring of six subunits. Each subunit contains two domains. In this model, the central hole encircles only one of the two DNA strands. Driven by ATP hydrolysis, the protein moves in a 5⬘ → 3⬘ direction along the strand to which it is bound, displacing the complementary strand and unwinding the duplex. DNA helicase activity is required for DNA replication. (COURTESY OF EDWARD H. EGELMAN, UNIVERSITY OF VIRGINIA.)

533

JWCL151_ch13_533-559.qxd

534

7/29/09

10:09 PM

Page 534

Chapter 13 DNA REPLICATION AND REPAIR

13.1DNA REPLICATION The proposal for the structure of DNA by Watson and Crick in 1953 was accompanied by a suggested mechanism for its “self-duplication.” The two strands of the double helix are held together by hydrogen bonds between the bases. Individually, these hydrogen bonds are weak and readily broken. Watson and Crick envisioned that replication occurred by gradual separation of the strands of the double helix (Figure 13.1), much like the separation of two halves of a zipper. Because the two strands are complementary to each other, each strand contains the information required for construction of the other strand. Thus once the strands are separated, each can act as a template to direct the synthesis of the complementary strand and restore the double-stranded state.

1

Parental DNA molecules

DNA molecules from 1st generation progeny

DNA molecules from 2nd generation progeny

Semiconservative Replication The Watson-Crick proposal shown in Figure 13.1 made certain predictions concerning the behavior of DNA during replication. According to the proposal, each of the daughter duplexes should consist of one complete strand inherited from the parental duplex and one complete strand that has been newly synthesized. Replication of this type (Figure 13.2, scheme 1) is said to be semiconservative because each daughter duplex contains one strand from the parent structure. In the absence of information on the mechanism responsible for replication, two other types of replication had to be considOld

SEMICONSERVATIVE REPLICATION 2



Old T

A

T

A A

T


G C

G C

G A

T C G

CONSERVATIVE REPLICATION

T A T G C C G A T

3


T T A G

C

A

T

C

G

C C

G

New

T A A

G A

Old New


T A A

T T

T

New

G

A C

T T

G G

A T C T

G A

A C

G G

A T C


New Old

FIGURE 13.1 The original Watson-Crick proposal for the replication of a double-helical molecule of DNA. During replication, the double helix unwinds, and each of the parental strands serves as a template for the synthesis of a new complementary strand. As discussed in this chapter, these basic tenets have been borne out.

DISPERSIVE REPLICATION

FIGURE 13.2 Three alternate schemes of replication. Semiconservative replication is depicted in scheme 1, conservative replication in scheme 2, and dispersive replication in scheme 3. A description of the three alternate modes of replication is given in the text.

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 535

13.1 DNA REPLICATION

Transfer to 14N

Generations

Light

Parental

Hybrid Heavy

ered. In conservative replication (Figure 13.2, scheme 2), the two original strands would remain together (after serving as templates), as would the two newly synthesized strands. As a result, one of the daughter duplexes would contain only parental DNA, while the other daughter duplex would contain only newly synthesized DNA. In dispersive replication (Figure 13.2, scheme 3), the parental strands would be broken into fragments, and the new strands would be synthesized in short segments. Then the old fragments and new segments would be joined together to form a complete strand. As a result, the daughter duplexes would contain strands that were composites of old and new DNA. At first glance, dispersive replication might seem like an unlikely solution, but it appeared to Max Delbrück at the time as the only way to avoid the seemingly impossible task of unwinding two intertwined strands of a DNA duplex as it replicated (discussed on page 538).

535

To decide among these three possibilities, it was necessary to distinguish newly synthesized DNA strands from the original DNA strands that served as templates. This was accomplished in studies on bacteria in 1957 by Matthew Meselson and Franklin Stahl of the California Institute of Technology who used heavy (15N) and light (14N) isotopes of nitrogen to distinguish between parental and newly synthesized DNA strands (Figure 13.3). These researchers grew bacteria in medium containing 15N-ammonium chloride as the sole nitrogen source. Consequently, the nitrogen-containing bases of the DNA of these cells contained only the heavy nitrogen isotope. Cultures of “heavy” bacteria were washed free of the old medium and incubated in fresh medium with light, 14N-containing compounds, and samples were removed at increasing intervals over a period of several generations. DNA was extracted from the samples of bacteria and subjected to equilibrium density-gradient centrifugation (see Figure 18.35). In this procedure, the DNA is mixed with a concentrated solution of cesium chloride and centrifuged until the double-stranded DNA molecules reach equilibrium according to their density.

I

Generations

Light 14N DNA (a) Hybrid 14N15N DNA

II

0 (parental) Heavy 15N DNA

III

0.3

Semiconservative

0.7

I

(b)

1.0 II

1.1 III

1.5

Conservative

1.9

I (c)

2.5 II

3.0 III Dispersive

4.1

(a)

(b)

FIGURE 13.3 Experiment demonstrating that DNA replication in bac-

on the ratio of 15N/14N that is present in their nucleotides. The greater the 14N content, the higher in the tube the DNA fragment is found at equilibrium. (a) The results expected in this type of experiment for each of the three possible schemes of replication. The single tube on the left indicates the position of the parental DNA and the positions at which totally light or hybrid DNA fragments would band. (b) Experimental results obtained by Meselson and Stahl. The appearance of a hybrid band and the disappearance of the heavy band after one generation eliminates conservative replication. The subsequent appearance of two bands, one light and one hybrid, eliminates the dispersive scheme. (B: FROM

teria is semiconservative. DNA was extracted from bacteria at different stages in the experiment, mixed with a concentrated solution of the salt cesium chloride (CsCl), placed into a centrifuge tube, and centrifuged to equilibrium at high speed in an ultracentrifuge. Cesium ions have sufficient atomic mass to be affected by the centrifugal force, and they form a density gradient during the centrifugation period with the lowest concentration (lowest density) of Cs at the top of the tube and the greatest concentration (highest density) at the bottom of the tube. During centrifugation, DNA fragments within the tube become localized at a position having a density equal to their own density, which in turn depends

M. MESELSON AND F. STAHL, PROC. NAT ’L. ACAD. SCI. U.S.A. 44:671, 1958.)

JWCL151_ch13_533-559.qxd

536

7/29/09

10:09 PM

Page 536


In the Meselson-Stahl experiment, the density of a DNA molecule is directly proportional to the percentage of 15N or 14 N atoms it contains. If replication is semiconservative, one would expect that the density of DNA molecules would decrease during culture in the 14N-containing medium in the manner shown in the upper set of centrifuge tubes of Figure 13.3a. After one generation, all DNA molecules would be 15N-14N hybrids, and their buoyant density would be halfway between that expected for totally heavy and totally light DNA (Figure 13.3a). As replication continued beyond the first generation, the newly synthesized strands would continue to contain only light isotopes, and two types of duplexes would appear in the gradients: those containing 15N–14N hybrids and those containing only 14N. As the time of growth in the light medium continued, a greater and greater percentage of the DNA molecules present would be light. However, as long as replication continued semiconservatively, the original heavy

Chromosome

DNA strand

parental strands would remain intact and present in hybrid DNA molecules that occupied a smaller and smaller percentage of the total DNA (Figure 13.3a). The results of the density-gradient experiments obtained by Meselson and Stahl are shown in Figure 13.13b, and they demonstrate unequivocally that replication occurs semiconservatively. The results that would have been obtained if replication occurred by conservative or dispersive mechanisms are indicated in the two lower sets of centrifuge tubes of Figure 13.3a.1 By 1960, replication had been demonstrated to occur semiconservatively in eukaryotes as well. The original experiments were carried out by J. Herbert Taylor of Columbia University. The drawing and photograph of Figure 13.4 show the results of a more recent experiment in which cultured mammalian cells were allowed to undergo replication in bromodeoxyuridine (BrdU), a compound that is incorporated into DNA in place of thymidine. Following replication, a chromosome is made up of two chromatids. After one round of replication in BrdU, both chromatids of each chromosome contained BrdU (Figure 13.4a). After two rounds of replication in BrdU, one chromatid 1

Anyone looking to explore the circumstances leading up to this heralded research effort and examine the experimental twists and turns as they unfolded might want to read the book Meselson, Stahl, and the Replication of DNA by Frederick Lawrence Holmes, 2001. A discussion of the experiment can also be found in PNAS 101:17889, 2004, which is on the Web.

Chromosome contains only thymidine Replicates in BrdU

Chromatid

Both chromatids contain one strand with BrdU and one strand with thymidine Continued replication in BrdU-containing medium

One chromatid of each chromosome contains thymidine (a)

(b)

FIGURE 13.4 Experimental demonstration that DNA replication occurs semiconservatively in eukaryotic cells. (a) Schematic diagram of the results of an experiment in which cells were transferred from a medium containing thymidine to one containing bromodeoxyuridine (BrdU) and allowed to complete two successive rounds of replication. DNA strands containing BrdU are shown in red. (b) The results of an experiment similar to that shown in a. In this experiment, cultured mammalian cells were grown in BrdU for two rounds of replication before mitotic chromosomes were prepared and stained by a procedure using fluorescent dyes and

Giemsa stain. Using this procedure, chromatids containing thymidine within one or both strands stain darkly, whereas chromatids containing only BrdU stain lightly. The photograph indicates that, after two rounds of replication in BrdU, one chromatid of each duplicated chromosome contains only BrdU, while the other chromatid contains a strand of thymidine-labeled DNA. (Some of the chromosomes are seen to have exchanged homologous portions between sister chromatids. This process of sister chromatid exchange is common during mitosis but is not discussed in the text.) (B: COURTESY OF SHELDON WOLFF.)

JWCL151_ch13_533-559.qxd

7/30/09

9:49 PM

Page 537


of each chromosome was composed of two BrdU-containing strands, whereas the other chromatid was a hybrid consisting of a BrdU-containing strand and a thymidine-containing strand (Figure 13.4a,b). The thymidine-containing strand had been part of the original parental DNA molecule prior to addition of BrdU to the culture.

Replication in Bacterial Cells We will focus in this section of the chapter on replication in bacterial cells, which is better understood than the corresponding process in eukaryotes. The early progress in bacterial research was driven by genetic and biochemical approaches including: ■

■

The availability of mutants that cannot synthesize one or another protein required for the replication process. The isolation of mutants unable to replicate their chromosome may seem paradoxical: how can cells with a defect in this vital process be cultured? This paradox was solved by the isolation of temperature-sensitive (ts) mutants, in which the deficiency only reveals itself at an elevated temperature, termed the nonpermissive (or restrictive) temperature. When grown at the lower (permissive) temperature, the mutant protein can function sufficiently well to carry out its required activity, and the cells can continue to grow and divide. Temperature-sensitive mutants have been isolated that affect virtually every type of physiologic activity (see also page 269), and they have been particularly important in the study of DNA synthesis as it occurs in replication, DNA repair, and genetic recombination. The development of in vitro systems in which replication can be studied using purified cellular components. In some studies, the DNA molecule to be replicated is incubated with cellular extracts from which specific proteins suspected of being essential have been removed. In other studies, the DNA is incubated with a variety of purified proteins whose activity is to be tested.

Taken together, these approaches have revealed the activity of more than 30 different proteins that are required to replicate the chromosome of E. coli. In the following pages, we will discuss the activities of several of these proteins whose functions have been clearly defined. Replication in bacteria and eukaryotes occurs by very similar mechanisms, and thus most of the information presented in the discussion of bacterial replication applies to eukaryotic cells as well. Replication begins at a specific site on the bacterial chromosome called the origin. The origin of replication on the E. coli chromosome is a specific sequence called oriC where a number of proteins bind to initiate the process of replication.2 Once initiated, replication proceeds outward from the origin in both directions, that is, bidirectionally (Figure 13.5). The sites in Figure 13.5 where the pair of replicated segments come toReplication Forks and Bidirectional Replication

2

The subject of initiation of replication is discussed in detail on page 547 as it occurs in eukaryotes.

537

Replication fork

Origin

Daughter strand Parental strand Replication fork

FIGURE 13.5 Model of a circular bacterial chromosome undergoing bidirectional, semiconservative replication. Two replication forks move in opposite directions from a single origin. When the replication forks meet at the opposite point on the circle, replication is terminated, and the two replicated duplexes detach from one another. New DNA strands are shown in red.

gether and join the nonreplicated DNA are termed replication forks. Each replication fork corresponds to a site where (1) the parental double helix is undergoing strand separation, and (2) nucleotides are being incorporated into the newly synthesized complementary strands. The two replication forks move in opposite directions until they meet at a point across the circle from the origin, where replication is terminated. The two newly replicated duplexes detach from one another and are ultimately directed into two different cells. Separation of the strands of a circular, helical DNA duplex poses major topological problems. To visualize the difficulties, we can briefly consider an analogy between a DNA duplex and a two-stranded helical rope. Consider what would happen if you placed a linear piece of this rope on the ground, took hold

Unwinding the Duplex and Separating the Strands

JWCL151_ch13_533-559.qxd

538

7/29/09

10:09 PM

Page 538


of the two strands at one end, and began to pull the strands apart just as DNA is pulled apart during replication. It is apparent that separation of the strands of a double helix is also a process of unwinding the structure. In the case of a rope, which is free to rotate around its axis, separation of the strands at one end would be accompanied by rotation of the entire fiber as it resisted the development of tension. Now, consider what would happen if the other end of the rope were attached to a hook on a wall (Figure 13.6a). Under these circumstances, separation of the two strands at the free end would generate increasing torsional stress in the rope and cause the unseparated portion to become more tightly wound. Separation of the two strands of a circular DNA molecule (or a linear molecule that is not free to rotate, as is the case in a large eukaryotic chromosome) is analogous to having one end of a linear molecule attached to a wall; in all of these cases, tension that develops in the molecule cannot be relieved by rotation of the entire molecule. Unlike a rope, which can become tightly overwound (as in Figure 13.6a), an overwound DNA molecule becomes positively supercoiled (page 391). Consequently, movement of the replication fork generates positive supercoils in the unreplicated portion of the DNA ahead of the fork (Figure 13.6b). When one considers that a complete circular chromosome of E. coli contains approximately 400,000 turns and is replicated by two forks within 40 minutes, the magnitude of the problem becomes apparent.

(a) Point of attachment of DNA

Replication machinery (b)

FIGURE 13.6 The unwinding problem. (a) The effect of unwinding a two-stranded rope that has one end attached to a hook. The unseparated portion becomes more tightly wound. (b) When a circular or attached DNA molecule is replicated, the DNA ahead of the replication machinery becomes overwound and accumulates positive supercoils. Cells possess topoisomerases, such as the E. coli DNA gyrase, that remove positive supercoils. (B: REPRINTED WITH PERMISSION FROM J. C. WANG, NATURE REVIEWS MOL. CELL BIOL. 3:434, 2002; © LAN MAGAZINES LIMITED.)

COPYRIGHT

2002, BY MACMIL-

It was noted on page 391 that cells contain enzymes, called topoisomerases, that can change the state of supercoiling in a DNA molecule. One enzyme of this type, called DNA gyrase, a type II topoisomerase, relieves the mechanical strain that builds up during replication in E. coli. DNA gyrase molecules travel along the DNA ahead of the replication fork, removing positive supercoils. DNA gyrase accomplishes this feat by cleaving both strands of the DNA duplex, passing a segment of DNA through the double-stranded break to the other side, and then sealing the cuts, a process that is driven by the energy released during ATP hydrolysis (shown in detail in Figure 10.14b). Eukaryotic cells possess similar enzymes that carry out this required function. We begin our discussion of the mechanism of DNA replication by describing some of the properties of DNA polymerases, the enzymes that synthesize new DNA strands. Study of these enzymes was begun in the 1950s by Arthur Kornberg at Washington University. In their initial experiments, Kornberg and his colleagues purified an enzyme from bacterial extracts that incorporated radioactively labeled DNA precursors into an acid-insoluble polymer identified as DNA. The enzyme was named DNA polymerase (and later, after the discovery of additional DNA-polymerizing enzymes, it was named DNA polymerase I). For the reaction to proceed, the enzyme required the presence of DNA and all four deoxyribonucleoside triphosphates (dTTP, dATP, dCTP, and dGTP). The newly synthesized, radioactively labeled DNA had the same base composition as the original unlabeled DNA, which strongly suggested that the original DNA strands had served as templates for the polymerization reaction. As additional properties of the DNA polymerase were uncovered, it became apparent that replication was more complex than previously thought. When various types of template DNAs were tested, it was found that the template DNA had to meet certain structural requirements if it was to promote the incorporation of labeled precursors (Figure 13.7). An intact, double-stranded DNA molecule, for example, did not stimulate incorporation. This was not surprising considering the requirement that the strands of the helix must be separated for replication to occur. It was less obvious why a single-stranded, circular molecule was also devoid of activity; one might expect this structure to be an ideal template to direct the manufacture of a complementary strand. In contrast, addition of a partially double-stranded DNA molecule to the reaction mixture produced an immediate incorporation of nucleotides. It was soon discovered that a single-stranded DNA circle cannot serve as a template for DNA polymerase because the enzyme cannot initiate the formation of a DNA strand. Rather, it can only add nucleotides to the 3⬘ hydroxyl terminus of an existing strand. The strand that provides the necessary 3⬘ OH terminus is called a primer. All DNA polymerases—both prokaryotic and eukaryotic—have these same two basic requirements (Figure 13.8a): a template DNA strand to copy and a primer strand to which nucleotides can be added. These requirements explain why certain DNA structures fail to promote DNA synthesis (Figure 13.7a). An intact, linear double helix provides the 3⬘ hydroxyl terminus but lacks a template. A The Properties of DNA Polymerases

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 539


circular single strand, on the other hand, provides a template but lacks a primer. The partially double-stranded molecule (Figure 13.7b) satisfies both requirements and thus promotes nucleotide incorporation. The finding that DNA polymerase cannot initiate the synthesis of a DNA strand raises a critical question: how is the synthesis of a new strand initiated in the cell? We will return to this question shortly. The DNA polymerase purified by Kornberg had another property that was difficult to understand in terms of its presumed role as a replicating enzyme: it only synthesized DNA in a 5⬘-to-3⬘ (written 5⬘ → 3⬘) direction. The diagram first presented by Watson and Crick (see Figure 13.1) depicted events as they would be expected to occur at the replication fork. The diagram suggested that one of the newly synthesized strands is polymerized in a 5⬘ → 3⬘ direction, while the other strand is polymerized in a 3⬘ → 5⬘ direction. Is there some other enzyme responsible for the construction of the 3⬘ → 5⬘ strand? Does the enzyme work differently in the cell than under in vitro conditions? We will return to this question as well. During the 1960s, there were hints that the “Kornberg enzyme” was not the only DNA polymerase in a bacterial cell. Then in 1969, a mutant strain of E. coli was isolated that had less than 1 percent of the normal activity of the enzyme, yet was able to multiply at the normal rate. Further studies revealed that the Kornberg enzyme, or DNA polymerase I, was only one of several distinct DNA polymerases present in bacterial cells. The major enzyme responsible for DNA replication (i.e., the replicative polymerase) is DNA polymerase III. A typical bacterial cell contains 300 to 400 molecules of DNA polymerase I but

5' 3'

3'

5' 3'

5'

3' 5'

5' 3'

5' 3' 3'

5' 3'

5'

3'

(b)

FIGURE 13.7 Templates and nontemplates for DNA polymerase activity. (a) Examples of DNA structures that do not stimulate the synthesis of DNA in vitro by DNA polymerase isolated from E. coli. (b) Examples of DNA structures that stimulate the synthesis of DNA in vitro. In all cases, the molecules in b contain a template strand to copy and a primer strand with a 3⬘ OH on which to add nucleotides.

5'

3'

Primer P

P

Primer 5' O

Base

Base

Base

Base

Base

Base

Template 3'

O

(a)

539

O

5'

O

O

G

C

S

OHO

P

γP O

A

OH

T

OO P O

β

O

NH2

O

OP O O

α

OH

N N

N

N O

Base

Base

P

Base Base

5'

(a)

A

T

HN

N

S

3' Growing DNA strand

O CH3 O P

O

Mg2+

Mg2+

Base

DNA template

Base

(b)

A FIGURE 13.8 The activity of a DNA polymerase. (a) The poly-

merization of a nucleotide onto the 3⬘ end of the primer strand. The enzyme selects nucleotides for incorporation based on their ability to pair with the nucleotide of the template strand. (b) A simplified model of the two-metal ion mechanism for the reaction in which nucleotides are incorporated into a growing DNA strand by a DNA polymerase. In this model, one of the magnesium ions draws the proton away

_

O

3' 5'

O

O

P

5'

DNA polymerase New DNA strands under construction

5' 3'

O

O-

O

S OH .. P P P S

3'

O

P

S

O

T

O

A

O

S

5'

(c)

from the 3⬘ hydroxyl group of the terminal nucleotide of the primer, facilitating the nucleophilic attack of the negatively charged 3⬘ oxygen atom on the ␣ phosphate of the incoming nucleoside triphosphate. The second magnesium ion promotes the release of the pyrophosphate. The two metal ions are bound to the enzyme by highly conserved aspartic acid residues of the active site. (c) Schematic diagram showing the direction of movement of each polymerase along the two template strands.

JWCL151_ch13_533-559.qxd

540

7/29/09

10:09 PM

Page 540


only about 10 copies of DNA polymerase III. The presence of DNA polymerase III had been masked by the much greater amounts of DNA polymerase I in the cell. But the discovery of other DNA polymerases did not answer the two basic questions posed above; none of the enzymes can initiate DNA chains, nor can any of them construct strands in a 3⬘ → 5⬘ direction. The lack of polymerization activity in the 3⬘ → 5⬘ direction has a straightforward explanation: DNA strands cannot be synthesized in that direction. Rather, both newly synthesized strands are assembled in a 5⬘ → 3⬘ direction. During the polymerization reaction, the —OH group at the 3⬘ end of the primer carries out a nucleophilic attack on the 5⬘ ␣-phosphate of the incoming nucleoside triphosphate, as shown in Figure 13.8b. The polymerase molecules responsible for construction of the two new strands of DNA both move in a 3⬘-to-5⬘ direction along the template, and both construct a chain that grows from its 5⬘-P terminus (Figure 13.8c). Consequently, one of the newly synthesized strands grows toward the replication fork where the parental DNA strands are being separated, while the other strand grows away from the fork. Although this solves the problem concerning an enzyme that synthesizes a strand in only one direction, it creates an even more complicated dilemma. It is apparent that the strand that Semidiscontinuous Replication

grows toward the fork in Figure 13.8c can be constructed by the continuous addition of nucleotides to its 3⬘ end. But how is the other strand synthesized? Evidence was soon gathered to indicate that the strand that grows away from the replication fork is synthesized discontinuously, that is, as fragments (Figure 13.9). Before the synthesis of a fragment can be initiated, a suitable stretch of template must be exposed by movement of the replication fork. Once initiated, each fragment grows away from the replication fork toward the 5⬘ end of a previously synthesized fragment to which it is subsequently linked. Thus, the two newly synthesized strands of the daughter duplexes are synthesized by very different processes. The strand that is synthesized continuously is called the leading strand because its synthesis continues as the replication fork advances. The strand that is synthesized discontinuously is called the lagging strand because initiation of each fragment must wait for the parental strands to separate and expose additional template (Figure 13.9). As discussed on page 542, both strands are probably synthesized simultaneously, so that the terms leading and lagging may not be as appropriate as thought when they were first coined. Because one strand is synthesized continuously and the other discontinuously, replication is said to be semidiscontinuous. The discovery that one strand was synthesized as small fragments was made by Reiji Okazaki of Nagoya University, Japan, following various types of labeling experiments. Okazaki found

3' 5' Leading strand template Leading strand Replication fork Lagging strand template 3' 5'

Lagging strand

3' 5'

3' 5' (a)

(b)

A FIGURE 13.9 The two strands of a double helix are synthesized

by a different sequence of events. DNA polymerase molecules move along a template only in a 3⬘ → 5⬘ direction. As a result, the two newly assembled strands grow in opposite directions, one growing toward the replication fork and the other growing away from it. One strand is assembled in continuous fashion, the other as fragments that are joined together enzymatically. (a) Schematic diagram depicting the differences in synthesis of the two strands. (b) Electron micrograph of a

replicating bacteriophage DNA molecule. The left two limbs are the replicated duplexes, and the right end is the unreplicated duplex. The lagging strand of the newly replicated DNA is seen to contain an exposed, single-stranded (thinner) portion, which runs from the replication fork to the arrow. (B: FROM J. WOLFSON AND DAVID DRESSLER, ANNUAL REVIEW OF MICROBIOL29; © 1975, BY ANNUAL REVIEWS, INC.)

REPRINTED WITH PERMISSION FROM THE OGY, VOLUME

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 541

541


that if bacteria were incubated in [3H]thymidine for a few seconds and immediately killed, most of the radioactivity could be found as part of small DNA fragments 1000 to 2000 nucleotides in length. In contrast, if cells were incubated in the labeled DNA precursor for a minute or two, most of the incorporated radioactivity became part of much larger DNA molecules (Figure 13.10). These results indicated that a portion of the DNA was constructed in small segments (later called Okazaki fragments) that were rapidly linked to longer pieces that had been synthesized previously. The enzyme that joins the Okazaki fragments into a continuous strand is called DNA ligase. The discovery that the lagging strand is synthesized in pieces raised a new set of perplexing questions about the initiation of DNA synthesis. How does the synthesis of each of these fragments begin when none of the DNA polymerases are capable of strand initiation? Further studies revealed that initiation is not accomplished by a DNA polymerase but, rather, by a distinct type of RNA polymerase, called primase, that conSedimentation velocity 20

40

60 S

120 sec

structs a short primer composed of RNA, not DNA. The leading strand, whose synthesis begins at the origin of replication, is also initiated by a primase molecule. The short RNAs synthesized by the primase at the 5⬘ end of the leading strand and the 5⬘ end of each Okazaki fragment serve as the required primer for the synthesis of DNA by a DNA polymerase. The RNA primers are subsequently removed, and the resulting gaps in the strand are filled with DNA and then sealed by DNA ligase. These events are illustrated schematically in Figure 13.11. The formation of transient RNA primers during the process of DNA replication is a curious activity. It is thought that the likelihood of mistakes is greater during initiation than during elongation, and the use of a short removable segment of RNA avoids the inclusion of mismatched bases. Replication involves more than incorporating nucleotides. Unwinding the duplex and separating the strands require the aid of two types of proteins that bind to the DNA, a helicase (or DNA unwinding enzyme) and single-stranded DNAbinding (SSB) proteins. DNA helicases unwind a DNA duplex in a reaction that uses energy released by ATP hydrolysis to move along one of the DNA strands, breaking the hydrogen bonds that hold the two strands together and exposing the single-stranded DNA templates. E. coli has at least 12 different helicases for use in various aspects of DNA (and RNA)

The Machinery Operating at the Replication Fork

Radioactivity (103 cts/min per 0.1 ml)

3' 5' Leading strand 60 sec

5' 3'

Lagging strand 1 Primer synthesis

3' 5'

by primase

2

30 sec

Elongation by DNA polymerase III

15 sec 7 sec 2 sec 1

P OH

2

3

Distance from top

3 Primer removal and

gap filling by DNA polymerase I

FIGURE 13.10 Results of an experiment showing that part of the DNA is synthesized as small fragments. Sucrose density gradient profiles of DNA from a culture of phage-infected E. coli cells. The cells were labeled for increasing amounts of time, and the sedimentation velocity of the labeled DNA was determined. When DNA was prepared after very short pulses, a significant percentage of the radioactivity appeared in very short pieces of DNA (represented by the peak near the top of the tube on the left). After periods of 60–120 seconds, the relative height of this peak falls as labeled DNA fragments become joined to the ends of high-molecular-weight molecules. (FROM R. OKAZAKI ET AL., COLD SPRING HARBOR SYMP. QUANT. BIOL. 33:130, 1968.)

4 Strand sealed

by DNA ligase

FIGURE 13.11 The use of short RNA fragments as removable primers in initiating synthesis of each Okazaki fragment of the lagging strand. The major steps are indicated in the drawing and discussed in the text. The role of various accessory proteins in these activities is indicated in the following figures.

JWCL151_ch13_533-559.qxd

542

7/29/09

10:09 PM

Page 542

Chapter 13 DNA REPLICATION AND REPAIR Primase DNA helicase

5'

3'

Lagging strand RNA Primer

5'

Movement of helicase Single-stranded DNA binding protein (SSB)

DNA Duplex

Leading strand 3'

(a)

A

DNA Helicase

FIGURE 13.12 The role of the DNA helicase, single-stranded

DNA-binding proteins, and primase at the replication fork. (a) The helicase moves along the DNA, catalyzing the ATP-driven unwinding of the duplex. As the DNA is unwound, the strands are prevented from reforming the duplex by single-stranded DNA-binding proteins (SSBs). The primase associated with the helicase synthesizes the RNA primers that begin each Okazaki fragment. The RNA primers, which are about 10 nucleotides long, are subsequently removed. (b) A

metabolism. One of these helicases—the product of the dnaB gene—serves as the major unwinding machine during replication. The DnaB helicase consists of six subunits arranged to form a ring-shaped protein that encircles a single DNA strand (Figure 13.12a). The DnaB helicase is first loaded onto the DNA at the origin of replication (with the help of the protein DnaC) and translocates in a 5⬘ → 3⬘ direction along the lagging-strand template, unwinding the helix as it proceeds (Figure 13.12). A three-dimensional model of a similar shaped bacteriophage helicase engaged in strand separation during replication is depicted on page 533. DNA unwinding by the helicase is aided by the attachment of SSB proteins to the separated DNA strands (Figure 13.12). These proteins bind selectively to single-stranded DNA, keeping it in an extended state and preventing it from becoming rewound or damaged. A visual portrait of the combined action of a DNA helicase and SSB proteins on the structure of the DNA double helix is illustrated in the electron micrographs of Figure 13.12b. Recall that an enzyme called primase initiates the synthesis of each Okazaki fragment. In bacteria, the primase and the helicase associate transiently to form what is called a “primosome.” Of the two members of the primosome, the helicase moves along the lagging-strand template processively (i.e., without being released from the template strand during the lifetime of the replication fork). As the helicase “motors” along the lagging-strand template, opening the strands of the duplex, the primase periodically binds to the helicase and synthesizes the short RNA primers that begin the formation of each Okazaki fragment. As noted above, the RNA primers are subsequently extended as DNA by a DNA polymerase, specifically DNA polymerase III. A body of evidence suggests that the same DNA polymerase III molecule synthesizes successive fragments of the lagging strand. To accomplish this, the polymerase III molecule is recycled from the site where it has just completed one Okazaki fragment to the next site along the lagging-strand template closer to the site of DNA unwinding. Once at the new site, the

(b)

Unwound DNA Strand with SSB Proteins

200 nm

series of five electron micrographs showing DNA molecules incubated with a viral DNA helicase (T antigen, page 550) and E. coli SSB proteins. The DNA molecules are progressively unwound from left to right. The helicase appears as the round particle at the fork, and the SSB proteins are bound to the single-stranded ends, giving them a thickened appearance. (B: FROM RAINER WESSEL, JOHANNES SCHWEIZER, AND HANS STAHL, J. VIROL. 66:807, 1992; COPYRIGHT © 1992, AMERICAN SOCIETY FOR MICROBIOLOGY.)

polymerase attaches to the 3⬘ OH of the RNA primer that has just been laid down by a primase and begins to incorporate deoxyribonucleotides onto the end of the short RNA. How does a polymerase III molecule move from one site on the lagging-strand template to another site that is closer to the replication fork? The enzyme does this by “hitching a ride” with the DNA polymerase that is moving in that direction along the leading-strand template. Thus even though the two polymerases are moving in opposite directions with respect to the linear axis of the DNA molecule, they are, in fact, part of a single protein complex (Figure 13.13). The two tethered polymerases can replicate both strands by looping the DNA of the lagging-strand template back on itself, causing this template to have the same orientation as the leading-strand template. Both polymerases then can move together as part of a single replicative complex without violating the “5⬘ → 3⬘ rule” for synthesis of a DNA strand (Figure 13.13). Once the polymerase assembling the lagging strand reaches the 5⬘ end of the Okazaki fragment synthesized during the previous round, the laggingstrand template is released and the polymerase begins work at the 3⬘ end of the next RNA primer toward the fork. The model depicted in Figure 13.13 is often referred to as the “trombone model” because the looping DNA repeatedly grows and shortens during the replication of the lagging strand, reminiscent of the movement of the brass “loop” of a trombone as it is played.

The Structure and Functions of DNA Polymerases DNA polymerase III, the enzyme that synthesizes DNA strands during replication in E. coli, is part of a large “replication machine” called the DNA polymerase III holoenzyme (Figure 13.14). One of the noncatalytic components of the holoenzyme, called the ␤ clamp, keeps the polymerase associated with the DNA template. DNA polymerases (like RNA polymerases) possess two somewhat contrasting properties: (1) they must remain associated with the template over long

JWCL151_ch13_533-559.qxd

7/30/09

9:50 PM

Page 543

13.1 DNA REPLICATION β subunit

DNA polymerase III SSB

Lagging-strand template

Leading-strand template

Unreplicated parental DNA

5'

5' 3'

3'

Leading strand

DNA Helicase 5′ 3′

5' RNA primer #2

(a)

543

RNA primer #1

Growing Okazaki fragment of lagging strand

Lagging strand

5' 3'

5' 3' 3' OH 5' RNA primer RNA primer #2 #3 5'

Polymerase released from template

3' 5'

Completed Okazaki fragment

RNA primer #1 to be replaced with DNA by DNA Polymerase I; nick sealed by DNA ligase

(b)

5'

5' 3'

3'

RNA primer #2 RNA primer #3

5'

(c)

FIGURE 13.13 Replication of the leading and lagging strands in E. coli is accomplished by two DNA polymerases working together as part of a single complex. (a) The two DNA polymerase III molecules travel together, even though they are moving toward the opposite ends of their respective templates. This is accomplished by causing the lagging-strand template to form a loop. (b) The polymerase releases the lagging-strand template when it encounters the previously synthesized Okazaki fragment.

3'

5'

Newly initiated Okazaki fragment

Old Okazaki fragment

(c) The polymerase that was involved in the assembly of the previous Okazaki fragment has now rebound the lagging-strand template farther along its length and is synthesizing DNA onto the end of RNA primer #3 that has just been constructed by the primase. (AFTER D. VOET AND J. G. VOET, BIOCHEMISTRY, 2D ED.; COPYRIGHT © 1995, JOHN WILEY AND SONS, INC. REPRINTED BY PERMISSION OF JOHN WILEY AND SONS, INC.)

β clamp

stretches if they are to synthesize a continuous complementary strand, and (2) they must be attached loosely enough to the template to move from one nucleotide to the next. These contrasting properties are provided by the doughnut-shaped ␤

Leading strand

Core polymerase

τ τ

FIGURE 13.14 Schematic representation of DNA polymerase III holoenzyme. The holoenzyme contains ten different subunits organized into several distinct components. Included as part of the holoenzyme are (1) two core polymerases which replicate the DNA, (2) two or more ␤ clamps, which allow the polymerase to remain associated with the DNA, and (3) a clamp loading (␥) complex, which loads each sliding clamp onto the DNA. The clamp loader of an active replication fork contains two t subunits, which hold the core polymerases in the complex and also bind the helicase. Another term, the replisome, is often used to refer to the entire complex of proteins that are active at the replication fork, including the DNA polymerase III holoenzyme, the helicase, SSBs, and primase. (BASED ON DRAWINGS BY M. O’DONNELL.)

5'

γ-clamp loader (ready to load β clamp for next Okazaki fragment) Lagging strand

DNA helicase

JWCL151_ch13_533-559.qxd

544

7/29/09

10:09 PM

Page 544


clamp that encircles the DNA (Figure 13.15a) and slides freely along it. As long as it is attached to a ␤ “sliding clamp,” a DNA polymerase can move processively from one nucleotide to the next without diffusing away from the template. The polymerase on the leading-strand template remains tethered to a single ␤ clamp during replication. In contrast, when the polymerase on the lagging-strand template completes the synthesis of an Okazaki fragment, it disengages from the ␤ clamp and is cycled to a new ␤ clamp that has been assembled at an RNA primer–DNA template junction located closer to the replication fork (Figure 13.15b). But how does a highly elongated DNA molecule get inside of a ring-shaped clamp as in Figure 13.15a? The assembly of the ␤ clamp around the DNA requires a multisubunit clamp loader that is also part of the DNA polymerase III holoenzyme (Figures 13.14, 13.15c). In the ATP-bound state, the clamp loader binds to a primer-template junction while holding the ␤ clamp in an open conformation as illustrated in Figure 13.15c. Once the DNA has squeezed

through the opening in the clamp wall, the ATP bound to the clamp loader is hydrolyzed, causing the release of the clamp, which closes around the DNA. The ␤ clamp is then ready to bind polymerase III as depicted in Figure 13.15b. DNA polymerase I, which consists of only a single subunit, is involved primarily in DNA repair, a process by which damaged sections of DNA are corrected (page 552). DNA polymerase I also removes the RNA primers at the 5⬘ end of each Okazaki fragment during replication and replaces them with DNA. The enzyme’s ability to accomplish this feat is discussed in the following section. Exonuclease Activities of DNA Polymerases Now that we have explained several of the puzzling properties of DNA polymerase I, such as the enzyme’s inability to initiate strand synthesis, we can consider another curious observation. Kornberg found that DNA polymerase I preparations always contained exonuclease activities; that is, they were able to degrade DNA

5' 3'

5' 3'

β clamp

β clamp

Polymerase III Leading strand template

(a)

FIGURE 13.15 The ␤ sliding clamp and clamp loader. (a) Spacefilling model showing the two subunits that make up the doughnut-shaped ␤ sliding clamp in E. coli. Double-stranded DNA is shown in blue within the ␤ clamp. (b) Schematic diagram of polymerase cycling on the lagging strand. The polymerase is held to the DNA by the ␤ sliding clamp as it moves along the template strand and synthesizes the complementary strand. Following completion of the Okazaki fragment, the enzyme disengages from its ␤ clamp and cycles to a recently assembled clamp “waiting” at an upstream RNA primer–DNA template junction. The original ␤ clamp is left behind for a period on the finished Okazaki fragment, but it is eventually disassembled and reutilized. (c) A model of a complex between a sliding clamp and a clamp loader from an archaean prokaryote based on electron microscopic image analysis. The clamp loader (shown with red and green subunits) is bound to the sliding clamp (blue), which is held in an open, spiral conformation resembling a lock-washer. The DNA has squeezed through the gap in the clamp. The primer strand of the DNA terminates within the clamp loader whereas the template strand extends through an opening at the top of the protein. The clamp loader has been described as a “screw-cap” that fits onto the DNA in such a way that the subdomains of the protein form a spiral that can thread onto the helical DNA backbone. (A: FROM JOHN KURIYAN, CELL, 69:427, 1992; © CELL PRESS; B: AFTER P. T. STUKENBERG,

Lagging strand template

(b)

Previously synthesized Okazaki fragment

(c)

J. TURNER, AND M. O’DONNELL, CELL 78:878, 1994; BY PERMISSION OF CELL PRESS; C: FROM T. MIYATA ET AL., PROC. NAT ’L. ACAD. SCI. U.S.A. 102:13799, 2005; COURTESY OF K. MORIKAWA)

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 545

545


polymers by removing one or more nucleotides from the end of the molecule. At first, Kornberg assumed this activity was due to a contaminating enzyme because the action of exonucleases is so dramatically opposed to that of DNA synthesis. Nonetheless, the exonuclease activity could not be removed from the polymerase preparation and was, in fact, a true activity of the polymerase molecule. It was subsequently shown that all of the bacterial DNA polymerases possess exonuclease activity. Exonucleases can be divided into 5 → 3 and 3 → 5 exonucleases, depending on the direction in which the strand is degraded. DNA polymerase I has both 3 → 5 and 5 → 3 exonuclease activities, in addition to its polymerizing activity (Figure 13.16). These three activities are found in different domains of the single polypeptide. Thus, remarkably, DNA polymerase I is three different enzymes in one. The two exonuclease activities have entirely different roles in replication. We will consider the 5 → 3 exonuclease activity first. Most nucleases are specific for either DNA or RNA, but the 5 → 3 exonuclease of DNA polymerase I can degrade either type of nucleic acid. Initiation of Okazaki fragments by the primase leaves a stretch of RNA at the 5 end of each fragment (see RNA primer #1 of Figure 13.13b), which is removed by the 5 → 3 exonuclease activity of DNA polymerase I (Figure 13.16a). As the enzyme removes ribonu5' 3' Exonuclease hydrolysis site

5'

C

T

X

C 3'

T

C

T

A

C

G

T

A

...

G

G

T

...

G

5'

G

C A

... ... ... ...

A

... ...

A

p

...

... ... ... G

C A

T

A

cleotides of the primer, its polymerase activity simultaneously fills the resulting gap with deoxyribonucleotides. The last deoxyribonucleotide incorporated is subsequently joined covalently to the 5 end of the previously synthesized DNA fragment by DNA ligase. The role of the 3 → 5 exonuclease activity will be apparent in the following section. Ensuring High Fidelity during DNA Replication The survival of an organism depends on the accurate duplication of the genome. A mistake made in the synthesis of a messenger RNA molecule by an RNA polymerase results in the synthesis of defective proteins, but an mRNA molecule is only one shortlived template among a large population of such molecules; therefore, little lasting damage results from the mistake. In contrast, a mistake made during DNA replication results in a permanent mutation and the possible elimination of that cell’s progeny. In E. coli, the chance that an incorrect nucleotide will be incorporated into DNA during replication and remain there is less than 109, or fewer than 1 out of 1 billion nucleotides. Because the genome of E. coli contains approximately 4 106 nucleotide pairs, this error rate corresponds to fewer than 1 nucleotide alteration for every 100 replication cycles. This represents the spontaneous mutation rate in this bacterium. Humans are thought to have a similar spontaneous mutation rate for replication of protein-coding sequences. Incorporation of a particular nucleotide onto the end of a growing strand depends on the incoming nucleoside triphosphate being able to form an acceptable base pair with the nucleotide of the template strand (see Figure 13.8b). Analysis of the distances between atoms and bond angles indicates that A-T and G-C base pairs have nearly identical geometry (i.e., size and shape). Any deviation from those pairings results in a different geometry, as shown in Figure 13.17. At each site along the template, DNA polymerase must discriminate

3'

Single-strand nick

(a)

CH3

O

A

A

C

C

A G

C

T

C

... ... ... ...

T

G

T

... ... ...

T

C G

...

... ... ... G

C

A

T

O

T

H

p

OH

3' 5' Exonuclease hydrolysis site

3'

5 → 3 exonuclease function removes nucleotides from the 5 end of a single-strand nick. This activity plays a key role in removing the RNA primers. (b) The 3 → 5 exonuclease function removes mispaired nucleotides from the 3 end of the growing DNA strand. This activity plays a key role in maintaining the accuracy of DNA synthesis. (FROM D. VOET J. G. VOET, BIOCHEMISTRY, 2D

ED.; COPYRIGHT

O H N

N

© 1995, JOHN WILEY

AND SONS, INC. REPRINTED BY PERMISSION OF JOHN WILEY AND SONS, INC.)

C N C 1'

N C 1' 54˚

10.8 CH3 H

N

H N

O

H N+ A

O T

O

N H

N

N N C 1'

N 46˚ 10.3

H

52˚

N H

68˚

G N

H N

O

11.1

FIGURE 13.16 The exonuclease activities of DNA polymerase I. (a) The

AND

N C 1' C 1'

51˚

5'

(b)

N H

N C

N

N

H

50˚

C

G

G

N A

N H

C 1'

A

T

T

5'

G

H

H N

Mismatched bases 3'

H N

C 1'

O 69˚

H N G

N

N

H N H

C 1' 42˚

10.3

FIGURE 13.17 Geometry of proper and mismatched base pairs.

(FROM H. ECHOLS AND M. F. GOODMAN, REPRODUCED WITH PERMISSION FROM THE ANNUAL REVIEW OF BIOCHEMISTRY, VOLUME 60; © 1991, BY ANNUAL REVIEWS INC.)

JWCL151_ch13_533-559.qxd

546

7/29/09

10:09 PM

Page 546


among four different potential precursors as they move in and out of the active site. Among the four possible incoming nucleoside triphosphates, only one forms a proper geometric fit with the template, producing either an A-T or a G-C base pair that can fit into a binding pocket within the enzyme. This is only the first step in the discrimination process. If the incoming nucleotide is “perceived” by the enzyme as being correct, a conformational change occurs in which the “fingers” of the polymerase rotate toward the “palm” (Figure 13.18a), gripping the incoming nucleotide. This is an example of an induced fit as discussed on page 98. If the newly formed base pair exhibits improper geometry, the active site cannot achieve the conformation required for catalysis and the incorrect nucleotide is not incorporated. In contrast, if the base pair exhibits proper geometry, the incoming nucleotide is covalently linked to the end of the growing strand. On occasion, the polymerase incorporates an incorrect nucleotide, resulting in a mismatched base pair, that is, a base pair other than A-T or G-C. It is estimated that an incorrect pairing of this sort occurs once for every 105–106 nucleotides incorporated, a frequency that is 103–104 times greater than the spontaneous mutation rate of approximately 109. How is the mutation rate kept so low? Part of the answer lies in the second of the two exonuclease activities mentioned above, the 3 → 5 activity (Figure 13.16b). When an incorrect nucleotide is incorporated by DNA polymerase I, the enzyme stalls and the end of the newly synthesized strand has an increased tendency to separate from the template and form a single-stranded 3 terminus. When this occurs, the frayed end of the newly synthesized strand is directed into the 3 → 5 exonuclease site (Figure 13.18), which removes the mismatched nucleotide. This job of “proofreading” is one of the most remarkable of all enzymatic activities and illustrates the sophistication to which biological molecular machinery has evolved. The 3 → 5 exonuclease activity removes approximately 99 out of every 100 mismatched bases, raising the fidelity to about 107–108. In addition, bacteria possess a mechanism called mismatch repair that operates after replication (page 554) and corrects nearly all of the mismatches that escape the proofreading step. Together these processes reduce the overall observed error rate to about 109. Thus the fidelity of DNA replication can be traced to three distinct activities: (1) accurate selection of nucleotides, (2) immediate proofreading, and (3) postreplicative mismatch repair. Another remarkable feature of bacterial replication is its rate. The replication of an entire bacterial chromosome in approximately 40 minutes at 37C requires that each replication fork move about 1000 nucleotides per second, which is equivalent to the length of an entire Okazaki fragment. Thus the entire process of Okazaki fragment synthesis, including formation of an RNA primer, DNA elongation and simultaneous proofreading by the DNA polymerase, excision of the RNA, its replacement with DNA, and strand ligation, occurs within a few seconds. Although it takes E. coli approximately 40 minutes to replicate its DNA, a new round of replication can begin before the previous round has been completed. Consequently, when these bacteria are growing at their maximal rate, they double their numbers in about 20 minutes.

Fingers Palm

Thumb Newly synthesized strand

3' OH

3'

3'

5'

5'

Template strand Mismatched base 3' 5' Exonuclease site (a)

Mismatched base to be removed

(b)

FIGURE 13.18 Activation of the 3 → 5 exonuclease of DNA polymerase I. (a) A schematic model of a portion of DNA polymerase I known as the Klenow fragment, which contains the polymerase and 3 → 5 exonuclease active sites. The 5 → 3 exonuclease activity is located in a different portion of the polypeptide, which is not shown here. The regions of the Klenow fragment are often likened to the shape of a partially opened right hand, hence the portions labeled as “fingers,” “palm,” and “thumb.” The catalytic site for polymerization is located in the central “palm” subdomain. The 3 terminus of the growing strand can be shuttled between the polymerase and exonuclease active sites. Addition of a mismatched base to the end of the growing strand produces a frayed (single-stranded) 3 end that enters the exonuclease site, where it is removed. (The polymerase and exonuclease sites of polymerase III operate similarly but are located on different subunits.) (b) A molecular model of the Klenow fragment complexed to DNA. The template DNA strand being copied is shown in blue, and the primer strand to which the next nucleotides would be added is shown in red. (A: AFTER T. A. BAKER AND S. P. BELL, CELL 92:296, 1998; AFTER A DRAWING BY C. M. JOYCE CELL PRESS; B: COURTESY OF THOMAS A. STEITZ.)

AND T. A. STEITZ, BY PERMISSION OF

Replication in Eukaryotic Cells As noted in Chapter 10, the nucleotide letters of the human genome sequence would fill a book roughly one million pages in length. While it took several years for hundreds of researchers to sequence the human genome, a single cell nucleus of approximately 10 m diameter can copy all of this DNA within a few hours. Given the fact that eukaryotic cells have large genomes

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 547


547

and complex chromosomal structure, our understanding of replication in eukaryotes has lagged behind that in bacteria. This imbalance has been addressed by the development of eukaryotic experimental systems that parallel those used for decades to study bacterial replication. These include ■

■

■

The isolation of mutant yeast and animal cells unable to produce specific gene products required for various aspects of replication. Analysis of the structure and mechanism of action of replication proteins from archaeal species (as in Figure 13.15c). Replication in these prokaryotes begins at multiple origins and requires proteins that are homologous to those of eukaryotic cells but are less complex and easier to study. The development of in vitro systems where replication can occur in cellular extracts or mixtures of purified proteins. The most valuable of these systems has utilized Xenopus, an aquatic frog that begins life as a huge egg stocked with all of the proteins required to carry it through a dozen or so very rapid rounds of cell division. Extracts can be prepared from these frog eggs that will replicate any added DNA, regardless of sequence. Frog egg extracts will also support the replication and mitotic division of mammalian nuclei, which has made this a particularly useful cell-free system. Antibodies can be used to deplete the extracts of particular proteins, and the replication ability of the extract can then be tested in the absence of the affected protein.

Replication in E. coli begins at only one site along the single, circular chromosome (Figure 13.5). Cells of higher organisms may have a thousand times as much DNA as this bacterium, yet their polymerases incorporate nucleotides into DNA at much slower rates. To accommodate these differences, eukaryotic cells replicate their genome in small portions, termed replicons. Each replicon has its own origin from which replication forks proceed outward in both directions (see Figure 13.24a). In a human cell, replication begins at about 10,000 to 100,000 different replication origins. The existence of replicons was first demonstrated in autoradiographic experiments in which single DNA molecules were shown to be replicated simultaneously at several sites along their length (Figure 13.19). Approximately 10 to 15 percent of replicons are actively engaged in replication at any given time during the S phase of the cell cycle (see Figure 14.1). Replicons located close together in a given chromosome tend to undergo replication simultaneously (as evident in Figure 13.19). Moreover, those replicons active at a particular time during one round of DNA synthesis tend to be active at a comparable time in succeeding rounds. In mammalian cells, the timing of replication of a chromosomal region is roughly correlated with the activity of the genes in the region and/or its state of compaction. The presence of acetylated histones, which is closely correlated with gene transcription (page 516), is a likely factor in determining the early replication of active gene loci. The most highly compacted, least acetylated regions of the chromosome are packaged into heterochromatin (page 485), and they are the last regions to be replicated. This difference in timing of replica-

Initiation of Replication in Eukaryotic Cells

FIGURE 13.19 Experimental demonstration that replication in eukaryotic chromosomes begins at many sites along the DNA. Cells were incubated in [3H] thymidine for a brief period before preparation of DNA fibers for autoradiography. The lines of black silver grains indicate sites that had incorporated the radioactive DNA precursor during the labeling period. It is evident that synthesis is occurring at separated sites along the same DNA molecule. As indicated in the accompanying line drawing, initiation begins in the center of each site of thymidine incorporation, forming two replication forks that travel away from each other until they meet a neighboring fork. (MICROGRAPH COURTESY OF JOEL HUBERMAN.)

tion is not related to DNA sequence because the inactive, heterochromatic X chromosome in the cells of female mammals (page 486) is replicated late in S phase, whereas the active, euchromatic X chromosome is replicated at an earlier stage. The mechanism by which replication is initiated in eukaryotes has been a focus of research over the past decade. The greatest progress in this area has been made with budding yeast because the origins of replication can be removed from the yeast chromosome and inserted into bacterial DNA molecules, conferring on them the ability to replicate either within a yeast cell or in cellular extracts containing the required eukaryotic replication proteins. Because these sequences promote replication of the DNA in which they are contained, they are referred to as autonomous replicating sequences (ARSs). Those ARSs that have been isolated and analyzed share several distinct elements. The core element of an ARS consists of a conserved sequence of 11 base pairs, which functions as a specific binding site for an essential multiprotein complex called the origin recognition complex (ORC) (see Figure 13.20). If the ARS is mutated so that it is unable to bind the ORC, initiation of replication cannot occur. Replication origins have proven more difficult to study in vertebrate cells than in yeast. Part of the problem stems from the fact that virtually any type of purified, naked DNA is suitable for replication using extracts from frog eggs. These studies suggested that, unlike yeast, vertebrate DNA does not possess specific sequences (e.g., ARSs) at which replication is initiated. However, studies of replication of intact mammalian chromosomes in vivo suggest that replication does begin within defined regions of the DNA, rather than by random selection as occurs in the amphibian egg extract. It is thought that a DNA molecule contains many sites where DNA replication can be initiated, but only a subset of these sites are actually used at a given time in a given cell. Cells that reproduce via shorter cell cycles, such as those of early amphibian embryos, utilize a greater number of sites as origins of replication than cells with longer cell cycles. The actual selection of sites for initiation of replication is thought to be governed by local epigenetic factors

JWCL151_ch13_533-559.qxd

548

7/30/09

9:51 PM

Page 548


(page 496), such as the positions of nucleosomes, the types of histone modifications, the state of DNA methylation, the degree of supercoiling, and the level of transcription. Restricting Replication to Once Per Cell Cycle It is essential that each portion of the genome is replicated once, and only once, during each cell cycle. Consequently, some mechanism must exist to prevent the reinitiation of replication at a site that has already been duplicated. The initiation of replication at a particular origin requires passage of the origin through several distinct states. Some of the steps that occur at an origin of replication in a yeast cell are illustrated in Figure 13.20. Similar steps requiring homologous proteins take place in plants and animals, suggesting that the basic mechanism of initiation of replication is conserved among eukaryotes. 1. In step 1 (Figure 13.20), the origin of replication is bound

by an ORC protein complex, which in yeast cells remains associated with the origin throughout the cell cycle. The ORC has been described as a “molecular landing pad” because of its role in binding the proteins required in subsequent steps. 2. Proteins referred to as “licensing factors” bind to the ORC (step 2, Figure 13.20) to assemble a protein–DNA complex, called the prereplication complex (pre-RC), that is “licensed” (competent) to initiate replication. Studies of the molecular nature of the licensing factors have focused on a set of six related Mcm proteins (Mcm2–Mcm7). The Mcm proteins are loaded onto the replication origin at a late stage of mitosis, or soon thereafter. Studies indicate that the Mcm2–Mcm7 proteins are capable of associating into a

ring-shaped complex that possesses helicase activity (as in step 4, Figure 13.20). Most evidence suggests that the Mcm2–Mcm7 complex is the eukaryotic replicative helicase; that is, the helicase responsible for unwinding DNA at the replication fork (analogous to DnaB in E. coli). 3. Just before the beginning of S phase of the cell cycle, the activation of key protein kinases leads to the activation of the Mcm2–Mcm7 helicase and the initiation of replication (step 3, Figure 13.20). One of these protein kinases is a cyclin-dependent kinase (Cdk) whose function is discussed at length in Chapter 14. Cdk activity remains high from S phase through mitosis, which suppresses the formation of new prereplication complexes. Consequently,

1

ARS

ORC Licensing factors bind soon after mitosis

Requires Cdc6 and Cdt1 Licensing factors (Mcm2-Mcm7)

Prereplication complex 2

Protein kinases (Cdk and DDK) phosphorylate and activate pre-RC complex at beginning of S phase 3

Initiation of replication

FIGURE 13.20 Steps leading to the replication of a yeast replicon. Yeast origins of replication contain a conserved sequence (ARS) that binds the multisubunit origin recognition complex (ORC) (step 1). The presence of the bound ORC is required for initiation of replication. The ORC is bound to the origin throughout the yeast cell cycle. In step 2, licensing factors (identified as Mcm proteins) bind to the origin during or following mitosis, establishing a prereplication complex that is competent to initiate replication, given the proper stimulus. Loading of Mcm proteins at the origin requires additional proteins (Cdc6 and Cdt1, not shown). In step 3, DNA replication is initiated following the activation of specific protein kinases, including a cyclin-dependent kinase (Cdk). Step 4 shows a stage where replication has proceeded a short distance in both directions from the origin. In this model, the Mcm proteins form a replicative DNA helicase that unwinds DNA of the oppositely directed replication forks. The other proteins required for replication are not shown in this illustration but are indicated in the next figure. In step 5, the two strands of the original duplex have been replicated, an ORC is present at both origins, and the replication proteins, including the Mcm helicases, have been displaced from the DNA. In yeast, the Mcm proteins are exported from the nucleus, and reinitiation of replication cannot occur until the cell has passed through mitosis. [In vertebrate cells, several events appear to prevent reinitiation of replication, including (1) continued Cdk activity from S phase into mitosis, (2) phosphorylation of Cdc6 and its subsequent export from the nucleus, and (3) inactivation of Cdt1 by a bound inhibitor.]

Newly synthesized DNA strand

Mcm2-Mcm7 (helicase?)

4

ORC 5

+

Mcm proteins

Nuclear envelope

Daughter DNAs

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 549


549

TABLE 13.1 Some of the Proteins Required for Replication E. coli protein

Eukaryotic protein

DnaA Gyrase DnaB DnaC SSB -complex pol III core

ORC proteins Topoisomerase I/II Mcm Cdc6, Cdt1 RPA RFC pol /

clamp

PCNA

Primase ——— DNA ligase pol I

Primase pol DNA ligase FEN-1

Function

Recognition of origin of replication Relieves positive supercoils ahead of replication fork DNA helicase that unwinds parental duplex Loads helicase onto DNA Maintains DNA in single-stranded state Subunits of the DNA polymerase holoenzyme that load the clamp onto the DNA Primary replicating enzymes; synthesize entire leading strand and Okazaki fragments; have proofreading capability Ring-shaped subunit of DNA polymerase holoenzyme that clamps replicating polymerase to DNA; works with pol III in E. coli and pol or in eukaryotes Synthesizes RNA primers Synthesizes short DNA oligonucleotides as part of RNA–DNA primer Seals Okazaki fragments into continuous strand Removes RNA primers; pol I of E. coli also fills gap with DNA

each origin can only be activated once per cell cycle. Cessation of Cdk activity at the end of mitosis permits the assembly of a pre-RC for the next cell cycle. 4. Once replication is initiated at the beginning of S phase, the Mcm proteins move with the replication fork (step 4) and are essential for completion of replication of a replicon. The fate of the Mcm proteins after replication depends on the species studied. In yeast, the Mcm proteins are displaced from the chromatin and exported from the nucleus (step 5). In contrast, the Mcm proteins in mammalian cells are displaced from the DNA but apparently remain in the

nucleus. Regardless, Mcm proteins cannot reassociate with an origin of replication that has already “fired.” Overall, the activities that occur at replication forks are quite similar, regardless of the type of genome being replicated—whether viral, bacterial, archaeal, or eukaryotic. The various proteins in the replication “tool kit” of eukaryotic cells are listed in Table 13.1 and depicted in Figure 13.21. All replication systems require helicases, single-stranded DNA-binding proteins, topoisomerases, primase, DNA polymerase, sliding clamp and clamp

The Eukaryotic Replication Fork

PCNA Pol ε RPA

Leading-strand template

DNA Ligase

PCNA

PCNA Pol ε

RFC RNA primer Pol α

RFC

RFC

Helicase (T antigen)

Helicase (T antigen) Pol δ Topoisomerase

Topoisomerase

Primase RPA

FEN-I Pol δ

Lagging-strand template

(a)

FIGURE 13.21 A schematic view of the major components at the eukaryotic replication fork. (a) The proteins required for eukaryotic replication. The viral T antigen is drawn as the replicative helicase in this figure because it is prominently employed in in vitro studies of DNA replication. DNA polymerases and are thought to be the primary DNA synthesizing enzymes of the lagging and leading strands, respectively. PCNA acts as a sliding clamp for both polymerases and . The sliding clamp is loaded onto the DNA by a protein called RFC (replication factor C), which is similar in structure and function to the -clamp loader of E. coli. RPA is a trimeric single-stranded DNA-binding protein comparable in function to that of SSB utilized in E. coli replication.

Primase (b)

The RNA-DNA primers of the lagging strand that are synthesized by the polymerase -primase complex are displaced by the continued movement of polymerase , generating a flap of RNA-DNA that is removed by the FEN-1 endonuclease. The gap is sealed by a DNA ligase. As in E. coli, a topoisomerase is required to remove the positive supercoils that develop ahead of the replication fork. (b) A proposed version of events at the replication fork illustrating how the replicative polymerases on the leading- and lagging-strand templates might act together as part of a replisome. To date there is no firm evidence that the leading and lagging strands are replicated by a single replicative complex as in E. coli.

JWCL151_ch13_533-559.qxd

550

7/30/09

9:51 PM

Page 550


loader, and DNA ligase. When studying the initiation of eukaryotic replication in vitro, researchers often combine mammalian replication proteins with a viral helicase called the large T antigen, which is encoded by the SV40 genome. The large T antigen induces strand separation at the SV40 origin of replication and unwinds the DNA as the replication fork progresses (as in Figure 13.21a). As in bacteria, the DNA of eukaryotic cells is synthesized in a semidiscontinuous manner, although the Okazaki fragments of the lagging strand are considerably smaller than in bacteria, averaging about 150 nucleotides in length. Like DNA polymerase III of E. coli, the eukaryotic replicative DNA polymerase is present as a dimer, suggesting that the leading and lagging strands are synthesized in a coordinate manner by a single replicative complex, or replisome (Figure 13.21b). To date, five “classic” DNA polymerases have been isolated from eukaryotic cells, and they are designated ␣, ␤, ␥, ␦, and ␧. Of these enzymes, polymerase ␥ replicates mitochondrial DNA, and polymerase ␤ functions in DNA repair. The other three polymerases have replicative functions. Polymerase ␣ is tightly associated with the primase, and together they initiate the synthesis of each Okazaki fragment. Primase initiates synthesis by assembly of a short RNA primer, which is then extended by the addition of about 20 deoxyribonucleotides by polymerase ␣. Polymerase ␦ is thought to be the primary DNA-synthesizing enzyme during replication of the lagging strand, whereas polymerase ␧ is thought to be the primary DNA-synthesizing enzyme during replication of the leading strand. Like the major replicating enzyme of E. coli, both polymerase ␦ and ␧ require a “sliding clamp” that tethers the enzyme to the DNA, allowing it to move processively along a template. The sliding clamp of eukaryotic cells is very similar in structure and function to the ␤ clamp of E. coli polymerase III illustrated in Figure 13.14. In eukaryotes, the sliding clamp is called PCNA. The clamp loader that loads PCNA onto the DNA is called RFC and is analogous to the E. coli polymerase III clamp loader complex. After synthesizing an RNA-DNA primer, polymerase ␣ is replaced at each template–primer junction by the PCNA–polymerase ␦ complex, which completes synthesis of the Okazaki fragment. When polymerase ␦ reaches the 5⬘ end of the previously synthesized Okazaki fragment, the polymerase continues along the lagging-strand template, displacing the primer (shown as a green flap in Figure 13.21a). The displaced primer is cut from the newly synthesized DNA strand by an endonuclease (FEN-1) and the resulting nick in the DNA is sealed by a DNA ligase. FEN-1 and DNA ligase are thought to be recruited to the replication fork through an interaction with the PCNA sliding clamp. In fact, PCNA is thought to play a major role in orchestrating events that occur during DNA replication, repair, and recombination. Because of its ability to bind a diverse array of proteins, PCNA has been referred to as a “molecular toolbelt.” Like bacterial polymerases, all of the eukaryotic polymerases elongate DNA strands in the 5⬘ → 3⬘ direction by the addition of nucleotides to a 3⬘ hydroxyl group, and none of them is able to initiate the synthesis of a DNA chain without a primer. Polymerases ␥, ␦, and ␧ possess a 3⬘ → 5⬘ exonuclease, whose proofreading activity ensures that replication oc-

curs with very high accuracy. Several other DNA polymerases (including ␩, ␬ and ␫) have a specialized function that allows cells to replicate damaged DNA as described on page 558. Up to this point in the chapter, the illustrations of replication depict a replicating polymerase moving like a locomotive along a stationary DNA track. But the replication apparatus consists of a huge complex of proteins that operates within the confines of a structured nucleus. Considerable evidence suggests that the replication machinery is present in association with both the nuclear lamina (page 477) and the nuclear matrix (page 499). When cells are given very short pulses of radioactive DNA precursors, over 80 percent of the incorporated label is associated with the nuclear matrix. If, instead of fixing the cells immediately after a pulse, the cells are allowed to incorporate unlabeled DNA precursors for an hour or so before fixation, most of the radioactivity is chased from the matrix into the surrounding DNA loops. This latter finding suggests that rather than remaining stationary, replicating DNA moves like a conveyer belt through an immobilized replication apparatus (Figure 13.22). Further studies suggest that replication forks that are active at a given time are not distributed randomly throughout the cell nucleus, but instead are localized within 50 to 250 sites, called replication foci (Figure 13.23). It is estimated that each of the bright red regions indicated in Figure 13.23 contains approximately 40 replication forks incorporating nucleotides into DNA strands simultaneously. The clustering of replication forks may provide a mechanism for coordinating Replication and Nuclear Structure

3'

5'

3'

5'

3'

5'

FIGURE 13.22 The involvement of the nuclear matrix in DNA replication. The origins of replication are indicated by the black dots, and the arrows indicate the direction of elongation of growing strands. According to this schematic model, it isn’t the replication machinery that moves along stationary DNA tracks, but the DNA that is spooled through the replication apparatus, which is firmly attached to the nuclear matrix.

JWCL151_ch13_533-559.qxd

7/29/09

10:09 PM

Page 551


551

the replication of adjacent replicons on individual chromosomes (as in Figure 13.19). The chromosomes of eukaryotic cells consist of DNA tightly complexed to regular arrays of histone proteins that are present in the form of nucleosomes (page 481). Movement of the replication machinery along the DNA is thought to displace nucleosomes that reside in its path. Yet, examination of a replicating DNA molecule with the electron microscope reveals nucleosomes on both daughter duplexes very near the replication fork (Figure 13.24a), indicating that the reassembly of nucleosomes is a very rapid event. Collectively, the nucleosomes that form during the replication process are comprised of a roughly equivalent mixture of histone molecules that are inherited from parental chromosomes and histone molecules that have been newly synthesized. Recall from page 482 that the core histone octamer of a nucleosome consists of an (H3H4)2 tetramer together with a pair of H2A/H2B dimers. The way in which parental nucleosomes are distributed during replication has been an area of recent debate. According to one line of research, the (H3H4)2 tetramers present prior to replication remain intact and are distributed randomly between the two daughter duplexes. As a result, old and new (H3H4)2 tetramers are thought to be intermixed on each daughter DNA molecule as indicated in the model shown in Figure 13.24b. According to this model, the two H2A/H2B dimers of each parental nucleosome fail to remain together as the replication fork moves through the chromatin. Instead, the H2A-H2B dimers of a nucleosome separate from one another and bind randomly to the new and old (H3H4)2 tetramers already present on the daughter duplexes (Figure 13.24b). According to another viewpoint, the (H3H4)2 tetramer from parental nucleosomes can be split into two H3-H4 dimers, each Chromatin Structure and Replication

FIGURE 13.23 Demonstration that replication activities do not occur randomly throughout the nucleus but are confined to distinct sites. Prior to the onset of DNA synthesis at the start of S phase, various factors required for the initiation of replication are assembled at discrete sites within the nucleus, forming prereplication centers. These sites appear as discrete red objects in the micrograph, which has been stained with a fluorescent antibody against replication factor A (RPA), which is a singlestranded DNA-binding protein required for the initiation of replication. Other replication factors, such as PCNA and the polymerase–primase complex, are also localized to these foci. (FROM YASUHISA ADACHI AND ULRICH K. LAEMMLI, EMBO J. VOL. 13, COVER NO. 17, 1994.)

5' 3'

3' 5' (H3H4)2 Nucleosome

(b)

H2A/H2B H2A/H2B

FIGURE 13.24 The distribution of histone core complexes to daughter strands following replication. (a) Electron micrograph of chromatin isolated from the nucleus of a rapidly cleaving Drosophila embryo showing a pair of replication forks (arrows) moving away from each other in opposite directions. Between the two forks one sees regions of newly replicated DNA that are already covered by nucleosomal core particles to the same approximate density as the parental strands that have not yet undergone replication. (b) Schematic model showing the distribution of core histones after DNA replication. Each nucleosome core particle is shown schematically to be composed of a central (H3H4)2 tetramer flanked by two H2A/H2B dimers. Histones that were present in parental nucleosomes prior to replication are indicated in blue; newly synthesized histones are indicated in red. According to this model, the parental (H3H4)2 tetramers remain intact and are distributed randomly to both daughter duplexes. In contrast, the pairs of H2A/H2B dimers present in parental nucleosomes separate and recombine randomly with the (H3H4)2 tetramers on the daughter duplexes. Other models have been presented in which the parental (H3H4)2 tetramer is split in half by a histone chaperone, and the two resulting H3-H4 dimers are distributed to different DNA strands (discussed in Cell 128:721, 2007 and Trends Cell Biol. 19:29, 2009). (A: COUR(a)

TESY OF

STEVEN L. MCKNIGHT AND OSCAR L. MILLER, JR.)

JWCL151_ch13_533-559.qxd

10:10 PM

Page 552


2.

3. 4.

5.

6.

7.

8. 9.

10.

11.

12. 13.

The original Watson-Crick proposal for DNA replication envisioned the continuous synthesis of DNA strands. How and why has this concept been modified over the intervening years? What does it mean that replication is semiconservative? How was this feature of replication demonstrated in bacterial cells? in eukaryotic cells? Why are there no heavy bands in the top three centrifuge tubes of Figure 13.3a? How is it possible to obtain mutants whose defects lie in genes that are required for an essential activity such as DNA replication? Describe the events that occur at an origin of replication during the initiation of replication in yeast cells. What is meant by replication being bidirectional? Why do the DNA molecules depicted in Figure 13.7a fail to stimulate the polymerization of nucleotides by DNA polymerase I? What are the properties of a DNA molecule that allow it to serve as a template for nucleotide incorporation by DNA polymerase I? Describe the mechanism of action of DNA polymerases operating on the two template strands and the effect this has on the synthesis of the lagging versus the leading strand. Contrast the role of DNA polymerases I and III in bacterial replication. Describe the role of the DNA helicase, the SSBs, the ␤ clamp, the DNA gyrase, and the DNA ligase during replication in bacteria. What is the consequence of having the DNA of the lagging-strand template looped back on itself as in Figure 13.13a? How do the two exonuclease activities of DNA polymerase I differ from one another? What are their respective roles in replication? Describe the factors that contribute to the high fidelity of DNA replication. What is the major difference between bacteria and eukaryotes that allows a eukaryotic cell to replicate its DNA in a reasonable amount of time?

5'

3' OH O

O

O

O

OH

3'

O

1.

?

Life on Earth is subject to a relentless onslaught of destructive forces that originate in both the internal and external environments of an organism. Of all the molecules in a cell, DNA is placed in the most precarious position. On one hand, it is essential that the genetic information remain mostly unchanged as it is passed from cell to cell and individual to individual. On the other hand, DNA is one of the molecules in a cell that is most susceptible to environmental damage. When struck by ionizing radiation, the backbone of a DNA molecule is often broken; when exposed to a variety of reactive chemicals, many of which are produced by a cell’s own metabolism, the bases of a DNA molecule may be altered structurally; when subjected to ultraviolet radiation, adjacent pyrimidines on a DNA strand have a tendency to interact with one another to form a covalent complex, that is, a dimer (Figure 13.25). Even the absorption of thermal energy generated by metabolism is sufficient to split adenine and guanine bases from their attachment to the sugars of the DNA backbone. The magnitude of these spontaneous alterations, or lesions, can be appreciated from the estimate that each cell of a warm-blooded mammal loses approximately 10,000 bases per day! Failure to repair such lesions produces permanent alterations, or mutations, in the DNA. If the mutation occurs in a cell destined to become a gamete, the genetic alteration may be passed on to the next generation. Mutations also have effects in somatic cells (i.e., cells that are not in the germ line): they can interfere with transcription and replication, lead to the malignant transformation of a cell, or speed the process by which an organism ages. Considering the potentially drastic consequences of alterations in DNA molecules and the high frequency at which they occur, it is essential that cells possess mechanisms for repairing DNA damage. In fact, cells have a bewildering arsenal of repair

O

REVIEW

13.2 DNA REPAIR

O

of which may combine with a newly synthesized H3-H4 dimer to form a “mixed” (H3H4)2 tetramer, which then assembles with H2A-H2B dimers. Regardless of the pattern by which it occurs, the stepwise assembly of nucleosomes and their orderly spacing along the DNA is facilitated by a network of accessory proteins. Included among these proteins are a number of histone chaperones that are able to accept either newly synthesized or parental histones and transfer them to the daughter strands. The best studied of these histone chaperones, CAF-1, is recruited to the advancing replication fork through an interaction with the sliding clamp PCNA.

O

552

7/29/09

5'

FIGURE 13.25 A pyrimidine dimer that has formed within a DNA duplex following UV irradiation.

JWCL151_ch13_533-559.qxd

7/29/09

10:10 PM

Page 553

13.2 DNA REPAIR

systems that correct virtually any type of damage to which a DNA molecule is vulnerable. It is estimated that less than one base change in a thousand escapes a cell’s repair systems. The existence of these systems provides an excellent example of the molecular mechanisms that maintain cellular homeostasis. The importance of DNA repair can be appreciated by examining the effects on humans that result from DNA repair deficiencies, a subject discussed in the Human Perspective on page 556. Both prokaryotic and eukaryotic cells possess a variety of proteins that patrol vast stretches of DNA, searching for subtle chemical modifications or distortions of the DNA duplex. In some cases, damage can be repaired directly. Humans, for example, possess enzymes that can directly repair damage from cancer-producing alkylating agents. Most repair systems, however, require that a damaged section of the DNA be excised, that is, selectively removed. One of the great virtues of the DNA duplex is that each strand contains the information required for constructing its partner. Consequently, if one or more nucleotides is removed from one strand, the complementary strand can serve as a template for reconstruction of the duplex. The repair of DNA damage in eukaryotic cells is complicated by the relative inaccessibility of DNA within the folded chromatin fibers of the nucleus. As in the case of transcription, DNA repair involves the participation of chromatin-reshaping machines, such as the histone modifying enzymes and nucleosome remodeling complexes discussed on page 517. Although presumably important in DNA repair, the roles of these proteins will not be considered in the following discussion.

553

on the Web at www.wiley.com/college/karp). Included among the various subunits of TFIIH are two subunits (XPB and XPD) that possess helicase activity; these enzymes separate the two strands of the duplex (step 2, Figure 13.26) in preparation for removal of the lesion. The damaged strand is then cut on both sides of the lesion by a pair of endonucleases (step 3), and the segment of DNA between the incisions is released (step 4). Once excised, the gap is filled by a DNA polymerase (step 5), and the strand is sealed by DNA ligase (step 6). CSB

RNA polymerase

T=T 1

RNA XPC

Transcription-coupled pathway T=T

Global pathway

2 T=T

3

3'OH 5' P

3'OH 5'P T=T

Nucleotide Excision Repair

4 3'OH

5' P

T

T=

Nucleotide excision repair (NER) operates by a cut-andpatch mechanism that removes a variety of bulky lesions, including pyrimidine dimers and nucleotides to which various chemical groups have become attached. Two distinct NER pathways can be distinguished: 1. A transcription-coupled pathway in which the template

strands of genes that are being actively transcribed are preferentially repaired. Repair of a template strand is thought to occur as the DNA is being transcribed, and the presence of the lesion may be signaled by a stalled RNA polymerase. This preferential repair pathway ensures that those genes of greatest importance to the cell, which are the genes the cell is actively transcribing, receive the highest priority on the “repair list.” 2. A slower, less efficient global genomic pathway that corrects DNA strands in the remainder of the genome. Although recognition of the lesion is probably accomplished by different proteins in the two NER pathways (step 1, Figure 13.26), the steps that occur during repair of the lesion are thought to be very similar, as indicated in steps 2–6 of Figure 13.26. One of the key components of the NER repair machinery is TFIIH, a huge protein that also participates in the initiation of transcription. The discovery of the involvement of TFIIH established a crucial link between transcription and DNA repair, two processes that were previously assumed to be independent of one another (discussed in the Experimental Pathways, which can be accessed

5

DNA polymerase δ/ε

3'OH

5' P

6

FIGURE 13.26 Nucleotide excision repair. The following steps are depicted in the drawing and discussed in the text: (1) damage recognition in the global pathway is mediated by an XPC-containing protein complex, whereas damage recognition in the transcription-coupled pathway is thought to be mediated by a stalled RNA polymerase in conjunction with a CSB protein; (2) DNA strand separation (by XPB and XPD proteins, two helicase subunits of TFIIH); (3) incision (by XPG on the 3 side and the XPF–ERCC1 complex on the 5 side); (4) excision, (5) DNA repair synthesis (by DNA polymerase and/or ); and (6) ligation (by DNA ligase I).

JWCL151_ch13_533-559.qxd

554

7/29/09

10:10 PM

Page 554


Base Excision Repair A separate excision repair system operates to remove altered nucleotides generated by reactive chemicals present in the diet or produced by metabolism. The steps in this repair pathway in eukaryotes, which is called base excision repair (BER), are shown in Figure 13.27. BER is initiated by a DNA glycosylase that recognizes the alteration (step 1, Figure 13.27) and removes the base by cleavage of the glycosidic bond holding the base to the deoxyribose sugar (step 2). A number of different DNA glycosylases have been identified, each more-or-less specific for a particular type of altered base, including uracil (formed by the hydrolytic removal of the amino group of cytosine), 8-oxoguanine (caused by damage from oxygen free radicals, page 34), and 3-methyladenine (produced by transfer of a methyl group from a methyl donor, page 431). Structural studies of the DNA glycosylase that removes the highly mutagenic 8-oxoguanine (oxoG) indicate that this enzyme diffuses rapidly along the DNA “inspecting” each of the G-C base pairs within the DNA duplex (Figure 13.28, step 1). In step 2, the enzyme has come across an oxoG-C base pair. When this occurs, the enzyme inserts a specific amino acid side chain into the DNA helix, causing the nucleotide to rotate (“flip”) 180 degrees out of the DNA helix and into the body of the enzyme (step 2). If the nucleotide does, in fact, contain an oxoG, the base fits into the active site of the enzyme (step 3) and is cleaved from its associated sugar. In contrast, if the extruded nucleotide contains a normal guanine, which only differs in structure by two atoms from oxoG, it is unable to fit into the enzyme’s active site (step 4) and it is returned to its appropriate position within the stack of bases. Once an altered purine or pyrimidine is removed by a glycosylase, the “beheaded” deoxyribose phosphate remaining in the site is excised by the combined action of a specialized (AP) endonuclease and a DNA polymerase. AP endonuclease cleaves the DNA backbone (Figure 13.27, step 3) and a phosphodiesterase activity of polymerase ␤ removes the sugar–phosphate remnant that had been attached to the excised base (step 4). Polymerase ␤ then fills the gap by inserting a nucleotide complementary to the undamaged strand (step 5), and the strand is sealed by DNA ligase III (step 6). The fact that cytosine can be converted to uracil may explain why natural selection favored the use of thymine, rather than uracil, as a base in DNA, even though uracil was presumably present in RNA when it served as genetic material during the early evolution of life (page 448). If uracil had been retained as a DNA base, it would have caused difficulty for repair systems to distinguish between a uracil that “belonged” at a particular site and one that resulted from an alteration of cytosine.

5'

It was noted earlier that cells can remove mismatched bases that are incorporated by the DNA polymerase and escape the enzyme’s proofreading exonuclease. This process is called mismatch repair (MMR). A mismatched base pair causes a distortion in the geometry of the double helix that can be recognized by a repair enzyme. But how does the enzyme “recognize”

P

P

P

P

G

U

C

A

C

G

G

T

P

3'

1

3'

P

P

P

P

P

P

5'

Uracil–DNA glycosylase 5'

P

P

P

P

G

P

C

A

G

T

P

3'

2 C 3'

P

G P

P

P

P

P

5'

AP endonuclease OH P 5'

P

P

P

G

P

C

A

G

T

P

3'

3 C 3'

P

G P

P

P

P

P

5'

Phosphodiesterase activity of DNA polymerase β OH 5'

P

P

P

G

P

C

A

G

T

P

3'

4 G

C 3'

P

P

P

P

P

P

5'

Polymerase activity of DNA polymerase β OH 5'

P

P

P

P

P

G

C

C

A

C

G

G

T

P

3'

5

3'

P

P

P

P

P

P

5'

DNA ligase III

5'

Mismatch Repair

P

P

P

P

P

P

G

C

C

A

C

G

G

T

P

3'

6

3'

P

P

P

P

P

P

5'

FIGURE 13.27 Base excision repair. The steps are described in the text. Other pathways for BER are known, and BER also has been shown to have distinct transcription-coupled and global repair pathways.

JWCL151_ch13_533-559.qxd

7/29/09

10:10 PM

Page 555

13.2 DNA REPAIR 1

2

h0GG1 o

G C

G C

3

4

o

G

o

G

C

555

G C

C

G G C

C

C

FIGURE 13.28 Detecting damaged bases during BER. In step 1, a DNA glycosylase (named hOGG1) is inspecting a nucleotide that is paired to a cytosine. In step 2, the nucleotide is flipped out of the DNA duplex. In this case, the base is an oxidized version of guanine, 8-oxoguanine, and it is able to fit into the active site of the enzyme (step 3) where it is cleaved from its attached sugar. The subsequent steps in BER were

shown in Figure 13.27. In step 4, the extruded base is a normal guanine, which is unable to fit into the active site of the glycosylase and is returned to the base stack. Failure to remove oxoG would have resulted in a G-to-T mutation. (BASED ON S. S. DAVID, WITH PERMISSION FROM

which member of the mismatched pair is the incorrect nucleotide? If it were to remove one of the nucleotides at random, it would make the wrong choice 50 percent of the time, creating a permanent mutation at that site. Thus, for a mismatch to be repaired after the DNA polymerase has moved past a site, it is important that the repair system distinguish the newly synthesized strand, which contains the incorrect nucleotide, from the parental strand, which contains the correct nucleotide. In E. coli, the two strands are distinguished by the presence of methylated adenosine residues on the parental strand. DNA methylation does not appear to be utilized by the MMR system in eukaryotes, and the mechanism of identification of the newly synthe-

sized strand remains unclear. Several different MMR pathways have been identified and will not be discussed.

1

Ku 2

DNA-PKcs

NATURE 434:569, 2005; © LIMITED.)

COPYRIGHT

2005,

BY

MACMILLAN MAGAZINES

Double-Strand Breakage Repair X-rays, gamma rays, and particles released by radioactive atoms are all described as ionizing radiation because they generate ions as they pass through matter. Millions of gamma rays pass through our bodies every minute. When these forms of radiation collide with a fragile DNA molecule, they often break both strands of the double helix. Double-strand breaks (DSBs) can also be caused by certain chemicals, including several (e.g., bleomycin) used in cancer chemotherapy, and free radicals produced by normal cellular metabolism (page 34). DSBs are also introduced during replication of damaged DNA. A single double-strand break can cause serious chromosome abnormalities, which can have grave consequences for the cell. DSBs can be repaired by several alternate pathways. The predominant pathway in mammalian cells is called nonhomologous end joining (NHEJ), in which a complex of proteins bind to the broken ends of the DNA duplex and catalyze a series of reactions that rejoin the broken strands. The major steps that occur during NHEJ are shown in Figure 13.29a and described in the accompanying legend. Figure 13.29b shows the nuclei of human fi-

3

DNA Ligase IV

4

(a)

FIGURE 13.29 Repairing double-strand breaks (DSBs) by nonhomologous end joining. (a) In this simplified model of double-strand break repair, the lesion (step 1) is detected by a heterodimeric, ring-shaped protein called Ku, that binds to the broken ends of the DNA (step 2). The DNA-bound Ku recruits another protein, called DNA-PKcs, which is the catalytic subunit of a DNA-dependent protein kinase (step 3). Most of the substrates phosphorylated by this protein kinase have not been identified. These proteins bring the ends of the broken DNA together in such a way that they can be joined by DNA ligase IV to regenerate an intact DNA duplex (step 4). The NHEJ pathway may also

(b)

involve the activities of nucleases and polymerases (not shown) and is more error prone than is the homologous recombination pathway of DSB repair. (b) Time course analysis of Ku localization at sites of DSB formation induced by laser microbeam irradiation at a site indicated by the arrowheads. The NHEJ protein Ku becomes localized at the damage site immediately following irradiation but remains there just briefly as the damage is presumably repaired. Micrographs were taken (1) immediately, (2) 2 hours, and (3) 8 hours after irradiation. (B: FROM JONG-SOO KIM ET AL, COURTESY OF KYOKO YOKOMORI, J. CELL BIOL. 170:344, 2005; BY THE ROCKEFELLER UNIVERSITY PRESS.)

COPYRIGHT PERMISSION OF

JWCL151_ch13_533-559.qxd

556

7/30/09

3:34 PM

Page 556


broblasts that had been treated with a laser to induce a localized cluster of double-strand breaks and then stained for the presence of the protein Ku at various times after laser treatment. This NHEJ repair protein is seen to localize at the site of the DSBs immediately following their appearance. Another DSB repair pathway known as homologous recombination requires a homologous chromosome to serve as a template for repair of the broken strand. The steps that occur during homologous recombination are similar to those of genetic recombination depicted in Figure 14.47. Defects in both repair pathways have been linked to increased cancer susceptibility.

REVIEW

?

Contrast the events of nucleotide excision repair and base excision repair. 2. Why is it important in mismatch repair that the cell distinguish the parental strands from the newly synthesized strands? How is this accomplished? 1.

THE HUMAN PERSPECTIVE The Consequences of DNA Repair Deficiencies We owe our lives to light from the sun, which provides the energy captured during photosynthesis. But the sun also emits a constant stream of ultraviolet rays that ages and mutates the cells of our skin. The hazardous effects of the sun are most dramatically illustrated by the rare recessive genetic disorder, xeroderma pigmentosum (XP). Patients with XP possess a deficient nucleotide excision repair system that cannot remove segments of DNA damaged by ultraviolet radiation. As a result, persons with XP are extremely sensitive to sunlight; even very limited exposure to the direct rays of the sun can produce large numbers of dark-pigmented spots on exposed areas of the body (Figure 1) and a greatly elevated risk of developing disfiguring and fatal skin cancers. Some help for XP patients may be on the way in the form

FIGURE 1 Darkly pigmented regions of the skin are evident in this boy with xeroderma pigmentosum. The area of skin below the chin, which is protected from the sun, is relatively devoid of the lesions. (KEN GREER/ VISUALS UNLIMITED.)

of a skin cream (Dimericine) that contains a bacterial DNA repair enzyme. The enzyme is contained in liposomes that can apparently penetrate the outer layer of the skin and participate in DNA repair. XP is not the only genetic disorder characterized by nucleotide excision repair deficiency. Cockayne syndrome (CS) is an inherited disorder characterized by acute sensitivity to light, neurological dysfunction due to demyelination of neurons, and dwarfism, but no evident increase in the frequency of skin cancer. Cells from persons with CS are deficient in the pathway by which transcriptionally active DNA is repaired (page 553). The remainder of the genome is repaired at the normal rate, presumably accounting for the normal levels of skin cancer. But why are persons with a defective repair mechanism subject to specific abnormalities such as dwarfism? Most cases of CS can be traced to a mutation in one of two genes, either CSA or CSB, which are thought to be involved in coupling transcription to DNA repair (see Figure 13.26). Mutations in these genes, in addition to impacting DNA repair, may also disturb the transcription of certain genes, leading to growth retardation and abnormal development of the nervous system. This possibility is strengthened by the finding that, in rare cases, the symptoms of CS can also occur in persons with XP who carry specific mutations in the XPD gene. As noted on page 553, XPD encodes a subunit of the transcription factor TFIIH required for transcription initiation. Mutations in XPD could lead to defects in both DNA repair and transcription. Certain other mutations in the XPD gene are responsible for another disease, trichothiodystrophy (TTD), which also combines symptoms suggestive of both DNA repair and transcription defects. Like CS patients, individuals with TTD exhibit increased sun sensitivity without the increased risk of development of cancer. TTD patients have additional symptoms, including brittle hair and scaly skin. These findings indicate that three distinct disorders—XP, CS, and TTD—can be caused by defects in a single gene, with the particular disease outcome determined by the specific mutation present in that gene. Structural studies of mutant XPD molecules suggest that these different mutations affect different functions of the protein. Elsewhere in this text, we have described circumstances that lead to premature (or accelerated) aging in humans or animal models: as the result of (1) increased free radicals (page 34), (2) increased mitochondrial DNA mutations (page 202), and (3) mutations in a protein of the nuclear envelope (page 477). In 2006, a 15-year-old boy who suffered from frequent sunburns and certain characteristics of prema-

JWCL151_ch13_533-559.qxd

7/29/09

10:10 PM

Page 557

13.3 BETWEEN REPLICATION AND REPAIR

ture aging came to the attention of clinical researchers. Genetic analysis determined that the boy carried a mutation in the XPF gene whose encoded protein makes one of the cuts during the NER pathway (Figure 13.26). Patients with mild mutations in XPF develop XP and have impaired NER. This individual had a more severe mutation in the XPF gene, causing his cells to be unable to repair covalent cross-links that form occasionally between the two strands of a DNA duplex. Studies on the cells of this individual, and on mice with a corresponding mutation, suggested that the unrepaired cross-links lead to increased cell death (apoptosis), which either directly or indirectly promotes premature aging. According to one hypothesis, defects in DNA repair systems that result primarily in an increased mutation rate in the body’s cells are associated with an increased susceptibility to cancer, whereas defects in DNA repair systems that result primarily in cell death are associated with accelerated aging.a Whether any of these premature-aging syndromes provides insight into the mechanisms of normal aging remains a matter of debate. Persons with DNA-repair disorders are not the only individuals who should worry about exposure to the sun. Even in a skin cell whose repair enzymes are functioning at optimal levels, a small fraction of the lesions fail to be excised and replaced. Alterations in DNA lead to mutations that can cause a cell to become malignant. Thus, one of the consequences of the failure to correct UV-induced damage is the risk of skin cancer. Consider the following statistics: more than one million persons develop one of three forms of skin cancer every year in the United States, and most of these cases are attributed to overexposure to the sun’s ultraviolet rays. Fortunately, the two most common forms of skin cancer—basal cell carcinoma and squamous cell carcinoma—rarely spread to other parts of the body and can usually be excised in a doctor’s office. Both of these types of cancer originate from the skin’s epithelial cells. a

It has not been mentioned in this discussion, for a number of reasons, that two of the genes most often responsible for premature aging syndromes encode members of a particular type of DNA helicase family called RecQ helicases. The genes in question are WRN and BLM which, when mutated, are responsible for the inherited diseases Werner Syndrome and Bloom Syndrome, respectively, which are characterized by both increased cancer risk and features of accelerated aging. It is suggested that these helicases are involved in certain types of base excision and DSB repair pathways. They appear to be particularly important in resolving situations where a replicative DNA polymerase becomes stalled at a lesion and the replication fork “collapses” (disassembles). The subject is discussed in Trends Biochem. Sci. 33:609, 2008.

13.3 BETWEEN REPLICATION AND REPAIR The human perspective describes an inherited disease— xeroderma pigmentosum (XP)—that leaves patients with an inability to repair certain lesions caused by exposure to ultraviolet radiation. Patients described as having the “classical” form of XP have a defect in one of seven different genes involved in nucleotide excision repair (page 553). These genes are designated XPA, XPB, XPC, XPD, XPE, XPF, and XPG, and some of their roles in NER are indicated in the legend of Figure 13.26. Another group of patients were identified that, like those with XP, were highly susceptible to developing skin cancer as the result of sun exposure. However, unlike the cells from XP sufferers, cells from these patients were capable of

557

However, malignant melanoma, the third type of skin cancer, is a potential killer. Unlike the others, melanomas develop from pigment cells (melanocytes) in the skin. The number of cases of melanoma diagnosed in the United States is climbing at the alarming rate of 4 percent per year due to the increasing amount of time people have spent in the sun over the past few decades. Studies suggest that one of the greatest risk factors to developing melanoma as an adult is the occurrence of a severe, blistering sunburn as a child or adolescent. Individuals at greatest risk are Caucasians with extremely light skin. Many of these individuals have pigment cells whose surfaces lack a functioning receptor (called MC1R) for a hormone that is secreted by nearby epithelial cells of the skin in response to ultraviolet radiation. Melanocytes respond to MC1R activation by producing the dark pigment melanin, thereby providing the individual with a tan. Tanned skin is more protected from UV rays than is light, untanned skin, even though it is UV radiation that is responsible for triggering the tanning response. What if it were possible to develop a tanned skin without having to suffer UV exposure? A number of research groups are working on such an approach by using various means other than exposure to UV-containing sunlight to stimulate the tanning response in pigment cells. Whether any of these approaches will prove safe and effective remains to be seen. Skin cancer is not the only disease that is promoted by deficient or overworked DNA repair systems. It is estimated that up to 15 percent of colon cancer cases can be attributed to mutations in the genes that encode the proteins required for mismatch repair. Mutations that cripple the mismatch repair system inevitably lead to a higher mutation rate in other genes because mistakes made during replication are not corrected. Cancer is also one of the consequences of double-strand DNA breaks that have either gone unrepaired or been repaired incorrectly. Breaks in DNA can be caused by a variety of environmental agents to which we are commonly exposed, including X-rays, gamma rays, and radioactive emissions. The most serious environmental hazard in this regard is probably radon (specifically 222Rn), a radioactive isotope formed during the disintegration of uranium. Some areas of the planet contain relatively high levels of uranium in the soil, and houses built in these regions can contain dangerous levels of radon gas. When the gas is breathed into the lungs, it can lead to doublestrand DNA breaks that increase the risk of lung cancer. A significant fraction of lung-cancer deaths in nonsmokers is probably due to radon exposure.

nucleotide excision repair and were only slightly more sensitive to UV light than normal cells. This heightened UV sensitivity revealed itself during replication, as these cells often produced fragmented daughter strands following UV irradiation. Patients in this group were classified as having a variant form of XP, designated XP-V. We will return to the basis of the XP-V defect in a moment. We have seen in the previous section that cells can repair a great variety of DNA lesions. On occasion, however, a DNA lesion is not repaired by the time that segment of DNA is scheduled to undergo replication. On these occasions, the replication machinery arrives at the site of damage on the template strand and becomes stalled there. When this happens, some type of signal is emitted that leads to the recruit-

JWCL151_ch13_533-559.qxd

558

7/29/09

10:10 PM

Page 558


ment of a specialized polymerase that is able to bypass the lesion.3 Suppose the lesion in question is a thymidine dimer (Figure 13.25) in a skin cell that was caused by exposure to UV radiation. When the replicative polymerase (pol ) reaches the obstacle, the enzyme is temporarily replaced by a “specialized” DNA polymerase designated pol , which is able to insert two A residues into the newly synthesized strand across from the two T residues that are covalently linked as part of the dimer. Once this “damage bypass” is accomplished, the cell switches back to the normal replicative polymerase and DNA synthesis continues without leaving any trace that a serious problem had been resolved. As you might have guessed from the juxtaposition of topics, patients afflicted with XP-V have 3

A cell has other options to deal with a stalled replication fork, but they are more complex and poorly understood, and will not be discussed.

a mutation in the gene encoding pol and thus have difficulty replicating past thymidine dimers. Discovered in 1999, polymerase is a member of a family of DNA polymerases in which each member is specialized for incorporating nucleotides opposite particular types of DNA lesions in the template strand. The polymerases of this family are said to engage in translesion synthesis (TLS). X-ray crystallographic studies reveal that the TLS polymerases have an unusually spacious active site that is able to physically accommodate altered nucleotides that would not fit in the active site of a replicative polymerase. These TLS polymerases are only capable of incorporating one to a few nucleotides into a DNA strand (they lack processivity); they have no proofreading capability; and they are much more likely to incorporate an incorrect (i.e., noncomplementary) nucleotide when copying undamaged DNA than the classic polymerases.

SYNOPSIS DNA replication occurs semiconservatively, which indicates that one intact strand of the parent duplex is transmitted to each of the daughter cells during cell division. This mechanism of replication was first suggested by Watson and Crick as part of their model of DNA structure. They suggested that replication occurred by gradual separation of the strands by means of hydrogen bond breakage, so that each strand could serve as a template for the synthesis of a complementary strand. This model was soon confirmed in both bacterial and eukaryotic cells by showing that cells transferred to labeled media for one generation produce daughter cells whose DNA has one labeled strand and one unlabeled strand. (p. 534) The mechanism of replication is best understood in bacterial cells. Replication begins at a single origin on the circular bacterial chromosome and proceeds outward in both directions as a pair of replication forks. Replication forks are sites where the double helix is unwound and nucleotides are incorporated into both newly synthesized strands. (p. 537) DNA synthesis is catalyzed by a family of DNA polymerases. The first of these enzymes to be characterized was DNA polymerase I of E. coli. To catalyze the polymerization reaction, the enzyme requires all four deoxyribonucleoside triphosphates, a template strand to copy, and a primer containing a free 3 OH to which nucleotides can be added. The primer is required because the enzyme cannot initiate the synthesis of a DNA strand. Rather, it can only add nucleotides to the 3 hydroxyl terminus of an existing strand. Another unexpected characteristic of DNA polymerase I is that it only polymerizes a strand in a 5 → 3 direction. It had been presumed that the two new strands would be synthesized in opposite directions by polymerases moving in opposite directions along the two parental template strands. This finding was explained when it was shown that the two strands were synthesized quite differently. (p. 538) One of the newly synthesized strands (the leading strand) grows toward the replication fork and is synthesized continuously. The other newly synthesized strand (the lagging strand) grows away from the fork and is synthesized discontinuously. In bacterial cells, the lagging strand is synthesized as fragments approximately 1000 nucleotides long, called Okazaki fragments, that are covalently joined to one another by a DNA ligase. In contrast, the leading strand is synthesized as a single continuous strand. Neither the con-

tinuous strand nor any of the Okazaki fragments can be initiated by the DNA polymerase but instead begin as a short RNA primer that is synthesized by a type of RNA polymerase called primase. After the RNA primer is assembled, DNA polymerase continues to synthesize the strand or fragment as DNA. The RNA is subsequently degraded, and the gap is filled in as DNA. (p. 540) Events at the replication fork require a variety of different types of proteins having specialized functions. These include a DNA gyrase, which is a type II topoisomerase required to relieve the tension that builds up ahead of the fork as a result of DNA unwinding; a DNA helicase that unwinds the DNA by separating the strands; single-stranded DNA-binding proteins that bind selectively to single-stranded DNA and prevent reassociation; a primase that synthesizes the RNA primers; and a DNA ligase that seals the fragments of the lagging strand into a continuous polynucleotide. DNA polymerase III is the primary DNA-synthesizing enzyme that adds nucleotides to each RNA primer, whereas DNA polymerase I is responsible for removing the RNA primers and replacing them with DNA. Two molecules of DNA polymerase III are thought to move together as a complex along their respective template strands. This is accomplished as the lagging-strand template loops back on itself. (p. 541) DNA polymerases possess separate catalytic sites for polymerization and degradation of nucleic acid strands. Most DNA polymerases possess both 5 → 3 and 3 → 5 exonuclease activities. The first acts to degrade the RNA primers that begin each Okazaki fragment, and the second removes inappropriate nucleotides following their mistaken incorporation, thus contributing to the fidelity of replication. It is estimated that approximately one in 109 nucleotides is incorporated incorrectly during replication in E. coli. (p. 544) Replication in eukaryotic cells follows a similar mechanism and employs similar proteins to those of prokaryotes. All of the DNA polymerases involved in replication elongate DNA strands in the 5 → 3 direction. None of them initiates the synthesis of a chain without a primer. Most possess a 3 → 5 exonuclease activity, ensuring that replication occurs with high fidelity. Unlike bacteria, replication in eukaryotes is initiated simultaneously at many sites along a chromosome, with replication forks proceeding outward in both directions from each site of initiation. Studies on yeast indicate that

JWCL151_ch13_533-559.qxd

7/30/09

3:35 PM

Page 559

ANALYTIC QUESTIONS

origins of replication contain a specific binding site for an essential multiprotein complex called ORC. Events at the origin ensure that replication of each DNA segment occurs once and only once per cell cycle. (p. 546) Replication in eukaryotic cells is intimately associated with nuclear structures. Evidence indicates that much of the machinery required for replication is associated with the nuclear matrix. In addition, replication forks that are active at any given time are localized within about 50 to 250 sites called replication foci. Newly synthesized DNA is rapidly associated with nucleosomes. According to one model, (H3H4)2 tetramers present prior to replication remain intact and are passed on to the daughter duplexes, whereas H2A/H2B dimers separate from one another and bind randomly to new and old (H3H4)2 tetramers on the daughter duplexes. (p. 550) DNA is subject to damage by many environmental influences, including ionizing radiation, common chemicals, and ultraviolet radiation. Cells possess a variety of systems to recognize and repair the resulting damage. It is estimated that less than one base change in a thousand escapes a cell’s repair systems. Four major types of DNA repair systems are discussed. Nucleotide excision repair (NER) systems operate by removing a small section of a DNA strand containing a bulky lesion, such as a pyrimidine dimer. During NER, the strands of DNA containing the lesion are separated by a helicase;

559

paired incisions are made by endonucleases; the gap is filled by a DNA polymerase; and the strand is sealed by a DNA ligase. The template strands of genes that are actively transcribed are preferentially repaired by NER. Base excision repair removes a variety of altered nucleotides that produce minor distortions in the DNA helix. Cells possess a variety of glycosylases that recognize and remove various types of altered bases. Once the base is removed, the remaining portion of the nucleotide is removed by an endonuclease, the gap is enlarged by a phosphodiesterase activity, and the gap is filled and sealed by a polymerase and ligase. Mismatch repair is responsible for removing incorrect nucleotides incorporated during replication that escape the proofreading activity of the polymerase. In bacteria, the newly synthesized strand is selected for repair by virtue of its lack of methyl groups compared to the parental strand. Double-strand breaks are repaired as proteins bind to the broken strands and join the ends together. (p. 552) In addition to the classic DNA polymerases involved in DNA replication and repair, cells also possess an array of DNA polymerases that facilitate replication at sites of DNA lesions or misalignments. These polymerases, which engage in translesion synthesis, lack processivity and proofreading capability and are more error-prone than classic polymerases. (p. 557)

ANALYTIC QUESTIONS 1. Suppose that Meselson and Stahl had carried out their experi-

2.

3.

4.

5. 6.

ment by growing cells in medium with 14N and then transferring the cells to medium containing 15N. How would the bands within the centrifuge tubes have appeared if replication were semiconservative? If replication were conservative? If replication were dispersive? Suppose you isolated a mutant strain of yeast that replicated its DNA more than once per cell cycle. In other words, each gene in the genome was replicated several times between successive cell divisions. How might you explain such a phenomenon? How would the chromosomes from the experiment on eukaryotic cells depicted in Figure 13.4 have appeared if replication occurred by a conservative or a dispersive mechanism? We have seen that cells possess a special enzyme to remove uracil from DNA. What do you suppose would happen if the uracil groups were not removed? (You might consider the information presented in Figure 11.44 on the pairing properties of uracil.) Draw a partially double-stranded DNA molecule that would not serve as a template for DNA synthesis by DNA polymerase I. Some temperature-sensitive bacterial mutants stop replication immediately following elevation of temperature, whereas others continue to replicate their DNA for a period of time before they cease this activity, and still others continue until a round of replication is completed. How might these three types of mutants differ?

7. Suppose the error rate during replication in human cells were

8.

9.

10.

11. 12.

13.

the same as that of bacteria (about 10⫺9). How would this impact the two cells differently? Figure 13.19 shows the results from an experiment in which cells were incubated with [3H]thymidine for less than 30 minutes prior to fixation. How would you expect this photograph to appear after a one-hour labeling period? Can you conclude that the entire genome is replicated within an hour? If not, why not? Origins of replication tend to have a region that is very rich in A-T base pairs. What function do you suppose these sections might serve? What advantages might you expect for DNA replication to occur in conjunction with the nuclear matrix as opposed to the nucleoplasm? What are the advantages of replication occurring in a small number of replication foci? What are some of the reasons you might expect human cells to have more efficient repair systems than those of a frog? Suppose you were to compare autoradiographs of two cells that had been exposed to [3H]thymidine, one that was engaged in DNA replication (S phase) and another that was not. How would you expect autoradiographs of these cells to differ? Construct a model that would explain how transcriptionally active DNA is repaired preferentially over transcriptionally silent DNA.

JWCL151_ch14_560-604.qxd

7/30/09

3:30 PM

Page 560

14 Cellular Reproduction 14.1 14.2 14.3

The Cell Cycle M Phase: Mitosis and Cytokinesis Meiosis

The Human Perspective:

Meiotic Nondisjunction and Its Consequences Experimental Pathways:

The Discovery and Characterization of MPF

A

ccording to the third tenet of the cell theory, new cells originate only from other living cells. The process by which this occurs is called cell division. For a multicellular organism, such as a human or an oak tree, countless divisions of a single-celled zygote produce an organism of astonishing cellular complexity and organization. Cell division does not stop with the formation of the mature organism but continues in certain tissues throughout life. Millions of cells residing within the marrow of your bones or the lining of your intestinal tract are undergoing division at this very moment. This enormous output of cells is needed to replace cells that have aged or died. Although cell division occurs in all organisms, it takes place very differently in prokaryotes and eukaryotes. We will restrict discussion to the eukaryotic version. Two distinct types of eukaryotic cell division will be discussed in this chapter. Mitosis leads to production of cells that are genetically identical to their parent, whereas meiosis leads to production of cells with half the genetic content of the parent. Mitosis serves as the basis for producing new cells, meiosis as the basis for producing new sexually reproducing organisms. Together, these two types of cell division form the links in the chain between parents and their offspring and, in a broader sense, between living species and the earliest eukaryotic life forms present on Earth. ■ Fluorescence micrograph of a mitotic spindle that had assembled in a cell-free extract prepared from frog eggs, which are cells that lack a centrosome. The red spheres consist of chromatin-covered beads that were added to the extract. It is evident from this micrograph that a bipolar spindle can assemble in the absence of both chromosomes and centrosomes. In this experiment, the chromatin-covered beads served as nucleating sites for the assembly of the microtubules that subsequently formed this spindle. The mechanism by which cells construct mitotic spindles in the absence of centrosomes is discussed on page 575. (REPRINTED WITH PERMISSION FROM

560

R. HEALD

ET AL.,

NATURE

VOL.

382,

COVER OF

8/1/96; ©

COPYRIGHT

1996, MACMILLAN MAGAZINES LIMITED.)